Regulation of Actin Ring Formation by Rho GTPases in Osteoclasts*

Actin ring formation is a prerequisite for osteoclast bone resorption. Although gelsolin null osteoclasts failed to exhibit podosomes, actin ring was observed in these osteoclasts. Wiscott-Aldrich syndrome protein (WASP) was observed in the actin ring of gelsolin null osteoclast. Osteoclasts stimulated with osteopontin simulated the effects of Rho and Cdc42 in phosphatidylinositol 4,5-bisphosphate (PIP2) association with WASP as well as formation of podosomes, peripheral microfilopodia-like structures, and actin ring. To explore the potential functions of Rho and Cdc42, TAT-mediated delivery of Rho proteins into osteoclasts was performed. Although Rho and Cdc42 are required for actin ring formation, transduction of either one of the proteins alone is insufficient for this process. Addition of osteopontin to osteoclasts transduced with Cdc42Val12 or transduction of osteoclasts with both RhoVal14 and Cdc42Val12 augments the formation of WASP-Arp2/3 complex and actin ring. Neomycin, an antibiotic, blocked the effects of osteopontin or TAT-RhoVal14 on PIP2 interaction with WASP. WASP distribution was found to be cytosolic in these osteoclasts. Depletion of WASP by short interfering RNA-mediated gene silencing blocked actin polymerization as well as actin ring formation in osteoclasts. These results suggest that Rho-mediated PIP2 interaction with WASP may contribute to the activation and membrane targeting of WASP. Subsequent interaction of Cdc42 and Arp2/3 with WASP may enhance cortical actin polymerization in the process of actin ring formation in osteoclasts.

Phosphoinositides are involved in modulating a variety of actin regulatory proteins (1) as well as promoting filament cross-linking to form stable and bundled actin fibers (2). Phosphoinositides have been identified to have a major role in gelsolin function, in the regulation of actin organization, and podosome assembly/disassembly in both mouse and avian osteoclasts (3)(4)(5). The sequence QRLFQVKGRR in the second phosphoinositides-binding domain (PBD) 2 of gelsolin has been shown to compete with the function of endogenous gelsolin domains for binding phosphoinositides when introduced into fibroblasts, platelets, and neutrophils. These cells exhibited a block in actin assembly as well as motility (6 -8). We have used these peptides to delineate phosphoinositides-mediated signaling in actin reorganization, podosomes assembly/disassembly, and bone resorption. Transduction of PBD peptides of gelsolin into osteoclasts produced clusters of podosomes and disrupted the formation of the actin ring. Hence, these osteoclasts were hypomotile and less resorptive (9). Actin ring formation is critical for osteoclast bone resorption but is not required for motility. Transduction of PBD of gelsolin not only blocked the interaction of PIP 2 with endogenous gelsolin but also with WASP and ezrin proteins (9).
WASP has been identified as a phosphoinositide-binding protein that regulates actin ring organization in podosomes and lamellipodia (10 -12). WASP and N-WASP are both activated by the combination of PIP 2 and Cdc42. However, they have different responses when treated with either alone. Although there are differences in the response, PIP 2 has a regulatory role in the activation of both WASP and N-WASP (13). Structural and biochemical studies have shown that coordinated binding of PIP 2 and Cdc42-GTP causes the activation of WASP (14,15), which, in turn, stimulates the actin-nucleating function of the Arp2/3 complex (2,(15)(16)(17)(18). Arp2 and -3 mediate nucleation and cross-linking of actin filaments in vitro. It has a role in protrusion as well as cortical actin remodeling (19).
Podosome assembly at the leading edge of polarized cells is critically dependent upon WASP, because macrophages from patients with Wiscott-Aldrich syndrome fail to assemble podosomes (20). Macrophages from patients with Wiscott-Aldrich syndrome exhibited defects in actin structure formation because of the inability of cells to localize the actinnucleating Arp2/3 complex (11). N-WASP plays a role in actin ring formation in fission yeast (21,22). More recently, it was identified that osteoclasts from WASP-null mice are markedly depleted of podosomes and failed to exhibit actin rings at sealing zones. Complementation of WASP-null osteoclasts with an enhanced green fluorescent protein-WASP fusion protein restores normal cytoarchitecture (23). Osteoclasts transduced with PBD of gelsolin failed to demonstrate WASP in the newly formed podosomes or in the actin ring area (9). Although podosomes are absent in gelsolin null (GsnϪ/Ϫ) osteoclasts, the actin ring was observed. WASP distribution was observed in the actin ring of GsnϪ/Ϫosteoclasts (24). Hence, these osteoclasts were able to resorb bone, but the resorption pits were simple due to the deficiency of podosomes and the resultant hypomotile nature of osteoclasts (5).
Although distinct pathways and signaling molecules have been described to play roles in the organization of actin ring (25)(26)(27)(28)(29), the actual target molecule(s) involved in actin ring formation remain unknown. Hence, in the present study, our objective is to determine the intracellular pathways and the mechanisms by which the WASP-mediated actin ring organization is induced in response to OPN/␣v␤3 signaling in osteoclasts. An aminoglycoside antibiotic, neomycin, which binds to phosphoinositides, was shown to affect the activation of proteins such as phospholipase D, Ezrin, Radixin, Moesin proteins (ERM), and sodium (Na) hydrogen exchanger as well as binding of PIP 2 to ments as described above, osteoclasts were washed three times with cold PBS and lysed in a Triton-containing lysis buffer (10 mM Tris-HCl, pH 7.05, 50 mM NaCl, 0.5% Triton X-100, 30 mM Na 4 P 2 O 7 , 50 mM NaF, 100 M Na 3 VO 4 , 5 mM ZnCl 2 , 1% aprotinin, and 2 mM phenylmethylsulfonyl fluoride). Cells were rocked on ice for 15 min and scraped off with a cell scraper. Cell lysates were centrifuged at 15,000 rpm for 15 min at 4°C, and the supernatant was saved as a Triton-soluble fraction. The pellet was resuspended in RIPA buffer (10 mM Tris-HCl, pH 7.2, 150 mM NaCl, 1% deoxycholate, 1% Triton X-100, 0.1% SDS, 1% aprotinin, and 2 mM phenylmethylsulfonyl fluoride) and centrifuged at 15,000 rpm for 15 min at 4°C. The supernatant is the Triton-insoluble fraction. Protein contents were measured using the Bio-Rad protein assay reagent.
Small Interfering RNA for WASP-Small interfering RNA (siRNA) for WASP (sc-36830) as well as control siRNA (sc-36869; sc37007) were purchased from Santa Cruz Biotechnology, and Lipofectamine 2000 (Invitrogen) was used to transfect osteoclasts. Because the silencing was 30 -40%, we have used permeabilization with streptomycin O to deliver the siRNAs into osteoclasts as described previously (4). Osteoclasts were washed twice with permeabilization buffer (120 mM KCl, 30 mM NaCl, 10 mM Hepes, pH 7.2, 10 mM EGTA, and 10 mM MgCl 2 ). Freshly prepared DTT (5 mM), ATP (1 mM), and 0.5 units/ml of streptolysin O and siRNA (0.5 and 1 M) were added to the buffer at the time of permeabilization. Cells were incubated with the above-mentioned solution for 2-3 min, and resealing was achieved by the addition of ␣-MEM containing 10% fetal bovine serum. Incubation was continued for 48 h. Additionally, control cells were permeabilized as above in the absence of siRNA. After 48 h, lysates were made and subjected to immunoblotting with an antibody to WASP. WASP distribution was determined by immunostaining analysis as described below.
Immunoprecipitation and Western Analysis-Equal amounts of lysate proteins were precleared with protein A-Sepharose, presoaked in lysis buffer containing BSA, and with nonimmune IgG coupled to Sepharose. The precleared supernatants were incubated with antibodies of interest, and the immune complexes were adsorbed onto protein A-Sepharose beads. The beads were pelleted and washed three times for 5 min each with ice-cold PBS. The immune complexes were then eluted in electrophoresis sample buffer and subjected to SDS-PAGE. The proteins were transferred to a polyvinylidene difluoride membrane for Western analyses. Blots were blocked with 10% milk in PBS containing 0.5%Tween (PBS-T) for 2-3 h and then incubated with 1:100 dilution of primary antibody of interest for 2-3 h. After three washes for 10 min each with PBS-T, the blot was incubated with a 1:1000 dilution of peroxidase-conjugated species-specific respective secondary antibody for 2 h at room temperature. After three washes for 10 min each with PBS-T, protein bands were visualized by chemiluminescence using the ECL kit (Pierce) (3)(4)(5). 32 P Labeling of Cells and TLC Analysis of Phosphoinositides-After 4 days in culture, osteoclasts were kept in phosphate (PO 4 Ϫ ) and serumfree medium for 2 h. The cells were then labeled with [ 32 P]PO 4 for 2 h in the above-mentioned medium at 37°C as described previously (3). After labeling, cells were washed twice with serum-free medium and subjected to various treatments as shown in Figs. 2 and 4. Triton-insoluble lysate fraction was made as described above and previously (3). Equal amounts of lysate proteins were immunoprecipitated with an antibody to WASP or nonimmune serum. Lipids were extracted from the immunoprecipitates and dried in the speed vacuum concentrator. The dried lipids were reconstituted in 100 l of chloroform/methanol (1:1) and spotted on silica gel TLC plates pretreated with 1.2% potassium oxalate in methanol and water (2:3). The plates were developed as described (3), and lipid spots were visualized by autoradiography. Nonradioactive phosphoinositide standards (Sigma) were used to identify the 32 P-labeled phosphoinositide(s) associated with WASP. In Vitro Phospholipid Kinase Assay for PI4P 5-Kinase-In vitro phospholipid kinase assay was performed with Triton-soluble lysate (about 200 g of lysate protein) made from osteoclasts subjected to various treatments as shown in Fig. 4. Some lysates were pretreated with PI 3-kinase inhibitors such as WM (100 nM) or LY294002 (10 M) prior to in vitro kinase assay for 45 min on ice. In vitro kinase assay was performed with minor modifications as described (3). Substrates such as PI4P and phosphatidylserine were used. About 0.16 mol of total phospholipid (80 nmol each) was used for the assay. Phospholipid was dried from a stock solution in chloroform and then sonicated in kinase assay buffer containing 1% cholate. Phospholipid micelles were added to lysates and incubated in a water bath at 37°C for 5 min. Ten microliters of solution containing 20 mM Hepes/NaOH, pH 7.4, 5 mM MgCl 2 , 5 M ATP, 1 mM DTT, and 5 Ci of [␥-32 P]ATP was then added. The mixture was vortexed gently, and the incubation was continued for 30 min at 37°C (3,40). Incubation was terminated by the addition of 400 l of chloroform/methanol/water (5:10:2, v/v). Lipids were then extracted and analyzed by TLC as described above and previously (3). Cdc42 Activation Assay-Cdc42 activation assay was performed using the GST-fused WASP-Cdc42 binding domain (CBD) coupled protein beads (catalog number WS03, Cytoskeleton. Inc., Denver, CO). The assay was performed as directed by the manufacturers' guidelines using the reagents provided in the Cdc42 activation assay biochem kit (BK034; Cytoskeleton. Inc.). Triton-insoluble lysates (200 g of protein) made from osteoclasts subjected to various treatments as indicated in Fig. 6a were incubated with WASP-CBD bound beads and subjected to SDS-PAGE. Subsequently, immunoblotting with an antibody to Cdc42 (provided in the kit) was performed. Cdc42-GTP molecule pull down with WASP-CBD peptide was visualized by chemiluminescence using the ECL kit (Pierce) (3)(4)(5). Total lysate (about 100 g of protein) not subjected to either immunoprecipitation or GST-pull down assay was also probed with anti-Cdc42 antibody to determine total cellular levels of Cdc42.
Immunocytochemistry-Osteoclast precursors (10 5 cell/coverslips) were seeded on coverslips or osteologic disc for 4 -5 days. At day 5, osteoclasts were fixed with 3% paraformaldehyde for 20 min and permeabilized with 0.1% Triton X-100 in PBS for 5 min as described previously. Background fluorescence was blocked by incubating cells with either 5% horse serum or 5% BSA in PBS for 30 -45 min at 4°C. The cells were washed and incubated with primary antibodies of interest (WASP, Arp2, or Arp3, and PIP 2 ; 1:100 dilution) in the blocking solution for 2 h at 4°C. The primary antibodies were detected with either Cy2-or Cy3conjugated secondary antibodies. Actin was visualized using rhodamine phalloidin (1:100 dilution; Sigma) as described previously (4,35). Negative controls were performed with nonimmune mouse and goat sera for the double stainings. The cells were washed and mounted on a slide in a mounting solution (Vector Laboratories) and sealed with nail polish. Immunostained osteoclasts were photographed with a Bio-Rad confocal laser-scanning microscope. Images were stored in TIF image format and processed by the Adobe Photoshop software program (Adobe System Inc., Mountain View, CA).
Pyrene Actin Polymerization Assay-Actin polymerization assay was performed using the lysates made from PBS, OPN-, and neomycin/ OPN-treated osteoclasts. 100 -150 g (10 -15 l) cytosol and 25 nM Arp2/3 complex as well as ATP and polymerization buffer provided in the kit were added to the assay mixture. 2.0 M unlabeled and 0.4 M pyrene-labeled G-actin was used for the assay. 50 nM purified Arp2/3 and WASP proteins were used in assays with no cytosol. The assay was performed in a total volume of 100 l. Purified Arp2/3 complex (RP01-A) and WASP-VCA (VCG-03) proteins as well as actin polymerization kit (BK003) were purchased from Cytoskeleton, Inc. GSTfused full-length WASP (FL-WASP) was purified using pGEX/FL-WASP construct by following the method as described previously (41). In vitro actin polymerization assay was performed essentially by following the manufacturer's instructions (Cytoskeleton Inc.) and as described previously (15,42). Cytosol or purified WASP proteins (FL-WASP or WASP-VCA) were preincubated with Arp2/3 complex for 15 min to facilitate WASP-Arp2/3 interaction. Lipid vesicle was prepared essentially as described by Rohatgi et al. (43), added (100 M total lipid) to the incubation mixture containing purified proteins (FL-WASP or WASP-VCA), and incubated for 15 min at 4°C. Similarly for neomycin treatment, purified FL-WASP was treated with neomycin (1 mM) for 15 min at 4°C prior to the addition of either PIP 2 vesicle or Arp2/3 complex. PIP 2 vesicles (100 M) and Arp2/3 complex (50 nM) were added sequentially to the incubation mix with 15 min of incubation at 4°C after each addition. All polymerization assays contained 1 M unlabeled G-actin and 0.3 M pyrene-labeled G-actin and 0.2 mM ATP (15,42) in 100 l of actin polymerization buffer provided in the actin polymerization kit (BK003). Actin polymerization was measured for 10 -15 min at 30-s to 1-min intervals at the excitation of 350 nm (with bandwidth of 20 nm) and emission at the 405 nm in a luminescence spectrofluorometer (Fluoroskan Ascent Lab Systems type 374; software version 2.4.1) at room temperature. Data collected were analyzed and plotted with Microsoft excel (Fig. 9A). F-actin content was measured as described previously (3,4). Statistical significance was calculated as mentioned below.
Data Analysis-All comparisons were made as "% control," which refers to vehicle-treated cells. The other treatment groups in each experiment were normalized to each control value. Data presented are the means Ϯ S.E. of experiments done at different times normalized to intra-experimental control values. For statistical comparisons, analysis of variance was used with the Bonferroni corrections (Instat for IBM, version 2.0; GraphPad software).

Analysis of Interaction of PIP 2 with WASP in Osteoclasts
We have demonstrated previously that OPN/␣v␤3-stimulated PIP 2 association with gelsolin mediates actin polymerization and actin filament reorganization in both avian and mouse osteoclasts (3,5,9). Subsequently, the effect of OPN on PIP 2 association with WASP ( Fig. 1, A-D) was explored. One-half of the WASP immunoprecipitates were immunoblotted with an antibody to PIP 2 (Fig. 1A). OPN stimulation of PIP 2 association with WASP was observed (Fig. 1A, lane 3). The ability of C3 exoenzyme to block OPN-induced PIP 2 interaction with WASP (Fig. 1A, lane 2) suggests the role of Rho GTPase in this process. The level of WASP in each lane was determined by immunoblotting with a WASP antibody as shown in Fig. 1B. The second half of the immunoprecipitates was used for immunoblotting with an antibody to Arp2 (Fig. 1C). Coprecipitation of Arp2 (43 kDa) with WASP was more in OPN-treated osteoclasts (Fig. 1C, lane 3), and this is reduced in osteoclasts treated with C3 transferase prior to OPN stimulation (lane 2). Immunoprecipitation with nonimmune serum is shown in lane 4 ( Fig. 1, A-C). In contrast to the previous observations shown in the interaction of PIP 2 with gelsolin in Triton-soluble fraction, WASP interaction with PIP 2 in response to OPN was observed mostly in the Triton-insoluble fraction. Very small PIP 2 interaction was observed in the Triton-soluble fraction (data not shown). However, Arp2/3 interaction with WASP was found only in the Triton-insoluble fraction (Fig. 1C).
To determine the profile of phospholipids (PI, PIP, PIP 2 , or PIP 3 ) associated with WASP, osteoclasts were labeled with [ 32 P]PO 4 , and lipids associated with WASP were subjected to TLC analysis (Fig. 1D). We have demonstrated previously the binding of PI, PIP, PIP 2 , and PIP 3 with Triton-soluble gelsolin (3). Only PIP 2 interaction was observed in the WASP immunoprecipitates, and this interaction is increased in OPNtreated osteoclasts (Fig. 1D, lane 3). Consistent with the immunoblotting analysis shown in Fig. 1A, C3 pretreatment significantly decreases OPN-induced PIP 2 interaction with WASP (Fig. 1D, lane 4). Furthermore, OPN stimulation of PIP 2 association with WASP was limited to the Triton-insoluble fraction of the osteoclast lysate. Absence of PIP 3 in the TLC analysis implicates the selective role of PIP 2 in the activation process of WASP. The above observations also identify the perceptible role of Rho GTPase in the interaction of PIP 2 with WASP and WASP-Arp2/3 complex formation.
Next, the OPN effect in the formation of WASP/Arp3 complex formation was also confirmed by immunostaining osteoclasts with antibodies to WASP (Fig. 1, E and F, red) and Arp3 (green). Osteoclasts The immunofluorescence localization of WASP with PIP 2 in resorbing osteoclasts plated on CaPO 4 -coated osteologic quartz discs is shown in Fig. 1H. These osteoclasts were not kept in serum-free medium or subjected to any treatment. Localization of WASP (Fig. 1H, red) and PIP 2 (green) was determined by immunostaining with the respective antibodies. Multiple resorption areas exhibiting ring-like structure were observed. Colocalization (Fig. 1H, yellow) of WASP and PIP 2 (indicated by arrows) was observed in the actin ring area. Fig. 1H, asterisks, indicates resorption pits observed underneath these osteoclasts. Immunostaining with nonimmune serum is shown in Fig. 1I.

Determination of the Levels of Transduced Rho Proteins in Osteo-
clasts-To elucidate the potential role of Rho GTPase in the interaction of PIP 2 with WASP, osteoclasts were transduced with TAT-fused Rho GTPases. The following TAT fusion proteins were used: Rho Val14 , Rho Asn19 , Cdc42 Val12 , Cdc42 Asn17 , herpes simplex virus-thymidine kinase (negative control protein), and HA-TAT (vector control protein). The uptake of TAT-Rho fusion proteins was determined by Western analysis using an antibody to HA. About 500 g of osteoclast lysate protein was used for Western analysis (Fig. 2). An increase in Rho uptake was observed in a time-dependent manner. The uptake reaches maximal levels at 40 -60 min and decreases after 2 h. But the protein appears to be stable for up to 6 h and becomes reduced from 12 h onward in the osteoclasts. The Effects of Rho GTPases on the Interaction of PIP 2 with WASP-Osteoclasts subjected to various treatments were immunoprecipitated with a WASP antibody and immunoblotted with an antibody to PIP 2 as shown in Fig. 3, A and D. An antibody to GAPDH was added to each immunoprecipitate to determine the loading levels ( Fig. 3, B and E). The effects of Rho and Cdc42 proteins are shown in Fig. 3, A and D, respectively. These results also demonstrate the effect of OPN. Densitometric scans of PIP 2 interaction with WASP were expressed as percent of control in Fig. 3G. Osteoclasts treated with OPN (Fig. 3A, lane 2; Fig. 3D, lane 5) and transduced with Rho Val14 (Fig. 3A, lane 6) demonstrated an increase in PIP 2 association with WASP. OPN or Rho Val14 -induced PIP 2 interaction was blocked in osteoclasts pretreated with C3 transferase (Fig. 3A, lanes 3 and 7). Basal level PIP 2 interaction with WASP was observed in osteoclasts treated with PBS, Rho Asn19 , and HA-TAT (Fig.  3A, lanes 1, 4, and 8). To compare the results quantitatively, densitometric scanning of three different blots was performed. The binding efficiency was determined (mean Ϯ S.E.; n ϭ 3) after normalizing the WASP-associated PIP 2 levels to GAPDH levels. The data were provided  as a histogram in Fig. 3G. An average of 3-4-fold increase in PIP 2 interaction with WASP was observed in OPN-treated and TAT-Rho Val14transduced osteoclasts.
Cdc42 was shown to activate N-WASP and WASP in combination with PIP 2 (14,17). Hence, we have analyzed the effects of constitutively active Cdc42 Val12 and dominant negative Cdc42 Asn17 on PIP 2 association with WASP (Fig. 3D). Although transduction of Cdc42 Val12 increases PIP 2 interaction with WASP (Fig. 3D, lane 4), the level was lower than that observed in OPN-treated (Fig. 3D, lane 5; Fig. 3A, lane 2) or Rho Val14 -transduced osteoclasts (Fig. 3A, lane 6). However, OPN treatment of osteoclasts transduced with Cdc42 Val12 increases PIP 2 interaction with WASP (Fig. 3D, lane 6). An average of 2.5-3-fold increase was observed in osteoclasts transduced with Cdc42 Val12 and subsequently treated with OPN (Fig. 3G). Osteoclasts transduced with Rho Asn19 prior to transduction with Cdc42 Val12 had no effect in PIP 2 interaction with WASP. Similarly, OPN treatment could not release the inhibitory effect mediated by Rho Asn19 in the interaction of PIP 2 with WASP (Fig. 3A, lane 5). But a significant increase in PIP 2 association with WASP was observed in osteoclasts first transduced with Cdc42 Asn17 and subsequently with Rho Val14 (Fig. 3D, lane 7). These results strongly suggest the role of Rho GTPase in PIP 2 interaction with WASP.

TLC Analyses of Phosphoinositides
In Vitro Phospholipid Kinase Assay Analysis-As there is an increase in PIP 2 association with WASP in osteoclasts transduced with Rho Val14 or treated with OPN, we proceeded to determine whether there was an increase in PI4P 5-kinase activity in these osteoclasts. Rho GTPase was identified as an upstream regulator of PI4P 5-kinase and PI 3-kinase (35, 44 -46). The direct role of Rho in the activation of PI4P 5-kinase was determined by in vitro phospholipid kinase assay analysis. Triton-soluble lysate from osteoclasts subjected to various treatments as indicated in Fig. 4A was used for this analysis. Because both PI 3-kinase and PI 5-kinase are activated by Rho or OPN treatment (35), some lysates were pretreated with PI 3-kinase inhibitors such as wortmannin (WM) or LY29004 prior to in vitro kinase assay in the presence of PI4P as substrate and [␥-32 P]ATP. Labeled lipids were extracted and subjected to TLC analysis. The products generated by either PI 3-kinase (e.g. PI(3,4)P 2 and PI(3,4,5)P 3 (PIP 3 )) or PI 5-kinase (PI(4,5)P 2 (PIP 2 )) was determined by TLC analysis. Lysates made from osteoclasts treated with OPN ( Fig. 4A, lanes 2, and 8) or transduced with TAT-Rho Val14 (lanes 4 and 9) exhibited an increase in the levels of PIP 2 through activation of PI 5-kinase as compared with PBS-treated (lane 1) or Herpes simplex virus-thymidine kinase-transduced (lane 3) osteoclasts. Very negligible or no increase in the PIP 2 levels was observed in osteoclasts transduced with either TAT-Cdc42 Val12 (Fig. 4A, lane 7) or Rho Asn19 (lane 5), respectively. Pretreatment of osteoclasts with C3 prior to transduction with TAT-Rho Val14 blocked Rho-induced PI 5-kinase activity as well as the synthesis of PIP 2 (Fig. 4A, lane 6). Lysates untreated with either WM or LY294002 exhibited both PIP 3 and PIP 2 (Fig. 4A, lanes 10 -14). Formation of PI(3,4)P 2 was not observed in these assays despite the formation of PIP 3 . Formation of both PIP 2 and PIP 3 in osteoclasts treated with OPN or transduced with TAT-Rho Val14 indicates the activation of PI 5-kinase and PI 3-kinase, respectively. Lysates treated with either WM (Fig. 4A, lanes 1-7) or LY294002 (Fig. 4A, lanes 8 and 9) failed to exhibit PI 3-kinase products PI(3,4)P 2 or PI(3,4,5)P 3 (PIP 3 ). PI 5-kinase mediated accumulation of PI(4,5)P 2 (PIP 2 ) was unaffected by these inhibitors (Fig. 4A, lanes 1-9). The above observations suggest that OPN-mediated increase in PIP 2 interaction with WASP occurs through Rho GTPase-mediated pathway. Rho plays a key role in the activation of PI 5-kinase and PI 3-kinase, which are involved in the formation of both PIP 2 and PIP 3 , respectively, in osteoclasts.
Analysis of Interaction of Phospholipids with WASP in Vivo by TLC Analysis-We have shown interaction of PIP 2 with WASP in Fig. 1. In order to determine the role of PIP 2 in the activation of WASP, we have used neomycin, which has been shown to competitively inhibit PIP 2 interaction with ezrin, sodium (Na) hydrogen exchanger, and gelsolin (31)(32)(33)(34). Determination of interaction of PIP 2 with WASP in response to various treatments was determined by TLC analysis (Fig. 4B). Densitometric scans of PIP 2 spots of three experiments were expressed as percent of control in Fig. 4C. An increase in PIP 2 association with WASP was observed in OPN-treated (Fig. 4B, lane 2) or TAT-Rho Val14 -transduced (lane 8) osteoclasts. Neomycin at a dose of 1 mM decreased OPNinduced association of PIP 2 with WASP in a time-dependent manner (Fig. 4B, lanes 3-6). Similarly, Rho Val14 transduction of osteoclasts preincubated with neomycin had no effect in increasing PIP 2 interaction with WASP (Fig. 4B, lane 7). As shown in Fig. 4B, a minimal increase in PIP 2 association with WASP was observed (lane 11) in TAT-Cdc42 Val12 -transduced osteoclasts, and this interaction is increased by treatment with OPN (lane 12). The increase was found to be equal to the levels observed in OPN-treated (Fig. 4B, lane 2) or TAT-Rho Val14 -transduced (lane 8) osteoclasts. The levels of PIP 2 associated with WASP in HA-TAT (Fig. 4B, lane 9) and TAT-Rho Asn19 (lane 10)-transduced osteoclasts are equal to the basal level association observed in PBStreated osteoclasts (Fig. 4B, lane 1). Immunoprecipitation with a nonimmune serum is shown in Fig. 4B, lane 13. Observations shown in Figs. 1-4 elucidate the upstream regulatory role of the Rho GTPase in the interaction of PIP 2 with WASP.

Immunostaining Analysis of the Effects of Neomycin on the Membrane Targeting of WASP Protein
Recombinant WASP, expressed in Escherichia coli, was observed to bind strongly to Cdc42, weakly to Rac, and not at all to Rho (47). Although, Rho was not identified as a direct WASP binding partner, observations shown in Figs. 1-4 indicate that WASP interaction with PIP 2 as well as its membrane targeting is Rho-dependent. To clarify further the roles of Rho GTPase and PIP 2 in detail in the activation and membrane targeting of WASP, immunostaining analysis was performed using an antibody to WASP (Fig. 5a, green) and rhodamine phalloidin for actin (Fig. 5a, red). Osteoclasts transduced with constitutively active Rho Val14 increases clusters of podosomes formation as well as colocalization of WASP and actin in the podosomes (Fig. 5a, indicated by  arrows). Although OPN-treated osteoclasts exhibit numerous podosomes throughout the subsurface of osteoclasts (Fig. 5a, panel D), podosome clusters as observed in TAT-Rho Val14 -transduced osteoclasts (Fig.  5a, panel A) were not observed. OPN treatment resulted in the formation of numerous lateral microfibrillar-like extensions (Fig. 5a, panel D) from the plasma membrane. Podosomes, actin ring, as well as the lateral microfibrillar protrusions from the plasma membrane exhibit colocalization of WASP and actin in OPN-treated osteoclasts (Fig. 5a, panel D). Even though TAT-Rho Val14 transduction increases podosome cluster formation, the extent of actin ring formation, as well as colocalization of WASP/actin in the actin ring area (Fig. 5a, panel A, indicated by wiggly  arrows), is smaller than that of OPN-treated osteoclasts. Microfibrillarlike extensions from the cell periphery were absent in these osteoclasts (Fig. 5a, panel A). Formation of numerous podosomes and peripheral filopodia-like structures suggests the activation of both Rho and Cdc42 in OPN-treated osteoclasts. A time-dependent inhibition of clusters of podosomes formation as well as colocalization of WASP and actin in the actin ring was observed in osteoclasts treated with neomycin for 45 min (Fig. 5a, panels B and E) and 2 h (Fig. 5a, panels C and F). Transduction of constitutively active Rho Val14 or stimulation with OPN had no effect on the formation of podosomes or actin ring in these osteoclasts. Distribution of WASP was observed throughout the cytoplasm of these osteoclasts.

Effects of Neomycin on the Interaction of WASP with Arp2 and PIP 2
Next distribution of WASP/Arp2 (Fig. 5b, panels A-C) and WASP/ PIP 2 (Fig. 5b, panels D-F) was observed by immunostaining and confocal microscopy analyses (Fig. 5b). OPN stimulated colocalization of WASP (Fig. 5b, green) and Arp2 (red) in the actin ring (Fig. 5b, panel B) as compared with PBS-treated osteoclast (Fig. 5b, panel A). The yellow color (Fig. 5b, panel B, shown by white arrows) indicates colocalization of WASP and Arp2 at the periphery in the actin ring. Very minimal colocalization of these proteins was observed in the actin ring of neomycin/OPN-treated osteoclasts (Fig. 5b, panel C). Neomycin causes diffused distribution of both Arp2 and WASP in the cytoplasm of these osteoclasts. Similarly, a significant increase in colocalization (Fig. 5b, panel E, yellow and indicated by black arrow) of PIP 2 (Fig. 5b, green) and WASP (Fig. 5b, red) was observed in OPN-treated osteoclasts (Fig. 5b, panel E) as compared with PBS-treated control cells (Fig. 5b, panel D). Consistent with the biochemical evidence of inhibition of PIP 2 interaction with WASP by neomycin (Fig. 4), immunostaining analysis also exhibits inhibition of interaction of WASP (Fig. 5b, red) and PIP 2 (Fig.  5b, green) in the actin ring (Fig. 5b, panel F). Diffused distribution of PIP 2 was observed in the cytoplasm of these osteoclasts (Fig. 5b, panel F). Basal localization of WASP (Fig. 5b, panel R) was observed in the actin ring of these osteoclasts (Fig. 5b, panel F). Observations shown here substantiate the role of PIP 2 in the activation of WASP and WASP-Arp2/3 complex formation.

Analysis of the Effects of OPN and Rho on the Interaction of Cdc42 with WASP
Cdc42 was shown to activate N-WASP and WASP in combination with PIP 2 (14,17). By having established the interaction of PIP 2 with WASP in osteoclasts, the interaction of Cdc42 with WASP was then determined (Fig. 6a, panel A). WASP immunoprecipitates were immunoblotted with an antibody to Cdc42. OPN treatment (Fig. 6a, lane 4) and TAT-Rho Val14 transduction (lane 6) increases Cdc42 interaction with WASP. More intriguingly, neither of these effects occurs in osteoclasts pretreated with neomycin (Fig. 6a, lanes 2 and 5). In neomycinpretreated osteoclasts, the levels of Cdc42 associated with WASP were found to be lower than the PBS-treated control (Fig. 6a, lane 3) or TAT-herpes simplex virus-thymidine kinase (a nonspecific protein control; lane 7)-transduced osteoclasts. Immunoprecipitation with a nonimmune serum is shown in Fig. 6a, lane 1. WASP level in each immunoprecipitate is shown in Fig. 6a, panel B. A decrease in binding of Cdc42 with WASP in osteoclasts treated with neomycin prior to OPN stimulation (Fig. 6a, lane 2) or Rho Val14 -transduction suggests that PIP 2 interaction with WASP may take place preceding Cdc42 interaction with WASP. Binding of PIP 2 to WASP increased the effectiveness of Cdc42 binding to WASP. In view of the fact that only GTP-Cdc42 binds to WASP-Cdc42-binding domain, the associated Cdc42 ought to be in the form GTP-Cdc42 (14,48).
Inhibition of interaction of GTP-Cdc42 with WASP in neomycintreated osteoclasts raises the following two possibilities: 1) WASP is present in the inactive state; 2) PIP 2 may be required for the activation of Cdc42 as shown by others (43,49). Hence, to determine the second possibility, lysates made from osteoclasts treated with neomycin prior to OPN or Rho Val14 were used for GST-pull down analysis. Given that only GTP-Cdc42 binds to Cdc42 binding domain of WASP (WASP-CBD; see Refs. 14 and 48), GST-fused WASP-CBD protein was used for pull down analysis. Western analysis with an antibody to Cdc42 is shown in Fig. 6a, panel C. Immunoprecipitation with an antibody to Cdc42 was used as identification control (Fig. 6a, panel C, lane 10), and pull down with GST alone (vector protein) was used as negative control (lane 1). 100 g of lysate from osteoclasts subjected to various treatments (Fig.  6a, panel D) was immunoblotted with an antibody to Cdc42 to determine the cellular levels of Cdc42. Our observations indeed demonstrated an increased pull down of GTP-Cdc42 in Rho Val14 -transduced Although PIP 2 is a product of a downstream regulator of Rho GTPase, their role in Cdc42 activation is essentially unknown. It is likely to be an indirect activator of Cdc42 as shown by others (43,49). Taken together, these findings suggest that PIP 2 participates in the interaction of Cdc42 with WASP.

Actin Ring Formation Depends on the Interaction of PIP 2 and Cdc42 with WASP in Osteoclasts
Next, we performed experiments to determine the role of Cdc42 in actin ring formation by immunofluorescence analyses (Fig. 6b). Consistent with the observations shown by Linder et al. (50) in macrophages, increasing the intracellular levels of Cdc42 Val12 in osteoclasts decreases the actin core present in podosomes (Fig. 6b, panel A, arrows). Although WASP staining was observed in podosomes and plasma membrane (in Fig. 6b, panel G), F-actin staining was not observed in the podosomes of these osteoclasts (Fig. 6b, panel R). These osteoclasts exhibited numerous microfibrillar-like extensions at the periphery with minimal or no distribution of actin. Addition of OPN to osteoclasts transduced with TAT-Cdc42 Val12 not only increases F-actin core in podosomes but also actin polymerization in the peripheral microfibril-lar extensions. These structures display colocalization of actin and WASP (Fig. 6b, panel B). Minimal colocalization of WASP and actin in the actin ring of osteoclasts transduced with TAT-Cdc42 Val12 (Fig. 6b, panel A) indicates that Cdc42 alone has negligible effect in the activation of WASP in osteoclasts. Pretreatment of osteoclasts with neomycin for 2 h prior to TAT-Cdc42 Val12 /OPN treatments blocked colocalization of actin and WASP at the periphery in the actin ring but had no effect on the filopodia-like extension formation. Although these osteoclasts exhibited filopodia-like extensions at the periphery (Fig. 6b,  panel B), colocalization of F-actin and WASP was not observed in these structures. WASP distribution was mostly cytoplasmic in these osteoclasts. Osteoclasts transduced with TAT-Rho Val14 increases the number as well as clusters of podosomes formation (Fig. 6b, panel C). But actin ring formation is not equal to the levels observed in OPN-alone treated osteoclasts as shown in Fig. 5a. Cotransduction of constitutively active forms of TAT-Cdc42 Val12 with Rho Val14 induced prominent cytoskeletal changes such as podosomes formation throughout the surface of osteoclasts, striking band of actin ring, and microspike extensions with adhesion structures (Fig. 6b, panel E). In the presence of neomycin, about 80 -90% inhibition of colocalization of WASP and actin in podosomes as well as in the actin ring was observed. Long smooth filopodia-like extensions were observed as a replacement for microspike-like extensions with adhesion structures in these osteoclasts. Colocalization of actin and WASP is minimal in these extensions as compared with cells in Fig. 6b, panels B and E. Hence, these observations indicate that due to the failure of activation of WASP by PIP 2 , its interaction with Cdc42 as well as actin polymerization in the actin ring area is inhibited, although diffuse distribution of WASP throughout the osteoclasts and in the smooth filopodia-like extensions were observed in osteoclasts pretreated with neomycin (Fig. 6b, panels C and F).

Effects of siRNA to WASP on Actin Ring Formation in Osteoclasts
We then introduced siRNAs into osteoclasts to reduce endogenous levels of WASP in osteoclasts. Transfection of WASP siRNA into osteoclasts by Lipofectamine 2000 (Invitrogen) produced 30 -40% silencing of WASP mRNA, although introduction of siRNA to WASP into streptolysin O-permeabilized osteoclasts exhibited a significant decrease (78 Ϯ 9%; n ϭ 3) in the endogenous levels of WASP protein. Immunoblot analysis with an antibody to WASP showed a decrease in the WASP expression (Fig. 7a, panel A, lanes 4 and 5) but not in scrambled (Sc) siRNA-treated (Fig. 7a, panel A, lane 3) osteoclasts. Streptolysin O treatment alone had no effect on the endogenous WASP levels (Fig. 7a,   panel A, lane 2). Immunoblotting of the same blot with an antibody to GAPDH (Fig. 7a, panel B) was used as loading control.
Localizations of WASP and actin were determined in osteoclasts treated with siRNA to WASP (Fig. 7b). A significant decrease in actin ring matched the WASP level in osteoclasts treated with siRNA nucleotides (Fig. 7b, panels A-D and AЈ-DЈ) was observed. Actin ring as well as WASP distribution was barely detectable at 1 M siRNA (Fig. 7b,  panel D). Neither actin clumps nor bundles were observed in these osteoclasts. Cell shape or the distribution of actin and WASP was not affected by permeabilization with streptolysin O (Fig. 7b, panel A) or transfection of scrambled siRNA to WASP (1 M; Fig. 7b, panel B). siRNA-mediated attenuation of WASP levels not only reduced the formation of actin ring but also bone resorption in osteoclasts (data not shown).

Immunoblotting Analysis of Interaction of Arp2 with WASP
The effect of various treatments on the formation of WASP/Arp2 complex was subsequently analyzed by immunoblotting of WASP immunoprecipitates with Arp2 antibody (Fig. 8). Rho Val14 transduction increases PIP 2 interaction with WASP (Fig. 3) and augments F-actin content equal to the levels observed in OPN-treated osteoclasts (Fig.  9B). However, in Rho Val14 -transduced osteoclasts, Arp2 (Fig. 8, lane 2) association with WASP is not increased to the levels observed in OPN (lane 5)-treated osteoclasts. Similar increase in Arp2 interaction with WASP was observed in osteoclasts transduced with TAT-Cdc42 Val12 (Fig. 8, lane 10). Interaction of Arp2 with WASP was enhanced in osteoclasts treated with TAT-Cdc42 Val12 /TAT-Rho Val14 (Fig. 8, lane 4), TAT-Cdc42 Val12 /OPN (lane 11), and TAT-Cdc42 Asn17 /OPN (lane 12). The augmented effect observed in these osteoclasts is equal to levels observed in OPN-alone treated osteoclasts (Fig. 8, lane 5). OPN-induced Arp2 interaction is blocked by pretreatment with neomycin (Fig.  8, lane 8) or C3 (lane 13). Basal level interaction of WASP/Arp2 was observed in PBS-treated (Fig. 8, lane 6) as well as HA-TAT-(lanes 1), Rho Asn19 -(lane 3), and Cdc42 Asn17 (lane 14)-transduced osteoclasts. Immunoprecipitation with a nonimmune serum is shown in Fig. 8, lanes 9 and 15. The above observations indicate that both Rho-and Cdc42mediated events are required for WASP activation, interaction of  In vitro polymerization assay was performed as described under "Materials and Methods." Results shown are means of three experiments, and each assay was performed in quadruplicate. B, measurement of F-actin content. F-actin content was measured by rhodamine phalloidin binding in osteoclasts treated as given below the figure. Cells were grown in 24-well tissue culture plates, and about 3-5 wells were used for each treatment. The data presented are mean Ϯ S.E. for three experiments. ***, p Ͻ 0.0001 versus PBS or HA-TAT-transduced osteoclasts; **, p Ͻ 0.001 versus OPNtreated or Rho Val14 -transduced osteoclasts; *, p Ͻ 0.05 versus Cdc42 Val12 -transduced osteoclasts.
Arp2/3 with WASP, and cortical actin polymerization. Immunoblotting of the same blot with an antibody to WASP (Fig. 8B) was used as loading control.

Actin Polymerization in Vitro
We then examined the role of PIP 2 in the activation of WASP and the subsequent effect on actin polymerization mediated by the Arp2/3 complex. It has been demonstrated in vitro that WASP is activated by PIP 2 and prenylated GTP␥S in order to arbitrate actin polymerization through its binding to Arp2/3 (51,52). Hence, in vitro actin polymerization assay was performed in the presence of unlabeled and pyrenelabeled G-actin (42) (Fig. 9A). Purified Arp2/3 complex was used to a final concentration of 25 nM. An increase in actin polymerization was observed in lysate made from OPN-treated osteoclasts. This increase was blocked by neomycin treatment, and the effect was equal to the level observed in PBS-treated osteoclasts. The basal levels of Arp2/3-mediated actin polymerization in PBS-and neomycin/OPN-treated osteoclasts may be due to the WASP that is present in the masked unstimulated form. A similar inhibitory effect of neomycin was observed in the assay with purified proteins. The effects of Arp2/3-mediated actin polymerization in the presence of WASP-VCA and FL-WASP protein is more than the levels observed in assays performed with lysates made from OPN-treated osteoclasts. Maximum actin polymerization effect was observed with WASP-VCA protein, and addition of PIP 2 micelle had no extra stimulatory effect on the effects mediated by WASP-VCA. Although WASP-VCA is more potent than FL-WASP, addition of PIP 2 to FL-WASP increased polymerization to the level equivalent to WASP-VCA protein. The actin polymerization observed in reactions containing FL-WASP may be because some of the purified protein is present in active unmasked conformation. Addition of PIP 2 to the reaction containing FL-WASP may activate the inactive masked WASP and actin polymerization as well. However, neomycin blocked PIP 2 -mediated activation of FL-WASP. Similar inhibition was observed in lysate made from neo/OPN-treated osteoclasts. Observations shown here as well as in Figs. 4 -6 and 8 suggest that activation of WASP is mediated by direct binding of PIP 2 .

Measurement of F-actin Content
Next, the F-actin content in osteoclasts subjected to various treatments ( Fig. 9B) was measured because this is compatible with the activation of WASP as well as the actin nucleating and polymerization function of Arp2/3. An increase in F-actin was observed in OPN, TAT-Rho Val14 , TAT-Cdc42 Asn17 /TAT-Rho Val14 , and TAT-Cdc42 Val12 /OPNtreated osteoclasts. Osteoclasts transduced with Cdc42 Val12 can activate actin polymerization to a minimal extent, and it requires TAT-Rho Val14 to reach maximal activation. Pretreatment of osteoclasts with neomycin or C3 transferase significantly blocked the OPN-or Rho-induced F-actin levels in osteoclasts. Although OPN treatment increases F-actin content in osteoclasts transduced with TAT-Cdc42 Val12 , it does not have any effect in osteoclasts transduced with TAT-Rho Asn19 . Osteoclasts transduced with TAT-Cdc42 Val12 , TAT-Rho Asn19 , and TAT-Cdc42 Asn17 did not have any effect on F-actin levels. PBS-treated or herpes simplex virus-thymidine kinase-transduced osteoclasts were used as controls.

DISCUSSION
Podosomes bring about motility and bone resorption functions of osteoclasts. We have demonstrated previously (5,9,35) the roles of gelsolin and the phosphoinositides binding domains of gelsolin in the podosome assembly/disassembly. Also, previous studies from our laboratory and other laboratories have demonstrated the role of several signaling molecules in osteoclast cytoskeletal organization, podosomes assembly, actin ring formation, and bone resorption (27, 28, 35, 37, 53-57, 59, 62-64). It is not known at this time, how these signaling molecules regulate cytoskeletal remodeling toward actin ring formation. Also unknown are the target proteins involved in this course of action. WASP has been identified to have a regulatory role in the assembly and maintenance of podosomes in macrophages (50). WASP-null osteoclasts generated in vitro through the siRNA silencing technique failed to exhibit podosomes, actin ring, and actin plaques (Fig. 7). The most recent report by Calle et al. (23) has shown that in cultures of WASP-null osteoclasts, the formation of actin ring was markedly impaired. Although WASP-null osteoclasts were able to polymerize actin and form actin patches, highly dynamic podosomes or actin rings were not organized as in normal osteoclasts (23). Osteoclasts from gelsolin null (GsnϪ/Ϫ) mice failed to exhibit distinct podosomes, but these osteoclasts demonstrated actin ring in the clear zone area (5). WASP distribution was observed in the actin ring of GsnϪ/Ϫ osteoclasts (24). In the absence of WASP, gelsolin did not function to form the actin ring in WASP-null (Ϫ/Ϫ) osteoclasts. Failure of formation of actin rings in WASPϪ/Ϫ osteoclasts and the localization of WASP in the actin ring of GsnϪ/Ϫ osteoclasts indicate that WASP could be the most probable candidate protein involved in the actin ring formation.
WASP functions as a scaffolding protein for Arp2/3, which is required for actin nucleation and polymerization. WASP binding with Arp2/3 takes place after full activation of WASP by PIP 2 and Cdc42 (14,15,17,18,48,65). Several laboratories have studied the regulatory molecules involved in the activation of WASP (14,15,48,(65)(66)(67). We have demonstrated previously that PIP 2 regulates uncapping of gelsolin, actin polymerization, and podosomes assembly in osteoclasts. Observations shown here have demonstrated colocalization of both PIP 2 and WASP in the actin ring of resorbing osteoclasts. Cytochemical characterization of RAW 264.7 osteoclast-like cells and bone marrow osteoclasts in which Arp2 was knocked down revealed fewer podosomes and no actin rings in resorbing osteoclasts, although many cells remained well spread (68). The above observations provide an explanation for the interaction of WASP with PIP 2 and then Arp2/3 in the configuration of actin ring during bone resorption.
Participation of members of the WASP superfamily (WASP and Scar) in the formation of cortical actin polymerization to induce filopodia or membrane ruffling has been shown to be mediated by the members of Rho GTPases such as, Rho, Cdc42, and Rac (51, 69 -71). An increase in the formation of PIP 2 from PIP (substrate) was observed in in vitro kinase assay using lysates made from osteoclasts treated with OPN or transduced with Rho Val14 . Osteoclast lysates treated with LY29004 or wortmannin blocked PI 3-kinase activity and not PI 4P 5-kinase activity. However, osteoclasts pretreated with C3 transferase prior to OPN addition or TAT-Rho Val14 transduction inhibited the formation of both PIP 2 and PIP 3 . This indicates that both phosphatidylinositol 4-phosphate 5-kinase (PI4P 5-kinase) and PI 3-kinase activation are Rho-mediated. Moreover, inhibition of PIP 3 synthesis and not PIP 2 in wortmannin-or LY29004-treated osteoclasts indicates that the single phospholipids spot in the TLC analysis (Fig. 4A) is PI(4,5)P 2 (synthesized by PI4P 5 kinase) and not PI(3,4)P 2 (synthesized by PI3K). Physical association of the small GTPase Rho with a 68-kDa PI4P 5-kinase was observed in Swiss 3T3 cells (72). PI4P 5-kinase has been identified as direct regulator of PIP 2 levels and CD44-ERM-actin complex formation (31,44). Overexpression of PI4P 5-kinase, which synthesizes PIP 2 , was shown to induce actin-based movement of raft-enriched vesicles through WASP-Arp2/3 complex formation (17). An increase in the interaction of PIP 2 with WASP in the Triton-insoluble fraction of OPN-treated or Rho Val14 -transduced osteoclasts (Figs. 1 and 3 -5) are compatible with the functional profile favoring membrane targeting of WASP protein as well as its interaction with Arp2/3 complex. Inhibition of this interaction in C3 transferase-treated osteoclasts and a very small or no increase in the interaction of PIP 2 with WASP in osteoclasts transduced with Cdc42 Val12 supported the role of Rho GTPase in the activation of WASP.
Subsequently, the role of PIP 2 in the activation of WASP was also confirmed by using neomycin. Neomycin, an aminoglycoside antibiotic, was identified to efficiently probe PIP 2 on the cell membranes where binding of large molecules to PIP 2 could be inhibited (73). Binding of neomycin to PIP 2 was shown to affect the activation of ERM proteins regardless of cell type (31). Similarly, neomycin inhibited the effect of PIP 2 on gelsolin function (34). Hence, we used this approach to determine the role of PIP 2 in the membrane localization of WASP as well as actin ring formation. The results presented here identify PIP 2 as regulator of podosomes assembly and actin ring formation in osteoclasts. Membrane targeting of WASP, its interaction with Arp2/3, and actin ring formation are inhibited in neomycin-treated osteoclasts (Figs. 5, 6, and 8). WASP distribution was mostly cytoplasmic in these osteoclasts. As the number of podosomes is reduced in these osteoclasts, we propose that neomycin not only blocks WASP activation but also gelsolin function as well. This is consistent with our previous observations of the role of PIP 2 in the gelsolin-mediated assembly of functional podosomes in osteoclasts (5,9).
Although Rho Val14 transduction increases PIP 2 interaction as well as membrane targeting of WASP (Figs. 3 and 4), it is not sufficient to form actin ring in osteoclasts (Fig. 5a). PIP 2 is required for the activation of WASP-and neomycin treatment-blocked WASP activation in osteoclasts. Hence, it was not surprising that the level of GTP-Cdc42 coimmunoprecipitated with WASP immunoprecipitates is lesser in osteoclasts treated with either neomycin/Rho Val14 or neomycin/OPN than Rho Val14 -transduced or OPN-treated osteoclasts (Fig. 6a, panel A). However, a decrease in binding of GTP-Cdc42 to GST/WASP-CBD in the GST-pull down analysis (Fig. 6a, panel B) using lysates from neomycin/Rho Val14 or neomycin/OPN was unanticipated if one envisaged that PIP 2 has no effect on the activation of Cdc42. Most interestingly, activation of Cdc42 by PIP 2 has been described recently by others (43,49,74). In addition to its role in actin polymerization (3,14,66), phosphoinositides are shown to have several other functions as follows: displacement of the guanine nucleotide dissociation inhibitor protein (75); activation of guanine nucleotide exchange factors by binding to their pleckstrin homology domain (76,77); and stimulation of release of GDP from G-proteins (60). In the stimulation of guanine nucleotide exchange factor and release of GDP, PIP 2 was shown to be more effective than PI (60). Although the mechanism is not clear, it appears that PIP 2 regulates activation of Cdc42 through its binding to Dbs, a Dbl family member that has been shown to have a specific regulatory role in the activation of Cdc42 and RhoA. Dbl family members are guanine nucleotide exchange factors for Rho GTPases and possess tandem Dbl (DH) and pleckstrin homology domains. Phosphoinositides binding to the pleckstrin homology domain of Dbs were shown to be required for the gross subcellular distribution leading to activation of GTPases (74).
Transduction of Cdc42 Val12 has lesser effects in PIP 2 (Figs. 3 and 4) and Arp2/3 (Fig. 8) interactions with WASP. WASP is an important regulator of podosomes assembly, and WASP distribution was observed in podosomes of macrophages (50) and osteoclasts (Figs. 5, 6, and 9). It has been demonstrated that coinjection of WASP with Cdc42 Val12 destroyed podosomal localization and increased prominent filopodia formation in macrophages (12). Osteoclasts transduced with Cdc42 Val12 failed to exhibit podosome structures enriched in F-actin as well as colocalization of WASP and actin. Moreover, WASP distribution was observed in the filopodia-like structures in osteoclasts transduced with Cdc42 Val12 . However, actin distribution was very minimal in these structures. Immunostaining analysis has demonstrated localization of WASP and not F-actin in podosomes of osteoclasts transduced with Tat-Cdc42 Val12 . It is possible that the WASP present in podosomes of TAT-Cdc42 Val12 -transduced osteoclasts is also inactive, and hence Cdc42 could not activate through its binding to the CBD domain of WASP. Although Cdc42 Val12 has no effect in inducing podosomes formation in osteoclasts or macrophages, others have shown in endothelial and dendritic immature cells that Cdc42 is sufficient to induce podosomes formation (58,71). Although there is a small increase in PIP 2 levels in Cdc42-transduced osteoclasts (Figs. 3 and 4), it is not sufficient for full activation of WASP as well as its interaction with the Arp2/3 complex (Fig. 8). Minimal interaction of PIP 2 with WASP in these osteoclasts may be due to sequential activation of Cdc42, Rac, and then Rho GTPases (61). The increase in actin ring formation in addition to filopodia-like extensions in osteoclasts transduced with both Rho Val14 and Cdc42 Val12 indicates that PIP 2 interaction with WASP is indeed a key step in the full activation and membrane targeting of WASP. Rho transduction increases podosomes formation throughout the surface of osteoclasts, but these cells do not exhibit microfibrillar or microfilopodialike structures at the periphery as observed in osteoclasts transduced with Cdc42 Val12 . Osteoclasts treated with OPN exhibit both peripheral microfilopodia-like projections and podosomes throughout the surface of osteoclasts. These structures demonstrate colocalization of WASP and actin. Formation of podosomes throughout the surface of osteoclasts and microfibrillar structures at the periphery indicate the activation of both Cdc42 and Rho in OPN-stimulated osteoclasts.
Although stimulation of actin polymerization has been shown to be independent of PIP 2 synthesis in polymorphonuclear leukocytes (60), an increase in actin polymerization was observed in response to PIP 2 interaction with WASP in osteoclasts. The in vitro polymerization assay (Fig.  9A) provides a tool to measure the role of PIP 2 in the WASP activation process. An increase in actin polymerization was observed in the presence of PIP 2 vesicle and FL-WASP. Inhibition of actin polymerization in neomycin-treated osteoclasts and by neomycin in assays with purified protein (FL-WASP) and PIP 2 micelle suggests that PIP 2 interaction with WASP is required for its activation. This observation supported others who have shown that addition of PIP 2 vesicle to WASP and Arp2/3 complex accelerated actin polymerization and produced an increase in actin filaments. Furthermore, addition of GTP␥S-Cdc42 doubled the effect of PIP 2 . Nonlipidated Cdc42 was found to be ineffective because it does not bind vesicles (14).
Actin ring formation requires cooperative interaction of PIP 2 and Cdc42 with WASP. The actin remodeling in the organization of podosomes is highly dynamic as opposed to actin ring formation during bone resorption. Hence, actin-remodeling processes during podosomes assembly/disassembly and actin ring formation are spatially separated. The present study provides evidence that Rho-mediated PIP 2 interaction with WASP may contribute to the activation and membrane targeting of WASP. Subsequent interaction of Cdc42 and Arp2/3 with WASP may enhance cortical actin polymerization in the process of actin ring formation in osteoclasts. There seems to be no competition existing between PIP 2 and Cdc42 in their binding to WASP. PIP 2 is also required for the activation of Cdc42. The findings from this study are our first step to determine the regulatory mechanisms involved in actin ring formation. Further investigation on the spatially and temporally regulated functions of both WASP and gelsolin is necessary to identify their role in actin ring formation as well as podosomes assembly/disassembly during bone resorption and osteoclast motility.