J Biol Chem, Vol. 275, Issue 3, 1739-1748, January 21, 2000

Reactive Oxygen Species Activate p90 Ribosomal S6 Kinase via Fyn and Ras^*

Jun-ichi Abe, Masanori Okuda^§, Qunhua Huang, Masanori Yoshizumi, and Bradford C. Berk^¶

From the Center for Cardiovascular Research, University of Rochester, Rochester, New York 14642 and the ^§ Department of Internal Medicine (First Division), Kobe University School of Medicine, Kobe 650-0017, Japan

	ABSTRACT

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Reactive oxygen species and growth factors stimulate similar intracellular signal transduction events including activation of Src kinase family members and extracellular signal-regulated kinases (ERK1/2). A potentially important downstream effector of Src and ERK1/2 is p90 ribosomal S6 kinase (p90RSK), which plays an important role in cell growth by activating several transcription factors as well as the Na⁺/H⁺ exchanger. In the present study, we determined whether H₂O₂ activates p90RSK to gain insight into signal transduction mechanisms activated by reactive oxygen species. H₂O₂ (200 µM) stimulated ERK1/2 and p90RSK activity in lymphocytes, endothelial cells, and fibroblasts. The MEK-1 inhibitor, PD98059 (30 µM), inhibited H₂O₂-mediated activation of ERK1/2 but not of p90RSK. An essential role for Fyn and Ras in p90RSK activation was suggested by five findings. 1) The tyrosine kinase inhibitor, herbimycin A, and the specific Src kinase family inhibitor, PP1, blocked p90RSK activation by H₂O₂ in a concentration-dependent manner. 2) p90RSK activation by H₂O₂ was significantly reduced in fibroblasts derived from transgenic mice deficient in Fyn, but not c-Src. 3) H₂O₂ rapidly activated Ras (peak at 2-5 min), which preceded p90RSK activation (peak at 20 min). 4) Dominant negative Ras completely blocked H₂O₂-induced activation of p90RSK. 5) In Fyn/ fibroblasts, activation of Ras by H₂O₂ was significantly attenuated. These results show essential roles for Fyn and Ras in H₂O₂-mediated activation of p90RSK and establish redox-sensitive regulation of Ras and p90RSK as a new function for Fyn.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Reactive oxygen species (ROS)¹ including H₂O₂, O₂, and OH have been implicated in the pathogenesis of cardiovascular disease, especially atherosclerosis and hypertension (1-3). Recently, it has been shown that ROS stimulate intracellular signal events similar to those activated by growth factors, including activating kinases such as c-Src and ERK1/2 (4, 5). Several transcription factors, including nuclear factor-B and c-fos, also appear to be activated by ROS (6, 7). Furthermore, several investigators have shown that H₂O₂ stimulates both ERK1/2 and the Na⁺-H⁺ exchanger (NHE-1) in cardiac myocytes (8). Previously, we showed that p90RSK was a putative NHE-1 kinase and that increased p90RSK activity correlated with increased Na/H⁺ exchange activity in hypertensive rats (9, 10). These findings suggest that p90RSK may be an important mediator of ROS function.

p90RSK has multiple functions. In quiescent cells, inactive p90RSK resides in the cytoplasm and is partially complexed with its upstream regulator, ERK1/2 (11). Stimulation of cells by growth factors operating through the Ras/Raf-1/MEK pathway leads to the activation of ERK1/2, phosphorylation and activation of p90RSK, and the import of these kinases into the nucleus. Whereas p90RSK was initially shown to phosphorylate transcription factors, including c-Fos (12) and Nur77 (13), it has also been proposed that p90RSK is involved in activation of nuclear factor-B by phosphorylation of I-B (14). Recently, p90RSK has been reported to have both positive and negative regulatory effects on cyclic AMP response element-binding proteins (CREBs). Specifically, p90RSK2, one of three closely related rsk genes (rsk1, rsk2, and rsk3), has been shown to function as a stimulus-coupled CREB kinase by phosphorylating serine 133, which regulates CREB activity (15). Thus, the upstream signal mechanisms by which H₂O₂ activates p90RSK should provide valuable insights into pathways of redox-sensitive signal transduction.

Previous studies have shown that c-Src is involved in signal events stimulated by ROS (4). Several receptors that lack intrinsic tyrosine kinase activity stimulate tyrosine phosphorylation through association with Src family kinases such as Lck, Lyn, and Fyn (16). It has also been reported that Src and Ras are critical for ERK1/2 activation by H₂O₂ (17, 18). The predominant pathway for RSK activation by growth factors has been proposed to involve Src-Ras-Raf-Mek-ERK1/2. Interestingly, several investigators have reported that there is an alternative pathway for p90RSK activation in addition to ERK1/2 (19). Specifically, Li et al. (20) reported that B cell receptor activation stimulates the tyrosine kinase Syk, which activates p90RSK via an ERK1/2-independent pathway in DT40 cells.

To determine the role of Src family kinases in H₂O₂-mediated activation of ERK1/2 and p90RSK, we utilized specific Src inhibitors and cells derived from animals deficient in c-Src and Fyn. We show here that activation of p90RSK by H₂O₂ is positively regulated by Fyn but not by c-Src. Furthermore, we demonstrate that Fyn is required for H₂O₂-mediated activation of Ras. Thus, the Fyn-Ras-p90RSK signaling pathway described here may represent a new redox-sensitive mechanism.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Cell Lines, Culture, and Transfection-- Fibroblasts deficient in c-Src (Src/) or Fyn (Fyn/) were isolated from mouse embryo fibroblasts homozygous for disruption of the Src and Fyn genes, immortalized with large T antigen (21). Cells were kindly provided by Sheila M. Thomas (Fred Hutchinson Cancer Center, Seattle, WA). Fibroblasts were maintained in Dulbecco's modified Eagle's medium supplemented with 10% calf serum as described previously (21). The dominant negative Ras (generously provided by Dr. Neil Nathanson, University of Washington, Seattle, WA) (22) was subcloned into pcDNA3.1(+) (Invitrogen Inc.) to transfect mammalian cells. LipofectAMINE transfections were performed as described previously (23). Cells at 70-80% confluence in 100-mm dishes were growth-arrested by incubation in RPMI 1640 medium for 24 h prior to use. Human peripheral blood mononuclear cells (MNCs) were obtained from healthy, normal subjects by Ficoll-Paque Plus (Amersham Pharmacia Biotech) density gradient centrifugation as described previously (24). By Giemsa and -naphthyl butyrate esterase staining, we found 78.4 ± 6.2% lymphocytes, 20.2 ± 7.5% monocytes, occasional granulocytes (0.8 ± 0.4%), and clumps of platelets (n = 5) contained in the MNC preparation. Jurkat cells were serum-starved for 16 h before being harvested and washed in 4 °C RPMI medium using minimal manipulation of the cells, because it was noted that extensive manipulation of the cells resulted in high basal ERK activity. The cells were resuspended at 6 × 10⁷ cells/ml and quickly stimulated with H₂O₂, phorbol 12-myristate 13-acetate, or thapsigargin.

NHE-1 and Myelin Basic Protein Phosphorylation (In-gel Kinase Assay)-- Equal amounts of cell lysate protein (20 µg) were separated by 7.5% SDS-PAGE in a gel containing 0.15 mg/ml of GST-NHE-1 (625-747) or 1.0 mg/ml of myelin basic protein. We have previously shown that results for in-gel kinase assay and immune complex kinase assay for ERK1/2 are highly correlated (R² = 0.92) (25). The gel was then incubated twice in Buffer A (50 mmol/liter HEPES, pH 7.4, and 5 mmol/liter -mercaptoethanol) containing 20% isopropyl alcohol for 30 min, once in Buffer A for 1 h, twice in Buffer A containing 6 M guanidine HCl for 30 min, twice in Buffer A containing 0.04% Tween 20 at 4 °C for 16 and 2 h, once in Buffer A containing 100 µmol/liter Na₃VO₄ and 10 mmol/liter MgCl₂ at 30 °C for 30 min, and once in Buffer A containing 100 µmol/liter Na₃VO₄ and 10 mmol/liter MgCl₂, 50 µmol/liter ATP, and 50 µCi of [-³²P]ATP for 1 h at 30 °C. The reaction was terminated by washing the gel 6-8 times in fixative solution containing 10 mmol/liter sodium pyrophosphate and 5% trichloroacetic acid for 15 min. The gel was dried and subjected to autoradiography. Autoradiographic signal intensity was quantified by densitometry in the linear range of film exposure by using a LaCie scanner and NIH Image 1.49.

Immunoprecipitation and Western Blot Analysis-- After treatment, the cells were washed with phosphate-buffered saline, harvested in 0.5 ml of lysis buffer (50 mM sodium pyrophosphate, 50 mM NaF, 50 mM NaCl, 5 mM EDTA, 5 mM EGTA, 100 µM Na₃VO₄, 10 mM HEPES, pH 7.4, 0.1% Triton X-100, 500 µM phenylmethylsulfonyl fluoride, and 10 µg/ml leupeptin), and flash-frozen on a dry ice/ethanol bath. After allowing the cells to thaw, cells were scraped off the dish and centrifuged at 14,000 × g (4 °C for 30 min), and protein concentration was determined using the Bradford protein assay (Bio-Rad). For immunoprecipitation, cell lysates were incubated with goat anti-p90RSK2 antibody for 3 h at 4 °C and then incubated with 20 µl of protein A-Sepharose CL-4B (Amersham Pharmacia Biotech) for 1 h on a roller system at 4 °C. The beads were washed two times with 1 ml of lysis buffer, two times with 1 ml of LiCl wash buffer (500 mM LiCl, 100 mM Tris-Cl, pH 7.6, 0.1% Triton X-100, 1 mM dithiothreitol), and two times in 1 ml of washing buffer (20 mM HEPES, pH 7.2, 2 mM EGTA, 10 mM MgCl₂, 1 mM dithiothreitol, 0.1% Triton X-100). For Western blot analysis, immunoprecipitates were subjected to SDS-PAGE, and proteins were transferred to nitrocellulose membranes (Hybond^TM-ECL, Amersham Pharmacia Biotech) as described previously (26). The membrane was blocked for 1 h at room temperature with a commercial blocking buffer from Life Technologies, Inc. The blots were then incubated for 4 h at room temperature with the p90RSK or ERK1/2 (Santa Cruz Biotechnology) antibody, followed by incubation for 1 h with secondary antibody (horseradish peroxidase-conjugated). For ERK1/2 activation, the blots were incubated for 12 h with anti-phospho-specific ERK1/2 (New England Biolabs) or nonspecific ERK1 and ERK2 antibodies (Santa Cruz Biotechnology). Immunoreactive bands were visualized using enhanced chemiluminescence (ECL, Amersham Pharmacia Biotech).

p90RSK Kinase Assays-- p90RSK kinase activity was measured by GST-NHE-1 phosphorylation as described previously with slight modification (9, 26). Cells were harvested in lysis buffer at 4 °C and then flash-frozen on a dry ice/ethanol bath. After allowing the cells to thaw, cells were scraped off the dish and centrifuged at 14,000 × g (4 °C for 30 min), and protein concentration was determined. p90RSK was immunoprecipitated by incubating 400 µg of protein from each sample with 3 µl of the goat anti-p90RSK2 (Santa Cruz Biotechnology) antibody for 3 h, adding 40 µl of a 1:1 slurry of protein A-Sepharose (Amersham Pharmacia Biotech) beads to the extract-antibody mixture, and incubating for 1 h at 4 °C. The beads were washed two times with 1 ml of lysis buffer, two times with 1 ml of LiCl wash buffer (500 mM LiCl, 100 mM Tris-Cl, pH 7.6, 0.1% Triton X-100, 1 mM dithiothreitol), and two times in 1 ml of modified Buffer A (20 mM HEPES, pH 7.2, 2 mM EGTA, 10 mM MgCl₂, 1 mM dithiothreitol, 0.1% Triton X-100). Immunoprecipitated RSK was resuspended in 25 mmol/liter HEPES, pH 7.4, 10 mmol/liter MgCl₂, 10 mmol/liter MnCl₂, and the kinase reaction was initiated by addition of 200 pmol of GST-NHE-1 (625-747), 15 µmol/liter ATP, and 0.5 mCi/ml of [-³²P]ATP. After the reaction proceeded for 20 min at 30 °C, it was terminated by addition of Laemmli sample buffer, and proteins were analyzed by 10% SDS-PAGE, followed by autoradiography.

Measurement of Intracellular [Ca²⁺]_i by Fura 2-- Measurement of [Ca²⁺]_i was performed as described previously (27, 28). Briefly, a Jurkat T-cell suspension (approximately 2 × 10⁶ cells/ml) was loaded with 2 µM fura 2/AM for 30 min at room temperature. After the loading period, the T-cells were washed twice with Krebs-Ringer-HEPES (KRH) medium containing 0.2% bovine serum albumin. The composition of KRH is as follows: 120 mM NaCl, 5.4 mM KCl, 1.0 mM CaCl₂, 0.8 mM MgCl₂, 11.1 mM glucose, 20 mM HEPES, pH 7.4. The fluorescence of the fura 2-loaded cell suspension was monitored in the cuvette with a PTI Delta scan spectrofluorometer. The excitation wavelengths were 340 and 380 nm, and emission was measured at 510 nm. [Ca²⁺]_i was estimated as described by Grynkiewicz et al. (29). [Ca²⁺]_i (in nM) = 225 nM × (F  F_min)/(F_max  F), where F was the measured fluorescence intensity, F_max was the fluorescence intensity after cell lysis, and F_min was the fluorescence intensity in the presence of 10 mM EGTA and 75 µM BAPTA-AM.

Activated Ras Affinity Precipitation Assay-- The expression vector encoding the fusion protein GST-Raf binding domain (GST-RBD) was obtained by ligation of cDNA encoding the first 149 amino acids of Raf-1 into the SmaI site of the pGEX 2T vector (Amersham Pharmacia Biotech). GST-RBD expression was induced in transformed bacteria with 1 mM isopropyl -D-thiogalactoside for 1 h, after which time, bacteria were harvested and lysed by sonication. The GST-RBD fusion protein was then purified on glutathione-Sepharose beads. Affinity precipitation of activated Ras was performed as described previously (30). Briefly, lysates were incubated on a rocker plate at 4 °C for 1 h with GST fusion protein bound to glutathione-Sepharose beads. Then, the supernatant containing equal amounts of protein after centrifugation was incubated with 50-60 µg of GST-RBD bound to beads at 4 °C for 4 h. The beads were then extensively washed with 20 mM HEPES, pH 7.5, 120 mM NaCl, 10% glycerol, 0.5% Triton X-100, 2 mM EDTA, 10 µg/ml leupeptin, and 10 µg/ml aprotinin. The eluted proteins were resolved on 12.5% polyacrylamide gel. Coomassie Brilliant Blue was used to stain the fusion protein in the gel (molecular mass, 42 kDa). Bound Ras was quantified by Western blot analysis as described above.

Analysis of Ras-bound GDP and GTP-- Cells at 70-80% confluence in 60-mm dishes were growth-arrested by incubation in phosphate-free Dulbecco's modified Eagle's medium for 18 h prior to use. The quiescent cells were washed once with 2 ml of phosphate-free Dulbecco's modified Eagle's medium. The cells were incubated with phosphate-free Dulbecco's modified Eagle's medium supplemented with 0.1-0.2 mCi/ml [³²P]orthophosphate for 6 h. Sodium orthovanadate (50 µM) was added to the cells during the last 30 min to increase GTP-bound Ras as described (31). The cells were then stimulated with H₂O₂ for the indicated periods of time. After lysis of the cells, Ras was immunoprecipitated from the cell lysates with anti-Ras monoclonal antibody Y13-259 and analyzed as described by Satoh et al. (32).

Materials-- All materials were from Sigma except as indicated. H₂O₂ was from Fisher Scientific, and PP1 was from Pfizer.

Statistical Analysis-- Data are reported as mean ± S.D. Statistical analysis was performed with the StatView 4.0 package (ABACUS Concepts, Berkeley, CA). Differences were analyzed with unpaired two-tailed Student's t test or with Welch's t test as appropriate.

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

H₂O₂ Stimulates p90RSK Activity in Human Blood Peripheral Mononuclear Cells-- A role for p90RSK in ROS signal transduction was first suggested by experiments to identify kinases that phosphorylate NHE-1. MNCs were stimulated with H₂O₂ or sorbitol, and cell lysates were subjected to an in-gel-kinase-assay using GST-NHE-1 (625-747) as substrate. H₂O₂ stimulated the activity of several NHE-1 kinases in human MNCs, including kinases of 90, 60, 44, and 42 kDa (Fig. 1). The kinase activated to the greatest extent by H₂O₂ was a 90-kDa kinase. Of interest, the activity of the 90 kDa kinase was stimulated by H₂O₂ but not by sorbitol. In rat vascular smooth muscle cells, we showed previously that p90RSK is the predominant 90-kDa NHE-1 kinase by immunodepletion of RSK with RSK2 antibody (9). To identify which RSK isoforms were present in human MNCs, Western blot analysis was performed. Immunoreactive 90-kDa protein bands were readily observed with each p90RSK isoform antibody (Fig. 1B). To confirm that p90RSK was activated by H₂O₂, we performed an immune complex kinase assay using GST-NHE-1 as substrate. Immunoprecipitated p90RSK showed a concentration-dependent increase in activity at 5 min after stimulation by H₂O₂ (Fig. 1C). To confirm that the increase in p90RSK activity (measured by immune complex GST-NHE-1 phosphorylation) was not due to ERK1/2 that co-precipitated with p90RSK, we immunoprecipitated with p90RSK antibody and then performed a Western blot with anti-ERK1/2 antibody (Fig. 1D). We could not detect ERK1/2 in p90RSK immunoprecipitates, indicating that p90RSK is the dominant NHE-1 kinase present.

View larger version (43K): [in this window] [in a new window]   Fig. 1.   p90RSK is activated by oxidative stress. A, human peripheral mononuclear cells were stimulated for 5 min with 0.4 M sorbitol or 200 µM H2O2. Cells were harvested in lysis buffer as described under "Materials and Methods," and in-gel kinase assay was performed with GST-NHE-1 (625-747) as substrate. B, human peripheral mononuclear cells were harvested, cell lysates were prepared, and Western blot analysis was performed with polyclonal goat antibodies against RSK1, RSK2, and RSK3. C, human peripheral mononuclear cells were stimulated for 5 min with the indicated concentrations of H2O2. Cells were harvested in lysis buffer, and proteins were immunoprecipitated with RSK2 antibody and phosphorylation of GST-NHE-1 (625-747) was determined by in vitro kinase assay. D, human peripheral mononuclear cells were stimulated for 5 min with 200 µM H2O2. Cells were harvested in lysis buffer, and proteins were immunoprecipitated with RSK2 antibody and Western blot analysis was performed with RSK2 (top panel) or ERK1/2 (bottom panel) antibody. TCL, total cell lysates. Data shown are representative of experiments repeated two times with different samples.

p90RSK Can Be Regulated via a MEK1/2-independent Pathway-- To determine whether activation of p90RSK by H₂O₂ was ubiquitous, we measured the response to H₂O₂ in several different cell types. Growth-arrested Jurkat T-cells, bovine aortic endothelial cells (BAEC), and mouse fibroblasts were analyzed by Western blotting with p90RSK antibody. There was equal expression of p90RSK in all cell types studied (Fig. 2A). Exposure to 200 µM H₂O₂ stimulated p90RSK activity in both Jurkat cells (12.6-fold at 5 min) and BAEC (3.6-fold at 20 min). A much smaller response was observed in mouse fibroblasts.

View larger version (49K): [in this window] [in a new window]   Fig. 2.   p90RSK and ERK1/2 is activated by H2O2 in multiple cell types. A, Western blot analysis. Jurkat cells, BAEC, mouse fibroblasts, and mononuclear cells were obtained and grown as described under "Materials and Methods." Western blot analysis was performed on whole cell lysates using RSK2 antibodies. B, ERK1/2 and p90RSK activation by H2O2. The indicated cells were growth-arrested for 48 h as described under "Materials and Methods" and then exposed to 200 µM H2O2 for the indicated times. ERK1/2 activity was measured by in-gel kinase assay as described under "Materials and Methods" in fibroblasts (left, top panel) and Western blot analysis with a phospho-specific ERK antibody that recognizes only the catalytically activated forms (phosphorylated on Thr202/Tyr204) of ERK1/2 (left, second panel from top; right, top panel of Jurkat and BAEC). Levels of total ERK1/2 in the same samples were measured with anti-ERK1/2 antibodies (IB: ERK1/2). Bottom panels of each set, p90RSK activity was measured by in vitro kinase assay using GST-NHE-1 (625-747) as substrate. Data shown are representative of experiments repeated two times with different samples.

We previously reported that activation of p90RSK by angiotensin II and platelet-derived growth factor was inhibited by PD98059 (MEK1/2 inhibitor), suggesting that this kinase is regulated by the MEK1/2-ERK1/2 pathway. To determine the role of MEK1/2-ERK1/2 in p90RSK activation, the time course for ERK1/2 activation by 200 µM H₂O₂ was determined. In fibroblasts, we used both an in-gel kinase assay and immunoblotting for dually phosphorylated, active ERK1/2 in whole cell lysates. Stimulation of fibroblasts with H₂O₂ caused maximal activation of ERK1/2 at 5 min (Fig. 2B, left, top panel). Activated ERK1/2 could also be detected in immunoblots of lysates from fibroblasts stimulated with H₂O₂, confirming the results of in-gel kinase assay. Griffith et al. (33) have reported that an immune complex kinase assay correlates well with immunoblotting for active phosphorylated ERK1/2 in Jurkat T cells. In addition, we found that activation of ERK1/2 by in-gel kinase assay and immunoblotting for active ERK1/2 were highly correlated (R² = 0.90) in endothelial cells (data not shown). Therefore, we used immunoblotting for phospho-ERK1/2 to evaluate ERK1/2 activation as described previously (33-35). Because in Jurkat cells ERK1 (44-kDa) is not expressed to the same extent as ERK2 (42-kDa), phosphorylated ERK1 is less prominent than ERK2 as reported by others (33, 36-40). After H₂O₂ stimulation, ERK1/2 activity rapidly increased with a peak at 2-5 min (Fig. 2B), which preceded peak activity of p90RSK in both Jurkat cells (5 min) and mouse fibroblasts (20 min). In contrast, p90RSK activity in BAEC increased rapidly at 5 min after H₂O₂ stimulation, but ERK1/2 activity did not increase until 20 min, significantly slower than the peak of p90RSK activity. These results suggest that there is an alternative pathway, besides ERK1/2, for activation of p90RSK in BAEC.

To define the role of MEK1/2-ERK1/2 in H₂O₂-mediated activation of p90RSK, we determined the effect of PD98059 on both ERK1/2 and p90RSK activity in Jurkat cells. PD98059 caused a concentration-dependent inhibition of H₂O₂-mediated ERK1/2 activation with an IC₅₀ value of ~1 µM. At 30 µM PD98059, ERK1/2 inhibition was ~78%. In contrast, p90RSK activation was only partially inhibited by PD98059, ~26% at 30 µM (Fig. 3B).

View larger version (23K): [in this window] [in a new window]   Fig. 3.   MEK1/2 inhibition prevents H2O2-mediated activation of ERK1/2 to a greater extent than activation of p90RSK. Growth-arrested Jurkat T-cells were treated with varying concentrations of PD98059 or vehicle (0.1% dimethyl sulfoxide) for 30 min and then stimulated with 200 µM H2O2 for 5 min. Cell lysates were prepared, lysate proteins were immunoprecipitated with RSK2 antibody, and in vitro kinase assay with GST-NHE-1 (625-747) as substrate was performed. ERK1/2 activity was measured by Western blot analysis with a phospho-specific ERK antibody. After SDS-PAGE, phosphorylation of GST-NHE-1 (625-747) and ERK1/2 was measured by densitometry in the linear range of film development. A, representative Western blot with anti-phospho-specific ERK1/2 antibody (top panel) and anti-ERK1/2 (middle panel) and autoradiogram for p90RSK activity (bottom panel). B, densitometric analysis of ERK1/2 and GST-NHE-1 (625-747) phosphorylation. Results were normalized to control (time = 0), which was arbitrarily set to 1.0 (shown is mean ± S.D., n = 3).

To determine whether ERK1/2 inhibition by PD98059 was sufficient to attenuate p90RSK activity, we compared the effect of PD98059 on ERK1/2 and p90RSK activities stimulated by H₂O₂ and the Ca²⁺ ionophore A23187 (which is known to stimulate p90RSK activity in an ERK1/2-dependent manner (41)). A23187 (60 µM) activated both ERK1/2 and p90RSK (Fig. 4). PD98059 inhibited ERK1/2 and p90RSK activation by A23187 in a dose-dependent manner, with complete inhibition by 30 µM PD98059. These results show that inhibiting ERK1/2 activation with PD98059 significantly attenuates p90RSK activation. In combination, the results suggest that H₂O₂-mediated p90RSK activation is regulated, at least partially, via MEK1/2-independent pathways, whereas A23187-mediated p90RSK activation is completely MEK1/2-dependent.

View larger version (23K): [in this window] [in a new window]   Fig. 4.   A23187 activates ERK1/2 and p90RSK: inhibition by PD98059. Growth-arrested Jurkat cells were treated with varying concentrations of PD98059 or vehicle (0.1% dimethyl sulfoxide) for 30 min and then stimulated with 60 µM A23187 for 5 min. Cell lysates were prepared, lysate proteins were immunoprecipitated with RSK2 antibody, and in vitro kinase assay with GST-NHE-1 (625-747) as substrate was performed. ERK1/2 activity was measured by Western blot analysis with a phospho-specific ERK antibody. After SDS-PAGE, phosphorylation of GST-NHE-1 (625-747) and ERK1/2 was measured by densitometry in the linear range of film development. A, representative Western blot with anti-phospho-specific ERK1/2 antibody (top panel) and anti-ERK1/2 (middle panel) and autoradiogram for p90RSK activity (bottom panel). B, densitometric analysis of ERK1/2 and GST-NHE-1 (625-747) phosphorylation. Results were normalized to control (time = 0), which was arbitrarily set to 1.0 (shown is mean ± S.D., n = 3).

Role of PKC and Calcium in H₂O₂-mediated p90RSK Activation in Jurkat Cells-- Increases in intracellular calcium modulate long-term cellular responses by regulating the function of several transcription factors, thus leading to new gene expression (42). In particular, CREB, which is phosphorylated and regulated by p90RSK, is a critical mediator of Ca²⁺-dependent gene expression (43). Because H₂O₂ increases [Ca ²⁺]_i, and because we previously found that H₂O₂-mediated c-fos expression was dependent on both calcium and protein kinase C (7), we determined whether p90RSK activation by H₂O₂ was calcium- and protein kinase C-dependent (Fig. 5).

View larger version (48K): [in this window] [in a new window]   Fig. 5.   H2O2-mediated activation of p90RSK is calcium-dependent and partially dependent on PKC. A, growth-arrested Jurkat cells were treated with 1 µM phorbol 12,13-dibutyrate (PDBu) for 16 h and then stimulated with 200 µM H2O2 for 2-5 min. ERK1/2 and p90RSK activities were determined as described under "Materials and Methods." Representative Western blots with anti-phospho-specific ERK1/2 antibody (top panel) and anti-ERK1/2 (middle panel) and autoradiogram for p90RSK activity (bottom panel) of three performed are shown. B, growth-arrested Jurkat cells were treated with Hanks' buffered salt solution or Ca2+-free Hanks' buffered salt solution containg BAPTA-AM for 30 min and then stimulated with 200 µM H2O2 for 2-5 min. Cell lysates were prepared, lysate proteins were immunoprecipitated with RSK2 antibody, and in vitro kinase assay with GST-NHE-1 (625-747) as substrate was performed. ERK1/2 activity was measured by Western blot analysis with a phospho-specific ERK antibody. Representative Western blots with anti-phospho-specific ERK1/2 antibody (top panel) and anti-ERK1/2 (middle panel) and autoradiogram for p90RSK activity (bottom panel) of three performed are shown.

We previously found that p90RSK does not require the activity of phorbol ester-responsive PKC isoforms for stimulation by angiotensin II in vascular smooth muscle cells, as shown by the failure of PKC down-regulation (incubation with 1 µM phorbol 12,13-dibutyrate for 24 h) to inhibit activity. As shown in Fig. 5A, down-regulation of PKC partially inhibited ERK1/2 and p90RSK activation by H₂O₂ in Jurkat cells. In contrast, phorbol 12-myristate 13-acetate-mediated activation of ERK1/2 and p90RSK was completely inhibited by phorbol 12,13-dibutyrate treatment.

To prevent increases in intracellular calcium, we used BAPTA-AM. Following BAPTA-AM treatment, H₂O₂ was no longer able to stimulate p90RSK in Jurkat cells (Fig. 5B, bottom panel). In contrast, Ca²⁺ chelation did not inhibit ERK1/2 activation by H₂O₂ (Fig. 5B, top two panels) as described previously (17). These results confirm that in addition to MEK1/2-ERK1/2, there is an alternative pathway for H₂O₂-mediated p90RSK activation via a calcium-dependent pathway. A23187-induced increases in cytosolic Ca²⁺ are very different from physiological changes. Therefore, we performed two sets of experiments to define the role of Ca²⁺ in H₂O₂-induced p90RSK activation. First, we determined the effect of BAPTA-AM, H₂O₂ (200 µM), thapsigargin (a microsomal Ca²⁺-ATPase inhibitor), and A23187 on [Ca²⁺]_i. Second, we compared p90RSK activation by thapsigargin and H₂O₂ in Jurkat cells.

The basal level of [Ca²⁺]_i was 85 ± 12 nM in Jurkat T-cells, and pretreatment with 75 µM BAPTA-AM chelated [Ca²⁺]_i to 3 ± 2 nM. 100 nM thapsigargin elevated [Ca²⁺]_i to 320 ± 28 nM, and 60 µM A23187 increased [Ca²⁺]_i to 1944 ± 387 nM (in the presence of 1.0 mM extracellular CaCl₂). In contrast, H₂O₂ (200 µM) did not increase [Ca²⁺]_i significantly (116 ± 24 versus 85 ± 12 nM (p = 0.12; n = 3)) in Jurkat T-cells.

To define the the role of "physiological" increases in [Ca²⁺]_i on p90RSK activation, we also determined the effect of thapsigargin on p90RSK activation. Whereas H₂O₂ (200 µM) stimulated p90RSK very strongly (Fig. 6), 100 nM thapsigargin only weakly activated p90RSK, despite increasing [Ca²⁺]_i to a greater extent. These results suggest that Ca²⁺-sensitive proteins that require a basal level of [Ca²⁺]_i are involved in H₂O₂-induced p90RSK activation, because pretreatment of BAPTA-AM completely inhibited p90RSK activation by H₂O₂. However, increases in [Ca²⁺]_i are not required for H₂O₂-mediated p90RSK activation in Jurkat-T-cells, because 200 µM H₂O₂ did not increase [Ca²⁺]_i significantly, and [Ca²⁺]_i increased by thapsigargin did not activate p90RSK to the same extent as H₂O₂.

View larger version (21K): [in this window] [in a new window]   Fig. 6.   H2O2 and thapsigargin stimulation of p90RSK in Jurkat cells. Growth-arrested Jurkat T-cells were treated with two different concentrations of thapsigargin (TG) or 200 µM H2O2 for 5 min. Cell lysates were prepared, lysate proteins were immunoprecipitated with RSK2 antibody, and in vitro kinase assay with GST-NHE-1 (625-747) as substrate was performed. Data shown are representative of experiments repeated two times with different samples.

Both Herbimycin A and PP1 Inhibit p90RSK Activation by H₂O₂ in Jurkat Cells-- Previous investigators have suggested that several tyrosine kinases in addition to c-Src may be upstream mediators of redox-sensitive signal transduction (4). Western blot analysis of total cell lysates showed that H₂O₂ stimulated tyrosine phosphorylation of 40-, 50-, 60-, 66-, and 110-kDa proteins (Fig. 7A). To determine whether a Src family kinase is an upstream mediator of p90RSK, we studied the effect of two different inhibitors, herbimycin A (a nonspecific tyrosine kinase inhibitor) and PP1 (a specific Src family kinase inhibitor) on H₂O₂-mediated p90RSK activity (Fig. 7B). Both herbimycin A and PP1 inhibited H₂O₂-mediated p90RSK activation in Jurkat cells in a dose-dependent manner with an IC₅₀ value of ~0.3 µM (Fig. 7B).

View larger version (48K): [in this window] [in a new window]   Fig. 7.   H2O2-stimulated tyrosine phosphorylation and p90RSK activation by H2O2 is inhibited by herbimycin A and PP1 in Jurkat cells. A, growth-arrested Jurkat cells were stimulated with 200 µM H2O2 for 5 min and lysed, and Western blot was performed with horseradish peroxidase-conjugated PY20 (anti-phosphotyrosine antibody). Data shown are representative of experiments repeated two times with different samples. B, growth-arrested Jurkat cells were pretreated with 0.1% Me2SO for 16 h (control) and the indicated concentrations of herbimycin A for 16 h or PP1 for 15 min. p90RSK activity was determined by immune complex kinase assay using [-32P]ATP and measured after SDS-PAGE by autoradiography. No difference in the amount of p90RSK was observed in lysates from control and treated cells by Western blot analysis with anti-p90RSK2 antibody (data not shown). Data shown are representative of experiments repeated two times with different samples.

Activation of p90RSK in Response to H₂O₂ Is Dependent on Fyn, but Not on c-Src-- The ability of herbimycin A and PP1 to inhibit p90RSK activation by H₂O₂ suggested an important role for Src family kinases. To verify the role of a Src family kinase-dependent mechanism, we utilized cells derived from mice deficient in Src or Fyn (21). There was no immunoreactive c-Src in Src/ cells, whereas immunoreactive Fyn was expressed to the same extent as in wild type cells. Likewise, there was no immunoreactive Fyn in Fyn/ cells, whereas there was no change in expression of c-Src in Fyn/ cells compared with wild type cells (Fig. 8A). H₂O₂ stimulated p90RSK activity in wild type fibroblasts, which was maximal at 20 min (Fig. 8B, top panel). In Src/ fibroblasts, p90RSK activity increased with a maximum 5.10 ± 0.95-fold increase 5 min after H₂O₂ stimulation (Fig. 8B, middle panel). In contrast, in Fyn/ fibroblasts, H₂O₂ failed to stimulate p90RSK activity at any time (Fig. 8B, bottom panel). These results indicate that H₂O₂ activation of p90RSK is dependent on Fyn but not c-Src in fibroblasts.

View larger version (35K): [in this window] [in a new window]   Fig. 8.   H2O2 activation of p90RSK is inhibited in Fyn/ cells but not in c-Src/ cells. A, wild type mouse fibroblast (WT), Src/, and Fyn/ cells were harvested, and Western blot analysis was performed on whole cell lysates using anti-Src antibody (left panel) and anti-Fyn antibody (right panel). B, cells were stimulated for the indicated times with 1 mM H2O2, and p90RSK kinase activity was measured by in vitro kinase assay using GST-NHE-1 (625-747), as shown by the arrow. No difference in the amount of p90RSK was observed in lysates from any of the cell lines by Western blot analysis with anti-p90RSK2 (data not shown). C, densitometric analysis of p90RSK kinase activity. Results were normalized by arbitrarily setting the densitometry of control cells (time = 0) to 1.0 (shown is mean ± S.D., n = 3).

To evaluate the role of ERK1/2 in p90RSK activation by H₂O₂ in these fibroblasts, we investigated ERK1/2 activation stimulated by H₂O₂ in Src/, Fyn/, and wild type fibroblasts. Peak ERK1/2 kinase activation occurred at 20 min in wild type fibroblasts (Fig. 9A, top panel). The ability of H₂O₂ to stimulate ERK1/2 was attenuated in Src/ and Fyn/ cells, when analyzed 5 min after H₂O₂ addition. However, this inhibition was entirely transient in nature. Maximal H₂O₂ activation of ERK1/2 occurred at 40 min in Fyn/ and Src / cells and was similar to ERK1/2 activation at 20 min in wild type cells (Fig. 9, B-D). These data suggest that there is an alternative pathway for H₂O₂-mediated p90RSK activation in addition to ERK1/2.

View larger version (34K): [in this window] [in a new window]   Fig. 9.   H2O2 activation of ERK1/2 is delayed in Src/ and Fyn/ cells. Cells were stimulated for the indicated times with 1 mM H2O2, and ERK1/2 activity was measured by Western blot analysis with a phospho-specific ERK1/2 antibody or anti-ERK1/2 antibody. A-C, representative Western blots with anti-phospho-specific ERK1/2 antibody (top panels) and anti-ERK1/2 (bottom panels) are shown. D, densitometric analysis of ERK1/2 phosphorylation. Results were normalized to control (time = 0), which was arbitrarily set to 1.0 (shown is mean ± S.D., n = 3).

Role of Ras in p90RSK and ERK1/2 Activation by H₂O₂-- Many studies indicate that tyrosine kinases activate ERK1/2 via a Ras-dependent mechanism (6, 44). To determine whether Ras is a component of the signaling cascade leading to p90RSK or ERK1/2 activation by H₂O₂, we examined the effect of dominant negative Ras (RasN17) on ERK1/2 and p90RSK activation in Chinese hamster ovary cells. Overexpression of RasN17 completely blocked ERK1/2 and p90RSK activation by H₂O₂, suggesting that Ras is required for H₂O₂-induced activation of p90RSK and ERK1/2 (Fig. 10).

View larger version (42K): [in this window] [in a new window]   Fig. 10.   Effect of dominant negative Ras (RasN17) on H2O2-mediated ERK1/2 and p90RSK activation in Chinese hamster ovary (CHO) cells. Chinese hamster ovary cells were transfected with RasN17 and 24 h later washed with medium and maintained in serum-free medium for 24 h. Cells were then treated with 1 mM H2O2 for 5 and 20 min. ERK1/2 and p90RSK activities were assayed as described in Fig. 2. Representative Western blots with anti-phospho-specific ERK1/2 antibody (top panel) and anti-ERK1/2 (middle panel) and autoradiogram for p90RSK activity (bottom panel) of three performed are shown.

Role of Fyn in Ras Activation-- The activity of Ras is regulated through binding of guanine nucleotides, the GTP-bound active form of protein being able to interact with effector proteins such as Raf-1 and PI3-K (45). The kinetics of H₂O₂ stimulated Ras activation have, as yet, not been studied. In the present study, Ras activation was assessed by monitoring the association between Ras-GTP and the RBD of Raf (30). Cell lysates of mouse fibroblasts stimulated with H₂O₂ were subjected to affinity precipitation with GST-RBD protein. The eluted proteins were then subjected to SDS-PAGE and immunoblotted with anti-Ras antibody. Stimulation of fibroblasts with H₂O₂ caused maximal activation of Ras at 2 min, as measured by GST-RBD association (Fig. 11A). In contrast, in Fyn/ fibroblasts, activation of Ras by H₂O₂ was significantly decreased (Fig. 11B).

View larger version (26K): [in this window] [in a new window]   Fig. 11.   H2O2-induced Ras activation: effect in Fyn/ cells. Time course for H2O2-induced Ras activation assayed by GST-RBD (A and B) or Ras-GTP/(GTP+GDP) (C and D). GTP-bound active Ras was isolated from lysates of quiescent and H2O2-stimulated wild type (WT) mouse fibroblast (A) and Fyn/ cells (B) by affinity precipitation with a GST-RBD fusion protein (equal loading) was detected by Coomassie Brilliant Blue staining of the polyacrylamide gel (bottom panels). GTP-bound Ras was analyzed by Western blotting with anti-Ras antibody (top panels). TCL, total cell lysate. There were no differences in the amount of Ras present in the cell extracts as determined by direct immunoblotting (data not shown). Shown is a representative result of three experiments. C and D, serum-starved subconfluent cultures of wild type fibroblasts (C) and Fyn/ cells (D) were labeled in vivo with [32P]orthophosphate and then treated for 2-20 min with H2O2, and Ras-bound guanine nucleotides were isolated. Shown is an autoradiogram of the extracted guanine nucleotides after thin-layer chromatography. The positions of labeled GDP and GTP are indicated. The percentage of Ras in the GTP-bound state is shown under the panels. Shown is a representative result of two experiments.

To confirm that GST-RBD accurately measures endogenous Ras activity, we also determined the formation of RasGTP by the GTP labeling technique (46). In response to H₂O₂, the fraction of GTP-bound Ras was increased from 23% to ~ 80% with peak at 2-5 min (Fig. 11C). There was complete inhibition of Ras activation by H₂O₂ in Fyn/ cells (Fig. 11D), similar to results for binding to GST-RBD (Fig. 11B). These data indicate that Fyn is required for Ras activation by H₂O₂.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

The major findings of this study are that H₂O₂ stimulates p90RSK in a Fyn- and Ras-dependent manner. Redox-sensitive regulation of p90RSK is thus a new function for Fyn. Data that support an essential role for Fyn and Ras in H₂O₂-mediated p90RSK activation include the following. 1) Tyrosine kinase and specific Src family kinases inhibitors, herbimycin A and PP1, blocked p90RSK activation by H₂O₂ at concentrations consistent with a specific effect on Src family kinases. 2) In Fyn/ fibroblasts, there was no p90RSK activation in response to H₂O₂. In contrast, in Src/ fibroblasts, H₂O₂-mediated p90RSK activation was not inhibited. 3) Transfection of dominant negative Ras blocked H₂O₂ activation of p90RSK. 4) In Fyn/ fibroblasts, H₂O₂ failed to activate Ras. Our results are the first to show that Fyn and Ras but not Src are involved specifically in oxidative stress-mediated p90RSK activation.

Based on the present study and previous work from our lab and other investigators we propose a scheme (Fig. 12) for ROS-mediated signal transduction leading to activation of mitogen-activated protein kinases and p90RSK. A novel aspect of this model is that activation of the tyrosine kinases c-Src and Fyn stimulates specific pathways leading to selective activation of BMK1/ERK5 by c-Src and activation of ERK1/2 and p90RSK by Fyn. Previous investigators have suggested that Src family kinases are upstream mediators of redox-sensitive signal transduction based on findings with UV irradiation (4). There are at least nine members of the Src family of cytoplasmic protein kinases. The function of each individual member has been revealed by genetic analysis in mice, which showed partial overlapping function among the Src family members during development (21, 47, 48). Previously we found that c-Src but not Fyn was required for H₂O₂-mediated BMK1/ERK5 activation in fibroblasts (49). In contrast, in the present study, we found that Fyn but not c-Src was required for H₂O₂-mediated p90RSK activation in fibroblasts. These results indicate that c-Src and Fyn have separate roles in ROS-mediated signal transduction (Fig. 12). Future studies will be required to define the precise nature of the downstream substrates for c-Src and Fyn.

View larger version (19K): [in this window] [in a new window]   Fig. 12.   Diagram of H2O2-mediated signal transduction pathways. Previously (49), we found that c-Src but not Fyn was required for H2O2-mediated BMK1/ERK5 activation. In contrast, in the present study, we found that Fyn but not c-Src was required for H2O2-mediated p90RSK activation. These results indicate that c-Src and Fyn have separate roles in oxidative stress-mediated signal transduction. Based on the dominant negative Ras experiments, Ras is essential for both p90RSK and ERK1/2 activation by H2O2.

In addition to activation of p90RSK by MEK1/2-ERK1/2 (path 1 in Fig. 12), the present study demonstrates the presence of another ROS-stimulated pathway that activates p90RSK independently of MEK1/2 (path 2 in Fig. 12). A MEK1/2-independent ERK1/2 activation pathway has been reported previously (50). It appears likely that H₂O₂-mediated p90RSK activation may also occur via an ERK1/2-independent pathway, because PD98059 inhibited H₂O₂-mediated ERK1/2 activation, but not p90RSK activation. A MEK1/2-ERK1/2-independent pathway is supported by findings that phosphorylation of Ser²²² and Ser⁷³³ in p90RSK, which are required to activate p90RSK, is independent of ERK1/2 (19). A candidate kinase is 3-phosphoinositide-dependent protein kinase-1 based on the following information. Ser²²² of p90RSK is in a motif homologous to residues of which the phosphorylation is important for activation of p70RSK, protein kinase B, and protein kinase C. Of interest, the residue equivalent to Ser²²² of p90RSK in protein kinase B is phosphorylated by 3-phosphoinositide dependent protein kinase-1 (19). In fact, recently published data show that p90RSK activation is regulated by both the ERK1/2 pathway and a 3-phosphoinositide-dependent protein kinase-1-dependent pathway (51, 52).

The mechanisms by which specific Src family kinases are selectively activated remains unknown. Src family kinases share several functional domains, including an amino-terminal myristoylation sequence for membrane targeting, a unique domain, SH2 and SH3 domains, a kinase domain, and a carboxyl-terminal noncatalytic domain. Functional properties of the amino-terminal unique region have not been well defined. Current data indicate that this domain may be required for specific interactions between particular Src family kinases and downstream targets. Recently, several laboratories (including ours) found that JAK activation is specifically regulated by Fyn but not Src (53). We have also found that Src and Fyn specifically regulate BMK1 and JAK2, respectively (49, 53). In addition, Wary et al. (54) reported that caveolin-1 functions as a membrane adapter to link the integrin alpha subunit to Fyn, not to Src, and Puceat et al. (55) have found that the anion Cl/HCO₃ exchanger (AE1) associates with Fyn but not Src. It is not clear whether another Src family kinase participates in ROS-induced p90RSK activation, but based on our results, Fyn but not Src is an important regulator of p90RSK activation. Future studies will be required to determine the nature of this specificity for Src family kinases, focusing on the unique amino-terminal domain.

It is not clear why the time course of H₂O₂-mediated p90RSK activation is delayed in Src/ cells (Fig. 8, B and C). Previously, we found that Fyn-JAK2 regulates the early phase of ERK1/2 activation, thus altering the time course for ERK1/2 activation (53). In a similar manner, we believe that Src regulates the early phase of p90RSK activation, and thus causes changes in activation time of p90RSK. As another example, Oh et al. (56) have recently shown that Src and Fyn both regulate SIRP-SHP-2 signaling, and SHP-2 has been reported to regulate ERK1/2 activation. Therefore, it is possible that SHP-2 may be involved in the early phase of p90RSK activation.

Understanding signal transduction pathways activated by H₂O₂ is important, because H₂O₂ is increased in hypertension, inflammation, ischemia-reperfusion injury, and atherosclerosis. It has become clear that the lesions of atherosclerosis represent a series of highly specific cellular and molecular responses that can be described as an inflammatory disease (57). Monocytes, leukocytes, and T lymphocytes are markedly increased in regions of human atheroma (58, 59). How much H₂O₂ is produced by leukocytes and monocytes in vivo is unknown; however, at sites of inflammation, the production of ROS is enhanced, and H₂O₂ can reach relatively high levels (0.1-1.0 mM) (60-64). Antioxidants such as vitamin E and probucol have had significant beneficial effects on coronary events (65) and restenosis of atherosclerotic vessels that underwent angioplasty (66). Therefore, to define the effect of H₂O₂ on T-lymphocytes and other cells in the vessel wall will be important to clarify the mechanisms that contribute to the initiation and progression of cardiovascular disease (57).

	ACKNOWLEDGEMENTS

We thank Dr. D. Shalloway for providing the GST-RBD construct and Dr. Sheila M. Thomas for providing us Src/ and Fyn/ fibroblasts. We also thank Drs. C. Yan, E. Takahashi, H. Ueba, B. Gallis, and G. Daum for invaluable assistance and critical reading of this manuscript.

	FOOTNOTES

^* This study was supported by a grant from the Japanese Heart Foundation, Bayer Yakuhin Research Grant Abroad (to J.-i. A.), and National Institutes of Health Grants HL44721 and HL49192 (to B. C. B.).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.

These authors contributed equally to this work.

^¶ To whom correspondence should be addressed: Cardiology Unit, Box 679, 601 Elmwood Ave., University of Rochester School of Medicine and Dentistry, Rochester, NY 14642. Tel.: 716-273-1947; Fax: 716-273-1497; E-mail: bradford_berk@urmc.rochester.edu.

	ABBREVIATIONS

The abbreviations used are: ROS, reactive oxygen species; BMK1, big mitogen-activated protein kinase 1; CREB, cyclic AMP-response element-binding protein; ERK, extracellular signal-regulated kinase; MNC, human blood peripheral mononuclear cell; PAGE, polyacrylamide gel electrophoresis; RSK, ribosomal S6 kinase; NHE, Na⁺-H⁺ exchanger; GST, glutathione S-transferase; BAPTA-AM, bis-(O-aminophenoxy)ethane-N,N,N',N'-tetraacetic acid acetoxymethyl ester; RBD, Raf binding domain; BAEC, bovine aortic endothelial cells.

	REFERENCES

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES