Disabled-2 Is a Negative Regulator of Integrin αIIbβ3-mediated Fibrinogen Adhesion and Cell Signaling*

Disabled-2 (DAB2) is an adapter protein that is up-reg-ulated during megakaryocytic differentiation of hematopoietic cells and is abundantly expressed in platelets. In this study, the role of DAB2 in integrin αIIbβ3-mediated matrix protein fibrinogen adhesion and cell signaling was investigated. In K562 cells differentiating to the megakaryocytic lineage, down-regulation of DAB2 by DAB2 small interfering RNA augmented integrin αIIbβ3 activation and resulted in an increase in cell adhesion to fibrinogen. Ectopic expression of DAB2 reversed the DAB2 small interfering RNA effect or, by itself, decreased fibrinogen adhesion of K562 cells. Mutational analysis revealed that a DAB2 Ser24 phosphorylation mutant (S24A) abrogated the inhibitory function of DAB2. The spatial and temporal association/interaction of DAB2 and platelet integrin αIIbβ3 (CD61) in both megakaryocytic cells and platelets led us to examine the effect of Ser24 phosphorylation on the interaction between DAB2 and integrin β3. Through cellular localization and co-immunoprecipitation analysis, we demonstrate for the first time that Ser24 phosphorylation promotes membrane translocation of DAB2 and its subsequent interaction with integrin β3, thereby defining a mechanism for DAB2 in regulating integrin αIIbβ3 activation and inside-out signaling. Consistent with the effect on fibrinogen adhesion, Ser24 phosphorylation of DAB2 was also involved in the negative regulation of αIIbβ3-induced T cell factor transcriptional activity. In contrast, the S24A mutant acted like wild-type DAB2 and inhibited both β-catenin- and plakoglobin-mediated T cell factor transactivation. Hence, DAB2 elicits distinct regulatory mechanisms in αIIbβ3 and β-catenin/plakoglobin signaling in a Ser24 phosphorylation-dependent and -independent manner, respectively. These findings indicate Ser24 phosphorylation as a molecular basis for DAB2 acting as a negative regulator in αIIbβ3 inside-out signaling and contribute to our understanding of DAB2 in megakaryocytic differentiation and platelet function.

In cells, protein phosphorylation of DAB2 modulates its functional activity during growth factor signaling, megakaryocytic differentiation, macrophage spreading, and cell cycle progression. To date, protein kinase C (PKC) and Cdc2 are the only two known DAB2 kinases (14,15). The major PKC phosphorylation site has been mapped to Ser 24 . This phosphorylation site is essential for the inhibitory function of DAB2 in 12-O-tetradecanoylphorbol-13-acetate (TPA)-induced AP-1 gene transcription (14). In addition, the cell cycle-dependent phosphorylation of DAB2 by Cdc2 promotes the association of DAB2 with Pin1, a peptidylprolyl isomerase that regulates the rate of DAB2 dephosphorylation. These findings indicate that DAB2 phosphorylation is important as a regulator of cell proliferation and differentiation.
An appropriate level of DAB2 expression is necessary for maintaining normal cellular and physiological function. In tumors, abnormal expression of DAB2 has been reported in ovarian, breast, prostate, colon, and metastatic pancreatic cancers (16 -19). Homozygotic disruption of DAB2 by knockout experiments demonstrates that DAB2 is essential for embryonic development and kidney transport (20,21). The up-regulation of DAB2 during megakaryocytic differentiation of several hematopoietic leukemic cell lines, including K562, HEL, and MEG-01, has been observed in recent studies in our laboratory; these results indicate a potential function of DAB2 in hematopoietic cell differentiation (8). DAB2 has also been shown to play a role in modulating cell adhesion and the kinetics of MAPK activation using vector-based RNA interference methodology (7). Among the adhesion-related molecules, integrins are a family of major receptors involved in cell adhesion to extracellular matrix proteins and cell-cell adhesion (22)(23)(24). In mammalian cells, all integrins are heterodimers composed of 18 ␣ and 8 ␤ subunits noncovalently linked (23). It is clear that integrins bind to other cell-surface receptors to activate a large variety of signaling pathways that modulate many cellular biological functions, including proliferation, survival/apoptosis, shape, polarity, motility, gene expression, and differentiation (22,23).
In megakaryocytes and platelets, the major integrin ␣ IIb ␤ 3 is expressed at high density in an inactive state (25). Upon platelet activation, ␣ IIb ␤ 3 is activated from within the cells (insideout signaling) so that platelets can bind their major ligand in plasma, fibrinogen, leading to platelet aggregation and thrombosis (26,27). A model for talin-induced inside-out signaling activation and ␣ IIb ␤ 3 conformational changes has been proposed (28). This is mediated through binding of the talin FERM domain or the PTB-like subdomain to the ␤ 3 cytoplasmic tail NPXY motif (29,30). Consequently, the outside-in signaling due to the binding of extracellular ligand further activates a variety of ␣ IIb ␤ 3 -associated signaling molecules, including focal adhesion kinase, integrin-linked kinase, and Grb2, and transmits signals that modulate gene transcription and cytoskeleton organization (23,31). One such protein, the integrin-linked kinase, is known to associate with the integrin ␤ 3 subunit in an activation-dependent manner, and it has been speculated to be part of the process controlling affinity and avidity changes in integrin ␣ IIb ␤ 3 (32). In addition, integrin-linked kinase overexpression leads to the activation of the T cell factor (TCF)/␤catenin signaling pathway, which in turn regulates integrinmediated cell signaling and gene transcription (33).
In this study, the function of DAB2 in integrin ␣ IIb ␤ 3 -mediated fibrinogen adhesion and cell signaling was investigated. We report that DAB2 negatively regulates integrin ␣ IIb ␤ 3 activation, leading to the inhibition of ␣ IIb ␤ 3 -mediated fibrinogen adhesion and TCF transcriptional activity. We further demonstrate that Ser 24 phosphorylation is crucial for membrane translocation of DAB2 and its interaction with the integrin ␤ 3 cytoplasmic tail, thereby defining Ser 24 phosphorylation as a key step in eliciting DAB2 inhibitory function in integrin ␣ IIb ␤ 3 activation and inside-out signaling. These findings contribute to our understanding of DAB2 in megakaryocytic differentiation and platelet function.

EXPERIMENTAL PROCEDURES
Materials-The anti-p96 and anti-PAC-1 antibodies were purchased from BD Biosciences. The anti-T7 tag antibody was purchased from Novagen (Madison, WI). The anti-integrin ␤ 3 antibody was purchased from Santa Cruz Biotechnology (Santa Cruz, CA). The anti-human CD61 antibody was purchased from Serotec (Oxford, United Kingdom). The Dual-Luciferase assay system reagents and pCI-neo vector were purchased from Promega (Madison, WI). Fibrinogen, TPA, and ingenol 3,20-dibenzoate were purchased from Sigma. The LipofectAMINE 2000 reagent was purchased from Invitrogen. The direct CD34 progenitor cell isolation kit was purchased from Miltenyi Biotec (Auburn, CA). The PKC inhibitors GF 109203X, Gö 6976, and rottlerin were purchased from Calbiochem.
Establishment of DAB2 Small Interfering RNA (siDAB2) Stable Cell Lines-The K562 subline K562-9 was transfected with 6 g of pCI-siDAB2 or pCI-neo with LipofectAMINE 2000 reagent. 48 h after transfection, cells were diluted with RPMI 1640 medium containing 600 g/ml G418 for single cell cloning in 96-well plates. Following 4 weeks of selection and expansion, the stable lines were characterized by Western blot analysis and maintained in selective medium containing 300 g/ml G418 for biological and functional analysis.
Cell Aggregation and Cell Adhesion Assay-The assay for cell aggregation and cell adhesion to culture plates was performed as described previously (7). For assay of cell adhesion to fibrinogen, 5 ϫ 10 5 cells/well in 500 l of medium plus 500 l of 0.5% bovine serum albumin and TSM buffer (20 mM Tris-HCl (pH 8.0), 150 mM NaCl, 1 mM CaCl 2 , and 2 mM MgCl 2 ) were added to a 24-well non-tissue culture plate precoated with fibrinogen (10 g). After incubation for 3 h at 37°C, the medium was removed and washed three times with 2 ml of warmed 0.5% bovine serum albumin and TSM buffer. The adhesive cells were fixed with 1% glutaraldehyde and stained with 0.5% crystal violet for 15 min. Finally, the dye was eluted with Sorenson's buffer, and A 570 nm was determined for quantification of cell adhesion to fibrinogen.
Flow Cytometry Analysis-For analysis of PAC-1 binding and CD61 expression, K562 cells were incubated with the fluorescein isothiocyanate-conjugated anti-PAC-1 antibody (20 l) and with the mouse antihuman CD61 monoclonal antibody (1:500) followed by the fluorescein isothiocyanate-conjugated goat anti-mouse antibody, respectively. Flow cytometry analysis was then performed using the FACScan system with CellQuest software (BD Biosciences).
GST Pull-down Analysis-The recombinant GST fusion proteins were expressed and affinity-purified from bacteria as described by the manufacturer (Amersham Biosciences). The GST fusion protein (ϳ25 g) was then incubated overnight with 1 mg of cell lysates in the presence of glutathione-Sepharose beads. After three washes with lysis buffer, excess glutathione (20 mM) was added to elute the GST fusion protein and any associated protein from the beads. The eluted proteins were then resolved by SDS-PAGE followed by Western blot analysis.
Luciferase Activity Assay-Following transient transfection of the indicated luciferase reporter plasmid and expression constructs, cell density was adjusted to 5 ϫ 10 4 cells/well in 24-well culture plates, and cells were cultured for 24 h. The transfected cells were lysed with passive lysis buffer, and an equal amount of cell extract was mixed with luciferase assay reagent II. The firefly luciferase activity was measured with a Lumat LB 9507 tube luminometer.
Immunofluorescence Staining-Cells were directly incubated with antibody (for CD61) or fixed with 3.7% formaldehyde solution at 37°C for 30 min and permeabilized with 0.1% Triton X-100 (for DAB2). After several washes with 1ϫ phosphate-buffered saline, the fixed cells were blocked with 10% fetal bovine serum and incubated with the indicated antibody at room temperature for 1 h. The cells were washed three times with 1ϫ phosphate-buffered saline and incubated with fluorescein isothiocyanate-conjugated secondary antibody (1:500) at room temperature for 1 h. The nucleus was visualized by staining with 4Ј,6diamidino-2-phenylindole. Finally, the cells were cytospun on a glass slide and mounted, ready for observation by confocal microscopy.
Isolation of CD34 ϩ Cells from Cord Blood and ex Vivo Megakaryocytic Differentiation-Mononuclear cells were isolated from umbilical cord blood using Ficoll-Paque Plus (1.077 g/ml). After centrifugation at 1600 rpm for 30 min, the mononuclear cells were transferred to a 50-ml tube, followed by several washes with 1ϫ phosphate-buffered saline. The cell suspension was mixed with Fc receptor-blocking reagents and Microbeads conjugated to the mouse anti-human CD34 monoclonal antibody and subjected to an autoMACS separator (Miltenyi Biotec). The enriched CD34 ϩ cells were induced to megakaryocytic differentiation by culture in Iscove's modified Dulbecco's medium supplemented with 200 g/ml transferrin, 10 ng/ml thrombopoietin, 10 g/ml insulin, 0.8 g/ml lecithin, 0.78 g/ml cholesterol, 0.28 g/ml linoleic acid, 1% bovine serum albumin, and 28 ng/ml CaCl 2 . Megakaryocytic differentiation was confirmed by the presence of a CD61 surface marker using flow cytometry analysis.
Platelet Isolation-Peripheral blood was drawn from healthy drugfree volunteers. Approximately 60 ml of blood was mixed with the anticoagulant solution (acid/citrate/dextrose) at a ratio of 9:1 and centrifuged at 2000 rpm for 10 min to obtain platelet-rich plasma. The platelets were gently resuspended in 4 ml of prostaglandin E1 buffer supplemented with 50 units/ml heparin and centrifuged at 2800 rpm for 8 min. After several washes with Tyrode's buffer, the platelet lysates were isolated for Western blot analysis.
Cellular Fractionation-K562 cells were resuspended in extraction buffer (20 mM Tris-HCl (pH 7.4), 2 mM EDTA (pH 7.4), 10 mM EGTA (pH 7.4), 0.25 M sucrose, 50 mM ␤-mercaptoethanol, 0.4 mM leupeptin, and 0.3 mM phenylmethylsulfonyl fluoride) and homogenized with a Dounce homogenizer. The cell lysates were centrifuged at 100,000 ϫ g for 1 h to obtain the soluble cytosolic fraction. The resulting pellets were washed twice and dissolved in extraction buffer supplemented with 0.5% Triton X-100 for 30 min. Following centrifugation at 49,000 ϫ g for 30 min, the supernatant was collected as the particulate fraction. Statistical Analysis-Student's t test was used for statistical analysis. p Ͻ 0.05 was considered statistically significant.

Effects of siDAB2 on K562 Cell Adhesion to Culture Plates
and Fibrinogen-To elucidate the role of DAB2 in blood cell differentiation and adhesion, K562-9 cells, a clonal subline of K562 retaining the properties of TPA-induced DAB2 and integrin ␤ 3 expression as well as the characteristics of megakaryocytic morphological changes, were stably transfected with the human-specific siDAB2 expression plasmid. The siDAB2 stable line K562-si19 had significantly reduced basal and TPA-induced DAB2 protein levels in comparison with the vector control cells (K562-V7) and the parental cells (K562-9) (Fig. 1A). Both K562-si19 and K562-V7 cells displayed identical cell growth rates and did not adhere to the culture plates under general culture conditions (Fig. 1, B and C). Upon TPA-mediated megakaryocytic differentiation, K562-si19 cells did not undergo cell-cell aggregations, as observed in K562-V7 and FIG. 1. Characterization of K562 cells stably expressing siDAB2. A, down-regulation of DAB2 in K562-si19 cells. The vector control (K562-V7) and siDAB2 (K562-si19) stable clones were seeded at a density of 3 ϫ 10 4 cells/ml and treated with 10 ng/ml TPA (T) or a vehicle control (ethanol (E)) for 48 h. The total cell lysates were collected and analyzed by Western blotting using anti-p96 (DAB2) antibody. The expression of ␤-actin was used for as a control for equal protein loading. B, growth curves of siDAB2 and vector control K562 stable lines. The K562-V7 (V7) and K562-si19 (si19) cells were plated at a density of 3 ϫ 10 5 cells/10-cm culture dish. The total cell numbers were counted at 12, 24, 48, 80, and 100 h after seeding. The data represent the mean Ϯ S.E. of six assays in a representative experiment. C, decrease in K562-si19 cell-cell aggregation and increase in K562-si19 cell adhesion to cell culture plates upon TPA-mediated megakaryocytic differentiation. The K562-V7 and K562-si19 cells were plated at a density of 3 ϫ 10 4 cells/ml, followed by ethanol or TPA treatment for 48 h. The cell adhesive properties were observed by phase-contrast microscopy, and representative fields are shown (magnification ϫ400). D, increase in K562-si19 cell adhesion to fibrinogen during TPA-induced megakaryocytic differentiation. The indicated K562 stable lines (K562-V7 and K562-si19) were cultured and treated with ethanol or TPA for 48 h. The cells were counted, and a total of 5 ϫ 10 5 cells were seeded on the 24-well plates precoated with 10 g of fibrinogen. After 3 h of incubation, the number of cells adhering to fibrinogen was determined by crystal violet staining and quantified using a value of A 570 nm . The data represent the mean Ϯ S.E. of three assays in a representative experiment. Similar results were obtained in three independent experiments. *, p Ͻ 0.05 compared with K562-V7-T cells.

DAB2 and Integrin-mediated Cell Adhesion
K562-9 cells. In contrast, K562-si19 cells adhered to the cell culture plates (Fig. 1C). These results are consistent with our previous findings (7) and indicate that DAB2 plays a role in cell adhesion during megakaryocytic differentiation of K562 cells.
Fibrinogen is a major matrix protein involved in the function of megakaryocytes and platelets in peripheral blood. Hence, the effects of siDAB2 on fibrinogen adhesion were determined using both K562-si19 and K562-V7 cells. Under usual growing conditions, these two stable lines have little fibrinogen adhesive activity and do not adhere well to plates precoated with fibrinogen. In the presence of TPA, K562-V7 fibrinogen adhesion increased by Ͼ30-fold. For K562-si19 cells, the adhesive activity for fibrinogen was increased by Ͼ111-fold compared with K562-V7-E control cells and by 2.5-fold compared with K562-V7-T cells (Fig. 1D).
The effect of siDAB2 on fibrinogen adhesion was further demonstrated in transient transfection experiments. The transfection efficiency was ϳ58% according to a control transfection with an enhanced green fluorescent protein expression plasmid (data not shown). Similar to the results with K562-si19 cells, transfection of the siDAB2 plasmid Dab2-2112 significantly increased fibrinogen adhesion of K562 cells in the presence of TPA ( Fig. 2A). However, small interfering RNAs targeting firefly luciferase (Ff1) and the HCV 5Ј-noncoding region (HCVsi) did not alter cell adhesion to fibrinogen. Most significantly, the effect of Dab2-2112 on cell adhesion could be reversed by coexpression of rat DAB2 p82, which was not the target of and was not down-regulated by Dab2-2112 (Fig. 2B). These data indicate that DAB2 is a negative regulator that modulates fibrinogen adhesion of K562 cells.
siDAB2 Modulates Fibrinogen Adhesion through Activation of Integrin ␣ IIb ␤ 3 -Integrin ␣ IIb ␤ 3 is usually involved in the fibrinogen adhesion of megakaryocytes and platelets. Whether or not DAB2 has any effect on integrin ␣ IIb ␤ 3 expression and activation was investigated in the following experiments. At first, to provide direct evidence for the involvement of integrin ␣ IIb ␤ 3 in the fibrinogen adhesion of K562 cells and their sublines, the anti-PAC-1 antibody, which recognizes the active form of ␣ IIb ␤ 3 and inhibits specifically ␣ IIb ␤ 3 -mediated cell adhesion (35,36), was added to the medium during the fibrinogen adhesion assay. Increasing concentrations of the anti-PAC-1 antibody (but not the control antibody) resulted in a dose-dependent inhibition of TPA-induced fibrinogen adhesion of both K562 and K562-V7 cells (Fig. 3, A and B). The anti-PAC-1 antibody also caused inhibition of the siDAB2-mediated increase in fibrinogen adhesion in K562-si19 cells (Fig. 3B). These data indicate that integrin ␣ IIb ␤ 3 is involved in the fibrinogen adhesion of K562 cells and their sublines.
The expression and activation status of integrin ␣ IIb ␤ 3 were then compared in K562-V7 and K562-si19 cells. Flow cytometry analysis revealed that 81.27% of the K562-V7 cells and 72.37% of the K562-si19 cells expressed the megakaryocytic differentiation surface marker integrin ␣ IIb ␤ 3 (CD61) (Fig. 3C). No significant difference in CD61 expression was observed between these two cell lines, which is consistent with our previous observation that siDAB2 does not affect integrin ␤ 3 expression in Western blot analysis (7). In contrast, the activation status of integrin ␣ IIb ␤ 3 revealed a remarkable increase in cells with siDAB2 as analyzed by PAC-1 binding (Fig. 3C). These results indicate that siDAB2-mediated integrin ␣ IIb ␤ 3 activation accounts for the increase in fibrinogen adhesion and support the role of DAB2 as a negative regulator of ␣ IIb ␤ 3 .
Spatial and Temporal Association of DAB2 and Integrin ␤ 3 during Megakaryocytic Differentiation of Hematopoietic Cells and in Platelets-In our previous study, we observed that DAB2 and integrin ␤ 3 are both up-regulated during TPA-induced megakaryocytic differentiation of human K562, MEG-01, and HEL cells (7). The following experiments were performed to demonstrate the spatial and temporal association of DAB2 and integrin ␤ 3 . By immunofluorescence staining and confocal microscopy analysis, both DAB2 and integrin ␤ 3 were found to distribute mainly surrounding the plasma membrane in TPA-treated K562 cells (Fig. 4A). The data indicate a spatial association between DAB2 and integrin ␤ 3 during megakaryo- cytic differentiation of K562 cells. To further reveal the temporal association of DAB2 and integrin ␤ 3 , selective pharmaceutical agents implicated in the PKC signaling pathways, including GF 109203X (conventional and novel PKC inhibitor), Gö 6976 (novel PKC inhibitor), rottlerin (PKC␦ inhibitor), and the phorbol ester ingenol 3,20-dibenzoate (selective PKC⑀ activator), were used to treat K562 cells, and their effects on DAB2 and integrin ␤ 3 expression were examined. The PKC inhibitors suppressed TPA-induced DAB2 and integrin ␤ 3 in a dose-dependent manner, whereas the selective PKC⑀ activator ingenol 3,20-dibenzoate, which induces K562 cell megakaryocytic differentiation (37), increased both DAB2 and integrin ␤ 3 with a maximal induction at a concentration of 0.1 M (Fig. 4B). Similarly, the MEK1 (mitogen-activated protein kinase/extracellular signal-regulated kinase kinase-1) inhibitor U0126 (8) and the platelet-derived growth factor receptor inhibitor STI571 2 also inhibited TPA-induced DAB2 and integrin ␤ 3 expression concomitantly (data not shown). These results indicate that DAB2 and integrin ␤ 3 are coexpressed and co-regulated during megakaryocytic differentiation.
The temporal association of DAB2 and integrin ␤ 3 was also analyzed using an ex vivo megakaryocytic differentiation model. CD34 ϩ hematopoietic pluripotent progenitor cells were isolated from umbilical cord blood and induced by thrombopoietin to undergo megakaryocytic differentiation. In the CD34 ϩ progenitor cells, DAB2 and integrin ␤ 3 (CD61) were barely detected by Western blotting and flow cytometry analysis, respectively (Fig. 4C). Following 10 days of thrombopoietin treatment of CD34 ϩ cells, both DAB2 and integrin ␤ 3 were induced, with ϳ87% of the cells expressing the CD61 megakaryocytic differentiation surface marker (Fig. 4C). In accord with these findings, platelets (but not neutrophils) isolated from the blood of healthy volunteers expressed highly abundant amounts of DAB2 and integrin ␤ 3 (Fig. 4D). Given that the proteins that participate in the same pathway or that are part of the same protein complex are often coexpressed and co-regulated (38), these results indicate that DAB2 and integrin ␤ 3 are likely to have a functional link that may account for the effects of siDAB2 on the argumentation of integrin ␣ IIb ␤ 3 activation and fibrinogen adhesion. regulate integrin ␣ IIb ␤ 3 activity. The temporal and spatial association of DAB2 and integrin ␤ 3 has prompted us to examine whether or not there is a direct interaction between them. Immunoprecipitation of the TPA-treated K562 cells using anti-DAB2 antibody H-110, but not the control antibody, revealed that integrin ␤ 3 associated with DAB2 in vivo (Fig. 5A, left  panels). The association occurred in a time-dependent manner (Fig. 5A, right panels). Strong association of DAB2 and integrin ␤ 3 occurred in the platelets as well (Fig. 5B). Ectopic expression of T7-tagged DAB2 in K562 cells also revealed a time-dependent association of DAB2 and integrin ␤ 3 (Fig. 5C). These results indicate that endogenous and exogenous DAB2 proteins bind integrin ␤ 3 in both megakaryocytic cells and platelets. Further analysis with recombinant proteins revealed that the GST-⌬B fusion protein, containing the N terminus of DAB2, bound integrin ␤ 3 present in TPA-treated K562 cell lysates (Fig. 5D), whereas the GST-␤ 3 fusion protein, containing the integrin ␤ 3 cytoplasmic tail, bound DAB2 in TPA-treated K562 cell lysates (Fig. 5E). Hence, the PTB domain of DAB2 and the cytoplasmic tail of integrin ␤ 3 are involved in the association of DAB2 with integrin ␤ 3 .
To determine whether or not TPA/PKC-mediated Ser 24 phosphorylation affects the binding of DAB2 to integrin ␤ 3 , either T7-tagged wild-type DAB2 (p82) or the S24A mutant was transfected into K562 cells, followed by TPA treatment for 24 h. Co-immunoprecipitation analysis with the anti-T7 tag antibody revealed that S24A bound less to integrin ␤ 3 compared with the p82 control (Fig. 5F). These results demonstrate that Ser 24 phosphorylation is important in the binding of DAB2 to integrin ␤ 3 .

Ser 24 Phosphorylation Promotes DAB2 Membrane Translocation and Is a Key
Step in DAB2 Regulation of Integrin ␣ IIb ␤ 3mediated Fibrinogen Adhesion-The decrease in binding of S24A to integrin ␤ 3 prompted us to determine the effect of Ser 24 phosphorylation on the cellular localization of DAB2. Either p82 or S24A was transfected into K562 cells for immunofluorescence staining with the anti-T7 tag antibody. In the absence of TPA, both p82 and S24A displayed cytoplasmic staining (Fig.   FIG. 4. Spatial and temporal association of DAB2 and integrin ␤ 3 during megakaryocytic differentiation of hematopoietic cells and in platelets. A, co-localization of DAB2 and integrin ␤ 3 . K562 cells were transfected with DAB2 expression plasmid and treated with TPA for 48 h prior to immunostaining with the anti-DAB2 (T7 tag) and anti-CD61 antibodies. The cells were then cytospun onto a glass slide and observed by confocal microscopy. FITC, fluorescein isothiocyanate; DAPI, 4Ј,6-diamidino-2-phenylindole. B, co-regulation of DAB2 and integrin ␤ 3 in K562 cells. K562 cells were pretreated with the indicated PKC inhibitors and activator for 30 min, followed by TPA treatment for 48 h. The cell lysates were subjected to Western blot analysis with anti-p96 (DAB2), anti-integrin ␤ 3 , and anti-␤-actin antibodies. IDB, ingenol 3,20-dibenzoate. C, coexpression of DAB2 and integrin ␣ IIb ␤ 3 during ex vivo megakaryocytic differentiation of human CD34 ϩ hematopoietic pluripotent progenitor cells. The CD34 ϩ stem cells (D 0 ) were isolated from cord blood and induced to megakaryocytic differentiation by thrombopoietin for 10 days (D 10 ). Total cell lysates were collected and subjected to Western blot analysis with anti-p96 (DAB2) antibody. The expression of ␤-actin was used as a control for equal protein loading. The expression of integrin ␤ 3 was determined by flow cytometry analysis using anti-CD61 antibody. D, expression of DAB2 and integrin ␤ 3 in platelets and neutrophils. The total lysates of platelets (P) and neutrophils (N) were isolated from whole blood in the appropriate lysis buffer and subjected to Western blot analysis with anti-p96 (DAB2), anti-integrin ␤ 3 , and anti-␤-actin antibodies. The total lysates of K562 cells treated with a vehicle control (ethanol (KE)) or TPA (KT) for 48 h were also included for comparison. 6A). After 1 h of TPA stimulation, p82-transfected cells displayed a membrane and cluster staining pattern, indicating that DAB2 translocated from the cytosol to the membrane compartment. S24A remained in the cytosol and displayed even cytoplasmic staining. Similar results were observed when comparing the protein localization of the C-terminally truncated form of DAB2, ⌬B, and ⌬B-S24A (data not shown). To further confirm these observations, K562 cells were transfected with p82 and S24A, and the cell lysates were fractionated into cytosolic and particulate fractions. After TPA treatment, p82 was translocated to the particulate fraction (Fig. 6B). In contrast, S24A was found only in the cytosolic fraction. These data reveal the essential role of Ser 24 phosphorylation in the membrane translocation of DAB2 and its subsequent interaction with integrin ␣ IIb ␤ 3 .
To determine whether or not Ser 24 phosphorylation of DAB2 plays a role in regulating integrin ␣ IIb ␤ 3 -mediated fibrinogen adhesion, various DAB2 constructs were cotransfected with Dab2-2112 into K562 cells to assess their effect on reversing the siDAB2-mediated increase in fibrinogen adhesion (Fig. 7, A  and B). In the presence of TPA, the two DAB2 deletion mutants, ⌬B and ⌬N, behaved like p82 and reduced fibrinogen adhesion to the level of vector control cells (Fig. 7C). In contrast, coexpression of S24A did not alter the siDAB2mediated increase in fibrinogen adhesion, indicating the involvement of DAB2 Ser 24 phosphorylation in fibrinogen adhesion. To confirm this, K562 cells were transfected with p82 and S24A, respectively. In the presence of TPA, p82 (but not S24A) reduced fibrinogen adhesion (Fig. 7D). These results reveal that Ser 24 phosphorylation of DAB2 is important to elicit its negative regulatory role in ␣ IIb ␤ 3 -mediated fibrinogen adhesion.
Ser 24 Phosphorylation of DAB2 Is Required for Its Inhibitory Function in ␣ IIb ␤ 3 -mediated (but Not ␤-Cateninand Plakoglobin-mediated) TCF Transcriptional Activity-The importance of DAB2 Ser 24 phosphorylation in integrin ␣ IIb ␤ 3 signaling was further delineated for the activation of the TCF/␤-catenin pathway, which has been implicated in cell adhesion signaling. We first characterized the expression of a luciferase reporter construct, pTcf4RE-luc, which contains four copies of the TCF4responsive element at the 5Ј-end of the ␤-globin basic promoter, during megakaryocytic differentiation of the siDAB2 stable line. An increase in luciferase activity was observed in both K562-si19 and K562-V7 cells in the presence of TPA (Fig. 8A).

FIG. 5. DAB2 is a binding partner of integrin ␤ 3 , and Ser 24 is an important amino acid involved in this binding.
A and B, in vivo association of DAB2 and integrin ␤ 3 in K562 cells. The lysates (5 mg) from K562 cells treated with TPA for 48 h (A, left panels) and from K562 cells (2 mg) treated with ethanol (E) and TPA (T) for the indicated times (A, right panels) and platelet lysates from healthy human volunteers (B) were subjected to immunoprecipitation (IP) with anti-DAB2 and anti-integrin ␤ 3 antibodies (Ab) and/or control IgG antibody (C). The immunoprecipitated proteins were dissolved in sample buffer, and Western blot (WB) analysis was performed with the anti-integrin ␤ 3 and anti-p96 (DAB2) antibodies. C, in vivo association between transfected DAB2 and integrin ␤ 3 in a time-dependent manner. pCI-neo-DAB2 was transfected into K562 cells with LipofectAMINE 2000 and treated with TPA for the indicated times. The total cell lysates were collected for immunoprecipitation with anti-T7 tag antibody (2 g). The immunoprecipitated proteins were subjected to Western blot analysis with the anti-T7 tag and anti-integrin ␤ 3 antibodies. D and E, in vitro binding between DAB2 and integrin ␤ 3 . K562 cells were treated with TPA (10 ng/ml) for 48 h, and the total cell lysates were collected for GST pull-down analysis with the GST control, the GST-DAB2 PTB domain fusion protein (GST-⌬B), and the GST-integrin ␤ 3 cytoplasmic tail fusion protein (GST-␤ 3 ). The pull-down lysates were then probed for the indicated proteins by Western blot analysis. The input amounts of integrin ␤ 3 , GST, GST-⌬B, DAB2, and GST-integrin ␤ 3 are also shown. F, reduced binding between integrin ␤ 3 and the DAB2 S24A mutant. Wild-type DAB2 (p82) and S24A was transfected into K562 cells, followed by TPA treatment for 24 h. T7-tagged DAB2 was immunoprecipitated with the anti-T7 tag antibody and subjected to Western blot analysis with the anti-T7 tag and anti-integrin ␤ 3 antibodies.
The luciferase activity was significantly higher in K562-si19 cells than in K562-V7 control cells. To determine whether or not the ␣ IIb ␤ 3 signaling relates to the activation of TCF transcriptional activity, we transiently coexpressed ␣ IIb and ␤ 3 with pTcf4RE-luc in K562 cells. Although integrin ␣ IIb or ␤ 3 alone did not affect the level of luciferase expression (data not shown), coexpression of integrins ␣ IIb and ␤ 3 resulted in an ϳ2-3-fold increase in luciferase activity, which was further enhanced in the presence of TPA (Fig. 8B). In addition, expression of p82 resulted in inhibition of the ␣ IIb ␤ 3 -mediated increase in pTcf4RE-luc luciferase activity. These results indicate that TCF activation is part of the integrin ␣ IIb ␤ 3 downstream signals and that DAB2 acts as a negative regulator in ␣ IIb ␤ 3 -mediated TCF signaling.
The role of DAB2 Ser 24 phosphorylation in integrin ␣ IIb ␤ 3mediated TCF transactivation was also characterized by cotransfection experiments with various DAB2 mutants. The results show that ⌬B and ⌬N behaved like wild-type DAB2 p82 and inhibited integrin ␣ IIb ␤ 3 -mediated increases in pTcf4REluc luciferase activity (Fig. 8B). The inhibitory effect is specific because expression of the membrane protein phospholipid scramblase-His 9 did not alter the luciferase activity. In contrast, S24A did not abrogate integrin ␣ IIb ␤ 3 -mediated pTcf4RE-luc expression. These data indicate that Ser 24 phosphorylation is necessary for DAB2 to elicit its inhibitory function in ␣ IIb ␤ 3 signaling.
In transient transfection experiments, we also found that DAB2 repressed both ␤-cateninand plakoglobin-mediated increases in pTcf4RE-luc luciferase activity (Fig. 8, C and D). To further demonstrate whether or not Ser 24 phosphorylation of DAB2 also plays a role in its inhibitory function in ␤-catenin-and plakoglobin-mediated TCF transactivation, mutant forms of DAB2 were cotransfected with pTcf4RE-luc in the presence of either ␤-catenin or plakoglobin. Analysis of the luciferase activity revealed that all DAB2 mutants, including S24A, ⌬B, and ⌬N, reduced ␤-cateninand plakoglobininduced luciferase activity (Fig. 8, C and D). These data indicate that Ser 24 phosphorylation is not crucial for DAB2 to inhibit ␤-cateninand plakoglobin-mediated TCF transcriptional activity. Therefore, DAB2 regulates ␣ IIb ␤ 3 and ␤-catenin/plakoglobin signaling through different mechanisms in a Ser 24 phosphorylation-dependent and -independent manner, respectively. DISCUSSION DAB2 is up-regulated during megakaryocytic differentiation of human leukemic cell lines (7,8) and CD34 ϩ hematopoietic pluripotent stem cells. Accordingly, DAB2 is highly expressed in human platelets. It has been shown that the platelet integrin ␣ IIb ␤ 3 is a differentiation marker of megakaryocytes and platelets and mediates cell adhesion. Using siDAB2, we have shown previously that DAB2 is involved in the regulation of cell-cell adhesion and MAPK activity (7). The spatial and temporal association of DAB2 and integrin ␣ IIb ␤ 3 as revealed in this study further suggests a pivotal role of DAB2 in integrinmediated cell adhesion. In this study, we have demonstrated for the first time that DAB2 possesses anti-adhesive activity. It negatively regulated megakaryocytic cell adhesion to fibrinogen, an extracellular matrix protein that is crucial for platelet activation and aggregation during blood coagulation. This conclusion is supported by the observation that expression of siDAB2 increased fibrinogen adhesion. Moreover, ectopic expression of DAB2 reversed the siDAB2 effect or, by itself, decreased fibrinogen adhesion of K562 cells. It has been reported that, similar to DAB2, the platelet protein plasminogen activator inhibitor type-1 also possesses anti-adhesive activity and interferes with platelet and megakaryoblastic cell adhesion (39). Due to the fact that DAB2 is highly enriched in platelets, our data imply that the anti-adhesive activity of DAB2 may have its physiological function during platelet activation and aggregation.
In contrast to our observations, DAB2 has been shown to play a role in enhancing integrin-mediated adhesion to collagen IV and laminin and non-integrin-mediated adhesion to poly-Dlysine substrate (6). It is not yet completely understood how DAB2 mediates such distinct effects on integrin-mediated cell adhesion. A recent study reported on the differential interaction of the DAB2 PTB domain with the cytoplasmic tail of various integrin ␤ subunits: strong interaction with ␤ 3 and ␤ 5 and no interaction with ␤ 1 and ␤ 2 (11). This may provide a basis to explain the dual effects of DAB2 on integrin-mediated cell adhesion. In this study, we have provided the first experimental evidence to demonstrate that the binding of integrin ␤ 3 by DAB2 negatively regulates the activation of integrin ␣ IIb ␤ 3 . This was revealed by the change in binding of anti-PAC-1 antibody, which recognizes the active form of ␣ IIb ␤ 3 that mediates high affinity and avidity fibrinogen binding (35), in the K562 stable line expressing siDAB2. We thus conclude that DAB2 serves as an important negative signaling element in megakaryocytic cells and platelets in the control of the activa-FIG. 6. Ser 24 phosphorylation plays an essential role in DAB2 membrane translocation. A, K562 cells were transfected with wildtype DAB2 (p82) or the DAB2 S24A mutant. Following a 1-h treatment with a vehicle control (ethanol) or 10 ng/ml TPA, immunofluorescent staining was performed with the anti-T7 tag antibody, followed by observation by confocal microscopy. FITC, fluorescein isothiocyanate; DAPI, 4Ј,6-diamidino-2-phenylindole. B, K562 cells were transfected with p82 and S24A. After 3 h of the indicated treatment, the cell lysates were fractionated into cytosolic (C) and particulate (P) fractions as described under "Experimental Procedures." The protein levels of DAB2 were determined by Western blot analysis with the anti-T7 tag antibody. tion and signaling of ␣ IIb ␤ 3 . The ␣ IIb ␤ 3 activation results from many cellular events that converge in integrin inside-out signaling (40). For example, activation of conventional PKC isoforms, phosphatidylinositol 3-kinase-mediated activation of novel and atypical PKCs, members of the Ras GTPase superfamily, and G-protein-coupled receptor signaling have all been reported to play a role in inside-out signaling and ␣ IIb ␤ 3 activation (41,42). These signals direct the interaction between integrin and regulatory proteins, including Src, Syk, Shc, myosin, Pyk2, integrin-associated protein, CD9, talin, and ␤ 3endonexin, that stimulate conformational changes in the receptor to expose binding sites for fibrinogen (43). Of these integrin ␤ cytoplasmic tail-interacting proteins, DAB2 binds to Src and myosin and modulates their activity (4,10). It is plausible that the binding of DAB2 to integrin ␤ 3 alters the protein complexes associated with the cytoplasmic tail of ␤ 3 . Although the effects of DAB2 on integrin signaling still need to be characterized, our study provides clues to the biological effect of DAB2-integrin ␤ 3 interaction and may explain the dual functions of DAB2 based on the fact that DAB2 has distinct binding activity for different integrin ␤ subunits.
We have extended our study to demonstrate the molecular basis for DAB2 acting as a negative regulator of ␣ IIb ␤ 3 activation. The evolutionary conservation of Ser 24 and its flanking sequences in DAB2 from mammalian species such as human, rat, and mouse implies that Ser 24 is an important amino acid for DAB2. Indeed, TPA/PKC-dependent Ser 24 phosphorylation of DAB2 is crucial for inhibiting TPA-induced AP-1 activity (14). We have further revealed in this study that Ser 24 phosphorylation of DAB2 is involved in the decrease in integrinmediated fibrinogen adhesion and TCF transactivation. In addition, Ser 24 phosphorylation is critical for membrane translocation of DAB2. These findings are in accord with previous studies that reported the phosphorylation-dependent redistribution of key signaling proteins such as focal adhesion kinase, Ras GTPase-activating protein, and HSP25 (44 -46). As a result, the inability of S24A to control ␣ IIb ␤ 3 signaling is caused by the restricted distribution of DAB2 in the cytosol and the lack of interaction with integrin ␤ 3 . Ser 24 phosphorylation of DAB2 serves as a key step in the negative regulation of integrin ␤ 3 -mediated cell adhesion and signaling. The DAB2 N-terminal deletion mutant ⌬N also acts like wild-type DAB2 in the regulation of TCF transcription. The multiprotein signaling network complexes of integrin ␤ 3 and DAB2 thus appear to be more complicated than expected.
In addition to acting as a negative regulator of integrin signaling, DAB2 negatively regulates ␤-catenin-mediated Wnt signaling. In addition, we have revealed that plakoglobin-mediated Wnt signaling is also diminished by DAB2. By mutationalanalysis,wefoundthat,incontrasttoSer 24 phosphorylation-FIG. 7. Ser 24 phosphorylation is essential for DAB2 modulation of K562 cell adhesion to fibrinogen. A and B, schematic representation and expression of T7-tagged DAB2 and its deletion mutants. K562 cells were transfected with the indicated plasmids and subjected to Western blot (WB) analysis with the anti-T7 tag antibody. PID, phosphotyrosine interacting domain; PRD, proline-rich domain. C, reversing the effect of human siDAB2 on fibrinogen adhesion by various rat DAB2 mutants. pTOPO-U6 (pU6) or Dab2-2112 (2112) was cotransfected with the indicated rat DAB2 constructs into K562 cells. 24 h after transfection, the cells were plated at 6 ϫ 10 5 cells/10-cm culture dish and treated with ethanol (E) or 10 ng/ml TPA (T) for 48 h, followed by the fibrinogen adhesion assay. **, p Ͻ 0.01 compared with Dab2-2112-T. D, wild-type DAB2 (p82), but not S24A, reduces fibrinogen adhesion. The pCI-neo vector control, p82, and S24A were transiently transfected into K562 cells. After 48 h treatment with ethanol or TPA as described for C, fibrinogen adhesion was determined and quantified. The data represent the mean Ϯ S.E. of three assays in a representative experiment. Similar results were obtained in two independent experiments. **, p Ͻ 0.01 compared with pCI-neo-T. dependent regulation of integrin signaling, the inhibition of ␤-catenin and plakoglobin signaling by DAB2 is Ser 24 phosphorylation-independent. These observations indicate that DAB2 participates in the control of cell signaling through multiple mechanisms. We postulate that Ser 24 phosphorylation is a key step in determining the DAB2 mode of action. Because DAB2 with Ser 24 phosphorylated tends to translocate to the plasma membrane (particulate) fraction, its interaction with transmembrane proteins and receptor protein-tyrosine kinases such as integrin ␤ subunits appears to be regulated in a Ser 24 phosphorylation-dependent manner. On the other hand, interaction with other cytosolic proteins may be mediated in a Ser 24 phosphorylation-independent manner. Hence, the reported interaction between DAB2 and the cytosolic protein Dvl-3, a signaling mediator of the ␤-catenin/Wnt pathway leading to TCF transactivation (12), appears to be mediated in a Ser 24 phosphorylation-independent manner. This could explain why the S24A mutant of DAB2 still elicits inhibitory activity for ␤-catenin-mediated increases in TCF-luciferase activity. Together, DAB2 has multiple modes of action and elicits distinct regulatory mechanisms in ␣ IIb ␤ 3 and ␤-catenin/plakoglobin signaling in a Ser 24 phosphorylation-dependent and -independent manner, respectively.
In summary, this study shows for the first time the antiadhesive activity of DAB2 in integrin ␣ IIb ␤ 3 -mediated fibrinogen adhesion. In addition, the Ser 24 phosphorylation-dependent membrane translocation and the subsequent interaction with integrin ␤ 3 provide a mechanism for DAB2 to regulate activation of integrin ␣ IIb ␤ 3 and its downstream signaling, including fibrinogen adhesion and TCF transactivation (Fig. 9). Such information can lead to defining a role and expanding our understanding of DAB2 in megakaryocytic differentiation and platelet function. Such interaction decreases ␣ IIb ␤ 3 activation, ␣ IIb ␤ 3 -mediated fibrinogen adhesion (inside-out signaling), and TCF transcriptional activity. ILK, integrin-linked kinase; FAK, focal adhesion kinase.
FIG. 8. Ser 24 phosphorylation is a key step in the DAB2 regulation of integrin ␣ IIb ␤ 3 -mediated (but not ␤-cateninand plakoglobin-mediated) TCF transcriptional activity. A, shown is TCF transactivation in siDAB2 stable lines. The TCF reporter plasmid pTcf4REluc was transfected into K562-V7 (V7) and K562-si19 (si19) cells. After TPA (T) and vehicle control (ethanol (E)) treatment for 48 h, the luciferase activity was determined using the Dual-Luciferase assay system. B-D, the TCF reporter plasmid pTcf4RE-luc (0.1 g) and the indicated DAB2 expression plasmids (0.7 g) were cotransfected into K562 cells in the presence of 0.4 g of ␣ IIb ␤ 3 , ␤-catenin, and plakoglobin expression plasmids, respectively. The transfected cells were treated with solvent vehicle (ethanol) or TPA for 48 h. The luciferase activity was then determined by the Dual-Luciferase assay system. The data represent the mean Ϯ S.E. of three assays in a representative experiment. * and **, p Ͻ 0.05 and p Ͻ 0.01, respectively, compared with vector/ ␣ IIb ␤ 3 (B), vector/␤-catenin (C), and vector/plakoglobin (D).