Focal Adhesion Kinase Regulation of N-WASP Subcellular Localization and Function*

N-WASP is a member of the WASP family of proteins that regulate actin cytoskeleton remodeling. FAK is a cytoplasmic tyrosine kinase implicated in integrin signaling during cell migration. Here we identify a direct interaction between N-WASP and FAK and show that N-WASP is phosphorylated by FAK at a conserved tyrosine residue, Tyr256. We found that phosphorylation of Tyr256 affected N-WASP nuclear localization, suggesting that phosphorylation of N-WASP by FAK may regulate its activity in vivo by altering its subcellular localization. We also showed that the nuclear localization of N-WASP is dependent on its being in the open conformation either after its activation by Cdc42 or the truncation of the C-terminal VCA domain. Phosphorylation of Tyr256 of N-WASP could reduce its interaction with nuclear importin NPI-1, which might be responsible for its decreased nuclear localization. Lastly, we show that phosphorylation of Tyr256 plays an important role in promoting cell migration. Together, these results suggest a novel regulatory mechanism of N-WASP by tyrosine phosphorylation and subcellular localization and its potential role in the regulation of cell migration.

Cell migration plays an essential role in many biological processes such as embryogenesis, wound repair, angiogenesis, inflammatory immune response, and cancer metastasis (1,2). Cell migration is a complex multistep process that involves protrusion of the leading edge of the cell, formation of adhesion complexes, myosin/actin-mediated cell contraction, and the release of adhesions at the cell rear. Consistent with the complexity of the cell migration process, a variety of intracellular signaling molecules and their associated biochemical pathways have been identified in the regulation of this process. Members of the Rho subfamily of small G proteins (Rho, Rac, and Cdc42) modulate cell migration by affecting the dynamic reorganization of the actin cytoskeleton (3,4). While Rho stimulates formation of stress fibers necessary for actin-myosin mediated cell contraction (5), Rac and Cdc42 regulate lamellopodia and filopodia formation, respectively, during the initiation of cell migration (6). The driving force for the formation of these protrusive structures is actin polymerization regulated by the actin nucleation machinery such as the Arp2/3 complex (7). The Arp2/3 complex is activated by the Wiskott-Aldrich syndrome protein (WASP) 1 family of proteins through direct binding of their conserved verprolin-homology, cofilin-homology, acidic (VCA) domain to the Arp2/3 complex (8). The WASP family member N-WASP is an effector of Cdc42 mediated regulation of actin cytoskeleton and filopodia formation. Binding of Cdc42 to the Cdc42/Rac1 interactive binding (CRIB) motif within the G protein binding domain (GBD) of N-WASP releases it from an autoinhibitory conformation, which exposes its VCA domain allowing it to activate the Arp2/3 complex (8,9).
Recent studies also suggest that focal adhesion kinase (FAK), a cytoplasmic tyrosine kinase involved in signaling by integrins and other cell surface receptors, plays an important role in the regulation of cell migration (10,11). FAK Ϫ/Ϫ fibroblasts derived from FAK-knockout mouse embryos exhibit a significant decrease in cell migration compared with the cells from wild-type mice (12). In addition, inhibition of FAK by overexpression of FRNK (a dominant-negative FAK fragment) decreases cell spreading and migration (13,14). Conversely, overexpression of FAK in a number of cell lines including the FAK Ϫ/Ϫ cells promotes their migration on fibronectin (15)(16)(17). A number of downstream signaling pathways have been identified to mediate FAK-stimulated cell migration. One such pathway involves FAK complex formation with Src and subsequent phosphorylation of the adaptor molecule Cas by the FAK/Src complex (15,18). Another pathway involves interactions between FAK and PI3 kinase and the adaptor molecule Grb7 (19,20). Despite success in identifying these and other potential FAK downstream signaling pathways involved in the regulation of cell migration, the mechanisms by which FAK signaling connects to the cellular machinery in motility are still poorly understood.
To further explore potential mechanisms of FAK regulation of cell migration, we searched for interactions between FAK and proteins involved in the regulation of actin polymerization by co-immunoprecipitation and in vitro binding assays. In this study, we show a novel interaction between FAK and N-WASP and demonstrate FAK-dependent phosphorylation of N-WASP at Tyr 256 . Furthermore, we demonstrate that the subcellular localization of N-WASP is dependent on its activation state and that phosphorylation of Tyr 256 of N-WASP can regulate its subcellular localization by affecting N-WASP interaction with nuclear importin, NPI-1. Lastly, we provide evidence that phosphorylation of N-WASP and its consequent increased localization in the cytoplasm increased the ability of N-WASP to stimulate cell migration. Taken together, these data suggest that FAK may promote N-WASP function by regulating its subcel-lular localization, which may be involved in the regulation of cell migration by FAK.

MATERIALS AND METHODS
Antibodies-The mouse mAb 12CA5 (␣-HA) and rabbit antiserum ␣-KC against FAK have been described previously (21). Affinity-purified antibody against GST were prepared from anti-GST serum using GST immobilized on glutathione-Sepharose as an affinity matrix. The mouse monoclonal ␣-phosphotyrosine antibody (4G10) and the mouse monoclonal ␣-vinculin antibody were obtained from Upstate Biotechnology (Lake Placid, NY). The mouse monoclonal ␣-Cdc42 antibody was obtained from Transduction Laboratories (Lexington, KY). The rabbit antiserum against N-WASP was a generous gift of Dr. H. Miki (University of Tokyo). The rabbit polyclonal ␣-HA (Y11) antibody, the mouse monoclonal ␣-c-Myc-tag (9E10) antibody, rabbit polyclonal ␣-GFP, and rabbit polyclonal ␣-FAK (C20) antibody were obtained from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA). Antisera against KLF8 were described previously (22).
Cell Culture-The FAK Ϫ/Ϫ fibroblasts derived from the FAK-null mouse embryos were a generous gift of Dr. D. Ilic (UCSF) and were maintained in DMEM supplemented with 10% FBS as described previously (12). SYF cells (23) were generous gifts from Drs. L. Cary, R. Klinghoffer and P. Soriano and were maintained in DMEM with 10% FBS. 293 cells, COS7 cells, and CHO cells were cultured in DMEM supplemented with 10% FBS. NIH3T3 cells were maintained in DMEM supplemented with 10% CS. Tet-off CHO cell lines that overexpress HA tagged N-WASP or N-WASP Y256F mutant were generated using plasmids pTet-Splice-HA-N-WASP and pTet-Splice-HA-N-WASP-Y256F (see below) essentially as described (21). The cells were maintained in DMEM with 10% FBS, 0.5 mg/ml G418 and 1 g/ml tetracycline to suppress exogenous gene expression until experiments as indicated.
Immunoprecipitation and Western Blotting-For most experiments, subconfluent cells were washed twice with ice-cold PBS and then lysed with 1% Nonidet P-40 lysis buffer (20 mM Tris, pH 8.0, 137 mM NaCl, 1% Nonidet P40, 10% glycerol, 1 mM NaVO 4 , 1 mM phenylmethylsulfonyl fluoride, 10 mg/ml aprotinin, and 20 mg/ml leupeptin). For experiments to detect in vivo phosphorylation of N-WASP, 0.5 mM pervanadate was used in lysis buffer instead of sodium vanadate as phosphatase inhibitor. Pervanadate was made as described previously (30). Briefly, 20 mM sodium vanadate was mixed with 20 mM H 2 O 2 (1:1), then incubated at room temperature for 10 -15 min before adding to lysis buffer. Lysates were cleared by centrifugation for 20 min at 4°C, and total protein concentration was determined using Bio-Rad Protein Assay. GST fusion proteins were precipitated by incubating cell lysates with glutathione agarose beads for 120 min at 4°C. After washing five times with lysis buffer, complexes were resolved using SDS-PAGE. Immunoprecipitations were carried out by incubating cell lysates with appropriate antibodies as indicated for more than 2 h at 4°C, followed by incubation for another 2 h with protein A-Sepharose. After washing, immune complexes were resolved using SDS-PAGE. Western blotting was carried out using horseradish peroxidase-conjugated IgG as a secondary antibody and the Amersham Biosciences ECL system for detection.
In Vitro Kinase Assay-Recombinant His-tagged N-WASP was kind gift of Dr. H. Miki (University of Tokyo). His-tagged FAK were purified from Sf21 insect cell lysates that had been infected with recombinant baculovirus encoding the His-tagged FAK (kind gift of Dr. F. Matsumura of Rutgers University), as described previously (31). These two recombinant proteins were incubated in the kinase buffer (10 mM HEPES, pH 7.4, 3 mM MnCl 2 , 3 mM MgCl 2 , 1 mM Na 3 VO 4 ) with or without 100 M ATP for 20 min at room temperature. The kinase reactions were then stopped by the addition of SDS sample buffer. After boiling for 5 min, the samples were resolved on SDS-PAGE and analyzed by Western blotting using anti-phosphotyrosine antibody 4G10 to detect tyrosine-phosphorylated proteins.
Nuclear and Cytoplasmic Fractionation-Fractionation was performed essentially as described (32). Briefly, cells were washed with PBS, then lysed in a lysis buffer (20 mM HEPES, pH 7.2, 10 mM KCl, 2 mM MgCl 2 , 0.5% Nonidet P40, 1 mM Na 3 VO 4 , 1 mM phenylmethylsulfonyl fluoride, 0.15 units ml Ϫ1 aprotinin) and homogenized by a tightly fitting Dounce homogenizer. The homogenate was centrifuged at 1,500 ϫ g for 5 min to sediment nuclei. The supernatant was then resedimented at 15,000 ϫ g for 10 min, and the resulting supernatant formed the cytoplasmic fraction. The nuclear pellet was washed three times with lysis buffer and resuspended in the same lysis buffer supplemented with 0.5 M NaCl to extract nuclear proteins. The extracted material was sedimented at 15,000 ϫ g for 10 min, and the resulting supernatant was harvested as the nuclear fraction.
Preparation of GST Fusion Proteins and in Vitro Binding Assays-GST fusion proteins were produced and purified as described previously (19). GST fusion proteins (20 g) were immobilized on glutathioneagarose beads and then incubated for 90 min at 4°C with different amount of recombinant His-tagged FAK (0.6 -10 g) in 300 l of radioimmune precipitation assay buffer (50 mM HEPES, pH 7.4, 150 mM NaCl, 1.5 mM MgCl 2 , 10% glycerol, 1 mM EGTA, 1% Triton X-100, 0.1% SDS, 1% sodium deoxycholate, 10 g/ml aprotinin, 10 g/ml leupeptin, 1 mM phenylmethylsulfonyl fluoride). After washing, the bound proteins were analyzed by SDS-PAGE followed by Coomassie Blue staining. The amount of bound FAK was quantified by densitometry. In other experiments, recombinant Cdc42 loaded with GTP␥S (30 g) was used in the in vitro binding assays with FAK (1 g) and GST fusion proteins (20 g). The bound proteins were analyzed with Western blotting with polyclonal anti-FAK ␣-KC (1:1000) or anti-Cdc42 (1:500).
GST-importin protein pull-down experiments were performed similarly except that 293 cell lysates from cells that had been transfected with pEGFP-N-WASP or its mutants were used to bind the immobilized GST fusion proteins. Bound proteins were analyzed by SDS-PAGE followed with immunoblotting with anti-GFP antibody.
Fluorescent Microscopy-Cells were processed for immunofluorescence staining as described previously (21). The primary antibodies used were anti-N-WASP (1:50), and fluorescein isothiocyanate-conjugated goat-anti-rabbit antibody (1:300; Jackson ImmunoResearch Laboratory, West Grove, PA) was used as the secondary antibodies. In other experiments, live cells that had been transiently transfected with vectors encoding GFP fusion proteins were observed directly by fluorescent microscopy. In some experiments, the fluorescent intensity in the nuclear (Fn) and cytoplasmic (Fc) parts of the cells was quantified using the NIH Image J program. The ratio of Fn-Fc versus Fc was then calculated and used to quantify nuclear localization of the proteins.
Cell Migration Assays-Cell migration assays for tet-off CHO cells were performed as described previously (28). Briefly, cells were plated on 35-mm tissue culture dish in complete growth media. After cells reached confluency, wounds were generated by manual scraping of the cell monolayer with a pipette tip. The dishes were then washed, replenished with media, and photographed using a phase contrast microscope. They were then placed in the tissue culture incubator, and the matched wound regions were photographed again 6 or 8 h later. In other experiments, cell migration assays for transiently transfected FAK Ϫ/Ϫ cells were performed using image analysis as described previously (20), except that plasmids encoding GFP-N-WASP or GFP-Y256F mutant was used instead of a vector encoding GFP in some experiments.

Association of FAK with N-WASP-
To investigate a potential interaction between FAK and N-WASP, we first performed co-immunoprecipitation experiments of the two endogenous proteins. Cell lysates were prepared from NIH 3T3 cells and immunoprecipitated by antibodies against N-WASP, and the immune complexes were subjected to Western blotting with anti-FAK to detect associated FAK. Fig. 1A shows the presence of FAK in anti-N-WASP immunoprecipitates, but not in the control immunoprecipitates. Conversely, immunoprecipitation using anti-FAK, but not the control IgG, also brought down N-WASP in the immune complex (Fig. 1B). These results suggest a specific association of endogenous FAK with N-WASP.
To determine the N-WASP binding site on FAK, we coexpressed a Myc-tagged N-WASP with HA-tagged wild-type FAK or several FAK mutants in 293 cells. Immunoprecipitations were performed with an anti-HA antibody, followed by Western blotting with an anti-Myc antibody ( Fig. 2A). As expected, N-WASP was co-precipitated with the wild-type FAK, but not the empty pKH3 vector. It was also co-precipitated with the Y397F mutant, suggesting that tyrosine phosphorylation of FAK is not required for the association. The N-terminal domain of FAK (residues 1-400) showed strong binding to N-WASP, whereas the kinase domain or the C-terminal domain of FAK showed little or no binding. A similar strategy was used to define the FAK-binding site within N-WASP by co-expression of HA-tagged FAK with GST-tagged N-WASP fragments in 293 cells. Fig. 2C showed that FAK associated with the N-WASP fragment containing the Basic and GBD domains (residues 148 -273), but not the other fragments corresponding to the WH1, proline-rich, or VCA domains (see Fig. 2B). Together these results suggested that FAK interaction with N-WASP is mediated by the N-terminal domain of FAK and N-WASP residues 148 -273, respectively.
To determine whether the interaction between these two molecules is direct or not, we prepared a GST fusion protein containing the N-WASP residues 148 -273 (GST-N-WASP2) in bacteria and a His-tagged recombinant FAK protein from insect Sf21 cells. As shown in Fig. 2D, recombinant FAK bound to immobilized purified GST-N-WASP2 in a dose-dependent manner, but did not bind to GST alone control. These results suggested that FAK was capable of interacting with N-WASP directly.
Because Cdc42 also binds to the same region of N-WASP as FAK, we examined whether inclusion of purified Cdc42 could affect association of FAK and N-WASP using the in vitro binding assays. Recombinant FAK was incubated with immobilized GST-N-WASP2 in the presence or absence of His-tagged recombinant Cdc42 prepared from Sf9 insect cells and loaded with GTP␥S. The bound proteins were analyzed by SDS-PAGE followed by Western blotting analysis. Fig. 2E shows that, as expected, Cdc42 bound to GST-N-WASP (middle panel) and that it did not affect FAK binding to GST-N-WASP (top panel). Staining of the same membrane with Ponceau S showed similar expression levels of the fusion proteins in all samples (bottom panel). These results suggested that Cdc42 did not affect FAK association with N-WASP.
FAK Phosphorylation of N-WASP at Tyr 256 -A number of tyrosine kinases have been shown to associate with WASP family proteins and phosphorylate them (29,(33)(34)(35)(36). To determine whether N-WASP is a substrate for FAK, we first tested whether purified FAK can phosphorylate N-WASP in vitro. Fig.  3A shows that incubation of purified FAK with recombinant His-tagged N-WASP led to its phosphorylation in the presence of ATP (middle lanes), but not in the absence of ATP (left lanes). This phosphorylated band was not observed when FAK alone was incubated with ATP (right lanes), suggesting that this is not a FAK degradation product, which co-migrated with Histagged N-WASP. Co-expression of FAK, but not its kinasedefective mutant, with N-WASP also induced tyrosine phosphorylation of N-WASP in 293 cells, suggesting that N-WASP is also a substrate for FAK in vivo (Fig. 3B). The Y397F mutant also induced N-WASP phosphorylation, although to a lesser extent than the wild-type FAK. This could be due to a reduced kinase activity of the Y397F mutant because phosphorylated Tyr 397 is a binding site for Src family kinases that can subsequently further activate FAK (37). Alternatively, this could be due to a lack of additional phosphorylation of N-WASP by the associated Src family kinases. However, phosphorylation of N-WASP by FAK is also observed in SYF cells that are deficient in the expression of Src family kinases Src, Yes, and Fyn (23) (Fig. 3C), suggesting phosphorylation of N-WASP by FAK is independent of the Src family kinases.
We also investigated whether cell adhesion could induce phosphorylation of endogenous N-WASP, as it is well known that cell adhesion lead to FAK activation. Lysates were prepared from NIH 3T3 cells that had been suspended from plates and then replated on FN-coated dishes, and they were analyzed by for tyrosine phosphorylation of N-WASP. Fig. 3D shows a transient increase of N-WASP phosphorylation in cells plated on FN compared with that in the suspended cells. Quantitative analysis of three independent experiments showed that cell adhesion to FN-induced tyrosine phosphorylation of N-WASP in a time dependent manner, reaching a peak ϳ2.5-fold of the basal level (in suspended cells) at 2 h after cell adhesion (Fig.  3E). To verify that cell adhesion induced N-WASP phosphoryl-ation requires FAK, we examined N-WASP phosphorylation in FAK Ϫ/Ϫ cells. Fig. 3F shows that induction of tyrosine phosphorylation of N-WASP by cell adhesion is significantly decreased in FAK Ϫ/Ϫ cells when compared with that in the control FAK ϩ/ϩ or NIH 3T3 cells. Together, these results suggest that N-WASP is a substrate for FAK.
Because Tyr 291 , the conserved WASP residue corresponding to Tyr 256 of N-WASP, has been identified as the major phosphorylation site for a number of tyrosine kinases (33), we tested whether Tyr 256 of N-WASP is the major site of FAK phosphorylation. Fig. 4A shows that mutation of Tyr 256 to Phe (Y256F) significantly reduced N-WASP phosphorylation by FAK, suggesting that Tyr 256 of N-WASP is the major FAK phosphorylation site. Recent structural and biochemical studies indicated that Cdc42 binding to WASP can increase the accessibility of Tyr 291 for phosphorylation by the tyrosine kinases, Lyn and Btk (35,38). We therefore tested whether phosphorylation of Tyr 256 of N-WASP by FAK is dependent on its activation state. Fig. 4B shows that co-expression of the dominant-negative Cdc42 (Cdc42N17) inhibited N-WASP phosphorylation by FAK whereas the constitutively active Cdc42 (Cdc42L61) enhanced its phosphorylation. We also examined the effect of Cdc42L61 on FAK phosphorylation of the N-WASP mutant H208D, which has been shown to be defective in binding to and activation by Cdc42 (29). Fig. 4C shows that Cdc42L61 could not stimulate FAK phosphorylation of the H208D mutant whereas it enhanced FAK phosphorylation of the wild-type N-WASP. Together, these results suggest that only the activated N-WASP could be phosphorylated by FAK.
Differential Subcellular Localization of N-WASP and Its Mutations at Tyr 256 -To evaluate the functional significance of the interaction of FAK with N-WASP and FAK-dependent phosphorylation of Tyr 256 of N-WASP, we first examined whether FAK can affect N-WASP directly in its ability to induce actin polymerization. We found that neither overexpression of FAK nor co-transfection of FAK and N-WASP could induce filopodia formation, whereas transfection of Cdc42L61 readily induced it under the same conditions in quiescent fibroblasts (data not shown). We then examined the effect of FAK phosphorylation of N-WASP on its ability to stimulate Arp2/3 complex-mediated actin polymerization using pyrene actin assays as described previously (29). Fig. 5 shows that FAK phosphorylation of N-WASP (N-WASP ϩ FAK ϩ ATP, red filled circles) only caused a very small increase in N-WASP activity (compare with N-WASP alone, green triangle; or N-WASP ϩ FAK, red open circle). In contrast, phosphorylation by Fyn (N-WASP ϩ Fyn ϩ ATP, black filled circle) or addition of the VCA domain only (blue cross) showed a strong stimulation of actin polymerization as observed previously (29). The difference between FAK and Fyn is probably due to association of Fyn SH2SH3 domains (which is absent in FAK) with the phosphorylated N-WASP, which could further stimulate N-WASP activity (39). These results as well as the observation that FAK could only phosphorylate the activated N-WASP suggested that FAK interaction and phosphorylation of N-WASP is not sufficient to activate N-WASP directly and that it is more likely that FAK may regulate N-WASP function in a synergistic manner with Cdc42.
To explore such a potential role, we prepared two point mutants of N-WASP at Tyr 256 , Y256F that cannot be phosphorylated by FAK and Y256D that may mimic the negative charge created by phosphorylation of Tyr 256 . Interestingly, when GFP fusion proteins containing the wild-type, Y256F, or Y256D mutants of N-WASP were transfected into fibroblasts, we observed a striking difference in the subcellular localization of these constructs. The wild-type N-WASP and Y256F mutant localized at both the nucleus and the cytoplasm, whereas the Y256D mutant showed only cytoplasmic localization (Fig. 6,  A-C). We also observed the localization of endogenous N-WASP in both the nucleus and the cytoplasm by both immunofluores-cent staining (Fig. 6D) and subcellular fractionation (Fig. 6E), thus excluding the possibility that the GFP fusion proteins localization was an artifact of transfection or overexpression.
The Open Conformation of N-WASP Facilitates Its Nuclear Localization and Interaction with Importin ␣ NPI-1-Given its role in the regulation of actin polymerization, N-WASP localization in the nucleus was very surprising. Indeed, N-WASP has been observed in the nucleus in several previous reports (40 -42), but the regulation and significance of N-WASP localization in both the cytoplasm and nucleus is unknown. Close inspection of the N-WASP sequence revealed three nuclear localization sequences (NLSs) in residues 128 -144, 131-147, and 194 -197 within the basic domain. It is therefore possible that the NLSs are masked in the autoinhibitory conformation, and only become accessible to the nuclear translocation machinery when N-WASP is in its open conformation. To test this possibility, we created a truncated N-WASP mutant with the VCA domain deleted. Because the cofilin motif is both required and sufficient for maintaining N-WASP in the autoinhibitory conformation, the N-WASP-⌬VCA mutant is expected to be in an open conformation. The subcellular localization of this mutant and wild-type N-WASP was examined as GFP fusion proteins. We found that the N-WASP-⌬VCA mutant was almost exclusively localized in the nucleus whereas the wild-type N-WASP localized in both the nucleus and cytoplasm (Fig. 7, A  and B). These results provided support for our hypothesis that opening of N-WASP from the autoinhibitory conformation leads to exposure of the NLSs and allows its transport to the nucleus. Activation of N-WASP by Cdc42 is achieved by Cdc42 binding to the CRIB motif of N-WASP, which causes a conformational change in N-WASP from the autoinhibitory to the open conformation (8,38). To examine the physiological relevance of our above hypothesis, we examined the effects of Cdc42 on the subcellular localization of N-WASP. We found that co-expression of the constitutively active Cdc42L61 increased the nuclear localization of N-WASP whereas co-expression of the dominant-negative Cdc42N17 decreased it as examined by immunofluorescence (Fig. 8, A-C, compare with Fig. 7A). This was confirmed by using biochemical fractionation to examine whether the co-transfected N-WASP was present in the nuclear fractions. Fig. 8D shows that N-WASP was detected in the nuclear fraction. Co-transfection of Cdc42L61 increased it significantly whereas Cdc42N17 decreased it (top panels). Western blotting of parallel samples shows that the cytoplasmic protein vinculin was exclusively localized in the cytoplasmic fraction and the transcription factor KLF8 was present only in the nuclear fraction, indicating that there was little or no cross-contamination of the two fractions. Together, these results suggest that activation of N-WASP by Cdc42, which is associated with its conversion to an open conformation, exposes the NLSs and allows the activated N-WASP to be transported from the cytoplasm to the nucleus.
To investigate the effect of Tyr 256 phosphorylation by FAK on N-WASP nuclear translocation, we introduced the Y256F and Y256D mutants into the N-WASP-⌬VCA background to generate the N-WASP-⌬VCA-Y256F and N-WASP-⌬VCA-Y256D mutants, respectively. We then examined the subcellular distribution of these two mutants as GFP fusion proteins. Fig. 7, C and D show that like N-WASP-⌬VCA, the N-WASP-⌬VCA-Y256F mutant localized exclusively in the nucleus. Interestingly, the N-WASP-⌬VCA-Y256D mutant was present in both the nucleus and the cytoplasm. Together with the observations of the Y256F and Y256D mutation in the full-length N-WASP (see Fig. 6), these results suggest that phosphorylation of Tyr 256 could inhibit nuclear transport of the N-WASP in the open conformation, although it did not completely abolish it.
Protein translocation to the nucleus is mediated by the conserved importins ␣ and ␤, and the former recognizes the NLS in the cargo protein directly (43). In vitro binding assays showed a strong association between N-WASP with importin ␣ NPI-1 (Fig. 9) and weak interaction with the other major importin ␣ Rch1 (data not shown). The effects of the N-WASP mutations on its association with NPI-1 were further examined. As shown in Fig. 9, the deletion mutant N-WASP-⌬VCA, which relieves the autoinhibition, showed increased binding to NPI-1 (compare lanes 3 and 6). Mutation of the Tyr 256 to Asp inhibited the N-WASP ability to interact with NPI-1 in both the full-length and the N-WASP-⌬VCA background whereas Tyr 256 to Phe mutation did not have any effect (compare lane 1 with lanes 2 and  3, and lane 4 with lanes 5 and 6). None of the N-WASP constructs showed binding to GST control (panel B) although more GST protein was present in the samples than the GST-NPI-1 fusion protein (panels C and D). Western blotting using anti-GFP showed similar expression levels of N-WASP and its various mutants used in these experiments (panel E). Together, these results provided further support for our hypothesis that the open conformation of N-WASP is required for its nuclear translocation and that phosphorylation of Tyr 256 inhibits it.  Effects of Tyr 256 Phosphorylation of N-WASP on Its Ability to Stimulate Cell Migration-Remodeling of the actin cytoskeleton plays a critical role in cell motility, and N-WASP has been shown to be able to promote hepatocyte growth factor-induced cell migration (44). To investigate whether tyrosine phosphorylation of N-WASP is important in the regulation of cell migration, we created stable CHO cell lines with inducible expression of wild-type N-WASP or the N-WASP Y256F mutant, respectively. Fig. 10A shows induction of N-WASP and Y256F mutant to similar expression levels in the cell lines upon removal of tetracycline, as expected. The cells were then subjected to wound closure assays to assess the effects of N-WASP and the Y256F mutant on cell migration. Fig. 10B shows that induction of N-WASP expression accelerated the relative rate of cell migration by about 3-fold in comparison to the Mock cells under the same conditions. In contrast, cells with induced Y256F expression exhibited only a slightly increased rate of cell migration over that of Mock cells. These results suggested that phosphorylation of Tyr 256 plays a role in the regulation of cell migration by N-WASP, which is consistent with its effects on sustaining the cytoplasmic localization of N-WASP to mediate actin remodeling in the leading edge of migrating cells.
We next used FAK Ϫ/Ϫ cells to further evaluate the role of N-WASP phosphorylation by FAK in the regulation of cell migration. Fig. 10C shows that, as expected, transient transfection of FAK into the FAK Ϫ/Ϫ cells stimulated cell migration by about 2-fold. Interestingly, co-transfection of the Y256F mutant with FAK significantly decreased stimulation of cell migration by FAK whereas co-transfection of N-WASP with FAK slightly increased the migration, suggesting that FAK phosphorylation of N-WASP may be one of the important downstream pathways of FAK in the regulation of cell migration. Consistent with this, transiently transfection of the Y256D mutant also significantly stimulated cell migration in the FAK Ϫ/Ϫ cells (Fig. 10D). Co-transfection of FAK with the Y256D mutant further increased cell migration slightly, suggesting that additional FAK downstream pathways may be necessary to fully restore cell migration of the FAK Ϫ/Ϫ cells. Together, these results suggested that FAK phosphorylation of N-WASP plays a role in the regulation of cell migration.  A and B). Panel C shows quantitative results for multiple cells in panels A and B as described under "Materials and Methods." Alternatively, they were fractionated and the distribution of GFP-N-WASP in the nuclear and cytoplasmic fractions was analyzed by Western blotting (D), as described in Fig. 6.   FIG. 9. Regulation of N-WASP interaction with importin ␣ NPI-1 by its conformation and phosphorylation. 293 cells were transfected with pEGFP-N-WASP or its various mutants, as indicated. Two days after transfection, lysates were prepared from the cells and incubated with GST-NPI-1 (A and C) or GST (B and D) immobilized on glutathione-agarose beads. The samples were analyzed by Western blotting using anti-GFP to detect the bound proteins (A and B) or using anti-GST to verify similar amounts of the GST fusion proteins in the samples (C and D). Aliquots of lysates (WCL) were also analyzed directly by Western blotting using anti-GFP to confirm similar expression levels of GFP-N-WASP and mutants (E).

DISCUSSION
FAK is known to play a critical role in the regulation of cell migration stimulated by ECM proteins and growth factors (10,11,45). Although a number of downstream signaling pathways are implicated in the regulation of cell migration by FAK, the mechanisms by which FAK and its associated signaling pathways regulate cell migration are still not completely understood. N-WASP is a member of the WASP family of proteins that play key roles in the regulation of actin cytoskeletal remodeling in a variety of cellular processes including cell migration (44,46). In the present study, we identified a direct interaction between N-WASP and FAK and show that N-WASP is phosphorylated at Tyr 256 in a FAK-dependent manner. We also provide evidence that suggests a novel mechanism of N-WASP regulation through nuclear translocation and tyrosine phosphorylation. These results suggest an interesting mechanism for FAK regulation of cell migration by affecting subcellular localization of N-WASP, and thus its functions in actin cytoskeletal remodeling at the leading edge of motile cells.
The functional significance of N-WASP phosphorylation by FAK is not completely understood at present. A recent study indicated that phosphorylation of N-WASP by Fyn can increase its ability to induce actin polymerization in vitro (29), raising the possibility that N-WASP phosphorylation by FAK may also stimulate N-WASP activity directly. Although we cannot ex-clude this possibility completely, several lines of evidence from our studies make it less likely. First, in vitro actin polymerization assays suggested that FAK only caused a very weak increase in N-WASP activity, whereas Fyn showed a strong stimulation as observed previously (29). The observed differences in the effect of FAK and Fyn on N-WASP activity is consistent with a recent model proposed by Torres and Rosen (39). This model suggests that binding of SH2 and SH3 domains of Src family kinases to N-WASP (rather than phosphorylation of N-WASP per se) is largely responsible for the stimulation of N-WASP activity. Although it also binds to N-WASP, FAK does not have either SH2 or SH3 domain to mediate these interactions and it binds to a different region of N-WASP. This might account for the relatively weak stimulation of N-WASP activity by FAK in vitro. Second, transient transfection of FAK into quiescent fibroblasts did not induce filopodia formation whereas Cdc42L61 stimulated it under the same conditions (data not shown). Lastly, previous structural studies show that Tyr 256 lies in the hydrophobic core of the protein when it is in the autoinhibited conformation. This site may only be accessible to a tyrosine kinase after N-WASP is activated and in an open conformation. Consistent with this, we found that tyrosine phosphorylation of N-WASP by FAK is dependent on its activation state in vivo (see Fig. 4B). Thus FAK phosphorylation of N-WASP is unlikely to cause N-WASP activation, but rather is a consequence of initial N-WASP activation.
Data presented here suggest a novel N-WASP regulatory mechanism upon Tyr 256 phosphorylation by FAK through its influence on intracellular localization. We found that a fraction of N-WASP is present in the nucleus, which was observed in a number of previous reports although its significance was not investigated (40 -42). Here we show that the activation of N-WASP by Cdc42L61 increased its nuclear localization whereas inhibition of N-WASP by Cdc42N17 decreased it (Fig.  8). Furthermore, truncation of the VCA domain, which also leads to an open conformation for N-WASP, significantly increased its nuclear translocation (Fig. 7). These results are consistent with the presence of NLSs in the N-terminal Basic motif of N-WASP. These NLSs are masked in the autoinhibitory conformation of N-WASP, and would only become exposed after activation of N-WASP by Cdc42, which convert it to an open conformation. Our studies also demonstrated that Tyr 256 becomes more accessible in the open conformation and its phosphorylation by FAK serves to negatively regulate its interaction with the nuclear importin complex and inhibit its nuclear translocation (Figs. 6 and 9). Therefore, our results suggest that FAK may augment Cdc42 activation of N-WASP and influence its activity in vivo by regulating its subcellular localization as summarized in our current working model (Fig. 11).
Given that the major target of N-WASP, the Arp2/3 complex, and other actin remodeling components have not been described in the nucleus, it is not likely that the nuclear N-WASP will function to promote actin polymerization in the nucleus. On the other hand, our observation of the regulated nuclear localization of N-WASP raised the interesting possibility that the nuclear N-WASP may perform some as yet unknown function in the nucleus. While this intriguing possibility deserves further experimental examination, we would like to propose that one of the functions of the regulated N-WASP localization by FAK phosphorylation at Tyr 256 is to sustain N-WASP function in the cytoplasm to stimulate actin remodeling in the leading edge of migrating cells (Fig. 11).
It is well established that, in quiescent cells, N-WASP is inactive and exists in an autoinhibitory conformation due to an intracellular interaction between the VCA and GBD domains. Binding of activated Cdc42 (GTP-loaded) to the GBD domain of N-WASP converts it to an open and active conformation, which allows the VCA domain to interact with the Arp2/3 complex and induce actin polymerization (8,9,38). Data presented here showed that the open (and active) conformation will also expose the NLSs that mediates interactions with importin ␣ and allow nuclear translocation through the nuclear pore complex. This may serve as a novel negative feedback mechanism that functions to tune down N-WASP signaling after its activation. FAK and potentially other tyrosine kinases can phosphorylate the exposed Tyr 256 in N-WASP when it is in the open conformation. This phosphorylation inhibits N-WASP interaction with importin ␣, and thus reduces nuclear translocation of the active N-WASP. This will sustain N-WASP in its functional site in the cytoplasm.
A number of previous studies have suggested interactions between FAK signaling pathways and the Rho sub-family of small G proteins, both of which play critical roles in the regu- FIG. 11. A working model for nuclear translocation of N-WASP and its regulation by FAK. In quiescent cells, N-WASP is in an autoinhibitory conformation due to an intramolecular interaction between the VCA and the GBD domains, which masks both the NLSs in the basic region and Tyr 256 . Upon activation by Cdc42, N-WASP assumes an open and active conformation that allows its VCA domain to interact with the Arp2/3 complex and initiate actin polymerization (Step I). Data presented in this paper extend this model by suggesting the following negative feedback loop and its regulation. The open conformation exposes the NLSs that mediates interactions with importin ␣ and allows nuclear translocation through the nuclear pore complex (Step IIa). FAK and potentially other tyrosine kinases can phosphorylate the exposed Tyr 256 in N-WASP when it is in the open conformation. This phosphorylation inhibits N-WASP interaction with importin ␣, and thus reduces nuclear translocation of the active N-WASP (Step IIb). Steps IIa and IIb are mutually exclusive: an increase in Step IIb is proposed here to sustain N-WASP function in the cytoplasm to stimulate actin polymerization in the leading edge of cells whereas increased Step IIa decreases N-WASP activity in the cytoplasm and may promote as yet unknown functions of N-WASP in the nucleus. lation of cell migration. For example, FAK has been implicated in the down-regulation of Rho during cell adhesion and spreading (47,48). FAK has also been shown to bind Graf, a GAP for Rho and Cdc42 (49,50). More recently, FAK and the related kinase Pyk2 have been shown to associate with a novel protein PSGAP containing a rhoGAP domain (51). However, Pyk2, but not FAK, can activate Cdc42 by inactivating the GAP activity of PSGAP. On the other hand, FAK has been implicated in the regulation of guanine exchange factors for Rho (52). Our results provide another interesting link between FAK and Cdc42 at the level of the Cdc42 effector N-WASP. Interestingly, a recent report suggested an interaction between N-WASP and integrin ␤1 subunit in the breast cancer cells (53) and another recent study showed interaction of Arp2/3 complex with vinculin at the focal complexes in the peripheral of cells (54). Together, these results raised the interesting possibility of a direct connection between integrin signaling and regulation of actin remodeling through the N-WASP and Arp2/3 complex system.
Sequence alignments of the WASP family members indicated that both the NLS and the residue corresponding to Tyr 256 of N-WASP are conserved in the WASP family. Interestingly, another WASP family member, Wave, has also been showed to localize in the nucleus, although the function of this localization is unknown (36). Furthermore, Tyr 291 in WASP, which corresponds to Tyr 256 of N-WASP, has been shown to be phosphorylated by several other tyrosine kinases including Fyn, Btk, Lyn, and Hyc. Thus, it is possible that the negative feedback loop and its regulation by tyrosine phospohrylation as outlined in Fig. 11 may be a conserved feature for other members of the WASP family, which are used by various tyrosine kinases to regulate the actin cytoskeleton in different cells.