The Mammalian Target of Rapamycin (mTOR) Partner, Raptor, Binds the mTOR Substrates p70 S6 Kinase and 4E-BP1 through Their TOR Signaling (TOS) Motif

the mTOR Substrates p70 S6 Kinase and 4E-BP1 Their TOR Signaling The mammalian target of rapamycin (mTOR) controls multiple cellular functions in response to amino acids and growth factors, in part by regulating the phosphorylation of p70 S6 kinase (p70S6k) and eukaryotic initiation factor 4E-binding protein 1 (4E-BP1). Raptor (reg-ulatory associated protein of mTOR) is a recently identified mTOR binding partner that also binds p70S6k and 4E-BP1 and is essential for TOR signaling in vivo . Herein we demonstrate that raptor binds to p70S6k and 4E-BP1 through their respective TOS (conserved TOR signaling) motifs to be required for amino acid- and mTOR-dependent regulation of these mTOR substrates in vivo . A point mutation of the TOS motif also eliminates all in vitro mTOR-catalyzed 4E-BP1 phosphorylation are separated by SDS-PAGE, transferred onto a PVDF membrane, and analyzed by autoradiography ( Autoradiogra- phy ). They are subsequently immunoblotted with the anti-mTOR antibody ( anti-mTOR blot ), with the anti-raptor antibody ( anti-raptor blot ), and with the anti-GST antibody ( anti-GST blot ). 32 P incorporated into GST-4E-BP1 is quantified by BAS 2500 in arbitrary units: lane 1 , 20.6; lane 2 , 343; lane 3 , 296; lane 4 , 30.3; lane 5 , 48.4; lane 6 , 928; lane 7 , 8150. These results are representative of three reproducible experi- ments. B , HEK293 cells are lysed, immunoprecipitated, and immunoblotted as described in A and then subjected to mTOR kinase assay using GST-p70S6k-F28A ( lanes 1–3 ), vehicle ( lane 4 ), or GST-p70S6k ( lanes 5–7 ) as a substrate. The phosphorylation of threonine 412 of GST-p70S6k and GST-p70S6k-F28A is examined by immunoblotting with the anti-phosphothreonine 412 of p70S6k-specific antibody ( anti- 412-P blot ). 32 P incorporated into GST-p70S6k is quantified by BAS 2500 in arbitrary units: lane 1 , 200; lane 2 , 1580; lane 3 , 2060; lane 4 , 115; lane 5 , 292; lane 6 , 1940; lane 7 , 7440. These results are representative of three reproducible experiments. subscribers these corrections and insert the at the appropriate the article to be corrected originally appeared. are urged to introduce these corrections into any reprints they distribute. Secondary (abstract) urged notice of these corrections as prominently as they carried the original abstracts.

The mammalian target of rapamycin (mTOR) controls multiple cellular functions in response to amino acids and growth factors, in part by regulating the phosphorylation of p70 S6 kinase (p70S6k) and eukaryotic initiation factor 4E-binding protein 1 (4E-BP1). Raptor (regulatory associated protein of mTOR) is a recently identified mTOR binding partner that also binds p70S6k and 4E-BP1 and is essential for TOR signaling in vivo. Herein we demonstrate that raptor binds to p70S6k and 4E-BP1 through their respective TOS (conserved TOR signaling) motifs to be required for amino acid-and mTOR-dependent regulation of these mTOR substrates in vivo. A point mutation of the TOS motif also eliminates all in vitro mTOR-catalyzed 4E-BP1 phosphorylation and abolishes the raptor-dependent component of mTOR-catalyzed p70S6k phosphorylation in vitro. Raptor appears to serve as an mTOR scaffold protein, the binding of which to the TOS motif of mTOR substrates is necessary for effective mTOR-catalyzed phosphorylation in vivo and perhaps for conferring their sensitivity to rapamycin and amino acid sufficiency.
The p70S6k is activated through a sequential multisite phosphorylation in response to insulin or mitogens in vivo (11). In addition, nutrients, especially amino acids, have been shown to regulate the phosphorylation of p70S6k and 4E-BP1 and to be necessary for insulin or mitogen regulation (12)(13)(14)(15)(16)(17). The activity of p70S6k␣1 in vivo is most closely related to the phosphorylation at Thr-412, situated in a hydrophobic motif C-terminal to the canonical catalytic domain (18,19). The identity of the kinase(s) acting on this site in vivo is uncertain; however, this site can be phosphorylated directly by mTOR in vitro (9,10). Recently, site-specific mutagenesis was employed to define a five-amino acid sequence called the TOS (TOR signaling) motif as the minimal functionally important region within this p70S6k noncatalytic N-terminal segment (21). As with N-terminal deletion, mutation of a single Phe within the TOS motif to Ala causes marked inhibition of activity of full-length p70S6k and a loss of sensitivity to rapamycin and amino acid withdrawal in the p70S6k-⌬CT104, lacking C-terminal noncatalytic tail, background. In addition, a TOS motif was identified in the 4E-BPs, wherein mutation of 4E-BP1 Phe-114 to Ala inhibits amino acid-and serum-induced 4E-BP1 phosphorylation.
Raptor is a recently discovered, highly conserved 150-kDa TOR-binding protein that also binds p70S6k and 4E-BPs (22,23). All raptor homologues (22)(23)(24) contain a unique conserved region in their N-terminal half (the RNC domain), followed by three HEAT repeats and seven WD repeats near the C terminus. The binding of TOR to raptor or its S. cerevisiae homologue KOG1 (24) is necessary for TOR signaling in vivo in Caenorhabditis elegans and S. cerevisiae (22,24). Moreover, mTORcatalyzed phosphorylation of 4E-BP1 in vitro is entirely dependent on the presence of raptor, whereas mTOR-catalyzed phosphorylation of p70S6k in vitro, although stimulated ϳ5fold by the addition of raptor, proceeds in the absence of raptor (22).
Herein we show that the TOS motif is necessary for the binding of p70S6k and 4E-BP1 to raptor. Mutation of the TOS motif abolishes mTOR-catalyzed 4E-BP1 phosphorylation in vitro in the presence of raptor and eliminates the raptor-dependent stimulation of mTOR-catalyzed p70S6k phosphorylation. Thus the inhibitory effect of TOS deletion or mutation on 4E-BP1 and p70S6k phosphorylation in vivo can be attributed to the inability of these mutants to bind raptor. Moreover we suggest that the binding of these mTOR substrates to raptor may also be necessary for their sensitivity to regulation by amino acids and rapamycin.
Transfection-Transient transfection was performed by the lipofection method using LipofectAMINE according to the manufacturer's protocol (Invitrogen).
Cell Culture-HEK293 cells were cultured in Dulbecco's modified Eagle's medium with 10% fetal bovine serum.

RESULTS AND DISCUSSION
To examine whether the TOS motif in p70S6k is important to the binding of p70S6k to raptor, wild-type GST fusion proteins of full-length p70S6k␣1 (GST-p70S6k) or an F28A mutant (GST-p70S6k-F28A) were coexpressed with FLAG-tagged raptor (FLAG-raptor) in HEK293 cells. The F28A mutation of p70S6k␣1 corresponds to the F5A mutation of p70S6k␣2, a mutation that recapitulates the phenotype of N-terminal deletion (21); thus, like p70S6k⌬2-46 lacking a short segment N terminus (20), hemagglutinin-tagged p70S6k␣2-F5A is inactive in HEK293 cells, and its activity can be partially rescued by deletion of the p70S6k C-terminal noncatalytic tail. As expected, we find that GST-p70S6k-F28A is also inactive in HEK293 cells (data not shown). Moreover, although p70S6k binds specifically to coexpressed FLAG-raptor (Fig. 1A, lane 4), p70S6k-F28A does not (Fig. 1A, lane 5). Inasmuch as p70S6k-F28A is inactive in HEK293 cells, we inquired whether p70S6k activity is required for the association between p70S6k and raptor. We find that the ATP site mutant, p70S6k-KM/⌬CT104, which is kinase-inactive and lacks C-terminal 104 residues (20,26), binds raptor to an extent similar to p70S6k (Fig. 1B, compare lane 4 with lane 6). These results indicate that the recognition of p70S6k by raptor requires an intact p70S6k TOS motif but neither kinase activity nor the carboxyl-terminal tail of p70S6k.
The TOS motif of 4E-BP1 is also required for mTOR signaling to 4E-BP1 in vivo (21). We therefore next examined the effects of TOS motif mutation on the ability of 4E-BP1 to bind raptor; GST fusions of wild-type 4E-BP1 (GST-4E-BP1) or an F114A mutant of 4E-BP1 (GST-4E-BP1-F114A) were coexpressed with FLAG-tagged raptor (FLAG-raptor) in HEK293 cells. As observed with p70S6k (Fig. 1A), wild-type 4E-BP1 binds specifically to raptor (Fig. 2, lane 4), whereas 4E-BP1-F114A does not (Fig. 2, lane 5). These results clearly indicate that the TOS motif is critical for the binding of p70S6k and 4E-BP1 to raptor and may be a common recognition element by which raptor couples mTOR targets to the mTOR kinase.
To identify the region of raptor that interacts with the TOS motif of p70S6k, we coexpressed several FLAG-tagged raptor fragments with GST-p70S6k or GST-p70S6k-F28A in HEK293 cells. The N-terminal region of raptor (amino acids 1-904,

Raptor Binds to the TOS Motif 15462
raptor-⌬CT), which contains the unique raptor N-terminal conserved (RNC) region and the HEAT repeats, binds to p70S6k as strongly as does the full-length raptor (Fig. 3, lanes 4 and 6); neither bind to p70S6k-F28A (Fig. 3, lanes 5 and 7). The Cterminal region of raptor (amino acids 904 -1335, raptor-⌬NT) and the isolated WD repeat domain of raptor (amino acids 1009 -1335, raptor-WD) do not bind to p70S6k (Fig. 3, lanes 8  and 10). These results suggest that the RNC region and/or the HEAT repeats of raptor appear to be involved in the regulation of the TOS motif of p70S6k. The selective binding of p70S6k to the N-terminal portion of raptor contrasts with the requirements for raptor binding to mTOR, which appears to involve multiple sites in raptor (22,23); thus raptor deleted of its N terminus (236 -1335) or C terminus (1-904) fails to bind mTOR. We reported previously that the overexpression of the full-length of raptor or raptor-⌬CT-(1-904) in HEK293 cells severely inhibited the kinase activity of coexpressed p70S6k (22). In contrast, the overexpression of raptor-⌬NT-(904 -1335) or raptor-WD-(1009 -1335) in HEK293 cells does not significantly inhibit the kinase activity of coexpressed p70S6k (data not shown). Thus, the inhibitory effect of overexpressed fulllength raptor and raptor-⌬CT-(1-904) may be caused by the ability of these elements to bind coexpressed p70S6k and sequester it from endogenous mTOR, thereby disrupting the ternary complex of mTOR-raptor-p70S6k necessary for effective mTOR-catalyzed p70S6k phosphorylation. Moreover, the ability of p70S6k overexpression to interfere with the phosphorylation of 4E-BP1 is likely to be due to the competition of these elements for a common binding site on raptor.
We had shown previously (22) that the association of raptor with mTOR is absolutely required for the ability of mTOR to catalyze 4E-BP1 phosphorylation in vitro. It was not clear, however, whether this reflected a stimulatory effect of raptor on the catalytic activity of mTOR or on the ability of raptor to properly configure or "present" the 4E-BP1 substrate. The finding that point mutation of the substrate TOS motif eliminates the binding of the substrates to raptor enables a test of these alternatives. Removal of raptor from mTOR by washing with 1% Nonidet P-40 essentially eliminates the ability of mTOR to phosphorylate GST-4E-BP1 in vitro (Fig. 4A, compare lane 6  with lane 7), as shown previously (22). In addition, it is evident that mTOR-catalyzed phosphorylation of GST-4E-BP1-F114A is virtually eliminated despite the presence of raptor (Fig. 4A,  lane 3). Thus, the ability of 4E-BP1 to bind raptor is indispensable for mTOR-catalyzed 4E-BP1 phosphorylation.

Raptor Binds to the TOS Motif 15463
GST-p70S6k-F28A, each purified from rapamycin-treated HEK293 cells; mTOR-catalyzed phosphorylation is monitored by both 32 P incorporation into the recombinant GST fusion protein and by anti-p70S6k Thr(P)-412 immunoreactivity. As shown previously (22), overall mTOR kinase activity toward GST-p70S6k and the specific phosphorylation of Thr(P)-412 (estimated by immunoblot) are enhanced by coimmunoprecipitation of mTOR with raptor (Fig. 4B, lane 7); washing the mTOR immunoprecipitates with detergent so that endogenous raptor is removed substantially reduces the mTOR-catalyzed overall 32 P incorporation into GST-p70S6k as well as the anti-Thr(P)-412 immunoreactivity achieved, as compared with that catalyzed by the same mTOR immunoprecipitate that had not been washed with Nonidet P-40 so as to remove coprecipitating endogenous raptor. In contrast, when GST-p70S6k-F28A is employed as substrate for these same mTOR immunoprecipitates, the overall 32 P incorporation and phosphorylation of Thr-412 is low and independent of whether or not raptor had been removed by Nonidet P-40 washing. The phosphorylation catalyzed by the mTOR-raptor complex of GST-p70S6k-F28A is only about 20% that of wild-type GST-p70S6k and very close to the phosphorylation of wild-type GST-p70S6k achieved by the raptor-free mTOR immunoprecipitate. Nevertheless, the inability of raptor to alter the residual mTOR-catalyzed phosphorylation of GST-p70S6k-F28A establishes that the stimulatory effects of raptor on the phosphorylation of wild-type p70S6k (and probably 4E-BP1) are not due to a stimulation of intrinsic catalytic activity of the mTOR kinase but entirely to the ability of raptor to bind and present these two substrates in a more effective way. The continued ability of mTOR to phosphorylate p70S6k-F28A (or wild-type p70S6k in the absence of raptor) is consistent with the earlier demonstration that mTOR can phosphorylate prokaryotic recombinant fragments of p70S6k that lack entirely the N-terminal region containing the TOS motif on a variety of sites, including Thr-412 (9). The present demonstration of a persistent TOS-and raptor-independent component of mTOR-catalyzed p70S6k phosphorylation in vitro raises the possibility that this activity may be responsible for the insulinresponsive but rapamycin-and amino acid-insensitive phosphorylation in vivo of the various N-terminal mutant p70S6ks (20,21). Thus the F28A mutation of the TOS motif, like Nterminal deletion, makes p70S6k Thr-412 phosphorylation and thus catalytic activity very low; deletion of the C terminus partially restores Thr-412 phosphorylation, which is now insulin-responsive but insensitive to rapamycin and amino acid sufficiency (12,19,21). The insensitivity to rapamycin and amino acid sufficiency were interpreted initially to indicate that the persistent insulin-responsive Thr-412 phosphorylation could not be due to an insulin-induced increase in mTORcatalyzed Thr-412 phosphorylation (12). Nevertheless, the present demonstration that mTOR-catalyzed phosphorylation of GST-p70S6k-F28A in vitro persists, although at a greatly reduced efficiency as compared with wild-type GST-p70S6k, suggests that such a reaction may account for the low but persistent in vivo activity of the p70S6k ⌬2-46/⌬CT104 variant. The p70 variants that are unable to bind to raptor may be very poorly phosphorylated by mTOR in vivo, but the persistent modest ability of mTOR to catalyze this modification is facilitated by deletion of the p70S6k C-terminal tail, as is seen with PDK1-catalyzed phosphorylation of the p70S6k activation loop (26). The insulin stimulation of these variants in vivo may reflect disinhibition from the TSC1-TSC2 complex (27,28). Finally, insensitivity to amino acids and rapamycin suggests that the association of p70S6k with raptor is necessary for p70 regulation to be responsive to these inputs. Thus, the rapamycin-FKBP12 complex may interfere with mTOR-catalyzed p70 phosphorylation only when p70S6k is complexed with raptor. This hypothesis implies that the model of action of the rapamycin-FKBP12 complex in vivo at moderately inhibitory concentrations of rapamycin (e.g. 2-20 nM) is not through suppression of the mTOR catalytic activity but by interference with the effective interaction between mTOR and its substrates. Such a proposal has been offered previously, based in part on the nearly hundredfold higher concentrations of rapamycin (in the presence of excess FKBP12) required for inhibition of mTOR kinase activity in vitro (8,10).
Evaluating this hypothesis will require much additional insight into the mechanism by which the mTOR-raptor complex interacts with its substrates, as well as an understanding of the contributions of mLst8 (24) and the TSC1-TSC2 (27,28) complex to the signaling functions and kinase activity of mTOR. Nevertheless, the present results establish one critical mechanism by which the protein raptor participates in coupling mTOR to its cellular substrates.