Roles of Necl-5/Poliovirus Receptor and Rho-associated Kinase (ROCK) in the Regulation of Transformation of Integrin αVβ3-based Focal Complexes into Focal Adhesions*

Focal complexes are continuously formed and transformed into focal adhesions during cell movement. We previously demonstrated that Necl-5 co-localizes with integrin αVβ3 at focal complexes, whereas Necl-5 does not localize at focal adhesions in moving NIH3T3 cells, suggesting that Necl-5 may be dissociated from integrin αVβ3 during the transformation of focal complexes into focal adhesions, but the underlying mechanism remains unknown. Here, we explore the roles of Necl-5 and Rho-associated kinase (ROCK) in the regulation of the transformation of focal complexes into focal adhesions. We found that inhibition of Necl-5 expression and expression of a constitutively active mutant of ROCK1 enhanced, whereas treatment with a ROCK inhibitor Y-27632 inhibited the transformation of focal complexes into focal adhesions. In HEK293 cells ectopically expressing Necl-5 and integrin αVβ3, treatment of cells with Y-27632 increased the binding of Necl-5 to clustered integrin αVβ3. The experiments using inhibitors of myosin ATPase and actin polymerization revealed that actomyosin-driven contractility exerts a similar function as ROCK. The phosphorylation of integrin β3 at Tyr747, which is known to be critical for the formation of focal adhesions, plays a pivotal role for the interaction between Necl-5 and integrin αVβ3. These results indicate that the transformation of focal complexes into focal adhesions is negatively and positively regulated by Necl-5 and ROCK, respectively, and that ROCK-dependent actomyosin-driven contractility is a critical determinant for the regulation of the interaction between Necl-5 and integrin αVβ3.

Cell movement has a pivotal role for physiological processes such as morphogenesis during development and the mobilization of immune cells to sites of infection as well as pathological events including metastasis of cancer cells. While moving in the direction of increasing or decreasing concentrations of external signaling molecules, cells spatiotemporally form special structures at leading edges, which include protrusions such as filopodia and lamellipodia, ruffles, focal complexes, and focal adhesions. Continuous formation and disassembly of these leading edge structures are necessary for facilitating cells to keep moving. Cells adhere to the extracellular matrix via integrins, cell-matrix adhesion molecules which regulate cell movement and exert their regulatory functions in cooperation with cell surface receptors for chemoattractants. Focal complexes are small dot-like integrin adhesion complexes formed at the edges of lamellipodia (1)(2)(3). After binding to matrix, integrins recruit cytoskeletal and cytoplasmic proteins, which leads to the cytoskeletal remodeling and the formation of more stable adhesive structures called focal adhesions. Focal adhesions contain proteins such as Src and focal adhesion kinase, integrin ␤ 3 , vinculin, and paxillin, and tyrosine phosphorylation of these proteins is implicated in the formation and turnover of focal adhesions (for review, see Refs. 4 and 5). Reorganization of the actin cytoskeleton, including actin polymerization and stress fiber formation, is required for cell spreading and locomotion and plays a central role in the formation of leading edge structures. During the migration process, the maturation of small focal complexes to larger focal adhesions at the leading edges requires the tension provided by actomyosin-driven contractility (for review, see Ref. 6).
Rho family small G proteins play a central role in the regulation of the formation of leading edge structures, cytoskeletal dynamics, and the maturation of focal complexes to focal adhesions (3,7,8). Rac1 triggers the formation of lamellipodia, ruffles, and focal complexes. Cdc42 is critical for the formation of filopodia and focal complexes. RhoA inhibits the formation of these leading edge structures but enhances the formation of focal adhesions and stress fibers. The transformation of focal complexes into focal adhesions depends on the activation of RhoA and its effector Rho-associated kinase (ROCK) 2 and is negatively regulated by Cdc42 and Rac1. It has been shown that ROCK is involved in the stress fiber formation (9) and regulates the contractility required for the cytoskeletal reorganization by two mechanisms, directly by the phosphorylation of myosin light chain (10) and indirectly by the inactivation of myosin phosphatase through the phosphorylation of the myosin binding subunit of myosin phosphatase (11). Although the ROCK-mediated actin-myosin-driven contractility is likely involved in the transformation of focal complexes into focal adhesions, the molecular mechanism by which ROCK regulates this process remains largely unknown.
We previously reported that Necl-5, an immunoglobulin-like adhesion molecule, also termed as poliovirus receptor/CD155/ Tage4, co-localizes with integrin ␣ V ␤ 3 at the leading edges in moving NIH3T3 cells and enhances cell movement (12). We recently showed that Necl-5 forms a complex with integrin ␣ v ␤ 3 , and overexpression of Necl-5 enhances its clustering. This results in the augmentation of the formation of focal complexes under peripheral ruffles over lamellipodia at the leading edges of NIH3T3 cells which are cultured on dishes pre-coated with vitronectin, known as an extracellular matrix for integrin ␣ v ␤ 3 (13), and directionally stimulated by PDGF. Over-expression of Necl-5 reduces the formation of focal adhesions, where integrin ␣ v ␤ 3 , but not Necl-5, is concentrated (14). These results indicate that Necl-5 enhances the formation of focal complexes and suggest that Necl-5 may negatively regulate the transformation of focal complexes into focal adhesions and that Necl-5 may be dissociated from integrin ␣ v ␤ 3 during this transformation. Knockdown of Necl-5 decreased the formation of focal complexes and focal adhesions (14). This result emphasizes that Necl-5 is required for the formation of focal complexes and suggests that focal complexes may be needed for their transformation into focal adhesions. Therefore, we explored here the roles of Necl-5 and ROCK in the regulation of the transformation of focal complexes into focal adhesions.
Cell Culture and Transfection-NIH3T3 cells were maintained in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% calf serum. HEK293 cells were maintained in DMEM supplemented with 10% fetal bovine serum. For transient expression experiments, cells were transfected with various expression vectors by the use of Lipofectamine 2000 reagent (Invitrogen) according to the manufacturer's protocol.
Down-regulation of Necl-5-A recombinant extracellular fragment of nectin-3 fused to the human IgG Fc (Nef-3) that efficiently interacts with Necl-5 was used to induce down-regulation of Necl-5 from the cell surface (15). Nef-3 was prepared as described previously (19) and cross-linked by the goat antihuman IgG Fc pAb (Jackson Immuno Research) before use. NIH3T3 cells were sparsely plated on dishes, cultured in the serum-containing medium in the presence or absence of Nef-3 for 18 h, washed, re-plated on vitronectin-coated glass coverslips, and then cultured in the serum-containing medium in the absence of Nef-3 for 6 h. In another series of experiments, NIH3T3 cells were sparsely plated on vitronectin-coated glass coverslips, cultured in the serum-containing medium for 18 h, and then cultured in the serum-containing medium in the presence or absence of Nef-3 for 1 h.
Cell Surface Biotinylation-To analyze the cell surface expression of Necl-5 and integrin ␣ V ␤ 3 , cells were incubated with 0.2 mg/ml sulfosuccinimidyl 2-(biotinamido) ethyl-dithiopropionate (sulfo-NHS-SS-biotin) (Pierce) in PBS at 4°C for 30 min and washed by sulfo-NHS-SS-biotin blocking reagent (50 mM NH 4 Cl in PBS) to quench free sulfo-NHS-SS-biotin followed by several further washes in PBS. Cells were then frozen by liquid N 2 , scraped off, and lysed in 500 l of IP buffer (20 mM Tris-HCl, pH 7.5, 150 mM NaCl, 0.1% SDS, 1% Triton X-100, 1 mM EDTA, 10 g/ml leupeptin, 2 g/ml aprotinin, and 10 M 4-amidinophenylmethanesulfonyl fluoride hydrochloride). Cell extracts were rotated at 4°C for 30 min and then divided into a detergent-insoluble pellet and a detergent-soluble supernatant by centrifugation. The supernatant was incubated with streptavidin beads (Amersham Biosciences) at 4°C for 4 h to collect biotinylated proteins. These samples were eluted by boiling the beads in SDS sample buffer (60 mM Tris-HCl, pH 6.7, 3% SDS, 2% 2-mercaptoethanol, and 5% glycerol) for 5 min and subjected to SDS-PAGE followed by Western blotting.
Directional Stimulation with PDGF-To generate a concentration gradient of PDGF, a -Slide VI flow (uncoated; Ibidi) was used (14). In brief, the -Slide VI flow has six parallel channels which were coated with 5 g/ml vitronectin according to the manufacturer's protocol. Cells were plated at a density of 5 ϫ 10 3 cells per square centimeter, cultured for 16 h, and starved of serum with DMEM containing 0.5% bovine serum albumin for 1 h. The concentration gradient of PDGF was applied using DMEM containing 0.5% bovine serum albumin and 30 ng/ml PDGF for 30 min according to the manufacturer's protocol.
Time-lapse Fluorescence Microscopy-Time-lapse fluorescence microscopy was performed as follows. NIH3T3 cells transfected with pEGFP-N3-integrin ␤ 3 or co-transfected with pEGFP-N3-integrin ␤ 3 and pBS-H1-Necl-5 were sparsely plated on vitronectin-coated glass-bottomed dishes, cultured in the serum-containing medium for 18 h, and then cultured in DMEM containing 0.5% bovine serum albumin in the presence or absence of Y-27632 for 1 h. An aliquot of PDGF was applied to the area close to the periphery of cells. Fluorescence signals were visualized by a confocal laser scanning microscope (Digital Eclipse C1si-ready, Nikon), and images were recorded in intervals of 1 min.
Co-immunoprecipitation Assay-Co-immunoprecipitation assays were performed as described (14). In brief, HEK293 cells were transfected with various combinations of plasmids, plated on vitronectin-coated dishes, and cultured overnight. After a 1-h serum starvation or treatment with each reagent, cells were washed with ice-cold PBS and lysed with buffer A (20 mM Tris-HCl at pH 8.0, 150 mM NaCl, 1 mM MnCl 2 , 10% glycerol, 1% Nonidet P-40, 10 mM NaF, 1 mM Na 3 VO 4 , 10 g/ml leupeptin, 2 g/ml aprotinin, and 10 M 4-amidinophenylmethanesulfonyl fluoride hydrochloride). In some experiments, NIH3T3 cells, instead of HEK293 cells, were sparsely plated vitronectincoated dishes and cultured overnight. After a 1-h serum starvation or treatment with Y-27632, cells were washed with ice-cold PBS and treated with 2 mM 3,3Ј-dithiobis-sulfosuccinimidyl propionate (DTSSP, Pierce) in PBS for 2 h. After the addition of Tris-HCl at pH 7.4, cells were washed with ice-cold PBS and lysed with buffer A. The lysates were rotated for 30 min and subjected to centrifugation at 12,000 ϫ g for 20 min. In the experiments using HEK293 cells, the supernatant was preincubated with protein G-Sepharose 4 Fast Flow beads (Amersham Biosciences) at 4°C for 1 h and then incubated with protein G-Sepharose beads pre-bound to the anti-FLAG mAb at 4°C for 4 h. In the experiments using NIH3T3 cells, the supernatant was preincubated with protein A-Sepharose beads (Amersham Biosciences) at 4°C for 1 h and incubated with protein A-Sepharose beads pre-bound to the anti-integrin ␣ V pAb at 4°C for 4 h. After the beads were extensively washed with buffer A, bound proteins were eluted by boiling the beads in SDS sample buffer for 5 min and subjected to SDS-PAGE followed by Western blotting.

RESULTS
Inhibition by Necl-5 of the Transformation of Focal Complexes into Focal Adhesions-We first examined whether Necl-5 is needed for the formation of focal complexes as well as whether the formation of focal complexes is required for the formation of focal adhesions. For this purpose, we used Nef-3 (the extracellular fragment of nectin-3 fused to the Fc portion of IgG) that trans-interacts with Necl-5 and induces its downregulation from the cell surface (15). When NIH3T3 cells, which were sparsely plated on dishes and cultured in the serumcontaining medium for 18 h, were re-plated on vitronectincoated glass coverslips and then cultured in the medium containing serum for 6 h, most cells formed polarized structures with lamellipodia at the leading edges, consistent with our earlier observations (Figs. 1, Aa and 1Ba) (12,14). The immunofluorescence signal for phosphotyrosine, an excellent marker for focal complexes and focal adhesions (20), was observed as dot-like structures under the peripheral ruffles and at sites to the rear of the leading edges. The dot-like structures under the peripheral ruffles were smaller in size than those at sites to the rear of the leading edges. These smaller dot-like structures under the peripheral ruffles corresponded to focal complexes, whereas larger dot-like structures at sites to the rear of the leading edges corresponded to focal adhesions. The signals for integrin ␤ 3 and Necl-5 were concentrated and co-localized with the signal for phosphotyrosine at focal complexes. In addition, the signal for integrin ␤ 3 , but not that for Necl-5, was concentrated and co-localized with the signal for phosphotyrosine at focal adhesions. Essentially the same results were obtained when integrin ␣ v was stained instead of integrin ␤ 3 . 3 Indeed, unless otherwise specified, essentially identical results were obtained for both integrin ␣ v and integrin ␤ 3 in the experiments that follow. In NIH3T3 cells stably expressing Necl-5 (Necl-5-NIH3T3 cells), most of the signals for phosphotyrosine, integrin ␤ 3 , and Necl-5 were seen at focal complexes, and only faint signals were observed at focal adhesions (Fig. 1Ab), consistent with our earlier results (14). Here, we determined that the smaller dot-like structures observed under the peripheral ruffles, which were immunopositive for Necl-5, integrin ␣ v , integrin ␤ 3 , and phosphotyrosine, corresponded to focal complexes, whereas the relatively larger dot-like structures, which were immunopositive for integrin ␣ v , integrin ␤ 3 , and phosphotyrosine, but not for Necl-5, corresponded to focal adhesions. However, when wild-type NIH3T3 cells, which were sparsely plated on dishes and cultured in the serum-containing medium in the presence of Nef-3 for 18 h, were re-plated on vitronectincoated glass coverslips and then cultured in the medium containing serum, but not Nef-3, for 6 h, most cells did not form polarized structures with lamellipodia (Fig. 1Bb). The signals FIGURE 1. Inhibition of the transformation of focal complexes into focal adhesions by Necl-5. A, immunofluorescence images of wild-type NIH3T3 and Necl-5-NIH3T3 cells. Cells plated on vitronectin-coated glass coverslips were stained with the anti-phosphotyrosine mAb (PY), the anti-integrin ␤ 3 mAb, and the anti-Necl-5 mAb. a, wild-type NIH3T3 cells; b, Necl-5-NIH3T3 cells. Yellow lines, focal complexes; arrowheads, focal adhesions; insets, higher magnification images; scale bars, 10 m. B, immunofluorescence images of NIH3T3 cells. NIH3T3 cells were preincubated with or without Nef-3 and/or the anti-Necl-5 mAb, washed, and plated on vitronectin-coated glass coverslips. Cells were stained with the anti-Necl-5 mAb, the anti-integrin ␤ 3 mAb, and the anti-vinculin mAb. a, NIH3T3 cells; b, Nef-3-treated NIH3T3 cells; c, Nef-3-and anti-Necl-5 mAb-treated NIH3T3 cells. Arrowheads, leading edges; scale bars, 10 m. C, cell surface expression levels of Necl-5 and integrin ␣ V ␤ 3 . NIH3T3 cells were preincubated with or without Nef-3, washed, and plated on vitronectin-coated dishes. The amounts of cell surface Necl-5 and integrin ␣ V ␤ 3 were analyzed by the biotinylation method. Results shown are representative of three independent experiments. D, immunofluorescence images of NIH3T3 cells. NIH3T3 cells were plated on vitronectin-coated glass coverslips and incubated with or without Nef-3 and/or the anti-Necl-5 mAb in the presence of serum. Cells were stained with the anti-Necl-5 mAb and the anti-integrin ␤ 3 mAb. a, NIH3T3 cells; b, Nef-3-treated NIH3T3 cells; c, Nef-3-and anti-Necl-5 mAb-treated NIH3T3 cells. Arrowheads, leading edges; scale bars, 10 m. E, cell surface expression levels of Necl-5 and integrin ␣ V ␤ 3 . NIH3T3 cells were plated on vitronectin-coated dishes and incubated with or without Nef-3. The amounts of cell surface Necl-5 and integrin ␣ V ␤ 3 were analyzed by the biotinylation method. F, expression levels of Necl-5 and integrin ␣ V ␤ 3 . NIH3T3 cells were transfected with control non-silencing siRNA or Necl-5 siRNA, and expressions of Necl-5 and integrin ␣ V ␤ 3 were analyzed. Results shown are representative of three independent experiments. G, immunofluorescence images of NIH3T3 cells. NIH3T3 cells transfected with Necl-5 siRNA or co-transfected with Necl-5 siRNA and GFP-Necl-5 R were plated on vitronectin-coated glass coverslips. Cells were stained with the anti-Necl-5 mAb, the anti-integrin ␤ 3 mAb, and the anti-vinculin mAb. a, Necl-5 siRNA-transfected NIH3T3 cells; b, NIH3T3 cells co-transfected with Necl-5 siRNA and GFP-Necl-5 R . Note that Necl-5-knockdown cells showed reduced immunofluorescence signals of Necl-5 and integrin ␤ 3 (*). Arrowheads, leading edges; scale bars, 10 m.
for Necl-5 and integrin ␤ 3 were hardly detected at any peripheral regions. The signal for integrin ␤ 3 was markedly reduced, but the signal for vinculin, another marker for focal adhesions, was clearly observed, indicating that integrin ␤ 3 -positive focal adhesions were reduced, but vinculin-positive focal adhesions were definitely formed. When cells were co-incubated with the anti-Necl-5 mAb that inhibits the interaction of Nef-3 with Necl-5 (15), the effect of Nef-3 was absent; the signals for Necl-5 and integrin ␤ 3 were concentrated and co-localized at focal complexes, which was similar to those in cells untreated with Nef-3 (Fig. 1Bc). A biotinylation assay showed that Nef-3 decreased the amount of cell surface Necl-5 but did not affect the amount of cell surface integrin ␣ v or integrin ␤ 3 , indicating that integrin ␣ v and integrin ␤ 3 remained on the cell surface without being endocytosed (Fig. 1C). These results indicate that when the amount of cell surface Necl-5 is reduced by downregulation, the formation of both focal complexes and focal adhesions are markedly reduced.
We then examined whether Necl-5 inhibits the transformation of focal complexes into focal adhesions. When NIH3T3 cells were sparsely plated on vitronectin-coated glass coverslips and cultured in the presence of serum for 19 h, many cells formed polarized structures with protrusive lamellipodia at the leading edges (Fig. 1Da). However, when cells were sparsely plated on vitronectin-coated glass coverslips and cultured for 18 h and then incubated with Nef-3 for 1 h, they did not form polarized structures with lamellipodia. The signals for Necl-5 and integrin ␤ 3 were hardly observed at any peripheral regions (Fig. 1Db). The signal for integrin ␤ 3 was concentrated at focal adhesions and markedly increased in comparison with that in cells cultured without Nef-3. The signal for Necl-5 was not observed at focal adhesions. Furthermore, when cells were coincubated with the anti-Necl-5 mAb, the effect of Nef-3 was cancelled; the signals for Necl-5 and integrin ␤ 3 were concentrated and co-localized at focal complexes (Fig. 1Dc). The amounts of integrin ␣ v and integrin ␤ 3 remained unchanged by Nef-3 despite the reduction of the amount of Necl-5 (Fig. 1E). These results indicate that when Necl-5 is downregulated by Nef-3 treatment after both focal complexes and focal adhesions are formed, the Nef-3-induced down-regulation of Necl-5 disassembles the clustering of integrin ␣ v ␤ 3 at focal complexes. This leads to diminish focal complexes but enhances focal adhesion formation. We applied another approach, the RNA interference method, to confirm the results of the experiment using Nef-3. We successfully knocked down Necl-5 by the use of Necl-5 siRNA without silencing integrin ␣ v ␤ 3 expression (Fig. 1F). Necl-5-knockdown cells did not form polarized structures with lamellipodia, and the signal for integrin ␤ 3 was reduced (see Fig. 1Bb); however, this phenotype was reversed by an expression of a GFP-tagged siRNA-resistant mutant of Necl-5 (GFP-Necl-5 R ) (Fig. 1G, a and b). Taken together, these results indicate that Necl-5 inhibits the transformation of focal complexes into focal adhesions and suggest that Necl-5 is dissociated from integrin ␣ v ␤ 3 during this transformation.
Inhibition of the Transformation of Focal Complexes into Focal Adhesions by a ROCK Inhibitor-To examine whether ROCK activation is necessary for the transformation of focal complexes into focal adhesions during directional cell movement, cells were sparsely plated on -slide VI flow dishes precoated with vitronectin, starved of serum, and directionally stimulated by PDGF in the presence or absence of a ROCK inhibitor Y-27632. In the absence of Y-27632, most cells became polarized and formed lamellipodia with peripheral ruffles at the leading edges toward higher concentrations of PDGF. The immunofluorescence signals for Necl-5 and integrin ␤ 3 were concentrated and co-localized at peripheral ruffles of the leading edges in the middle section of the cells (results not shown), consistent with our earlier observations (12,14). In the basal section of the cells, the signals for Necl-5 and integrin ␤ 3 were observed at focal complexes under peripheral ruffles ( Fig.  2A). In the presence of Y-27632, however, the signal for integrin ␤ 3 was enhanced and accumulated at focal complexes with a linear staining pattern (Fig. 2B). The signal for integrin ␤ 3 at focal adhesions became weaker than that in the absence of Y-27632 (see Fig. 2A). These results indicate that ROCK activation is necessary for the transformation of focal complexes into focal adhesions.  with ROCK1-CA decreased the focal complex formation and conversely increased the focal adhesion formation (Fig. 3A). Identical results were obtained when Necl-5-NIH3T3 cells, where the formation of focal complexes were enhanced (14), were used instead of wild-type NIH3T3 cells (Fig. 3B). These results indicate that ROCK activation enhances the transformation of focal complexes into focal adhesions.

Effects of a ROCK Inhibitor and Knockdown of Necl-5 on Integrin
␣ v ␤ 3 Dynamics-To investigate the roles of ROCK and Necl-5 in integrin ␣ v ␤ 3 dynamics in living cells, we performed time-lapse fluorescence microscopy in NIH3T3 cells transfected with EGFP-tagged integrin ␤ 3 (integrin ␤ 3 -EGFP), which was reported to behave like endogenous integrin ␣ v ␤ 3 in association with endogenous integrin ␣ v (7). When cells were sparsely plated on vitronectin-coated glass-bottomed dishes, clusters of integrin ␤ 3 were observed. When cells moved forward in response to PDGF, small clusters of integrin ␤ 3 appeared at the leading edge (Fig. 4A). These clusters remained stationary but maturated into larger clusters when cells moved, indicating that focal complexes were transformed into focal adhesions. In contrast, in the presence of Y-27632, small clusters continuously appeared at peripheral areas but immediately disappeared, indicating that focal complexes failed to mature into focal adhesions (Fig. 4B). In Necl-5knockdown-NIH3T3 cells, the formation of integrin clusters was hardly observed (Fig. 4C). This is consistent with the result of the experiment using Nef-3 (see Fig.  1Bb). Taken together, these results indicate that ROCK regulates the transformation of focal complexes into focal adhesions and that Necl-5 is needed for the formation of both focal complexes and focal adhesions.
Involvement of ROCK in the Dissociation of Necl-5 from Clustered Integrin ␣ v ␤ 3 -We then examined whether ROCK regulates the dissociation of Necl-5 from clustered integrin ␣ v ␤ 3 that binds to vitronectin. When NIH3T3 cells were cultured on vitronectin-coated dishes and endogenous integrin ␣ v was immunoprecipitated by an anti-integrin ␣ v pAb from cell lysates, the amount of co-immunoprecipitated Necl-5 was increased in the presence of Y-27632 compared with that in the absence of the ROCK inhibitor (Fig. 5A). When HEK293 cells ectopically expressing integrin ␣ v , integrin ␤ 3 -EGFP, and FLAG-tagged Necl-5 (FLAG-Necl-5) were cultured on vitronectin-coated dishes and FLAG-Necl-5 was immunoprecipitated by an anti-FLAG mAb from cell lysates, integrin ␣ v and integrin ␤ 3 -EGFP were co-immunoprecipitated (Fig. 5B). The amounts of co-immunoprecipitated integrin ␣ v and inte-grin ␤ 3 -EGFP were increased in the presence of Y-27632 compared with those in the absence of the ROCK inhibitor. These results indicate that ROCK activation may induce the dissociation of Necl-5 from clustered integrin ␣ v ␤ 3 .
To examine whether the cytoplasmic region of Necl-5 is necessary for the ROCK-induced dissociation of Necl-5 from the clustered integrin ␣ v ␤ 3 , we expressed integrin ␣ v , integrin ␤ 3 -EGFP, and FLAG-Necl-5 in which the cytoplasmic region was deleted (FLAG-Necl-5-⌬CP) in HEK293 cells and performed similar experiments to those described above. Both integrin ␣ v and integrin ␤ 3 -EGFP were co-immunoprecipitated with FLAG-Necl-5-⌬CP, and the amounts of co-immunoprecipitated integrin ␣ v and integrin ␤ 3 -EGFP were increased by treatment with Y-27632 (Fig. 5B). This result suggests that integrin ␣ v ␤ 3 , but not Necl-5, is likely a target for the ROCK-induced dissociation of Necl-5 from the clustered integrin ␣ v ␤ 3 .

Involvement of Myosin in the Dissociation of Necl-5 from Clustered
Integrin ␣ v ␤ 3 -Because ROCK induces phosphorylation of myosin light chain through the phosphorylation and inactivation of myosin phosphatase (11), we then examined whether myosin-driven contractility is involved downstream of ROCK in the dissociation of Necl-5 from clustered integrin ␣ v ␤ 3 that binds to vitronectin. We showed that the amounts of co-immunoprecipitated integrin ␣ v and integrin ␤ 3 -EGFP with FLAG-Necl-5 were increased in the presence of BDM, a myosin ATPase inhibitor (21) (Fig.  5C). In the presence of calyculin A that inhibits myosin phosphatase (22), the amounts of co-immunoprecipitated integrin ␣ v and integrin ␤ 3 -EGFP were decreased (Fig. 5D). Taken together, these results indicate that myosin-driven contractility plays a role downstream of ROCK in the dissociation of Necl-5 from clustered integrin ␣ v ␤ 3 .
Involvement of Actin Polymerization in the Dissociation of Necl-5 from Clustered Integrin ␣ v ␤ 3 -To determine the role of actin polymerization in the dissociation of Necl-5 from clus- Times after the PDGF application are shown in minutes. Note that an integrin ␤ 3 cluster (A, arrowhead) appeared at the leading edge (12Ј) and remained stationary but grew in size (15Ј-24Ј) during lamellipodium extension. In B, integrin ␤ 3 clusters continuously appeared at the leading edge (arrowhead) but immediately disappeared without moving to inward. In C integrin ␤ 3 clusters did not appear at the leading edge. Scale bars, 20 m. tered integrin ␣ v ␤ 3 that binds to vitronectin, we tested the effect of cytochalasin D that disrupts actin filament assembly. The amount of immunoprecipitated FLAG-Necl-5 was unchanged, but the amounts of co-immunoprecipitated integrin ␣ v and integrin ␤ 3 -EGFP with FLAG-Necl-5 were increased in the presence of cytochalasin D in comparison with those in the absence of this agent (Fig. 5E). This result indicates the functional importance of actin polymerization in the dissociation of Necl-5 from clustered integrin ␣ v ␤ 3 .
Role of the Phosphorylation of Integrin ␤ 3 at Tyr 747 in the Dissociation of Necl-5 from Clustered Integrin ␣ v ␤ 3 -To clarify the molecular mechanism of the Necl-5-induced inhibition of the transformation of focal complexes into focal adhesions, we compared the tyrosine phosphorylation of integrin ␤ 3 between wild-type and Necl-5-NIH3T3 cells. It is known that the phosphorylation of integrin ␤ 3 at Tyr 747 is critical for the integrinmediated cytoskeletal assembly and formation of focal adhesions (23). The phosphorylation level of integrin ␤ 3 at Tyr 747 was reduced in Necl-5-NIH3T3 cells as compared with that in wild-type NIH3T3 cells (Fig. 6A). This result suggests that the Necl-5-induced suppression of the phosphorylation of integrin ␤ 3 at Tyr 747 may be one of the mechanisms of the Necl-5-induced inhibition of the transformation of focal complexes into focal adhesions.
To test the possibility that the phosphorylation of integrin ␤ 3 at Tyr 747 is critical for the interaction with Necl-5, we compared the effects of wild-type integrin ␤ 3 and the Y747A mutant of integrin ␤ 3 in which Tyr 747 was replaced to Ala (23). When wild-type integrin ␤ 3 -EGFP or EGFP-tagged Y747A mutant of integrin ␤ 3 was co-expressed with integrin ␣ v and FLAG-Necl-5 in HEK293 cells and FLAG-Necl-5 was immunoprecipitated by the anti-FLAG mAb from cell lysates, the amounts of co-immunoprecipitated integrin ␣ v and integrin ␤ 3 were markedly increased in Y747A mutant-transfected cells (Fig. 6B). Notably, Y-27632 had no effect on the amounts of co-immunoprecipitated integrin ␣ v and integrin ␤ 3 in Y747A mutanttransfected cells. Collectively, these results indicate that the phosphorylation of integrin ␤ 3 at Tyr 747 is critical for the inter- Integrin ␣ V was immunoprecipitated (IP) by the anti-integrin ␣ V pAb from cell lysates, and immunoprecipitated samples were subjected to Western blotting (WB) using the anti-Necl-5 mAb and the anti-integrin ␣ V mAb. Pre-immune rabbit serum was used as a control. The results shown are representative of two independent experiments. B-E, enhancement by Y-27632, BDM, and cytochalasin D and inhibition by calyculin A of co-immunoprecipitation of integrin ␣ V ␤ 3 with Necl-5. HEK293 cells were transfected with various combinations of the indicated plasmids and cultured on vitronectin-coated dishes in the presence or absence of 10 M Y-27632, 20 mM BDM, 5 nM calyculin A, or 2 M cytochalasin D for 1 h. Dimethyl sulfoxide (DMSO) was used as a solvent (C-E). FLAG-tagged Necl-5 or Necl-5-⌬CP was immunoprecipitated by the anti-FLAG mAb from cell lysates, and immunoprecipitated samples were subjected to Western blotting using the anti-FLAG mAb, the anti-integrin ␣ V mAb, and the anti-GFP pAb. B, Y-27632; C, BDM; D, calyculin A; E, cytochalasin D. The results shown are representative of two independent experiments. FIGURE 6. Role of integrin ␤ 3 phosphorylation at Tyr 747 in the dissociation of Necl-5 from clustered integrin ␣ v ␤ 3 . A, inhibition of the phosphorylation of integrin ␤ 3 at Tyr 747 in Necl-5-NIH3T3 cells. Wild-type NIH3T3 and Necl-5-NIH3T3 cells were cultured on vitronectin-coated dishes and starved of serum for 1 h. Cell lysates were subjected to Western blotting (WB) using the anti-phospho (p)-integrin ␤ 3 (Tyr 747 ) pAb and the anti-integrin ␤ 3 pAb. B, effects of the Y747A mutant of integrin ␤ 3 and Y-27632 on co-immunoprecipitation of integrin ␣ V ␤ 3 with Necl-5. HEK293 cells were transfected with various combinations of the indicated plasmids and cultured on vitronectin-coated dishes in the presence or absence of 10 M Y-27632 for 1 h. FLAG-tagged Necl-5 was immunoprecipitated by the anti-FLAG mAb from cell lysates, and immunoprecipitated samples were subjected to Western blotting using the anti-FLAG mAb, the anti-integrin ␣ V mAb, and the anti-GFP pAb. The results shown are representative of two independent experiments. action of Necl-5 with integrin ␣ v ␤ 3 . In other words, Necl-5 prefers to bind unphosphorylated integrin ␤ 3 and prevents its phosphorylation. Once ROCK is activated, activated ROCK induces the dissociation of Necl-5 from clustered integrin ␣ v ␤ 3 .