Background: Cell surface receptor for the epidermal growth factor (EGFR) and cytoplasmic tyrosine kinase c-Src co-operate in several cellular functions such as proliferation and apoptosis. Our previous studies have shown that ectopic expression of the adaptor protein p52shc or p66shc, but not p46shc, and EGF stimulation lead to the activation of c-Src that is accompanied by phosphorylation of signal transducers and activators of transcription (Stat) in A431 cells.
Results: Here, we show that by using A431 cells as a model system, expression of p52shc, or cell stimulation with EGF or H2O2 leads to phosphorylation of EGFR on Tyr 845 that is located to the activation segment of the catalytic domain. The phosphorylation of Tyr 845 can be inhibited by PP2, but not by AG1478, and is associated with Src activation and Stat 3/5 phosphorylation, but not with MAP (mitogen-activated protein) kinase phosphorylation. Phosphorylation of Stat 3/5 in response to p52shc expression, EGF or H2O2 could also be inhibited by introduction into cells of phospho-Tyr 845-specific antibody or by expression of dominant-negative version of c-Src. Co-incubation of purified c-Src and EGFR results in phosphorylation of Tyr 845 in vitro, indicating that c-Src can directly phosphorylate EGFR on Tyr 845.
Conclusion: These results indicate that multiple signals for c-Src activation can promote Stat 3/5 phosphorylations through Src-dependent phosphorylation of EGFR on Tyr 845.
The epidermal growth factor receptor (EGFR), an 1186-amino acid residue transmembrane glycoprotein, is the prototypical member of receptor tyrosine kinases that are activated by EGF (Ullrich & Schlessinger 1990; Jorissen et al. 2003). Upon EGF binding to the extracellular domain, EGFR undergoes dimerization and enzymatic activation followed by autophosphorylation of multiple tyrosine residues in a cytoplasmic carboxyl-terminal region of the molecule. The autophosphorylated tyrosine residues serve as docking sites for cytosolic signalling molecules containing Src homology (SH) 2 domains and phosphotyrosine-binding (PTB) domains (van der Geer et al. 1994; Pawson 1995). c-Src, a 536-amino acid residue membrane-associated protein, is a prototypical member of non-receptor tyrosine kinases that work as a co-transducer of transmembrane signals eliciting from various cell surface receptors, including EGFR (Thomas & Brugge 1997; Abram & Courtneidge 2000; Parsons & Parsons 1997). Tyrosine kinase activity of c-Src is regulated by an intramolecular interaction that involves SH3 and SH2 domains, an SH2-kinase domain linker, an activation segment in the kinase domain, and a phosphotyrosine residue (Tyr-527) in the carboxyl terminus (Sicheri & Kuriyan 1997). Dephosphorylation of phospho-Tyr-527 or protein-protein interaction that relieves the inactive conformation will activate c-Src.
Both EGFR and c-Src have their oncogenic versions from the viral origins, v-erbB and v-Src, respectively, that can cause malignant cell transformation by themselves (Bishop 1983; Frame 2002). On the other hand, several lines of evidence indicate that EGFR and c-Src co-operate in the processes of normal cell growth and malignant cell transformation. Namely, cells expressing high level of both EGFR and c-Src exhibit higher potencies in their growth and malignant transformation than cells with high expression of either EGFR or c-Src alone (Luttrell et al. 1988; Maa et al. 1995; Tice et al. 1999). More recent studies have shown that in certain kinds of cells, EGFR and c-Src co-operate in promoting anti-apoptotic or pro-apoptotic signalling (Kitagawa et al. 2002; Karni & Levitzki 2000; Sato et al. 2002). Together, these findings raise the question that ‘what kinds of signalling events are working for the co-operation of EGFR and c-Src? The objectives of this study were to evaluate whether several types of cell stimuli activating EGFR and c-Src are accompanied by phosphorylation of signal transducers and activators of transcription (Stat), and to determine whether this involves Src-dependent phosphorylation of EGFR. We have previously shown that in human epidermoid carcinoma A431 cells, application of EGF or over-expression of the SH2/PTB-containing adaptor protein p52shc or p66shc, but not p46shc, leads to c-Src activation (Sato et al. 2002). The EGF- or Shc-induced Src activation involves a direct interaction of Shc and the activation segment of c-Src. We have also shown that Stat, but not MAP kinase, is a downstream target of c-Src activated by EGF or Shc. In this paper, we present evidence that in addition to EGF and Shc, hydrogen peroxide induces phosphorylation of EGFR, Src, and Stat proteins, and that Src-dependent phosphorylation of the EGFR Tyr-845, a tyrosine residue locating in the activation segment, plays a critical role in the Stat phosphorylation in A431 cells.
Direct phosphorylation of EGFR on Tyr-845 by c-Src
To determine whether c-Src phosphorylates directly EGFR on Tyr-845, we performed an in vitro kinase assay using purified c-Src and EGFR. Figure 1 shows that coincubation of EGFR and c-Src (lanes 3 and 4, pY845), but neither EGFR nor c-Src alone (lanes 1, 2, 5 and 6, pY845), promoted phosphorylation of EGFR Tyr-845. It is worth noting that the Tyr-845 phosphorylation was augmented in the presence of EGF, suggesting that EGF-induced conformational change has a positive effect on the Src-dependent Tyr-845 phosphorylation. On the other hand, phosphorylation of EGFR Tyr-1068 was solely EGF-dependent (see lanes 2 and 4, pY1068). PP2, but not AG1478, could inhibit Tyr-845 phosphorylation (lanes 7-12, pY845), indicating again that EGFR/kinase activity is not required for Tyr-845 phosphorylation. Consistent with its EGF dependency, Tyr-1068 phosphorylation was inhibited by AG1478, but not by PP2 (lanes 7-12, pY1068).
Phosphorylation of EGFR on Tyr-845 in A431 cells stimulated with EGF, p52shc expression and H2O2
We next determined whether Src activation leads to phosphorylation of EGFR on Tyr-845. A431 cells were stimulated with 10 nm of EGF, ectopic expression of green fluorescent protein (GFP)-p52shc or 1 mm of H2O2. All treatments induced Tyr-845 phosphorylation as determined by immunoblotting with phospho-specific antibody of immunoprecipitated EGFR (pY845, Fig. 2A). The Src-specific inhibitor PP2, but not the EGFR inhibitor AG1478, effectively inhibited the phosphorylation. EGF, GFP-p52shc and H2O2 also promoted phosphorylation of other tyrosine residues in EGFR such as pY992 (not shown), pY1068 (Fig. 2A), and pY1173 (not shown). As opposed to the Tyr-845 phosphorylation, however, these phosphorylation events were only sensitive to AG1478 (Fig. 2A, pY1068). In H2O2-treated cells, phosphorylation of Tyr-845 was partly inhibited by AG1478 (Fig. 2A, lane 9). This could be due to that EGFR catalytic activity, which would be blocked by AG1478 and could affect the conformation of the molecule (see also Fig. 1A, lanes 3 and 4, pY845), is partly involved in Src-catalysed phosphorylation of EGFR Tyr-845.
To evaluate the specificity of GFP-p52shc function, we analysed EGFR phosphorylation in A431 cells expressing either GFP or GFP-p46shc, both of which have been shown to be inert toward Src activity (Sato et al. 2002). Neither expression of GFP nor GFP-p46shc by itself promoted Tyr-845/Tyr-1068 phosphorylation (Fig. 2B), confirming that the GFP-p52shc function is specific. We also observed EGF-dependent and PP2-sensitive Tyr-845 phosphorylation as well as EGF-dependent and AG1478-sensitive Tyr-1068 phosphorylation in cells expressing GFP or GFP-p46shc (Fig. 2B), as we observed it in intact A431 cells (Fig. 2A).
H2O2 promotes Phospho-Tyr-527-independent activation of c-Src in A431 cells
The results obtained in Fig. 2 are consistent with the idea that the activated c-Src is responsible for Tyr-845 phosphorylation in A431 cells stimulated with EGF or p52shc. On the other hand, H2O2-induced c-Src activation has not been well characterized. In A431 cells, EGF stimulation or GFP-p52shc expression leads to activation of c-Src in a Tyr-527 dephosphorylation-independent manner (Sato et al. 2002). Therefore we examined tyrosine phosphorylation state of c-Src in H2O2-treated A431 cells. H2O2 promoted phosphorylation of c-Src Tyr-416 (Fig. 3A). The Tyr-416 phosphorylation was effectively blocked by PP2. In contrast, c-Src Tyr-527 was stably phosphorylated irrespective of the presence of H2O2 and PP2, suggesting that H2O2-induced c-Src activation occurs without Tyr-527 dephosphorylation. In addition, H2O2-induced c-Src activation was not accompanied by an increase in the formation of the Src-Shc complex (not shown), indicating that the c-Src activation is independent of the Shc action.
Src-dependent up-regulation of Stat-p21waf1 pathway in H2O2-treated A431 cells
We next examined the effect of H2O2 on the cellular signalling pathway, because in A431 cells stimulated with EGF or p52shc, phosphorylation of Stat proteins, but not that of MAP kinase pathway, has been shown to be up-regulated in a Src-dependent manner (Sato et al. 2002). As shown in Fig. 3B, H2O2 induced tyrosine phosphorylation of Stat3, Stat5b and MAP kinase, but not Stat1. H2O2 also induced PP2-sensitive protein expression of p21waf1. PP2 effectively blocked the H2O2-induced phosphorylation of Stat3 and Stat5b. However, PP2 did not inhibit the MAP kinase. These results indicate that, as in A431 cells stimulated with EGF or p52shc, Src-dependent up-regulation of Stat-p21waf1 pathway, but not MAP kinase pathway, is taking place in H2O2-treated A431 cells.
Inhibition of phosphorylation of EGFR on Tyr-845 by kinase-negative Src in A431 cells
We next employed the kinase-negative c-Src construct (kn-Src) to establish the importance of Src activity in cell signalling pathway. Ectopic expression of kn-Src inhibited phosphorylation of EGFR Tyr-845, c-Src Tyr-416, and Stat3 in A431 cells stimulated with EGF, p52shc or H2O2 in a dose-dependent manner (Fig. 4). The effect of kn-Src was specific for EGFR Tyr-845 phosphorylation, since under the same conditions, phosphorylation of other tyrosine residues of EGFR such as Tyr-1068 was not affected (Fig. 4).
An antibody specific to Phospho-Tyr-845 blocks phosphorylation of Stat3 in A431 cells
Finally, to confirm further the relationship between Src-dependent phosphorylation of EGFR Tyr-845 and phosphorylation of Stat3/5b, we wanted to determine whether specific blockage of EGFR pY845 affects Stat phosphorylation. To this end, we performed transfection of A431 cells with anti-pY845 IgG. As shown in Fig. 5, anti-pY845 IgG-transfected cells showed a marked decrease in cell stimulation-dependent phosphorylation of Stat3, whereas neither anti-pY1068 IgG nor control IgG showed such effect (Fig. 5, pStat3). A similar reduction of cell stimulation-dependent expression of p21waf1 was seen in cells transfected with anti-pY845 IgG (Fig. 5, p21waf1). On the other hand, MAP kinase phosphorylation was not affected by any transfection conditions (Fig. 5, pMAPK). Transfection of control IgG, anti-pY845 IgG, or anti-pY1068 IgG did not affect cell stimulation-induced phosphorylation of EGFR (Fig. 5, pY845, pY1068) and c-Src (Fig. 5, pY416). IgG could be recovered from the extracts of transfected cells by using protein A-Sepharose (Fig. 5, IgG, lanes 3-11), whereas IgG could not be recovered from non-transfected cells (Fig. 5, IgG, lanes 1 and 2), indicating that successful incorporation of IgG occurred in this transfection system.
The results presented here provide a possible mechanism whereby c-Src and EGFR co-operate in phosphorylating Stat proteins in A431 cells. Namely, it involves activation of c-Src by Shc association (in case of EGF and over-expression of p52shc) or by direct chemical modification mechanism in case of H2O2 (Akhand et al. 1999), phosphorylation of EGFR Tyr 845 by the activated c-Src, and phosphorylation of Stat by c-Src and/or other tyrosine kinases activated by c-Src (see Summary figure). Specifically, the role for c-Src activity and phosphorylated Tyr-845 of EGFR in Stat phosphorylation was established with the aid of pharmacological inhibitors, kinase-negative construct of c-Src, and site-specific anti-phosphotyrosine antibodies.
We employed H2O2 as an alternative method to stimulate c-Src in A431 cells. Although its physiological relevance is currently unclear, we have shown that the H2O2-induced activation of Src and phosphorylation of EGFR Y845 resulted in induction of protein expression of p21waf1 (Figs 2 and 5). Therefore, it is attractive to surmise that not only EGF but redox signal, if any, could modulate the cell cycle and survival response in A431 cells in a manner that depends on Src/EGFR action.
In several cell systems, Tyr-845 of EGFR has been reported to be the Src-dependent phosphorylation site, although functional importance of the phosphorylation has not been demonstrated. They include transformed cells such as A431 cells (Sato et al. 1995a; this study) and MDA468 breast cancer cells (Biscardi et al. 1999), C3H10T1/2 mouse embryonic fibroblasts engineered to over-express EGFR and c-Src (Luttrell et al. 1988; Maa et al. 1995; Biscardi et al. 1999), the integrin-dependent adherent cells ECV304 (Moro et al. 2002), and B82L fibroblasts exposed to Zn2+ ions (Wu et al. 2002). More recent study by Kloth et al. (2003) has demonstrated that Stat5b is tyrosine-phosphorylated in response to EGF in EGFR-over-expressing cells but that this tyrosine phosphorylation event is dependent, in part, on Src activity and the presence of the Src-mediated phosphorylation of EGFR Tyr-845. Our present results are consistent with their findings in that Src-catalysed phosphorylation of EGFR Tyr-845 is required for subsequent up-regulation of the Stat pathway.
Tyr-845 of EGFR locates in the activation segment of the kinase domain. In general, activation segments of protein kinases contain phosphorylation site(s) that are important for regulation of catalytic activity (Fukami et al. 1999). Many of them are autophosphorylation sites; for example, MAP kinase and c-Src. In the case of EGFR, however, it has been reported that Tyr-845 is not required for the regulation of catalytic activity of EGFR, but required for the DNA synthesis induced by EGF, serum, and lysophosphatidic acid (Biscardi et al. 1999; Tice et al. 1999). Under these circumstances, our present studies shed light on the role of Src phosphorylation of EGFR Tyr-845 in cellular signalling accompanied by Stat phosphorylation. Specifically, data obtained suggest that Src phosphorylation of EGFR Tyr-845 induces up-regulation of Stat-p21waf1 pathway, which would trigger cell cycle arrest and apoptosis in A431 cells (Cao et al. 2000; Chin et al. 1997). It is well established that Stat proteins can be activated by a number of interferons, cytokines and growth factors, including EGF, and are involved in a variety of cellular functions including mitogenesis, differentiation, apoptosis and oncogenesis (Leonard & O’Shea 1998). Thus, our present study has demonstrated a functional linkage between Src/EGFR pathway and Stat/waf1 pathway in apoptotic signalling involving EGF-dependent Stat activation. Our results are also consistent with a more recent report showing that sequence specific peptide aptamers, interacting with the intracellular domain of the EGF receptor, interfere with phosphorylation of EGFR Tyr-845 and Stat3 activation (Buerger et al. 2003).
Regarding to the direct interaction between EGFR and Stat proteins, we have conducted immunoprecipitation analysis to detect interaction between Stat3/5 and EGFR, however, we could not detect any interaction in the condition that has allowed us to detect Src-EGFR interaction (data not shown). This might be due to that (i) their interaction is too transient and/or weak to catch, and/or (ii) antibodies used in this trial were not appropriate for collecting the Stat-EGFR complex. We favour the first possibility because Stat protein, when they are tyrosine-phosphorylated and stimulated in terms of the transcriptional activity, should translocate from the membranes to the nucleus. One more possibility is that signalling event(s) other than direct or indirect interaction between EGFR and Stat proteins operate for induction of p21waf1. Therefore, further study will be directed to dissect the molecular mechanism of signal transduction involving phosphorylated EGFR Tyr-845 and Stat3 activation, and its physiological consequences.
Mouse EGF was purchased from Wako Pure Chemicals (Osaka, Japan). Plasmids for bacterial or mammalian expression of green fluorescent protein (GFP) or glutathione-S-transferase (GST) fusion forms of mouse p46shc and p52shc were constructed as described (Sato et al. 1997, 2002). DNA encoding a kinase-negative version of the chicken c-src gene, in which codons for Lys-295 are substituted by those for Met, was generated by polymerase chain reaction and subcloned into p3xFLAG-CMV-14 vector (Sigma). H2O2 was obtained from Santoku Chemical Industries (Tokyo, Japan). The Src-specific inhibitor PP2 and the EGFR-specific inhibitor AG1478 were from Calbiochem. Phospho-specific anti-EGFR antibodies (pY845, pY992, pY1045, pY1068, pY1173) were from Cell Signalling Technology and Upstate Biotechnology. Anti-EGFR antibody was raised against a synthetic peptide covering residues 1155-1175 of EGFR (Sato et al. 1995a). Anti-Src (mAb327), anti-phospho-Tyr-416, and anti-phospho-Tyr-527 were from Oncogene Research Products. Anti-phosphotyrosine antibody PY99 was from Santa Cruz Biotechnology. Anti-Shc antibody was from Upstate Biotechnology. Phospho-specific antibodies to Stat1 (pY701), Stat3 (pY705), and Stat5b (pY694) were from Cell Signalling Technology. Anti-phosphorylated MAP (mitogen-activated protein) kinase antibody was from New England Biolabs.
Cell preparations, extractions and biochemical analyses
A431 cells were grown in Dulbecco's Modified Eagle's Medium supplemented with 10% foetal calf serum. Cell transfection with plasmid DNAs expressing GFP-fusion Shc proteins was done by using Effectene reagents (Qiagen, Germany) as previously described (Sato et al. 2000, 2002). Cell transfection with IgGs was performed using Profect reagents (Targeting Systems) according to the manufacturer's instruction. All cell preparations were serum-starved for 24 h, and then stimulated with or without of either 10 nm EGF or 1 mm H2O2 for 5 min. Cells were extracted with buffer containing 1% Triton X-100, 20 mm Tris-HCl, pH 7.5, 1 mm EDTA, 1 mm EGTA, 10 mmβ-mercaptoethanol, 1 mm sodium orthovanadate, 10 µg/mL leupeptin, and 20 µm (p-amidinophenyl)methanesulphonyl fluoride. Specified amounts of the Triton X-100-solubilized proteins, as determined by the dye-binding assay using a Bio-Rad reagent, were analysed for protein phosphorylation and molecular interaction by immunoprecipitation and/or immunoblotting as described (Sato et al. 2002).
In vitro kinase assay
Phosphorylation of EGFR was assessed by in vitro kinase assay in the presence of 10 mm MgCl2, 3 mm MnCl2, and 1 mm ATP. EGFR and c-Src were purified from A431 cells and bovine brain, respectively, as described (Fukami et al. 1993; Sato et al. 1995b). Kinase reaction was initiated by addition of MgCl2/MnCl2/ATP to EGFR (100 ng protein), c-Src (10 ng protein), or mixture of both in the absence of the specified amounts of EGF, PP2 and AG1478. The reaction was allowed to proceed for 20 min at 30 °C and terminated by the addition of SDS-containing buffer (Laemmli 1970). The SDS-treated samples were separated by SDS-polyacrylamide gel electrophoresis (SDS-PAGE) (Laemmli 1970) and analysed as indicated in the text.
This work was supported by grants-in-aid for science research from the Ministry of Education, Culture, Sports, Science and Technology of Japan, and by a grant of the Japan Society for the Promotion of Science for exchange under the US-Japan Cooperative Cancer Research Program.