Swiprosin-1 Is a Novel Actin Bundling Protein That Regulates Cell Spreading and Migration

Protein functions are often revealed by their localization to specialized cellular sites. Recent reports demonstrated that swiprosin-1 is found together with actin and actin-binding proteins in the cytoskeleton fraction of human mast cells and NK-like cells. However, direct evidence of whether swiprosin-1 regulates actin dynamics is currently lacking. We found that swiprosin-1 localizes to microvilli-like membrane protrusions and lamellipodia and exhibits actin-binding activity. Overexpression of swiprosin-1 enhanced lamellipodia formation and cell spreading. In contrast, swiprosin-1 knockdown showed reduced cell spreading and migration. Swiprosin-1 induced actin bundling in the presence of Ca2+, and deletion of the EF-hand motifs partially reduced bundling activity. Swiprosin-1 dimerized in the presence of Ca2+ via its coiled-coil domain, and a lysine (Lys)-rich region in the coiled-coil domain was essential for regulation of actin bundling. Consistent with these observations, mutations of the EF-hand motifs and coiled-coil region significantly reduced cell spreading and lamellipodia formation. We provide new evidence of how swiprosin-1 influences cytoskeleton reorganization and cell spreading.


Introduction
Motile cells display dynamic movement by lamellipodia-or filopodia-based membrane extensions at the leading cell edge. The lamellipodium is a sheet-like protrusion that contains an extensively branched network or meshwork of actin filaments [1,2]. Filopodia are rod-like extensions composed of long, unbranched, parallel bundles of actin [3]. The shape and dynamics of protrusive organelles are dependent on actin filament nucleation and polymerization as well as the assembly of actin filaments into bundles and networks by numerous actin-binding proteins [4][5][6]. Actin-filament bundles and networks in the lamellipodia of motile cells contain several actin cross-linking proteins, including fascin, fimbrin, filamin, and a-actinin [7]. Proteins such as fascin and fimbrin are particularly suited for forming strong bundles because they are short, compact, and form monomeric cross-links between adjacent actin filaments. In cells, fascin is the major bundling protein in filopodia, and in fascin-depleted cells, the few remaining filopodia are wavy, loosely bundled, and parallel to the membrane [8]. Fimbrin represents the simplest actin-bundling protein structure, originally identified as a component of the actin bundle in intestinal microvilli. Filamin localizes to the actin filament network of lamellipodia and human cell lines deficient in filamin isoforms and FLNa, and filamin spreads poorly at the edges of these cells [9]. The a-actinin knockdown inhibits actin orientation and adhesion elongation in protrusions [10]. In summary, crosslinking of actin filaments is critical for cell motility and is a fundamental process during filopodia protrusion and lamellipodia formation.
Swiprosin-1 was first identified in human lymphocytes, predominantly in CD8 + lymphocytes [11] and later in immature, resting or activated B cells [12,13], and non-lymphoid tissue, especially in the brain [14][15][16]. Swiprosin-1 also has been identified in mast cells and is upregulated through the protein kinase CbI/g pathway [17]. Recent reports demonstrated that swiprosin-1 is involved in the B cell receptor-induced calcium flux, controlling B cell receptor signaling [18][19][20]. However, the swiprosin-1 function in lymphoid or non-lymphoid cells is still poorly understood. We found that swiprosin-1 is highly accumulated in actin cytoskeletonrich regions and modulates mast cell activation through actin remodeling [21]. Another group reported that Drosophila swiprosin-1 often overlaps with F-actin foci during embryonic myoblast fusion [22]. Furthermore, swiprosin-1 is found in the cytoskeleton fraction in NK-like cells with actin and actin-binding proteins such as a-actinin, plastin, and filamin [23]. In addition, swiprosin-1 has been identified in caspase-9 complexes with the cytoskeletal protein ezrin, or with microtubule-associated tau proteins [14,24]. Consistent with these findings, a previous report demonstrated that swiprosin-1 exhibits phosphorylation cycles similar to those of gelsolin and the Arp2/3 complex, which are actin-binding proteins that participate in actin dynamics after epidermal growth factor (EGF) stimulation [25]. However, direct [13] relationship between swiprosin-1 and the actin cytoskeleton, and its related functions have not been reported yet. Here, we study the interaction between swiprosin-1 and actin as well as the important role of swiprosin-1 in mediating the structural changes during cell adhesion and spreading.
In the present study, we asked if swiprosin-1 binds to F-actin. If so, what is the functional consequence of this binding? We demonstrated that swiprosin-1 directly binds to F-actin through multiple actin-binding sites and that swiprosin-1 functions as a structural protein for F-actin bundling in vitro and in vivo. We investigated the mechanism whereby swiprosin-1 induces actin bundling. Database mining revealed that swiprosin-1 contains 2 putative EF-hand motifs and a coiled-coil domain at the Cterminus [11,13]. At the leading edge of cells, calcium plays a critical role in the action of many actin cross-linking proteins, including a-actinin [26], villin [27], and the 34-kDa actin crosslinking protein [28]. Calcium is an essential regulator of actin bundling but not actin binding by swiprosin-1. The bundling function may be mediated through both EF-hand motifs and coiled-coil domain, because deletion of amino acids in the two EFhand motifs or coiled-coil domain significantly reduced the actin bundling activity of swiprosin-1. The coiled-coil domain facilitates self-dimerization in the presence of calcium and is critical for bundling activity. Accordingly, a swiprosin-1 variant with a mutation in the coiled-coil region lost the activity to enhance cell spreading and lamellipodium.

Cell Culture
Jurkat T (ATCC CRL-1651) and Raji B (ATCC CCL-86) cells obtained from American Type Culture Collection (Manassas, VA) were grown in RPMI medium supplemented with 10% heatactivated fetal bovine serum (FBS), penicillin G (100 IU/ml), streptomycin (100 mg/ml), and L-glutamine (2 mM). CHO-K1 (ATCC CCL-61), HeLa (ATCC CCL-2), and HEK293T cells (ATCC CRL-1573) were grown in DMEM medium supplemented with 10% heat-inactivated FBS. After written informed consent, human primary PBLs were isolated from healthy donors by dextran sedimentation and centrifugation through a discontinuous Ficoll gradient (Amersham Biosciences, Piscataway, NJ). The cell lines and human PBLs mentioned above were cultured at 37uC in a humidified incubator containing 5% CO 2 and 95% air. All experiments using human PBLs were approved by Ethics Committee of the School of Life Sciences, GIST.

Purification of Recombinant Fusion Proteins
His tagged wild-type swiprosin-1 cloned into pET-28a were transformed into E. coli strain BL21 (DE3) and the protein lysates were obtained as described above. The soluble supernatant was loaded onto an equilibrated gravity-flow column (Bio-Rad, Hercules, CA) packed with Ni-NTA agarose resin (Peptron, Korea) and subsequently washed with lysis buffer. The protein was eluted with lysis buffer supplemented with 300 mM imidazole. During purification, the presence of swiprosin-1 protein was confirmed by SDS-PAGE.
were collected and spin-infected into Jurkat T cells by centrifugation at 8006g for 30 min in the presence of 8 mg/ml polybrene. The infection efficiency was checked by western blot 48 h after infection.

Conjugation Assay
For conjugation with anti-CD3/28-coated beads, Jurkat T cells transfected with GFP_Swip-1 or Actin_GFP were incubated for 30 min with anti-CD3/28-coated beads. The conjugates were then imaged by FV1000 confocal laser scanning microscope (Olympus, Japan). For superantigen stimulation, Raji B cells were incubated with SEE (5 mg/ml) for 30 min, washed, and resuspended in RPMI medium, and then equal numbers of B and T cells were mixed and incubated at 37uC for 30 min [29].

Spreading and Migration Assay
For cell spreading assays, CHO-K1 or HeLa cells were harvested with phosphate-buffered saline (PBS)/EDTA, washed with serum-free DMEM, and re-plated on 10 mg/ml FN-coated glass coverslip in serum-free medium. After 60 min, they were fixed with 4% paraformaldehyde. Images were captured using a FV1000 confocal laser scanning microscope (Olympus, Japan). The cell size was determined from digital images of nine randomly selected fields using FLUOVIEW software. For T cell spreading assay, Jurkat T cells expressing GFP or GFP_Swip-1 were placed on 10 mg/ml FN-coated glass coverslip for 10 min. The cells were then treated with SDF-1a (100 nM) and observed random cell migration or spreading using time laps imaging for 20 min. Images were captured and processed as described above.
For wound-healing assays, cells were cultured for 1 d. They were then scratched with micropipette tips, and images were captured at 0 and 10 h after wounding using a Nikon Eclipse TE300 microscope and a Nikon Plan Fluor 4_0.13 objective. For each experiment, migration of cells in nine random fields were measured, and three independent filters were analyzed.

Immunofluorescence Staining and Confocal Imaging
Cells (Jurkat T or CHO-K1 cells expressing wild-type of various mutants of swiprosin-1) were incubated for 1 h on FN (10 mg/ml) or PLL-coated glass coverslips. The cells were fixed, washed twice with PBS, and blocked with 5% goat serum (DAKO, Denmark) in PBS for 30 min. The cells were then incubated with primary antibodies in blocking buffer for 3 h at room temperature, rinsed 3 times with PBS, incubated with secondary antibody in blocking buffer for 1 h, rinsed 3 times with PBS, and mounted with antifade solution (Molecular Probes, OR). To stain swiprosin-1 and actin, cells were permeabilized for 5 min with PBS containing 0.1% Triton X-100 at room temperature before incubation with primary antibody or Phalloidin-TRITC. The slides were examined with an FV1000 confocal laser scanning microscope (Olympus, Japan) equipped with 406, 636, and 1006 objectives. For the co-localization analysis, the Z section cutting area through the bottom field was chosen. Images captured at different wavelengths were superimposed and the intensity of each fluorochrome in the field was then plotted in a scattergram by FLUOVIEW software. Overlapping degree of intensity (ODI) values were calculated by FLUOVIEW software. Co-localization percentage was then derived from the ODI, multiplying by 100. Lamellipodium formation (LPF) in CHO-K1 cells was quantified: 0 = no lamellipodium, 1 = isolated areas of lamellipodium covering no more than 50% of the peripheral area, 2 = extensive lamellipodium covering more than 50% of the peripheral area. More than 100 cells were counted, and the results were represented as % of cells with score 0, 1, and 2, individually.
For GST pull-down, equal amounts of GST or GST_ Swip-1s (wild-type and coiled-coil mutant), and His_Swip-1 were mixed in lysis buffer (50 mM Hepes, pH 7.4, containing 150 mM NaCl, 1.5 mM MgCl 2 , 10% glycerol, 1% triton X-100, 5 mM sodium fluoride, 1 mM sodium orthovanadate, 1 mM 4-nitrophenyl phosphate, 10 mg/ml leupeptin, 10 mg/ml pepstatin A, and 1 mM 4-(2-aminoethyl benzenesulfonyl fluoride), and incubated with glutathione-Sepharose 4B beads (GE Healthcare, UK) for 2 h at 4uC with rotation. If the GST pull-down assay requires special conditions that need to remove residual Ca 2+ or to maintain Ca 2+ , the lysis or wash buffers were supplied with 2 mM EGTA or 2 mM CaCl 2 during the whole process of experiment. For western blot analysis, all precipitates were washed and eluted with SDS sample buffer (100 mM Tris-HCl, pH 6.8, 4% SDS, 20% glycerol with bromophenol blue), and heated for 5 min. The proteins were separated through 10% SDS-PAGE gels and were transferred into a nylon membrane. The membrane was blocked in 5% skim milk (1 h), rinsed, and incubated with intended antibodies (anti-GFP, anti-swiprosin-1, anti-actin, and anti-myc) in TBS containing 0.1% Tween 20 (TBS-T) and 3% skim milk for 2 h. The membrane was then incubated with 0.1 mg/ml peroxidase-labeled secondary antibody (against rabbit or mouse) for 1 h. After three washes in TBS-T, bands were visualized by ECL reagents and were then exposed to x-ray film.

In vitro Actin Co-sedimentation Assays
Actin co-sedimentation assays were performed as follows [30] : non-muscle actin derived from human platelets were purchased (Cytoskeleton Inc., Denver, CO). Actin was mixed in G-buffer (0.5 mM CaCl 2 , 5 mM Tris-HCl, pH 8.0) to produce an actin stock. Actin was polymerized in actin polymerization buffer (100 mM KCl, 2 mM MgCl 2 , 0.5 mM ATP, 0.2 mM Tris-HCl, pH 8.0) at room temperature for 1 h and then incubated with Hisor GST-tagged Swip-1 or its mutants for 30 min at room temperature. Actin filaments with bound proteins were pelleted by centrifugation at 100,0006g for 2 h at room temperature (for the actin binding assay) or 15,0006g for 10 min at room temperature (for the actin bundling assay). Equal amounts of pellet and supernatant were resolved by SDS-PAGE and proteins were visualized by Coomassie blue staining. The percent of actin in supernatant (sup) and pellet (pet) was quantified by densitometry using ImageJ 1.44p (Wayne Rasband).

Quantification of Binding Data
Quantitation of actin binding affinity was performed as previously described [31]. Briefly, the intensity ratio of recombinant protein to F-actin in each pellet was converted to a molar ratio (mole of His_Swip-1 or GST_Swip-1/mole of actin) using standard curves run on each gel that contained known amounts of His_Swip-1 or GST_Swip-1 (1-15 mM) and actin (4 mM) in mol/ mol ratios. Cosedimentation binding data were plotted versus the free His_Swip-1 or GST_Swip-1 concentration added and fit according to the following equation, where [C bound ] and [C free ] are the bound and free concentrations of His_Swip-1 or GST_Swip-1 proteins, respectively, [Actin] is the total actin concentration, B max is the maximal molar binding ratio, and K d is the dissociation constant (mM). Data were fit using the Microsoft Excel Solver package by varying the values of B max and K d and minimizing the sum of squares between the actual and predicted binding ratios. Significance was calculated using analysis of variance and t tests. Results were considered significant at p,0.01.

Protein Cross-linking Assay
Cells were transfected with the indicated plasmids and lysed with 0.5% Nonidet P-40 lysis buffer (0.5% Nonidet P-40, 50 mM Tris-HCl (pH 7.5), 150 mM NaCl, 50 mM NaF, 1 mM NaVO3, 1 mM DTT, 1 mM PMSF, and protease inhibitor mixture) for 48 h after transfection. Glutaraldehyde (0.1-1610 220 %) was added to the lysate at the indicated concentrations. After incubating the lysate on ice for 20 min, the glutaraldehyde reactions were stopped by adding 26loading buffer, and the samples were heated at 100uC for 5 min and resolved by SDS PAGE. Western blot analysis was performed with anti-GFP antibody.

Electron Microscopy
Non-muscle actin (cytoskeleton) was polymerized in F-actin buffer containing 2 mM MgCl 2 , 50 mM KCl, and 10 mM ATP at pH 7.5. Mixtures of F-actin (2 mM) and wild-type swiprosin-1 (4 mM) in the presence or absence of EGTA (1 mM) were allowed to react for 60 min. The reaction mixtures were stained using a solution of uranyl acetate in DW. Formval and Metal coating grids were immersed in the stain solution for 20 min, drained of excess stain with filter paper, and air-dried. The samples were analyzed using FEL Tecnai G2 TEM operated at 120 KV.

Statistics
The mean values were calculated from data taken from at least three (usually three or more) separate experiments conducted on separate days. Where significance testing was performed, an unpaired Student's t-test was used. We considered differences between groups significant at P,0.05.

Swiprosin-1 is Associated with the Actin Cytoskeleton and Enhances Lamellipodia and Cell Migration
Swiprosin-1 was initially identified in human lymphocytes [11], thus, we first investigated the localization of swiprosin-1 in Jurkat T or human primary T cells ( Fig. 1A-a and b). Endogenous Swip-1 in Jurkat T cells was recruited to the actin-rich region during T cell interaction with superantigen SEE-pulsed antigen-presenting cells ( Fig. 1A-a). In addition, exogenous Swip-1 in Jurkat T or human primary T cells transfected with cDNA encoding GFPtagged swiprosin-1 (GFP_Swip-1) showed similar localization with actin during T cell interaction with anti-CD3/28-coated beads ( Fig. 1A-b). These results suggest that swiprosin-1 is associated with actin function. Next, we tested whether overexpression of swiprosin-1 enhanced T cell adhesion and spreading in response to a chemokine SDF-1a. SDF-1a is a general chemokine for T cells, and most lymphocytes express CXCR4 (C-X-C chemokine receptor type 4), a SDF-1a receptor. In addition, lymphocyte stimulation by SDF-1a induces cell adhesion and spreading [32]. Jurkat T cells that stably overexpressed over 98% GFP or GFP_Swip-1 were established ( Fig. 1A-c, top panel). A time-lapse study showed that overexpression of swiprosin-1 significantly enhanced SDF-1a-mediated T cell spreading on FN (Fig. 1A-c, bottom panel). To better monitor swiprosin-1 localization in various cell lines, we checked the protein level of swiprosin-1 (Fig. 1B). Swiprosin-1 was fairly well detected in Jurkat T, 293T, and HeLa cells, but not in CHO-K1 cells. The fact that CHO-K1 cells express little amount of swiprosin-1 let us to test whether swiprosin-1 exerts a gain-of-function role in CHO-K1 cells. FN is required for the integrin-dependent cell adhesion and migration via regulation of actin cytoskeleton. Moreover, the responsiveness of CHO-K1 cells on FN has been well established [33]. Interestingly, CHO-K1 cells overexpression of GFP_Swip-1 significantly enhanced the spreading of CHO-K1 cells on FN, thereby resulting in an increased area of adhered cells (Fig. 1C-a). In accordance with this observation, the average cell area (Fig. 1Cb), lamellipodium formation (Fig. 1C-c), and Rac1 activity (Fig. 1C-d) were also increased.
Because overexpression of swiprosin-1 enhanced cell spreading and lamellipodia formation in CHO-K1 cells, we examined whether knockdown of this protein also altered cell spreading and hence cell migration. We used HeLa cells, which can spread on FN and have been used for the cell adhesion assay [34]. HeLa cells, transfected with siRNA targeting swiprosin-1, significantly reduced the ability to spread onto FN, 30 and 60 min after plating, compared with scrambled siRNA ( Fig. 2A-a-d). As a result, the average cell area ( Fig. 2A-b) and cell spreading ( Fig. 2A-c) were reduced. Moreover, lamellipodium formation as observed by the F-actin staining was considerably diminished in cells with the siRNA targeting swiprosin-1 ( Fig. 2A-d). In addition, the cells had significantly reduced wound healing compared with the cells transfected with scrambled siRNA (Fig. 2B-a and b). These results clearly demonstrate that swiprosin-1 is involved in cell spreading and migration through remodeling of the actin cytoskeleton.

Swiprosin-1 Binds to the Actin Cytoskeleton in vitro
To investigate the direct interaction between swiprosin-1 and actin in vitro, we performed an in vitro high-speed co-sedimentation assay. High speed actin co-sedimentation assay is a general method of analyzing the specific proteins or protein domains binding to F-actin [30]. We incubated the purified protein (GST_Swip-1 or His_Swip-1) with F-actin, and checked the protein co-sedimentation with F-actin after high-speed centrifugation. Co-sedimentation with purified non-muscle actin and His_Swip-1 revealed that swiprosin-1 binds to F-actin (Fig. 3A). As positive and negative controls, actin-binding protein a-actinin cosedimented with F-actin, while BSA did not. Binding was saturated at a 1.9:1 molar ratio of swiprosin-1 to actin (B max = 2.36360.544 mol/mol) and a K d of 1.91260.788 mM (Fig. 3B). Collectively, these results unambiguously demonstrated that swiprosin-1 is an actin-binding protein. However, in some experiments (Fig. 3C, Fig. 6, Fig. S1A and B, and Fig. S2), we used glutathione-S-transferase (GST)-tagged swiprosin-1 (GST_Swip-1) because some His-tagged mutants of swiprosin-1 were not properly expressed in E. coli, presumably due to structural instability or improper folding. The use of the GST tag facilitated the expression of soluble protein in E. coli. As shown in Fig. 3B and Fig. S1A, co-sedimentation of swiprosin-1 and F-actin was similar regardless of the protein tag, suggesting that GST does not significantly influence the interaction between swiprosin-1 and Factin under the assay conditions. Next, we performed mutation analysis to identify the actin binding site(s) of swiprosin-1. According to our actin binding assay using various deletion mutants, swiprosin-1 contains 3 actin binding sites (69-96, 96-163, and 163-199) situated between 70-199 amino acids (Fig. 3C and Fig. S2A-D). However, no (or a very weak) actin-binding site was identified in the first N-terminal region (amino acids 1-69) and C-terminal coiled-coil region (amino acids 200-240), nevertheless, these sites were required for the actin-binding affinity regulation of swiprosin-1.

Swiprosin-1 Induces Actin Bundling in vitro
Recruitment of swiprosin-1 at the lamellipodium during CHO-K1 cell adhesion on FN (Fig. 1C), and the significantly reduced cell spreading observed by targeted knockdown of swiprosin-1, led us to hypothesize that swiprosin-1 acts as an actin cross-linking protein. To test this hypothesis, we conducted a low-speed actinbundling assay [35] using various concentrations of wild-type Histagged swiprosin-1 (His_Swip-1; 0.25-4 mM). Low-speed actin cosedimentation assay allows efficient pelleting of only heavy, bundled F-actin after incubation of purified swiprosin-1 with Factin. The maximum amount of F-actin in the pellet (P) was recovered with the same concentration of His_Swip-1, as determined by densitometry (Fig. 4A), suggesting that bundling of actin needed a similar molar actin:swiprosin-1ratio. GST-fusion to the swiprosin-1 did not alter the actin bundling activity, but rather slightly increased the activity (Fig. S1B). In addition to the actin-bundling assay, we visualized F-actin bundling by electron microscopy. In the presence of His_Swip-1 or GST_Swip-1, Factin formed clear bundles ( Fig. 4B and C). In contrast, F-actin alone appeared as single filaments, and no bundles were observed.

Calcium Influences the Swiprosin-1-induced Actinbundling Activity but has Little Effect on F-actin Binding
Calcium can affect cellular dynamics by regulating actin bundling by binding to actin bundling proteins [36]. Because swiprosin-1 contains 2 EF-hand motifs, we determined whether calcium ions regulate the F-actin binding affinity or actin bundling activity of swiprosin-1. Co-incubation of His_Swip-1 and F-actin in the presence of 0.1-1 mM EGTA significantly reduced the actin bundling activity (Fig. 5A). Excessive calcium ions had little effect on actin bundling (Fig. 5B). To examine whether the inhibition of actin bundling by EGTA was due to a direct effect on the affinity of swiprosin-1 to F-actin, we performed F-actin binding assays in the presence of EGTA (1-2 mM) or CaCl 2 (1-2 mM). Chelation or excess Ca 2+ had little effect on swiprosin-1 binding to F-actin (Fig. 5C). Furthermore, the inhibition of actin bundling by EGTA was also visualized by electron microscopy (Fig. 5D). In summary, these results demonstrate that calcium is an important regulator of actin bundling activity but has no effect on swiprosin-1 binding to F-actin.
Both EF-hand Motifs and the Coiled-coil Domain are Important for the Actin Bundling Activity of Swiprosin-1 The importance of calcium ions in the regulation of actin bundling by swiprosin-1 led us to investigate which domain of swiprosin-1 is essential for actin bundling. Interestingly, deletion of the EF-hand motifs (D97-163; M2) of swiprosin-1 resulted in considerably less actin bundling activity than that observed with wild-type swiprosin-1, suggesting that Ca 2+ changed conformation of swiprosin-1 that affected the bundling activity. More strikingly, deletion of the coiled-coil domain (1-163; M1) abolished the actinbundling activity, suggesting that the coiled-coil domain is essential for swiprosin-1-induced actin bundling. In contrast to the deletion mutants of the EF-hand motifs and coiled-coil domain, mutants of the N-terminus (97-240 = M3) showed normal actin-bundling activity (Fig. 6A).
Sequence analysis revealed that the coiled-coil region of swiprosin-1 could be clustered to two regions, i.e., the glutamate (Glu)-rich region and the lysine (Lys)-rich region (Fig. 6B-a). To identify which region of the coiled-coil domain is essential for actin bundling by swiprosin-1, we only deleted Glu-or Lys-rich regions in the coiled-coil domain and compared the actin-bundling activity with wild-type swiprosin-1. Surprisingly, deletion of the Lys-rich region (218-240) dramatically reduced actin bundling to a level comparable to the variant with a deletion of the entire coiled-coil domain (200-240), suggesting that the Lys-rich region of the coiled-coil domain is critical for actin bundling (Fig. 6B-b). In contrast, the actin binding activity was not significantly different among the three coiled-coil mutants (Fig. 6B-c).
In Fig. 5A, we demonstrated the requirement of Ca 2+ for swiprosin-1-mediated actin bundling but did not establish the mechanism of its effect. We, therefore, determined the effect of Ca 2+ on the dimerization of swiprosin-1 in cells. Co-immunoprecipitation was performed by treating cells with BAMTA-AM, an intracellular calcium chelator, or ionomycin, a calcium ionopore, before cell lysis. The experiments were performed in the presence of cytochalasin D to exclude the effect of actin polymerization. Treatment with BAMTA-AM significantly reduced the interaction of myc_Swip-1 with GFP_Swip-1 (Fig. 7D), corroborating that calcium affects the self-dimerization of swiprosin-1. However, ionomycin had little effect, thereby suggesting that the endogenous Ca 2+ concentration may be enough to sustain self-dimerization of swiprosin-1 in vivo. To further test whether the dimerization via coiled-coil domain is affected directly by Ca 2+ , we examined the effect of EGTA on the dimerization of wild-type swiprosin-1 (a) and two coiled-coil domain containing mutants (h and i). Interestingly, the binding of all three constructs (a, h, and i) to the His_Swip-1 was reduced in the presence of EGTA (Fig. 7E), suggesting that EGTA directly influences the dimerization activity via affecting the coiled-coil domain and thereby suppresses actinbundling activity of swiprosin-1.

Deletion of the EF-hand Motifs or Coiled-coil Domain Attenuated Swiprosin-1-induced Membrane Lamellipodia
We showed that actin bundling by swiprosin-1 is regulated by EF-hand motifs and the coiled-coil domain (Fig. 6). Then we asked if in vitro function of these domains correlates with the phenotypes observed in vivo. To address this question, CHO-K1 cells were transfected with wild-type or deletion mutants of the coiled-coil domain (M1), EF-hand motifs (M2), and N-terminal region (M3), and the cells were then placed on FN to measure cell spreading and lamellipodia formation. Compared with the wild-type protein, the M1 and M2 mutants revealed a dramatic decrease of lamellipodia formation (Fig. 8A-a and d) and cell spreading ( Fig. 8A-b). Interestingly, these two mutants (M1 and M2) were observed mainly in the cytosolic region and showed no significant localization at the F-actin rich region at the cell periphery (Fig. 8Ac), suggesting a connection between actin bundling activity and lamellipodium formation at the cell periphery. However, the M3 mutant, without the N-terminal region (1-95), showed impaired lamellipodia formation, but the M3 mutant localized well with the F-actin bundles at the cell periphery (Fig. 8A-c). This result suggests that the N-terminal region may have a regulatory role in the lamellipodia formation process, while other regions, including thetwo EF-hand motifs and the coiled-coil domain, facilitate actinbundling in vivo. Fig. 8B summarizes the specific functions of each domain of swiprosin-1. Swiprosin-1 contains three actin-binding sites formed by amino acids 70-200 of swiprosin-1 (Fig. 3C and Fig. S2A-D). Although the three-dimensional structure of swiprosin-1 is unknown, Ca 2+ and the EF-hand motifs may affect the dimer conformation of swiprosin-1, mediated by coiled-coil interactions, which is critical for the regulation of actin bundling. In conclusion, our results demonstrate that the N-terminal 1-70 amino acids may regulate lamellipodia formation. In contrast, the two EF-hand motifs and the coiled-coil domain are indispensable for swiprosin-1-mediated actin bundling; the loss of either of these domains results in the dysfunction of swiprosin-1.

Discussion
Protein functions are oftenly revealed by their localization to specialized cellular sites [37][38][39]. The highest concentration of swiprosin-1 in the microvilli-like protrusive structures and in lamellipodia led us to hypothesize that swiprosin-1 is involved in actin remodeling. This hypothesis was also supported by the evidence that swiprosin-1 is phosphorylated during growth factor-dependent actin remodeling, like other actin-binding proteins such as the Arp2/3 complex or cofilin [25]. Swiprosin-1 is also involved Swiprosin-1 Regulates Actin Bundling in cytokine regulation, potentially through actin remodeling in mast cells [21]. Therefore, our initial approach was to determine whether swiprosin-1 is involved in the regulation of cell spreading and migration. Through a combination of loss-of-function and gain-of-function approaches, we found that swiprosin-1 is important for lamellipodial protrusion and that swiprosin-1 induces cell spreading and migration. We, therefore, obtained an insight into the fundamental mechanism of how swiprosin-1 regulates actin Swiprosin-1 Regulates Actin Bundling remodeling in vitro and in vivo. Mutagenic analysis and in vitro actin-binding assays revealed that swiprosin-1 contains $3 actinbinding sites (N-terminus, EF-hands, and C-terminus) and one coiled-coil domain required for self-dimerization. Swiprosin-1 binds to actin in an apparent molecular ratio of 1.9:1, similar to other actin-binding proteins such as plectin, dystrophin, a-actinin, Swiprosin-1 Regulates Actin Bundling and dystonin [40][41][42]. Similarly to these actin-binding proteins, we found that swiprosin-1 has an actin-bundling activity through the actin binding sites and a coiled-coil domain.
Among various swiprosin-1 constructs, mutants M1 (C-terminal deletion) and M3 (N-terminal deletion) increased the binding affinity of swiprosin-1 to the F-actin. This phenomenon was intriguing because these two mutants have only two actin binding sites, while full-length of swiprosin-1 has at least three regions. One possible explanation is that although swiprosin-1 has three actin binding sites, it also has other sites that can potentially influence the strength of actin binding. Deletion of potential regulatory regions in mutants (M1 = deletion of coiled-coil region+intermediate region between EF and CC domain; M2 = deletion of Nterminal regulatory region) may increase the strength of actin binding.
Interestingly, calcium chelation by EGTA reduced swiprosin-1mediated actin-bundling activity; however, EGTA did not significantly affect the actin-binding activity. Thus, Ca 2+ may change the conformation of swiprosin-1 to preferentially induce actin bundling. Indeed, we found that the coiled-coil domain is more essential than the EF-hand motifs for the actin bundling activity because the loss of one or two actin-binding site(s) can also induce actin bundling if there is a coiled-coil domain. Nonetheless, we also consider the EF-hand motif to be important because a mutant without two EF-hand motifs showed reduced actinbundling activity, even if the protein did not lose its actin-binding activity. This suggests that a Ca 2+ -induced conformational change of the EF-hand motifs may be one of the mechanisms underlying swiprosin-1 actin-bundling activity. However, the current results showing that EGTA significantly reduces the dimerization of coiled-coil domain suggest that Ca 2+ also directly influences the dimerization activity via affecting the coiled-coil domain and thereby regulates the actin-bundling activity of swiprosin-1. Therefore, Ca 2+ may have at least one of the following two roles: 1) influencing the structural conformation of EF-hand motifs, thereby affecting dimerization; 2) stabilizing the dimeric conformation via coiled-coil domain independently of the EF-hand motifs.
Dimerization is crucial for the actin cross-linking function of most actin-bundling proteins, including a-actinin, filamin, and proteins of the b-spectrin family [4,6,43,44]. For example, the in vitro formation of plectrin dimers and oligomers are likely mediated via coiled-coil interactions of rod domains [45]. Similarly, a coiled-coil domain of swiprosin-1 served as a dimerization site and was essential for actin bundling in vitro. In Fig. 8, we showed that GFP_Swip-1 induced formation of lamellipodia in the cell membrane region, but coiled-coil mutant M1 and EF-hand mutant M2 exhibited poor formation of lamellipodia and these mutants were mainly localized in the cytosol. Coiled-coil domain mutant M1 also exhibited poor formation of lamellipodia, suggesting that dimerization is essential for swiprosin-1 function in vivo. More detailed functional analysis revealed that swiprosin-1 dimerization and actin cross-linking are inseparable events that are indispensable for actin bundling by swiprosin-1.Furthermore, Ca 2+ is important for actin bundling as it stabilizes the swiprosin-1 dimer conformation.
In Fig. 8 we showed that the truncated version EF hand and CC are mainly localized in the cytosol. This is comparable to data published recently [22]. In the report, authors demonstrated that truncated versions (EF-hand and coiled-coil region) of Drosophila Swiprosin-1 are also localized in the cytosol, whereas the endogenous protein localizes to specific foci at the fusion site at the membrane.
Signals induced by cell adhesion to the extracellular matrix regulate important physiological events, including cell motility and growth, which most often involve reorganization of the actin cytoskeleton. The lamellipodium is the thin, sheet-like, leading edge of a cell that is moving on a solid substrate and is the result of many structural proteins and signaling proteins working in concert [44]. The activity of swiprosin-1 to induce lamellipodia foremation on the FN matrix suggests that this protein may also function as a small adaptor protein. Similarly, the actin-bundling protein filamin serves as a molecular linker between the membrane and the cytoskeleton, recruiting signaling proteins that regulate actin polymerization and remodeling [46,47]. Filamin contains a docking site for RhoGEF and a small GTPase [46,48]. Accordingly, we found that overexpression of swiprosin-1 increases the activity of the small GTPase, Rac1 (Fig. 1D-d). This activity may be important for lamellipodia induction in living cells. Previous studies reported that swiprosin-1 has three phosphorylation sites; serine phosphorylation sites at serine residue 74 and 76, and tyrosine phosphorylation at residue 83 [49][50][51]. In addition, phosphorylation of swiprosin-1 is important for its association with signaling molecules [20]. Most actin regulatory molecules are regulated by specific kinases or phosphatases. For example, bundling of actin by fascin is inhibited by serine phosphorylation in vitro [52,53] and in vivo [54]. In Fig. 8A, we showed deletion of the N-terminal region also decreases lamellipodia formation without the loss of F-actin localization at the cell periphery even though they had less actin bundling activity. It suggests that phosphorylation site(s) in N-terminal region of swiprosin-1 has regulatory function of lamellipodia formation, thereby modulating cell spreading.
In conclusion, we found that swiprosin-1 is a novel actinbinding protein that modulates actin bundling. We identified that swiprosin-1 has three actin-binding sites in the N-terminal, EFhand, and C-terminal regions. The occupation of the EF-hand motifs by Ca 2+ may induce a conformational change that regulates dimerization via the coiled-coil domain. In other words, calcium may directly affect swiprosin-1 dimer stability, and this structural change may regulate the in vitro and in vivo function of swiprosin-1. Actin-bundling proteins can influence the initiation, organization, and stabilization of actin bundles; however, how the actinbundling proteins regulate these functions is poorly understood. Our findings provide new insights into actin assembly by the novel actin-regulating protein, swiprosin-1. Figure S1 Actin-binding and actin-bundling activity of GST-tagged swiprosin-1. (A) GST_SW-1 (1-15 mM) was mixed with F-actin (4 mM) and subjected to the F-actin binding assay as described in Fig. 3B. The actin-binding affinity was measured as described in the Materials and methods. The  Fig. 3C. F-actin (2 mM) was incubated with GST, GST_SW-1, or the indicated mutants, and the actin-binding activity was then determined in the co-sedimentation assay. The actin-binding strength was quantified and scored as shown in Fig. 3C. (TIF) Author Contributions