Vascular Endothelial Growth Factor Causes Translocation of p47 to Membrane Ruffles through WAVE1
2003; Elsevier BV; Volume: 278; Issue: 38 Linguagem: Inglês
10.1074/jbc.m302251200
ISSN1083-351X
AutoresRu Feng Wu, Ying Gu, You Xu, Fiemu E. Nwariaku, Lance S. Terada,
Tópico(s)Phagocytosis and Immune Regulation
ResumoGrowth factors initiate cytoskeletal rearrangements tightly coordinated with nuclear signaling events. We hypothesized that the angiogenic growth factor, vascular endothelial growth factor (VEGF), may utilize oxidants that are site-directed to a complex critical to both cytoskeletal and mitogenic signaling. We identified the WASP-family verprolin homologous protein-1 (WAVE1) as a binding partner for the NADPH oxidase adapter p47phox within membrane ruffles of VEGF-stimulated cells. Within 15 min of VEGF stimulation, p47phox coprecipitated with WAVE1, with the ruffle and oxidase agonist Rac1, and with the Rac1 effector PAK1. VEGF also increased p47phox phosphorylation, oxidant production, and ruffle formation, all of which were dependent upon PAK1 kinase activity. The antioxidant Mn (III) tetrakis(4-benzoic acid) porphyrin and ectopic expression of either the p47-binding WAVE1 domain or the WAVE1-binding p47phox domain decreased VEGF-induced ruffling, whereas the active mutant p4-(S303D, S304D,S328D) stimulated oxidant production and formation of circular dorsal ruffles. Both kinase-dead PAK1-(K298A) and Mn (III) tetrakis(4-benzoic acid) porphyrin decreased c-Jun N-terminal kinase (JNK) activation by VEGF, whereas dominant-negative JNK did not block ruffle formation, suggesting a bifurcation of mitogenic and cytoskeletal signaling events at or distal to the oxidase but proximal to JNK. Thus, WAVE1 may act as a scaffold to recruit the NADPH oxidase to a complex involved with both cytoskeletal regulation and downstream JNK activation. Growth factors initiate cytoskeletal rearrangements tightly coordinated with nuclear signaling events. We hypothesized that the angiogenic growth factor, vascular endothelial growth factor (VEGF), may utilize oxidants that are site-directed to a complex critical to both cytoskeletal and mitogenic signaling. We identified the WASP-family verprolin homologous protein-1 (WAVE1) as a binding partner for the NADPH oxidase adapter p47phox within membrane ruffles of VEGF-stimulated cells. Within 15 min of VEGF stimulation, p47phox coprecipitated with WAVE1, with the ruffle and oxidase agonist Rac1, and with the Rac1 effector PAK1. VEGF also increased p47phox phosphorylation, oxidant production, and ruffle formation, all of which were dependent upon PAK1 kinase activity. The antioxidant Mn (III) tetrakis(4-benzoic acid) porphyrin and ectopic expression of either the p47-binding WAVE1 domain or the WAVE1-binding p47phox domain decreased VEGF-induced ruffling, whereas the active mutant p4-(S303D, S304D,S328D) stimulated oxidant production and formation of circular dorsal ruffles. Both kinase-dead PAK1-(K298A) and Mn (III) tetrakis(4-benzoic acid) porphyrin decreased c-Jun N-terminal kinase (JNK) activation by VEGF, whereas dominant-negative JNK did not block ruffle formation, suggesting a bifurcation of mitogenic and cytoskeletal signaling events at or distal to the oxidase but proximal to JNK. Thus, WAVE1 may act as a scaffold to recruit the NADPH oxidase to a complex involved with both cytoskeletal regulation and downstream JNK activation. Following stimulation with growth factors, parenchymal cells enter a proliferative state and migrate directionally. Accordingly, growth factors initiate both mitogenic signaling through the MAPKs 1The abbreviations used are: MAPK, mitogen-activated protein kinase; PDGF, platelet-derived growth factor; GFP, green fluorescent protein; JNK, c-Jun N-terminal kinase; HUVEC, human umbilical vein endothelial cell; HA, hemagglutinin; EGFP, enhanced GFP; GST, glutathione S-transferase; Ad, adenovirus; VEGF, vascular endothelial growth factor; DCF, dichlorofluorescein diacetate; SH, Src homology; PX, Phox; DN, dominant-negative; WASP, Wiskott-Aldrich syndrome protein; PAK, p21-activated kinase; BD, binding domain; DsRad, Discoma sp. Red; SOD, superoxide dismutase.1The abbreviations used are: MAPK, mitogen-activated protein kinase; PDGF, platelet-derived growth factor; GFP, green fluorescent protein; JNK, c-Jun N-terminal kinase; HUVEC, human umbilical vein endothelial cell; HA, hemagglutinin; EGFP, enhanced GFP; GST, glutathione S-transferase; Ad, adenovirus; VEGF, vascular endothelial growth factor; DCF, dichlorofluorescein diacetate; SH, Src homology; PX, Phox; DN, dominant-negative; WASP, Wiskott-Aldrich syndrome protein; PAK, p21-activated kinase; BD, binding domain; DsRad, Discoma sp. Red; SOD, superoxide dismutase. as well as actin cytoskeletal rearrangements associated with migration, such as membrane ruffling. However, the relationship between these two processes and the signal pathways leading to these events have not been clearly defined.Reactive oxidants have been linked to both proliferative pathways and cytoskeletal dynamics and thus may present a common mediator of these two facets of growth factor signaling. PDGF, epidermal growth factor, and insulin initiate a small burst of oxidants critical for protein tyrosine phosphorylation, MAPK activation, and DNA synthesis, whereas oxidant scavengers suppress these events (1Bae Y.S. Kang S.W. Seo M.S. Baines I.C. Tekle E. Chock P.B. Rhee S.G. J. Biol. Chem. 1997; 272: 217-221Abstract Full Text Full Text PDF PubMed Scopus (1089) Google Scholar, 2Sundaresan M. Yu Z.X. Ferrans V.J. Irani K. Finkel T. Science. 1995; 270: 296-299Crossref PubMed Scopus (2300) Google Scholar, 3Heffetz D. Zick Y. J. Biol. Chem. 1989; 264: 10126-10132Abstract Full Text PDF PubMed Google Scholar). Rac1, through its involvement with the NADPH oxidase, also mediates H-RasV12-induced mitogenesis and PDGF-dependent cyclin D1 expression (4Irani K. Xia Y. Zweier J.L. Sollott S.J. Der C.J. Fearon E.R. Sundaresan M. Finkel T. Goldschmidt-Clermont P.J. Science. 1997; 275: 1649-1652Crossref PubMed Scopus (1421) Google Scholar, 5Page K. Li J. Hodge J.A. Liu P.T. Vanden Hoek T.L. Becker L.B. Pestell R.G. Rosner M.R. Hershenson M.B. J. Biol. Chem. 1999; 274: 22065-22071Abstract Full Text Full Text PDF PubMed Scopus (123) Google Scholar). The use of oxidants is not limited to pathways linked to tyrosine kinase receptors, because smooth muscle NADPH oxidase is activated during the proliferative signaling of both angiotensin II and the ganglioside GD3 (6Bhunia A.K. Schwarzmann G. Chatterjee S. J. Biol. Chem. 2002; 277: 16396-16402Abstract Full Text Full Text PDF PubMed Scopus (61) Google Scholar, 7Ushio-Fukai M. Alexander R.W. Akers M. Yin Q. Fujio Y. Walsh K. Griendling K.K. J. Biol. Chem. 1999; 274: 22699-22704Abstract Full Text Full Text PDF PubMed Scopus (495) Google Scholar).Oxidants have also been linked to cytoskeletal reorganization. High levels of exogenous oxidants increase recruitment of actin to the cytoskeleton, and conversely actin cytoskeletal disruption precipitates oxidant-dependent NF-κβ activation (8Zhao Y. Davis H.W. J. Cell Physiol. 1998; 174: 370-379Crossref PubMed Scopus (91) Google Scholar, 9Kheradmand F. Werner E. Tremble P. Symons M. Werb Z. Science. 1998; 280: 898-902Crossref PubMed Scopus (329) Google Scholar). Oxidant production is also most prominent in actively migrating endothelial cells at the margin of a monolayer wound and, in particular, is concentrated in membrane ruffles of these highly motile cells (10Moldovan L. Moldovan N.I. Sohn R.H. Parikh S.A. Goldschmidt-Clermont P.J. Circ. Res. 2000; 86: 549-557Crossref PubMed Scopus (155) Google Scholar). This localization of oxidant production is comparable to that found in phorbol-ester-stimulated neutrophils whose oxidative burst is concentrated within ruffling membranes (11Heyworth P.G. Robinson J.M. Ding J. Ellis B.A. Badwey J.A. Histochem. Cell Biol. 1997; 108: 221-233Crossref PubMed Scopus (67) Google Scholar). Consistent with these observations, we recently found constitutive localization of the NADPH oxidase subunit p47phox to the cytoskeleton of endothelial cells with translocation of p47-GFP to membrane ruffles (12Gu Y. Xu Y.C. Wu R.F. Souza R.F. Nwariaku F.E. Terada L.S. Exp. Cell Res. 2002; 272: 62-74Crossref PubMed Scopus (68) Google Scholar).In the present study, we identify the ruffle catalyst WAVE1/suppressor of cAMP receptors as a binding partner of p47phox. Upon stimulation of endothelial cells with VEGF, p47phox associates with WAVE1 as well as with Rac1 and PAK1, proteins involved in both ruffle formation and oxidase activation. p47phox-dependent oxidants appear to mediate VEGF-induced ruffle formation, tyrosine phosphorylation, and JNK activation.EXPERIMENTAL PROCEDURESPlasmid Construction—Full-length WAVE1 was PCR-amplified from a HUVEC library between exogenous EcoRI and SalI sites and captured in pCR4-Blunt-TOPO (Invitrogen) and then subcloned into pCIneo (pCIN-WAVE1). Full-length WAVE1, WAVE1-(1-442), WAVE1-(1-362), and WAVE1-(1-269) were amplified from pCIN-WAVE1 with the addition of terminal EcoRI and BamHI sites and ligated into the yeast shuttle vector pGADT7. All other WAVE1 truncations were similarly amplified from the original library clone retrieved from the two-hybrid screen and ligated into pGADT7. pCIN-HA was produced by the insertion of a 3× HA tag upstream of the EcoRI site of pCI-neo. WAVE1-(270-559), WAVE1-(1-269), and WAVE1-(100-442) were then subcloned into pCIN-HA with frame correction. WAVE1-(270-442) and p47-(153-286) were PCR-amplified with the addition of terminal EcoRI and SalI sites and ligated into pCIN-HA and pCI-neo, respectively. Gal4-BD-p47 fusions in pGBKT7 were derived as previously described (13Xu Y.C. Wu R.F. Gu Y. Yang Y.S. Yang M.C. Nwariaku F.E. Terada L.S. J. Biol. Chem. 2002; 277: 28051-28057Abstract Full Text Full Text PDF PubMed Scopus (71) Google Scholar) or through amplification between exogenous EcoRI and SalI sites with ligation into pGADT7 (Clontech). p47-(1-298) was amplified with the addition of terminal EcoRI sites to subclone into pGEX-2TK. pGEX-p47-(full-length) was described previously (13Xu Y.C. Wu R.F. Gu Y. Yang Y.S. Yang M.C. Nwariaku F.E. Terada L.S. J. Biol. Chem. 2002; 277: 28051-28057Abstract Full Text Full Text PDF PubMed Scopus (71) Google Scholar). WAVE1 was fused to the N terminus of EGFP through amplification of pCIN-WAVE1 with the addition of terminal EcoRI and BamHI sites and removal of the endogenous stop codon and ligation into pEGFP-N3 (Clontech). p47phox was fused to the C terminus of DsRed by ligation of full-length p47phox into the EcoRI site of pDsRed2-C1 (Clontech). HA-JNK1, HA-JNK2, p47-GFP, and pCINF-p47 were previously described (12Gu Y. Xu Y.C. Wu R.F. Souza R.F. Nwariaku F.E. Terada L.S. Exp. Cell Res. 2002; 272: 62-74Crossref PubMed Scopus (68) Google Scholar, 13Xu Y.C. Wu R.F. Gu Y. Yang Y.S. Yang M.C. Nwariaku F.E. Terada L.S. J. Biol. Chem. 2002; 277: 28051-28057Abstract Full Text Full Text PDF PubMed Scopus (71) Google Scholar). pCINF-WAVE1-(100-442) was derived by amplification of the WAVE1 fragment between exogenous EcoRI and SalI sites and ligation into pCINF. The HA-tagged APF mutants of human JNK1 and JNK2 were a gift of Dr. Lynn Heasley. Actin-GFP was obtained from Clontech. Myc-PAK1-(K298A) was a gift of Dr. Melanie Cobb. All of the PCR reactions were performed with either PfuTurbo (Stratagene) or Pfx Platinum (Invitrogen), and all of the constructs were confirmed by sequencing.Yeast Two-hybrid Screening and Mating—AH109 yeast (Clontech) were stably transformed with pGBKT7-p47-(1-298) and tested negative for autonomous transactivation. A HUVEC library previously dropped out in pGAL4-AD (13Xu Y.C. Wu R.F. Gu Y. Yang Y.S. Yang M.C. Nwariaku F.E. Terada L.S. J. Biol. Chem. 2002; 277: 28051-28057Abstract Full Text Full Text PDF PubMed Scopus (71) Google Scholar) was used to secondarily transform bait-containing yeast, which were selected under high stringency for bait and library vectors (reversal of Trp and Leu auxotrophy) and interaction (reversal of His and Ade auxotrophy and lacZ expression). Surviving colonies were recloned and retested, and library plasmids were extracted, passaged through Escherichia coli, characterized, and transformed into AH109 yeast. Library plasmid-containing AH109 yeast were mated with bait-containing Y187 yeast, and diploids were plated and assessed for lacZ expression with a filter lift assay. Truncated forms of p47phox and WAVE1 were tested for interaction using a similar yeast mating strategy. Strongly positive interactions caused blue color development within 2 h. Negative interactions remained white for >24 h. Weak positive interactions caused distinct blue coloration between 2 and 24 h.GST Pull Down Assay—BL21-RP E. coli (Stratagene) harboring pGEX-2TK, pGEX-p47, or pGEX-p47-(1-289) were induced with isopropyl-1-thio-β-d-galactopyranoside for 4 h, and the soluble proteins were captured on glutathione-Sepharose 4B (Amersham Biosciences). Full-length or truncated WAVE1 proteins were transcribed and translated in vitro from pCIN-WAVE1 in the presence of [35S]methionine using the TnT Quick Coupled system (Promega). 5 μg of the fusion proteins were incubated with 10 μl of translation mixture for 2 h at 4 °C prior to extensive washing.Adenoviral Construction—Adenoviruses harboring lacZ and p47phox were described previously (12Gu Y. Xu Y.C. Wu R.F. Souza R.F. Nwariaku F.E. Terada L.S. Exp. Cell Res. 2002; 272: 62-74Crossref PubMed Scopus (68) Google Scholar, 13Xu Y.C. Wu R.F. Gu Y. Yang Y.S. Yang M.C. Nwariaku F.E. Terada L.S. J. Biol. Chem. 2002; 277: 28051-28057Abstract Full Text Full Text PDF PubMed Scopus (71) Google Scholar). Sequential site-directed mutations were introduced with PCR to create p47-(S303D,S304D,S328D), which was then subcloned into pShuttle and then pAdX (Clontech). Ad-p47-(W193R) and Ad-p47-(S303D,S304D,S328D) were generated, cloned, and characterized in human embryonic kidney 293 cells. Efficiency of transgene expression in HUVEC assessed by lacZ expression was >95%. Expression of p47-wild type, p47-(W193R), and p47-(S303D,S304D,S328D) proteins were equivalent by immunoblot at equal m.o.i.Cell Culture and Microscopy—HUVEC (Clonetics) used in this study were propagated in EGM-2 medium containing 2% fetal calf serum, epidermal growth factor, fibroblast growth factor, VEGF, insulin growth factor, ascorbic acid, hydrocortisone, heparin, and gentamicin. Transfection was performed after cell synchronization 6-8 h after thymidine release using electroporation with a constant amount of total plasmid DNA. For adenoviral transduction, cells were infected for 1 h at an m.o.i. of 200:1, and harvested 24 h later. For microscopy experiments, cells were plated on fibronectin-coated chambered glass cover-slips and examined live. Prior to VEGF stimulation, cells were serum-starved in medium containing 0.1% fetal calf serum without growth factors for 24 h. For some experiments, cells were treated with 100 μm SOD mimetic MnTBAP (Calbiochem) for 1 h at 37 °C. The number of ruffling cells was counted in all of the actin-GFP-expressing cells in at least ten random fields in at least three separate slides between 15 and 20 min after the addition of VEGF (30 ng/ml, PeproTech). Confocal images were obtained with a Zeiss Axiovert S100TV and LSM 410 laser-scanning system. In some experiments, cells were fixed, permeablized, and stained with 5 units of rhodamine phalloidin (Molecular Probes) prior to examination (12Gu Y. Xu Y.C. Wu R.F. Souza R.F. Nwariaku F.E. Terada L.S. Exp. Cell Res. 2002; 272: 62-74Crossref PubMed Scopus (68) Google Scholar). Fluorescence ratio imaging was performed using Zeiss LSM software, version 3.98.Immunoprecipitation and Western Immunoblot—FLAG-p47-transfected HUVEC were rocked in ice-cold lysis buffer (20 mm Tris-HCl, pH 7.5, 150 mm NaCl, 1 mm Na2 EDTA, 1 mm EGTA, 1% Triton X-100, 2.5 mm sodium pyrophosphate, 1 mm β-glycerophosphate, 1 mm Na3VO4, 1 μg/ml leupeptin, and 1 mm phenylmethylsulfonyl fluoride) for 20 min, harvested, and sonicated once for 3 s. Insoluble material was removed by centrifugation at 4 °C for 20 min at 10,000 × g prior to immunoprecipitation with anti-FLAG (M2, Sigma). Endogenous and adenovirally transduced p47phox were immunoprecipitated using mouse anti-p47phox (Upstate Biotechnology). Immunoblots were performed with antibodies against WAVE1 (Upstate Biotechnology), p47phox (Upstate Biotechnology and a gift of Dr. Bernard Babior), PAK1 (Santa Cruz Biotechnology), Rac1 (Transduction Laboratories), and phosphotyrosine (Santa Cruz Biotechology).p47phox Phosphorylation and Oxidant Production—HUVEC were transfected with pCINF-p47, recovered for 24 h, and then serum-starved for an additional 20 h. Culture medium was replaced with phosphate-free Dulbecco's modified Eagle's medium (Sigma) containing 0.5 mCi/ml [32P]orthophosphate for 4 h. Cells were then stimulated with VEGF (30 ng/ml) for 15 min, and cell lysates were subjected to immunoprecipitation with anti-FLAG, PAGE separation, and autoradiography. A third of the immunoprecipitate was immunoblotted for p47phox to assess capture and loading. Oxidant production was assessed as oxidation of 2′,7′-dichlorofluorescin diacetate (DCF, Molecular Probes). HUVEC were washed and loaded with DCF (10 μm) in Hanks' balanced salt solution for 20 min at 37 °C. They were then washed and exposed to VEGF (30 ng/ml for 15 min), and DCF fluorescence was measured as described previously (12Gu Y. Xu Y.C. Wu R.F. Souza R.F. Nwariaku F.E. Terada L.S. Exp. Cell Res. 2002; 272: 62-74Crossref PubMed Scopus (68) Google Scholar).JNK and PAK Kinase Assays—Following stimulation, cells were washed with ice-cold phosphate-buffered saline, scraped, pelleted, and lysed in lysis buffer. HA-JNK or PAK was immunoprecipitated with anti-HA or anti-PAK1 (Santa Cruz Biotechnology), and the immunocomplexes were washed twice with lysis buffer and once with kinase buffer. The immunocomplex kinase assay contained either 2 μg of GST-Jun (Santa Cruz Biotechnology) for JNK or 5 μg of myelin basic protein (Upstate Biotechnology) for PAK as a substrate in the presence of [γ-32P]ATP (PerkinElmer Life Sciences) at 30 °C for 30 min (13Xu Y.C. Wu R.F. Gu Y. Yang Y.S. Yang M.C. Nwariaku F.E. Terada L.S. J. Biol. Chem. 2002; 277: 28051-28057Abstract Full Text Full Text PDF PubMed Scopus (71) Google Scholar).RESULTSWAVE1 Binds p47phox—p47phox in its unstimulated state is thought to exist in a folded conformation with intramolecular interactions between its C terminus and its core SH3 domains (14Inanami O. Johnson J.L. McAdara J.K. Benna J.E. Faust L.R. Newburger P.E. Babior B.M. J. Biol. Chem. 1998; 273: 9539-9543Abstract Full Text Full Text PDF PubMed Scopus (150) Google Scholar). Phosphorylation of multiple serines within this C terminus results in the unmasking of these core binding domains. Therefore, we truncated the autoinhibitory C terminus and fused the remaining protein (p47-(1-298)) to Gal4-BD. Upon screening our HUVEC library, a clone representing the C-terminal 459 residues of WAVE1 in-frame with Gal4-AD was recovered. This clone survived auxotrophic and lacZ expression selection and was strongly positive upon retesting using a yeast-mating strategy. In addition, full-length WAVE1 bound robustly to GST-p47-(1-298) in vitro (Fig. 1a), indicating a direct interaction between the two proteins and suggesting the masking of the WAVE1 binding domain by the full-length p47phox protein. The two smaller labeled proteins in the WAVE1 translation reaction are presumably incompletely translated or partially proteolyzed WAVE1 truncations and were not present in reactions not containing the WAVE1 template (data not shown).Using the yeast-mating technique, general binding domains were identified on both proteins. The WAVE1-binding region of p47phox was contained within the central tandem SH3 domains with the N-proximal domain (SH3a) in isolation binding more robustly than the isolated distal (SH3b) domain (Fig. 1c). Both p47-(1-205), containing the N-terminal Phox (PX) and SH3a domains, and p47-(205-390), containing the SH3b domain and C-terminal region, failed to bind WAVE1, presumably because of SH3 surface masking by proline-rich sequences within the adjacent PX and C-terminal regions, respectively. Interestingly, p47-(153-219,Arg-193), containing a Trp→Arg mutation disrupting the SH3a binding surface (12Gu Y. Xu Y.C. Wu R.F. Souza R.F. Nwariaku F.E. Terada L.S. Exp. Cell Res. 2002; 272: 62-74Crossref PubMed Scopus (68) Google Scholar, 15de Mendez I. Homayounpour N. Leto T.L. Mol. Cell Biol. 1997; 17: 2177-2185Crossref PubMed Scopus (84) Google Scholar) also bound WAVE1, suggesting a nonconventional mechanism of interaction perhaps unrelated to the SH3a surface.p47phox in turn appeared to bind to the mid-region of WAVE1 (Fig. 1c). Again, the binding mechanism was not easily deduced. Neither residues 1-269 nor 270 -599, which together span the entire protein, bound p47-(1-298). Similarly, neither 1-362 nor 363-559 bound p47-(1-298). An intact proline-rich region (residues 278-442) appeared to be necessary for binding as truncations not containing the full proline-rich region did not bind p47. GST pull down experiments also confirmed the ability of WAVE1-(278-442) and WAVE1-(100-442) to bind p47-(1-298) (Fig. 1d). Curiously, WAVE1-(270-559) bound p47-(1-298) in vitro but not in the yeast two-hybrid system, possibly due to interference with p47 binding by the N-terminal fused Gal4-AD in the WAVE1 truncation of the latter system.VEGF Stimulates Interaction between WAVE1 and p47phox within Membrane Ruffles—As with its myeloid paralog WASP, WAVE1 functionally catalyzes Arp2/3-dependent actin nucleation and polymerization (16Westphal R.S. Soderling S.H. Alto N.M. Langeberg L.K. Scott J.D. EMBO J. 2000; 19: 4589-4600Crossref PubMed Scopus (179) Google Scholar) and is thus found concentrated within membrane ruffles (17Miki H. Suetsugu S. Takenawa T. EMBO J. 1998; 17: 6932-6941Crossref PubMed Scopus (565) Google Scholar). In live HUVEC, we found that a DsRed-p47 fusion consistently colocalized with actin-GFP within membrane ruffles following VEGF stimulation, whereas DsRed-p47 was diffusely localized within the cytosol in unstimulated cells (Fig. 2). Specific localization to ruffles was demonstrated by fluorescence ratio imaging of cells cotransfected with DsRed-p47 and free EGFP (Fig. 3, a-c). In addition, the reversal of chromophores using p47-GFP and free DsRed also demonstrated relative concentration of p47phox in membrane ruffles (Fig. 3, d-f). Thus, localization of p47phox to ruffles did not appear to be a function of DsRed itself or of the increased cytosolic volume within ruffles. Interestingly, DsRed-p47 translocated to the nucleus in 15-30% of cells following VEGF stimulation (Fig. 2, d and f). The significance of this finding is unclear at this point but may reflect cotransport of p47phox with other mitogenic or survival factors to the nucleus.Fig. 2Localization of DsRed-p47 to membrane ruffles. HUVEC were cotransfected with DsRed-p47 and actin-GFP and were left unstimulated (a-c) or stimulated with VEGF (30 ng/ml, panels d-f) for 15-20 min, and fusion proteins observed in situ in living cells. Ruffles appeared as distinctive rapidly moving, sinuous, linear, and dorsally projected actin-GFP-rich structures. In all coexpressing cells, DsRed-p47 colocalized avidly with ruffles (arrow, panel f). Some cells displayed prominent translocation of DsRed-p47 into the nucleus (d). Red (a and d) and green (b and e) channels were imaged sequentially.View Large Image Figure ViewerDownload (PPT)Fig. 3Ratio imaging of DsRed-p47 in ruffles. HUVEC were cotransfected with DsRed-p47 and GFP (a-c) or with DsRed and p47-GFP (d-f) and then stimulated with VEGF (30 ng/ml). Images were taken in live cells. Red/green (c) and green/red (f) ratio images demonstrate relative concentration of DsRed-p47 and p47-GFP, respectively, within VEGF-stimulated ruffles. In contrast, perinuclear accumulation of p47-chromophore fusions appears to be secondary to increased volume of cells imaged in planar sections.View Large Image Figure ViewerDownload (PPT)In addition, endogenous WAVE1 coprecipitated with FLAG-p47 and recovery of WAVE1 was maximal at 15 min after stimulation with VEGF (Fig. 4a). This time course coincided with the onset of ruffling activity, which peaked at 10-15 min following VEGF addition, and subsided by 30 min. Furthermore, VEGF also increased coprecipitation of endogenous WAVE1 with endogenous p47phox (Fig. 4b). As expected, spatial colocalization within ruffles of VEGF-stimulated cells was demonstrated using WAVE-GFP and DsRed-p47 (Fig. 5). Colocalization within ruffles was consistent and reproducible with all of the ruffles visualized containing both fusion proteins. Finally, ectopic expression of either the WAVE1 binding domain of p47phox (p47-(153-286)) or the p47 binding domain of WAVE1 (WAVE1-(100-442)) decreased VEGF-induced ruffling (Fig. 6), suggesting that the interaction between WAVE1 and p47phox facilitates such ruffling.Fig. 4Coprecipitation of WAVE1 and FLAG-p47. a, HUVEC transfected with FLAG-p47 were stimulated with VEGF (30 ng/ml) for the indicated times, immunoprecipitated with anti-FLAG or irrelevant isotype control (Irrel), and immunoblotted (IB) for endogenous WAVE1 (upper panel). Coprecipitation of WAVE1 with FLAG-p47 was maximal at 15 min. Blot was reprobed with anti-p47phox (lower panel). b, untransfected HUVEC were stimulated with VEGF for 15 min and immunoprecipitated with either irrelevant antibody or mouse anti-p47phox and probed with antisera for WAVE1 and then p47phox. Coprecipitation of endogenous proteins was increased by VEGF.View Large Image Figure ViewerDownload (PPT)Fig. 5Colocalization of WAVE1-GEP and DsRed-p47. HUVEC coexpressing WAVE1-GFP and DsRed-p47 were examined in situ. a-c, unstimulated. Two separate cells (d-f and g-i) are shown following stimulation with VEGF (30 ng/ml), demonstrating colocalization of WAVE1-GFP and DsRed-p47 within membrane ruffles (arrows in f and i). The red channel (b, e, and h) was recorded 15-20 s prior to the green channel (a, d, and g) to allow separate filters, thus avoiding bleed-through. Consequently, ruffle movement is seen as a slight displacement of green on red within ruffles.View Large Image Figure ViewerDownload (PPT)Fig. 6Suppression of ruffles by p47-(153-286) and WAVE1-(100-442). HUVEC were cotransfected with actin-GFP and empty vector (a, b, e, and f), pCIN-p47-(153-286) (c and d), or pCIN-WAVE1-(100-442) (g and h). Following stimulation with VEGF (b, d, f, and h), ruffles were diminished in pCIN-p47-(153-286) or WAVE1-(100-442) expressing cells. i, number of ruffling cells was quantified using the actin-GFP transfectants. *, p < 0.001 compared with control (first column). **, p < 0.001 compared with VEGF + empty vector (second column).View Large Image Figure ViewerDownload (PPT)PAK1 Acts Upstream of VEGF-induced p47phox Phosphorylation, Oxidant Production, and Ruffle Formation—Activated Rac1 and its effector PAK1 are potent inducers of ruffles, and both proteins respond to growth factors and concentrate in linear ruffles (18Kraynov V.S. Chamberlain C. Bokoch G.M. Schwartz M.A. Slabaugh S. Hahn K.M. Science. 2000; 290: 333-337Crossref PubMed Scopus (560) Google Scholar, 19Dharmawardhane S. Schurmann A. Sells M.A. Chernoff J. Schmid S.L. Bokoch G.M. Mol. Biol. Cell. 2000; 11: 3341-3352Crossref PubMed Scopus (226) Google Scholar). Active Rac1 has also been shown to coprecipitate with WAVE1 (17Miki H. Suetsugu S. Takenawa T. EMBO J. 1998; 17: 6932-6941Crossref PubMed Scopus (565) Google Scholar), although the latter protein lacks a functional Cdc42/Rac1-interaction binding domain, suggesting an indirect interaction. We found that within 15 min of VEGF stimulation, endogenous forms of both PAK1 and Rac1 coprecipitated with FLAG-p47 (Fig. 7a). WAVE1 also appeared to be part of the PAK1-p47 complex (Fig. 7b), consistent with the formation of a ternary complex in which WAVE1 and associated protein(s) may act as a scaffold for the recruitment of Rac1, PAK1, and p47phox. Rac1 is an essential subunit of the NADPH oxidase, and PAK1 serves as an effective p47phox kinase in vitro (20Knaus U.G. Morris S. Dong H.J. Chernoff J. Bokoch G.M. Science. 1995; 269: 221-223Crossref PubMed Scopus (355) Google Scholar), although its capacity to perform this function in intact cells has not been demonstrated. The formation of such a complex might therefore be expected to facilitate p47phox phosphorylation and consequent oxidant production. Indeed, we found that VEGF led to PAK1 activation within 5 min and caused phosphorylation of p47phox within 15 min, whereas kinase-dead dominant-negative (DN) PAK1-(K298A) blocked p47phox phosphorylation (Fig. 8, a-c). This ability of PAK1 to facilitate p47phox phosphorylation, either directly or indirectly, is consistent with recent studies of the NADPH oxidase suggesting that Rac1 activates unknown proximal signaling events leading to oxidase assembly independent from its direct involvement with the oxidase (21Price M.O. Atkinson S.J. Knaus U.G. Dinauer M.C. J. Biol. Chem. 2002; 277: 19220-19228Abstract Full Text Full Text PDF PubMed Scopus (99) Google Scholar). In parallel, PAK1-(K298A) also blocked VEGF-induced oxidant production (Fig. 8d) and ruffle formation (Fig. 9). Furthermore, the membrane-permeant SOD mimetic MnTBAP also reduced VEGF-induced ruffling (Fig. 10), suggesting that the effects of endogenous p47phox occurred at least in part through oxidant production. Taken together, these data suggest a linear sequence of complex formation, p47phox phosphorylation, oxidant production, and membrane ruffl
Referência(s)