Peroxynitrite Induces Covalent Dimerization of Epidermal Growth Factor Receptors in A431 Epidermoid Carcinoma Cells*

Irreversible tyrosine modifications by inflammatory oxidants such as peroxynitrite (ONOO−) can affect signal transduction pathways involving tyrosine phosphorylation. The epidermal growth factor receptor (EGFR), a member of the c-ErbB receptor tyrosine kinase family, is involved in regulation of epithelial cell growth and differentiation, and possible modulation of EGFR-dependent signaling by ONOO− was studied. Exposure of epidermoid carcinoma A431 cells to 0.1–1.0 mm ONOO− resulted in tyrosine nitration on EGFR and other proteins but did not significantly affect EGFR tyrosine autophosphorylation. A high molecular mass tyrosine-phosphorylated protein (∼340 kDa) was detected in A431 cell lysates after exposure to ONOO−, most likely representing a covalently dimerized form of EGFR, based on immunoprecipitation and/or immunoblotting with α-EGFR antibodies and co-migration with ligand-induced EGFR dimers cross-linked with 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide. Covalent EGFR dimerization by ONOO− probably involved intermolecular dityrosine cross-linking and was enhanced after receptor activation with epidermal growth factor. Furthermore, irreversibly cross-linked EGFR was more extensively tyrosine-phosphorylated compared with the monomeric form, indicating that ONOO− preferentially cross-links activated EGFR. Exposure of A431 cells to ONOO− markedly reduced the kinetics of tyrosine phosphorylation of a downstream EGFR substrate, phospholipase C-γ1, which may be related to covalent alterations in EGFR. Alteration of EGFR signaling by covalent EGFR dimerization by inflammatory oxidants such as ONOO− may affect conditions of increased EGFR activation such as epithelial repair or tumorigenesis.

Irreversible tyrosine modifications by inflammatory oxidants such as peroxynitrite (ONOO ؊ ) can affect signal transduction pathways involving tyrosine phosphorylation. The epidermal growth factor receptor (EGFR), a member of the c-ErbB receptor tyrosine kinase family, is involved in regulation of epithelial cell growth and differentiation, and possible modulation of EGFR-dependent signaling by ONOO ؊ was studied. Exposure of epidermoid carcinoma A431 cells to 0.1-1.0 mM ONOO ؊ resulted in tyrosine nitration on EGFR and other proteins but did not significantly affect EGFR tyrosine autophosphorylation. A high molecular mass tyrosinephosphorylated protein (ϳ340 kDa) was detected in A431 cell lysates after exposure to ONOO ؊ , most likely representing a covalently dimerized form of EGFR, based on immunoprecipitation and/or immunoblotting with ␣-EGFR antibodies and co-migration with ligandinduced EGFR dimers cross-linked with 1-ethyl-3-(3dimethylaminopropyl)carbodiimide. Covalent EGFR dimerization by ONOO ؊ probably involved intermolecular dityrosine cross-linking and was enhanced after receptor activation with epidermal growth factor. Furthermore, irreversibly cross-linked EGFR was more extensively tyrosine-phosphorylated compared with the monomeric form, indicating that ONOO ؊ preferentially cross-links activated EGFR. Exposure of A431 cells to ONOO ؊ markedly reduced the kinetics of tyrosine phosphorylation of a downstream EGFR substrate, phospholipase C-␥1, which may be related to covalent alterations in EGFR. Alteration of EGFR signaling by covalent EGFR dimerization by inflammatory oxidants such as ONOO ؊ may affect conditions of increased EGFR activation such as epithelial repair or tumorigenesis.
Inflammatory conditions are associated with increased production of reactive oxygen metabolites, such as superoxide anion (O 2 . ) and hydrogen peroxide (H 2 O 2 ), as well as induction of nitric oxide (NO ⅐ ) synthesis, which is thought to play major roles in host defense but is also presumed to contribute to the development of tissue injury associated with chronic inflammation. The nearly diffusion-limited reaction between nitric oxide (NO ⅐ ) and superoxide (O 2 . ) produces peroxynitrite (ONOO Ϫ ), a powerful oxidant and cytotoxic species that has been proposed to importantly contribute to the pathophysiology of a large variety of diseases associated with inflammation (1). Peroxynitrite is reactive toward all classes of biomolecules, including lipids, proteins and nucleic acids, and oxidative reaction products in proteins primarily include modifications of cysteine, methionine, tryptophan, and tyrosine residues (2). Irreversible covalent modifications of tyrosine residues by ONOO Ϫ include the formation of 3-nitrotyrosine and 3,3Ј-dityrosine, which represent characteristic "markers" for inflammatory (NO ⅐ -derived) oxidants (5)(6)(7)(8), but these modifications may also have functional consequences and could contribute to the pathophysiological effects of NO ⅐ -derived reactive nitrogen oxides such as ONOO Ϫ . Chemical studies with the nitrating agent tetranitromethane have shown that nitration of critical tyrosine residues inactivates a wide variety of enzymes (9 -11) and affects structural proteins (1,5). Studies with isolated tyrosine kinases have demonstrated that substrate tyrosine residues are not phosphorylated when they are nitrated (12,13), which implies that tyrosine nitration may interfere with signal transduction pathways involving tyrosine phosphorylation. Alternatively, formation 3,3Ј-dityrosine by inflammatory oxidants such as ONOO Ϫ can result in inter-or intramolecular covalent cross-linking in e.g. membrane proteins (7,8) and thereby affect signaling pathways. Despite various recent investigations (13)(14)(15)(16), the potential pathophysiological significance of such protein modifications has, however, not been demonstrated conclusively in intact cellular systems.
One group of proteins that may be significantly affected by inflammatory oxidants are membrane receptors for various growth factors or cytokines that are induced and/or activated during inflammation to promote wound healing processes. Among these is the epidermal growth factor receptor (EGFR), 1 which is overexpressed and activated in response to epithelial injury to promote epithelial repair processes (17,18). The EGFR is a member of the receptor tyrosine kinase superfamily and is involved in the regulation of proliferation and differentiation of primarily epithelial cell types (19,20). The EGFR is a 170-kDa glycoprotein that spans the membrane via one ␣-hel-ical segment of 23 amino acids connecting a large, heavily glycosylated extracellular ligand binding domain and an intracellular tyrosine kinase domain. Stimulation of EGFR by its endogenous ligands (transforming growth factor-␣ or epidermal growth factor (EGF)) involves ligand-induced receptor dimerization (21,22), which is now recognized as a general activation mechanism common to a large number of receptors for growth factors, cytokines, and hormones (20,23). The activated EGFR-ligand complex is then internalized and eventually degraded in lysosomes or dissociated for EGFR recycling. Activation of EGFR causes phosphorylation on intracellular serine, threonine, and tyrosine residues and activation of tyrosine kinase activity, which results in tyrosine phosphorylation and activation of various effector proteins, ultimately leading to stimulation of DNA replication and cell division.
Recently, the signaling mechanism of EGFR has been demonstrated to involve generation of hydrogen peroxide (24), and various reports have indicated that oxidative modification of a cysteine residue in EGFR may reversibly affect its activation (25)(26)(27). Alternatively, other covalent modifications by ONOO Ϫ , such as nitration of one or more of the tyrosine residues involved in the EGFR signaling cascade or the formation of intra-or intermolecular 3,3Ј-dityrosine cross-links within EGFR could more irreversibly interfere with EGFR signaling pathways. The present study was undertaken to investigate the potential consequences of ONOO Ϫ -induced tyrosine modifications on tyrosine phosphorylation and activation of EGFR in intact cells. The results indicate that exposure of A431 epidermoid carcinoma cells to ONOO Ϫ causes irreversible dimerization of EGFR, most likely via dityrosine cross-links, which appears to be related to the extent of EGFR activation. Formation of ONOO Ϫ during inflammatory processes may thereby alter EGFR-mediated signaling pathways in target cells and affect cell proliferation or differentiation, which may have implications in wound repair or tumor formation.

EXPERIMENTAL PROCEDURES
Cell Culture-Experiments were performed using the human epidermoid carcinoma cell line A431, which overexpresses EGFR and contains 1-3 ϫ 10 6 EGF receptors/cell (28). A431 cells were grown to 80 -90% confluence in 100-mm culture dishes in Dulbecco's modified Eagle's medium (Sigma) supplemented with 1.2 g/liter NaHCO 3 and 7 ml/liter penicillin/streptomycin (Sigma) and 5% newborn bovine serum (Life Technologies, Inc.). Before experimentation, cells were serum-starved for 2 h, washed twice with phosphate-buffered saline (PBS), and subsequently placed in 2 ml of modified PBS (50 mM Na 2 HPO 4 , 90 mM NaCl, 5 mM KCl, 0.8 mM MgCl 2 , 1 mM CaCl 2 , and 5 mM glucose, pH 7.4). This modified PBS was used to avoid interfering reactions of ONOO Ϫ and other oxidant systems used in our experiments with media constituents and to provide sufficient buffering capacity to avoid pH changes after the addition of alkaline solutions of ONOO Ϫ .
Oxidant Treatments-Peroxynitrite was prepared via quenched-flow reaction of H 2 O 2 with HNO 2 and subsequently passed over MnO 2 to remove residual H 2 O 2 , and quantitated spectrophotometrically prior to experimentation (⑀ 302 ϭ 1,670 M Ϫ1 cm Ϫ1 ; Ref. 29). Following a 10-min preincubation of A431 cells in modified PBS, ONOO Ϫ was added in small aliquots (Ͻ6 l) on the side of the culture dish and immediately mixed with the incubation buffer by rapid swirling, in order to assure optimal exposure of the cells to ONOO Ϫ before decomposition. In control experiments, cells were treated in a similar fashion with equal volumes of 1.2 M NaOH (vehicle control) or with aliquots of ONOO Ϫ stock solutions that were allowed to decompose completely by overnight storage at room temperature but still contain contaminants such as nitrite and nitrate (29). Similar experiments were performed with reagent H 2 O 2 or sodium hypochlorite (NaOCl) (Sigma), which were quantitated spectrophotometrically before experimentation, using ⑀ 240 of 43.6 M Ϫ1 cm Ϫ1 and ⑀ 290 of 350 M Ϫ1 cm Ϫ1 at pH 12, respectively. For experiments involving horseradish peroxidase (HRP, type I; Sigma), this enzyme was added to the incubation medium shortly before the addition of H 2 O 2 . EGFR activation was performed by the addition of EGF (20 nM; Upstate Biotechnology Inc.) either before or after oxidant treatments. In experiments with the chemical cross-linker 1-ethyl-3-[3-(dimethyl-amino)propyl]-carbodiimide (EDAC), cells were incubated with 10 mM EDAC for 40 min at 37°C following other treatments. To remove surface-bound EGF, cells were rinsed twice for 2 min at 4°C with acid wash buffer (0.1 M NaAc, 0.15 M NaCl, pH 4.5), after which the cells were washed twice with PBS.
After the various cell treatments, culture dishes were placed on ice, and the cells were washed twice with cold PBS and lysed in 250 l of solubilization buffer (50 mM HEPES, 250 mM NaCl, 1.5 mM MgCl 2 , 2 mM Na 3 VO 4 , 1% Triton, 10% glycerol, 1 mM phenylmethylsulfonyl fluoride, 10 g/ml aprotinin, and 10 g/ml leupeptin, pH 7.5). Dishes were kept on ice for 20 min on a rocker platform to allow complete solubilization of EGFR, and cell lysates were collected and centrifuged (10 min at 10,000 ϫ g) to remove insoluble cellular debris.
Analysis of EGFR Tyrosine Phosphorylation and Protein Nitration-Aliquots of cell lysates (100 l) were mixed with 20 l of 6ϫ sample reducing buffer (125 mM Tris-HCl, pH 6.8, 20% glycerol, 4% SDS, 10% ␤-mercaptoethanol) and boiled for 5 min to denature proteins. Proteins were separated on 4% polyacrylamide-SDS gels, and electroblotted on polyvinylidene difluoride (Immobilon-P; Sigma) membranes for Western analysis. Membranes were blocked with 1% bovine serum albumin and then incubated with a monoclonal antibody against phosphotyrosine (PY-20; Upstate Biotechnology), a polyclonal antibody against 3-nitrotyrosine (Upstate Biotechnology) or polyclonal antibodies against the intracellular domain of EGFR (kindly donated by Drs. D. Cadena and G. N. Gill (clone 1964) and by Dr. J. Schlessinger (RK-2)), followed by incubation with secondary antibodies conjugated with horseradish peroxidase. Antibodies were detected by diaminobenzidine staining (Vector Laboratories, Burlingame, CA) or by enhanced chemiluminescence (Amersham Pharmacia Biotech, Buckinghamshire, UK).
Experiments with A431 Plasma Membranes and Purified EGFR-A431 plasma membranes were prepared from approximately 10 8 cells, which were washed with PBS and homogenized in 20 ml of HEPES buffer (20 mM HEPES, 1.5 mM MgCl 2 , 1 mM EDTA, 1 mM phenylmethylsulfonyl fluoride, and 100 g/ml leupeptin, pH 7.4) at 4°C by brief sonication. Cell homogenates were centrifuged at 1,500 ϫ g for 10 min to remove nuclei and unlysed cells, the supernatant was further centrifuged at 25,000 ϫ g for 30 min, and the pellet containing plasma membranes was washed with PBS and finally resuspended in PBS at approximately 10 mg/ml protein (determined according to Bradford (30) with bovine serum albumin as a standard) and stored at Ϫ80°C until use. Plasma membranes were diluted to 1 mg/ml protein in 50 mM sodium phosphate buffer (pH 7.4), containing 5 mM MgCl 2 , 1 mM MnCl 2 and 100 M ATP, reacted with ONOO Ϫ or other oxidants, and solubilized for SDS-PAGE and Western analysis. Similar experiments were performed with EGFR, purified from A431 cells by affinity chromatography (Sigma), which was suspended to 20 g/ml in 50 mM sodium phosphate buffer (pH 7.4), containing 5 mM MgCl 2 , 1 mM MnCl 2 , and 100 M ATP. Aliquots of 50 l were reacted with ONOO Ϫ (500 M) before or after receptor stimulation with 20 nM EGF. Incubations were terminated by the addition of 10 l of 6ϫ sample reducing buffer, and samples were subjected to SDS-PAGE and Western blotting.
Immunoprecipitation of EGFR-Immunoprecipitation of EGFR from A431 cell lysates was performed using monoclonal ␣-EGFR antibody 528 (kindly provided by Dr. J. Mendelsohn), which was complexed with protein A-Sepharose (PAS; Sigma). The preformed PAS⅐mAb complex was washed three times with HNTG buffer (20 mM HEPES, 150 mM NaCl, 0.1% Triton X-100, 10% glycerol, pH 7.5), and aliquots of the PAS⅐mAb (containing 4 g of mAb 528) were incubated overnight with 200 l of A431 cell lysate (up to 1 mg of protein) at 4°C (31). The PAS⅐mAb receptor complex was washed three times with HNTG and mixed with 3ϫ sample reducing buffer for analysis by SDS-PAGE.
Analysis of Tyrosine Phosphorylation of Phospholipase C-␥1 (PLC-␥1)-After treatment of A431 cells with EGF and/or ONOO -, cells were solubilized, and aliquots containing 2 mg of protein were immunoprecipitated by overnight incubation at 4°C with 2 g of a monoclonal antibody against PLC-␥1 (clone D-7-3; Upstate Biotechnology) followed by incubation for 2 h with 5 mg of protein A-Sepharose (Sigma). The PAS-antibody complex was washed three times with HNTG buffer, boiled for 5 min in sample reducing buffer, and subsequently subjected to SDS-PAGE and Western blot analysis with ␣-phosphotyrosine (PY-20) or with a polyclonal antibody against PLC-␥1 (Upstate Biotechnology).
Analysis of Tyrosine Modification Products by HPLC-Cellular proteins were precipitated from cell lysates with 5% trichloroacetic acid, and washed twice with ethanol/ethyl acetate 50/50 (v/v) to remove contaminating lipids. Protein pellets were dissolved in 6 M HCl and hydrolyzed under vacuum for 24 h at 110°C, after which hydrolysates were dried under N 2 and reconstituted in 50 mM phosphate buffer (pH 3.0)/methanol (93/7) (v/v). Tyrosine and its oxidation products were analyzed by HPLC with tandem UV (274 nm) and fluorescence detection (excitation, 284 nm; emission, 410 nm) (3) and quantitated by comparison with external standards.
Assays of Cell Viability or Proliferation-A431 cells were seeded at 5 ϫ 10 4 cells/well in six-well plates in Dulbecco's modified Eagle's medium with 1% newborn bovine serum and incubated for 24 h before the addition of ONOO Ϫ (after replacing the medium with modified PBS as described above; final volume 1 ml) or 20 nM EGF. Cell proliferation was assessed on four consecutive days after these treatments by collection of cells by trypsinization and cell counting using a hemocytometer.
In separate experiments, cell viability was assessed 1 h after the addition of ONOO Ϫ using Alamar Blue ® (Alamar Bioscience, Sacramento, CA), which is metabolized by respiring cells, resulting in increased absorbance at 600 nm. Cell viability was assessed by comparing A 600 for untreated cells (100% viable) with those of ONOO Ϫ -treated cells.

RESULTS AND DISCUSSION
Effects of ONOO Ϫ on Cell Viability and Proliferation-The addition of 0.25-1.0 mM ONOOto subconfluent A431 cells in six-well plates (50,000 cells/well) caused only marginal loss of cell viability as determined using Alamar Blue ® . Cell viability was decreased by 20 -30%, measured 24 h after the addition of 1 mM ONOO -. As illustrated in Fig. 1, ONOOcauses inhibition of cell proliferation only when added at concentrations higher than 0.5 mM. Continuous stimulation of A431 cells with 5 nM EGF also resulted in inhibition of cell growth, as shown previously (32). Thus, overstimulation of EGFR and the addition of ONOOboth cause a stress response in A431 cells that results in growth arrest and/or apoptosis (33)(34)(35)(36).
Nitration of Cellular Proteins and Tyrosine Phosphorylation-The addition of 0.1-1.5 mM ONOO Ϫ to confluent A431 cells in 100-mm culture dishes caused a dose-dependent increase in tyrosine nitration of cellular proteins as detected by both HPLC and immunoblotting with an ␣-nitrotyrosine antibody. The 3-nitrotyrosine content in cellular proteins increased from Ͻ0.02 to 0.23 Ϯ 0.06, 0.32 Ϯ 0.06, and 0.65 Ϯ 0.19 mmol/mol of tyrosine residues (n ϭ 4) after the addition of 0.5, 1.0, and 1.5 mM ONOO Ϫ , respectively. SDS-PAGE followed by immunoblotting with ␣-nitrotyrosine also demonstrated dosedependent nitration of various proteins including EGFR. Nitration of EGFR was confirmed by immunoprecipitation of EGFR (using mAb 528) followed by SDS-PAGE and immunoblotting with ␣-nitrotyrosine ( Fig. 2A).
The EGFR in nonstimulated A431 cells is partly phosphoryl-ated on tyrosine residues, as detected with an ␣-phosphotyrosine antibody, and the extent of this basal autophosphorylation was only marginally but not significantly affected after exposure to ONOO Ϫ for 30 min (Fig. 2B). Stimulation of A431 cells with 20 nM EGF for 10 -30 min caused a dramatic increase in EGFR autophosphorylation, but EGF-induced autophosphorylation was not significantly affected after treatment of A431 cells with up to 1.5 mM ONOO Ϫ (not shown). Together, these results suggest that the extent of protein tyrosine nitration by nontoxic concentrations of ONOO Ϫ is not sufficient to cause significant alteration in EGFR autophosphorylation. Moreover, EGFR tyrosine residues other than the intracellular phosphorylation substrates may have been nitrated. Hence, unless certain tyrosine residues are preferentially nitrated, tyrosine nitration to an extent that is feasible in vivo (37, 38) appears unlikely to cause significant inhibition of tyrosine phosphorylation. In similar recent studies, exposure of various cell types to ONOO Ϫ was found to cause variable changes in tyrosine phosphorylation as well as nitration of tyrosine residues, although these events could not be directly related (13)(14)(15)(16).
Peroxynitrite-induced Formation of Dimeric EGFR-Analysis of cell lysates of ONOOexposed A431 cells by SDS-PAGE revealed the presence of a large phosphorylated protein (approximately 340 kDa) that was not detected in untreated cells FIG. 1. Inhibition of A431 cell proliferation by ONOO ؊ . A431 cells were seeded at 10% confluence in six-well plates (50,000 cells/well) and allowed to attach overnight. The next day (day 0) cells were treated with 0 (Ⅺ), 0.25 (f), 0.5 (q), or 1.0 mM (OE) ONOO Ϫ or with 5 nM EGF (ࡗ). Cells were subsequently kept in Dulbecco's modified Eagle's medium/F-12 medium containing 1% newborn bovine serum for the next 4 days (EGF was kept in the medium throughout the experiment). Cell proliferation was followed by cell counting on four consecutive days after these treatments.  (Fig. 3). This protein was also detected by various polyclonal antibodies against EGFR (Fig. 3), indicating that this protein is a covalent aggregate of EGFR with one or more other proteins and most likely represents a homodimer of EGFR. The presumed EGFR dimer was also detected after immunoprecipitation of A431 lysates with mAb 528, followed by SDS-PAGE and immunoblotting with PY-20 or ␣-EGFR antibodies. Control experiments showed that formation of this putative EGFR dimer was not due to pH changes after the addition of alkaline ONOO Ϫ solutions or to contaminating components in the ONOO Ϫ stock solution, since similar additions of either 1.2 M NaOH, previously decomposed ONOO Ϫ , or H 2 O 2 were ineffective (Fig. 3).
The putative EGFR homodimer was detectable within 1 min after the addition of ONOO Ϫ to A431 cells, and the extent of dimerization did not change significantly for at least 60 min after the addition of ONOO Ϫ (not shown), suggesting that EGFR dimerization occurs by direct chemical modification of EGFR, rather than via secondary intracellular pathways induced by ONOO Ϫ . This notion was confirmed by similar experiments with (partially) purified EGFR; the addition of ONOO Ϫ to either A431 plasma membranes or to purified EGFR also resulted in the formation of dimeric EGFR, as assessed by SDS-PAGE (not shown). Furthermore, EGFR dimerization was also detected after the addition of ONOO Ϫ to A431 cells at 4°C, conditions that would have minimized metabolic pathways.
Possible Involvement of Dityrosine Cross-linking-The extracellular domain of EGFR contains cysteine-rich motifs (19,39), and thiol-reactive reagents have been demonstrated to affect EGFR activation by modification of one or more cysteine-residues in the intracellular domain (25)(26)(27). Since cysteine residues represent important targets for ONOO Ϫ (29), formation of intermolecular disulfide bonds might have been involved in EGFR cross-linking by ONOO Ϫ . However, SDS-PAGE analyses were performed under reducing conditions that would have cleaved intermolecular disulfide bonds. Furthermore, neither H 2 O 2 (Fig. 3) nor the inflammatory oxidant HOCl (not shown), oxidants capable of oxidizing cysteine residues to the disulfide, were able to induce similar EGFR cross-linking, indicating that ONOO Ϫ acts by a unique mechanism distinct from these other oxidants. Characteristic products from reaction of ONOO Ϫ with cysteine residues may include S-nitrosothiols (40 -42), which may somehow have been involved in ONOO Ϫ -induced EGFR cross-linking. However, the addition of 1 mM S-nitrosoglutathione to subconfluent A431 cells, which would be ex- pected to cause the formation of protein S-nitrosothiols by transnitrosation reactions, did not result in EGFR dimerization, nor did it significantly affect EGFR autophosphorylation (not shown). ONOO Ϫ -induced EGFR dimerization and tyrosine nitration were both effectively inhibited in the presence of 1 mM GSH, although depletion of intracellular GSH by overnight preincubation of A431 cells with 100 M buthionine sulfoximine to inhibit GSH synthesis (43) did not significantly affect the extent of EGFR cross-linking by ONOO Ϫ . This latter finding further substantiates that ONOO Ϫ causes EGFR cross-linking by direct receptor modification at the cell surface rather than by modifications of intracellular targets, which would be prevented by intracellular GSH.
The EGF receptor contains 15 tyrosine residues in its extracellular region, some of which are exposed and involved in ligand binding (44,45) and hence may represent susceptible targets for oxidation by ONOO Ϫ . A major product of reaction of ONOO Ϫ with tyrosine residues is 3,3Ј-dityrosine, formed via combination of two tyrosyl radicals (3,4); hence, EGFR could theoretically be covalently dimerized via intermolecular dityrosine cross-links. Proteins recovered from ONOO Ϫ -treated A431 cells were found to have increased levels of 3,3Ј-dityrosine, and protein 3,3Ј-dityrosine content and the extent of EGFR cross-linking were both found to increase proportionately with increasing concentrations of ONOO Ϫ (Fig. 4, A and  B), consistent with the notion that EGFR cross-linking is due to intermolecular dityrosine formation. Moreover, irreversible EGFR cross-linking was also observed after incubation of A431 cells with 100 M H 2 O 2 in the presence of HRP (Fig. 4C), or other heme peroxidases such as myeloperoxidase (not shown).
In the presence of these heme peroxidases, H 2 O 2 is known to be capable of oxidizing tyrosine via a one-electron mechanism to form 3,3Ј-dityrosine (46). In contrast, no detectable covalent EGFR cross-linking was observed in response to H 2 O 2 or HOCl, oxidants that are unable to cause significant oxidation of tyrosine to 3,3Ј-dityrosine. Collectively, these findings strongly suggest involvement of intermolecular 3,3Ј-dityrosine cross-links in covalent EGFR dimerization by ONOO Ϫ and HRP/H 2 O 2 . Formation of protein tyrosyl radicals was recently demonstrated in human blood plasma after reaction with ONOO Ϫ using electron paramagnetic resonance spectroscopy (47). As protein tyrosyl radicals can be relatively long lived with halflives of up to several minutes, protein cross-linking by combination of two protein tyrosine radicals is highly feasible.
Oxidant-induced EGFR Cross-linking Depends on Receptor Stimulation-Ligand-induced activation of EGFR is known to be mediated via receptor dimerization, which enhances ligand affinity (48 -50). A431 cells release transforming growth factor-␣ as an endogenous ligand for EGFR (51), which stimulates autocrine receptor activation and cell proliferation. The forma-tion of dimerized EGFR during receptor activation has been visualized with the use of chemical cross-linking agents such as disuccinimidyl suberate, EDAC, or glutaraldehyde (21,52,53). The addition of 10 mM EDAC to unstimulated A431 cells was found to result in cross-linking of a small fraction of the EGFR population, reflecting basally activated EGFR by autocrine receptor stimulation. Furthermore, SDS-PAGE analysis demonstrated that EDAC-cross-linked EGFR comigrates with the ONOO Ϫ -induced EGFR dimer, and EGFR-cross-linking by ONOO Ϫ and EDAC were found to be additive (Fig. 5A). These findings indicate that both agents act by covalently cross-linking ligand-induced EGFR dimers. Similar results were also obtained with A431 plasma membranes treated with ONOO Ϫ , HRP/H 2 O 2 , or EDAC (not shown).
Increased EGFR dimerization by cell stimulation with EGF has been demonstrated with the use of various cross-linking agents (21,52,53) and was confirmed in the present study. The extent of ONOO Ϫ -induced EGFR cross-linking was similarly found to be enhanced after prestimulation of A431 cells with 20 nM EGF (Fig. 5B). Additionally, significantly less EGFR crosslinking by ONOO Ϫ was detected after prolonged serum starvation of A431 cells, which diminishes EGFR activation and autophosphorylation. Finally, experiments in which receptorbound ligands were removed by acid wash treatment revealed that, whereas EGF-dependent receptor autophosphorylation and dimerization was diminished after acid wash treatment, ONOO --induced EGFR cross-linking could not be reversed (not shown).
Potential Effects on EGFR Activation-The apparent relationship between ligand-induced EGFR activation/dimerization and irreversible cross-linking by ONOO Ϫ or HRP/H 2 O 2 suggests that these oxidant systems preferentially cross-link activated EGFR. Consistent with this notion, determination of the extent of EGFR tyrosine phosphorylation, by quantitation of immunostaining of monomeric and dimeric EGFR with either PY-20 or ␣-EGFR antibodies, demonstrated that covalently cross-linked EGFR, by either ONOO Ϫ or HRP/H 2 O 2 , was more extensively tyrosine-phosphorylated than the monomeric form (Fig. 6). Similarly, covalently cross-linked EGFR using EDAC has also been found to have a higher degree of autophosphorylation and possesses higher tyrosine kinase activity than the monomeric form (21,22). It is recognized, however, that this argument is based on assumed similar binding affinities of these antibodies to both the monomeric and dimeric form of EGFR.
To study potential changes in EGFR signaling by exposure of A431 cells to ONOO -, we investigated the kinetics of EGFR phosphorylation as well as phosphorylation of a downstream substrate for EGFR, PLC-␥1. As demonstrated in Fig. 7, EGFinduced EGFR tyrosine phosphorylation displayed similar ki- netics before or after exposure of A431 cells to ONOO -. At the same time, the kinetics of PLC-␥1 phosphorylation by EGF stimulation was dramatically affected by ONOOpreexposure, indicating significant changes in downstream signal propagation. As shown, the t 1/2 of EGF-induced phosphorylation of PLC-␥1 is increased from Ͻ1 min to 2-5 min after exposure to ONOO -. Analysis with an ␣-nitrotyrosine antibody indicated that exposure of A431 cells to up to 1 mM ONOOdid not result in detectable tyrosine nitration of PLC-␥1, indicating that the altered kinetics of PLC-␥1 tyrosine phosphorylation was not due to tyrosine nitration within the enzyme (not shown). Recently, H 2 O 2 has also been shown to affect the kinetics of PLC-␥1 activation and its downstream signaling (54), indicating a general susceptibility of EGFR-mediated signaling pathways to oxidant stress. Therefore, the change in the kinetics of PLC-␥1 tyrosine phosphorylation may be independent of covalent EGFR cross-linking by ONOOand may represent additional effects of ONOOon EGFR. Future studies are under way to more rigorously investigate the possible routes by which ONOOand related oxidants may affect EGFR and its downstream signaling.
Conclusions-In summary, the results in this study demonstrate that exposure of A431 cells to nontoxic concentrations of ONOO Ϫ or other oxidizing systems, such as H 2 O 2 in the presence of heme peroxidases, causes formation of covalent EGFR dimers, most likely via intermolecular dityrosine cross-linking. Moreover, these oxidant systems appear to preferentially crosslink activated EGFR, as indicated by the enhanced degree of oxidant-induced covalent cross-linking after EGFR activation and by the increased extent of tyrosine autophosphorylation in covalent EGFR dimers. Hence, formation of inflammatory oxidants such as ONOO Ϫ or reactive intermediates formed by peroxidase-dependent mechanisms (55,56) may result in significant covalent modification and dimerization of EGFR. Although it is too early to speculate on the potential implications of such EGFR modifications by inflammatory oxidants, covalent EGFR dimerization may be of particular significance in conditions of increased EGFR expression and activation, such as after epithelial injury or in malignant tumors (57,58), situations that are also commonly associated with increased production of NO ⅐ and/or inflammatory oxidants. Irreversible EGFR dimerization may conceivably affect receptor activation and downstream signaling and could thereby alter epithelial repair processes or tumor development.  2, 4, and 6) and subsequently stimulated with 20 nM EGF (lanes 3-6) for the indicated periods of time. PLC-␥1 was immunoprecipitated from A431 cell lysates and immunoblotted with either ␣-phosphotyrosine (PY-20) or ␣-PLC-␥1. B and C, A431 cells were treated with ONOOand subsequently stimulated with 20 nM EGF for the indicated time period. The extent of tyrosine phosphorylation was assessed by densitometry, which was performed on Western immunoblots with PY-20 of PLC-␥1 (immunoprecipitated from A431 cell lysates; B) or EGFR (C), and normalized for immunostaining of either ␣-PLC-␥1 or ␣-EGFR (RK-2). Results represent the average of two or three separate experiments.