Identification of a JAK2-independent Pathway Regulating Growth Hormone (GH)-stimulated p44/42 Mitogen-activated Protein Kinase Activity GH ACTIVATION OF Ral AND PHOSPHOLIPASE D IS Src-DEPENDENT*

We have demonstrated here that growth hormone (GH) stimulates the formation of the active GTP-bound form of both RalA and RalB in NIH-3T3 cells. Full activation of RalA and RalB by GH required the combined activity of c-Src and JAK2, both kinases activated by GH independent of the other. Activation of RalA and RalB by growth hormone did not require the activity of JAK2 per se . Ras was also activated by GH and was required for the GH-stimulated formation of GTP-bound RalA and RalB. Activation of RalA by GH subsequently resulted in increased phospholipase D activity and the formation of its metabolite, phosphatidic acid. GH-stim-ulated RalA-phospholipase D-dependent formation of phosphatidic acid was required for activation of p44/42 MAPK and subsequent Elk-1-mediated transcription stimulated by GH. Thus we report the identification of a JAK2-independent pathway regulating GH-stimulated p44/42 MAPK activity. NIH-3T3 cells were stimulated with the indicated doses of hGH for the indicated time periods, and the GST-linked probe RLIP76-RBD, which recognizes only the active GTP-bound form of RalA and RalB, was used to separate RalA-GTP and RalB-GTP from the inactive GDP-bound form of these molecules. GTP-bound RalA ( A and E ) and RalB ( C and G ) were visualized by Western blot analysis. Total cellular RalA ( B and F ) and RalB ( D and H ) were also determined in total cell lysate by Western blot analysis as protein loading control. The results presented are representative of a minimum of three (usually five) independent experiments.

Due to lack of intrinsic kinase activity, members of the cytokine receptor superfamily, including the growth hormone (GH) 1 receptor, recruit and activate non-receptor tyrosine kinases of the JAK family to relay their cellular signals (1). JAK2 has been reported to be the predominant JAK required for the initiation of GH signal transduction upon ligand binding to the GH receptor (2)(3)(4). To date, all identified downstream signaling pathways utilized by GH apparently require JAK2 activity (2)(3)(4). The only reported JAK2-independent effect of GH is Ca 2ϩ entry via L-type calcium channels (5), although this has been disputed (6,7). However, it is likely that other, as yet uncharacterized, signal transduction pathways stimulated by GH are activated independent of JAK2 activity.
The major groups of signaling molecules thus far identified to be activated by GH include the following: 1) other receptor (EGF receptor) (8) and non-receptor (c-Src, c-Fyn (9), and FAK (10)) kinases, although as in the case of the EGF receptor it may be used simply as an adapter protein; 2) members of the MAP kinase family including p44/42 MAP kinase (11,12), p38 MAP kinase (13), and c-Jun N-terminal kinase/stress-activated protein kinase (9) and the respective downstream pathways; 3) members of the insulin receptor substrate (IRS) group including IRS-1, -2, and -3 which may act as docking proteins for further activation of signaling molecules including phosphatidylinositol 3-kinase (14); 4) small Ras-like GTPases (15); and 5) STAT family members including STATs 1, 3, 5a, and 5b (16,17), which constitute one major mechanism for transcriptional regulation by GH.
Ras is a member of the Ras-like GTPase family (18,19). This family is characterized by similarities in the effector domain which Ras utilizes to interact with downstream target molecules. The Ras-like GTPases play a critical role in multiple signaling pathways leading from various cell-surface receptors. The activation and inactivation of the Ras-like GTPases are controlled by conformational change because of a GTP-GDP binding cycle that is controlled by the following three different regulatory proteins: guanine nucleotide exchange factors (GEFs), GTPase-activating proteins, and guanine nucleotide dissociation inhibitors. In its GTP-bound state, Ras in turn interacts with distinct downstream effectors and initiates multiple signaling pathways, which include at least three downstream signaling cascades mediated by the Raf protein kinase (i.e. A-Raf, B-Raf, and Raf-1), RalGEF (i.e. RalGDS, Rlf, and Rgl), and phosphatidylinositol (PI) 3-kinases (18 -20).
Recent reports (20,21) suggest that two other members of the Ras-like small GTPase family, namely RalA and RalB, possess pivotal roles in the control of cell proliferation, migration, differentiation, cytoskeletal organization, vesicular transport, and receptor endocytosis. Ral is also the member of the Ras-like GTPases family, and its activity is regulated by cycling between active GTP-bound and inactive GDP-bound states controlled through the direct binding of active Ras to Ral-specific GEFs. However, additional Ras-independent mechanisms also exist to stimulate the formation of GTPbound Ral. For example (22,23), RalA can be activated independently of Ras activation via its direct binding to Ca 2ϩ alone (24) or to Ca 2ϩ -bound calmodulin in response to the elevated level of intracellular calcium (23). Moreover, PI 3-kinase (25) and Src-like kinases (25,26) have also been implicated in Ral activation. Once activated, Ral further interacts with several other proteins that may function as its downstream effectors. RalA has been demonstrated to associate directly with phospholipase D1 (PLD1) via its N-terminal sequence and operates synergistically with another PLD1-interacting small GTPase, Arf, to activate PLD1 activity (27). Phospholipase D (PLD, including PLD1 and PLD2) is a widely expressed phospholipidspecific phosphodiesterase that hydrolyzes phosphatidylcholine, a major phospholipid in the cell membrane, to form phosphatidic acid (PA) and choline. PA can be further converted to diacylglycerol (DAG) and lyso-PA, both of which are the well known intracellular mediators and extracellular messengers of multiple biological activities (28,29). Two other proteins have also been reported that are to known to interact with the GTP-bound form of RalA, leading to RalA-dependent cellular effects. The first is Ral-binding protein 1 or RalBP1 (also called RLIP76) (30), which is involved in receptor-mediated endocytosis (30,31). RalBP1 is also a GTPase-activating protein for Cdc42, a Rho family member involved in actin cytoskeleton organization and filopodia formation in fibroblasts (30). The second is filamin, which serves as a downstream intermediate in Cdc42-mediated filopod production by its association with RalA (32).
We have demonstrated here that GH stimulates the formation of the active GTP-bound form of both RalA and RalB in NIH-3T3 cells. Activation of RalA and RalB by growth hormone did not require the activity of JAK2 per se. However, full activation of RalA and RalB by GH required the combined activity of both c-Src and JAK2, both kinases activated by GH independent of the other. Activation of RalA by GH subsequently resulted in the activation of PLD and the formation of phosphatidic acid that was required for activation of p44/42 MAP kinase by GH. Thus we report the identification of a JAK2-independent pathway regulating GH-stimulated p44/42 MAP kinase activity.

EXPERIMENTAL PROCEDURES
Materials-Recombinant human growth hormone (hGH) was a generous gift of Novo Nordisk (Singapore). Src kinase inhibitors PP1 and PP2 and phosphatidic acid were obtained from Biomol Research Laboratories (Plymouth Meeting, PA). The JAK2 inhibitor tyrphostin AG490, negative control of Src kinase inhibitor PP3, and brefeldin A (BFA) were purchased from Calbiochem. RalA monoclonal antibody and RalB polyclonal antiserum were purchased from Transduction Laboratories (Lexington, KY). c-Src polyclonal antiserum, HA monoclonal antibody, and protein A/G plus-agarose were obtained from Santa Cruz Biotechnology (Santa Cruz, CA). JAK2 polyclonal antiserum, anti-JAK2 IgG covalently coupled to protein A-agarose, Ras monoclonal antibody, and the Src kinase assay kit were purchased from Upstate Biotechnology, Inc. (Lake Placid, NY). Secondary anti-IgG antibodies, the ECL kit, [␥-32 P]ATP, and [ 3 H]palmitic acid were purchased from Amersham Biosciences. The PLD assay kit was obtained from Molecular Probes (Eugene, OR). The p44/42 MAP kinase assay kit was purchased from New England Biolabs (Beverly, MA). The transfection reagent Effectene was purchased from Qiagen (Hilden, Germany). All other reagents were purchased from Sigma.
pGEX4T3-GST-RalBD construct for GST-RLIP-RBD (33) containing amino acids 397-518 of human RLIP76 and pGEX 2T Ϯ RBD construct for GST-Raf1-RBD (34) containing amino acids 51-131 of Raf1 were the generous gifts of Dr. Johannes L. Bos (Utrecht, Netherlands). The wild type and dominant negative RalA plasmids were kindly provided by Dr. Yasutaka Ohta (Boston, MA). The wild type and dominant negative c-Src expression plasmids were obtained from Dr. Joan S. Brugge (Boston, MA). The dominant negative JAK2 expression plasmid was a kind gift of Dr. Olli Silvennoinen (Tampere, Finland). The wild type and dominant negative PLD1 and PLD2 plasmids were generously provided by Dr. Michael Frohman (Stony Brook, NY). The fusion trans-activator plasmid (pFA2-Elk-1) consisting of the DNA binding domain of Gal4 (residues 1-147) and the trans-activation domain of Elk-1 were purchased from Stratagene (La Jolla, CA). pFC2-dbd plasmid is the negative control for the pFA plasmid to ensure the observed effects are not due to the Gal4 DNA binding domain and was also obtained from Stratagene (La Jolla, CA). The dominant negative Ras plasmid was purchased from Upstate Biotechnology, Inc. (Lake Placid, NY). All plasmids were prepared with the plasmid megaprep kit from Qiagen (Hilden, Germany).
Cell Culture and Treatment-NIH-3T3 cells were cultured at 37°C in 5% CO 2 in Dulbecco's modified Eagle's medium supplemented with 10% heat-inactivated fetal bovine serum, 100 units/ml penicillin, 100 g/ml streptomycin, and 2 mM L-glutamine. Prior to treatment, cells were deprived of serum for 12-16 h in serum-free medium. Unless otherwise indicated, the final concentration of the PP1 was 50 M; PP2 was 20 M; PP3 was 50 M; AG490 was 100 M; and hGH was 50 nM. This concentration of GH is within the physiological range for circulating rodent GH.
Ral and Ras Activation Assay-NIH-3T3 cells were grown to subconfluence, incubated for 16 h in serum-free medium, washed once in serum-free medium, and incubated with 50 nM hGH for the indicated times. After stimulation with hGH and lysis with 1ϫ Ral buffer (10% glycerol, 1% Nonidet P-40, 50 mM Tris-HCl, pH 7.4, 200 mM NaCl, 2.5 mM MgCl 2 , 1 mM phenylmethylsulfonyl fluoride, 1 M leupeptin, 10 g of soybean trypsin inhibitor per ml, and 0.1 M aprotinin), samples were put on ice for 10 min and centrifuged at 14,000 rpm at 4°C for 10 min. Glutathione-Sepharose beads that had been precoupled to recombinant glutathione S-transferase (GST)-RalBP1-RBD or GST-Raf1-RBD were prepared as described (33,34). After preclearance of the lysates with glutathione-agarose, 15 g of GST-RalBP1-RBD or GST-Raf-1-RBDagarose precoupled to glutathione beads was added to 500 g of cell lysate per assay with gentle rocking at 4°C for 45 min. Samples were washed 3 times in lysis buffer, and bound proteins were eluted in 20 l of Laemmli sample buffer. Samples were separated by SDS-PAGE (12.5% polyacrylamide), immunoblotted, and probed with the respective antibodies.
JAK2 Immunoprecipitation-Cells were lysed at 4°C in 1 ml of lysis buffer (50 mM Tris-HCl, pH 7.4, 1% Triton X-100, 150 mM NaCl, 1 mM EGTA, 1 mM EDTA, 1 mM sodium orthovanadate, 0.5% Nonidet P-40, 0.1% phenylmethylsulfonyl fluoride) for 30 min with regular vortices. Cell lysates were centrifuged at 14,000 ϫ g for 15 min, and the resulting supernatants were collected, and protein concentration was determined. 800 g of protein was used for each immunoprecipitation. Immunoprecipitations were performed by incubating lysates with 20 l of gel slurry of anti-JAK2/protein A-agarose. The reaction mixture was gently rocked at 4°C for 1 h. Immunoprecipitations were washed 3 times with ice-cold lysis buffer. The pellet was resuspended in 1ϫ SDS sample buffer containing 50 mM Tris, pH 6.8, 2% SDS, 2% ␤-mercaptoethanol, and bromphenol blue, boiled for 10 min, and centrifuged at 14,000 ϫ g for 5 min. The supernatant was collected and subjected to 8% SDS-PAGE. Proteins were transferred to nitrocellulose membranes using standard electroblotting procedures.
Immunoblotting-After preincubation with inhibitors for the indicated times and/or incubation with the indicated concentration of hGH for the appropriate duration, the cells were washed once with ice-cold PBS and lysed at 4°C in an appropriate amount of lysis buffer. Cell lysates were dissolved and denatured in 1ϫ SDS-PAGE sample buffer, and separation was achieved on 8 -12% SDS-polyacrylamide gels and transferred to nitrocellulose membranes. The membranes were blocked with 5% non-fat dry milk in phosphate-buffered saline with 0.1% Tween 20 (PBST) for 1 h at 22°C. The blots were then treated with the primary antibody in PBST containing 1% non-fat dry milk at 4°C overnight. After three washes with PBST, immunolabeling was detected by ECL according to the manufacturer's instructions. For reblotting, membranes were stripped by incubation for 30 min at 50°C in a solution containing 62.5 mM Tris-HCl, pH 6.7, 2% SDS, and 0.7% mercaptoethanol. Blots were then washed for 30 min with several changes of PBST at room temperature. Efficacy of stripping was determined by re-exposure of the membranes to ECL. Thereafter, blots were reblocked and immunolabeled as described above.
Src Kinase Assay-Src kinase assays were performed as described according to the manufacturer's instructions (Upstate Biotechnology, Inc.). In brief, supernatant containing 150 g of protein per sample derived from cells stimulated with hGH was incubated with 1 g of Src polyclonal antibody at 4°C for 2-4 h in a final volume of 500 l. Immunocomplexes were collected by incubation with 20 l of protein A/G plus-agarose for 1 h. Immunoprecipitates were washed 3 times with ice-cold lysis buffer. 10 l (150 M final concentration) of the substrate peptide, 10 l of Src reaction buffer, and 10 l of [␥-32 P]ATP stock were added to a microcentrifuge tube and incubated for 10 min at 30°C with agitation. 20 l of 40% trichloroacetic acid was then added to precipitate peptides, and a 25-l aliquot was transferred onto the center of a numbered P81 paper square. The assay squares were washed 5 times for 5 min each with 0.75% phosphoric acid and once with acetone. The assay squares were transferred to a scintillation vial, and 5 ml of scintillation mixture was added, and the level of radioactivity was determined in a scintillation counter. The sample that contains no enzyme serves as the background control.
p44/42 MAP Kinase Assay-p44/42 MAP kinase assays were per-formed according to the manufacturer's instructions. In brief, cells were lysed at 4°C in lysis buffer (20 mM Tris, pH 7.4, 150 mM NaCl, 1 mM EDTA, 1 mM EGTA, 1% Triton X-100, 2.5 mM sodium pyrophosphate, 1 mM glycerol phosphate, 1 mM Na 3 VO 4 , 0.1% phenylmethylsulfonyl fluoride, 1 g/ml leupeptin). The lysates were centrifuged at 15,000 ϫ g for 15 min at 4°C. The supernatant containing 200 g of protein per sample was incubated overnight at least 4 h with an immobilized phospho-specific p44/42 MAP kinase (Thr 202 /Tyr 204 ) monoclonal antibody (1:300 dilution) in a final volume of 500 l in 1ϫ lysis buffer. The pellets were washed twice with 500 l of lysis buffer containing 1 mM phenylmethylsulfonyl fluoride and twice with 500 l of kinase buffer (25 mM Tris, pH 7.5, 5 mM glycerol phosphate, 2 mM dithiothreitol, 0.1 mM Na 3 VO 4 , 10 mM MgCl 2 ). The kinase reactions were performed in the presence of 2 g of Elk-1 fusion protein and 200 M ATP at 30°C for 30 min. Elk-1 phosphorylation was selectively detected by Western immunoblotting using a chemiluminescence detection system and a specific phospho-Elk-1 (Ser 383 ) antibody (1:1000 dilution). Measurement of Phosphatidic Acid and PLD Activity-Subconfluent cells were serum-starved and labeled overnight with [ 3 H]palmitic acid (5 Ci/ml) in Dulbecco's modified Eagle's medium. After cells were stimulated with 50 nM hGH for the indicated time, the samples were placed on ice, rinsed with cold PBS, and the lipids extracted by the method of Bligh and Dyer (35). The dried samples were resuspended in chloroform/methanol (2:1) and developed by TLC on Silica Gel 60 plates (Merck) using chloroform/methanol/acetic acid (90:10:10, v/v/v) as the solvent with the unlabeled PA and phosphatidylethanol as lipid standards (Avanti Polar Lipids). For PLD activity assays, cells were stimulated with hGH in the presence of 1% ethanol to determine the total activity of PLD by the standard transphosphatidylation assay. The plates were air-dried, treated with EN 3 HANCE (PerkinElmer Life Sciences), and exposed to a Kodak X-Omat AR film. The appropriate lipid spots were marked and scraped from the TLC plates and counted via liquid scintillation. Levels of PA or phosphatidylbutanol were normalized to total fatty acid label incorporated into lipid. PLD activity assays were also performed by use of the amplex red phospholipase kit from Molecular Probes (Eugene, OR). In brief, cells were lysed in the 1ϫ reaction buffer with 1% Triton X-100 by several quick freeze-thaw cycles at Ϫ80°C (10 -15 min each). 100 l of diluted samples containing 20 g of total whole lysate was used to perform the assay. The fluorescence in a fluorescent microplate reader was measured using excitation detection at 540 nm and emission detection at 595 nm. Each point was triplicated, and the reading was corrected by subtracting the values derived from the non-PLD controls.
Transient Transfection and Elk-1 Reporter Assay-NIH-3T3 cells were cultured to 60 -80% confluence for transfection experiments in 6-well plates (16). 0.2 g of pCMV␤ and 0.2 g of reporter plasmid pFR-Luc were transfected together with 4 ng of the respective fusion trans-activator plasmid (pFA-Elk-1 or pFC2-dbd). For each well, 10 -20 l of Effectene for each g of DNA was used as per the manufacturer's instructions. DNA-lipid complex was diluted to a final volume of 2 ml (for triplicate samples) with 2% fetal bovine serum medium and cells allowed to grow for 12-16 h. 50 nM hGH was added for an additional 24 h. The cells were washed in PBS and lysed with 200 l of 1ϫ lysis buffer (25 mM Tris-phosphate, pH 7.8, 2 mM EDTA, 2 mM dithiothreitol, 10% glycerol, 1% Triton X-100) by a freeze-thaw cycle, and lysate was collected by centrifugation at 14,000 rpm for 15 min. The supernatant was used for the assay of luciferase and ␤-galactosidase activity. The luciferase activities were normalized on the basis of protein content as well as on the ␤-galactosidase activity of pCMV␤ vector. The ␤-galactosidase assay was performed with 20 l of precleared cell lysate according to a standard protocol (13).
Statistical Analysis and Presentation of Data-All experiments were performed at least 3 times. Numerical data are expressed as mean Ϯ S.D. Data were analyzed using the two-tailed t test or analysis of variance.

RESULTS
hGH Stimulation of NIH-3T3 Cells Increases the Level of GTP-bound RalA and RalB-We employed the GST-linked probe RLIP76-RBD (33), which recognizes only the active GTPbound form of RalA and RalB and not the inactive GDP-bound form of these molecules, to determine the level of RalA-GTP and RalB-GTP in lysates of NIH-3T3 cells stimulated by GH. hGH stimulation of NIH-3T3 cells resulted in the rapid formation of GTP-bound RalA and RalB which could be observed within 30 s of cellular stimulation with hGH ( Fig. 1, A and C).
The hGH-stimulated formation of both RalA-GTP and RalB-GTP was biphasic, with the first peak of activity observed at 1-2 min after stimulation with hGH, followed by a decline to 15 min, and a second peak of GTP-bound RalA and RalB observed at 30 min, again followed by a decline to 60 min. hGH stimulation of NIH-3T3 cells did not alter RalA or RalB protein levels over the examined period of stimulation ( Fig. 1, B and D). The hGH-stimulated formation of both RalA-GTP and RalB-GTP was also dose-dependent with stimulation of the GTP-bound forms of RalA and RalB first observed at 0.005 nM hGH and maximal stimulation from 5 to 50 nM hGH (Fig. 1, E and G). Thus RalA and RalB are two signaling molecules utilized by hGH to exert its effect on cellular function.
hGH-stimulated Activations of JAK2 and c-Src Are Independent-We next wished to determine the upstream kinases responsible for the hGH-stimulated conversion of RalA and RalB to the GTP-bound form. It was necessary, however, to first examine the potential interdependence between two GH activated kinases, namely JAK2 and c-Src (9, 36), after cellular stimulation with hGH. Neither the generic Src family kinase inhibitor PP1, nor the more specific Src kinase inhibitor PP2, nor the structurally related non-inhibitory PP3 affected JAK2 tyrosine phosphorylation stimulated by hGH ( Fig. 2A). Simi- larly, forced expression of either wild type c-Src or a kinasedeficient c-Src by transient transfection of the respective cDNAs into NIH-3T3 cells did not alter the level of hGHstimulated JAK2 tyrosine phosphorylation (Fig. 2C). Such a result was obtained despite marked overexpression of the respective c-Src molecules (Fig. 2E). Both PP1 and PP2 inhibited hGH-stimulated c-Src kinase activity (Fig. 2F) under the conditions in which it failed to affect tyrosine phosphorylation of JAK2. As expected, PP3 did not affect c-Src kinase activity. Forced expression of the kinase-deficient c-Src in NIH-3T3 cells also inhibited c-Src kinase activity (Fig. 2G). Therefore, hGHstimulated tyrosine phosphorylation of JAK2 in NIH-3T3 cells is independent of the activity of c-Src.
We next examined whether the activity of JAK2 was required for hGH-stimulated activation of c-Src. We therefore first utilized the JAK2-specific inhibitor AG490. AG490, even when utilized at a high concentration more than sufficient to that reported to inhibit JAK2 activity (100 M was utilized even though 20 M is sufficient to inhibit hGH-stimulated activation of JAK2), failed to inhibit hGH-stimulated c-Src kinase activity. To demonstrate the efficacy of AG490, we also examined the effect of the same concentration of AG490 on the ability of hGH to stimulate tyrosine phosphorylation of JAK2. Pretreatment of NIH-3T3 cells with AG490 before hGH stimulation completely prevented hGH-stimulated tyrosine phosphorylation of JAK2 (Fig. 2H). Similarly, forced expression of a kinasedeficient JAK2 (K882E) (37) also failed to inhibit c-Src kinase activity (Fig. 2F), despite the ability of this kinase-deficient JAK2 to abrogate hGH-stimulated JAK2 tyrosine phosphorylation under the same conditions. Thus, hGH-stimulated activations of JAK2 and c-Src kinases are two independent and parallel events.
Full Activation of RalA and RalB by hGH Requires Both JAK2 and c-Src-We proceeded to determine which of the two independently activated kinases described above were responsible for hGH-stimulated conversion of RalA and RalB to the GTP-bound form. We therefore first examined the effect of the JAK2 kinase inhibitor AG490 on hGH-stimulated conversion of RalA and RalB to the GTP-bound form. AG490 treatment of NIH-3T3 cells decreased the basal level of both RalA-GTP and RalB-GTP without alteration in the total cellular level of either RalA or RalB (Fig. 3, A-D). Treatment of NIH-3T3 cells with AG490 slightly diminished, but did not prevent, RalA-GTP and RalB-GTP formation stimulated by hGH. In contrast, AG490

FIG. 2. HGH-stimulated activation of JAK2 and c-Src are independent and parallel.
A-E, hGH-stimulated activation of JAK2 is independent of c-Src activity. NIH-3T3 cells were stimulated with hGH after pretreatment and in the continued presence of vehicle, PP1, PP2, or PP3 as indicated, and cell extracts were prepared and processed for determination of tyrosine-phosphorylated JAK2 (A). Total JAK2 present in the JAK2 immunoprecipitates is indicated (B). NIH-3T3 cells were transiently transfected with the empty vector or an expression vector containing either wild type c-Src or kinase-dead c-Src (c-Src-KD) and stimulated with hGH, and cell extracts were prepared and processed for determination of tyrosine-phosphorylated JAK2 (C). Total JAK2 present in the JAK2 immunoprecipitates is indicated (D). The efficacy of wild type c-Src and kinase-dead c-Src overexpression is indicated by Western blot (E). F and G, hGH-stimulated activation of c-Src is independent of JAK2 activity. NIH-3T3 cells were stimulated with hGH after pretreatment and in the continued presence of vehicle, PP1, PP2, PP3, or AG490 as indicated, and cell extracts were prepared and processed for determination of Src kinase activity (F). NIH-3T3 cells were transiently transfected with the empty vector or an expression vector containing either kinase dead c-Src (c-Src-KD) or kinase-dead JAK2 (K882E) and stimulated with hGH, and cell extracts were prepared and processed for determination of Src kinase activity (G). H-K, AG490 and dominant negative JAK2 inhibit hGHstimulated JAK2 tyrosine phosphorylation. NIH-3T3 cells were stimulated with hGH after pretreatment and in the continued presence of vehicle or AG490 as indicated, and cell extracts were prepared and processed for determination of tyrosine-phosphorylated JAK2 (H). Total JAK2 present in the JAK2 immunoprecipitates is indicated (I). NIH-3T3 cells were transiently transfected with the empty vector or an expression vector for dominant negative JAK2 (K882E) and stimulated with hGH, and cell extracts were prepared and processed for determination of tyrosinephosphorylated JAK2 (J). Total JAK2 present in the JAK2 immunoprecipitates is indicated (K). The results presented are representative of a minimum of three (usually five) independent experiments. completely prevented hGH-stimulated tyrosine phosphorylation of the JAK2 substrate STAT5 (data not shown). Similarly forced expression of a kinase-deficient JAK2 slightly diminished, but did not prevent, RalA-GTP and RalB-GTP formation stimulated by hGH. Thus JAK2 is not required for hGH-stimulated conversion of RalA and RalB to the GTP-bound form. We next examined the effect of the generic Src family kinase inhibitor PP1, the more specific Src kinase inhibitor PP2, and the structurally related non-inhibitory PP3 on RalA-GTP and RalB-GTP formation stimulated by hGH. Both PP1 and PP2 abrogated, but did not completely prevent, RalA-GTP and RalB-GTP formation stimulated by hGH (Fig. 3E). PP3 did not affect hGH-stimulated RalA-GTP and RalB-GTP formation. The effect of PP1 and PP2 on inhibition of RalA-GTP and RalB-GTP formation stimulated by hGH was more potent compared with the effect of AG490 on hGH-stimulated GTP-bound RalA and RalB (compare Fig. 3, E and G to A and C). Forced expression of kinase-inactive c-Src also diminished, and to a greater extent than kinase-inactive JAK2, RalA-GTP, and RalB-GTP formation stimulated by hGH (Fig. 3, I and K). Because removal of the kinase activities of either JAK2 or c-Src only partially inhibited the formation of GTP-bound RalA and RalB stimulated by hGH, we therefore examined whether combined inhibition of JAK2 and c-Src would completely prevent RalA-GTP and RalB-GTP formation stimulated by hGH. Transient transfection of both kinase-inactive JAK2 and kinaseinactive c-Src cDNAs completely prevented hGH-stimulated RalA-GTP and RalB-GTP formation (Fig. 3, I and K). Thus, full activation of RalA and RalB by hGH required the combined activities of JAK2 and c-Src kinases.
Ras Activity Is Required for hGH-stimulated RalA-GTP and RalB-GTP Formation-Activation of RalA and RalB has been demonstrated previously (23,38) to require Ras-dependent activation of RalGEFs thereby defining Ral as a Ras effector molecule. We therefore first examined the ability of hGH to stimulate the formation of Ras-GTP in NIH-3T3 cells. hGH stimulation of NIH-3T3 cells resulted in the rapid appearance of the GTP-bound form of Ras, maximal at 2 min and followed by a decline in the level of Ras-GTP such that at 30 min after stimulation with hGH the level of Ras-GTP returned to basal activity (Fig. 4A).
We next proceeded to determine the kinase dependence of the hGH-stimulated conversion of Ras to the GTP-bound form. We first examined the effect of the JAK2 kinase inhibitor AG490 on hGH-stimulated conversion of Ras to the GTP-bound form. AG490 treatment of NIH-3T3 cells completely prevented hGH-stimulated activation of Ras without alteration in the total cellular level of Ras ( Fig. 4C and Fig. 4D). The Src kinase inhibitors PP1 and PP2 used above were without effect on the ability of hGH to stimulate the formation of GTP-bound Ras (Fig. 4E). Concordantly forced expression of a kinase-deficient JAK2 prevented Ras-GTP formation stimulated by hGH, whereas kinase-inactive c-Src was without effect on hGH-stimulated Ras activation (Fig. 4G). Thus, Ras activation by GH is JAK2-dependent.
To determine whether Ras activity was required for hGHstimulated formation of RalA-GTP and RalB-GTP, we examined the ability of hGH to stimulate RalA-GTP and RalB-GTP formation in the presence of forced expression of RasN17. RasN17 forms a nonproductive complex with RasGEFs (22,23) and therefore inhibits endogenous Ras activity. Forced expression of RasN17 abrogated the ability of hGH to stimulate the formation of both GTP-bound RalA and RalB (Fig. 4, K and M). Thus hGH-stimulated formation of RalA-GTP and RalB-GTP is Ras-dependent.
RalA Is Required for hGH-stimulated p44/42 MAP Kinase Activity and Elk-1-mediated Transcription-p44/42 MAP kinase has been demonstrated previously (15,39) to be activated by GH in a Ras-dependent manner. Because activation of RalA and RalB by GH is also Ras-dependent, we reasoned that Ral C-F, AG490 but not PP1, PP2, and PP3 inhibits hGH-stimulated formation of GTP-bound Ras. NIH-3T3 cells were stimulated with hGH after pretreatment and in the continued presence of vehicle or AG490, PP1, PP2, PP3, and GTP-bound Ras were visualized by Western blot analysis as indicated. Total cellular Ras (D and F) was also determined in total cell lysate by Western blot analysis as protein loading control. G-J, kinase-dead JAK2 but not kinase-dead c-Src inhibits hGH-stimulated formation of GTP-bound Ras. NIH-3T3 cells were transiently transfected with the empty vector or an expression vector containing either kinase-dead c-Src (c-Src-KD) or kinase-dead JAK2 (K882E) and stimulated with hGH. GTP-bound Ras (H) were visualized by Western blot analysis as indicated. Total cellular Ras (H) was also determined in total cell lysates by Western blot analysis as protein loading control. K-O, dominant negative Ras (RasN17) inhibits hGH-stimulated formation of the GTP-bound form of RalA and RalB in NIH-3T3 cells. NIH-3T3 cells were transiently transfected with either empty vector or the expression vector for RasN17 and stimulated with hGH as described under "Experimental Procedures." The GST-linked probe RLIP76-RBD was used to separate RalA-GTP and RalB-GTP from the inactive GDP-bound form of these molecules. GTP-bound RalA (K) and RalB (M) were visualized by Western blot analysis as indicated. Total cellular RalA (L) and RalB (N) were also determined in total cell lysates by Western blot analysis as protein loading control. The results presented are representative of a minimum of three (usually five) independent experiments. may be upstream of, and influence, p44/42 MAP kinase activity stimulated by GH. We therefore examined the effect of forced expression of either wild type RalA or a dominant negative form of RalA (RalA28N) on the ability of hGH to activate p44/42 MAP kinase activity. As reported previously (40), hGH stimulation of NIH-3T3 cells resulted in a rapid time-dependent increase in the activity of p44/42 MAP kinase. Thus maximal activation of p44/42 MAP kinase activity was observed between 5 and 15 min after stimulation with hGH followed by a decline in activity to 60 min (Fig. 5A). Forced expression of wild type RalA slightly increased the basal activity of p44/42 MAP kinase activity and resulted in a prolonged activation of p44/42 MAP kinase activity in response to hGH. Thus, little or no dimunition in p44/42 MAP kinase activity was observed 30 -60 min after stimulation with hGH in comparison to the vector-transfected control (Fig. 5A). Forced expression of RalA at the different time points was demonstrated by the appearance of RalA at a slightly higher molecular weight (due to the presence of a FLAG tag) than endogenous RalA on Western blot analysis for RalA on the same whole cell lysates used for estimation of p44/42 MAP kinase activity (Fig. 5B). Forced expression of RalA28N resulted in decreased basal p44/42 MAP kinase activity, significantly less activation of p44/42 MAP kinase activity after GH stimulation, and shorter duration of the hGHstimulated p44/42 MAP kinase activity (Fig. 5C). Similarly, forced expression of RalA28N at the different time points was demonstrated by the appearance of RalA28N at a slightly higher molecular weight (due to the presence of a FLAG tag) than endogenous RalA on Western blot analysis for RalA on the same whole cell lysates used for estimation of p44/42 MAP kinase activity (Fig. 5D). Thus, RalA is required for full activation of p44/42 MAP kinase activity by hGH in NIH-3T3 cells.
GH has been reported previously (13,41) to stimulate transcription via Elk-1 in a p44/42 MAP kinase-dependent manner (13,42). Furthermore, it has been reported that sustained activation of p44/42 MAP kinase is required for activation of Elk-1-mediated transcription (43). Because RalA overexpression resulted in sustained activation of GH-stimulated p44/42 MAP kinase, we examined the effect of forced expression of RalA and a RalA dominant negative mutant (RalA28N) on the ability of hGH to stimulate Elk-1-mediated transcription in NIH-3T3 cells. Forced expression of RalA increased the basal level of Elk-1-mediated transcription and dramatically increased the ability of hGH to stimulate transcription via Elk-1 (Fig. 5E). Forced expression of RalA28N reduced the basal level of Elk-1-mediated transcription and completely prevented hGH-stimulated Elk-1-mediated transcription (Fig. 5E). Thus RalA is required for full p44/42 MAP kinase activation by hGH and subsequent Elk-1-mediated transcription.
hGH Activates Phospholipase D-One of the proteins proposed to mediate the effects of Ral on cellular function is phospholipase D (27,38,44). We therefore examined whether hGH stimulation of NIH-3T3 cells would also result in an increase of phospholipase D activity. For determination of the effect of hGH on PLD activity, cells were stimulated with hGH in the presence of 1% ethanol, and PLD activity measured by the standard transphosphatidylation assay (27,38,44). hGH stimulation of NIH-3T3 cells resulted in a rapid rise in PLD activity, maximal at 5 min, and followed by a decline in activity to 60 min (Fig. 6A). We also measured the effect of cellular stimulation with hGH on PLD activity by use of the commercially available amplex red phospholipase D kit. As observed in Fig. 6, A and B, hGH stimulated an increase in PLD activity similar to that observed by use of TLC to determine PLD activity. Thus cellular stimulation with hGH resulted in an increase in PLD activity.
PLD catalyzes the hydrolysis of phosphatidylcholine and phosphatidylethanolamine to form phosphatidic acid (28,29). We consequently next examined the effect of hGH stimulation of NIH-3T3 cells on PA production. Cells were serum-deprived and concomitantly incubated with [ 3 H]palmitic acid before stimulation with hGH. The migration of PA in thin layer chromatography was identified against lipid standards, and the appropriate spot was removed and radioactivity determined. As is observed in Fig. 6C, hGH stimulation of NIH-3T3 cells resulted in a rapid rise in the level of PA with peak levels of PA observed at 5 min after stimulation. A sustained increase in the level of PA was observed to at least 60 min after stimulation with hGH. Thus hGH stimulates PA production in NIH-3T3 cells (Fig. 6C).
hGH Activation of PLD Is RalA-dependent-To determine whether hGH-stimulated activation of phospholipase D is Raldependent, we determined the effect of forced expression of either wild type RalA or a dominant negative form of RalA (RalA28N) on the ability of hGH to activate PLD. As observed in Fig. 6D, transfection of wild type RalA cDNA did not significantly alter the basal level of PLD activity but significantly enhanced the hGH-stimulated increase in PLD activity. Conversely, transfection of NIH-3T3 cells with a dominant negative form of RalA, although not altering the basal level of PLD activity, abrogated the ability of hGH to stimulate increases in PLD activity (Fig. 6D). Thus hGH-stimulated activation of PLD is RalA-dependent.
PLD Activity Is Required for hGH-stimulated p44/42 MAP Kinase Activity and Elk-1-mediated Transcription-PLD1 has been demonstrated to exist in a complex with RalA and the small G protein ADP-ribosylation factor-1 (ARF1) (27). Inhibition of ARF1 with the fungal metabolite BFA has been demonstrated to prevent the activation of PLD by extracellular stimuli (28,29,46). We therefore first examined the effect of BFA on the ability of hGH to stimulate p44/42 MAP kinase activity. As observed in Fig. 7A, BFA in concentrations ranging from 1 to 50 g/ml effectively inhibited the activation of p44/42 MAP kinase by hGH. To demonstrate that the inhibition of hGH-stimulated p44/42 MAP kinase by BFA was specifically due to inhibition of PLD-dependent PA production, we added exogenous PA at the same time as BFA and examined p44/42 MAP kinase activity in response to cellular stimulation with hGH. Exogenously added PA, in the presence of 50 g of BFA which effectively prevented p44/42 MAP kinase activation by hGH, restored hGH-stimulated p44/42 MAP kinase activity (Fig. 7B). Thus BFA inhibition of hGH-stimulated p44/42 MAP kinase activity by hGH was specifically due to inhibition of PA production.
To verify the results obtained with BFA, we therefore examined the effect of forced expression of PLD1, an enzymatically inactive form of PLD1 (PLD1-K898R), PLD2, and an enzymatically inactive form of PLD2 (PLD2-K758R) on hGH-stimulated p44/42 MAP kinase activity. Forced expression of PLD1 increased the basal level of p44/42 MAP kinase activity and also increased hGH-stimulated p44/42 MAP kinase activity (Fig.  7C). The enzymatically inactive form of PLD1 unexpectedly also increased basal p44/42 MAP kinase activity but prevented any stimulation of p44/42 MAP kinase activity by hGH (Fig.  7C). Forced expression of PLD2 did not significantly alter the basal activity of p44/42 MAP kinase but significantly enhanced p44/42 MAP kinase activity stimulated by hGH (Fig. 7C). The enzymatically inactive form of PLD2 did not significantly alter basal p44/42 MAP kinase activity and prevented the hGHstimulated increase in p44/42 MAP kinase activity (Fig. 7C). The forced expression of PLD1, PLD1-K898R, PLD2, and PLD2-K758R was verified by Western blot analysis (Fig. 7D).
We next examined the effect of forced expression of PLD1, an enzymatically inactive form of PLD1 (PLD1-K898R), PLD2, and an enzymatically inactive form of PLD2 (PLD2-K758R) on hGH-stimulated Elk-1-mediated transcription. Forced expression of PLD1 increased both basal and hGH-stimulated Elk-1mediated transcription (Fig. 7F). Similar to the result observed with the effect of PLD1-K898R on basal p44/42 MAP kinase activity, forced expression of this enzymatically inactive PLD1 also increased the basal level of Elk-1-mediated transcription and prevented an hGH-stimulated increase in Elk-1-mediated transcription. Forced expression of PLD2 did not increase the basal level of Elk-1-mediated transcription but substantially enhanced hGH-stimulated Elk-1-mediated transcription when compared with the vector-transfected control. Forced expression of PLD2-K758R completely prevented hGH-stimulated Elk-1-mediated transcription (Fig. 7G).
A RalA-PLD Pathway Is Required for hGH-stimulated Activation of Elk-1-mediated Transcription-We have demonstrated above that forced expression of RalA dramatically enhanced the ability of hGH to stimulate Elk-1-mediated transcription and that activation of PLD in response to cellular stimulation with hGH was RalA-dependent. It was therefore required to demonstrate that the RalA enhancement of hGHstimulated Elk-1-mediated transcription was PLD-dependent. We therefore tested whether the enzymatically inactive PLD2 (PLD2-K758R) could inhibit the increase in hGH-stimulated Elk-1-mediated transcription consequent to forced expression of RalA. As observed above, transfection of RalA cDNA dramatically enhanced both the basal and hGH-stimulated Elk-1mediated transcription, and PLD2-K758R prevented any hGHstimulated increase in Elk-1-mediated transcription. When both RalA and PLD2-K758R were transfected together, an increase in the basal level of Elk-1-mediated transcription was evident as is observed for forced expression of RalA alone, but no hGH-stimulated increase in Elk-1-mediated transcription was observed (Fig. 8). Thus, RalA enhancement of hGH-stim-FIG. 6. hGH activates phospholipase D and stimulates phosphatidic acid production. A and B, hGH activates phospholipase D activity in NIH-3T3 cells. NIH-3T3 cells were stimulated with hGH for the indicated times and processed for determination of PLD activity by either the standard transphosphatidylation assay (A) or by use of the commercially available amplex red phospholipase D kit (B). C, hGH stimulates phosphatidic acid production. NIH-3T3 cells were serumdeprived and concomitantly incubated with [ 3 H]palmitic acid before stimulation with hGH. The migration of PA in thin layer chromatography was identified against lipid standards, and the appropriate spot was removed and radioactivity determined. D, hGH activation of PLD is RalA-dependent. NIH-3T3 cells were transiently transfected with the expression vectors for wild type RalA or dominant negative RalA (RalA28N), stimulated with hGH, and processed for PLD activity as indicated under "Experimental Procedures." Data presented are mean Ϯ S.E. of triplicate determinations. Experiments were repeated a minimum of 3 (usually 5) times.
ulated Elk-1-mediated transcription requires the activity of at least PLD2. DISCUSSION We have demonstrated here that hGH stimulation of NIH-3T3 results in a rapid and biphasic activation of the small Ras-like GTPases RalA and RalB. Similar biphasic activation of Ral (30) and other small Ras-like GTPases including Ras (25,47) and Rap1 (48) has been reported in other systems. It has been postulated that this phenomenon is due to the differential input dynamics of upstream signals and/or differential feedback from downstream effector molecules. For example, it has been reported that biphasic activation of Ras by endothelin-1 was linked to the sequential activation of two downstream pathways, p44/42 MAP kinase and PI 3-kinase (47). The decrease in Ras activity following the first peak of Ras activity stimulated by endothelin-1 elicits a negative feedback through p44/42 MAP kinase-dependent Sos1 phosphorylation, whereas the second peak of Ras activation by endothelin-1 is facilitated by persistent tyrosine phosphorylation of SHC (47). The implication of such a bipartite activation of Ral and its potential contribution to GH signal transduction requires further investigation.
We have demonstrated here that hGH-stimulated activation of c-Src is independent of the activity of JAK2. This is the first reported example that a kinase utilized by GH does not require the activity of JAK2. Other kinases utilized by GH, such as focal adhesion kinase, have been reported to associate with and require the activity of JAK2 (10). We have reported previously (9) that GH also activates another Src kinase, c-Fyn, although the dependence of its activation on JAK2 was not investigated. It is also likely that other members of the Src family of kinases activated by GH, such as c-Fyn, are activated independent of JAK2. The related hormone prolactin, which also utilizes JAK2 for its signal transduction, has also been demonstrated previously (6) to stimulate c-Src-independent activation of JAK2. Other examples of JAK-independent activation of kinases by members of the cytokine receptor superfamily includes interleukin-3 activation of Src (49) and erythropoietin activation of Lyn (50). Interestingly it has been reported (51) that angiotensin II stimulates an association between the N terminus of JAK2 and the SH2 domain of c-Src which is dependent on the activity of JAK2. Whether such an association was also required for activation of Src by angiotensin II was not demonstrated (51). In any case such an association would allow JAK2 and c-Src to be spatially co-located and may facilitate synergistic interactions where common signaling molecules are involved. In this regard it is relevant to note that GH stimulates the association of JAK2 and FAK, and FAK is one component of a multiprotein signaling complex centered around CrkII and also including c-Src (9). FAK has also been reported to be a Src substrate (10). In this case, however, at least the activity of JAK2 was not required for GH-stimulated activation of c-Src. We cannot exclude the possibility that the JAK2 molecule itself, and not JAK2 kinase activity, may be required for the activation of c-Src. However, prolactin is able to activate c-Src in the absence of the proline-rich Box1 region of the prolactin receptor required for the activation of JAK2, and therefore JAK2 association with the prolactin receptor is not required for prolactin-stimulated Src activity (6). Thus, there may exist multiple independent parallel pathways by which GH could affect cellular function. It remains to be determined what the contribution of JAK2-independent signaling pathways will be to the final cellular effects of GH. Analysis of the genetic targets of the different signaling pathways by cDNA microarray may prove useful in this regard and is in progress.
We have demonstrated here that full activation of RalA and RalB by GH requires the combined activity of both c-Src and JAK2. It is interesting to note however that the impairment in Ral activation by inhibition of c-Src is considerably greater than that observed by inhibition of JAK2. This phenomenon was observed with utilization of both pharmacological inhibitors and cellular expression of the respective kinase-deficient molecules. It is therefore apparent that formation of GTPbound Ral by GH is predominantly mediated by GH-stimulated c-Src activity. The requirement for JAK2 activity for full GHstimulated activation of RalA and RalB is apparently due to the exclusive JAK2-dependent activation of Ras by GH (see below). Thus we have identified two signaling molecules (RalA and RalB) that can be activated by GH, albeit to a lesser extent, in the absence of JAK2 activity. Both fMet-Leu-Phe and plateletactivating factor activation of Ral in neutrophils have also been demonstrated to be partially dependent on c-Src (25). Furthermore, RalA has been demonstrated previously (44,58) to mediate activation of PLD in v-Src-transformed cells. It is interesting to note that Ral has also been demonstrated to regulate the activity of c-Src in response to cellular stimulation by EGF (52). It is therefore possible that Ral participates in regulating the final "output" of the GH-stimulated multiprotein signaling complex centered around CrkII and containing c-Src (9) in addition to functioning in the linear pathway we have described here. Further support for a role of c-Src in GH signal transduction is the ability of Csk (Src-inactivating kinase) to inhibit GH-stimulated p44/42 MAP kinase activity (53), and this observation is likely to be mediated by the Src-Ral-PLD pathway we have described here (also see below for discussion).
Another small GTPase, Ras, has been demonstrated previously (15,39) to be activated by GH, and we have also observed here that GH stimulates the rapid formation of GTP-bound Ras in NIH-3T3 cells. Ral proteins are activated by RalGEFs which are themselves activated by direct binding to Ras (18,20). Transient transfection of the dominant negative Ras mutant RasN17 attenuated GH-stimulated formation of GTP-bound RalA and RalB suggestive that GH activation of RalA and RalB is also Ras-dependent. Ral has further been demonstrated to be activated by Ras-independent pathways (22)(23)(24), and the failure of RasN17 to inhibit completely GH-stimulated formation of RalA-GTP and RalB-GTP indicates that GH also utilizes Ras-independent pathways to participate in Ral activation. Ral has been reported to be activated by Rap1 (33). In addition, both Rap1 (54) and RalA (23,24,33) can be activated in response to an elevated level of intracellular calcium. It has been reported that Src-like kinase activity is required for GH-stimulated calcium influx (7,55). As GH-stimulated Ral activation is also Src kinase-dependent, it would be reasonable to propose that GH-stimulated Ral activity might also be mediated via Ca 2ϩ influx. By use of calcium channel inhibitors, we have indeed demonstrated that GH activation of RalA and RalB is also dependent on Ca 2ϩ influx via L-type calcium channels. 2 Whether Rap1 is also involved in GH-stimulated Ral activation requires further investigation. We have observed that GH stimulation of Chinese hamster ovary cells stably transfected with GH receptor cDNA (CHO-GHR-( 1-638 )) results in a potent activation of Rap1, whereas minimal activation of Rap1 by GH is observed in NIH-3T3 cells. 3 We have also observed a similar preferential activation of c-Jun N-terminal kinase by GH in CHO-GHR-( 1-638 ) cells in comparison to NIH-3T3 cells due to a relative deficiency of CrkII (9,40), and Rap1 activation has been demonstrated previously (56) to be CrkII-dependent. In any case, it remains to be determined if Rap1 will participate in the activation of RalA or RalB in NIH-3T3 cells. Other Rasrelated molecules such as TC21 have also been demonstrated to activate Ral (57). Further work should delineate the signaling molecules downstream of JAK2 and c-Src utilized by GH to stimulate the formation of GTP-bound RalA and RalB.
It is interesting to note in this study that the overexpression of wild type RalA resulted in an extended activation of p44/42 MAP kinase activity but did not increase the maximal level of GH-stimulated p44/42 MAP kinase activity. An analogous situation has been described for nerve growth factor-stimulated activation of p44/42 MAP kinase in PC12 cells (43). In that example, Ras was required for the initial activation of p44/42 MAP kinase by nerve growth factor, and the small GTPase Rap1 maintained the activation of p44/42 MAP kinase (43). Similarly, the Rap1-sustained activation of p44/42 MAP kinase by nerve growth factor was required for full activation of Elk-1-mediated transcription (43). We also observed that overexpression of RalA resulted in dramatically increased Elk-1-mediated transcription stimulated by GH indicative that RalA is a pivotal component in the mediation of the effects of p44/42 MAP kinase activation by GH. Analogously, the decreased GHstimulated activation of p44/42 MAP kinase observed upon overexpression of the dominant negative RalA28N may simply be due to the inability of the cell to maintain p44/42 MAP kinase in an activated form rather than any deficit in activation. In any case, overexpression of dominant negative RalA28N resulted in the absence of GH-stimulated Elk-1-mediated transcription. Thus RalA regulation of p44/42 MAP kinase activity to produce sustained high level activation would be required for the full transcriptional program initiated upon activation of p44/42 MAP kinase by GH.
We have demonstrated here that GH stimulation of NIH-3T3 cells results in the activation of PLD and the subsequent generation of phosphatidic acid in the cells. RalA has been demonstrated previously (44) to mediate activation of PLD in v-Srctransformed cells. Thus, overexpression of RalA potentiated PLD activation by v-Src, and dominant negative RalA inhibited PLD activity in both v-Src-and v-Ras-transformed cells (55). We have analogously demonstrated that hGH stimulates the activation of RalA in both a c-Src-and Ras-dependent manner and that the hGH-stimulated activation of PLD is indeed RalAdependent. The association of RalA and Arf has been demonstrated previously (27) to be required for increased PLD activity. PLD-catalyzed hydrolysis of phospholipids results in the generation of PA. The generation of PA by GH was demonstrated to be essential for GH-stimulated p44/42 MAP kinase activation as BFA (which prevents PLD activation and subsequent PA production by inhibiting Arf GTP-GDP exchange) dramatically diminished GH-stimulated p44/42 MAP kinase activation. Furthermore, the pretreatment of cells with PA significantly reversed the inhibition of GH-stimulated p44/42 MAP kinase activation by BFA. Transfection of dominant negative mutants of PLD (PLD1-K898R or PLD2-K758R) also prevented GH-stimulated p44/42 MAP kinase activation and Elk-1-mediated transcription. It is therefore apparent that PA serves as an effector generated as a result of PLD activation for p44/42 MAP kinase activation by GH. It has been proposed that PLD and its PA product mediate agonist-dependent Raf-1 translocation to the plasma membrane and the subsequent activation of the p44/42 MAP kinase pathway (46). The recruitment of Raf-1 to membranes is mediated by direct interaction of Raf-1 with PA and is independent of association with Ras (46,59). It remains to be determined whether the requirement of PA for GH-stimulated p44/42 MAP kinase activation is due to a similar mechanism. It is possible that Ral and Raf-1 may independently activate the p44/42 MAP kinase pathway as both Raf and RalGDS signaling independently stimulate hTBP promoter activity in a mitogen-activated protein kinase/extracellular signal-regulated kinase kinase activation-dependent manner (60). Other GH-stimulated direct PA-dependent cellular events remain to be determined. PA has been proposed to be a potent activator of signaling molecules such as tyrosine kinases, GTPase-activating protein, PI 4-kinase (which produces the PLD cofactor phosphatidylinositol 4,5-bisphosphate), in addition to Raf (61). PA has further linked to Ca 2ϩ signaling (62) and to superoxide anion production through NADPH oxidase (63).
Alternatively, once produced, PA may be hydrolyzed by PA phosphohydrolase to produce DAG with resultant activation of PKC (28,29,61). Although PKC has been demonstrated to be required for GH-stimulated activation of p44/42 MAP kinase in other cellular systems (3), we have observed no PKC dependence of GH-stimulated p44/42 MAP kinase activation here. 4 Thus, Ral is unlikely to regulate GH-stimulated p44/42 MAP kinase activity by DAG generation from PA and subsequent PKC activation. PA can also be converted to lyso-PA and arachidonic acid (AA) by the action of phospholipase A 2 (28,29). GH has been demonstrated previously (64) to activate PLA 2 , and activation of PLA 2 by GH increases the level of AA and subsequent formation of AA metabolites. Inhibition of PLA 2 partially inhibits GH-stimulated p44/42 MAP kinase activation, 2 suggestive that the catalytic action of PLA 2 on PA is involved in the Ral-PLD-p44/42 MAP kinase pathway described here.
Ral has been implicated in the control of cell proliferation and Ras-mediated oncogenic transformation (19,44,57). For example, expression of RalGEFs or activated Ral proteins can cooperate with activation of other Ras effector cascades to result in cellular transformation (65). Although GH stimulation of NIH-3T3 cells results in a marked increase in p44/42 MAP kinase activity, there is little increase in cell number in response to exogenous GH. Thus activation of Ras and Ral per se will not necessarily result in mitogenesis nor oncogenic transformation. In other cellular systems, such as the mammary carcinoma cell, both autocrine and exogenously added hGH result in p44/42 MAP kinase-dependent mitogenesis (66,67). In this regard it is interesting that autocrine hGH production by mammary carcinoma cells results in a dramatic increase in cyclin D1 transcription (68), and Ral has been demonstrated previously to regulate cyclin D1 gene transcription through NF-B (65). Furthermore, overexpression in NIH-3T3 cells of a Ras effector mutant that activates RalGEF but not Raf or PI 3-kinase resulted in the formation of a metastatic and invasive phenotype (69). We have observed that autocrine production of hGH in mammary carcinoma cells also results in an invasive phenotype, 5 and it is likely that Ral may be required for such an effect. PLD1 has also been demonstrated to contribute to cellular proliferation as transfected PLD1 results in the oncogenic transformation of 3Y1 cells overexpressing the EGF receptor (45).
In conclusion, we have demonstrated that GH stimulation of NIH-3T3 cells results in JAK2-independent formation of GTPbound RalA and RalB with subsequent regulation of GH-stimulated p44/42 MAP kinase activity through PLD. A diagrammatic summary of this pathway is provided in Fig. 9. The identification of JAK2-independent activation of specific signaling pathways by GH will dramatically increase our understanding of the repertoire of signaling molecules utilized by GH to achieve its pleiotropic cellular effects.