Phosphorylation of Raptor by p38β Participates in Arsenite-induced Mammalian Target of Rapamycin Complex 1 (mTORC1) Activation*

Cell growth is influenced by environmental stress. Mammalian target of rapamycin (mTOR), the central regulator of cell growth, can be positively or negatively regulated by various stresses through different mechanisms. The p38 MAP kinase pathway is essential in cellular stress responses. Activation of MK2, a downstream kinase of p38α, enhances mTOR complex 1 (mTORC1) activity by preventing TSC2 from inhibiting mTOR activation. The p38β-PRAK cascade targets Rheb to inhibit mTORC1 activity upon glucose depletion. Here we show the activation of p38β participates in activation of mTOR complex 1 (mTORC1) induced by arsenite but not insulin, nutrients, anisomycin, or H2O2. Arsenite treatment of cells activates p38β and induces interaction between p38β and Raptor, a regulatory component of mTORC1, resulting in phosphorylation of Raptor on Ser863 and Ser771. The phosphorylation of Raptor on these sites enhances mTORC1 activity, and contributes largely to arsenite-induced mTORC1 activation. Our results shown here and in previous work demonstrate that the p38 pathway can regulate different components of the mTORC1 pathway, and that p38β can target different substrates to either positively or negatively regulate mTORC1 activation when a cell encounters different environmental stresses.

the p38 pathway is considered to be a major stress-activated signaling pathway (9). The activation of transcription factors and subsequent gene expression is a major mechanism by which the p38 pathway mediates biological responses (10 -12). The activation of other types of cellular proteins is also essential for the p38 pathway to execute its function (13).
The mammalian target of rapamycin (mTOR) 3 is a serine/ threonine kinase that acts as an environmental sensor to regulate a plethora of cellular biosynthetic processes (14). mTOR activation promotes cell growth and proliferation, whereas mTOR inhibition stops cell growth and initiates catabolic processes (15). The p70 S6 kinase (S6K) and eIF4E-binding protein 1 (4EBP1) are key regulators of mRNA translation, and are the most well characterized targets of mTOR (16,17). Phosphorylation of S6K and 4EBP1 by mTOR leads to increased levels of translation of specific mRNAs (18,19). mTOR exists in two distinct functional complexes, mTOR complex 1 (mTORC1) and mTORC2. mTORC1 is potently and specifically inhibited by rapamycin, and it regulates cell size, autophagy, ribosome biogenesis, protein translation, transcription, and cellular viability (15). mTORC2, which is much less sensitive to rapamycin, is important for cytoskeletal regulation and AKT activation (20). mTORC1 is a central signal integrator that receives signals arising from growth factors, nutrients, cellular energy metabolism, and environmental stresses (21). Growth factors, insulin, and energy depletion work via different pathways to activate or inhibit the TSC1/TSC2 complex (22)(23)(24)(25)(26). The tumor suppressors TSC1 and TSC2 negatively regulate mTOR by acting as the GTPase-activating protein for the small GTPase Ras-homology enriched in brain (Rheb), and thus they decrease Rheb-GTP and prevent Rheb from stimulating mTOR activity (27,28). TSC1/TSC2 is essential for growth factor-, nutrient-, and energy-mediated regulation of mTORC1; however, recent studies have shown that other proteins can also play an essential role. The small G protein RAG can sense nutrients and directly mediate mTORC1 relocalization and activation (29,30). mTOR can self-enhance its activity by phosphorylation of Rap-tor (regulatory associated protein of TOR), a component of the mTORC1 complex (31). It was suggested that the mTOR-mediated Raptor phosphorylation is a biochemical rheostat that modulates mTORC1 activity in accordance with environmental cues (32).
Cross-talk between the p38 and mTOR pathways has been reported. Phosphorylation of serine 1210 on TSC2 by MK2, a downstream kinase of p38␣, creates a 14-3-3 binding site, which has been implicated to play a role in inhibiting TSC2 and thus promote anisomycin-induced mTOR activation (33). A recent paper suggested that p38␣ or p38␤ and Drosophila p38b promote mTOR activation under various conditions (34). Our recent study shows that the p38␤-PRAK (p38-regulated/activated protein kinase, or MK5) cascade selectively participates in glucose depletion-induced inactivation of mTORC1. PRAK directly phosphorylates Rheb on serine 130, which in turn impairs the GTP-binding ability of Rheb and prevents Rheb from stimulating mTOR activity (35). It is clear that there are multiple different interplays between the p38 pathway and mTOR pathway.
Although energy stress leads to inactivation of mTORC1, other cell stresses can either inhibit or activate mTORC1 (14). Arsenite is a contaminant toxicant and is also used as a medicinal compound. It activates mTORC1, but the mechanism of this activation is unknown. Here we report that p38␤ selectively participates in arsenite-induced mTORC1 activation, but not in other stress-induced activation of mTOR. Arsenite-induced activation of mTOR is independent from AMPK and TSC2, and occurs at least partially via direct phosphorylation of Raptor by p38␤ to activate mTORC1.

EXPERIMENTAL PROCEDURES
Materials-Sodium arsenite, rapamycin, insulin, and anisomycin were obtained from Sigma. PD98059 and SB203580 were obtained from Calbiochem. Immobilon-P PVDF membrane and Immobilon Western Chemiluminescent HRP Substrate were obtained from Millipore. Autoradiography film was obtained from Kodak.
For insulin or H 2 O 2 stimulation, cells were serum-starved for 24 h followed by insulin or H 2 O 2 stimulation, as indicated. For glucose starvation, cells were cultured with D-glucose-free DMEM (Invitrogen) containing 25 mM HEPES and 10% dialyzed FBS. Cells were stimulated with 25 mM glucose for 15-60 min as indicated. For amino acid starvation, cells were serumstarved for 24 h and incubated in DPBS (PBS containing 25 mM glucose) for 30 min. DMEM containing all amino acids was then added for 10 -30 min as indicated.
293T cells were transfected with Lipofectamine 2000 and the corresponding empty vector was used for adjustment to ensure an equal amount of total DNA. 36 h after transfection, cells were collected using lysis buffer as indicated.
For immunoprecipitations, anti-FLAG M2 beads or anti-HA beads were added to the lysates and incubated with rotation for 3 h or overnight at 4°C. The immunoprecipitates were then washed three times each with lysis buffer followed by addition of 1ϫ SDS sample buffer, boiling for 10 min, resolved by SDS-PAGE, and analyzed by immunoblotting.
Reactions were terminated by washing twice in CIAP buffer and boiling in 1ϫ SDS sample buffer.
In Vitro Kinase Assay-p38␤ kinase assays were performed in a final volume of 20 l consisting of p38 kinase buffer with 250 M ATP and bacterial-derived GST-Raptor-(741-887) fragments as substrates for 30 min at 30°C. The reactions were stopped by 2ϫ SDS sample buffer, and subjected to SDS-PAGE and immunoblotting.
For the mTORC1 kinase assay, 293T cells were transfected with HA-Raptor WT or S771A, S863A, and S771A/S863A mutants. Transfected cells were treated with 1 mM arsenite for 30 min as indicated in the figures and then lysed in CHAPS lysis buffer followed by immunoprecipitation by HA antibodies. As described (39), the immunoprecipitates were washed three times in CHAPS lysis buffer and then twice in 25 mM HEPES, pH 7.4, 20 mM potassium chloride. Kinase assays were carried out in a volume of 15 l consisting of mTORC1 kinase buffer (25 mM HEPES, pH 7.4, 50 mM KCl, 10 mM MgCl 2 , 250 M ATP) and bacterial-derived GST-S6K as a substrate for 30 min at 30°C. The reactions were terminated by 2ϫ SDS sample buffer, and subjected to SDS-PAGE and immunoblotting.
It was a surprise to us that p38␤ plays a promoting role in arsenite-induced mTORC1 activation, because our recent study showed that the p38␤-PRAK cascade negatively regulates mTORC1 in 2-deoxyglucose-treated cells (35). We did a side by side comparison and confirmed that p38␤ is required for 2-deoxyglucose-induced inactivation of mTORC1 and arsenite-induced activation of mTORC1 (supplemental Fig. S2). In contrast to p38␤, PRAK is required for 2-deoxyglucose-induced inactivation of mTORC1 but has no role in arsenite-induced mTORC1 activation (supplemental Fig. S3, A and B). It appears that p38␤ targets different downstream proteins to execute different functions in 2-deoxyglucose-and arsenitetreated cells.
p38␤ Activation Is Selectively Involved in Arsenite-induced Activation of mTORC1-Activation of mTORC1 has been reported in a number of conditions. Replenishment of insulin, glucose, amino acid, or serum all activates mTORC1. We found that p38␤ was not required for mTORC1 activation under these conditions (Fig. 2, A-D). The protein synthesis inhibitor anisomycin can activate a variety of intracellular signaling pathways, including the p38 and mTORC1 pathways. It was shown that serum refurnishing or anisomycin treatment both activated MK2, which in turn phosphorylates TSC2 on Ser 1210 and enables TSC2 binding to 14-3-3, and thus promotes mTORC1 activation (33). We found that, as in serum refurbishment, p38␤ has no role in anisomycin-induced S6K1 phosphorylation, excluding the possibility that p38␤ functions via MK2 to activate mTORC1 (Fig. 2E). Under serum-starved conditions, H 2 O 2 can activate mTOR, and we found that p38␤ has no role in H 2 O 2 -induced mTOR activation (Fig. 2F). Collectively, of all the activation of mTORC1 tested, arsenite-induced mTORC1 activation is the only one that requires p38␤.
It is known that arsenite activates all p38 group members, including p38␤ (7). Because of the similar size of p38 family members, there is currently no method available to specifically detect p38␤ activation. To determine whether p38␤ is activated in our experimental setting, we used transiently expressed FLAG-p38␤ to determine whether arsenite activates p38␤. 293T cells expressing FLAG-p38␤ were treated with arsenite and the phosphorylation of FLAG-p38␤ was detected by Western blotting with anti-phospho-p38 antibodies (Fig. 3A). Arsenite-induced p38␤ phosphorylation and rapamycin had no effect on arsenite-induced p38␤ activation, suggesting that p38␤ activation is upstream of mTORC1. Ectopically expressing constitutively active MKK6 (MKK6b(E)) in cells can also lead to activation of all p38 group members (40). We tested whether expression of MKK6b(E) in MEF cells can mimic arsenite treatment and found a low level increase in rapamycinsensitive S6K1 phosphorylation (Fig. 3B), and found this phosphorylation to be p38␤but not PRAK-dependent (Fig. 3C). It appears that overactivation of p38␤ alone is sufficient to increase mTORC1 activity, but the increase of mTORC1 activity is limited. This data are consistent with the results showing that p38␤ deletion or SB203580 treatment did not completely inhibit arsenite-induced mTORC1 activation (Fig. 1).
Arsenite-induced Activation of mTORC1 Is Independent from TSC2 and AMPK-Because AMPK and TSC2 are important regulators of mTORC1, we determined their role in arseniteinduced mTORC1 activation. To examine whether AMPK is involved in arsenite-induced activation of mTORC1, we used AMPK␣1/␣2 double knock-out MEFs, and found that arseniteinduced mTORC1 activation was not affected by AMPK deletion (Fig. 4A). We then used TSC2 Ϫ/Ϫ cells to determine the involvement of TSC2 and found that arsenite-induced mTORC1 activation was not affected by TSC2 deletion (Fig.  4B). As anticipated, arsenite-induced mTORC1 activity can be inhibited in TSC2 Ϫ/Ϫ cells by SB203580 and rapamycin (Fig.  4C). Thus, arsenite-induced mTORC1 activation is independent from AMPK and TSC2.
Arsenite Induces Interaction between p38␤ and Raptor-To elucidate the pathway of p38␤-mediated mTORC1 activation, we coexpressed p38␤ with different components of the mTORC1 complex in 293T cells and treated the cells with or without arsenite. We immunoprecipitated HA-tagged Raptor, mLST8/G␤L, Deptor, and Rheb, finding that p38␤ did not interact with any of these proteins in the absence of arsenite treatment, but interacted with Raptor but not mLST8/G␤L, Deptor, or Rheb in the presence of arsenite stimulation (Fig.  5A). To test whether the interaction between p38␤ and Raptor is p38␤-specific, we compared p38␤ with p38␣ in their interaction with Raptor and found that only p38␤ can interact with Raptor in arsenite-treated cells (Fig. 5B). We then checked whether arsenite treatment induces interaction between endogenous p38␤ and Raptor. 293T cells were treated with or without 1 mM arsenite for 30 min and p38␤ was immunoprecipitated with anti-p38␤ antibodies. We detected Raptor in the immunoprecipitates from arsenite-treated cells, but not in the immunoprecipitates from control non-treated cells (Fig. 5C). The specificity of p38␤ immunoprecipitation was demon-  strated by the absence p38␤ and Raptor in the immunoprecipitates using control IgG (Fig. 5C). Because arsenite induces p38␤ activation, we examined whether the activity of p38␤ is required for its interaction with Raptor. We coexpressed Raptor with wild-type p38␤ or p38␤ inactive mutant (p38␤(AF)) in which the regulatory phosphorylation sites TGY were mutated to AGF, and treated the cells with or without arsenite. Both p38␤ and p38␤(AF) could co-immunoprecipitate with Raptor after the cells were treated with arsenite (Fig. 5D). Consistently, treatment of the cells with SB203580 also did not affect arsenite-induced interaction between Raptor and p38␤ (Fig. 5D). We also examined whether the MKK6-p38␤ module has any role in arseniteinduced p38␤-Raptor interaction and found that MKK6 was not associated with the p38␤-Raptor complex (Fig. 5E). Our data indicated that arsenite-induced interaction between p38␤ and Raptor does not require p38␤ activating phosphorylation and p38␤ activity.
p38␤ Phosphorylates Raptor-The inducible interaction between p38␤ and Raptor suggests that Raptor is a target of p38␤ in arsenite-treated cells. To explore the effect of p38␤ on Raptor, we coexpressed Raptor with p38␤ and included MKK6b(E) in some samples to activate p38␤ in 293T cells. We found that activation of p38␤ by coexpressed MKK6b(E) led to a band shift of Raptor, and this shift is at least partially due to phosphorylation, because treatment of the sample with CIAP reduced the band shift (Fig. 6A). To identify the phosphorylation site(s) we created a number of deletion mutants of Raptor and determined which fragment could be modified by MKK6b(E)-activated p38␤. Western blotting analysis of Raptor mutants that coexpressed with MKK6b(E) and p38␤ in 293T cells revealed that the band shift of Raptor is primarily caused by a fragment of 741-887 amino acids (supplemental Fig. S4, A-C). The band shift of this fragment is due to phosphorylation, because treatment of immunoprecipitated Raptor with CIAP eliminated the band shift (supplemental Fig. S4D).
We mutated each of the serine and threonine sites (41) to alanine in the 741-887 region of Raptor and determined which sites are involved in the phosphorylation-mediated band shift. Converting Ser 863 and Ser 771 to Ala both slightly reduced the band shift, and simultaneous mutation of Ser 863 and Ser 771 completely abolished the shift (Fig. 6B). We then checked whether mutation of Ser 771 and Ser 863 to Ala affected the MKK6b(E)/p38␤-mediated shift of full-length Raptor and found that the S771A/S863A mutant has much less band shift in comparison with wild-type Raptor (Fig.  6C). Thus, Ser 863 and Ser 771 constitute the major phosphorylation sites that are responsible for the p38␤-induced band shift of Raptor.
We obtained anti-Ser 863 phospho-Raptor and anti-Ser 771 phospho-Raptor antibodies. An in vitro kinase reaction using recombinant His-p38␤ and GST-Raptor-(741-887) was performed and the reaction mixtures were analyzed by immunoblotting using the anti-phospho-Raptor antibodies. Phosphorylation of Ser 863 and Ser 771 were both detected in the products of the kinase reaction (Fig. 6D). The weak signal of Ser 771 phosphorylation is most likely due to the titer of the anti-Ser 771 phospho-Raptor being low, and we concentrated anti-Ser 771 antibody 40-fold by using an Amicon ultra filter for further studies. We determined that the concentrated antibody can specifically detect Ser 771 phospho-Raptor and the anti-Ser 863 phospho-Raptor antibody can specifically detect Ser 863 phospho-Raptor in arsenite-treated cells expressing HA-Raptor (supplemental Fig. S5A), although the titers of these antibodies are still not high enough to detect endogenous protein. By using these antibodies, we analyzed wild-type and p38␤ Ϫ/Ϫ MEF cells that had infected HA-Raptor expressing lentivirus, and found that arsenite treatment induced Ser 771 and Ser 863 phosphorylation of Raptor in wild-type but not p38␤ Ϫ/Ϫ cells (Fig. 6E). The requirement of p38␤ in arsenite-induced Raptor phosphorylation was also observed when different doses of arsenite were used (supplemental Fig. S5B). Thus, arsenite induces p38␤-dependent phosphorylation of Raptor on Ser 771 and Ser 863 .
Arsenite-induced Raptor Ser 863 Phosphorylation Is mTORC1 Independent-Ser 863 phosphorylation of Raptor was previously reported to be executed by mTORC1 during insulin stimulation (31). Here we need to determine whether mTORC1 has any role in arsenite-induced Ser 863 phosphorylation of Raptor. To determine whether arsenite-induced phosphorylation of Raptor is mTORC1-dependent, we treated the cells with the mTORC1 inhibitor rapamycin. Rapamycin inhibited arsenite-induced S6K1 phosphorylation, but it did not block arsenite-induced Raptor Ser 863 phosphorylation (Fig. 7A). Similar results were obtained in MKK6b(E)-expressing cells (Fig. 7B). We also evaluated whether inhibition of mTOR by overexpressing inactive mutant of mTOR (mTOR(KD)) in 293T cells can affect arsenite-induced phosphorylation of Raptor, and found that overexpression of FLAG-mTOR(KD) inhibited arsenite-induced S6K1 phosphorylation but not Ser 863 and Ser 771 phosphorylation on Raptor (Fig. 7C). Collectively, these data demonstrated that arsenite-induced Raptor phosphorylation is independent from mTORC1 activity. mTOR-mediated Ser 863 phosphorylation was shown to enhance mTORC1 activity (31,32). We examined whether S863A, S771A, and S863A/S771A mutations affect on the function of Raptor in cells treated with arsenite. We expressed wildtype and mutants of HA-Raptor in 293T cells and stimulated the cells with or without arsenite. The cells were lysed with CHAPS buffer, because in this buffer mTOR is still associated with Raptor (38). Raptor was immunoprecipitated with anti-HA antibodies and the Raptor-mTOR complexes were subjected to a kinase assay using recombinant GST-S6K1 as a FIGURE 5. Arsenite induces interaction between p38␤ and Raptor. A, FLAG-p38␤ expression vector was cotransfected with the expression vector of HA-Raptor, HA-mLST8/G␤L, HA-Deptor, or HA-Rheb as indicated. 36 h later, the cells were treated with or without 1 mM arsenite for 30 min. The cells were harvested in 0.3% CHAPS lysis buffer and then immunoprecipitated (IP) with anti-HA antibodies. The immunoprecipitates and lysates were immunoblotted with anti-HA or anti-FLAG antibodies. B, p38␤ but not p38␣ can interact with Raptor upon arsenite treatment. 293T cells cotransfected with expressing vectors of FLAG-p38␤, FLAG-p38␣, and HA-Raptor in different combinations as indicated. The immunoprecipitation (IP) and immunoblotting were performed as described in A. C, endogenous p38␤ interacts with Raptor. 293T cells were treated with arsenite as indicated and harvested in 0.3% CHAPS lysis buffer. Cell lysates were immunoprecipitated with anti-p38␤ antibodies or control IgG. The immunoprecipitates and lysates were immunoblotted with anti-Raptor and anti-p38␤ antibodies. D, the interaction between p38␤ and Raptor is not dependent on p38␤ kinase activity. 293T cells were cotransfected with expression vectors of FLAG-p38␤(WT), FLAG-p38␤(AF), and HA-Raptor as indicated. 12 h later, 20 M SB203580 was added into the culture medium as indicated. At 24 h after transfection, the cells were harvested and analyzed as described in A. E, MKK6 is not in p38␤-Raptor complex. 293T cells were cotransfected with expression vectors of FLAG-p38␤, Myc-MKK6b, and HA-Raptor as indicated. At 36 h after transfection, the cells were harvested and analyzed as described in A.
substrate as described (39). Wild type-Raptor complex isolated from arsenite-treated cells had increased activity in the phosphorylation of S6K1 (Fig. 8A), which reflects the activation of mTOR by arsenite treatment. S771A, S863A, and S863A/ S771A mutations all reduced arsenite-induced mTOR activity in the immunocomplex (Fig. 8A).
Because fusion of Raptor with the C-terminal 20-amino acid portion of Rheb creates a dominant-active Raptor (Raptor-Rheb(C20)), we mutated Ser 863 and Ser 771 to Ala in this dominant-active Raptor and analyzed the effect of these mutations on the activity of Raptor-Rheb(C20). The S792A mutant was included as a control. Mutation of Ser 863 or Ser 771 , but not  Ser 792 , reduced Raptor-Rheb(C20)-induced increase of S6K1 phosphorylation (Fig. 8B).
Knockdown of Raptor and then reconstituting the cells with wild-type or phosphorylation site mutants of Raptor is an approach to evaluate the effect of the phosphorylation site mutations, and was used in our study. We designed a number of shRNA to target 3Ј untranslated region (3-UTR) of Raptor and identified effective shRNAs. As shown in Fig. 8C, short hairpin Raptor 10 effectively reduced Raptor protein in 293T cells. We transfected expression plasmid containing the coding region of wild-type Raptor, S771A Raptor, S863A Raptor, or S771A/ S863A Raptor into Raptor knockdown cells, and examined arsenite-induced phosphorylation of S6K1. The cells with the reconstitution of S771A Raptor, S863A Raptor, or S771A/ S863A Raptor had a lower level of arsenite-induced phospho-S6K1 in comparison with the cells that were reconstituted with wild-type Raptor (Fig. 8D). Collectively, the data shown in Fig. 8 suggest the role of phosphorylation on Ser 863 and Ser 771 of Raptor in arsenite-induced activation of mTORC1.

DISCUSSION
Cellular stresses have profound effects on cell growth. One of the mechanisms by which stresses influence cell growth is modulation of mTOR activity. In this report, we demonstrate that a specific p38 group member, p38␤, plays an important role in arsenite-induced activation of mTORC1. We show that p38␤ is activated by arsenite treatment and can directly phosphorylate Raptor on Ser 863 and Ser 771 . Phosphorylation of Raptor enhances activity of mTORC1. The data presented here and in previous publications (33,35) demonstrate that the p38 pathway can regulate the mTOR pathway at multiple sites to enhance or inhibit the activity of mTORC1 (supplemental Fig.  S6). Although not being specifically addressed, p38␣␣ may be able to regulate mTOR (33), for example, through activation of MK2. MK2 can phosphorylate TSC2 to sequester TSC2 at the 14-3-3 site and thus release the inhibition of mTORC1 (33). It was known that MK2 plays a role in anisomycin-induced mTORC1 activation (33), whereas it has no role in arseniteinduced mTORC1 activation (supplemental Fig. S7, A and B). Consistently, arsenite-induced Ser 863 phosphorylation of Raptor was not affected by MK2 deletion (supplemental Fig. S7C). We recently showed that p38␤ plays an essential role in the energy depletion-induced inactivation of mTORC1 (35). The specificity of this regulation is conferred by the downstream kinase PRAK (supplemental Fig. S3A). PRAK is essential for p38␤-mediated inactivation of mTORC1 in response to energy depletion, but has no role in arsenite-induced mTORC1 activation and p38␤-dependent phosphorylation of Raptor (supplemental Fig. S3, B and C). Distinct from MK2, deletion of PRAK had no effect on anisomycin-induced mTORC1 activation (supplemental Fig. S3D). It is clear that the p38 pathway can regulate mTOR in many different ways, and the mechanisms used are stimuli-dependent (supplemental Fig. S6).
It is very challenging to address how p38␤ has different roles in regulating mTORC1 in response to different cellular stresses.  SEPTEMBER 9, 2011 • VOLUME 286 • NUMBER 36

JOURNAL OF BIOLOGICAL CHEMISTRY 31509
We showed that PRAK and Raptor are substrates of p38␤ in response to energy starvation and arsenite, respectively. It is unknown how p38␤␤ selects different substrates in response to different stimuli. It is clear that formation of different signaling complexes plays a role in determining the divergent functions of p38␤ in regulating mTOR upon different stresses, as we observed that interaction between p38␤ and Raptor only can be induced by arsenite treatment but not 2-deoxyglucose or anisomycin treatments (supplemental Fig. S8). It is possible that arsenite and energy starvation activate different populations of p38␤ located in either different protein complexes or subcellular locations and thus can access different substrates. It is also possible that the time and levels of p38␤ activation play a role in determining the different complex formations of p38␤. It is well accepted that a cell can achieve temporal, spatial, and stress-specific regulation in response to different stresses, and the paradoxical role of p38␤ in regulating mTOR in response to different stresses could be the result from these complicated activation mechanisms.
Raptor is believed to be a key mTORC1 scaffolding protein that binds mTOR substrates via the TOR signaling (TOS) motif, facilitating their phosphorylation by mTOR. A number of recent studies have shown phosphorylation of Raptor on various sites in the up-or down-regulation of mTORC1 activity. Phosphorylation at Ser 722 and Ser 792 by AMPK create 14-3-3 binding sites and inhibit mTORC1 activity in response to energy stress (41); phosphorylation of Raptor by Rsk at Ser 721 and by mTOR at Ser 863 have been shown to enhance mTORC1 activity (31,32,42); phosphorylation of Raptor on Ser 696 and Thr 706 by Cdc2 was observed during mitosis (36). We show in this report that arsenite induces a p38␤-dependent Raptor phosphorylation on Ser 863 and Ser 771 , and phosphorylation of Raptor on these sites enhances mTORC1 activity. Our data and others suggested that Ser 863 on Raptor can be phosphorylated by more than one kinase, and we observed that arsenite induces phosphorylation of Ser 863 and Ser 771 simultaneously, whereas anisomycin induces only Ser 863 phosphorylation (supplemental Fig. S8). The mechanism of augmentation or inhibition of mTOR activity by Raptor phosphorylation remains elusive at present; however, we can speculate that phosphorylation may change the conformation of Raptor, which affects its interaction with mTOR substrates and thus positively or negatively influences the phosphorylation of mTOR substrates by mTOR.
The complexity of mTOR regulation in response to stresses provides an example of different stress pathways being integrated within a cell, and cells engaging in different strategies to use the same protein to differentially respond to various stresses. It is clear that new technologies and strategies need to be developed to study the spatial dependence of signaling transduction.