Lipopolysaccharide induction of THP-1 cells activates binding of c-Jun, Ets, and Egr-1 to the tissue factor promoter.

These studies examine the molecular basis for increased transcription of tissue factor (TF) in THP-1 cells stimulated with lipopolysaccharide (LPS). DNase I footprinting identified six sites of protein-DNA interaction between −383 and the cap site that varied between control and induced extracts. Four footprints show qualitative differences in nuclease sensitivity. Footprints I (−85 to −52) and V (−197 to −175) are induction-specific and localize to regions of the promoter that mediate serum, phorbol ester, partial LPS response (−111 to +14), and the major LPS-inducible element (−231 to −172). Electrophoretic mobility shift assays with the −231 to −172 probe demonstrate JunD and Fos binding in both control and induced nuclear extracts; however, binding of c-Jun is only detected following LPS stimulation. Antibody inhibition studies implicate binding of Ets-1 or Ets-2 to the consensus site between −192 and −177, a region that contains an induction-specific footprint. The proximal region (−85 to −52), containing the second inducible footprint, binds Egr-1 following induction. These data suggest that LPS stimulation of THP-1 cells activates binding of c-Jun, Ets, and Egr-1 to the TF promoter and implicates these factors in the transcriptional activation of TF mRNA synthesis.

Factor VII/VIIa bound to the cellular receptor tissue factor (TF), 1 initiates both the intrinsic and extrinsic coagulation pathways by activating Factors IX and X (reviewed in Refs. 1,2). Constitutive expression of TF is detected in many cell types that do not normally contact blood, providing a hemostatic barrier to initiate coagulation following injury (3). Ischemic tissue damage following intravascular clotting (Schwartzman reaction) is produced when LPS-stimulated monocytes are introduced into leukopenic rabbits (4). LPS induction of TF expression in monocytes of patients with sepsis causes disseminated intravascular coagulation, which has a high mortality rate in humans. In the baboon model, administration of either anticoagulants (5) or neutralizing antibodies against TF (6) prevents LPS-induced disseminated intravascular coagulation.
Monocytes are the only circulating cells that modulate expression of TF, and synthesis can be up-regulated by a number of agonists including tumor necrosis factor-␣ (7), lipopolysaccharide (LPS) (8,9), and phorbol ester (PMA, Ref. 10). TF mRNA is very unstable with a half-life of 60 -90 min. In LPSinduced monocytes, the increase in TF mRNA results from transcriptional activation (8), whereas, in the monocytic cell line THP-1, LPS induces both an increase in transcription and changes in mRNA stability (11). Functional studies have demonstrated that promoter sequences between Ϫ383 to ϩ121 support high level constitutive expression (12). In the context of the TF promoter, two regions appear to be involved in maximal induction in LPS-stimulated THP-1 cells (Ϫ227 to Ϫ172 and Ϫ96 to Ϫ33); however, a short oligonucleotide from the TF promoter (Ϫ192 to Ϫ172) that contains an isolated NF-B/Rel motif is sufficient for LPS induction. A minimal promoter from Ϫ111 to ϩ121 mediates basal expression and serum induction in Cos-7 cells (13), serum and PMA induction in HeLa cells (14), and tumor necrosis factor-␣ induction in endothelial cells (15). We have also determined that the Ϫ111 to ϩ8 promoter mediates LPS induction in THP-1 cells and PMA induction in HL-60 cells. 2 In the present study, we have characterized protein interactions with the TF promoter using nuclear extracts derived from control and LPS-induced THP-1 cells. DNase I footprint analysis showed six regions of nuclease protection and two sites of LPS-induced protein binding. Electrophoretic mobility shift assays (EMSAs), oligonucleotide competitions, and antibody supershift studies were used to characterize the specific proteins that interact with DNA at these sites. Sp1 and Sp3 interact with the promoter in both control and induced extracts. Upon induction there is a change in the AP-1 binding from JunD/Fos in control cells to c-Jun/Fos and JunD/Fos in LPS-induced cells. In contrast to other reports, we have no evidence for NF-B or Rel binding to the Ϫ192 to Ϫ172 site; rather, our data suggest that either Ets-1 or Ets-2 binds to this sequence following induction. Finally, in induced nuclear extract, we detect binding of Egr-1 to the consensus site between Ϫ85 and Ϫ52. These data suggest that LPS stimulation of TF expression results from induction of c-Jun, Ets, and Egr-1 binding to the TF promoter. monitor induction of TF expression, mRNA was examined by Northern blot analysis using a TF cDNA probe.
DNase I Footprinting-The DNA fragments used as probes in the footprinting analysis were excised by restriction digestion of plasmids that have been described previously (20). The 167-bp fragment (Ϫ280 to Ϫ113) was excised from pTFPSma, and the 224-bp fragment containing the upstream Ϫ383 to Ϫ280 region fused to the downstream Ϫ113 to ϩ8 sequence was cut from pTFPX⌬Sma. These fragments were uniquely end-labeled with the large fragment of DNA polymerase I (Klenow) at either the EcoRI site with [␣-32 P]dATP or the XbaI site with [␣-32 P]dCTP. DNase I footprinting was performed as described previously (20). DNA binding reactions (50 l; 12 mM Hepes, pH 7.9, 60 mM NaCl, 4.8 mM MgCl 2 , 10% glycerol, 2.4 mM dithiothreitol, 0.12 mM EDTA, 0.4 mM PMSF, and 0.4 g/ml each leupeptin, aprotinin, and antipain) contained 20 g of nuclear extract, 0.15-0.25 ng (10,000 -20,000 cpm) of end-labeled probe, 1 g of dI-dC⅐dI-dC (Pharmacia LKB Biotechnology, Inc.), and 0.1 g/ml bovine serum albumin and were incubated for 1 h on ice followed by a 2-min incubation at room temperature. The samples were digested with DNase I (Worthington) freshly diluted in DNase I dilution buffer (25 mM Hepes, pH 7.9, 25 mM MgCl 2 , 1 mM DTT, and 0.1 mg/ml bovine serum albumin) for 60 s, and the reaction was terminated by the addition of 100 l of stop solution (20 mM EDTA, pH 8.0, 1% SDS, 0.2 M NaCl, 100 g/ml yeast tRNA, and 0.1 mg/ml proteinase K) followed by a 1-h incubation at 45°C. Digestion products were extracted with phenol:chloroform, ethanol-precipitated, dried briefly, and resuspended in loading buffer (95% formamide, 0.1% bromphenol blue, 0.1% xylene cyanol). Samples were heated to 95°C for 3 min and cooled on ice. The products were resolved on 6% acrylamideurea gels run in 1 ϫ TBE (0.089 M Tris borate, 1 mM EDTA), transferred to Whatman No. 3MM paper, dried, and visualized by autoradiography.

RESULTS
DNase I Footprint Analysis-DNase I footprinting analysis was performed to visualize protein-DNA interactions within the TF promoter. In Fig. 1, protein interactions with the 224-bp fragment containing the upstream region (Ϫ383 to Ϫ280) fused to the downstream region (Ϫ113 to ϩ8) demonstrate two footprints. Footprint I (Ϫ85 to Ϫ52) is inducible when control and LPS-induced extracts are compared. Footprint II (Ϫ105 to Ϫ90) is present with both control and induced extract on the coding strand (Fig. 1A); however, stronger protection is seen with induced extract on the noncoding strand (Fig. 1B). Although other regions of this probe display differences in nuclease sensitivity when extracts containing protein are compared with the DNase control lane, the patterns of cleavage with control and induced extracts are very similar, suggesting that those proteins do not change with induction. Reduced cleavage of nucleotides between Ϫ50 and the mRNA start site (Fig. 1A) may result from binding of TATA binding protein and RNA polymerase to sequences in this region.
Footprints generated with the l67-bp probe (Ϫ280 to Ϫ113) are shown in Fig. 2. Protein(s) derived from both control and induced extracts generated footprint III (Ϫ150 to Ϫ120), although induced extracts protected this region of DNA slightly more efficiently. Proteins that protected the nucleotides within footprints IV and VI were present in both control and induced extracts. In contrast, footprint V was only detected when LPSinduced extracts were used. In addition, a strong DNase I-hypersensitive site, located at Ϫ197, appeared at the junction of footprints V and VI on the coding strand with the LPS-induced extracts. As summarized in Fig. 3, footprints IV (Ϫ175 to Ϫ153), V (Ϫ197 to Ϫ175), and VI (Ϫ220 to Ϫ197) contain consensus binding sites for numerous proteins. The perturbations in DNase I cleavage visualized when control and induced nuclear extracts are compared suggest that LPS may qualitatively or quantitatively affect the proteins available to interact with these sequences.
Characterization of Proteins That Bind to the TF Promoter-To identify nuclear proteins that interacted with the footprinted regions, EMSAs were performed. These studies focused on two regions that contained inducible footprints and appeared to change most dramatically following LPS induction, the distal region (Ϫ231 to Ϫ172) and the proximal region (Ϫ99 to Ϫ47). The probes used in these studies are indicated in Fig.  3, which also shows the region of the promoter, transcription factor consensus sites, and footprints contained in each oligonucleotide.
Oligo 60, which contains two AP-1 sites and an overlapping NF-B/Rel/Ets site, interacts with proteins in both control and induced extracts, and there was a qualitative difference in the pattern of migration of the complexes (data not shown). In order to improve these binding studies, Oligo 87 (Ϫ231 to Ϫ145) was synthesized to extend the region slightly and include an Sp1 sequence (Ϫ167 to Ϫ162) and a CACCC sequence (Ϫ157 to Ϫ153). Competitive gel shift analysis performed with Oligo 87 is displayed in Fig. 4. Specific binding to this probe was detected with both control and induced extracts, although the induced extract produced a broader area of protein-DNA complex migration. Binding was specific as indicated by the effective competition of a 100 ϫ molar excess of unlabeled probe. The addition of a 100 ϫ molar excess of an AP-1 consensus oligo competed for most of the binding activity; an Sp1 consensus oligo competed for some of the binding activity in the LPS-induced extract, and the combination of both the AP-1 and Sp1 effectively competed for binding in both control and LPSinduced extracts. No competition was observed with the NF-B or Ets consensus oligos.
Most of the binding to Oligo 60 was also competed with the AP-1 consensus oligo, although no competition was observed when consensus oligos for NF-B, Ets, or Sp1 were included in the binding reaction (data not shown). Antibody supershift assays were performed with Oligos 60 and 87 to characterize the transcription factors interacting with this region (Fig. 5). Induced extract contains an increased level of binding to these probes, which may contribute to the difference in mobility of protein-DNA complexes seen when control and induced extracts are compared. In Fig. 5A, proteins bound to Oligo 60 were recognized by antibodies against both the Jun and Fos families of transcription factors since these antibodies supershifted the protein-DNA complexes (indicated by arrows to the right of the EMSA). The ␣-Jun antibody effectively supershifted the entire binding complex with the control extract and also produced a supershifted complex with the LPS-induced extract, although this complex was more diffuse and there was residual binding that was not completely supershifted. The ␣-Fos antibody also produced supershifted complexes with both the control and induced extracts. These supershifted complexes appeared to be similar in both the control and induced extracts; however, this complex appeared to slightly more abundant with the LPS-induced extract. ␣-Jun and ␣-Fos antibodies together supershifted all of the binding complexes that recognized Oligo 60.
The binding complexes formed with Oligo 87 were tested with the same ␣-Jun and ␣-Fos antisera (Fig. 5B). Although both antibodies produced supershifted complexes when tested separately and in combination, there was residual binding activity present in both control and induced extracts. The protein complexes formed with Oligo 87 were tested with additional antibodies (Fig. 5C), and the incubations were extended to 90 min to increase the sensitivity of the assay. The longer incubation resulted in the appearance of a slower migrating complex which was much stronger with induced extract (indicated by the dot to the right of the EMSA). Both ␣-Sp1 and ␣-Jun antisera produced a supershifted complex indicated by the arrow at the right of C. Neither ␣-Rel nor ␣-NF-B (data not shown) antibodies were able to recognize proteins that interacted with Oligo 87. In contrast, the ␣-Ets antibody abrogated the binding of the upper complex which formed upon longer incubations. This higher mobility complex was also missing in reactions containing LPS-induced extract and the ␣-Sp1 and ␣-Jun antisera.
Since the TF promoter contains two AP-1 consensus sequences, we were unable to determine whether or not Jun homodimers, Jun-Fos heterodimers, or a mixture of both were present in the control and induced nuclear extracts using either Oligo 60 or Oligo 87 (Fig. 5). Fig. 6 demonstrates the results of EMSA with a probe containing a single AP-1 consensus site. These data suggest that all of the AP-1 binding activity consists of Jun-Fos heterodimers, since the ␣-Fos antisera supershifts all of the complexes in both control and induced extracts with both Oligo 60 (Fig. 5A) and a probe containing a single AP-1 site (Fig. 6). Fig. 6 also demonstrates the activation of c-Jun binding activity in THP-1 cells induced with LPS. In order to determine whether different Jun family members were binding to the TF promoter in control and induced nuclear extracts, a panel of ␣-Jun antibodies was tested for reactivity against AP-1 complexes that bound to Oligo 87 (Fig. 7). The pan-Jun antiserum supershifted complexes in both control and induced extracts. The ␣-JunD effectively supershifted the entire complex with control extracts (indicated by the bracket at the left) but left residual binding with LPS-induced extracts (indicated by the dotted line at the right). Incubation of induced extracts with a combination of ␣-c-Jun and ␣-JunD effectively supershifted the entire complex.
As summarized in Fig. 3, Oligo 53, which spans the inducible footprint I, contains consensus sequences for Egr-1/2, AP-2, and two Sp1 sites, one of which overlaps the Egr-1/2 site. EMSAs with Oligo 53 are shown in Fig. 8. With control extract, two specific bands are produced indicated by the solid arrows. The lower band was competed by cold Oligo 53, and an Sp1 consensus oligo competed for binding to both bands. The induced extract produced an additional complex that migrated between these two bands, indicated by the open arrow. The addition of cold Oligo 53 effectively competed for the binding activity and reduced the intensity of this LPS-inducible complex. As a nonspecific competitor, an oligo for the Ets binding site was included in the binding reaction and was ineffective. A combination of the Ets and Sp1 competitor oligos was as effective as Sp1 alone at titrating the binding to this oligo. Since one Sp1 consensus site (Ϫ78 to Ϫ71) overlaps the Egr consensus site (Ϫ81 to Ϫ73), an additional competition experiment was performed to test the effect of increasing the concentration of cold Sp1 consensus oligo on the binding of the complex that was observed with LPS induction (Fig. 8B). As the concentration of   4), the intensity of this inducible band appeared to increase. A further increase in concentration of cold Sp1 to 800 ϫ actually reduced binding of the inducible band, which may be due to nonspecific effects of excess DNA in the binding reaction. Antibody supershifts were performed with Oligo 53 (Fig. 8C). ␣-Sp1 produced a supershifted complex with both the control and induced extracts. With LPS-induced extract, ␣-Egr-1 abrogated the binding of the inducible complex. As a nonspecific control, ␣-Ets-1/2 had no effect on any of the complexes formed with this probe. DISCUSSION LPS activation of monocytes stimulates protein tyrosine kinases, protein kinase C, and protein kinase A (21). Both LPS and TPA signal through protein kinase C to activate Raf-1 and mitogen activated protein kinase (22), and activation of this pathway is required for monocyte differentiation and induction of c-Jun and c-Fos mRNA expression in response to LPS (23,24). Jun and Fos represent families of transcription factors that bind AP-1 sites, 5Ј TGA G/C TCA 3Ј (reviewed in Ref. 25). These transcription factors share common structural motifs: members dimerize through interaction at the leucine zipper, they bind to DNA via a basic region adjacent to the leucine zipper, and they contain additional modulatory domains (26). Although the Jun proteins can bind to DNA as homodimers or heterodimers, the Fos proteins bind DNA only as obligate heterodimers (27,28). Heterodimerization of Jun and Fos with select members of transcription factor families that share the leucine zipper motif, i.e. ATF (29) and C/EBP␤ (30), has also been reported. We have evaluated C/EBP binding to the TF promoter and have not detected any (data not shown).
Phosphorylation of c-Jun on serine and threonine at three sites in the carboxyl terminus just upstream of the basic region, by casein-kinase II and glycogen synthase kinase-3, prevents AP-1 binding (31). In human epithelial cells and fibroblasts, activation of a phosphatase by protein kinase C results in dephosphorylation of the latent form of c-Jun within 15-60 min following incubation with TPA, resulting in an increase in binding to the TPA response element which is recognized by AP-1 (32). Changes in phosphorylation of sites in the amino terminus modulate the function of the A1 domain to mediate transcriptional activation, and the ␦ and ⑀ domains that interact with the cell-specific inhibitory protein IP-1 (33,34). Serine phosphorylation by a Ha-Ras-induced c-Jun nuclear kinase causes an increase in DNA binding (35), whereas phosphorylation by mitogen activated protein kinase enhances transcriptional activation (36). Expression of TF in THP-1 cells is rapidly activated within 60 min after LPS addition and is mediated by protein kinase C (12,37). The TF promoter contains two AP-1 sites that are required for optimal induction (12). Since binding studies with a single AP-1 consensus site detect Jun-Fos heterodimers exclusively in both control and induced nuclear extracts, we conclude that AP-1 heterodimers bind to the TF AP-1 consensus sequences (Fig. 6). In addition, these studies demonstrate the induction of c-Jun binding activity in THP-1 cells incubated with LPS. In order to evaluate AP-1 binding in the context of the TF promoter, supershift assays with antibodies that recognize Jun and Fos family members were performed with Oligo 60 (Fig. 5A) and Oligo 87 (Fig. 5, B and C, and Fig. 7). These studies clearly demonstrate that the AP-1 binding in control extracts is due to binding of JunD/Fos heterodimers, whereas, in induced extract, both JunD/Fos and c-Jun/Fos heterodimers interact with the Ϫ220 to Ϫ197 region of the promoter. To characterize the additional binding activity detected with Oligo 87, several different antibodies were tested (Fig. 5C). In order to increase the sensitivity of the antibody-protein interaction, the binding time was increased in these supershift assays and, as a result, a slower migrating complex appeared with the LPS-induced extract. The specificity of this complex was confirmed by a self-competition assay (data not shown). The binding that resulted in the slower migrating complex was abrogated when incubated with the ␣-Ets antibody that crossreacts with Ets-1 and Ets-2. There are also two potential Sp1 binding sites downstream of the Ets site within this region including a GC box and a CACCC sequence (38). These studies also demonstrate Sp1 binding to this probe. Interestingly, both the ␣-Jun and the ␣-Sp1 antisera interfered with the formation of this higher mobility complex. Neither ␣-c-Rel nor ␣-NF-B antibodies (data not shown) interfered with any of the binding complexes. In the TF promoter, the Ets consensus core (Ϫ181 to Ϫ177) overlaps the NF-B consensus sequence (Ϫ187 to Ϫ179) and is located between the AP-1 and Sp1 binding sites. With induced extract, interference with either AP-1, Sp1, or Ets binding by including specific antisera in the reaction prevents formation of the slower migrating complex, suggesting that there might be some cooperative interactions among the proteins that interact with Oligo 87.
The Ets family of transcription factors contains many members, including Ets-1, Ets-2, Elk, Erg, PU.1, and PEA3, that share a common Ets domain and bind to a number of consensus sequences, all of which contain a GGAA core (reviewed in Refs. 39,40). Ets-1 (p39-p51) and Ets-2 (p56) are nuclear phosphoproteins that are highly conserved between species (41). Ets-1 and Ets-2 bind weakly to DNA, although this binding appears to be strengthened, and transcriptional activation enhanced, either by the presence of more than one binding site (42,43) or by interactions with other adjacent transcription factors like AP-1, GABP␤, SRE:SRF, and Sp1 (44 -47). Synergism between AP-1 and Ets-2 activates keratin 18 expression in embryonal carcinoma cells differentiated with retinoic acid (48). In activated T-cells, cooperative binding of AP-1 and the Ets-related protein Elf-1 are required to activate transcription of the granulocyte macrophage colony-stimulating factor gene through the PB-1 element (49). Interestingly, intracellular signals that activate TF expression, such as protein kinase C, and increases in intracellular Ca 2ϩ also appear to induce Ets-2 mRNA synthesis and increase the stability of the protein (43,50). In fact, induction of the macrophage scavenger receptor in TPA-stimulated THP-1 cells results from activation of ras and induction of c-Jun, JunB, and Ets-2 binding to several juxtaposed AP-1/Ets binding sites in the promoter (51). We have suggested that the B site contains an overlapping PEA3 element that might be important for TF function (20). In fact, we have recently determined that the Ϫ188 to Ϫ175 region contains very strong homology (11/12 nucleotides) to sequences recognized by Ets-2. 3 In addition, we have demonstrated that an antibody to Ets-1/Ets-2 interferes with induced nuclear protein binding to Oligo 87 (Fig. 5). These data suggest that either Ets-1 or Ets-2 interacts with the TF promoter following LPS stimulation and may contribute to LPS induction.
The identity of the proteins that interact with the TF B-like site following LPS induction remains controversial. It has been well documented that LPS activation of monocytes induces a rapid translocation of NFB-related proteins from the cytoplasm into the nucleus (12,20,52). Comparison of the TF sequence with optimal B/Rel DNA-binding motifs suggests that two differences from the p50 consensus precludes binding of NF-B p50; however, the sequence does conform to the optimal binding sites for both p65 and c-Rel (53). We have reported weak binding (20), and others (12,15,54) have shown that B/Rel proteins from induced cells can bind to the TF site when isolated from the context of the promoter. More importantly, the data from this study suggest that these proteins do not bind to the site in the context of the TF promoter. Neither oligonucleotide competition nor antibody supershift experiments with both p50-specific and p65/c-Rel-specific antisera have demonstrated any interaction of these proteins with either Oligo 60 or Oligo 87.
Functional studies have clearly demonstrated the importance of the Ϫ227 to Ϫ172 region for promoter inducibility. Furthermore, the data suggest that cooperative interactions in this region are required for optimal induction (12). When linked to a heterologous promoter, the Ϫ192 to Ϫ172 oligonucleotide containing the TF NFB-like site (Ϫ188 to Ϫ179) confers LPS induction on reporter constructs that are 80% of the levels observed with a larger fragment (Ϫ227 to Ϫ172) containing the two upstream AP-1 sites and the NFB-like site. These data are in conflict, however, with other experiments in the same report that suggest that the Ϫ188 to Ϫ179 region is unable to support high level induction when considered in the context of the TF promoter. In fact, those studies suggest that the AP-1 motifs (Ϫ244 to Ϫ194) are responsible for 75% of the LPS inducibility, and the other 25% actually resides between Ϫ96 and Ϫ33. Finally, mutation of either the AP-1 sites or the B site dramatically reduces induction by LPS in monocytes and tumor necrosis factor-␣ in endothelial cells (15). Even a 3-bp mutation introduced into the B site of the Ϫ227 to Ϫ172 fragment (pTFM1) reduces inducibility by almost 80%. Interestingly, those three nucleotides are in the critical core sequence (GGAA) common to all Ets-related consensus sequences. This mutation, which prevents B-related binding, would also certainly abrogate Ets binding.
When considered as an independent element, the Ϫ192 to Ϫ172 sequence may demonstrate a different pattern of protein binding and regulation from this sequence in the context of the TF promoter. The ability of proteins to interact with a consensus sequence can depend on the length of the probe and the effects of neighboring binding sites. For example, AP-1 binding affinity can either be strengthened or weakened depending on the sequences that flanked the binding site (55). These methodological differences between our analysis and other published reports may account for the differences in results, because our findings suggests that Ets and AP-1 interact with the TF promoter following LPS stimulation of THP-1 cells and mediate activation at this site.
The proximal promoter contains consensus binding sites for several additional proteins including AP-2, Sp1, Egr-1, and Wilms' tumor protein, which has been reported to bind to the Egr-1 consensus (56). When the Oligo 53 probe was used no AP-2 binding was detected by either oligonucleotide competition or antibody supershift studies and no Wilm's tumor protein binding was detected by antibody supershift studies (data not shown). The TF promoter contains five Sp1 sites between Ϫ231 and the cap site and two Sp1 consensus sequences are located in the Ϫ80 to Ϫ57 region. The zinc finger protein Sp1 is important for expression of a number of genes, including CD11b and CD14 in myeloid cells (57,58). Another family member, Sp3, acts as a repressor by binding to both GC and GT boxes and antagonizing Sp1 activation at these sites (59). Our studies have determined that both Sp1 (Fig. 8) and Sp3 (data not shown) bind to Oligo 53 in both control and induced extracts, although the effect of Sp3 binding on TF expression remains to be resolved.
Finally, we have identified Egr-1 binding exclusively in induced nuclear extracts. Egr-1 is a member of a gene family of zinc finger-containing nuclear phosphoproteins that bind to the consensus sequence GCGGGGGCG (reviewed in Ref. 60). Expression of Egr-1 is rapidly up-regulated in cells as an immediate-early response gene (61,62). Expression and DNA binding is stimulated in fibroblasts by serum, phorbol ester, and the protein phosphatase inhibitor okadaic acid (63), and in monocytes by serum and cytokines (64). In HL-60 cells, Egr-1 is induced by PMA, and expression restricts the cells to macrophage differentiation (65,66). In other studies, we have demonstrated an association between Egr-1 expression and induction of TF synthesis in HL-60 cells. 2 The proximal promoter (Ϫ111 to ϩ8) linked to the CAT reporter gene is LPS-inducible in both THP-1 and PMA-inducible in HL-60 cells (data not shown). In other reports, sequences between Ϫ96 to Ϫ33 support partial induction at 25% of the level of the fully inducible promoter fragment (Ϫ44 to ϩ121) (12). These data suggest that this site, in addition to the upstream region, contributes to LPS induction of TF in response to LPS. We have identified an inducible binding activity that interacts between Ϫ85 to Ϫ52, a region of the promoter that contains two Sp1 sites, one of which overlaps an Egr-1 site. EMSA and antibody supershift studies indicate that both Sp1 and Egr-1 bind to this DNA element; however, the inducible binding activity that appears in LPSstimulated cells is Egr-1.
We have used a combination of DNase I footprinting and EMSA to characterize protein-DNA interactions with the TF promoter, using nuclear extracts prepared from control and LPS-induced THP-1 cells. Sp1 binding is detected in both control and induced extracts. Following induction, we see a conversion in AP-1 binding activity from JunD/Fos in control extract to c-Jun/Fos and JunD/Fos in induced extract. In addition, we detect inducible binding to two regions of the promoter (Ϫ197 to Ϫ175) and (Ϫ85 to Ϫ52). Our antibody studies suggest that Ets-1 or Ets-2 interacts with the distal site and Egr-1 interacts with the proximal site. In conclusion, we would like to suggest that AP-1/Ets interactions synergize to enhance expression from the Ϫ227 to Ϫ172 element. In addition, the binding of Egr-1 to the proximal site makes a significant contribution to the transcriptional activation of TF in activated monocytes.