当前位置: 首页 > 医学版 > 期刊论文 > 内科学 > 内分泌学杂志 > 2005年 > 第8期 > 正文
编号:11168736
Integrin Activates Receptor-Like Protein Tyrosine Phosphatase , Src, and Rho to Increase Prolactin Gene Expression through a Final Phosphati
     Department of Pharmacology (A.I.V., K.K.J., F.M.S.) and Kaplan Cancer Center (F.M.S.), New York University School of Medicine, New York, New York 10016

    Address all correspondence and requests for reprints to: Dr. Frederick M. Stanley, Department of Pharmacology, MSB407, New York University Medical Center, 550 First Avenue, New York, New York 10016. E-mail: stanlf01@med.nyu.edu.

    Abstract

    We previously showed that receptor-like protein tyrosine phosphatase (RPTP)- inhibited insulin-increased prolactin gene transcription. Others suggested that RPTP was a key intermediary between integrins and activation of Src. We present evidence that inhibition of insulin-increased prolactin gene transcription was secondary to RPTP activation of Src, reflecting its role as mediator of integrin responses. Src kinase activity was increased in GH4 cells transiently or stably expressing RPTP and cells plated on the integrin-5?1 ligand fibronectin. C-terminal Src kinase inactivated Src and blocked RPTP inhibition of insulin-increased prolactin gene transcription. Expression of dominant-negative Src also prevented the RPTP-mediated inhibition of insulin-increased prolactin gene expression. Low levels of a constitutively active Src mutant (SrcY/F) stimulated whereas higher expression levels of Src Y/F inhibited prolactin gene expression. Src-increased prolactin gene transcription was inhibited by expression of a blocking Rho-mutant (RhoN19), suggesting that Src acted through or required active Rho. Experiments with an activated Rho-mutant (RhoL63) demonstrated a biphasic activation/repression of prolactin gene transcription that was similar to the effect of Src. The effects of both Src and Rho were phosphatidylinositol 3-kinase dependent. Expression of SrcY/F or RhoL63 altered the actin cytoskeleton and morphology of GH4 cells. Taken together, these data suggest a physiological pathway from the cell matrix to increased prolactin gene transcription mediated by RPTP/Src/Rho/phosphatidylinositol 3-kinase and cytoskeletal change that is additive with effects of insulin. Over activation of this pathway, however, caused extreme alteration of the cytoskeleton that blocked activation of the prolactin gene.

    Introduction

    THE INSULIN-INCREASED prolactin gene transcription seen in GH4 cells, a prolactin-producing pituitary tumor cell line, is an important model for studying gene transcription activated by insulin (1). Autophosphorylation of the insulin receptor upon ligand binding activates its kinase activity (1). The insulin receptor then phosphorylates substrates, e.g. the insulin receptor substrates (IRSs) and Shc that activate downstream signaling pathways to mediate the diverse effects of insulin (2). Insulin-activated prolactin gene transcription depends on activation of phosphatidyl inositol 3-kinase (PI 3-kinase) (2). The subsequent phosphatidyl inositol-3,4,5-trisphosphate-dependent kinase(s) have not yet been identified, but this cascade results in the activation of Elk-1. Elk-1 was identified as the transcription factor that binds the insulin response element of the prolactin promoter and is required for insulin-increased prolactin gene expression (3).

    RPTP blocks insulin-increased prolactin gene transcription in GH4 cells (2). RPTP is a member of the protein tyrosine phosphatase (PTP)ase family that includes the receptor-like, membrane-spanning PTPases and cytosolic PTPases. Receptor-like PTPases are characterized by an extracellular domain of variable length, a single membrane-spanning domain, and one or two catalytic domains on the intracellular portion of the molecule (4). RPTP does not dephosphorylate the insulin receptor or its immediate substrates shc and IRS-1 in our experiments (2), but others have found some effects of RPTP on the association of PI 3-kinase with IRS-1, although this did not affect downstream signaling (5). This suggests that its effects on insulin-increased prolactin gene expression are mediated by downstream signaling molecules. Src represents a potential mediator downstream of RPTP because RPTP dephosphorylates and activates c-Src (6) and RPTP–/– fibroblasts had a phenotype similar to Src–/– fibroblasts (6, 7, 8).

    Studies with v-Src and activated c-Src demonstrate that the Src family of nonreceptor tyrosine kinases activate numerous signaling pathways, many of which are also activated by receptor tyrosine kinases (9). A triple knockout of Src, Fyn, and Yes, however, causes defective integrin-signaling but does not affect signaling by receptor tyrosine kinases. This indicates that Src might be important for integrin signaling. Src’s importance to integrin signaling is supported by the role that many Src substrates [e.g. cortactin, focal adhesion kinase (FAK), paxillin, p130CAS, and vinculin] play in cytoskeletal organization and cell adhesion (9). In response to integrin activation, perinuclear localized c-Src translocates to the cell membrane in which it is activated by dephosphorylation or protein-protein interaction. A study of integrin signaling in mouse fibroblasts showed the importance of RPTP for force generation by Src family tyrosine kinases (10). Active Src causes a rearrangement in cytoskeletal organization. Thus, Src-related kinases are likely to be crucial mediators of cell-cell/cell-substrate interactions.

    Insulin/IGF-I and integrin signaling are interdependent (11, 12). Integrins are essential for insulin secretion by isolated pancreatic ?-cells (13). This may be due to their importance for calcium release from intracellular stores. Insulin receptor directly interacts with integrin-V?3 (11) and FAK and Src directly phosphorylate the insulin receptor (14). Internalization of insulin receptor is also dependent on interaction of cells with the extracellular matrix (15). Ligation of 6?4-integrin in breast cancer cell lines results in phosphorylation of IRS-1 and IRS-2 and PI 3-kinase activation (12, 16). IRS-1 expression is dependent on cell adhesion and FAK (17, 18). Conversely, insulin activates PI 3-kinase and ERK through FAK (19). SHP-2, a down-regulator of insulin signaling (20), is also important for integrin signaling in some systems (21). Blocking V?3-integrin binding to its ligand vitronectin blocks insulin signaling in NIH3T3 cells (22) and a 24-kDa fragment of fibronectin promotes insulin-induced differentiation of fat cells (23). Both insulin and integrin are required for optimal stimulation of MAPK activation in REF52 cells (24). GH3B6 cells that are derived from the same tumor as the GH4 cells express 5?1 (fibronectin receptor) 6?1 (laminin receptor), and CD44 (hyaluronic acid receptor) (25).

    The experiments presented in this paper support an integrin/RPTP/Src signaling pathway that affects prolactin gene expression. We first show that RPTP or plating on fibronectin activates Src in GH4 cells. Inhibition of Src blocks the effect of RPTP, suggesting that Src is required for the effects of RPTP. Low levels of activated Src stimulate insulin-increased prolactin gene transcription, whereas higher levels of activated Src block it. These effects of Src depend on Rho and PI 3-kinase. Finally, overexpression of either activated Src or activated Rho alter the cytoskeleton.

    Materials and Methods

    Materials

    [32P]ATP, 3000 Ci/mmol, was obtained from ICN Biochemicals Corp. (Costa Mesa, CA). Fast chloramphenicol acetyltransferase (CAT), reagent for fluorescent assay of CAT activity, was purchased from Molecular Probes (Eugene OR). Reagents for assay of luciferase were from Promega (Madison, WI). Acetyl-CoA, enolase, and silica gel plates for thin-layer chromatography were obtained from Sigma (St. Louis, MO). DMEM containing 4.5 g/liter glucose and iron-supplemented calf serum was obtained from Hyclone Laboratories (Logan, UT). LY294002 and PD98059 were from Calbiochem (La Jolla, CA). All other reagents were of the highest purity available and were obtained from Sigma, Pierce (Rockford, IL), Behring Diagnostics (Deerfield, IL), Bio-Rad Laboratories (Hercules, CA), Eastman (Rochester, NY), Fisher (Fair Lawn, NJ), or Roche Molecular Biochemicals (Indianapolis, IN).

    Antibodies

    Antibody to Src was from Oncogene Science (Cambridge, MA). Antibody against human influenza virus hemagglutinin was purchased from Covance (Vienna, VA). Antisera against RPTP and RPTP were described previously (26, 27). Anti-Src Y416 was from Cell Signaling Technology (Beverly, MA). Anti-myc antibody was a gift from Dr. J. Sap (New York University, New York, NY). Anti-green fluorescent protein (GFP) was from Molecular Probes. Anti-Flag was purchased from Sigma. Antihuman insulin receptor monoclonal antibody 83–14 was the generous gift of Dr. Kenneth Siddle (University of Cambridge, Cambridge, UK).

    Plasmids

    The construction of pPrl-CAT plasmids containing –173/+75 of prolactin (Prl) 5'-flanking DNA and mutant of the prolactin promoter were described (28, 29, 30). The human insulin expression vector, pRT3 human insulin receptor (HIR)2, was the gift of Dr. J. Whittaker (Haegdorn Institute, Copenhagen, Denmark). Dr. J. Sap gave us the expression plasmids for RPTP (31) and RPTP (27). An HA-tagged RPTP was given by Dr. D. Shalloway (Cornell, Ithaca, NY) (32) and was recloned into pcDNA3. Plasmids expressing Src wild-type (wt), Src Y527F, Src K295M, Fyn K299M, and C-terminal Src kinase (CSK) were the gift of Dr. Robert Schnieder (New York University School of Medicine) (33, 34). These were subcloned to express Myc and/or GFP epitope tags. Cytomegalovirus (CMV) RhoN19 and CMV RhoL63 that were Myc tagged were the gift of Dr. A. Hall (University College, London, UK) and have been described (35, 36, 37). GFP-PTB-associated splicing factor (PSF) was the gift of Dr. M. Mathur (New York University School of Medicine). The expression of transfected plasmids was verified by Western blotting of epitope tags.

    Analysis of prolactin promoter responsiveness using transient transfection

    Electroporation experiments and reporter assays were performed as described (38). GH4 cells were harvested with an EDTA solution, and 20 to 40 x 106 cells were used for each electroporation. All electroporations contained 5 μg of the plasmid pHIR-RT3 that expresses high levels of the human insulin receptor (30). This is necessary to achieve the high levels of insulin stimulation seen in these studies and is consistent with numerous other systems in which cotransfection of receptors has been necessary to achieve physiological regulation of transfected genes (30). Experiments with GH4 cells stably transfected with the human insulin receptor give similar results. Trypan blue exclusion before electroporation ranged from 95 to 99%. The voltage of the electroporation was 1550 V. This gives trypan blue exclusion of 70–80% after electroporation. The transfected cells were then plated in multiwell dishes (Falcon Plastics, Oxnard, CA) at 5 x 106 cells per 9 cm2 tissue culture well in DMEM with 10% hormone-depleted serum (29, 39). Cells were refed at 24 h with DMEM with 10% hormone depleted serum ± 1 μg/ml bovine insulin (Calbiochem). This concentration of insulin was used to avoid problems with insulin degradation. Previous studies showed that insulin activates prolactin-CAT expression at physiological insulin concentrations (30). Furthermore, despite reports by others that IGF-I activates prolactin expression (40) through the insulin response element, we have not observed this, even using a high concentration of IGF-1 (100 ng/ml). After 48 h, the flasks were washed three times with normal saline and frozen. The cells were harvested and reporter activity was assayed. CAT activity was assayed using acetyl-CoA and BODIPY chloramphenicol (Molecular Probes) as described previously (30). Fluorescence intensity was measured using a FluorImager 575 (Amersham, Piscataway, NJ) with ImageQuant (Amersham Biosciences, Piscataway, NJ) software. CAT activity was normalized for variability of transfections using ?-galactosidase as described below.

    In later experiments, a Prl (–173/+75)luciferase reporter plasmid was used in place of Prl (–173/+75)CAT. Luciferase assays were performed on GH4 cell lysates using reagents and protocols from Promega. Luciferase activity was normalized for variability of transfections using ?-galactosidase as described below. Control experiments demonstrated that stimulation of Prl (–173/+75)luciferase by insulin and other hormones was identical with that seen with the corresponding CAT reporter (data not shown).

    A Rous sarcoma virus-?-galactosidase expression plasmid was included in the electroporations. This plasmid is not expressed and its inclusion has no effect on the overall results of the experiments, but it was included to control for minor variations in transfection efficiency. Briefly, 2 μg of Rous sarcoma virus-?-galactosidase expression plasmid was included in the electroporations. The ?-galactosidase activity in the cell lysates was determined using o-nitrophenyl-?-D-galactopyranoside. Transfection efficiency did not vary significantly among transfections performed at the same time. The percent acetylation was then corrected for minor variations in ?-galactosidase activity by converting the percent acetylation to percent acetylation/OD430 ?-galactosidase activity per milligram protein. The fold stimulation or inhibition was then determined.

    Immunoprecipitation and Western immunoblot analysis

    GH4 cells were harvested in a lysis buffer consisting of 50 mM HEPES (pH 7.5), 1% Triton X-100, 150 mM NaCl, 1.5 mM MgCl2, 1 mM EGTA, 10% glycerol, 1 mM Na3VO4, 50 mM Na4P2O7, 1 mM NaF, 1 mM [4-(2-aminoethyl)-benzenesulfonylfluoride, HCl], and 10 μg/ml aprotinin. Protein was determined using the Bradford reagent (Bio-Rad Laboratories). Equal amounts of protein were then analyzed by SDS-PAGE using 10% gels. The proteins were then transferred to nitrocellulose membranes (Micron Separations, Westboro, MA). The membranes were first stained with Ponceau S to assess the quality of the transfer and verify that all lanes had been loaded equally. Immunoblots were visualized using enhanced chemiluminescence (Pierce). The human insulin receptor was immunoprecipitated from lysates of transfected cells using an human insulin receptor specific antibody, 83–14 (41) as described previously (1). This antibody bound to an epitope on the extracellular surface of the insulin receptor -subunit (42, 43).

    Src kinase assay

    Cytosolic lysates (as above) were made from GH4 cells and GH4 cells that stably expressed RPTP (A23 cells) or RPTP (K5 cells) (2). The cSrc was immunoprecipitated from 1 mg of crude lysate using anti Src antibody (Oncogene Sciences). Immunoprecipitations were performed for 2–16 h at 4 C. They were then incubated an additional 2 h with protein A or protein G agarose (1.5 mg/immunoprecipitation) and washed extensively. The kinase assay was performed with immunoprecipitated c-Src. The assay contained 10 mM HEPES (pH 7.4), 50 μM [32P]ATP, 25 μM ATP, 10 mM MgCl2, 1 mM dithiothreitol, and 0.1 μg of enolase that had first been denatured. The reaction was stopped by addition of 2 x electrophoresis sample buffer and the supernatant was resolved on a 10% sodium dodecyl sulfate-polyacrylamide gel.

    Cell staining and fluorescence microscopy

    GH4 cells were electroporated with GFP-PSF alone or together with CMV-SrcY/F or CMV-RhoL63. Control electroporations contained GFP-PSF and pcDNA3 (vector). The cells were inoculated onto poly-L-lysine-treated coverslips at a density of 250,000 cells/coverslip. The media were exchanged at 24 h, and the cells were incubated an additional 24 h. The coverslips were then washed twice with PBS. They were fixed using 4% paraformaldehyde in PBS and permeabilized with 0.1% Triton X-100. The cells were then stained with Alexa Fluor 647 Phalloidin (Molecular Probes) and mounted using Mowiol. The cells were digitally scanned using a 510 scanning confocal microscope (Zeiss, Thornwood, NY) equipped with helium/neon and argon lasers. Zeiss Laser Scanning Microscope software version 2.8 was used in data acquisition.

    Statistical analysis

    InStat software (GraphPad Software, San Diego, CA) was used for all statistical analysis. The data were subjected to one-way ANOVA and the Tukey-Kramer multiple comparisons test was used to determine the significance of the observed differences.

    Results

    Plating cells on fibronectin increased activated Src and basal and insulin-increased prolactin gene transcription

    Others had suggested that integrin signaling was coupled to Src activation through RPTP (10). We tested a number of different coating materials to determine whether basal or insulin-increased prolactin gene transcription was affected by the substrate. Tissue culture plates were treated overnight at 4 C with collagen, spermidine, or fibronectin. They were washed once with PBS and plated with GH4 cells that had been electroporated with Prl-luciferase reporter (Fig. 1A). Cells plated on fibronectin had significantly increased basal (0.05 > P > 0.01) and insulin-increased (0.001 > P) prolactin gene expression, whereas no significant differences were noted with the other treatments. Plating on fibronectin also increased activated Src levels in GH4 cells as assessed with an activation site-specific antibody (Fig. 1, B and C). Levels of activated Src were relatively constant in control cells throughout the experiment. The level of activated Src was significantly higher in fibronectin-plated cells generally reaching a peak at about 18–24 h before declining to control levels at 48 h. Insulin and other plate-coating materials (collagen and spermidine) had no effect on Src activation (data not shown). These results are consistent with an effect mediated by integrin-5?1, the fibronectin receptor, reported to be present on GH cells (25).

    FIG. 1. Src activation and basal and insulin-increased prolactin gene expression in GH4 cells plated on fibronectin. A, GH4 cells were electroporated with 10 μg Prl (–173/+75) luciferase. They were plated into wells precoated overnight at 4 C with 0.1 mg/ml of the agent indicated. The coating material was aspirated and the wells were washed with PBS immediately before use. The cells were refed at 24 h and then treated with/without insulin for an additional 24 h. The cells were harvested and luciferase enzyme activity was determined as directed by the manufacturer (Promega). The average relative light units per 100 μg protein in control and insulin-treated cultures were determined and the insulin incubations were compared with control levels to determine the fold-stimulation by insulin (fold-control). The average + SEM from three separate experiments done in triplicate is shown. B, GH4 cells were plated in wells treated with fibronectin or PBS as controls. They were harvested at the times indicated and frozen at –80 C. Equal amounts of protein were then used for SDS-PAGE analysis and Western blotting first with an antibody to SrcPY416 and next with an antibody to total Src. C, Western blots of three experiments performed as in B were scanned using a densitometer (Amersham Biosciences, Piscataway, NJ). The bands were quantitated using ImageQuant software and the relative Src activity ± SEM is shown.

    Src activity in GH4 cells is increased by RPTP

    A number of experiments showed that Src activity was increased by RPTP. Src was immune precipitated from the parental GH4 cells and from GH4 cells that stably expressed high levels of RPTP (A23 cells) or RPTP (K5 cells). The immune precipitated Src was then used in a kinase assay with enolase as a substrate (Fig. 2A). Phosphorylation of enolase was higher in immune precipitates from the A23 cells than in K5 or parental GH4 cells. Quantitation of these experiments showed a significant elevation (0.01 > P) of Src kinase activity in the A23 cells (Fig. 2C) but not in the K5 cells. Western blot analysis showed that the amount of Src in the immunoprecipitated extracts was not significantly different in the various cell lines (Fig. 2B). This indicated that Src activity not amount was increased in the cell line that expressed elevated levels of RPTP. Similar results were obtained with transiently transfected cells (Fig. 2D). Src activity was increased in a dose-dependent manner in RPTP- but not in RPTP-transfected cells as determined with an activation state-specific antibody. The actual increase in Src activity in this experiment is much greater than the figure shows because only 10–20% of the cells express RPTP/ in these experiments.

    FIG. 2. Src kinase activity in WT- and RPTP-expressing cell lines. Cell lysates were prepared from GH4, A23, and K5 cells. A23 cells are GH4 cells that were stably transfected with an RPTP expression vector and were selected by Western blotting for maximum RPTP expression (2 ). K5 cells are GH4 cells that were stably transfected with an expression vector for RPTP and were selected by Western blotting for maximum expression of RPTP (2 ). The lysates were immunoprecipitated with anti-Src antibody and the immunoprecipitates were used in kinase assays with the substrates enolase and [32P]ATP (see Materials and Methods). A, Results of a typical kinase assay are shown. 32P-labeled enolase was separated on SDS-PAGE and visualized using a phosphoimager. B, Western blot of Src protein levels in extracts of GH4, A23, and K5 cells used for kinase assay shown in A. C, Three kinase experiments done in duplicate were quantitated using ImageQuant software (Amersham Biosciences). The average kinase activity for control, A23, and K5 cells was calculated. The activity of the control was set to 1 for each experiment and the fold increase in Src kinase activity in the A23 and K5 cells was calculated. Results are the average + SEM. D, GH4 cells electroporated with vector alone (control) or the indicated amounts of expression vectors for RPTP or RPTP. They were harvested 24 h after electroporation, and equal amounts of cell protein were analyzed by SDS-PAGE and blotted to nitrocellulose. Antibody to Src phosphotyrosine 416 was used to identify the amount of activated Src. The blot was stripped and reprobed with an antibody to total Src. These results are representative of four similar experiments.

    CSK reverses RPTP inhibition of insulin-increased prolactin-CAT expression

    Phosphorylation of tyrosine 527 inactivates Src. The kinase that phosphorylates Src on tyr 527 is CSK (9). The expression of CSK should counteract the inhibition of insulin-increased prolactin gene transcription by RPTP if RPTP works by activating Src. CSK alone significantly increased the effect of insulin on prolactin gene expression (0.001 > P, Fig. 3A). This might be expected if CSK inactivated Src-related kinases that reduced basal and insulin-increased prolactin gene transcription, although it could also be due to independent effects of CSK. RPTP inhibited insulin-increased prolactin gene expression 85% to levels that were not significantly higher than cultures incubated without insulin (P > 0.05). Expression of CSK along with RPTP completely counteracted this effect of RPTP. GH4 cells that were cotransfected with expression vectors for both CSK and RPTP had higher levels of insulin-increased prolactin transcription than seen in GH4 cells electroporated with the vector alone (0.001 > P). This could result from effects of CSK on endogenous Src-related kinases or from CSK activation of independent processes. CSK did not affect the expression of RPTP in GH4 cells (Fig. 3B) and neither CSK nor RPTP levels were altered by insulin (insulin did not affect the expression of any transiently expressed protein used in these studies). More CSK is expressed in RPTP-cotransfected cells in this experiment. This suggests a large activation of Src by RPTP because the stimulation by CSK is reduced despite higher levels of CSK that would be expected to stimulate insulin-increased prolactin gene expression.

    FIG. 3. CSK reverses the effect of RPTP-GH4 cells were transfected with 10 μg of Prl (–173/+75) CAT with 5 μg of pRK5-PTP, 15 μg of pCSK, or 5 μg of pRK5-PTP and 10 μg of pCSK as indicated. DNA concentration in all electroporations was kept constant by the use of vector DNA. The cultures were then incubated in hormone depleted medium for 24 h. The medium was exchanged and 1 μg/ml of insulin was added for an additional 24 h. A, The cells were harvested and CAT enzyme activity was determined as described (30 ). The average percent acetylation per 100 μg protein in control and insulin-treated cultures was determined, and the insulin incubations were compared with control levels to determine the fold stimulation by insulin (fold-control). The average + SEM from three separate experiments done in duplicate is shown. B, Parallel cultures from the above electroporation were set up and incubated with insulin for 24 h. The cells were washed and harvested in lysis buffer. Cytoplasmic proteins were separated by SDS-PAGE and blotted to nitrocellulose. The top row shows expression of CSK tagged with GFP, whereas the bottom row shows the expression of RPTP tagged with HA.

    Inhibition of Src activity blocked the effect of RPTP

    The previous result strongly implied that a Src-related kinase mediated the inhibition of insulin-increased prolactin gene transcription in RPTP-expressing cells. It did not indicate which of the Src-related kinases might be responsible for this response because CSK can phosphorylate many Src-related kinases. Dominant negatives of the individual Src kinases might indicate which of the Src kinases mediated this response in GH4 cells. An expression vector for a kinase-inactive Src or kinase-inactive Fyn was transfected along with RPTP and the prolactin reporter plasmid to resolve this (Fig. 4). RPTP inhibited insulin increased prolactin-CAT expression more than 90% (0.001 > P). Kinase-inactive Src was partially able to block the effect of RPTP to inhibit insulin-increased prolactin gene transcription at 10 μg (0.01 > P) and completely blocked the effect of RPTP when 30 μg was used (0.001 > P). Dominant-negative Fyn did not effect RPTP inhibition of insulin-increased prolactin-CAT expression at 30 μg. This suggested that the response was specific to Src and that the activation of Src by RPTP explained the effect of RPTP on prolactin gene transcription. The expression of kinase inactive Src and kinase inactive Fyn did not affect the expression of RPTP protein (Fig. 4B). Again, insulin did not affect the amount of protein expressed by the transfected plasmids.

    FIG. 4. Kinase inactive Src blocks RPTP inhibition of insulin-increased prolactin gene transcription. GH4 cells were transfected with 10 μg of Prl (–173/+75) CAT and kinase-inactive Src or kinase inactive Fyn at concentrations shown in the figure. DNA concentration was kept constant in all electroporations by using vector DNA. The cultures were then incubated in hormone-depleted medium for 24 h. The medium was exchanged and 1 μg/ml of insulin was added for an additional 24 h. A, The cells were harvested and CAT enzyme activity was determined as described (30 ). The average percent acetylation per 100 μg protein in control and insulin-treated cultures was determined, and the insulin incubations were compared with control levels to determine the fold stimulation by insulin (fold-control). The average + SEM from four separate experiments done in duplicate is shown. B, Parallel cultures from the above electroporation were set up and incubated with insulin for 24 h. The cells were washed and harvested in lysis buffer. Cytoplasmic proteins were separated by SDS-PAGE and blotted to nitrocellulose. The top panel shows expression of myc-tagged Src DN (dominant negative) or Fyn DN, whereas the bottom panel shows the expression of RPTP tagged with HA.

    Activated Src blocked hormone-increased prolactin-CAT expression

    If RPTP activation of c-Src were the mechanism for the RPTP inhibition of insulin-increased prolactin-CAT expression, then we would expect that over expression of a constitutively active form of c-Src might mimic the effect of RPTP. Thus, expression of activated Src would block insulin-activated prolactin gene expression. Mutation of the regulatory tyrosine of c-Src, tyr527, to phenylalanine produces a constitutively active c-Src (Src Y/F). Src Y/F exerts a biphasic effect on prolactin gene expression. Electroporation of increasing amounts of expression plasmid resulted in a concomitant increase in the level of Src Y/F protein in GH4 cells (Fig. 5B) that was not affected by insulin treatment. Interestingly, low concentrations of Src Y/F stimulated both basal and insulin-increased prolactin gene expression in an additive fashion (SrcY/F 1 μg vs. control, 0.001 > P, and SrcY/F 1 μg + insulin vs. insulin, 0.001 > P, fig 5A) as was previously reported (44). Higher amounts of Src Y/F also increased basal prolactin gene expression, but the additive effect of insulin was lost. Finally, no increase in prolactin-CAT expression, basal or hormone activated, was observed in insulin-treated cells when 10 μg of Src Y/F was transfected into GH4 cells along with the prolactin-CAT reporter plasmid (Fig. 5A). This was not a general effect of Src Y/F to increase transcription. A similar experiment in which GH luciferase was used as the reporter showed that expression of Src Y/F had no effect on either basal or thyroid hormone-increased GH transcription at any level of Src Y/F (data not shown). Furthermore, the expression of Src Y/F did not affect the expression of the human insulin receptor cotransfected in these experiments or its phosphorylation in response to insulin (Fig. 5C).

    FIG. 5. Expression of Src Y/F blocks hormone-regulated prolactin-CAT expression. GH4 cells were transfected with 10 μg of Prl (–173/+75) CAT without or with Src Y/F at the concentrations indicated in the figure. Vector was added to all electroporations so that the amount of DNA was constant. The cultures were then incubated in hormone-depleted medium for 24 h. The medium was exchanged and 1 μg/ml insulin was added for an additional 24 h. A, The cells were harvested and CAT enzyme activity was determined as described (30 ). The average percent acetylation per 100 μg protein per hour in control and insulin-treated cultures was determined, and the insulin incubations were compared with control levels to determine the fold stimulation by insulin (fold-control). The average + SEM from three separate experiments done in duplicate is shown. B, Parallel cultures from the above electroporations were set up and incubated with insulin for 24 h. The cells were washed and harvested in lysis buffer. Cytoplasmic proteins were separated by SDS-PAGE and blotted to nitrocellulose. Antibody to GFP was used to visualize the GFP-SrcY/F present in each extract. C, GH4 cells were transfected with 5 μg of human insulin receptor and the amount of GFP-Src Y/F or vector shown. Cell lysates were prepared and an aliquot was separated on SDS-PAGE, blotted to nitrocellulose, and probed with an antibody to GFP (top row, GFP-Src y/f). Equal amounts of the remaining lysates were immunoprecipitated using an antihuman insulin receptor antibody, 83–14. The immunoprecipitates were divided into two aliquots that were resolved on parallel SDS-PAGE and blotted to nitrocellulose. One blot (middle row, phospho-HIR) was probed with a polyclonal antiphosphotyrosine (gift of J. Sap, New York University School of Medicine). The parallel blot was probed with the human insulin receptor-specific antibody 83–14 (bottom row, total HIR). The results are typical of three experiments.

    Expression of RPTP eliminated insulin-increased prolactin gene transcription, whereas there was no effect on transcription increased by epithelial growth factor (EGF) or cAMP (2). Src should affect EGF- and cAMP-activated prolactin gene transcription differently from insulin-increased transcription if RPTP acts through Src. GH4 cells were electroporated with increasing amounts of wild-type Src (Src WT) and treated with either insulin or cAMP to test this (Fig. 6A). Prolactin gene transcription increased by Src WT peaked with 3 μg of transfected plasmid and then decreased to baseline levels with 12 μg of plasmid. Insulin and cAMP both increased prolactin gene transcription and produced additive effects with Src WT, which were also maximal at 3 μg. Interestingly, insulin-increased prolactin gene expression was inhibited at the highest concentration of Src WT, whereas cAMP-activated transcription was not significantly different from levels induced without added Src WT (vector + cAMP vs. Src WT 12 μg + cAMP, P > 0.05). This implied that insulin signaling was more sensitive than signaling by cAMP to effects of Src WT overexpression/activation. Alternatively, cAMP might have protected the cells from the adverse effects of high levels of Src WT. EGF gave results similar to those obtained with cAMP (data not shown).

    FIG. 6. Src inhibition of insulin-increased prolactin gene transcription is specific. A, GH4 cells were transfected with 10 μg of Prl (–173) luciferase without or with the concentrations of Src WT indicated in the figure. Vector was added to all electroporations so that the amount of DNA was constant. The cultures were then incubated in hormone-depleted medium for 24 h. The medium was exchanged and 1 μg/ml insulin or 0.1 mM 8-chlorophenylthio-cAMP (Sigma) was added as indicated for an additional 24 h. The cells were harvested and luciferase activity was determined. The relative light units per 100 μg protein in control and insulin-treated cultures were determined, and the insulin incubations were compared with basal levels (untreated, vector cotransfected cells) to determine the fold stimulation by insulin or EGF (fold-basal). The average + SEM from three separate experiments done in duplicate is shown. B, Parallel cultures from the above electroporations were set up and incubated with insulin for 24 h. The cells were washed and harvested in lysis buffer. Cytoplasmic proteins were separated by SDS-PAGE and blotted to nitrocellulose. Antibody to GFP was used to visualize the GFP-Src WT present in each extract.

    PI 3-kinase mediated the effects of activated Src on prolactin gene expression

    Src has been linked to the activation of numerous signaling pathways. Src was shown to activate both the stress-activated protein kinase/c-Jun N-terminal kinase and p38 MAPK pathways in Src transformed fibroblasts (45), but Src-activated prolactin gene expression was not decreased by an inhibitor of p38, SB203580 (data not shown). The expression of an interfering MAPK kinase 4 that blocked the stress-activated protein kinase/c-Jun N-terminal kinase signaling pathway was not effective either (data not shown). Similarly, Src was shown to be upstream of MAPK kinase 1/ERK1/2 activation in MCF-7 cells (46). Finally, Src reportedly activated PI 3-kinase (47, 48). Electroporation of 3 μg of an expression vector for Src WT increased both basal and insulin-increased prolactin gene expression (0.001 > P, Fig. 7A). LY294002, a competitive inhibitor of PI 3-kinase, significantly reduced both the insulin- and Src WT-activated prolactin gene transcription (0.01 > P, Fig. 7A). Treatment of GH4 cells with the MAPK kinase 1/2 inhibitor PD98059 failed to block activation of prolactin gene expression by Src WT, although it significantly increased the effect of Src WT (0.001 > P) and Src WT + insulin (0.01 > P) (Fig. 7A). Neither LY294002 nor PD98059 affected the expression of Src WT in GH4 cells (Fig. 7B). This experiment showed that Src activation of the prolactin promoter is PI 3-kinase dependent and confirmed earlier studies that demonstrated PI 3-kinase dependence of insulin-increased prolactin gene transcription (2).

    FIG. 7. Insulin- and Src-increased prolactin gene transcription is dependent on PI 3-kinase. GH4 cells were transfected with 10 μg of Prl (–173/+75) CAT and 1 μg of GFP-Src. WT DNA concentration was kept constant in all electroporations by using vector DNA. The cultures were then incubated in hormone-depleted medium for 24 h. The medium was exchanged and 10 μM PD98059 or 10 μM Ly294002 were added to the indicated cultures for 2 h. Insulin was added for an additional 24 h at 1 μg/ml. The cells were harvested and CAT enzyme activity was determined as described (30 ). The average percent acetylation per 100 μg protein in control and insulin-treated cultures were determined, and the insulin incubations were compared with control levels to determine the fold stimulation by insulin (fold-control). The average ± SEM from three separate experiments done in duplicate is shown. B, Parallel cultures from the above electroporations were set up and incubated with insulin for 24 h. The cells were washed and harvested in lysis buffer. Cytoplasmic proteins were separated by SDS-PAGE and blotted to nitrocellulose. Antibody to GFP was used to visualize the GFP-Src WT present in each extract.

    Src activation of prolactin gene expression is Rho dependent

    The Src-related kinase Lck was shown to mediate some of its effects by activating Rho-related GTPases (49). This suggested that Src-increased Rho activation could mediate Src-increased prolactin gene transcription. Expression of Rho with an N19 mutation (RhoN19) was shown to block effects mediated by Rho (36). Expression of this plasmid caused a significant reduction in Src-activated prolactin gene transcription and prolactin gene transcription activated by both Src and insulin (0.01 > P, Fig. 8A) and suggested that Src activation of prolactin gene expression might be mediated through Src-activated Rho. Interestingly, expression of a constitutively active Rho (RhoL63) resulted in a total inhibition of Src-activated prolactin gene transcription (0.001 > P). This was similar to results seen with high levels of Src Y/F, and this would be expected if Src-increased prolactin gene transcription were Rho mediated. Western blot analysis demonstrated that expression of RhoN19 and RhoL63 did not inhibit Src activity by decreasing the amount of Src protein expressed in transfected cells (Fig. 8B).

    FIG. 8. Src-increased prolactin gene expression depended on Rho. GH4 cells were transfected with 10 μg of Prl (–173/+75) luciferase. The electroporations also contained 1 μg of CMV-Src WT alone or with 10 μg of CMV-RhoN19 or CMV-RhoL63 expression plasmids as indicated. DNA concentration was kept constant in all electroporations by using vector DNA. A, The cultures were then incubated in hormone-depleted medium for 24 h. The medium was exchanged and 1 μg/ml of insulin was added for an additional 24 h. The cells were harvested and luciferase enzyme activity was determined as directed by the manufacturer (Promega). The average relative light units per 100 μg protein in control and insulin-treated cultures was determined and the insulin incubations were compared with control levels to determine the fold-stimulation by insulin (fold-control). The average + SEM from three separate experiments done in duplicate is shown. B, Parallel cultures from the above electroporations were set up and incubated with insulin for 24 h. The cells were washed and harvested in lysis buffer. Cytoplasmic proteins were separated by SDS-PAGE and blotted to nitrocellulose. Antibody to GFP was used to visualize the GFP-Src WT (bottom row), whereas Myc-tagged RhoN19 and L63 were detected with antimyc monoclonal antibody (top row).

    Activated Rho, RhoL63, mimics the effect of Src

    A dose-response experiment was performed with activated RhoL63 to determine whether its effect reflected that of Src (Fig. 9). Electroporation of increasing amounts of the RhoL63 expression vector produced concomitant increases in RhoL63 detected by Western blotting (Fig. 9B). This produced a biphasic effect on prolactin transcription that was almost identical with the response to Src Y/F (Fig. 5). Low levels of RhoL63 caused an increase in both basal and insulin-increased prolactin gene transcription (Fig. 9A). At higher concentrations of expression vector, basal and insulin-increased transcription was both increased to the same extent by RhoL63. Finally, the highest amount of RhoL63 completely blocked insulin-increased prolactin gene expression.

    FIG. 9. RhoL63 mimics Src inhibition of insulin-increased prolactin gene expression. GH4 cells were transfected with 10 μg of Prl (–173/+75) luciferase. The electroporations also contained RhoL63 at the concentrations indicated. DNA concentration was kept constant in all electroporations by using vector DNA. A, The cultures were then incubated in hormone-depleted medium for 24 h. The medium was exchanged and 1 μg/ml of insulin was added for an additional 24 h. The cells were harvested and luciferase enzyme activity was determined as directed by the manufacturer (Promega). The average relative light units per 100 μg protein in control and insulin-treated cultures were determined and the insulin incubations were compared with control levels to determine the fold stimulation by insulin (fold-control). The average + SEM from three separate experiments done in duplicate is shown. B, Parallel cultures from the above electroporations were set up and incubated with insulin for 24 h. The cells were washed and harvested in lysis buffer. Cytoplasmic proteins were separated by SDS-PAGE and blotted to nitrocellulose. A monoclonal antimyc antibody was used to detect myc tagged RhoL63.

    PI 3-kinase is downstream of activated Rho

    Because RhoL63 mimicked the response of Src Y/F, it seemed possible that activation of PI 3-kinase mediated the response to RhoL63. Alternatively, some reports suggested that PI 3-kinase activated Rho and related GTPases (50). To discriminate between these alternatives, we treated cultures that had been transfected with RhoL63 with LY294002 (Fig. 10). As with Src (Fig. 7), the effect of RhoL63 was reduced more than 50% by inhibition of PI 3-kinase. RhoL63 protein levels were not affected by LY294002 (Fig. 10B). Furthermore, neither PD98059 nor SB2003580 reduced RhoL63-increased prolactin gene transcription (data not shown). This indicated that RhoL63-activated prolactin gene transcription was mediated through activation of PI 3-kinase, at least partially.

    FIG. 10. RhoL63 activated prolactin gene expression is PI 3-kinase dependent. GH4 cells were transfected with 10 μg of Prl (–173/+75) luciferase and 1 μg CMV-RhoL63. DNA concentration was kept constant in all electroporations by using vector DNA. The cultures were then incubated in hormone-depleted medium for 24 h. The medium was exchanged and 10 μM LY294002 were added to half of the cultures for 2 h. Insulin was added for an additional 24 h at 1 μg/ml. The cells were harvested and luciferase enzyme activity was determined as directed by the manufacturer (Promega). The average relative light units per 10 μg protein in control and insulin-treated cultures were determined, and the insulin incubations were compared with control levels to determine the fold stimulation by insulin (fold-control). The average ± SEM from three separate experiments done in duplicate is shown. B, Parallel cultures from the above electroporations were set up and incubated with insulin for 24 h. The cells were washed and harvested in lysis buffer. Cytoplasmic proteins were separated by SDS-PAGE and blotted to nitrocellulose. Antibody to myc tag was used to visualize the myc-RhoL63 present in each extract (indicated by the solid arrow). A strong band of nonspecific binding that remained constant was seen in all samples just below the RhoL63 signal (indicated by the black and white arrow).

    Activated Src and activated Rho alter the cytoskeleton of GH4 cells

    Src and Rho affect the cytoskeleton. One study suggested that Src localization to focal adhesions was dependent on GTP-bound Rho (51), and this might explain the similarity in the activation of prolactin gene transcription by Src Y/F and RhoL63. Further support for this mechanism would be provided if Src Y/F and RhoL63 induced a similar alteration of actin cytoskeleton of GH4 cells. GH4 cells were electroporated with Src Y/F or RhoL63. The electroporations also included a plasmid that expressed GFP-labeled PSF. PSF is an RNA splicing factor that is localized to the nucleus and serves as an indicator of transfected cells (52). Expression of PSF does not affect prolactin gene transcription but indicates the Src and Rho transfected cells. The cells were incubated for 48 h to allow any morphological changes to occur, and filamentous actin was labeled with Alexa Fluor 647 phalloidin (Molecular Probes). Control GH4 cells are rounded with an abundant cortical actin ring (see Fig. 12, top panel). GFP-PSF has no effect on the structure of the actin cytoskeleton because the cell with the green nucleus is indistinguishable from the untransfected cells that neighbor it. All of the GH4 cells electroporated with Src Y/F (Fig. 11 middle panel) or RhoL63 (Fig. 11, bottom panel) have an altered appearance that is never seen in control GH4 cells plated on uncoated plates. The Src transfected cells were spindle shaped with prominent focal adhesions at the poles. RhoL63-transfected cells characteristically had filopodia, but they were often more severely affected, having the spindle-shaped appearance more commonly seen in Src Y/F-transfected cells (Fig. 11, bottom panel, arrow). This may imply that Src Y/F acted more rapidly than RhoL63. Alternatively, Src may affect pathways in addition to the Rho pathway that result in a more pronounced morphological change. Because none of the control cells exhibit an extended phenotype, the changes seen in all of the transfected cells are highly significant.

    FIG. 12. A model for cooperative activation of prolactin gene transcription by integrin and insulin. Integrin ligation (on the left) activates RPTP that then recruits and activates Src. This activates Rho through a guanine nucleotide exchange factor that has not been identified but could be a Vav isoform (55 ). Insulin (on the right) activates the insulin receptor kinase to phosphorylate substrate molecules such as IRS-1/2. Phosphorylated IRS then activates PI 3-kinase and cytoskeletal rearrangement. This results in fusion of prolactin secretory vesicles with the plasma membrane and prolactin (PRL) release (64 ). PI 3-kinase was shown to be required for activation of prolactin gene expression (2 ) that is Elk-1 dependent (3 ). The PI 3-kinase-dependent kinase that mediates Elk-1 phosphorylation has not been identified. Inactive enzymes are illustrated in red and active enzymes are in green. Adapter proteins are yellow. Rho-GDP is red to indicate its inactive state, whereas Rho-GTP is green to indicate active Rho.

    FIG. 11. SrcY/F and RhoL63 alter the cytoskeleton and morphology of GH4 cells. GH4 cells were electroporated with 5 μg of GFP-PSF as a marker for transfected cells and with 10 μg of CMV-RhoL63, 10 μg of CMV-SrcY/F, or 10 μg of control vector. After 48 h, the cells were processed and treated with Alexa Fluor 647 phalloidin to stain F-actin. The cells were observed using a Zeiss 510 scanning confocal microscope. The signal from the Alexa Fluor 647 phalloidin is in the left column, the GFP is in the center column, whereas the combined image is in the right column. GH4 cell membranes had a slight green autofluorescence that is sometimes visible in the GFP and combined images. Representative cells are shown. The top row is control cells; the middle row is Src Y/F transfected cells; the bottom row is RhoL63 transfected cells.

    Discussion

    Normal physiological functioning of cells results from integration of complex sets of signals that indicate the condition of the cell, the immediate milieu, and the organism. It seems clear from these experiments that insulin, signaling the requirements of the organism, and integrin, signaling the cellular milieu, are interdependent activators of prolactin gene transcription. Figure 12 presents a model of insulin and integrin signaling that is supported by these data, our unpublished data, and the literature, although some of the steps need further resolution. Ligation of integrins activates Src through RPTP. Src activates Rho and then PI 3-kinase to alter the cortical actin network that results in prolactin release and stimulation of prolactin gene transcription. Similarly, insulin acts to activate PI 3-kinase in GH4 cells (3), perhaps through phosphorylation of IRS proteins (2). This then results in prolactin secretion (29) and increased prolactin gene transcription by activation of Elk-1 (3). The kinase that directly phosphorylates Elk-1 in response to insulin has not yet been identified.

    Extracellular matrix components profoundly affect prolactin production and secretion in GH cells and GH cells produce both extracellular matrix and integrins (25). Integrins were shown to directly affect insulin signaling by reducing IRS-1 (17), phosphorylation of the insulin receptor (14), and affecting insulin receptor internalization (15). Plating GH4 cells on fibronectin increased the level of active Src (Fig. 1, B and C). Others had shown that v-Src increased prolactin gene expression (53). Thus, it was not surprising that plating on fibronectin also increased basal- and insulin-increased prolactin gene transcription (Fig. 1A). Many studies (reviewed in Ref. 25) demonstrate that prolactin secretion and production by GH3B6 cells and by primary pituitary cells is dependent on extracellular matrix. These studies suggest that an appropriate physiological milieu is also necessary for maximum stimulation by insulin and perhaps other hormones due to cooperativity between the signaling pathways.

    Integrins were shown to activate Src through RPTP (10). Our previous studies with RPTP (2) did not identify any targets of RPTP in the insulin-signaling pathway and suggested that physiological pathways activated by RPTP affect insulin-increased prolactin gene transcription. Thus, it was possible that the effects of RPTP were due to overstimulation of Src, its normal physiological target. RPTP activates Src, and activated Src is required for the effects of RPTP prolactin gene expression (Figs. 2–4). The A23 cell line was derived from the GH4 cells by stable expression of RPTP. The 4-fold activation of Src kinase in these cells (Fig. 2, A–C) suggests that an RPTP activation of Src might explain RPTP inhibition of insulin-increased prolactin gene expression. Src is also activated in cells transiently expressing RPTP that are also unresponsive to insulin (Fig. 2D). CSK, the kinase that inactivates Src, reversed the RPTP-mediated inhibition of insulin-increased prolactin gene expression as would be expected if RPTP worked by activating Src (Fig. 3), and a kinase-inactive Src also restored insulin-increased prolactin gene expression to normal in RPTP-expressing cells (Fig. 4). Finally, high levels of activated Src block insulin activated prolactin gene transcription as is observed in RPTP-expressing cells (Fig. 5) How integrins activate RPTP remains to be explained. RPTP is constitutively phosphorylated on tyrosine 789 (54). It is possible that interaction of RPTP with liganded integrin unmasks this residue so that it can recruit Src (6).

    Src inhibition of insulin-increased prolactin transcription was surprising because others reported that v-Src activated prolactin gene expression in GH4 cells (53). Dose-response experiments, however, confirmed that a low level of wt Src (Fig. 6A) or Src Y/F (Fig. 5A) increased prolactin gene transcription, whereas expression of high levels of Src inhibited prolactin gene transcription (Figs. 5A and 6A). It was important to determine how this occurred. Src has a large number of well-defined substrates (e.g. cortactin, FAK, paxillin, p130CAS, and vinculin) that activate cytoskeletal organization and cell adhesion (9). This suggested that the effects of Src might be mediated by Rho-related GTPases. Src was shown to activate Vav that is a guanine nucleotide exchange factor for Rho-related GTPases (55). Src had also been shown to activate PI 3-kinase (47). Rac was shown to physically interact with PI 3-kinase (56), and PI 3-kinase activated a subset of Rho-dependent functions (50). These studies suggested that a sequence from Src to Rho/Rac/Cdc42 and then PI 3-kinase was possible. The experiment shown in Fig. 7 demonstrate that the effect of low amounts of Src is PI 3-kinase dependent. The effect of Src is at least partly dependent on Rho because a dominant-negative mutant of Rho reduces the effect of Src (Fig. 8). The experiments with low levels of activated RhoL63 (Fig. 9) suggested that activation of Src activated Rho and then PI 3-kinase to increase prolactin gene transcription. The effect of RhoL63 was also inhibited by LY294002 (Fig. 10).

    High levels of Src WT (Fig. 6A), SrcY/F (Fig. 5A), or RhoL63 (Figs. 8 and 9) inhibited insulin-increased prolactin gene transcription, and these effects were not mediated by PI 3-kinase. Both Src and RhoL63 profoundly altered the actin cytoskeleton and changed the morphology of GH4 cells (Fig. 11). Others have observed similar morphological changes when GH3B6 cells are plated on extracellular matrix in which activated Src would presumably be elevated (25). Cytochalasin D, jasplakinolide, and swinolide that are inhibitors of actin treadmilling also inhibited insulin-activated prolactin gene transcription (Vulin, A. I., and F. M. Stanley, manuscript in preparation). The concentration-dependent effects of these inhibitors were directly related to their concentration-dependent disruption of the cortical actin skeleton, and these were also relatively specific because the GH promoter was not equally affected. Others also demonstrated that remodeling of the actin cytoskeleton altered the expression of a subset of promoters (57). This was thought to result from either binding of a transcriptional activator to G actin, preventing the activator from entering the nucleus, or repression of transcription directly by G actin. The profound alteration of GH4 cell morphology and cytoskeletal organization caused by Src Y/F or RhoL63 indicated that the levels of G- and F-actin in these cells were altered. This might inhibit prolactin promoter activation. Increasing expression of ?-actin produced a biphasic effect on prolactin gene transcription that was similar to that of Src WT or RhoL63 (Vulin, A. I., and F. M. Stanley, manuscript in preparation).

    There are other explanations for the inhibition of prolactin gene transcription by high levels of activated Src. For example, reorganization of the actin cytoskeleton might disrupt the scaffolding that assembles signaling complexes into functional units. Alternately, Src might activate pathways such as the MAPK pathway that is activated by Src through phosphorylated FAK to inhibit insulin-signaling to prolactin gene expression. This might provide an explanation for the significant activation of the Src-wt and Src-wt + insulin by PD98059 in Fig. 7. Other explanations are also credible, but only future experiments are likely to discriminate among them. The interesting implication of these data is that part of the signal from insulin to the nucleus is dependent on cytoskeletal integrity. This could explain the failure of PI 3-kinase inhibitors to completely block insulin- or Src-increased prolactin gene expression. It also suggests that cytoskeletal change mediates insulin-induced prolactin secretion. Thus, a process that depletes prolactin stores, secretion, is coupled to prolactin gene transcription, a process that restores prolactin stores. Future experiments will use real-time RT-PCR of the endogenous prolactin mRNA to confirm the results obtained with overexpressed proteins because it is possible that the our experimental conditions (transfection of insulin receptor and other signaling components) could influence the normal physiological interactions of these molecules.

    Interestingly, the effects of Src on the morphology of GH4 cells are opposite those usually seen with Src. Src overexpression results in activation of p190 Rho-GAP with subsequent conversion of GTP to GDP-Rho and loss of stress fibers (58, 59). This causes a more rounded morphology. The elongation of the GH4 cells in response to Src or RhoL63 suggests that Src does not reduce GTP-bound Rho. This might occur if GH4 cells have low levels of p190 Rho-GAP or if GH4 cells contain a protein that associates with phosphorylated p190 Rho-GAP so that it does not function efficiently to increase the GTPase activity of Rho.

    The effect of activated Src/Rho on prolactin gene transcription suggests interesting parallels between insulin activation of Glut4 translocation in adipocytes and insulin activation of prolactin secretion and gene transcription in GH4 cells (Fig. 12). Actin reorganization and PI 3-kinase initiated kinase cascade were both required for translocation of Glut4 to the plasma membrane in 3T3 adipocytes (60). These have been considered two separate signaling pathways that converge in Glut4 translocation (61). Other studies, however, demonstrate that TC10, a Rho-related GTPase, activates PI 3-kinase and Glut 4 translocation in 3T3-L1 adipocytes (62) just as integrin ligation in GH4 cells activates Rho leading to activation of PI 3-kinase and cytoskeletal reorganization. Glut4 gene expression is also increased by insulin (63) and a similar sequence might link Glut4 translocation and gene transcription. This implies that a common pathway might mediate both the metabolic and mitogenic effects of insulin at least in this case.

    Acknowledgments

    We thank A. Hall (University College, London, UK); M. Mathur, J. Sap, and R. Schnieder (New York University School of Medicine); K. Siddle (University of Cambridge, Cambridge, UK); D. Shalloway (Cornell University, Ithaca, NY); and J. Whittaker (Haegdorn Institute, Copenhagen, Denmark) for plasmids and antibodies used in these studies.

    References

    Jacob KK, Whittaker J, Stanley F 2002 Insulin receptor tyrosine kinase activity and phosphorylation of tyrosines 1162 and 1163 are required for insulin-increased prolactin gene expression. Mol Cell Endocrinol 186:7–16

    Jacob KK, Sap J, Stanley FM 1998 Receptor-like protein-tyrosine phosphatase- specifically inhibits insulin-increased prolactin gene expression. J Biol Chem 273:4800–4809

    Jacob KK, Stanley FM 2001 Elk-1, C/EBPa, and Pit-1 confer an insulin responsive phenotype on prolactin promoter expression in Chinese hamster ovary cells and define factors required for insulin-increased transcription. J Biol Chem 276:24931–24936

    Walton KM, Dixon JE 1993 Protein tyrosine phosphatases. Annu Rev Biochem 62:101–120

    Le HT, Ponniah S, Pallen CJ 2004 Insulin signaling and glucose homeostasis in mice lacking protein tyrosine phosphatase-. Biochem Biophys Res Commun 314:321–329

    Su J, Muranjan M, Sap J 1999 Receptor protein tyrosine phosphatase- activates Src-family kinases and controls integrin-mediated responses in fibroblasts. Curr Biol 9:505–511

    Schlaepfer DD, Broome MA, Hunter T 1997 Fibronectin-stimulated signaling from a focal adhesion kinase-c-Src complex: involvement of the Grb2, p130cas, and Nck adaptor proteins. Mol Cell Biol 17:1702–1713

    Kaplan KB, Swedlow JR, Morgan DO, Varmus HE 1995 c-Src enhances the spreading of src-/-fibroblasts on fibronectin by a kinase-independent mechanism. Genes Dev 9:1505–1517

    Abram CL, Courtneidge SA 2000 Src family tyrosine kinases and growth factor signaling. Exp Cell Res 254:1–13

    von Wichert G, Jiang G, Kostic A, De Vos K, Sap J, Sheetz MP 2003 RPTP- acts as a transducer of mechanical force on v/?3-integrin-cytoskeleton linkages. J Cell Biol 161:143–153

    Schneller M, Vuori K, Ruoslahti E 1997 v?3 Integrin associates with activated insulin and PDGF? receptors and potentiates the biological activity of PDGF. EMBO J 16:5600–5607

    Vuori K, Ruoslahti E 1994 Association of insulin receptor substrate-1 with integrins. Science 266:1576–1578

    Bosco D, Meda P, Halban PA, Rouiller DG 2000 Importance of cell-matrix interactions in rat islet ?-cell secretion in vitro: role of 6?1 integrin. Diabetes 49:233–243

    El Annabi S, Gautier N, Baron V 2001 Focal adhesion kinase and Src mediate integrin regulation of insulin receptor phosphorylation. FEBS Lett 507:247–252

    Boura-Halfon S, Voliovitch H, Feinstein R, Paz K, Zick Y 2003 Extracellular matrix proteins modulate endocytosis of the insulin receptor. J Biol Chem 278:16397–16404

    Shaw LM 2001 Identification of insulin receptor substrate 1 (IRS-1) and IRS-2 as signaling intermediates in the 6?4 integrin-dependent activation of phosphoinositide 3-OH kinase and promotion of invasion. Mol Cell Biol 21:5082–5093

    Lebrun P, Baron V, Hauck CR, Schlaepfer DD, Van Obberghen E 2000 Cell adhesion and focal adhesion kinase regulate insulin receptor substrate-1 expression. J Biol Chem 275:38371–38377

    Yujiri T, Nawata R, Takahashi T Sato Y Tanizawa Y, Kitamura T, Oka Y 2003 MEK kinase 1 interacts with focal adhesion kinase and regulates insulin receptor substrate-1 expression. J Biol Chem 278:3846–3851

    Goel HL, Dey CS 2002 Insulin stimulates spreading of skeletal muscle cells involving the activation of focal adhesion kinase, phosphatidylinositol 3-kinase and extracellular signal regulated kinases. J Cell Physiol 193:187–198

    Kharitonenkov A, Chen Z, Sures I, Wang H, Schilling J, Ullrich A 1997 A family of proteins that inhibit signalling through tyrosine kinase receptors. Nature 386:181–186

    Inagaki K, Noguchi T, Matozaki T, Horikawa T, Fukunaga K, Tsuda M, Ichihashi M, Kasuga M 2000 Roles for the protein tyrosine phosphatase SHP-2 in cytoskeletal organization, cell adhesion and cell migration revealed by overexpression of a dominant negative mutant. Oncogene 19:75–84

    Lopez-Alemany R, Redondo JM, Nagamine Y, Munoz-Canoves P 2003 Plasminogen activator inhibitor type-1 inhibits insulin signaling by competing with v?3 integrin for vitronectin binding. Eur J Biochem 270:814–821

    Kamiya S, Kato R, Wakabayashi M, Wakabayashi M, Tohyama T, Enami I, Ueki M, Yajima H, Ishii T, Nakamura H, Katayama T, Takagi J, Fukai F 2002 Fibronectin peptides derived from two distinct regions stimulate adipocyte differentiation by preventing fibronectin matrix assembly. Biochemistry 41:3270–3277

    Slack-Davis JK, Eblen ST, Zecevic M, Boerner SA, Tarcsafalvi A, Diaz HB, Marshall MS, Weber MJ, Parsons JT, Catling AD 2003 PAK1 phosphorylation of MEK1 regulates fibronectin-stimulated MAPK activation. J Cell Biol 162:281–291

    de Carvalho DF, Silva KL, de Oliveria DA, Villa-Verde DM, Coelho HS, Silva LC, Nasciutti LE 2000 Characterization and distribution of extracellular matrix components and receptors in GH3B6 prolactin cells. Biol Cell 92:351–362

    Su J, Yang LT, Sap J 1996 Association between receptor protein-tyrosine phosphatase RPTP and the Grb2 adaptor. Dual Src homology (SH) 2/SH3 domain requirement and functional consequences. J Biol Chem 271:28086–28096

    Jiang Y-P, Wang H, D’Eustachio P, Musacchio JM, Schlessinger J, Sap J 1993 Cloning and characterization of R-PTP-, a new member of the receptor protein tyrosine phosphatase family with proteolytically cleaved cellular adhesion molecule-like region. Mol Cell Biol 13:2942–2951

    Jacob KK, Stanley FM 1994 The insulin and cAMP response elements of the prolactin gene are overlapping sequences. J Biol Chem 269:25515–25520

    Stanley FM 1988 Stimulation of prolactin gene expression by insulin. J Biol Chem 263:13444–13448

    Stanley FM 1992 An element in the prolactin promoter mediates the stimulation of insulin on transcription of the prolactin gene. J Biol Chem 267:16719–16726

    Su J, Batzer A, Sap J 1994 Receptor tyrosine phosphatase R-PTP- is tyrosine-phosphorylated and associated with the adaptor protein Grb2. J Biol Chem 269:18731–18734

    Zheng XM, Resnick RJ, Shalloway D 2000 A phosphotyrosine displacement mechanism for activation of Src by PTP. EMBO J 19:964–978

    Klein NP, Schneider RJ 1997 Activation of Src family kinases by hepatitis B virus HBx protein and coupled signaling to Ras. Mol Cell Biol 17:6427–6436

    Klein NP, Bouchard MJ, Wang LH, Kobarg C, Schneider RJ 1999 Src kinases involved in hepatitis B virus replication. EMBO J 18:5019–5027

    Lamarche N, Tapon N, Stowers L, Burbelo PD, Aspenstrom P, Bridges T, Chant J, Hall A 1996 Rac and Cdc42 induce actin polymerization and G1 cell cycle progression independently of p65PAK and the JNK/SAPK MAP kinase cascade. Cell 87:519–529

    Nobes CD, Hall A 1995 Rho, rac, and cdc42 GTPases regulate the assembly of multimolecular focal complexes associated with actin stress fibers, lamellipodia, and filopodia. Cell 81:53–62

    Olson MF, Ashworth A, Hall A 1995 An essential role for Rho, Rac, and Cdc42 GTPases in cell cycle progression through G1. Science 269:1270–1272

    Stanley FM 1989 Transcriptional regulation of prolactin gene expression by thyroid hormone—alternate suppression and stimulation in different GH cell lines. Mol Endocrinol 3:1627–1633

    Samuels H, Stanley F, Casanova J 1979 Depletion of L-3,5,3'-triiodothyronine and L-thyroxine in euthyroid calf serum for use in cell culture studies of the action of thyroid hormone. Endocrinology 105:80–85

    Castillo AI, Tolon RM, Aranda A 1998 Insulin-like growth factor-1 stimulates rat prolactin gene expression by a Ras, ETS and phosphatidylinositol 3-kinase dependent mechanism. Oncogene 16:1981–1991

    Soos M, Siddle K, Baron M, Heward JM, Luzio JP, Bekkatin J, Lennox ES 1986 Monoclonal antibodies reacting with multiple epitopes on the human insulin receptor. Biochem J 235:199–208

    Wilden PA, Kahn CR, Siddle K, White MF 1992 Insulin receptor kinase domain autophosphorylation regulates receptor enzymatic function. J Biol Chem 267:16660–16668

    Krook A, Whitehead JP, Dobson SP, Griffiths MR, Ouwens M, Baker C, Hayward AC, Sen SK, Massen JA, Siddle K, Tavare JM, O’Rahilly S 1997 Two naturally occurring insulin receptor tyrosine kinase domain mutants provide evidence that phosphoinositide 3-kinase activation alone is not sufficient for the mediation of insulin’s metabolic and mitogenic effects. J Biol Chem 272:30208–30214

    Pickett CA, Gutierrez-Hartmann A 1994 Ras mediates Src but not epidermal growth factor-receptor tyrosine kinase signaling pathways in GH4 neuroendocrine cells. Proc Natl Acad Sci USA 91:8612–8616

    Turkson J, Bowman T, Adnane J, Zhang Y, Djeu JY, Sekharam M, Frank DA, Holzman LB, Wu J, Sebti S, Jove R 1999 Requirement for Ras/Rac1-mediated p38 and c-Jun N-terminal kinase signaling in Stat3 transcriptional activity induced by the Src oncoprotein. Mol Cell Biol 19:7519–7528

    Nguyen DH, Webb DJ, Catling AD, Song Q, Dhakephalkar A, Weber MJ, Ravichandran KS, Gonias SL 2000 Urokinase-type plasminogen activator stimulates the Ras/extracellular signal-regulated kinase (ERK) signaling pathway and MCF-7 cell migration by a mechanism that requires focal adhesion kinase, Src, and Shc. Rapid dissociation of GRB2/Sps-Shc complex is associated with the transient phosphorylation of ERK in urokinase-treated cells. J Biol Chem 275:19382–19388

    Daulhac L, Kowalski-Chauvel A, Pradayrol L, Vaysse N, Seva C 1999 Src-family tyrosine kinases in activation of ERK-1 and p85/p110-phosphatidylinositol 3-kinase by G/CCKB receptors. J Biol Chem 274:20657–20663

    Daulhac L, Kowalski-Chauvel A, Pradayrol L, Vaysse N, Seva C 1999 Gastrin stimulates the formation of a p60Src/p125FAK complex upstream of the phosphatidylinositol 3-kinase signaling pathway. FEBS Lett 445:251–255

    Han J, Das B, Wei W, Van Aelst L, Mosteller RD, Khosravi-Far R, Westwick JK, Der CJ, Broek D 1997 Lck regulates Vav activation of members of the Rho family of GTPases. Mol Cell Biol 17:1346–1353

    Reif K, Nobes CD, Thomas G, Hall A, Cantrell DA 1996 Phosphatidylinositol 3-kinase signals activate a selective subset of Rac/Rho-dependent effector pathways. Curr Biol 6:1445–1455

    Fincham VJ, Unlu M, Brunton VG, Pitts JD, Wyke JA, Frame MC 1996 Translocation of Src kinase to the cell periphery is mediated by the actin cytoskeleton under the control of the Rho family of small G proteins. J Cell Biol 135:1551–1564

    Mathur M, Tucker PW, Samuels HH 2001 PSF is a novel corepressor that mediates its effect through Sin3A and the DNA binding domain of nuclear hormone receptors. Mol Cell Biol 21:2298–2311

    Pickett CA, Gutierrez-Hartmann A 1995 Epidermal growth factor and Ras regulate gene expression in GH4 pituitary cells by separate, antagonistic signal transduction pathways. Mol Cell Biol 15:6777–6784

    den Hertog J, Hunter T 1996 Tight association of GRB2 with receptor protein-tyrosine phosphatase is mediated by the SH2 and C-terminal SH3 domains. EMBO J 15:3016–3027

    Marignani PA, Carpenter CL 2001 Vav2 is required for cell spreading. J Cell Biol 154:177–186

    Bokoch GM, Vlahos CJ, Wang Y, Knaus UG, Traynor-Kaplan AE 1996 Rac GTPase interacts specifically with phosphatidylinositol 3-kinase. Biochem J 315:775–779

    Sotiropoulos A, Gineitis D, Copeland J, Treisman R 1999 Signal-regulated activation of serum response factor is mediated by changes in actin dynamics. Cell 98:159–169

    Chang J-H, Gill S, Settleman J, Parsons SJ 1995 c-Src regulates the simultaneous rearrangement of actin cytoskeleton, p190RhoGAP, and p120RasGAP following epidermal growth factor stimulation. J Cell Biol 130:355–368

    Fincham VJ, Chudleigh A, Frame MC 1999 Regulation of p190 Rho-GAP by v-Src is linked to cytoskeletal disruption during transformation. J Cell Sci 112:947–956

    Watson RT, Pessin JE 2001 Subcellular compartmentalization and trafficking of the insulin-responsive glucose transporter, GLUT4. Exp Cell Res 271:75–83

    Jiang ZY, Chawla A, Bose A, Way M, Czech MP 2002 A phosphatidylinositol 3-kinase-independent insulin signaling pathway to N-WASP/Arp2/3/F-actin required for GLUT4 glucose transporter recycling. J Biol Chem 277:509–515

    Maffucci T, Brancaccio A, Piccolo E, Stein RC, Falasca M 2003 Insulin induces phosphatidylinositol-3-phosphate formation through TC10 activation. EMBO J 22:4178–4189

    Armoni M, Quon MJ, Maor G, Avigad S, Shapiro DN, Harel C, Esposito D, Goshen Y, Yaniv I, Karnieli E 2002 PAX3/forkhead homolog in rhabdomyosarcoma oncoprotein activates glucose transporter 4 gene expression in vivo and in vitro. J Clin Endocrinol Metab 87:5312–5324

    Yoneda M, Nishizaki T, Tasaka K, Kurachi H, Miyake A, Murata Y 2000 Changes in actin network during calcium-induced exocytosis in permeabilized GH3 cells: calcium directly regulates F-actin disassembly. J Endocrinol 166:677–687(Anthony I. Vulin, Kirsten)