Osteopontin and Protein Kinase C Regulate PDLIM2 Activation and STAT1 Ubiquitination in LPS-treated Murine Macrophages*

The molecular pathways regulating signal transducer and activator of transcription 1 (STAT1) levels in states of inflammation are incompletely understood. The suppressor of cytokine signaling, protein inhibitor of STAT, and SHP-1/2 tyrosine phosphatases ultimately regulate activity of STAT molecules. However, these mechanisms do not degrade STAT proteins. In this regard, using a murine macrophage model of LPS stimulation, we previously demonstrated that osteopontin (OPN) increased STAT1 ubiquitination and 26 S proteasome degradation via the ubiquitin E3 ligase, PDLIM2. In this study, we further characterize OPN-dependent activation of PDLIM2 in a model of LPS-stimulated RAW264.7 murine macrophages. We identify serine 137 as a protein kinase C-phosphorylation site in PDLIM2 that is required for ubiquitination of STAT1. PDLIM2 phosphorylation requires OPN expression. Using phospho-mutants and phospho-mimetic constructs of PDLIM2, our in vivo and in vitro ubiquitination studies confirm the role of PDLIM2 in formation and degradation of Ub-STAT1. The functional consequences of PDLIM2-mediated STAT1 degradation were confirmed using an IFN-γ-regulated transcription factor STAT1α reporter construct and chromatin immunoprecipitation assay for the inducible nitric-oxide synthase promoter. In a murine cecal ligation and puncture model of sepsis in wild-type and OPN (−/−) animals, OPN was necessary for PDLIM2 serine phosphorylation and STAT1 ubiquitination in bone marrow macrophages. We conclude that OPN and PDLIM2 are important regulators of STAT1-mediated inflammatory responses.

Macrophage inducible nitric-oxide synthase (iNOS) 2 expression is central to many of the systemic effects associated with endotoxin (LPS) stimulation (1). Utilizing both in vivo and in vitro murine models of endotoxin (LPS) stimulation, we have previously demonstrated that osteopontin (OPN) is a potent NO feedback-regulated trans-repressor of iNOS expression (2,3). However, the underlying molecular pathway has not been fully characterized. We have previously examined the OPN-dependent effector arm of this NO-regulated negative feedback loop, focusing on iNOS transcriptional regulation. In this regard, signal transducer and activator of transcription 1 (STAT1) is an essential activator of LPS and/or pro-inflammatory cytokine-mediated iNOS transcription in murine, rat, and human cells (4). All mammalian iNOS promoters contain several sites with homologies to the interferon (IFN)-␥-regulated STAT1␣-binding sites (GAS). STAT signaling is tightly regulated, and several mechanisms have been proposed to account for this control (5). Post-translational modifications of STAT proteins via ubiquitination have been suggested as an important means to regulate STAT signaling (6).
Ubiquitination plays a major role in regulating many cellular processes by marking proteins, including transcription factors, for degradation through the 26 S proteasome-dependent pathway (7). Conjugation of ubiquitin to target proteins requires three enzymes, ubiquitin-activating enzyme (E1), ubiquitinconjugating enzyme (E2), and ubiquitin ligase (E3). E3 interacts with both E2 and the target protein to facilitate transfer of ubiquitin to the substrate. It is E3 ligase that confers specificity to the reaction. In the case of STAT1, Kim and Maniatis (6) were the first to demonstrate that IFN-activated STAT1 levels may be regulated by the ubiquitin-proteasome pathway. Tanaka et al. (8) were the first to identify PDLIM2 protein (previously known as SLIM or mystique) as a STAT ubiquitin E3 ligase and demonstrate a central role for ubiquitination in regulation of the IFN-STAT signaling pathway. Previously, using a murine macrophage model of LPS stimulation, we found that OPN increased STAT1 ubiquitination and degradation through the ubiquitin E3 ligase, PDLIM2 (9). In this study, we build upon our former observations to demonstrate that LPS-and OPNdependent activation of PDLIM2 and subsequent STAT1 ubiquitination require protein kinase C (PKC)-regulated phosphorylation of PDLIM2 Ser-137.
Assay of NO Production-NO released from cells in culture was quantified by measurement of the NO metabolite, nitrite (12). After stimulation, 50 l of culture medium was mixed with 50 l of 1% sulfanilamide in 0.5 N HCl. After a 5-min incubation at room temperature, an equal volume of 0.02% N-(1-naphthyl)ethylenediamine was added. Following incubation for 10 min at room temperature, the absorbance at 540 nm was compared with that of NaNO 2 standard on a MAXLINE micro-plate reader.
Construction and Characterization of AdV-PDLIM2 Expression Virus-PDLIM2 cDNA was inserted into the multiple cloning site of pShuttle-CMV vector (Stratagene, La Jolla, CA) to obtain pShuttle-CMV-PDLIM2. pShuttle-CMV-PDLIM2 and pAdEasy-1 vector (Stratagene, La Jolla, CA) were cotransfected into BJ5183 cells for homologous recombination. The recombinants were screened to obtain pAdV-PDLIM2. The verification of the AdV-PDLIM2 genome was carried out by DNA sequencing. The verified recombinant AdV-PDLIM2 was propagated in HEK 293 cells and then purified and stored at Ϫ80°C until use. Functional characterization was done by monitoring PDLIM2 expression using Western blot analysis.
AdV-PDLIM2-infected RAW264.7-RAW264.7 cells were plated on 150-mm dishes at a density of 1.5 ϫ 10 7 per dish. After 24 h, cells were infected at a 1000 multiplicity of infection. At 48 h post-infection, the cells were treated with LPS (50 ng/ml) for 6 h. The dishes were washed two times with PBS, and cells were harvested with lysate buffer.
Phosphopeptide Mapping of PDLIM2-WT-PDLIM2 was isolated from cell lysates and separated by PAGE. The resolving gel was stained; the WT-PDLIM2 band was excised and stored at Ϫ80°C until use. The gel slices were subjected to in-gel tryptic and chymotryptic digestions after reduction and carboxyamidomethylation. The resultant digests were pooled just prior to liquid chromatography MS/MS (tandem MS) injection. The phosphorylated peptide sequence analyses were performed at Harvard Microchemistry Facility, Cambridge, MA, by microcapillary reverse-phase HPLC nano-electrospray tandem mass spectrometry on a Finnigan LCQ DeCA XP plus quadrupole ion trap mass spectrometer.
Western Blot Analysis-Cells were lysed in buffer (0.8% NaCl, 0.02% KCl, 1% SDS, 0.1% Triton X-100, 0.5% sodium deoxycholic acid, 0.144% Na 2 HPO 4 , 0.024% KH 2 PO 4 , and 2 mM phenylmethylsulfonyl fluoride, pH 7.4), put on ice for 30 min, homogenized, and centrifuged at 12,000 ϫ g for 10 min at 4°C. The protein concentration was determined by the protein assay kit (Bio-Rad). The protein samples were separated by 4 -20% SDS-PAGE and electrotransferred onto polyvinylidene difluoride membranes (Amersham Biosciences) by semi-dry transfer. The membrane was blocked with 5% skim milk in PBS, 0.05% Tween for 1 h at room temperature and probed using primary antibodies for 1 h at room temperature, using appropriate horseradish peroxidase-conjugated secondary antibody. Bound peroxidase activity was detected by West Pico chemiluminescent kit (Pierce).
Immunoprecipitation Studies-Immunoprecipitation studies were performed as described previously (12). Cells were washed with cold PBS and lysed in IP buffer (10 mmol of Tris-HCl, pH 7.5, 3 mmol of EGTA, 20 mmol of NaCl, 0.02% Triton X-100, 1ϫ protease inhibitors mixture, 0.2 mmol of dithiothreitol, 1 mmol of phenylmethylsulfonyl fluoride). After homogenization, cell lysates were clarified by centrifugation at 13,000 rpm for 20 min at 4°C. Aliquots (500 g) were precleared by addition of 1.0 g of normal IgG and 20 l of appropriate agarose conjugate for 1 h at 4°C. After centrifugation, the supernatant was collected and incubated with 2 g of corresponding Ab for 1 h at 4°C. A total of 20 l of appropriate agarose conjugate was then added, and the mixture was incubated overnight. After washing five times with lysis buffer, immunoblotting was performed as described above.
Transient Transfection and Luciferase Assay-1.5 ϫ 10 5 RAW264.7 cells were plated on a 24-well plate and allowed to grow for 24 h before transfection. DNA transfections were carried out in 24-well plates using JetPEI macrophage reagent (Genesee, San Diego). At least 24 h later, the medium was changed, and the cells were stimulated with 50 ng/ml LPS for 12 h. To control transfection efficiency between groups, 0.1 g of pRL-TK vector was added to each well. Cells were harvested in 0.1 ml of reporter lysis buffer (Promega, Madison, WI), and Dual-Luciferase reporter assays were performed using the manufacturer's protocol (Promega, Madison, WI).
In Vitro Phosphorylation Assay-His-tagged PDLIM2, mutant His-tagged PDLIM2-S137A, and mutant His-tagged PDLIM2-S137D vectors were expressed in COS-1 cells. The recombinant proteins were purified with MagneHis protein purification system (Promega, Madison, WI). Prior to phos-phorylation, recombinant proteins were dialyzed (24 h, 4°C) against 10 mM Tris-HCl, pH 7.5, containing 1 mM DTT and 100 mM NaCl. The reaction for PKC activity was carried out in a solution containing 20 mM Tris-HCl, pH 7.5, 10 mM MgCl 2 , 1 mM DTT, 0.25 mM EDTA, 0.4 mM CaCl 2 , 0.3 mg/ml phosphatidylserine, 0.03 mg/ml diacylglycerol, 0.1 mg/ml BSA, and 10 g of recombinant PDLIM2 protein, 5 Ci of [␥-32 P]ATP, and 0.1 unit of PKC␣ (Calbiochem). A negative control was performed in the absence of PKC. After 30 min at 30°C, reactions were stopped with 2ϫ SDS-loading buffer, and the proteins were separated by SDS-PAGE. Gels were dried and subjected to autoradiography.
In Vitro Ubiquitination Assay-The in vitro ubiquitination assay was described previously (9). Recombinant wild-type PDLIM2, mutant His-tagged PDLIM2, or empty His vector was expressed in RAW264.7 cells. Following LPS stimulation for 6 h, recombinant proteins were purified with the MagneHis protein purification system (Promega, Madison, WI) as the source of E3 ligase. Presence of PDLIM2 was confirmed by rabbit anti-PDLIM2 Ab. FLAG-tagged P-STAT1 was expressed in LPS-treated RAW264.7 cells; the cells were lysed, and whole cell extracts were subjected to immunoprecipitation with anti-STAT1 or M2 anti-FLAG-agarose as the source of substrate for in vitro ubiquitination. Anti-FLAG STAT1 immunoprecipitates were washed five times with lysis buffer and once with ubiquitination buffer (50 mM Tris-Cl, pH 7.5, 2 mM MgCl 2 , 2 mM ATP, and 1 mM DTT) and incubated in 40 l of ubiquitination buffer with 5 g of Ub, 100 ng of Ub-activating enzyme (E1), and 100 ng of GST-UbcH5a (Calbiochem) in the absence or presence of recombinant PDLIM2 for 3 h at 30°C. After in vitro ubiquitination, all samples were washed thee times with lysis buffer, eluted with SDS-sample buffer, resolved by SDS-PAGE, and transferred electrophoretically to PVDF membrane. STAT1 ubiquitination was detected using Western blotting.
In Vivo Ubiquitination Assay-In vivo ubiquitination assays were carried out in RAW264.7. Plasmids were cotransfected with HA-ubiquitin and FLAG-STAT1 expression vectors into cells using JetPEI macrophage reagent (Genesee, San Diego). After 24 h, RAW264.7 cells were treated with 10 M MG132 for 2 h and incubated for 6 h with 50 ng/ml LPS prior to harvest. Cells lysates were clarified by centrifugation at 14,000 rpm for 20 min at 4°C, and the supernatant was precleared with protein G-agarose for 1 h at 4°C. The lysates were subsequently immunoprecipitated using anti-STAT1 or M2 anti-FLAG-agarose at 4°C for 12 h. Immunoprecipitated proteins were subjected to SDS-PAGE analysis and analyzed with anti-HA antibody using Western blot analysis.
Chromatin Immunoprecipitation/Real Time-PCR-Chromatin was fixed and immunoprecipitated using the chromatin immunoprecipitation (ChIP) assay kit (Upstate Biotechnology, Billerica, MA) as recommended by the manufacturer. Sequences were identified for the mouse iNOS promoter (GenBank TM accession number L09126). Purified chromatin was immunoprecipitated using anti-STAT1 (Santa Cruz Biotechnology) or 5 l of rabbit nonimmune serum. Eluted DNA fragments were purified to serve as templates. The input fraction corresponded to 0.1 and 0.05% of the chromatin solution before immunoprecipitation. The average size of the sonicated DNA fragments subjected to immunoprecipitation was 500 bp as determined by ethidium bromide gel electrophoresis. After DNA purification, the presence of the selected DNA sequence was quantified by real time-PCR. ChIP assays for activated Stat1 binding to its IFN-␥-regulated transcription factor STAT1␣ (GAS) site on the iNOS promoter used primers ACACGAGGCTGAGCTGACTT and CACACATGGCATG-GAATTTT, resulting in a 186-bp fragment. Serial dilutions of genomic DNA were used to generate standard curves for RT-PCR using the corresponding primer sets. Copy numbers of the eluted promoter were normalized against the copy numbers in the corresponding inputs and expressed in arbitrary units.
Murine Cecal Ligation and Puncture Model of Sepsis-Wildtype (WT) and OPN(Ϫ/Ϫ) 129 male mice (12-16 weeks old) were utilized. OPN(Ϫ/Ϫ) animals were a gift from David Denhardt, Rutgers University. They were maintained on sterile standard laboratory chow and water ad libitum in individual ventilated cages under specific pathogen-free conditions in the animal facility at Duke University. All animal experiments were approved by the Duke University Institutional Animal Care and Use Committee. Via a midline laparotomy incision, the cecum was ligated by 4.0 silk distal to the ileocecal junction; a 18-gauge needle was passed through both walls of the cecum. After a small amount of fecal matter was expressed, the cecum was returned to the peritoneal cavity, and the abdomen was closed. Sham-operated animals served as controls. A saline bolus (20 ml/kg) was administered subcutaneously to provide resuscitation. At 24 h, animals were euthanized, and bone marrow macrophages (BMM) were isolated.
Isolation of BMM-BMM were isolated by lavage of excised femurs, washed in PBS, plated in DMEM with 0.2% BSA into two 100-mm tissue culture dishes, and allowed to adhere for 2 h at 37°C. After washing out nonadherent cells, cells were cultured in DMEM with 10% FBS and 20% L cell-conditioned medium (as a source of M-CSF) to generate confluent BMM. Primary BMM were maintained at 37°C under 5% CO 2 in DMEM supplemented with 10% heat-inactivated FCS, 100 units/ml penicillin, and 100 g/ml streptomycin. After incubation for the designated time period at 37°C in 5% CO 2 , the cells or media were harvested for assays.
Isolation of Peritoneal Macrophages-Mice were sacrificed, and macrophages were harvested by peritoneal lavage. Physiological saline (10 ml) was injected into the peritoneal cavity. The peritoneum was then massaged for ϳ2 min to bring cells into suspension, and the fluid was extracted using a hypodermic needle. Cells were then centrifuged at 200 ϫ g and pooled into RPMI 1640 culture medium supplemented with 10% FBS, 2 mM L-glutamine, 20 mM HEPES, penicillin, and streptomycin at 2 ϫ 10 6 cells/ml. The cell suspension was then seeded into polystyrene Petri dishes (10 ml per dish) and incubated overnight in an atmosphere of 5% (v/v) CO 2 in air at 37°C to allow selective macrophage adherence. Contaminating nonadherent cells were removed by thoroughly washing each dish with phosphate-buffered saline (without Ca 2ϩ and Mg 2ϩ ) at pH 7.4. Purified macrophages were then removed by incubation with 0.5% Statistical Analysis-All data are presented as mean Ϯ S.D. Analysis was performed using the Student's t test. Values of p Ͻ 0.05 were considered significant.

Identification of LPS-dependent Serine/Threonine Phosphorylation Sites in PDLIM2-
In the setting of LPS stimulation of RAW264.7 macrophages, OPN increases STAT1 ubiquitination and degradation through PDLIM2. Phosphopeptide mapping was performed to identify the specific PDLIM2 amino acid phosphorylation targets in the setting of LPS exposure. Sequence analysis using PhosphoScan-site 2.0 indicates the potential presence of 28 serine, 8 threonine, and/or 2 tyrosine phosphorylation sites. To determine whether LPS and OPN induce PDLIM2 phosphorylation, we treated RAW264.7 cells with LPS (50 ng/ml) for 6 h and performed IP to measure the extent of phosphorylated serine, threonine, and tyrosine in PDLIM2 (Fig. 1). Controls consisted of untreated cells and LPS ϩ siRNA OPN-treated cells. The results demonstrate that LPS increased PDLIM2 serine phosphorylation compared with untreated controls and LPS ϩ siRNA OPN. Repletion of OPN to LPS ϩ siRNA OPN cells restored serine phosphorylation in PDLIM2. An adenovirus PDLIM2 expression vector pAdV-PDLIM2 was constructed. RAW264.7 cells were infected at a 1000 multiplicity of infection by AdV-PDLIM2. After 24 h, cells were treated with LPS or LPS ϩ siRNA OPN for 6 h, and PDLIM2 protein was purified by PAGE and submitted to the Harvard Microchemistry and Proteomics Analysis Facility for characterization. Untreated cells and cells infected with empty pAdV served as additional controls. When compared with control and LPS ϩ siRNA OPN, LPS induces two potentially unique serine/threonine phosphorylations sites, Ser-137 and Thr-138, within the sequence RPCSPFSTPP. Otherwise, there were no additional phosphorylated serine, threonine, or tyrosine residues noted under any of the treatment conditions. To determine the potential functional relevance of these phosphorylation sites, each was mutated to a phosphoresistant (Ser/Thr to Ala) or phosphomimetic (Ser to Asp) amino acid, expressed, isolated, and tested in subsequent studies.
E3 Ligase Function and PDLIM2 Mutants-Three PDLIM2 phosphorylation mutants, S137A, T138A, and S137D, were created and examined for E3 ligase activity using in vivo and in vitro ubiquitination assays (Fig. 2). Recombinant His-tagged PDLIM2, His-tagged PDLIM2 mutants, or empty His vectors were transfected and expressed in RAW264.7 cells and purified to use as the source of E3 ligase. In selected instances, cells were stimulated with LPS (50 ng/ml) for 6 h. FLAG-tagged STAT1  was expressed in RAW264.7 cells and used as the source of substrate for in vitro ubiquitination. (Fig. 2a) A Ub-STAT1 ladder was found in LPS ϩ WT-PDLIM2 and LPS ϩ T138A but was absent in the LPS ϩ S137A. Similarly, Ub-STAT1 was also found in the control S137D setting.
We then performed in vivo ubiquitination assays in which RAW264.7 cells were cotransfected with WT-PDLIM2 or mutant forms of PDLIM2 with HA-Ub and FLAG-STAT1 vectors. After 24 h, cells were treated with MG132 (10 M) and in selected instances were stimulated with LPS (50 ng/ml) for 6 h (Fig. 2b). The Ub-STAT1 complex was immunoprecipitated and subjected to Western blot analysis using anti-HA Ab. Ub-STAT1 was found in LPS ϩ WT-PDLIM2, LPS ϩ T138A, and LPS ϩ T138D cells, but it was absent in the LPS ϩ S137A. Ub-STAT1 was not found in the control S137D setting, secondary to the absence of activated STAT1. These data suggest that phosphorylation of PDLIM2 at Ser-137 is required for OPN-dependent formation of ubiquitinated P-STAT1 in LPS-treated macrophages.
OPN-dependent PDLIM2 Ser-137 Phosphorylation and STAT1-Given its defined role as a STAT1 E3 ligase, the functional role of LPS-dependent PDLIM2 Ser-137 phosphorylation was then determined in the context of STAT1 degradation using PDLIM2-S137A as an antagonist or PDLIM2-S137D as an agonist (8,9). Using RAW cells, total cellular and phosphorylated nuclear STAT1s (P-STAT1) were measured in the presence and absence of LPS (Fig. 3a). In selected instances, cells were transfected with PDLIM2-S137A, PDLIM2-S137D, and/or siRNA OPN prior to LPS stimulation. In the presence of LPS, nuclear P-STAT1 was readily detected in contrast to untreated controls. When PDLIM2-S137A or siRNA OPN were added with LPS, the levels of both P-STAT1 (ϳ4-fold) and total STAT1 (ϳ9-fold) were significantly increased compared with LPS alone (p Ͻ 0.05 LPS versus LPS ϩ S137A and LPS ϩ siRNA OPN for both P-STAT1 and total STAT1). Exogenous OPN restored expression of both STAT1 signals to base-line levels in LPS ϩ siRNA OPN, but it had no effect in LPS ϩ PDLIM2-S137A (p Ͻ 0.05 LPS versus LPS ϩ S137A ϩ OPN for both P-STAT1 and total STAT1). When PDLIM2-S137D phosphomimetic was added to LPS or LPS ϩ OPN groups, P-STAT1 and total STAT1 were both detected at levels equivalent or less than that of LPS alone. LPS ϩ mismatch siRNA did not alter the STAT1 profile in comparison to LPS alone; OPN added to unstimulated control cells did not alter STAT1 expression in comparison with control alone (data not shown).
IP studies were then performed to assess the extent of Ubassociated P-STAT1 in the presence of cycloheximide and MG132 (Fig. 3b). Immunoblot analysis for P-STAT1 was performed as a control. In the setting of LPS stimulation, P-STAT1 and Ub-P-STAT1 were readily detected compared with unstimulated controls. However, when OPN expression was ablated in the LPS ϩ siRNA OPN group, the ubiquitinated P-STAT1 was not seen despite the presence of P-STAT1. Restoration of OPN to this treatment condition resulted in a dramatic increase in Ub-P-STAT1 to a level no different from that of LPS. When PDLIM2-S137A was added to LPS cells, Ub-P-STAT1 was significantly decreased (ϳ8-fold) compared with LPS alone (p Ͻ 0.02), although the level of P-STAT1 was increased. Addition of exogenous OPN in this setting does not significantly alter the levels of P-STAT1 or Ub-P-STAT1. The addition of S137D to LPS cells did not alter their P-STAT1 and Ub-P-STAT1 profiles. LPS ϩ mismatch siRNA and OPN ϩ control cells did not alter the P-STAT1 profile in comparison with LPS or control cells, respectively (data not shown). Taken as a whole, these data suggest that LPS and OPN mediate PDLIM2 Ser-137 phosphorylation, and PDLIM2 Ser-137 phosphorylation was required for formation of Ub-P-STAT1 and its subsequent degradation via the 26 S proteasome pathway.
OPN Signal Transduction and PDLIM2 Phosphorylation-The amino acid sequence surrounding Ser-137 was submitted for protein kinase substrate analysis by NetPhosK 1.0 (13). Based upon these results, we utilized pharmacological inhibitors for PKC, p38, PKA, PI3K, and MEK1/2 in LPS-stimulated RAW264.7 cells and performed IP for serine-phosphorylated PDLIM2 (Fig. 4a). Administration of the general PKC inhibitor, G06850, was associated with significantly decreased serine- a, functional role of LPS-dependent PDLIM2 Ser-137 phosphorylation was determined in the context of STAT1 degradation using PDLIM2-S137A as an antagonist or PDLIM2-S137D as an agonist. Immunoblot analysis was performed using RAW cells to measure total cellular and phosphorylated nuclear STAT1 (P-STAT1) in the presence and absence of LPS. In selected instances, cells were transfected with PDLIM2-S137A, PDLIM2-S137D, and/or siRNA OPN prior to LPS stimulation. Blot is representative of three experiments. b, IP studies were performed to assess the extent of Ub-associated P-STAT1 in the presence of cycloheximide (10 g/ml) and MG132 (10 M). Cell lysates (500 g) were immunoprecipitated with P-STAT1 Ab and immunoblotted (IB) with anti-Ub Ab. In selected instances, cells were transfected with PDLIM2-S137A, PDLIM2-S137D, and/or siRNA OPN prior to LPS stimulation. Immunoblot analysis for P-STAT1 was performed as a control. Blot is representative of three experiments. phosphorylated PDLIM2. This was not reversed by addition of exogenous OPN. We then used several PKC inhibitors to identify the roles of PKC isoforms in LPS-induced PDLIM2 phosphorylation (Fig. 4b). RAW264.7 cells were transfected with AdV-PDLIM2 for 24 h and then pretreated with G06850 (general PKC inhibitor), G06976 (PKC␣ and PKC␤ inhibitor), rottlerin (PKC␦ and PKC inhibitor), and pseudosubstrate peptide (PKC inhibitor). DMSO treatment served as a control; cells were stimulated with LPS. We found that LPS-induced PDLIM2 phosphorylation was not altered by a pseudosubstrate peptide inhibitor of PKC, but it was significantly suppressed by G06976 and G06850; also, it was slightly inhibited by rottlerin, which has reduced potency for inhibiting PKC␣ and PKC␤. In the presence of cycloheximide and MG132, we then performed in vivo ubiquitination assays for Ub-STAT1 in LPS-treated cells using the various PKC inhibitors (Fig. 4c). Again, LPS induced a Ub-P-STAT1 ladder; this was significantly diminished in the presence of G06850 and G096976, suggesting significantly decreased PDLIM2 activity in the presence of these PKC inhibitors.
In vitro phosphorylation assays were performed (Fig. 4d). PDLIM2 was immunoprecipitated from RAW cells and incubated in the presence of classical PKC␣. S137A and S137D were used as controls. PKC isoform phosphorylated only wild-type PDLIM2. In contrast, we saw no phosphorylation of the S137A or S137D mutants. Total PDLIM2 protein in the immunoprecipitates, detected by immunoblot analysis, showed similar protein levels for all of the mutants tested. These results implicate serine 137 as the only PKC-dependent PDLIM2 phosphorylation site in the presence of LPS. These results indicate that PKC␣ plays a significant role in PDLIM2 phosphorylation and P-STAT1 ubiquitination in LPStreated macrophages.
PDLIM2 and STAT1 Function-A GAS luciferase reporter construct was then used to assess the potential impact of PDLIM2 function on P-STAT1 transcriptional activity (Fig. 5a). Transient transfection assays were repeated using a pGL3 luciferase reporter plasmid construct bearing a quadruple repeat of the GAS enhancer element, agtttcataTTACTCTAAatc. LPS stimulation significantly increased STAT1-dependent luciferase activity in both LPS and LPS ϩ PDLIM2 when compared with unstimulated controls. OPN knockdown in LPS ϩ PDLIM2 ϩ siRNA OPN resulted in an additional 2.2-fold increase in luciferase activity. When PDLIM2-S137A was added to LPS cells, luciferase activity was significantly increased compared with LPS ϩ PDLIM2 and LPS alone. In LPS ϩ S137D, luciferase activity was 75% that found in LPS ϩ PDLIM2; in LPS ϩ T138A, luciferase activity was not statistically different from LPS ϩ PDLIM2 and LPS alone. We next measured nitrite levels, an NO metabolite, as an end point of P-STAT1 transactivation of the iNOS promoter (Fig.  5b). The murine iNOS promoter contained two NF-B-binding sites, NF-B1 (nucleotides Ϫ1044 to Ϫ1034) and NF-B2 (nucleotides Ϫ114 to Ϫ104), and a single STAT1-binding site (GAS; nucleotides Ϫ942 to Ϫ934), deemed critical for iNOS FIGURE 4. Protein kinase C and PDLIM2 phosphorylation. a, pharmacological inhibitors of PKC, p38, PKA, PI3K, and MEK1/2 were incubated with LPS-stimulated RAW264.7 cells. IP was performed for PDLIM2 and then blotted with anti-phosphoserine Ab to determine serine phosphorylated PDLIM2. Immunoblot for PDLIM2 was performed as a control. Blot is representative of three experiments. b, PKC pharmacological inhibitors, G06850 (general PKC inhibitor), G06976 (PKC␣ and PKC␤ inhibitor), rottlerin (PKC␦ and PKC inhibitor), and pseudosubstrate peptide (PKC inhibitor), were incubated with LPS-stimulated RAW264.7 cells. IP was performed for PDLIM2 and then blotted with anti-phosphoserine Ab to determine serine-phosphorylated PDLIM2. Immunoblot for PDLIM2 was performed as a control. Blot is representative of three experiments. c, in vivo ubiquitination assay. RAW264.7 cells were cotransfected with WT-PDLIM2, HA-Ub, and FLAG-STAT1 vectors. The PKC pharmacological inhibitors, G06850 (general PKC inhibitor), G06976 (PKC␣ and PKC␤ inhibitor), rottlerin (PKC␦ and PKC inhibitor), and pseudosubstrate peptide (PKC inhibitor), were utilized. After 24 h, cells were treated with MG132 (10 M) and stimulated with LPS (50 ng/ml) for 6 h. The Ub-STAT1 complex was immunoprecipitated and subjected to Western blot analysis using anti-HA Ab. The blot is representative of three experiments. d, in vitro phosphorylation assay. PDLIM2, S137A, and S137D were immunoprecipitated from RAW cells and incubated in the presence and absence of classical PKC␣ and [␥-32 P]ATP. S137A and S137D were used as controls. Blot is representative of two experiments. expression (9). RAW264.7 cells were transfected using empty vector, wild-type PDLM2, PDLIM2-S137A, PDLIM2-S137D, or PDLIM2-T138A; after 24 h, cells were treated with LPS and nitrite measured in the culture medium following a 6-h incubation. Compared with wild-type PDLIM2, PDLIM2-S137A significantly increased nitrite levels, reflecting increased P-STAT1. Ablation of OPN with siRNA also resulted in significantly increased nitrite in the culture medium. PDLIM2-S137D and PDLIM2-T138A did not significantly alter nitrite production. These data suggest that increased levels of P-STAT1 functionally translate into increased NO production. We then proceeded to perform ChIP/RT-PCR assays to determine the extent of LPS-mediated STAT1 binding to the iNOS promoter (Fig. 5c). These results demonstrate significantly increased STAT1 binding in the presence of S137A and siRNA to OPN, reflecting increased STAT1 as the result of decreased Ub-dependent degradation.
PDLIM2 Phosphorylation in a Murine CLP Model of Sepsis-To determine the in vivo relevance of these observations, we utilized a CLP model of sepsis utilizing WT and OPN(Ϫ/Ϫ) mice. Animals were sacrificed after 24 h, and BMMs were studied (Fig. 6). PDLIM2 was immunoprecipitated and Westernblotted for serine phosphorylation. In WT animals, CLP was associated with significantly increased PDLIM2 serine phosphorylation. In OPN(Ϫ/Ϫ), serine-phosphorylated PDLIM2 was essentially absent in CLP. Western blot analysis in BMM was then performed to determine iNOS, P-STAT1, and OPN expression in CLP in WT and OPN null animals at 0 and 24 h. WT expression of iNOS and P-STAT1 in BMM was significantly less (ϳ10-fold) than that of OPN(Ϫ/Ϫ) (p Ͻ 0.01 for iNOS and P-STAT1). These results show that iNOS and P-STAT1 were significantly increased in the absence of OPN in a CLP model of sepsis. WT-sham and OPN(Ϫ/Ϫ) sham expression of iNOS, P-STAT1, and OPN at 0 and 24 h was statistically equivalent to the expressions seen in WT-CLP and OPN null FIGURE 5. Functional assessment of PDLIM2-dependent STAT1 degradation. a, transient transfection analysis of consensus GAS promoter activity. RAW cells were transiently transfected with WT-PDLIM2, mutant forms of PDLIM2, siRNA OPN or the mismatch control siRNA MM for 24 h. The cells then underwent repeat transfection using a pGL3 luciferase reporter plasmid construct bearing a quadruple repeat of the GAS enhancer element. At least 24 h later, the medium was changed, and cells were exposed to LPS (50 ng/ml) for 6 h. Wild-type cells served as controls. Data are presented as mean Ϯ S.E. of four experiments. *, p Ͻ 0.01 versus wild type; **, p Ͻ 0.01 versus wild type, GAS, GAS ϩ PDLIM2, GAS ϩ S137D, and GAS ϩ T138A; @, p Ͻ 0.05 versus wild type, GAS, GAS ϩ PDLIM2, and GAS ϩ T138A. b, nitrite production. RAW cells were transiently transfected with empty vector, WT-PDLIM2, mutant forms of PDLIM2, siRNA OPN, or the mismatch control (siRNA MM) for 24 h. At least 24 h later, the medium was changed, and cells were exposed to LPS (50 ng/ml) for 6 h. Nitrite was measured in the culture medium. Wild-type cells served as controls. Data are presented mean Ϯ S.E. of four experiments. *, p Ͻ 0.01 versus wild type, empty vector, PDLIM2, S137D, and T138A. c, ChIP-RT-PCR assay for activated STAT1 binding to the iNOS promoter. RAW cells were transiently transfected with empty vector, WT-PDLIM2, mutant forms of PDLIM2, siRNA OPN, or the mismatch control (siRNA MM) for 24 h. At least 24 h later, the medium was changed, and cells were exposed to LPS (50 ng/ml) for 6 h. Wild-type cells served as controls. Purified chromatin was immunoprecipitated using anti-STAT1 (Santa Cruz Biotechnology) or 5 l of rabbit nonimmune serum; eluted DNA fragments were purified to serve as templates. ChIP assays for P-Stat1 binding to its GAS site in the iNOS promoter used primers ACACGAGGCTGAGCTGACTT and CACACATGGCATGGAATTTT, resulting in a 186-bp fragment. Serial dilutions of genomic DNA were used to generate standard curves for real time PCR using the corresponding primer sets. Copy numbers of the eluted promoter were normalized against the copy numbers in the corresponding inputs and expressed in arbitrary units. Data are presented mean Ϯ S.E. of three experiments. *, p Ͻ 0.01 versus wild type, empty vector, PDLIM2, S137D, and T138A. NOVEMBER 26, 2010 • VOLUME 285 • NUMBER 48 CLP at 0 h for both BMM and liver, respectively (data not shown). To address the role of OPN in ubiquitination of P-STAT1 in this CLP model, P-STAT1 was immunoprecipitated to determine Ub-P-STAT1 expression in BMM at 24 h.

Osteopontin and PDLIM2
In BMM from WT animals, CLP was associated with ubiquitination of P-STAT1; in OPN(Ϫ/Ϫ), Ub-P-STAT1 was significantly less than that noted in WT (p Ͻ 0.01; WT versus OPN (Ϫ/Ϫ)), although overall P-STAT1 levels were increased in the OPN(Ϫ/Ϫ) animals. These in vivo studies indicate that OPN is necessary for PDLIM2 serine phosphorylation and STAT1 ubiquitination to regulate downstream STAT1-dependent functions such as iNOS expression.
We then addressed PDLIM2 serine phosphorylation, P-STAT1 formation, and nitrite expression in primary murine peritoneal macrophages (Fig. 7). Serine-phosphorylated PDLIM2 was readily detected on IP with LPS treatment; in the presence of PDLIM2-S137A or G06850, LPS-induced serinephosphorylated PDLIM2 was significantly decreased (p Ͻ 0.01 LPS versus LPS ϩ S137A or LPS ϩ G06850). Nuclear P-STAT1 was also significantly increased in the presence of LPS with PDLIM2-S137A or G06850. Nitrite production in BMM and peritoneal macrophages was measured to support functional biological significance. LPS-induced NO was significantly increased with exogenous PDLIM2 and, conversely, decreased with PDLIM2-S137A (p Ͻ 0.01 LPS versus LPS ϩ PDLIM2 or LPS ϩ PDLIM2-S137A).
In summary, our results indicate that OPN functions as a feedback inhibitor of iNOS expression by PKC-mediated activation of the E3 ligase, PDLIM2, which targets STAT1 for degradation by the 26 S proteasome.

DISCUSSION
Evidence indicates that LPS-mediated iNOS gene transcription is an exceedingly complex and redundant signal transduction pathway with varying signal-and cell-dependent responses. Specifically, in systemic inflammation induced by LPS, the macrophage is responsible for the majority of the circulating NO metabolites. Macrophage iNOS expression is central to many of the systemic effects associated with LPS stimulation. However, although the molecular pathways that up-regulate iNOS expression have been extensively studied in multiple cell types, including the macrophage, little is known of the parallel counter-regulatory pathways that repress or inhibit macrophage iNOS expression in the context of endotoxemia and sepsis. Previously, using a system of LPS-treated RAW264.7 macrophages, we demonstrated that OPN regulates ubiquitin-dependent degradation of STAT1 through the activity of PDLIM2 as the ubiquitin E3 ligase (9). This regulation of STAT1 degradation underlies the effect of OPN as an inhibitor of iNOS gene transcription. These results define OPN as a unique and as yet incompletely characterized trans-activator of STAT1 degradation by the Ub-proteasome system. In this study, we demonstrate that LPS-and OPN-dependent activation of PDLIM2 and subsequent STAT1 ubiquitination require PKC phosphorylation of PDLIM2 Ser-137.
STAT signaling is tightly regulated, and several mechanisms have been proposed to account for this control (14). The suppressor of cytokine signaling (SOCS) and protein inhibitor of STAT (PIAS) families of proteins have been shown to bind to and inhibit either the cytokine receptor-associated Janus kinase (JAK) or activated STAT molecule, respectively. SOCS proteins are induced following cytokine stimulation, bind to JAK kinases, and inhibit activated JAK kinases from further phosphorylation of STAT proteins (15,16). As a result of inhibition by SOCS proteins, STAT signaling becomes a transient response, and levels of phosphorylated STAT pro- A, serine-phosphorylated PDLIM2. IP studies were performed to assess the extent of Ub-associated P-STAT1 in BMM. Cell lysates (500 g) were immunoprecipitated with PDLIM2 Ab and immunoblotted (IB) with anti-phosphoserine Ab. Immunoblot analysis for PDLIM2 was performed as a control. Blot is representative of three experiments. B, Western blot of iNOS, phosphorylated STAT1 (P-STAT1), and OPN expression in CLP in WT and OPN(Ϫ/Ϫ) animals at 0 and 24 h. Western blot analysis was performed for iNOS, OPN, and P-STAT1 proteins, as described under "Experimental Procedures." Blot is representative of four experiments. C, immunoprecipitation and ubiquitination of P-STAT1 in CLP. Cells were lysed and cleared by centrifugation. Aliquots were precleared by addition of normal IgG and 20 l of appropriate agarose conjugate. After centrifugation, the supernatant was collected and incubated with 2 g of P-Stat1 Ab. After washing, immunoblotting was performed using Ub AB. Blot is representative of four experiments.
teins decrease within hours of activation. However, the SOCS feedback mechanism does not result in reduced levels of STAT proteins. Each member of the PIAS family has been shown to inhibit STAT-mediated gene activation. Another mechanism that down-regulates STAT signaling involves tyrosine phosphatases, such as SHP-1 and SHP-2. These tyrosine phosphatases have been shown to down-regulate the activity of STAT1, STAT3, or STAT5 following activation by IFN-␥, leukemia inhibitory factor, or IL-2 either by dephosphorylating STAT proteins directly or through dephosphorylation of JAK kinases (17)(18)(19). Like the SOCS and PIAS mechanisms, the down-regulation through tyrosine phosphatases also does not lead to reduced levels of STAT proteins.
PDLIM2 (also known as Mystique and SLIM) is a postsynaptic density-95 (PSD-95)/discs large (DLG)/zonula occludens-1 (ZO-1)-Lin-11, Isl-1, Mec-3 (PDZ-LIM) domain protein that was identified in cells transformed by overexpression of the insulin-like growth factor-I receptor (20). It is a member of the actinin-associated LIM protein family of proteins that contains a single N-terminal PDZ and C-terminal LIM domain. PDLIM2 contains a PDZ domain and a LIM domain and interacts in the nucleus with tyrosine-phosphorylated STAT molecules (21). The LIM domain forms a zinc finger structure related to the RING finger and PHD structures; similar proteins have been shown to possess E3 ligase activity (8). For a specific substrate, E3 ligase is the only member of the E1, E2, and E3 sequence that undergoes regulation (21,22). The PDZ and LIM domains mediate protein-protein interactions. The cDNA encoding mouse PDLIM2 is 1509 bp (GenBank TM BC024556) for mouse with an open reading frame of 349 amino acids (8,21). It is expressed as a 38-kDa nuclear protein. The human transcript (GenBank TM NM_021630) is homologous to mouse with 77% identity at the amino acid level and 79% identity at the cDNA level. It is highly expressed in lung, spleen, thymocytes, and primary hematopoietic cells, including macrophages (8,9). PDLIM2 is the first identified ubiquitin ligase with specificity for STAT proteins.
The relationship between OPN and PDLIM2 has not been previously characterized. OPN is a secreted glycoprotein that is rich in aspartate and sialic acid residues and contains functional domains for calcium binding, phosphorylation, glycosylation, and extracellular matrix adhesion (23). OPN appears to mediate cell-matrix interactions and cellular signaling through binding with integrin, primarily ␣ v ␤ 3 , and CD44 receptors. A variety of stimuli, including phorbol 12-myristate 13-acetate, 1,25-dihydroxyvitamin D, basic fibroblast growth factor (bFGF), TNF-␣, IL-1, IFN-␥, and LPS appears to regulate OPN expression. OPN has multiple molecular functions that mediate cell adhesion, chemotaxis, macrophage-directed IL-10 suppression, stress-dependent angiogenesis, prevention of apoptosis, and anchorage-independent growth of tumor cells (23). Our studies have demonstrated that OPN functions as a feedback regulated trans-repressor of iNOS expression (2,3). Our data indicate that the underlying molecular pathway is LPS-and OPN-dependent PKC-regulated phosphorylation of PDLIM2 Ser-137 with subsequent ubiquitination and proteasome degradation of STAT1. A link between OPN and PKC has been previously established. In PC-3 prostate cancer cells, Jain et al. (24)   . PDLIM2 phosphorylation and P-STAT1 in primary murine peritoneal macrophages. A, serine-phosphorylated PDLIM2. IP studies were performed to assess the extent of phosphorylated PDLIM2. We treated primary murine peritoneal macrophages with LPS (50 ng/ml) for 6 h. In selected instances, G06850 (general PKC inhibitor) or PDLIM2-S137A was also added. IP was performed for PDLIM2 and then blotted with anti-phosphoserine Ab to determine serine-phosphorylated PDLIM2. Blot is representative of three experiments. B, P-STAT1 formation. Western blot analysis was performed for P-STAT1 proteins, as described under "Experimental Procedures." Blot is representative of four experiments. c, nitrite production in BMM and primary peritoneal macrophages (PM). Data are presented as mean Ϯ S.E. of four experiments. *, p Ͻ 0.01 versus control and LP; **, p Ͻ 0.01 versus control. IB, immunoblot.