MEK1-dependent delayed expression of Fos-related antigen-1 counteracts c-Fos and p65 NF-kappaB-mediated interleukin-8 transcription in response to cytokines or growth factors.

Binding sites for the dimeric transcription factor activator protein (AP)-1 are found in numerous immunoregulatory and inflammatory genes. The precise mechanisms by which AP-1 activates or represses immune response genes and in particular the roles of individual AP-1 subunits in inflammatory responses are largely unknown. We report here that c-Fos and Fos-related antigen-1 (Fra-1), two inducible components of AP-1, are recruited to the endogenous interleukin (IL)-8 promoter in an IL-1-dependent manner. c-Fos activates IL-8 transcription and synergizes in this effect with p65 NF-kappaB. In contrast, Fra-1 strongly inhibits inducible IL-8 transcription. Fra-1 activation involves its stabilization, ubiquitination, and interaction with histone deacetylase-1. Blockade of MEK1 by PD98059 suppresses c-Fos and Fra-1 expression and, thus, affects two counteractive signals for IL-8 mRNA synthesis simultaneously. This disturbs the inducible recruitment of TATA box-binding protein and RNA polymerase II to the IL-8 promoter. Additional experiments reveal that, in conjunction with p65 NF-kappaB, the MEK1-ERK-dependent synthesis of c-Fos and Fra-1 serves to adjust the overall expression level of IL-8 in response to two of its physiological inducers, IL-1 and epidermal growth factor. Relative to c-Fos, the delayed recruitment of Fra-1 to the IL-8 promoter provides an example how AP-1 subunits may dampen excessive chemokine synthesis.

bers of the Jun, Fos, and ATF protein families (2). Jun (c-Jun, JunB, and JunD) and Fos (c-Fos, FosB, Fos-related antigen (Fra-1, Fra-2)) proteins act as dimeric transcription factors that can bind to the promoter regions of numerous mammalian genes. Fos and Jun proteins can also dimerize to many other basic leucine zipper proteins such as ATF, C/EBP, Maf, and NF-E2, increasing the number of the potential AP-1 factors that bind to a given AP-1 site (3). Furthermore, activity of AP-1 proteins is controlled at multiple levels, first by changes in gene transcription, mRNA turnover, and protein stability (4), second by interaction with non-basic leucine zipper transcription factors (3,5), and third by phosphorylation of their transactivation domains by mitogen-activated protein (MAP) kinases (4).
The number of AP-1 proteins and their complex control make it difficult to assess the impact of particular individuals on gene expression. Recently, the usage of genetically altered mice and cells derived from them has unraveled crucial functions for AP-1 as a regulator of cell life and death (6). However, very little information is available from these model systems on the role of AP-1 in inflammation and infection despite the fact that AP-1 binding sites are found in many genes activated during innate (and also adaptive) immune reactions (7). Because of the structural and regulatory complexity of AP-1, at present a prerequisite for defining its role in inflammation is to determine the composition of AP-1 complexes as a function of time and stimulus for individual genes, in particular those that are crucial for the inflammatory process (7). In this regard, a family of secreted proteins called chemokines is exceptionally important, because these factors control the process of leukocyte invasion into injured tissue, a hallmark of inflammation (8).
Interleukin (IL)-8, a prototypic human chemokine, was detected more than a decade ago as the founding member of the chemokine superfamily (9). IL-8 secretion and mRNA synthesis are strongly regulated by more than 100-fold by a plethora of external stimuli (10). A major part of this regulation occurs at the transcriptional level. A relatively small region of the 5Јflanking region of the IL-8 gene (nucleotides Ϫ1 to Ϫ133) is essential and sufficient for its transcriptional induction by most stimuli (10). This promoter element contains an NF-B element that is required for activation in all cell types studied as well as a single consensus AP-1 site whose sequence is tgactca (3,10). Unlike the binding site for NF-B, the AP-1 site as well as additional sites contained within the IL-8 core promoter, namely those for C/EBP, OCT-1 (11,12), or for NF-Brepressing factor (13) are not essential for induction but are required for maximal gene expression (10).
We and others have provided numerous lines of functional evidence that coordinated activation of the MAP kinases ERK, c-Jun NH 2 -terminal kinase, and p38 is important for IL-8 reg-ulation in response to proinflammatory cytokines, such as interleukin (IL)-1 (14), viruses (15), bacteria (16 -18), and various stress inducers (10). However, the precise molecular mechanisms by which all three types of MAP kinases regulate onset, duration, or extent of IL-8 mRNA synthesis have been largely elusive. The fact that MAP kinases regulate AP-1 activity suggests that they may induce IL-8 transcription via the single AP-1 site. However, MAP kinase also regulates other steps in the overall process of gene expression such as chromatin remodeling (19) or, as found by us (20,21), IL-8 mRNA stability. In addition, as for many other AP-1 target genes, there is a gap in the understanding of the actual composition of AP-1 complexes that bind to the endogenous IL-8 gene promoter (7). Several groups have addressed this point by studying the interaction of nuclear AP-1 proteins with IL-8 promoter-derived oligonucleotides in vitro using band-shift assays (22,23). However, these experiments cannot explain which AP-1 components interact with the two endogenous, chromatin-embedded alleles of the IL-8 gene in vivo (10).
In addressing these issues we have used the chromatin immunoprecipitation (ChIP) technique to analyze the downstream mechanisms by which the ERK MAP kinase pathway contributes to IL-1-inducible IL-8 gene expression. Hereby, we have focused on the role of c-Fos and Fra-1, two important AP-1 proteins that are expressed in many cell types and whose role in regulation of inflammatory genes has been largely elusive.
In this study, we show that IL-1 up-regulates c-Fos and Fra-1 in an ERK-dependent manner, a process that involves Fra-1 stabilization and ubiquitination. Furthermore, we show that c-Fos and Fra-1 bind to the endogenous IL-8 promoter and provide compelling evidence for a hitherto unrecognized role of Fra-1 in inducible expression of IL-8. Importantly, we find that Fra-1 counteracts the positive regulatory role of c-Fos and p65 NF-B on IL-8 gene expression. These data provide a mechanism of how c-Fos and Fra-1 enable transcriptional regulation of IL-8 to different levels in response to two physiological inducers of IL-8, namely IL-1 and EGF. Thus, our study establishes an example of how an activated MAP kinase pathway connects to two AP-1 proteins that regulate inflammatory gene expression.
ChIP-One 175-cm 2 flask of confluent KB cells treated as described in the figure legends was used for each condition. Proteins bound to DNA were cross-linked in vivo by replacement of the medium with warm PBS including 1% formaldehyde. After incubation for 1 min at room temperature, this solution was replaced by warm PBS including 0.125 M glycine to stop cross-linking. Then the supernatant was removed, and the cells were washed in and scraped into ice-cold PBS. Cells were collected at 500 ϫ g at 4°C, washed again in ice-cold PBS, and then lysed in ChIP-radioimmune precipitation assay buffer (10 mM Tris, pH 7.5, 150 mM NaCl, 1% Nonidet P-40, 1% deoxycholate, 0.1% SDS, 1 mM EDTA, and freshly added 1% aprotinin). Lysates were cleared by sonification (4 ϫ 1 min on ice) and centrifuged at 15,000 ϫ g at 4°C for 20 min. Supernatants were collected and stored in aliquots at Ϫ80°C for subsequent ChIP. 4 -10 l of antibodies were added to 250 -500 l of lysates, and the mixture was rotated at 4°C overnight. Then 40 l of a protein A/G mixture preequilibrated in ChIP radioimmune precipitation assay buffer was added to the lysates, and incubation was continued for 1 h at 4°C. Beads were collected by centrifugation and washed 2 times in 1.4 ml of ChIP-radioimmune precipitation assay buffer, once in high salt buffer (10 mM Tris, pH 7.5, 2 M NaCl, 1% Nonidet P-40, 0.5% deoxycholate, 1 mM EDTA), once in ChIP-radioimmune precipitation assay buffer, and once in TE buffer (10 mM Tris, pH 7.5, 1 mM EDTA) and finally resuspended in 55 l of elution buffer (10 mM Tris, pH 7.5, 1 mM EDTA, 1% SDS). Samples were vigorously mixed for 15min at 30°C and then centrifuged. 50 l of supernatant was diluted to 200 l with TE buffer including RNase A (50 g/ml). Similarly, 50 l of the initial lysate (input samples) were diluted to 200 l with TE buffer including 1% SDS and 50 g/ml RNase A. After 30 min at 37°C, proteinase K was added (0.5 mg/ml), and both input and immunoprecipitates were incubated for at least 6 h at 37°C followed by at least 6 h at 65°C. Samples were diluted to 500 l in loading buffer PB (Qiagen), and DNA was purified using Qiaquick spin columns (Qiagen) according to the manufacturer's instructions. DNA was eluted with TE buffer and stored at Ϫ20°C until further use. PCR was performed on input and immunoprecipitated DNA using 2.5 units of HotStart Taq polymerase (Qiagen), 1 M sense and antisense primer, 2.5-3 l of template DNA, 0.2 mM dNTPs in a total volume of 30 l. PCR cycles were 15 min at 95°C and 34 -36 cycles of 94°C (20 s), 55-60°C (20 s), 72°C (20 s) followed by a final extension reaction at 72°C for 7 min. PCR products were separated by agarose gel electrophoresis, visualized by ethidium bromide staining, and fluorescence intensities of bands were quantified using a Biometra TI5 system (Biometra) and the BioDocAnalyse software, Version 1.0 (Biometra). Selected ChIP results were quantified by real time PCR using a light cycler 1.0 instrument (Roche Applied Science) as indicated in the legends to Figs. 5 and 7C. The reaction mixture contained 3-4 l of ChIP or input DNA, 0.5 M primers, and 4 l of the light cycler DNA Master SYBR Green I kit (Roche Applied Science) in a total volume of 20 l. PCR cycles were 10 min at 95°C, 40 cycles of 95°C for 10 s, 60°C for 10 s, and 72°C for 10 s. Melting curve analysis revealed a single PCR product. Serial dilutions of input DNA revealed that PCR reactions were linear from 100 to 0.1 ng and were used to calculate absolute amounts of PCR products by the light cycler software. ChIP primers for the IL-8 gene (accession number M28130) were as follows: IL-8 promoter (sense, nucleotides 1303-1325, 5Ј-aagaaaactttcgtcatactccg-3Ј; antisense, nucleotides 1450 -1473, 5Ј-tggctttttatatcatcaccctac-3Ј); IL-8 upstream region (sense, nucleotides 407-430, 5Ј-atcatgggtcctcagaggtcagac-3Ј; antisense, nucleotides 519 -543, 5Ј-ggtgggagggaggtgttatctaatg-3Ј).
Reverse Transcription-PCR-1 g of total RNA was prepared as described below and transcribed into cDNA using Moloney murine leukemia virus reverse transcriptase (Invitrogen) in a total volume of 40 l. 5 l of this reaction mixture were used to amplify cDNAs using assays on demand (Applied Biosystems) for IL-8 (Hs00174103), c-Fos (Hs00170630), Fra-1 (Hs00759776), and ␤-actin (Hs00174103) and an ABI7500 real time PCR instrument. The threshold cycle (ct) for each individual PCR product was calculated by the instrument software, and ct values obtained for IL-8, c-Fos, and Fra-1 were normalized by subtracting the ct values obtained for ␤-actin. The resulting ⌬ct values were then used to calculate relative changes of mRNA expression as the ratio (R) of mRNA expression of stimulated/unstimulated cells according to the equation, R ϭ 2 Ϫ(⌬ct(stimulated)Ϫ⌬ct(unstimulated)) .
In Vivo Ubiquitination Assay-This assay was performed essentially as described in Treier et al. (26). Briefly, one 75-cm 2 flask of cells were transfected with 12.5 g of pMT107-His 6 -ubiquitin and expression vectors as indicated in the legend of For input controls proteins from 100 l of cell lysates were precipitated by adding 100 l of 10% trichloroacetic acid. Samples were incubated at 4°C overnight. The protein pellet was collected by centrifugation at 15,000 ϫ g at 4°C for 15 min, washed twice in ethanol, and dried in a Speed Vac. Proteins were redissolved by boiling in SDS-PAGE sample buffer. Purified and input proteins were separated by SDS-PAGE, and Fra-1 and MEK1 were detected by immunoblotting with the appropriate antibodies.
Coimmunoprecipitation-Cells were transfected or treated as indicated in the figure legends. After 1 day cells were lysed in Triton cell lysis buffer. 2 g of Fra-1 or HDAC1 antibody was added to 500 -1000 g of cell extract. After 16 h at 4°C, protein A-Sepharose was added for another h. Pellets were washed 3ϫ in immunoprecipitation buffer (20 mM Tris, pH 7.4, 154 mM NaCl, 50 mM NaF, 1 mM Na 3 VO 4 , 1% Triton X-100). Proteins were eluted with SDS sample buffer and detected by Western blot.
Oligonucleotide DNA Microarray Experiments-The microarray used in this study was the first version of the Inflammation Array (MWG Biotech) and contained 110 oligonucleotide probes for inflammatory genes that were previously validated by our laboratory (24) as well as 5 "housekeeping" genes. Total RNA from cells treated as indicated in the legend of Table I was purified with a Qiagen RNeasy kit followed by "on column" DNase I digestion (Qiagen). RNA was used to prepare Cy3labeled cRNA by oligo(dT)-T7-primed double-stranded cDNA synthesis (cDNA synthesis system, Roche Applied Science) followed by in vitro transcription with T7 polymerase (MEGAscript T7 kit, Ambion) as directed by the manufacturer. cRNA yield was determined photometrically. Equal amounts of cRNAs derived from ϳ2.5 g of total RNA were hybridized individually to microarrays in pre-prepared hybridization solution (MWG Biotech) at 42°C overnight and then washed sequentially in 2ϫ SSC (1ϫ SSC ϭ 0.15 M NaCl and 0.015 M sodium citrate), 0.1% SDS, 1ϫ SSC, and 0.5ϫ SSC. Hybridized arrays were scanned at maximal resolution on an Affymetrix 428 scanner at variable photomultiplier voltage settings. Fluorescence intensity values from Cy3 channels were processed using Imagene 4.2 software (Biodiscovery). To obtain maximal signal intensities without saturation effects, intensity values from TIFF images were integrated into one value per probe and normalized by the MAVI software (Version Pro 2.5.1, MWG Biotech).
Enzyme-linked Immunosorbent Assay-IL-8 protein concentrations in the cell culture medium were determined using the human IL-8 duo set kit (R&D Systems) exactly as directed in the manufacturer's instructions.

TABLE I
Sensitivity of inflammatory genes to MAP kinase inhibition in epidermal carcinoma cells and in primary fibroblasts KB epidermal carcinoma cells or human gingival primary fibroblasts cells were left untreated (0) or treated with 2 M SB203580 (SB), 20 M SP600125 (SP), 50 M PD98059 (PD), or all three inhibitors for 30 min and then stimulated with IL-1 (10 ng/ml) for 4 h. Subsequently, total RNA from each sample was used to prepare Cy3-labeled cRNA. cRNAs were hybridized individually to a customized DNA oligonucleotide microarray containing probes for 110 inflammatory genes. Gene expression in cells subjected to the indicated treatments was compared to that of unstimulated cells and is expressed by the indicated ratios of normalized fluorescence intensities. The relative mRNA expression for 22 genes that are induced by at least 2-fold by IL-1 is shown. Details of the microarray experiments are described under "Experimental Procedures." The complete set of data can be obtained from the corresponding author.

Partial Suppression of IL-1-induced IL-8 mRNA and Protein
Expression by PD98059 -Initially we compared the effects of the MAP kinase inhibitors SB203580, SP600125, and PD98059 on IL-1-inducible gene expression in the human epidermal carcinoma cell line, KB, and in human primary fibroblasts. As shown in Table I, we detected 22 genes that were induced by at least 2-fold in both cell types. As expected from the cooperative role of MAP kinases in activation of IL-1-induced genes (7, 10), treatment with individual MAP kinase inhibitors alone or in combination partially suppressed many genes; however, to variable extents. The most strongly induced gene was IL-8 (cxcl8). Interestingly, in both cell types the MEK1 inhibitor PD98059, a well characterized and highly specific allosteric inhibitor of the protein kinase MEK1 (28), had the strongest inhibitory effect on IL-8 gene expression compared with SP600125 or SB203580. Kinetic analysis by real time PCR confirmed that PD98059 partially suppressed IL-1-induced IL-8 mRNA expression in KB cells (Fig. 1A) as well as in human gingival fibroblasts and HEK293IL-1R cells (data not shown). PD98059 also partially suppressed the accumulation of secreted IL-8 protein (Fig. 1B). Thus, in three IL-1-responsive cell types PD98059 affected a mechanism that was not essential for IL-1-inducible IL-8 expression but was required to achieve maximal gene expression.
PD98059 Suppresses IL-1-induced mRNA and Protein Expression of c-Fos and Fra-1-We reasoned that the rapid inhibitory effect of PD98059 treatment was very likely to affect IL-1-induced IL-8 transcription via AP-1. We, therefore, tried to identify proteins that bind to the IL-8 promoter and that are affected by PD98059. In parallel to the increase in IL-8 mRNA, IL-1 induces a transient expression of c-fos (Fig. 1C) and of fra-1 mRNA (Fig. 1D). c-Fos mRNA expression peaked at 30 min of IL-1 stimulation, preceding the expression of Fra-1 mRNA, which peaked at around 1 h of IL-1 stimulation (Fig. 1, C and D). Virtually all of the c-Fos and Fra-1 protein was expressed in the nucleus (Fig. 1E). As at the mRNA level, the inducible expression of c-Fos protein preceded that of Fra-1 (Fig. 1E). Of note, both c-Fos and Fra-1 proteins separated by SDS-PAGE showed at least two forms of the proteins with different mobility, suggesting that IL-1 induces substantial posttranslational modifications of both proteins (Fig. 1E). Compared with the other major AP-1 components (c-Jun, JunB, JunD), c-Fos and Fra-1 mRNA and protein expression as well as their post-translational modifications were strongly inhibited by PD98059 (Fig. 1E) or by another MEK1 inhibitor, UO126 (data not shown).
c-Fos and Fra-1 Are Recruited to the Endogenous IL-8 Promoter-We, therefore, designed experiments to determine the effects of PD98059 on the recruitment of c-Fos and Fra-1 proteins to the endogenous IL-8 promoter. For this we amplified in ChIP experiments a fragment of 170 bp that covers the whole known regulatory region of the IL-8 gene including the AP-1 and NF-B sites, the TATA box region, and the putative polymerase II binding site ( Fig. 2A). ChIP conditions were optimized so that this fragment was only amplified after immunoprecipitation with specific antibodies directed at factors that bind to sites within this promoter region, e.g. p65 NF-B, or Where indicated, cells were stimulated for different length of times with IL-1 (10 ng/ml). In A, C, and D, IL-8, c-fos, and fra-1 mRNA expression levels were determined by real time PCR normalized to the expression of ␤-actin, and values are expressed relative to the unstimulated control. The mean values from duplicate determinations are shown. In B, IL-8 protein secreted into the cell culture supernatant was determined by specific enzyme-linked immunosorbent assay. Shown are the mean IL-8 levels from three independent experiments ϮS.E. E, for detection of protein expression of AP-1 components, proteins from nuclear extracts were separated by SDS-PAGE and c-Fos, Fra-1, JunD, c-Jun, and JunB proteins detected by Western blotting. Antibodies against ERK were used to confirm equal loading.
c-Fos and Fra-1 in IL-1-and EGF-induced IL-8 Expression polymerase II but not with irrelevant antibodies. In addition, no PCR product was obtained with p65 or polymerase II antibodies by using a primer set that covers a non-regulatory region far 5Ј within the IL-8 promoter (Fig. 2B).
A detailed kinetic analysis revealed that both c-Fos and Fra-1 proteins were induced and recruited to the IL-8 promoter in response to IL-1, c-Fos slightly preceding the maximum of Fra-1 recruitment (Fig. 2, C-F). Quantification of Western blot and ChIP results showed that the time-delayed IL-1-inducible recruitment of c-Fos and Fra-1 to the IL-8 promoter exactly mimicked the inducible expression pattern of both proteins (Fig. 2, D and F). We also observed some constitutive binding of Fra-1 to the IL-8 promoter (Fig. 2E), whereas there was very little c-Fos binding in unstimulated cells (Fig. 2E). Again, this was reflected at the protein level, as in whole cell lysates we observed a weak constitutive expression of a faster migrating form of Fra-1, whereas c-Fos expression was undetectable (Figs. 1C, 2C, and 3C). Importantly, PD98059 suppressed significantly the IL-1-inducible recruitment of c-Fos and Fra-1 to the IL-8 promoter by about 30 -50% (Fig. 2, G and H).
c-Fos and Fra-1 Have Opposing Effects on IL-8 Transcription-To assess the functional significance of c-Fos and Fra-1 for IL-8 transcription we expressed both proteins in HEK293IL-1R cells together with a luciferase reporter gene construct that contains the regulatory region of the IL-8 promoter as shown in Fig. 2. Expression of c-Fos, but not of Fra-1, induced IL-8 transcription (Fig. 3A). c-Fos synergized in IL-8 transcription when co-transfected with low amounts of a p65   (Fig. 3B). Similarly, c-Fos induced and Fra-1 inhibited IL-8 transcription in KB cells (Fig. 3C). These data strongly suggest that c-Fos is a positive regulator, whereas Fra-1 is a negative regulator of IL-8 transcription. To further test this hypothesis, we suppressed expression of endogenous Fra-1 in KB cells by RNA interference. As shown in Fig. 3D, suppression of IL-1induced Fra-1 protein expression (Fig. 3E) resulted in an ϳ2fold increase in activity of the cotransfected IL-8 promoter reporter gene. Collectively, these strongly indicate that the MEK1-ERK pathway utilizes c-Fos and Fra-1 to modulate IL-8 transcription. Because c-Fos and Fra-1 serve distinct functions, an activated ERK pathway contributes by a positive as well as by a negative signal to IL-8 mRNA synthesis. This mechanism may serve to modulate the overall IL-8 expression level in response to various external signals.
Fra-1 Interacts with HDAC1 and Is Stabilized and Ubiquitinated in a MEK1-dependent Manner-To search for potential mechanisms of the negative regulatory role of Fra-1, we inves- c-Fos and Fra-1 in IL-1-and EGF-induced IL-8 Expression tigated its interaction with putative corepressors of IL-8 transcription. HDAC1 is a histone deacetylase that interacts with p65 NF-B (29). p65 NF-B is the major NF-B subunit that regulates IL-8 transcription (10,13,29). Like p65 NF-B, HDAC1 binds to the IL-8 promoter (see Fig. 5). When Fra-1 and Myc-tagged HDAC1 were coexpressed, HDAC1 coimmunoprecipitated with Fra-1, indicating that HDAC1 and Fra-1 interact in intact cells (Fig. 4A). To analyze the relevance of ERK activation for this interaction, we coexpressed active MEK1R4F together with Fra-1 and HDAC1. These experiments revealed that MEK1R4F massively increases the amount of overexpressed Fra-1 protein (see Fig. 4B, third and sixth lanes). Longer exposures of the Western blot result also showed that MEK1R4F also substantially increases the amount of endogenous Fra-1 (Fig. 4B, second and fourth lanes). A surprising finding was that MEK1R4F strongly increased the interaction of MYC-HDAC1 with FRA1 (Fig. 4B, fourth and  sixth lanes). This very likely depends on posttranslational modifications of Fra-1 induced by MEK1R4F, because the relatively low amount of endogenous Fra-1 induced by MEK1R4F (Fig.  4B, fourth lane) coprecipitated the same amount of Myc-HDAC1 as the MEK1R4F-stabilized Fra-1 protein (Fig. 4B,  sixth lane). Moreover, without coexpression of MEK1R4F, the amount of MYC-HDAC1 coprecipitating with overexpressed Fra-1 was relatively low (compare Fig. 4B, the fourth and sixth  lanes with the fifth lane). Controls included in the experiments showed that the interaction of Fra-1 with Myc-HDAC1 was specific, as no coprecipitating bands were detected by the Myc antibody from cells expressing Fra-1 or MEK1R4F or MYC-HDAC1 alone or from the vector control. We also found that endogenous HDAC1 interacted with IL-1-induced Fra-1 by coimmunoprecipitation experiments (Fig. 4C). Although it is known that c-Fos is phosphorylated in an ERK-dependent manner (30), very little is known about posttranslational modifications of Fra-1. We noticed on Western blots that coexpression of Fra-1 with MEK1R4F not only resulted in an increase in Fra-1 but also in many Fra-1 forms with very high apparent M r on SDS-PAGE (data not shown). This suggested that Fra-1 might be ubiquitinated. Therefore, Fra-1 was coexpressed together with HIS-ubiquitin in the presence or absence of active MEK1R4F. As shown in Fig. 4D, affinity purification of ubiquitinylated proteins clearly shows that the overexpressed, stabilized Fra-1 is substantially modified by polyubiquitination in a MEK1R4F-dependent manner. Collectively, the results presented in Figs. 1 and 4 are intriguing because they suggest that the ERK pathway (via MEK1) induces the delayed expression of Fra-1, stabilizes (presumably by ubiquitination events) the protein, and at the same time promotes its interaction with HDAC1.
PD98059 Affects Assembly of the Preinitiation Complex at the IL-8 Promoter-To further assess the impact of PD98059 on the putative multiprotein complex that regulates IL-8 transcrip-

FIG. 4. Fra-1 interacts with HDAC1 and is stabilized and ubiquitinated in an MEK1-dependent manner.
HEK293IL-1R cells were transiently transfected as indicated with 2.5 g of the expression plasmids for HA-MEK1R4F, FLAG-Fra-1, MYC-HDAC1, and combinations thereof or empty vector control (first lane). A and B, 24 h after transfection cells were lysed, and Fra-1 was immunoprecipitated (IP) from cell extracts. Coprecipitated Myc-HDAC1 was detected by anti-Myc antibodies. Lysates were probed with antibodies against Fra-1, Myc, and HA epitopes to detect expression of Fra-1, MYC-HDAC1, and HA-MEK1RF4, respectively. B, two different exposure times for the Fra-1 Western blot (WB) are shown to indicate the massive increase in Fra-1 protein in the presence of active MEK1R4F. IgGhc and IgGlc; heavy and light chains of immunoglobulins. C, KB cells were treated for 3 h with IL-1 (10 ng/ml) or left untreated. Thereafter cells were lysed, and HDAC1 was immunoprecipitated. Co-precipitated Fra-1 was detected by Western blotting. Expression of Fra-1 and HDAC1 was confirmed by Western blots of lysates (lower panel). D, HEK293IL-1R cells were transfected with 12.5 g of the expression plasmids for HA-MEK1R4F, FLAG-Fra-1, or His-ubiquitin as indicated. 48 h later cells were lysed in denaturing buffer, precipitated, and redissolved in sample buffer (Input). His-ubiquitinated proteins were purified by adsorption to Ni 2ϩ -Sepharose (Ni 2ϩpulldown). Proteins were separated by SDS-PAGE, and Fra-1 and MEK1R4F were detected in the input sample and the bound fraction using antibodies against Fra-1 and HA, respectively. Ub (n) indicates the polyubiquitinated form of Fra-1. tion, we analyzed its effects on histone acetylation, on p65 NF-B, and on assembly of the preinitiation complex. As shown in Fig. 5 in unstimulated cells the overall acetylation level of histones H3 and H4 within the IL-8 promoter region is high and is only weakly induced by IL-1 by about 1.5-2-fold. There is also constitutive binding of HDAC1. In addition, there is a strong transient increase in binding of p65 NF-B and of TBP, respectively. As a result IL-1 induces a Ͼ10-fold increase in polymerase II binding within 30 min. Together with the ChIP experiments shown in Fig. 2, these data suggest that strong polymerase II recruitment requires cumulative binding of c-Fos, Fra-1, p65, and TBP. Cooperative binding and dynamic exchange of factors is facilitated by nucleosome remodeling, as indicated by the substantial histone tail acetylation.
PD98059 weakly suppressed binding of HDAC1 at 8 h of IL-1 stimulation and had no significant effect on the overall histone acetylation level or on p65 NF-B. However, in addition to preventing binding of c-Fos and Fra-1 (Fig. 2G) PD98059 significantly suppressed binding of TBP of IL-1 stimulation. Finally, PD98059 suppressed recruitment of polymerase II by 30 -50% after 3 and 8 h of IL-1-stimulation. Hence, the partial inhibition of c-Fos, Fra-1, TBP, and polymerase II recruitment reflects the partial inhibition of IL-1-induced IL-8 mRNA as observed in the experiments shown in Fig. 1. Collectively, these results provide a mechanism for the downstream action of PD98059 and suggest that c-Fos and Fra-1 serve to fine-tune IL-8 transcription.
Regulation of IL-8 Expression by EGF-To further follow this hypothesis, we investigated the regulation of IL-8 gene expression by EGF, which like IL-1 is also a physiological inducer of IL-8. However, EGF causes only a 10 -20-fold increase in IL-8 expression, which is significant but much weaker than that obtained with IL-1 (Fig. 6A). A direct comparison revealed that IL-1 activates transcription from an IL-8 reporter gene construct strongly, more than 20-fold, through the NF-B and AP-1 cis elements (Fig. 6B), whereas EGF induces IL-8 transcription weakly by about 2-3-fold and mainly through the AP-1 site (Fig. 6B). This difference does not reflect an inadequate response of the cells to EGF as c-Fos and Fra-1 proteins were induced by EGF with the same strength and kinetics compared with IL-1 (compare Fig. 1C with Fig. 7A). As for IL-1, the EGF-induced expression of c-Fos and Fra-1 were suppressed by PD98059 as was their EGF-induced recruitment to the IL-8 promoter (Fig. 7).
In contrast to IL-1, EGF-induced only a weak, delayed recruitment of p65 NF-B to the IL-8 promoter (Fig. 8)  c-Fos and Fra-1 in IL-1-and EGF-induced IL-8 Expression not affected by PD98059 (Fig. 8). In agreement with its low potency to induce IL-8 expression and p65 NF-B recruitment, EGF induced only a modest increase in polymerase II binding, which was almost completely inhibited by PD98059 at 0.5 h and 24 h of treatment (Fig. 9, A and B). Finally, we found that the EGF-induced IL-8 secretion was significantly suppressed by PD98059 (Fig. 9C), suggesting that EGF-induced IL-8 gene expression is almost entirely controlled by the ERK pathway via c-Fos and Fra-1. Collectively, the data shown in Figs. 6 -9 suggest that EGF activates IL-8 weakly because it mainly utilizes the AP-1 site of the IL-8 promoter and provides a molecular mechanism that explains the weaker ability of EGF compared with IL-1 to induce the endogenous IL-8 gene.

DISCUSSION
The transcription factor AP-1 was one of the first mammalian transcription factors to be identified, but its physiological functions are still being uncovered. The precise mechanisms by which AP-1 activates or represses genes in response to external challenge, the target genes that mediate AP-1 function, and the relevance of AP-1 for human disease are still not clear. Furthermore, until recently research on AP-1 has been centered around the role of c-Jun and c-Fos in cell death and survival (1, 2, 6).
In this study we have focused on the role of AP-1 in regulation of the human chemokine interleukin-8 by the proinflammatory cytokine interleukin-1. Both IL-1 and IL-8 are central for the inflammatory response, and elucidating the mechanisms by which IL-1 regulates IL-8 provides a model of general significance for the expression of inflammatory genes. As reported recently, we have not found a role for c-Jun in IL-1inducible IL-8 secretion as assessed by cell-permeable peptides that inhibit c-Jun NH 2 -terminal kinase-mediated c-Jun phosphorylation or by RNAi-mediated c-Jun suppression (24).
In this study we have focused on the role of c-Fos and Fra-1 in regulation of the interleukin-8 gene. c-Fos (encoded by fos) and Fra-1 (encoded by fos-like gene 1, fosl1) show high homology in their leucine zipper and DNA binding domains, but outside of these domains both proteins show little similarity (31). Both, c-Fos and Fra-1 have long been known as growth factor regulated genes (32)(33)(34)(35), and likewise c-Fos was found a long time ago to be induced by the proinflammatory cytokines IL-1 and tumor necrosis factor (36,37).
Despite this fact, the role of c-Fos in IL-1-mediated gene expression has not been defined. We found that c-Fos is recruited to the endogenous IL-8 promoter in an IL-1-dependent manner, transactivates an IL-8 promoter-driven reporter gene, and synergizes in this effect with p65 NF-B, a known positive regulator of IL-8 transcription (10). Moreover, IL-1-inducible expression of c-Fos was completely suppressed by the ERK pathway inhibitor PD98059, suggesting that IL-1-induces c-Fos mainly through ERK (Fig. 1). We also found that PD98059 suppressed IL-1-induced shifts in the apparent M r of c-Fos on SDS-PAGE (Fig. 1). Presumably, this shift is caused by multisite phosphorylation of the carboxyl-terminal transactivation domain of c-Fos as observed by many groups (30, 38 -40), and in fact two reports have recently reported that c-Fos is a direct substrate of ERK in vitro and in vivo (30,40). Phosphorylation potentiates c-Fos-mediated transactivation of AP-1 reporter genes (30,40). In agreement with these data we show here that selective activation of the ERK pathway by active MEK1 synergizes with coexpressed c-Fos to induce IL-8 transcription (Fig. 3B). Thus, our data clearly show that a MEK1-ERK-c-Fos pathway connects the IL-1 receptor to the IL-8 promoter.
In contrast to c-Fos, very little is known about a role of Fra-1 in IL-1 signaling. The induction of Fra-1 by IL-1 occurs mainly in the nucleus and is very strong and slightly delayed compared with that of c-Fos (Fig. 3, 5). A possible explanation for this delay in expression might be that c-Fos is required for cytokineinduced Fra-1 transcription, a mechanism that has been so far only observed in fibroblasts or osteoclasts by conditional activation of c-Fos-ER fusion proteins (41,42) or during mouse osteoclast differentiation induced by RANKL (42).
As for c-Fos, PD98059 completely suppressed IL-1-induced Fra-1 mRNA and protein synthesis, suggesting that Fra-1 expression is controlled by ERK (Figs. 3 and 4). This finding is in agreement with findings in other systems where PD98059 suppressed Fra-1 expression mediated by the mitogen lysophosphatidic acid or conditional activation of the ERK pathway by RAF or MEK1 mutants, which also activate Fra-1 gene expression (35,43). Another study recently implied the c-Jun NH 2terminal kinase pathway in tumor necrosis factor-mediated Fra-1 expression (44). Our results clearly show that ERK is essential for IL-1-inducible Fra-1 expression but do not rule out an additional role of other IL-1-induced signaling pathways such as c-Jun NH 2 -terminal kinase.
Apart from its role in bone metabolism and certain aspects of tumor biology, little is known about Fra-1 functions or about putative Fra-1 target genes (41,(45)(46)(47)(48). Fra-1 transgenic mice develop osteosclerosis (49), whereas Fra-1 ablation in mice is embryonic lethal (50). Interestingly, Fra-1 insertion into the c-Fos locus can fully rescue the osteopetrosis phenotype caused by c-Fos gene inactivation in mice (51,52). In contrast, Fra-1 cannot substitute for c-Fos target genes in mouse embryonic fibroblasts (51). These data suggest that both proteins may have overlapping but also distinct functions.
Controversy exists about the transactivation potential of Fra-1 and the molecular basis for its actions. Several reports have found that Fra-1 lacks an autonomous transcriptional activation domain (41,46,48). In agreement with this, Fra-1 repressed an AP-1 reporter gene induced by c-Fos, c-Jun, or TPA, and it was concluded that Fra-1 counteracts AP-1 activation in tumor promotion (53). However, Fra-1 knock-in rescues the c-Fos Ϫ/Ϫ phenotype in vivo, and it is hard to reconcile how Fra-1 can compensate for loss of c-Fos if it were only a negative regulator of transcription (45,51). Thus, it is likely that Fra-1 can also positively regulate gene expression. In fact, others have found that GAL4-Fra-1 fusion proteins can be stimulated by ERK pathway activation (25,54,55). Furthermore, the transcriptional activity of a tethered Fra-1-c-Jun heterodimer FIG. 9. Weak EGF-induced IL-8 secretion correlates with weak recruitment of polymerase II to the IL-8 promoter and is suppressed by PD98059. KB cells were treated as described in the legend of Fig. 7. In A recruitment of polymerase II to the IL-8 promoter was determined by ChIP. PCR products were visualized by ethidium bromide staining after gel electrophoresis (A). In B the intensities of bands were quantified, and the results from four ChIP experiments are expressed as mean fold change ϮS.E. In C IL-8 protein secreted into the cell culture supernatant was determined by specific enzyme-linked immunosorbent assay. Shown are the mean IL-8 levels from four independent experiments ϮS.E. Gray bars represent treatment with PD98059. Pol II, polymerase II. c-Fos and Fra-1 in IL-1-and EGF-induced IL-8 Expression was higher than that of a c-Jun homodimer and varied among three distinct promoter constructs tested, suggesting that the capacity of Fra-1 to activate gene transcription is promoterdependent (56).
We report here that Fra-1 strongly inhibits IL-8 transcription in response to coexpressed c-Fos or p65 NF-B or both (Fig.  6A). Furthermore, MEK1-induced IL-8 transcription is potentiated by coexpressed c-Fos or p65 NF-B but strongly inhibited by Fra-1 (Fig. 6B). These data clearly show that Fra-1 is a negative regulator of IL-8 transcription. Thus, apparently IL-1 induces the Fra-1 protein, which in turn limits IL-8 gene expression. Thus, our data support the model of Fra-1 as a negative regulator of gene expression (53) and highlight the importance of analyzing the specific contribution of Fra-1 to expression of individual genes. Our data propose a negative regulatory role for Fra-1 in regulation of a natural chemokine promoter and thereby identify a gene relevant for innate or adaptive immune responses as a bona fide Fra-1 target gene. We also demonstrate that at least for the IL-8 gene, c-Fos and Fra-1 have distinct functions. As these findings contrast with the seemingly functional equivalence of c-Fos and Fra-1 in vivo (45,51), it is tempting to speculate that in mouse transgenic models differences in c-Fos and Fra-1 function may only become apparent after inflammatory or infectious challenge of the animals.
Fra-1 repression is likely to involve several mechanisms. First, an AP-1 complex containing Fra-1 may displace a transcriptionally active AP-1 heterodimer containing c-Fos at the AP-1 site of the IL-8 promoter. Although we have so far not found recruitment of c-Jun (data not shown), we have preliminary evidence that at the early times of IL-1 stimulation JunB binds to the IL-8 promoter, whereas later an increase in JunD is found. 2 Second, we report here that Fra-1 interacts with histone deacetylase 1 (HDAC). The interaction of HDAC1 with Fra-1 may contribute to the overall level of HDAC1 recruitment to the IL-8 gene. In agreement with that we found that PD98059, which suppresses Fra-1 expression, also suppressed HDAC1 binding to the IL-8 promoter after 8 h of IL-1 stimulation. We also observed an interaction of p65 NF-B with HDAC1 (data not shown), confirming the results of others (29). Of note, HDAC1 is present at the IL-8 promoter already in unstimulated cells. It is, therefore, likely that other factors in addition to Fra-1 and p65 NF-B contribute to HDAC1 recruitment.
Similar to the interaction of Fra-1 with HDAC1 as reported here, it was shown recently that the interaction of c-Jun with a histone deacetylase mediates gene repression (57). Posttranslational modifications of c-Jun by its upstream activator c-Jun NH 2 -terminal kinase result in dissociation of the interaction and thereby activate c-Jun-mediated transactivation (57). In the presence of active MEK1 the interaction of coexpressed Fra-1 with HDAC1 was enhanced (Fig. 4). Active MEK1 drastically stabilizes Fra-1 and produced species of higher M r detectable on SDS-PAGE (Fig. 4). Equally, IL-1 induced a mobility shift of endogenous Fra-1 that was PD98059-sensitive (Figs. 1 and 2). Both types of experiments indicate that modifications of Fra-1 largely depend on the MEK1-ERK pathway. These modifications may represent phosphorylation of Fra-1, which was recently described in Ras-transformed thyroid cells (55). However, the experiment shown in Fig. 4 provides the first evidence that active MEK1 induces strong polyubiquitination of ectopically expressed Fra-1. This mechanism may help to stabilize the protein. In support of our findings, ERK was recently found to protect Fra-1 against proteasomal degrada-tion in colon carcinoma cells (58). These observations distinguish the regulation of Fra-1 by the MEK1-ERK pathway from the short-lived c-Fos protein whose ERK-dependent transient expression results from proteasomal degradation (59 -61). Our data also suggest that in contrast to the c-Jun NH 2 -terminal kinase-c-Jun example, MEK1-ERK-mediated modifications stabilize the interaction with negative regulators of gene expression such as histone deacetylases. Further experiments will be designed to reveal in detail the role of ubiquitination for gene-repressive functions of Fra-1.
The role of the Fra-1-HDAC1 complex in IL-8 transcription awaits further investigation. HDAC1 might affect acetylation of individual lysines of histones H3 or H4 or, alternatively, might deacetylate other proteins that are important for IL-8 transcription. The detection of substantial histone H3 and histone H4 acetylation already in unstimulated cells, the permanent presence of histone deacetylases, and the rapid recruitment of transcription factors and RNA polymerase II within half an hour of IL-1-stimulation suggest that the chromatin structure of the IL-8 promoter is kept in an open, accessible conformation to allow very fast IL-8 mRNA synthesis. In fact, the IL-8 gene behaves like an immediate early gene as our results show that IL-8 mRNA is rapidly increased in parallel to c-fos mRNA. Low, constitutively expressed amounts of AP-1 proteins in conjunction with the IL-1induced nuclear translocation of p65 NF-B might determine IL-8 transcription within the first 30 min of IL-1 treatment. Thereafter de novo synthesis of c-Fos and various Jun proteins might serve to maintain high levels of mRNA synthesis. According to our data, in this model at 5-8 h after stimulation by IL-1 the decline in p65 NF-B binding and the increase in Fra-1 binding slowly start to shut off transcription. In this situation the stabilization of IL-8 mRNA may primarily maintain IL-8 mRNA steady state levels.
The data obtained during the first part of the study suggested that PD98059 suppressed two counteractive signals for 2 E. Hoffmann, unpublished data. c-Fos and Fra-1 in IL-1-and EGF-induced IL-8 Expression IL-8 mRNA synthesis, a positive via c-Fos and a negative via Fra-1. The substance does not affect p65 NF-B. Hence, on balance PD98059 can only partially suppress IL-1-inducible IL-8 transcription, as exemplified by its partial inhibition of recruitment of key molecules of the preinitiation complex, namely TBP and polymerase II (Fig. 5). As a net effect PD98059 leads to the partial inhibition of IL-1-induced synthesis of IL-8 mRNA and protein that was the starting point for this study (Fig. 1) and cannot suppress IL-1-induced IL-8 mRNA expression completely (Fig. 1A). Another prediction from these experiments was that sole activation of ERK and subsequent c-Fos and Fra-1 expression should not result in maximal IL-8 expression levels. To test the physiological significance of this hypothesis more directly, we analyzed the transcriptional regulation of the IL-8 gene by EGF. Compared with IL-1, EGF is a weak inducer of IL-8 (Fig. 6). In contrast, Fra-1 and c-Fos genes were induced by EGF with a strength and kinetics comparable with that obtained with IL-1 (Fig. 7A). EGF induces c-Fos and Fra-1 recruitment to the endogenous IL-8 promoter followed by a weak recruitment of RNA polymerase II (Figs. 7 and 9). In addition, EGF-induced IL-8 transcription was largely dependent on the AP-1 rather than on the NF-B site (Fig. 6B). In line with this, EGF failed to cause a strong and rapid recruitment of p65 NF-B to the endogenous IL-8 promoter (Fig. 8). Finally, the EGF-induced expression of c-Fos and Fra-1, the recruitment of polymerase II, and the activation of the endogenous IL-8 gene were significantly suppressed by PD98059, suggesting that regulation of IL-8 by EGF is almost entirely controlled by the MEK1-ERK-AP-1 pathway (Figs. 7 and 9). Thus, the data in this study provide a model that explains how two different inducers of IL-8, i.e. IL-1 or EGF, activate the gene to different levels within the same cell.
Chemokine-dependent leukocyte recruitment is required during all phases of wound healing (62). The molecular mechanism of IL-8 gene expression described here is likely to be of (patho)physiological importance as extensive leukocyte infiltration during the initial phase of tissue damage requires high levels of IL-8 and other chemokines induced by proinflammatory cytokines such as IL-1, whereas at later stages during wound healing chemokine production may be mainly under the control of locally produced growth factors such as EGF. In this phase lower amounts of leukocytes are likely to be required for tissue remodeling (62).
Collectively the data presented in this study define a hitherto unrecognized important role of Fra-1 in IL-1-mediated gene expression and also demonstrate how the ERK MAP kinase pathway modulates p65 NF-B-mediated chemokine gene expression via c-Fos and Fra-1. We further demonstrate that an inflammatory gene can be expressed within the same cell to various levels in response to physiological stimuli by combinatorial usage of two signaling pathways (ERK and NF-B) that modulate the expression and the activity of positive and negative transcriptional regulators (Fig. 10).
The number of signaling pathways that is utilized by infectious or inflammatory stimuli to induce hundreds of inflammatory genes is small, mainly composed of NF-B and the MAP kinases ERK, c-Jun NH 2 -terminal kinase, and p38 (7). The time-delayed expression of Fra-1 may, therefore, also represent a "dampening" or "switch off" mechanism relevant for many other inflammatory proteins whose levels of expression are under control of the NF-B and MAP kinase pathways.