INTRODUCTION

Top Abstract Introduction Procedures Results Discussion References

The concerted up-regulation of transcription and translation is required for a cell, responding to mitogenic stimuli, to grow and enter the cell cycle (1-3). Recent studies have implicated increased S6 phosphorylation in the selective translation of a subset of essential mRNAs containing an oligopyrimidine tract at their transcriptional start site (4, 5). This event is regulated by p70^s6k/p85^s6k (6), two mitogen-stimulated protein kinase isoforms that rely on multiple phosphorylation as a principal mechanism for activation (7). The p85^s6k isoform is expressed from the same transcript as p70^s6k through an alternative translational initiation start site,¹ adding a 23-amino acid extension at the amino terminus and constitutively targeting it to the nucleus (8). Little is known about the role of p85^s6k in the nucleus; however, S6 has been shown to reside in both the nucleoplasm, in a free form, and in the nucleolus, in preribosomal particles, where it is also phosphorylated in response to mitogens (9). In studies conducted to date, regulation of the nuclear isoform parallels that of p70^s6k, so a concomitant role for p85^s6k in mitogenesis is predicted (10, 11), perhaps involving transcription or processing of RNA (8).

The identification of intramolecular p70^s6k regulatory elements, in the form of domains and phosphorylation sites, has increased our understanding of the mechanism by which the kinase autoregulates (7, 12, 13). To date, eight phosphorylation sites have been identified in the endogenous kinase (14, 15). In initial studies, Ser⁴¹¹, Ser⁴¹⁸, Thr⁴²¹, and Ser⁴²⁴, residing within a potential autoinhibitory domain at the carboxyl terminus of the kinase (16, 17), were found to be principal sites of mitogen-induced phosphorylation (14). These sites are characterized by a proline in the +1 position and a hydrophobic residue in the 2 position. More recently, studies have led to the identification of Ser³⁷¹ (18) as a phosphorylation site that shares the same motif, and three additional sites, Thr²²⁹, Thr³⁸⁹, and Ser⁴⁰⁴, which are flanked in the +1 and 1 positions by bulky aromatic amino acids (15). The phosphorylation of these latter sites in response to mitogens is blocked by treatment of cells with the immunosuppressant rapamycin or the fungal metabolite wortmannin (15, 19). Based on mutation studies, Thr²²⁹, in the activation loop, as well as Ser³⁷¹ and Thr³⁸⁹, in the linker region coupling the catalytic and autoinhibitory domains, appear to be critical for kinase activation (15, 18). The activation loop phosphorylation site, Thr²²⁹ in p70^s6k, is a common regulatory element found in many kinases (20, 21). In parallel studies, we have shown that the phosphorylation of this site is regulated by the newly described phosphoinositide-dependent protein kinase, PDK1² (22). Thr²²⁹, Ser³⁷¹, and Thr³⁸⁹, as well as the domains in which they reside, are strikingly conserved in most members of the AGC (protein kinases A, G, and C) family of Ser/Thr kinases (21).

In contrast to the domains containing Thr²²⁹, Ser³⁷¹, and Thr³⁸⁹, the autoinhibitory domain, as well as the carboxyl terminus of p70^s6k, are conspicuously absent in the other members of the AGC family of Ser/Thr kinases (21). We have reported that mutation of the S/TP sites within the autoinhibitory domain, as well as Ser⁴⁰⁴, to alanines or acidic residues modulates kinase activity (15, 19). However, others have claimed little to no effect of similar mutations on the activity of the kinase (23, 24). Indeed, the deletion of the p70^s6k carboxyl terminus, containing the S/TP sites, has little effect on either basal or mitogen-induced kinase activity (12, 13, 25), which has also been used to conclude that these sites are not involved in regulating kinase activation at either the G₀/G₁ or M/G₁ transition state of the cell cycle (23, 24). Despite these observations, this domain is completely conserved in all the mammalian forms of p70^s6k and is also present in the recently cloned Drosophila homolog, Dp70^s6k (26, 27). More notably, a p70^s6k amino-terminal truncation, which blocks kinase activation (12, 13, 25) and mitogen-induced Thr²²⁹ and Thr³⁸⁹ phosphorylation (13), is rescued by the same carboxyl-terminal deletion that removes the autoinhibitory domain (12, 13, 25). These observations suggest instead that the autoinhibitory domain, and possibly the phosphorylation sites residing within this domain, play a critical role in regulating p70^s6k activity in the intact kinase through the modulation of Thr²²⁹ and Thr³⁸⁹ phosphorylation.

In this study, we utilized phosphorylation site and truncation mutations to elucidate the role of the autoinhibitory domain S/TP sites in regulating mitogen-induced p70^s6k activation. Next, we examined the nature of this event as it relates to the ability of the S/TP sites, along with Thr³⁸⁹, to regulate Thr²²⁹ phosphorylation in vivo. Finally, by employing PDK1 in vitro, we have determined that the mechanism by which these carboxyl-terminal phosphorylation sites control Thr²²⁹ phosphorylation in vivo is synergistic and intrasteric.

	EXPERIMENTAL PROCEDURES

Top Abstract Introduction Procedures Results Discussion References

Plasmid Construction and Mutagenesis-- Generation of the amino- and carboxyl-terminal truncation mutants of p70^s6k, as well as point mutations at phosphorylation sites, was achieved using the Altered Site II Mutagenesis System (Promega), as described previously (15). All constructs were tagged with a myc epitope and placed immediately following the p70^s6k or PDK1 initiator ATG codon (15, 22). Phosphorylation site mutants were placed in the appropriate background by BglII-PstI fragment exchanges. All constructs were subcloned into a cytomegalovirus-driven expression vector after being verified by DNA sequencing.

Cell Culture, Transfection, and Metabolic Labeling-- Human embryonic kidney 293 cells were maintained in Dulbecco's modified Eagle's medium supplemented with 10% FCS. Twenty-four hours before transfection, cells were seeded at a density of 10⁶ cells per 10-cm-diameter plate. The cells were then transiently transfected, using a modified calcium phosphate procedure (28, 29), with 1-5 µg of the appropriate construct. Total DNA transfected was kept at 10 µg for all experiments using the empty vector. After 12 h, the transfected cells were placed into serum-free Dulbecco's modified Eagle's medium for an additional 24 h. Metabolic labeling was carried out in phosphate-free Dulbecco's modified Eagle's medium with 1-2 mCi ³²P_i per 5 ml of medium followed by extraction either with or without serum stimulation as described previously (30). Extracts were centrifuged at 12,000 × g for 5 min at 4 °C, and the supernatants were quickly frozen in liquid N₂ and stored at 70 °C.

Immunoblotting, Kinase Assays, and Two-dimensional Phosphopeptide Mapping-- Protein concentrations were measured using the Bio-Rad D/C protein assay. For Western blot analysis, 20 µg of extract protein were resolved by SDS-polyacrylamide gel electrophoresis before transfer onto an Immobilon P membrane (Millipore). Expression of the epitope-tagged proteins was detected by decorating the membrane with the monoclonal 9E10 antibody followed by a secondary rabbit anti-mouse antibody and finally by a fluorescein isothiocyanate-conjugated swine anti-rabbit tertiary antibody. Expression levels were quantified using fluorimetry (Molecular Dynamics) and ImageQuant software (Molecular Dynamics). Activities of the ectopically expressed mutants, immunoprecipitated with the 9E10 antibody, were measured against S6 as a substrate (15). Kinase activities were quantitated using phosphorimagery (Molecular Dynamics) and ImageQuant software. Activities were normalized to the level of expressed kinase and compared with other mutants only when expression levels were similar. Two-dimensional phosphopeptide mapping of ³²P-labeled, ectopically expressed p70^s6k was performed as described previously (13). The resulting phosphopeptide maps were visualized by phosphorimagery.

In Vitro Phosphorylation-- Myc-tagged p70^s6k and myc-tagged PDK1 were independently transfected in MEK 293 cells as described above, quiesced, and extracted in Buffer A (50 mM Tris, pH 7.5, 50 mM NaCl, 10 mM NaF, 10 mM -glycerolphosphate, 10 mM NaPPi, 0.5 mM EGTA, 1 mM DTT, 1 mM benzamidine, 0.5 mM phenylmethylsulfonyl fluoride, 0.1% TX-100, and 10 µg/ml leupeptin and aprotinin). Expression was determined by Western blotting as above, and total extracts containing PDK1 (2.5 µg) and p70^s6k constructs (40 µg) were mixed in Buffer A before co-immunoprecipitation with 9E10 and protein G-Sepharose (22). The immunoprecipitates were washed twice with Buffer A, twice with Buffer A containing 500 mM NaCl, and finally with Buffer B (50 mM Tris, pH 7.5, 10 mM NaCl, 1 mM DTT, 10% glycerol, 1 mM benzamidine, and 0.2 mM phenylmethylsulfonyl fluoride). The washed immunoprecipitates were incubated in Buffer B containing 10 mM MgCl₂ and 10 µCi [-³²P]ATP (20 µM) for 30 min at 30 °C. ³²P-Labeled proteins were resolved by SDS-polyacrylamide gel electrophoresis and visualized by autoradiography on a PhosphorImager.

	RESULTS

Top Abstract Introduction Procedures Results Discussion References

p70^s6kN₅₄ Activation Is Rescued by Acidic Substitutions in the Carboxyl Terminus-- The inability of the p70^s6k amino-terminal truncation mutant, p70^s6kN₅₄ (Fig. 1), to respond to mitogens can be largely rescued by removing the carboxyl terminus (12, 13, 25). To test whether mitogen-induced phosphorylation of the four autoinhibitory S/TP sites could mimic removal of the carboxyl terminus and rescue activity, acidic amino acids were substituted for these residues in p70^s6kN₅₄, and the activity of the newly generated variant, p70^s6kN₅₄·D₃E (Fig. 1), was measured following transient expression in 293 cells. In contrast to the double truncation mutant (12, 13, 25), p70^s6kN₅₄·D₃E displayed high basal activity (Fig. 2), resembling the elevated basal activity previously reported for the same acidic amino acid substitutions placed in wild-type p70^s6k (15, 19). Serum stimulation increased p70^s6kN₅₄·D₃E activity to about 50% of the value obtained for the double truncation mutant and about 10-fold over that detected for the parent, p70^s6kN₅₄ (Fig. 2). Because the basal activity of p70^s6kD₃E can be further augmented by placing an acidic residue at Thr³⁸⁹ (13, 15), the same mutation was placed in p70^s6kN₅₄·D₃E (N₅₄·D₃E·T389E), which further increased both basal and serum-stimulated activities (Fig. 2). Together, the results demonstrate that substitution of acidic amino acids for phosphorylation sites residing in the carboxyl terminus of p70^s6k largely rescues the activity of the amino-terminal truncation mutant.

View larger version (13K): [in this window] [in a new window]   Fig. 1.   Schematic diagram of p70s6k domains and phosphorylation sites. Shown are the carboxyl- and amino-terminal domains (black and cross-hatched regions, respectively) flanking the catalytic domain (white). The autoinhibitory domain is also shown (diagonally hatched). The phosphorylation sites flanked by a proline in the +1 position are shown along the outside surface, whereas the rapamycin and wortmannin sensitive sites are shown on the inside surface. Acidic substitutions for the S/TP sites in the autoinhibitory domain phosphorylation sites in p70s6kD3E mutant are indicated. The positions of the carboxyl- and amino-terminal truncations are noted.

View larger version (19K): [in this window] [in a new window]   Fig. 2.   Rescue of p70s6k N54 activity with acidic mutations in the carboxyl terminus. 293 cells were transiently transfected with myc-tagged p70s6kN54 constructs harboring different mutations in indicated phosphorylation sites (Fig. 1). After serum deprivation, cells were extracted directly or stimulated with 10% FCS for 45 min before extraction. The ectopically expressed mutants were immunoprecipitated using the 9E10 antibody and assayed for kinase activity against 40S ribosomes as described under "Experimental Procedures." Activities are shown as the percentage of serum-stimulated p70s6kN54C104 activity when expressed at the same level. The error bars represent the standard error of three independent experiments.

Deregulation of Thr²²⁹ Phosphorylation-- Removal of the amino and carboxyl termini increases basal Thr²²⁹ phosphorylation, whereas removal of the amino terminus alone suppresses Thr²²⁹ phosphorylation (13). These findings suggest that access of the Thr²²⁹ kinase is restricted by the p70^s6k carboxyl terminus, possibly through the phosphorylation state of the S/TP sites in the autoinhibitory domain. To examine this possibility, Thr²²⁹ phosphorylation and kinase activity were first determined for p70^s6kC₁₀₄. In quiescent cells, p70^s6kC₁₀₄ had high levels of Thr²²⁹ phosphorylation but undetectable Thr³⁸⁹ phosphorylation (Fig. 3B), consistent with results obtained for the double truncation mutant (13). Despite the high levels of Thr²²⁹ phosphorylation, this variant has low basal kinase activity (Fig. 3A). Upon serum stimulation, Thr²²⁹ phosphorylation increased approximately 2-fold, whereas Thr³⁸⁹ phosphorylation was greatly enhanced (Fig. 3C), correlating with increased kinase activity (Fig. 3A). Therefore, truncation of the carboxyl terminus alone is sufficient to disrupt the regulation of Thr²²⁹ phosphorylation in resting cells. To determine whether acidic mutations of the four S/TP sites in the autoinhibitory domain could mimic this effect, a variant harboring these mutations, p70^s6kD₃E, was transiently expressed in 293 cells. In quiescent cells, this mutant displayed elevated kinase activity, which could be further stimulated with serum (Fig. 4A). However, in contrast to the p70^s6kC₁₀₄ truncation mutant, basal Thr²²⁹ phosphorylation in p70^s6kD₃E was very low, although it could be stimulated with serum (Fig. 4, compare B and C). It should be noted that stronger exposures of the chromatogram depicted in Fig. 4B showed low levels of Thr³⁸⁹ phosphorylation (data not shown), reflecting the elevated basal kinase activity and rapamycin sensitivity of this construct (13). Thus, the acidic mutations alone are insufficient to increase basal Thr²²⁹ phosphorylation, indicating that another element is necessary for regulating Thr²²⁹ phosphorylation.

View larger version (42K): [in this window] [in a new window]   Fig. 3.   Effects of carboxyl-terminal truncation on p70s6k activity and Thr229 phosphorylation. 293 cells transiently expressing myc-tagged p70s6kC104 were quiesced prior to labeling with 32Pi and either directly extracted or extracted following a 45-min stimulation with 10% FCS. Samples were immunoprecipitated and either assayed for kinase activity (A) or subjected to two-dimensional tryptic/chymotryptic phosphopeptide mapping (B and C) as described under "Experimental Procedures." The relevant phosphopeptides and the origin are indicated on the figure. The results shown are typical of at least two independent experiments.

View larger version (34K): [in this window] [in a new window]   Fig. 4.   Effect of autoinhibitory domain S/TP mutations on Thr229 phosphorylation. 293 cells were transiently transfected with myc-tagged p70s6kD3E, quiesced, and metabolically labeled with 32Pi as in Fig. 3. The cells were then extracted directly (B) or stimulated for 45 min with 10% FCS before extraction (C). S6 kinase activity was determined as in Fig. 2 (A), or the 32Pi labeled kinase was immunoprecipitated and analyzed by phosphopeptide mapping as in Fig. 3 (B and C). The results are typical of at least two independent experiments.

Regulation of Thr²²⁹ Phosphorylation by Thr³⁸⁹ Phosphorylation-- An acidic residue substituted for Thr³⁸⁹ potentiates the ability of the S/TP mutations to rescue the p70^s6k amino-terminal truncation mutant (Fig. 2), and activation of p70^s6kD₃E is paralleled by increased Thr³⁸⁹ phosphorylation (Fig. 4C). These findings suggest that Thr³⁸⁹ phosphorylation may be the additional element required to bring about Thr²²⁹ phosphorylation. To test this possibility, phosphopeptide analyses of p70^s6k were compared with those from p70^s6k variants harboring acidic mutations at Thr³⁸⁹ and in the S/TP sites. Phosphopeptide analysis of wild-type p70^s6k from quiescent cells revealed low levels of Thr²²⁹ phosphorylation, which were dramatically increased by serum stimulation (Fig. 5, compare B and C), as was activity (Fig. 5A). Substitution of a glutamate for Thr³⁸⁹ in the wild-type p70^s6k background raised basal kinase activity levels (Fig. 5A) and Thr²²⁹ phosphorylation (Fig. 5, compare B and D) approximately 2-fold over that of the wild-type enzyme. The corresponding mutation in the p70^s6kD₃E variant led to a dramatic increase in basal Thr²²⁹ phosphorylation, reaching a value equivalent to the serum-stimulated wild-type kinase (Fig. 5, compare C with E). As previously shown (15), both constructs were further activated by serum (Fig. 5A). In contrast, substitution of an alanine for Thr³⁸⁹ in either p70^s6k or p70^s6kD₃E completely abolished kinase activity (Fig. 5A and Ref. 15) and Thr²²⁹ phosphorylation in response to serum (Fig. 6). These results support the hypothesis that in the wild-type kinase, phosphorylation of the autoinhibitory domain sites functions to up-regulate Thr²²⁹ phosphorylation by cooperating with Thr³⁸⁹ phosphorylation.

View larger version (42K): [in this window] [in a new window]   Fig. 5.   Role of autoinhibitory domain S/TP sites in regulating Thr229 phosphorylation. 293 cells were transiently transfected with the indicated p70s6k construct and deprived of serum for 24 h before metabolic labeling with 32Pi. Cells were then extracted directly (B, D, and E) or extracted following a 45-min stimulation with 10% FCS (C). Kinase activity was determined by immunoprecipitation as in Fig. 2 (A) or processed for two-dimensional phosphopeptide mapping as in Fig. 3 (B-E). The results are typical of at least two independent experiments.

View larger version (44K): [in this window] [in a new window]   Fig. 6.   The effect of Thr389 mutations on Thr229 phosphorylation. 293 cells were transiently transfected with either p70s6k T389A (A) or p70s6kT389A·D3E (B) and deprived of serum for 24 h before metabolic labeling with 32Pi. Following stimulation with 10% FCS, the cells were extracted, and the p70s6k mutants were immunoprecipitated and processed for two-dimensional phosphopeptide mapping as in Fig. 3. The results are typical of at least two independent experiments.

Intrasteric Regulation of Thr²²⁹ Phosphorylation-- The results suggest that phosphorylation sites at the carboxyl terminus synergistically regulate Thr²²⁹ phosphorylation. However, the data do not address whether the observed synergy on Thr²²⁹ phosphorylation is through a single ordered intrasteric mechanism or is instead regulated in vivo through the interplay of multiple effector molecules. To obtain insight into this issue, advantage was taken of the recently described Thr²²⁹ kinase PDK1 (22, 32) and p70^s6k mutants harboring the different acidic amino acid substitutions. When tested in vitro against the wild-type p70^s6k (Fig. 7A), PDK1 only poorly phosphorylated the kinase (Fig. 7B) and had no effect on activity (data not shown and Ref. 22). Furthermore, even though p70^s6kT389E was a slightly better substrate for PDK1 than p70^s6kD₃E, the response was only marginally enhanced for either variant over that obtained with wild-type p70^s6k (Fig. 7B). This finding supports the hypothesis that neither of these individual sets of mutations is sufficient to allow PDK1 to access Thr²²⁹. In contrast, substitution of both sets of acidic mutations resulted in synergistic phosphorylation of p70^s6kD₃E·E389 by PDK1 (Fig. 7B), consistent with the observed effect of the combined mutations on Thr²²⁹ phosphorylation in vivo (Fig. 5E). This effect was abolished when an alanine was substituted for Thr³⁸⁹ in the wild-type p70^s6k and the p70^s6kD₃E variant (Fig. 7B and data not shown). The data support the hypothesis that Thr²²⁹ phosphorylation is regulated by the collective efforts of the carboxyl-terminal phosphorylation sites, through an intrasteric mechanism, presumably by modulating a domain in p70^s6k that blocks access of PDK1 to the activation loop phosphorylation site.

View larger version (21K): [in this window] [in a new window]   Fig. 7.   In vitro phosphorylation of Thr229 by PDK1. 293 cells independently transfected with myc-tagged PDK1 or the indicated p70s6k construct were serum starved and extracted directly. For PDK1, 15 µg of total cell extract, and for p70s6k, 40 µg of cell extract, were analyzed by Western blotting (A); then extracts of each p70s6k variant and PDK1 were co-immunoprecipitated and subjected to an in vitro kinase reaction (B, upper panel), and the data were quantitated and presented in the form of a histogram (B, lower panel) as described under "Experimental Procedures." The asterisk in B indicates the position of autophosphorylated PDK1.

	DISCUSSION

Top Abstract Introduction Procedures Results Discussion References

Of the eight known p70^s6k phosphorylation sites, the first to be identified were the S/TP sites in the carboxyl-terminal autoinhibitory domain of the kinase (14). Although we reported that neutral and acidic mutations at these sites lower and raise basal kinase activity, respectively (15, 19), others found little to no effect of similar mutations (23, 24). This study emphasizes the importance of the autoinhibitory domain phosphorylation sites in p70^s6k activation. First, acidic amino acid substitutions at these sites largely rescue the activity of an amino-terminal truncation mutant, and second, these sites cooperate with an acidic mutation at Thr³⁸⁹ to synergistically regulate phosphorylation of the activation loop site, Thr²²⁹. Although the p70^s6kC₁₀₄ mutant is still regulated by mitogens, indicating that other elements are involved in p70^s6k activation, it does not exclude a role for the autoinhibitory domain S/TP in the activation of the intact kinase. Indeed, the synergistic effect conferred by the S/TP and Thr³⁸⁹ acidic mutations on Thr²²⁹ phosphorylation suggests a possible mechanism for the sensitive control of Thr²²⁹ phosphorylation, which is dependent on the stoichiometry of S/TP site phosphorylation in the autoinhibitory domain (see below). This is further supported by the observation that truncation of the carboxyl terminus largely deregulates Thr²²⁹ phosphorylation (Fig. 3).

The ability of the acidic mutations of the S/TP sites to rescue p70^s6kN₅₄ activity was unexpected, because deletion of this domain did not significantly affect the serum-induced phosphorylation of these sites (13). However, this may be due to the degree of phosphorylation at the autoinhibitory domain S/TP sites in response to mitogen stimulation. Mitogen stimulation is hypothesized to first lead to an increase in S/TP site phosphorylation in the autoinhibitory domain, which functions together with the amino terminus to facilitate Thr³⁸⁹ phosphorylation (Fig. 8). In the absence of the amino terminus, the level of mitogen-induced S/TP site phosphorylation may not be sufficient to promote a net increase in Thr³⁸⁹ phosphorylation and subsequent Thr²²⁹ phosphorylation, attenuating kinase activation. However, substitution of an acidic amino acid at each of the S/TP sites would raise the overall of negative charge of this domain and overcome the effect of the amino-terminal truncation, triggering Thr³⁸⁹ phosphorylation. In support of this model, phosphopeptide maps show that Thr³⁸⁹ phosphorylation is rescued when acidic S/TP site mutations are placed in the p70^s6kN₅₄ background.³ Thus, Thr³⁸⁹ phosphorylation would act as an intermediary step between autoinhibitory S/TP and activation loop site phosphorylation, which would be the final step in mitogen-induced p70^s6k activation (Fig. 8).

View larger version (11K): [in this window] [in a new window]   Fig. 8.   Mechanism of p70s6k activation. p70s6k subdomains and phosphorylation sites are denoted as in Fig. 1. The schematic represents the proposed mechanism leading to wild-type p70s6k activation, outlining the sequential steps in the process (see under "Discussion").

Although a number of candidate autoinhibitory domain S/TP kinases have been suggested, including cyclin-dependent kinase-1 and the mitogen-activated kinases p42^mapk/p44^mapk (33), their requirement has not been substantiated to date. Indeed, utilization of interfering mutants of p21^ras and p74^raf, as well as deletion mutants of the platelet-derived growth factor receptor, demonstrated that p42^mapk/p44^mapk were not effectors of the p70^s6k signaling pathway (11). It has been shown that over-expression of a kinase dead p70^s6k blocks Thr²²⁹, Thr³⁸⁹, and Ser⁴⁰⁴ phosphorylation as well as strongly suppressing Ser⁴¹¹ and Thr⁴²¹ phosphorylation in the autoinhibitory domain (5, 34), a pattern of inhibition resembling that induced by rapamycin (15). Furthermore, it was found that over-expression of kinase dead or wild-type p70^s6k blocks the same sites of phosphorylation in the suppressor of protein synthesis initiation factor 4E, eIF4E-binding protein-1, as does rapamycin (34). These results suggest that overexpression of p70^s6k might sequester a common upstream kinase that is also responsible for phosphorylating eIF4E-binding protein-1 (34). Consistent with this hypothesis, the phosphorylation sites in eIF4E-binding protein-1 also display S/TP motifs (35), and recently, it was shown that these sites are phosphorylated in vitro by the mammalian target of rapamycin (36). However, it is unlikely that mammalian target of rapamycin is the S/TP kinase for p70^s6k, because rapamycin has no effect on serum-induced S/TP phosphorylation in the amino-terminal truncation mutant, p70^s6kN₅₄.⁴ Interest in the identity of the S/TP kinase has been further increased by the recent observation that Pin1, the conserved mitotic peptidyl-prolyl isomerase, binds to p70^s6k, apparently through phosphorylated Ser⁴¹¹ (37).

The finding that in the carboxyl-terminal deletion mutant, p70^s6k C₁₀₄, kinase activity and Thr³⁸⁹ phosphorylation are tightly regulated suggests the existence of an additional regulatory element in p70^s6k that is controlled by the phosphorylation of the S/TP and Thr³⁸⁹ sites. This element would function to modulate Thr²²⁹ phosphorylation and activity. An obvious candidate for such an element is the linker region, which couples the carboxyl and catalytic domains of p70^s6k (Fig. 1). Previously, it was noted that many members of the AGC family of protein kinases (21) contain a site homologous to Thr³⁸⁹, as well as the conserved motif surrounding this site (15). In a more recent study, it was also pointed out that this conservation extends through the entire linker region (18). Within this region we identified a novel site, Ser³⁷¹, the phosphorylation of which appears to be critical for kinase activation (18). The site equivalent to Ser³⁷¹ has been identified as a major autophosphorylation site in protein kinase C II (38, 39) and protein kinase C  (40), Thr⁶⁴¹ and Thr⁶³⁸, respectively. In both cases, there is a proline in the +1 position and a hydrophobic residue in the 2 position, as with Ser³⁷¹ (18). Furthermore, modeling studies with protein kinase C have suggested that the conserved linker region may interact with the amino terminus and that Thr⁶⁴¹ is juxtaposed to the active site allowing autophosphorylation by an intramolecular reaction (41-43). Although Ser³⁷¹ is not an autophosphorylation site in p70^s6k (18), the modeling studies suggest that it could be strategically placed to modulate potential interactions between the amino terminus, the catalytic domain, and the autoinhibitory region. Indeed, mutation of this site to either an alanine or an aspartic acid blocked both Thr³⁸⁹ phosphorylation and kinase activation, but surprisingly, it did not affect Thr²²⁹ phosphorylation (18). This may indicate that mutations at Ser³⁷¹ disrupt the normal function of the linker region in regulating Thr³⁸⁹ and Thr²²⁹ phosphorylation.

PDK1 has been identified as the common activation loop kinase for Thr²²⁹ phosphorylation in p70^s6k (22) and Thr³⁰⁸ in protein kinase B (32). In a manner similar to the p70^s6k S/TP and Thr³⁸⁹ phosphorylation sites, it has been shown that the access of PDK1 to the activation loop of protein kinase B is controlled by the binding of specific phosphatidylinositides to its PH domain (31, 32). Therefore, the two kinases are linked by a common, constitutively active upstream kinase that appears to catalyze the final event in activation. However, distinct internal regulatory elements control the differential activity of each kinase by regulating the access of PDK1 to the activation loop. This mechanism may provide an economical way for the two kinases to share a common upstream activator without sacrificing their ability to be independently regulated.