c-Src is required for stimulation of gelsolin-associated phosphatidylinositol 3-kinase.

We have shown that osteopontin binding to integrin alphav beta3 in osteoclasts stimulates gelsolin-associated phosphatidylinositol (PtdIns) 3-hydroxyl kinase (PI 3-kinase), leading to increased levels of gelsolin-bound PtdIns 3,4-P2, PtdIns 4,5-P2, and PtdIns 3, 4,5-P3, uncapping of barbed end actin, and actin filament formation. Inhibition of PI 3-kinase activity by wortmannin blocks osteopontin stimulation of actin filament formation, suggesting that activation of gelsolin-associated PI 3-kinase is an important pathway in cytoskeletal regulation. To study the mechanism of gelsolin-associated PI 3-kinase activation, we analyzed anti-gelsolin immunoprecipitates for the association of protein kinases. We demonstrated that c-Src co-immunoprecipitates with gelsolin, and that osteopontin stimulates its activity. Elimination of osteopontin-stimulated Src activity associated with gelsolin through antisense oligodeoxynucleotides blocked the stimulation of PI 3-kinase activity associated with gelsolin and the gelsolin-dependent cytoskeletal reorganization induced by osteopontin, including increased F-actin levels. In addition, treatment of osteoclasts with antisense oligonucleotides to Src reduced bone resorption. Our results demonstrate that osteopontin stimulates gelsolin-associated Src, leading to increased gelsolin-associated PI 3-kinase activity and PtdIns 3,4,5-P3 levels, which facilitate actin filament formation, osteoclast motility, and bone resorption.

Osteoclasts are multinucleated, giant cells, responsible for bone resorption. Osteoclast adhesion to bone leads to the formation of the osteoclast clear zone: a ring-like adhesion zone circumscribing an area of bone resorption. The cytoplasm of the clear zone contains numerous actin filaments perpendicular to the bone matrix, which are anchored in podosomes (1)(2)(3)(4)(5). Podosomes are small cell processes specific to cells of monocytic origin. Podosomes contain numerous proteins observed in the focal adhesions of other cells. Although podosomes and focal adhesions are related, there are important functional differences. Podosomes are less tightly associated with the substratum and are more highly dynamic, changing in size and location and appearing and disappearing with life spans of 2-12 min (4). Osteoclasts are highly motile cells, and podosomes appear to be a preferred cell/matrix attachment mechanism for motility in cells such as macrophages, monocytes, and osteoclasts.
The adhesion of osteoclasts through podosomes involves interaction of cell surface receptors, integrins, with matrix components. In osteoclasts, integrin ␣ v ␤ 3 1 is responsible for adhesion associated with bone resorption (3,6). The ligands of ␣ v ␤ 3 are many, and they contain the RGD cell adhesion sequence present in several serum and bone matrix proteins. Osteopontin (OP) is an RGD-containing bone matrix protein, which plays a key role in anchoring osteoclasts to bone surfaces. Besides its role as an anchorage protein, OP is produced by osteoclasts in large quantities (7), 2 and its ␣ v ␤ 3 integrin receptor, besides location in the podosomes, is found in the osteoclast membrane opposite to the bone matrix (4,8). We have shown that osteopontin binding to basolateral ␣ v ␤ 3 is an autocrine motility factor regulating the shape and depth of osteoclast resorption (9). The intracellular biochemical pathways that integrins regulate, and the cellular functions that they control, have recently been the focus of careful scrutiny. Association of focal adhesion kinase (FAK) with the cell surface at focal adhesions directs interactions with the cytoplasmic domains of integrins and their participation in signal transduction (10 -12). Phosphorylation of FAK in response to cell adhesion and other stimuli induce the formation of complexes between FAK and other signaling molecules in vivo, including Src (13,14), Grb2 (15), Nck (16), and PI 3-kinase (17,18).
A unique aspect of signaling through ␣ v ␤ 3 is its property of responding to soluble ligands (8,19,20). We have previously demonstrated the mechanisms of ␣ v ␤ 3 signaling in response to soluble OP (19,21). Binding of OP-or RGD-containing peptides to ␣ v ␤ 3 stimulated formation of signal generating complexes consisting of FAK, c-Src, and PI 3-kinase associated with ␣ v ␤ 3 (19,21). OP stimulated PtdIns 3,4-P 2 and PtdIns 4,5-P 2 (PtdIns-P 2 ) and PtdIns 3,4,5-P 3 (PtdIns-P 3 ) levels in osteoclasts (19). We further defined one specialized domain of increased PtdIns 3,4-P 2 and PtdIns 3,4,5-P 3 levels as an actincapping protein found in the podosome, gelsolin (21). We demonstrated that the increase of PtdIns-P 2 and PtdIns-P 3 associated with gelsolin, uncapped actin oligomers leading to an increase in F-actin content and actin filament formation (21). PtdIns-P 2 regulates several actin-binding proteins, including gelsolin (22), profilin (23), ␣-actinin (24), and vinculin (25,26). Osteopontin affected PtdIns-P 2 and -P 3 levels associated specifically with gelsolin and not the other proteins (21). A marked increase in the level of PtdIns-P 2 with gelsolin leads to the hypothesis that PtdIns-P 2 synthesis may be essential for podosome assembly and disassembly. OP treatment also re-sulted in the increased activity of PI 3-kinase associated with gelsolin. Phosphorylation of PtdIns 4,5-P 2 by PI 3-kinase associated with gelsolin leads to the formation of PtdIns 3,4,5-P 3 in OP-treated cells. The physiological function of the association of PI 3-kinase with gelsolin is stimulation of actin polymerization during cell motility, and inhibition of PI 3-kinase activity with wortmannin, a specific inhibitor of PI 3-kinase, blocks the increase in F-actin and actin polymerization stimulated by osteopontin (21).
Examination of the signaling pathways leading to integrinmediated activation of PI 3-kinase suggests that members of the Src family of nonreceptor kinases may play a role (27)(28)(29)(30). PI 3-kinase activity is associated with c-Src, and the associated kinase activity increases quickly upon stimulation with thrombin in platelets (31). Several lines of evidence reveal that PI 3-kinase is activated by c-Src (32,33). c-Src itself plays an important role in the osteoclast. Deletion of the gene for c-Src in mice results in impaired osteoclast polarization, failure of bone resorption, and osteopetrosis (34,35), Osteoclasts from c-Src deficient mice lack ruffled borders and have impaired bone resorptive activity in vitro (36). Wortmannin inhibited PI 3-kinase activity in osteoclasts both in vivo and in vitro and also ruffled border formation and bone resorption (37). These results suggest that both Src and PI3-kinase activity are important in osteoclastic bone resorption.
To examine the role of c-Src in the pathway to activation of gelsolin-associated PI 3-kinase from OP/␣ v ␤ 3 integrin-mediated signaling, we have utilized an antisense strategy to disrupt the function of Src in the avian osteoclast system. Our results demonstrate that c-Src activity associated with the actin-binding protein gelsolin was increased upon treatment with OP. Furthermore, this increase in Src activity was necessary for the activation of gelsolin-associated PI 3-kinase, and for the subsequent increase in F-actin content and actin filament formation stimulated by liganding of ␣ v ␤ 3 . Thus, c-Src is upstream of gelsolin-associated PI 3-kinase, and it activates PI 3-kinase.
Since the c-Src antisense oligodeoxynucleotides-treated osteoclasts failed to form an organized podosome-containing clear zone, and were deficient in bone resorption, our data address an important second issue regarding the Src knockout phenotype. Our results, at least in vitro, demonstrate that the osteoclast defect of the Src Ϫ/Ϫ mouse is not species-specific. They suggest that the avian osteoclasts also possess cellular sites where Src function cannot be substituted for by another member of the Src superfamily.

EXPERIMENTAL PROCEDURES
Materials and Methods-[␥-32 P]ATP, rainbow molecular weight markers for proteins were obtained from Amersham Pharmacia Biotech. Herbimycin A was obtained from Life Technologies, Inc. Protein A-Sepharose, mouse IgG, anti-gelsolin antibody, phospholipid standards, and most of the chemicals were purchased from Sigma. Polyvinylidene difluoride membranes were obtained from Millipore Corp. (Bedford, MA). Protein assay reagent kit and the reagents for polyacrylamide gel electrophoresis were purchased from Bio-Rad. Antibody to c-Src was purchased from Oncogene (Uniondale, NY). Recombinant c-Src was obtained from UBI (Lake Placid, NY). RNAzol, background quencher, and hybridization solutions were purchased from Tel-Test (Friendswood, TX). Turboblotter transfer system and nylon membranes for RNA transfer were purchased from Schleicher & Schuell. Prime-A-Gene labeling kit and restriction enzymes were purchased from Promega (Madison, WI). Rhodamine-phalloidin was obtained from Molecular Probes (Eugene, OR).
Preparation of Osteoclast Precursors-Avian osteoclast precursors were prepared as described previously (21,38). Briefly, osteoclast precursors were isolated from bone marrow of egg-laying hens maintained on Ca 2ϩ -deficient diets. Partially purified preparations of mononuclear cells were recovered from the interface of Ficoll/Hypaque gradients. Nonadherent cells were separated from the adherent population after 18 -24 h in culture. The nonadherent cells were sedimented, resuspended in fresh medium (5 ϫ 10 6 cells/ml), and cultured in the presence of cytosine arabinoside (5 g/ml). Multinucleated osteoclast precursor cells formed between 3 and 6 days in culture, and the preparations were 70 -90% pure multinucleated tartrate-resistant acid phosphatase-positive cells.
Lysate Preparation, Immunoprecipitation, and Western Blot Analysis-After 4 days in culture, osteoclast precursors were incubated in serum-free media for 2 h. Subsequently cells were treated with one of the following: OP (25 g/ml for 15 min at 37°C) or GRGDS (100 g/ml for 15 min at 37°C). In some cases herbimycin A was added at a concentration of 100 ng/ml for 14 -16 h prior to treatment. Following treatment, cells were washed three times with ice-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 sodium pyrophosphate, 5 mM NaF, 0.1 mM Na 3 VO 4 , 5 mM ZnCl 2 , and 2 mM phenylmethylsulfonyl fluoride. Lysates were pelleted by centrifugation (15,000 rpm, 15 min, 4°C), and the pellets were solubilized by trituration in radioimmune precipitation 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). Protein concentrations were measured using the Bio-Rad protein assay reagent kit, and equal amounts of protein lysates were used for immunoprecipitations. Immunoprecipitations and Western blotting were carried out as described (21).
Immune Complex Kinase Assay Analysis for Src-In vitro immune complex protein kinase assays were performed as described previously (39). Equal amounts of protein lysates were immunoprecipitated with anti-gelsolin or c-Src antibodies. The immune complexes collected by the addition of protein A-Sepharose were used for kinase assays. The Sepharose beads, after washing several times with the buffers described (39), were resuspended in 20 l of kinase buffer (20 mM Hepes, pH 7.4, 5 mM MgCl 2 , and 0.1 mM Na 3 VO 4 ) containing [␥-32 P]ATP (10 Ci) and casein (1 mg/ml) as an exogenous substrate. The mixture was incubated at 25°C for 20 min, and the reaction was stopped by the addition of SDS-sample buffer. The samples were boiled and subjected to SDS-polyacrylamide gel electrophoresis, and radiolabeled proteins were detected by autoradiography.
Oligonucleotides-Experiments were performed with synthetic phosphorothioate oligodeoxynucleotides (ODNs). 24-mer sequences corresponding to positions 1-24 of c-Src cDNA (40) were constructed. The targeted sequences include the presumed translation initiation site. ODNs were made in both sense, and antisense orientations and the sequences are as follows: sense (5Ј-ATG GGG AGC AGC AAG AGC AAG CCC 3Ј) and antisense (5Ј-GGG CTT GCT CTT GCT GCT CCC CAT 3Ј). Scrambled ODNs containing the antisense nucleotides were also synthesized and used as a control: 5Ј-TTT GTT ATC CTC CGT GGC CTC CCG 3Ј.
Cell Permeabilization-Osteoclast precursor cells, cultured for 4 -6 days were used for permeabilization with streptolysin O (41,42). Osteoclasts were washed twice with permeabilization buffer (120 mM KCl, 30 mM NaCl, 10 mM Hepes, pH 7.2, 10 mM EGTA, 10 mM MgCl 2 ) (43). Freshly prepared dithiothreitol (5 mM), ATP (1 mM), and 0.5 unit/ml streptolysin O and ODNs in sense, antisense, or scrambled orientation at various concentrations were added to the buffer at the time of permeabilization. Resealing was achieved by the addition of ␣-minimal essential medium containing 10% fetal bovine serum, and incubation was continued for 8 h. Before the end of the 8 h, the cells were serumstarved for 2 h and stimulated with OP as above. ODNs were maintained at the indicated concentration throughout the time mentioned. Additionally, control cells were permeabilized and stimulated with OP as above but in the absence of ODNs.
Northern Analysis-Total RNA was isolated using RNAzol and processed according to the manufacturer's guidelines (Tel-Test). Total RNA (10 g) was denatured, electrophoresed on 1% agarose/formaldehyde gels and transferred to nylon membranes using a Turboblotter transfer system. The membranes were exposed to UV light to cross link the RNA. 32 P-Labeled pp60 c-src cDNA probes (1.6 kilobase pairs) were random prime-labeled using [␣-32 P]dCTP. The blots were prehybridized at 37°C in a solution containing 1 M NaCl, 1% sodium dodecyl sulfate (SDS), 50% formamide, and 1ϫ background quencher for 1-2 h and hybridized overnight using the solution in high efficiency hybridization system, as directed by the manufacturers' instructions. Blots were then washed once at room temperature in 5ϫ SSPE, 0.1% SDS; once at 37°C in 1ϫ SSPE, 0.1% SDS; and twice at 65°C in 0.1ϫ SSPE, 0.1% SDS. Blots were then analyzed by exposure to x-ray films.
PI 3-Kinase Assay-Lysate preparation and immunoprecipitations were carried out as described above. Equal amounts of proteins were used for immunoprecipitation. The immune complexes, adsorbed to protein A-Sepharose pellets were assayed for lipid kinase activity. PI 3-kinase was assayed by following the method (44) described previously (21). Briefly, the immune complexes, adsorbed to protein A-Sepharose pellets, were washed successively as follows: once with Triton lysis buffer; twice with 0.5 M LiCl, 0.1 M Tris-HCl, pH 8.0; once with 0.1 M NaCl, 1 mM EDTA, 10 mM Tris-HCl, pH 7.6; and finally once with kinase assay buffer (1 mM dithiothreitol, 20 mM Hepes/NaOH, pH 7.4 at 25°C, 5 mM MgCl 2 ). The pellets were then assayed for inositol lipid kinase activity. Freshly prepared lipid mix (20 l; Ref. 44) was added to the beads; the mix was vortexed gently and then placed at 37°C for 5 min. Ten microliters of kinase buffer (20 mM Hepes/NaOH, pH 7.4, 5 mM MgCl 2 , 1 mM dithiothreitol), containing [␥-32 P]ATP (5 Ci/assay), and 5 M Na 2 ATP was then added. The mixture was gently vortexed and the incubation continued at 37°C for an additional 15 min. Incubations were terminated by the addition of 0.425 ml of chloroform: methanol:water (5:10:2, v/v). Lipids were then extracted as described (45) and dried under N 2 . 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 in chloroform:methanol:acetic acid:acetone:water (40:15:13:12:7) and dried. Bands were visualized by autoradiography and quantitated by scanning in GS 300 Transmittance/Reflectance scanning densitometer (Hoefer Scientific Instruments, San Francisco, CA). Phospholipid standards were visualized by exposure to iodine vapors.
Measurement of F-actin Using Rhodamine-Phalloidin Binding-Factin measurement was performed as described (21,46). Cells were cultured in 24-well culture plates for 4 days, incubated with ODNs after permeabilization with streptolysin O, and treated with OP or vehicle as described above. Eight wells were used for each treatment. The cells were fixed with 1.5% formaldehyde in PBS for 15 min, then permeabilized with 0.1% Triton X-100 in PBS for 5 min. The cells were rinsed and incubated with rhodamine-phalloidin in PBS for 30 min. After washing quickly several times with PBS, the cells were extracted with absolute methanol. The fluorescence of each sample was measured using fluorimetry (Gilford Fluoro IV). To assess nonspecific binding, a 10-fold excess of unlabeled phalloidin was used. The nonspecific binding was subtracted from the total binding to get the specific binding.
Actin Staining-Actin staining was carried out as described (47). Cells were rinsed briefly with PBS containing 5 mM EGTA (PBS-EGTA) and fixed in 4% (w/v) paraformaldehyde in PBS-EGTA for 20 min at 37°C. Coverslips were immersed in 47.5% ethanol containing 5 mM EGTA for 15 min at room temperature and rinsed with several changes of PBS-EGTA before staining with 1:20 dilution of rhodamine-phalloidin in PBS-EGTA for 30 min at 37°C. After rinsing several times with PBS-EGTA, coverslips were mounted on a mounting solution (Vector Laboratories Inc., Burlingham, CA). Cells were viewed in a Meridian ACAS 570 (Okemos, MI) with confocal option. Rhodamine-phalloidin images were recorded in 514 argon excitation line with a 40ϫ/1.3 numerical aperture oil objective. Confocal images were processed by the Adobe Photoshop software program (Adobe Systems, Inc., Mountain View, CA).
Bone Resorption Assay-Avian osteoclasts were cultured in 24-well plates for 4 -6 days. On day 5 in culture, cells were permeabilized and treated with or without ODNs as described above. Bone resorption assays were carried out as described previously (38). Cells were incubated with ODNs and [ 3 H]proline-labeled bone particles (2 ϫ 10 4 cpm/ ml) in serum-containing media at 37°C for 12 h. For each treatment, four to six wells were used. Aliquots of supernatants (400 l) were removed at 12 h and replaced with fresh media containing ODNs. Further aliquots were obtained after 24 h. Released isotope from [ 3 H]proline-labeled bone particle by osteoclasts was measured by liquid scintillation counting. The cells were then washed three times with ice-cold PBS and lysed in Triton-containing lysis buffer as described above, and the protein content was determined as described above. Radioactivity was then normalized to protein content.
Statistical Analysis-All comparisons were made to "control," which refers to mock permeabilized cells or mock permeabilized and vehicletreated cells. Data presented are mean Ϯ S.E. of experiments performed at different times, normalized to intra-experimental control values. Statistical comparisons between the treatment groups were done using analysis of variance with the Bonferroni correction for multiple comparisons.

RESULTS
Osteopontin Stimulates c-Src Activity Associated with Gelsolin-We have previously reported osteopontin stimulation of gelsolin-associated PI 3-kinase in avian osteoclasts. The activation of PI 3-kinase associated with Triton-soluble gelsolin was required for the osteopontin stimulation of F-actin (21). PI 3-kinase has been identified in a variety of systems associated with receptor tyrosine kinases (45,48), and it is known to be a substrate for c-Src. Therefore, anti-gelsolin immunoprecipitates made from lysates of vehicle or OP-treated cells were analyzed for Src association with gelsolin by Western blotting (Fig. 1). c-Src was observed in anti-gelsolin immunoprecipitates of Triton-soluble fractions of cell lysates (Fig. 1A, lanes 1  and 2). Negligible amounts of Src were associated with the anti-gelsolin immunoprecipitates of Triton-insoluble fractions (Fig. 1A, lanes 3 and 4). The effect of OP was an increase in gelsolin-associated Src of 75% in three experiments similar to the representative experiment shown in Fig. 1A. The control immunoblots (Fig. 1B, lanes 1-4) revealed equal amounts of gelsolin (90 kDa) loaded between vehicle and OP-treated conditions in the Triton-soluble and -insoluble fractions. To analyze the nature of Src association with gelsolin, we prepared GST-gelsolin fusion proteins and affinity columns. We were unable to demonstrate protein-protein interactions between Src and gelsolin. However, both proteins were associated with phosphatidylinositol 4,5-bisphosphate affinity columns. 2 To further analyze activation of gelsolin-associated Src by OP, in vitro protein kinase assays were performed using casein as the exogenous substrate. The results demonstrated stimulation of c-Src autophosphorylation and phosphorylation of casein when the anti-gelsolin immunoprecipitates from OPtreated cells served as the kinase source (Fig. 1C, lane 3). This observation was specific to the Triton-soluble fraction of cellular gelsolin. Quantitation of six different experiments demonstrated that the increase in phosphorylation of Src was 225 Ϯ 17% and casein was 161 Ϯ 15% (mean Ϯ S.E., p Ͻ 0.01). There was an increase in Src autophosphorylation and kinase activity in immunoprecipitates from GRGDS-treated cells (Fig. 1C,  lane 2), but the stimulation was less than that observed in immunoprecipitates from OP-treated cells. Much less Src activity was observed associated with gelsolin in immunoprecipitates from Triton-insoluble fractions of cell lysates (Fig. 1C,  lanes 4 -6). Immunoprecipitation with nonimmune serum from OP-treated cells did not reveal phosphorylation of the 60-kDa Src (Fig. 1C, lane 7). Herbimycin A, a Src tyrosine kinase inhibitor, inhibited all of the 60-kDa kinase auto-phosphorylation associated with gelsolin and its kinase activity. 2 Further-more, the gelsolin-associated 60-kDa phosphoprotein was confirmed as Src by V8 protease peptide analysis. Between 20 and 25% of the total cellular Src was observed in anti-gelsolin immunoprecipitates from Triton-soluble fractions, and 2-5% of Src kinase activity was observed in anti-gelsolin immunoprecipitates from the Triton-insoluble fractions.
Effect of c-Src ODNs on mRNA Expression-The above observations prompted us to investigate whether the c-Src associated with gelsolin contributed to OP stimulation of PI 3-kinase associated with gelsolin. We utilized an antisense strategy to inhibit Src activity in osteoclasts. Sense and antisense ODNs (24-mer) were constructed to the Src translation initiation region, including the ATG initiation codon. Cells were permeabilized with streptolysin O (S) and incubated with the ODNs (Fig.  2). Antisense ODNs at doses of 10 -40 M decreased Src message levels. As shown in Fig. 2 (A, lane 8, and B) (the mean Ϯ S.E. of three experiments), the reduction in Src levels was 80% at 40 M antisense ODNs, whereas sense ODNs did not decrease Src message levels (Fig. 2, A, lane 3, and B). The bottom panel of Fig. 2A demonstrates the amount of total RNA loaded in each lane. Although the amount of RNA loaded in lanes 6 and 7 was slightly less than the other lanes, the lane corresponding to 40 M antisense ODNs revealed equal loading to controls ( Fig. 2A, lanes 1-3).
Effect of Src ODNs on Western Analyses of Anti-Src Immunoprecipitates-To determine if the decrease in Src mRNA levels induced by antisense Src ODNs was coordinate with decreased protein expression, lysates were made from osteoclasts treated in various ways. Lysates were subjected to immunoprecipitation using anti-Src antibodies followed by immunoblotting with a monoclonal antibody to Src. Levels of Src protein were dramatically reduced by antisense ODNs to Src (Fig. 3, lanes 5 and 6), and no effect was observed with similar doses of sense ODNs (lanes 3 and 4). OP stimulation had no effect on intracellular Src levels (lane 2). Fig. 3 represents Western analyses of immunoprecipitates from two experiments. Antisense ODNs reduced Src protein levels to different degrees in each experiment. Densitometric analysis of five separate experiments revealed that the reduction of Src protein was 83 Ϯ 11%, as compared with control cells.
Effect of Src ODNs on in Vitro Kinase Assays Using Anti-Src Immunoprecipitates-Osteoclast lysates from ODNs-treated (40 M) and untreated cells were subjected to immunoprecipi-  Fig. 4 was also reflected in a reduction in Src activity associated with gelsolin. Treatment with Src antisense ODNs eliminated the OP-induced increase in kinase activity observed in anti-gelsolin immunoprecipitates of cell lysates (Fig. 5A, lane 8). There was no effect of sense (S/OP) (lane 4) or scrambled (Sc/OP) (lane 6) ODNs on OP-stimulated gelsolin-associated Src activity and phosphorylation of casein compared with OP treatment alone (lane 2). Stimulation of phosphorylation of two other proteins of molecular sizes 56 and 52 kDa, in addition to the 60-kDa c-Src, by OP was not affected by antisense to Src (lane 8). The presence of these phosphorylated proteins in the antigelsolin immunoprecipitates was variable. Their identity is unknown despite attempts at their identification (see Fig. 2 for an experiment where they were absent). It is possible that these proteins could be Src family proteins associated with gelsolin. Fig. 5B shows the results of densitometric scans of five experiments expressed as percent of vehicle-treated cells (control). A significant decrease in Src activity (p ϭ Ͻ0.001) was Effect of Src Antisense Oligonucleotides on Gelsolin-associated PI 3-Kinase Activity-To assess the effect of Src inhibition on gelsolin-associated PI 3-kinase activity, anti-gelsolin immunoprecipitates from lysates of variously treated cells were subjected to in vitro phospholipid kinase assays. Consistent with our previous observations (21), OP stimulated phosphorylation of PtdIns 4,5-P 2 to PtdIns 3,4,5-P 3 (Fig. 6A, lane 2) as compared with immunoprecipitates from vehicle-treated cells (lane 1). Likewise, OP increased the PI 3-kinase activity in anti-gelsolin immunoprecipitates formed from osteoclasts pretreated with sense (lane 4) or scrambled ODNs (lane 6) prior to OP treatment. Treatment with Src antisense ODNs significantly reduced both basal (lane 7) and blocked OP-stimulated PI 3-kinase activity (lane 8). Densitometric scans of three experiments expressed as percent of control and mean Ϯ S.E. for all of the treatments is shown in Fig. 6B. The decrease in kinase activity was significant in antisense/OP-treated cells (p Ͻ 0.001) as compared with OP or sense/OP or scrambled/OP-treated cells.

Effect of Src ODNs on in Vitro Kinase Assay Using Antigelsolin Immunoprecipitates-The inhibition of total cellular Src activity by Src antisense ODNs demonstrated in
Since the experiments in Figs. 4 -6 so clearly demonstrate basal activity of gelsolin-associated Src and PI 3-kinase in immunoprecipitates from vehicle-treated cells, we point out that the avian osteoclasts secrete large amounts of OP, and that there is no quiescent cell state in our osteoclast model.
Quantitation of F-actin Content-Previously, we have shown that inhibition of PI 3-kinase by wortmannin inhibited the osteopontin-induced increase in F-actin content in osteoclasts (21). In order to determine whether antisense ODNs to Src had similar effects, we measured the F-actin content of OP-treated osteoclasts. The values plotted in Fig. 7 are means Ϯ S.E. from three experiments. Consistent with our previous observations (21), OP stimulated F-actin content of osteoclasts (Fig. 7, OP). Cells pretreated with sense ODNs for 8 h prior to osteopontin treatment also had increases in F-actin (Fig. 7, S/OP) in response to OP treatment. Quantitative analysis indicated 78 Ϯ 6.5% or 65 Ϯ 8% increases in F-actin content in OP-treated or sense/OP-treated cells, respectively. Cells treated with antisense/OP (AS/OP) did not show an increase in F-actin content in response to osteopontin.
Actin Staining-To further demonstrate that the changes in F-actin were associated with changes in organization and distribution of the actin cytoskeleton, cells were stained with rhodamine-phalloidin to visualize actin filaments. F-actin was enriched in dot-like structures in the periphery of the vehicletreated cells (Fig. 8A), representing short perpendicular actin filaments in podosomes. OP stimulation for 15 min, the time of incubation used in the studies described above, induced an increase in actin stress fiber formation along with a reorganization of the peripheral adhesion zone demonstrating either actin filaments (Fig. 8B, left) or membrane ruffling (Fig. 8B,  right) replacing the podosomes. A similar response is seen in sense/OP-treated cells (Fig. 8C). These stress fibers were absent in antisense/OP-treated cells (Fig. 8D); instead, membrane ruffles were seen at the periphery of the cell, and the clear zone organization was absent.
Effect of Antisense Oligomers to Src on Bone Resorption-Observations made by others have demonstrated that oste- oclasts of Src deficient mice fail to form ruffled borders and bone resorption lacunae (36,49). To determine whether our experiments using antisense ODNs to Src would result in a decrease in osteoclast-mediated bone resorption, we performed bone resorption assays using bone particles labeled with [ 3 H]hydroxyproline. Cells were permeabilized with streptolysin O and incubated with or without ODNs. After 12 h, media were removed, fresh media with or without ODNs were added, and incubation was continued for another 12 h. Media collected after 12 and 24 h were filtered and counted. Six to eight wells were used for each treatment. The values shown in Table I are means Ϯ S.E. from three experiments. Cells permeabilized only (control) or cells-treated with sense ODNs demonstrated equal levels of bone resorption either at 12 or 24 h, whereas treatment with antisense ODNs to Src significantly decreased bone resorption as compared with controls. Under similar conditions, we have previously shown that OP significantly stimulated bone resorption by 2-fold. There was no response to OP in cells pretreated with antisense ODNs. Pit assays of bone resorption were not performed because the avian osteoclast precursor model used in these experiments produces only superficial resorption pits on dentine slices as reported by several laboratories. DISCUSSION In this report, we demonstrate that Src is associated with gelsolin and its stimulation is required for OP stimulation of gelsolin-associated PI 3-kinase. Src immunoprecipitates with gelsolin, and OP stimulates the kinase activity of Src associated with gelsolin in a Triton-soluble fraction of cell lysates. The cellular function of gelsolin-associated Src was demonstrated by the absence of podosomes and cytoskeletal reorganization stimulated by osteopontin when Src was missing in the gelsolin-associated complex.
Activation of nonreceptor protein tyrosine kinases play a central role in assembly of focal adhesions and the propagation of signals triggered by integrin receptors (50,51). Src-dependent phosphorylation of FAK and the focal adhesion-associated proteins contributes to their assembly and activates signaling events (52). Activation of cell surface integrins has been shown to rapidly increase the tyrosine phosphorylation of FAK and the focal adhesion-associated proteins, paxillin and tensin (53)(54)(55). Phosphorylation of FAK on tyrosine 397 creates a binding site for the SH2 domain of Src (and Fyn) and results in the formation of a FAK/Src complex, which is an early event in assembly of focal adhesion complexes and the activation of integrin signaling pathways (16,51,56). In agreement with this concept, our data suggest that Src is required for podosome assembly, since they were absent in the Src antisense-treated cells.
We have previously shown that treatment of osteoclasts with OP stimulated formation of signal-generating complexes consisting of ␣ v ␤ 3 , FAK, PI 3-kinase, and Src (19). OP stimulates synthesis of phosphoinositides such as PtdIns-P 2 , PtdIns 3,4-P 2 , and PtdIns-P 3 through the activation of the respective phospholipid kinases in osteoclasts (19). The increase in PtdIns-P 2 and PtdIns-P 3 levels associated with Triton-soluble gelsolin leads to actin uncapping, stimulation of F-actin formation, and actin filament formation in osteoclasts. The increase in PtdIns-P 3 levels associated with gelsolin is due to translocation of PI 3-kinase to gelsolin in a Triton-soluble cell fraction (21), and the increase in PI 4,5-P 2 may be due to activation of PI 4-phosphate 5-kinase. Chong et al. (57) have suggested that liganding of integrins by matrix activates PI 4-phosphate 5-kinase. The increase in the levels of PtdIns 4,5-P 2 associated with gelsolin in OP-treated cells are decreased by pretreatment with antibody to integrin ␣ v ␤ 3 (LM609) (21) or herbimycin A. The findings of gelsolin-associated Src stimulated by osteopontin suggest that Src might activate gelsolin-associated PI 3-kinase. This implication was substantiated by our findings that antisense ODNs to Src blocked OP-induced PI 3-kinase activation. Translocation and activation of gelsolin-associated PI 3-kinase may have occurred through binding of the p85 subunit of PI 3-kinase to gelsolin-associated Src (30), resulting in p85 phosphorylation and subsequent formation and activation of p85/ p110 heterodimers (the active form of PI 3-kinase).
The PI 3-kinases are a family of phospholipid kinases that phosphorylate PtdIns in the 3-hydroxyl position of the inositol ring. The prototypical PI 3-kinase is a heterodimer comprised of a regulatory subunit, p85, and a catalytic subunit, p110. The catalytic subunit is activated by association with p85 and association is stimulated by tyrosine phosphorylation of the p85 SH 2 domain. p85 is a known substrate of Src. In recent years, several isoforms of p85 have been discovered, and isoforms of p110 have also been found. In addition, catalytic subunits of PI 3-kinase that do not require association with the regulatory subunit have been discovered. The forms of PI 3-kinase in the osteoclast have not been determined. However, the PI 3-kinase associated with gelsolin is a prototypical form (21), and gelsolin-associated p85 phosphorylation is stimulated by osteopontin/␣ v ␤ 3 (21).
Several lines of evidence suggest that c-Src plays a role in regulating actin rearrangement in normal cells. Analysis of murine fibroblasts over expressing c-Src (Kϩ) and dominant negative variants of Src (KϪ) have demonstrated pronounced actin rearrangement in Kϩ cells, which was absent in KϪ cells (58). These results suggest a role for c-Src in actin rearrangement. Moreover, it has been shown that oncogenic function of v-Src requires its translocation to the cytoskeleton at sites of focal adhesion (59). High levels of c-Src expression are part of the osteoclast phenotype, and other members of the Src family, which account for redundancy in other cells, do not compensate for the absence of Src in the Src Ϫ/Ϫ mice (34). Therefore, we sought to determine whether Src activation is essential for changes in the actin cytoskeleton during stimulation of bone resorption by OP in avian osteoclasts. Failure of osteopontinstimulated actin stress fiber formation by Src antisense ODNtreated cells demonstrates that activation of gelsolin-associated Src is an important regulatory mechanism in osteoclast function. Src has been shown to be required for the formation of ruffled borders and bone resorption, but Src is not required for osteoclast development (60 -62). Inhibition of ruffled border formation and the in vitro pit forming activity of osteoclast by wortmannin confirms the role of PI 3-kinase in pathways controlling osteoclastic bone resorption (37,63). Thus, the antisense Src knock-out experiments may explain not only the role of gelsolin-associated PI 3-kinase in cellular responses, but also how it is activated and the pathway coupled to this response.
That activation of gelsolin-associated PI 3-kinase is dependent upon Src activation is evident from the in vitro Src knockout experiments. Although osteoclast gelsolin appears to be central to localization of tyrosine kinases involved in regulation of the actin cytoskeleton, the exact mechanism of gelsolin/Src interaction is unknown. The activation of gelsolin-associated PI 3-kinase by Src does not necessarily mean that it is directly associated with gelsolin. An intriguing question is by what mechanisms are Src and PI 3-kinase associated with gelsolin. There is now compelling evidence that PI 3-kinase lipid products directly associate with the SH2 domain of kinases (64). This observation suggests that the stimulation of Src association with gelsolin may serve to organize the gelsolin/PI 3-kinase complex. We have previously demonstrated that OP-stimulated translocation of PI 3-kinase from Triton-insoluble to Triton-soluble gelsolin (21). OP stimulated the levels of gelsolin-associated PtdIns 4,5-P 2 which is a known substrate for PI 3-kinase. The increase in the synthesis of PtdIns 3,4,5-P 3 in Triton-soluble gelsolin by OP may function in the recruitment of signaling molecules. Further studies need to be directed at determining how these interactions occur and are regulated in vivo.