Leupaxin (LPXN), which belongs to the paxillin extended family of adaptor proteins, was previously identified as a component of the sealing zone in osteoclasts. LPXN was found to associate with several podosomal proteins, such as the protein tyrosine kinase Pyk2, the protein-tyrosine phosphatase-PEST (PTP-PEST), actin-binding proteins, and regulators of actin cytoskeletal reorganization. It was previously demonstrated that inhibition of LPXN expression resulted in reduced osteoclast-mediated resorption. In the current study, overexpression of LPXN in murine osteoclasts resulted in both enhanced resorptive activity and cell adhesion, as assessed by in vitro resorption assays. The overexpression of LPXN resulted in an increased association of Pyk2 with LPXN. In an attempt to determine an additional biochemical basis for the observed phenomenon in increased osteoclast activity, a coimmunoprecipitation screen for additional binding partners revealed that Src, a protein tyrosine kinase that is critical to both podosome formation and osteoclast function, was also associated with LPXN. After exposure to the pro-inflammatory and osteoclastogenic cytokine TNF-α, there was an increase in the level of Src that coimmunoprecipitated with LPXN. Our data indicate that association of the scaffold protein LPXN with Src adds further complexity to the organization of the podosomal signaling complex in osteoclasts.
- sealing zone
- LD2 domain
osteoclasts attach to the extracellular matrix through integrin-dependent dynamic adhesion structures that are structurally related to focal adhesions (3, 14, 46). The sealing zone, which is composed of podosomes, isolates the resorption apparatus of the osteoclast. The actin-rich podosomes contain various structural proteins, nonreceptor cytoplasmic protein tyrosine kinases, and regulators of the actin cytoskeleton (9, 23, 45).
Leupaxin (LPXN), which was originally characterized as a member of the paxillin-extended family of multifunctional docking (i.e., adaptor) phosphoproteins in cells of hematopoietic origin (28), was further identified as an additional component of the sealing zone in osteoclasts (19). In both hematopoietic cells and the osteoclast, LPXN was found to associate with the nonreceptor protein tyrosine kinase Pyk2, the cytoplasmic protein tyrosine phosphatase-PEST (PTP-PEST), and cytoskeletal regulatory proteins, such as the actin-binding protein actopaxin and the Arf-GAP protein p95PKL (5, 19, 49). In the osteoclast podosomal complex, Pyk2 has been suggested to functionally replace the focal adhesion kinase (FAK) that is expressed at low levels (43, 51, 52). The association of both Pyk2 and PTP-PEST with LPXN may regulate podosome turnover in osteoclasts, since PTP-PEST can control the activities of p130Cas, paxillin, and Pyk2 (13, 33). This mechanism can presumably ensure a continuous turnover of phosphorylated and dephosphorylated forms of LPXN between the cytoplasmic pool and that in podosomal complexes. Therefore, it is reasonable to hypothesize that, similar to paxillin, LPXN may serve as an additional regulator of osteoclast adhesion-dependent cytoskeletal organization and signal transduction (19).
LPXN, which shares ∼37% structural homology with paxillin, also contains several protein-binding modules (8, 47). These multiple protein-protein interaction motifs include repeated NH2-terminal leucine (L)- and aspartate (D)-rich sequences known as LD domains and COOH-terminal LIM (for Lin-11 Isl-1 Mec-3) domains (7). In general, the LD motifs function in recruitment of multiple signaling partners, whereas the LIM domains are involved in localization of these adaptor proteins to focal adhesions (8, 19, 42). With respect to the best-characterized member, paxillin, LD motif-binding partners include the structural protein vinculin, integrin-linked kinase, regulators of the actin cytoskeleton, and the protein tyrosine kinases FAK/Pyk2 and Src (8, 35). The redistribution of Pyk2, which is itself a substrate for Src (16), from a cytoplasmic pool to focal adhesions, is presumably dependent upon its association with paxillin (29). An additional consequence of the association with protein tyrosine kinases may be increased phosphorylation of these molecular adaptor proteins, resulting in further recruitment of other signaling molecules (8). Although the core consensus sequence of individual LD motifs among the paxillin-related scaffold proteins is largely conserved, unique flanking sequences can provide additional specificity of function in response to different external stimuli, resulting in recruitment of a diverse repertoire of actin-binding and signaling proteins in various cell types (47). In the osteoclast, the extent of redundancy and/or specificity of functions that paxillin and LPXN separately execute within the podosomal signaling complex remains to be determined.
We had previously reported that inhibition of LPXN expression (in the murine osteoclast) greatly reduced osteoclast-mediated resorption without any apparent loss of cell adhesion (19). In the current study, we have examined the consequences of LPXN overexpression on osteoclast resorption and cell adhesion. Furthermore, we have determined Src to be an additional binding partner for LPXN. Finally, we provide evidence that exposure of osteoclasts to the pro-inflammatory and osteoclastogenic cytokine TNF-α (6) results in an increased association of Src with LPXN. These results provide further evidence that LPXN is part of the dynamic signaling complex resident in the osteoclast sealing zone.
MATERIALS AND METHODS
The chicken anti-mouse and anti-rabbit LPXN antibodies (IgY) were generated with the help of Aves Laboratories (Tigard, OR). Antibodies to Pyk2 and Src were purchased from Santa Cruz Biotechnology (Santa Cruz, CA). The fluorescent secondary antibodies Alexa 488 and Alexa 568 were purchased from Invitrogen (Carlsbad, CA). Rhodamine phalloidin was purchased from Sigma (St. Louis, MO). The AdEasy-XL adenoviral vector system was purchased from Stratagene (La Jolla, CA). The Chariot transfection reagent was purchased from Active Motif (Carlsbad, CA). Cell culture medium components were purchased from Invitrogen and Atlanta Biologicals, (Lawrenceville, GA). The mouse macrophage cDNA library (catalog number 936304; λ ZAP II vector) was purchased from Stratagene. Cyclo-RGD peptides [catalog number PCI-3661-P1; cyclo (Arg-Gly-Asp-d-Phe-Lys, cRGDfK)] were purchased from Peptides International (Louisville, KY). Osteologic discs were purchased from BD Biosciences (Rockville, MD). The SuperSignal West Dura Extended Duration Substrate chemiluminescence reagent was purchased from Pierce Biotechnology (Rockford, IL). All other chemicals and reagents were purchased from Sigma.
Isolation of osteoclast-like cells from bone marrow.
The tibiae and femurs of 7-wk old mice were used to isolate bone marrow stromal cells, as described previously (19, 24). Bone marrow cells were suspended in α-minimal essential medium (α-MEM, Invitrogen) supplemented with 10% fetal bovine serum (designated α-10 MEM), and cultured at 37°C in a 5% CO2 incubator. After 24 h, nonadherent cells were layered on Histopaque-1077 (Sigma), and centrifuged at 300 g for 15 min at room temperature. The cell layer between the Histopaque and the medium was removed and washed with α-10 MEM and centrifuged at 2,000 rpm for 7 min. Both recombinant macrophage colony-stimulating factor-1 (M-CSF) (R&D Systems, Minneapolis, MN) and the receptor-activated NF-κB ligand (RANKL) were added to the cultures at concentrations of 10 ng/ml and 100 ng/ml, respectively, as described previously (19). Multinucleated osteoclasts were seen to form and mature between days 4 and 5. The percentage of tartrate-resistant acid phosphatase (TRACP)-positive cells was assessed to be ∼99%, using the TRACP+ assay, as described previously (19). All procedures were approved by the Institutional Animal Care and Use Committee at the University of Maryland (Baltimore, MD).
PCR-based cloning of the murine ortholog of leupaxin (mLPXN).
The open reading frame (ORF) for mLPXN was amplified from a murine macrophage cDNA library (Stratagene) using the following PCR conditions: an initial heat denaturation at 94°C for 3 min followed by 35 cycles of 94°C for 1 min; annealing at 65°C for 1 min; extension at 72°C for 1 min; and a final extension cycle at 72°C for 10 min. A 1,161-bp fragment representative of the ORF for mLPXN was amplified; the forward primer (F) included an NheI site (underlined), 5′-GCTAGCATGGAAGAGCTGGATGCCT-3′, and the reverse primer (R) included a XhoI site (underlined), 5′-CTCGAGGGCCCGTTTGTACTCCATTA-3′. To confirm overexpression of the rabbit ortholog of LPXN (rLPXN) after adenoviral infection of murine osteoclasts, a 219-bp fragment of rLPXN was amplified with the following primers: F, 5′-CCGTCCAGGATAGCACAAAT-3′ and R, 5′-TCTCCGATACGGGTTTCTTG-3′. The PCR products were analyzed on ethidium bromide-stained 2% agarose gels and photographed under UV light. The sequence was confirmed by automated sequencing (Biopolymer Laboratory, University of Maryland, Baltimore, MD). The derived nucleotide sequence of the ORF for mLPXN confirmed 100% identity with mLPXN (GenBank accession number AB053936).
Construction of recombinant adenoviral vectors for murine and rabbit orthologs of LPXN.
Separate adenoviruses for LPXN orthologs (murine and rabbit) were generated, according to methods described previously (2). Previously, we had cloned LPXN from a rabbit osteoclast cDNA library (19). In brief, the murine and rabbit orthologs of the LPXN gene were ligated into pShuttle-IRES-hrGFP2 (AdEasy-XL adenoviral vector system, Stratagene) using the NheI-XhoI multiple cloning sites. Recombinant adeno-LPXN-GFP2 was confirmed by digestion with PacI. High titer (2 × 1010 plaque forming units/ml) viral stocks were made as described here, and according to methods previously published (2, 24, 44). Approximately 15 μg of shuttle plasmid and 5 μg of pVQAd 9.2–100 backbone plasmid were digested with PacI. Both linear fragments were cotransfected into the HEK-293 cell line on a 60-mm plate at ∼50% confluency by a standard CaCl2 method (2). The initial viral lysate was amplified in HEK-293 cells, and the final viral lysate was purified over two rounds of CsCl gradient ultracentrifugation. Virus particles were dialyzed against 3% sucrose/PBS buffer, diluted to 1 × 1012 particles/ml, and then frozen at −80°C. We assayed for infectious particle concentration by a cell-based plaque assay on HEK-293 cells. Presence of replication-competent adenovirus was checked by plaquing on the wild-type permissive cell line A549 for at least 14 days. As negative controls, osteoclasts were separately infected with an adenovirus encoding enhanced green fluorescent protein (rAd-EGFP), as described previously (30).
PCR-based amplification of LD2 domain in rLPXN.
The cDNA of rLPXN (GenBank accession number AF118146), which was cloned earlier from a rabbit osteoclast cDNA library, served as the template for PCR amplification of the LD2 domain. In brief, the primers used to amplify the region encoding the LD2 (underlined) amino acid sequence (SRK-ESNLDETSKMLSV-QDSTNPF) were as follows: F, 5′-CCGGAATTCTTCCAGAAAGGAGAGTAACCTGG-3′ and R, 5′-CCGCTCGAGAAAGGGATTTGTGCTATCCTG-3′. The following PCR conditions were used: denaturation at 94°C for 30 s; annealing at 58°C for 30 s; extension at 72°C for 30 s, for 10 cycles; and a final extension at 72°C for 10 min. The product size was ∼90 bp.
Construction of a rLPXN glutathione S-transferase (GST)-LD2 fusion protein.
After PCR-based amplification of the LD2 domain of rLPXN, it was directionally cloned into the pGEX-5X-1 vector at the EcoRI and XhoI multiple cloning sites. Positive clones were identified and confirmed by restriction digestion; Escherichia coli BL-21 cells were then subsequently transformed with these positive clones. Overnight cultures (∼150 ml) of recombinant GST-LD2 fusion protein in BL-21 cells were transferred into larger volumes of LB broth + ampicillin for 2 h in the presence of isopropyl-β-d-thiogalactopyranoside (0.1 mM) for an additional 3 to 4 h at 37°C. The cultures were centrifuged at 5,000 rpm for 10 min, and the pellet (∼2 ml) was washed with 25 ml of GST buffer 1 (20 mM Tris, pH 8.0, 150 mM NaCl and 1 mM EDTA). The pellet was resuspended in GST buffer 1 with 100 μg/ml of lysozyme, in the presence of protease inhibitors (Complete Inhibitor, Roche, Basel, Switzerland), and kept at room temperature for 15 min. Triton X-100 was added to a final concentration of 1%, and the pellet was incubated on ice for 10 min. The lysed pellet was sonicated at 30-s pulse intervals and then centrifuged at 10,000 rpm at 4°C for 15 min. The supernatant was loaded onto a pre-equilibrated glutathione-Sepharose column (Pierce Biotechnology). The column was first washed with GST buffer 1 and further washed with GST buffer 2 (20 mM Tris, pH 8.0, 1 mM EDTA) until the OD280 = 0.00. This column was incubated/rocked overnight at 4°C. GST buffer 2 (5 ml), containing 5 mM glutathione, was added to elute the GST fusion protein, and the fractions (1 ml) were examined for absorption at OD280. The eluted protein was subsequently separated by SDS-PAGE and confirmed for size (∼28 kDa, GST plus LD2) and identity using Western blot analyses with a monoclonal antibody against GST (Upstate Biotechnology, Charlottesville, VA).
Both endogenous and overexpressed mLPXN in murine osteoclasts were immunoprecipitated as follows. Murine osteoclasts were lysed in a modified RIPA lysis buffer for 1 h at 4°C [10 mM Tris·HCl, pH 7.5, 50 mM NaCl, 0.5% Triton X-100, 30 mM sodium pyrophosphate, 5 mM NaF, 0.1 mM Na3VO4, 5 mM ZnCl2, and 2 mM phenylmethylsulfonyl fluoride (PMSF)] and further supplemented with a protease-inhibitor cocktail (Complete Protease Inhibitor Cocktail, Roche Bioscience, Palo Alto, CA). The lysates were centrifuged at 14,000 rpm for 10 min, and protein measurements were performed on the supernatants. Approximately 500 μg of lysates were precleared with chicken preimmune serum (50 μl of IgY) for 2 h at 4°C, to which 50 μl of a 50% suspension of PrecipHen (agarose-conjugated goat anti-chicken IgG, Aves Labs) was added. The lysates were precleared overnight at 4°C. The precleared lysates were centrifuged at 8,000 rpm at 4°C, and the supernatants were transferred to fresh microfuge tubes. Approximately 4 μg of affinity-purified chicken IgY against mLPXN was added for specific immunoprecipitation. After incubation at 4°C overnight, 50 μl of a 50% suspension of PrecipHen was added to the individual tubes for 2 h at 4°C. The immunoprecipitates were centrifuged at 8,000 rpm for 10 min at 4°C. The supernatants were transferred into fresh tubes and incubated with preimmune chicken IgY for 2 h at 4°C, following which ∼50 μl of a 50% suspension of PrecipHen beads was added. The beads were washed at least four times with lysis buffer, and once with PBS (10 min each, at 4°C). After the last wash, the beads were centrifuged at 8,000 rpm for 10 min, the supernatant was discarded, and the immunoprecipitates were eluted by boiling in ∼100 μl of 2× Laemmli sample buffer with 2% β-mercaptoethanol. The immunoprecipitates (both specific and nonspecific) were separated on 8% SDS-PAGE gels and subsequently transferred to nitrocellulose membranes (Pierce), using a semi-dry transfer apparatus (Bio-Rad, Hercules, CA). The nitrocellulose membranes were blocked in 5% nonfat dry milk for at least 2 h at room temperature (or overnight at 4°C) before Western blot analysis with antibodies for Src and Pyk2.
Affinity-precipitation/binding assays with GST-LD2 fusion protein.
Affinity-precipitation/binding assays were performed with the GST-LD2 proteins to determine interacting proteins as follows, and as described previously (19, 25). First, glutathione-Sepharose beads were washed five times for 10 min each using 10 volumes of binding buffer (50 mM Tris·HCl, pH 8.0, 120 mM NaCl, 0.5% Nonidet P-40, 1 mM DTT, 1 mM PMSF, plus a protease inhibitor cocktail). Second, ∼50 μg of GST-LD2 fusion protein was incubated with a 25-μl aliquot of a 50% suspension of Sepharose beads. Third, Sepharose-bound GST-LD2 fusion proteins or Sepharose-bound GST protein alone (as a control) were separately incubated with ∼2 mg of murine osteoclast lysates for 4 h at 4°C. The beads were centrifuged at 500 × g for 2 min and washed with binding buffer four times at 4°C. Finally, 1 to 2 volumes of 2× SDS-PAGE sample buffer were added to the washed beads, the beads were boiled, and proteins were separated by SDS-PAGE, before transfer to PVDF membranes for subsequent Western blot analyses with the appropriate antibodies.
Transfection of GST-LD2 fusion protein into murine osteoclasts.
The GST-LD2 protein was transfected into murine osteoclasts according to the following protocol. Murine osteoclasts cultured in 100-mm2 plates were used on day 5 when multinucleated osteoclasts were evident. Approximately 1.5 μg of the fusion protein was first diluted in 200 μl of PBS; in parallel, ∼6 μl of Chariot reagent was diluted in 200 μl of dH2O. These two reagents were mixed together and incubated at room temperature for 30 min. The culture medium was aspirated (from a 100-mm2 culture dish) of murine osteoclasts, and the Chariot-GST-LD2 protein mixture was added to the cells with 4.6 ml of serum-free (α-0) medium, and incubated for 2 h at 37°C in a 5% CO2 incubator. After transfection, the osteoclasts were replated for 24 h for indirect immunofluorescence as described previously (19).
Osteoclasts were rinsed twice in ice-cold PBS, fixed with 4% paraformaldehyde, and permeabilized with PBS containing 0.2% Triton X-100. The cells were blocked with PBS containing 3% bovine serum albumin overnight. The cells were incubated for 2 h in a 1:500 dilution of the appropriate antibody (LPXN or Src) at room temperature in the blocking solution described above. The chicken IgY antibodies were affinity purified with peptides with amino acid sequences specific for either mLPXN (CH LDQQSTEESKIPQTPKT, IgY 4247) or rLPXN (CZEELERSTLQDSDEYSNS, IgY 4554). The primary antibodies were detected with either Alexa-488- or Alexa 562-conjugated secondary antibodies (1:1,000 dilution, Molecular Probes). Actin was labeled with rhodamine phalloidin. Confocal microscopy (Zeiss LSM Meta 510) was performed, with all images shown being taken with either a ×63 or ×100 oil-immersion objective lens.
In vitro resorption assays.
Mature murine osteoclasts were plated on calcium phosphate (CaPO4, 1 μm)-coated quartz discs (Osteologic discs), according to previously published methods (19). Briefly, osteoclasts were treated with Cellstripper (Mediatech, Herndon, VA), a nonenzymatic cell dissociation solution, for 10 min at 37°C; the osteoclasts were gently triturated and resuspended in α-10 MEM, and centrifuged for 7 min at 1,200 rpm, similar to methods previously published (10, 11, 39). Approximately 48 h later, the effects on osteoclast-mediated resorption were assessed by rinsing the discs in distilled water to lyse the osteoclasts, as described previously (19, 24). The extent of resorption could be assessed by clear areas on the quartz discs where the CaPO4 film had disappeared. The discs were photographed (×20 objective), and the area and number (of pits) were counted using the trace and fill measurement module in SigmaScan Pro 5 (SPSS, Chicago, IL). The resorption areas in each image were computed as “percentage of the total area of the image” as determined by the total number of pixels of each image (1.92 million). Statistical analyses were performed using ANOVA, and significance of differences between means was assessed at P < 0.05.
Culture plates (12-well) were coated with 10 μg/ml cyclo-RGD peptide in sterile PBS overnight at 4°C. The plates were washed briefly and then blocked in 10 mg/ml BSA in PBS for 2 h at 37°C. Osteoclasts, which had been infected with either rAd-EGFP (control) or rAd-mLPXN, were replated onto the RGD-coated surfaces in serum-free medium for 2 h at 37°C. Briefly, osteoclasts were treated with Cellstripper, for 10 min at 37°C, the osteoclasts were gently triturated and resuspended in α-10 MEM, and centrifuged for 7 min at 1,200 rpm, similar to methods previously published (39). After adhesion for 2 h, the plates were washed three times with PBS, and adherent cells were fixed in 3% paraformaldehyde. After staining, the number of adherent cells was counted under the microscope. Statistical analyses were performed using ANOVA and significance of differences between means was assessed at P < 0.05.
Statistical analyses were performed as follows: the raw data points for either the resorption or cell adhesion assays were used to obtain the mean for the control group of osteoclasts. Each raw data point within that group was divided by the mean and subsequently transformed (multiplying the quotient by 100) to get percentage values for each data point within that group. From these transformed data points, we determined the mean ± SE for the control group, giving values of 100 ± SE. To obtain similar data for the experimental treatment groups (i.e., mLPXN-overexpressing osteoclasts), each raw data point was divided by the original mean of control group; the quotients were similarly transformed into “percentage of control”. Therefore, similar to that for the control group, means ± SE were obtained for the experimental groups, as described previously (18).
Cloning and overexpression of mLPXN in osteoclasts.
We have previously demonstrated that decreased expression of LPXN was correlated with decreased osteoclast resorptive activity, without an apparent loss of adhesion (19). In the current study, we wanted to examine the consequence(s) of overexpression of LPXN on osteoclast function. The murine ortholog of LPXN was cloned (from a mouse macrophage cDNA library) by PCR-based amplification of the ORF; sequence analysis confirmed ∼100% identity with LPXN cloned earlier from the mouse spleen (GenBank accession number NM_134152), as shown in Fig. 1A; the size of the ORF for mLPXN was 1,161 bp. Next, an affinity-purified chicken IgY antibody (IgY 4247) directed against mLPXN was generated, which recognized a single band at ∼50 kDa in osteoclasts, as shown in Fig. 1B. This antibody was specific for mLPXN, because it did not cross-react with the human ortholog of LPXN present in human prostate cancer cells (PC-3 cell line, data not shown). Having generated an mLPXN-specific chicken antibody (IgY), we constructed a recombinant adenovirus (rAd) for mLPXN (titer ≈2 × 1010 pfu/ml, rAd-mLPXN) for overexpression in osteoclasts. Murine osteoclasts, which were formed on day 5 after initiation of bone marrow culture in the presence of M-CSF and RANKL, were infected with rAd-mLPXN [at a multiplicity of infection (MOI) of 100] for 72 h; the control infection was performed with an adenovirus encoding EGFP alone (rAd-EGFP), as described previously (24, 30, 44). Subsequently, the cells were separately processed for either Western blot analysis, immunofluorescence, or resorption assays. Western blot analyses demonstrated that the mLPXN IgY antibody was able to detect the adenovirus-mediated overexpression of mLPXN (72 h postinfection) in murine osteoclasts, as shown in Fig. 1C, a, lane 2; normally, the levels of LPXN are extremely low in osteoclasts, compared with those in PC-3 cells, a prostate cancer cell line (unpublished observations, Gupta et al.). In addition, the mLPXN 4247 IgY was also successful in immunoprecipitating LPXN (Fig. 1C, b, lane 3). After adenoviral infection of murine osteoclasts, which were formed on day 5 in the presence of M-CSF and RANKL, there were no apparent changes in osteoclast phenotype or morphology, as shown in Fig. 1D, a, frames 1 and 2. The efficiency of adenoviral infection was determined by both EGFP-derived fluorescence and Western blot analysis of mLPXN protein expression in adenoviral-infected osteoclasts, as shown in Fig. 1D, b and c; as estimated from the EGFP-derived fluorescence, the infection efficiency was determined to be at least 60%, similar to that previously reported for a plasma membrane zinc transporter in murine osteoclasts (24). Western blot analysis showed that graded increases in rAd-mLPXN infection (MOI ranging from 0 to 200), resulted in overexpression of mLPXN protein in osteoclasts, reaching maximal levels at MOI ranging between 50 and 100. The protein levels of β-actin were used for normalization of the signal for mLPXN (Fig. 1D, c). In subsequent experiments that examined overexpression of mLPXN protein, its effects on osteoclast function, and subcellular distribution, an MOI of ∼100 was used. Having ascertained adenovirus-mediated overexpression of mLPXN in osteoclasts, we examined the subcellular distribution and colocalization of mLPXN with that of actin and Pyk2, respectively, as shown in Fig. 1E, frames a–d. In control osteoclasts, which were infected with rAd-EGFP alone, LPXN was diffusely distributed throughout the cytoplasm, whereas actin decorated the sealing zone (Fig. 1E, a). The subcellular distribution of LPXN and Pyk2 in control osteoclasts was also diffuse throughout the cytoplasm (Fig. 1E, frame b); it should be noted that although Pyk2 has been characterized as an adhesion kinase in the osteoclast (16), the majority of Pyk2 is diffusely distributed in the cytoplasmic pool in most cell types (4, 27). After overexpression of mLPXN, the colocalization of LPXN with actin appeared more pronounced at the sealing zone (Fig. 1E, frame c), whereas the distribution of LPXN with Pyk2 became more diffuse throughout the cytoplasm (Fig. 1E, frame d).
Effects of mLPXN overexpression in osteoclast-mediated resorption and adhesion.
The effects of adenovirus-mediated overexpression of mLPXN on osteoclast function were assessed using in vitro resorption and adhesion assays. For this purpose, osteoclasts were assayed for changes in resorptive activity (following 72 h postinfection with rAd-mLPXN) by replating them (10, 11, 39) on CaPO4-coated Osteologic discs, which, although are not ideal for perfect osteoclast polarization, still allow for rapid quantification of in vitro resorption, as described previously (12, 19, 24). Osteoclasts were allowed to resorb the CaPO4 matrix for 48 h; resorption areas were determined as clear areas (as indicated by arrows) where the CaPO4-coated surfaces had been cleared or resorbed, as shown in Fig. 2A, frames 1 and 2). The resorption areas were computed, as described in materials and methods, and according to previous reports (19, 24). The resorptive activity of rAd-mLPXN infected osteoclasts showed a significant increase (∼2.25-fold), as shown in Fig. 2B, possibly reflecting a “gain-of-function” following overexpression of LPXN. For changes in cell adhesion, osteoclasts, which were either infected with rAd-EGFP (controls) or rAd-mLPXN, were replated on cyclo-RGD-coated (10 μg/ml) 12-well plates in serum-free medium for 2 h at 37°C, according to previous methods for replating of osteoclasts (39). The adherent cells were fixed, stained, and counted under the microscope. The increase in resorption following overexpression of mLPXN was mirrored by a similar increase in cell adhesion, as shown in Fig. 2C. Previous data have demonstrated that Pyk2 is the major tyrosine kinase associated with LPXN in leukocytes, macrophages, and osteoclasts (19, 28). We demonstrated that when osteoclasts were treated with cyclo-RGD peptides, there was an increased tyrosine phosphorylation of LPXN-associated Pyk2 (i.e., Pyk2Y402), as shown in Fig. 2D, similar to previous reports on phosphorylation of Pyk2 in osteoclasts (16). Given these data, part of the increase in cell adhesion (on RGD-coated surfaces) in mLPXN-overexpressing osteoclasts can be explained by a possible “reinforcement” of the LPXN-Pyk2 complex.
Association of LPXN with Pyk2 in LPXN-overexpressing osteoclasts.
To determine a biochemical basis for the increased osteoclast resorptive activity and adhesion, we next examined the possibility of whether overexpression of mLPXN effected changes in LPXN association with Pyk2. Following adenovirus-mediated overexpression of mLPXN in osteoclasts (which was confirmed by immunoblotting the LPXN-immunoprecipitate for LPXN, as shown in Fig. 3A, bottom, lane 4), the levels of LPXN-associated Pyk2 increased ∼2-fold, as shown in Fig. 3A (top, n = 3 separate experiments). The increased association of Pyk2 with LPXN was apparent even when the previously cloned rabbit ortholog of LPXN (rLPXN) was overexpressed (Fig. 3A, bottom, lane 5). The selective overexpression of rLPXN in murine osteoclasts was confirmed by both RT-PCR (Fig. 3B, a, lane 2) and Western blot analyses. Western blot analysis was performed using a chicken polyclonal antibody directed against the rabbit ortholog of LPXN, (Fig. 3B, b, lane 2).
Association of LPXN with Src.
Having determined that overexpression of LPXN in osteoclasts resulted in increased association with Pyk2, we next screened for additional binding partners for LPXN, such as other tyrosine kinases, which have been shown to play an important role in osteoclast function (21, 34). Previous studies have clearly established the role of Src as a critical signaling molecule in osteoclast function (15, 31), and it has been well documented that paxillin can serve as a docking platform for Src, either directly or indirectly through FAK and Pyk2 (8). Pyk2 has previously been shown to coimmunoprecipitate with Src, and Src-dependent tyrosine phosphorylation of Pyk2 is an important mechanism in formation of the sealing zone upon osteoclast adhesion to the extracellular matrix (26). LPXN was immunoprecipitated from murine osteoclast lysates with the chicken polyclonal antibody directed specifically against mLPXN (as shown previously in Fig. 1B), and the LPXN immunoprecipitates were subsequently immunoblotted for Src, as shown in Fig. 3C. A positive signal for Src was evident in osteoclast total cell lysates (Fig. 3C, lane 1), and in lysates that were only precleared with PrecipHen (Fig. 3C, lane 2) prior to immunoprecipitation of mLPXN with the chicken polyclonal antibody (Fig. 3C, lane 4). After nonspecific immunoprecipitation (NS-IP) with nonimmune chicken serum, the signal for Src was absent (Fig. 3C, lane 3). There was very little association of Src with LPXN in osteoclast precursors (i.e., in the absence of RANKL in culture medium; data not shown). Finally, the overexpression of either the murine or rabbit orthologs of LPXN resulted in an increased association of LPXN with Src (∼2-fold, Fig. 3D, lanes 4 and 5, n = 4 separate experiments). We next examined the possibility of whether the levels of Src associated with LPXN were altered following treatment with an osteoclast-activating cytokine such as TNF-α (1, 45). LPXN was immunoprecipitated from osteoclasts after a 24-h exposure to the osteoclast-activating cytokine TNF-α (10 and 50 ng/ml), and the LPXN-immunoprecipitates were immunoblotted for Src, as shown in Fig. 3E; the positive signal for Src was evident in osteoclast lysates (Fig. 3E, lane 1, top), absent in the nonspecific immunoprecipitation (NS-IP, Fig. 3E, lane 2), but present at higher levels (∼1.5-fold, n = 3 separate experiments) in LPXN immunoprecipitates from TNF-α-treated osteoclasts (Fig. 3E, lanes 4 and 5), compared with untreated controls (Fig. 3E, lane 3). The membranes were subsequently stripped and immunoblotted for LPXN (Fig. 3E, bottom).
Subcellular distribution of LPXN and Src in osteoclasts.
The subcellular distribution of both the overexpressed murine and rabbit orthologs of LPXN and Src was examined using confocal microscopy, as shown in Fig. 4A, a–c. The immunofluorescence data showed that, after overexpression of either the murine (Fig. 4A, b) or rabbit (Fig. 4A, c) orthologs of LPXN, Src was colocalized with both orthologs of LPXN at the sealing zone (arrows). These data corroborated previous Western blot analysis that showed an increased association of Src with LPXN in LPXN-overexpressing osteoclasts.
Interaction of Src with the LD2 domain of LPXN.
Previous reports have demonstrated that several proteins can bind to the LD2 domain of paxillin; these include vinculin, Src, FAK/Pyk2, and talin (40). It has been demonstrated that paxillin is capable of binding Src either directly near the proline-rich NH2 terminus or indirectly (via FAK or Pyk2) near the LD2 domain (40). To determine whether similar Src-binding sites exist in LPXN, we examined the amino acid sequences in species orthologs of LPXN. Our analysis of the amino acid sequences derived from human, rabbit, and mouse LPXN orthologs revealed that the LD motifs were highly conserved, with the exception of the second LD motif (LD2), which was completely absent in the murine ortholog of LPXN, as shown in Fig. 4B. Previously, it has been reported that the paxillin paralog Hic-5 differs in the number of LD motifs; specifically, an LD1 motif is present in the murine Hic-5 sequence but not in the human ortholog (7, 40). Because the LD2 motif was highly conserved in both the human and rabbit orthologs (75% sequence identity), we designed and purified an rLPXN GST-LD2 fusion protein (with the sequence ESNLDETSKMLSV) that was used for GST-affinity precipitation assays. Our data indicate that Src is able to bind both to the rLPXN (GST-rLPXN, Fig. 4C, lane 3) and the LD2 protein (Fig. 4C, lane 4); the positive and negative controls included osteoclast lysate (Fig. 4C, lane 1) and GST protein only (Fig. 4C, lane 2). Finally, we examined the subcellular distribution of the LD2 domain of LPXN; for this purpose, the GST-LD2 protein was transfected into murine osteoclasts using a protein transfection reagent. Murine osteoclasts were transfected with 500 ng of the GST-LD2 fusion protein for 2 h and subsequently replated. The subcellular distribution of LD2-LPXN was determined via indirect immunofluorescence, using a monoclonal antibody against GST, as shown in Fig. 4D (arrow, frame 2, green), compared with control/untransfected osteoclasts (frame 1). Actin was labeled with rhodamine phalloidin. The immunofluorescent signal for the GST-LD2 fusion protein (in GST-LD2 transfected osteoclasts) was colocalized with the actin ring, presumably at the sealing zone (Fig. 4D, frame 2, yellow).
The biological functions of the paxillin-extended family of adaptor proteins include regulation of cell spreading and motility, and coordinating the recruitment of structural proteins, kinases, GTPase-activating proteins, and other adaptor proteins to specific cellular compartments, such as focal adhesions and podosomes (40). Several studies have previously identified paxillin as an adhesion-dependent tyrosine-phosphorylated component of the osteoclast podosomal complex (15, 32). Much less information exists on LPXN, which was identified as yet another constituent of the osteoclast sealing zone (19).
Previous studies have suggested that dramatic changes in the molecular composition of podosomes occur during polarization and adhesion before assembly of the resorptive apparatus (17, 32). A fraction of the total pool of Pyk2, which is predominantly cytoplasmic, is both tyrosine-phosphorylated and translocated to podosomes upon cell adhesion (16, 36, 51). After overexpression of LPXN, the apparent “gain-of-function”, as reflected by an increase in both resorptive activity and cell adhesion, was accompanied by an increase in the level of LPXN-associated Pyk2. These data were similar to LPXN-associated Pyk2 levels in PC-3 cells, a prostate cancer cell line, where overexpression of LPXN also resulted in an increase of LPXN-associated Pyk2 levels, accompanied by an increase in cell migration (unpublished observations, Gupta et al.).
Since the initial detection of paxillin as a tyrosine-phosphorylated substrate for Src, a direct association of the SH2 domains of the Src family of tyrosine kinases with sequences closer to the NH2-terminal domain of paxillin has been well established (41). In our efforts to explore an additional biochemical basis for the increased resorptive activity following overexpression of LPXN, we have also presented evidence that Src is an additional tyrosine kinase binding partner for the adaptor protein. Similar to paxillin, LPXN may either provide a platform for or a direct substrate for Src (48). Because the levels of Src associated with LPXN were increased after overexpression of LPXN in osteoclasts, we have provided additional data to support the role of LPXN as an important scaffold protein in osteoclast activity and adhesion. The cross-species conservation of LD protein-protein interaction domains between members of the paxillin family of adaptor proteins indicates that these peptide sequences are a critical feature in the specificity of function (7). In the current study, we have determined that both the human and rabbit orthologs of LPXN contain an LD2 domain that is ∼75% homologous to that of paxillin, whereas, surprisingly, the murine homolog of LPXN did not. Our GST-affinity precipitation data have demonstrated that the LD2 domain of the rabbit ortholog of LPXN was capable of directly binding Src. Furthermore, the LD2 domain of LPXN was localized predominantly with actin at the sealing zone, although we have not examined the functional consequences of this localization in terms of resorptive activity or cell adhesion. Evidently, because mLPXN was also found to be capable of binding Src in the absence of any consensus LD2 domain, other binding sites may exist for Src, an issue that has not been addressed in the current study.
Given that Src kinase activity is essential for normal osteoclast function, LPXN may provide an additional platform for Src, thereby regulating downstream effectors, such as Pyk2 and possibly LPXN itself. It has previously been reported that the osteoclastogenic and enhanced osteoclast-survival action of TNF-α (through induction of NF-κB activity) is partly mediated by Src (1, 22, 50). Previously, we have demonstrated modest changes in tyrosine phosphorylation upon treatment of osteoclasts with TNF-α (19). The modest increase in levels of Src that coimmunoprecipitated with LPXN upon exposure of osteoclasts to TNF-α demonstrates that LPXN may act as a dynamic scaffold protein in the process of osteoclast activation by the appropriate stimuli. In addition, there are several consensus sites for serine/threonine kinases, such as protein kinase C on LPXN, which can also provide downstream effector signals for TNF-α (37, 38), possibly resulting in “activation” of the scaffold protein.
In summary, we have demonstrated that overexpression of LPXN results in enhanced osteoclast function, such as resorptive activity mirrored by increased cell adhesion. We have further demonstrated that Src is another protein tyrosine kinase that can associate with LPXN in the sealing zone of the osteoclast. These data reinforce our previous observations that LPXN has an important role as an adaptor protein to coordinate the appropriate localization of key podosomal proteins in the sealing zone of the osteoclast.
This work was supported by National Institutes of Health Grant AR-44792 (to A. Gupta).
S. M. Núñez was supported by scholarships from the Barry M. Goldwater, MARC, Meyerhoff programs, and the Howard Hughes Medical Institute.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2007 the American Physiological Society