Dual Regulation of Glycogen Synthase Kinase-3β by the α1A-Adrenergic Receptor*

Catecholamines, acting through adrenergic receptors, play an important role in modulating the effects of insulin on glucose metabolism. Insulin activation of glycogen synthesis is mediated in part by the inhibitory phosphorylation of glycogen synthase kinase-3 (GSK-3). In this study, catecholamine regulation of GSK-3β was investigated in Rat-1 fibroblasts stably expressing the α1A-adrenergic receptor. Treatment of these cells with either insulin or phenylephrine (PE), an α1-adrenergic receptor agonist, induced Ser-9 phosphorylation of GSK-3β and inhibited GSK-3β activity. Insulin-induced GSK-3β phosphorylation is mediated by the phosphatidylinositol 3-kinase/Akt signaling pathway. PE treatment does not activate phosphatidylinositol 3-kinase or Akt (Ballou, L. M., Cross, M. E., Huang, S., McReynolds, E. M., Zhang, B. X., and Lin, R. Z. (2000) J. Biol. Chem. 275, 4803–4809), but instead inhibits insulin-induced Akt activation and GSK-3β phosphorylation. Experiments using protein kinase C (PKC) inhibitors suggest that phorbol ester-sensitive novel PKC and Gö6983-sensitive atypical PKC isoforms are involved in the PE-induced phosphorylation of GSK-3β. Indeed, PE treatment of Rat-1 cells increased the activity of atypical PKCζ, and expression of PKCζ in COS-7 cells stimulated GSK-3β Ser-9 phosphorylation. In addition, PE-induced GSK-3β phosphorylation was reduced in Rat-1 cells treated with a cell-permeable PKCζ pseudosubstrate peptide inhibitor. These results suggest that the α1A-adrenergic receptor regulates GSK-3β through two signaling pathways. One pathway inhibits insulin-induced GSK-3β phosphorylation by blocking insulin activation of Akt. The second pathway stimulates Ser-9 phosphorylation of GSK-3β, probably via PKC.

Glycogen synthase kinase-3 (GSK-3) 1 is an evolutionarily conserved signaling molecule that plays an important role in diverse biological processes, including metabolism, differentiation, and development. GSK-3 was initially identified as a kinase that phosphorylates glycogen synthase and has been found to play a key role in the regulation of glycogen metabolism in insulin-responsive cells (reviewed in Ref. 1). GSK-3 also regulates cell fate determination in Dictyostelium and has been intensively studied as a component of the Wg/Wnt signaling pathway in invertebrates and vertebrates that controls development (2)(3)(4).
The activity of glycogen synthase, the rate-limiting enzyme in glycogen synthesis, is increased by dephosphorylation of at least four serine residues in response to insulin treatment (5). Activation of PP1G, a protein phosphatase that can dephosphorylate all the sites in glycogen synthase, is one mechanism used by insulin to increase the activity of the enzyme (6). In addition, a protein kinase cascade has also been identified that appears to play an important role in insulin regulation of glycogen synthase (reviewed in Ref. 1). Two pharmacological inhibitors of phosphatidylinositol 3-kinase (PI3K), wortmannin and LY 294002, block insulin stimulation of glycogen synthase (7), suggesting that PI3K is required for this response. Insulin activation of PI3K and its downstream effector, the protein kinase Akt, induces the phosphorylation and inhibition of GSK-3. GSK-3 phosphorylates glycogen synthase at four sites (3a, 3b, 3c, and 4), provided that site 5 has already been phosphorylated by casein kinase II (1,5). Phosphorylation of sites 3a and 3b has been shown to potently inactivate the enzyme. Labeling studies with 32 P have shown that phosphate turnover is relatively rapid in glycogen synthase, so inhibition of GSK-3 activity by the PI3K/Akt signaling pathway could lead to net dephosphorylation of sites 3a and 3b, resulting in increased glycogen synthase activity.
Two isoforms of GSK-3 (GSK-3␣ and GSK-3␤) have been identified in mammalian tissues, and both are negatively regulated by insulin (1). Both isoforms were found to be phosphorylated at a single tyrosine residue in unstimulated cells, and dephosphorylation of GSK-3␤ with a protein tyrosine phosphatase in vitro inactivates the enzyme (8). These results led to the suggestion that insulin inactivates GSK-3 through tyrosine dephosphorylation. However, subsequent studies showed that insulin-induced inhibition of GSK-3 is due to an increase in phosphorylation of Ser-21 in GSK-3␣ and Ser-9 in GSK-3␤ (9,10). Several kinases have been shown to phosphorylate these inhibitory sites in vitro (11)(12)(13)(14)(15)(16), but Akt is thought to be the physiologically relevant kinase in vivo that inactivates GSK-3 in response to insulin. This conclusion is based on the observations that insulin-induced inhibition of GSK-3 is blocked in cells treated with wortmannin or LY 294002 (9,17); coexpression of GSK-3␤ with wild-type or constitutively active Akt results in the inhibition of GSK-3 activity (10); expression of a dominant-negative mutant of Akt blocks the effect of insulin on GSK-3␤ (18); and Akt phosphorylates and inactivates both GSK-3 isoforms in vitro (9). In addition, expression of a constitutively active Akt mutant leads to increased glycogen synthase activity only in cells that contain GSK-3 (19).
It is well known that catecholamines act through both ␣and ␤-adrenergic receptors to modulate the effects of insulin on glucose metabolism. For example, stimulation of ␣ 1 -adrenergic receptors in the livers of fed rats inhibits glycogen synthase and promotes gluconeogenesis (20). Adrenergic receptors are members of the G protein-coupled receptor family, and three subtypes of the ␣ 1 -adrenergic receptor (␣ 1A , ␣ 1B , and ␣ 1D ) have been identified by molecular cloning (21). All three subtypes are widely expressed in a number of tissues and are coupled to ␣-subunits in the G q/11 family that are insensitive to inhibition by pertussis toxin (21)(22)(23). It is important to note that, in addition to ␣ 1 -adrenergic receptors, other receptors that couple to G q/11 , including vasopressin, purinergic, and angiotensin II receptors, also inhibit glycogen synthase (24,25).
In contrast to what is known about how insulin affects GSK-3 activity, relatively little is known about how G q -coupled receptors regulate GSK-3 in mammalian cells. In this study, we investigated the regulation of GSK-3␤ by the ␣ 1A -adrenergic receptor stably expressed in Rat-1 fibroblasts. We found that activation of the ␣ 1A -adrenergic receptor leads to a modest increase in phosphorylation of GSK-3␤ at Ser-9 and inhibition of its catalytic activity. However, activation of the ␣ 1A -adrenergic receptor does not activate PI3K or Akt (26), but instead attenuates insulin-induced GSK-3␤ phosphorylation by inhibiting Akt. These results may explain how catecholamines modulate insulin activation of glycogen synthase.
DNA Constructs-The GSK-3␤ cDNA was isolated by reverse transcriptase polymerase chain reaction with primers 5Ј-CGCCTCGAGAT-GTCAGGGCGGCCCAGA (forward) and 5Ј-CGCGAATTCTCAGGTGG-AGTTGGAAGCTG (reverse) using RNA from Swiss mouse 3T3 cells as template. A hemagglutinin (HA) epitope tag was introduced at the 5Ј-end of the GSK-3␤ cDNA by polymerase chain reaction using primers 5Ј-CGCCTCGAGGCCACCATGGCATACCCCTACGACGTGCCCGAC-TACGCCTCAGGGCGGCCCAGAACC and the reverse primer shown above to make HA-GSK-3␤. The cDNA fragment was digested with EcoRI and XhoI and subcloned into pcDNA3.1/Zeo (Invitrogen, Carlsbad, CA). The HA-GSK-3␤ S9A phosphorylation site mutant was constructed using the QuikChange site-directed mutagenesis kit (Stratagene, La Jolla, CA) and the forward primer 5Ј-CCCAGAACCACCG-CTTTGCGGAGAGC. The fidelity of all cDNA constructs was verified by sequencing. Myc-tagged wild-type PKC and kinase-dead K273A PKC were obtained from Dr. Feng Liu (University of Texas Health Science Center, San Antonio, TX) (27).
Cell Culture-Rat-1 fibroblasts stably transfected with the human ␣ 1A -adrenergic receptor were a gift from Pfizer (28). Rat-1 and COS-7 cells were maintained in Dulbecco's modified Eagle's medium (Mediatech, Herndon, VA) containing 10% fetal bovine serum (Sigma). Rat-1 cells stably expressing the human ␣ 1A -adrenergic receptor and HA-GSK-3␤ or HA-GSK-3␤ S9A were selected using the antibiotic zeomycin. Cells were incubated in serum-free medium for 16 -18 h before treatments. 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 before adding PE to the buffer. COS-7 cells were seeded at 2 ϫ 10 5 cells/six-well dish in growth medium and the next day were transfected with 2 g of DNA in 1.25 ml of Opti-MEM plus 5 l of LipofectAMINE (Life Technologies, Inc.). After 5 h, the transfection solutions were replaced with growth medium, and the cells were allowed to grow for 1 day. The cells were then placed in serum-free medium overnight, and cell extracts were prepared as described below.
Immunocomplex Kinase Assays-After treatment, cells were rinsed with ice-cold phosphate-buffered saline and then scraped into 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 each aprotinin and leupeptin). Homogenates were centrifuged at 12,000 ϫ g for 10 min at 4°C, and supernatants were retained for kinase assays and Western blotting. Akt-1 activity was assayed following a method described previously (26). To assay GSK-3␤ activity, equal amounts of lysate protein were incubated with antibody to either GSK-3␤ (Transduction Laboratories, Lexington, KY) or HA (Covance) for 3 h on ice and then with protein G-agarose for 1 h. The immunocomplexes were washed three times with lysis buffer and once with kinase assay buffer (40 mM Hepes, pH 7.4, 20 mM 2-glycerophosphate, and 10 mM MgCl 2 ). The reaction was initiated with 25 l of assay buffer containing 40 M ATP, 2.5 Ci of [␥-32 P]ATP, 1 M protein kinase A (PKA) inhibitor peptide (Sigma), and 1 g of phospho-GS2 substrate peptide (Upstate Biotechnology, Inc.). After incubation at 30°C for 20 min, an aliquot was spotted onto P-81 phosphocellulose paper; and following extensive washing with 75 mM phosphoric acid, radioactivity on the papers was quantified by scintillation counting.
PKC activity was assayed following a previously described method with modifications (29). Equal amounts of lysate protein were incubated with anti-PKC antibody (Transduction Laboratories) for 3 h on ice and then with protein G-agarose for 1 h. Following extensive washing, the reaction was initiated with 25 l of assay buffer containing 50 mM Tris, pH 7.5, 5 mM MgCl 2 , 1 mM NaHCO 3 , 1 mM phenylmethylsulfonyl fluoride, 40 g/ml phosphatidylserine, 40 M ATP, 2.5 Ci of [␥-32 P]ATP, and 40 M [Ser 159 ]PKC⑀-(149 -164) substrate peptide (Quality Controlled Biochemicals). The reactions were incubated for 10 min at 30°C. An aliquot was spotted onto P-81 phosphocellulose paper, and radioactivity was quantified as described above.
Immunoblotting-Cell lysates were prepared as described above, and proteins were subjected to SDS gel electrophoresis and Western blotting as previously described (26).
Adenoviral Constructs-An epitope-tagged Akt construct (Akt-HA) was obtained from Dr. Richard Roth (Stanford University, Stanford, CA). Polymerase chain reaction was used to incorporate the myristoylation/palmitoylation sequence from the Lck tyrosine kinase into the 5Ј-end of Akt-HA to produce m-Akt-HA. The primers used were 5Ј-CG-CCTCGAGGCCACCATGGGCTGCGTGTGCAGCAGCAACCCCGAGG-ACGACAGCGACGTGGCTATTGTGAAG (forward) and 5Ј-CGCGATA-TCTCACAGTCCAGGTCCCAGAC (reverse). Recombinant adenoviruses expressing m-Akt-HA (pAdm-Akt) were constructed using a method described by He et al. (30). Briefly, m-Akt-HA was first subcloned into the shuttle vector pAdTrack, and the resultant plasmid was linearized with PmeI and transformed into Escherichia coli BJ5183 together with the adenoviral backbone plasmid pAdEasy-1. Recombinants were selected, and recombination was confirmed by restriction analysis. Finally, the linearized recombinant plasmid was transfected into a packaging cell line (HEK 293 cells). The adenoviruses produced were used to infect additional HEK 293 cells, and a high titer adenovirus stock was made following multiple rounds of infection. The viruses were purified by CsCl banding. Concentrated viruses were incubated with Rat-1 cells under conditions that caused nearly 100% of the cells to be infected (data not shown). Adenoviruses expressing green fluorescent protein (pAdGFP) were also constructed to serve as a negative control.

Phenylephrine Antagonizes Insulin Inhibition of GSK-3␤-
Insulin activation of Akt leads to the inhibitory phosphorylation of GSK-3␤ at Ser-9, which promotes the dephosphorylation and activation of glycogen synthase. It is well known that catecholamine activation of adrenergic receptors opposes insulin action by inhibiting glycogen synthesis. We showed previously that stimulation of the ␣ 1A -adrenergic receptor with PE, a specific ␣ 1 -adrenergic receptor agonist, inhibits insulin-like growth factor I-induced activation of Akt (26). Therefore, we investigated if PE treatment also inhibits insulin-induced GSK-3␤ phosphorylation. Rat-1 cells stably expressing the ␣ 1Aadrenergic receptor were treated with or without insulin in the presence or absence of PE. Cell extracts were then subjected to Western blotting to measure Ser-9 phosphorylation. As expected, exposure of the cells to insulin alone induced GSK-3␤ phosphorylation (Fig. 1A). Unexpectedly, PE treatment alone also caused an increase in GSK-3␤ phosphorylation, although the response was weaker than that seen with insulin ( Fig. 1A). (The effect of PE alone on GSK-3␤ will be examined further below.) Cotreatment of the cells with insulin plus PE reduced the insulin-stimulated phosphorylation to the level seen with PE alone (Fig. 1A). To determine whether Ser-9 phosphorylation correlated with a decrease in GSK-3␤ kinase activity, Rat-1 cells stably expressing both the ␣ 1A -adrenergic receptor and HA-tagged wild-type GSK-3␤ were treated for 10 min with or without PE or insulin, and then GSK-3␤ activity was measured in HA immunoprecipitates ( Fig. 2A). GSK-3␤ activity was reduced by 40% in insulin-treated cells and by 20% in cells treated with PE alone (Fig. 2B). The insulin-induced inhibition of GSK-3␤ was relieved in the presence of PE (Fig. 2B).
The effect of PE on Akt, which acts upstream of GSK-3␤ in the insulin signaling pathway, was also examined. Akt activity in Rat-1 cell extracts was measured using an in vitro immunocomplex kinase assay. Insulin treatment increased Akt activity ϳ6-fold, and this response was totally suppressed in the presence of PE (Fig. 1B). PE by itself did not activate Akt. These results suggested that the ␣ 1A -adrenergic receptor attenuates the effect of insulin on GSK-3 activity by inhibiting Akt activation. To test this hypothesis further, cells were infected with either a control virus (pAdGFP) or a virus encoding a constitutively active form of Akt (pAdm-Akt). Similar to results reported by others using distinct Akt mutants (10,19), expres-sion of m-Akt-HA caused an increase in Ser-9 phosphorylation of GSK-3␤ in the absence of any agonist treatment (Fig. 1C). Treatment of these cells with PE did not cause a decrease in Ser-9 phosphorylation, implying that the ␣ 1A -adrenergic receptor inhibits insulin signaling to GSK-3 at or above the level of Akt. Together, these results show that activation of the ␣ 1Aadrenergic receptor opposes insulin activation of Akt and insulin inhibition of GSK-3␤. The change in GSK-3␤ activity could account for the counter-regulatory effect of catecholamines on insulin stimulation of glycogen synthase.

PE-induced GSK-3␤ Inhibition Is Mediated by Ser-9
Phosphorylation-The results showing that treatment of Rat-1 cells with PE alone induced Ser-9 phosphorylation (Fig. 1A) and GSK-3␤ inhibition (Fig. 2B) were surprising because we anticipated that activation of the ␣ 1A -adrenergic receptor would have effects opposite to those induced by insulin. To characterize further the effect of the ␣ 1A -adrenergic receptor on GSK-3␤, Rat-1 cells were treated with PE for increasing times, and GSK-3␤ activity was measured in cell extracts. Maximal inhibition of 25% was seen after 10 min of PE treatment, and the activity remained essentially the same for the next 20 min (Fig.  3A). In comparison, insulin treatment induced a 35% inhibition of GSK-3␤ activity at 10 min, and the inhibitory effect increased to ϳ47% after 30 min of treatment (Fig. 3A). The phosphorylation of Ser-9 seemed to correlate well with GSK-3␤ inhibition in cells exposed to PE, considering the semiquantitative nature of Western blotting. GSK-3␤ phosphorylation reached a maximum at 5 min and did not change for the remainder of the time course (Fig. 3B). The signal with insulintreated cells was stronger than that seen with PE-treated cells (Fig. 3B). Phosphorylation of GSK-3␤ in response to insulin also reached a maximum after 5 min and did not seem to change over the next 25 min. Thus, GSK-3␤ inhibition correlated poorly with Ser-9 phosphorylation in cells treated with insulin. The reason for this apparent discrepancy is not known.
We wondered if inhibition of GSK-3␤ by the ␣ 1A -adrenergic receptor is mediated solely by phosphorylation of Ser-9. For example, there is evidence suggesting that GSK-3 activity can also be regulated by tyrosine phosphorylation (8,31) and by protein-protein interactions (reviewed in Ref. 3). To determine the contribution of Ser-9 phosphorylation to PE-induced GSK-3␤ inhibition, an HA-tagged phosphorylation site mutant (S9A) of GSK-3␤ was stably expressed in Rat-1 cells containing the ␣ 1A -adrenergic receptor. Fig. 2A (upper panel) shows that Ser-9 in endogenous GSK-3␤ (lower band) was phosphorylated in insulin-treated parental cells and in cells expressing either wild-type HA-GSK-3␤ or the S9A mutant. Wild-type HA-GSK-3␤ (upper band), but not the S9A mutant, was also phosphorylated in insulin-treated cells. Fig. 2A (lower panel) shows that the HA-tagged GSK-3␤ proteins were expressed at equal levels in both cell populations.
These cells were then used to compare the inhibition of wild-type HA-GSK-3␤ and HA-GSK-3␤ S9A following PE treatment. The cells were treated with insulin or PE, and GSK-3␤ activity was measured in HA immunoprecipitates. HA-GSK-3␤ was inhibited by ϳ20% in cells treated with PE (Fig. 2B), whereas there was a 10 -15% increase in activity of HA-GSK-3␤ S9A under the same conditions (Fig. 2C). Likewise, insulin inhibited the activity of wild-type HA-GSK-3␤ (Fig. 2B), but slightly activated the S9A mutant (Fig. 2C). These results indicate that PE-induced GSK-3␤ inhibition is mediated by Ser-9 phosphorylation.
PE-induced Signaling Pathways That Mediate GSK-3␤ Ser-9 Phosphorylation-Several serine/threonine kinases, including Akt (9), p70 S6 kinase (11,12), PKA (13,14), p90 rsk (11,12), and some PKC isoforms (15,16), have been shown to phosphorylate and inhibit GSK-3 in vitro. We examined the involvement of these kinases in the PE-induced phosphorylation of GSK-3␤. Akt does not appear to be involved in this pathway, as PE did not activate Akt or its upstream regulator PI3K in Rat-1 cells (Fig. 1B) (26). We also tested if PE-induced Ser-9 phosphorylation is inhibited by a pharmacological inhibitor of PI3K. Cells were pretreated with 50 M LY 294002 for 30 min prior to stimulation with PE, and GSK-3␤ phosphorylation was analyzed by Western blotting. We found that PE-induced GSK-3␤ phosphorylation was not affected by LY 294002 treatment, whereas the inhibitor strongly antagonized the phosphorylation of GSK-3␤ induced by insulin (Fig. 4A). These results are consistent with our hypothesis that the PI3K/Akt signaling pathway does not mediate the stimulatory effect of the ␣ 1Aadrenergic receptor on GSK-3␤ phosphorylation.
We have previously shown that activation of the ␣ 1A -adrenergic receptor in these cells causes an increase in p70 S6 kinase activity (26) and cAMP levels (32). However, PE-induced GSK-3␤ phosphorylation was not blocked by pretreatment with , and GSK-3␤ activity was measured in cell lysates (see "Experimental Procedures"). The data are means Ϯ S.E. from three experiments and are plotted as a percentage of the basal GSK-3␤ activity measured in untreated control cells. The asterisk represents significant differences between PE versus vehicle and insulin versus vehicle (p Ͻ 0.05) at the 10-min time point. The data were analyzed by one-way analysis of variance, and pairwise comparisons were made using Fisher's post-hoc tests. B, equal amounts of cell lysate protein from A were subjected to Western blotting, and GSK-3␤ phosphorylated at Ser-9 was detected with a phospho-specific antibody (upper panels). The blots were reprobed with a general anti-GSK-3␤ antibody (lower panels). The experiment was repeated with similar results.

FIG. 4. Effect of inhibitors on PE-induced GSK-3␤ phosphorylation.
Equal amounts of lysate protein from Rat-1 cells treated as described below were analyzed on Western blots to visualize GSK-3␤ Ser-9 phosphorylation (upper panels). The blots were reprobed with a general anti-GSK-3␤ antibody (lower panels). Experiments were performed two or three times, yielding similar results, and representative data are shown. A, cells were preincubated for 30 min with or without 50 M LY 294002 (LY) prior to stimulation for 5 min with 10 M PE or 1 M insulin (INS). B, cells were pretreated for 30 min with or without 5 M Gö 6983 prior to stimulation with 10 M PE for 5 min. C, cells were preincubated in serum-free medium with 0.1% dimethyl sulfoxide or 100 nM PMA for 24 h. Then the cells were washed and incubated for 30 min in fresh serum-free medium prior to treatment with 10 M PE or 100 nM PMA for 5 min as indicated. D, cells were preincubated in buffer containing EGTA or Ca 2ϩ (see "Experimental Procedures") and then stimulated for 5 min with or without 10 M PE. GSK-3␤ Ser-9 phosphorylation and total GSK-3␤ were visualized on a Western blot (left panels). The blot was stripped and reprobed with antibodies to phospho-Thr-389 p70 S6 kinase (p70S6k) and total p70 S6 kinase (right panels). Con, control. either rapamycin (data not shown) or LY 294002 (Fig. 4A) to inhibit p70 S6 kinase (26) or with 8-Br-cAMP-S to inhibit PKA (data not shown). We have also shown that treatment of Rat-1 cells with catecholamines does not activate p90 rsk or its upstream regulators ERK1 and ERK2 (32). Together, these results suggest that p70 S6 kinase, PKA, and p90 rsk do not play a role in the phosphorylation of GSK-3␤ in response to stimulation of the ␣ 1A -adrenergic receptor.
Like all G q -coupled receptors, activation of the ␣ 1A -adrenergic receptor causes an increase in phospholipase C␤ activity that leads to activation of diacylglycerol-dependent PKCs and an increase in intracellular Ca 2ϩ . To examine the role of PKCs in mediating ␣ 1A -adrenergic receptor signaling to GSK-3␤, we first determined if PE-induced GSK-3␤ phosphorylation is blocked by a broad-spectrum PKC inhibitor. Rat-1 cells were pretreated with 5 M Gö 6983 for 30 min prior to stimulation with PE, and Ser-9 phosphorylation was visualized by Western blotting. Pretreatment with Gö 6983 completely inhibited PEinduced GSK-3␤ phosphorylation (Fig. 4B), suggesting that PKCs are involved in this response. Next, the role of diacylglyceroldependent PKCs was examined in cells treated with or without 100 nM PMA for 24 h to down-regulate these enzymes. Acute PMA treatment of control cells strongly stimulated the phosphorylation of GSK-3␤ (Fig. 4C), similar to results previously reported for GSK-3␣ in Swiss 3T3 cells (33). In cells pretreated with PMA, the basal level of GSK-3␤ phosphorylation increased slightly, but increased phosphorylation induced by a subsequent challenge with PMA was abolished (Fig. 4C). More important, PE-induced phosphorylation of GSK-3␤ was partially blocked in cells depleted of diacylglycerol-dependent PKCs (Fig. 4C). We next examined the role of Ca 2ϩ in ␣ 1Aadrenergic receptor signaling to GSK-3␤ by incubating cells in medium containing EGTA under conditions known to completely abolish the PE-induced increase in intracellular Ca 2ϩ concentration (26). The cells were then treated with or without PE, and Ser-9 phosphorylation of GSK-3␤ was measured by Western blot analysis. We found that there was no difference in Ser-9 phosphorylation in the presence or absence of intracellular Ca 2ϩ release, whereas PE-induced phosphorylation of p70 S6 kinase was significantly decreased in Ca 2ϩ -depleted cells (Fig. 4D). Together, these results suggest that diacylglyceroldependent novel PKCs, but not conventional Ca 2ϩ -dependent PKCs, play a role in signaling by the ␣ 1A -adrenergic receptor to GSK-3␤. By contrast, we showed previously that PE-induced activation of p70 S6 kinase and phosphorylation of 4E-BP1 still occurs in Rat-1 cells subjected to prolonged treatment with PMA, but not in cells depleted of intracellular Ca 2ϩ (34).
Effect of PKC on GSK-3␤ Ser-9 Phosphorylation-Since PEinduced GSK-3␤ phosphorylation was only partially blocked by down-regulation of diacylglycerol-dependent PKCs, but totally inhibited by treatment with Gö 6983, this suggested that atypical PKCs might also be involved. The angiotensin II receptor, a G q -coupled receptor, has been shown to signal through PKC (35), but it is unknown if this atypical PKC can also be activated by the ␣ 1A -adrenergic receptor. We therefore measured PKC activity with an immunocomplex kinase assay after treating Rat-1 cells with or without PE for 5 min. PKC activity increased ϳ2-fold in cells treated with PE compared with untreated control cells (Fig. 5A). We next investigated if increased PKC activity stimulates an increase in GSK-3␤ phosphorylation. Because Rat-1 cells undergo transient transfection at a very low efficiency and do not overexpress exogenous proteins at a high level (for example, see Fig. 2A, upper panel), COS-7 cells were used for these experiments. The cells were cotransfected with HA-GSK-3␤ and either Myc-tagged wild-type PKC or kinase-dead PKC. Extracts were made after the cells were serum-starved overnight, and proteins were immunoprecipitated with anti-HA antibody. Western blot analysis was then done on the immunoprecipitates to detect GSK-3␤ phosphorylated at Ser-9. We found that expression of wild-type PKC caused a small increase in HA-GSK-3␤ phosphorylation compared with cells transfected with HA-GSK-3␤ alone or with HA-GSK-3␤ together with kinase-dead PKC (Fig. 5B, upper  panel). The membrane was stripped and reprobed with anti-HA antibody to confirm that similar amounts of HA-GSK-3␤ were present in each lane (Fig. 5B, middle panel), and Western blotting was done to confirm that the Myc-tagged PKC proteins were expressed (Fig. 5B, lower panel).
To further investigate the role of PKC in signaling by the ␣ 1A -adrenergic receptor to GSK-3␤, Rat-1 cells were treated with a cell-permeable pseudosubstrate peptide of PKC that has been shown to inhibit agonist-induced activation of the enzyme (36). The stimulation of Ser-9 phosphorylation by PE was reduced in the presence of this peptide inhibitor (Fig. 5C). Together with the data using Gö 6983 to inhibit PKC (Fig. 4B) (37), these results suggest that PKC might be partly respon- sible for mediating Ser-9 phosphorylation of GSK-3␤ in response to stimulation of the ␣ 1A -adrenergic receptor. DISCUSSION The results presented here demonstrate that the ␣ 1A -adrenergic receptor can modulate GSK-3␤ activity in Rat-1 cells in two ways. One mechanism antagonizes insulin-induced Ser-9 phosphorylation to cause an increase in GSK-3␤ activity, and the second promotes the phosphorylation of GSK-3␤ at Ser-9 to cause a decrease in GSK-3␤ activity.
Recent experiments strongly suggest that Akt is the kinase that phosphorylates and inhibits GSK-3 in response to insulin (1). We show here that stimulation of the ␣ 1A -adrenergic receptor with PE antagonizes insulin activation of Akt (Fig. 1). As a consequence, insulin-induced Ser-9 phosphorylation and inhibition of GSK-3␤ are also reduced ( Figs. 1 and 2). The inhibitory effect of the ␣ 1A -adrenergic receptor on insulin signaling to GSK-3␤ is overcome by expression of constitutively active Akt (Fig. 1C) or by mutation of Ser-9 in GSK-3␤ to Ala (Fig. 2). These results indicate that the effect of the ␣ 1A -adrenergic receptor on insulin regulation of GSK-3␤ is most likely due to inhibition of the PI3K/Akt signaling pathway. This mechanism could explain the counter-regulatory effect of catecholamines on insulin activation of glycogen synthase. We showed previously that treatment of Rat-1 cells with PE inhibits the activation of PI3K and Akt induced by insulin-like growth factor I and platelet-derived growth factor (26). We have found that PE also inhibits insulin-induced PI3K activation (data not shown). The mechanism by which the ␣ 1A -adrenergic receptor inhibits PI3K activation by these diverse tyrosine kinase receptors is under investigation.
Surprisingly, treatment of Rat-1 cells with PE alone induced a modest increase in phosphorylation of GSK-3␤ and inhibition of its activity (Fig. 3). We suspect that this effect of the ␣ 1Aadrenergic receptor on GSK-3 would not result in the activation of glycogen synthase, as catecholamines have been shown to decrease the activity of glycogen synthase in hepatocytes from fasted rats (38). Presumably, dephosphorylation of glycogen synthase by PP1G could counteract the effect of GSK-3 under these conditions. On the other hand, we cannot rule out the possibility that inhibition of GSK-3 by the ␣ 1A -adrenergic receptor could play a role in a cellular process not related to glucose metabolism, such as development. The inhibitory effect of the ␣ 1A -adrenergic receptor on GSK-3␤ activity in Rat-1 cells was abolished using the S9A mutant (Fig. 2), indicating that inhibition is dependent on Ser-9 phosphorylation. Unlike insulin, increased Ser-9 phosphorylation of GSK-3␤ in response to PE is independent of PI3K/Akt signaling since PE did not activate PI3K or Akt (Fig. 1) (26), and the effect of PE on GSK-3␤ phosphorylation was not blocked by LY 294002 (Fig.  4A). These results indicate that the signaling pathways used by the tyrosine kinase receptor and the G q -coupled receptor to regulate GSK-3␤ activity are distinct.
We found that PE-induced phosphorylation of GSK-3␤ was blocked in the presence of a general PKC inhibitor (Fig. 4B), suggesting that the ␣ 1A -adrenergic receptor uses a PKC-dependent pathway to regulate the enzyme. It was shown previously that acute treatment of mouse fibroblasts or A431 cells with PMA inactivates GSK-3 by 40 -50% (39,40), suggesting that mammalian cells contain a PKC-dependent pathway that can regulate GSK-3. Indeed, several purified recombinant Ca 2ϩdependent PKCs have been reported to phosphorylate and inhibit GSK-3␤ in vitro (PKC␣ ϭ PKC␤1 ϭ PKC␥ Ͼ PKC␤2) (15), and transgenic mice that express elevated levels of PKC␤2 in the intestinal epithelium also exhibit decreased GSK-3␤ activity (41). The PMA-sensitive novel PKC␦ has also been reported to phosphorylate GSK-3 (16). We found that PMA treatment by itself promoted the phosphorylation of GSK-3␤ in Rat-1 cells, but Ca 2ϩ -dependent PKC isoforms do not appear to be involved in GSK-3␤ phosphorylation induced by the ␣ 1A -adrenergic receptor (Fig. 4, C and D). Instead, our results suggest that activation of a PMA-sensitive novel PKC (Fig. 4C) and the atypical PKC (Fig. 5) by the ␣ 1A -adrenergic receptor may play a role in signaling to GSK-3␤. PKC may act directly on GSK-3␤ in cells, as incubation of GSK-3␤ (but not the S9A mutant) with PKC in vitro has been demonstrated to inhibit GSK-3 kinase activity (42).
Our results indicate that the signaling pathways used by the insulin and ␣ 1A -adrenergic receptors to induce GSK-3␤ phosphorylation are clearly distinct, as the former depends on PI3K and Akt, whereas the latter does not. The pathways used by some other G protein-coupled receptors to regulate GSK-3 are also different from the one described here for the ␣ 1A -adrenergic receptor. For example, agonist binding to the CB 1 cannabinoid receptor activates Akt and increases GSK-3␣ Ser-21 phosphorylation. Treatment of the cells with pertussis toxin or wortmannin blocks the activation of Akt (43), indicating that signaling is through G i /G o and PI3K. Other receptors coupled to G i /G o have also been shown to activate the PI3K/Akt signaling pathway (44,45), and one would expect that they would also induce the wortmannin-sensitive phosphorylation of GSK-3. Treatment of Rat-1 cells with isoproterenol, a ␤-adrenergic receptor agonist, increases the phosphorylation and decreases the activity of both isoforms of GSK-3. Unlike our findings with the ␣ 1A -adrenergic receptor (data not shown), the effect of isoproterenol is reversed in the presence of a PKA inhibitor, and PKA has been proposed to directly phosphorylate and inactivate GSK-3 in response to ␤-adrenergic receptor stimulation (13). By contrast, activation of ␤ 3 -adrenergic receptors in rat fat cells increases Akt activity and decreases GSK-3 activity in a manner that is independent of cAMP and insensitive to wortmannin or LY 294002 (46), suggesting that, in these cells, the ␤ 3 -adrenergic receptor might activate Akt by a novel PI3K-independent pathway that results in the inhibition of GSK-3.
Several GSK-3 signaling pathways that regulate development also utilize heptahelical cell-surface receptors that resemble G protein-coupled receptors. In the Wg/Wnt signaling pathway in Drosophila and vertebrates, the extracellular Wg/ Wnt signal is transduced from the membrane receptor Frizzled through Dsh/Dvl to inactivate GSK-3 (3,40,47,48). Frizzled-1 has been reported to signal through G␣ o and G␣ q (49), and Wg-dependent inhibition of GSK-3 requires a PMA-sensitive isoform of PKC (40). Conflicting results have been obtained as to whether the inhibition of GSK-3 involves phosphorylation of the protein or the activation of Akt (40,47,48,50), but it is well accepted that GSK-3 binding to axin and other proteins is critical for its regulation by the Wg/Wnt signaling pathway (3,4). Dictyostelium uses two heptahelical cAMP receptors (CAR3 and CAR4) that differentially modulate GSK-3 to regulate development (reviewed in Ref. 3). Stimulation of CAR4 inhibits GSK-3 by an unknown mechanism to promote the pre-stalk cell fate, whereas cAMP binding to CAR3 activates GSK-3 to allow the formation of pre-spore cells. In the latter pathway, CAR3 activates the tyrosine kinase ZAK1, which is thought to phosphorylate and activate GSK-3 (31).
Our understanding of the physiological role of GSK-3 has evolved significantly since its identification as an insulin-regulated glycogen synthase kinase. Additional studies to elucidate the signaling pathways used by the ␣ 1A -adrenergic receptor and other G protein-coupled receptors to regulate the function of this enzyme will add to our understanding of the complexity of GSK-3 regulation.