Role of phosphoinositide 3-kinase in activation of ras and mitogen-activated protein kinase by epidermal growth factor

Abstract

The paradigm for activation of Ras and extracellular signal-regulated kinase (ERK)/mitogen-activated protein (MAP) kinase by extracellular stimuli via tyrosine kinases, Shc, Grb2, and Sos does not encompass an obvious role for phosphoinositide (PI) 3-kinase, and yet inhibitors of this lipid kinase family have been shown to block the ERK/MAP kinase signalling pathway under certain circumstances. Here we show that in COS cells activation of both endogenous ERK2 and Ras by low, but not high, concentrations of epidermal growth factor (EGF) is suppressed by PI 3-kinase inhibitors; since Ras activation is less susceptible than ERK2 activation, PI 3-kinase-sensitive events may occur both upstream of Ras and between Ras and ERK2. However, strong elevation of PI 3-kinase lipid product levels by expression of membrane-targeted p110alpha is by itself never sufficient to activate Ras or ERK2. PI 3-kinase inhibition does not affect EGF-induced receptor autophosphorylation or adapter protein phosphorylation or complex formation. The concentrations of EGF for which PI 3-kinase inhibitors block Ras activation induce formation of Shc-Grb2 complexes but not detectable EGF receptor phosphorylation and do not activate PI 3-kinase. The activation of Ras by low, but mitogenic, concentrations of EGF is therefore dependent on basal, rather than stimulated, PI 3-kinase activity; the inhibitory effects of LY294002 and wortmannin are due to their ability to reduce the activity of PI 3-kinase to below the level in a quiescent cell and reflect a permissive rather than an upstream regulatory role for PI 3-kinase in Ras activation in this system.

FIG. 1

PI 3-kinase inhibitors block ERK2 phosphorylation and activation induced by low concentrations of EGF. (A) Serum-starved COS-7 cells were either left untreated or preincubated with LY294002 (LY; 20 μM) or wortmannin (WM; 100 nM) for 5 min. The cells were then stimulated with increasing concentrations of EGF for 5 min, and 50 μg of lysates was analyzed for ERK2 phosphorylation by mobility shift assays (upper panel) or by using anti-phospho-ERK antibodies (lower panel). (B) Cells, transfected with Myc-ERK2 in the absence or presence of dominant-negative PI 3-kinase Δp85, were serum starved, and some cells were also preincubated with LY294002 (20 μM) for 5 min prior to stimulation with increasing concentrations of EGF for 5 min. The activity of immunoprecipitated ERK2 was analyzed in in vitro kinase assays with MBP as substrate. Data are expressed as fold stimulation, where basal ERK2 activity is defined as 1.0. Representative experiments are shown. (C) Cells were treated as described for panel B, except that dominant-negative N17 Ras was used instead of Δp85. Values are means ± standard errors from at least two separate experiments.

FIG. 2

The PI 3-kinase inhibitor LY294002 blocks Ras activation induced by low concentrations of EGF. (A) Serum-starved COS-7 cells were either left untreated or preincubated with LY294002 (LY; 20 μM) or wortmannin (WM; 100 nM) for 5 min and then stimulated with increasing concentrations of EGF for 5 min. Active GTP-bound Ras was extracted from lysates with the GST-Raf-Ras-binding domain (RBD) coupled to glutathione agarose and analyzed by immunoblotting with pan-Ras antibodies. (B) The effect of 20 μM LY294002 on endogenous Ras GTP-loading and endogenous ERK2 activation expressed as percent inhibition of activation seen at different concentrations of EGF. (C) Ras activation is not affected by matrix composition. COS-7 cells were attached to uncoated tissue culture dishes (−) or to dishes that had been precoated with either fibronectin (FN) or collagen (CO). Serum-starved cells were then either left untreated or preincubated for 5 min with LY294002 (LY; 20 μM) before being challenged with 0.05 ng of EGF per ml for 5 min. The amount of GTP-bound Ras in lysates was determined as described for panel A. Representative experiments are shown.

FIG. 3

An activated form of PI 3-kinase, p110-CAAX, is unable to induce activation of ERK2 or Ras. (A) COS-7 cells were transfected with either Myc-ERK2 alone or in combination with p110-CAAX or Drosophila Sos (dSos). The activity of immunoprecipitated ERK2 was analyzed as described for Fig. 1B. Data are expressed as fold stimulation over basal ERK2 activity. Values are means ± standard errors from at least two separate experiments. (B) Cells, transfected with either Ras alone or in combination with p110-CAAX or v-Src, were incubated for 12 h in medium containing low serum and [³²P]orthophosphate. Ras proteins were immunoprecipitated, and bound nucleotides were eluted, resolved by thin-layer chromatography, and quantitated. The amount of Ras-GTP is expressed as percentage of total guanyl nucleotides bound to Ras. Values are means ± standard errors from at least three separate experiments. (C) Cells were transfected as described for panel B and serum starved, and the amount of GTP-bound Ras in lysates was determined as described for Fig. 2A (upper panel). Fifty micrograms of lysates (1/20 of the total amount for each point) was also analyzed for expression of transfected Ras (lower panel). A representative experiment is shown. RBD, Ras-binding domain. (D) Cells were transfected with either HA-PKB alone or in combination with p110-CAAX or v-Src. Some HA-PKB-transfected cells were stimulated with 20 ng of EGF per ml for 5 min. The activity of immunoprecipitated PKB was analyzed in in vitro kinase assays with histone H2B as substrate. Data are expressed as fold stimulation over basal PKB activity. Values are means ± standard errors from at least two separate experiments.

FIG. 4

The PI 3-kinase inhibitor LY294002 does not block EGF-induced EGF receptor (EGFR) autophosphorylation, Shc phosphorylation, or association between Shc and Grb2. Serum-starved COS-7 cells, some preincubated for 5 min with LY294002 (LY; 20 μM), were stimulated with increasing concentrations of EGF for 5 min. (A) Fifty micrograms of lysate was analyzed for EGF receptor autophosphorylation by immunoblotting with phosphotyrosine antibodies. Numbers are expressed as percentages of receptor phosphorylation seen at 5 ng of EGF per ml. (B) EGF-induced association between EGF receptors and Grb2 or Shc was analyzed by anti-Grb2 or anti-Shc immunoprecipitation followed by immunoblotting with EGF receptor antibodies. (C) EGF-induced Shc phosphorylation and association between Shc and Grb2 were analyzed by anti-Shc immunoprecipitation followed by immunoblotting with phosphotyrosine and Grb2 antibodies, respectively. (D) Association between Grb2 and Sos was analyzed by anti-Sos immunoprecipitation followed by immunoblotting with Grb2 antibodies. Representative experiments are shown.

FIG. 5

Low concentrations of EGF do not induce phosphorylation on serine 473 in PKB. Serum-starved COS-7 cells were left untreated or preincubated for 5 min with LY294002 (LY; 20 μM) before being challenged with increasing concentrations of EGF for 5 min. Fifty micrograms of lysates was analyzed for PKB phosphorylation by immunoblotting with anti-phospho-PKB antibodies. A representative experiment is shown.

FIG. 6

Low concentrations of LPA induce LY294002-sensitive Ras activation and ERK2 phosphorylation and phosphorylation on serine 473 in PKB. Serum-starved COS-7 cells were either left untreated or preincubated with LY294002 (LY; 20 μM) for 5 min and then stimulated with increasing concentrations of LPA for 5 min. (A) The amount of active Ras in lysates was determined as described for Fig. 2A. RBD, Ras-binding domain. (B) LPA-induced ERK2 phosphorylation analyzed by mobility shift assay. (C) LPA-induced PKB phosphorylation, analyzed as described for Fig. 5. Representative experiments are shown.

FIG. 7

Model for the role of PI 3-kinase in the activation of Ras by EGF. At high concentrations of EGF, stable complexes of Sos form with Grb2, Shc, and the EGF receptor; their activation of Ras and subsequently ERK is not influenced by PI 3-kinase. At low concentrations of EGF, activation of Ras is by Shc-Grb2-Sos complexes which are not EGF receptor associated. Low concentrations of EGF do not stimulate PI 3-kinase activity, but their activation of Ras is dependent on basal PI 3-kinase activity. Another point at which PI 3-kinase may influence ERK activation by low or intermediate concentrations of EGF lies at the level of MEK activation. Ras-independent pathways also link EGF receptor to ERK activation: these may be sensitive to PI 3-kinase inhibition. See the text for details. Thick arrows indicate stable complex formation; thin arrows indicate transient interactions involved in signal transduction or indirect effects.