F-actin Binding Region of SPIN90 C-terminus Is Essential for Actin Polymerization and Lamellipodia Formation

Abstract

We recently reported that SPIN90 is able to bind with several proteins involved in regulating actin cytoskeleton networks, including dynamin, WASP, β PIX, and Nck. Based on these findings, we investigated how SPIN90 regulates the actin cytoskeleton and promotes actin assembly. This study demonstrated that aluminium fluoride-induced localization of SPIN90 to lamellipodia requires amino acids 582-722 at the SPIN90 C-terminus, which is also essential for F-actin binding and Arp2/3 complex mediated polymerization of actin into branched actin filaments. Furthermore, after deletion of the F-actin binding region (582-722 SPIN90) failed to localize at the membrane edge and was unable to promote lamellipodia formation, suggesting that the F-actin binding region in the SPIN90 C-terminus is essential for the formation of branched actin networks and regulation of the actin cytoskeleton at the leading edge of cells.

Keywords:

• SPIN90
• Arp2/3 complex
• lamellipodia
• actin polymerization
• F-actin
• actin filament branching

INTRODUCTION

Cells move by extending a leading edge via the coordinated assembly of actin filaments, which are then organized into mechanically resilient networks (24). The Arp2/3 complex localizes to regions of actin assembly and forms dendritic actin networks in several motile cellular structures, including the lamellae of the leading edge and the actin comet tails found behind endosome-derived vesicles (2, 15). In these structures, the evolutionarily conserved seven-subunit Arp2/3 complex, which is composed of the two large subunits (Arp2 and Arp3) and five smaller subunits (p40, p35, p21, p20 and p19), is thought to provide a nucleation for actin assembly (14, 21). The Arp2/3 complex is activated by binding with a number of proteins including N-WASP, Scar/WAVE, CARMIL, and Cortactin (4, 7, 23). The activated complex binds to the sides of preexisting F-actin filaments, promoting de novo development of a branching network called the dendritic nucleation (18, 20).

SPIN90, an Nck-binding protein that contains an SH3 domain, three proline-rich regions, a serine/threonine rich region, and a long C-terminus containing a hydrophobic region of unknown function, has been strongly implicated in the regulation of the actin cytoskeleton. For example, we recently reported that SPIN90 can bind to several proteins, including dynamin, WASP, βPIX, and Nck (9, 10, 12, 13). SPIN90 has also been shown to binds to G-actin and Arp2/3 complex via its C-terminus and co-localizes with the Arp2/3 complex at the cell periphery when stimulated by growth factor (9). Lamellipodia formation was inhibited in SPIN90-knockdown cells, suggesting that SPIN90 participates in actin assembly of motile cellular structures (9). However, the molecular mechanisms by which SPIN90 mediates actin assembly and lamellipodia formation have not yet been fully elucidated.

In this study, we investigated how SPIN90 regulates the actin cytoskeleton and interacts with F-actin to promote actin assembly. Our experiments demonstrated that the aluminium fluoride-induced localization of SPIN90 to lamellipodia requires amino acids 582-722 at the SPIN90 C-terminus, which is also essential for F-actin binding and Arp2/3 complex mediated actin polymerization in the presence of G-actin binding domain.

METHODS

Plasmid Construction and Antibodies

cDNAs encoding full-length SPIN90 (amino acids 1-722) and various deletion mutants (an N-terminus truncation containing amino acids 1-268, a C-terminus truncation containing amino acids 269-722, and C-term truncation variants containing amino acids 269–656, 269–557, 269–581, 543–722, 582-722, 657-722) were subcloned into pGEX4T-1 (Amersham Pharmacia Biotech, Piscataway, NJ), HA pcDNA 3.0, His C pcDNA 3.1 (Invitrogen, Carlsbad, CA). The polyclonal anti-SPIN90 antibody was generated in rabbits immunized with purified GST-SPIN90-N-terminus (amino acids 1–268). The antibody was affinity-purified using the same GST fusion protein that served as the immunogen. The anti-β -actin antibody was purchased from Sigma (St. Louis, MO); the monoclonal anti-HA antibody (16B12) and the anti-His antibodies were obtained from Covance (Berkeley, CA) and Cell Signaling (Beverly, MA), respectively, and the anti-Src and anti-tubulin antibodies were purchased from Upstate (Charlottesville, VA) and Sigma, respectively.

Protein Preparation and Purification

The SPIN90 constructs were transformed into Escherichia coli strain BL21(DE3). Bacteria were grown to an optical absorbance of 0.6 A₆₀₀ and induced by incubation with 0.5 mM isopropyl-1-thio-β -D-galactosidase (IPTG) for 3 h at 30°C. The bacteria were harvested by centrifugation and resuspended in 1/10 the starting culture volume of phosphate-buffered saline (PBS) containing 2.5 mM dithiothreitol and 20 μ g/ml phenylmethylsulfonyl fluoride. The cells were lysed by sonication, and the lysate was clarified by centrifugation at 12,000 × g for 30 min at 4°C, followed by filtration through a 0.22-μ m sterilized filter. Each sample was loaded onto glutathione-Sepharose 4B (Amersham Biosciences, Piscataway, NJ), which was then washed extensively with PBS. The bound proteins were eluted with 10 mM glutathione in PBS, and the purity of the eluted proteins was confirmed by SDS-PAGE followed by Coomassie Blue taining.

Cell Culture and Immunofluorescence

MDA-MB-231 and HEK 293T cells were grown in DMEM supplemented with 10% FBS at 37°C in 5% CO₂. Aluminium fluoride treatment of cells was performed by supplementing the culture medium with 30 mM NaF and 50 μ M AlCl₃. For immunofluorescence analysis, cells were washed 3 times with PBS containing 1 mM CaCl₂ and 1 mM MgCl₂, fixed for 10 min in 3.5% paraformaldehyde, and then permeabilized for 10 min in 0.2% Triton X-100 . Once permeabilized, the cells were incubated at room temperature, first with primary antibodies for 1 h, and then with fluorophore-conjugated secondary antibodies for an additional 50 min. For assessment of the cytoskeleton, filamentous actin was labeled with Texas Red-phalloidin (Molecular Probes, Eugene, OR) and visualized using a Leica DMRBE microscope equipped with a 63 × (1.4 NA) oil objective lens and FITC-or Texas Red-optimized filter sets (Omega® Optical Inc., Brattleboro, VT).

Immunoblotting

HEK 293T cells expressing his-tagged SPIN90 fragments were washed with cold PBS, extracted with cytoskeleton isolation buffer (1% Triton X-100 in 80 mM PIPES, pH 6.8, 5 mM EGTA, 1 mM MgCl₂) supplemented with protease inhibitors and immediate centrifugation at 3000 × g for 2 min yielding detergent-soluble and insoluble fraction. The detergent insoluble pellets were sonicated for 30 s, centrifuged at 100,000 × g for 30 min, and then subjected to SDS-PAGE and Coomassie Blue staining. For depolymerization of F-actin, latrunculin B (Sigma) was added to HEK 293T cells transfected with his-tagged SPIN90 fragments at the indicated concentrations.

Actin Polymerization Assay

Actin polymerization was measured by monitoring the change in fluorescence intensity of pyrene-labeled actin as described previously (16, 25). Briefly, rabbit skeletal muscle monomeric actin was labeled with pyrene with 10% efficiency and then mixed with unlabeled G-actin (Cytoskeleton Inc., Denver, CO) at a ratio of 1:4 in G-actin buffer (5 mM Tris-HCl, pH 8.0, 0.2 mM CaCl₂, 0.5 mM DTT and 0.2 mM ATP) and then centrifuged at 200,000 × g for 1 h to remove residual filamentous actin. 50 nM Arp2/3 complex (Cytoskeleton Inc., Denver, CO) and 200 nM of the appropriate SPIN90-C-terminus fragments were mixed with polymerization buffer (5 mM Tris, pH 7.5, 1 mM EGTA, 0.1 mM CaCl₂, 0.5 mM DTT, 3mM NaN₃, 50 mM KCl, 2 mM MgCl₂, and 0.2 mM ATP), and polymerization was initiated by adding G-actin stock. The final concentration of G-actin in the polymerization reaction mixture was 2.5 μ M. Changes in fluorescence were then monitored for 30 min using a spectrofluorometer (SHIMADZU RF-5301 PC, Kyoto, Japan) with excitation at 365 nm and emission at 407 nm.

Branching Assay

Branching assay was modified as described previously (1). Briefly, GST-SPIN90-C-terminus fragments (600 nM), GST-VCA (400 nM), Arp2/3 complex (100 nM), and actin monomers were mixed in Mg-G buffer containing 1 × KMEI (100 mM KCl, 2 mM MgCl₂, 1 mM EGTA, 20 mM imidazole, pH 7.0, 0.2 mM ATP, 1 mM DTT). Polymerization was allowed to proceed for 15 min at 22°C, and then an equivalent molar ratio of Texas Red-phalloidin was added with gentle mixing. Samples were diluted 200–400-fold in fresh fluorescence buffer (100 mM KCl, 1 mM MgCl₂, 100 mM DTT, 20 mM imidazole, pH 7.0, 0.5% methylcellulose, 20 μ g ml⁻¹ catalase, 100 μ g ml⁻¹ glucose oxidase, 3 mg ml⁻¹ glucose, 0.2 mM ATP, 3 mM NaN₃), and 2 μl of the mixture was applied to coverslips previously coated either with 0.1% nitrocellulose in amyl acetate or for 10 min with 20 μ g ml⁻¹ poly-L-lysine. Actin filaments were examined using a Leica DMRBE microscope. We performed quantitative analyses of actin filament branching. The pictures were digitalized and measured manually using MetaMorph imaging software (Universal Imaging Corporation, Downingtown, PA).

F-actin Cosedimentation Assay

The GST-tagged SPIN90 fragment proteins (1 μM) were mixed with different concentrations of F-actin in 10 mM imidazole, pH 7.0, 75 mM KCl, 0.5 mM ATP, 0.5 mM EGTA, 0.5 mM DTT, and 5 mM MgCl₂. Mixtures were incubated at room temperature for 30 min and then centrifuged at 160,000 × g for 45 min at 4°C. Supernatants and pellets were resolved by SDS-PAGE and the bands were visualized by Coomassie Blue staining. The band intensities were analyzed with a scanning densitometer (GS-710, Bio-Rad, Hercules, CA). The binding constants were calculated with the data from a scanning densitometer using ORIGIN software® (Microcal, Northampton, MA).

RESULTS

The C-Terminus Fragment of SPIN90 (Amino Acids 582-722) Associates with the Actin Cytoskeleton

To determine the precise binding domain responsible for the association of SPIN90 with actin cytoskeleton, we transfected HEK 293T cells with constructs encoding his-tagged SPIN90-full, -C-terminus (269-722), and -N-terminus (1-268). The cytoskeleton fractions were isolated with cytoskeleton extraction buffer, which yields Triton X-100 (Tx-100) soluble and insoluble (actin cytoskeleton enriched) fractions. This result revealed that SPIN90-N-terminus (1–268) did not accumulated in the Tx-100 insoluble fraction, indicating that the SPIN90-N-terminus does not associate with the actin cytoskeleton (Figure 1A). In contrast, SPIN90-full, -C-term (269-722) and -C-term (582-722) all showed positive associations with the actin cytoskeleton (Figure 1B), indicating that the actin cytoskeleton binding region lies within the C-terminal 141 amino acids (582-722) of SPIN90. This was further confirmed by the finding that deletion of amino acids (582-722) disrupted the association with actin cytoskeleton (Figure 1B).

Figure 1 The SPIN90-C-term (582-722) accumulates in the detergent insoluble cellular fraction. (A) HEK 293T cells expressing his-tagged SPIN90-full, -N-terminus, and -C-terminus were lysed in cytoskeleton isolation buffer containing 1% Triton X-100 and directly subjected to centrifugation. The detergent soluble and insoluble fraction were inmmunoblotted to anti-his and anti-β -actin antibodies. (B) His-tagged-SPIN90-full, SPIN90-C-terminus (269-722) and SPIN90-C-term fragments (269-581 and 582-722) were expressed in HEK 293T cells and the detergent insoluble and soluble fractions were detected as described in (A). Arrow indicates the SPIN90-C-term fragments. Non-specific binding is shown by the arrow with the asterisk. (C) Mapping of the actin cytoskeleton binding site of SPIN90 from (A) and (B). The amino acids of SPIN90 are indicated (Full length: amino acids 1-722, N-terminus: amino acids 1-268, and C-terminus: amino acids 269-722). Actin cytoskeleton association represents the band present in Tx-100 insoluble fraction (closed rectangle) and absent in Tx-100 insoluble fraction (open rectangle). (D) Cells expressing the actin cytoskeleton-associated SPIN90-C-term fragment (582-722) treated with 5 μ M latrunculin B for 30 min, lysed and fractionated as described in (A), and immunoblotted with anti-his, -β -actin, tubulin, and Src antibodies.

To determine whether this actin cytoskeleton binding was specific to filamentous actin (F-actin), we performed Tx-100 solubility assays followed by immunoblotting. c-Src, which is known to associate with actin cytoskeleton, was present mainly in the Tx-100 insoluble fraction and tubulin was exclusively present in the Tx-100 soluble fraction, (Figure 1D). Without treatment with latrunculin B, an F-actin depolymerizing drug (3), his-tagged SPIN90-C-term (582-722) was found in the Tx-100 insoluble fraction, indicating the association of SPIN90 and actin cytoskeleton. Treatment of latrunculin B for 30 min dramatically released filamentous actin, Src, and his-tagged SPIN90-C-term (582-722) into the Tx-100 soluble fraction, due to the efficient depolymerization of the actin cytoskeleton (Figure 1D), supporting further that SPIN90 closely associates with F-actin.

SPIN90-C-Term (657-722) is Required for F-Actin Binding

To identify the region of SPIN90 responsible for F-actin binding, we examined the abilities of fusion proteins containing GST and full-length or various truncated forms of SPIN90 to co-sediment with F-actin in vitro, which is a commonly used assay for F-actin binding (25, 26). Our results revealed that GST and GST-SPIN90-N-terminus failed to cosediment with F-actin (Figure 2A), whereas GST-SPIN90-full and -C-terminus (269-722) co-sedimented with F-actin. From there, we focused on SPIN90-C-terminus (269-722), which contains an Arp2/3 complex binding domain (amino acids 336-407), a G-actin binding domain (amino acids 527-549), and additional sequences of unknown functions. Similar F-actin co-sedimentation assays were performed to avoid overlapping of the band representing GST-SPIN90-C-term (582-722; ∼ 45kDa) with that of actin (43 kDa), we generated a GST-construct including amino acids 657-722 containing a highly basic and hydrophobic region (19, 22, 27). Our results revealed that GST alone and GST-C-term (269-656) failed to co-sediment with F-actin, whereas GST-C-terminus (269-722) and GST-C-term (657-722) both bound F-actin (Figure 2B). These results suggest that the region responsible for SPIN90 binding to F-actin is located at the SPIN90-C-term (657-722), implicating that the basic and hydrophobic amino acids are essential for F-actin binding of SPIN90 (11, 19, 22, 27). When we estimated the dissociation constant (K_d) for a SPIN90-full, -C-terminus (269-722) and -C-term (657-722) with respect to F-actin, we found that the association of the GST-SPIN90-full, -C-terminus (269-722), and -C-term (657-722) with F-actin became saturated as the F-actin concentration increased. Plotting the association of the GST-SPIN90-full, -C-terminus (269-722), and -C-term (657-722) with respect to F-actin revealed that the K_d values for actin filament association were 0.72, 0.93 and 1.81 μ M, respectively (Figure 2C), indicating that the affinity of SPIN90-C-term (657-722) for F-actin was weaker than those of -C-terminus (269-722) and -full length. This result suggests that the F-actin binding region alone is insufficient for the optimum F-actin binding activity and requires additional domains of SPIN90.

Figure 2 Binding of SPIN90 to F-actin. GST-SPIN90-full (1-722), -N-terminus (1-268), -C-terminus (269-722) (A) and GST-SPIN90-C-term (269-722, 269-656, and 657-722) (B) fusion proteins (1 μ M) were mixed with or without polymerized F-actin (3 μM) and co-sedimented for 45 min at 180,000 × g, and pellets were analyzed by SDS-PAGE, followed by staining with Coomassie Blue. (C) GST-SPIN90-full, -C-terminus and -C-term fragment (657-722) proteins were cosedimented with various concentrations of F-actin. Coomassie staining was quantified densitometrically and normalized to the percentage of depletion, and these values were used to calculate the indicated apparent K_d values for the various SPIN90 fusion proteins. Values are given as mean and S.D. of three independent experiments.

The F-Actin Binding Region of SPIN90 is Required for Arp2/3 Complex Activation

Previously, we reported that actin polymerization could be observed in vitro by detection of increased fluorescence as pyrene-labeled actin monomers were assembled into filaments in the presence of SPIN90 protein (9). Since our results indicated that the SPIN90-C-terminus contains Arp2/3 complex, G-actin and F-actin binding regions, we next performed the fluorescent-based assay to examine which domain of SPIN90 is essential for Arp2/3 complex-mediated actin polymerization. When C-term fragments containing the F-actin binding region (543-722, 582-722, and 657-722) were mixed with Arp2/3 complex, we were unable to detect more than basal fluorescence (Figure 3B). Similarly, a C-term construct (269-581) containing a G-actin and an Arp2/3 complex binding regions, did not activate Arp2/3 complex-mediated actin polymerization. In contrast, introduction of the whole SPIN90-C-terminus (269-722) significantly induced Arp2/3 complex-mediated actin polymerization (Figure 3B). Addition of the SPIN90-C-terminus in the absence of the Arp2/3 complex failed to stimulate actin assembly. These results suggest that Arp2/3 complex-mediated actin polymerization requires the G-actin and Arp2/3 complex binding region as well as the F-actin binding region of SPIN90.

Figure 3 The F-actin binding region of SPIN90 is important for Arp2/3 complex-mediated actin polymerization. (A) Purified GST-SPIN90-C terminus protein fragments (269-722, 269-581, 543-722, 582-722 and 657-722) were separated by SDS-PAGE and stained with Coomassie Blue for visualization of GST fusion proteins. (B) Time course of actin polymerization in the presence of 50 nM Arp2/3 complex and 200 nM SPIN90-C-terminus fragments with 10% pyrene labeled G-actin. Fluorescence intensity is shown in arbitary units.

The F-Actin Binding Region of SPIN90 Activates Filament Branching via the Arp2/3 Complex

Previously studies have shown that WASP family proteins containing VCA (verprolin (V) homology domain, a cofilin (C) homology domain and acidic (A) region) are capable of Arp2/3 complex-induced actin filament branching (5). To determine whether SPIN90-C-terminus (269-722) activates the Arp2/3 complex in a similar manner, the filament branching activity assay was performed. Microscopic observation for actin filament assembly revealed that 59% of the actin filament contained branches in the presence of N-WASP VCA (positive control) and the Arp2/3 complex (Figure 4B). In the presence of the SPIN90-C-terminus (269-722) and the Arp2/3 complex, 39% of actin filaments were branched (Figure 4B). Various types of branches were seen, including single filaments, single filaments with single branches, and single filaments with multiple branches (Figure 4C); there was a higher percentage of single-branched (28%) versus multiple branched (11%) filaments (Figure 4D). In contrast, the SPIN90-C-terminus (269–722) alone (data not shown) or Arp2/3 complex alone did not produce actin filament branches (Figure 4A). The Arp2/3 complex plus SPIN90-C-term fragments (269–557, 582-722, and 657–722) showed few branched filaments (6.1, 4.3, and 4%, respectively Figure 4A, Figure 4B). In conjunction with the results of our actin polymerization assay (Figure 3), these findings suggest that the F-actin binding region of SPIN90 as well as G-actin or Arp2/3 complex binding regions of SPIN90 is essential for activation of the Arp2/3 complex and formation of branched actin filaments.

Figure 4 Images of actin filament branch formation by SPIN90. (A) Actin monomers (4 μ M) were polymerized in the presence of 100 nM Arp2/3 complex with 600 nM of the indicated SPIN90-C-terminus fragments. The resulting actin filaments were stabilized, labeled with an equimolar amount of rhodamine-phalloidin, and immediately applied to coverslips for microscopy. Scale bar represents 5 μ m. (B) Quantitative analysis of filament branching. Arp2/3 complex was incubated with G-actin and SPIN90-C-terminus fragments or N-WASP VCA as described in the Material and Mehtods section. The values are presented as a percent of the total population of actin filaments that appear in branched structures (SPIN90-C-terminus, n = 6; N-WASP VCA, n = 6; ** p < 0.014, Student's t test). (C) Images of the different types of branched filaments formed in the presence of the SPIN90-C-terminus. (D) The actin filament branches were subcategorized into those having one branch versus those having two or more branches. Error bars represent the S.D from three independent assays.

The F-Actin Binding Region of SPIN90 is Required for Its Localization to Lamellipodia

Aluminium fluoride has been shown to act as a phosphate analogue stabilizing actin filaments and promoting lamellipodia formation (6, 8). In order to examine the role of SPIN90 in the formation of lamellipodia, MDA-MB-231 cells were treated with aluminium fluoride and the formation of lamellipodia was examined in the cell leading edges. Analysis of immunofluorescence images showed that endogenous SPIN90 was widely distributed in the cytoplasm of resting cells (Figure 5A). In contrast, aluminium fluoride-treated cells showed redistribution of SPIN90 to the tips of lamellipodia, where the SPIN90 co-localized with phalloidin-stained F-actin (Figure 5A). Comparison of lamellipodia formation revealed that ∼ 60% of aluminium fluoride-treated cells showed lamellipodia, versus only 10% of untreated cells (Figure 5B). Immunoblot experiment revealed that SPIN90 expression was significantly increased following aluminium fluoride treatment, peaking 20–30 min post-treatment (Figure 5C). To examine the involvement of the F-actin binding region in this effect, we transfected cells with a vector encoding an HA-SPIN90-full length, an HA-tagged F-actin binding region-truncated construct of SPIN90 (Δ 582-722), or empty vector. Cells expressing HA-SPIN90-full length showed dramatically increased lamellipodia in response to aluminium fluoride (Figure 5D, right panel) and prominent localization of SPIN90-full length with F-actin in the lamellipodia. In contrast, cells expressing the deletion construct (Δ582–722) did not differ from controls in their response to aluminium fluoride treatment. SPIN90 (Δ582–722) failed to localize at the cell edge and instead remained in the cytoplasm (Figure 5D, right panel). These results indicate that deletion of SPIN90-C-term (582-722) (including the F-actin binding region) greatly changes the ability of SPIN90 to localize to the cell edge in response to aluminium fluoride, suggesting that the F-actin binding region of SPIN90 is likely to participate in the regulation of actin dynamics as well.

Figure 5 Ablation of the SPIN90-C-term (582-722) abolishes aluminium fluoride induced lamellipodia formation. MDA-MB-231 cells were untreated or treated with aluminium fluoride for 20 min and then immunostained with rabbit anti-SPIN90 antibody followed by visualization with an FITC-conjugated secondary antibody (green) or Texas Red-phalloidin to visualize actin (red). Merged images are shown in yellow. SPIN90 protein was distributed along the plasma membrane at the peripheral edge in the aluminium fluoride treated cells. Scale bar represents 10 μ m. (B) Quantification of cells with lamellipodia versus those without in the presence or absence of aluminium fluoride. The values shown represent the means ± S.D. of triplicate experiments. (C) Immunoblot analysis with anti-SPIN90 antibody at the indicated times after aluminium fluoride treatment. (D) Cells expressing empty HA-vector, HA-SPIN90-full and HA-SPIN90-Δ 582-722 (amino acids 1-581) were treated or not treated with aluminium fluoride for 20 min and then immunostained with anti-HA antibody followed by visualization with a fluorescein isothiocyanate-conjugated secondary antibody (green) and Texas Red-phalloidin staining of actin (red). Merged images are shown in yellow. Scale bar represent 10 μ m.

DISCUSSION

We previously reported that SPIN90 is able to bind with several proteins related to the regulation of actin cytoskeleton networks, including dynamin, WASP, β PIX, and Nck (9, 10, 12, 13). Furthermore, the SPIN90 protein sequence contains Arp2/3 complex and G-actin binding region known to mediate Arp2/3 complex activation and actin comet tail formation (9). SPIN90 is prominently localized to the lamellipodia of cells treated with growth factors such as PDGF (9). Collectively, these previous findings indicate that SPIN90 plays pivotal roles in reorganization of the actin cytoskeleton and in actin based motility. However, the binding region of SPIN90 with F-actin has not been identified yet. In this study, we show that aluminium fluoride treatment induced the relocalization of SPIN90 to cell edges and triggered the formation of lamellipodia, in which the association of SPIN90-C-term end (582-722) with F-actin is essential (Figure 1).

Other F-actin binding proteins, such as N-WASP, Btk, and p57 have been shown to possess basic and hydrophobic amino acid-enriched regions that interact with corresponding acidic amino acid clusters on actin (19, 22, 27). Consistent with this, we found that SPIN90 contains a highly hydrophobic and basic region at the end of its C-term (582-722). Deletion of this region in SPIN90 abolished the F-actin binding in vitro, Arp2/3 complex mediated actin polymerization, actin filaments branching, and formation of lamellipodia in aluminum fluoride-treated cells. SPIN90-C-term (269-581) containing the Arp2/3 complex and G-actin binding domains but not the F-actin binding region failed to initiate actin polymerization and formation of branched actin filaments, presumably due to the lack of the F-actin binding region and/or change of protein conformation that affected the binding affinity for the Arp2/3 complex. Our finding contrasts with previous reports that the Arp2/3 complex and G-actin binding domains are sufficient to induce actin polymerization of WASP family proteins (20, 23), perhaps due to differences in domain structure between SPIN90 (Arp2/3 complex binding region—G-actin binding region—F-actin binding region) and WASP family members such as WASP-VCA (G-actin binding region—Central region—Arp2/3 complex binding region) (23). The deletion of F-actin binding region (Δ582-722) might induce the change of protein conformation and thereby results in the decrease of binding affinity with Arp2/3 complex. Therefore, F-actin binding region is essential for proper localization and recruitment of Arp2/3 complex. For instance, NTA (N-terminal Arp2/3 complex binding region) of Cortactin failed to target to the cell cortex in PDGF-treated cells and abolished actin polymerization activity in a pyrene actin polymerization assay, whereas NTA plus F-actin binding region was successfully targeted to the cell cortex and showed actin polymerization activity under these conditions (26). Taken together, these previous studies and our present results suggest that the F-actin, the G-actin, and the Arp2/3 complex binding domain in SPIN90-C-terminus may act synergistically to recruit Arp2/3 complexes to actin filament sides, triggering Arp2/3 complex conformational change and subsequent formation of dendritic actin networks (Figure 6). Although future studies will be required to elucidate how SPIN90 interacts with the various subunits of the Arp2/3 complex (a seven protein complex containing Arp2, Arp3, and five unrelated polypeptides) (17), regulates actin dynamics, and /or modulates the action of various actin-binding regulatory proteins, our present findings provide new important insights into the mechanisms underlying SPIN90-mediated Arp2/3 complex activation, actin polymerization, and lamellipodia formation. It is also very important to elucidate the spatial and temporal control mechanism of branching networks between the various subunits of Arp2/3 complex and SPIN90.

Figure 6 Model for Arp2/3 complex activation by SPIN90. The model depicts SPIN90 associated with Arp2/3 complex (A), G-actin (V) and F-actin (F) via its C-terminus and with other actin regulatory proteins (WASP and NCK) via its N-terminus (SH3 and PRD domains).

ACKNOWLEDGMENTS

This study was supported in part by grants from the National Research Laboratory and Molecular & Cellular BioDiscovery Research Group (The Ministry of Science & Technology).