Proteomic Analysis of Tyrosine Phosphorylations in Vascular Endothelial Growth Factor- and Reactive Oxygen Species-Mediated Signaling Pathway Young Mee Kim, Eun Joo Song, Jawon Seo, Hee-Jung Kim, and Kong-Joo Lee* The Center for Cell Signaling & Drug Discovery Research, College of Pharmacy and Division of Life & Pharmaceutical Sciences, Ewha Womans University, Seoul 120-750, Korea Received July 5, 2006
Vascular endothelial growth factor (VEGF) mediates angiogenic signaling by activating tyrosine kinase receptors. Endothelial cells treated with VEGF are known to increase reactive oxygen species (ROS) production and activate the MAPK pathway. To identify the target proteins of the VEGF receptor, we treated human umbilical vein endothelial cells (HUVECs) with VEGF or H2O2, and identified and semiquantified tyrosine-phosphorylated proteins, combining 2D-gel electrophoresis, Western analysis using antibody against phospho-tyrosine, and mass spectrometry. We detected 95 proteins that were differentially phosphorylated; some were specifically phosphorylated by VEGF but not by H2O2. 2D-gel electrophoresis revealed that heterogeneous populations of the same protein responded differently to H2O2 and VEGF. Bioinformatic studies examining the nature of the differential phosphorylation in various subpopulations of proteins should provide new insights into VEGF- and H2O2-induced signaling pathways. Keywords: proteomic analysis • VEGF • ROS • tyrosine phosphorylation • 2D-gel electrophoresis • mass spectrometry • tubule formation • HUVECs
Introduction Vascular endothelial growth factor (VEGF) is known to play a significant role in angiogenesis.1 In endothelial cells, VEGF binds to two homologous membrane receptors of tyrosine kinase, Flt-1 (VEGFR-1) and KDR (VEGFR-2). Tyrosine phosphorylation of KDR, mediated by VEGF, initiates angiogenic signaling via stimulation of the MAPK cascade.2 However, how these pathways are regulated is not well-understood. Recent studies showed that transient production of reactive oxygen species (ROS), including H2O2 and O2-, might serve as an important signaling event in the pathways triggered by receptor tyrosine kinases (RTKs) such as EGFR, FGFR, PDGFR, Tie2, and InsR.3,4 VEGFR has been reported to bind to VEGF and induce angiogenesis by increasing ROS production and activating various signaling molecules including MAPK, PI3 kinase, and Akt.5,6 Only a few of the signaling molecules involved in the kinase-mediated angiogenesis have been identified .7 We report here the identification and semiquantitation of the tyrosinephosphorylated proteins involved in VEGF- and ROS-signaling cascades by proteomic analysis combining 2D gel separation, Western analysis, and mass spectrometry.
Materials and Methods Cell Culture and Treatment with H2O2 or VEGF. Human umbilical vein endothelial cells (HUVECs) (AngioLab, Korea), * To whom correspondence should be adressed at College of Pharmacy and Division of Life & Pharmaceutical Sciences, Ewha Womans University, Seoul 120-750, Korea. Tel., 82-2-3277-3038; fax, 82-2-3277-3760; e-mail,
[email protected]. 10.1021/pr060326s CCC: $37.00
2007 American Chemical Society
primary endothelial cells, were prepared 80% confluent in 100 mm dishes at passage 4-6, in Medium 199 (GIBCO BRL) containing 10% fetal bovine serum (FBS), heparin (100 µg/mL, Sigma), endothelial cell growth supplement (ECGS, 50 µg/mL, Sigma), 100 µg/mL streptomycin, and 100 units/mL penicillin G, at 37 °C, in an atmosphere of 5% CO2 and 95% air. The cells used in the present study were always obtained at passage 6. The HUVECs were incubated in Medium 199 containing 2% FBS overnight (>16 h), incubated with 50 ng/mL VEGF or 0.5 mM H2O2 for 15 min at 37°C, and washed twice with HBSS. The treated cells were trypsinized and washed twice with icecold phosphate-buffered saline (PBS) prior to use. Measurements of Intracellular ROS. VEGF- or H2O2-treated HUVECs were incubated for 5 min with 5 µM 2′,7′-dichlorofluorescein diacetate (DCFH-DA) which is converted to 2′,7′dichlorofluorescein by intracellular esterase. The latter was then oxidized by ROS to the highly fluorescent 2′,7′-dichlorofluorescein (DCF). The fluorescence of each dish was immediately analyzed by Zeiss confocal laser-scanning fluorescence microscope (LSM 510) at excitation wavelength, 488 nm, and emission at 515 or 540 nm. All measurements were done at least in triplicate. Immunoblot Analysis. VEGF- or H2O2-treated cells were lysed and immunoprecipitated with anti-phosphotyrosine antibody. Cell lysates or immunoprecipitants were detected with Western analysis using various antibodies including antiphosphorylated ERK and anti-phosphorylated p38 (New England Biolabs, Inc.), anti-phosphotyrosine (4G10, Upstate Journal of Proteome Research 2007, 6, 593-601
593
Published on Web 01/09/2007
research articles Biotechnology, Inc.), anti-GAPDH (Labfrontier, Inc.), anti-PDI (Stressgen), anti-non-POU (BD Transduction Labs.), and antihnRNP A2 at a 1:1000 dilution in PBST at 4 °C overnight. The protein-antibody complex was visualized with horseradish peroxidase-conjugated secondary antibody at a 1:2000 dilution. The immunoblots were then incubated for 30 s in the ECL plus kit (Amersham Biosciences) and detected by LAS 1000 (Fuji photofilm Co.) or exposed to X-ray film (Hyper-film, Amersham Biosciences). In Vitro Kinase Assay. Immunoprecipitates obtained by treating HUVECs with anti-JNK antibody (Santa Cruz Biotechnology) were assayed for SAPK/JNK activity, using glutathioneS-transferase (GST)-c-Jun 1-89 peptide (provided by Dr. J. R. Woodgett) as a substrate, as described previously.8 In Vitro Tubule Formation. In vitro tubule formation of HUVECs was assayed on Matri-gel (Becton Dickinson). A total of 200 µL of Matri-gel was coated on a 24-well plate and allowed to form a solid gel by incubation for 30 min at 37 °C. The serum-starved cells were incubated in the presence or absence of 200 µM DPI for 30 min and then treated with H2O2 (0.5, 0.25, or 0.125 mM) or VEGF (25 ng/mL) for 15 min. The stimulated cells were seeded onto Matri-gel coated wells in M-199 containing 10% FBS and incubated for 7-24 h. The images of cells were observed with an Olympus digital camera and analyzed for tubule length with Image Pro plus software (Media Cybernetics, Inc.). Two-Dimensional Gel Electrophoresis. The protein mixtures were separated in IPG Ready DriStrips gel by electrofocusing in 7-cm Immobiline DryStrips (pH 4-7, Amersham Biosciences) with IPGphor as described previously.9 The protein mixtures in pH 6-10 were separated in 10 cm microtube gel (9.2 M urea, 2% Triton X-100, 4% acrylamide, 1% 3.5-10 ampholyte, and 1% 6-8 ampholyte) by electrofocusing. After focusing, the gels were soaked in cold 1× gel sample buffer, loaded onto 11% two-dimensional SDS-PAGE, and stained with silver or electroblotted onto a PVDF membrane. Quantitative Analysis of Tyrosine Phosphorylation. 2D-gel images were scanned using an image scanner (Amersham Biosciences) and semiquantitatively analyzed using ImageMaster software (Amersham Biosciences). Each sample for antiphosphotyrosine antibody was assayed in triplicate together with internal standards of phosphotyrosine of known molecular weights (Upstate Biotechnology, Inc.), and the ratio of intensity of sample to internal standard was used for quantitation. The quantified phosphorylated spots having a p-value less than 0.05 in their intensity were identified with mass spectrometry. Protein Analysis with Mass Spectrometry. Spots from 1∼10 sets of 2D gels, depending on protein abundance, were digested in-gel and analyzed using MS peptide fingerprinting using Voyager DE-STR MALDI-TOF MS (Applied Biosystem, Boston, MA) as described previously.9 The autolytic peaks of trypsin (842.5099 and 2211.1046) were used for internal calibration. This procedure typically results in mass accuracies of 30 ppm. For interpretation of the mass spectra, we used the MS-FIT program available on the World Wide Web site of the University of California at San Fransisco (http://prospector.ucsf.edu/). For confirmation, selected samples were separated in IPG Ready DriStrips gel by electrofocusing in 18-cm Immobiline DryStrips (pH 4-7, Amersham Biosciences) with IPGphor and were sequenced with nanoLC-ESI-Q-TOF tandem MS (Q-TOF API US, Micromass/Waters Corp.). For nanoLC-ESI-Q-TOF MS, the rest of the digested peptides were reconstituted in 10 µL of 0.1% trifluoroacetic acid (TFA) and desalted with ZipTip C18 (Mil594
Journal of Proteome Research • Vol. 6, No. 2, 2007
Kim et al.
lipore Co.) according to the manufacturer’s instructions. The peptides were fractionated with 10 µL of 5, 30, 50, and 80% acetonitrile (ACN)/0.1% TFA. ESI-MS/MS coupled with reversephase liquid chromatography (15 cm × 75 µm, C18 column) experiments were performed with a Q-TOF equipped with a nanoflow Z-spray ion source. For phosphoprotein identification, fragment masses of MS/MS spectra were matched with the theoretical fragment masses of all human proteins in the Swiss-Prot database using Proteinlynx 2.1 software program (Waters Inc., MA) and MODi.10 For nonredundant MS acquisition, selected exclusion and inclusion lists were applied to phosphoprotein analysis. Bioinformatic Prediction of Tyrosine Phosphorylated Proteins. The phosphorylation sites of proteins were predicted by two programs: NetPhos, which has low stringency (www. cbs.dtu.dk/services/NetPhos), and ScanProsite, which has high stringency (us.expasy.org/tools/scanprosite/). PTKs involved in these phosphorylations, and the proteins potentially binding to tyrosine-phosphorylated proteins, were predicted by Scansite program (scansite.mit.edu/). The protein domains were predicted by Motif Scan (hits.isb-sib.ch/cgi-bin/index).
Results VEGF- and H2O2-Mediated Signaling Pathways. We first examined the global phosphorylation patterns of VEGF- or H2O2-treated HUVECs. To confirm that VEGF is involved in ROS signaling pathways, we exposed HUVECs to 50 ng/mL VEGF or 0.5 mM H2O2 as positive control for 15 min (Figure 1A) and measured DCF fluorescence produced immediately after stimulation. We found that ROS was generated by VEGF in a concentration-dependent manner (data not shown) and the fluorescence with VEGF was roughly equivalent to that observed with H2O2 treatment. We next assessed the effect of scavenging ROS induced by VEGF. We found that pretreatment of the cells with diphenylene-iodonium chloride (DPI), an NADPH oxidase inhibitor, blocked the VEGF-induced ROS generation. This suggests that VEGF induces ROS in an NADPH oxidase-dependent manner in HUVECs as reported previously.6 To determine whether ROS generation by VEGF is sufficient for angiogenic signaling pathway, we examined the activations of ERK, p38 MAPK, and SAPK/JNK, and the up-regulation of VEGF production by VEGF or H2O2 in the presence or absence DPI pretreatment (Figure 1B). ERK and p38 MAPK showed activation in response to both VEGF and H2O2. Phosphorylation of p38 MAPK was dependent on ROS generation, while the activation of ERK by H2O2 occurred in an ROS-independent manner. We also determined the activation of SAPK/JNK in an in vitro kinase assay using GST-c-jun as substrate. We found that SAPK/JNK was activated significantly by H2O2 in an ROSdependent manner, but only slightly by VEGF. Next, we compared the up-regulation of VEGF production by H2O2 or VEGF. HUVECs, in the presence or absence of DPI treatment for 30 min, were exposed to 0.5 mM H2O2 or 50 ng/mL VEGF for 15 min and recovered for 20 h. VEGF production was significantly up-regulated by VEGF, but only weakly by H2O2, in an ROS-dependent manner (Figure 1B). To determine whether the VEGF- and H2O2-induced differentiations of HUVECs into tubules were different, cells exposed to VEGF or H2O2 with or without DPI were incubated on Matri-gel. As shown in Figure 1C, tubule formation by VEGF partially occurred in an ROS-dependent manner, while H2O2 treatment significantly reduced tubule formation in a concentration-dependent manner. It appears that VEGF induced both ROS generation and tubule formation in HUVECs, but ROS generation is not
Proteomic Analysis of VEGF and ROS Signaling
Figure 1. Effect of H2O2 and VEGF treatment on (A) ROS generation, (B) MAPK activation, and (C) tubule formation. (A) Cells incubated with media in the presence or absence of 20 µM DPI for 30 min were treated with 0.5 mM H2O2 or 50 ng/mL VEGF for 15 min at 37 °C. Cells were washed twice with HBSS and incubated with 5 µM DCFH-DA, and DCF fluorescence of each dish was immediately analyzed as described under Materials and Methods. All experiments were at least triplicated. (B) Cells incubated in the presence or absence of 20 µM DPI for 30 min and then treated with H2O2 (0.5 mM) or VEGF (50 ng/mL) for 15 min. Phosphorylated ERK and p38 were detected with antibodies against phosphorylated ERK and phosphorylated p38, respectively, followed by HRP-conjugated secondary antibody. SAPK/ JNK activities were assayed using glutathione-S-transferase (GST)-c-Jun 1-89 peptide as a substrate after immunoprecipitation with anti-JNK antibody. Cells incubated with media with or without 20 µM DPI for 30 min were treated with H2O2 (0.5 mM) or VEGF (50 ng/mL) for 15 min and then recovered for 20 h. The expression of VEGF was detected with VEGF monoclonal antibody followed by HRP-conjugated secondary antibody. (C) The serumstarved cells were incubated with or without 20 µM DPI for 30 min and then treated with H2O2 (0.5, 0.25, and 0.125 mM) or VEGF (25 ng/mL) for 15 min. The stimulated cells were seeded onto Matri-gel-coated wells in M-199 containing 10% FBS. The cells were incubated for 7-24 h.
research articles sufficient for tubule formation. This suggests that the responses of endothelial cells to VEGF and H2O2 have some common, as well as distinguishing, characteristics. Identification of Phosphorylated Proteins by VEGF or H2O2 Treatment. We then examined tyrosine-phosphorylated proteins involved in VEGF and H2O2 signaling pathway using proteomic analyses. HUVECs were treated with 0.5 mM H2O2 or 25 ng/mL VEGF for 15 min and recovered for various times. This treatment resulted in no changes in protein expression level on 2D gel at this time point. Tyrosine phosphorylations in response to both VEGF and H2O2 immediately increased but rapidly recovered to control levels (data not shown). Cellular proteins harvested immediately after each treatment were separated on 2D gel and detected with silver staining or with immunostaining using anti-phosphotyrosine monoclonal antibody, shown in Figure 2. We detected 95 protein spots that showed significant changes in phosphorylation in response to H2O2 or VEGF (p-value less than 0.05) at two pI ranges. Fiftyfive protein spots were detected in the pI range of 4-7 (Figure 2A,B) and 40 protein spots in the pI range 6-10 (Figure 2C,D). The gels visualized by silver staining are shown in Figure 2A,C. Those detected with immunostaining using anti-phosphotyrosine antibody and ECL plus chemiluminescence detection kit are shown in Figure 2B,D. The detailed immunoblots of tyrosine-phosphorylated proteins induced by VEGF or H2O2, respectively, are presented in Figure 2E. We performed proteomic analysis of the tyrosine-phosphorylated proteins involved in VEGF- and H2O2-mediated signaling. Protein spots detected on the immunoblots were excised from the silver-stained gels and subjected to in-gel digestion with trypsin followed by mass spectrometric analyses using a sensitive and accurate detection protocol developed in this laboratory. We identified 63 proteins from 92 of 95 spots using MALDI-TOF MS (Table 1); some proteins were confirmed by sequencing using nanoLC-ESI-Q-TOF tandem MS (Supporting Information Table 1). The results were analyzed for various functions using gene ontology (www.godatabase.org) and for predicted domains of the proteins using Motif Scan (hits.isbsib.ch/cgi-bin/index). We identified 14 proteins (26 spots) acting in nucleic acid binding, 15 proteins (24 spots) with cell motility and structural function, 15 proteins (18 spots) acting in cell growth or maintenance, 5 proteins (9 spots) functioning as chaperones/mediators of protein folding, 3 (3 spots) with signaling function, 9 proteins (9 spots) relating to proteasome degradation, 2 proteins (2 spots) acting as ligand binders or carriers, and 1 protein (1 spot) functioning in cell communication. Twenty-four of these have been previously reported to function in tyrosine-phosphorylations, while 39 proteins are new. Three proteins (spots 69, 74, 76) were present in low abundance (
