Angiotensin II Activates at Least Two Tyrosine Kinases in Rat Liver Epithelial Cells SEPARATION OF THE MAJOR CALCIUM-REGULATED TYROSINE KINASE FROM p125 FAK *

,

In analyzing EGF 1 -dependent tyrosine phosphorylation in rat liver epithelial cell, two waves of Tyr(P) substrate phosphorylation occurring at 5 and 60 s were noted (1). The latter group appeared to be phosphorylated in part due to activation of a second process (e.g. a downstream tyrosine kinase) (2). Subse-quently, we identified several of the same substrates in Ang II, vasopressin, or epinephrine-treated cells, and determined that these G-protein-coupled receptors stimulated tyrosine phosphorylation in a calcium-dependent, protein kinase C-independent manner (2)(3)(4). Since virtually all tyrosine kinases autophosphorylate on tyrosine residues, immune complex tyrosine kinase activity could be assessed in anti-Tyr(P) immunoprecipitates from control and Ang II-treated cells. Ang II increased tyrosine kinase activity within 15-30 s; maximal activation was seen within 1 to 2 min. Activation was abrogated by intracellular chelators that blunt the Ang II-induced calcium signal (3). The mechanism by which calcium increases tyrosine kinase activity remains unclear, but the schema is presumably indirect because tyrosine kinase activity can not be stimulated by adding calcium to cell-free extracts of these cells. Thus, our previous work showed that G-protein-coupled receptors generating a calcium signal rapidly, but indirectly, activate one or more tyrosine kinase(s) (2)(3)(4).
Many of the hormones noted above that activate tyrosine phosphorylation also stimulate cell proliferation and the immediate-early gene expression that accompanies growth factor action. The latter is likely to occur by activation of the MAP kinase (ERK 1 and ERK 2) pathway (33,34), or the newly characterized c-Jun N-terminal kinase (JNK) pathway (35). In fact, both G q -coupled receptors (in some instances via a calcium-dependent mechanism) and G i -coupled receptors have been shown to activate MAP kinase (15, 16, 20, 36 -39). In addition, our laboratory has shown recently that Ang II treatment of rat liver epithelial cells produces a 200-fold activation of JNK through a calcium-and tyrosine kinase-dependent process. 2 In summary, data from several cell types suggest that the selected G q -and G i -coupled receptors as well as other agonists (e.g. integrin stimulation) can activate p125 FAK , MAPK, and JNK. The potential stimulation of several pathways thought to be controlled by tyrosine phosphorylation led us to ask whether Ang II stimulated p125 FAK or another tyrosine kinase(s) or both.
We report that Ang II stimulates at least two cytosolic tyrosine kinases. The first, pp125 FAK , is tyrosine-phosphorylated and activated in these cells in an Ang II-dependent manner; however, greater than 80% of the tyrosine kinase activity resides in another molecule that can be distinguished from p125 FAK . Molecular means were used to identify 20 tyrosine kinases in the cells, but none appear to be the calcium-dependent cytosolic tyrosine kinase. In contrast, we have partially purified an autophosphorylating kinase from Ang II-treated cells by sequential affinity chromatography. This activity migrates as a 115-120-kDa Tyr(P) protein that can be separated electrophoretically and distinguished from pp125 FAK . The possibility is raised that distinct Ang II-dependent intracellular signaling pathways are activated through separate tyrosine kinases.

EXPERIMENTAL PROCEDURES
Materials-EGF was purified from mouse salivary glands as described previously (40). Human Ang II (Asp-Arg-Val-Tyr-Ile-His-Pro-Phe) was obtained from Sigma. Ang II was prepared in 50 mM acetic acid. EGF was prepared as 100-fold-concentrated solutions in 10 mM NaP i (pH 7.4), 150 mM NaCl containing 0.1% bovine serum albumin. Cytochalasin D (Calbiochem) was prepared as stock solutions in dimethyl sulfoxide (final concentration Յ 0.1%, v/v). Anti-p125 FAK , monoclonal antibody 2A7, and polyclonal antibody BC-2 were generously provided by Tom Parson, University of Virginia. Antibodies to JAK1, JAK2, and JAK3 were generously provided by Dr. James Ihle, St. Jude's. Monoclonal anti-Tyr(P) antibody, PT66, and PT66-agarose were purchased from Sigma.
Cell Cultures-WB and GN4 cell were maintained at 37°C in Richter's improved minimal essential medium containing 10% fetal bovine serum and 0.1 M insulin in a humidified 5% CO 2 atmosphere as described previously (3). Seven to 10 days before each experiment, WB cells of passages 19 -26 or GN4 cell passages 5-15 were seeded onto plastic culture dishes (Costar) and used when they reached confluence.
Anti-Tyr(P) Immunoblotting-Cell treatments and anti-Tyr(P) immunoblotting were performed essentially as described previously (2). Cell lysates of WB or GN4 cells grown in 35-mm diameter culture dishes, subcellular fractions, PT66 immunoprecipitates, or affinity resin fractions were probed in a similar fashion. Briefly, in some experiments treatment incubations were terminated by rapid aspiration of the medium and addition of 250 l of ice-cold RIPA buffer. Cell extracts were scraped from culture dishes and solubilized by addition of 125 l of 3-fold-concentrated sodium dodecyl sulfate (SDS)-polyacrylamide gel electophoresis (PAGE) sample buffer and heating to 100°C for 3 min. For experiments in which homogenate, cytosol, and membrane fractions were separated, four 60-mm plates preincubated with 200 M vanadate for 15 min and were then treated with vehicle or Ang II for 1 min. Cells were scraped into vanadate-containing immunoprecipitation buffer (see below) and homogenized in a 2-ml Teflon glass homogenizer with motor driven pestle on ice (30 strokes). One-ml aliquots were ultracentrifuged in a Beckman TL100 centrifuge at 105,000 ϫ g for 45 min. Proteins from cell lysates, subcellular fraction, or affinity procedures were separated by electrophoresis on 7, 8, or 10% polyacrylamide gels and transferred to nitrocellulose membranes. Blots were probed by sequential incubation with anti-Tyr(P) antibodies and 125 I-protein A (2-10 Ci/g; DuPont NEN) or using ECL reagent. Molecular weights were estimated by using prestained standards (Sigma).
Anti-Tyr(P), p125 FAK , JAK, and EGF Receptor Immunoprecipitation-Confluent cultures of WB or GN4 cells in 60-mm dishes, subcellular fractions or affinity column fractions, were immunoprecipitated in ice-cold lysis buffer, referred to as NLB (20 mM HEPES (pH 7.3) con-taining 500 mM NaCl, 50 mM NaF, 5 mM EDTA, 1 mM Na 3 VO 4 , 1% Triton X-100, 10% (v/v) glycerol, and 20 g/ml phenylmethylsulfonyl fluoride, 10 g/ml leupeptin, and 100 kallikrein inhibitor units of aprotinin per ml). Lysates were clarified prior to immunoprecipitation by centrifugation at 14,000 ϫ g for 10 min at 4°C. Tyrosine-phosphorylated proteins were precipitated by incubation with monoclonal anti-Tyr(P) antibody PT66, 5 g of rabbit anti-mouse IgG, and 10 l of Pansorbin (Calbiochem), or 25 l of protein A/G-agarose (Santa Cruz), for 2 h at 4°C. Immune complexes were collected by centrifugation at 14,000 ϫ g for 3 min at 4°C, were washed twice with this lysis buffer, and if used for immune complex tyrosine kinase activity, the precipitates were washed twice with 2-fold concentrated tyrosine kinase assay buffer (100 mM sodium HEPES (pH 7.6), 60 mM MgCl 2 , 2 mM MnCl 2 , 0.2 mM Na 3 VO 4 , 0.2% Nonidet P-40), and then were resuspended with 70 l of this buffer. In some experiments, an aliquot (10 -20 l) of the washed immune complex was removed for anti-Tyr(P) immunoblotting after SDS-PAGE on 8% gels as described. Efficient anti-Tyr(P) recovery of tyrosine phosphoproteins and kinase activity required high NaCl concentration (500 mM) in the cell lysis buffer. We postulate that Tyr(P) residues are dissociated from endogenous binding domains (e.g. Src homology domains) at high ionic strength and thereby are made accessible for immunoprecipitation. In cell lysates or affinity fractions, p125 FAK was precipitated using 2-10 l of monoclonal 2A7; EGF receptor was precipitated with anti-rat polyclonal EGF receptor antibody 1382 prepared in this laboratory. JAKs 1, 2, and 3 were immunoprecipitated after lysis in NLB with specific antisera provided by James Ihle.
Immune Complex Tyrosine Kinase Assay and In Vitro Autophosphorylation Reactions-Routinely, 10 -20-l aliquots of immune complex suspension were preincubated for 5 min at 4°C with 160 g of the synthetic tyrosine kinase substrate poly(Glu 80 -Tyr 20 ) or the control substrate poly(Glu) (Sigma). Reactions (80 l total reaction volume) were initiated by addition of 5 M [␥-32 P]ATP (2-10 Ci/reaction) for cell lysate experiments. Affinity fractions were assayed with 60 M ATP, 5-20 Ci of [␥-32 P]ATP to equalize the ATP concentration in various fractions. After 4 -10 min at 25°C, 50 l of the reaction mix was spotted onto Whatman No. 3MM paper. The papers were washed with trichloroacetic acid, air-dried, and assayed by liquid scintillation for acid-insoluble 32 P. Tyrosine kinase activity was defined as 32 P incorporation in counts/min occurring in the presence of poly(Glu 80 -Tyr 20 ) minus that occurring in the presence of poly(Glu 80 -Tyr 20 ). This defined phosphorylation above the background of endogenous protein phosphorylation. The rates of tyrosine phosphorylation so measured were linear for at least 15 min and were proportional to the amount of cell lysate used for immunoprecipitation. In experiments in which endogenous substrate autophosphorylation was assessed, the poly(Glu 80 -Tyr 20 ) was not added to the reaction mixture. The reactions were terminated at 4 -10 min by boiling in SDS-sample buffer. Samples were subjected to 7-8% SDS-PAGE; gels were stained, destained, dried, and subjected to autoradiography on Kodak AR film. In some instances gels were transferred to Immobilon and subjected to two dimensional phosphoamino acid analysis as described previously (2).
Degenerate PCR Cloning of Tyrosine Kinase Domains-Unknown tyrosine kinase domains were amplified with degenerate oligonucleotides designated TK forward (cca gtt ctc gag cat cgn gat ttn gcn gcn cg) and TK reverse (ctg can acc tgg atg ccn ta tag cta ttg acc), encoding the conserved sequences HRDLAAR and DVWSFGV, respectively, as described (41). The first strand cDNA template was prepared from poly(A) mRNA made by cesium chloride isolation of total RNA from the WB cell line followed by oligo(dT) selection. The cDNA was synthesized using Moloney murine leukemia virus reverse transcriptase in RT buffer (50 mM Tris-HCl (pH 8.3), 75 mM KCl, 10 mM dithiothreitol, and 3 mM MgCl 2 ), 15 units of RNA guard (Pharmacia Biotech Inc.), 10 pmol of random hexamers, and 1.25 mM each dNTP in a 20-l reaction. The volume of the cDNA was adjusted to 25-60 l with TE following first strand synthesis. PCR amplifications were carried out with 5 l of random-primed first strand cDNA in a 50-l reaction volume of PCR buffer (10 mM Tris-HCl (pH 9.0), 50 mM KCl, 1.5 mM MgCl 2 , 0.01% gelatin, and 0.1% Triton X-100), 0.2 mM of each dNTP, 50 pmol of each primer, and 2.5 units of Taq DNA polymerase (Life Technologies, Inc.) under the following conditions: 5 min at 94°C, 1 min at 55°C, and 2 min at 72°C (1 cycle), 1 min at 94°C, 1 min at 55°C, and 2 min at 72°C (22 cycles), and 1 min at 94°C, 1 min at 55°C, and 10 min at 72°C (1 cycle). The PCR product was then gel-purified, cut with XhoI and ClaI at sites included in the PCR primers, and subcloned into pBSII SKϩ (Stratagene). In later experiments, the PCR product was TA-cloned (Invitrogen) as per manufacturer's instructions. The 190 cloned PCR products were sequenced by dideoxy termination sequencing (Seque-2 I. Zohn, H. Yo, X. Li, A. Cox, and H. S. Earp, unpublished results. nase 2.0, U. S. Biochemical Corp.), grouped into families (40 of 190 were Abl), and the sequences compared to the GenBank data base with the Blast algorithm.
Expression Cloning of WB Cell Tyrosine Kinases-A cDNA library was constructed in gt11 using oligo(dT)-primed poly(A)-containing RNA purified from confluent cultures of WB cells. Recombinants were screened for cDNAs encoding active tyrosine kinases by a modification of the method of Letwin et al. (42). A total of 480,000 plaque-forming units from the unamplified phage library was used to infect Escherichia coli strain Y1090. Infected cells were plated in LB top agarose on 10 150-mm LB agar plates and incubated for 3 h at 42°C. When plaques had become visible, expression of insert sequences was induced by applying nitrocellulose filters, prewet with 10 mM IPTG, to the surface of the agarose and continuing incubation at 37°C. After 4 h, filters were rinsed briefly with 50 mM Tris-Cl, 150 mM NaCl (pH 7.5), plus 0.05% sodium azide ("RB") and blocked overnight in RB containing 3% bovine serum albumin (Boehringer Mannheim). Plaques containing proteins phosphorylated on tyrosine were identified by immunoblotting the filters with purified polyclonal anti-phosphotyrosine antibodies followed by 125 I-protein A as described. Twenty positive plaques were detected. Phage from the 12 plaques giving rise to the strongest isopropyl-1-thio-␤-D-galactopyranoside induction and anti-phosphotyrosine immunoblotting, yielding 8 purified clones with Lambdasorb reagent (Promega). Insert DNAs were excised by restriction and endonuclease digestion with EcoRI and ligated into EcoRI-cut pBluescript II(KSϩ) (Stratagene). Partial nucleotide sequences of insert cDNAs, obtained using the Sequenase 2.0 kit (U. S. Biochemical Corp.), were compared to known sequences using the GCG analysis package.
Affinity Purification-In a typical preparation, 50 150-mm tissue culture plates of confluent GN4 cells were treated with 1 M Ang II for 1 min and scraped into 1.0 ml of NLB/plate. After centrifugation, the lysate was tumbled with 1 ml of PT66-agarose, washed five times, and eluted with 10 ml of NLB with 10 mM phenylphosphate. The Tyr(P) elute was brought to 50 mM MgCl 2 in 150 mM NaCl and incubated with 0.5 ml of ATP-Sepharose synthesized by one of us (L. Graves). Briefly, 4-aminoanilido-ATP was synthesized by reacting ATP with N-ethyl-NЈ-(3-dimethyllaminopropyl)carbodiimide and phenylenediamine. The product, isolated by high performance liquid chromatography, was analyzed by mass spectrometry to confirm the identity. The 4-aminoanilido-ATP derivative was coupled by carbodiimide chemistry to Sepharose resin containing a carboxyl group attached to a 6-carbon linker (Pharmacia ECH Sepharose). After transfer to a Bio-Rad 1-ml column, the resin was washed with 150 mM and 500 mM NaCl containing buffers and the purified ATP-binding proteins were eluted with 1 mM ATP in six to eight 1-ml fractions. Fractions were combined and used for immunoprecipitation (2 h at 0°C, as described) with p125 FAK or EGF receptor antibodies or were concentrated by incubation with PT66agarose overnight at 0°C. Some aliquots were autophosphorylated in immune complex kinase assays, as described, and subjected to gel electrophoresis and autoradiography. Alternatively, aliquots were prepared for mini-gel electrophoresis, followed by silver staining, or after transfer to nitrocellulose, for Tyr(P) immunoblotting. For some gel electrophoresis analyses, the bisacrylamide concentration was reduced from 1 to 0.5% to increase the separation in the 120-kDa region.

RESULTS
The Ang II-dependent Tyr(P) Substrates and Tyrosine Kinase(s) Are Cytosolic Proteins-The subcellular localization of Ang II-dependent Tyr(P) substrates and tyrosine kinase activity was investigated in WB and rat liver epithelial cells. Previous experiments had assessed the time course of substrate phosphorylation and kinase activity in extracts of cells lysed in detergent-containing buffers with 500 mM NaCl, protease, and phosphatase inhibitors (2). To determine subcellular localization, cells preincubated with 200 M vanadate to preinhibit cellular Tyr(P) phosphatases were incubated with Ang II for 1 min, the time of maximal tyrosine kinase activation. Preincubation with vanadate prolonged the Ang II-dependent tyrosine phosphorylation response, but did not detectably alter the substrates phosphorylated. Cells were scraped, homogenized on ice, and centrifugated at 105,000 ϫ g for 45 min. After resuspension of the pelleted membranes, Triton X-100 was added to a portion of the original homogenate (kept on ice during subcellular fractionation), as well as to the cytosol and membrane fractions. Following electrophoresis of the samples and transfer to nitrocellulose, an anti-Tyr(P) immunoblot was performed (Fig. 1). The majority of the Tyr(P) substrates in Ang II-treated WB cells were localized to the cytosolic fraction. In contrast, EGF-stimulated tyrosine phosphorylation of substrates both in the cytosolic and membrane fractions; the major substrate in the membrane fraction was the 170-kDa EGF receptor.
To determine the localization of the tyrosine kinase(s) activated in Ang II-treated cells, the experiment was repeated using GN4 cells, a chemically transformed line derived from WB, that expresses ϳ3-fold more Ang II-stimulated tyrosine kinase activity. After separating homogenate, cytosol, and membrane fractions, detergent was added, followed by immunoprecipitation with anti-phosphotyrosine monoclonal antibody PT66. Immune complex tyrosine kinase activity was performed as described under "Experimental Procedures" assessing the transfer of 32 P from [␥-32 P]ATP to the exogenous substrate poly(Glu 80 -Tyr 20 ). In immunoprecipitates from the homogenate fraction, Ang II treatment increased tyrosine kinase activity by ϳ2.5-fold (2,637 cpm for control to 6,199 cpm for Ang II-treated). At least 75-80% of this activity was found in the immunoprecipitates of cytosolic Tyr(P) proteins (740 cpm for control, 4,596 cpm for Ang II-treated). In contrast, little activity was found in the immunoprecipitates from the membrane fraction (699 cpm for control, 313 cpm for Ang II-treated). Thus, the majority of the Tyr(P) substrates and virtually all of the increased tyrosine kinase activity were found in the cytosolic fraction.
p125 FAK Is Activated in Ang II-treated Cells, but Is Not the Major Ang II-dependent Tyrosine Kinase-The fact that the Ang II-dependent tyrosine kinase activity was soluble and that there were Tyr(P) substrates in the p115-130-kDa region led us to test whether p125 FAK was the major kinase activated. GN4 cells treated with vehicle or Ang II for 1 min were subjected to single or sequential immunoprecipitation protocols using control mouse Ig (mIg), the anti-Tyr(P) monoclonal antibody PT66, or the FAK monoclonal antibody 2A7. Immunoprecipitates from treated cells were subjected to electrophoresis, transfer to nitrocellulose, and Tyr(P) immunoblotting with PT66 as described under "Experimental Procedures" (Fig. 2). Immunoprecipitation with irrelevant mIg did not precipitate Tyr(P)-containing proteins from either control or Ang II- Procedures." Following ultracentrifugation at 105,000 ϫ g for 45 min, the particulate fraction was resuspended in an equivalent volume of buffer and each fraction (homogenate, cytosol, and membrane) was boiled in SDS sample buffer. Following SDS-PAGE and transfer to nitrocellulose a phosphotyrosine immunoblot was performed, as under "Experimental Procedures." Ang II-dependent Tyr(P) substrates were found predominantly in the cytosolic fraction. treated cells. Immunoprecipitation of p125 FAK (Fig. 2, lanes 3  and 6) demonstrated tyrosine phosphorylated p125 FAK in control GN4 cells; Ang II slightly increased p125 FAK Tyr(P) content at 1 min, the time of maximal Ang II-dependent tyrosine phosphorylation. However, FAK was not the major component in Ang II-treated cells. Immunoprecipitation with anti-Tyr(P) PT66 showed the previously described large increase in Tyr(P) substrates after Ang II treatment (Fig. 2, lanes 2 and 5). To determine more precisely the portion of Tyr(P) phosphoprotein in the p120 -125 region that was pp125 FAK , a sequential immunoprecipitation was performed. The first immunoprecipitation with either irrelevant mIg or 2A7 anti-FAK was performed at 0°C in vanadate-containing buffers. The supernatants remaining from the first immunoprecipitation were then reimmunoprecipitated with PT66. Preclearance of nearly all immunoreactive FAK by 2A7 had little effect on the amount of PT66 immunoprecipitable Tyr(P) substrate from Ang II-treated cells (Fig. 2, lanes 10 and 11).
While the above indicated that p125 FAK is only a small fraction of the Tyr(P) substrate, it did not determine whether p125 FAK was the major autophosphorylating kinase in immune complexes. Tyrosine kinase activity assays were performed on 2A7 (p125 FAK monoclonal) and PT66 (anti-Tyr(P) antibody) immunoprecipitates. In five experiments, the immune complex p125 FAK activity measured by the ability of p125 FAK immunoprecipitates to phosphorylate poly(Glu 80 -Tyr 20 ) nearly doubled when isolated from cells after Ang II treatment (1 min), from an average of 6,747 cpm for control to 11,730 cpm for Ang II. The immune complex tyrosine kinase activity of the PT66 immunoprecipitate paired with a representative p125 FAK experiment rose from a control of 15,656 cpm to 66,342 cpm after Ang II treatment (1 min). Autophosphorylation activity was also assessed by repeating the single and sequential immunoprecipitation protocol in control or Ang II-treated GN4 cells and assessing incorporation from [␥-32 P]ATP into precipitated protein in vitro, followed by gel electrophoresis and autoradiography of dried gels as described under "Experimental Procedures." Fig. 3 demonstrates that p125 FAK autokinase activity in 2A7 immunoprecipitates is slightly elevated (lanes 3 and 6) in Ang II-treated cells (1 min), corresponding to the increase in immune complex poly(Glu 80 -Tyr 20 ) activity noted above. Ang II-dependent PT66 immune complex auto kinase activity was much greater (Fig. 3, lanes 2 and 5). In the sequential protocol, immunoprecipitation removing nearly all immunoreactive p125 FAK from the supernatant did little to diminish the subsequent PT66-immunoprecipitable autokinase activity seen in the 115-125-kDa region (Fig. 3, lanes 10 and 11). Examination of lanes 5, 10, and 11 in Fig. 3 revealed a diffuse band above 120 kDa of phosphorylated protein and one sharp band of in vitro phosphorylated protein at ϳ115-120 kDa. On shorter exposures of the autoradiograph, this single band is even more prominent. This is observed either before (lane 10) or after (lane 11) the removal of p125 FAK by immunoprecipitation, suggesting that a major autophosphorylation protein was not removed by p125 FAK antibody. 32 P-Phosphoamino acid analysis by two dimensional thin layer chromatography has shown that all [ 32 P]phosphate added in vitro was on tyrosine residues (data not shown). This analysis suggested that an autophosphorylating kinase at 115-120 kDa may be the most prominent in vitro labeled protein, a supposition substantiated by the purification discussed below.
In studies of p125 FAK activation via G-protein-coupled receptors, others tested whether the cytoskeletal was involved in the signal transduction pathway (19,24). Preincubation of cells with cytochalasin D, an agent that disrupts actin microfilaments and cytoskeletal movement, prevented the tyrosine phosphorylation by G-protein-coupled receptor agonists (e.g. bombesin) in Swiss 3T3 cells (19,24). GN4 cells were preincubated with 2 M cytochalasin D for 2 h prior to stimulation with Ang II (1 min). Fig. 4 demonstrates that cytochalasin D had little or no effect on Ang II-dependent tyrosine phosphorylation of substrates in the 115-130-kDa region. We did not assess p125 FAK tyrosine phosphorylation in this experiment, but pretreatment did, however, distinctly inhibit tyrosine phosphorylation of 65-75-kDa substrates in Ang II-treated GN4 cells. In other cell types, Tyr(P) substrates in this region have proved to be the cytoskeletal protein paxillin, which is tyrosine-phosphorylated in cells stimulated by hormones binding to G-protein-coupled receptors (24). In conclusion, Ang II stimulates rapid tyrosine phosphorylation of two groups of substrates: one sensitive and one insensitive to cytochalasin D. This may occur because Ang II activates at least two tyrosine kinase pathways.
Attempts to Identify an Ang II-regulated Tyrosine Kinase Other than p125 FAK -Molecular technologies and available antibodies were used to identify the range of tyrosine kinases present in WB rat liver epithelial cells. First, a cDNA library that had been constructed using WB cell poly(A) mRNA was used for cDNA expression cloning to identify novel tyrosine kinases that might not have canonical kinase domain sequences but would be detected if expressed in E. coli, which have neither tyrosine kinases nor tyrosine phosphatases. Anti-Tyr(P) cDNA expression cloning identified three tyrosine kinases (Fer, Fyn, Bek). Failing to identify a novel cytosolic tyrosine kinase, degenerative PCR cloning using the tyrosine kinase domain primers was performed (41,43). A total of 20 tyrosine kinases were identified in WB cells, including the EGF receptor and Src (identified using antibodies), the three kinases noted above detected by expression cloning, and 15 more identified by PCR. The latter group included three that were novel when detected (JAK3 (44), Tyro 3 (45), and an Elk-like kinase). Ten of the kinases were membrane-bound receptors (EGF receptor, Bek, Axl, Tyro 3, Tyro 10, Eck, Elk-like, Flk, Flt, and IGF-I receptor), and two were Src family members (Fyn and Src); neither group would likely be the Ang II-regulated kinase whose activity is Ͼ80% soluble. Of the eight other intracellular tyrosine kinases (Fer, Abl, Abl-related, Csk-1, FAK, JAK1, JAK2, and JAK3), all but Csk-1 and Abl-related were ruled out as the major calcium-dependent kinase in double immunoprecipitation experiments similar to those done in Figs. 2 and 3 (data not shown). Csk-1 is not autophosphorylated, and thus would not be a kinase activity immunoprecipitated with anti-Tyr(P) antibodies (46). This left the abl-related sequence (47) (which we have not formally ruled out, but which is not in the correct molecular weight range). Since a recent report demonstrated Ang II-dependent JAK2 activation in rat smooth muscle cells (48), we specifically tested for the ability of Ang II and a known JAK2 stimulus, growth hormone (49), to stimulate JAK2 autophosphorylation in GN4 cells. Fig. 5 shows that Ang II failed to increase JAK2 tyrosine phosphorylation at 1, 15, or 30 min, while human growth hormone clearly increases JAK2 tyrosine phosphorylation.
Purification-GN4 cells grown to confluence were treated with Ang II for 60 s, and cell lysates were harvested by detergent lysis. Lysates were incubated with anti-phosphotyrosine PT66-agarose, washed, and eluted with phenylphosphate, resulting in a substantial purification with respect to protein.
Multiple Tyr(P) substrates were isolated. This fraction (termed Tyr(P) elute in Fig. 6) was passed over an ATP-Sepharose column (ATP linked with a spacer through the ␥-phosphate), allowing efficient isolation of ATP-binding Tyr(P) proteins. Greater than 90% of the silver-staining protein and Tyr(P) substrate was not adsorbed and was found in the flow-through fractions of the ATP column. Fig. 6 (panel A) shows an analysis at the end of a typical purification using a standard gel polyacrylamide gel. At least five Tyr(P)-containing proteins adsorbed to the ATP column (p170, p140, p125, p115-120, and p75) were eluted with 1 mM ATP and were concentrated by overnight reincubation with PT66-agarose. This last step was followed by washing and extraction of the purified proteins in SDS sample buffer (final Tyr(P) step). The lower band in the 115-120-kDa region corresponded to the most heavily autophosphorylated protein seen in Fig. 3. A second purification was examined using a gel with a low bisacrylamide concentration to enhance separation of phosphoproteins in the 120-kDa region (Fig. 6B). The gel allows visualization of the 115-120-kDa Tyr(P) protein below the 120-kDa prestained marker separated from the preponderance of Tyr(P) substrates above the 120-kDa marker. Most of these did not adsorb to or elute from the ATP column. This can be seen even more clearly in Fig. 6C, which shows the results from a third purification in which the ATP eluate was concentrated ϳ20-fold using a final Tyr(P) step. Again the low bisacrylamide gel was used. This third purification was also used to confirm the specificity of the ATP affinity column by identifying two of the trace Tyr(P) proteins in final ATP eluate as the EGF receptor and pp125 FAK . The eluate from the ATP affinity resin was aliquoted into three fractions, which were concentrated by precipitation with either PT66-agarose (labeled the final Tyr(P) step), rat EGF receptor polyclonal antiserum, 1382, or the p125 FAK monoclonal 2A7. The PT66 lane was loaded with only 1/25 the sample used in the EGF receptor or p125 FAK lanes.

FIG. 4. Lack of complete inhibition of Ang II-dependent tyrosine phosphorylation in cells preincubated with cytochalasin D.
GN4 cells were pretreated with 2 M cytochalasin D for 2 h and then treated with Ang II for 1 min. Cells were lysed, and lysates were subjected to 8% SDS-PAGE and transferred to nitrocellulose as described under "Experimental Procedures." Phosphotyrosine immunoblotting (PT66) shows the presence of tyrosine phosphorylation in Ang II-treated cells that is resistant to cytochalasin inhibition (115-130-kDa region). The tyrosine phosphorylation in the 65-75-kDa region stimulated by Ang II is inhibited by cytochalasin. FIG. 5. JAK2 is not significantly tyrosine-phosphorylated in Ang II-treated GN4 cells. GN4 cells were treated with Ang II (AII, 1 M), growth hormone (GH, 100 ng/ml), or interferon ␥ (IFN␥, 500 units/ml) for the indicated times. Cells were lysed and immunoprecipitated with anti-phosphotyrosine (PT66) antibodies. Immune complexes were subjected to SDS-PAGE, transfer to nitrocellulose, and immunoblotting with anti-JAK2 antibody (generously provided by James Ihle, St. Judes). Growth hormone stimulated JAK2 autophosphorylation, while Ang II did not. The effect of IFN␥ was minimal. Lane 1 was immunoprecipitated with anti-JAK2, showing total immunoprecipitable JAK2 from GN4 cells.
p170 eluted from the ATP column was precipitated by the EGF receptor antibody. The Tyr(P)-containing protein band that migrated just above the 120-kDa molecular mass marker on low bisacrylamide gels was immunoprecipitated by p125 FAK monoclonal. Two separate immunoblots of the concentrated final Tyr(P) step were performed using the p125 FAK monoclonal antibody 2A7 or the polyclonal antibody BC-2 (raised against the p125 FAK tyrosine kinase domain, amino acids 311-701). These confirmed that the band migrating above the 120-kDa marker on the low bisacrylamide gel was p125 FAK ; the substrate migrating below the 120-kDa marker did not react with either the p125 FAK monoclonal or antisera even though there was greater than 20 times the silver staining protein in the 115-120-kDa band (data not shown). Fig. 7 shows that the tyrosine autokinase activity from an aliquot of the same immune complexes used for the immunoblots depicted in Fig. 6C. The auto kinase activity is almost totally confined to the band running just below the 120-kDa marker, with an apparent molecular mass of 115-120 kDa. The EGF receptor immunoprecipitate shows little or no autokinase activity, and the p125 FAK 2A7 immune complex exhibited barely detectable kinase activity. A reimmunoprecipitation with PT66 after immunodepleting EGF receptor or p125 FAK , respectively, showed that PT66 could still precipitate the 115-120-kDa autokinase activity. While the purification yielded tyrosine-phosphorylated EGF receptor and FAK that bind the ATP column, their tyrosine kinase activity is minimal either by virtue of their low abundance (compared to the 115-120-kDa protein) or due to differential loss of activity. The p140 and p75 substrates were not phosphorylated in vitro in this preparation under these conditions. Thus, 115-120-kDa protein appears to be the major kinase from the Ang II-treated cells. DISCUSSION Downstream signaling from G-protein-coupled receptors is extraordinarily diverse as these receptors are involved in some manner in most physiologic processes. In general, stimulation of G-protein-coupled receptors results in activation of Ser/Thr kinases (21,22). However, in certain cell types, hormones or agonist lipids that bind to G-protein-coupled receptors stimulate tyrosine phophorylation, and sometimes even cell proliferation. In addition, this subset of hormone G-protein-coupled receptors can activate the MAP kinase pathway (15, 16, 20, 36 -39) and can lead to immediate early gene expression 2 (50).
FIG. 6. Purification of a 120-kDa tyrosine kinase activity in GN4 cells by phosphotyrosine and ATP affinity column chromatography. Cells were treated with Ang II (1 M) for 1 min, lysed in detergent buffer, and purified over phosphotyrosine-agarose (PT66) as described under "Experimental Procedures." Proteins were eluted from the PT66-agarose with 10 mM phenyl phosphate and passed over a ATP-Sepharose column, from which they were eluted with 1 mM ATP. The eluates were immunoprecipitated with PT66-agarose at 0°C and eluted by boiling in SDS sample buffer. Aliquots were saved at each step for Tyr(P) immunoblot analysis. A, phosphotyrosine immunoblot of a standard 3%/8% SDS-PAGE gel showing proteins eluted from phosphotyrosine (PT66)-agarose (P-tyr eluate), the proteins not bound by ATP-agarose (ATP flow thru), the eluate from the ATP-agarose (ATP eluate), and concentration of the ATP eluate by a second phosphotyrosine immunoprecipitation (Final  p-tyr step). B, a second purification was analyzed by phosphotyrosine immunoblotting as in A, but using a gel with a low bisacrylamide ratio designed to better separate the 120-kDa region phosphoproteins. This panel indicates the presence of two tyrosine phosphorylated proteins just below the mass of Tyr(P) substrates that are selectively retained by ATP-Sepharose, suggesting that they are ATP-binding proteins. One of these Tyr(P) proteins migrates just below the 120-kDa marker. C, in a third purification, the fraction eluted from the ATP resin by ATP was aliquoted, precipitated with either anti-Tyr(P) (PT66), EGF receptor (EGFR(1382)), or p125 FAK (Fak(2A7)) antibodies, and run on a gel with low bisacrylamide ratio in preparation for Tyr(P) immunoblotting. Twenty-five-fold more immune complex was loaded into the EGFR and p125 FAK lanes. The immunoprecipitates demonstrate that at least two known kinases (EGFR and p125 FAK ) were purified by sequential Tyr(P) and ATP affinity chromatography. These two known kinases were present in lower amounts than the major Tyr(P)-containing protein, which migrated just below the 120-kDa marker.

FIG. 7. Autophosphorylation of Tyr(P) proteins immunoprecipitated from fractions eluted from ATP-Sepharose with ATP.
A portion of the purification used for the Tyr(P) immunoblot in Fig. 6C was used for the following immunoprecipitations and immune complex kinase assays. Proteins eluted from the ATP resin were immunoprecipitated: (i) directly with PT66-agarose (Final tyr-p step), (ii) with EGF receptor antibody, 1382 (EGFR IP), (iii) following the EGF receptor precipitation the remaining supernatant was reimmunopreciptated with PT66-agarose (Final Tyr(P) step-EGFR), (iv) with anti-p125 FAK antibody, 2A7 (FAK IP), or (v) following the 2A7 precipitation the remaining supernatant was reimmunoprecipitated with PT66-agarose (Final p-tyr step-FAK). All immune complexes were incubated with [␥-32 P]ATP 60 M (10 Ci) in phosphorylation buffer as described under "Experimental Procedures." Following a 1-min incubation, the samples were boiled in SDS sample buffer, subjected to SDS-PAGE, and autoradiographed. The EGF receptor and p125 FAK constitute a small fraction of the autophosphorylating kinase activity in the final Tyr(P) step, the majority of the activity was found in the 115-120-kDa protein.
In rat liver epithelial cells, Ang II alone not only stimulates proliferation, it also activates the MAP kinase and JNK kinase pathways and modifies transcription factor activity as adjudged by increased AP-1 binding. The Ang II actions on MAP kinase and JNK are protein kinase C-independent, as well as calcium/calmodulin-independent, yet both actions are inhibited by genistein, implicating the involvement of a tyrosine kinase. 2 Activation of JNK in a calcium-dependent, tyrosine kinase-dependent manner has not been observed in other cells (51). Thus, binding to the Ang II receptor and the resultant ␤␥ and calcium signals may have differential consequences depending upon the downstream signaling elements expressed in the cell types under study. The expression of a novel tyrosine kinase activated in a G-protein-and/or calcium-dependent manner may allow additional Ang II-dependent actions in some cells, i.e. Ang II may stimulate multiple pathways in cells in which it activates more than one tyrosine kinase.
Subcellular fractionation demonstrated that many of the Tyr(P) substrates and virtually all of the activated tyrosine kinase in Ang II-treated cells were soluble. Our catalogue of rat liver epithelial cell tyrosine kinases yielded several intracellular candidates (p125 FAK and the JAKs) as potential calciumregulated tyrosine kinases in the 115-130-kDa range. p125 FAK is both tyrosine phosphorylated and minimally activated in Ang II-treated cells, but it is a minor component of the Ang II-dependent response (Figs. 2 and 3). The JAK kinases, while often found in the particulate fraction, presumably non-covalently bound to their associated cytokine receptor (52), could be released from the membrane and activated by a calcium-dependent step. In addition, another group reported that Ang II activated JAK2 in smooth muscle cells (48), and a second group showed Ang II stimulation of serum-inducible element binding (53), often regarded as evidence of JAK pathway activation (52). We have not seen evidence of increased serum-inducible element binding in Ang II-treated GN4 cells 2 , nor do we detect significant JAK2 tyrosine phosphorylation in an Ang II-treated rat liver epithlelial cells (Fig. 7). Evidence for JAK2 activation in smooth muscle cells and its absence in rat liver epithelial cells confirms the cell type complexity of intracellular tyrosine kinase signaling.
The p125 FAK immunoprecipitation experiments (Figs. 2 and 3) suggested that another tyrosine kinase was activated in Ang II-treated cells; the kinase purification supports this conclusion. The sharp band of activity seen in immune complexes (Fig. 3) appears to be the major autophosphorylating kinase activity isolated by large scale Tyr(P) and ATP affinity chromatography (Figs. 6 and 7). After the ATP affinity step, the p115-120-kDa Tyr(P) autophosphorylating protein is the dominant band on silver-stained gels, and in some preparations, it is the only band (data not shown). By Tyr(P) immunoblotting, a much more sensitive technique than silver staining, most preparations eluted from the ATP resin and concentrated with PT66-agarose contain varying proportions of five Tyr(P) proteins: p170, p140, p125, p115-120, and p75. The p115-120 band is always dominant. The selectivity of the ATP affinity step was confirmed by the identity of p170 and p125 as known tyrosine kinases. The presence of the EGF receptor (p170) was surprising because even with immunoprecipitation of a 60-mm GN4 culture dish with anti-EGF receptor antibody, we have not observed Ang II-dependent EGF receptor phosphorylation. However, in purifications using 50 150-mm culture dishes, enough tyrosine-phosphophorylated EGF receptor is present to be specifically purified by the dual affinity procedure. The contribution of the EGF receptor and p125 FAK to the autokinase activity in the final preparation is dwarfed by the activity in p115-120-kDa protein (Fig. 7). Since the p115-120-kDa protein is much more abundant, we cannot speculate on the specific activities of the putative calcium-regulated tyrosine kinase (p115-120) and p125 FAK . The p115-120-kDa protein does not immunologically cross-react with either p125 FAK monoclonal 2A7, or a polyclonal raised against the p125 FAK tyrosine kinase domain. This does not preclude the putative novel kinases's similarity to p125 FAK , but suggests that any relationship may be distant. Finally, there is little 32 P incorporation into p140 and p75 at this stage of the purification. We do not know whether these are low specific activity kinases, other ATP binding proteins, or Tyr(P) proteins specifically bound to one of the ATP binding kinases.
In summary, Ang II activates at least two tyrosine kinases in rat liver epithelial cells, p125 FAK and a second activity, which appears to migrate at p115-120-kDa. In addition, Ang II activates at least two downstream signaling pathways: the Jun N-terminal kinase, activated in a calcium-and tyrosine kinasedependent manner, and MAP kinases whose activation by Ang II is less dependent on a calcium signal. 2 It will be of interest to determine whether the Ang II-dependent activation of distinct tyrosine kinases regulates separate intracellular signaling cascades.