Src-dependent Phosphorylation of the Epidermal Growth Factor Receptor on Tyrosine 845 Is Required for Zinc-induced Ras Activation*

Previous studies have shown that exposure of cells to Zn 2 (cid:1) ions induces Ras and MAPK activation through the EGF receptor (EGFR). To further determine the role of EGFR in Zn 2 (cid:1) -induced signaling, mouse B82L fibroblasts expressing no detectable EGFR protein (B82L-par), wild type EGFR (B82L-wt), kinase-deficient EGFR (B82L-K721M), or COOH-truncated EGFR (B82L-c’958) were tested. Exposure to Zn 2 (cid:1) induced Ras activity in B82L-wt, B82L-K721M, and B82L-c’958 but not in B82L-par cells, indicating that the tyrosine kinase domain and the auto-phosphorylation sites of the EGFR were not required for Zn 2 (cid:1) -induced Ras activation. Zn 2 (cid:1) induced Src activation in all B82L CA). Monoclonal anti-EGFR antibody Ab14 was Neomarkers CA). PP1 BIOMOL Inc. PA). LipofectAMINE transfection reagents were purchased from Invitrogen. Cell Culture and Exposure— The B82L parental fibroblasts (B82L-par) and B82L fibroblasts expressing the human wild type EGFR (B82L-wt), kinase-defective EGFR (B82L-K721M), and the COOH-ter-minally truncated EGFR at Tyr-958 (B82L-c ’ 958) have been described previously 37). B82L-par fibroblasts were maintained in Dulbecco ’ s modified Eagle ’ s medium (DMEM) (Invitrogen) containing 10% fetal bovine serum. A mutant dihydrofolate reductase gene provides a dom-inant selectable marker for the B82L cells overexpressing the EGFR 39). These cells were maintained in the same medium containing 2 (cid:2) M methotrexate (39, 40). Prior to experiments methotrexate was re- moved for at least 24 h. Confluent B82L cells were then incubated in serum-free DMEM for 18 – 20 h before stimulation with zinc sulfate or EGF.Asolution of 100 m M zinc sulfate was prepared in distilled water and used as a stock for dilution into serum-free DMEM. The final concen-tration of Zn 2 (cid:1) was 500 (cid:2) M , and that of EGF was 100 ng/ml. Transfection the EGFR-Y845F Construct— The EGFR construct The B82L-par cells were 80% transfected EGFR

The EGFR family of receptor tyrosine kinases in mammals is composed of four members: EGFR (ErbB1), ErbB2, ErbB3, and ErbB4 (14). Of these, EGFR, an 1186-amino acid residue transmembrane glycoprotein (15), is the prototypal member of this superfamily and is expressed in many cell types (16). Structurally, EGFR consists of an extracellular ligand binding domain, a single ␣-helical transmembrane pass, an intracellular tyrosine kinase domain, and a COOH-terminal region that contains autophosphorylation sites (15,(17)(18)(19)(20). Upon binding of specific polypeptide ligands, including EGF, transforming growth factor-␣, betacellulin, heparin-binding EGF, epiregulin, and amphiregulin, EGFR undergoes homo-or heterodimerization and activation of its intrinsic tyrosine kinase activity (17,21). These events lead to the autophosphorylation of multiple tyrosine residues in the COOH-terminal tail of the molecule that serve as binding sites for cytosolic signaling proteins containing Src homology 2 (SH2) domains and phosphotyrosine binding domains (20,22).
Five sites of in vivo autophosphorylation have been identified in the EGFR: three major (tyrosines 1068, 1148, and 1173) and two minor (tyrosines 992 and 1086) (18,(23)(24)(25). The binding of phosphorylated EGFR tyrosines with downstream signaling proteins initiates multiple signaling cascades that culminate in cell proliferation, migration, protein secretion, differentiation, and/or oncogenesis (16). Among these signaling pathways, the Shc/Grb2/Sos/Ras/MAPK cascade is a major mitogenic signaling pathway initiated by the EGFR (26). EGFR is also part of signaling networks transactivated by stimuli that do not directly interact with this receptor. These stimuli include agonists that specifically bind to other membrane receptors, membrane depolarization agents, and environmental stressors (27). The mechanisms for the transactivation of EGFR vary with the cell type and stimuli (27).
Cellular Src functions as a co-transducer of transmembrane signals emanating from a variety of growth factor receptors, including EGFR (28 -30 the nonreceptor tyrosine kinase Src cooperate in both mitogenesis and transformation (28 -32). Novel Src-phosphorylated tyrosine residues on the EGFR have been identified (33)(34)(35).
Our previous study showed that Zn 2ϩ ions induced EGFR signaling in human airway epithelial cells (12). In this study, we examined the mechanisms involved in Zn 2ϩ -induced EGFRmediated signaling, especially the possible cooperation of EGFR with Src. We found that an intact tyrosine at 845 of EGFR (Tyr-845) was necessary for Zn 2ϩ -induced Ras activation. However, the tyrosine kinase and autophosphorylation sites within EGFR were dispensable. Furthermore, Zn 2ϩ was found to activate Src kinase and induce its association with EGFR. These results indicate that Ras activation triggered by Zn 2ϩ is dependent on Src-mediated phosphorylation of EGFR at Tyr-845.
A solution of 100 mM zinc sulfate was prepared in distilled water and used as a stock for dilution into serum-free DMEM. The final concentration of Zn 2ϩ was 500 M, and that of EGF was 100 ng/ml.
Stable Transfection with the EGFR-Y845F Construct-The EGFR (Y845F) construct was described previously (34,35). The B82L-par cells were grown to 80% confluence and transfected with the EGFR (Y845F) construct using LipofectAMINE (Invitrogen). Specifically, 10 g of the plasmid DNA was incubated with 60 l of LipofectAMINE reagent for 30 min at room temperature. Transfection mixtures were diluted into Opti-MEM I reduced serum medium (Invitrogen), and the diluted complex solution was overlaid onto the subconfluent cells in 100-mm culture dishes. After incubating with B82L-par cells for 6 h at 37°C, the transfection mixture was removed and replaced with fresh DMEM containing 10% fetal bovine serum. The transfected cells were passaged in DMEM containing 1000 g/ml Geneticin (Invitrogen). After five passages, several colonies were picked and characterized for EGFR protein expression and phosphorylation using specific antibodies.
Ras Activation Assay-A Ras activation assay kit was purchased from Upstate Biotechnology (Lake Placid, NY). Ras activity was determined according to the supplier's instruction. B82L cells treated with zinc sulfate were lysed with 5 ϫ Mg 2ϩ lysis/wash buffer (MLB) (125 mM HEPES, pH 7.5, 750 mM NaCl, 5% Nonidet P-40, 50 mM MgCl 2 , 5 mM EDTA, and 10% glycerol). 500 -1000 g of cell lysate was precleared with glutathione agarose. 5 l of a 50% slurry of Raf-1 RBD (Ras binding domain)-agarose was incubated with the lysate at 4°C for 30 min. The agarose was collected by centrifugation and washed with MLB three times and once with cold PBS, then boiled in 25 l of reducing sample loading buffer. GTP-bound Ras protein was resolved by electrophoresis and transferred to nitrocellulose before being probed with a mouse monoclonal anti-Ras (clone Ras10) antibody (1 g/ml). Protein bands were visualized using a goat anti-mouse secondary antibody conjugated to HRP and an enhanced chemiluminescence detection system (12).
Co-immunoprecipitation of the EGFR with Src-Confluent B82L cells were treated with zinc sulfate for 20 min and lysed with RIPA lysis buffer. The lysates were pre-cleared with Protein A-Sepharose and immunoprecipitated with agarose-conjugated anti-EGFR antibody for 1 h at 4°C. Immune complexes were washed twice with RIPA buffer and once with cold PBS. The immunoprecipitates were suspended in 25 l of 4ϫ sample loading buffer (62.5 mM Tris-HCl, pH 6.8, 10% glycerol, 2% SDS, 0.7 M ␤-mercaptoethanol, 0.05% bromphenol blue) and boiled for 5 min before separation on 4 -15% Tris-HCl ready gels. The protein was transferred to a nitrocellulose membrane and incubated with anti-Src mouse monoclonal IgG 1 at 4°C overnight. The EGFR-associated Src bands were detected with HRP-conjugated goat anti-mouse secondary antibody using ECL chemiluminescence regents as described above.

RESULTS
EGFR Mutants-Schematic representations of human wild type EGFR and three EGFR mutants are shown in Fig. 1. The three EGFR mutants include a construct encoding a kinaseinactive receptor that contains a lysine to methionine substitution at position 721 (B82L-K721M), which was involved in ATP binding, a construct encoding a kinase-active receptor truncated at Y958 (B82L-c'958), which lacks all five EGFR EGFR-wt, wild type receptor that has a functional kinase domain, an intact Src phosphorylation site, and all five autophosphorylation sites; EGFR-c'958, a COOHterminally truncated receptor at Tyr-958 that lacks all five autophosphorylation sites; EGFR-K721M, a kinase-deficient receptor that has a lysine to methionine point mutation at Tyr-721 that is involved in ATP binding; EGFR-Y845F, a mutant receptor that has a tyrosine to phenylalanine point mutation at Tyr-845 that is phosphorylated by nonreceptor tyrosine kinase Src. autophosphorylation sites, and a construct encoding a mutant EGFR at Tyr-845 where the tyrosine residue is replaced with phenylalanine (34,35). No endogenous EGFR protein was detected in B82L-par cells while other transfected B82L cells express abundant EGFR protein (data not shown), which is consistent with previous studies (37,39,41).
Zn 2ϩ -induced Ras Activation in B82L Cells-To assess the functional activation of EGFR, we used Ras as a downstream marker of EGFR signaling. Ras activation assays showed that exposure of B82L-par cells to Zn 2ϩ or to EGF did not increase GTP-bound Ras after 20 min (Fig. 2A). In contrast, Zn 2ϩ induced a significant increase in GTP-bound Ras in B82L-wt, B82L-c'958 cells, or B82L-K721M (Fig. 2, B-D), confirming the requirement for EGFR, as suggested by our previous study. Interestingly, Zn 2ϩ increased GTP-bound Ras in cells expressing kinase-inactive and COOH-truncated EGFR demonstrating that neither the tyrosine kinase nor the autophosphorylation sites of EGFR were required for Zn 2ϩ -induced Ras activation. These results suggested that EGFR was transactivated in Zn 2ϩ -treated cells. In contrast, no increase in GTPbound Ras was detected in Zn 2ϩ -treated B82L-Y845F cells (Fig. 2E), indicating that EGFR tyrosine 845, which is phosphorylated by c-Src (35,42), is required for Zn 2ϩ -induced Ras activation. EGF increased GTP-bound Ras in wt, kinase-inactive K721M and to a lesser extent in c'958 EGFR-expressing cells as previously repeated (39,43,44). EGFR Tyr-845 was not required for EGF-induced Ras activation (Fig. 2E).
Requirement of Src Kinase Activity for Zn 2ϩ -induced Ras Activation-To investigate whether Src kinase mediated transactivation of EGFR in Zn 2ϩ -exposed cells, we exposed cells to PP1, a potent and selective Src kinase inhibitor (45). Pretreatment with PP1 significantly blocked Zn 2ϩ -induced Ras activa-tion in B82L-wt (Fig. 3), whereas PP1 had no inhibitory effect on EGF-induced Ras activation. These results indicate that Src activity is required for Zn 2ϩ -but not for EGF-induced Ras activation. Similar results were obtained with B82L-c'958 or B82L-K721M cell lines used in this study (data not shown).
To exclude the possibility that PP1 affected the intrinsic kinase activity of EGFR, EGF-induced receptor autophosphorylation was examined in B82L-wt cells that were pretreated with PP1. Consistent with previous reports (41,46), PP1 had no significant effect on EGF-induced EGFR autophosphorylation (data not shown).
Zn 2ϩ -induced Phosphorylation of Src on Y416 -To determine whether Zn 2ϩ affected c-Src, we examined the effects of Zn 2ϩ on Src using a phospho-specific anti-Src (Y416) antibody. Phosphorylation of Src Y416 is part of the enzyme activation mechanism (47). As shown in Fig. 4, exposure to Zn 2ϩ for 20 min induced the phosphorylation of Y416 on Src in B82L-par (Fig. 4A), B82L-wt (Fig. 4B), B82L-c'958 (Fig. 4C), and B82L-K721M (Fig. 4D) cell lines. These results further support the hypothesis that Src activation is involved in EGFR phosphorylation. Src phosphorylation was also detected in the B82L-Y845F cell line (Fig. 4E) wherein the known Src phosphorylation site Tyr-845 on EGFR was replaced with phenylalanine. These data indicated that activation of Src occurred independent of the existence of EGFR Tyr-845. In some cell types (B82Lwt, B82L-K721M, or B82L-Y845F) EGF also increased phosphorylation of Src, and the magnitude of the increase was always less than those induced by Zn 2ϩ (Fig. 4).
Zn 2ϩ -induced Association of the EGFR with Src-Next, we determined whether Zn 2ϩ induces a physical association between EGFR and Src in B82L-wt (Fig. 5A), B82L-c'958 ( Fig.  5B), B82L-K721M (Fig. 5C), and B82L-Y845F (Fig. 5D) cell lines. Immunoprecipitation of EGFR followed by immunoblotting with anti-Src antibody revealed that treatment with Zn 2ϩ for 20 min caused a co-localization of EGFR with Src in these cell lines. In contrast, treatment with EGF for 20 min did not result in an interaction between Src and EGFR (Fig. 5).
Zn 2ϩ -induced Phosphorylation of EGFR at Tyr-845 and Tyr-1068 -The finding that mutation of EGFR Tyr-845 prevented Zn 2ϩ -induced Ras activation suggested that phosphorylation of this site was essential for transactivation of EGFR in cells exposed to Zn 2ϩ . To further test this hypothesis we examined the phosphorylation of this site using a phospho-specific antibody to Tyr-845 of the EGFR. Zn 2ϩ induced phosphorylation of EGFR at Tyr-845 in B82L-wt (Fig. 6A), B82L-c'958 (Fig. 6B), and B82L-K721M (Fig. 6C) cells. EGF treatment also increased the phosphorylation of Tyr-845 on EGFR in these cell lines. To determine whether the phosphorylation of Tyr-845 is related to Src activation induced by Zn 2ϩ , the effect of PP1 on Zn 2ϩinduced phosphorylation of EGFR Tyr-845 was also examined in these cell lines. PP1 significantly blocked Zn 2ϩ -induced phosphorylation of EGFR Tyr-845 in all of the cell lines. In contrast, PP1 only had a slight inhibitory effect on EGF-induced EGFR phosphorylation at Tyr-845 in B82L-wt and B82L- , and B82L-Y845F (E) cells were starved in serum-free DMEM for 24 h before treatment with 500 M zinc sulfate or 100 ng/ml EGF for 20 min. Cells were lysed with MLB, and lysates were immunoprecipitated with Raf-1 Ras binding domain (RBD)-conjugated agarose. The immunoprecipitates were subject to 4 -15% Tris-HCl ready gels. The membranes were blotted with anti-Ras IgG 2ak overnight and then incubated with anti-mouse HRPconjugated rabbit IgG. The optical densities of GTP-bound Ras bands were quantified and normalized to that of control. Bar graphs summarize at least three separate experiments.

FIG. 3. Effect of the Src kinase inhibitor, PP1, on Zn 2؉ -induced
Ras activation. Subconfluent B82L-wt cells were pretreated with 5 M PP1 for 30 min and then stimulated with 500 M zinc sulfate or 100 ng/ml EGF for 20 min, respectively. Lysates were immunoprecipitated with Raf-1 Ras binding domain (RBD)-conjugated agarose. The immunoprecipitates were subject to 4 -15% Tris-HCl ready gels, and the membranes were blotted with anti-Ras IgG 2ak overnight and then incubated with anti-mouse HRP-conjugated rabbit IgG. Data shown are representative of three independent experiments. K721M cells. As expected, both Zn 2ϩ and EGF failed to induce phosphorylation of EGFR Tyr-845 in B82L-Y845F cells. Additionally, the phosphorylation of EGFR Tyr-1068, a major EGFR autophosphorylation site (23), was also examined in these cell lines stimulated with Zn 2ϩ or EGF. As expected, no phosphorylation of Tyr-1068 in B82L-c'958 cells was observed (Fig. 7B). In addition, Zn 2ϩ failed to induce the phosphorylation of EGFR Tyr-1068 in B82L-wt (Fig. 7A), B82L-K721 (Fig. 7C), and B82L-Y845F (Fig. 7D) cells. In comparison, EGF induced phos-phorylation of EGFR Tyr-1068 in each cell type (Fig. 7, A, C, and D).

DISCUSSION
Cross-communication between heterologous signaling systems is essential to integrate a variety of extracellular stimuli into a limited number of signaling pathways (48,49). EGFR has been identified as a key element in the complex signaling network that is transactivated by G protein-coupled receptors, cytokine receptors, estrogen receptors, integrins, ion channels, or stress-inducing agents (49 -52). Accumulated evidence indicates that Src kinases, calcium, the Ca 2ϩ -regulated focal adhesion kinase family kinase Pyk2, protein kinase C, Janus, and tyrosine kinase Jak2 play crucial roles in the process of EGFR transactivation (48,53,54). Moreover, the recent identification of Zn 2ϩ -dependent metalloproteinases and transmembrane growth factor precursors as critical elements in G proteincoupled receptor-induced EGFR transactivation pathways has defined new components of a cellular communication network of rapidly increasing complexity (51,55,56).
In the present study, the mechanisms for Zn 2ϩ -induced EGFR signaling were examined in B82L cells expressing wild type, null, or modified versions of EGFR. These results show that Zn 2ϩ ions induce Ras activation through a mechanism that requires EGFR but not EGFR tyrosine kinase activity or autophosphorylation sites. Furthermore, an intact tyrosine at Tyr-845 on EGFR is required for Zn 2ϩ -induced Ras activation. Zn 2ϩ -induced phosphorylation of the EGFR Tyr-845 residue requires Src activation. These data suggest that Zn 2ϩ -induced EGFR transactivation occurs through an Src-dependent pathway.
Src phosphorylation of EGFR has been mapped to most of the five autophosphorylation sites as well as other novel sites (50). Tyr-845 in the EGFR is not a known autophosphorylation site and has been shown to be phosphorylated by Src both in vivo and in vitro (34,57). Our data support a direct phosphorylation of EGFR by c-Src for the following reasons: First, Zn 2ϩ ions induce co-localization of EGFR with Src and Src phosphorylation of Tyr-845 has been reported to occur in an EGFR⅐Src complex (57). The c-Src SH2 domain can bind activated EGFR specifically and directly (58 -60), which suggests other kinases may not be necessary to mediate the phosphorylation of Tyr-845 (34). Other Src-dependent or -independent residues on EGFR may act as docking sites for Src or facilitate the association of EGFR with Src, leading to Src phosphorylation of Tyr-845 (33,34). However, phosphorylation of Tyr-845 is only one of multiple Src effectors, because Zn 2ϩ still induced phosphorylation of Src in the absence of Tyr-845 in B82L-Y845F cells. Second, Src activity is required for the phosphorylation of Tyr-845 as demonstrated with the use of a pharmacological inhibitor of Src kinase activity (PP1). Overexpression of wild type or kinase-deficient Src constructs by other investigators produced results that agreed with these findings (34,35). The observation that EGF induced a greater extent of tyrosine phosphorylation at Tyr-845 than Zn 2ϩ is likely due to ligandinduced conformational changes that enhance accessibility of this residue (61).
Zn 2ϩ ions induced Ras activation in cells expressing kinaseinactive or truncated EGFR where Tyr-845 of EGFR is structurally intact and can be phosphorylated, suggesting that Tyr-845 provides a site of interaction with downstream effectors in Zn 2ϩ -induced Ras activation. ErbB2, a preferred dimerization partner of all ErbB proteins (62, 63), may be involved in EGFdependent signaling in cells overexpressing aberrant EGFR (36,44,64). Phosphorylation of EGFR Tyr-1068 by EGF in B82L-K721M cells implied that ErbB2 might play a certain role in EGF-induced signaling (44). We have observed that Zn 2ϩ ions and EGF can activate ErbB2 in B82L-wt and B82Lc'958 cells (data not shown). This hypothesis will be examined using specific cell lines with or without expression of ErbB2 (65) in our laboratory. Although Tyr-845 of EGFR was necessary for Zn 2ϩ -induced Ras activation, this site is dispensable for Zn 2ϩ -induced EGFR-Src co-localization, because Zn 2ϩ still induced EGFR-Src association in B82L-Y845F cells. Thus, we hypothesize that other sites on EGFR may act as the docking sites for EGFR-Src interaction. Whether there exists a scaffolding protein to assemble these elements remains unknown, although caveolin-1 has been reported to serve as an anchoring site for various signaling molecules in B82L fibroblasts (41,66).
Differences in Ras activation by Zn 2ϩ ions and EGF likely reflect differences in the manner in which EGFR is utilized by these stimuli. First, the autophosphorylation sites in the EGFR are unnecessary for Ras activation induced by Zn 2ϩ . Unlike EGF, Zn 2ϩ ions do not induce the autophosphorylation of Tyr-1068 in the EGFR. Second, the Tyr-845 residue is required for Zn 2ϩ -but not for EGF-induced Ras activation. The phosphorylation of Tyr-845 induced by Zn 2ϩ ions, but not that induced by EGF, is dependent on Src activity. Therefore, Src activation and subsequent phosphorylation of Tyr-845 are crucial steps in the process of Zn 2ϩ -induced Ras activation. Other tyrosine phosphorylation sites play a significant role in EGF-induced Ras activation. We have observed a baseline of GTP-bound Ras in all B82L cell lines, which could be suppressed by the Src kinase inhibitor, suggesting that constitutive Src activity may be responsible for the basal activation of Ras.
In summary, this study describes a novel mechanism for metal-induced EGFR transactivation, which is likely to be mediated by Src through the phosphorylation site of Tyr-845 on EGFR.