An Intracellular Calcium Signal Activates p70 but Not p90 Ribosomal S6 Kinase in Liver Epithelial Cells*

In the rat liver epithelial cell lines GN4 and WB, angiotensin II (Ang II) activates the G q class of regulatory G-proteins, increasing intracellular calcium, protein kinase C activity, and protein tyrosine phosphorylation. We compared the ability of Ang II and other compounds that increase intracellular calcium ( i.e. the calcium ionophore A23187 and thapsigargin) or protein kinase C activity (the phorbol ester 12- O -tetradecanoylphorbol-13-acetate) to activate p70 ribosomal S6 kinase (p70 S6K ) and p90 ribosomal S6 kinase (p90 RSK ). In GN4 cells, increasing intracellular calcium stimulated p70 S6K activity in a rapamycin- and wortmannin- sensitive manner, but did not affect p90 RSK activity. In contrast, 12- O -tet-radecanoylphorbol-13-acetate strongly activated p90 RSK but only weakly stimulated p70 S6K . The ability of calcium to activate p70 S6K was confirmed by blocking the A23187-dependent activation through chelation of extracellular calcium with EGTA; the effect of thapsigargin was inhibited by the cell permeant chelator bis-( o -aminophenoxy)ethane- N,N,N (cid:42) , N (cid:42) -tetraacetic acid tetraacetoxymethyl ester (BAPTA-AM). Similarly, BAPTA-AM prevented the activation of p70 S6K by Ang II, suggesting

The peptide hormone angiotensin II (Ang II) 1 initiates multiple intracellular processes that in some cell types influence cell proliferation. The seven-transmembrane-spanning Ang II receptor is coupled to the G q class of regulatory G-proteins, which, when activated, stimulates inositol phosphate-mediated calcium release, protein kinase C activity, and additional signals regulated by ␤␥ subunits (1). In rat liver epithelial cells, an additional response to Ang II is a rapid (15-60 s), calciumdependent increase in tyrosine phosphorylation of proteins (2,3). One of the tyrosine-phosphorylated proteins has been purified from Ang II-treated rat liver epithelial cells using ATP affinity chromatography as the penultimate step and shown to be the major calcium-dependent protein tyrosine kinase (CADTK) in these cells (4). Although not directly activated by calcium in vitro, this kinase is stimulated by treatment of cells with agonists and hormones that elevate intracellular calcium (3,4). Peptide sequencing of the purified kinase and cDNA cloning using rat liver epithelial cell mRNA revealed a novel cytoplasmic tyrosine kinase highly related to the p125 focal adhesion kinase (5). The kinase has also been recently identified by three other groups. One group using a polymerase chain reaction strategy to search for novel kinases cloned the human sequence PYK-2 and have also shown its activation in a calcium-and PKC-dependent manner (6). Two other groups screening for kinases related to p125 focal adhesion kinase have reported the sequence as cell adhesion kinase ␤ and related adhesion focal tyrosine kinase (7,8).
In addition to stimulation of calcium and calmodulin, PKC, MAPK, c-Jun N-terminal protein kinase (JNK), and CADTK/ PYK-2 (1,9,10), Ang II has been shown to activate the ribosomal p70 S6 kinase (p70 S6K ) (11). This ubiquitous class of mitogen-activated protein kinases is best known for their ability to phosphorylate the S6 protein of the 40 S ribosome (reviewed in Ref. 12). Although related in sequence to another S6 kinase (p90 RSK ), the regulation of p70 S6K is clearly distinct from the p90 RSK , the latter being activated as part of the Ras-dependent MAPK cascade (13). In contrast to p90 RSK , activation of the p70 S6K pathway is inhibited by rapamycin (14,15). In mammals, the intracellular target for rapamycin is the binding protein FKB12, which forms a complex with the recently identified protein kinases known as FRAP and RAFT (16,17). Similar in sequence homology to the phosphatidylinositol-3 kinases (PI-3 kinases) (18), the FRAP/RAFT kinase is required for activation of p70 S6K (16) but does not appear to phosphorylate phosphatidylinositol, PI-3 kinase, or p70 S6K . Thus the mechanism by which FRAP/RAFT regulates p70 S6K signaling remains to be elucidated.
The p70 S6K can be activated by numerous stimuli, including growth factors, inhibitors of protein synthesis (cycloheximide or anisomycin), and, as noted above, hormones such as Ang II (11,19,20). Complete activation of p70 S6K appears to require multisite phosphorylation by several protein kinases, at least one of which is proline-directed and phosphorylates a sequence(s) in the C-terminal region of the S6 kinase (21)(22)(23)(24). Despite the knowledge of specific phosphorylation sites, the activating "S6 kinase kinases" remain to be identified. In addition to the involvement of the rapamycin-sensitive FRAP/RAFT mentioned above, p70 S6K is inhibited by wortmannin and LY294002, implicating PI-3 kinase in the signaling to p70 S6K (25,26). These results are further supported by studies involving expression of a constitutively active PI-3 kinase (27). A potential intermediary for PI-3 kinase in p70 S6K activation is the serine and threonine kinase PKB, also known as Akt (28,29), which has been shown to activate p70 S6K when co-expressed in cells (30).
The p70 S6K is involved in mitogenesis; studies with neutralizing antibodies to p70 S6K or the inhibitor rapamycin have shown that this kinase is required for G 1 progression (31)(32)(33). The ability of p70 S6K to phosphorylate the ribosomal S6 protein has been well characterized (34), but the contribution of this event to the regulation of protein synthesis is less clear (reviewed in Ref. 12). Phosphorylation of the S6 protein has been found to enhance the translation of mRNAs containing 5Ј-polypyrimidine tracts (35,36), and it is likely that p70 S6K plays a role in growth-related control of translation. For example, translation of specific messages such as those encoding elongation factor 1␣ (36) or insulin-like growth factor II (37) appear to be regulated by the S6 pathway; these may be members of a class of proteins required for progression through the G 1 phase of the cell cycle.
Ang II also stimulates protein synthesis in a number of cell types (11,38), an event that may be dependent on calcium (39) and p70 S6K (11). Because of our interest in calcium and Ang II signaling, we compared the ability of these stimuli to activate the p70 S6K in liver epithelial cells, cells known to have a proliferative response to Ang II (1). In addition, Ang II leads to a protein kinase C-independent increase in AP1 binding, as well as stimulation of MAP kinase and c-Jun N-terminal kinase (10) and a calcium-dependent activation of a novel; calcium-dependent tyrosine kinase (CADTK). In this article we describe the finding that agents and hormones that increase intracellular calcium activate p70 S6K but not p90 RSK . Furthermore, these studies suggest that the initial steps in Ang II action are well correlated with activation of CADTK.
Cell Culturing and Harvesting-Rat liver epithelial cells (GN4 or WB) were grown at 37°C in Richter's minimal essential media containing 0.1 M insulin and supplemented with 10% fetal bovine serum in a humidified 5% CO 2 atmosphere as described earlier (3). Cells were serum starved (0.1% fetal bovine serum) for 20 -24 h prior to agonist stimulation. Cells were washed twice with phosphate-buffered saline (PBS) and once with Buffer H (50 mM ␤-glycerophosphate, pH 7.4, 1.5 mM EGTA, 0.15 mM sodium orthovanadate, 1 mM dithiothreitol, 25 g/ml leupeptin, and 1 mM phenylmethylsulfonyl fluoride), after which they were scraped in 0.5 ml of Buffer H. The cells were then sonicated with two 5-s pulses (Fisher MDL 550 Sonic Dismembrator), and the lysates were centrifuged at 100,000 ϫ g (4°C).
p70 S6K Assays-The activity of p70 S6K in cell lysates was determined as described earlier (43). The 100,000 ϫ g supernatants were assayed for p70 S6K activity by monitoring the phosphorylation of the ribosomal protein S6 (2 g) in a buffer (30 l) containing 25 mM ␤-glycerophosphate (pH 7.4), 1.5 mM EGTA, 0.1 mM Na 3 VO 4 , 1 mM dithiothreitol, 10 mM MgCl 2 , 10 M calmidizolium, 2 M PKI peptide, and 100 M [␥-32 P]ATP (2000 Ci/mmol) for 20 min at 30°C. The reactions were terminated by the addition of SDS-PAGE sample buffer (30 l), heated, and applied to SDS-PAGE (10% acrylamide). The radioactive band corresponding to the ribosomal S6 protein was identified by staining with Coomassie Blue and autoradiography; the radioactive band was excised from the dried gel and quantitated by liquid scintillation counting. Calmidizolium (10 M) was included in the assay mixture to prevent the phosphorylation of the ribosomal S6 protein by calcium-and calmodulin-dependent kinases.
Immunoprecipitation Assays for p70 S6K and p90 RSK -In some experiments, p70 S6K was assayed by immunoprecipitation of the p70 S6K after lysis of the cells in Tris/Triton lysis buffer (20 mM Tris, pH 7.5, 137 mM NaCl, 1% Triton X-100, 10% glycerol, 2 mM EDTA, 0.5 mM phenylmethylsulfonyl fluoride, 5 g/ml leupeptin, 10 nM calyculin A, and 148 M Na 3 VO 4 ). The samples were immunoprecipitated as described below and the immunoprecipitate/protein A-agarose beads were washed once with Tris/Triton lysis buffer and twice with PBS (1 ml). The S6 kinase assays were performed using a modification of the above assay procedure. In these assays, the immunoprecipitate/protein A-agarose beads (20 L) were assayed in 20 mM Hepes, pH 7.3, 10 mM ␤-glycerophosphate, 1.5 mM EGTA, 0.1 mM Na 3 VO 4 , 1 mM dithiothreitol, 10 mM MgCl 2 , and 50 M [␥-32 P]ATP (1000 Ci/mmol) plus 250 M S6 peptide in a final reaction volume of 60 l. The assay was performed at 30°C for 20 min with constant shaking in an Eppendorf 5436 Thermomixer. The reaction was terminated by the addition of 20 l of 100 mM EDTA (pH 7.0); the samples were centrifuged for 5 min at 13,000 ϫ g, and the supernatant (40 l) was spotted on Whatman P-81 paper. The papers were washed in 10% phosphoric acid, and the radioactivity incorporated into S6 peptide was determined by liquid scintillation counting. The assays for p90 RSK were performed exactly as described above, except the cell lysates were immunoprecipitated with a rabbit polyclonal antibody developed against the C terminus of p90 RSK as described previously (Ref. 44; kindly provided by Dr. Edwin G. Krebs, Department of Pharmacology, University of Washington).
p70 S6K Immunoprecipitation and Immunoblotting-For immunodetection of p70 S6K , cells were washed as described above and lysed in RIPA buffer containing 20 mM Tris (pH 7.4), 137 mM NaCl, 10% glycerol, 0.1% SDS, 0.5% deoxycholate, 1.0% Triton X-100, and 2.0 mM EDTA plus 10 nM calyculin A and 0.5 mM phenylmethylsulfonyl fluoride. The lysates were centrifuged at 13,000 ϫ g for 10 min (4°C). The samples were immunoprecipitated using a rabbit polyclonal antibody developed against a C-terminal peptide (QAFPMISKRPEHLRMNL) of the p70 S6K (kindly provided by Dr. J. Weiel, Glaxo-Wellcome). Immunoprecipitation was facilitated by the addition of protein A-agarose (20 l, packed beads), and the immunoprecipitates were washed three times with RIPA buffer prior to addition of SDS-PAGE sample buffer. Samples were applied to SDS-PAGE and blotted to polyvinylidene difluoride, and the immunoprecipitated p70 S6K was detected by incubating blots with the C-terminal p70 S6K antibody (diluted 1:2500 in TBST). The immunoblot was incubated with goat anti-rabbit (alkaline phosphatase-conjugated) antibodies (diluted 1:5000 in TBST), and the color was developed according to the manufacturer's procedure (Promega).
Measurement of CADTK Tyrosine Autophosphorylation-The activity of CADTK was assayed by determining the amount of tyrosine autophosphorylation by a modification of the procedure described earlier (4). Briefly, the cells were washed twice with PBS and harvested in RIPA or Tris/Triton lysis buffer as described above for the S6 kinase immunoblotting experiments. The CADTK was immunoprecipitated with a rabbit polyclonal antibody developed against a glutathione Stransferase-CADTK fusion protein (5). The CADTK immunoprecipitates were washed three times with RIPA or Tris/Triton lysis buffer and applied to SDS-PAGE (8% acrylamide). After electrophoresis the pro-teins were transferred electrophoretically to nitrocellulose, and the amount of phosphotyrosine on CADTK was quantitated by antiphosphotyrosine immunoblotting (PT66) and ECL as described earlier (4).

RESULTS
Calcium Activates p70 S6K but Not p90 RSK -Incubation of GN4 rat liver epithelial cells with compounds that increase intracellular calcium (i.e. the ionophore A23187, thapsigargin (which releases calcium from internal stores; Ref. 45), or Ang II), stimulates intracellular protein tyrosine phosphorylation (2). To investigate the influence of calcium-stimulated tyrosine phosphorylation on "downstream" kinase signaling pathways, we examined the ability of calcium to activate the ribosomal S6 kinase p70 S6K . Incubation of GN4 cells with 1 M A23187 or 2 M thapsigargin increased S6 kinase activity as measured by the phosphorylation of the 40 S ribosomal S6 protein (Fig. 1A). A typical response to these stimuli was a 2-3-fold increase over basal levels, similar to that produced by EGF (See Fig. 3). The calcium-stimulated increase in S6 kinase activity was inhibited by the calcium chelators EGTA and BAPTA-AM. Preincubation with EGTA prevented activation of p70 S6K by A23187; the cell-permeant chelator BAPTA-AM (BAPTA) inhibited both the thapsigargin-and ionophore-stimulated S6 kinase activity (Fig. 1A). These results indicate that for these agonists the increase in p70 S6K activity is calcium-dependent, and with thapsigargin, the effect is primarily dependent on intracellular calcium.
In addition to the increased the phosphorylation of the ribosomal S6 protein, calcium activated p70 S6K as demonstrated by SDS-PAGE and immunoblotting analysis (Fig. 1B). In these experiments, p70 S6K was immunoprecipitated from cell lysates and immunoblotted for p70 S6K as described under "Experimental Procedures." Four immunoreactive bands corresponding to the multiple phosphorylated forms of p70 S6K were detected (46). In all experiments, increased p70 S6K activity (as determined by S6 protein phosphorylation) correlated with an upward mobility shift of both the slowest and fastest migrating immunoreactive bands on SDS-PAGE. The increased upward mobility shift of p70 S6K produced by A23187, or thapsigargin was prevented by preincubation with calcium chelators (Fig.  1B). BAPTA preincubation reduced the mobility shift of p70 S6K from untreated and A23187-or thapsigargin-treated cells, demonstrating that BAPTA affected both the stimulated and basal S6 kinase activity. In contrast, EGTA inhibited only the A23187-stimulated S6 kinase mobility shift, similar to the results obtained by assaying S6 protein phosphorylation. (Fig. 1,  A and B).
Like thapsigargin, Ang II rapidly increases intracellular calcium and tyrosine phosphorylation in rat liver epithelial cells (2). As shown in Fig. 2, incubation of GN4 cells with Ang II rapidly stimulated p70 S6K activity comparable with that observed with thapsigargin. The increase in p70 S6K activity was apparent as early as 2-5 min after Ang II addition, with the peak of S6 phosphorylation occurring after 10 -20 min. The effect of Ang II and thapsigargin was sustained for more than 60 min (data not shown). Similar results were found by analyzing immunoblots of p70 S6K ; Ang II, thapsigargin, and A23187 stimulated a time-dependent increase in the p70 S6K mobility shift on SDS-PAGE, which paralleled the increase in S6 protein phosphorylation (see Fig. 4A).
In cardiac myocytes, Ang II and calcium have been reported to stimulate the activity of MAPK and the ribosomal S6 kinase p90 RSK (47). Although Ang II activated MAPK in GN4 cells by a PKC-dependent mechanism, thapsigargin does not stimulate MAPK in these cells (5,10). We therefore compared the ability of thapsigargin and EGF to activate p70 S6K and p90 RSK in GN4 cells. As expected, both agonists stimulated p70 S6K (Fig. 3A), but only EGF stimulated p90 RSK in GN4 cells (Fig. 3B). The kinase assay on p90 RSK immunocomplexes using the S6 peptide (RRLSSLRA) as a substrate also demonstrated that A23187 did not activate p90 RSK , whereas Ang II potently activated this enzyme, as previously reported (10) (data not shown).
Calcium-dependent p70 S6K Activation Is Inhibited by Wortmannin and Rapamycin-To further rule out the possibility that p90 RSK was contributing to the phosphorylation of the S6 protein in cell lysates and to investigate the mechanism by which calcium stimulates p70 S6K , we examined the effect of compounds known to inhibit p70 S6K activation. Rapamycin has been previously shown to inhibit p70 S6K without affecting p90 RSK activity (14,15). As shown in (Fig. 4A), incubation of cells with 10 nM rapamycin completely inhibited the Ang II-and calcium ionophore (A23187)-stimulated increase in p70 S6K S6 peptide kinase activity. Similarly, rapamycin blocked the thapsigargin-dependent activation of p70 S6K ; (unstimulated, 35.7 Ϯ 0.4 pmol/min/ml; thapsigargin, 94.3 Ϯ 7.1 pmol/min/ml; thapsigargin and rapamycin, 23.9 Ϯ 0.2 pmol/min/ml). Rapamycin also blocked Ang II-, A23187-, and thapsigargin-dependent p70 S6K activation, as assessed by the gel mobility immunoblotting (Fig. 4B). These results demonstrate that Ang II and calcium stimulate p70 S6K through a rapamycin-sensitive pathway and confirm that p70 S6K and not p90 RSK is the major ribosomal S6 kinase activated by calcium stimuli in GN4 cells.
Studies with the inhibitor wortmannin and constitutively active forms of PI-3 kinase have demonstrated a role for this enzyme in the regulation of p70 S6K (25)(26)(27). To determine whether calcium stimulated p70 S6K through a PI-3-kinase-dependent pathway, the effect of wortmannin on p70 S6K activity was examined. As shown in Fig. 4B, incubation of GN4 cells with wortmannin (50 nM) inhibited the mobility shift of p70 S6K stimulated by Ang II, A23187, or thapsigargin. The inhibition FIG. 3. Thapsigargin activates p70 S6K , but not p90 RSK . Serumstarved GN4 cells were incubated with 2 M thapsigargin (Thaps) or 6.6 nM EGF for the lengths of time indicated. The cells were washed two times with PBS and harvested in Tris/Triton lysis buffer as described under "Experimental Procedures." The lysate from each 10-cm dish was immunoprecipitated with antibodies specific for p70 S6K (A) or p90 RSK (B) and assayed for S6 peptide kinase activity as described above. The S6 peptide activity is plotted as the amount of activity (pmol/min/ml), and the results represent the mean Ϯ S.E. of p70 S6K by wortmannin occurred at low concentrations (10 -50 nM), consistent with the effects of this compound on PI-3 kinase (48). Similar inhibitory affects of wortmannin were found on the calcium-dependent activation of S6 kinase, as determined by S6 protein phosphorylation (data not shown). Thus, this calcium-stimulated signaling pathway appears to be mediated through both a PI-3 kinase and FRAP/RAFT-dependent pathway.
Angiotensin II Activates p70 S6K in a Calcium-dependent Manner-Ang II stimulates an inositol phosphate-mediated increase in intracellular calcium (reviewed in Ref. 49), and since our results demonstrated that calcium stimulates p70 S6K activity, we examined whether the activation by Ang II was calcium-dependent. Incubation of GN4 cells with the calcium chelator BAPTA inhibited the Ang II-stimulated S6 kinase activity in a dose-dependent manner, with Ͼ90% inhibition occurring at 50 M BAPTA (Fig. 5A). Like the results found with S6 kinase activity, BAPTA inhibited the Ang II-stimulated p70 S6K mobility shift (Fig. 5B) at concentrations similar to those required to inhibit the activation by A23187 or thapsigargin (Fig. 1, A and B). Chelation of extracellular calcium with EGTA (5 mM) only minimally affected the Ang II-stimulated S6 kinase activity, further supporting the thesis that intracellular calcium was required for the activation of p70 S6K (Fig. 5B). The calcium-dependent activation of p70 S6K occurred independently of calmodulin, since the calmodulin inhibitor calmidizolium (50) did not inhibit this event. Similarly, neither calmidizolium nor W-7 prevented the autophosphorylation of CADTK, although calmidizolium (15 M) inhibited the Ang II-dependent activation of the myosin light chain peptide kinase activity, a known calmodulin-dependent process (data not shown). In comparison, preincubation of GN4 cells with BAPTA did not inhibit but significantly stimulated the activation of p90 RSK by Ang II in a dose-dependent manner (Fig. 5C). Furthermore, incubation with BAPTA alone (50 M) increased the basal level of p90 RSK to levels equivalent to that of Ang II stimulation in these cells. Similar to the results obtained with p90 RSK , BAPTA pretreatment potently enhanced the basal and Ang II-stimulated level of MAPK activity (data not shown).
PKC Activates p90 RSK but Not p70 S6K in GN4 Cells-Since Ang II increases diacylglycerol formation and PKC activity in GN4 cells, we investigated whether the activation of p70 S6K by Ang II was also PKC-dependent. Incubation of cells with phorbol ester (TPA, 100 nM) for 20 min did not stimulate p70 S6K activity in these cells, although p90 RSK was stimulated by TPA (100 nM, 5 min) (Fig. 6, A and B). To further investigate the involvement of PKC, PKC activity was down-regulated by chronic incubation with TPA (1 M, 24 h). The activation of p70 S6K by Ang II was not inhibited by this treatment, indicating that PKC did not play a substantial role in regulating p70 S6K in these cells (Fig. 6A). In comparison, this treatment effectively eliminated the activation of p90 RSK by TPA as expected (Fig. 6B).
The Activation of p70 S6K by Calcium Requires a Tyrosine Kinase-In epithelial cells, the Ang II-and calcium-stimulated increase in protein tyrosine phosphorylation is inhibited by the tyrosine kinase inhibitor genistein (3,51). GN4 cells were briefly incubated with genistein (10 -200 M) to test the ability of this compound to prevent the activation of p70 S6K by Ang II or calcium stimuli. As shown in Fig. 7A, genistein inhibited the Ang II-stimulated p70 S6K activity in a dose-dependent manner, suggesting the involvement of a tyrosine kinase in the calcium and Ang II signaling to p70 S6K . Similarly, genistein inhibited the activation of p70 S6K by thapsigargin (data not shown). Recently, we have purified the major calcium-and Ang IIstimulated tyrosine kinase (CADTK) from GN4 rat liver epi- thelial cells (4), cloned the cDNA, and raised specific antisera to the protein (5). We compared the ability of genistein to inhibit the activity of CADTK by assaying the tyrosine autophosphorylation of CADTK, as described under "Experimental Procedures." The amount of tyrosine autophosphorylation of this kinase appears to be proportional to the enzymatic activity. 2 As shown in Fig. 7B, incubating GN4 cells with concentrations of genistein from 50 to 400 M led to a progressive decrease in the Ang II-stimulated CADTK tyrosine autophosphorylation. Similar concentrations of genistein were required to inhibit the tyrosine autophosphorylation of the EGF receptor in these cells. In comparison, the calcium-dependent tyrosine phosphorylation of an exogenous substrate (e.g. paxillin, a 68 -70-kDa protein) tyrosine-phosphorylated by CADTK was inhibited by even lower concentrations of genistein (100 -200 M) than was the CADTK autophosphorylation. 3 The results with genistein implicate a tyrosine kinase in the activation of p70 S6K by calcium or Ang II, suggesting a role for CADTK.
The Stimulation of p70 S6K Correlates with Increased CADTK Activity-To further investigate whether CADTK was involved in the regulation of p70 S6K , we compared the activation of p70 S6K in cells known to exhibit different levels of Ang II-and thapsigargin-dependent CADTK activation. Previously we showed that transformed GN4 rat liver epithelial cells exhibit approximately 3-4 times more Ang II-stimulated tyrosine phosphorylation than the parental cell type from which it was derived (WB) (3). We have recently confirmed that GN4 cells exhibit approximately 5 times more Ang II-dependent activation and 2-3-fold more CADTK protein than WB cells (5). Treatment with Ang II or the ionophore A23187 resulted in a significantly larger activation of p70 S6K in GN4 cells compared with WB cells (Fig. 8A). Typically, the ability of Ang II or A23187 to activate p70 S6K was at least 2-3-fold higher in GN4 cells than in WB cells and was independent of the length of After incubating cells with Ang II for 20 min, the cells were harvested, and the cell lysates were assayed for S6 kinase activity. The activity is plotted as radioactivity (cpm, 32 P) incorporated into the ribosomal S6 protein and represents the mean Ϯ S.E. (bars) of duplicate samples. Data plotted are representative of n ϭ 3 experiments. Cont, control. B, CADTK, GN4 cells grown in 6-cm plates were serum-starved and incubated with the concentrations of genistein indicated for 15 min. After incubation with genistein, the cells were treated with 1 M Ang II for 90 s, and the cells were washed twice with PBS and harvested in RIPA buffer (1 ml). The amount of tyrosine autophosphorylation on CADTK was determined by immunoprecipitation with CADTK-specific antisera followed by PT66 antiphosphotyrosine immunoblotting as described under "Experimental Procedures." exposure to these compounds. Similarly, thapsigargin stimulated a similar 2-3-fold higher increase in S6 kinase activity in GN4 cells (data not shown).
To eliminate the possibility that the enhanced calcium-dependent activation was simply due to increased expression of the p70 S6K in GN4 cells, the amount of p70 S6K activity and protein was compared in the two cell lines by immunoblotting and activity assays as described earlier. As determined by immunoblotting, the amount of the p70 S6K protein and the nuclear form of this kinase (p85 S6K ) (33) was comparable in the two cell types (Fig. 8B). Importantly, in GN4 cells, Ang II, thapsigargin, and A23187 stimulated a pronounced mobility shift of both the p70 and p85 S6 kinases; by contrast, much less mobility shift was observed in WB cells, in agreement with the S6 kinase activity assays. Furthermore, the protein synthesis inhibitor anisomycin (20) strongly stimulated the mobility shift (Fig. 8B) and activity of p70 S6K (and p85 S6K ) equivalently in WB and GN4 cells (Fig. 8C). These results demonstrate that differences in p70 S6K or p85 S6K expression do not account for the enhanced activation by calcium in GN4 cells. Instead these results suggest that the increased expression of CADTK in GN4 cells facilitates the activation of p70 S6K by Ang II or calcium stimuli. DISCUSSION Ang II has numerous effects on rat liver epithelial cells. It is a weak mitogen, a response that is initiated via the AT 1 receptor, transduced by G q protein stimulation of phospholipase C, and effected by serine, threonine, and tyrosine phosphorylation (1). Elevating intracellular calcium with thapsigargin or ionophore (A23187) mimics the Ang II-stimulated increase in protein tyrosine phosphorylation in GN4 cells (2), and our laboratory has recently isolated a likely candidate for the calciumstimulated entity that regulates tyrosine phosphorylation, a novel CADTK. Incubating GN4 cells with Ang II or agonists that raise intracellular calcium (i.e. thapsigargin and A23187) activates CADTK (4), whereas incubation with the calcium chelator BAPTA-AM inhibits activation of this enzyme. The effect of calcium on CADTK is indirect; adding calcium or calcium and calmodulin to cell lysates does not activate CADTK, and thus the mechanism by which calcium and other signals regulate this enzyme remains to be established.
In addition to activating PKC, the calcium-and calmodulindependent protein kinase, and CADTK in GN4 cells, Ang II also activates MAPK and JNK and increases AP-1 binding (the latter can be accomplished in a PKC-independent manner) (10). We now demonstrate Ang II and calcium-dependent activation of p70 S6K . The challenge is to discern which of these multiple Ang II-dependent pathways are downstream of PKC, calcium and calmodulin, CADTK, or even G-protein ␤␥ subunits.
We began this process by purifying CADTK; peptide and cDNA cloning has identified this enzyme as the rat homologue of a novel human nonreceptor tyrosine kinase, PYK-2 (4 -6). Using PC12 cells, PYK-2 was shown to be activated by elevating intracellular calcium or PKC activity. In these cells, PYK-2 stimulation increased MAPK activity, providing a potential mechanism for calcium-dependent regulation of MAPK in these and other cells (6,52). In contrast, our studies in rat epithelial cells demonstrate that a calcium signal (i.e. thapsigargin and A23187) does not significantly activate MAPK and that Ang II-dependent MAPK activation is primarily a PKC-dependent process. 2 Instead, Ang II and a calcium signal (thapsigargin) substantially activate JNK in a calcium-dependent manner (10). In fact, the Ang II effect is PKC-independent and is amplified in cells depleted of PKC. JNK activation by Ang II and thapsigargin correlates with increased expression and tyrosine autophosphorylation of CADTK in our rat liver epithelial cells. We have shown that the chemically transformed GN4 cell line expresses more CADTK than the parental cell line (WB), and the activation of CADTK by Ang II is approximately 5-fold greater in confluent GN4 cells when compared with WB cells. Activation of JNK by Ang II and thapsigargin is also 5-fold higher in GN4 cells, whereas the total JNK activity induced by stress (e.g. anisomycin) is equal in WB and GN4 cells. Thus in GN4 cells, CADTK is not involved in MAPK activation but is likely to be a mediator of signaling to JNK. Because some of the stimuli known to activate JNK also increase p70 S6K activity in other cell types, (i.e. Ang II (11) and anisomycin (20)), we compared the ability of a calcium signal to regulate p70 S6K in epithelial cells. In this study we find that like JNK (10), p70 S6K is regulated by agonists that raise intracellular calcium, whereas neither p90 RSK nor MAPK is affected by the calcium-activated pathway. This is contrasted to results in PC12 cells (6,52) and suggests that the ability of calcium to regulate these signaling pathways is cell type-specific.
The role of intracellular calcium and Ang II (through its effects on intracellular calcium) in stimulating p70 S6K has not been studied extensively. In one article, A23187 weakly activated p70 S6K in Swiss 3T3 fibroblasts (53), a cell line that does not appear to contain CADTK. 2 Earlier, the work of Meier et al. (54) reported the activation of an S6 kinase in kidney epithelial cells in response to calcium stimuli (A23187), although the identity of this kinase was not established. It is clear that in this study we were measuring the calcium-dependent activation of p70 S6K in cell lysates. First, although the cell lysate assay could detect p90 RSK phosphorylation of the 40 S ribosomes, we have used a specific immune complex assay to show that there is little if any thapsigargin-dependent p90 RSK activation. Second, all our assay data have been confirmed by immunoblot analysis of the p70 S6K mobility shift, an assay that highly correlates with the activation. Third, rapamycin, which inhibits p70 S6K but not p90 RSK activation, abolished the Ang IIor thapsigargin-dependent activation of 40 S ribosome phosphorylation in cell lysates.
Importantly, our results suggest that most of the Ang II-dependent activation of p70 S6K is calcium-dependent. Preincubation with BAPTA prevented the Ang II-dependent activation of p70 S6K but not p90 RSK . Instead, p90 RSK and MAPK were substantially activated by BAPTA, alone or in the presence of Ang II. At this point, the mechanism of the BAPTA effect is unclear; we can only speculate that calcium may regulate one or more phosphatases in the MAPK cascade or that BAPTA has a non-calcium-dependent effect on the MAPK pathway. We also cannot exclude that additional G-protein-coupled signals, e.g. ␤␥ subunit or other G-protein-dependent responses, contribute to the stimulation of p70 S6K by Ang II. Recently the activation of p70 S6K by receptors coupled to G i has been described, supporting a potential role for ␤␥ subunits in the regulation of p70 S6K (55).
In this study we also found that agonists that increase intracellular calcium (A23187, thapsigargin, and Ang II) in GN4 cells activate p70 S6K following a rapid stimulation of CADTK. The cell-permeable chelator BAPTA-AM inhibited the activation of both CADTK and p70 S6K by thapsigargin, A23187, or Ang II, consistent with a role for CADTK upstream of p70 S6K in these cells. In further support of a role for CADTK, the calciumand Ang II-dependent activation of p70 S6K was prevented by genistein at concentrations similar to those required to inhibit the thapsigargin-dependent JNK activation and the thapsigargin-and Ang II-dependent tyrosine phosphorylation of CADTK substrates (e.g. paxillin) in vivo. 2 Other findings linking CADTK to the p70 S6K pathway come from a comparison of p70 S6K activity in WB and GN4 cells. The GN4 cell line expresses 2-3-fold greater CADTK and has 5-fold greater Ang-II-dependent CADTK and JNK activation (5). GN4 cells also exhibit 2-3-fold greater p70 S6K activation in response to Ang II and thapsigargin. The failure of p70 S6K to match the 5-fold difference in CADTK and JNK activation between GN4 and WB cells may well stem from the complex pathway to p70 S6K , which appears to involve regulation of at least two upstream kinase pathways (see below).
One possibility is that a calcium signal may be activating PI-3 kinase, since low concentrations of wortmannin (50 nM) inhibited the activation of p70 S6K by A23187, thapsigargin, or Ang II. Numerous studies have suggested that PI-3 kinase is an upstream mediator of p70 S6K activity (reviewed in Ref. 12), although much of the evidence for PI-3 kinase (including our study) has been obtained with the inhibitor wortmannin (48), the selectivity of which has been recently questioned (56). With the caveat of specificity, the activation of CADTK by Ang II or thapsigargin is not inhibited by wortmannin, 2 suggesting that PI-3 kinase is downstream or unrelated to the calcium-dependent activation of CADTK. Wortmannin also does not inhibit JNK activation, another pathway linked to the regulation of CADTK in GN4 cells. 2 Thus, if CADTK is a common activator of both p70 S6K and JNK in GN4 cells, as our results suggest, at some point the signaling pathways diverge such that a wortmannin-sensitive step is specific to the p70 S6K pathway.
Our results are also consistent with the calcium-dependent activation of another limb of the p70 S6K pathway, the one that activates the rapamycin-sensitive kinase, i.e. FRAP/RAFT. Rapamycin inhibits the activity of the FRAP/RAFT kinases by binding to its cognate binding protein, FKB12, thus inhibiting p70 S6K stimulation (16). However, like the PI-3 kinases, the role of FRAP/RAFT in regulating p70 S6K is poorly understood, largely because substrates for these kinases remain to be identified. Again the calcium-dependent pathway to JNK and p70 S6K is divergent in GN4 cells, since rapamycin blocks Ang II-and thapsigargin-dependent p70 S6K activation but does not effect the stimulation of CADTK or JNK by these agonists. 2 In addition to increasing intracellular calcium, Ang II also stimulates PKC activity (1,5). Despite the fact that PKC has been shown to activate p70 S6K in other cell types (57), our results do not support a role for PKC in the regulation of p70 S6K in GN4 cells. Although Ang II stimulated p70 S6K in a manner similar to thapsigargin or A23187, TPA alone did not increase p70 S6K activity in GN4 cells, although MAPK and p90 RSK were activated as expected.
TPA also increases CADTK activity in these cells, albeit less potently than Ang II and more slowly than thapsigargin. These results present a paradox, since our studies suggest that CADTK is upstream of p70 S6K , and yet TPA does not activate p70 S6K . We have found that a brief (5 min) TPA pretreatment inhibited the activation of p70 S6K by Ang II or thapsigargin by approximately 40% in GN4 cells. 4 Similarly, an even more profound inhibitory effect of PKC has been found on the activation of JNK by calcium stimuli or Ang II in GN4 cells. 2 Taken together, these studies demonstrate that PKC alone is incapable of activating p70 S6K and may inhibit the downstream calciumdependent signals to both JNK and p70 S6K in GN4 cells.
In summary, we have demonstrated a calcium-dependent pathway that results in substantial activation of p70 S6K . This calcium-dependent pathway has characteristics similar to to the calcium-dependent pathway to JNK that we have previously studied in this cell line (5,10). Intriguingly, this calciumdependent pathway does not activate MAPK or p90 RSK ; instead these enzymes are activated in PKC-dependent manner. Considerable circumstantial evidence points to CADTK as an initial mediator of the calcium signal to both JNK and p70 S6K . However, the pathway to these enzymes clearly diverges at some point such that the calcium-dependent activation of JNK is insensitive to wortmannin and rapamycin, whereas the activation of p70 S6K is completely inhibited by these compounds. The exact signaling downstream from CADTK and, in fact, unequivocal proof of CADTK involvement must await further studies.