Reactive oxygen species signaling through regulation of protein tyrosine phosphorylation in endothelial cells.

Tyrosine phosphorylation of proteins, controlled by tyrosine kinases and protein tyrosine phosphatases, plays a key role in cellular growth and differentiating. A wide variety of hormones, growth factors, and cytokines modulate cellular tyrosine phosphorylation to transmit signals across the plasma membrane to the nucleus. Recent studies suggest that reactive oxygen species (ROS) also induce cellular protein tyrosine phosphorylation through receptor or nonreceptor tyrosine kinases. To determine whether protein tyrosine phosphorylation by ROS regulates endothelial cell (EC) metabolism and function, we exposed vascular ECs to H2O2 or H2O2 plus vanadate. This resulted in a time- and dose-dependent increase in protein tyrosine phosphorylation of several proteins (M(r) 21-200 kDa), as determined by immunoprecipitation and Western blot analysis with antiphosphotyrosine antibody. Immunoprecipitation with specific antibodies identified increased tyrosine phosphorylation of mitogen-activated protein kinases (42-44 kDa), paxillin (68 kDa), and FAK (125 kDa) by ROS. An immediate signaling response to increased protein tyrosine phosphorylation by ROS was activation of phospholipases such as A2, C, and D. Suramin pretreatment inhibited ROS stimulation of phospholipase D (PLD), suggesting a role for growth factor receptors in this activation. Further, PLD activation by ROS was attenuated by N-acetylcysteine, indicating that intracellular thiol status is critical to ROS-mediated signal transduction. These results provide evidence that ROS modulate EC signal transduction via a protein tyrosine phosphorylation-dependent mechanism.

Tyrosine phosphorylation of proteins, a balance between tyrosine kinases and protein tyrosine phosphatases (PTPs), is modulated by a variety of hormones, growth factors, and cytokines (9,10). Recent reports indicate that ROS induce cellular protein tyrosine phosphorylation through receptor and nonreceptor tyrosine kinases (11,12). H202 alone or in combination with vanadate, which generates peroxovanadium compounds, modulates intracellular calcium (13), and activates phospholipases (14)(15)(16) and mitogenactivated protein kinases (17) through modulation of protein kinase/phosphatases. However, the physiologic significance of protein tyrosine phosphorylation by ROS in endothelial cell (EC) metabolism and function is not well understood. This study was undertaken to examine the ability of ROS to modulate EC function through protein tyrosine phosphorylationdependent signaling pathways. Our data show that H202 and H202 plus vanadate (diperoxovanadate [DPV]) stimulate tyrosine phosphorylation of several EC proteins. Furthermore, our results suggest that ROS-induced stimulation of protein tyrosine phosphorylation alters downstream signaling pathways such as Ca2+ signaling and activation of phospholipases A2 (PLA2), phospholipase C (PLC), and phospholipase D (PLD), and generation of lipid-derived second messengers.  (19,20). About 2% of the added radioactivity was incorporated into total phospholipids. Labeled cells were washed in serum-free MEM and were incubated with MEM or MEM containing ROS or other agents, at concentrations and time periods as indicated, in the presence of 0.05% butanol. Incubations were terminated by the addition of 1 ml of methanol: HCI (100:1 v/v) followed by extraction of lipids in chloroform: methanol (19).

Materials
[32P]PBt formed as a result of PLD activation (19) (19). The methanolwater phase of the lipid extract containing the [3H]inositol phosphates was subjected to anion exchange chromatography in AG 1 x 8 column as described earlier (21)  The medium was removed and radioactivity was quantified by scintillation counting. In some experiments the media containing the [3H]AA metabolites were acidified with formic acid to pH 3.0 and extracted 3 times with 4.5 ml of ethylacetate. The ethylacetate fractions were evaporated under nitrogen and 6-keto prostaglandin F,a (6-keto PGFl.), the stable metabolite of prostacyclin, was separated by TLC using the upper phase of ethylacetate:isooctane:glacial acetic acid:water (90:50:20:100 by vol). Areas corresponding to free AA and 6-keto PGFia were visualized under iodine vapors using standards and radioactivity was determined. About 40% of the added [3H]AA was incorporated into the total lipids of the ECs. TLC analysis showed that 70 to 80% of the released radioactivity comigrated with authentic 6keto PGFla and 10 to 20% with AA. In some experiments, the cells were treated with MeOH:HCl (100:1 v/v) and lipids were extracted as described earlier (19).
The chloroform phase was dried under N2 and radioactivity in [3H]AA and DAG was determined after TLC, using hexane: diethylether:glacial acetic acid (50:50:1 by vol) as the developing solvent.
Immunoprecipitation, SDS-PAGE and Western Blot Analysis Cells treated with MEM or MEM containing DPV or vanadate were rinsed 3 times in ice-cold phosphate-buffered saline (PBS) containing 1 mM vanadate. Cells were scraped in 250 pl of lysis buffer (20 mM Tris-HCI, pH 7.4; 1% NP-40, 137 mM NaCI, 0.5% Triton X-100) supplemented with 2 pg/ml leupeptin, 2 pg/ml pepstatin, 1 pg/ml aprotinin, 1 mM phenylmethylsulfonyl fluoride (PMSF), and 1 mM vanadate. Cell lysates were sonicated and centrifuged at 14,000 xgfor 15 min at 4°C. An aliquot of the supernatant was used for protein estimation by the Pierce bieinchoninic acid assay. The cell lysates (equal protein of about 0.5-1 mg) were subjected to immunoprecipitation with anti-FAK or antipaxillin or anti-ERK-1 plus ERK-2 (1-2 pg/ml) at 40C for 4 to 12 hr. Protein A/G (20 jil) was then added and incubated for an additional 4 to 6 hr at 40C. The antibody complex was pelleted and one portion of the immunocomplex was dissociated by boiling in 1 x sodium dodecyl sulfate (SDS) sample buffer for 5 min. Another aliquot of the immunocomplex was washed twice with kinase buffer (50 mM PIPES, pH 7.0; 10 mM MgCl2; 3 mM MnCI2, and 0.1 mM dithiothreitol). The kinase assays were initiated by the addition of 1 pg myelin basic protein and 50 M [^y-32P]ATP (10 Ci/mmol) in a final volume of 100 pl. The reaction was terminated after various time periods at 30°C by the addition of 10 mM ATP and Laemmli sample buffer. The phosphorylation of myelin basic protein was examined by SDS-polyacrylamide gel electrophoresis (PAGE) followed by autoradiography. For Western blot analysis, 40 pl of 6 x Laemmli SDS-PAGE buffer was added to 200 pl of the lysate (20,23), and samples were boiled for 5 min and stored at -200C. Cell lysates adjusted to equal protein were subjected to SDS-PAGE on 8 or 14% gels and were electrotransferred onto polyvinylidene difluoride (PVDF) membranes for Western blot analysis. Membranes were blocked with blocking buffer (GIBCO-BRL, Gaithersburg, MD) for 1 hr, followed by incubation with 4G10 monoclonal antiphosphotyrosine antibody (1:1000 dilution) for 2 hr. The blots were washed with TBST (50 mM Tris base, 200 mM NaCI, and 0.1% Tween 20) and were then incubated with goat antimouse IgG heavy and light chains horseradish peroxidase (1:3000 dilution) for 1 hr. Subsequently, the blots were washed in TBST and the phosphotyrosine-containing proteins were immunodetected using ECL.

Determination ofTyrosine Kinase and PTP Activities
Tyrosine kinase activity in control and ROS-treated BPAEC lysates was determined as described previously (24)

Changes in [Ca2i
BPAECs grown on glass cover slips as monolayers were loaded with 10 pM Fura-2 acetoxymethylester (AM) in the dark for 45 min as described earlier (16). The loaded cells were rinsed 3 times and incubated for an additional 30 min in the dark in buffer containing: 4.8 mM KCl, 130 mM NaCI, 1.0 mM MgCl2, 1.5 mM CaCl2, 1.0 mM Na2HPO4, 15 mM glucose, and 10 HEPES (pH 7.4) without albumin. The Fura-2 loaded cells were inserted diagonally in the 1.0-cm acryl cuvettes filled with 2 ml buffer. The cells were challenged with varying concentrations of DPV and Fura-2 fluorescence was measured with an Aminco-Bowman Series 2 luminescence spectrometer (SLM/ Aminco, Urbana, IL) at excitation wavelengths of 340 and 380 nm and an emission wavelength of 510 nm. The 340/380 ratio was taken every half second and measurements were corrected for autofluorescence by measuring the fluorescence of the cells not loaded with Fura-2 AM. Intracellular free calcium [Ca2+1] was calculated with software provided by the manufacturer.

ROS Modulates Tyrosine Kinase and PIP Activities
As ROS are potent inhibitors of PTPs (13,26), it was necessary to determine the effect of H202 and H202 plus vanadate on tyrosine kinase and phosphatase activities in ECs. As shown in Figure 1, H202 treatment of BPAECs increased tyrosine kinase activity without altering the PTP activity. However, 10 pM vanadate inhibited the PTP activity without altering the tyrosine kinase activity. A combination of vanadate plus H202 was a potent inhibitor of PTP (> 80% inhibition) and activator of tyrosine kinase (3-fold over control). These data suggest that modulation of tyrosine kinase and PTP activities in ECs is dependent on the nature of ROS used.

ROS Increases Tyrosine Phosphorylation ofProteins in ECs
Modulation of tyrosine kinases/PTPs by ROS may increase protein tyrosine phosphorylation of ECs in a manner similar to that in other cell types (27). Treatment of BPAECs with varying concentrations of H202 resulted in a dose-dependent increase in tyrosine phosphorylation of several proteins, as determined by immunoblotting with antiphosphotyrosine antibody (Figure 2). At lower concentrations of H202, the only prominent tyrosine phosphorylated band that was immunodetected was 1 10 to 130 kDa. Under similar experimental conditions, treatment of BPAECs with ATP, bradykinin (BRK), or TPA showed no significant increase in tyrosine phosphorylation of proteins. Similarly, treatment of BPAECs with 1 to 5 pM vanadate showed no increase in tyrosine phosphorylation of EC proteins. However, vanadate at 10 pM exhibited a small increase in tyrosine phosphorylation of proteins at 20 to 35 kDa and 110 to 180 kDa (   (24). To further examine the effect of ROS on protein tyrosine phosphorylation in ECs, BPAECs were exposed to varying concentrations of 50 to 200 pM H202 for 30 min. Cell lysates from control and H202-treated cells were subjected to immunoprecipitation with anti-FAK or antipaxillin monoclonal bodies and the immunoprecipitates were separated by SDS-PAGE and probed with A-PY (or anti-FAK) or antipaxillin. As shown in Figures 5 and 6, H202 significantly increased tyrosine phosphorylation of FAK and paxillin in a dose-dependent manner compared to cells exposed to medium alone. Similar activation of tyrosine phosphorylation of FAK was observed when BPAECs were exposed to a xanthine/xanthine oxidase system that generates superoxide anion radicals. In addition to FAK and paxillin, 5 pM DPV   H202 (100 PM) caused activation of tyrosine kinases, as evidenced by increased protein tyrosine phosphorylation (Western blot analysis). In addition to 1P3, an accumulation of DAG was observed with H202 (100 pM) (2to 5-fold over control) and H202 (100 pM) plus vanadate (10 pM) (7-fold over control). However, vanadate (10-100 pM) failed to enhance DAG levels. These results suggest that ROS activate PLA2, PLC, and PLD pathways that may involve stimulation of tyrosine kinases. As protein tyrosine phosphorylation is a balance between tyrosine kinases and PTPs, inhibition of phosphatases should Tyrosine phosphorylated proteins were detected by ECL and 32P-labeled myelin basic protein was detected on X-ray film after SDS-PAGE (5% gel).
upregulate tyrosine kinase-mediated activation of phospholipases and protein tyrosine phosphorylation. To investigate whether PTP inhibitors modulate H202induced stimulation of phospholipases and protein tyrosine phosphorylation, BPAECs were challenged with MEM, MEM containing H202, or H202 plus vanadate. As shown in Table 3  accumulation as well as protein tyrosine phosphorylation (Table 4). These results suggest a role for tyrosine kinase/PTP in ROS-mediated activation of signal transduction pathways in ECs.  Figure 9. Effect of BAPTA on DPV-induced protein tyrosine phosphorylation. BPAECs were pretreated with medium alone or medium containing 25 pM BAPTA for 30 min. Cells were washed and challenged with medium or medium containing 5 pM DPV for 15 min. Cell lysates were prepared as described under "Materials and Methods." Cell lysates from each of the treatments (10 pg protein) were subjected to SOS-PAGE, transferred onto PVDF membrane and immunoblotted with antiphosphotyrosine antibody 4610. Tyrosine phosphoryated proteins were detected using ECL.    factors, we explored the role of growth factor receptors in ROS-mediated activation of PLD. Suramin blocks agonist-growth factor receptor interactions and therefore inhibits PLD activation by ROS when growth factor receptors are involved. As shown in Table 5, suramin pretreatment of BPAECs blocked DPV-induced [32P]PBt formation. The effect of suramin was specific for DPV-induced PLD activation, as it had no effect on TPA-mediated [32P]PBt formation ( with 50% inhibition at 100 pM suramin (Table 6). These results suggest that growth factor receptors may be involved in ROS-mediation PLD activation.

Role ofGrowth Factor
Effct ofAntioxidants on ROS-Mediated PLD Activation As proteins and lipids are likely targets through which ROS can modulate cell signaling, we investigated the effect of antioxidants on DPV-induced PLD activation. Enhancing the EC redox state by pretreatment with N-acetylcysteine abolished the ability of DPV to activate PLD activity (Table 7). These results suggest that modulation of sulfhydryl reactivity by DPV is blocked by N-acetylcysteine.

Discussion
In this study, the effects of ROS on activation of PLA2, PLC, and PLD and protein tyrosine phosphorylation in ECs were examined. PLC, and PLD pathways, respectively. It was also observed that pretreatment of ECs with vanadate potentiated the H202induced activation of PLA2 and PLD. Vanadate, a known inhibitor of phosphatases, also interacts with H202 to generate peroxovanadium compounds (11,12). A major product generated by mixing equimolar amounts of H202 and vanadate under neutral pH condition was identified as DPV (18). Our data also indicate that ROS-induced stimulation of PLA2 and PLD was attenuated by genistein, an inhibitor of tyrosine kinases. These results suggest that ROS-induced activation of phospholipases involves protein tyrosine phosphorylation. This finding further supports the earlier observation that H202 plus vanadate mimics insulin (12), stimulates P13 kinase (28), and enhances protein tyrosine phosphorylation (29). The mechanism(s) involved in tyrosine kinase-mediated activation of PLA2, PLC, and PLD is unclear. There is increasing evidence for the involvement of protein tyrosine phosphorylation in growth factor-, oxidant-, and IgE-mediated PLD activities. However, the nature of tyrosine phosphorylated proteins involved in PLD activation has not been identified. Earlier studies by Vepa et al. (24) indicate that caveolin and FAK (125 kDa) are targets for H202 in ECs. In this study, we have identified ERK-1 and ERK-2 and focal adhesion proteins (FAK and paxillin) as potential targets for ROS-induced tyrosine phosphorylation. Exogenous addition of oxidants induced tyrosine phosphorylation of ERK-1 and ERK-2, and this activation was mediated in Table 7. Effect of N-acetylcysteine on ROSand TPA-induced PLD activation.
[32P]PBt, dpm/dish  part by MEK (30,31). Similarly, exposure of NIH3T3 cells to H202 differentially activated ERK-2, JNK, and p38 MAP kinases (32). Activation of MAP kinases culminates in phosphorylation of downstream targets including enzymes and nuclear factors (33). Although the data presented here do not implicate a direct role for MAP kinases in the activation of EC phospholipases, recent studies suggest that p38 MAP kinase and phosphorylation of heat shock protein HSP27 involve MAPKAPK-2 (34). Suramin, a known inhibitor of ligand-receptor interaction, attenuated ERK-2 activation by epidermal growth factor (EGF) (32). A similar inhibitory effect of suramin on DPVinduced PLD activation was observed, suggesting a role for growth factor receptors rich in cysteine in ROS-mediated cell signaling. Indeed, a role for sulfhydryl groups in ROS-mediated signal transduction was confirmed by the inhibitory effect of N-acetylcysteine on DPVinduced PLD activation (Table 7) and protein tyrosine phosphorylation (16).
ROS increase [Ca2+]i through 1P3dependent release, enhanced Ca2, transport, and through Ca2+-dependent channels (7). Although there are several studies to indicate that oxidative stress induces cell toxicity (2)(3)(4)(5)(6)(7)(8), recent evidence, including the results of this study, suggests that ROS modulate EC function by altering Ca2+ signaling, generation of second messengers and protein kinases/phosphatases. Although the precise mechanism(s) and regulation of ROSmediated activation of phospholipases remain to be clearly established, this study provides support for changes in [Ca2%] and redox status of the cell in ROSmediated protein tyrosine phosphorylation and EC signaling pathways ( Figure 10). Although it is well known that exposure of mammalian cells including ECs to elevated levels of ROS induces toxicity, there is no direct correlation between ROS-induced alterations in signaling pathways and ROS-mediated toxicity. It has been hypothesized that ROS-mediated signaling alterations precede toxicity and by blocking changes in cell signaling by free radical scavengers and antioxidants, the ROS-mediated toxicity can be partially blocked. Further studies on ROS-mediated cellular signaling and toxicity should provide a better understanding of the complexities of cell signaling mechanisms under normal and pathologic conditions. Environmental Health Perspectives * Vol 106, Supplement 5 -October 1998