Mitogen-activated Protein Kinases Mediate Activator Protein-1-dependent Human Inducible Nitric-oxide Synthase Promoter Activation*

Inducible nitric-oxide synthase (iNOS) is an important signaling protein involved in the regulation of biological processes (e.g. vasodilation, inflammation) and is subject to transcriptional regulation by cytokines and lipopolysaccharide (LPS). Full activation of the human iNOS (hiNOS) promoter by cytokines (i.e., tumor necrosis factor-α, interleukin-1β, interferon-γ (IFN-γ)) required downstream and upstream nuclear factor-κB (−115, −8283) and activator protein-1 (AP-1) (−5115, −5301) transcription factor binding sites. Human lung epithelial (A549) cells were transiently transfected with luciferase reporter plasmids containing an 8.3-kilobase human iNOS promoter to examine the molecular signaling events necessary for hiNOS transcriptional activation. The combination of LPS and IFN-γ, but neither alone, increased hiNOS promoter activity 28-fold, in a reaction requiring two critical AP-1 (JunD·Fra-2) promoter binding sites. Mitogen-activated protein kinases (MAPKs) were assessed as potential activators of AP-1 and the hiNOS promoter. Both pharmacological and molecular inhibitors of the extracellular signal-related kinase (ERK) and p38 pathways reduced cytokine mixture (CM)- and LPS/IFN-γ-induced promoter activation. By gel retardation analysis, the addition of MAP/ERK kinase-1 and p38 inhibitors significantly diminished AP-1 binding in both CM- and LPS/IFN-γ-stimulated cells. Thus, p38- and ERK-dependent pathways, through effects on the AP-1 complex, activate the hiNOS promoter in cells stimulated with CM or LPS/IFN-γ.

Inducible nitric-oxide synthase (iNOS) is an important signaling protein involved in the regulation of biological processes (e.g. vasodilation, inflammation) and is subject to transcriptional regulation by cytokines and lipopolysaccharide (LPS). Full activation of the human iNOS (hiNOS) promoter by cytokines (i.e., tumor necrosis factor-␣, interleukin-1␤, interferon-␥ (IFN-␥)) required downstream and upstream nuclear factor-B (؊115, ؊8283) and activator protein-1 (AP-1) (؊5115, ؊5301) transcription factor binding sites. Human lung epithelial (A549) cells were transiently transfected with luciferase reporter plasmids containing an 8.3-kilobase human iNOS promoter to examine the molecular signaling events necessary for hiNOS transcriptional activation. The combination of LPS and IFN-␥, but neither alone, increased hiNOS promoter activity 28-fold, in a reaction requiring two critical AP-1 (JunD⅐Fra-2) promoter binding sites. Mitogen-activated protein kinases (MAPKs) were assessed as potential activators of AP-1 and the hiNOS promoter. Both pharmacological and molecular inhibitors of the extracellular signal-related kinase (ERK) and p38 pathways reduced cytokine mixture (CM)-and LPS/IFN-␥-induced promoter activation. By gel retardation analysis, the addition of MAP/ERK kinase- 1

and p38 inhibitors significantly diminished AP-1 binding in both CM-and LPS/IFN-␥-stimulated cells. Thus, p38-and ERK-dependent pathways, through effects on the AP-1 complex, activate the hiNOS promoter in cells stimulated with CM or LPS/IFN-␥.
Nitric oxide (NO) 1 is a signaling molecule produced by a family of heme-containing nitric-oxide synthases (NOS) concomitant with conversion of L-arginine to L-citrulline (1). Neuronal NOS (NOS1) and endothelial NOS (NOS3) are constitutively present in a variety of cell types, with their activities sensitive to changes in intracellular calcium. Much larger amounts of NO are produced in response to numerous extracellular stimuli after induction of the calcium-insensitive iNOS (NOS2) (2). The transient burst of NO production by iNOS has been implicated in numerous biological processes (e.g. immune modulation, control of vascular tone) (3)(4)(5).
Clues to the role of iNOS in physiologic and pathophysiologic states resulted from studies of its cellular and subcellular localization as well as the stimulatory elements and signaling pathways required for its induction (6). Unfortunately, many of these studies were performed in animal models, which clearly do not reflect human physiology. For instance, NO production by iNOS in macrophages appears to be more important in mice than humans (7). Differences among species in NO production and NOS expression in response to a given inflammatory stimulus may reflect differences in the cell signaling mechanisms that lead to iNOS gene transcription and/or in promoter structure.
Transcriptional regulation is a crucial checkpoint in the initiation of cytokine-stimulated NO production by human iNOS (hiNOS) (8). 7.2-kb and 8.3-kb 5Ј-flanking regions of the human iNOS gene were found to contain NF-B and activator protein-1 (AP-1) binding regulatory regions (9,10). JunD and Fra-2 were identified in the heterodimers that bound to upstream and downstream AP-1 sites in the 8.3-kb hiNOS promoter. The upstream and downstream NF-B sites bound RelA/RelA and RelA/p50, respectively (10). Because both lipopolysaccharide (LPS) and cytokines can activate AP-1 and NF-B (11)(12)(13)(14), the promoter provides a molecular model with which to define the signaling determinants of iNOS activation in response to inflammatory mediators.
MAPKs are signaling proteins rapidly activated by tyrosine phosphorylation, and tyrosine kinase inhibitors have been studied as potential therapeutic targets in inflammatory and proliferative disorders (15)(16)(17)(18)(19). Activation of MAPK pathways by LPS and cytokines (20,21) represents a potential signaling mechanism for NO production during the inflammatory response. Furthermore, the extracellular signal-regulated kinases (ERKs), c-jun N-terminal kinase/stress-activated protein kinases (JNK/SAPKs), and p38 MAPK pathways have been implicated in the activation of AP-1 (22), which is involved in hiNOS promoter activation by cytokines (10).
LPS, tumor necrosis factor-␣, interferon-␥ (IFN-␥), and interleukin-1␤ (IL-1␤), which are present in human airways during Gram-negative infection, cause inflammation in part by initiating NO production by iNOS (23). Because AP-1 and NF-B activation are major determinants of hiNOS induction in A549 cells, we hypothesized that MAPK proteins were upstream signaling effectors of hiNOS promoter activation in response to LPS and cytokines. Using both pharmacological and molecular approaches, the involvement of p38 MAPK and ERK pathways in hiNOS promoter activation by cytokines and LPS/IFN-␥ via binding of AP-1 to specific promoter sequences was demonstrated.

EXPERIMENTAL PROCEDURES
Cell Culture and Cytokine Induction-A549 cells (American Type Culture Collection (ATCC) CCL 185), a human alveolar type II epithelial cell-like lung adenocarcinoma cell line, were grown at 37°C at 5% CO 2 in Ham's F-12 K medium supplemented with 10% heat-inactivated fetal bovine serum, 2 mM glutamine, 100 units/ml penicillin, and 100 g/ml streptomycin (all from Biofluids). To initiate experiments, cells were washed with serum-free medium and treated with or without inhibitors for 1 h before incubation for 6 h with cytokine mixture (CM) containing 100 units/ml IFN-␥ (Roche Molecular Biochemicals), 0.5 ng/ml IL-1␤ (Genzyme), and 10 ng/ml tumor necrosis factor-␣ (Roche Molecular Biochemicals), or a mixture of 100 g/ml LPS (Sigma) and 100 units/ml IFN-␥. Erbstatin, genistein, SB203580, and PD98059 (Calbiochem) dissolved in dimethyl sulfoxide (Me 2 SO) were added directly to the cells.
Transient Transfection and Determination of hiNOS Promoter Activity-A549 cells were transfected with hiNOS promoter constructs linked to luciferase cDNA (PGL3-hiNOS) (10) without or with mammalian expression vectors containing cDNAs for various forms of MAP/ ERK kinase 1 (MEK1), MAPK kinase 3 (MKK3), and MKK6. Expression vectors for dominant negative MEK1 (pcDNA3-MEK1 dn), MKK6 (pcEFLGST-MKK6 dn), MKK3 (pcDNA3-MKK3 dn), as well as those for wild-type and constitutively active MKK3 (pcDNA-MKK3 wt, pcDNA-MKK3 ca) were kind gifts from Dr. J. S. Gutkind, NIDCR, National Institutes of Health, Bethesda, MD. In separate experiments, expression of the above proteins in A549 cells was confirmed by Western blot analysis (data not shown). After 36 h, cells were induced with cytokines for 6 h before harvest of cells for luciferase activity determination (Luciferase Assay System Kit, Promega). hiNOS promoter regional sequence requirements for LPS/IFN-␥-stimulated activation were assessed using deletion mutants as reported previously (24). Luminescence units were expressed relative to total lysate protein, determined using the bicinchoninic acid protein assay (Pierce).
Luciferase activities of both stimulated (CM or LPS/IFN-␥) and unstimulated cells are the means of values from at least two independent experiments with assays in triplicate. The fold induction was calculated by dividing the stimulated by the unstimulated luciferase activity. For pharmacological inhibition experiments, the percentage of control is the fold induction in the presence of inhibitor divided by that in the absence times 100. For MKK cotransfection experiments, the percentage of control is the hiNOS promoter activity of cells cotransfected with a mammalian expression vector containing MKK cDNA divided by that of cells cotransfected with the corresponding empty vector times 100. Me 2 SO (up to 1% final concentration) did not affect LPS/IFN-␥-or CM-stimulated promoter activity (data not shown).
Preparation of Nuclear Extracts-Nuclear extracts were prepared from cells grown near confluence in 100-mm 2 dishes, incubated for 1 h without or with inhibitors, before stimulation with CM or LPS/IFN-␥ for 3 h, and harvested as described (10). Protein concentrations were determined using a Bradford assay kit with bovine serum albumin as standard.
Electrophoretic Mobility Gel Shift Assay-NF-B and AP-1 binding activity was determined by electrophoretic mobility gel shift assay using the Promega Gel Shift Assay System. Samples (5 g) of nuclear proteins were incubated with the indicated radiolabeled oligonucleotide for 20 min at room temperature (ϳ21°C). Specificities of the binding reactions were tested in competition assays in which a 100-fold excess of unlabeled wild-type or mutated oligonucleotide was added to extracts from stimulated cells 15 min before the labeled probe. In assays without nuclear protein, none of the probes caused any discernible band shift (data not shown). Protein-nucleotide complexes were separated by electrophoresis in a 6% DNA-retardation gel (Novex) with Tris borate/ EDTA (50 mM Tris-HCl, 50 mM boric acid, 1 mM EDTA, pH 8.3) at constant current (30 mA) at room temperature. Photographic film was exposed to dried gels at Ϫ70°C and scanned using an Epson Expression 636 scanner. Band densities were measured using Scion Image beta 3b software. Because of their variable nature, band density results are mean gray value for a given area minus a measured gray background value for each gel. Me 2 SO alone (up to 1% final concentration) did not affect LPS/IFN-␥or CM-stimulated transcription factor binding (data not shown).
Measurement of Cell Viability-Cell viability was assessed by the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay (25). Briefly, cells plated in 96-well plates (2.5 ϫ 10 5 cells/well) were incubated for 1 day before addition of LPS/IFN-␥ or CM with and without inhibitors or Me 2 SO alone (eight wells/group), and 6 h later, MTT (500 g/ml). After 4 h in the dark, plates were centrifuged (450 ϫ g for 10 min); the precipitate was dissolved in Me 2 SO and shaken for 10 min. Absorbance at 540 nm was directly proportional to the number of cells plated (data not shown). Viability was assessed as the ratio of absorbances for inhibitor-treated and untreated cells. No differences in viability among differently treated cells were found (data not shown).

Effects of LPS and IFN-␥ on hiNOS Promoter Activity in A549
Cells-After incubation of A549 cells with different concentrations of LPS without or with IFN-␥ for 6 h, neither IFN-␥ nor LPS alone significantly activated the hiNOS promoter. In the presence of IFN-␥, however, LPS increased promoter activity in a concentration-dependent manner to a maximum of 28-fold the control value (Table I). Results were similar for 12-h incubations, with a 45-fold increase in hiNOS promoter activity attributable to LPS/IFN-␥ (data not shown).
Role of AP-1 and NF-B Binding Sites in hiNOS Promoter Activation by LPS/IFN-␥-It had been shown that removal of the upstream NF-B binding site reduced the promoter response to the cytokine mixture by 82% (24); removal of promoter regions upstream of AP-1 binding sites (deletions Ϫ6796, Ϫ6534, and Ϫ5774 bp), had no further effect (Fig. 1A). For the response to LPS/IFN-␥, the Ϫ7277-bp deletion mutant retained 50% of the native promoter activity (Fig. 1B), suggesting that  upstream NF-B binding (RelA/RelA) was more important for CM activation than for that by LPS/IFN-␥. The region between Ϫ5774 bp and Ϫ7277 bp accounted for a 4.5-fold increase in LPS/IFN-␥-stimulated promoter activity. In an attempt to identify a discrete binding site responsible for enhancement of hiNOS activation by LPS/IFN-␥, we constructed promoter fragments containing serial truncations at short intervals (Ϫ6796bp, Ϫ6534-bp, Ϫ5774-bp). Reductions in promoter activity due to successive small deletions were incremental but small, suggesting heterogeneity in the unknown enhancer elements responsible for full hiNOS activation by LPS/IFN-␥ and the absence of a discrete site (Fig. 1, B and C).
In cells transfected with the hiNOS promoter construct lacking the region upstream from the two AP-1 sites (Ϫ5774 bp), CM caused a 7.6-fold stimulation, twice that caused by LPS-IFN-␥. Removal of the two AP-1 sites (Ϫ3665 bp) further decreased LPS/IFN-␥ stimulation 3.5-fold. Because this region accounted for the largest discrete drop in LPS/IFN-␥-stimulated hiNOS promoter activation, and the presence of intact AP-1 binding sites was critical for CM stimulation, we examined the signaling pathways that might lead to hiNOS promoter activation via AP-1.
Role of Tyrosine Kinase Activity in CM-induced hiNOS Promoter Activation-The general tyrosine kinase inhibitors genistein and erbstatin each attenuated CM-induced hiNOS promoter activity in a concentration-dependent manner (Table II). The effects of the more specific tyrphostins, AG126 and AG1288, as well as the JAK2 inhibitor AG490, were then investigated. All three tyrphostins decreased CM-and LPS/ IFN-␥-induced activation in concentration-dependent manner ( Table III), suggesting that MAPK and JAK pathways are involved in hiNOS promoter induction.
To define better the MAPKs involved in activation of the hiNOS promoter, we used a molecular approach by coexpressing dominant negative, wild-type, and constitutively active forms of ERK and p38 MAPK kinases. Complementary results were obtained with pharmacological inhibitors of MEK1 (PD98059) and p38 MAPK (SB203580). PD98059 produced 25 and 42% reductions in CM-and LPS/IFN-␥-stimulated hiNOS induction, respectively ( Fig. 2A). As was the case for PD98059, coexpression of dominant negative MEK1 lead to a significant reduction in basal as well as CM-and LPS/IFN-␥-stimulated hiNOS activation (Fig. 2B), indicating a role for ERK activation in hiNOS transcriptional activation.
SB203580 caused 43 and 45% inhibition of CM-and LPS/ IFN-␥-stimulated hiNOS promoter activity, respectively (Fig.  3A). Cotransfection of dominant negative forms of MKK3, but not of MKK6, with the hiNOS promoter led to a significant reduction in CM-or LPS/IFN-␥-stimulated hiNOS promoter activity (Fig. 3B). Coexpression of wild-type or constitutively active MKK3 enhanced basal hiNOS promoter activity, confirming the involvement of the p38 MAPK pathway (Fig. 4). hiNOS promoter activation by CM and LPS/IFN-␥ was not   enhanced further by the overexpression of wild-type and constitutively active MKK3. Administration of SB203580 reversed the hiNOS promoter activation attributed to expression of constitutively active MKK3, demonstrating specificity of SB203580 for p38 MAPK in the context of cytokine-stimulated hiNOS activation. Together the two pharmacological inhibitors reduced CM-and LPS/IFN-␥-stimulated hiNOS promoter activity 86 and 89%, respectively (Fig. 5). Apparently additive effects of the p38 and MEK1 inhibitors suggest that the two kinases contribute independently to the activation of the hi-NOS promoter.

Effects of p38 MAPK and MEK1 Inhibition on Transcription Factor Binding to hiNOS Promoter Regulatory Elements-Us-
ing DNA probes corresponding to the respective NF-Bu, NF-Bd, AP-1u, and AP-1d binding sites, we assessed the effects of ERK and p38 MAPK pathway inhibition on transcription factor binding by gel retardation analysis. Exposure to CM or LPS/ IFN-␥ for 3 h increased to similar extents the binding of NF-B and AP-1 complexes (Fig. 6, A and B). The gel migration dis-tance of the DNA probe-transcription factor complex was the same for CM-and LPS/IFN-␥-stimulated cells, consistent with the binding of identical dimer complexes to the oligonucleotide probes regardless of the agonist used (data not shown). In unstimulated cells, SB203580 and PD98059 did not affect binding to NF-B sites, whereas the combination of SB203580 and PD98059 increased binding to AP-1 sites. SB203580 and PD98059 inhibited LPS/IFN-␥-but not CM-stimulated NF-B downstream (Ϫ115) binding. SB203580 and PD98059 augmented both LPS/IFN-␥-and CM-stimulated NF-B upstream (Ϫ8283) binding. SB203580 or PD98059, singly or together, decreased CM-and LPS/IFN-␥-stimulated AP-1 binding. Of note, the inhibitory effects of PD98059 and SB203580 on the CM-and LPS/IFN-␥-induced increases in AP-1 downstream binding were additive. These data point to a crucial role for MAPKs in AP-1-dependent human iNOS induction. DISCUSSION The production of NO by iNOS has a variety of important biological effects including oxidative stress and signaling (3,23). Production of iNOS protein is tightly regulated at the transcriptional level, and the signaling pathways by which human iNOS transcription is controlled in response to inflammatory stimuli are poorly understood. Here, we report that LPS, in combination with IFN-␥, activated specific NF-B-and AP-1-binding elements in an 8.3-kb hiNOS promoter. The ERK and p38 MAPK pathways played a crucial role in human iNOS transcriptional activation via modulation of AP-1 binding to specific promoter sequences. The 8.3-kb hiNOS promoter construct has several distinct advantages over measures of NO production or NOS expression, most importantly, the ability to study the molecular determinants of transcriptional activation (i.e. NF-B, AP-1 binding). Introduction of an exogenous iNOS promoter makes it possible to focus directly on transcriptional activation in intact cells, avoiding confounding factors that affect measures of iNOS production. Furthermore, in intact cells, promoter activation is amenable to molecular characterization, and iNOS transcription is the rate-limiting step in cytokine-induced NO production. However, the validity of conclusions regarding signaling pathways involved in iNOS induction clearly depends on the structural characteristics of the putative promoter sequence. Taylor et al. (9) cloned a 7.2-kb hiNOS promoter whose activity was increased 4-fold by cytokines in A549 cells. The upstream region contained in the 8.3-kb promoter construct appears also to be important for full hiNOS transcriptional activation. In addition, MAPK-dependent hiNOS promoter activation was related to the enhancement of AP-1 binding activity at the Ϫ5115 and Ϫ5301 positions of the 8.3-kb hiNOS 5Ј-flanking region.
Animal models of iNOS induction do not accurately reflect signaling mechanisms in human cells. In a region containing critical cytokine response elements (Ϫ5774 to Ϫ3665), two identical AP-1 sites are present in the hiNOS promoter which are absent in its murine counterpart (24). The mouse iNOS promoter contains an upstream regulatory region with a cluster of four IFN-␥-response enhancer elements that account for the potentiation of LPS-stimulated promoter induction by IFN-␥ (26 -28). In the human promoter, the upstream putative IFN-␥-response elements (Ϫ8296 to Ϫ7227) apparently do not contribute to cytokine-stimulated promoter activation (24). Other transcription factors (e.g. octomer motif, NF-IL6) enhance cytokine-induced murine iNOS expression (29,30) but are unlikely to play a direct role in human iNOS induction (24).
A role for tyrosine kinases in iNOS induction has been established (31)(32)(33)(34)(35)(36). The importance of early gene activation was stressed in animal models of sepsis, in which tyrphostins decreased mortality only when administered before, or soon after, the initiation of sepsis (16,17,19). The JAK2 inhibitor AG490 attenuated CM stimulation of hiNOS promoter activity. However, none of the contributory interferon response elements usually responsible for JAK/STAT-activated gene transcription had been identified in a previous characterization of hiNOS promoter activation by cytokines (24). One possible explanation is that IFN-␥-activated factors modified hiNOS transactivation by NF-B or AP-1 without binding directly to the promoter. Alternatively, JAK2 may be responsible for the direct activation of the MAPK pathway (37). In agreement, AG490 attenuated IL-2-induced T cell proliferation by inhibiting JAK3-dependent MAPK activation and AP-1 binding (38). The additive stimulation via these pathways may explain the requirement for full activation of the hiNOS promoter by tumor necrosis factor-␣/IL-1␤ to observe an IFN-␥ effect. In contrast to CM activation, LPS/IFN-␥ stimulation of hiNOS promoter activity was reduced by the deletion of promoter regions containing putative interferon response elements (Ϫ5774 bp to Ϫ7727 bp). AG490 may have reduced LPS/IFN-␥ stimulation of promoter activity by inhibiting STAT binding to one of these elements.
Both p38 and ERK pathways involve upstream protein tyrosine phosphorylation by specific MAPK kinases and can lead to activation of AP-1 elements (39 -42). Others have reported a role for MAPK-mediated AP-1 activation of gene transcription (43,44). In our study, the use of both pharmacological and molecular inhibitors of MAPK pathways was instrumental in dissecting the mechanisms of hiNOS transcriptional activation. The p38 inhibitor SB203580 and MEK1 inhibitor PD98059 reduced LPS/IFN-␥-and CM-mediated hiNOS promoter activation in additive fashion, suggesting a potential role for MAPKs in AP-1-dependent hiNOS induction. Similarly, coexpression of dominant negative MEK1 led to a 50% attenuation of both basal and CM-or LPS/IFN-␥-stimulated hiNOS promoter activity. Because the pattern of inhibition by dominant negative MEK1 expression was similar to that by PD98059, the latter is likely specific for the ERK pathway, insofar as hiNOS promoter activation is concerned. MKK6 and MKK3 are upstream kinases with demonstrated specificity for p38 MAPK (45,46). Consistent with the action of SB203580, overexpression of dominant negative MKK3 reduced CM-and LPS/IFN-␥-stimulated, but not basal, hiNOS promoter activity. Overexpression of wild-type and constitutively active MKK3 enhanced basal hiNOS promoter activity, confirming the involvement of the p38 MAPK pathway. The increase in CM-and LPS/IFN-␥-stimulated hiNOS promoter activity observed upon expression of constitutively active MKK3 was reversed by SB203580, demonstrating that the inhibition of hiNOS activation by SB203580 was via p38 MAPK.
The lack of effect of dominant negative MKK6 on hiNOS activation suggests selectivity for MKK3 in CM-and LPS/IFN-␥-stimulated hiNOS transcriptional activation via p38 MAPK. Similar signaling specificity was demonstrated when overexpression of PYK2 activated p38 MAPK via MKK3 and not MKK6 (47). Although both MKK3 and MKK6 can activate p38 MAPK, each exhibited p38 MAPK isoform selectivity (45), and each was activated differentially by specific extracellular stimuli (48,49).
The decrease in hiNOS promoter activity caused by MAPK inhibitors was accompanied by reduced CM-and LPS/IFN-␥stimulated AP-1 binding, whereas binding to the upstream NF-B (RelA/RelA) sequence was increased. These data are consistent with an essential role for MAPKs in AP-1-but not NF-B-dependent hiNOS activation. The mechanism by which binding to the NF-B upstream site was augmented by SB203580 and PD98059 is unclear, although others have observed reduction of NF-B activation by activated MAPK (50). The oligonucleotide sequence-specific differences in agonistinduced NF-B binding may reflect the different dimers that bind to upstream (RelA/RelA) and downstream (RelA/p50) sites.
Although MAPKs can also activate NF-B (51-53), the p38 and MEK1 inhibitors did not reduce CM-induced NF-B binding. NF-B binding to the hiNOS promoter, which is required for full hiNOS promoter activation, can, therefore, be stimulated via pathways other than those involving MAPKs. Similarly, others demonstrated that MAPK-dependent AP-1 activation was partially responsible for human immunodeficiency virus type 1 long terminal repeat activation and that MAPK-independent NF-B induction was required for full promoter activation (54). A potential MAPK-independent NF-B inducer, the recently characterized Toll-like receptor-2, or TLR2 (55) has an intracytoplasmic IL-1 receptor-like domain and is activated by LPS. In fact, LPS activation of NF-B was inhibited by transfection with dominant negative mutant components of the IL-1 signaling pathway (56), and a TLR2 homolog, TLR4, plays a role in LPS-induced NF-B activation (57). TLR-and IL-1␤ receptor-mediated activation of NF-B in A549 cells may be responsible for the increased NF-B binding seen in response to LPS/IFN-␥ and cytokines.
The p38 and MEK1 inhibitors SB203580 and PD98059 caused additive inhibition of hiNOS promoter activity and AP-1 binding to downstream, but not upstream sites, possibly reflecting the different binding site sequences. Although additive reduction in AP-1 binding likely explains the additive inhibition of hiNOS promoter activity, transcription factor binding to promoter regulatory elements is one of several potential mechanisms for control of gene transactivation (58). Consistent with AP-1 binding as the critical step in hiNOS promoter activation, cytokines stimulated de novo JunD and Fra-2 expression, coinciding with increased AP-1 binding (10). However, in addition to affecting AP-1 binding, MAPKs can also influence posttranslational modification of AP-1 subunits (22). In addition, TATA-binding protein phosphorylation plays a role in MAPKinduced gene transcription. ERK2 directly phosphorylated the TATA-binding protein amino-terminal domain during phorbol 12-myristate 13-acetate-induced differentiation and G 0 -G 1 transition (59).
The data presented here highlight several similarities and differences between LPS-and cytokine-induced signal transduction. Both p38 MAPK and MEK1 activation were responsible for augmented AP-1 binding in both CM-and LPS/IFN-␥stimulated cells. CM-and LPS/IFN-␥-stimulated hiNOS promoter activation required intact ERK and p38 MAPK pathways. Deletion analysis of the hiNOS promoter confirmed that the two AP-1 sites, as well as the upstream NF-B site, are important for CM (10) and LPS/IFN-␥ stimulation of hiNOS induction. Finally, as shown for CM (10), LPS/IFN-␥ increased NF-B and AP-1 binding to the four key regulatory sequences in the hiNOS promoter.
Despite these similarities, LPS/IFN-␥ alone did not activate all of the signaling pathways required for full activation of the hiNOS promoter, as demonstrated by the much greater response to CM. Future studies can look for the coactivators, or additional signaling pathways, required for optimal gene transcription. Furthermore, in contrast to their effects in cells stimulated with CM, inhibitors of MEK1 and p38 MAPK reduced binding to the downstream NF-B site. Other investigators have suggested potential cross-talk between the MAPK and NF-B pathways (50,60,61). Although MAPK-dependent NF-B activation might be important for the induction of other genes, AP-1 binding is apparently the rate-limiting step in hiNOS promoter activation. The AP-1 complex JunD⅐Fra-2, or its binding site, serves as a molecular target for the control of diverse biological processes.