Involvement of Nectin in Inactivation of Integrin αvβ3 after the Establishment of Cell-Cell Adhesion*

Integrin plays an essential role in the formation of cell-matrix junctions and is also involved in the fundamental cellular functions. In the process of the formation of cell-cell junctions, an immunoglobulin-like cell-cell adhesion molecule nectin initially trans-interacts together and promotes the formation of adherens junctions (AJs) cooperatively with another cell-cell adhesion molecule cadherin. The activation of integrin αvβ3 is critically necessary for this nectin-induced formation of AJs. However, after the establishment of AJs, integrin αvβ3 becomes inactive and retains the association with nectin at AJs. The molecular mechanism of this dynamic regulation of integrin αvβ3 during the formation of AJs remains unclear. We found here that the expression of phosphatidylinositol-phosphate kinase type Iγ90 (PIPKIγ90), which is involved in the regulation of integrin activation, in Madin-Darby canine kidney cells, preferentially reversed the inactivation of integrin αvβ3 at cell-cell adhesion sites and partially disrupted E-cadherin-based AJs. The activation of PIPKIγ is correlated with its phosphorylation state. The tyrosine phosphatase protein-tyrosine phosphatase μ (PTPμ) effectively dephosphorylated PIPKIγ and thus canceled the PIPKIγ-dependent activation of integrin αvβ3 by blocking the interaction of integrin αvβ3 with talin. Moreover, PTPμ associated with nectin, and its phosphatase activity was enhanced by the trans-interaction of nectin, leading to the decrease in PIPKIγ90 phosphorylation. Therefore, the trans-interaction of nectin essentially functions in the inactivation of integrin at AJs through the PTPμ-induced inactivation of PIPKIγ.

Integrin is a key cell-cell adhesion molecule at cell-matrix junctions and comprises heterodimers with ␣ and ␤ subunits (1). Integrin exhibits intracellular conformational changes between the low and high affinity forms (2). The low affinity form shows weak adhesion activity for extracellular matrix proteins and is inactive, whereas the high affinity form has increased adhesion activity for its extracellular ligands and is active (3). It was reported that integrin is essential for the formation of specialized subcellular apparatuses, such as focal complexes and focal adhesions, and for cell movement, proliferation, and differentiation (1,4,5). We recently demonstrated that integrin ␣ v ␤ 3 interacts with Necl-5 at the leading edge of moving cells and that this complex enhances cell movement and proliferation together with platelet-derived growth factor receptor by stimulation of platelet-derived growth factor (6,7). Necl-5 is an Ig-like cell adhesion molecule and resembles nectin in its structure: three Ig-like loops at the extracellular region, a single transmembrane domain, and one cytoplasmic region.
When moving cells collide with each other, the initial cellcell contact occurs with the trans-interaction of Necl-5 with nectin-3 (8). Nectin is an emerging Ig-like cell-cell adhesion molecule that localizes at adherens junctions (AJs) 2 and is involved in the formation of AJs (9). Nectin exerts its cell-cell adhesion activity in a Ca 2ϩ -independent manner and consists of four members: nectin-1, nectin-2, nectin-3, and nectin-4 (9). However, the trans-interaction of Necl-5 with nectin-3 is tentative, and Necl-5 is down-regulated from the cell surface by clathrin-dependent endocytosis (10). The downregulation of Necl-5 impairs the integrin ␣ v ␤ 3 -and plateletderived growth factor receptor-dependent intracellular signaling for cell movement and proliferation, resulting in the reduction of cell movement and proliferation. The phenomenon that moving and proliferating normal cultured cells arrest both movement and proliferation after they grow confluent and form cell-cell junctions has been well known for a long time (11,12), but its molecular mechanism is poorly understood. The down-regulation of Necl-5 is likely to be at least partly one of the underlying mechanisms of contact inhibition of cell movement and proliferation. On the other hand, nectin-3 dissociated from Necl-5 is retained on the cell surface and subsequently trans-interacts with nectin-1, which most feasibly trans-interacts with nectin-3 among the nectin family members (8). This trans-interaction of nectins promotes the recruitment of cadherin, a major cell-cell adhesion molecule at AJs, to the nectin-based cell-cell adhesion sites, eventually establishing AJs (9,13). * This work was supported by grants-in-aid for Scientific Research and for During the nectin-induced formation of cadherin-based AJs, several intracellular signaling molecules including Rap1, Cdc42, and Rac small G proteins are activated, and actin cytoskeleton is reorganized by the trans-interaction of nectin in cooperation with the high affinity form of integrin ␣ v ␤ 3 (14 -17). In this process, the activation of protein kinase C and FAK, downstream molecules of integrin ␣ v ␤ 3 , is also required (17,18). However, after the establishment of AJs, the high affinity form of integrin ␣ v ␤ 3 is converted into the low affinity form that also continues to associate with nectin (17,18). Although the molecular mechanism by which integrin ␣ v ␤ 3 is inactivated after the formation of AJs remains to be elucidated, this inactivation seems to be beneficial for the maintenance of AJs, because the sustained activation of integrin renders cells highly motile, which tends to disrupt cell-cell junctions.
Integrin is activated by binding of talin to the cytoplasmic tail of integrin ␤ subunit (3), which causes the structural change of the integrin ␣/␤ dimer from the bent to the extended conformation. This change allows integrin to gain the higher affinity to the extracellular matrix. The binding of talin to integrin is up-regulated by increasing the amount of phosphatidylinositol 4,5-bisphosphate (19), which is generated by phosphatidylinositol-phosphate kinases such as phosphatidylinositol-phosphate kinase type I␥90 (PIPKI␥90). Integrin that is activated in this way induces the activation of c-Src and FAK, both of which phosphorylate and activate PIPKI␥90 (20,21). Moreover, phosphorylated PIPKI␥90 correlates with an increase in its interaction with talin, and this interaction further stimulates the kinase activity of PIPKI␥90 itself (20,22). These combined mechanisms result in the enhancement of phosphatidylinositol 4,5-bisphosphate synthesis and thus the further promotion of talin binding to integrin, suggesting the positive feedback loop of integrin activation. Thus, the phosphorylation state of PIPKI␥ is important for the regulation of integrin activation.
Based on these lines of evidence, we examined in this study how integrin ␣ v ␤ 3 is inactivated after the nectin-induced formation of AJs by exploring the phosphatase that suppresses the phosphorylation of PIPKI␥ and whether nectin actually associates with this phosphatase and regulates its phosphatase activity.
Cell Lines and Transfection-MDCK cells, HEK293 cells, and L cells were cultured in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal calf serum. For DNA transfection, Lipofectamine 2000 or Lipofectamine Plus (Invitrogen) was applied following the manufacturer's instructions.
Immunofluorescence Microscopy-Immunofluorescence microscopy was performed as described (17). Briefly, the cells were fixed with ice-cold acetone-methanol (1:1) solution for 1 min. After being blocked with 1% bovine serum albumin, the cells were immunostained with the indicated first Abs for 1 h, followed by the incubation with fluorophore-labeled secondary Abs for 30 min. The samples were analyzed by LM510 META confocal microscope (Carl Zeiss).
Immunoprecipitation Assay-MDCK and HEK293 cells expressing various combinations of indicated molecules were lysed with Buffer A (20 mM Tris-HCl, pH 7.5, 150 mM NaCl, 1 mM MgCl 2 , 1 mM Na 3 VO 4 , 1% Nonidet P-40, 1 mM phenylmethylsulfonyl fluoride, 3 g/ml leupeptin, 5 g/ml aprotinin). The cell lysates were centrifuged at 100,000 ϫ g at 4°C for 15 min, and then the supernatant was incubated with the anti-FLAG mAb at 4°C for 2 h followed by incubation with protein G-Sepharose beads at 4°C for 2 h. After the beads were extensively washed with Buffer A, the bound proteins were eluted from the beads by boiling with Laemmli buffer for 5 min and subjected to SDS-PAGE (25), followed by Western blotting with the indicated Abs. To investigate the association of endogenous PTP with nectin-3, MDCK cells cultured on the 0.4-m pored transwell plate (Corning) were treated with a membraneimpermeable cross-linker bis(sulfosuccinimidyl) suberate (Pierce) according to the manufacturer's instructions. After the treatment, the cells were lysed with buffer A, and the cell lysates were incubated with the anti-nectin-3 pAb or the control goat IgG, followed by the incubation with protein G-Sepharose. The immunoprecipitated samples were then analyzed by Western blotting.
In Vitro Binding of PTP and Nectin-3-For the preparation of the purified protein of His-PTP-EC, HEK293 cells were transfected with pFLAG-CMV-5-PTP-EC. At 48 h after the transfection, the culture supernatant containing soluble His-PTP-EC was collected and then was applied to nickel-Sepharose 6 fast flow beads (GE Healthcare) equilibrated with Buffer B (25 mM Tris-HCl, pH 8.0, 200 mM NaCl, 1 mM CaCl 2 , 1 mM MgCl 2 , and 20 mM imidazole at pH 8.0). After the beads were extensively washed with Buffer B, bound His-PTP-EC was eluted with Buffer C (25 mM Tris-HCl, pH 8.0, 200 mM NaCl, 1 mM CaCl 2 , 1 mM MgCl 2 , and 500 mM imidazole at pH 8.0). The protein concentration of His-PTP-EC was determined with bovine serum albumin as a reference protein on SDS-PAGE. The extracellular region of nectin-3 fused to IgG Fc (Nef-3) was prepared as described (26).
To examine the direct binding of His-PTP-EC and Nef-3, His-PTP-EC (6 pmol) was immobilized on nickel-Sepharose beads, and Nef-3 (60 pmol) was incubated with these beads or nickel-Sepharose beads alone as a control in 0.3 ml of Buffer B containing 0.5 mg/ml bovine serum albumin for 1 h. After the beads were extensively washed with Buffer B, the bound proteins were eluted with Buffer C. The eluate was then subjected to SDS-PAGE, followed by Western blotting. Bound Nef-3 was determined by the anti-human IgG Fc pAb conjugated with horseradish peroxidase (GE Healthcare).
Sucrose Density Gradient Centrifugation-The assay for isolation of plasma membrane fraction was performed as described previously (28). Briefly, the MDCK cells were washed with phosphate-buffered saline and then sonicated in Buffer D (10 mM HEPES-NaOH at pH 7.5, 100 mM KCl, 1 mM MgCl 2 , and 25 mM NaHCO 3 ) on ice for 15 s six times at 3-min intervals. The homogenate was centrifuged at 1,000 ϫ g at 4°C for 5 min. The supernatant was diluted with Buffer D into 5 mg/ml of protein, and 0.2 ml was applied on a 4.8 -ml continuous sucrose density gradient (10 -50% sucrose in Buffer D), followed by centrifugation at 100,000 ϫ g at 4°C for 1 h with a swing rotor (P55ST2; Hitachi). After the centrifugation, fractions of 0.3 ml each were collected. Each fraction was subjected to SDS-PAGE, followed by Western blotting with the anti-E-cadherin and anti-talin mAbs.
Assessment for Integrin ␣ v ␤ 3 Activity-MDCK cells cultured on 18-mm coverslips in a 12-well dish were used for the Ca 2ϩ switch assay as described previously (18). Briefly, the cells were washed with phosphate-buffered saline and incubated in serum-free DMEM (Normal Ca 2ϩ medium) for 1 h. Next, the cells were incubated in serum-free DMEM containing 5 mM EGTA (low Ca 2ϩ medium) for 3 h. The cells were then incubated in serum-free DMEM (normal Ca 2ϩ medium) for indicated period. To detect the high affinity form of integrin ␣ v ␤ 3 , we used His-tagged recombinant WOW-1 Fab as described previously (29). Briefly, the cells were incubated with Histagged recombinant WOW-1 Fab for 30 min before the end of the Ca 2ϩ assay. The cells were then washed twice with DMEM and were lysed with Laemmli buffer. The amount of WOW-1 bound to the high affinity form of integrin ␣ v ␤ 3 was detected by Western blotting with the anti-His pAb.
Assay for PTP Phosphatase Activity-The PTP phosphatase activity was assessed using a Universal tyrosine phosphatase assay kit (Takara Bio) as previously described (30). Briefly, HEK293 cells transiently expressing Myc-nectin-3 or HA-Ecadherin were cultured in confluent and lysed with Lysis buffer attached to this kit and then centrifuged at 100,000 ϫ g at 4°C for 15 min. The supernatant was precleared by protein G-Sepharose beads, and the precleared supernatant was incubated with the anti-PTP mAb at 4°C for 2 h, followed by incubation with protein G-Sepharose beads at 4°C for 2 h. After the beads were extensively washed with Buffer E (0.5% Tween 20, 50 mM PIPES, pH 7.0) three times, these beads were suspended into PTP buffer attached to this kit, and the samples were subjected to the phosphatase assay according to the manufacturer's instructions. A paired Student t test was performed for statistical analysis.

RESULTS
Activation of Integrin ␣ v ␤ 3 by PIPKI␥ at Cell-Cell Junctions-We previously showed that after the achievement of AJs, integrin ␣ v ␤ 3 becomes inactive and localizes at cell-cell adhesion sites as well as focal adhesions (17,18). Consistent with this, integrin ␣ v ␤ 3 was concentrated at the cell-cell adhesion sites and co-localized with E-cadherin in confluent MDCK cells, whereas talin was distributed throughout the cytoplasm and did not co-localize with E-cadherin (Fig. 1A), indicating the different localization of integrin ␣ v ␤ 3 from talin. Because talin is involved in the final step of the activation of integrin by directly binding to the cytoplasmic tail of integrin ␤ 3 subunit (31), this different localization of integrin ␣ v ␤ 3 from talin represents the accumulation of the low affinity form of integrin ␣ v ␤ 3 at AJs in confluent MDCK cells.
However, when GFP-PIPKI␥90 was transfected into MDCK cells, talin as well as GFP-PIPKI␥90 was preferentially targeted to the plasma membrane of the cell-cell adhesion sites where the immunofluorescence signal for integrin ␣ v ␤ 3 was concentrated (Fig. 1B, arrowheads), leading to the notion that PIPKI␥90 induces the reactivation of integrin ␣ v ␤ 3 through talin. Interestingly, E-cadherin-based AJs were partially disrupted, probably because of this reactivation of integrin ␣ v ␤ 3 (Fig.  1B, arrow), which was not observed in GFP-transfected (Fig. 1A) or untransfected MDCK cells. These results indicate the critical role of PIPKI␥ in integrin ␣ v ␤ 3 reactivation that causes the instability of AJs. Conversely, the inactivation of PIPKI␥ seems to be at least one of the important underlying mechanisms in the inactivation of integrin ␣ v ␤ 3 after the establishment of AJs. We further examined by knockdown of PIPKI␥ whether endogenous PIPKI␥ is indeed involved in the association of integrin with talin. The expression of PIPKI␥ was markedly reduced in MDCK cells using siRNA against PIPKI␥ (Fig. 1C). Although integrin ␤ 3 and talin clustered well and co-localized at focal adhesions of wild-type MDCK cells, this clustering or co-localization was not observed in PIPKI␥ knockdown MDCK cells (Fig. 1D), indicating the necessity of PIPKI␥ for the association of integrin with talin even at the endogenous level.
Identification of PTP as a Phosphatase for PIPKI␥-It was reported that the kinase activity of PIPKI␥90 is enhanced by its tyrosine phosphorylation (20 -22), resulting in the increased binding of talin to integrin and consequent integrin activation. Thus, the phosphorylation state of PIPKI␥90 is likely to be closely correlated with the regulation of integrin activation. We then examined the implication of the tyrosine phosphatase in the dephosphorylation and inhibition of PIPKI␥90. To explore which phosphatases most effectively dephosphorylate PIPKI␥90, HEK293 cells were co-transfected with FLAG-PIPKI␥90, c-Src, and several tyrosine phosphatases including SHP-1, PTP, and LMW-PTP, and the phosphatase-induced decrease in PIPKI␥90 phosphorylation was monitored. In the presence of SHP-1, the phosphorylation of PIPKI␥90 was slightly reduced ( Fig.  2A). As compared with SHP-1, PTP more markedly decreased the phosphorylation level of PIPKI␥90. In contrast, LMW-PTP did not attenuate PIPKI␥90 phosphorylation. This result indicates that PTP is the most promising candidate for the inhibitor of PIPKI␥. Then we confirmed the association of endogenous PTP with PIPKI␥ in MDCK cells (Fig. 2B). To further specify the role of PTP in the dephosphorylation of PIPKI␥, we knocked down PTP in MDCK cells and examined whether the phosphorylation level of PIPKI␥ is affected by the expression of PTP. The expression of PTP was markedly reduced in MDCK cells using siRNA against PTP (Fig. 2C). The tyrosine phosphorylation of PIPKI␥ was actually enhanced in PTP knockdown MDCK cells (Fig. 2D). These results indicate that PTP specifically acts as a tyrosine phosphatase for PIPKI␥ in MDCK cells.
Inhibitory Effect of PTP on the PIPKI␥-dependent Recruitment of Talin and Activation of Integrin ␣ v ␤ 3 -We next investigated whether PTP is involved in the inhibition of the PIPKI␥-dependent assembly of talin and the inactivation of integrin ␣ v ␤ 3 at the cell-cell adhesion sites in MDCK cells. Co-transfection of FLAG-PIPKI␥90 and GFP-PTP into MDCK cells clearly canceled the PIPKI␥90-dependent translocation of talin to the cell-cell adhesion sites and prevented the disruption of AJs, whereas the phosphatase-inactive mutant of GFP-PTP (GFP-PTPC/S) did not exert such inhibitory effects (Fig. 3A). These results indicate that PTP inhibits the PIPKI␥-induced translocation of talin and its binding to integrin ␣ v ␤ 3 at the cell-cell adhesion sites.
The involvement of endogenous PTP in the localization of talin was further certified biochemically in MDCK cells by the knockdown of PTP. The subcellular localization of talin as well as FAK, a binding protein of integrin (32), was changed from the cytoplasmic to the cytoskeletal membrane fraction by knockdown of PTP (Fig. 3B). In addition, when the total cell lysates were subjected to sucrose density gradient centrifugation, the shift of talin to high density fractions where the membrane marker E-cadherin exists was observed in PTP knockdown MDCK cells (Fig. 3C, lanes 3-6). These results provide another line of evidence that PTP suppresses the translocation of talin to the plasma membrane.
To gain the distinct evidence for the involvement of PTP in the inactivation of integrin ␣ v ␤ 3 , the high affinity form of integrin ␣ v ␤ 3 was monitored using WOW-1 Fab during the formation of cell-cell junctions induced by the Ca 2ϩ switch assay. WOW-1 Fab specifically detects the high affinity form of integrin ␣ v ␤ 3 (33). As compared with wild-type MDCK cells, the high affinity form of integrin ␣ v ␤ 3 in PTP knockdown MDCK cells was increased at 0.5 and 1.5 h after the reculture with normal Ca 2ϩ medium (Fig. 3D). Taken together, these results indicate that PTP perturbs the interaction of talin with integrin ␣ v ␤ 3 and inactivates integrin ␣ v ␤ 3 by the inhibition of PIPKI␥.
Interaction of PTP with Nectin-We next examined the relationship between PTP and the nectin-afadin system in the inactivation of integrin ␣ v ␤ 3 , because nectin and integrin ␣ v ␤ 3 physically and functionally associate together (17). In MDCK cells, PTP co-localized with nectin-3 and afadin endogenously at cell-cell junctions (Fig. 4A). The association of endogenous PTP with nectin-3 in MDCK cells was also confirmed by co-immunoprecipitation assay by chemically cross-linking cell surface proteins (Fig. 4B). Moreover, PTP was co-immunoprecipitated with not only nectin-3 but also other nectin family members including nectin-1, nectin-2, and nectin-4 in HEK293 cells ectopically expressing GFP-PTP with FLAG-nectin-1, FLAG-nectin-2, FLAG-nectin-3, or FLAG-nectin-4, when each FLAG-nectin molecule was immunoprecipitated with the anti-FLAG mAb (Fig. 4C,  panel a). However, another tyrosine phosphatase LMW-PTP was not co-immunoprecipitated with FLAG-nectin-3 (Fig.  4C, panel b), suggesting that the co-immunoprecipitation of PTP with nectin is not nonspecific. In addition, PTP was not co-immunoprecipitated with afadin in HEK293 cells overexpressing these molecules (Fig. 4C, panel c). These results indicate that PTP associates with nectin at the cellcell adhesion sites.
We next prepared the recombinant proteins of the Histagged extracellular region of PTP (His-PTP-EC) and the extracellular region of nectin-3 fused to IgG Fc (Nef-3) to investigate the direct binding of PTP and nectin-3. When Nef-3 was incubated with His-PTP-EC immobilized on nickel beads, the interaction of His-PTP-EC with Nef-3 was detected (Fig. 5B). These results indicate that PTP and nectin-3 physically interact with each other through their extracellular regions at cell-cell junctions.
To examine whether the interaction of nectin with PTP affects the localization of PTP at the cell-cell adhesion sites, GFP-PTP was transfected into L cells expressing full-length nectin-3 (nectin-3-L cells) or nectin-3-⌬EC (nectin-3-⌬EC-L cells) as well as wild-type L cells. The assembly of GFP-PTP at the cell-cell adhesion was markedly increased in nectin-3-L cells compared with wild-type L cells (Fig. 5C). Such an increase in the assembly of GFP-PTP at the cell-cell adhesion was not observed in nectin-3-⌬EC-L cells. Thus, the interaction of nec- and Triton X-100-insoluble (cytoskeletal) fractions. The samples were subjected to SDS-PAGE, followed by Western blotting with the anti-talin and anti-FAK Abs. WT, wild type; KD, knockdown. C, increased recruitment of talin to membrane fraction in PTP knockdown MDCK cells. The cell lysates from wild-type or PTP knockdown MDCK cells were used for sucrose density gradient centrifugation. The fractions were collected from bottom to top of the centrifugation tubes, followed by Western blotting with the anti-talin and anti-E-cadherin Abs. D, necessity of PTP for the inactivation of integrin ␣ v ␤ 3 after the formation of cell-cell junctions. The Ca 2ϩ switch assay was performed in wild-type or PTP knockdown MDCK cells as described under "Experimental Procedures." The cells were incubated with His-tagged WOW-1 Fab for 30 min before the end of the Ca 2ϩ switch assay. The cells were then lysed with Laemmli buffer, and WOW-1 Fab-bound integrin ␣ v ␤ 3 was determined with the anti-His pAb. Actin was also immunoblotted (IB) for the loading control. The results shown in this figure are representative of three independent experiments. tin with PTP is important for the effective recruitment of PTP to the cell-cell adhesion sites where integrin ␣ v ␤ 3 and PIPKI␥90 also localize.
Enhancement of PTP Phosphatase Activity by Trans-interacting Nectin-We finally examined whether the transinteraction of nectin actually enhances the phosphatase activity of PTP for PIPKI␥90. When HEK293 cells ectopically expressing FLAG-PIPKI␥90 and c-Src with or without GFP-PTP and Myc-nectin-3 were cultured in confluence to form the trans-interaction of nectin-3, the c-Src-induced phosphorylation of PIPKI␥90 was reduced by co-transfection of PTP (Fig. 6A). Intriguingly, nectin-3 remarkably increased the dephosphorylation of PIPKI␥90 in the presence of PTP. This nectin-induced dephosphorylation of PIPKI␥90 was not observed when GFP-PTP was not transfected. In contrast, when HEK293 cells expressing FLAG-PIPKI␥90 and c-Src with or without GFP-PTP and HA-Ecadherin were cultured in confluence, the phosphorylation of PIPKI␥90 was not affected by the trans-interaction of E-cadherin, although there is a report that E-cadherin associates with PTP (34). Moreover, we confirmed that the trans-interaction of nectin, but not E-cadherin, significantly raised the phosphatase activity of PTP (Fig. 6B). Taken together, these results indicate that the trans-interaction of nectin preferentially reduces the phosphorylation of PIPKI␥90 mediated by the nectin-induced activation of PTP, eventually resulting in the inactivation of integrin ␣ v ␤ 3 after the establishment of AJs.

DISCUSSION
Nectin and the high affinity form of integrin ␣ v ␤ 3 cooperatively play a pivotal role in the formation of AJs by activating signaling molecules that are necessary for the formation of AJs (13). However, after the achievement of AJs, the high affinity form of integrin ␣ v ␤ 3 is converted into the low affinity form that continues to localize at AJs. The molecular mechanism for how integrin ␣ v ␤ 3 becomes inactive during the formation of AJs has not been elucidated yet to date. In this manuscript, we successfully proposed the molecular mechanism by which trans-interacting nectin also functions to inactivate integrin ␣ v ␤ 3 at mature AJs through the PTP-mediated dephosphorylation of PIPKI␥90. The schematic representation of this molecular mechanism is depicted in Fig. 7. PTP is one of the receptor type protein-tyrosine phosphatases expressing in several epithelial cells and endothelial cells (35). PTP contains a MAM (Merpin/A5/PTP) domain, an Ig-like domain, four fibronectin type III repeats in its extracellular region, and two phosphatase domains in its cytoplasmic region. PTP itself trans-interacts through its MAM and Ig-like domains and localizes at AJs (36 -38). It was previously demonstrated that PTP interacts with the cytoplasmic region of cadherin, ␤-catenin, and p120 ctn and reduces the phosphorylation of these molecules (39,40). In addition to this, we found here that nectin physically associ-ates with PTP through their extracellular region and increases the phosphatase activity of PTP. This increase is somewhat small, but statistically significant, indicating the essential involvement of nectin in the up-regulation of the PTP phosphatase activity. Moreover, the assembly of PTP at the cell-cell adhesion sites is dependent on the association of PTP with nectin, because the signal intensity of GFP-PTP at cell-cell junctions are higher in L fibroblasts expressing full-length nectin-3 than those expressing nectin-3-⌬EC that is incapable of interacting with PTP. Notably, this nectin-mediated PTP assembly does not depends on cadherin, because L fibroblasts do not express any cadherins. However, it remains unknown whether these transmembrane proteins, nectin, cadherin, and PTP form a ternary complex and how these proteins communicate with each other to increase the phosphatase activity of PTP at The cell lysates of HEK293 cells transfected with GFP-PTP and the indicated FLAG-nectin-3 mutants (nectin-3-⌬CP, nectin-3 without its cytoplasmic region; nectin-3-⌬EC, nectin-3 without its extracellular region; nectin-3-⌬C, nectin-3 lacking C-terminal four amino acids that is necessary for binding of nectin to afadin) were immunoprecipitated (IP) with the anti-FLAG mAb. The immunoprecipitants were subjected to Western blotting with the anti-PTP and anti-FLAG mAbs. B, direct interaction of PTP with nectin-3. The recombinant protein of His-PTP-EC was immobilized on nickel-Sepharose beads and incubated with Nef-3. After the incubation, the eluates were subjected to SDS-PAGE, followed by Western blotting with the anti-human IgG Fc for the detection of Nef-3 and anti-His Abs. C, recruitment of PTP to the cell-cell adhesion sites dependent on the association of PTP with nectin-3. Wild-type L cells or L cells ectopically expressing nectin-3 or nectin-3-⌬ EC were transfected with GFP-PTP and then stained for F-actin with phalloidin. Scale bars, 10 m. The results shown in this figure are representative of three independent experiments. IB, immunoblotting. FIGURE 6. Nectin-induced dephosphorylation of PIPKI␥90 and up-regulation of the phosphatase activity of PTP. A, increased dephosphorylation of PIPKI␥90 by the trans-interaction of nectin, but not E-cadherin. Confluent HEK293 cells expressing FLAG-PIPKI␥90 and c-Src with or without GFP-PTP, Myc-nectin-3, and HA-E-cadherin were lysed and immunoprecipitated (IP) with the anti-FLAG mAb, followed by Western blotting with the anti-FLAG and anti-phosphotyrosine (pY) mAbs. B, enhancement of PTP phosphatase activity by nectin. The cell lysates from HEK293 cells expressing Myc-nectin-3 or HA-E-cadherin were immunoprecipitated with the anti-PTP mAb. Untransfected HEK293 cells were used as a control. The phosphatase activity of PTP in the immunoprecipitants were analyzed using tyrosine phosphatase assay kit (Takara Bio). The data shown in this graph are the relative PTP phosphatase activity as compared with the value of the control, which is expressed as 1. *, p Ͻ 0.05. The results shown in this figure are representative of three independent experiments. IB, immunoblotting.
AJs. A detailed investigation to address these issues would be needed in the future.
There are several reports that PIPKI␥90 is involved in the regulation of integrin activation by generating phosphatidylinositol 4,5-bisphosphate and targeting talin to integrin and that the kinase activity of PIPKI␥90 is up-regulated by its phosphorylation (19,20,22). Thus, the phosphorylation level of PIPKI␥90 seems to be important for the regulation of integrin activation. c-Src and FAK are shown to phosphorylate PIPKI␥90 (20,21), but little is known regarding how PIPKI␥90 is dephosphorylated. We demonstrated here that PIPKI␥90 is one of the substrates of PTP and that among several phosphatases, PTP most efficiently dephosphorylates PIPKI␥90 in a trans-interacting nectin-dependent manner. Actually, PTP inhibits the PIPKI␥90-induced recruitment of talin to the cellcell adhesion sites and the reactivation of integrin ␣ v ␤ 3 . Thus, PTP and PIPKI␥ are involved in the dynamic regulation of integrin ␣ v ␤ 3 activation/inactivation during the formation of AJs.
Balanced and controlled phosphorylation of cell-cell adhesion molecules and signaling molecules related to AJs are essential for the formation and maintenance of AJs. Tyrosine phosphorylation of cadherins and their binding proteins catenins as well as nectin-induced phosphorylation of intracellular signaling molecules including c-Src effectively promote the formation of AJs (13,41). In contrast, continuous phosphorylation of E-cadherin leads to its ubiquitination, which enhances its endocytosis and the disassembly of AJs (42). Consistent with this, knockdown of PTP impairs the barrier function in human lung microvascular endothelia because of the improper tyrosine phosphorylation state of VE-cadherin (43). It is of note that we newly show the contribution of PTP to the stability of AJs by regulating the phosphorylation state of PIPKI␥ and inhibiting integrin ␣ v ␤ 3 in addition to its role in dephosphorylation of cadherins and catenins for the maintenance of AJs.
The necessity of PIPKI␥ for the formation of AJs was recently reported (44). In that study, PIPKI␥ interacts with E-cadherin and the subunit of the clathrin adaptor protein complex and facilitates the E-cadherin transport to the plasma membrane through the adaptor protein complex to efficiently form AJs. Downregulation of PIPKI␥ by RNA interference impairs the E-cadherin target to the plasma membrane and inhibits the formation of AJs. Combined with our findings that overexpression of PIPKI␥90 in MDCK cells disrupts E-cadherinbased AJs by the reactivation of integrin ␣ v ␤ 3 , the proper amount of PIPKI␥ and its controlled kinase activation and inactivation are critical for the formation and maintenance of AJs. Because the detailed mechanisms for the prolonged maintenance of AJs remain to be elucidated, further extensive studies are necessary in the future.