Leukocyte extravasation is an important step of inflammation, in which integrins have been demonstrated to play an essential role by mediating the interaction of leukocytes with the vascular endothelium and the subendothelial extracellular matrix. Previously, we identified an integrin-linked kinase (ILK)-binding protein affixin (β-parvin), which links initial integrin signals to rapid actin reorganization, and thus plays critical roles in fibroblast migration. In this study, we demonstrate that γ-parvin, one of three mammalian parvin family members, is specifically expressed in several lymphoid and monocytic cell lines in a complementary manner to affixin. Like affixin, γ-parvin directly associates with ILK through its CH2 domain and colocalizes with ILK at focal adhesions as well as the leading edge of PMA-stimulated U937 cells plated on fibronectin. The overexpression of the C-terminal fragment containing CH2 domain or the depletion of γ-parvin by RNA interference inhibits the substrate adhesion of MCP-1-stimulated U937 cells and the spreading of PMA-stimulated U937 cells on fibronectin. Interestingly, the overexpression of the CH2 fragment or the γ-parvin RNA interference also disrupts the asymmetric distribution of PTEN and F-actin observed at the very early stage of cell spreading, suggesting that the ILK-γ-parvin complex is essential for the establishment of cell polarity required for leukocyte migration. Taken together with the results that γ-parvin could form a complex with some important cytoskeletal proteins, such as αPIX, α-actinin, and paxillin as demonstrated for affixin and actopaxin (α-parvin), the results in this study suggest that the ILK-γ-parvin complex is critically involved in the initial integrin signaling for leukocyte migration.

Leukocyte migration from the vasculature into tissues, called extravasation, is an essential event in the process of inflammation. The regulation of these processes can be a key target for controlling inflammatory diseases, including bronchial asthma, collagen diseases, and atherosclerosis. The selective recruitment of leukocytes from the peripheral blood to the inflammatory tissues and the recirculation of lymphocytes between the blood and lymph nodes consist of sequential multiple processes, such as weak adhesion to and rolling on endothelial cells by selectins, firm attachment between leukocyte integrins and Ig superfamily members on endothelium, and extravasation, in which leukocytes migrate through the endothelium into tissues (1). After the transendothelial migration, the leukocytes further migrate into the surrounding tissues within several hours by their association with the extracellular matrix (ECM)3 through their corresponding integrins. During these steps, specific chemokines presented on the endothelium or classical chemoattractants, like C5a, modulate leukocyte affinity for endothelial and matrix ligands via rapid activation of leukocyte β1 and/or β2 integrins (2). In particular, adhesion to VCAM-1 on the endothelium or to ECM glycoproteins, such as fibronectin (FN), via α4β1 and α5β1 integrins contributes to leukocyte migration independently of β2 integrin (αL/Mβ2) (3, 4). Thus, the adhesion control by β1 integrin is one of the basic elements in migration and localization of leukocytes.

Integrin-linked kinase (ILK), a ubiquitously expressed serine/threonine protein kinase capable of interacting with the cytoplasmic domain of integrins β1 and β3 at focal adhesions (FA), has been involved in various cellular signalings, including integrin activation (inside-out signal), cell survival induced by appropriate cell-matrix interactions (outside-in signal), and differentiation (5, 6, 7). More importantly, the critical involvement of ILK in cell migration has been reported in recent analyses of null models. Gene disruption studies of ILK in Drosophila melanogaster and Caenorhabditis elegans revealed its essential role as an interface between integrin and actin cytoskeleton (8, 9). In mice, conditional or complete disruption of the ILK gene clarified its function as a possible regulator of integrin-cytoskeleton cross-links. For instance, ILK-deficient fibroblasts and chondrocytes showed incomplete adhesion and spreading with abnormal cytoskeletal reorganization (10, 11). Endothelial cells from ILK-null mice formed a poor vascular network possibly due to the defect in their migration (12). However, despite a growing body of evidence about the critical importance of ILK in cell spreading and motility, the role of ILK in leukocyte migration has not been well documented as yet. Interestingly, Friedrich et al. (13) showed that ILK, which is highly expressed in human mononuclear leukocytes, is activated by exposure to chemokine in a PI3K-dependent manner and diminishes β1 integrin-dependent firm adhesion to endothelial cells by its overexpression, suggesting that ILK in monocytic cells is involved in the regulation of leukocyte adhesion via integrins.

Recently, we have identified an ILK-binding FA protein named affixin, which consists of two tandem calponin homology (CH) domains found in the N-terminal actin-binding domain of α-actinin superfamily proteins (14). Affixin not only interacts with ILK but also with other FA proteins, such as α-actinin and αPIX, and is indispensable for the establishment of attachment and spreading of fibroblasts on ECM (15, 16). Several lines of evidence indicate that it is a dominant downstream target of ILK, which transmits integrin outside-in signals to actin reorganization. Consistently, null mutants of the C. elegans ortholog of affixin PAT-6 (an embryonic phenotype called paralyzed and arrested elongation at the 2-fold stage) cause defects in the assembly of integrin-actin complexes and cell-ECM attachment of body wall muscles, similar to results with null mutants of PAT-4/ILK or PAT-3integrin (8, 17, 18). In mammals, there are three paralogs of affixin, which are now collectively called parvins: actopaxin/CH-ILKBP/α-parvin (α-parvin), affixin/β-parvin (affixin), and γ-parvin (14, 19, 20, 21). Although affixin as well as α-parvin have been extensively investigated in their expression and their functional importance for integrin signaling, the analysis of γ-parvin was limited to the report on the specific expression pattern of its mRNA in human lymphatic tissues (22). Although γ-parvin is the most diverse member of mammalian parvins, the conservation in its molecular organization implies that γ-parvin may associate with ILK and participate in the leukocyte adhesion, spreading, and motility as a dominant downstream target of ILK.

In the present work, we demonstrate that the γ-parvin protein, specifically expressed in several lymphoid/monocytic cell lines and human PBMC, associates with ILK endogenously in U937 cells and directly in vitro. Yeast two-hybrid binding assay, the overexpression of γ-parvin mutants in mammalian cells, and the RNA interference (RNAi) experiments of γ-parvin reveal that γ-parvin binds to and colocalizes with ILK at FA through its second CH domain and is engaged in leukocyte cell adhesion, polarization, and spreading. These results suggest that the ILK-γ-parvin complex is critically involved in the initial steps in leukocyte migration as well as affixin and α-parvin.

Anti-vinculin, anti-actin, anti-α-actinin, anti-PINCH-1, and anti-FLAG monoclonal and polyclonal Abs were obtained from Sigma-Aldrich; anti-ILK Ab was from Upstate Biotechnology; anti-T7 Ab was from Novagen; anti-PTEN and anti-Omni (T7) Abs were from Santa Cruz Biotechnology; anti-hemagglutinin (HA) rat Ab was from Roche; anti-α-parvin Ab was from Abcam; anti-paxillin Ab was from Transduction Laboratories. R-PE-conjugated anti-CD11b, anti-CD49d, and anti-CD49e Abs were from BD Pharmingen. Rhodamine- and FITC-phalloidin were purchased from Molecular Probes and Sigma-Aldrich, respectively. Anti-αPIX Ab was generated in rabbits using GST fusion protein containing αPIX residues 155–545 as an Ag by the method described by Manser et al. (23), and the absence of its cross-reactivity with βPIX was confirmed before use. Anti-affixin and anti-γ-parvin Abs were generated as previously described (14).

Chinese hamster ovary (CHO)-K1 cells were cultured at 37°C in a humidified atmosphere of 5% CO2 in F-12 medium containing 10% FBS, 100 U/ml penicillin, and 100 μg/ml streptomycin. COS-7 cells and U937 cells were cultured under the same conditions as those for CHO-K1 cells, except for the use of DMEM and RPMI 1640 medium, respectively, instead of the F-12 medium. U937 tet-on cells stably expressing γ-parvin or its deletion mutants were obtained by transfecting cells with pOSTet14MCS carrying the appropriate cDNA and selecting G418-resistant cells by incubating with 700 μg/ml G418 disulfate salt (Sigma-Aldrich) for 3 days without further cloning. Cells were maintained by incubating with 400 μg/ml G418.

For CHO-K1 cells, cDNA transfection was performed by lipofection using Polyfect (Qiagen) for immunoprecipitation assays or Fugene 6 transfection reagent (Roche) for immunofluorescence analyses. For COS-7 and U937 cells, we performed electroporation with Gene Pulser II (Bio-Rad) or Nucleofector II (Amaxa).

Full-length γ-Parvin was subcloned into pGEX-6P-1 to produce GST-tagged fusion proteins (GST-FL). GST-tagged full-length, induced in Escherichia coli with isopropyl β-d-thiogalactopyranoside, was purified with glutathione-Sepharose 4B beads (Amersham Biosciences). The purified proteins were dialyzed against the appropriate buffers before use.

We used the TNT T7 Quick Coupled Transcription/Translation System (Promega) to obtain the ILK protein mixture. Fifty microliters of mixture, containing pcDNA3.1 (Invitrogen Life Technologies) carrying the full-length ILK cDNA or pcDNA3.1 vector as a control (1 μg), methionine (20 μM), and TNT Quick Master Mix (40 μl), was incubated at 30°C for 90 min. The GST- or GST-tagged protein-conjugated glutathione-Sepharose 4B beads were incubated with the reaction mixture in 200 μl of binding buffer (100 mM HEPES (pH 7.5), 40 mM NaCl, 3% Triton X-100, and 0.3% BSA) at 4°C for 3 h. After the extensive washing by a washing buffer (50 mM HEPES (pH 7.5), 20 mM NaCl, 1% Triton X-100, and 0.1% BSA), GST alone or the GST full-length was precipitated with the beads, and coprecipitated ILK was detected by Western blotting.

In overexpression experiments, CHO-K1 or COS-7 cells transfected with the expression plasmids were lysed with the buffer described with the figures. After clarification by centrifugation at 20,000 × g for 30 min, these lysates were incubated with the protein G-Sepharose (Amersham Biosciences) conjugated with 2 μg of anti-T7 or anti-FLAG Ab for 1 h at 4°C. After extensive washing with the lysis buffer, the immunocomplexes were solubilized by adding SDS sample buffer to the resin and subjected to Western blot analysis using a chemiluminescence ECL system (Amersham Biosciences). For the analyses of endogenous interactions, ∼3 × 108 U937 cells or 1 × 108 PBMC were suspended in 1000 μl of the lysis buffer, and the lysates were incubated with the protein G-Sepharose conjugated with 50 μg of affinity-purified anti-γ-parvin Ab or control normal rabbit IgG for 4 h at 4°C.

U937 cells or those transfected with expression plasmids were cultured on human FN-coated coverslips in the medium containing 100 nM PMA for 48 h. After washing with PBS, they were fixed with 2% paraformaldehyde in PBS for 15 min at room temperature and then permeabilized with PBS containing 0.1% Triton X-100 for 15 min. In the staining with anti-ILK Ab (see Fig. 4), cells were fixed with 100% methanol. After blocking, the cells were treated with appropriate primary Abs for 45 min at 37°C in a moist chamber, washed with PBS containing 0.05% Tween 20, and incubated with secondary Abs (Cy3-linked goat anti-rabbit IgG (Amersham Biosciences) and Alexa Fluor 488 goat anti-mouse, anti-rabbit, or anti-rat IgG Ab (Molecular Probes)). After washing, the samples were observed under a fluorescence microscope (BX50; Olympus) equipped with a cooled CCD camera (Photometrics). Differential interference contrast images were collected using an Axio Imager Z1 microscope (Carl Zeiss).

FIGURE 4.

Endogenous γ-parvin shows colocalization at FA with ILK and concentrates at the leading edge during early spreading phase. A, Immunofluorescence staining of PMA-stimulated well-spread U937 cells. Cells were double stained with anti-γ-parvin Ab and anti-vinculin Ab (a–c), anti-ILK Ab (g–i), or rhodamine-phalloidin (d–f). Note that endogenous γ-parvin colocalizes with vinculin at FA (a–c, arrowheads), which correspond to the both ends of the actin stress fiber (d–f, arrows) and that γ-parvin is also colocalized with ILK at the tip of the leading edge (g–i, arrowheads). Intense signals from nuclei or perinuclear regions represent nonspecific staining depending on the fixation conditions. Scale bar, 10 μm. B, Immunofluorescence stainings for U937 cells during the early stage of cell spreading were performed using anti-γ-parvin Ab (b, g, and l). Cells were stained simultaneously with rhodamine-phalloidin (c), anti-PTEN Ab (h), or anti-ILK Ab (m). Differential interference contrast (DIC) images are shown (a, f, and k). Fluorescence intensities calculated along lines (d, i, and n) drawn from anterior (A) to posterior (P) are shown (e, j, and o) using Axio Vision 4.2 software (Carl Zeiss). Note that γ-parvin concentrates on one side of the cell periphery with F-actin and ILK, which is the opposite side of PTEN localization. Scale bar, 10 μm.

FIGURE 4.

Endogenous γ-parvin shows colocalization at FA with ILK and concentrates at the leading edge during early spreading phase. A, Immunofluorescence staining of PMA-stimulated well-spread U937 cells. Cells were double stained with anti-γ-parvin Ab and anti-vinculin Ab (a–c), anti-ILK Ab (g–i), or rhodamine-phalloidin (d–f). Note that endogenous γ-parvin colocalizes with vinculin at FA (a–c, arrowheads), which correspond to the both ends of the actin stress fiber (d–f, arrows) and that γ-parvin is also colocalized with ILK at the tip of the leading edge (g–i, arrowheads). Intense signals from nuclei or perinuclear regions represent nonspecific staining depending on the fixation conditions. Scale bar, 10 μm. B, Immunofluorescence stainings for U937 cells during the early stage of cell spreading were performed using anti-γ-parvin Ab (b, g, and l). Cells were stained simultaneously with rhodamine-phalloidin (c), anti-PTEN Ab (h), or anti-ILK Ab (m). Differential interference contrast (DIC) images are shown (a, f, and k). Fluorescence intensities calculated along lines (d, i, and n) drawn from anterior (A) to posterior (P) are shown (e, j, and o) using Axio Vision 4.2 software (Carl Zeiss). Note that γ-parvin concentrates on one side of the cell periphery with F-actin and ILK, which is the opposite side of PTEN localization. Scale bar, 10 μm.

Close modal

U937 tet-on cells and parental U937 cells incubated with or without 100 ng/ml doxycycline for 24 h were cultured on human FN-coated coverslips with 100 nM PMA for 48 h, and then immunofluorescence staining was performed. The mean size of the cells was quantified by analyzing 50 cells from randomly selected fields using Scion Image software. For adhesion assays, the tet-on cells were cultured on FN-coated 24- or 48-well plates with or without 50 nM recombinant human MCP-1 (R&D Systems). The wells were gently washed with PBS twice, and the adherent cells were counted under a microscope.

U937 tet-on cells incubated with 100 ng/ml doxycycline for 24 h and parental U937 cells were washed with PBS containing 0.5% BSA and 20 mM EDTA, blocked with human γ-globulin, and incubated with PE-conjugated Abs or control IgG for 30 min on ice. After washing with PBS containing 0.5% BSA and 20 mM EDTA, the cells were fixed with 2% paraformaldehyde in PBS and analyzed by FACScan with CellQuest Pro program (BD Biosciences).

Before performing actin cosedimentation assays, all samples were centrifuged in a Beckman Coulter airfuge (150,000 × g) for 1 h at 4°C to eliminate protein aggregates. F-actin was prepared by polymerization of G-actin (Cytoskeleton) in the presence of F-actin buffer (15 mM Tris-HCl (pH 8.0), 50 mM KCl, 0.2 mM CaCl2, 2 mM MgCl2, and 1 mM ATP) for 1 h at room temperature. F-actin (19 μM) was incubated with GST-γ-parvin (9 μM), α-actinin (2 μM; Cytoskeleton), GST (1 μM), or BSA (3 μM) for 30 min at room temperature. After incubation, the samples were centrifuged at 150,000 × g for 1.5 h at 24°C to pellet the F-actin with associated proteins. The supernatant and pellet fractions were resolved by 10% SDS-PAGE and detected by Coomassie brilliant blue stain.

To establish γ-parvin-depleted U937 cells, the EBV-based expression vectors, pEB6-Super encoding short hairpin RNA (shRNA) sequence for γ-parvin RNAi or a scramble sequence (Dharmacon), were transfected by electroporation and the cells were selected using RPMI 1640 containing 800 μg/ml G418 disulfate salt (24). RNAi target sequences for γ-parvin (no. 1) 5′-GGACGTCTTTGATGAATTA-3′ and (no. 4) 5′-CAGAAAT GCTGCACAACGT-3′ were used in these experiments.

Recently, we have reported the identification and characterization of affixin, which plays an essential role in cell-substrate interaction as an interface between integrin and cytoskeleton (14, 16). During the course of these studies, we identified two other kinds of proteins, which exhibit significant overall homology with affixin and were later revealed to correspond to α- and γ-parvin (21). In contrast with affixin (β-parvin) and α-parvin (also called actopaxin or CH-ILKBP), functional characterization of γ-parvin including its interaction with ILK has not been described (14, 19, 20, 21). We report a comprehensive analysis on this parvin member.

The lymphoid tissue-specific parvin, γ-parvin, was identified as an ILK-interacted protein in yeast two-hybrid screening against human bone marrow and fetal liver cDNA libraries using full-length ILK as bait. The screening isolated three positive clones, which involved full-length γ-Parvin (331 aa, and the predicted molecular mass is 37.5 kDa) and C-terminal portion of γ-Parvin (BM7; aa 181–331). On the basis of the obtained positive clones, the full-length cDNA of human γ-parvin was subcloned into pcDNA4 vector with T7 tag and its expression confirmed in CHO-K1 cells. To investigate the expression profile of the γ-parvin protein, we generated an affinity-purified rabbit polyclonal Ab against full-length γ-parvin. As shown in Fig. 1,A, the anti-γ-parvin Ab reacted with the γ-parvin overexpressed in CHO-K1 cells, but not with the α-parvin and affixin. This Ab detected endogenous γ-parvin in lysates from several cell lines as a single band with a molecular mass of 40.4 kDa, which agrees with that of overexpressed protein (Fig. 1 B; compare lane 1 with the others).

FIGURE 1.

Specificity of anti-γ-parvin Ab and expression of γ-parvin protein in various cell lines. A, Anti-γ-parvin Ab specifically recognizes T7-tagged γ-parvin. Total lysates of CHO-K1 cells transfected with vector alone, T7-tagged α-parvin, l-affixin, and γ-parvin were subjected to 10% SDS-PAGE, followed by immunoblotting with anti-γ-parvin (top) or anti-T7 Ab (bottom). B, γ-Parvin is preferentially expressed in mononuclear cell lines. The total lysates of various cell lines and CHO-K1 cells overexpressed with γ-parvin as a marker (lane 1) were probed with anti-γ-parvin Ab (top). The same blot was stripped and reprobed with anti-actin Ab to indicate relative loading amounts (bottom). C, Expression of α-parvin and affixin in leukocyte cell lines. The total lysates of the cell lines were loaded and subjected to immunoblotting with anti-affixin Ab (middle). The same blot was stripped and reprobed with anti-α-parvin Ab (upper) or anti-actin Ab (lower) to show relative expression and loading amounts, respectively. D, γ-Parvin is expressed in human PBMC. The lysates from human leukocyte subsets were prepared by Ficoll-Hypaque density gradient centrifugation from healthy human volunteers with informed consent, and subjected to SDS-PAGE and Western blot with anti-γ-parvin, anti-affixin, and anti-actin Abs. The endogenous expressions of γ-parvin and affixin were confirmed in human PBMC.

FIGURE 1.

Specificity of anti-γ-parvin Ab and expression of γ-parvin protein in various cell lines. A, Anti-γ-parvin Ab specifically recognizes T7-tagged γ-parvin. Total lysates of CHO-K1 cells transfected with vector alone, T7-tagged α-parvin, l-affixin, and γ-parvin were subjected to 10% SDS-PAGE, followed by immunoblotting with anti-γ-parvin (top) or anti-T7 Ab (bottom). B, γ-Parvin is preferentially expressed in mononuclear cell lines. The total lysates of various cell lines and CHO-K1 cells overexpressed with γ-parvin as a marker (lane 1) were probed with anti-γ-parvin Ab (top). The same blot was stripped and reprobed with anti-actin Ab to indicate relative loading amounts (bottom). C, Expression of α-parvin and affixin in leukocyte cell lines. The total lysates of the cell lines were loaded and subjected to immunoblotting with anti-affixin Ab (middle). The same blot was stripped and reprobed with anti-α-parvin Ab (upper) or anti-actin Ab (lower) to show relative expression and loading amounts, respectively. D, γ-Parvin is expressed in human PBMC. The lysates from human leukocyte subsets were prepared by Ficoll-Hypaque density gradient centrifugation from healthy human volunteers with informed consent, and subjected to SDS-PAGE and Western blot with anti-γ-parvin, anti-affixin, and anti-actin Abs. The endogenous expressions of γ-parvin and affixin were confirmed in human PBMC.

Close modal

Consistent with the previous report by Korenbaum et al. (22) that the γ-parvin mRNA is predominantly expressed in human lymphatic tissues as determined by Northern blot analysis, the expression of the γ-parvin protein was solely detected in cell lines derived from monocytes (U937 and THP-1) and a T cell subset (Jurkat) and not significantly detected in other cell lines, including fibroblasts (IMR90, NIH3T3, and CHO-K1), epitheliocytes (MDCK, HeLa, and 293), and neurocytes (PC-12), or in human platelets (Fig. 1,B). Interestingly, this expression pattern was highly complement with that of affixin. In contrast with γ-parvin, affixin was expressed in various cell lines including fibroblasts and epitheliocytes (data not shown). In addition, as shown in Fig. 1,C, affixin was expressed in leukocyte cell lines, such as Raji and K562, which were devoid of γ-parvin expression, but not in cell lines in which γ-parvin was predominantly expressed (Jurkat, U937, and THP-1). α-Parvin showed little expression in the leukocyte cell lines compared with in HeLa cells (Fig. 1,C). The endogenous expressions of γ-parvin and affixin were also confirmed in human PBMC, which includes lymphocytes and monocytes (Fig. 1 D). These results suggest that γ-parvin is a member of the parvin family that predominantly functions in leukocytes.

Among the positive clones we previously obtained from the yeast two-hybrid screening using full-length ILK as bait, the shortest cDNA fragment of γ-parvin (BM7), corresponded to the C-terminal half of the molecule (Fig. 2,A), implying that γ-parvin associates with ILK through the CH2 domain similar to affixin and α-parvin (14, 20). Subsequent yeast two-hybrid assays supported this speculation by revealing that the whole part of the CH2 domain (aa 209–311) is necessary and sufficient for interacting with the full-length ILK (Fig. 2,A). These results were then supplemented by coimmunoprecipitation assay from CHO-K1 cells cotransfected with T7-tagged γ-parvin, or its deletion mutants, and FLAG-ILK expression vectors. Full-length γ-parvin and its C-terminal fragment mainly containing the CH2 domain (CH2), but not the N-terminal fragment mainly containing the CH1 domain (CH1), were coprecipitated with ILK (Fig. 2,B). Additionally, the reciprocal immunoprecipitation assay using anti-FLAG Ab revealed that T7 full-length γ-parvin interacted with FLAG-tagged wild-type (wt) ILK and a kinase-deficient mutant of ILK (K220M) but not with a mutant in the activation loop of the kinase domain, E359K, implying that the activation loop and not the kinase activity of ILK is essential for their binding as well as for affixin and α-parvin (Fig. 2,C) (14, 25). We next examined whether the interaction between γ-parvin and ILK is direct by incubating bacterially expressed GST-tagged full-length γ-parvin with in vitro translated full-length ILK. Fig. 2 D shows the specific association of ILK with GST full-length γ-parvin, but not with GST alone, indicating that γ-parvin directly interacts with ILK.

FIGURE 2.

γ-Parvin interacts with ILK through CH2 domain. A, Domain structure of γ-parvin and mapping of ILK-binding region on γ-parvin molecule by yeast two-hybrid assays. The full-length human γ-parvin cDNA (FL; aa 1–331) and two truncated mutants CH1 and CH2 (corresponding to aa 1–169 and aa 170–331, respectively) were subcloned into each expression vector as indicated. BM7 (aa 181–331) was isolated with full-length human γ-parvin cDNA in the yeast two-hybrid screening against human bone marrow and fetal liver cDNA libraries using full-length ILK as bait (top). In the yeast two-hybrid binding assays, the cDNA fragments of γ-parvin and full-length ILK (wt ILK) were subcloned into pGAD424 and pAS2–1C vectors, respectively (bottom). These vectors were cotransformed into yeast Y187(a), and 4 days later the interactions were examined by β-galactosidase filter assays as previously described (14 ). B, Overexpressed full-length human γ-parvin or CH2 γ-parvin is coimmunoprecipitated with ILK in CHO-K1 cells. The T7-tagged full-length, CH1, and CH2 γ-parvin or the control vector was overexpressed with FLAG-tagged wt ILK into CHO-K1 cells. The immunoprecipitates using anti-T7 Ab were subjected to immunoblotting with anti-T7 or anti-FLAG Abs. C, The point mutant in the activation loop of ILK is unable to bind to γ-parvin. FLAG-wt ILK, an ILK point mutant in which glutamate 359 in the activation loop is replaced with lysine (E359K), or another ILK point mutant in which lysine 220 in the ATP-binding site of the kinase domain is substituted with methionine (K220M), or the control vector was overexpressed into CHO-K1 cells with T7 full-length γ-parvin. The immunoprecipitates using anti-FLAG Ab were subjected to immunoblotting with anti-T7 or anti-FLAG Abs. D, γ-Parvin binds directly to ILK in vitro. GST-FL or GST alone was incubated with ILK or control mixture produced by in vitro translation. Coprecipitated proteins were collected using glutathione-Sepharose, resolved by SDS-PAGE, and transferred to a polyvinylidene difluoride membrane (W.B., Western blot). The membrane was probed with anti-ILK Ab (top) and stained with Coomassie brilliant blue (CBB) to confirm the amount of GST or the GST fusion proteins (bottom). E, Coimmunoprecipitation of ILK with endogenous γ-parvin in vivo. The lysates from U937 cells incubated in the presence or the absence of PMA for 24 h (a), or human PBMC from healthy donors were immunoprecipitated with control IgG or the anti-γ-parvin Ab (b), and subjected to SDS-PAGE followed by Western blotting with anti-ILK, anti-γ-parvin, and anti-actin Abs. The negative interaction of actin shows that the interaction is not the result of contamination of large cytoskeletal pellets. The bands of IgG L chain (∗) are indicated. B, C, and E, Cells were lysed in the lysis buffer containing 50 mM HEPES (pH 7.5), 150 mM NaCl, 1 mM EDTA, 10 μg/ml leupeptin, 1 mM PMSF, and 1.5% Triton X-100.

FIGURE 2.

γ-Parvin interacts with ILK through CH2 domain. A, Domain structure of γ-parvin and mapping of ILK-binding region on γ-parvin molecule by yeast two-hybrid assays. The full-length human γ-parvin cDNA (FL; aa 1–331) and two truncated mutants CH1 and CH2 (corresponding to aa 1–169 and aa 170–331, respectively) were subcloned into each expression vector as indicated. BM7 (aa 181–331) was isolated with full-length human γ-parvin cDNA in the yeast two-hybrid screening against human bone marrow and fetal liver cDNA libraries using full-length ILK as bait (top). In the yeast two-hybrid binding assays, the cDNA fragments of γ-parvin and full-length ILK (wt ILK) were subcloned into pGAD424 and pAS2–1C vectors, respectively (bottom). These vectors were cotransformed into yeast Y187(a), and 4 days later the interactions were examined by β-galactosidase filter assays as previously described (14 ). B, Overexpressed full-length human γ-parvin or CH2 γ-parvin is coimmunoprecipitated with ILK in CHO-K1 cells. The T7-tagged full-length, CH1, and CH2 γ-parvin or the control vector was overexpressed with FLAG-tagged wt ILK into CHO-K1 cells. The immunoprecipitates using anti-T7 Ab were subjected to immunoblotting with anti-T7 or anti-FLAG Abs. C, The point mutant in the activation loop of ILK is unable to bind to γ-parvin. FLAG-wt ILK, an ILK point mutant in which glutamate 359 in the activation loop is replaced with lysine (E359K), or another ILK point mutant in which lysine 220 in the ATP-binding site of the kinase domain is substituted with methionine (K220M), or the control vector was overexpressed into CHO-K1 cells with T7 full-length γ-parvin. The immunoprecipitates using anti-FLAG Ab were subjected to immunoblotting with anti-T7 or anti-FLAG Abs. D, γ-Parvin binds directly to ILK in vitro. GST-FL or GST alone was incubated with ILK or control mixture produced by in vitro translation. Coprecipitated proteins were collected using glutathione-Sepharose, resolved by SDS-PAGE, and transferred to a polyvinylidene difluoride membrane (W.B., Western blot). The membrane was probed with anti-ILK Ab (top) and stained with Coomassie brilliant blue (CBB) to confirm the amount of GST or the GST fusion proteins (bottom). E, Coimmunoprecipitation of ILK with endogenous γ-parvin in vivo. The lysates from U937 cells incubated in the presence or the absence of PMA for 24 h (a), or human PBMC from healthy donors were immunoprecipitated with control IgG or the anti-γ-parvin Ab (b), and subjected to SDS-PAGE followed by Western blotting with anti-ILK, anti-γ-parvin, and anti-actin Abs. The negative interaction of actin shows that the interaction is not the result of contamination of large cytoskeletal pellets. The bands of IgG L chain (∗) are indicated. B, C, and E, Cells were lysed in the lysis buffer containing 50 mM HEPES (pH 7.5), 150 mM NaCl, 1 mM EDTA, 10 μg/ml leupeptin, 1 mM PMSF, and 1.5% Triton X-100.

Close modal

In Fig. 2,E, the endogenous interaction between γ-parvin and ILK was assessed in U937 human promonocytic cells and human PBMC from healthy donors. Because their interaction may be dependent on cell-substrate interaction, we immunoprecipitated γ-parvin from cells with or without PMA stimulation, which enhances the ability of U937 cells to bind to FN and causes cell spreading on substrates (26, 27). ILK was coprecipitated with γ-parvin using anti-γ-parvin Ab in U937 cells (Fig. 2,Ea) and human PBMC (Fig. 2,Eb), and the amount of the precipitated ILK was not dependent on PMA stimulation (Fig. 2 Ea). Therefore, these results suggest that γ-parvin endogenously interacts with ILK in leukocytes in an adhesion-independent manner.

Previous studies revealed that affixin and α-parvin are localized at FA, where actin stress fibers and ECM are connected through accumulated integrins, and affixin is involved in the initial formation of FA with ILK (14, 20). To examine the localization of γ-parvin at FA in leukocytes, we overexpressed GFP-tagged full-length γ-parvin (GFP-FL) or its deletion mutants (GFP-CH1, GFP-CH2) in PMA-stimulated U937 cells. As shown in Fig. 3, GFP-tagged full-length γ-parvin and GFP-CH2 were localized at FA identified by vinculin staining (Fig. 3, A–C and G–I; arrowheads), although most GFP-CH2-expressing cells tended to show small round morphology with a diffuse distribution of GFP signals within a whole cell body (see below). Together with the result that GFP-CH1 did not localize at FA (Fig. 3, D–F; arrowheads), these results suggest that γ-parvin associates with FA by way of its CH2 domain. To confirm the endogenous localization of γ-parvin in PMA-stimulated U937 cells, we examined U937 cells spread on FN-coated coverslips stained with anti-γ-parvin Ab. As shown in Fig. 4,A, γ-parvin staining was localized at vinculin-positive FA (Fig. 4,A, a–c; arrowheads) from which actin stress fibers emanate (Fig. 4,A, d–f; arrows). ILK was colocalized with γ-parvin at FA (data not shown) and concentrated on the cell surface blebs with ILK during the early stages of cell spreading (Fig. 4 A, g–i; arrowheads). These results are well consistent with those obtained from affixin and α-parvin in epithelial or fibroblast cells, suggesting functional similarity of these proteins (14, 20).

FIGURE 3.

γ-Parvin localizes at FA through CH2 domain. Immunofluorescence staining of U937 cells overexpressed with GFP-tagged full-length γ-parvin (GFP-FL) (A–C) or its deletion mutants (GFP-CH1, D–F or GFP-CH2, G–I). Cells were stained with anti-vinculin Ab as an FA marker (B, E, and H). Note that GFP-tagged full-length γ-parvin and GFP-CH2 colocalize with vinculin at FA (arrowheads). Intense signals from nuclei or perinuclear regions represent nonspecific staining depending on the fixation conditions. Scale bar, 10 μm.

FIGURE 3.

γ-Parvin localizes at FA through CH2 domain. Immunofluorescence staining of U937 cells overexpressed with GFP-tagged full-length γ-parvin (GFP-FL) (A–C) or its deletion mutants (GFP-CH1, D–F or GFP-CH2, G–I). Cells were stained with anti-vinculin Ab as an FA marker (B, E, and H). Note that GFP-tagged full-length γ-parvin and GFP-CH2 colocalize with vinculin at FA (arrowheads). Intense signals from nuclei or perinuclear regions represent nonspecific staining depending on the fixation conditions. Scale bar, 10 μm.

Close modal

Migration requires the induction of a distinct polarized cell morphology driven by integrins and cytoskeletons to form a leading edge, which later develops into a lamellipodium (28). In migrating leukocytes, F-actin and various receptors are concentrated on the leading edge, whereas several molecules, such as PTEN, accumulate at the trailing edge, resulting in the polarization of the leukocytes (29, 30). We found that PMA-stimulated U937 cells plated on FN also became polarized with a well-defined leading edge during the early spreading phase. As shown in Fig. 4,B, γ-parvin was concentrated mainly at the leading edge with ILK (Fig. 4,B, k–o) and F-actin (Fig. 4,B, a–e), whereas PTEN was localized at the opposite edge (Fig. 4 B, f–j).

Previously, we reported that the overexpression of the CH1 domain of affixin augments cell spreading after replating on FN, whereas the overexpression of its CH2 domain completely inhibits cell spreading after replating (14). Subsequent works have suggested that the CH1 domain facilitates cell spreading by constitutively activating Rac1 or Cdc42 through its binding with αPIX/ARHGEF6, whereas the CH2 domain acts as a dominant negative mutant that competes with endogenous affixin for ILK and/or α-actinin binding (15, 16). To examine the role of γ-parvin in cell-substrate interaction of monocyte, we established U937 cells that stably expressed T7-tagged full-length, CH1, or CH2 γ-parvin under the control of tetracycline-inducible transactivation. In these cells, the addition of 100 ng/ml doxycycline for 24 h significantly induced the expression of T7 full-length, T7-CH1, or T7-CH2 (Fig. 5,A). To investigate the effect of γ-parvin mutants on the surface expressions of integrins, the cells overexpressed with each γ-parvin mutant were subjected to flow cytometric analysis. As a result, the surface expression levels of α4β1, α5β1, and αMβ2 integrins were not affected by their overexpression, and the surface expression of CD11b (integrin αM), also known as a myelomonocytic differentiation marker (31), indicated that the monocytic differentiation of U937 cells was not influenced (Fig. 5 B).

FIGURE 5.

Expression of γ-parvin mutants modulates MCP-1-induced cell adhesion on FN. A, Western blot analysis showing relative expression levels of each construct in tet-on U937 cells. The cells were incubated in the absence (lanes 1, 3, and 5) or the presence (lanes 2, 4, and 6) of 100 ng/ml doxycycline (DC) for 24 h. The total lysates were analyzed using anti-T7 Ab (top). The same blot was stripped and reprobed with anti-actin Ab to indicate relative loading among lanes (bottom). B, Surface expression of integrins on U937 cells overexpressed with γ-parvin mutants. Parental U937 cells (a) or U937 cells overexpressed with full-length (FL; b), CH1 (c), or CH2 (d) γ-parvin by doxycycline induction were subjected to flow cytometric study using PE-labeled mAbs to CD11b (integrin αM; β2 integrin), CD49d (integrin α4; β1 integrin), and CD49e (integrin α5; β1 integrin), and PE-labeled mouse IgG1 as a control. Note that the overexpression of γ-parvin mutants has no effect on each integrin expression. C, The adherent cell counts of tet-on or tet-off U937 cells incubated with or without MCP-1 in cell adhesion assays. The parental (P) or tet-on U937 cells (1 × 105 cells per well) were incubated with or without doxycycline for 24 h and then MCP-1 for 2 h on FN-coated 24-well plates. After gently washing, the mean adherent cell numbers in three independent fields were calculated. Note that MCP-1 promotes U937 cell adhesion to FN and CH1 induction further augments cell adhesion, but CH2 overexpression completely blocks the effect of MCP-1. NS, Not significant. D, β1 Integrin blocking Ab entirely suppresses the effect of CH1 overexpression. Before the cell adhesion assay on FN-coated 48-well plates, 3 × 104 tet-on U937 cells overexpressed with CH1, CH2, or the control U937 cells (vector alone; Vec) were incubated for 30 min with or without the β1 integrin blocking Ab, P4C10 (Chemicon International) (36 ). C and D, Data represent the mean ± SD of three independent separate experiments. Statistically significance data (∗, p < 0.05) by Student’s t test, respectively.

FIGURE 5.

Expression of γ-parvin mutants modulates MCP-1-induced cell adhesion on FN. A, Western blot analysis showing relative expression levels of each construct in tet-on U937 cells. The cells were incubated in the absence (lanes 1, 3, and 5) or the presence (lanes 2, 4, and 6) of 100 ng/ml doxycycline (DC) for 24 h. The total lysates were analyzed using anti-T7 Ab (top). The same blot was stripped and reprobed with anti-actin Ab to indicate relative loading among lanes (bottom). B, Surface expression of integrins on U937 cells overexpressed with γ-parvin mutants. Parental U937 cells (a) or U937 cells overexpressed with full-length (FL; b), CH1 (c), or CH2 (d) γ-parvin by doxycycline induction were subjected to flow cytometric study using PE-labeled mAbs to CD11b (integrin αM; β2 integrin), CD49d (integrin α4; β1 integrin), and CD49e (integrin α5; β1 integrin), and PE-labeled mouse IgG1 as a control. Note that the overexpression of γ-parvin mutants has no effect on each integrin expression. C, The adherent cell counts of tet-on or tet-off U937 cells incubated with or without MCP-1 in cell adhesion assays. The parental (P) or tet-on U937 cells (1 × 105 cells per well) were incubated with or without doxycycline for 24 h and then MCP-1 for 2 h on FN-coated 24-well plates. After gently washing, the mean adherent cell numbers in three independent fields were calculated. Note that MCP-1 promotes U937 cell adhesion to FN and CH1 induction further augments cell adhesion, but CH2 overexpression completely blocks the effect of MCP-1. NS, Not significant. D, β1 Integrin blocking Ab entirely suppresses the effect of CH1 overexpression. Before the cell adhesion assay on FN-coated 48-well plates, 3 × 104 tet-on U937 cells overexpressed with CH1, CH2, or the control U937 cells (vector alone; Vec) were incubated for 30 min with or without the β1 integrin blocking Ab, P4C10 (Chemicon International) (36 ). C and D, Data represent the mean ± SD of three independent separate experiments. Statistically significance data (∗, p < 0.05) by Student’s t test, respectively.

Close modal

Using these stably expressed U937 cells, we evaluated the involvement of γ-parvin in cell adhesion under MCP-1 stimulation. Previous reports demonstrated that MCP-1, one of the β-chemokines, plays a critical role in the accumulation of macrophages in inflamed tissues by modifying their adhesive activity to VCAM-1 and FN, and suggested that MCP-1 influences directed migration (32, 33). Therefore, we examined the effect of MCP-1 treatment on U937 cells that have been shown to attach to FN via α4β1 or α5β1 integrins (34, 35, 36). Cells were allowed to adhere to FN-coated plates with or without MCP-1 (50 nM) for 2 h, and after gentle washing, the adherent cells were counted under a microscope. As shown in Fig. 5,C, the treatment with MCP-1 increased FN adhesion of parental control U937 cells ∼3.5-fold. The overexpression of the CH1 domain markedly increased cell attachment ∼1.7- and 2.0-fold in the presence and the absence of MCP-1 stimulation, respectively. In contrast, the induction of the CH2 domain overexpression completely suppressed MCP-1-induced enhancement of cell adhesion. These results did not change even when we prolonged the incubation time with MCP-1 up to 12 h (data not shown), indicating that γ-parvin mutants affected the adhesive property of the cells, but not their rate of attachment. In contrast, these reciprocal effects of the CH1 and CH2 overexpression on cell adhesion were already seen at 20 min during the MCP-1 induction (data not shown), suggesting that the γ-parvin mutants affected the very early steps of cell adhesion induced by MCP-1. In addition, as shown in Fig. 5 D, the effect of CH1 induction on cell adhesion was almost entirely blocked by mAb blockade of β1 integrins, indicating again that CH1 γ-parvin mainly affected β1 integrin-mediated cell adhesion system. Taken together with the results of the flow cytometric analyses, γ-parvin could be responsible for cell adhesion not by modulating integrin expression but rather by regulating integrin affinity and avidity.

After the cell adhesion process, migration needs the induction of polarized cell morphology and the development of lamellipodium at the leading edge to spread its cell body (28). To evaluate the involvement of γ-parvin in the cell spreading process next to cell adhesion, we examined the effects of the γ-parvin CH1 domain on PMA-induced enhancement of cell-substrate adhesion. To monitor the effects of γ-parvin overexpression on cell spreading, U937 stable cells plated on FN-coated coverslips were stimulated with 100 nM PMA for 48 h to make them spread on the substrate after preincubation with or without doxycycline for 24 h. The expression of γ-parvin or its mutant proteins did not affect the surface expressions of α4β1 and α5β1 integrins in comparison with parental U937 cells even under the PMA stimulation (Fig. 6,A). In contrast, by the function of PMA as a potent inducer of monocyte-macrophage differentiation, the expression of CD11b (integrin αM) was increased as previously reported (31), but the extent of the induction was the same level among the stable cells (Fig. 6,A). As shown in Fig. 6, B and Cc, CH1 overexpression exerted a dramatic effect on cell morphology and induced an elevated cell surface area as compared with the doxycycline-depleted cells. In contrast, CH2 overexpression induced significant change in a cell surface area (Fig. 6,B), and the CH2-overexpressing cells tended to have a blocked cell spread and showed a rounded cell morphology without any cell projections (Fig. 6,Cd). We hardly detected cells with developed FA and stress fibers in these CH2-overexpressing cells, which occupied a certain population after PMA treatment. Instead, the CH2-overexpressing cells exhibited faint and diffuse localizations of vinculin and F-actin as observed in CHO-K1 cells overexpressing the affixin CH2 domain (Fig. 6, Db, Fb, and Fc) (14). Importantly, the polarized distribution of F-actin and PTEN found in the control cells at a very early spreading stage was not observed in the CH2-overexpressing cells (Fig. 6,F; compare b with a). We performed TUNEL and Hoechst 33258 nuclear staining to confirm that the CH2 overexpression did not induce apoptosis within 72 h (Fig. 6,E, e–h). The TUNEL-positive apoptotic cells were <1%, irrespective of the induction for CH2 γ-parvin in U937 cells. Moreover, under the suspension culture without PMA, the cells overexpressing γ-parvin or its deletion mutants did not show any difference in its growth rate within 96 h (data not shown). As shown in Fig. 6 G, the expression levels of endogenous ILK and γ-parvin were not influenced by the induction of CH2 γ-parvin. Taken together, γ-parvin is involved in the regulation of cell adhesion and the spreading process under the induction of chemotaxis.

FIGURE 6.

Overexpression of γ-parvin deletion mutant CH1 or CH2 alters cell spreading in PMA-stimulated U937 cells. A, Surface expression of integrins on PMA-stimulated U937 cells overexpressed with γ-parvin mutants. Parental U937 cells (a) or U937 cells overexpressed with full-length (FL; b), CH1 (c), or CH2 (d) γ-parvin were incubated on FN-coated dishes with 100 nM PMA for 48 h and subjected to flow cytometric experiment. Note that the overexpression of γ-parvin mutants has no effect on the increase of CD11b expression caused by PMA stimulation and on the expression of FN receptors CD49d and CD49e. B, Relative cell sizes of tet-on U937 cells incubated with or without doxycycline (DC) in cell spreading assays. Each mean size was compared with that of the cells transfected with the control vector and incubated without doxycycline. Data represent the mean ± SD of three separate experiments. Statistical significance (∗, p < 0.05) determined by Student’s t test. NS, Not significant. C, The U937 cells overexpressed with T7-tagged full-length (FL; b), CH1 (c), or CH2 (d) γ-parvin and control U937 cells (Vec; a) were stained with FITC-phalloidin (a) or anti-T7 Ab (b–d). Note that cell spreading is promoted with well-developed lamellipodia by CH1 overexpression but inhibited without any cell projection by CH2 overexpression. Scale bar, 50 μm. D, The cells without doxycycline induction (a) or the cells overexpressed with T7-tagged CH2 (b) were stained with anti-vinculin Ab. Note that the CH2-overexpressing cells do not show the typical punctual localization of vinculin at FA, but its diffuse localization at the cell periphery as compared with the doxycycline-depleted cells. Bar, 10 μm. E, CH2 overexpression does not induce apoptosis. The cells overexpressed with T7-tagged CH2 were incubated with (e–h) or without (a–d) FBS as positive control for 48 h, were spread on FN-coated slips by PMA stimulation for 48 h, and triple stained with anti-T7 Ab (a and e), TUNEL reaction mixture (b and f), and Hoechst 33258 (c and g). For TUNEL assay, In situ Apoptosis Detection kit (Takara Bio) was used according to the manufacturer’s instructions. Merged views are demonstrated in d and h for anti-T7 Ab (red), TUNEL-positive (green), and Hoechst 33258 (blue). Scale bar, 10 μm. F, The cells overexpressed with (b and c) or without (a) T7-tagged CH2 were double stained with rhodamine-phalloidin (red) and anti-PTEN Ab (green). Note that the cells overexpressed with CH2 show diffuse distributions of F-actin and PTEN (b), which are also observed during the later phases of cell spreading (c), whereas the cells without CH2 expression reveal the well-polarized localization of F-actin and PTEN (a). Scale bar, 10 μm. G, The overexpression of CH2 γ-parvin does not influence the cellular levels of endogenous γ-parvin, ILK, and PINCH-1 proteins. The total lysates of the cells without doxycycline induction (lane 1) and the cells overexpressed with T7-tagged CH2 by doxycycline induction (lane 2) were analyzed using anti-γ-parvin, anti-ILK, anti-PINCH-1, and anti-T7 Abs.

FIGURE 6.

Overexpression of γ-parvin deletion mutant CH1 or CH2 alters cell spreading in PMA-stimulated U937 cells. A, Surface expression of integrins on PMA-stimulated U937 cells overexpressed with γ-parvin mutants. Parental U937 cells (a) or U937 cells overexpressed with full-length (FL; b), CH1 (c), or CH2 (d) γ-parvin were incubated on FN-coated dishes with 100 nM PMA for 48 h and subjected to flow cytometric experiment. Note that the overexpression of γ-parvin mutants has no effect on the increase of CD11b expression caused by PMA stimulation and on the expression of FN receptors CD49d and CD49e. B, Relative cell sizes of tet-on U937 cells incubated with or without doxycycline (DC) in cell spreading assays. Each mean size was compared with that of the cells transfected with the control vector and incubated without doxycycline. Data represent the mean ± SD of three separate experiments. Statistical significance (∗, p < 0.05) determined by Student’s t test. NS, Not significant. C, The U937 cells overexpressed with T7-tagged full-length (FL; b), CH1 (c), or CH2 (d) γ-parvin and control U937 cells (Vec; a) were stained with FITC-phalloidin (a) or anti-T7 Ab (b–d). Note that cell spreading is promoted with well-developed lamellipodia by CH1 overexpression but inhibited without any cell projection by CH2 overexpression. Scale bar, 50 μm. D, The cells without doxycycline induction (a) or the cells overexpressed with T7-tagged CH2 (b) were stained with anti-vinculin Ab. Note that the CH2-overexpressing cells do not show the typical punctual localization of vinculin at FA, but its diffuse localization at the cell periphery as compared with the doxycycline-depleted cells. Bar, 10 μm. E, CH2 overexpression does not induce apoptosis. The cells overexpressed with T7-tagged CH2 were incubated with (e–h) or without (a–d) FBS as positive control for 48 h, were spread on FN-coated slips by PMA stimulation for 48 h, and triple stained with anti-T7 Ab (a and e), TUNEL reaction mixture (b and f), and Hoechst 33258 (c and g). For TUNEL assay, In situ Apoptosis Detection kit (Takara Bio) was used according to the manufacturer’s instructions. Merged views are demonstrated in d and h for anti-T7 Ab (red), TUNEL-positive (green), and Hoechst 33258 (blue). Scale bar, 10 μm. F, The cells overexpressed with (b and c) or without (a) T7-tagged CH2 were double stained with rhodamine-phalloidin (red) and anti-PTEN Ab (green). Note that the cells overexpressed with CH2 show diffuse distributions of F-actin and PTEN (b), which are also observed during the later phases of cell spreading (c), whereas the cells without CH2 expression reveal the well-polarized localization of F-actin and PTEN (a). Scale bar, 10 μm. G, The overexpression of CH2 γ-parvin does not influence the cellular levels of endogenous γ-parvin, ILK, and PINCH-1 proteins. The total lysates of the cells without doxycycline induction (lane 1) and the cells overexpressed with T7-tagged CH2 by doxycycline induction (lane 2) were analyzed using anti-γ-parvin, anti-ILK, anti-PINCH-1, and anti-T7 Abs.

Close modal

The previous studies suggested that the affixin CH1 domain binds to αPIX and induces the guanine nucleotide exchange factor (GEF) activity of αPIX, which leads to the activation of Cdc42 and PAK, F-actin reorganization, and enhanced cell spreading (15, 37). Thus, we examined the interaction between γ-parvin and αPIX and found that, as shown for affixin, HA-αPIX was efficiently coprecipitated with CH1 γ-parvin (Fig. 7,Aa) and more weakly with full-length γ-parvin that could not be detected without longer exposure (data not shown). The interaction was confirmed endogenously by analyzing anti-γ-parvin immunoprecipitate from U937 cells (Fig. 7,Ab). To examine whether the effect of γ-parvin CH1 overexpression is dependent on the GEF activity of αPIX, the stable CH1-overexpressing cells were transfected with wt αPIX or a GEF activity-deficient (ΔGEF) mutant αPIX (L383R, L384S), and then plated on the FN-coated coverslips under the stimulation of PMA. The cell surface area was estimated by the measurements of αPIX-positive cells. As shown in Fig. 7, B and C, the coexpression of GEF activity-deficient αPIX, but not of wt αPIX, significantly suppressed the effect of CH1 overexpression, and showed small round morphology with few membrane protrusions (Fig. 7, Cd). These results indicate that the CH1-induced enhancement of cell spreading requires the GEF activity of αPIX, which mediates the activation of Cdc42 and results in PAK activation.

FIGURE 7.

γ-Parvin interacts with αPIX and its GEF activity is essential for the CH1-induced enhancement of cell spreading. A, Overexpressed CH1 γ-parvin is coimmunoprecipitated with αPIX in CHO-K1 cells (a). CHO-K1 cells overexpressed with HA-tagged αPIX and the T7-tagged full-length (FL), CH1 γ-parvin, the CH1 affixin, or the control vector were lysed in lysis buffer containing 50 mM HEPES (pH 7.5), 100 mM NaCl, 10 μg/ml leupeptin, 1 mM PMSF, 2 mM NaF, and 0.75% Triton X-100. The immunoprecipitates using anti-T7 Ab were subjected to immunoblotting with anti-T7 or anti-HA Abs. Although HA-αPIX is preferentially coimmunoprecipitated with T7-CH1 but not with T7 full-length γ-parvin, very slight amount of coprecipitated full-length γ-parvin was detected in this immunocomplex in the long exposure (data not shown). Bands of IgG L chain (∗) are displayed. Coimmunoprecipitation of αPIX with endogenous γ-parvin in vivo (b). U937 cells were lysed in lysis buffer containing 50 mM HEPES (pH 7.5), 100 mM NaCl, 10 μg/ml leupeptin, 1 mM PMSF, 2 mM NaF, 1.5% Triton X-100, 0.5% deoxycholate, and 0.1% SDS. The immunoprecipitates using control IgG or the anti-γ-parvin Ab from the lysate was subjected to SDS-PAGE followed by Western blotting with anti-αPIX and anti-γ-parvin Abs. B, Relative cell sizes of the CH1 tet-on or off U937 cells cotransfected with HA-wt or HA-ΔGEF αPIX. Each mean size was compared with that of cells incubated without doxycycline (DC) and cotransfected with HA-wt αPIX. Data represent the mean ± SD of three separate experiments. Statistical significance (∗, p < 0.05) determined by Student’s t test. NS, Not significant. C, The U937 cells overexpressed with (c and d) or without (a and b) T7-tagged CH1 were coexpressed with HA-wt (a and c) or HA-ΔGEF (b and d) αPIX, and stained with anti-HA Ab. Note that the effect of CH1 overexpression is inhibited by GEF activity-deficient (ΔGEF) αPIX coexpression. Scale bar, 50 μm.

FIGURE 7.

γ-Parvin interacts with αPIX and its GEF activity is essential for the CH1-induced enhancement of cell spreading. A, Overexpressed CH1 γ-parvin is coimmunoprecipitated with αPIX in CHO-K1 cells (a). CHO-K1 cells overexpressed with HA-tagged αPIX and the T7-tagged full-length (FL), CH1 γ-parvin, the CH1 affixin, or the control vector were lysed in lysis buffer containing 50 mM HEPES (pH 7.5), 100 mM NaCl, 10 μg/ml leupeptin, 1 mM PMSF, 2 mM NaF, and 0.75% Triton X-100. The immunoprecipitates using anti-T7 Ab were subjected to immunoblotting with anti-T7 or anti-HA Abs. Although HA-αPIX is preferentially coimmunoprecipitated with T7-CH1 but not with T7 full-length γ-parvin, very slight amount of coprecipitated full-length γ-parvin was detected in this immunocomplex in the long exposure (data not shown). Bands of IgG L chain (∗) are displayed. Coimmunoprecipitation of αPIX with endogenous γ-parvin in vivo (b). U937 cells were lysed in lysis buffer containing 50 mM HEPES (pH 7.5), 100 mM NaCl, 10 μg/ml leupeptin, 1 mM PMSF, 2 mM NaF, 1.5% Triton X-100, 0.5% deoxycholate, and 0.1% SDS. The immunoprecipitates using control IgG or the anti-γ-parvin Ab from the lysate was subjected to SDS-PAGE followed by Western blotting with anti-αPIX and anti-γ-parvin Abs. B, Relative cell sizes of the CH1 tet-on or off U937 cells cotransfected with HA-wt or HA-ΔGEF αPIX. Each mean size was compared with that of cells incubated without doxycycline (DC) and cotransfected with HA-wt αPIX. Data represent the mean ± SD of three separate experiments. Statistical significance (∗, p < 0.05) determined by Student’s t test. NS, Not significant. C, The U937 cells overexpressed with (c and d) or without (a and b) T7-tagged CH1 were coexpressed with HA-wt (a and c) or HA-ΔGEF (b and d) αPIX, and stained with anti-HA Ab. Note that the effect of CH1 overexpression is inhibited by GEF activity-deficient (ΔGEF) αPIX coexpression. Scale bar, 50 μm.

Close modal

To further confirm the physiological role of γ-parvin in cell adhesion and spreading, we established U937 cells that stably express shRNA for γ-parvin RNAi (sequence no. 1) under the control of the polymerase III H1 promoter subcloned in an EBV-based plasmid vector (24). As shown in Fig. 8,A, the expression of γ-parvin was significantly reduced in these cells as compared with the cells transfected with the vectors containing a scramble shRNA sequence. In these cells, the depletion of γ-parvin moderately suppressed the protein level of ILK (Fig. 8,A) and another ILK-interacting protein, PINCH-1 (data not shown). Considering the previous report that the assembly of ternary PINCH-ILK-α-parvin complexes is required to maintain their protein levels in cells, the complex formation of ILK with γ-parvin might be essential to retain its protein level as well as PINCH-1 (38). To further define the effect of the γ-parvin depletion, we performed similar cell adhesion or spreading assays under the MCP-1 or PMA stimulation. Consistent with the results of the CH2 overexpression, the MCP-1-induced cell adhesion of the γ-parvin-depleted cells on FN was greatly reduced to the level of noninduced cells (Fig. 8,B). In a similar way, the γ-parvin-depleted cells revealed significantly decreased PMA-stimulated cell spreading on FN (Fig. 8, C–E). As shown in Fig. 8,F, the depletion of γ-parvin inhibited the formation of vinculin-containing FA. In these γ-parvin-depleted cells, the development of the cell surface blebs and the polarization shown by PTEN/F-actin double staining were significantly blocked, suggesting that γ-parvin is required for the cell-substrate interaction during the early spreading phase (Fig. 8, D, F, and G). We also confirmed the similar phenotype in the cells transiently transfected with small interfering RNA encoding another sequence (no. 4) for γ-parvin RNAi, to exclude the possibility of off-target effects (data not shown). Taken together, these results not only support the CH2 overexpression data, but also demonstrate that the ILK-γ-parvin complex is physiologically essential for leukocyte migration and maintenance of their protein levels. To further characterize the effects of the γ-parvin depletion, we compared the growth rates between the γ-parvin-depleted U937 cells and the control cells. As a result, the depletion of γ-parvin significantly suppressed the growth rate of transfected cells in comparison with the scramble shRNA control (data not shown). As shown in Fig. 8, H and I, the γ-parvin-depleted U937 cells revealed a significant increase in apoptosis unlike the cells overexpressed with CH2 γ-parvin, suggesting that the ILK-γ-parvin complex could be involved in cell survival signaling. Because the overexpression of CH2 γ-parvin did not affect the protein levels of ILK and PINCH-1 (Fig. 6 G), the difference in the apoptosis induction might result from the instability of ILK or PINCH, which is considered to be involved in Akt signaling (38, 39).

FIGURE 8.

Depletion of γ-parvin inhibits cell adhesion, polarization, and spreading. A, Western blot data showing the nearly total loss of γ-parvin in U937 cells stably expressing shRNA for γ-parvin RNAi. Total extracts were probed with anti-γ-parvin (top), anti-ILK (middle), and anti-actin Ab (bottom) as indicated. U937 cells transfected with vector alone (lane 1) or stably expressing scramble shRNA (lane 2) were used as negative controls. B, Depletion of γ-parvin completely blocks the promoting effect of MCP-1 in cell adhesion to FN. U937 cells stably expressing γ-parvin shRNA or control cells (3 × 104 cells per well) were incubated with or without MCP-1 for 2 h on FN-coated 48-well plates. After gently washing, the mean adherent cell numbers were calculated in three independent fields. C, Relative cell sizes in cell spreading assays using U937 cells stably expressing each shRNA vector. Mean sizes were compared with that of the cells transfected with vector alone. NS, Not significant. D, The percentage of cells with round morphology, multiple blebs, or well-spread morphology was evaluated in cell spreading assays. E, U937 cells used in cell spreading assays were stained with anti-γ-parvin Ab (a, c, and e) and FITC-phalloidin (b, d, and f). Note that the depletion of γ-parvin inhibits cell spreading before the formation of multiple blebs. Scale bar, 50 μm. F, The control (a and b) or γ-parvin (c and d) RNAi cells were double stained with FITC-phalloidin (a and c) and anti-vinculin Ab (b and d). Note that the γ-parvin RNAi cells do not show the typical punctate localization of vinculin at FA but its diffuse localization at the cell periphery as compared with the control cells. Scale bar, 10 μm. G, The control (a) or γ-parvin (b) RNAi cells during the early spreading phase in spreading assay were double stained with rhodamine-phalloidin (red) and anti-PTEN Ab (green). Note that the γ-parvin RNAi cells show diffuse distributions of F-actin and PTEN, whereas the control cells reveal their well-polarized localization. Scale bar, 10 μm. H and I, Depletion of γ-parvin induces apoptosis. The percentages of TUNEL-positive cells were calculated by counting at least 300 cells from three randomly selected fields. NS, Not significant. I, These cells were double stained with Hoechst 33258 (a, c, and e) and TUNEL reaction mixture (b, d, and f). Data represent the mean ± SD of three separate experiments. Statistical significance (∗, p < 0.05) determined by Scheffe’s test (B, C, and H) or Student’s t test (D), respectively.

FIGURE 8.

Depletion of γ-parvin inhibits cell adhesion, polarization, and spreading. A, Western blot data showing the nearly total loss of γ-parvin in U937 cells stably expressing shRNA for γ-parvin RNAi. Total extracts were probed with anti-γ-parvin (top), anti-ILK (middle), and anti-actin Ab (bottom) as indicated. U937 cells transfected with vector alone (lane 1) or stably expressing scramble shRNA (lane 2) were used as negative controls. B, Depletion of γ-parvin completely blocks the promoting effect of MCP-1 in cell adhesion to FN. U937 cells stably expressing γ-parvin shRNA or control cells (3 × 104 cells per well) were incubated with or without MCP-1 for 2 h on FN-coated 48-well plates. After gently washing, the mean adherent cell numbers were calculated in three independent fields. C, Relative cell sizes in cell spreading assays using U937 cells stably expressing each shRNA vector. Mean sizes were compared with that of the cells transfected with vector alone. NS, Not significant. D, The percentage of cells with round morphology, multiple blebs, or well-spread morphology was evaluated in cell spreading assays. E, U937 cells used in cell spreading assays were stained with anti-γ-parvin Ab (a, c, and e) and FITC-phalloidin (b, d, and f). Note that the depletion of γ-parvin inhibits cell spreading before the formation of multiple blebs. Scale bar, 50 μm. F, The control (a and b) or γ-parvin (c and d) RNAi cells were double stained with FITC-phalloidin (a and c) and anti-vinculin Ab (b and d). Note that the γ-parvin RNAi cells do not show the typical punctate localization of vinculin at FA but its diffuse localization at the cell periphery as compared with the control cells. Scale bar, 10 μm. G, The control (a) or γ-parvin (b) RNAi cells during the early spreading phase in spreading assay were double stained with rhodamine-phalloidin (red) and anti-PTEN Ab (green). Note that the γ-parvin RNAi cells show diffuse distributions of F-actin and PTEN, whereas the control cells reveal their well-polarized localization. Scale bar, 10 μm. H and I, Depletion of γ-parvin induces apoptosis. The percentages of TUNEL-positive cells were calculated by counting at least 300 cells from three randomly selected fields. NS, Not significant. I, These cells were double stained with Hoechst 33258 (a, c, and e) and TUNEL reaction mixture (b, d, and f). Data represent the mean ± SD of three separate experiments. Statistical significance (∗, p < 0.05) determined by Scheffe’s test (B, C, and H) or Student’s t test (D), respectively.

Close modal

Previous studies showed that affixin and α-parvin differentially interact with α-actinin and paxillin, respectively, through the conserved N-terminal potion of the CH2 domain (16, 25). Thus, to elucidate the molecular mechanism of γ-parvin involvement in cell-substrate adhesion, we next examined the interaction of γ-parvin with paxillin and α-actinin by coimmunoprecipitation assays performed in COS-7 or CHO-K1 cells. As shown in Fig. 9, A and C, overexpressed T7-γ-parvin was coimmunoprecipitated with coexpressed FLAG-paxillin as well as FLAG-α-actinin, suggesting that unlike other members of the parvin family, γ-parvin has the ability to bind both α-actinin and paxillin. These interactions were essentially confirmed endogenously by analyzing anti-γ-parvin immunoprecipitate from U937 cells (Fig. 9, B and D). The reason for the very weak coprecipitation of α-actinin and αPIX with γ-parvin may indicate a requirement of some conformational change of γ-parvin to expose each binding site for full association with these proteins (Figs. 9,D and 7 Ab). Altogether, these results suggest that γ-parvin functions as a member of a multiple cytoskeletal complex with ILK, αPIX, paxillin, and α-actinin as demonstrated for other parvin family members.

FIGURE 9.

γ-Parvin forms complex with paxillin and α-actinin in U937 cells but does not directly associate with F-actin in vitro. COS-7 (A) or CHO-K1 (C) cells transfected with the appropriate combination of plasmid vectors or U937 cells (B and D) were lysed in lysis buffer containing 50 mM HEPES (pH 7.5), 100 mM NaCl, 10 μg/ml leupeptin, 1 mM PMSF, 2 mM NaF and appropriate concentration of detergents including 0.75–1.5% Triton X-100 (A–C) or 1.0% Nonidet P40 with 0.5% deoxycholate and 0.1% SDS (D). Note that T7-γ-parvin is coimmunoprecipitated with FLAG-paxillin (A) or FLAG-α-actinin (C). Similarly, endogenous paxillin and α-actinin are coimmunoprecipitated with γ-parvin (B and D). The negative interaction with actin in D shows that the interaction is not the result of contamination of large cytoskeletal pellets. E, γ-Parvin did not show actin-binding activity by actin cosedimentation assay using GST-γ-parvin fusion protein. The following recombinant proteins were incubated with skeletal muscle F-actin: GST-γ-parvin (lanes 3 and 4), α-actinin (lanes 5 and 6), GST (lanes 7 and 8), and BSA (lanes 9 and 10). GST-γ-parvin was also incubated in the absence of F-actin (lanes 1 and 2). The supernatants (S) and pellets (P) were analyzed by SDS-PAGE on 10% gel stained with Coomassie brilliant blue.

FIGURE 9.

γ-Parvin forms complex with paxillin and α-actinin in U937 cells but does not directly associate with F-actin in vitro. COS-7 (A) or CHO-K1 (C) cells transfected with the appropriate combination of plasmid vectors or U937 cells (B and D) were lysed in lysis buffer containing 50 mM HEPES (pH 7.5), 100 mM NaCl, 10 μg/ml leupeptin, 1 mM PMSF, 2 mM NaF and appropriate concentration of detergents including 0.75–1.5% Triton X-100 (A–C) or 1.0% Nonidet P40 with 0.5% deoxycholate and 0.1% SDS (D). Note that T7-γ-parvin is coimmunoprecipitated with FLAG-paxillin (A) or FLAG-α-actinin (C). Similarly, endogenous paxillin and α-actinin are coimmunoprecipitated with γ-parvin (B and D). The negative interaction with actin in D shows that the interaction is not the result of contamination of large cytoskeletal pellets. E, γ-Parvin did not show actin-binding activity by actin cosedimentation assay using GST-γ-parvin fusion protein. The following recombinant proteins were incubated with skeletal muscle F-actin: GST-γ-parvin (lanes 3 and 4), α-actinin (lanes 5 and 6), GST (lanes 7 and 8), and BSA (lanes 9 and 10). GST-γ-parvin was also incubated in the absence of F-actin (lanes 1 and 2). The supernatants (S) and pellets (P) were analyzed by SDS-PAGE on 10% gel stained with Coomassie brilliant blue.

Close modal

Finally, to examine whether γ-parvin associates with F-actin similarly to α-parvin (19), actin cosedimentation assays were conducted using GST fusion protein of full-length γ-parvin. Although F-actin cosedimented with α-actinin, GST-γ-parvin was not found in the precipitate of F-actin as GST and BSA (Fig. 9 E).

Parvins are recently identified CH domain protein family members that associate with FA or related structures in muscle (such as dense body or sarcolemma) (14, 18, 19, 20, 21, 40). All parvins so far examined bind ILK through their CH2 domain and play an indispensable role in transmitting integrin signals to actin reorganization in the cytoplasm. Genetic defects of C. elegans Parvin (PAT-6), ILK (PAT-4), and β Integrin (PAT-3) all result in the same embryonic phenotype called PAT (paralyzed and arrested elongation at the 2-fold stage) characterized by developmental arrest of the body wall muscle due to the abnormal integrin-actin assembly and cell-ECM attachment (8, 17, 18). In mammals, the knockdown of affixin, one of the three mammalian parvins, in fibroblasts caused the disruption of FA and lamellipodium formation (16). In this study, we comprehensively analyzed the expression and function of the most diverse member of the mammalian parvins, γ-parvin, which has not been well characterized. In this study, we demonstrated that γ-parvin shares many typical features with other mammalian parvins and plays critical roles for leukocyte adhesion and spreading on FN.

We cloned γ-parvin as an ILK-binding protein in our yeast two-hybrid screening together with other parvins, affixin and α-parvin from the cDNA libraries. The subsequent analysis demonstrated that, like other parvins, γ-parvin directly binds ILK through its CH2 domain and can form a protein complex with several cytoskeletal proteins involved in cell-substrate adhesion, including paxillin, α-actinin, and αPIX, which have been shown to bind to α-parvin or affixin (15, 16, 19). Considering the relatively high divergence of γ-parvin from other parvins, these results are somewhat unexpected. γ-Parvin only shows 45.9 and 51.4% amino acid identity with affixin in the CH1 and CH2 domains, respectively, and the values are even lower than those between affixin and D. melanogaster parvin (CG32528-PA; GenBank/EBI Data Bank accession no. AAF49016; 61.3 and 70.3%, respectively) or C. elegans PAT-6 (47.7 and 62.2%, respectively). These findings may indicate that the amino acid residues essential for binding to ILK, paxillin, α-actinin, and αPIX are all conserved in γ-parvin. In fact, the valine 282 and leucine 285 residues in α-parvin, critical for the binding with paxillin (19), are conserved in the corresponding region of γ-parvin. In contrast, the glutamate 359 of ILK was also essential for the interaction with γ-parvin as demonstrated for affixin and α-parvin (14, 25), indicating the presence of the same ILK interaction mode between γ-parvin and affixin. Interestingly, with respect to protein-protein interaction, the only difference found between γ-parvin and affixin was the paxillin binding of γ-parvin, which was not detected for affixin. Additionally, affixin and α-parvin show opposite interacting affinities with α-actinin and paxillin; nevertheless, almost all overlapping binding regions for these proteins are highly conserved in affixin and α-parvin (16, 25). At the present time, we do not know the molecular basis of the difference in property between affixin and α-parvin. Although further studies are required to clarify this issue, it is intriguing to speculate that a possible steric hindrance that restricts the interaction of affixin with paxillin and that of α-parvin with α-actinin is relieved in γ-parvin.

Consistent with the conserved interactions with several cell adhesion molecules, γ-parvin showed the same subcellular localization as other mammalian parvins at FA in well-spreading leukocytes through its CH2 domain. It was also colocalized with ILK at the cell surface blebs observed in the early stage of cell spreading, suggesting its essential roles in the initial cell-spreading process of leukocytes. Consistently, the overexpression of the CH2 domain of γ-parvin or the depletion of γ-parvin by RNAi resulted in the suppression of cell attachment and the spreading of PMA-stimulated U937 cells. In contrast, the overexpression of the CH1 domain of γ-parvin, which was shown to interact with αPIX, led to the significant enhancement of cell attachment and spreading of U937 cells independent of PMA stimulation. These results completely agree with the results obtained from affixin in fibroblasts (14). Taken together, our results support the notion that γ-parvin also plays critical roles in the initial cell-substrate interactions by transmitting integrin outside-in signals from ILK to cytoplasmic F-actin reorganization. Importantly, γ-parvin and ILK showed polarized distributions with F-actin in contrast to PTEN during the early stage of cell spreading, and the overexpression of γ-parvin CH2 or the depletion of γ-parvin by RNAi disrupted the polarized distribution of F-actin and PTEN. These results further suggest that the ILK-γ-parvin complex is indispensable for amplifying initial integrin signals to establish cell polarization required for leukocyte migration. Supporting this notion, Sakai et al. (10) reported that the ILK-deficient mice failed to polarize epiblast cells and suggested that a major function of ILK in epiblast cells is to organize the proper localization of F-actin in the cells. Furthermore, recently Liu et al. (39) reported that ILK-deficient T cells revealed the significant reduction in chemotaxis to chemokines and enhanced apoptosis using T cell-specific knockouts by breeding conditional ILK knockout mice. Considering the dominant expression of γ-parvin in leukocyte cell lines, our results suggest that γ-parvin is a parvin molecule that has developed to specifically function for leukocytes. Although it should further be confirmed for a wider variety of cell lines, our study results demonstrating that γ-parvin and affixin show mutually exclusive expression patterns are interesting. Because the human γ-parvin gene is located on chromosome 22, ∼12 kb downstream of the 3′ end of the human affixin gene (22), these patterns may result from the mutually interfering transcriptional regulation between these genes. Because these three structurally related parvins can bind to ILK through the highly conserved N-terminal CH2 regions, we need to know the mutual interactions among the members of the parvin family. Indeed, it has been reported that the α-parvin and affixin proteins are coexpressed in Hela cells, and each parvin could form a mutually exclusive complex with ILK (41). Thus, functional divergence and mutual interaction among the members of the parvin family, particularly if any between affixin and γ-parvin in leukocyte migration, should be examined in the future.

The migration of leukocytes consists of multistep sequential events in which many adhesive cell surface molecules are dynamically regulated (1). After the first tethering step by selectin, the development of a tight adhesion on the endothelium is predicted to use α4β1 integrin, followed by αL/Mβ2 integrin (42, 43, 44). Although it is unknown whether β2 integrin directly correlates with ILK, so far, the ILK activation is evoked at least via the stimulation of β1 integrin by the α4β1 ligand, VCAM-1 (13). In epithelial cells, the kinase activity of ILK is elevated in the cytoskeletal fraction, and the interaction of α-parvin with ILK within the cytoskeleton stimulates ILK activity (45). These results imply that the activated integrin outside-in signals induced by their ligands can be transmitted to ILK by its recruitment to the early integrin complex with the parvin family proteins, which results in their translocation to the cytoskeletal fraction during the remodeling of the integrin-cytoskeleton complex. This notion was further supported by the observation that the integrin-ILK-affixin complex is incorporated into the membrane skeletal fractions after the acute and transient activation of ILK triggered by thrombin in platelets (46). Thus, it is plausible that the integrin-ILK signals can finally result in some modification of γ-parvin, such as the phosphorylation of the CH2 domain by ILK, and play an essential role in the remodeling of the integrin-cytoskeletal interface by hierarchical mediation with the α-actinin-zyxin-Mena and αPIX-PAK complex through γ-parvin (15, 16). Indeed, we confirmed that full-length and CH2 γ-parvins are phosphorylated in vitro by wt ILK but not by K220M ILK immunoprecipitated from COS-7 cells overexpressing these proteins (our unpublished observation), suggesting a possibility that, like affixin (14), γ-parvin is not only an ILK-binding protein but also an in vivo substrate of ILK. However, several recent studies suggest that ILK may function as an adapter protein rather than a kinase, thus further studies including the determination of ILK phosphorylation sites on γ-parvin are needed to examine these hypotheses and to extend our comprehension for the ILK-γ-parvin signaling.

It has been shown that chemokines, such as MCP-1, enhance leukocyte accumulation during an inflammatory response by the rapid conversion of initial leukocyte tethering to firm adhesion (47). Chemokines are also suggested to be important for the subsequent migration of leukocytes into the subendothelium. In this report, we demonstrated that the overexpression of the CH1 domain of γ-parvin in U937 promonocytic cells significantly potentiates the MCP-1-induced augmentation of cell adhesion to FN, whereas the overexpression of the CH2 domain blocks it completely. These effects can be explained from their effects on outside-in signaling triggered by the interaction between α4β1 or α5β1 integrins and FN, which enhances cell spreading after substrate attachment. However, Friedrich et al. (13) have recently reported the possibility that ILK is involved in MCP-1-induced inside-out signals by showing that MCP-1 activates ILK within 30 s in THP-1 cells, and the overexpression of ILK expression entirely inhibits the MCP-1-induced activation of cell adhesion via β1 integrin-VCAM-1 interaction. Therefore, there is a possibility that γ-parvin is also involved in the inside-out signal of integrin by which chemokines dynamically regulate substrate interaction of leukocytes to promote extravasation. Although further study is needed to examine this hypothesis, the results in this study provide a novel machinery to link the cytoskeleton to the integrin cytoplasmic domains in leukocytes.

We thank Dr. Hisataka Sabe for providing the human paxillin cDNA and Dr. Yohei Kirino for technical assistance for flow cytometry and preparation of human PBMC.

The authors have no financial conflict of interest.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1

This work was supported in part by grants from the Yokohama City University Center of Excellence Program of the Ministry of Education, Sports, Science, and Technology of Japan (to Y.I.), the Uehara Memorial Foundation (to S.Y.), the Yokohama Foundation for Advancement of Medical Science (to S.Y.), the 2005 Strategic Research Project No. W17017 from Yokohama City University (to S.Y.), and the Ministry of Education, Culture, Sports, Science, and Technology of Japan (to S.Y.).

3

Abbreviations used in this paper: ECM, extracellular matrix; FN, fibronectin; ILK, integrin-linked kinase; FA, focal adhesion; CH, calponin homology; RNAi, RNA interference; shRNA, short hairpin RNA; HA, hemagglutinin; wt, wild type; CHO, Chinese hamster ovary; GEF, guanine nucleotide exchange factor.

1
Springer, T. A..
1994
. Traffic signals for lymphocyte recirculation and leukocyte emigration: the multistep paradigm.
Cell
76
:
301
-314.
2
Laudanna, C., J. Y. Kim, G. Constantin, E. Butcher.
2002
. Rapid leukocyte integrin activation by chemokines.
Immunol. Rev.
186
:
37
-46.
3
Chuluyan, H. E., A. C. Issekutz.
1993
. VLA-4 integrin can mediate CD11/CD18-independent transendothelial migration of human monocytes.
J. Clin. Invest.
92
:
2768
-2777.
4
Shang, X. Z., B. J. Lang, A. C. Issekutz.
1998
. Adhesion molecule mechanisms mediating monocyte migration through synovial fibroblast and endothelium barriers: role for CD11/CD18, very late antigen-4 (CD49d/CD29), very late antigen-5 (CD49e/CD29), and vascular cell adhesion molecule-1 (CD106).
J. Immunol.
160
:
467
-474.
5
Hannigan, G. E., C. Leung-Hagesteijn, L. Fitz-Gibbon, M. G. Coppolino, G. Radeva, J. Filmus, J. C. Bell, S. Dedhar.
1996
. Regulation of cell adhesion and anchorage-dependent growth by a new β1-integrin-linked protein kinase.
Nature
379
:
91
-96.
6
Huang, Y., C. Wu.
1999
. Integrin-linked kinase and associated proteins (review).
Int. J. Mol. Med.
3
:
563
-572.
7
Hannigan, G., A. A. Troussard, S. Dedhar.
2005
. Integrin-linked kinase: a cancer therapeutic target unique among its ILK.
Nat. Rev. Cancer
5
:
51
-63.
8
Mackinnon, A. C., H. Qadota, K. R. Norman, D. G. Moerman, B. D. Williams.
2002
. C. elegans PAT-4/ILK functions as an adaptor protein within integrin adhesion complexes.
Curr. Biol.
12
:
787
-797.
9
Zervas, C. G., S. L. Gregory, N. H. Brown.
2001
. Drosophila integrin-linked kinase is required at sites of integrin adhesion to link the cytoskeleton to the plasma membrane.
J. Cell Biol.
152
:
1007
-1018.
10
Sakai, T., S. Li, D. Docheva, C. Grashoff, K. Sakai, G. Kostka, A. Braun, A. Pfeifer, P. D. Yurchenco, R. Fassler.
2003
. Integrin-linked kinase (ILK) is required for polarizing the epiblast, cell adhesion, and controlling actin accumulation.
Genes Dev.
17
:
926
-940.
11
Grashoff, C., A. Aszodi, T. Sakai, E. B. Hunziker, R. Fassler.
2003
. Integrin-linked kinase regulates chondrocyte shape and proliferation.
EMBO Rep.
4
:
432
-438.
12
Friedrich, E. B., E. Liu, S. Sinha, S. Cook, D. S. Milstone, C. A. MacRae, M. Mariotti, P. J. Kuhlencordt, T. Force, A. Rosenzweig, et al
2004
. Integrin-linked kinase regulates endothelial cell survival and vascular development.
Mol. Cell. Biol.
24
:
8134
-8144.
13
Friedrich, E. B., S. Sinha, L. Li, S. Dedhar, T. Force, A. Rosenzweig, R. E. Gerszten.
2002
. Role of integrin-linked kinase in leukocyte recruitment.
J. Biol. Chem.
277
:
16371
-16375.
14
Yamaji, S., A. Suzuki, Y. Sugiyama, Y. Koide, M. Yoshida, H. Kanamori, H. Mohri, S. Ohno, Y. Ishigatsubo.
2001
. A novel integrin-linked kinase-binding protein, affixin, is involved in the early stage of cell-substrate interaction.
J. Cell Biol.
153
:
1251
-1264.
15
Mishima, W., A. Suzuki, S. Yamaji, R. Yoshimi, A. Ueda, T. Kaneko, J. Tanaka, Y. Miwa, S. Ohno, Y. Ishigatsubo.
2004
. The first CH domain of affixin activates Cdc42 and Rac1 through αPIX, a Cdc42/Rac1-specific guanine nucleotide exchanging factor.
Genes Cells
9
:
193
-204.
16
Yamaji, S., A. Suzuki, H. Kanamori, W. Mishima, R. Yoshimi, H. Takasaki, M. Takabayashi, K. Fujimaki, S. Fujisawa, S. Ohno, Y. Ishigatsubo.
2004
. Affixin interacts with α-actinin and mediates integrin signaling for reorganization of F-actin induced by initial cell-substrate interaction.
J. Cell Biol.
165
:
539
-551.
17
Williams, B. D., R. H. Waterston.
1994
. Genes critical for muscle development and function in Caenorhabditis elegans identified through lethal mutations.
J. Cell Biol.
124
:
475
-490.
18
Lin, X., H. Qadota, D. G. Moerman, B. D. Williams.
2003
. C. elegans PAT-6/actopaxin plays a critical role in the assembly of integrin adhesion complexes in vivo.
Curr. Biol.
13
:
922
-932.
19
Nikolopoulos, S. N., C. E. Turner.
2000
. Actopaxin, a new focal adhesion protein that binds paxillin LD motifs and actin and regulates cell adhesion.
J. Cell Biol.
151
:
1435
-1448.
20
Tu, Y., Y. Huang, Y. Zhang, Y. Hua, C. Wu.
2001
. A new focal adhesion protein that interacts with integrin-linked kinase and regulates cell adhesion and spreading.
J. Cell Biol.
153
:
585
-598.
21
Olski, T. M., A. A. Noegel, E. Korenbaum.
2001
. Parvin, a 42 kDa focal adhesion protein, related to the α-actinin superfamily.
J. Cell Sci.
114
:
525
-538.
22
Korenbaum, E., T. M. Olski, A. A. Noegel.
2001
. Genomic organization and expression profile of the parvin family of focal adhesion proteins in mice and humans.
Gene
279
:
69
-79.
23
Manser, E., T. H. Loo, C. G. Koh, Z. S. Zhao, X. Q. Chen, L. Tan, I. Tan, T. Leung, L. Lim.
1998
. PAK kinases are directly coupled to the PIX family of nucleotide exchange factors.
Mol. Cell
1
:
183
-192.
24
Suzuki, A., M. Hirata, K. Kamimura, R. Maniwa, T. Yamanaka, K. Mizuno, M. Kishikawa, H. Hirose, Y. Amano, N. Izumi, et al
2004
. αPKC acts upstream of PAR-1b in both the establishment and maintenance of mammalian epithelial polarity.
Curr. Biol.
14
:
1425
-1435.
25
Nikolopoulos, S. N., C. E. Turner.
2002
. Molecular dissection of actopaxin-integrin-linked kinase-paxillin interactions and their role in subcellular localization.
J. Biol. Chem.
277
:
1568
-1575.
26
Ferreira, O. C., Jr, J. E. Valinsky, K. Sheridan, E. A. Wayner, C. Bianco, A. Garcia-Pardo.
1991
. Phorbol ester-induced differentiation of U937 cells enhances attachment to fibronectin and distinctly modulates the α5β1 and α4β1 fibronectin receptors.
Exp. Cell Res.
193
:
20
-26.
27
Boles, B. K., J. Ritzenthaler, T. Birkenmeier, J. Roman.
2000
. Phorbol ester-induced U-937 differentiation: effects on integrin α5 gene transcription.
Am. J. Physiol.
278
:
L703
-L712.
28
Sánchez-Madrid, F., M. A. del Pozo.
1999
. Leukocyte polarization in cell migration and immune interactions.
EMBO J.
18
:
501
-511.
29
Li, Z., X. Dong, Z. Wang, W. Liu, N. Deng, Y. Ding, L. Tang, T. Hla, R. Zeng, L. Li, D. Wu.
2005
. Regulation of PTEN by Rho small GTPases.
Nat. Cell Biol.
7
:
399
-404.
30
Meili, R., A. T. Sasaki, R. A. Firtel.
2005
. Rho rocks PTEN.
Nat. Cell Biol.
7
:
334
-335.
31
Miller, L. J., R. Schwarting, T. A. Springer.
1986
. Regulated expression of the Mac-1, LFA-1, p150,95 glycoprotein family during leukocyte differentiation.
J. Immunol.
137
:
2891
-2900.
32
Weber, C., R. Alon, B. Moser, T. A. Springer.
1996
. Sequential regulation of α4β1 and α5β1 integrin avidity by CC chemokines in monocytes: implications for transendothelial chemotaxis.
J. Cell Biol.
134
:
1063
-1073.
33
Ashida, N., H. Arai, M. Yamasaki, T. Kita.
2001
. Distinct signaling pathways for MCP-1-dependent integrin activation and chemotaxis.
J. Biol. Chem.
276
:
16555
-16560.
34
Kovach, N. L., T. M. Carlos, E. Yee, J. M. Harlan.
1992
. A monoclonal antibody to β1 integrin (CD29) stimulates VLA-dependent adherence of leukocytes to human umbilical vein endothelial cells and matrix components.
J. Cell Biol.
116
:
499
-509.
35
Kato, K., Y. Hashimoto, H. Kanamori, T. Okubo, H. Mohri.
1998
. Additional recognition sites in the C-terminal heparin-binding domain of fibronectin promote adhesion of PMA-treated U937 cells.
Peptides
19
:
7
-13.
36
Hazlehurst, L. A., N. Valkov, L. Wisner, J. A. Storey, D. Boulware, D. M. Sullivan, W. S. Dalton.
2001
. Reduction in drug-induced DNA double-strand breaks associated with β1 integrin-mediated adhesion correlates with drug resistance in U937 cells.
Blood
98
:
1897
-1903.
37
Filipenko, N. R., S. Attwell, C. Roskelley, S. Dedhar.
2005
. Integrin-linked kinase activity regulates Rac- and Cdc42-mediated actin cytoskeleton reorganization via α-PIX.
Oncogene
24
:
5837
-5849.
38
Fukuda, T., K. Chen, X. Shi, C. Wu.
2003
. PINCH-1 is an obligate partner of integrin-linked kinase (ILK) functioning in cell shape modulation, motility, and survival.
J. Biol. Chem.
278
:
51324
-51333.
39
Liu, E., S. Sinha, C. Williams, M. Cyrille, E. Heller, S. B. Snapper, K. Georgopoulos, R. St-Arnaud, T. Force, S. Dedhar, R. E. Gerszten.
2005
. Targeted deletion of integrin-linked kinase reveals a role in T-cell chemotaxis and survival.
Mol. Cell. Biol.
25
:
11145
-11155.
40
Matsuda, C., K. Kameyama, K. Tagawa, M. Ogawa, A. Suzuki, S. Yamaji, H. Okamoto, I. Nishino, Y. K. Hayashi.
2005
. Dysferlin interacts with affixin (β-parvin) at the sarcolemma.
J. Neuropathol. Exp. Neurol.
64
:
334
-340.
41
Zhang, Y., K. Chen, Y. Tu, C. Wu.
2004
. Distinct roles of two structurally closely related focal adhesion proteins, α-parvins and β-parvins, in regulation of cell morphology and survival.
J. Biol. Chem.
279
:
41695
-41705.
42
Alon, R., P. D. Kassner, M. W. Carr, E. B. Finger, M. E. Hemler, T. A. Springer.
1995
. The integrin VLA-4 supports tethering and rolling in flow on VCAM-1.
J. Cell Biol.
128
:
1243
-1253.
43
Chan, J. R., S. J. Hyduk, M. I. Cybulsky.
2000
. α4β1 Integrin/VCAM-1 interaction activates αLβ2 integrin-mediated adhesion to ICAM-1 in human T cells.
J. Immunol.
164
:
746
-753.
44
May, A. E., F. J. Neumann, A. Schomig, K. T. Preissner.
2000
. VLA-4 (α4β1) engagement defines a novel activation pathway for β2 integrin-dependent leukocyte adhesion involving the urokinase receptor.
Blood
96
:
506
-513.
45
Attwell, S., J. Mills, A. Troussard, C. Wu, S. Dedhar.
2003
. Integration of cell attachment, cytoskeletal localization, and signaling by integrin-linked kinase (ILK), CH-ILKBP, and the tumor suppressor PTEN.
Mol. Biol. Cell
14
:
4813
-4825.
46
Yamaji, S., A. Suzuki, H. Kanamori, W. Mishima, M. Takabayashi, K. Fujimaki, N. Tomita, S. Fujisawa, S. Ohno, Y. Ishigatsubo.
2002
. Possible role of ILK-affixin complex in integrin-cytoskeleton linkage during platelet aggregation.
Biochem. Biophys. Res. Commun.
297
:
1324
-1331.
47
Imhof, B. A., M. Aurrand-Lions.
2004
. Adhesion mechanisms regulating the migration of monocytes.
Nat. Rev. Immunol.
4
:
432
-444.