Reactive oxygen species induce reversible PECAM-1 tyrosine phosphorylation and SHP-2 binding

Matthias Maas, Ronggang Wang, Cathy Paddock, Srigiridhar Kotamraju, Balaraman Kalyanaraman, Peter J. Newman, Debra K. Newman

Abstract

Platelet endothelial cell adhesion molecule-1 (PECAM-1, CD31) functions to control the activation and survival of the cells on which it is expressed. Many of the regulatory functions of PECAM-1 are dependent on its tyrosine phosphorylation and subsequent recruitment of the Src homology (SH2) domain containing protein tyrosine phosphatase SHP-2. The recent demonstration that PECAM-1 tyrosine phosphorylation occurs in cells exposed to the reactive oxygen species hydrogen peroxide (H2O2) suggested that this form of oxidative stress may also support PECAM-1/SHP-2 complex formation. In the present study, we show that PECAM-1 tyrosine phosphorylation in response to exposure of cells to H2O2 is reversible, involves a shift in the balance between kinase and phosphatase activities, and supports binding of SHP-2 and recruitment of this phosphatase to cell-cell borders. We speculate, however, that the unique ability of H2O2 to reversibly oxidize the reactive site cysteine residues of protein tyrosine phosphatases may result in transient inactivation of the SHP-2 that is bound to PECAM-1 under these conditions. Finally, we provide evidence that PECAM-1 tyrosine phosphorylation and SHP-2 binding in endothelial cells requires exposure to an “oxidative burst” of H2O2, but that exposure of these cells to sufficiently high concentrations of H2O2 for a sufficiently long period of time abrogates binding of SHP-2 to tyrosine-phosphorylated PECAM-1. These findings support a role for PECAM-1 as a sensor of oxidative stress, perhaps most importantly during the process of inflammation.

  • phosphotyrosine
  • hydrogen peroxide
  • SHP-2
  • CD31

in addition to its adhesive capabilities, platelet endothelial cell adhesion molecule-1 (PECAM-1, CD31) has an important function in controlling the activation and survival of cells that express this member of the immunoglobulin (Ig) superfamily. Within its cytoplasmic tail, PECAM-1 contains two immunoreceptor tyrosinebased inhibitory motifs (ITIMs) centered around tyrosine residues 663 and 686, each of which is tyrosine phosphorylated on cellular stimulation (14). Many different stimuli have been shown to induce tyrosine phosphorylation of PECAM-1, which enables PECAM-1 to recruit proteins that contain Src homology (SH2) domains. The preferred binding partner for tyrosine-phosphorylated PECAM-1 is the SH2 domain containing protein tyrosine phosphatase SHP-2, and PECAM-1/SHP-2 complexes function to inhibit signal transduction mediated by receptors that contain immunoreceptor tyrosine-based activating motifs (14).

During the past decade, an expanding body of evidence has established that reactive oxygen species are not only toxic products of aerobic metabolism but also function as highly controlled regulators of cell signaling that control physiological processes, such as cell development and proliferation, and pathophysiological processes, such as inflammation and ischemia-reperfusion injury (46, 13, 17). Many cells express enzyme systems that are comparable to the neutrophil NADPH oxidase, allowing them to generate reactive oxygen species (6). In the presence of water, the highly reactive superoxide anions that are produced are rapidly converted to the more mild oxidant H2O2. In contrast to superoxide anion, H2O2 can freely diffuse across the membrane of a cell and exert its oxidizing activity within the intracellular environment (4, 17). Recent reports have shown that protein tyrosine phosphatases contain reactive site cysteine residues that are reversibly oxidized by H2O2 (6, 17), resulting in their transient inactivation (1, 3, 12, 16). In this manner, the balance of tyrosine phosphorylation, which is catalyzed by protein tyrosine kinases, and dephosphorylation, which is catalyzed by protein tyrosine phosphatases, is shifted in favor of kinase activity, resulting in a higher level of tyrosine-phosphorylated proteins within cells exposed to oxidizing conditions (23).

The recent demonstration that PECAM-1 tyrosine phosphorylation occurs in endothelial cells exposed to the reactive oxygen species H2O2 (9) suggested to us that PECAM-1 might serve as a sensor of a cell's oxidative environment. The purpose of the present study, therefore, was to explore the mechanism underlying PECAM-1 tyrosine phosphorylation in cells exposed to H2O2, to determine whether SHP-2 is efficiently recruited to tyrosine-phosphorylated PECAM-1 under these conditions, and to establish the conditions required for PECAM-1 to become phosphorylated and bind SHP-2 in endothelial cells exposed to this type of oxidative stress. Our findings support a role for reactive oxygen species regulation of PECAM-1-mediated signal transduction.

MATERIALS AND METHODS

Antibodies. Anti-PECAM-1 murine monoclonal antibodies PECAM-1.3 (specific for Ig homology domain 1) and rabbit polyclonal antibody SEW-16 have been previously described (8). The phosphotyrosine-specific monoclonal antibody PY-20 directly conjugated with horseradish peroxidase was purchased from Zymed (South San Francisco, CA). Polyclonal antibodies against the NH2- and COOH-terminal SH2 domains of the protein-tyrosine phosphatase SHP-2 were obtained from Santa Cruz Biotechnology (Santa Cruz, CA). Peroxidase-conjugated goat anti-mouse and donkey anti-rabbit antibodies were purchased from Jackson Immunoresearch Laboratories (West Grove, PA)

Reagents. Protease inhibitors, including PMSF, leupeptin, aprotinin, prostaglandin E1, Triton X-100, bovine serum albumin, catalase, H2O2, glucose oxidase, and PMA, were purchased from Sigma (St. Louis, MO). Protein G-Sepharose was purchased from Amersham Biosciences (Piscataway, NJ). Phosphatase inhibitor cocktail (final concentration in mM: 2 imidazole, 1 sodium fluoride, 1.15 sodium molybdate, 1 sodium orthovanadate, and 4 sodium tartrate dihydrate) and protease inhibitor cocktail [final concentration 0.5 mM 4-(2-aminoethyl)benzenesulfonyl fluoride (AEBSF) hydrochloride, 150 nM aprotinin, 1 μM E64 protease inhibitor, 0.5 mM EDTA disodium, and 1 μM leupeptin hemisulfate] were obtained from Calbiochem (San Diego, CA).

Cell culture and construction of mutant forms of PECAM-1. Human embryonic kidney-293 (HEK-293) cells expressing wild-type or tyrosine → phenylalanine mutant forms of PECAM-1 have been described previously (8). A cDNA fragment encoding the two NH2-terminal SH2 domains of SHP-2 (amino acids 1–223) was generated by PCR and cloned into the green fluorescent protein (GFP) fusion vector pEGFP (Clontech; Palo Alto, CA). Plasmid constructs were stably transfected, separately or in various combinations, into REN cells, a human malignant mesothelioma cell line (20), using lipofectamine (GIBCO-BRL; Gaithersburg, MD). Human umbilical vein endothelial cells (HUVEC) were cultured at 37°C in a humidified atmosphere of 5% CO2-95% air in RPMI buffer containing 15% donor horse serum, 50 μg/ml endothelial cell growth supplement (BD Bioscience; Bedford, MA), and 6.5 U/ml heparin. Cells were used at third passage after growing to confluence. Before each experiment the medium was replaced by RPMI buffer containing 1% donor horse serum.

Stimulation of cells with H2O2. Transfected cells in suspension (2 × 106/ml) were incubated with the specified concentration of H2O2 for 10 min. After a short centrifugation, cell pellets were solubilized in lysis buffer containing 2% Triton X-100, 10 mM EGTA, 15 mM HEPES, 145 mM NaCl, 0.1 mM MgCl2, 1 mM PMSF, 20 μg/ml leupeptin, and 2 mM sodium orthovanadate (pH 7.4), at 4°C for 30 min. HUVEC were stimulated in 100-mm dishes by adding indicated amounts of H2O2 or glucose oxidase to the medium. The medium contained 11.1 mmol/l glucose as a substrate for glucose oxidase. The level of H2O2 generated by glucose-glucose oxidase was measured as described previously (22). After treatment, the medium was removed and cells were lysed in 1.5 ml lysis buffer (1% Triton X-100, 150 mM NaCl, 20 mM Tris, pH 8.0, phosphatase inhibitor cocktail, and protease inhibitor cocktail) for 30 min at 4°C. Lysates were clarified by centrifugation for 30 min at 14,000 g, and the supernatant was subjected to immunoprecipitation analysis.

Immunoprecipitation and Western blotting. Detergent cell lysates were immunoprecipitated with 10 μg PECAM-1.3 for transfected cells or 25 μg PECAM-1.3 for HUVEC overnight at 4°C. The immune complexes were collected on protein G Sepharose beads, separated by SDS-PAGE under reducing conditions, and transferred to polyvinylidene fluoride membrane (Millipore; Bedford, MA). After being blocked overnight at 4°C in 20 mM Tris, 100 mM NaCl, 0.05% Tween 20, and 3% BSA, the membrane was incubated with primary antibodies (PECAM-1.3, SHP2 or PY-20) for 2 h, washed, incubated with secondary antibodies (anti-mouse or anti-rabbit), and developed using enhanced chemiluminescence (Amersham; Arlington Heights, IL) or Super Signal West Pico (Pierce; Rockford, IL). Images were obtained on a Kodak Image Station 2000R or by exposure to film (Eastman Kokak; Rochester, NY).

Src in vitro kinase assay and determination of total cellular phosphatase activity. PECAM-1-transfected HEK-293 cells were starved for 24 h in media containing 0.5% serum. The cells were then stimulated with varying concentrations of H2O2 for 10 min at 37°C and lysed in lysis buffer [1% Triton X-100, 50 mM HEPES (pH 7.5) 150 mM NaCl, 10% glycerol, 1.5 mM MgCl2, 1 mM EGTA, 1 mM PMSF, 10 μg/ml aprotinin and leupeptin, 2 mM sodium orthovanadate]. Equivalent amounts of cell extract proteins were immunoprecipitated with 4 μg anti-Src monoclonal antibody, clone GD-11 (Upstate Biotech; Lake Placid, NY) for 3 h at 4°C. Immunoprecipitates were collected on BSA-blocked protein G agarose beads, washed twice with lysis buffer, and then twice with kinase buffer [20 mM HEPES (pH 7.4) 10% wt/vol glycerol, 150 mM NaCl, 10 mM MgCl2, and 10 mM MnCl2]. To initiate kinase reactions, the washed precipitates were incubated with 20 μl of kinase buffer containing 20 μM ATP and 30 μCi [γ-32P]ATP (6,000 Ci/mmol, NEN; Boston, MA). The reactions were incubated at 37°C for 15 min and stopped by the addition of 1 ml ice-cold TNE buffer [containing 110 mM Tris (pH 7.4) 100 mM NaCl, and 1 mM EDTA]. The 32P-labeled immunoprecipitates were eluted with 30 μl of 2 × SDS-PAGE sample buffer, heated at 100°C, separated by SDS-PAGE, and visualized by autoradiography. Band intensities were determined densitometrically using an AMBIS scanner (San Diego, CA). Cellular phosphatase activity was determined at 37°C in a buffer containing (in mM) 100 HEPES (pH 7.4) 150 NaCl, 1 EDTA, and 1 DTT using p-nitrophenyl phosphate (PNPP) as a substrate. Briefly, H2O2-stimulated cells were lysed in lysis buffer containing 145 mM NaCl, 15 mM HEPES (pH 7.4) 10 mM EGTA, 1% Triton X-100, 0.1 mM MgCl2, 1 mM PMSF, 20 μg/ml leupeptin, and 2 mM sodium orthovanadate and were incubated with 10 mM PNPP at 37°C for 30 min to 2 h, and the absorbance of the reaction product p-nitrophenolate was then measured at 405 nm in an ELISA plate reader. Phosphatase activity was expressed as optical density at 405 mm per microgram protein lysate.

Laser-scanning confocal microscopy. REN cell lines stably expressing wild-type PECAM-1 and/or a GFP fusion protein containing the SH2 domains of SHP-2 were seeded in Falcon culture slides (BD Biosciences; Bedford, MA), allowed to grow for 2–3 days until they reached confluence, and then treated with 1 mM H2O2 at 37°C for 30 min. After fixation with 2% paraformaldehyde for 20 min at 22°C, the cells were permeabilized with 0.5% Nonidet P-40 in PBS for 3 min and then incubated for 60 min with the anti-PECAM-1 monoclonal antibody PECAM-1.3 (5 μg/ml) followed by a Texas redconjugated antibody fragment [F(ab′)2] of goat anti-mouse IgG (H + L) (Jackson Immune Research; West Grove, PA). After being washed, the chamber walls were removed, and slides were mounted in Vectashield mounting media (Vector Laboratories; Burlingame, CA). Specimens were examined by using a MRC 600 confocal laser imaging equipped with a krypton-argon laser (Bio-Rad) and a Nikon epifluorescent microscope (Nikon; Melville, NY). Individual excitation filters were used for Texas red (PECAM-1) and FITC (SHP-2/GFP) fluorescence, and emissions were captured using dichroic reflector blocks (FITC excitation: 488 nm, emission 522 nm; Texas red excitation: 568 nm, emission 585 nm). Projections for each individual filter were digitally imaged using CoMOS software (Bio-Rad) by taking a series of 12–25 optical sections for each field. Confocal Assistant software (freeware by Todd Clark Brelje) was used to combine the entire Z series into a stacked projection, to assign colors, and to overlay and process the final image. Colocalized fluorophores are represented by yellow pixels.

Statistical analysis. In immunoprecipitation experiments, chemiluminescence intensity for each band was determined by using Image 1D software (Eastman Kodak) and normalized for PECAM-1 antigen loading control, and the maximal net intensity of each experiment was adjusted to 100%. The number of experimental determinations (n) is equal to the number of separate experiments. Comparisons between groups were performed with one-way analysis of variance and a Dunnett post hoc test. The acceptable level of significance was P < 0.05. Data in figures are presented as means ± SD.

RESULTS

PECAM-1 becomes tyrosine phosphorylated on exposure of cells to oxidizing conditions. Previous studies have shown that PECAM-1 can become tyrosine phosphorylated in response to variety of cellular stimuli and that following this event, the protein-tyrosine phosphatase SHP-2 associates with the PECAM-1 cytoplasmic domain (14). Among the stimuli recently shown to induce PECAM-1 tyrosine phosphorylation in endothelial cells is exposure to H2O2 (9). However, the mechanism by which PECAM-1 becomes tyrosine phosphorylated in cells exposed to H2O2 and whether such tyrosine phosphorylation events support recruitment of SHP-2 are currently unknown. To examine the first of these questions, PECAM-1 tyrosine phosphorylation was evaluated following exposure to H2O2 of HEK-293 cells expressing wild-type PECAM-1, or those expressing a mutant form of PECAM-1 in which the ITIM tyrosine residues at positions 663 and 686 were substituted with phenylalanine (PECAM-1-Y663,686F). As shown in Fig. 1A, H2O2 induced PECAM-1 tyrosine phosphorylation in a dose-dependent manner in that PECAM-1 tyrosine phosphorylation was observed on exposure of cells to as little as 10 μM H2O2 and increased with increasing concentrations of H2O2 up to 3 mM. PECAM-1 tyrosine phosphorylation was abrogated by the substitution of the tyrosine residues at positions 663 and 686 with phenylalanine (Fig. 1A), indicating that PECAM-1 tyrosine phosphorylation in cells exposed to H2O2 occurs on the two ITIM tyrosine residues, as previously reported (9).

Fig. 1.

Dose-dependent phosphorylation of PECAM-1 immunoreceptor tyrosine-based inhibitory motifs (ITIM) tyrosine residues in response to oxidative stress coincides with inhibition of cellular phosphatase activity but not with enhanced kinase activity. A: human embryonic kidney (HEK)-293 cells expressing either wild-type PECAM-1 (WT PECAM-1/293, left) or a mutant form of PECAM-1 in which the ITIM tyrosine residues at positions 663 and 686 were substituted with phenylalanine (PECAM-1-Y663,686F/293, right) were exposed for 10 min to concentrations of H2O2 ranging from 10 μM to 3 mM (lanes 3–6). PECAM-1 was immunoprecipitated (IP) from cell lysates using the PECAM-1-specific monoclonal antibody PECAM-1.3, and the resulting immunoprecipitates were analyzed sequentially for tyrosine phosphorylation using the phosphotyrosine-specific monoclonal antibody PY20 (top) and for antigen loading using the rabbit polyclonal anti-PECAM-1 antibody SEW-16 (bottom). Negative controls included nonspecific IgG immunoprecipitates from HEK-293 transfectants cultured for 10 min in the presence of 1 mM H2O2 (lane 1) and PECAM-1.3 immunoprecipitates from HEK-293 transfectants cultured for 10 min in the absence of H2O2 (lane 2). Note that WT PECAM-1, but not PECAM-1-Y663,686F, becomes tyrosine phosphorylated following exposure of cells to H2O2, suggesting that in PECAM-1 the two ITIM tyrosine residues are normally targets for phosphorylation in cells exposed to oxidative stress. B: HEK-293 cells expressing WT PECAM-1 were treated with the indicated concentrations of H2O2 for 10 min at 37°C and then lysed. Cellular phosphatase activity was assessed using total cell lysates, and Src kinase activity was assessed using anti-Src immunoprecipitates as described in materials and methods. Note that H2O2 inhibited cellular phosphatase activity in a dose-dependent manner but did not affect Src kinase activity. OD450, optical density at 450 nm.

The extent to which a given protein is phosphorylated on tyrosine residues is determined by the relative levels of tyrosine kinase and balancing tyrosine phosphatase activities. One of the best-described properties of H2O2 is its ability to transiently inactivate tyrosine phosphatases via reversible oxidation of their active site cysteine residues (13). Thus we hypothesized that the PECAM-1 tyrosine phosphorylation that we observed in cells exposed to H2O2 may be due to a low level of constitutive kinase activity that generated phosphotyrosine residues on PECAM-1 that were preserved due to the inactivation of cellular phosphatases by H2O2. To test this hypothesis, we measured tyrosine kinase and phosphatase activities in lysates of HEK-293 cells exposed to varying concentrations of H2O2. Src family kinases have previously been shown to phosphorylate PECAM-1 both in in vitro kinase assays and in overexpression systems (2, 10, 11, 15, 19), and the SH2 domain of Src has been shown to bind to a tyrosine-phosphorylated form of the PECAM-1 cytoplasmic domain (10, 15). Therefore, the kinase activity that we measured was that of Src. As shown in Fig. 1B, protein tyrosine phosphatase activity was inhibited at concentrations of H2O2 that induced PECAM-1 tyrosine phosphorylation (see Fig. 1A), whereas Src kinase activity was unaffected by H2O2. Furthermore, tyrosine-phosphorylated PECAM-1 became rapidly dephosphorylated following removal of H2O2 from the culture medium (Fig. 2A), consistent with the transient nature of phosphatase inactivation by H2O2 (Fig. 2B). From these data, we conclude that PECAM-1 tyrosine phosphorylation induced by exposure of cells to H2O2 results from inhibition of phosphatase rather than stimulation of Src kinase activity.

Fig. 2.

PECAM-1 ITIM tyrosine phosphorylation in cells exposed to oxidative stress is reversible. A: HEK-293 cells transfected with WT PECAM-1 (left, lanes 3–6) or PECAM-1-Y663,686F (right) were exposed to 1 mM H2O2 for 10 min and then cultured for varying amounts of time in the absence of H2O2. Untransfected HEK-293 cells (left, lanes 1 and 2) served as a negative control (CTR). PECAM-1 was immunoprecipitated from cell lysates using PECAM-1.3 and analyzed for tyrosine phosphorylation and antigen loading as described for Fig. 1. Note that the levels of PECAM-1 tyrosine phosphorylation are maximal immediately after exposure of cells to H2O2 and decline with increasing amounts of time after removal of H2O2 from the culture. B: HEK-293 cells expressing WT PECAM-1 were treated with 1 mM H2O2 for 10 min at 37°C and then cultured for varying amounts of time in the absence of H2O2. Cellular phosphatase activity was assessed using total cell lysates as described in materials and methods. Note that cellular phosphatase activity increases transiently coincident with the early phase of PECAM-1 dephosphorylation and SHP-2 dissociation.

SHP-2 coprecipitates with PECAM-1 and is recruited to the borders between PECAM-1 expressing cells on exposure to H2O2. Preliminary studies revealed that exposure of cells to H2O2 results in the tyrosine phosphorylation of multiple cellular proteins in addition to PECAM-1 (data not shown). To address the question as to whether, in the presence of potentially competitive docking proteins, SHP-2 preferentially associates with PECAM-1, we assessed the ability of anti-PECAM-1 antibodies to coprecipitate SHP-2 from cell lysates of H2O2-treated HEK-293 cells expressing wild-type PECAM-1 or PECAM-1-Y663,686F. As shown in Fig. 3, SHP-2 coprecipitated with tyrosine-phosphorylated PECAM-1 in both H2O2- and pervanadate-stimulated cells in a manner that required cytoplasmic tyrosine residues 663 and 686.

Fig. 3.

Protein-tyrosine phosphatase SHP-2 associates with tyrosine-phosphorylated PECAM-1 following exposure of cells to oxidative stress. HEK-293 cells transfected with WT PECAM-1/293 (left) or PECAM-1-Y663,686F/293 (right) were treated with buffer, 3.5 mM H2O2, or 2 mM pervanadate at 37°C for 5 min. Untreated cells served as negative controls. PECAM-1 immunoprecipitates were prepared as described in Fig. 1, probed sequentially with PY-20 to assess the tyrosine phosphorylation state of PECAM-1 (top), anti-SHP-2 to assess association of SHP-2 with PECAM-1 (middle), and SEW 16 to indicate the level of antigen loading (bottom). Note that SHP-2 coprecipitated with tyrosine-phosphorylated PECAM-1 from both pervanadate- and H2O2-treated WT PECAM-1-transfected 293 cells but not from untreated cells or cells expressing PECAM-1-Y663,686F.

As a second approach to assess the ability of SHP-2 to associate with PECAM-1 in cells exposed to H2O2, we performed two-color laser scanning confocal microscopic analysis of stably transfected REN cell lines that expressed either the two SH2 domains of SHP-2 fused to GFP (SHP-2/GFP), wild-type PECAM-1 and SHP-2/GFP, or PECAM-1-Y663,686F and SHP-2/GFP. Cells were grown to confluence, exposed briefly to 1 mM H2O2, and imaged to determine the ability of PECAM-1 (red) to recruit SHP-2 (green) to the intercellular junction. As shown in Fig. 4, SHP-2/GFP was distributed diffusely throughout the cytoplasm in cells lacking PECAM-1 or in quiescent cells expressing wild-type PECAM-1. However, on exposure of wild-type PECAM-1-containing cells to oxidative stress, SHP-2 redistributed to cell-cell borders (Fig. 4). Colocalization was dependent on the two ITIM tyrosine residues within PECAM-1, because cells expressing PECAM-1-Y663,686F were unable to effectively concentrate SHP-2/GFP at sites of cell contact. These data demonstrate that junctional PECAM-1 functions to recruit SHP-2 to the borders between PECAM-1-expressing cells on their exposure to H2O2, even in the presence of multiple tyrosine phosphorylated SHP-2 binding proteins.

Fig. 4.

PECAM-1-dependent recruitment of SHP-2 to sites of cell-cell contact following exposure of cells to oxidative stress. REN cells were stably transfected with a plasmid encoding the SH2 domains of SHP-2 fused to green fluorescent protein (GFP) (SHP-2/GFP) ± either WT PECAM-1 or PECAM-1-Y663,686F. Cells were seeded on culture slides and exposed to 1 mM H2O2 for 30 min. After fixation and permeabilization, PECAM-1 was stained using the murine monoclonal antibody PECAM-1.3 followed by Texas red-conjugated antibody fragments [F(ab′)2] of goat anti-mouse IgG. The location of the SH2 domains of SHP-2 was visualized using the green fluorescence of the GFP portion of the SHP-2/GFP fusion protein. Both green and red images were captured using an MRC 600 confocal laser-scanning microscope and then digitally processed using COMOS and Confocal Assistant software. Note that PECAM-1 (red), with or without its ITIM tyrosine residues, concentrates at cell-cell borders independently of exposure of cells to oxidizing conditions. SHP-2 (green), in contrast, redistributes from the cytoplasm to the cell junctions on exposure of cells to oxidative stress, provided that PECAM-1 is present and contains phosphorylatable tyrosine residues within its cytoplasmic ITIMs.

H2O2 induces association of SHP-2 with tyrosine phosphorylated PECAM-1 in endothelial cells in a time- and dose-dependent manner. After demonstrating that H2O2 induces PECAM-1 tyrosine phosphorylation and SHP-2 recruitment in transfected cells, we sought to confirm these findings in a more physiological system. We chose HUVEC for this purpose because they highly express PECAM-1. Dose-dependent PECAM-1 tyrosine phosphorylation and SHP-2 association were observed in HUVEC at concentrations of H2O2 ranging from 300 μM to 1 mM (Fig. 5, A and B). Interestingly, however, at concentrations of H2O2 >1 mM, PECAM-1 remained heavily tyrosine phosphorylated but coprecipitated increasingly less SHP-2. As shown in Fig. 5C, protein tyrosine phosphatase activity was inhibited at concentrations of H2O2 that induced PECAM-1 tyrosine phosphorylation. HUVEC required only 30 min of exposure to 1 mM H2O2 to exhibit PECAM-1 tyrosine phosphorylation and SHP-2 binding, which persisted as long as H2O2 remained in the culture medium (Fig. 6, A and B). As shown in Fig. 6C, the kinetics of inhibition of protein tyrosine phosphatase activity coincided with PECAM-1 tyrosine phosphorylation. As was observed in PECAM-1-transfected HEK-293 cells, H2O2-induced PECAM-1 tyrosine phosphorylation and SHP-2 coprecipitation in HUVEC were reversible because both of these events began to return to basal levels within 45 min after removal of H2O2 from the culture medium (Fig. 7, A and B).

Fig. 5.

Dose dependence of H2O2-induced PECAM-1 tyrosine phosphorylation, PECAM-1/SHP-2 complex formation, and inhibition of cellular phosphatase activity in human umbilical vein endothelial cells (HUVEC). HUVEC were treated with a bolus of H2O2 in increasing concentrations for 1 h at 37°C and lysed. PECAM-1 was immunoprecipitated using the monoclonal antibody PECAM-1.3 (A and B). A: immunoprecipitates were analyzed for PECAM-1 tyrosine phosphorylation using the phosphotyrosine-specific monoclonal antibody PY20 (top), for SHP-2 using a rabbit polyclonal anti-SHP-2 (middle), and for antigen loading using the mouse monoclonal antibody PECAM-1.3 (bottom). B: band intensity data from five independent experiments were obtained, combined, and subjected to statistical analysis. Statistical significance between negative control and treatment groups, respectively: *P < 0.05 and **P < 0.01. Note that PECAM-1 becomes tyrosine phosphorylated and binds SHP-2 in cells exposed to all concentrations of H2O2 tested but that reduced amounts of SHP-2 are bound to tyrosine-phosphorylated PECAM-1 in cells exposed to H2O2 concentrations >1 mM. %max, Percentage of maximum. C: HUVECs were treated with the indicated concentrations of H2O2 for 1 h at 37°C and then lysed. Cellular phosphatase activity was assessed in total cell lysates as described in materials and methods. Note that H2O2 inhibited cellular phosphatase activity in a dose-dependent manner, coincident with induction of PECAM-1 tyrosine phosphorylation and SHP-2 binding.

Fig. 6.

Kinetics of H2O2-induced PECAM-1 tyrosine phosphorylation, PECAM-1/SHP-2 complex formation, and inhibition of cellular phosphatase activity in HUVECs. HUVEC were treated with 1 mM H2O2 for the indicated amounts of time at 37°C and lysed. PECAM-1 was immunoprecipitated using the monoclonal antibody PECAM-1.3 (A and B). A: immunoprecipitates were analyzed for PECAM-1 tyrosine phosphorylation using the phosphotyrosine-specific monoclonal antibody PY20 (top), for SHP-2 using a rabbit polyclonal anti-SHP-2 (middle), and for antigen loading using the mouse monoclonal antibody PECAM-1.3 (bottom). B: band intensity data from four independent experiments were obtained, combined, and subjected to statistical analysis. Statistical significance between negative control and treatment groups, respectively: *P < 0.05 and **P < 0.01. Note that PECAM-1 tyrosine phosphorylation and SHP-2 binding requires at least 30 min of exposure of HUVECs to 1 mM H2O2. C: HUVECs were treated with 1 mM H2O2 for the indicated amounts of time at 37°C and then lysed. Cellular phosphatase activity was assessed in total cell lysates as described in materials and methods. Note that H2O2 inhibited cellular phosphatase activity in a time-dependent manner coincident with induction of PECAM-1 tyrosine phosphorylation and SHP-2 binding.

Fig. 7.

PECAM-1 ITIM tyrosine phosphorylation and SHP-2 binding in HUVECs exposed to oxidative stress is reversible. After stimulation with 1 mM H2O2 for 60 min, H2O2-containing medium was replaced with fresh medium lacking H2O2, and HUVEC were cultured for additional indicated amounts of time before lysis. Negative controls include cells not exposed to H2O2. PECAM-1 was immunoprecipitated using the monoclonal antibody PECAM-1.3 (A and B). A: immunoprecipitates were analyzed for PECAM-1 tyrosine phosphorylation using the phosphotyrosine-specific monoclonal antibody PY20 (top), for SHP-2 using a rabbit polyclonal anti-SHP-2 (middle), and for antigen loading using the mouse monoclonal antibody PECAM-1.3 (bottom). B: band intensity data from four independent experiments were obtained, combined, and subjected to statistical analysis. Statistical significance between negative control and treatment groups respectively: *P < 0.05 and **P < 0.01. Note that the PECAM-1 tyrosine phosphorylation and SHP-2 binding after H2O2 treatment is reversible with time. C: HUVECs were treated with 1 mM H2O2 for 1 h at 37°C and then cultured for varying amounts of time in the absence of H2O2. Cellular phosphatase activity was assessed using total cell lysates as described in materials and methods. The fold change in phosphatase activity relative to that present immediately after removal of H2O2 was calculated for each time point. Results are presented as means ± SD of three independent experiments. Note that cellular phosphatase activity increases slightly and transiently coincident with the early phase of PECAM-1 dephosphorylation and SHP-2 dissociation.

Glucose oxidase catalyzes the oxidation of d-glucose to d-gluconolactone in a reaction that results in the production of H2O2. Therefore, addition of glucose oxidase to glucose-containing media enables cells in the culture to be exposed to a continuous source of H2O2. Dose-dependent PECAM-1 tyrosine phosphorylation and SHP-2 binding were observed in HUVEC cultured in the presence of glucose oxidase ranging from 200 to 500 mU/ml (Fig. 8A). Independent experiments revealed that 200 mU/ml of glucose oxidase generated 15 μM H2O2/min (data not shown); therefore, culture of cells for 2 h in the presence of 200 or 500 mU/ml glucose oxidase might be expected to give rise to concentrations of H2O2 of 1.8 and 4.5 mM, respectively, assuming no detoxification of the generated H2O2 by the cultured cells. These concentrations are compatible with those required to induce PECAM-1 tyrosine phosphorylation and SHP-2 binding in HUVEC exposed to a bolus of H2O2 (see Fig. 5). Interestingly, however, in HUVEC exposed to 500 mU/ml glucose oxidase for >2 h, PECAM-1 tyrosine phosphorylation levels remained high, but its ability to coprecipitate SHP-2 decreased (Fig. 8B).

Fig. 8.

Continuous generation of H2O2 by glucose-glucose-oxidase (Glc Ox) results in PECAM-1 tyrosine phosphorylation and SHP-2 association. A: HUVEC were incubated with the indicated concentrations of glucose oxidase for 2 h. Cells were lysed, and PECAM-1 was immunoprecipitated using the monoclonal antibody PECAM-1.3. Immunoprecipitates were analyzed for PECAM-1 tyrosine phosphorylation using the phosphotyrosine-specific monoclonal antibody PY20 (top), for SHP-2 using a polyclonal rabbit anti-SHP-2 (middle), and for antigen loading using the mouse monoclonal antibody PE-CAM-1.3 (bottom). Note that SHP-2 association with tyrosine-phosphorylated PECAM-1 is observed only when HUVEC are exposed to 500 mU/ml of glucose oxidase. B: HUVEC were incubated with 200 or 500 mU/ml glucose oxidase for the indicated amounts of time. Cells were lysed, and immunoprecipitations and Western blotting were performed as described in A. C: band intensity data from three independent experiments were obtained, combined, and subjected to statistical analysis. Statistical significance between negative control and treatment groups: **P < 0.01. Note that SHP-2 association with tyrosine-phosphorylated PECAM-1 is dependent on both the dose of glucose oxidase to which the cells are exposed and the duration of exposure.

DISCUSSION

In addition to its adhesive capabilities, PECAM-1 has important intracellular signaling capabilities that function to control the activation and survival of cells (14). The best-described signaling properties of PECAM-1 are mediated by phosphorylation of two ITIM tyrosine residues at positions 663 and 686 of the PECAM-1 cytoplasmic domain, which occurs in response to exposure of cells to a number of different chemical and mechanical stimuli and enables recruitment of SH2 domain-containing proteins (14). Among the many SH2 domain-containing proteins that have been reported to associate with the PECAM-1 cytoplasmic domain, the protein tyrosine phosphatase SHP-2 is the preferred binding partner (14). The recent demonstration that PECAM-1 tyrosine phosphorylation can occur in response to exposure of endothelial cells to the reactive oxygen species H2O2 (9) suggested to us that PECAM-1 might also serve as a sensor of the oxidative environment in which a cell finds itself. In the present study, we reproduce the observation that PECAM-1 becomes phosphorylated on ITIM tyrosine residues in cells exposed to H2O2 (Fig. 1). We expand on this finding by showing that PECAM-1 tyrosine phosphorylation in cells exposed to H2O2 coincides with a shift in the balance between kinase and phosphatase activities (Fig. 2) and is reversible after the removal of H2O2 from the cellular environment (Figs. 2 and 5). We also show that, although exposure to H2O2 results in the creation of multiple phosphotyrosine-containing proteins, SHP-2 efficiently binds tyrosine-phosphorylated PECAM-1 (Figs. 3, 5, 7, and 8) and is recruited by PECAM-1 to cell-cell borders (Fig. 4). We establish that PECAM-1 tyrosine phosphorylation and SHP-2 binding in endothelial cells requires exposure to relatively high concentrations of H2O2, accomplished either with administration of a bolus of H2O2 (Figs. 5, 6, 7) or with provision of enzymatic conditions suitable for generation of a large amount of H2O2 within a brief period of time (Fig. 8). Higher concentrations of H2O2 were required for the induction of PECAM-1 tyrosine phosphorylation and SHP-2 binding in cultured endothelial cells relative to transfected HEK-293 cells, which may reflect differences in the culture conditions to which the cells were exposed. Alternatively, because the half-life of H2O2 is critically dependent on the redox equilibrium inside the cell (17), our findings may reflect differences in the ability of these cell types to detoxify H2O2. Although the concentrations of H2O2 to which endothelial cells might be exposed under physiological conditions are unknown, the high concentrations that result in PECAM-1 tyrosine phosphorylation and SHP-2 binding are likely to be present in pathological conditions such as ischemia-reperfusion injury or as a consequence of an “oxidative burst” encountered at a site of inflammation (4). Finally, we document, for the first time, a condition that supports PECAM-1 tyrosine phosphorylation but not SHP-2 binding, i.e., exposure of cells to sufficiently high concentrations of H2O2 for a sufficiently long period of time (Figs. 5, 6, and 8). These findings support a role for endothelial cell PECAM-1 as a sensor of oxidative stress, perhaps most importantly during the process of inflammation.

H2O2 is often used experimentally as an exogenous source of reactive oxygen species, which are normally generated during aerobic cellular metabolism, as a consequence of cardiovascular disease or at sites of inflammation (reviewed in Refs. 5, 7, 13, and 17). The amount of reactive oxygen species to which a cell is exposed under any of these conditions is tightly regulated by antioxidants present in and around the cell (17, 18). We found that different amounts of H2O2 were required to induce PECAM-1 tyrosine phosphorylation in different cell types and that endothelial cells, in particular, required exposure to a relatively high concentration of H2O2. This requirement predicts that extensive tyrosine phosphorylation of endothelial cell PECAM-1 would be found only when the H2O2 concentrations encountered are sufficiently high as to overcome the considerable antioxidant potential of the cell. Thus PECAM-1 is more likely to function as a redox sensor in endothelial cells under pathological conditions such as cardiovascular disease or inflammation as, for example, when infiltrating neutrophils generate an oxidative burst while traversing the endothelial cell barrier.

The observation that tyrosine phosphorylation of cellular proteins is increased when cells are exposed to reactive oxygen species has been interpreted as evidence that tyrosine kinase activity is redox sensitive; however, protein tyrosine phosphatases are more likely to be direct targets for the action of H2O2 than are kinases (reviewed in Refs. 5, 17, and 18). This is because tyrosine phosphatases possess a reactive site cysteine residue that, due to the nature of surrounding amino acids, exists as a thiolate anion at neutral pH (21). This property of the reactive site cysteine enables it to contribute to the catalytic activity of the protein tyrosine phosphatase, but it also renders the sulfhydryl group susceptible to oxidation by the rather mild oxidant H2O2. Inactivation of protein tyrosine phosphatases, which normally counterbalance the basal activity of protein tyrosine kinases, is therefore likely to provide the mechanistic explanation for the increased levels of protein tyrosine phosphorylation that one observes on exposure of cells to H2O2 (reviewed in Refs. 5, 17, and 18). Our studies, which revealed that PECAM-1 tyrosine phosphorylation and SHP-2 recruitment on exposure of cells to H2O2 coincided with decreased protein tyrosine phosphatase activity but not with increased protein tyrosine kinase activity, are consistent with the interpretation that inactivation of protein tyrosine phosphatases links H2O2 exposure to PECAM-1 tyrosine phosphorylation.

Our studies also revealed that H2O2-mediated PECAM-1 tyrosine phosphorylation was reversible on removal of H2O2 from the culture medium. We predict from this observation that the phosphatases normally responsible for PECAM-1 dephosphorylation are transiently inactivated in the presence of H2O2 and resume activity in its absence. This prediction has precedence in that H2O2-mediated oxidation of cysteine residues in the reactive sites of protein tyrosine phosphatases is readily reversible in the reducing environment of the cell, such that phosphatase inactivation by H2O2 is transient in nature (17, 18). Interestingly, this line of reasoning predicts that the SHP-2 that is recruited to PECAM-1 in an H2O2-containing environment would be inactive as long as H2O2 is present. Support for this possibility has been provided by the studies of Meng et al. (12), who showed that an oxidative burst resulted in recruitment of SHP-2 to the PDGF receptor in Rat-1 cells, and that the SHP-2 recruited under these conditions was reversibly oxidized at the cysteine residue in its catalytic domain. Thus, to the extent that the signaling properties of tyrosine-phosphorylated PECAM-1 depend on the catalytic activity of bound SHP-2, the presence of H2O2 or other reactive oxygen species in the cellular environment may influence the functionality of the PECAM-1/SHP-2 complex. Efforts to determine the oxidation state and activity of SHP-2 recruited to tyrosine-phosphorylated PECAM-1 in the presence and absence of H2O2 are currently underway in our laboratory.

Finally, we found that exposure of cells to concentrations of H2O2 >1 mM resulted in the failure of SHP-2 to bind to tyrosine-phosphorylated PECAM-1. Instead, SHP-2 preferentially associated with an 85-kDa phosphotyrosine-containing protein under these conditions (data not shown). These findings document, for the first time, a condition under which PECAM-1 tyrosine phosphorylation is dissociated from SHP-2 binding. The findings that PECAM-1 remains tyrosine phosphorylated (Fig. 5, A and B) and phosphatase activity is completely inhibited (Fig. 5C) at these high concentrations of H2O2 eliminate PECAM-1 dephosphorylation by SHP-2 as a possible explanation for the failure of SHP-2 to bind to tyrosine-phosphorylated PECAM-1 at high concentrations of H2O2. An alternative explanation for these findings might involve competitive binding of SHP-2 to an alternative tyrosine-phosphorylated binding partner. Experiments are currently underway to determine the impact of this competition for available SHP-2 on the formation and function of PECAM-1/SHP-2 complexes.

In summary, we have provided evidence that exposure of endothelial cells to the reactive oxygen species H2O2 leads to reversible tyrosine phosphorylation of the ITIM tyrosine residues within the PECAM-1 cytoplasmic domain and subsequent association with SHP-2. Ongoing studies will address the extent to which these findings generalize to other types of oxidative stress and will determine the effect of oxidative stress on PECAM-1-mediated signal transduction in normal blood and vascular cell function as well as during pathological states such as inflammation and cardiovascular disease.

DISCLOSURES

This work was supported by the Blood Center Research Foundation (to D. K. Newman and P. J. Newman), National Heart, Lung, and Blood Institute Grant HL-68769 (to D. K. Newman, P. J. Newman, and B. Kotamrju), and Innovative Medizinische Forschung (Münster, Germany) Grant MA 6 2 01 02 (to M. Maas).

Acknowledgments

The authors thank Mathew J. Armstrong for technical assistance, Steven Albelda (University of Pennsylvania School of Medicine) for supplying the REN cell line, and Benjamin Neel (Harvard Institute of Medicine) for providing the SHP-2 cDNA.

Footnotes

  • The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

  • * M. Maas and R. Wang contributed equally to this study.

References

View Abstract