Characterization of an Activation Protein-1-binding Site in the Murine Interleukin-12 p40 Promoter

Interleukin (IL)-12 is a heterodimeric cytokine produced by macrophages in response to intracellular pathogens and provides an obligatory signal for the differentiation of T-helper-1 cells. We previously reported an analysis of the IL-12 p40 promoter in RAW264.7 macrophages. Multiple control elements were involved in activation of transcription by bacterial products. A critical control element, located between −96 and −88, interacts with C/EBP family members. In this study, using a strategy to demonstrate functional activity in a minimal promoter context, three novel cis-acting elements are found to have an important role in IL-12 p40 promoter activation by lipopolysaccharide. One of these elements is characterized in detail. Mutations from −79 to −74 in the murine IL-12 p40 promoter significantly reduce lipopolysaccharide-induced promoter activity. Electrophoretic mobility shift assays demonstrate binding of AP-1 family members to this region. Spacing between the C/EBP and AP-1 site is important for promoter activation, suggesting cooperativity between these elements. c-Jun and a mutant c-Jun molecule activate the IL-12 p40 promoter and synergistically activate the promoter when co-expressed with C/EBPβ. Finally, this region of the promoter is demonstrated to be a target for mitogen-activated protein kinase and toll-like receptor signaling pathways.

IL-12 1 is a heterodimeric cytokine produced by macrophages and dendritic cells. Among its biologic activities, IL-12 production is necessary for the differentiation of T-helper-1 (Th1) cells and for the secretion of the Th1 cytokine, interferon-␥ (IFN-␥) (1). Accordingly, IL-12 and Th1 cells are required for cellmediated immunity against intracellular microbes (2)(3)(4). The protective role of IL-12 in human infectious diseases including leprosy (5), tuberculosis (6), and leishmaniasis (7) has been well characterized. Conversely, overexpression of Th1 cytokines and IL-12 may contribute to the development of chronic inflammatory disorders (8), including Crohn's disease and rheumatoid arthritis. Thus, the regulated expression of IL-12 in antigen-presenting cells is a critical event in the pathogenesis of infectious and inflammatory diseases.
IL-12 is composed of two covalently linked glycosylated chains, p40 and p35, that are encoded by separate genes and together form the biologically active p70 heterodimer (9 -11). The p35 gene is constitutively expressed in many tissue types (9). p40 mRNA is detected in macrophages and other cells that produce bioactive IL-12 (9) and is strongly induced by intracellular bacteria and bacterial products (12). Therefore, studies of IL-12 transcriptional regulation have focused on the p40 gene. Murphy and colleagues (13) identified an NF-B site between Ϫ122 and Ϫ132 relative to the transcription start site in the murine p40 gene that was necessary for induction of promoter activity by LPS. In another series of investigations, Ma and colleagues (14 -16) have identified an Ets protein DNA-binding sequence between Ϫ212 and Ϫ207 in the human p40 gene that was implicated in promoter activation by LPS. We have reported a comprehensive functional analysis of the murine and human IL-12 p40 promoters in RAW264.7 cells (17). The p40 promoter was strongly induced by heat-killed Listeria monocytogenes and LPS. An important control element defined by this analysis, located between Ϫ96 and Ϫ88 relative to the murine transcription start site, interacts with C/EBP proteins. This element was functionally synergistic with the NF-B site. However, mutations in several other elements had functional effects on promoter activation. Thus, induction of the IL-12 p40 promoter by bacterial products will be defined by a complex series of events.
In this study, a functional role for three novel cis-acting elements in the p40 promoter is demonstrated in reporter assays by utilizing a combination of multiple substitution and deletion mutants. One of these elements is characterized in detail. Mutations from Ϫ79 to Ϫ74 in the murine IL-12 p40 promoter significantly reduce LPS-induced promoter activation and correlate with binding of AP-1 family members by electrophoretic mobility shift assays. Spacing between the adjacent C/EBP and AP-1 sites and orientation of the AP-1 site are important for promoter activation, suggesting cooperativity between these two sites. Expression of a dominant negative AP-1 protein, A-Fos, in RAW264.7 cells inhibits p40 promoter activation by LPS, demonstrating a functional role for AP-1 family members. Furthermore, c-Jun and a mutant c-Jun molecule TAM67 activate the IL-12 p40 promoter and synergistically activate the promoter when co-expressed with C/EBP␤. Like-wise, c-Jun and TAM67 activate the tumor necrosis factor-␣ (TNF) promoter in RAW264.7 cells. Finally, experiments using inhibitors and activators of signal transduction pathways identify the region of the promoter from Ϫ101 to ϩ55 as a target for the p38 kinase, the N-terminal Jun kinase (JNK), and the toll-like receptor (TLR) signal transduction pathways.

EXPERIMENTAL PROCEDURES
Plasmid Constructs-The murine wild type IL-12 p40 promoter Ϫ350 to ϩ55 region was inserted into the chloramphenicol acetyltransferase (CAT) reporter vector pCAT basic (Promega) modified by addition of KpnI and BglII sites in the polylinker as described previously (17). All site-directed mutant plasmids described in this paper were generated by two-step polymerase chain reaction using overlapping internal primers containing a mutant site (18). Polymerase chain reaction-generated mutant promoters were cloned into pCAT basic vector (Promega) with a modified polylinker or pGL2B vector (Promega) and sequenced before use in transfection experiments. Mammalian expression plasmids for human wild type c-Jun, c-Fos, JunB (Tom Curran), TAM67 (Michael Birrer), and C/EBP␤ (from Steven McKnight) were described previously (17,19). An expression plasmid for the AP-1 dominant negative protein A-Fos was obtained from Charles Vinson (see Ref. 20 (27). As MAP kinase pathway controls, MEF2c-GAL4 was obtained from Jiahuai Han. pFA-Jun, and pFR-luciferase were from Stratagene. A multimerized NF-B element luciferase reporter plasmid and a human TNF-␣ promoter (Ϫ283 to ϩ113) luciferase reporter plasmid were obtained from Adrian Ting. A multimerized consensus AP-1 DNA-binding element CAT reporter gene (TRE-CAT) was provided by Andre Nel.
Cell Lines and Reagents-The RAW264.7 murine macrophage line (American Type Culture Collection) was maintained in DMEM supplemented with 10% fetal bovine serum and penicillin/streptomycin. LPS was purchased from Sigma. L. monocytogenes was obtained from Hao Shen, and Mycobacterium tuberculosis whole cell extract was from John Belisle. Antibodies against Pu.1, C/EBP, AP1, ATF, and Rel family members were purchased from Santa Cruz Biotechnology, Inc.
Transfections-RAW267.4 cells were transiently transfected using SuperFect Transfection Reagent (Qiagen) by the described protocol with modifications. For each transfection, 2.5 g of plasmid were mixed with 100 l of DMEM (without serum and antibiotics) and 10 l of SuperFect reagent. The mixture was incubated at room temperature for 10 min, and then 600 l of DMEM complete medium was added and immediately placed onto the cells in 6-well plates. After incubation for 3-4 h at 37°C, the cells were washed with PBS and split equally into two wells. One well was activated with LPS (to final concentration 5 g/ml) 24 h later. Sixteen to eighteen hours after activation, the cells were harvested by using 1ϫ Reporter Lysis Buffer (Promega). CAT assays were performed with 75 g of total protein from cell extracts. The conversion of [ 14 C]chloramphenicol to its acetylated forms was assessed by thin layer chromatography and quantified by a Phospho-rImager (Molecular Dynamics). Luciferase activity was determined from 20 l of cell extract, as described previously (17). Cells were co-transfected with a constitutively active HSP promoter that expresses ␤-galactosidase (from Bradley Cobb) to monitor for transfection efficiency. ␤-Galactosidase assays were performed with 25 g of total protein (for CAT assays) or 20 l (luciferase assays) from cell extracts, as described previously (17).
Nuclear Extracts and DNA Binding Assays-RAW267.4 cell nuclear extracts were prepared as described previously (28,29). EMSA probes were made by annealing equal amount of single-stranded oligonucleotides with 5Ј-GATC overhangs (Genosys Biotechnologies, Inc.). 200 ng of annealed probe was labeled by filling in with [␣-32 P]dGTP and[␣-32 P]dCTP by Klenow enzyme. Labeled probes were purified with Nuc-Trap purification columns (Stratagene). Sequences of the IL-12 p40 EMSA probes are demonstrated in the text (see Fig. 4A). The sequence of the canonical AP-1 EMSA probe is 5Ј gatcCGCTTGATGACTCAGC-CGGAA, and the canonical C/EBP probe is 5Ј gatcAAGCTGCAGATT-GCGCAATCTGCAGCTT 3Ј. Electrophoretic mobility shift assays were performed as described previously (17), using 10 5 cpm probe and 5 g of nuclear extracts per reaction. In supershift experiments, 1-2 l of antibody was added to each reaction, and the nuclear extract/antibody mixture was incubated at room temperature for 30 min before adding probes. DNA-binding complexes were separated by 5% polyacrylamide/ Tris glycine/EDTA gel run at 4°C for 4 h at 150 V. Gels were dried and exposed to Kodak X-AR 5 film at Ϫ80°C in the presence of an intensifying screen.

Demonstration of Novel Functional Cis-acting Control Elements in the IL-12 p40
Promoter-In our previous analysis of the murine and human IL-12 p40 promoters, the C/EBP element spanning Ϫ93 to Ϫ88, the NF-B element at Ϫ131 to Ϫ120, and the TATA box were the only sites convincingly shown to be important for p40 promoter function (17). However, two results suggested that other functionally important elements exist within the promoter. First, deletion or substitution mutations at other locations appeared to have less significant effects on promoter activity. Second, 25% of the wild type promoter activity was retained following disruption of the NF-B element. This activity appeared to be too strong to be directed solely by a C/EBP-binding site and a TATA box. In particular, of the 21 clustered substitution mutants analyzed in this previous study, four (Ϫ114/Ϫ109, Ϫ107/Ϫ102, Ϫ80/Ϫ75, and Ϫ62/Ϫ57) were identified that demonstrated a 2-3-fold decrease in promoter activation (17). These elements, which have relatively minor effects individually, may have additive or synergistic effects when combined with mutations in other sites.
As mutations in the C/EBP site abrogate promoter activity, the functional significance of other potentially important sites (Ϫ114/Ϫ109, Ϫ107/Ϫ102, Ϫ80/Ϫ75, Ϫ62/Ϫ57) needs to be evaluated in the presence of an intact C/EBP site. The Ϫ114/Ϫ109, Ϫ107/Ϫ102, and Ϫ62/Ϫ57 substitution mutants were described previously (17). However, the Ϫ80/Ϫ75 mutant from the initial study was re-derived by changing a 13-base pair sequence from Ϫ85 to Ϫ73. This more extensive mutant may demonstrate more significant functional effects. Furthermore, as random mutants may also have artifactual effects, a different random sequence was inserted to rule out this possibility.
Site-directed mutants in two elements other than the C/EBP site were constructed in the Ϫ355 to ϩ55 wild type p40 promoter-CAT reporter (Fig. 1A). The murine macrophage cell line RAW264.7 was transiently transfected with double mutant, single mutant, and wild type plasmids. Consistent with our previously reported results, a mutation in the NF-B element (Fig. 1B, Ϫ132/Ϫ127) reduces LPS-induced promoter activation by around 80%, and a mutation in the C/EBP site (Fig. 1B, Ϫ93/Ϫ88,) virtually eliminates promoter activity. Simultaneous disruption of the NF-B element and the sequence from Ϫ85 to Ϫ73 markedly reduces promoter activation by LPS compared with the respective single mutations ( Fig. 1B; compare the double mutant [Ϫ85/Ϫ73 and Ϫ132/Ϫ127 to the single mutant Ϫ85/Ϫ73 and the single mutant Ϫ132/Ϫ127). Three other mutations (Ϫ114/Ϫ109, Ϫ107/Ϫ102, and Ϫ62/Ϫ57), when combined with mutations in the NF-B site, also demonstrate significantly diminished p40 promoter activity (Fig. 1B). A double mutant in the Ϫ114/Ϫ109 and Ϫ107/Ϫ102 elements does not attenuate promoter activation to a greater extent than the respective single mutants (Fig. 1B). Therefore, one control element likely resides within these two adjacent sequences. In summary, although each of these single mutations had a relatively small effect on promoter activation, their combination with a mutation in the NF-B site demonstrates dramatic decreases in LPS-induced promoter activity. These results identify three novel elements, from Ϫ114 to Ϫ102, Ϫ82 to Ϫ73, and Ϫ62 to Ϫ57, important for activation of IL-12 p40 transcription by LPS, and provide clear evidence that an intact C/EBP site is not sufficient for promoter activation.
As a second approach, a minimal inducible IL-12 p40 pro-moter was defined. From our previous analysis, a promoter deletion of all elements upstream of the C/EBP site mediated strong LPS-inducible promoter activity of greater than 10% of the wild type Ϫ355 to ϩ55 promoter (17). Therefore, functional control elements likely exist downstream of the C/EBP element. To test this hypothesis, site-directed mutants were created in elements downstream of the C/EBP site in a promoter from Ϫ101 to ϩ55, which deletes all elements upstream of the C/EBP element. As expected, a mutation in the C/EBP site within this minimal promoter eliminates LPS-induced activation (Fig. 2, Ϫ101 to ϩ55 and Ϫ93/Ϫ88) in RAW264.7 cells. This experiment confirms the functional importance of the element from Ϫ85 to Ϫ73, as a mutation in this site markedly reduces promoter activity (Fig. 2, Ϫ101 to ϩ55 and Ϫ82/Ϫ74). Furthermore, the Ϫ62 to Ϫ57 mutant sequence reduces promoter activity by ϳ70% (Fig. 2, Ϫ101 to ϩ55 and Ϫ62/Ϫ57). Detailed Characterization of a Functional Element within the Ϫ82 to Ϫ73 Region-By two different approaches, the sequence between Ϫ82 and Ϫ73 is demonstrated to have an important functional role in IL-12 p40 promoter activation by LPS. This control element may contain a binding site for one transcription factor or may represent a composite binding site for two or more proteins. To define better critical nucleotides within this element, 3-base pair site-directed mutants were constructed within this sequence (Fig. 3A). Two mutants, Ϫ79/Ϫ77 and Ϫ76/Ϫ74, demonstrate significantly decreased LPS-induced promoter activity, whereas mutants in the flanking sequences retain wild type activity (Fig. 3B). Thus, an important regulatory element is localized to the region spanning Ϫ79 to Ϫ74.
The DNA sequence from Ϫ81 to Ϫ73, CTAGTCAGT, has homology to a canonical AP-1 DNA-binding element (30). The fos/jun proto-oncogenes encode proteins that are major components of the AP-1 transcription factor complex. Several Fos- Two g of each single and double mutant promoter CAT reporter plasmid and 0.5 g of HSP-␤-galactosidase reporter plasmid were transfected into RAW264.7 macrophage cells. Transfected cells were incubated for 24 h and then activated with LPS (5 g/ml) for 16 -18 h. CAT activity was quantitated following thin layer chromatography by PhosphorImager analysis, and results were expressed as percent conversion of [ 14 C]chloramphenicol to acetylated substrates, normalized to ␤-galactosidase activity obtained from a co-transfected control reporter plasmid. Each result represents the mean Ϯ the S.D. of data from three to five experiments.

FIG. 2. Functionally redundant cis-acting control elements are identified downstream of the C/EBP site in the IL-12 p40 promoter using a minimal LPS-inducible promoter.
The p40 promoter-CAT reporter plasmid Ϫ101 to ϩ55 contains 5Ј-elements beginning at the C/EBP site at position Ϫ101 relative to the transcription start site. Within this minimal promoter, 6 -10-base pair substitution mutants were created between Ϫ93 and Ϫ51. The plasmids were transfected into RAW264.7 cells, incubated for 24 h, and then activated with LPS (5 g/ml) for 16 -18 h. Extracts from LPS-activated cells were analyzed by CAT assay. Results (normalized for ␤-galactosidase activity) are expressed as the percentage of the CAT activity obtained with the Ϫ101 to ϩ55 plasmid. Plasmid Ϫ101 to ϩ55 and Ϫ93/Ϫ88 represents a mutation in the C/EBP element. Two mutations in plasmid Ϫ101 to ϩ55 (Ϫ82/Ϫ74 and Ϫ62/Ϫ57) had significant effects on LPSinduced promoter activation. Each result represents the mean Ϯ S.D. of data from four to five experiments. related (c-Fos, FosB, Fra-1, and Fra-2) and Jun-related (c-Jun, JunB, and JunD) proteins have been described. These leucine zipper proteins form the transcriptionally active AP-1 complex as Fos/Jun heterodimers or Jun/Jun homodimers (31,32). Competition with an unlabeled oligonucleotide containing a consensus AP-1-binding site in 100-fold molar excess completely inhibits formation of this complex in EMSAs, strongly suggesting that AP-1 family members interact with this sequence in the IL-12 p40 promoter (Fig. 4B, lanes 4 and 12). A consensus C/EBP-binding site partially competes AP-1 DNA binding activity in gel shift experiments, suggesting that either C/EBP family members are present in this complex or that AP-1 dimers may to some extent bind to a consensus C/EBP element (Fig. 4B, lanes 3 and 11). Competition with a consensus NF-B DNA-binding site oligonucleotide does not inhibit complex formation on all three radiolabeled probes (data not shown). Mutants within the 88/63 EMSA probe, 79/77m (mutated from Ϫ79 to Ϫ77) and 76/74m (mutated from Ϫ76 to Ϫ74) eliminate binding of the AP-1 complex (Fig. 5, lanes 3-6). These specific mutants correlate with decreased promoter activity in functional assays (Fig. 3B).
The presence of AP-1 family members in this DNA-binding complex was confirmed by supershift experiments. A polyclonal antibody to the AP-1 family Fos (Fig. 6, lanes 5 and 15, K-25) strongly inhibits and supershifts the DNA-protein complex. By using antibodies that recognize specific Fos family members, it appears that the most abundant Fos DNA binding activity is attributable to c-Fos (Fig. 6, lanes 6 and 16). Antibodies to Fra-1 and Fra-2 cause minor inhibition of complex formation (Fig. 6, lane 8, 9, 18, and 19). An antibody that recognizes the Jun family member, JunB, strongly inhibits and supershifts the DNA-binding complex (Fig. 6, lanes 3 and 13). Additionally, an antibody to c-Jun causes inhibition and supershift of the complex (Fig. 6, lanes 2 and 12). Antibodies directed against the leucine zipper proteins ATF-1, ATF-2, and C/EBP␤ demon-strate little, if any, inhibition of complex formation in nuclear extracts from LPS-treated cells (data not shown). Likewise, antibodies against unrelated transcription factors Pu.1 (Fig. 6,  lanes 10 and 20), NF-B p50, NF-B p65, and the glucocorticoid receptor (data not shown) do not supershift or inhibit the DNAbinding complex. Thus, the AP-1 DNA-binding complex on the IL-12 p40 promoter is composed of JunB and c-Jun homodimers and/or heterodimers with c-Fos. However, other AP-1 family members and leucine zipper proteins may be present as well.
Determination of Spacing Requirements between Control Elements in the IL-12 p40 Promoter-Proper spacing of cis-acting elements may be necessary for the interaction of transcriptional activators with the basal transcription machinery, or it may be required for cooperative binding of transcription factors (33,34). Spacing restrictions may involve distance or alignment of sites on the DNA double helix. The functional effects of spacing between the TATA box and regulatory elements as well as between control elements within the IL-12 p40 promoter were next determined. Insertion mutants were constructed within the Ϫ355 to ϩ55 p40 promoter that increase the distance between the AP-1 and the C/EBP site by 5, 10, and 15 bp, approximating one-half, one, and one and a half helical turns of DNA, respectively (Fig. 7A). In addition, insertion mutants adding 5 and 10 bp between the C/EBP site and NF-B element and 5 and 10 bp between the C/EBP site and the TATA box were created to determine spacing requirements between C/EBP, NF-B, and the basal transcription machinery. All mutations were inserted within sequences that were not critical for promoter activation (Ϫ45/Ϫ40 for mutants between TATA and C/EBP site, Ϫ121/Ϫ117 for mutants between NF-B site and C/EBP site, and Ϫ84/Ϫ85 for C/EBP and AP-1 elements) (17). These mutants were tested by transient transfection in unactivated and LPS-activated RAW264.7 cells. Spacing mutations between the TATA box and C/EBP site have little effect on promoter activity (Ͼ80% of wild type activity; data not shown). However, mutations between the NF-B and C/EBP sites, and the C/EBP and AP-1 elements reveal significant effects on promoter induction. A 10-bp insertion between the NF-B and C/EBP elements decreases LPS-induced promoter activation by 65% (Fig. 7B). A 10-bp insertion between the C/EBP and AP-1 sites diminishes LPS-induced promoter activation by one-third, whereas a 5-and 15-bp insertion decreases activation by 50% (Fig. 7B). As random sequences inserted within the promoter may fortuitously create a proteinbinding site, the 5-and 10-bp mutants between the C/EBP and AP-1 elements were re-derived with different sequences, and similar results were demonstrated in reporter assays (data not shown).
As the effect of spacing between the C/EBP and AP-1 sites may have been minimized by redundant control elements, insertion mutants were then constructed within the Ϫ101 to ϩ55 minimal promoter. In this context, the effect of 5-, 10-, and 15-bp insertions are more dramatic, reducing promoter activity more than 60% (Fig. 7C).
Based on comparison to consensus sequences, there are two potential AP-1-binding sites within the Ϫ79 to Ϫ74 region (Fig.  3). To localize better the AP-1 DNA-binding element, and to determine whether the sequence and orientation of this site in the IL-12 p40 promoter influences activity, additional mutants were created. Within each of these two potential AP-1 sites, a canonical AP-1-binding site was created, and the orientation of the AP-1 site was reversed. From Ϫ79 to Ϫ73, a canonical AP-1-binding site was created by changing the sequence AGT-CAGT to ACTGAGT. This mutation decreases LPS-induced promoter activation by almost 75% (data not shown). A second mutant reversed the orientation from Ϫ79 to Ϫ73, by changing AGTCAGT to TGACTGA. This mutation also diminishes LPSinduced promoter activity by 75% (data not shown). From Ϫ81 to Ϫ75, a canonical AP-1 site was constructed by changing the sequence CTAGTCA to TGAGTCA. When transfected into RAW264.7 cells, this mutation strikingly increases LPS-induced promoter activation to almost 6 times wild type activity (Fig. 7D). Significantly, this mutant demonstrates a strong basal signal (20% of LPS-induced wild type activity) in unactivated cells (Fig. 7D). Next, the IL-12 p40 AP-1 site at Ϫ81 to Ϫ75 was reversed in orientation from CTAGTCA to ACTGATC. This mutation reduces LPS-induced promoter activity by 5-fold (Fig. 7D). An EMSA probe was created containing this consensus AP-1 sequence from Ϫ81 to Ϫ75 in the p40 promoter. Compared with the wild type probe 88/63 (see Fig. 4), AP-1 DNA binding activity is markedly increased in unactivated and LPS-treated nuclear extracts from RAW264.7 cells (data not shown). Supershift experiments suggest that DNA binding activities on this mutant probe are attributable to JunB, c-Fos, and c-Jun (data not shown), as demonstrated for the wild type probe. The most plausible interpretation of these results is that the IL-12 p40 AP-1 site lies within the Ϫ81 to Ϫ75 region. Hence, the consensus AP-1 site created from Ϫ79 to Ϫ73 would actually weaken the natural AP-1 element at Ϫ81 to Ϫ75. Also, orientation of this AP-1 site is important for optimal promoter FIG. 5. AP-1 DNA binding activity is eliminated by mutations that decrease IL-12 p40 promoter activation by LPS in reporter assays. Two mutant EMSA probes (79/77m and 76/74m) were created within the wild type probe 88/63. 32 P-Labeled wild type probe 88/63 (lanes 1 and 2), mutant probe 79/77m (lanes 3 and 4), and mutant probe 76/74m (lanes 5 and 6) were incubated with 0.5 g of poly(dI-dC) and 5 g of nuclear extracts from untreated (lanes 1, 3, and 5) and LPSactivated (lanes 2, 4, and 6) RAW264.7 cells for 30 min prior to electrophoresis. Both mutant probes, which correlate with the decreased promoter activity in CAT assays (see Fig. 3B), do not demonstrate AP-1 DNA binding activity (arrow on left). activation, as reversal of the sequence of this element significantly diminishes promoter activity.

Expression of a Dominant Negative c-Fos Molecule Inhibits IL-12 p40
Promoter Activation by LPS-To demonstrate a functional role for AP-1 in IL-12 p40 promoter activity, a dominant negative AP-1 molecule, A-Fos (20), was transfected into RAW264.7 macrophages. A-Fos contains an acidic amphipathic protein sequence appended to the N terminus of the Fos leucine zipper and lacks a DNA binding domain (20). A-Fos inhibits LPS-induced Ϫ350 to ϩ55 IL-12 p40 promoter activation in a dose-dependent manner, with maximal inhibition of ϳ50% (data not shown). A-Fos also inhibits activation of the p40 promoter by C/EBP␤ by 50% (data not shown). In the context of the minimal Ϫ101 to ϩ55 p40 promoter, the effects of A-Fos are more significant, with inhibition of LPS-induced promoter activation by 70% and inhibition of C/EBP␤-induced activity by greater than 50% (Fig. 8). These inhibitory effects are dependent on the AP-1 DNA-binding element, as LPS-and C/EBP␤- FIG. 7. Determination of spacing requirements between control elements in the IL-12 p40 promoter. A, 5-, 10-, and 15-base pair insertion mutants were created at Ϫ119 and Ϫ84 within the IL-12 p40 Ϫ350 to ϩ55 promoter-CAT reporter plasmid. B, insertion mutant plasmids were transfected into RAW264.7 cells, incubated for 24 h, and then activated with LPS (5 g/ml) for 16 -18 h. Extracts from LPSactivated cells were analyzed by CAT assay. Results (normalized for ␤-galactosidase activity) are expressed as the percentage of the CAT activity obtained with the Ϫ350 to ϩ55 wild type promoter. These results suggest that spacing between the C/EBP and NF-B elements, as well as between the C/EBP and AP-1 sites, are important for promoter activation. Each result represents the mean Ϯ S.D. of data from four to five experiments. C, spacing mutants between the C/EBP and AP-1 elements have more dramatic effects on promoter activation by LPS in the context of a minimal deleted promoter, Ϫ101 to ϩ55. Results (normalized for ␤-galactosidase activity) were expressed as the percentage of the CAT activity obtained with the Ϫ101 to ϩ55 plasmid. Each result represents the mean Ϯ S.D. of data from four to five experiments. D, the specific sequence and orientation of the AP-1 site in the IL-12 p40 promoter is important for activation by LPS. Within the Ϫ350 to ϩ55 p40 promoter-CAT reporter plasmid, the AP-1-binding site from Ϫ81 to Ϫ75 was mutated to a canonical AP-1 site (Ϫ81 to Ϫ75 from CTAGTCA to TGAGTCA). This mutation strikingly increases LPSinduced promoter activation to almost 6 times wild type activity (Ϫ350 to ϩ55) and demonstrates a strong basal signal (20% of LPS-induced wild type activity) in unactivated cells. Also, the IL-12 p40 AP-1 site at Ϫ81 to Ϫ75 was reversed in orientation from CTAGTCA to ACTGATC. This mutation decreases LPS-induced promoter activity to 20% of wild type (Ϫ350 to ϩ55). Plasmids were transfected into RAW264.7 cells, incubated for 24 h, and then activated with LPS (5 g/ml) for 16 -18 h. Extracts from LPS-activated cells were analyzed by CAT assay. Results (normalized for ␤-galactosidase activity) are expressed as the percentage of the CAT activity obtained with the Ϫ350 to ϩ55 wild type promoter. Each result represents the mean Ϯ S.D. of data from three experiments. induced activity of the Ϫ350 to ϩ55 p40 promoter with a mutated AP-1 site (Ϫ85/Ϫ73) is not inhibited by A-Fos (data not shown). This result strongly suggests that A-Fos inhibits IL-12 p40 promoter activation by heterodimerizing with endogenous Jun family members but will not inhibit C/EBP-mediated transcriptional activation in the absence of an adjacent AP-1 element. Furthermore, the inhibitory effect of A-Fos does not represent a nonspecific phenomenon such as transcriptional squelching. A-Fos expression does not decrease the activity of two constitutively active promoters, a cytomegalovirus viral enhancer and an SP-1 multimer upstream of a minimal core promoter (data not shown).
c-Jun and a Mutant c-Jun Molecule Activate the IL-12 p40 Promoter and Synergize with C/EBP␤-To investigate a role for individual AP-1 proteins, mammalian expression plasmids for AP-1 family members were co-transfected with a Ϫ101 to ϩ55 p40 promoter-luciferase reporter into RAW264.7 cells. Neither JunB nor c-Fos expression activate the promoter (Fig.  9A). However, expression of c-Jun strongly activates the promoter (Fig. 9A) in a dose-dependent manner (data not shown). Co-transfection of Jun family members with c-Fos does not augment promoter activity (data not shown). Within the TNF promoter, a functional interaction between a C/EBP element immediately adjacent to an AP-1 site from Ϫ106 to Ϫ74 has been described (35). Expression of c-Jun and a mutant c-Jun molecule that lacks the C-terminal 67 amino acid transactivation domain, TAM67, were both capable of activating the TNF promoter (35). As TNF and IL-12 are critical mediators of inflammatory responses, we tested whether TAM67 activates the TNF and IL-12 p40 promoters in macrophages. Expression of TAM67 in RAW264.7 cells activates the Ϫ101 to ϩ55 IL-12 p40 promoter to the same extent as c-Jun (Fig. 9A). Furthermore, TAM67 activates a Ϫ283 to ϩ113 TNF promoter luciferase reporter in RAW264.7 cells (Fig. 9B). The effect of TAM67 on the TNF and IL-12 p40 promoters appears to be specific; c-Jun, but not TAM67, activates a multimerized consensus AP-1 DNA-binding element CAT reporter plasmid (TRE-CAT) in untreated RAW264.7 cells (data not shown). Furthermore, TAM67, originally described as a dominant negative inhibitor of AP-1-mediated transcription (19), strongly inhibits LPS-induced TRE-CAT activity (85%), whereas it augments LPSinduced IL-12 p40 promoter activity by 2-fold in RAW264.7 cells (data not shown).
To assess functionally additive or synergistic effects between C/EBP and AP-1 in IL-12 p40 promoter activation, C/EBP␤ was co-expressed with AP-1 family members in RAW264.7 cells. Expression of C/EBP␤ with c-Jun results in 8-fold greater promoter activation than with C/EBP␤ alone and 4-fold greater promoter activity than with c-Jun alone (Fig. 9C). Similar results are demonstrated with TAM67 (Fig. 9C). However, synergistic activation is not apparent when JunB and C/EBP␤ are co-expressed (Fig. 9C).
MAP Kinase and TLR Signaling Pathways Activate the IL-12 p40 Promoter through Control Elements in the Ϫ101 to ϩ55 Region-Recently, the p38 MAP kinase and the toll-like receptor (TLR) signal transduction pathways have been linked to IL-12 p40 gene expression (25,36). However, downstream targets of these pathways that mediate IL-12 p40 promoter activation have not been identified. Therefore, a series of dominant negative and constitutively active molecules from the p38 kinase, JNK, and TLR signal transduction pathways were tested for effects on IL-12 p40 promoter activity. All signal transduction molecules and soluble inhibitors display dose-dependent effects on promoter activity (data not shown). To demonstrate activation and inhibition of the p38 pathway in RAW264.7 cells, activity of a GAL4-luciferase reporter by a MEF2c-GAL4 DNA-binding domain fusion protein (37) was assessed. In RAW264.7 cells, LPS strongly induces p38 activity. The soluble p38 inhibitor SB202190 (1 M) inhibits LPS-induced p38-mediated transcription by 60%, and constitutively active MKK3 and MKK6 strongly induced this reporter in unactivated cells (data not shown). For the IL-12 p40 promoter, SB202190 inhibits LPS-induced activity of the Ϫ101 to ϩ55 promoter, whereas the structurally similar control SB202474 does not demonstrate inhibition (Fig. 10A). Furthermore, expression of constitutively active MKK3 or MKK6 induces Ϫ101 to ϩ55 p40 FIG. 9. c-Jun and a deleted c-Jun mutant, TAM67, activate the IL-12 p40 and the TNF-␣ promoters in RAW264.7 cells. A, Ϫ101 to ϩ55 IL-12 p40 promoter-luciferase reporter plasmid was co-transfected with 0.5 g of expression plasmids for c-Jun, TAM67, JunB, and c-Fos. Plasmid amounts were normalized to 2.5 g by the addition of empty expression plasmid. Following transfection, RAW264.7 cells were incubated for 36 h. Experiments with AP-1 expression plasmids were conducted in unactivated cells. As controls, cells were transfected with 2.5 g of empty expression plasmid and remained unactivated or were activated with LPS (5 g/ml) for 18 h prior to termination. B, Ϫto TNF-␣ promoter-luciferase reporter was co-transfected with 0.5 g of expression plasmids for c-Jun, TAM67, JunB, and c-Fos, as in A. C, c-Jun and TAM67 synergistically activate the IL-12 p40 promoter when co-expressed with C/EBP␤. A Ϫ101 to ϩ55 IL-12 p40 promoter-luciferase reporter plasmid was co-transfected with expression plasmids for C/EBP␤ (0.1 g) and AP-1 proteins (0.5 g), as in A. Results are expressed as fold activation compared with cells transfected with C/EBP␤ alone. A-C, results (relative light units normalized for ␤-galactosidase activity) are expressed as the percentage of luciferase activity obtained with the Ϫ101 to ϩ55 wild type promoter activated with LPS. Each result represents the mean Ϯ S.D. of three to five experiments.
promoter activity in untreated cells, although not to the same extent as LPS (Fig. 10A).
The role of the JNK pathway in IL-12 p40 promoter activation has not been determined, although the description of an AP-1 site suggests that this pathway may be involved. LPS activates the JNK pathway in RAW264.7 cells, as assessed by activation of a GAL4-luciferase reporter by a c-Jun-GAL4 DNA-binding domain fusion protein (data not shown). A dominant negative inhibitor of JNK, JIP-1, inhibits LPS-induced JNK-mediated transcription, while constitutively active MEKK1 strongly up-regulated activity in untreated cells (data not shown). Consistent with involvement of the JNK pathway in IL-12 p40 promoter activation, JIP-1 inhibits LPS-induced Ϫ101 to ϩ55 promoter activity, and active MEKK1 strongly activates this minimal p40 promoter (Fig. 10B). As MEKK1 also induces NF-B (38), it is important to note that this effect occurs with a p40 promoter that lacks an NF-B element.
Dominant negative signal transduction molecules in the Toll/ IL-1 receptor pathway inhibit IL-12 p40 promoter activation by bacterial products (25). As NF-B is a well described target of TLR signaling, we asked whether this signal transduction pathway affects IL-12 p40 promoter activation through elements other than NF-B. Expression of a dominant negative TLR2, TRAF6, and MyD88 inhibits LPS-induced activity of the IL-12 p40 Ϫ101 to ϩ55 promoter (Fig. 10C) to the same magnitude as they inhibit activation of a multimerized NF-B element luciferase reporter in RAW264.7 cells (data not shown). Similar results were obtained when RAW264.7 cells were activated with heat-killed L. monocytogenes and M. tuberculosis whole cell extract (data not shown). Furthermore, a constitutively active CD4-TLR4 fusion protein activates the Ϫ101 to ϩ55 promoter in untreated cells, although not to the same degree as LPS (Fig. 10C). Once again, similar results are obtained for IL-12 p40 promoter constructs containing and lacking the NF-B site (data not shown), suggesting that the TLR pathway may influence promoter activation through elements other than NF-B.

DISCUSSION
In summary, functional and DNA binding assays reveal complex interactions within a small region of the IL-12 p40 promoter. A C/EBP DNA-binding element was previously characterized at Ϫ96 to Ϫ88. In the current study, mutations from Ϫ79 to Ϫ74 in the murine IL-12 p40 promoter significantly reduce LPS-induced promoter activation and correlate with DNA binding of AP-1 by EMSA. AP-1 family members detected in the IL-12 p40 promoter DNA-binding complex include JunB, c-Jun, and c-Fos. Spacing between the C/EBP and AP-1 site and orientation of the AP-1 site are important for promoter activation, suggesting cooperativity between these two sites. Expression of a dominant negative AP-1 protein, A-Fos, in RAW264.7 cells inhibits p40 promoter activation by LPS and C/EBP␤. c-Jun and a c-Jun mutant TAM67 activate the IL-12 p40 promoter and synergize with C/EBP␤ in transient expression studies. c-Jun and TAM67 also activate the TNF promoter in RAW264.7 cells. Finally, this small region of the promoter downstream of Ϫ101 is identified as a target for MAP kinase and TLR signal transduction pathways.
This group had reported a comprehensive functional analysis of the murine and human IL-12 p40 promoters in RAW264.7 cells (17). A critical control element for promoter activation by bacterial products, located between Ϫ96 and Ϫ88, interacts with C/EBP family members. This element was functionally synergistic with the NF-B site between Ϫ131 and Ϫ120 (13). In this analysis, when the C/EBP site was mutated, virtually all LPS-induced promoter activity was lost. Thus, it was possible to conclude that the C/EBP site was both necessary and sufficient for promoter activation. To characterize effects of other cis-acting elements, a strategy was developed to assess the functional role of other elements in the presence of an intact C/EBP site. First, mutations were created in two cisacting elements other than the C/EBP site. A second approach utilized a minimal promoter where all elements upstream of the C/EBP site were deleted. These experiments demonstrate the existence of three important cis-acting elements at Ϫ114 to Ϫ102, Ϫ82 to Ϫ73, and Ϫ62 to Ϫ57. Of note, the functional regions described in this study correspond to constitutive and bacterial product-inducible DNase I footprints in the previous analysis (17). Furthermore, these regions likely correspond to DNA-protein interactions detected on the endogenous IL-12 p40 promoter in primary macrophages by in vivo genomic DNase I footprinting (39). Taken together, these results provide clear evidence that the C/EBP site may be necessary but is not sufficient for LPS-induced promoter activation.
A cis-acting control element at Ϫ207 to Ϫ211 in the human IL-12 p40 promoter that binds a multiprotein complex consisting of Ets-2, interferon regulatory factor-1, interferon consensus sequence binding protein, and c-Rel has been extensively characterized (14 -16, 40). Although we could not demonstrate a functional effect for a single site-directed substitution mutant in a previous analysis (17), a functional effect may be demonstrable by methodologies employed in the current study. Thus, this analysis and others suggest that at least six control elements within a small region of the IL-12 p40 promoter may be necessary for activated gene expression.
One novel site defined by this analysis from Ϫ84 to Ϫ73 was characterized in detail. A control element was identified from Ϫ79 to Ϫ74 that binds AP-1 family members. Supershift experiments detect predominantly c-Fos, JunB, and to a lesser extent, c-Jun in the IL-12 p40 EMSA complex. However, relative amounts of AP-1 family members cannot be accurately deduced from these results, as the degree of inhibition or supershift is a function of the antibody as well as the relative amount of protein in the DNA-binding complex. Furthermore, a recent study demonstrated that c-Rel, a relatively minor component of the IL-12 p40 NF-B DNA-binding complex, is the relevant activator of p40 transcription in cell line models and in vivo (41). Therefore, from our functional results, c-Jun may be the relevant activator of the IL-12 p40 promoter despite the apparent abundance of JunB and c-Fos in the EMSA complex.
To demonstrate a role for AP-1 in IL-12 p40 promoter activation, a dominant negative AP-1 molecule, A-Fos, was expressed in RAW264.7 macrophages. A-Fos contains an acidic amphipathic protein sequence appended onto the N terminus of the c-Fos leucine zipper. This extension physically interacts with the Jun basic region and prevents the basic region from binding DNA. This interaction was demonstrated to be specific. A-Fos inhibited transcriptional activation of Jun-dependent reporters but not C/EBP-dependent reporters (20). Expression of A-Fos significantly inhibits promoter activation by LPS or C/EBP␤ (Fig. 8), demonstrating that AP-1 DNA binding is critical for IL-12 p40 promoter activation. Then, the potential role of individual AP-1 family members in IL-12 p40 promoter activation was determined. Interestingly, c-Jun and a deleted c-Jun mutant, TAM67, activate the p40 promoter in untreated RAW264.7 to a similar magnitude as LPS and synergistically activate the promoter when co-expressed with C/EBP␤. Furthermore, JunB and c-Fos, although abundantly detected in EMSA complexes, are unable to activate the promoter or aug-ment C/EBP␤-induced activation.
Important information about functional synergy between the AP-1 and C/EBP element was obtained from experiments that altered spacing between these sites. A 5-, 10-, and 15-bp insertion between AP-1 and C/EBP markedly decreases promoter activation by LPS. In other promoter/enhancers, spatial arrangement of transcription factor binding sites on the double helix is critical for promoter activation (33,34). For the p40 promoter, it appears that distance rather than arrangement of these sites on the double helix is critical for optimal activation, as the addition of any length decreased reporter activity. A promoter mutant that created a canonical AP-1 site from Ϫ81 to Ϫ75 demonstrates markedly increased LPS-induced promoter activity and strong basal activity (Fig. 7D), suggesting that this region contains the AP-1 site. However, precise mapping of the IL-12 p40 AP-1 element will require a careful analysis of DNA-protein contacts using methylation interference assays and further mutagenesis of the region. When the AP-1 site is reversed in orientation, a marked decrease in LPS-induced activity is found. This result may suggest a protein-protein interaction with C/EBP or another factor or may be related to the ability of AP-1 to bend DNA, thus facilitating protein-protein interactions (42).
AP-1 family members are prototypic targets for the c-Jun N-terminal kinase (JNK) MAP kinase pathway. Although p38 and extracellular signal-regulated kinases have been implicated in IL-12 p40 gene expression (36,43), the role of the JNK pathway is unclear. In this study, the JNK pathway is implicated in IL-12 p40 promoter activation using a dominant negative inhibitor of JNK, JIP-1, and a constitutively active MEKK1 molecule. Interaction of conserved bacterial ligands with TLRs is important for activation of IL-12 p40 gene expression in macrophages (25). However, specific promoter region and transcription factor targets that modulate IL-12 p40 gene expression have not been identified. For TLR signaling, NF-B is a critical downstream target (26), although MAP kinases may be activated through TLRs (44). The p38, JNK, and TLR pathways, in part, appear to influence IL-12 p40 gene expression through elements downstream of Ϫ101. However, specific transcription factors activated by these pathways remain elusive. For example, in nuclear extracts from LPS-treated RAW264.7 cells and primary bone marrow-derived macrophages, protein expression and DNA binding for C/EBP or AP-1 family members is not affected by treatment with a soluble p38 inhibitor (data not shown).
The IFN-␤ (45) and the T-cell receptor ␣ (46) gene enhancers provide important biochemical details of how transcriptional activators are assembled into a nucleoprotein complex called the "enhanceosome" that promotes their interaction and cooperative binding to DNA. Extensive analysis of the IL-12 p40 promoter begins to describe an IL-12 p40 enhanceosome. In particular, a small region from Ϫ96 to Ϫ74 of the promoter is a critical focus for protein-DNA and possibly protein-protein interactions. This small sequence that contains a C/EBP and AP-1 DNA-binding element is of particular interest because it appears that the strategy used by macrophages to induce IL-12 p40 expression may define a common mechanism for other important inflammatory genes. As discussed, within the TNF-␣ promoter, there is a C/EBP element immediately adjacent to an AP-1 site. Functional synergy and cooperative DNA binding between c-Jun and C/EBP␤ has been demonstrated in activation of the TNF-␣ promoter (35). Similarly, AP-1 and C/EBP sites have been characterized at similar locations with respect to a TATA box in the IL-8 gene (47,48). Thus, transcriptional synergy and cooperativity between the C/EBP and AP-1 families of transcription factors may be involved in the regula-tion of several important inflammatory genes expressed in macrophages.
These studies, performed in cell lines, suggest mechanisms by which IL-12 p40 may be transcriptionally activated in vivo. Confirming the findings of this analysis in primary cells and in vivo will require a combination of approaches to determine transcription factor occupancy on the endogenous promoter and their responses to biologic inducers and inhibitors of IL-12 gene expression.