The Actin-Binding Protein PPP1r18 Regulates Maturation, Actin Organization, and Bone Resorption Activity of Osteoclasts

ABSTRACT

Osteoclasts resorb bone by attaching on the bone matrix and forming a sealing zone. In Src-deficient mice, osteoclasts cannot form the actin ring, a characteristic actin structure that seals the resorbed area, and resorb hardly any bone as a result. However, the molecular mechanism underlying the role of Src in the regulation and organization of the actin ring is still unclear. We identified an actin-regulatory protein, protein phosphatase 1 regulatory subunit 18 (PPP1r18), as an Src-binding protein in an Src-, Yes-, and Fyn-deficient fibroblast (SYF) cell line overexpressing a constitutively active form of Src. PPP1r18 was localized in the nucleus and actin ring. PPP1r18 overexpression in osteoclasts inhibited terminal differentiation, actin ring formation, and bone-resorbing activity. A mutation of the protein phosphatase 1 (PP1)-binding domain of PPP1r18 rescued these phenotypes. In contrast, PPP1r18 knockdown promoted terminal differentiation and actin ring formation. In summary, we showed that PPP1r18 likely plays a role in podosome organization and bone resorption.

INTRODUCTION

To maintain bone homeostasis by resorbing the bone, osteoclasts become differentiated from hematopoietic cells in response to stimulation by receptor activator NF-κB ligand (RANKL) and macrophage colony-stimulating factor (M-CSF) produced by osteoblasts or osteocytes (1). RANKL signaling promotes expression and activation of nuclear factor of activated T cells 1 (NFATc1), a transcription factor and master regulator of osteoclastogenesis, which upregulates the expression of various molecules that accelerate osteoclastic differentiation and bone resorption, such as dendrocyte-expressed seven transmembrane protein (DC-Stamp), osteoclast-associated, immunoglobulin-like receptor (OSCAR), β3 integrin, Src, and cathepsin K (1–3). During differentiation, osteoclast precursor cells fuse with each other, spread, and form the actin ring, a unique actin structure at the cell periphery (4–6). Osteoclasts strongly attach to the bone matrix, demarcate the bone-resorbing area by sealing it with the actin ring, and form a ruffled border to secrete bone-resorbing factors, such as protons and cathepsin K (4–10). Thus, the formation of the actin ring and ruffled border is necessary for bone resorption (7, 8, 11).

Tyrosine kinase Src plays important roles in the function of osteoclasts (7, 8, 11, 12). Although Src is ubiquitously expressed and involved in many cell functions, such as growth, expansion, proliferation, and migration, one of the few notable phenotypes of Src-deficient mice is osteopetrosis that occurs due to the dysfunction of osteoclasts (7, 8, 11, 12). Src-deficient osteoclasts cannot form the actin ring and ruffled border, and their bone resorption capacity is impaired (8, 9, 13). Src contains a unique region and three major domains: the Src homology 3 domain (SH3), the Src homology 2 domain (SH2), and a kinase catalytic domain (14, 15). Src is localized in the cell membrane, where it binds to actin regulatory proteins, such as focal adhesion kinase (FAK), cortactin, Crk-associated substrate (130Cas), protein tyrosine kinase 2 beta (Pyk2), and casitas B-lineage lymphoma (c-Cbl), through its SH3 and SH2 domains (10, 16–27). After binding, Src phosphorylates these proteins, regulates actin accumulation, and leads to the formation of the actin ring in osteoclasts (13, 25). Although p130Cas, Pyk2, or c-Cbl deficiencies in osteoclasts result in an osteopetrotic phenotype due to osteoclast dysfunction, it is milder than that observed in Src-deficient mice (19, 23, 26). This suggests that there are unknown proteins that interact with Src and regulate bone resorption by osteoclasts.

Protein phosphatase 1 regulatory subunit 18 (PPP1r18), also known as protein phosphatase 1 F-actin cytoskeleton-targeting subunit (phostensin), is encoded by the KIAA1949 gene in Homo sapiens. There are long (613-amino-acid) and short (165-amino-acid) isoforms of PPP1r18 (28, 29). PPP1r18 was identified as a PP1-binding protein by yeast two-hybrid assay and was revealed to be an actin-regulatory protein (28, 30, 31). PPP1r18 expression is decreased in malignant and metastatic breast cancer cells (32). These observations suggest that PPP1r18 regulates actin elongation and perhaps actin dynamics. However, the expression pattern of PPP1r18 and its role in actin dynamics are still unclear.

The actin ring in osteoclasts is formed by the accumulation of dot-like structures, the podosomes (4, 13, 33). Src plays an important role in podosome formation of the actin ring and turnover in osteoclasts (13). Expression of the constitutively active mutant forms of Src, namely, v-Src and Src Y527F, promotes podosome formation and migration in tumors, normal fibroblasts, epithelial cells, and many other cell types (34–36). Here, we overexpressed Src Y527F in Src-, Yes-, and Fyn-deficient fibroblast (SYF) cells (37) in an attempt to isolate Src-binding proteins responsible for podosome formation. We identified PPP1r18 as an Src-binding protein that has an important role in the maturation of osteoclasts and regulation of actin ring formation in osteoclasts.

RESULTS

Identification of PPP1r18 as an Src-binding protein in podosomes.

To identify Src-binding proteins involved in podosome formation, we first generated SYF cells that stably expressed Flag-tagged Src Y527F and formed podosomes instead of actin stress fibers (Fig. 1A and B). Lysates from control or Src Y527F-transfected cells were immunoprecipitated with an anti-Flag antibody. The resulting proteins were analyzed by nanoscale-microcapillary reversed-phase liquid chromatography-tandem mass spectrometry (LC-MS/MS) to identify Src-binding proteins (see Tables S1 and S2 in the supplemental material). Actin-binding proteins that were involved in podosome formation included cortactin (Cttn) (16–18), actin-related protein (Arp) 2/3 complex (Arp2, Arp3, and Arp 2/3 complex subunit 4 are shown as Actr2 and Apct4) (38), and tubulin (tubulin alpha 1a and tubulin beta 2b are shown as Tuba1a and Tubb2b) (Tables S1 and S2) (16). Because Src has no actin-binding domain, it likely regulates actin ring formation through actin-binding proteins. Thus, we examined the expression of different proteins identified by LC-MS/MS analysis, looking for proteins with an actin-binding site but for which no previous reports about the role in osteoclasts had been published. The actin-binding protein PPP1r18 (2310014H01Rik), to which a clear role in osteoclasts has not been attributed previously, was expressed at a level comparable to that of cortactin, according to LC-MS/MS data. The number of identified peptides indicates the expression level of the protein (Table S1). Therefore, we focused our further efforts on the characterization of PPP1r18 in osteoclasts.

FIG 1.

Identification of PPP1r18 in SYF cells and its expression in osteoclasts. (A) Rhodamine-phalloidin staining of control or stably Src Y527F-transfected SYF cells. The scale bar indicates 30 μm. (B) Control or Src Y527F-expressing SYF cells were lysed, and the expression of Src was analyzed by Western blotting.

Expression and localization of PPP1r18 in the actin ring of tartrate-resistant acid phosphatase-positive multinuclear cells regulated by Src.

To examine whether PPP1r18 was expressed in osteoclasts, we performed Western blotting. Spleen cells were isolated and cultured with M-CSF for 3 days to induce their differentiation into spleen macrophages, which act as osteoclast precursor cells. The latter cells were stimulated by RANKL and M-CSF for their differentiation into tartrate-resistant acid phosphatase-positive [TRAP(+)] multinuclear (containing >3 nuclei) cells (MNCs), regarded as osteoclasts. PPP1r18 was expressed in TRAP(+) MNCs, but at a lower level than that in precursor cells (Fig. 2A). We next examined the transition of the expression level of PPP1r18 during osteoclastogenesis. TRAP(+) MNCs appeared after 5 days of RANKL stimulation and became large, round cells that had actin rings (Fig. 2B). The expression level of PPP1r18 gradually decreased during osteoclast differentiation (Fig. 2C). To examine the intracellular localization of PPP1r18 in TRAP(+) MNCs, we performed immunofluorescence analysis and found that PPP1r18 was localized in the actin ring, where it colocalized with Src (Fig. 2D and E). PPP1r18 was also colocalized with cortactin and vinculin, which are localized in podosomes of the actin ring to regulate actin ring formation (18, 39) (Fig. 2F and G). These results suggest that PPP1r18 is involved in podosome and actin ring formation with Src in osteoclasts.

FIG 2.

Colocalization of PPP1r18 with Src in the actin rings of osteoclasts. Spleen cells were cultured in the presence of 100 ng/ml macrophage colony-stimulating factor (M-CSF) for 3 days and differentiated to spleen macrophages (Mφ). Mφ then were cultured in the presence of 100 ng/ml receptor activator NF-κB ligand (RANKL) and 30 ng/ml M-CSF for 6 days and differentiated to tartrate-resistant acid phosphatase-positive multinuclear cells [TRAP(+) MNCs]. (A) Bone marrow cells were flushed out with phosphate-buffered saline, centrifuged, and lysed in lysis buffer. Cultured Mφ and TRAP(+) MNCs were lysed after being washed with PBS. The expression of PPP1r18 was analyzed by Western blotting. (B) Mφ were cultured with 100 ng/ml RANKL and 30 ng/ml M-CSF for 6 days. Cells were fixed at days 3 to 6 after RANKL stimulation and stained with TRAP. (C) Mφ were cultured with 100 ng/ml RANKL and 30 ng/ml M-CSF for 6 days. Cells were lysed at days 3 to 6 after RANKL stimulation, and the expression level of PPP1r18 was evaluated by Western blotting. (D to G) TRAP(+) MNCs were differentiated on cover glasses in the presence of 100 ng/ml RANKL and 30 ng/ml M-CSF. Following fixation, TRAP(+) MNCs were incubated with anti-PPP1r18 and anti-rabbit Alexa Fluor 488-conjugated antibodies (D), anti-Src and anti-mouse Alexa Fluor 488-conjugated antibodies (E), anticortactin and anti-mouse Alexa Fluor 488-conjugated antibodies (F), or antivinculin and anti-mouse Alexa Fluor 488-conjugated antibodies (G), rhodamine-phalloidin, and 4′,6-diamidino-2-phenylindole (DAPI). The scale bars indicate 50 μm. (H) Cell lysates obtained from spleen Mφ and TRAP(+) MNCs were immunoprecipitated with an anti-Src antibody and protein G-Sepharose. The expression of PPP1r18 (upper panels) and Src (lower panels) was analyzed by Western blotting. IP, immunoprecipitation; WB, Western blotting. (I) PPP1r18 and either empty or Src Y527F vector were transfected into SYF cells. After 2 days in culture, proteins were harvested with cell lysis and immunoprecipitated with an anti-Myc tag antibody and protein G-Sepharose. Phosphotyrosine, PPP1r18, Src, and actin were detected by Western blotting with the respective antibodies. (J) TRAP(+) MNCs were differentiated from Src^−/− mouse spleen cells and transduced with empty (control) adenoviruses (upper panels) or Src Y527F-containing adenoviruses (Adx Src Y527F) (lower panels). Osteoclasts were stained with anti-PPP1r18 and anti-rabbit Alexa Fluor 488-conjugated antibodies and rhodamine-phalloidin. The scale bars indicate 50 μm. Representative data from at least two mice are shown for all experiments.

To confirm PPP1r18 binding to Src in osteoclasts, we performed immunoprecipitation experiments using cell lysates from spleen macrophages and TRAP(+) MNCs. Src was abundantly expressed in TRAP(+) MNCs (11) and could bind to PPP1r18, although the expression of PPP1r18 in TRAP(+) MNCs was low (Fig. 2H). Because Src is a tyrosine kinase, we tested whether it can lead to tyrosine phosphorylation of PPP1r18, using cells transduced with adenoviruses harboring constitutively active Src Y527F. However, no tyrosine residues in PPP1r18 were phosphorylated (Fig. 2I). Because constitutively active Src (Src Y527F) would affect actin accumulation and podosome formation, we next asked whether Src regulated PPP1r18 localization. In Src-deficient TRAP(+) MNCs, PPP1r18 was ubiquitously localized, whereas Src Y527 expression induced PPP1r18 accumulation in the actin rings of podosomes (Fig. 2J). These results suggest that PPP1r18 binds Src and is recruited to actin-rich regions, possibly contributing to podosome formation.

PPP1r18 overexpression negatively regulates maturation and actin ring formation in osteoclasts.

We next examined the role of PPP1r18 in TRAP(+) MNCs. To determine the effect of PPP1r18 overexpression, TRAP(+) MNCs were differentiated with 100 ng/ml recombinant human soluble RANKL (sRANKL) and 30 ng/ml M-CSF for 6 days and then transduced either with empty (control) adenoviruses or with those containing PPP1r18 sequence. After transduction, the cells were cultured for 1 day and examined by TRAP cell and actin staining (Fig. 3A and B). Although the number of TRAP(+) MNCs was not affected, the TRAP(+) MNCs became smaller following PPP1r18 overexpression (Fig. 3A to D). Consistent with the change in cell size, the number of nuclei in PPP1r18-overexpressing TRAP(+) MNCs was lower (Fig. 3E). Moreover, overexpression of PPP1r18 reduced the percentage of cells with formed actin ring (Fig. 3F). We next examined the effect of PPP1r18 on osteoclastic differentiation by real-time quantitative PCR (qPCR). Overexpression of PPP1r18 did not affect the expression levels of Acp5, Nfatc1, or Ctsk mRNA (Fig. 3G). However, the expression level of Dcstamp was inhibited by PPP1r18 overexpression (Fig. 3G). These results suggest that overexpression of PPP1r18 in TRAP(+) MNCs suppressed cell fusion, maturation, and actin ring formation in osteoclasts.

FIG 3.

Inhibition of osteoclast maturation and actin ring formation by PPP1r18 overexpression. TRAP(+) multinuclear cells (MNCs) were differentiated from spleen cells with macrophage colony-stimulating factor (M-CSF) and receptor activator NF-κB ligand (RANKL) and transduced with empty vector (control)- or PPP1r18-carrying adenoviruses at a multiplicity of infection value of 150. (A) The expression of PPP1r18 in control and PPP1r18-transduced osteoclasts was analyzed by Western blotting. (B) TRAP(+) MNCs were fixed and stained with TRAP and rhodamine-phalloidin. The scale bars indicate 50 μm. (C to F) The number of TRAP(+) MNCs (C), size of TRAP(+) MNCs (D), number of nuclei in TRAP(+) MNCs (E), and number of cells with an actin ring (F) were determined (mean ± SD; n = 4). *, P < 0.01. (G) The expression levels of osteoclast marker genes in spleen macrophages (Mφ) and TRAP(+) MNCs treated with either empty vector (control)- or PPP1r18-carrying adenoviruses for 1 day were examined by qPCR. Representative data from at least two mice are shown for all experiments.

The PPP1CA-binding site in PPP1r18 plays a key role in actin ring formation.

PPP1r18 binds to protein phosphatase 1 (PP1) via a PP1-binding motif, the Lys-Ile-Ser-Phe sequence (amino acid residues 539 to 542) (Fig. 4A), and this interaction likely regulates PP1 activity (28, 29). Mutation of PPP1r18 Ile540 and Phe542 to Gly (IGFG mutant) resulted in the loss of PPP1r18 binding to PP1 (Fig. 4A), as has also been previously reported (28). IGFG mutant PPP1r18 did not bind to PP1 phosphatase catalytic subunit alpha (PPP1CA), despite the fact that wild-type PPP1r18 could bind to PPP1CA in TRAP(+) MNCs (Fig. 4B). To examine the effect of PPP1r18 binding to PP1 on the maturation and actin ring formation of TRAP(+) MNCs, we overexpressed PPP1r18 with the IGFG mutation in TRAP(+) MNCs. Overexpression of IGFG mutant PPP1r18 did not affect the number of TRAP(+) MNCs. Furthermore, the mutant protein was localized in the nuclei, and the actin ring was similar to that seen in the presence of endogenous wild-type PPP1r18 (Fig. 5A and B). Although overexpression of wild-type PPP1r18 reduced cell size, decreased the number of nuclei in the cells, and suppressed actin ring formation, overexpression of IGFG mutant PPP1r18 did not have these effects (Fig. 5A to E). We next examined whether PPP1r18 regulates PP1 localization. PP1 was localized at the actin ring and nuclei in osteoclasts (Fig. 5F). Overexpression of wild-type PPP1r18 disturbed PP1 localization that was similar to PPP1r18 localization (Fig. 5F and G). In contrast, PP1 not only was localized at the actin ring and nuclear region but also was localized ubiquitously at low levels in osteoclasts overexpressing the PPP1r18 IGFG mutant, although the PPP1r18 IGFG mutant was localized at the actin ring (Fig. 5F and G). These results suggest that PPP1r18 regulates PP1 localization. To determine whether PPP1r18 and PP1 affect bone resorption, we performed the pit formation assay. TRAP(+) MNCs were differentiated by coculture with osteoblasts and bone marrow cells, because TRAP(+) MNCs differentiated from spleen cells with sRANKL and M-CSF are known to exhibit weak resorbing capacity (23). Overexpression of wild-type PPP1r18 suppressed pit formation in dentin slices, whereas overexpression of mutant IGFG PPP1r18 did not (Fig. 5H to J). These results suggest that PPP1r18 binding to the catalytic subunit of PP1 is important for the regulation of osteoclast maturation, actin ring formation, and bone resorption.

FIG 4.

Binding of PPP1r18 to PP1 through the PP1-binding motif. (A) Schematic representation of PPP1r18. PPP1r18 binds to PP1 at amino acid residues 539 to 542 (KISF sequence). Ile540 and Phe542 were mutated to Gly (PPP1r18 IGFG mutant). (B) TRAP(+) multinuclear cells (MNCs) were infected at a multiplicity of infection value of 150 with empty vector (shown as a control)-, wild-type Myc-tagged PPP1r18-, or Myc-tagged PPP1r18 IGFG mutant-carrying adenoviruses for 1 day. The cells were lysed and subjected to immunoprecipitation with an anti-Myc tag antibody. The expression of PPP1CA (upper panels) and PPP1r18 (anti-Myc tag) (lower panels) was analyzed by Western blotting. IP, immunoprecipitation; WB, Western blotting. Representative data from at least two mice are shown for all experiments.

FIG 5.

Importance of the PP1-binding domain of PPP1r18 in actin ring formation. (A to G) Overexpression of wild-type PPP1r18 (PPP1r18wt) or the PPP1r18 IGFG mutant in TRAP(+) multinuclear cells (MNCs) by using transfection with adenoviruses for 1 day. Cells transfected with empty adenoviruses served as a control. (A) The cells were fixed and incubated with anti-PPP1r18 and anti-rabbit Alexa Fluor 488-conjugated antibodies, phalloidin, and DAPI. The scale bars indicate 50 μm. (B to E) Number of TRAP(+) MNCs (B), size of TRAP(+) MNCs (C), number of nuclei in TRAP(+) MNCs (D), and number of cells with an actin ring (E) (mean ± SD; n = 4). *, P < 0.01; **, P < 0.05. (F) The cells were fixed and incubated with anti-PPP1CA and anti-rabbit Alexa Fluor 488-conjugated antibodies, phalloidin, and DAPI. The scale bars indicate 30 μm. (G) The cells were fixed and incubated with anti-Myc tag- and anti-rabbit Alexa Fluor 488-conjugated antibodies, phalloidin, and DAPI. The scale bars indicate 30 μm. (H to J) Osteoclasts were differentiated by coculture of primary osteoblasts and bone marrow stroma cells on collagen gels with PGE₂ and 1α,25-dihydroxyvitamin D₃. (H) Wild-type PPP1r18 or IGFG mutant PPP1r18 was introduced at day 7, and the cells were cultured for an additional day. Subsequently, the cells were harvested with a 0.2% collagenase solution in phosphate-buffered saline, and the same number of TRAP(+) MNCs were replated on dentin slices. After 48 h of culture, the cells were fixed and stained with TRAP. The scale bars indicate 50 μm. (I) The TRAP(+) MNCs were counted (mean ± SD; n = 4). After the cells were removed from the dentin slices, the pit was stained with hematoxylin. (J) The pit images were analyzed with ImageJ (mean ± SD; n = 4). *, P < 0.01. Representative data from at least two mice are shown for all experiments.

Suppression of PPP1r18 upregulates bone resorption by promoting NFATc1 activation.

Next, TRAP(+) MNCs differentiated with 100 ng/ml sRANKL and 30 ng/ml M-CSF were infected with adenoviruses containing either control or PPP1r18 short hairpin RNA (shRNA) to determine the effects of PPP1r18 depletion. After infection and culturing the cells for 1 day, we examined their characteristics by TRAP and actin staining. Although PPP1r18 expression was reduced by the infection with PPP1r18 shRNA adenoviruses, this treatment did not affect the number of TRAP(+) MNCs or actin ring formation in TRAP(+) MNCs (see Fig. S1A to D in the supplemental material). Thus, we examined the consequences of reduced concentrations of sRANKL used to stimulate TRAP(+) MNC differentiation. The fraction of TRAP(+) MNCs that formed actin rings after knockdown of PPP1r18 increased 3-fold following the stimulation with 25 ng/ml sRANKL and 1.5-fold following the incubation with 50 ng/ml and 75 ng/ml sRANKL (Fig. S1E to L and 6A to F). However, the number of TRAP(+) MNCs was significantly decreased by shPPP1r18 treatment when differentiation was stimulated by 25 ng/ml and 50 ng/ml sRANKL (Fig. S1K and 6C). In contrast, the size of TRAP(+) MNCs and the number of nuclei in TRAP(+) MNCs were increased by PPP1r18 knockdown (Fig. 6D and E). In addition, the expression levels of Nfatc1, Acp5, Ctsk, and Dcstamp mRNAs were enhanced by PPP1r18 knockdown (Fig. 6G). Moreover, PPP1r18 knockdown in TRAP(+) MNCs differentiated by coculture activated bone resorbing activity (Fig. 6H to J). These results suggest that a decrease in PPP1r18 levels promotes not only actin ring formation but also differentiation and maturation of osteoclasts as well as subsequent bone resorbing activity.

FIG 6.

Acceleration of osteoclast maturation and actin ring formation by PPP1r18 knockdown. Spleen macrophages were cultured in the presence of 50 ng/ml receptor activator NF-κB ligand (RANKL) and 30 ng/ml macrophage colony-stimulating factor (M-CSF) for 6 days. Cells were infected at a multiplicity of infection of 150 with control (shControl) or PPP1r18 (shPPP1r18) shRNA-harboring adenoviruses and cultured for 1 day in the presence of 50 ng/ml RANKL and 30 ng/ml M-CSF. The cells then were fixed and stained with TRAP and phalloidin. (A) TRAP(+) multinuclear cells (MNCs) were lysed, and the expression of PPP1r18 was analyzed by Western blotting. (B) TRAP(+) MNCs were fixed and stained with TRAP and rhodamine-phalloidin. The scale bars indicate 50 μm. (C to F) The number of TRAP(+) MNCs (C), size of TRAP(+) MNCs (D), number of nuclei in TRAP(+) MNCs (E), and number of cells with an actin ring (F) were determined (mean ± SD; n = 4). *, P < 0.01. (G) The expression levels of osteoclast marker genes in spleen macrophages (Mφ) and in TRAP(+) MNCs infected with control shRNA or PPP1r18 shRNA for 1 day were examined by qPCR. (H to J) Osteoclasts were differentiated by coculture of primary osteoblasts and bone marrow stroma cells on collagen gels with PGE₂ and 1α,25-dihydroxyvitamin D₃. Control or PPP1r18 shRNA was introduced at day 7, and the cells were cultured for an additional day. Subsequently, the cells were harvested with a 0.2% collagenase solution in phosphate-buffered saline, and the same number of TRAP(+) MNCs were replated on dentin slices. (H) After 48 h in culture, the cells were fixed and stained with TRAP. The scale bars indicate 50 μm. (I) The TRAP(+) MNCs were counted (mean ± SD; n = 4). After the cells were removed from dentin slices, the pit was stained with hematoxylin. (J) Pit images were analyzed by ImageJ, and pit areas were expressed as the mean ± SD (n = 4). *, P < 0.01. Representative data from at least two individual mice are shown for all experiments.

Regulation of c-Src Ser17 phosphorylation and its downstream signaling by PPP1r18 and PP1.

We finally investigated the molecular mechanisms by which PPP1r18 regulates actin ring formation. Upon overexpression of wild-type PPP1r18, phosphorylation of Src Ser17 and cortactin Tyr421, one of the catalytic targets of c-Src during actin ring formation, was decreased (Fig. 7A). In contrast, phosphorylation of both Src Ser17 and cortactin Tyr421 did not change upon PPP1r18 IGFG overexpression (Fig. 7A). Moreover, wild-type PPP1r18 overexpression inhibited c-Src binding to cortactin and p130Cas, whereas overexpression of the PPP1r18 IGFG mutant had no such effect (Fig. 7B). These results suggest that the complex between PPP1r18 and PP1 dephosphorylates c-Src Ser17 and inhibits Src signaling and the consequent formation of the actin ring.

FIG 7.

Regulation of c-Src by the complex between PPP1r18 and PP1. TRAP(+) MNCs were infected with empty vector-, wild-type PPP1r18-, or PPP1r18 IGFG mutant-harboring adenovirus. (A) The cells were lysed, and the phosphorylation of c-Src Ser17 (top panel) and cortactin Tyr421 (third panel) was assessed by Western blotting. (B) Cell lysates were immunoprecipitated with an anti-Src antibody and protein G, and the protein expression was analyzed by Western blotting. IP, immunoprecipitation; WB, Western blotting. Representative data from at least two mice are shown for all experiments.

DISCUSSION

Attachment of osteoclasts to the extracellular matrix in bone tissues is essential for bone resorption. Src activity plays an important role in this process by regulating actin ring formation, although the molecular mechanisms of Src effects are not fully understood. To resolve this, we performed mass spectrometry analysis of Src-binding proteins that were extracted from podosomes formed by SYF cells overexpressing Src Y527F. We have validated the accuracy of our approach by identifying known Src-binding proteins, such as cortactin (16–18), proteins comprising the actin-related protein 2/3 complex (38), and tubulin (16). Src does not have an actin-binding domain but can regulate actin dynamically. In addition, Src binds to actin-binding proteins. Thus, in the set of identified Src-binding partners, we specifically focused on the actin-binding protein PPP1r18, whose expression levels were similar to those of cortactin according to LC-MS/MS analysis. PPP1r18 was also localized very close to Src in podosomes formed by Src Y527F-overexpressing SYF cells. This suggests that PPP1r18 is involved in podosome formation by Src. Because the role of PPP1r18 in osteoclasts has not been well established, we examined it in the present study.

Although PPP1r18 has long and short isoforms, we analyzed only the former, because the short isoform was hardly detected in our setting. cDNA sequences of the long and short isoforms are redundant and cannot be easily distinguished. Thus, we could not examine the expression levels of the long and short isoforms separately by real-time qPCR. Only the long form of PPPr18 was recognized by the anti-PPP1r18 antibody (H-300). However, in general, there are no good antibodies for the detection of the short PPPr18 isoform.

Even though the expression level of PPP1r18 was low in TRAP(+) MNCs, PPP1r18 was colocalized with Src in the actin rings of osteoclasts. Furthermore, PPP1r18 was coimmunoprecipitated with Src in Western blots. In addition, PPP1r18 was ubiquitously expressed in the cytosol of Src-deficient TRAP(+) MNCs and was found to be localized to the podosomes in the actin ring formed by overexpressing constitutively active Src. Moreover, PPP1r18 overexpression inhibited actin ring formation and bone resorption, whereas PPP1r18 knockdown promoted these processes. In contrast, cooverexpression of Src and PPP1r18 did not promote phosphorylation of PPP1r18. These results suggest that PPP1r18 interacts with Src, is recruited to the cell periphery, and then becomes involved in actin ring formation in osteoclasts.

PPP1r18 inhibited not only actin ring formation but also maturation (fusion and spreading) of TRAP(+) MNCs. Examination of osteoclast differentiation markers by quantitative PCR revealed that PPP1r18 did not regulate Acp5, Nfatc1, and Ctsk expression but negatively regulated Dcstamp expression. The decrease of Dcstamp expression may be one of the reasons for the inhibition of TRAP(+) MNC maturation, because Dc-Stamp is essential for the fusion of osteoclast precursors (40). In addition, we believe that PPP1r18 may suppress the migration of osteoclast precursor cells because it has been reported to inhibit migration of breast cancer cells (32). These findings suggest that PPP1r18 plays two different roles in osteoclast precursors and osteoclasts. PPP1r18 is localized not only to the actin ring but also in the nucleus. Thus, PPP1r18 may regulate the transcription of Dcstamp and other genes. In addition, PPP1r18 may regulate cell migration and actin ring formation through the regulation of actin organization.

The recruitment of the catalytic unit of PP1 (PPP1CA) by PPP1r18 may be important for the regulation of actin ring formation and bone resorption because overexpression of the IGFG mutant PPP1r18 failed to suppress actin ring formation and osteoclastic bone resorption. Similar to PPP1r18 localization, PPP1CA was localized in the actin ring and the nuclei of TRAP(+) MNCs. After PPP1r18 overexpression, the localization of PPP1CA was disturbed and became more uniform across the whole cell, again showing similarity to the characteristic PPP1r18 localization under the same conditions. However, PPP1CA stayed largely in the actin ring in cells overexpressing the IGFG mutant PPP1r18 because of endogenous PPP1r18 bound to PPP1CA in the actin ring. The IGFG mutant PPP1r18 also remained in the actin ring. These results demonstrated that overly large amounts of PPP1r18 and PPP1CA disturb actin organization of osteoclasts. At the same time, the presence of an adequate PPP1CA amount maintained the actin ring in cells overexpressing the IGFG mutant PPP1r18. PP1 binds to nearly 200 validated interactors and works as a catalytic subunit (41, 42). In the present study, we revealed that the PPP1r18/PP1 complex may be important in osteoclast maturation and actin ring formation.

Phosphorylation of Src Ser17 regulates Src interactions and activation of its kinase activity (43). Thus, we examined whether the phosphorylation of Src Ser17 and Src kinase activity are regulated by the PPP1r18/PP1 complex. We found that Src Ser17 was dephosphorylated and that the level of phosphorylation of cortactin Tyr421, a well-known Src substrate and one of the actin ring-regulating proteins, was downregulated. However, the IGFG mutant PPP1r18 did not affect the phosphorylation status of Src Ser17 and cortactin Tyr421. Moreover, PPP1r18 overexpression interrupted binding of Src to cortactin and p130Cas. These results suggest that the PPP1r18/PP1 complex regulates Src activation through dephosphorylation of Ser17. Src Ser17 is also important for activation of small G protein Rap1, which regulates osteoclastic bone resorption through regulating actin organization (43, 44). Thus, Src Ser17 may play an important role in actin ring formation by upregulating cortactin, p130Cas, and Rap1 and negatively affecting the PPP1r18/PP1 complex. However, the biological significance of Src Ser17 phosphorylation is still unclear; therefore, additional experiments will be necessary to confirm it.

Phosphorylation and dephosphorylation of proteins regulate various cellular functions. Phosphorylation of various downstream proteins by Src kinase plays a key role in podosome organization and actin ring formation. Dephosphorylation of Src may also be important for podosome turnover and membrane ruffling during bone resorption. In our experiments, PPP1r18 localization in podosomes was regulated by Src, and it is possible that PPP1r18 inhibits Src attachment to podosomes or promotes its separation from the podosome core at the cell membrane. This suggests that the PPP1r18/PP1 complex is recruited to the cell membrane by Src to regulate Src signaling and subsequent podosome dynamics. The complex of PPP1r18 and PP1 is involved in the phosphorylation and dephosphorylation dynamics of Src downstream molecules. Csk-binding protein (Cbp/PAG) and C-terminal Src kinase (Csk) have been reported as Src substrates and negative regulators of Src (45–47). We therefore found a new mechanism of podosome regulation: the PPP1r18/PP1 complex and Src interact with each other and thereby regulate actin organization.

The localization of PPP1r18 in the nucleus is also important. Significantly inhibited Dcstamp mRNA expression may be one of the reasons why the fusion and maturation of osteoclasts were suppressed by PPP1r18 overexpression. Moreover, knockdown of PPP1r18 expression upregulated mRNA levels of Nfatc1 and its downstream osteoclast differentiation markers. Subsequently, TRAP(+) MNCs became enlarged and formed the actin ring with accentuated maturation. These results suggest that PPP1r18 also plays an important role in the maturation of osteoclasts by regulating the transcription of NFATc1 and its downstream molecules.

In conclusion, we found that PPP1r18 binds to PP1 and recruits it to the actin ring of osteoclasts, which may lead to dephosphorylation of target proteins and inhibition of actin ring formation. Thus, PPP1r18 may be one of the negative regulators of Src signaling that coordinates osteoclastic bone resorption. Moreover, it has been shown that PPP1r18 directly binds to actin and suppresses actin elongation (31). Thus, PPP1r18 probably directly regulates podosome formation and dynamics, in addition to its function as a PP1 regulator, as shown in this study. Moreover, PPP1r18 also regulates the transcription of NFATc1 and its downstream proteins and maturation of osteoclasts. To the best of our knowledge, this is the first report that shows the role of PPP1r18 in osteoclasts.

MATERIALS AND METHODS

Cells and animals.

SYF cells were purchased from the American Type Culture Collection (ATCC) (Manassas, VA) and cultured in Dulbecco's modified Eagle's medium (DMEM) (purchased from Wako, Osaka, Japan) containing 10% fetal bovine serum (FBS). HEK293 cells were purchased from the RIKEN Cell Bank (Ibaraki, Japan) and cultured in DMEM containing 10% FBS. Src^+/− mice (7) were purchased from the Jackson Laboratory (Bar Harbor, ME). Src^+/− mice were intercrossed to obtain wild-type and Src^−/− mice. Src^−/− mice were fed with milled food after weaning because of the loss of tooth eruption. All animal experimental protocols used in the present study were approved by the Animal Experimentation Ethics Committees of Tohoku University (authorization number 2012shidou-125) and Kyushu Dental University (authorization number 27-A002).

Reagents.

sRANKL was kindly provided by the Oriental Yeast Company, Ltd. (Shiga, Japan). Recombinant human M-CSF was purchased from Peprotech, Inc. (NJ). Antiphostensin antibody (PPP1r18, sc-99169) and antivinculin antibody (sc-73614) were purchased from Santa Cruz Biotechnology, Inc. (TX). Antibodies against v-Src (Ab-1), cortactin (4F11), and p130Cas (610272) were purchased from Merck Millipore (Billerica, MA). Anti-Myc tag antibody (017-21871) for Western blotting was purchased from Wako. Anti-Myc-tag antibody pAB (562) for immunofluorescence was purchased from MBL International Corporation (Woburn, MA). Antibodies against phosphorylated tyrosine (pTyr-100) and phosphorylated cortactin (Tyr421) were purchased from Cell Signaling Technology, Inc. (Danvers, MA). Anti-β-actin (A5441) antibody, horseradish peroxidase (HRP)-conjugated anti-mouse IgG secondary antibodies, and HRP-conjugated anti-rabbit IgG secondary antibodies were purchased from Sigma-Aldrich (MI). Rhodamine, phalloidin, Alexa Fluor 488-conjugated anti-rabbit IgG secondary antibodies, Alexa Fluor 555-conjugated anti-mouse IgG secondary antibodies, and 4′,6-diamidino-2-phenylindole (DAPI) were purchased from Thermo Fisher Scientific, Inc. (Waltham, MA).

Plasmids and constructs.

Src plasmid was purchased from Addgene (Cambridge, MA). PPP1r18 plasmid was purchased from GE Healthcare and cloned into the pCMV-Myc-N vector from TaKaRa (Shiga, Japan) by PCR, using Q5 high-fidelity DNA polymerase from New England BioLabs (Ipswich, MA). Ile540 and Phe542 of PPP1r18 were mutated to Gly (PPP1r18 IGFG mutant) by using a QuikChange site-directed mutagenesis kit from Agilent Technologies (Santa Clara, CA), the 5′-CACAAACAGCTCAAGGGCTCCGGTAGTGAGACAGCCC-3′ primer, and a complementary primer. Mouse shRNA against PPP1r18 and control shRNA were purchased from Santa Cruz Biotechnology, Inc. The adenovirus vector (pAxEFwtit2) was purchased from TaKaRa.

Transfection of SYF cells with constitutively active Src (Src Y527F).

Wild-type Src was cloned into the pcDNA3.1-hygro vector from Thermo Fisher Scientific Inc. Tyr527 of Src was mutated to Phe by using a QuikChange site-directed mutagenesis kit, the 5′-GAAGACTACTTTACGTCCACTGAGCCACAGTTCCAGCCCGGGGAGAACCTATAG-3′ primer (underlining indicates that the codon encodes a muted amino acid), and a complementary primer. Src Y527F-pcDNA-hygro and control vectors were digested with MfeI (New England BioLabs) and used for transfecting SYF cells, which were then selected with 500 μg/ml hygromycin B (Roche, Basel, Switzerland).

Identification of Src-binding proteins.

Src Y527F-expressing or control SYF cells were cultured on 250-mm square cell culture dishes from Corning (NY). When cells were 90% confluent, they were collected into 8 ml of phosphate-buffered saline (PBS). After centrifugation and washing with ice-cold PBS, the cells were solubilized in lysis buffer [20 mM HEPES (pH 7.4), 150 mM NaCl, 1 mM EGTA, 1.5 mM MgCl₂, 10% glycerol, 1% Triton X-100, 10 μg/ml aprotinin, 10 μg/ml leupeptin, 1 mM 4-(2-aminoethyl)-benzenesulfonylfluoride hydrochloride (AEBSF), 0.2 mM sodium orthovanadate]. The lysates were centrifuged at 16,000 × g for 20 min at 4°C. Thirty microliters of anti-Flag M2 affinity gel (Sigma-Aldrich) was added to the supernatant and incubated for 3 h. The lysates were centrifuged at 16,000 × g for 20 min at 4°C and washed with PBS twice. Src and its binding proteins were eluted four times with 5 mg/ml Flag peptide (Sigma-Aldrich) dissolved in PBS containing 10 μg/ml aprotinin, 10 μg/ml leupeptin, and 1 mM AEBSF for 15 min. The proteins were analyzed by LC-MS/MS as described previously (48, 49).

Adenovirus generation.

Recombinant adenoviruses carrying wild-type PPP1r18 or IGFG mutant PPP1r18 were constructed by the recombination between the expression cosmid cassettes. The cosmids were digested with PacI restriction enzyme (New England BioLabs) and used to transfect HEK293 cells. Virus titers were determined by the modified point assay (50).

In vitro osteoclast differentiation.

Spleen cells were isolated from 4- to 6-week-old wild-type mice or Src^−/− male mice and incubated with M-CSF (100 ng/ml; 1,000,000 cells/cm²) for 3 days to differentiate spleen macrophages. Subsequently, spleen macrophages were cultured with sRANKL (100 ng/ml) and M-CSF (30 ng/ml) on culture plates from Corning or cover glasses from Thermo Fisher Scientific (PA) for 7 days with a medium change every 2 days.

Differentiated cells were evaluated by TRAP staining. TRAP(+) MNCs were regarded as osteoclasts. Cells were fixed with 3.4% formaldehyde for 10 min at 25°C. Fixed cells were washed three times with PBS and permeabilized with an ethanol-acetone (1:1) mixture for 1 min. After washing with double-distilled water, the cells were stained for 30 min with a TRAP solution containing naphthol-MX phosphate, Fast Red violet LB salt, and N,N-dimethylformamide in 50 mM acetate buffer. The size of TRAP(+) MNCs was measured by using ImageJ (https://imagej.nih.gov/ij/).

Pit formation assay.

Primary osteoblasts were isolated from the calvariae of 3-day-old C57BL/6 mice with 0.2% dispase (Wako) and 0.1% collagenase (Wako). Primary osteoblasts (1 × 10⁶ cells) and bone marrow cells (1 × 10⁷ cells) from 4-week-old mice were cultured on collagen gels in 10-cm dishes (Corning). Cocultured cells were cultured in alpha minimum essential medium (α-MEM) containing 10% FBS, 1 × 10⁻⁸ M 1α,25-dihydroxy-vitamin D₃ (Wako), and 1 × 10⁻⁶ M prostaglandin E2 (PGE₂) (Sigma-Aldrich) (23). After culturing for 9 days, the cells were treated with 0.02% collagenase, washed with PBS, resuspended in α-MEM to a final concentration of 3 × 10³ cells/ml, and seeded on dentin slices (diameter, 4 mm; Kyushu-do, Kitakyushu, Japan). After 48 h of culture, the cells were scraped off and the slices were stained with Mayer's hematoxylin (Wako) (23). The area of each pit was measured from the photographs with ImageJ.

Immunofluorescence.

TRAP(+) MNCs were differentiated on cover glasses as described above. Cells were fixed with 3.4% formaldehyde for 10 min at room temperature. Fixed cells were washed three times with PBS, permeabilized with 0.2% Triton X-100 dissolved in 1% bovine serum albumin (BSA)-PBS for 20 min, and incubated with 1% BSA-PBS for 1 h. Subsequently, the cells were incubated with the primary antibodies for 2 h and then with Alexa Fluor-conjugated secondary antibodies, Alexa Fluor- or rhodamine-conjugated phalloidin, and DAPI. The subcellular localization of the indicated proteins was determined by fluorescence microscopy (DeltaVision Elite [GE Healthcare Japan, Tokyo, Japan] or BZ-9000 [Keyence Japan, Osaka, Japan]). The nuclei were counted by using ImageJ counter.

Real-time qPCR.

Total RNA of TRAP(+) MNCs or spleen macrophages was isolated with TRIzol reagent (Thermo Fisher Scientific Inc.). cDNA was synthesized from 1 μg of total RNA by using SuperScript II transcriptase and random primers (Thermo Fisher Scientific Inc.). Real-time quantitative PCR (qPCR) was performed by incubating cDNA, Thunderbird SYBR green PCR master mix purchased from Toyobo (Osaka, Japan), and the primers indicated below in a 7300 real-time PCR system purchased from Applied Biosystems (Foster City, CA) as described previously (23). All reactions were performed in triplicate and analyzed.

The primer sequences were as follows: Acp5 (TRAP), 5′-TCCTGGCTCAAAAAGCAGTT-3′ (forward) and 5′-ACATAGCCCACACCGTTCTC-3′ (reverse); Ctsk (cathepsin K), 5′-GGGAAGCAAGCACTGGATAA-3′ (forward) and 5′-CCGAGCCAAGAGAGCATATC-3′ (reverse); Dcstamp, 5′-TCCTCCATGAACAAACAGTTCCAA-3′ (forward) and 5′-AGACGTGGTTTAGGAATGCAGCTC-3′ (reverse); Gapdh, 5′-AACTTTGGCATTGTGGAAGG-3′ (forward) and 5′-ACACATTGGGGTAGGAACA-3′ (reverse); and Nfatc1, 5′-GGTGCTGTCTGGCCATAACT-3′ (forward) and 5′-GCGGAAAGGTGGTATCTCAA-3′ (reverse).

Western blotting.

The cells were washed twice with ice-cold PBS and solubilized in lysis buffer. The lysates were centrifuged at 16,000 × g for 20 min at 4°C. The supernatants were boiled in sample buffer containing 0.125 M Tris-HCl (pH 6.8), 40% glycerol, 4% sodium dodecyl sulfate (SDS), 0.2 M dithiothreitol, and 0.01% bromophenol blue, subjected to SDS-polyacrylamide gel electrophoresis, and transferred to a nitrocellulose membrane. The membrane was immunoblotted with the corresponding primary and horseradish peroxidase (HRP)-coupled anti-mouse or anti-rabbit secondary antibodies and developed with an Immobilon Western chemiluminescent HRP substrate purchased from Merck Millipore.

Data analysis and statistics.

All experiments except mass spectrum measurements were performed at least twice. The statistical significance of differences between groups was analyzed by one-way analysis of variance (ANOVA) followed by a post hoc test. Differences were considered to be statistically significant if the P value was <0.01. Data are expressed as the mean values ± standard deviation of the mean (mean ± SD), where n is the number of culture wells.