Toll-like Receptor 2 and Mitogen- and Stress-activated Kinase 1 Are Effectors of Mycobacterium avium-induced Cyclooxygenase-2 Expression in Macrophages*

Understanding how pathogenic mycobacteria subvert the protective immune response is crucial to the development of strategies aimed at controlling mycobacterial infections. Prostaglandin E2 exerts an immunosuppressive function in the context of mycobacterial infection. Because cyclooxygenase-2 (COX-2) is a rate-limiting enzyme in prostaglandin biosynthesis, there is a need to delineate the mechanisms through which pathogenic mycobacteria regulate COX-2 expression in macrophages. Our studies demonstrate that the NF-κB and CRE elements of the COX-2 promoter are critical to Mycobacterium avium-induced COX-2 gene expression. M. avium-triggered signaling originates at the Toll-like receptor 2 (TLR2). Ras associates with TLR2 and activates the mitogen-activated protein kinase (MAPK) extracellular signal-regulated kinase (ERK), whereas tumor necrosis factor receptor-associated factor 6 (TRAF6)/transforming growth factor β-activated kinase 1 (TAK1)-dependent signaling activates p38 MAPK. Both ERK and p38 MAPK activation converge to regulate the activation of mitogen- and stress-activated kinase 1 (MSK1). MSK1 mediates the phosphorylation of the transcription factor CREB accounting for its stimulatory effect on CRE-dependent gene expression. M. avium-triggered cytoplasmic NF-κB activation following IκB phosphorylation is necessary but not sufficient for COX-2 promoter-driven gene expression. MSK1 activation is also essential for M. avium-triggered NF-κB-dependent gene expression, presumably mediating nucleosomal modifications. These studies demonstrate that the nuclear kinase MSK1 is necessary in regulating the pathogen-driven expression of a gene by controlling two transcription factors. The attenuation of MSK1 may therefore have potential benefit in restricting survival of pathogenic mycobacteria in macrophages.

To design novel therapeutics and to rationalize vaccination strategies against mycobacterial diseases, it is of foremost importance to understand the mechanisms by which mycobacte-ria subvert the protective immune response. The response of macrophages to invasion by intracellular pathogens such as mycobacteria depends on the signals triggered during the initial contact of the bacterium with the macrophages. Prostaglandin E 2 (PGE 2 ) 1 is believed to exert an immunosuppressive function in the context of mycobacterial infections. PGE 2 interferes with T lymphocyte responses by inhibiting the production of interleukin-2 (1). PGE 2 also inhibits the secretion of interferon-␥, which is important in activating T cells and macrophages (2). Cooper and colleagues (3) reported that the ability of the virulent Mycobacterium avium to attenuate macrophage activation was dependent on its ability to trigger p38 mitogenactivated protein kinase (MAPK)-dependent production of PGE 2 . These observations were supported by a case report in which monocytes from a child with disseminated M. avium infection demonstrated a defect in bactericidal activity which could be offset by treatment with indomethacin (4). The critical role of PGE 2 is also supported by a report that human immunodeficiency virus envelope glycoprotein (gp120) can enhance M. avium replication within macrophages by induction of PGE 2 (5).
Cyclooxygenase-2 (COX-2) is a key enzyme catalyzing the rate-limiting step in the inducible production of prostaglandins. COX-2 catalyzes the conversion of arachidonic acid to PGH 2 , which is then metabolized by terminal synthases to a variety of biologically active prostanoids (6). COX-2 gene expression is regulated at the level of transcription as well as post-transcriptionally. The regulation of the COX-2 promoter in macrophages is complex and involves different promoter elements. The transcriptional activation of the COX-2 gene depends on the stimulus-coupled binding of transcription factors (7), including NF-B, C/EBP␤, and AP-1 (8 -12). The CREelement is the most generally required promoter element, being essential for both basal and induced COX-2 transcription in a variety of cell types (13)(14)(15)(16)(17). The specific factors regulating COX-2 induction depend on the cell type as well as on the stimulus.
Considering that COX-2 is likely to play a critical role in M. avium-triggered PGE 2 release from macrophages, we set out to explore the signaling cascades and the cis-regulatory elements governing COX-2 induction from M. avium-challenged macrophages. Our results demonstrate that TLR2dependent ERK and p38 activation converge upon the activation of mitogen-and stress-activated kinase 1 (MSK1), which is instrumental in controlling both NF-B-and CRE-driven transcription of the COX-2 gene. To the best of our knowledge, this is the first demonstration of the role of MSK1 in controlling pathogen-driven gene expression of COX-2, a molecule that plays a central role in pathogenesis.
Growth of M. avium-M. avium (NCTC 8562) was obtained from the All India Institute of Medical Sciences, New Delhi), grown in Middlebrook 7H9 broth (Difco, Detroit, MI) supplemented with oleic acidalbumin-dextrose supplement and 0.04% Tween 80 until mid-log phase, harvested, washed, and resuspended in a small volume of phosphatebuffered saline (PBS) containing 0.04% (v/v) Tween 80. The suspension was briefly sonicated until no bacterial clumps were visualized by microscopy. The preparation was diluted to a concentration of 2 ϫ 10 8 colony-forming units/ml in 10% glycerol and stored in aliquots at Ϫ80°C. Freshly thawed aliquots were used no more than once. Where indicated, cells were incubated with polymyxin B (10 g/ml) for 30 min prior to use.
Cell culture-The murine macrophage cell line RAW264.7 was obtained from the National Centre for Cell Science, Pune, India, and cultured in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum, 2 mM glutamine, and antibiotics. Where indicated, RAW264.7 cells were challenged with M. avium at a multiplicity of infection (m.o.i.) of 2.
Plasmid constructs-Using genomic DNA from RAW264.7 cells as template, the COX-2 promoter region (Ϫ891 to ϩ9) was amplified by PCR using the primer pair 5Ј-CGGGGTACCCAAACACTGTTTCT-GAAT-3Ј (sense, primer a) and 5Ј-ATAAAGCTTTTTGACAACTGGCT-GCTA-3Ј (antisense, primer b). The resulting COX-2 promoter region was cloned into the vector pGL2 Basic (Promega) harboring the promoter-less luciferase gene, using asymmetric KpnI and Hind III sites (underlined) to generate the plasmid pCOX301. NF-B(Ϫ402/Ϫ392), C/EBP␤(Ϫ135/Ϫ130), AP-1(Ϫ65/Ϫ60), or CRE(Ϫ60/Ϫ56) binding sites were mutated by deleting the underlined bases from the respective sites as indicated: 5Ј-GGGGATTCC-3Ј (for NF-B), 5Ј-TCTTGCGCAA-3Ј (for C/EBP␤), 5Ј-ACAGAGTCAC-3Ј (for AP1) and 5Ј-CACTACGTCA-3Ј(for CRE). Mutants were generated by overlap extension PCR. The primers used are depicted in Table I. The initial rounds of PCR were carried out using the primer pairs a and b for one reaction, c and d for the second reaction, and pCOX301 as template. The products of each PCR were purified and used as templates for the second round of PCR using the primers a and d. The final products were cloned between the KpnI and HindIII sites of pGL2 Basic. Vectors for wild type and dominant-negative Ras (Ras S17N); wild type and dominant-negative (S32A, S36A) IB␣ in pCMV were purchased from BD Biosciences. For generating FLAG-tagged Ras expression vectors, the respective genes were excised from the parent vectors described above using EcoRI and BamHI and cloned between the EcoRI and BamHI sites in pFLAG-CMV6c. Wild type and dominant-negative (amino acid residues 289 -522) TRAF6 in pRK-FLAG were gifts from Tularik Inc. Wild type and dominant-negative (K63W) HA-TAK1 were gifts from Dr. Kuni Matsumoto, Nagoya University, Japan; wild type and dominant-negative (D195A) MSK1 in pCMV-FLAG were gifts from Dr. Dario Alessi, University of Dundee, UK. Wild type TLR4 (mTLR4) in pEF-Bos was a gift from Dr. Yoshinori Nagai of the University of Tokyo, Japan. Dominant-negative mTLR4 (P712H) was generated by site-directed mutagenesis based on the principle of overlap extension PCR. Using pEF-BOS-TLR4 as template, the first round of PCR was performed with primer pairs a and b for one reaction and primer pairs c and d for the second reaction. The second round of PCR was performed using the products of the first round as templates, and primers a and d. Primers a, b, c, and d were 5Ј-ACAG-AATTCTCAGCGTTCTTATCCCTT-3Ј, 5Ј-GACTTTATTCATGGTGTA-GCC-3Ј, 5Ј-GGCTACACCATGAATAAAGTC-3Ј, and 5Ј-CGCGGATCC-GGTCCAAGTTGCCGTTTC-3Ј, respectively. The resulting 1.2-kb PCR product was cloned in pUC19 between the EcoRI and BamHI sites, sequenced on both strands, and an Xho-BamHI (468 bp) fragment containing the mutated TLR4 gene encoding amino acid residues 679 -835 was then excised and used to replace the Xho-BamHI fragment of wild type TLR4 in pEF-BOS. p4NF-B-luc containing four NF-B sites in tandem fused to the luciferase reporter gene, was a gift from Dr. A. Brent Carter of the University of Iowa. The p6CRE-luc construct containing 6 CRE sites in tandem fused to the luciferase reporter gene was a gift from Dr. Steve Rees of GlaxoSmithKline. pGEX/Raf-RBD was a gift from Dr. Robert J. Sheaff of the University of Minnesota. cDNA products for murine TLR2 and CREB were obtained by reverse transcription-PCR from a total RNA preparation from RAW264.7 cells using the primer pairs: 5Ј-CTAGCTAGCTAGATGCTACGAGCTCTT-3Ј (sense) and 5Ј-CCGCTCGAGCGGCTAGGCCTTTATTGC-3Ј (antisense) for TLR2; and 5Ј-ATAAAGCTTATGGAATCTGGA-3Ј (sense) and 5Ј-A-TGGATCCATCTGATTTGTGGCAGTA-3Ј (antisense) for CREB. The resulting PCR products were cloned in pcDNA3.1 (between NheI and XhoI sites) for TLR2 and pFLAG-CMV-6a (between HindIII and BamHI sites) for CREB. For generating dominant-negative dominant TLR2 (carrying a 13-amino acid deletion from the C terminus), a 2313-bp fragment was amplified by PCR using the sense primer mentioned above and the antisense primer: 5Ј-CCGCTCGAGCGGCTGGCCT-TCATC-3Ј. The resulting PCR product was cloned in pcDNA 3.1 between NheI and XhoI sites. Site-directed mutagenesis was performed to generate dominant-negative CREB (S133A). Primers b and c were 5Ј-AGGAGGCCTGCCTACAGGAAA-3Ј and 5Ј-TTTCCTGTAGGCAG-GCCTCCT-3Ј, respectively. All constructs were checked by sequencing on both strands.
Enzyme Immunoassay for PGE 2 -Supernatants from mock-treated and M. avium-treated macrophages were harvested at the indicated times and tested for the presence of PGE 2 using the BIOTRAK PGE 2 assay kit from Amersham Biosciences.
Western Blotting-Proteins were separated by SDS-PAGE and then transferred electrophoretically to polyvinylidene difluoride membranes. The blots were blocked with 5% nonfat dry milk in Tris-buffered saline (TBS) for 1 h at room temperature and subsequently incubated overnight at 4°C with primary antibodies in TBS-Tween 20 (1%, v/v) (TBST) with 5% (w/v) bovine serum albumin. Following three washes of 5 min each with TBST, the blots were incubated with horse radish peroxidase-conjugated goat anti-rabbit IgG (Santa Cruz Biotechnology, Inc.) (or appropriate secondary antibodies) in blocking buffer for 1 h at room temperature. After three washes with TBST, the blots were developed by chemiluminescence using the Phototope-HRP Western blot detection kit (Cell Signaling Technology) and exposed to x-ray film (Kodak XAR5).
RNA Isolation and Reverse Transcription-PCR-Total RNA was prepared from cells using the RNeasy Mini kit (Qiagen) according to the manufacturer's protocol. 100 ng of RNA was reverse-transcribed using the Titan 1-tube reverse transcription-PCR Kit (Roche Applied Sciences). The sense primer 5Ј-GGGTTGCTGGGGGAAATGTG-3Ј, and the antisense primer 5Ј-GGTGGCTGTTTTGGTAGGTG-3Ј, were used to amplify 479 bp of COX-2 mRNA. Glyceraldehyde-3-phosphate dehydrogenase was amplified using the primers: 5Ј-CCA TCA ATG ACC CCT TCA TTG ACC-3Ј (sense), and 5Ј-GAA GGC CAT GCC AGT GAG CTT CC-3Ј (antisense) to generate a 604-bp product. The PCR conditions for COX-2 mRNA were denaturation at 94°C for 30 s, annealing at 50°C for 1 min, and extension at 68°C or 1 min for 35 cycles.
Electrophoretic Mobility Shift Assay-To prepare nuclear extracts, plates with adhered cells were washed twice with ice-cold PBS and scraped in ice-cold TNE buffer (40 mM Tris HCl, pH 7.5, 0.1 M NaCl, 1 mM EDTA). Cells were pelleted at 600 ϫ g for 1 min, and pellets were resuspended in 400 l of lysis buffer (10 mM HEPES, pH 7.9, 10 mM KCl, 0.1 mM EDTA, 0.1 mM EGTA, 1 mM dithiothreitol, 0.5 mM Pefabloc, 2 g/ml leupeptin, 2 g/ml aprotinin, 0.5 mg/ml benzamidine). Cells were lysed by the addition of 23 l of 10% (v/v) Nonidet P-40 (Nonidet P-40) and the nuclear pellet was suspended in 25 l of ice-cold extraction buffer (20 mM HEPES, pH 7.9, 0.4 M NaCl, 1 mM EDTA, 1 mM EGTA, 1 mM dithiothreitol, 0.5 mM Pefabloc, 2 g/ml leupeptin, 2 g/ml aprotinin, 0.5 mg/ml benzamidine), kept on ice for 30 min with occasional mixing, and centrifuged at 10, 000 ϫ g for 5 min at 4°C. The supernatant (nuclear extract) was immediately frozen and stored at Ϫ70°C. Typically, 15 g of nuclear extract was incubated on ice for 5 min in the absence or presence of competitor DNA in a volume of 20 l containing 4 l of 5ϫ binding buffer (125 mM HEPES, pH 8, 2.5 mM EDTA, 2.5 mM dithiothreitol, 5% Nonidet P-40, 250 mM NaCl, 20% (v/v) glycerol), and 0.75 g of poly(dI-dC) as nonspecific competitor DNA. The 32 P-end-labeled double-stranded oligonucleotides (60,000 cpm per reac-tion) containing the COX-2 gene-specific NF-B binding site (5Ј-GGT-GAGGGGGATTCCCTTAGTTA-3Ј) (or the mutated oligonucleotide with the sites of mutation (GA 3 CC) in boldface) or the COX-2-specific CRE site (5Ј-GTCACCACTACGTCACGTGGAG-3Ј) (or the mutated oligonucleotide with the sites of mutation (ACG 3 GAT) in boldface) was added, and the mixture was incubated for 30 min at 25°C. The DNAprotein complex formed was separated from the oligonucleotide on a 4% native polyacrylamide gel. The gel was dried and analyzed by autoradiography. For supershift assays, nuclear extracts were incubated for 30 min at 25°C with rabbit polyclonal antibodies against p50, p65 (for NF-B), or CREB (for CRE) or no antibody before incubation with the radiolabeled probe.
Luciferase reporter assays-Cells were transfected with luciferase reporter plasmid. Transfected cells were challenged with M. avium, washed once with PBS, and scraped into luciferase lysis buffer (50 mM Tris, pH 8, 70 mM K 2 HPO 4 , 0.1% Nonidet P-40, 2 mM MgCl 2 , 1 mM dithiothreitol, 20 g/ml aprotinin, 10 g/ml pepstatin, 10 g/ml leupeptin). The lysates were rapidly mixed, and insoluble material was pelleted by centrifugation at 4°C. The supernatant was removed and stored at Ϫ80°C. For promoter activation analysis, luciferase activity Expression of COX-2 was analyzed as described in panel C. The blot is representative of results obtained in three independent experiments; panel F represents the results obtained from densitometric scanning of these gels (means Ϯ S.D.). E, cells were left untreated or pretreated with the inhibitors SB203580 (5 M) or U0126 (15 M) or a combination of U0126 and SB203580 at the same concentrations, or H89 (10 M) for 30 min followed by the treatment with M. avium at an m.o.i. of 2 for 6 h. Total RNA was isolated and reverse transcription-PCR was performed using primers specific for the COX-2 or the glyceraldehyde-3-phosphate dehydrogenase (gapdh) genes. The products were visualized on agarose gels by ethidium bromide staining. The data shown are representative of results obtained from three independent experiments. assays were performed in a luminometer, and the results were normalized for transfection efficiencies by assay of ␤-galactosidase activity.
Ras Activity Assay-Ras activity was measured by assaying for Ras proteins bound to the Ras-binding domain of Raf-1 expressed as a GST fusion protein (18). Briefly, cells (2 ϫ 10 6 ) after treatment were washed with ice-cold PBS, lysed in lysis buffer (25 mM HEPES, pH 7.5, 150 mM NaCl, 10 mM MgCl 2 , 5 mM EDTA, 25 mM NaF, 1 mM sodium orthovanadate, 10% glycerol, 10 g/ml aprotinin, and 10 g/ml leupeptin) and centrifuged. Lysate was incubated with 5 g of Ras-binding domain of Raf-1 (Raf-1 RBD) at 4°C for 3 h with gentle agitation, and the beads were washed thrice in lysis buffer. Bound Ras proteins were solubilized in Laemmli buffer and detected by SDS-PAGE followed by Western blotting using anti-Ras antibody.
Co-immunoprecipitation of Ras with TLR2-RAW 264.7 cells were transfected with FLAG-Ras expressing vector and challenged with M. avium for the indicated time periods. The cells were harvested, lysed in lysis buffer (40 mM Tris-HCl (pH 8.0), 500 mM NaCl, 0.1% Nonidet P-40, 6 mM EGTA, 10 mM sodium ␤-glycerophosphate, 10 mM NaF, 300 M sodium orthovanadate, 2 mM phenylmethylsulfonyl fluoride, 10 g/ml aprotinin, 1 g/ml leupeptin, and 1 mM dithiothreitol), and centrifuged. The supernatant was then immunoprecipitated with anti-FLAG-agarose. The immunoprecipitated beads were washed three times with lysis buffer. The samples were fractionated on 15% SDS-polyacrylamide gels, transferred to a polyvinylidene difluoride membrane, and subjected to immunoblot analysis with antibodies specific for TLR2.

M. avium Stimulates PGE
M. avium-stimulated PGE 2 Release Depends on the ERK and p38 MAPKs-PGE 2 release from M. avium-challenged cells was inhibited when cells were pretreated with pharmacological inhibitors of ERK and p38 MAPKs, U0126 and SB203580, respectively (Fig. 1A). Although other effectors such as bacterial peptidoglycan have been reported to modulate PGE 2 release by activating COX-2 expression in a phosphatidylinositol 3-kinase-dependent manner, the phosphatidylinositol 3-kinase inhibitor wortmannin had no effect on PGE 2 release in the present instance. The ERK and p38 MAPKs appeared to regu- late PGE 2 release in a synergistic manner, because a combination of the inhibitors U0126 and SB203580 led to a complete inhibition of PGE 2 release (Fig. 1A). Besides their ability to phosphorylate and activate transcription factors, both ERK and p38 MAPK also transduce signals by activating downstream kinases such as mitogen-and stress-activated protein kinase 1 (MSK1), which in turn phosphorylates transcription factors. Cells pretreated with H89, a pharmacological inhibitor of MSK1, at the dose used led to a complete inhibition of PGE 2 release (Fig. 1A), suggesting a role of MSK1 in PGE 2 release.
M. avium Stimulates COX-2 Expression-COX-2 converts arachidonic acid released from the cell lipids by the action of phospholipase A 2 to prostaglandin H2, the common precursor of all prostaglandins. To examine whether elevated PGE 2 synthesis is associated with increased COX-2 expression we studied the M. avium-induced expression of COX-2. Western analysis showed that M. avium challenge resulted in a timedependent increase in the expression of COX-2 in RAW264.7 (Fig. 1C) COX-2 expression is regulated by, among others, transcriptional mechanisms (7-12). We therefore analyzed the expression of the COX-2 gene in M. avium-challenged macrophages by reverse transcription-PCR. Compared with untreated cells, the steady-state COX-2 mRNA levels increased in cells treated with M. avium (Fig. 1E). Taken together, our results on activation of COX-2 at the level of the gene and the protein in cells challenged with M. avium suggested that this regulation is likely to be critical in controlling PGE 2 release from these cells.
M. avium-induced COX-2 Expression Is Dependent on the MAPKs-Considering the role of MAPK inhibitors in inhibiting PGE 2 release, we tested the role of MAPKs in M. aviuminduced COX-2 expression. As shown in Fig. 1 (D and F), M. avium-induced COX-2 expression could be attenuated partially either by U0126 or by SB203580 alone, suggesting a role of ERK as well as p38 MAPKs in M. avium-triggered COX-2 expression. As in the case of PGE 2 release, COX-2 expression was completely inhibited either by a combination of U0126 and SB203580, or by H89. The levels of COX-2 mRNA were also similarly regulated (Fig. 1E). These results suggested the possibility that COX-2 gene expression is coordinately regulated by ERK and p38 MAPKs regulating MSK1. This possibility has been tested in the following sections.

Ras Is Involved in M. avium-induced COX-2 Expression-
The small GTP-binding protein Ras is an important signal mediator in response to stimuli such as growth factors, cytokines, and hormones (22). In resting cells, Ras is maintained in . For panels C and D, RAW264.7 macrophages were transfected with empty vector (pcDNA3.1, pRK5, and pCMV for TLR2, TRAF6, and TAK1 constructs, respectively) or dominant-negative constructs followed by treatment with M. avium for 20 min. Cells were washed, lysed, and immunoblotted with phospho-ERK (C) or with phospho p38 (D) as described above.
an inactivated Ras-GDP form. After activation, Ras is converted into active Ras-GTP. Ras exerts its function through activating downstream effectors, among which the best studied is the MAPK cascade (23). Considering the role of ERK and p38 MAPKs in regulating PGE 2 release as well as COX-2 expression in response to M. avium challenge, we tested the effect of the Ras inhibitor manumycin A (24) on M. avium-induced COX-2 expression. Cells were pretreated with manumycin A prior to challenge with M. avium. Manumycin A partially inhibited M. avium-induced COX-2 expression (Fig. 1, D and F). To further confirm the role of Ras in M. avium-induced signaling in RAW264.7 cells, the activation of Ras was assessed by an activation assay relying on the interaction of activated Ras with Raf-1-RBD (18). Upon M. avium challenge, Ras activity increased in a time-dependent manner (Fig. 2, A and B). Taken together, these results suggested that M. avium-induced COX-2 expression depends on Ras activation.
TLR2 Associates with Ras in M. avium-stimulated Macrophages-Because M. avium-mediated PGE 2 release is inhibited by dominant-negative TLR2, we tested the hypothesis that rapid Ras activation could be linked to TLR2 signaling. It was observed that M. avium-induced Ras activation was blocked in cells transfected with dominant-negative TLR2 (Fig. 2, C and  D). To test whether rapid Ras activation is concurrent with Ras-TLR2 interaction, RAW264.7 cells were transfected with FLAG-tagged Ras and challenged with M. avium, and lysates were prepared and immunoprecipitated with anti-FLAG antibody and immunoblotted for the presence of TLR2 in the immunoprecipitated complex. The association of Ras and TLR2 occurred in a time-dependent manner (Fig. 2E). The immunoprecipitate showed no band corresponding to TLR2 when it was blotted with an irrelevant antibody. Control experiments were performed in which an irrelevant FLAG-tagged protein was transfected prior to challenge with M. avium. Co-immunoprecipitation of TLR2 with anti-FLAG antibody was not visible in this case. Taken together, all these results suggested that M. avium-induced Ras activation occurs through interaction of Ras with TLR2.
M. avium-induced ERK Activation Depends on TLR2/Ras Signaling-Our previous studies have already demonstrated that M. avium activates ERK and p38 MAPKs in macrophages (25). In the present study we observed that pretreatment with manumycin A (Fig. 3, A and B) inhibited the phosphorylation of ERK in a dose-dependent manner. This suggested that ERK was downstream of Ras signaling. Further, transfection of cells with dominant-negative TLR2 or dominant-negative Ras led to an inhibition of M. avium-induced ERK phosphorylation (Fig.  3C), supporting the view that M. avium activates ERK in a TLR2/Ras-dependent signaling pathway.
M. avium-induced p38 MAPK Activation Depends on TLR2/ TRAF6 Signaling-M. avium-induced p38 MAPK activation was inhibited by transfecting cells with dominant-negative constructs of TLR2, TRAF6, or TAK1 (Fig. 3D). TAK1 is known to signal downstream of the TLRs to activate p38 MAPK in response to a variety of stimuli (26,27). Because TLR2/TRAF6/ TAK1 signaling is a well established pathway of MAPK activation, we felt it safe to predict that this pathway controls the activation of p38 MAPK in the present instance also.
M. avium Stimulates COX-2 Promoter Activity-COX-2 expression is controlled by both transcriptional and post-transcriptional mechanisms. To determine whether M. avium regulates COX-2 promoter activity, RAW 264.7 cells were transfected with different constructs of the mouse COX-2 promoter coupled with the luciferase reporter gene. M. avium evoked at least a 10-fold increase in luciferase activity in cells transiently expressing COX-2 promoter (Ϫ891 to ϩ 9)-luc com- pared with untreated cultures (Fig. 4A). To rule out the effect of contaminating lipopolysaccharide (LPS) on COX-2 promoteractivated gene expression, M. avium cells were first incubated with polymyxin B before challenge of macrophages. The expression of luciferase was unchanged using polymyxin B-treated as compared with untreated M. avium (Fig. 5A).

The NF-B and CRE Sites Are Essential for Optimal COX-2 Promoter-induced Reporter Induction in M. avium-stimulated
Macrophages-In macrophages, several cis-acting elements have been found to play a role in COX-2 gene transcription (7)(8)(9)(10)(11)(12). To assess the role of the potential cis-acting elements in M. avium-induced COX-2 gene transcription, the NF-B, CRE, C/EBP␤, and AP-1 elements were individually mutated (as depicted in Fig. 4A) to evaluate the effects of such mutations on M. avium-stimulated COX-2 promoter activity. The COX-2 reporter constructs were transiently transfected into RAW264.7 macrophages. Cells were subsequently challenged with M. avium for 6 h, harvested, and lysed for luciferase activity reporter activity induced by M. avium. Our results demonstrate that the constructs mutated at the AP-1 and C/EBP␤ sites were as effective in stimulating luciferase gene expression as the wild-type COX-2 promoter (Fig. 4A). However, there was partial inhibition of luciferase gene expression when either the NF-B or the CRE site was mutated, suggesting that M. aviuminduced COX-2 promoter activation depends on the NF-B as well as the CRE elements. Luciferase expression was completely inhibited when both the NF-B and the CRE sites were mutated (Fig. 4A).
IB-␣ Phosphorylation Is Necessary for M. avium-driven COX-2 Promoter Activation-In most cell types, NF-B is maintained in an inactive form in the cytoplasm by association with IBs. In the better known mechanisms of NF-B activation, TLR signaling leads to the IB kinase-dependent phosphorylation of IB-␣, its ubiquitination and degradation by the proteasome, accompanied by the release and subsequent translocation of NF-B into the nucleus (28). The likely involvement of this classic pathway in M. avium-induced COX-2 promoter activation was tested. Transfection with dominant-negative TLR2 led to an inhibition of M. avium-stimulated luciferase activity (Fig. 5B). However, transfection with dominant-negative TLR4 (P712H) had no effect on M. avium-stimulated luciferase activity (Fig.  5B). The behavior of TLR4 (P712H) as dominant-negative was validated by the observation that this mutant could inhibit LPSdriven luciferase activation (Fig. 5E), corroborating the observa- . In addition to the constructs mentioned above, cells were co-transfected (B and E) with empty vectors or constructs as indicated; followed by challenge with M. avium (B) or with E. coli LPS (E) for 6 h. C and D, cells were transfected with the 4ϫ B-luciferase construct (C) or the 6ϫ CRE-luciferase construct (D) instead of the COX-2 promoter-luciferase construct. Transfected cells were treated with inhibitors (C and D) as described above. Cells were then left untreated or incubated with E. coli LPS or with M. avium (as indicated in the panels) for 6 h. Following incubations, luciferase reporter activities in the cell extracts were measured using luciferin as substrate. In each case, the level of induction of luciferase activity was compared with that of untreated cells. The activities shown represent data that have been normalized with ␤-galactosidase activity. Results represent the means Ϯ S.D. of three independent experiments. tions of Rhee and Hwang (29). Taken together, these results suggested that M. avium-mediated COX-2 gene expression is dependent on TLR2 but not on TLR4. The role of IB-␣ phosphorylation was tested by transfecting cells with a mutated IB-␣(S32A/S36A) construct in which the two critical serine residues, which are phosphorylated by the IB kinases, have been mutated; along with the COX-2 promoter-luciferase construct. Expression of the mutant IB␣ protein partially blocked M. aviumdependent activation of the COX-2 luc reporter (Fig. 5B), indicating that the mutant IB␣ protein was an effective repressor of COX-2 promoter-driven reporter expression. Considering that the phosphorylation and proteasome-mediated degradation of IB␣ is the classic pathway of cytoplasmic NF-B activation, the partial inhibition by pCMV-IB-␣(S32A/S36A) was expected and supported by our observations that transfection with the mutant IB-␣(S32A/S36A) construct inhibited M. avium-induced 4x B-driven luciferase expression (Fig. 5E). Canonical NF-B activation was therefore required for M. avium-induced COX-2 reporter activity.
There are results both in favor and against a role of NF-B in the stimulation of COX-2 gene expression. Whereas, NF-B activates a COX-2 promoter construct in IL-1␤-stimulated pulmonary type II A549 cells (30), it plays no role in COX-2 induction in tumor necrosis factor-␣-stimulated rat vascular smooth muscle cells (31). Our present studies definitively show that M. aviumstimulated gene expression driven by the COX-2 promoter depends on the NF-B element on the COX-2 promoter.
We also examined whether the NF-B activation correlates with differences in NF-B-specific DNA binding. Nuclear extracts were prepared from untreated cells as well as cells treated for 3.5 h with M. avium. Extracts from M. aviumtreated cells showed binding to an oligonucleotide sequence corresponding to the NF-B-binding site of the COX-2 promoter (Fig. 4B), and DNA binding was inhibited by excess unlabeled oligonucleotide. No binding was observed when a mutant NF-B-binding oligonucleotide sequence was used (data not shown). Supershift of the band obtained with the NF-B-specific oligonucleotide with anti-p65 or anti-p50 (Fig. 4B) suggested that p65 and p50 are the components of the protein complex that binds to the NF-B element of the COX-2 promoter.
The MAPKs ERK and p38 Regulate M. avium-induced COX-2 Promoter Activation-To analyze the effect of ERK and p38 MAPK activation on COX-2 promoter-driven luciferase expression in M. avium-challenged RAW264.7, cells were pretreated with ERK and p38 MAPK-specific inhibitors U0126 and SB203580, respectively. Both these inhibitors caused a partial inhibition of luciferase expression (Fig. 5A). However, a combination of both inhibitors completely inhibited luciferase expression, suggesting that ERK and p38 MAPK activation are critical in sustaining M. avium-driven COX-2 promoter activation.
MSK1 Is Activated in M. avium-stimulated Macrophages-The MAPKs phosphorylate several downstream kinases, one of which is MSK1, which is constitutively localized in the nucleus where it regulates transcription by phosphorylating histone H3, p65/RelA, and CBP/p300 (32,33) leading to modification of nucleosome conformation and increased accessibility of the transcription machinery to promoters. ERK and p38 are both known to activate MSK1 (34). To determine whether the effects of ERK and p38 MAPKs are transmitted through the activation of MSK1, we first investigated the ability of M. avium to trigger the activation of MSK1. MSK1 kinase activity was determined in M. avium-treated macrophages by immunoprecipitating MSK1 from cell lysates and assaying the ability of the immunoprecipitates to phosphorylate the peptide: EILSRRPSYRK (CREBTIDE) (34). M avium activated MSK1 in a time-dependent manner (Fig. 6A). MSK1 activity could be partially inhibited by pretreatment with either U0126 or SB203580 (Fig. 6B), suggesting a role of both ERK and p38 MAPK signaling in MSK1 activation. A combination of U0126 and SB203580 completely blocked MSK1 activation (Fig. 6B), suggesting that ERK and p38 MAPK act synergistically to activate MSK1. The finding that both kinases are required for MSK1 activity raises the possibility that they may phosphorylate non-overlapping sites that affect the kinase activity of MSK1, or that the two kinases regulate each other, or that the activity of neither of the two kinases is sufficient to fully phosphorylate MSK1. Whatever the mechanism, our findings suggest that MSK1 plays an essential role functioning as a downstream effector of M. avium signaling controlling both NF-B-and CRE-driven COX-2 promoter activation. The following sections detail our attempts to address this question.
CREB Plays a Role in M. avium-induced COX-2 Promoter Activation-CREB, a member of the ATF family of transcription factors, binds to the CRE element. Because mutation of the COX-2 CRE site inhibits COX-2 reporter activity, we explored CREB Is Phosphorylated in M. avium-stimulated Macrophages-CREB is a target of MSK1 (35). We examined whether CREB phosphorylation is differentially regulated in untreated and M. avium-challenged macrophages and whether MSK1 regulates this process. To test this, cells were transfected with wild type or dominant-negative MSK1 and challenged with M. avium. Lysates from M. avium-challenged and untreated macrophages were probed with an antibody that specifically recognizes CREB phosphorylated at Ser-133. Our results showed that CREB phosphorylation was enhanced in M. avium-challenged macrophages and that this was abrogated in cells transfected with dominant-negative MSK1 (Fig. 6C).
We also examined whether the enhanced CREB phosphorylation correlated with differences in CREB-specific DNA binding. Nuclear extracts were prepared from untreated cells as well as cells treated for 3.5 h with M. avium. Only extracts from M. avium-treated cells showed binding to an oligonucleotide sequence corresponding to the CREB-binding site of the COX-2 promoter (Fig. 4C). DNA binding was inhibited by excess unlabeled oligonucleotide. No binding was observed when a mutant CREB-binding oligonucleotide sequence was used (data not shown). Anti-CREB antibody supershifted the band obtained with the CREB-specific DNA probe suggesting that CREB binds to the CRE element of the COX-2 promoter.
MSK1 Is Necessary For Regulating COX-2 Promoter-driven Gene Expression in M. avium-challenged Macrophages-To assess the effects of MSK1 on COX-2-promoter-driven luciferase expression in M. avium-treated macrophages, cells pretreated with H89 (Fig. 5A) or transfected with dominant-negative MSK1 (Fig. 5B) were subsequently challenged with M. avium, and luciferase gene expression was assayed. In each case, suppression of MSK1 activity led to almost complete inhibition of luciferase expression, suggesting that MSK1 plays an important role in regulating activation of the COX-2 promoter following M. avium challenge. We further analyzed the role of M. avium-mediated MSK1 activation on NF-B and CREdriven luciferase expression by transfecting cells with a luciferase reporter gene fused downstream of either a 4ϫ B element or a 6ϫ CRE element. M. avium could activate both NF-B-driven (Fig. 5C) and CRE-driven (Fig. 5D) luciferase expression in these cells. However, such activation was abrogated in cells in which MSK1 activity was inhibited by treatment with H89. Transfection with dominant-negative MSK1 gave similar results (data not shown). These results definitively demonstrated that MSK1 regulates both NF-B-and CRE-driven gene expression triggered by M. avium. In this regard, the effect of M. avium was different from that observed in the case of LPS. We observed that Escherichia coli LPSstimulated NF-B-and CRE-driven luciferase expression (Fig.  5, C and D). The latter effect could be inhibited in cells pretreated with H89 (Fig. 5D). However, NF-B-driven luciferase expression could not be inhibited by pretreatment with U0126 or SB203580 or H89 (Fig. 5C).

DISCUSSION
High PGE 2 concentrations during mycobacterial infection cause down-regulation of cell-mediated immunity permitting disease progression (36). The immunosuppressive effects of prostaglandin E 2 in relation to mycobacterial infection underscore the need to understand the mechanisms regulating the expression of the rate-limiting enzyme in prostaglandin synthesis, COX-2. Animal models of tuberculosis have demonstrated that blocking of COX-2 produced high inflammation and expression of tumor necrosis factor-␣ and interferon-␥. Although previous reports have shown that virulent M. avium stimulates the expression of COX-2, the signaling pathways regulating COX-2 expression remain to be studied. The data presented in this report focus on the role of signaling pathways in M. avium-induced COX-2 gene transcription. Our data demonstrate that M. avium-induced gene expression driven by the COX-2 promoter depends on TLR2-dependent signaling converging on the NF-B and CRE-elements present in the promoter, but not on the C/EBP␤ or the AP-1 elements. Mutation at either the NF-B-binding site or at the CRE-binding site led to a partial inhibition of M. avium-induced COX-2 promoterdriven luciferase gene expression (Fig. 4A), whereas mutation at both these sites completely inhibited luciferase expression. At the same time, nuclear extracts from M. avium-challenged macrophages effected electrophoretic mobility shifts of oligonucleotides corresponding to the COX-2-gene specific NF-B and CRE sites (Fig. 4, B and C). Transcription factor NF-B has been shown to induce macrophage COX-2 in response to a variety of stimuli (37). Rhee and Hwang (29) have demonstrated that LPS triggers TLR4-mediated NF-B activation leading to COX-2 gene expression. Our studies demonstrate that TLR2 contributes to M. avium-induced COX-2 gene induction (Fig. 5). Suppressor IB-␣ transfection led to inhibition of COX-2 promoter activity as well as consensus NF-B-driven promoter activity (Fig. 5, B and E). These results supported a role of the canonical NF-B activation in M. avium-stimulated COX-2 promoter activation. M. avium-induced canonical NF-B activation was also observed by Giri et al. (38). In addition to the classic pathway, TLRs are known to activate downstream signaling cascades through small G proteins such as Rac and Ras to mediate NF-B activation. We observed that M. avium challenge led to association of Ras with TLR2 and a rapid activation of Ras (Fig. 2). COX-2 promoter activation, as well as activation of the consensus NF-B promoter relied on TLR2/Ras-dependent ERK activation (Fig. 3A). In the classic Toll signaling pathway, TLRs upon recognizing suitable ligands recruit MyD88/IRAK/TRAF6. This leads to the activation of MAPK kinase kinases such as TAK1. M. avium triggered TLR2/TRAF6/TAK1-dependent p38 MAPK activation (Fig. 3D). Both ERK and p38 MAPK were responsible for the activation of the nuclear kinase MSK1. A combination of pharmacological inhibitors of ERK and p38 completely inhibited MSK1 activation (Fig. 6B). Inhibitors of either ERK or p38 (U0126 or SB203580, respectively) partially inhibited NF-Bdriven luciferase expression, whereas H89 could block luciferase expression completely (Fig. 5C). These results clearly demonstrated that MSK1 plays an important role in M. aviumtriggered NF-B activation. This effect of M. avium on RAW264.7 macrophages was different from that reported in the case of LPS-driven COX-2 expression in RAW264.7, where LPS-driven NF-B activation has been shown to occur independent of ERK (39) or MSK1 (8) activation. Previous work by Deak et al. (34) demonstrated that the transcription factors CREB and ATF1 are physiological substrates of MSK1. In the present study we also observed that M. avium induced phosphorylation of CREB in an MSK1-dependent manner (Fig. 6C).
In summary, M. avium triggers TLR2-dependent signaling leading to the phosphorylation of ERK and p38 MAPKs, which converge to activate MSK 1 (Fig. 7). MSK1 regulates both the NF-B-and the CRE-driven gene transcription. Both phosphorylated CREB and NF-B are known to recruit histone acetylases such as CBP/p300, which in turn contribute to making the promoter more accessible to transcription factors and help in bridging the transcription factor-CBP complexes to components of the basal transcription machinery (40). MSK1 phosphorylates histone H3 (41) and has also been identified as a nuclear kinase for p65, phosphorylating p65 at serine 276 (42). Our results point to the fact that MSK1-dependent changes, likely at the level of the nucleosome, are prerequisite for M. avium-driven COX-2 gene expression. The nature of nucleosome remodeling following M. avium infection is presently under investigation.