Neural Wiskott-Aldrich syndrome protein is recruited to rafts and associates with endophilin A in response to epidermal growth factor.

Neural Wiskott-Aldrich syndrome protein (N-WASP) has been implicated in endocytosis; however, little is known about how it interacts functionally with the endocytic machinery. Sucrose gradient fractionation experiments and immunofluorescence studies with anti-N-WASP antibody revealed that N-WASP is recruited together with clathrin and dynamin, which play essential roles in clathrin-mediated endocytosis, to lipid rafts in an epidermal growth factor (EGF)-dependent manner. Endophilin A (EA) binds to dynamin and plays an essential role in the fission step of clathrin-mediated endocytosis. In the present study, we show that the Src homology 3 (SH3) domain of EA associates with the proline-rich domain of N-WASP and dynamin in vitro. Co-immunoprecipitation assays with anti-N-WASP antibody revealed that EGF induces association of N-WASP with EA. In addition, EA enhances N-WASP-induced actin-related protein 2/3 (Arp2/3) complex activation in vitro. Immunofluorescence studies revealed that actin accumulates at sites where N-WASP and EA are co-localized after EGF stimulation. Furthermore, studies of overexpression of the SH3 domain of EA indicate that EA may regulate EGF-induced recruitment of N-WASP to lipid rafts. These results suggest that, upon EGF stimulation, N-WASP interacts with EA through its proline-rich domain to induce the fission step of clathrin-mediated endocytosis.

Four principal steps in vesicular membrane traffic can be distinguished. First, the coated membrane is invaginated by the assembly of coat proteins and formation of the coated pits. This is followed by a pinching off of the coated vesicles, transport of the vesicles to the appropriate acceptor membrane, and fusion of the vesicles with this membrane. Numerous proteins and lipids are involved in and regulated properly in a molecular cascade (1). In mammalian cells, endocytosis of vesicles can occur via clathrin-coated pits, a clathrin-independent pathway, or caveolae. It is thought that endocytosis occurs at rafts, which are plasma membrane domains enriched in cholesterol and sphingolipids (2,3). Rafts also contain glycosylphosphatidylinositol-anchored proteins and various transmembrane proteins (4 -6).
Activation of receptors by ligands results in internalization of receptor complexes, and these complexes are rapidly recycled from endosomes back to the cell surface (7,8). Endocytosis of epidermal growth factor (EGF) 1 receptor served as a model system for studies of ligand-dependent receptor trafficking for many years (9). EGF-activated receptors are internalized via the clathrin-coated pit pathway (10) through interactions with the clathrin adaptor complex AP2 that recruits the clathrin triskelion (11). Various accessory proteins, such as Eps15 and Epsin, are recruited to the AP2-clathrin complex and subsequently form an endocytic clathrin coat (12). The clathrinadaptor coats undergo rearrangement, resulting in invagination of the coated membrane, i.e. the clathrin-coated pit (13). For constriction from shallow to deeply invaginated coated pits and fission, endophilin A (EA) and dynamin are recruited to the rim of the polymerizing clathrin coat (14). EA has lysophosphatidic acid acyltransferase (LPA-AT) activity that catalyzes transfer of fatty acids from co-enzyme A to LPA, thereby generating phosphatidic acid (PA) (15). This shift in the biophysical properties of phospholipids in the cytoplasmic leaflet of the membrane bilayer would cause inward distortion of the luminal leaflets and subsequent membrane fission (16). This finding was confirmed by a recent study (17) showing that microinjection of an antibody against EA into lamprey reticulospinal synapses interferes with synaptic vesicle recycling and clathrin-coated vesicle formation.
The 100-kDa GTPase dynamin acts as an essential factor in the fission stage of clathrin-mediated endocytosis (18). Dynamin assembles at the site of fission and garrotes the membrane in a process driven by GTP hydrolysis (19). It was reported that purified dynamin causes vesiculation of liposomes in vitro in a GTP-dependent fashion (20). The newly formed detached clathrin-coated vesicles were internalized and moved through the cytoplasm to early endosomes. Vesicular trafficking at the plasma membrane would require rearrangement of the cortical actin filaments to remove the barrier to vesicular fusion or budding events (21). Actin filaments and actin-based motor proteins play an essential role in endocytosis in yeast (22)(23)(24)(25). It was also reported that treat-ment of several actin-disrupting agents inhibit receptor-mediated endocytosis in mammalian cells (26,27). These data suggest that the actin cytoskeleton plays an essential role in endocytosis. The actin-related protein 2/3 (Arp2/3) complex enhances the nucleation and polymerization of actin filaments to promote filament assembly in vivo (28,29). Neural Wiskott-Aldrich syndrome protein (N-WASP) plays an essential role in Arp2/3-dependent actin dynamics by enhancing Arp2/3 complex-induced nucleation of actin filaments (30). N-WASP contains a WASP homology domain, a Cdc42 binding domain, a proline-rich domain, two G-actin-binding verprolin-homology domains, a cofilin-homology domain, and a carboxyl-terminal acidic segment (31). The verprolin-homology-cofilin-homologyacidic (VCA) domain of N-WASP is an essential minimal region for activation of Arp2/3 complex. At rest, N-WASP is thought to be an auto-inhibited through an intramolecular interaction between its Cdc42 interaction and COOH-terminal domains (32). Binding of GTP-bound Cdc42 and phosphatidylinositol 4,5-bisphosphate (PIP 2 ) to N-WASP causes a conformational change in N-WASP that allows the VCA domain to interact with the Arp2/3 complex and initiate actin polymerization (33).
The yeast homologue of WASPs, Las17, was implicated in endocytosis (34), and lymphocytes from WASP knockout mice exhibited reduced actin polymerization and defective T cell receptor endocytosis (35). These data suggest that N-WASP plays various roles at many steps of endocytosis. Here we show that EGF induces recruitment of N-WASP from the cytoplasm to rafts. We also show that N-WASP interacts with EA through its proline-rich domain, playing an essential role in the fission step of clathrin-mediated endocytosis.

EXPERIMENTAL PROCEDURES
Constructions-For expression in mammalian cells, several N-WASP constructs, inducing full-length N-WASP and proline-rich domain (amino acids 271-385)-deleted N-WASP (N-WASP⌬P), were constructed in pcDL-SR␣ and pEYFP (Clontech). EA3 cDNA (gift from Dr. T. Endo, University of Chiba, Japan) was amplified by PCR with primers that introduced 5Ј BamHI and 3Ј HindIII sites. PCR products were cloned into the BamHI-HindIII sites of pCMV-tag3B (Clontech) and into the BglII and HindIII sites of pEGFP (Clontech) to produce proteins tagged with Myc and green fluorescence protein (GFP). To obtain the glutathione S-transferase (GST) fusion protein of EA (GST-EA), full-length EA3 cDNA was subcloned into pCMV-tag3B, cut with BamHI and XhoI, and inserted into pGEX-2T (Amersham Biosciences). GST fusion proteins of EA3, GST-EA⌬SH3 (amino acids 1-254) and GST-SH3 (amino acids 278 -348), were produced by in-frame insertion of the PCR-amplified fragment corresponding to each sequence into the BamHI-EcoRI sites of pGEX-2T.
Antibodies-Anti-Myc antibody was purchased from Santa Cruz Biotechnology, and the anti-clathrin heavy chain antibody (Ab-1) was from Oncogene. The anti-dynamin antibody (mouse clone 41), the anti-caveolin 1, the anti-Rab5, and the anti-phosphotyrosine antibody (PY20) were from Transduction Laboratories. The anti-N-WASP antibody was prepared as described (36). The secondary antibodies linked to peroxidase were from Cappel. The secondary antibodies linked to fluorescein, Texas Red, and Cy5 were from Molecular Probes.
Fractionation of Cell Lysates by Sucrose Gradient Centrifugation-HeLa cells plated on 150-mm dishes were serum-starved for 24 h and then stimulated with 100 ng/ml EGF (Invitrogen). After two washes with ice-cold phosphate-buffered saline (PBS), HeLa cells were scraped into 800 l of 500 mM sodium carbonate (pH 11.0). The cell lysates were extruded through a 23-gauge needle 10 times and then sonicated for 5 min in a sonicator bath. One milliliter of the homogenate was then adjusted to 45% sucrose by the addition of 1 ml of 90% sucrose prepared in MES buffered saline (MBS) (25 mM MES, pH 6.5, 90% sucrose, 150 mM NaCl) and placed at the bottom of an ultracentrifuge tube (14 ϫ 89 mm, Beckman Instruments). A discontinuous sucrose gradient was formed (2 ml of 5% sucrose/3 ml of 25% sucrose/3 ml of 35% sucrose; all in MBS containing 250 mM sodium carbonate) and centrifuged at 100,000 ϫ g for 3 h in an SW41 rotor (Beckman Instruments). Proteins at the 5/25, 25/35, and 35/45% interfaces were collected and separated by SDS-PAGE (10% acrylamide) followed by Western blot analysis with the ECL detection system (Pierce).
Cell Culture and Immunofluorescence Microscopy-HeLa cells, A431 cells, and COS7 cells were cultured in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal bovine serum. For immunofluorescence microscopy, 1 ϫ 10 5 cells were plated on 2% gelatincoated coverslips in 35-mm dishes. Cells were serum-starved for 24 h and then stimulated with 100 ng/ml EGF and fixed with 3.7% formaldehyde in PBS. Fixed cells were permeabilized with 0.2% Triton X-100 in PBS for 5 min and then incubated with primary antibodies for 60 min. After washing, cells were incubated with secondary antibodies. To stain EGF receptors, biotin-conjugated anti-EGF receptor (EGFR1, Biogenesis) was used as primary antibody and ultra avidin-rhodamine (Leinca Technologies, Inc.) as secondary antibody. To visualize actin filaments, rhodamine-conjugated phalloidin (Molecular Probes) was also added during the incubation with secondary antibodies. After a 30-min incubation, coverslips were washed and mounted on glass slides. Cells were observed with a confocal laser scanning microscope (MRC 1024; Bio-Rad).
Methyl-␤-cyclodextrin Treatment-HeLa cells plated on coverslips were incubated with 10 mM methyl-␤-cyclodextrin (CDX, Sigma) in the serum-free DMEM, 50 mM HEPES (pH 7.6) at 37°C for 1 h. After incubation in serum-free DMEM without CDX for 15 min, they cells were stimulated with 100 ng/ml EGF and fixed.
Transfection-COS7 cells were transfected in Opti-MEM (Invitrogen) using 6 l of LipofectAMINE (Invitrogen) and 3 g of plasmid DNA per 35-mm dish according to the manufacturer's instructions. The DNA/ LipofectAMINE was maintained on the cells for 4 h, and the medium was then exchanged with maintenance medium. Twenty-four hours after transfection, cells were subjected to EGF internalization assay. For EGF stimulation, transfected cells were cultured for 2 h in maintenance medium and then for 24 h in serum-free DMEM. To obtain cell lysates, 20 g of recombinant plasmid was mixed with 10 7 cells, and the mixtures were subjected to electroporation with a Gene Pulser (Bio-Rad).
Immunoprecipitation-EGF-treated or transfected cells in 100-mm dishes were washed twice with ice-cold PBS and lysed in 200 l of TGH buffer (50 mM HEPES (pH 7.6), 50 mM NaCl, 5 mM EDTA, 1 mM orthovanadate, 10% glycerol, 1% Triton X-100, 1 mM phenylmethylsulfonyl fluoride). After addition of 800 l of IP buffer (50 mM HEPES (pH 7.6), 50 mM NaCl, 5 mM EDTA, 1 mM orthovanadate, 10% glycerol, 1 mM phenylmethylsulfonyl fluoride), lysates were extruded 10 times through a 23-gauge needle and centrifuged at 100,000 ϫ g for 5 min. Supernatants were mixed with 10 g of anti-N-WASP antibody or anti-Myc antibody (Santa Cruz Biotechnology) for 2 h. As a negative control, normal mouse IgG (Santa Cruz Biotechnology) or pre-immune rabbit serum was used. Then protein A-and G-agarose beads (Pierce) were added, and the mixtures were incubated for 1 h. Immunoprecipitates were washed three times with IP buffer and then analyzed by Western blotting.
Pull-down Assay-GST fusion proteins were expressed in Escherichia coli JM109 and purified from E. coli lysates with glutathione-Sepharose beads (Amersham Biosciences) according to standard methods. GST fusion proteins were eluted with 50 mM glutathione in 10 mM HEPES (pH 7.6). Glutathione in the samples was removed by dialysis with IP buffer. Protein concentrations were measured by Bradford assays with bovine serum albumin as a standard. Twenty micrograms of the GST fusion proteins were immobilized on glutathione-Sepharose beads and mixed with HeLa cell lysates or COS7 cell lysates. After the beads were washed with IP buffer, they were suspended in SDS sample buffer and subjected to SDS-PAGE and Western blot analysis.
Purification of Actin, Arp2/3 Complex, N-WASP, and GST Fusion Proteins-Actin was purified from rabbit skeletal muscle, and monomeric actin was isolated by gel filtration on Superdex 200 (Amersham Biosciences) in G buffer (2 mM Tris-HCl (pH 8.0), 0.2 mM ATP, 0.2 mM CaCl 2 , 0.5 mM DTT). Arp2/3 complex was purified from bovine brain extracts as described previously (37). N-WASP was prepared with a baculovirus system as described previously (30). A GST fusion protein containing the verplorin homology, cofilin homology, acidic region (VCA domains) and the GST fusion protein of the Ash/Grb2 NH 2 -terminal SH3 domain was used as a positive control (36). All GST fusion proteins used in the actin polymerization assay were eluted with 50 mM glutathione in 10 mM HEPES (pH 7.6). Actin Polymerization Assay-Actin polymerization was measured as the change in the fluorescence intensity of pyrene-labeled actin as described previously (36). To follow actin polymerization with purified components, pyrene-labeled G-actin or unlabeled G-actin was isolated by incubation with freshly thawed proteins in G buffer (5 mM Tris-HCl (pH 8.0), 0.2 mM CaCl 2 , 0.2 mM ATP, 0.2 mM DTT) for 12 h at 4°C followed by removal of residual F-actin by centrifugation at 40,000 ϫ g for 1 h. Polymerization reaction mixtures contained 60 nM Arp2/3 complex and various proteins and lipids in 95 l of the assay buffer (10 mM HEPES (pH 7.6), 100 mM KCl, 1 mM MgCl 2 , 0.1 mM EDTA, 1 mM DTT) and were preincubated for 5 min. The reaction was initiated by adding a 5-l mixture of 40 M unlabeled actin, 4 M labeled actin, and 4 mM ATP to the preincubated reaction mixtures. The change in fluorescence was measured at 407 nm with excitation at 365 nm in a fluorescence spectrometer (Jasco). All kinetic analyses were performed with the software provided by the manufacturer.

N-WASP Recycles between Rafts and the Cytoplasm in an
EGF-dependent Manner-To study the function of N-WASP, we examined localization of N-WASP during EGF-induced endocytosis. It is believed that endocytosis occurs at particular sites in the plasma membrane called rafts (2,3). EGF-induced clathrin-coated vesicles are sorted primarily to early endosomes. Both rafts and early endosomes are rich in cholesterol and sphingolipid and can be separated from the cytosol by discontinuous sucrose gradient fractionation (38). The components of rafts and early endosomes were fractionated from lysates of EGF-stimulated HeLa cells to determine the EGFdependent distribution of N-WASP. Raft components are enriched at the 5/25% interface of the gradient, whereas the early endosome marker early endosome antigen-1 (EEA1) is found at the 25/35% interface. The rest of the cell lysates, mainly cytosolic proteins, are localized at the 35/45% interface (38,39). In the present study, caveolins, marker proteins of rafts, were detected in all fractions including the 5/25% interface (Fig. 1A). This distribution did not change after EGF stimulation (data not shown). Rab5, a small GTPase associated with EEA1, was localized at the 25/35% interface (Fig. 1A). These data indicate that raft components were localized at the 5/25% interface, and those of early endosomes were at the 25/35% interface. Hereafter, the collections of proteins at the 5/25, 25/35, and 35/45% interfaces will be referred to as the rafts fraction, EEA1 fraction, and cytosol fraction, respectively.
Activated EGF-receptors were detected with an anti-phosphotyrosine antibody and used as a marker of internalized clathrin-coated vesicles. Phosphorylated receptors were detected in the rafts fraction at 1.5 min after EGF treatment and were then detected in the EEA1 fraction after 15 and 30 min (Fig. 1B). This change in distribution supports the idea that EGF receptors are internalized from rafts to early endosomes via a clathrin-dependent pathway. Clathrin, localized at early endosomes and cytosol in unstimulated cells, was recruited to rafts and then subsequently moved to early endosomes similar to EGF receptors. In contrast to clathrin, dynamin was not detected in early endosomes prior to EGF stimulation. Dynamin was also recruited to rafts after EGF stimulation, but this recruitment was later than that of clathrin. This time difference supports the hypothesis that dynamin is recruited to the edge of the neck of the clathrin-coated pit (14).
Like dynamin and many other proteins, N-WASP was detected in the cytosol fraction. After EGF treatment, N-WASP was recruited to rafts and then internalized into early endosomes (Fig. 1B). The results of sucrose fractionation indicate that after EGF treatment, N-WASP was recruited approximately at the same time as clathrin.
To visualize the recruitment of N-WASP induced by EGF, HeLa cells were stained with anti-N-WASP antibody (Fig. 1C) before and after EGF stimulation. In serum-starved cells, N-WASP was localized in a dot-like pattern around the perinu-

FIG. 1. N-WASP is recruited to lipid rafts in an EGF-dependent manner.
A and B, HeLa cell lysates treated with 100 ng/ml EGF were fractionated on a discontinuous sucrose gradient loaded with 5, 25, 35, and 45% sucrose. Fractions were collected and analyzed by SDS-PAGE and Western blotting with antibodies indicated. Proteins collected at the 5/25, 25/35, and 35/45% interfaces are referred to as the raft fraction, EEA1 fraction, and cytosol fraction, respectively. The indicated time is the period of incubation with EGF. C, immunofluorescence microscopy of N-WASP. Serum-starved HeLa cells were treated with 10 mM CDX at 37°C for 1 h. Untreated cells (left) and CDX-treated cells (right) were incubated with EGF (100 ng/ml) for the indicated times, fixed, and stained with anti-N-WASP antibody. White arrows indicate N-WASP recruited to the plasma membranes. D, immunofluorescence microscopy of N-WASP. Serum-starved A431 cells were treated with 100 ng/ml EGF for the indicated times, fixed, and stained with anti-EGF receptor (EGFR) and anti-N-WASP. The right column is the merged image of EGFR (red) and N-WASP (green).
White arrows indicate the site where N-WASP is co-localized with EGFR. clear region. At 1.5-5 min after EGF stimulation, N-WASP was located throughout the cytoplasm, and a portion was present at the plasma membrane (Fig. 1C, left). The translocation of N-WASP was well observed at the cell periphery (Fig. 1C), where membrane ruffles were induced upon EGF stimulation. After 15 and 30 min, N-WASP had returned to its resting localization pattern.
CDX is a useful tool to extract cholesterol from biological membranes with high preference over other lipid species (40,41). The major component of rafts is cholesterol, and therefore, CDX treatment should disrupt rafts in living cells. In the present study, CDX treatment inhibited uptake of the Texas Red-labeled EGF into HeLa cells (data not shown), suggesting that clathrin-mediated endocytosis is dependent upon rafts. In CDX-treated cells, N-WASP remained at the perinuclear region after EGF stimulation (Fig. 1C, right), and no translocation to the plasma membrane was observed. These data suggest that EGF induces the recruitment of N-WASP to rafts in the plasma membrane and confirm the sucrose fractionation data.
To examine whether N-WASP co-localizes with EGF receptor (EGFR) in rafts, EGF-treated A431 cells were double-stained with anti-EGFR and anti-N-WASP antibodies. At 1.5-5 min after EGF stimulation, EGFRs were accumulated to the plasma membranes together with N-WASP (Fig. 1D) and subsequently internalized with the clathrin-coated vesicles. The internalized EGFRs were co-localized clathrin and dynamin (data not shown). In CDX-treated cells, neither EGFR nor N-WASP showed recruitment to the plasma membranes (data not shown). These data suggest that both EGFR and N-WASP were recruited to raft fractions upon EGF stimulation.
N-WASP Is Associated with EA-Some components of the endocytic machinery, such as Pacsin and Intersectin (42,43), have been reported to bind to both dynamin and N-WASP. These proteins have SH3 domains and associate with dynamin and N-WASP through the proline-rich domains. EA also has an SH3 domain in its COOH terminus and binds to dynamin (44). EA has LPA-AT activity and plays an essential role in the fission step of clathrin-mediated endocytosis (45). We hypothesized that EA associates with N-WASP through its SH3 domain. To evaluate this hypothesis, we performed co-immunoprecipitation experiments.
Myc-tagged full-length EA expression plasmids (Myc-EA) and full-length N-WASP expression plasmids were co-transfected transiently into COS7 cells, and the EA/N-WASP interaction was detected by Western blot analysis after precipitation of cell lysates with anti-N-WASP antibody or anti-Myc antibody ( Fig. 2A). Both immunoprecipitates contained Myc-EA and N-WASP. Positive signals were specific, as preimmune serum and control IgG immunoprecipitates did not yield positive signals. These findings indicate that EA associates with N-WASP in vivo.
We then examined whether the SH3 domain of EA can associate with the proline-rich domain of N-WASP. As shown in Fig. 2B, we constructed a variety of GST fusion proteins of EA, including full-length (GST-EA), SH3 domain-deleted (GST-EA⌬SH3), and domain only (GST-SH3). These fusion proteins were mixed with COS7 cell lysates for pull-down assays. Western blotting revealed that the precipitates collected with GST-EA and GST-SH3 contained dynamin and N-WASP (Fig.  2C). When the same assay was performed with lysates of cells in which proline-rich domain-deleted N-WASP (N-WASP⌬P) was overexpressed, endogenous N-WASP but not N-WASP⌬P was detected in precipitates (Fig. 2D). These results show that the SH3 domain of EA interacts with the proline-rich domain.
EA Is Associated with N-WASP in an EGF-dependent Manner-As shown in Fig. 2, when both Myc-EA and N-WASP were expressed transiently, association of EA and N-WASP was revealed by co-immunoprecipitation. But surprisingly, when only Myc-EA was transfected into COS7 cells and anti-N-WASP antibody immunoprecipitates were blotted with anti- Myc antibody, however, no signal was detected (Fig. 3A). Because EA is a cytosolic enzyme, it may be recruited to rafts in an EGF-dependent manner together with N-WASP. Therefore, we tested the possibility that EGF might induce the interaction between N-WASP and EA. COS7 cells transfected with Myc-EA were stimulated with EGF, and immunoprecipitation with anti-N-WASP antibody or with anti-Myc antibody was performed. The time course of EA co-immunoprecipitation with N-WASP revealed that the association of EA with N-WASP occurred between 1.5 and 5 min after EGF treatment and disappeared by 15 min (Fig. 3A). The time course of this association corresponds to that of detection of N-WASP in the rafts fraction (Fig.  1B). Therefore, EGF might induce association of EA with N-WASP.
It has been reported that dynamin interacts with EA and acts as an essential factor in the fission stage of clathrin-mediated endocytosis. Despite these reports, dynamin was not found in the anti-Myc immunoprecipitates in Myc-EA-expressing cells under resting conditions (Fig. 3B). After EGF stimulation, dynamin co-precipitated with Myc-EA (Fig. 3B), suggesting this interaction is also EGF-dependent. Both immunoprecipitation studies with anti-Myc and anti-N-WASP indicated that Myc-EA, N-WASP, and dynamin formed a same complex upon EGF stimulation.
EA Stimulates N-WASP-induced Arp2/3 Complex Activation-Because EA associated specifically with N-WASP, we examined whether EA could further stimulate N-WASP-induced Arp2/3 complex activation in vitro. In the resting form, N-WASP exists in an auto-inhibited conformation that involves an intramolecular interaction between the Cdc42-interaction and the COOH-terminal domains (32). Binding of GTP-bound Cdc42 and PIP 2 to N-WASP causes a conformational change in N-WASP that exposes the VCA domain, which is the minimal essential region for activation of the Arp2/3 complex, and initiates actin polymerization. Actin polymerization can be monitored with pyrene-labeled actin, a fluorescent derivative of actin that yields higher fluorescence intensity when assembled into filaments. We used a cell-free system to measure the effect of N-WASP and EA on Arp2/3 complex-induced actin polymerization. The GST-VCA fusion protein of N-WASP maximally activated Arp2/3 complex-induced actin polymerization as described previously (37). In contrast, full-length N-WASP caused little acceleration of actin assembly by the Arp2/3 complex. The various GST fusion proteins of EA used in the present in vitro binding assays (Fig. 2) had no effect on actin polymerization (Fig. 4A). However, GST-EA and GST-SH3, which can bind N-WASP, enhanced N-WASP activation of Arp2/3 complexinduced actin polymerization, whereas GST-EA⌬SH3 did not (Fig. 4B). A similar phenomenon was observed when various histidine-tagged forms of EA were used in the place of the GST fusion proteins (data not shown). The effect of GST-SH3 on N-WASP activation to Arp2/3 complex was greater than that of GST-EA. The SH3 domain as well as full-length EA were far more effective than the SH3 domain of Grb2/Ash, one of known activators of N-WASP. GST-SH3 enhanced the ability of N-WASP in a dose-dependent manner to the level evoked by GST-VCA (Fig. 4D). Although the effect of GST-EA was also dose-dependent, it did not reach the maximal level like GST-SH3 (Fig. 4C).
PA Potentiates the Effect of EA on N-WASP-PIP 2 binds specifically to the basic domain of N-WASP and partially activates the Arp2/3 complex to induce actin polymerization (33). LPA and PA are the substrate and product of the LPA-AT activity of EA, respectively (15). Both are acidic lipids that bind electrostatically to the basic domain of N-WASP. Therefore, we speculated that LPA and PA should influence N-WASP activation of the Arp2/3 complex in the presence of EA. To examine this possibility, we added PA and LPA to N-WASP in the presence or absence of GST-EA, and we performed Arp2/3 complex-induced actin polymerization assays. In the absence of GST-EA, PIP 2 induced N-WASP activation of actin polymerization more effectively than LPA or PA (Fig. 4E). In the presence of GST-EA, PA-induced enhancement of N-WASP-induced actin polymerization was greater than that with PIP 2 (Fig. 4F). However, increasing concentrations of PA and GST-EA could not increase N-WASP activation of actin polymerization to the level evoked by GST-VCA (data not shown). In the presence of GST-SH3, PIP 2 enhanced N-WASP-induced actin polymerization more significantly than PA or LPA (data not shown). These data suggest that EA together with PA could produce greater activation of N-WASP to induce actin polymerization than PIP 2 , the major activator of N-WASP, and that the local productions of PA by the LPA-AT activity of EA cooperate in vivo inducing N-WASP-dependent actin polymerization.
To examine localization of EA, COS7 cells transfected with GFP-tagged EA (GFP-EA) were stimulated with EGF. In serum-starved cells, GFP-EA and N-WASP were localized in the cytoplasm. After EGF stimulation, GFP-EA was recruited to the plasma membrane and was co-localized with endogenous N-WASP (Fig. 5). Actin also accumulated at the plasma membrane where both GFP-EA and N-WASP were co-localized in an EGF-dependent manner. Fifteen minutes after EGF stimulation, GFP-EA and N-WASP returned to the cytoplasm. These findings suggest that EA may regulate the actin cytoskeleton by activating N-WASP with PA during EGF-induced endocytosis.
EA May Regulate EGF-induced Recruitment of N-WASP-As shown in Fig. 1, EGF induces a recruitment of N-WASP to rafts. To investigate the influence of clathrin-mediated endocytosis on the recruitment of N-WASP, COS7 cells were transfected with Myc-tagged ENTH domain-deleted Epsin (Myc-Epsin⌬ENTH) and GFP-tagged SH3 domain of EA (GFP-SH3), stimulated with EGF, and immunostained with anti-N-WASP antibody.
Epsin is a component of clathrin-coated pits that bind to both the clathrin heavy chain and AP2 (11). In cells overexpressing Epsin⌬ENTH, AP2 and clathrin were recruited to activated EGF receptors, but formation of clathrin-coated pits was inhibited (46). It was also reported that injection of the SH3 domain of EA inhibits association of endogenous EA with dynamin in an EGF-dependent manner, which then blocks fission of clathrin-coated pits (17,47). In the present study, uptake of Texas Red-labeled EGF was inhibited in cells overexpressing Myc-Epsin⌬ENTH or GFP-SH3 (data not shown).
We then examined whether N-WASP is localized normally when clathrin-mediated endocytosis is inhibited. Five minutes after EGF treatment, N-WASP was recruited to the plasma membranes of non-transfected cells but did not migrate in either Myc-Epsin⌬ENTH (Fig. 6A) or GFP-SH3 overexpressing cells (Fig. 6B). Quantification of both results confirmed that the effect was quite noticeable (Fig. 6C). In almost 50% of control cells (untransfected and GFP-transfected cells), N-WASP was recruited to the plasma membranes by EGF stimulation. In contrast, only 10% of cells transfected with GFP-SH3 or Myc-Epsin⌬ENTH showed the recruitment of N-WASP to the plasma membranes. These results indicate that both formation of clathrin-coated pits and fission are necessary for EGF-induced recycling of N-WASP.
Sucrose fractionation studies of lysates from N-WASP⌬Ptransfected COS7 cells revealed that endogenous N-WASP but not N-WASP⌬P was recruited to rafts in an EGF-dependent manner (Fig. 6D). To visualize the localization of N-WASP⌬P, HeLa cells were transfected with YFP-N-WASP⌬P and treated with 100 ng/ml EGF for 5 min. As is the case with endogenous protein (Fig. 1), YFP-N-WASP was recruited to the plasma membranes (Fig. 6E, upper panels), whereas few cells showed the translocation of YFP-N-WASP⌬P (Fig. 6E, lower panels). This also confirmed that recruitment of N-WASP to rafts is regulated through the proline-rich domain of N-WASP. DISCUSSION

N-WASP Is Recruited to Rafts after EGF Stimulation-Many
recent studies have focused on the connection between endocytosis and N-WASP; however, the role of N-WASP in endocytosis remains unclear. N-WASP activates the Arp2/3 complex and is essential for regulation of actin polymerization. Endosomes, pinosomes, and clathrin-coated and secretory vesicles are associated with actin comet tails in the cytoplasm as have endosomes and lysosomes in vitro (42,48). Actin comet formation is dependent on N-WASP (49,50); therefore, it is possible that N-WASP plays a role in vesicle transport. In addition, N-WASP interacts with proteins such as Pacsin, which binds to dynamin through the SH3 domain (51). These findings suggest that N-WASP could play various roles at different steps in endocytosis.
In the present study, we showed that N-WASP is recruited to rafts in an EGF-dependent manner (Fig. 1B). Because N-WASP was not detected in rafts in serum-starved cells, it is unlikely that N-WASP is involved in receptor-independent endocytosis. Upon EGF stimulation, N-WASP was quickly recruited to rafts and then internalized into early endosomes together with clathrin-coated vesicles containing activated EGF receptors.
Stimulation of glucose uptake by insulin in muscle and adipose tissue requires translocation of the glucose transporter protein GLUT4 from intracellular storage sites to lipid rafts (52). It was recently reported that activation of a small GTPbinding protein TC10 and rearrangement of the peripheral actin cytoskeleton are essential for insulin-stimulated glucose uptake and GLUT4 translocation (53-55). TC10 was found to interact with N-WASP, and it was suggested that recruitment of TC10 is regulated by N-WASP (56). These findings suggest that N-WASP may be recycled in an EGF-dependent manner and recruit some proteins to lipids rafts.
EA May Regulate EGF-induced Recruitment of N-WASP to Rafts-Immunofluorescence studies with anti-N-WASP antibody revealed that N-WASP, which is localized in a dot-like pattern at the perinuclear region under resting conditions, spreads throughout the cytoplasm and is recruited to the plasma membrane after EGF stimulation (Fig. 1C). In contrast, N-WASP remained in the perinuclear region in CDX-treated cells and did not move to the plasma membrane after EGF stimulation (Fig. 1C). These data suggest that EGF induces recruitment of N-WASP to rafts in the plasma membrane.
EGF induces activation of a tyrosine kinase leading to association of Ash/Grb2 and AP2 with the intracellular domain of EGF receptors. Clathrin-coated pits are then formed via AP2 bound to the activated receptors (11). At this step, it is possible that Epsin binds to AP2 through the DPW motifs of Epsin (46). In cells overexpressing ENTH domain-deleted Epsin (Epsin⌬ENTH), the activation of tyrosine kinase and recruitment of both AP2 and clathrin may occur normally, but clathrin-coated pit formation is inhibited (51). N-WASP was originally identified as Ash/Grb2-binding protein and was assumed to form a complex with the EGF receptor via Ash/Grb2 (31). However, EGF-induced recruitment of N-WASP was inhibited in cells overexpressing a deletion mutant of Epsin where Ash/Grb2 may be recruited normally to the EGF receptors. Therefore, EGF-induced recruitment of N-WASP to rafts appears to require clathrin-coated pit formation. Moreover, we have shown that overexpression of the EA SH3 domain inhibits EGF-induced recruitment of N-WASP and that a mutant of N-WASP lacking the proline-rich domain, which cannot associate with EA, does not move to rafts (Fig. 6). These data suggest that EGF-induced recruitment of N-WASP to rafts may be regulated by EA through the proline-rich domain of N-WASP.
N-WASP Associates with EA in an EGF-dependent Manner-Recently, it was reported that N-WASP associates with several proteins such as Pacsin and Intersectin-1, and those may be involved in the endocytic machinery (42,43). These proteins have SH3 domains that interact with the proline-rich domain of dynamin, and internalization of clathrin-coated vesicles is inhibited in cells overexpressing these SH3 domains (44). Overexpression of full-length Pacsin affected cortical actin organization, inducing formation of filopodia, suggesting that Pacsin activates N-WASP. However, overexpression of the SH3 domain alone had no effect on the actin cytoskeleton, indicating that Pacsin-induced cytoskeletal rearrangements are not due directly to Pacsin-N-WASP interactions (57). Intersectin-1 is a modular scaffolding protein that interacts with N-WASP through an SH3 domain and with Cdc42 through a Dbl homology domain (58). Recent studies (59) showed that Intersectin-1 binds to N-WASP and the activated Cdc42 through its action as a guanine nucleotide exchange factor.
In the present study, we show that N-WASP forms a complex together with EA and dynamin in an EGF-dependent manner (Fig. 3). EA and dynamin regulate the fission step of clathrin- White arrows indicate plasma membranes. N-WASP was recruited to the plasma membranes in non-transfected cells, but the localization did not change in transfected cells. C, quantification of the results by assessing the percentages of cells in which N-WASP was recruited to the plasma membranes by EGF stimulation. Untransfected cells: 58.0 Ϯ 13.2%, n ϭ 300; GFP: 47.9 Ϯ 9.5%, n ϭ 228; GFP-SH3: 9.3 Ϯ 0.5%, n ϭ 122; Epsin⌬ENTH: 11.8 Ϯ 1.7%, n ϭ 116. D, COS7 cells were transfected with plasmids encoding mutant N-WASP that is deleted for the proline-rich domain (N-WASP⌬P) and stimulated with 100 ng/ml EGF for 5 min. Lysates were separated into raft fraction, EEA1 fraction, and cytosol fraction as described in Fig.  1A. These fractions were analyzed by Western blotting with anti-N-WASP antibody. Note that endogenous N-WASP but not N-WASP⌬P was recruited to rafts in an EGF-dependent manner. E, HeLa cells were transfected with YFP-tagged N-WASP or YFP-tagged N-WASP⌬P. These cells were starved (left) and stimulated with EGF for 5 min (right) and then fixed. White arrows indicate YFP-N-WASP recruited to rafts in the plasma membrane. mediated endocytosis (45), and thus, N-WASP may have some role in clathrin-mediated endocytosis.
EA Enhances N-WASP Activation of Arp2/3 Complex-N-WASP is regulated by multiple activators. Individually most of these molecules, such as Cdc42 and PIP 2 , yield weak activation, and stimulation of N-WASP is accomplished by the proper combination of upstream signals. Thus, N-WASP integrates multiple signals to target actin polymerization precisely and regulate the actin cytoskeleton.
We confirmed that EA enhances N-WASP activation of the Arp2/3 complex in vitro (Fig. 4). Despite weak activation of polymerization by EA alone, EA in combination with PA activates N-WASP more strongly than PIP 2 (Fig. 4F). Thus, it is possible that EGF-induced N-WASP regulates the actin cytoskeleton at lipid rafts. Immunofluorescence studies of COS7 cells overexpressing EA revealed that actin accumulates at plasma membrane sites where EA and N-WASP co-localize in an EGF-dependent manner (Fig. 5).
In summary, we have shown that N-WASP is recruited to rafts via an EA-dependent pathway in response to EGF and associates with EA. Because EA activates N-WASP in the presence of PA, it seems reasonable to assume that N-WASP may have some role in EGF-induced endocytosis by binding to EA. Thus our results indicate that N-WASP interacts with EA to regulate the fission step of endocytosis.