Differential Regulation of the Phosphatidylinositol 3-Kinase/Akt and p70 S6 Kinase Pathways by the a 1A -Adrenergic Receptor in Rat-1 Fibroblasts*

Phosphatidylinositol (PI) 3-kinase and its downstream effector Akt are thought to be signaling intermediates that link cell surface receptors to p70 S6 kinase. We examined the effect of a G q -coupled receptor on PI 3-kinase/Akt signaling and p70 S6 kinase activation using Rat-1 fibroblasts stably expressing the human a 1A - adrenergic receptor. Treatment of the cells with phenylephrine, a specific a 1 -adrenergic receptor agonist, activated p70 S6 kinase but did not activate PI 3-kinase or any of the three known isoforms of Akt. Furthermore, phenylephrine blocked the insulin-like growth factor-I (IGF-I)-induced activation of PI 3-kinase and the phosphorylation and activation of Akt-1. The effect of phenylephrine was not confined to signaling pathways that include insulin receptor substrate-1, as the a 1 -adrener- gic receptor agonist also inhibited the platelet-derived growth factor-induced activation of PI 3-kinase and Akt-1. Although increasing the intracellular Ca 2 1 concentration with the ionophore A23187 inhibited the activation of Akt-1 by IGF-I, Ca 2 1 does not appear to play a role in the phenylephrine-mediated inhibition of the PI 3-kinase/Akt pathway. The differential ability of phenylephrine

Cellular growth requires the generation of new translational machinery to accommodate the increased demand for additional proteins. It has been shown that treatment of cells with growth-promoting agents induces the translational up-regulation of ribosomal proteins and protein synthesis elongation factors (1,2). This process is controlled in part by phosphorylation of the S6 protein of 40 S ribosomal subunits by the M r ϭ 70,000 S6 kinase (p70 S6 kinase 1 ; Refs. 3 and 4). p70 S6 kinase is activated upon treatment of cells with a variety of growth factors, hormones, mitogens, and phosphatase inhibitors, etc. This increase in activity is due to phosphorylation of p70 S6 kinase at multiple sites presumably by multiple kinases (5,6). Due to its importance in the growth response, the signal transduction pathways leading to activation of p70 S6 kinase have received considerable attention. A variety of experimental approaches have led to the identification of phosphatidylinositol (PI) 3-kinase and its downstream effector, the protein kinase Akt, as signaling intermediates that link cell surface receptors to p70 S6 kinase. First, treatment of cells with wortmannin or LY294002, two inhibitors of PI 3-kinase, prevents the activation of Akt (7-9) and p70 S6 kinase (10,11) in response to growth factors or hormones. Second, platelet-derived growth factor (PDGF) receptor mutants that cannot bind PI 3-kinase were unable to induce efficiently the activation of Akt (7,8) or p70 S6 kinase (12,13). Third, overexpression of constitutively active forms of PI 3-kinase induces the activation of Akt (14 -16) and p70 S6 kinase (15)(16)(17) in the absence of added extracellular ligands. Conversely, expression of dominant-negative mutants of the p85 subunit of PI 3-kinase inhibits the PDGFinduced activation of Akt (7) and p70 S6 kinase (12). Finally, overexpression of a dominant-negative mutant of Akt causes a reduction in insulin-induced activation of p70 S6 kinase (18), whereas constitutively or conditionally active forms of Akt stimulate p70 S6 kinase (7,8,16,19,20). Together, these results suggest the existence of a linear signaling pathway leading from receptors to PI 3-kinase, Akt and p70 S6 kinase. This pathway is also thought to involve additional components, as Akt does not phosphorylate p70 S6 kinase directly in vitro (21).
Like p70 S6 kinase, Akt is activated by a wide variety of hormones, growth factors, and other stimuli (7,9,22). Akt is thought to mediate many of the cellular effects of insulin and insulin-like growth factor-I (IGF-I) on glucose metabolism and cell survival (23,24). For example, Akt phosphorylates and inactivates glycogen synthase kinase-3 in cells treated with insulin or IGF-I, thus promoting glycogen synthesis (25). Similarly, the anti-apoptotic effect of insulin and IGF-I is partly mediated by Akt phosphorylation of the pro-apoptotic protein BAD (26,27). Akt is activated by phosphorylation of Thr-308 in the activation loop of the catalytic domain and Ser-473 in the carboxyl-terminal tail (22). It is believed that phosphorylation of both sites requires an interaction between the amino-termi-nal pleckstrin homology domain of Akt with membrane inositol phospholipids generated by PI 3-kinase (23,24,28). Translocation of Akt to the membrane is thought to induce a conformational change that permits phosphorylation of Thr-308 and Ser-473 by membrane-associated Akt kinases. The kinase that phosphorylates Thr-308 in Akt has been identified as 3-phosphoinositide-dependent protein kinase 1 (29,30). 3-Phosphoinositide-dependent protein kinase 1 also phosphorylates the equivalent site in some other protein kinases, including p70 S6 kinase (21,31).
In contrast to insulin and IGF-I, treatment of cells with catecholamines to activate ␣ 1 -adrenergic receptors (ARs) induces glycogenolysis. Three ␣ 1 -AR subtypes have been cloned (␣ 1A , ␣ 1B , and ␣ 1D ), and all three receptors activate phospholipase C ␤, which promotes increased production of inositol 1,4,5-trisphosphate and diacylglycerol, leading to elevation of the intracellular Ca 2ϩ concentration ([Ca 2ϩ ] i ) and activation of protein kinase C (PKC), respectively (32). Recent evidence indicates that ␣ 1 -ARs also stimulate additional physiologically relevant signaling pathways. For example, treatment of cultured rat neonatal cardiac myocytes with the ␣ 1 -AR agonist phenylephrine (PE) leads to activation of p70 S6 kinase and an increase in the rate of protein synthesis and cell growth (33).
Since activation of Akt promotes glycogen synthesis, it seemed inconsistent that ␣ 1 -ARs, which stimulate glycogen breakdown, would signal to p70 S6 kinase via an Akt-dependent pathway. In this study, we examined this question using Rat-1 fibroblasts expressing the human ␣ 1A -AR. We show that treatment of these cells with PE activates p70 S6 kinase but does not activate PI 3-kinase or Akt. Moreover, we show that the ␣ 1A -AR inhibits the activation of PI 3-kinase and Akt induced by other growth factors. Finally, we tested whether the differential ability of PE and IGF-I to activate Akt correlates with their ability to protect cells from UV-induced apoptosis.
Cell Culture-Rat-1 fibroblasts stably transfected with the human ␣ 1A -AR were a gift from Dr. G. Johnson of Pfizer (34). Cells were maintained in Dulbecco's modified Eagle's medium (Mediatech, Herndon, VA) with 10% fetal bovine serum (Sigma) in 5% CO 2 at 37°C. Before treatment, cells were incubated in serum-free medium for 16 -18 h. For experiments involving Ca 2ϩ , the cells were preincubated for 1 h in high salt glucose buffer (10 mM Hepes, pH 7.4, 140 mM NaCl, 4 mM KCl, 2 mM MgSO 4 , 1 mM KH 2 PO 4 and 10 mM glucose) plus either 2 mM EGTA or 1 mM Ca 2ϩ . Drugs were then added to the buffer at the indicated doses.
Immunocomplex Akt Kinase Assays-Akt activity was assayed following a method described previously (35). After treatment, cells were rinsed with ice-cold phosphate-buffered saline and then incubated with lysis buffer (50 mM Tris, pH 7.5, 120 mM NaCl, 1% Nonidet P-40, 1 mM EDTA, 50 mM NaF, 40 mM 2-glycerophosphate, 0.1 mM sodium orthovanadate, 1 mM benzamidine, 0.5 mM phenylmethylsulfonyl fluoride, and 10 g/ml aprotinin and leupeptin) for 15 min on ice. Cells were scraped off the plate, and insoluble material was removed by centrifugation. Equal amounts of lysate protein were incubated with antibody against Akt-1 (New England Biolabs), Akt-2 (Upstate Biotechnology, Inc., Lake Placid, NY), or Akt-3 (Upstate Biotechnology) overnight at 4°C and then with 20 l of protein A-agarose (Santa Cruz Biotechnology, Santa Cruz, CA) for 1 h. The beads were washed 2 times with lysis buffer and once with Akt assay buffer (50 mM Tris, pH 7.5, 10 mM MgCl 2 , 1 mM dithiothreitol, 1 mM benzamidine and 0.5 mM phenylmethylsulfonyl fluoride). The washed immunocomplexes were resuspended in 25 l of Akt assay buffer containing 40 M ATP, 2.5 Ci of [␥-32 P]ATP, 1 M protein kinase A inhibitor peptide (Sigma), and 100 M Akt peptide substrate (Santa Cruz Biotechnology). The reaction mixture was incubated for 30 min at 30°C, and then an aliquot was spotted onto P81 phosphocellulose paper. Following extensive washing in 75 mM phosphoric acid, radioactivity on the papers was quantified by scintillation counting. Protein concentration of cell lysates was determined using a Bradford assay (Bio-Rad).
p70 S6 Kinase Assay-S6 kinase activity in cell lysates was measured using 40 S ribosomal subunits as substrate as described previously (11). The amount of 32 P incorporated into S6 was quantitated by liquid scintillation counting.
Immunocomplex PI 3-Kinase Assay-To measure PI 3-kinase activity, cell lysate (0.5-1 mg of protein) was incubated with 4 g of PY20 phosphotyrosine antibody (Transduction Laboratories, Lexington, KY) overnight at 4°C and then with 20 l protein A-agarose for 1 h. The beads were washed twice with lysis buffer and twice with PI 3-kinase assay buffer (20 mM Hepes, pH 7.5, 100 mM NaCl, and 50 M EDTA). The immunocomplexes were assayed for PI 3-kinase activity essentially as described previously (36), except that micelles of PI were produced by sonication in assay buffer. Reaction products were resolved by thin layer chromatography on silica gel plates. Spots containing PI 3-phosphate were scraped off the plate and quantitated by scintillation counting.
Immunoblotting-Cell lysates were prepared as described above, and proteins were subjected to SDS-gel electrophoresis on 10% polyacrylamide gels. Proteins were transferred onto nitrocellulose, and the blots were incubated with primary antibody. After extensive washing the blots were incubated with secondary antibody coupled to horseradish peroxidase, and proteins were visualized with the ECL detection kit (NEN Life Science Products). Membranes were stripped by incubating them for 30 min at 50°C in 62.5 mM Tris, pH 6.7, 2% SDS, and 100 mM 2-mercaptoethanol and then reprobed with another primary antibody.
Phospholipid Analysis-Rat-1 cells were seeded in 10-cm culture dishes at a density of 5 ϫ 10 5 cells per dish. After 24 h the medium was changed to inositol-free Dulbecco's modified Eagle's medium (Life Technologies, Inc.) containing 10% dialyzed fetal bovine serum and 10 Ci/ml [ 3 H]myo-inositol. The cells were allowed to incorporate the label for 48 h, and then they were incubated in serum-free medium for 12 h before treatment with growth factors. Phospholipids were extracted, deacylated, and fractionated on a Whatman Partisphere SAX HPLC column essentially as described previously (37). [ 3 H]Inositol present in each column fraction was determined by liquid scintillation counting. Column standards were produced by incubating PI, L-␣-PI-4-phosphate (Sigma), or L-␣-PI-4,5-bisphosphate (Sigma) with [␥-32 P]ATP and PI 3-kinase immunoprecipitated from PDGF-treated cells essentially as described above for the PI 3-kinase assay. The 32 P-labeled lipids were extracted, deacylated, and fractionated on the HPLC column as described above.
Measurement of [Ca 2ϩ ] i -Cells were loaded with 1.2 M fura-2 AM in high salt glucose buffer plus 1 mM Ca 2ϩ at 37°C for 30 -45 min (38). Loaded cells were washed with fresh buffer to remove the extracellular dye and resuspended in high salt glucose buffer containing either 1 mM Ca 2ϩ or 100 M EGTA. The 340/380 excitation ratio was measured with a PTI Delta Scan spectrofluorimeter (Photon Technology International, Inc., South Brunswick, NJ) using 340 and 380 nm excitation and 510 nm emission. [Ca 2ϩ ] i values were calculated according to the equation: where R is the 340/380 fluorescent ratio; R min and R max are the minimal and maximal fluorescent ratios, respectively; and the K d (dissociation constant) of fura-2 for Ca 2ϩ is taken to be 224 nM (39). After a stable basal fluorescent ratio was established, the cells were subjected to various treatments.
Apoptosis-Rat-1 cells expressing the ␣ 1A -AR were seeded at 3 ϫ 10 5 cells per 10-cm dish. The next day the medium was removed, and the cells were serum-starved for 12 h and then pretreated with or without agonists for 5 min. The medium was then removed, and the cells were exposed to 10 mJ/cm 2 of 254 nm UV light, and the medium with or without agonists was replaced. Apoptotic cells were labeled 6 h after irradiation using the fluorescein-based In Situ Cell Death Detection kit (Roche Molecular Biochemicals) and analyzed by flow cytometry according to the manufacturer's instructions.

Differential Activation of p70 S6 Kinase, PI 3-Kinase, and
Akt by the ␣ 1A -AR-The Rat-1 fibroblasts used in this study stably express the human ␣ 1A -AR; they do not express endogenous ␣ 1 -ARs (40). Treatment of the cells with PE, a selective ␣ 1 -AR agonist, promoted an increase in S6 kinase activity as measured in whole cell lysates. Maximal activation of approximately 3-fold was reached after 30 min in the presence of 10 M PE (Fig. 1A). The degree of stimulation achieved with PE was about 70% that obtained using IGF-I. Concurrent treatment with PE plus IGF-I did not lead to an additive effect but rather to a level of kinase activity that was lower than that obtained with IGF-I alone (Fig. 1A). The PE-induced increase in S6 kinase activity was almost completely blocked by pretreatment of cells with the immunosuppressant rapamycin (Fig. 1B), indicating that the activated S6 kinase is p70 S6 kinase and not Rsk (41). Indeed, we have previously shown that PE does not activate Erk1/2 or Rsk2 in these cells (42). In addition, activation of p70 S6 kinase by PE was abolished in cells pretreated with the PI 3-kinase inhibitor LY294002 (Fig.  1B). We next asked whether the ␣ 1A -AR signals to p70 S6 kinase via a PKC-dependent pathway. Down-regulation of PKC by long term treatment with PMA had no effect on the subsequent activation of p70 S6 kinase by PE, whereas activation induced by a 30-min treatment with PMA was lost (Fig. 1C). Thus, signaling to p70 S6 kinase by the ␣ 1A -AR appears to require PI 3-kinase but not PKC.
As outlined in the Introduction, a variety of evidence suggests that PI 3-kinase and its downstream effector Akt act upstream of p70 S6 kinase in signaling pathways activated by extracellular ligands. If this is the case, then PE treatment of Rat-1 fibroblasts expressing the ␣ 1A -AR should cause an increase in Akt activity. This hypothesis was tested by performing Akt immunocomplex kinase assays on lysates prepared from cells treated for 5 min with or without PE. Surprisingly, no increase in Akt-1 activity was detected in cells treated with PE ( Fig. 2A). Exposure of cells to PE for a longer time (up to 30 min; see Fig. 5B) or at a higher concentration (100 M; data not shown) did not lead to a measurable activation of Akt-1. By contrast, treatment with IGF-I caused a large increase (approximately 11-fold) in Akt-1 activity ( Fig. 2A). To assess the possibility that the ␣ 1A -AR might activate one of the other two known isoforms of Akt, we also performed immunocomplex kinase assays of Akt-2 and Akt-3. The basal Akt-2 activity was about 4 times higher than that of Akt-1 or Akt-3 in untreated cells, and treatment with PE or IGF-I induced no further increase in Akt-2 activity ( Fig. 2A). Little or no activation of Akt-2 by insulin has been observed in other cell types (43). Western blot analysis confirmed the presence of Akt-2 in these Rat-1 cells (data not shown). Immunocomplex kinase assays revealed that, similar to Akt-1, IGF-I stimulated a 9-fold increase in Akt-3 activity, whereas PE had no effect ( Fig. 2A).
These results were unexpected because activation of p70 S6 kinase in response to PE treatment was sensitive to LY294002 (Fig. 1B), implying that PI 3-kinase/Akt signaling participates in this response. To resolve this apparent contradiction, PI 3-kinase activity was directly measured by analyzing the production of 3-phosphorylated phosphoinositides in cells treated with PE. Rat-1 fibroblasts grown in the presence of [ 3 H]myoinositol were treated with or without agonists and 3 H-labeled phospholipids were analyzed by HPLC. The amount of PI 3,4bisphosphate and PI 3,4,5-trisphosphate recovered from cells simulated with PE was not increased above the control levels, whereas PDGF induced more than a 2-fold increase in the amount of both of these phospholipids (Fig. 2B). Together, these results show that PE induces a significant activation of p70 S6 kinase without a corresponding increase in Akt or PI 3-kinase activity in these cells.
Effect of ␣ 1A -AR on IGF-I-induced Akt Activation-Akt has been proposed to play an important role in insulin-induced glycogen synthesis (25). Since physiological stress induces a relatively insulin-resistant state, we wondered whether catecholamines might negatively regulate insulin signaling to Akt. We found that IGF-I-induced Akt-1 kinase activation was strongly suppressed in cells co-treated with PE (Fig. 3A). Activation of Akt-1 is associated with phosphorylation of Thr-308 in the catalytic domain and Ser-473 in the carboxyl-terminal tail (22). We used Western blots probed with antibodies that specifically recognize these phosphorylated residues to examine further the activation state of Akt-1 in cells treated with IGF-I and PE. As shown in Fig. 3B (upper two panels), Akt-1 in untreated cells was not phosphorylated at either Thr-308 or Ser-473, and IGF-I strongly stimulated the phosphorylation of both residues. In contrast, cells treated with PE contained no detectable phospho-Thr-308 or phospho-Ser-473. In cells treated with IGF-I plus PE, the phosphate content of both residues was sharply reduced as compared with cells treated with IGF-I alone (Fig. 3B). The blot was stripped and reprobed with a general Akt-1 antibody to show that an equal amount of Akt-1 was present in each lane (Fig. 3B, lower panel). These results indicate that the ␣ 1A -AR initiates a pathway that negatively regulates IGF-I-induced phosphorylation and activation of Akt-1.
Effect of ␣ 1A -AR on PI 3-Kinase Activation-We next searched for an upstream component in the Akt-1 signaling pathway that might be a target for PE-induced inhibition. PI 3-kinase is thought to be a component in the pathway, so we measured the activity of this enzyme in phosphotyrosine immunoprecipitates. Consistent with the phospholipid analysis (Fig. 2B), treatment of cells with PE alone for 5 min caused a reduction in the basal level of PI 3-kinase activity (Fig. 4). By contrast, PI 3-kinase activity in cells treated with IGF-I alone was increased about 2-fold above the basal level. Finally, PE strongly antagonized the IGF-I-induced activation of PI 3-kinase, reducing PI 3-kinase activity almost to the level seen in control cells (Fig. 4). The behavior of PI 3-kinase in response to these cell treatments closely parallels that of Akt-1 kinase activity (Fig. 3A). Thus, the ␣ 1A -AR appears to induce a negative regulatory mechanism that opposes Akt activation by inhibiting its upstream regulator, PI 3-kinase.
Increased serine phosphorylation of insulin receptor substrate-1 (IRS-1) has been shown to regulate negatively the activation of PI 3-kinase by insulin/IGF-I (44 -46). We wondered whether the inhibitory effect of the ␣ 1A -AR on PI 3-kinase is restricted to signaling pathways that include IRS-1 or whether it might be a more general phenomenon. Therefore, we tested the effect of PE on PI 3-kinase activation induced by PDGF, which does not signal through IRS-1. Cells were treated for increasing times with PDGF in the presence or absence of PE, and PI 3-kinase activity was measured in phosphotyrosine immunoprecipitates. PDGF was much more effective than IGF-I at activating PI 3-kinase in these cells; after 5 min in the presence of PDGF the PI 3-kinase activity increased approximately 55-fold over the basal level in untreated control cells (Fig. 5A). We were surprised that the PDGF-induced increase in 3-phosphorylated phosphoinositides in vivo (Fig. 2B) was relatively small in comparison to the large increase in PI 3-kinase activity measured in vitro (Fig. 5A). It could be that PI 3-kinase substrates are limiting in vivo or that 3-phosphorylated phosphoinositides produced in vivo are rapidly dephosphorylated. Activation of PI 3-kinase by PDGF was reduced approximately 50% in the presence of PE at the 5-min time point (Fig. 5A). At all times examined, PI 3-kinase activity was lower in cells treated with PDGF plus PE than in cells treated with PDGF alone (Fig. 5A). As a consequence of the inhibitory effect of PE on PI 3-kinase, PDGF-induced Akt-1 activity was also reduced in cells co-treated for 30 min with PE and PDGF (Fig.  5B). Thus, the inhibitory mechanism initiated by the ␣ 1A -AR targets PI 3-kinase even in signaling pathways that do not include IRS-1.
Effect of [Ca 2ϩ ] i on Akt-1 Activation-We have observed that PE treatment of Rat-1 fibroblasts expressing the ␣ 1A -AR induces a sustained increase in [Ca 2ϩ ] i that is much larger than that produced by IGF-I (40). 2 We therefore investigated whether Ca 2ϩ might be involved in the PE-induced mechanism that inhibits activation of the PI 3-kinase/Akt-1 pathway. In the first experiment, cells were pretreated with or without the Ca 2ϩ ionophore A23187 and then exposed to IGF-I to stimulate 2 R. Z. Lin, unpublished data.

FIG. 3. Inhibitory effect of PE on IGF-I-induced Akt-1 activation.
A, cells were treated with 100 ng/ml IGF-I, 10 M PE, or both agents together for 5 min, and Akt-1 activity was measured in immunocomplexes. B, equal amounts of cell lysate protein were subjected to SDS-gel electrophoresis followed by Western blotting (see "Experimental Procedures"). Phosphorylated sites in Akt-1 were detected with antibodies specific to phospho-Thr-308 (upper panel) or phospho-Ser-473 (middle panel). The blot was stripped and reprobed with a general Akt-1 antibody (lower panel). Akt-1. Consistent with an inhibitory role for high [Ca 2ϩ ] i , immunoprecipitation kinase assays showed that the IGF-I-induced activation of Akt-1 was strongly inhibited in cells pretreated with A23187 (Fig. 6A).
By having established that high [Ca 2ϩ ] i seems to regulate Akt-1 activity negatively, we directly assessed whether the inhibitory effect of PE on IGF-I-induced activation of the kinase is mediated by Ca 2ϩ . Rat-1 fibroblasts were incubated in medium containing 1 mM Ca 2ϩ or in Ca 2ϩ -free medium containing 2 mM EGTA to deplete intracellular Ca 2ϩ stores. The cells were then treated with or without agonists, and Akt-1 activity was measured in immunoprecipitates. Akt-1 activity in control and PE-treated cells was slightly reduced in cells incubated in Ca 2ϩ -free versus Ca 2ϩ -containing medium (Fig. 6B). IGF-I stimulation of Akt-1 was unaffected by the absence of intracellular Ca 2ϩ . Finally, depletion of intracellular Ca 2ϩ only slightly alleviated the PE-induced inhibition of the IGF-I response (Fig.  6B). In a parallel control experiment, cells incubated in the presence or absence of Ca 2ϩ -containing medium were stimulated with PE, and the intracellular Ca 2ϩ response was measured by spectrofluorimetric analysis. As expected, cells incubated in the presence of Ca 2ϩ had a higher basal [Ca 2ϩ ] i than those incubated in Ca 2ϩ -free medium (Fig. 6C). In addition, cells incubated in the presence of Ca 2ϩ generated a robust release of intracellular Ca 2ϩ in response to PE treatment, whereas those incubated in Ca 2ϩ -free medium showed no Ca 2ϩ response (Fig. 6C). Thus, although high [Ca 2ϩ ] i appears to antagonize Akt-1 activation, the PE-induced inhibition of Akt-1 activation by IGF-I can occur even in the absence of intracellular Ca 2ϩ release.
Effect of PE and IGF-I on UV-induced Apoptosis-Recent evidence indicates that the PI 3-kinase/Akt pathway delivers an anti-apoptotic signal (26,27). We reasoned that PE and IGF-I, because of their differential ability to activate PI 3-kinase and Akt-1, would also show differential ability to protect cells from UV-induced apoptosis. To test this hypothesis, Rat-1 cells were pretreated with or without PE or IGF-I and then exposed to UV irradiation. TUNEL-positive cells were then quantitated 6 h later by flow cytometry. Approximately 3.5% of the cells in the control culture were apoptotic (Fig. 7). After exposure to UV light this value increased to 9%. PE potentiated the lethal effect of UV exposure, whereas cell cultures kept in the presence of IGF-I showed a reduced number of apoptotic cells both with and without UV treatment (Fig. 7). This biological response correlates well with the stimulatory and inhibitory effects of IGF-I and PE, respectively, on the PI 3-kinase/Akt pathway.

DISCUSSION
The results presented here demonstrate that activation of Akt is not required for activation of p70 S6 kinase. Treatment of Rat-1 cells expressing the ␣ 1A -AR with PE did not stimulate any of the three Akt isoforms even though there was a significant increase in p70 S6 kinase activity (Figs. 1 and 2). We further show that p70 S6 kinase can be activated even in cells in which PI 3-kinase is inhibited (Figs. 1A and 4). This result was unexpected because pharmacological approaches and coexpression studies using highly active or dominant-negative versions of these signaling molecules have indicated that PI 3-kinase and Akt are major upstream regulators of p70 S6 kinase (7, 8, 10 -13, 15-17, 19). However, not all data are consistent with this hypothesis. First, activation of p70 S6

FIG. 5. Inhibitory effect of PE on PDGF-induced PI 3-kinase activation.
A, cells were treated for the indicated times with 50 ng/ml PDGF in the presence or absence of 10 M PE, and PI 3-kinase activity was measured in phosphotyrosine immunoprecipitates (see "Experimental Procedures"). Data are plotted as the percentage increase over basal PI 3-kinase activity measured in untreated control cells. B, cells were treated for 30 min with 10 M PE, 50 ng/ml PDGF, or both agonists together, and Akt-1 activity was measured in immunoprecipitates.
FIG. 6. Effect of intracellular Ca 2؉ on Akt-1 activation. A, cells were preincubated for 20 min with or without 10 M A23187 and then treated for 5 min plus or minus 100 ng/ml IGF-I. Akt-1 activity was measured in immunoprecipitates. B, cells were incubated in high salt glucose buffer containing 2 mM EGTA or 1 mM Ca 2ϩ as described under "Experimental Procedures." Then they were treated for 5 min with 10 M PE, 100 ng/ml IGF-I, or both agonists together, and Akt-1 activity was measured in immunoprecipitates. C, cells were loaded with fura-2 AM and incubated in high salt glucose buffer with (curve a) or without (curve b) Ca 2ϩ as described under "Experimental Procedures." When the base line stabilized, infusion with 10 M PE was started (arrow).
[Ca 2ϩ ] i was determined by spectrofluorimetric analysis. kinase by some agonists is relatively resistant to wortmannin, suggesting that PI 3-kinase-independent signaling pathways might be involved in activation of the kinase (11). Second, a PDGF receptor mutant (Y740F) that failed to activate PI 3-kinase or Akt activated p70 S6 kinase almost as well as the wild-type receptor (7,13). Third, although expression of a dominant-negative mutant of Akt-1 in CHO cells inhibited insulininduced activation of p70 S6 kinase by 75%, the same mutant had only a small inhibitory effect on p70 S6 kinase when expressed in 3T3-L1 adipocytes (18). Finally, Thomas and coworkers (20) have shown that activated mutants of Akt must be constitutively targeted to the membrane in order to stimulate p70 S6 kinase. They have proposed that these mutants artificially induce p70 S6 kinase activation and that results obtained with these mutants may not reflect wild-type Akt signaling (20). Interestingly, membrane-bound versions of active PI 3-kinase were also more efficient than cytosolic mutants at activating p70 S6 kinase (15). When expressed in COS-7 cells, the two forms of PI 3-kinase generated distinct patterns of phosphoinositides (15), raising the possibility that novel phospholipids generated in cells expressing these mutants induce p70 S6 kinase activation. PI 3,4-bisphosphate and PI 3,4,5-trisphosphate levels in intact cells (Fig. 2B) and PI 3-kinase activity in phosphotyrosine immunoprecipitates (Fig. 4) do not increase in response to PE treatment, thus leading us to conclude that activation of p70 S6 kinase by the ␣ 1A -AR is PI 3-kinase-independent. Why then is the activation of p70 S6 kinase by PE inhibited by LY294002 (Fig. 1B)? One possibility is that LY294002 inhibits a protein distinct from PI 3-kinase that is required for p70 S6 kinase activation. It has been shown that LY294002 and wortmannin inhibit the in vitro autophosphorylation of the mammalian target of rapamycin (mTOR), a kinase that positively regulates p70 S6 kinase (47). Therefore, the inhibitory effect of LY294002 on p70 S6 kinase activation could be exerted through mTOR.
PKC and Ca 2ϩ have been suggested to participate in PI 3-kinase-independent pathways that lead to p70 S6 kinase activation. Similar to the results obtained here with PE ( Figs.  1 and 2), exposure of a T cell leukemia line to PMA resulted in activation of p70 S6 kinase with no increase in Akt-1 activity (16). As expected for a PI 3-kinase-independent response, activation of p70 S6 kinase by PMA is relatively resistant to inhibition by wortmannin (11,12). It was reported earlier that a PDGF receptor mutant that couples to the ␥ subtype of phospholipase C but not to PI 3-kinase induces partial activation of p70 S6 kinase in HEPG2 cells (12). This response was abolished after long term treatment of the cells with PMA, indicating that phospholipase C activates p70 S6 kinase through PKC in this cell type (12). By contrast, we found that down-regulation of PKC in Rat-1 cells by long term PMA treatment has no effect on the activation of p70 S6 kinase by PE (Fig. 1C). Therefore, it appears that PI 3-kinase-independent activation of p70 S6 kinase by the ␣ 1A -AR in Rat-1 cells is mediated by a pathway that does not involve PMA-sensitive isoforms of PKC.
It has been known for some time that treatment of cells with A23187 induces the activation of p70 S6 kinase (11,41), and recent work has demonstrated that Ca 2ϩ ionophores and the Ca 2ϩ -mobilizing agent thapsigargin activate p70 S6 kinase independently of Akt-1 (35). Consistent with this observation, treatment of cells with these agents causes little or no activation of PI 3-kinase (11,35). On the other hand, activation of p70 S6 kinase by Ca 2ϩ ionophores and thapsigargin is wortmanninsensitive (11,35). These cellular responses are similar to those we observed here using PE (Figs. 1, 2, and 4). Thus, activation of p70 S6 kinase by the ␣ 1A -AR in Rat-1 cells may be mediated by a Ca 2ϩ -dependent pathway (Fig. 6C). Recent reports using EGTA-containing medium or BAPTA-AM to deplete cells of free intracellular Ca 2ϩ have indicated that growth factor signaling to p70 S6 kinase is Ca 2ϩ -dependent, whereas Akt activation occurs via a Ca 2ϩ -independent pathway (35,48). Our results here (Fig. 6) and elsewhere (49) confirm and extend these findings. We have found that the PE-induced activation of p70 S6 kinase is greatly reduced in Rat-1 cells depleted of free intracellular Ca 2ϩ (49). By contrast, the response of Akt to IGF-I is normal in Ca 2ϩ -depleted cells (Fig. 6B). The Ca 2ϩ -dependent step in p70 S6 kinase activation has not yet been identified.
In contrast to its stimulatory effect on p70 S6 kinase (Fig. 1), we found that PE negatively regulates Akt-1 (Fig. 3). The IGF-I-induced activation of Akt was inhibited in cells treated with A23187, suggesting that a high [Ca 2ϩ ] i inhibits signaling to this kinase (Fig. 6A). Surprisingly, although PE induces an increase in [Ca 2ϩ ] i (Fig. 6C), the inhibitory effect of the ␣ 1A -AR on Akt activation does not seem to be exerted through a Ca 2ϩmediated pathway. In Ca 2ϩ -depleted cells, PE was still unable to activate Akt, and its inhibitory effect on the IGF-I response was largely intact (Fig. 6B). It is not known whether Akt activity might be inhibited in other physiological settings that involve intracellular Ca 2ϩ release.
Consistent with its effect on Akt-1, we also found that treatment of cells with PE reduced the IGF-I-induced PI 3-kinase activation measured in phosphotyrosine immunoprecipitates (Fig. 4). Prior studies have shown that increased serine phosphorylation of the adaptor protein IRS-1 induced by a variety of factors inhibits the ability of the protein to be tyrosine-phosphorylated by the insulin receptor, thus preventing IRS-1 from binding and activating PI 3-kinase (44 -46). Two mechanisms have been proposed to mediate the serine phosphorylation of IRS-1. Activators of PKC are thought to promote the phosphorylation of IRS-1 at serine 612 by mitogen-activated protein kinases (44,45), whereas other factors such as PDGF are thought to regulate IRS-1 function negatively through the phosphorylation of three other serines via an Akt-dependent pathway (45). Inhibition of PI 3-kinase by the ␣ 1A -AR does not appear to involve either of these two mechanisms because (a) PE treatment of Rat-1 cells does not activate Akt ( Fig. 2A) or the mitogen-activated protein kinases Erk1 and Erk2 (42) and (b) PE also inhibits the activation of PI 3-kinase induced by the PDGF receptor (Fig. 5A), which does not utilize IRS-1 for signaling. Alternative mechanisms to explain the inhibitory effect of the ␣ 1A -AR on PI 3-kinase activity might be that it inhibits tyrosine phosphorylation of the p85 subunit of PI 3-kinase or prevents association of p85 with the p110 catalytic subunit of PI 3-kinase. In preliminary experiments, we observed no difference on Western blots in the pattern of tyrosine-phosphorylated proteins from cells treated with or without PE. 3 Attempts to examine the p85-p110 interaction in co-immunoprecipitates 3 M. E. Cross and R. Z. Lin, unpublished data. FIG. 7. Effect of PE on UV-induced apoptosis. Cells pretreated with or without 10 M PE or 100 ng/ml IGF-I were exposed to UV irradiation as described under "Experimental Procedures." Apoptotic cells were identified by TUNEL labeling and flow cytometry. Data are plotted as the percentage of labeled cells detected in 10,000 cells analyzed.
were unsuccessful due to the low amount of these proteins expressed in Rat-1 cells. 3 We are continuing to investigate the mechanism by which the ␣ 1A -AR negatively regulates PI 3-kinase.
Our results indicate that the ␣ 1A -AR differs from the insulin/ IGF-I receptor in its ability to activate PI 3-kinase (Figs. 2B and 4). This result may have significant physiologic implications for cell survival. It is well recognized that G q -coupled receptors play an important role in the development and ultimate decompensation of cardiac hypertrophy (50). It was recently demonstrated that overexpression of G q leads to increased apoptosis of cardiac myocytes both in vitro and in vivo (51). In contrast, treatment with IGF-I stimulates cardiac myocyte hypertrophy but improves cardiac function in experimental models of heart failure (52). This disparity may be explained by the differential activation of PI 3-kinase and Akt by the IGF-I receptor versus the ␣ 1 -AR (or, possibly, G q -coupled receptors in general). Perhaps exposure of cardiac myocytes to ␣ 1 -AR agonists or IGF-I activates p70 S6 kinase, leading to increased protein synthesis and cellular hypertrophy, whereas only IGF-I activates Akt, providing a survival signal. Indeed, our results show that IGF-I treatment has a protective effect against UVinduced apoptosis, whereas PE treatment enhanced the apoptotic effect of UV irradiation (Fig. 7). Studies in rat neonatal cardiac myocytes are currently being done to test the validity of this clinically relevant hypothesis.
To our knowledge, the finding that activation of a G q -coupled receptor leads to inhibition of Akt and PI 3-kinase has not been previously reported. By using COS-7 cells overexpressing m1 muscarinic acetylcholine receptors and epitope-tagged Akt, Gutkind and co-workers (53) found that activation of this G qcoupled receptor very weakly stimulated Akt. The difference between that finding and the results reported here may be due to cell type differences and alteration in the regulation of Akt when it is overexpressed. Activation of the ␥ subtype of PI 3-kinase by G i -coupled receptors via G protein ␤␥ subunits is well described in the literature (54,55). Not surprisingly, recent reports indicated that G i -coupled receptors also activate Akt (56,57). Inhibition of Akt by ␣ 1 -ARs has important physiological implications for Akt-mediated glucose regulation. Akt is thought to be a critical molecular switch for insulin-mediated regulation of glucose metabolism (23). The possibility that G qcoupled receptors negatively regulate these insulin-induced effects opens new avenues for research examining the role of this large family of receptors in the development of insulin resistance and diabetes mellitus.