Specific Interaction of Egr1 and c/EBPβ Leads to the Transcriptional Activation of the Human Low Density Lipoprotein Receptor Gene*

The sterol-independent regulatory element (SIRE) of the LDL receptor (LDLR) promoter mediates oncostatin M (OM)-induced transcription of the LDLR gene through a cholesterol-independent pathway. Our prior studies have detected specific associations of the zinc finger transcription factor Egr1 with the SIRE sequence in OM-stimulated HepG2 cells. Because the SIRE motif is composed of a c/EBP binding site and a cAMP response element, both of which are quite divergent from the classical GC-rich Egr1 recognition sequences, we hypothesized that Egr1 may regulate LDLR transcription through interacting with members of the c/EBP and CREB families. Here, we show that treating HepG2 cells with OM specifically leads to prominent increases of the levels of c/EBPβ and Egr1 bound to the LDLR promoter in vivo. In vitro, the binding of Egr1 to the SIRE sequence is weak, but is strikingly enhanced in the presence of HepG2 nuclear extract. Mammalian two-hybrid assays demonstrate that the N-terminal transactivation domain of Egr1 specifically interacts with c/EBPβ but not with c/EBPα or CREB. The OM treatment further enhances this interaction, resulting in a large increase in the Egr1 transactivating activity. The direct protein to protein contact between Egr1 and c/EBPβ is also demonstrated by co-immunoprecipitation experiments. Furthermore, we show that a mutation of the phosphorylation motif of c/EBPβ diminished the OM-stimulated interaction of Egr1 and c/EBPβ. Taken together, we provide strong evidence that Egr1 regulates LDLR transcription via a novel mechanism of protein-protein interaction with c/EBPβ.

dependent mechanisms. The sterol-regulatory element (SRE-1) plays a key role in the cholesterol-mediated negative feedback regulation of the LDLR transcription by interacting with the SRE-1-binding proteins (SREBPs), whose nuclear translocation is controlled by intracellular cholesterol levels (1). The sterol-independent regulatory element (SIRE) has been recently identified and shown to mediate the stimulatory effects of cytokine oncostatin M (OM), cAMP, and c/EBP␤ on LDLR transcription in HepG2 cells in a cholesterol-independent fashion (2). The SIRE motif (TGCTGTAAATGACGTGG) is located at the promoter region of Ϫ17 to Ϫ1, downstream of the SRE-1 and Sp1 binding sites. It is composed of a c/EBP binding site (Ϫ17 to Ϫ9) and a CRE site (Ϫ8 to Ϫ1). Mutations within the SIRE sequence do not affect the cholesterol-mediated suppression and only slightly lower basal promoter activity to levels 60 -80% of the wild-type sequence. However, alterations of nucleotides (even a single base) within the SIRE motif can totally prevent the OM-induced increase in the LDLR promoter activity.
In vitro DNA binding assays (EMSA) have detected the binding of several basic leucine zipper DNA-binding proteins including c/EBP␤, CREB, ATF1, ATF2, ATF3, and c-Jun to the SIRE sequence (2). These DNA-protein complexes can be detected in unstimulated as well as OM-stimulated nuclear extracts of HepG2 cells. However, stimulation of HepG2 cells with OM induces the formation of a new DNA protein complex, previously referred as the C2 complex. By conducting EMSA, followed by UV-cross-linking and Western blot analysis, we have demonstrated that the unknown DNA-binding protein present in the OM-induced C2 complex is Egr1 (3). By using the chromatin immunoprecipitation (ChIP) assay we further showed the specific association of Egr1 with the LDLR promoter in intact HepG2 cells after OM stimulation. Northern blot analysis showed that the expression of Egr1 is rapidly induced by OM and its expression precedes the up-regulation of LDLR transcription by OM in HepG2 cells. Additional evidence to link Egr1 with the OM-induced LDLR transcription is the dependence of Egr1 induction and LDLR induction on the activation of the MEK/ERK signaling cascade by OM (3,4). Treatment of HepG2 cells with MEK1 inhibitor U0126 not only eliminated Egr1 expression but also abolished the OM activity on LDLR transcription. The functional role of Egr1 in LDLR transcription was assessed by cotransfection of an Egr1 expression vector with an LDLR promoter reporter construct. We showed that overexpression of Egr1 alone does not lead to a significant increase of the LDLR transcription. However, the LDLR promoter activity is markedly up-regulated by coexpression of Egr1 with c/EBP␤ in HepG2 cells. These results suggest a cofactor dependence of Egr1 in the regulation of LDLR transcription.
Egr1 is a zinc finger transcription factor that binds to a GC-rich motif (GCGT/GGGGCG) through its zinc finger domains (5,6). Identical or highly similar sites have been shown to be functional regulatory elements in the promoters of Egr1 regulated genes. Intriguingly, the LDLR promoter does not contain a consensus or a homologous Egr1 binding site (EBS). To our knowledge, the SIRE element in the human LDLR promoter is the only recognition sequence of Egr1 that has a GC content less than 50%. This raises an important question: does Egr1 bind to the SIRE DNA directly by itself? If not, what are the cofactors of Egr1 in the recognition of the SIRE motif?
In this study, we provide strong evidence from several lines of investigation to demonstrate that c/EBP␤ is the partner of Egr1. Egr1 transactivating activity is greatly enhanced by its interaction with c/EBP␤ but not with other members of the c/EBP family or the CREB family. Our studies suggest that upon OM stimulation, Egr1 and c/EBP␤ constitute an activation complex at the SIRE site that leads to the transcriptional activation of the LDLR gene.
The constructs of GAL4-Egr1 fusion protein, an Egr1 reporter plasmid pEBS 4 Luc, and pCITEzif, used in reactions of the in vitro translation of Egr1, have been previously described (7,8). The plasmid pEF-NFIL6 and plasmid pCMV-NFIL6-T235A were gifts from Dr. Shizuo Akira (Hyogo College of Medicine, Hyogo, Japan). The plasmids pCI-ATF2 and pCIATF3 were gifts from Dr. Shigetaka Kitajima at Tokyo Medical & Dental University, Tokyo, Japan. The plasmids RSV-CREB and pCMV-c/EBP␣ (LB1321) were kindly provided by Dr. Linda M. Boxer at the VA Palo Alto Health Care System. The plasmid MSV/EBP␦ encoding the mouse c/EBP␦ was generously provided by Dr. Steven McKnight at the University of Texas Southwestern Medical Center. The plasmid pFR-Luc was obtained from Stratagene. The plasmids of pAD-SV40T and pBD-p53 were obtained from Clontech.
Plasmid Constructions-For construction of the vp16AD and the full-length or truncated c/EBP␤ fusion proteins, different coding regions of c/EBP␤ were amplified by PCR using pEF-NFIL6 as the template and appropriate 5Ј-and 3Ј-primers flanked with BamH1 sites. The primer sequences for cloning and other assays are listed in Table I. The GC-rich PCR system (Roche Applied Science) was used to amplify c/EBP␤. The PCR reactions of 35 cycles were carried out at 95°C for 30 s, 62°C for 30 s, and 72°C for 60 s with a final extension at 72°C for 7 min. The PCR fragments were directly cloned into pCR2.1-Topo vector. The insert was isolated with BamH1 digestion and cloned into pVP16 vector (Clontech) at the BamH1 site. The clones encoding the vp16AD-c/EBP␤ fusion proteins with the right insert orientation and in the correct reading frame were verified by sequencing. For the construction of a His-tagged c/EBP␤, the coding region (amino acids 24 -345) was PCRamplified and cloned directly into the pcDNA4/HisMax-TOPO vector. The resulting plasmid pcDNA4/His-c/EBP␤-P1 expresses c/EBP␤ with its N terminus tagged with six histidine residues. The expression of His-c/EBP␤ in HepG2 cells is confirmed by immunoblotting with anti-His monoclonal antibody. The plasmid His-c/EBP␤-T235A was obtained by performing site-directed mutagenesis with pcDNA4/His-c/EBP␤-P1 as the template DNA. The expression vector pCMV-FLAG-Egr1-DN, encoding the dominant-negative mutant of Egr-1, was constructed by inserting a RsaI/BglII fragment of the murine Egr-1 cDNA (nucleotides 2640 -3382, GenBank TM accession no. M22326) into plasmid p3XFLAG-CMV-NLS. The protein expressed from this vector contains a triple FLAG sequence MDYKDHDGDYKDHDLDYKDDDDK on the N terminus, followed by a nuclear localization sequence derived from the SV40 large T antigen (sequence LPPKKKRKV), and amino acids 322-533 of Egr-1.
In Vitro Translation and Co-immunoprecipitation-Egr1 and c/EBP␤ were in vitro translated using the TNT T7-coupled reticulocyte lysate systems (Promega, L4610) with pCITEzif as the template for Egr1 and pcDNA4/His-c/EBP␤-P1 as the template for His-c/EBP␤ following the manufacturer's protocol. The plasmids were individually added to the rabbit reticulocyte lysate containing T7 polymerase and [ 35 S]methionine in a reaction volume of 50 l at 30°C for 90 min. The translation products were verified by SDS-PAGE and autoradiography.
For co-immunoprecipitation, 10 l of translation mixture of 35 Slabeled Egr1 was incubated with 10 l of translation mixture of 35 Slabeled His-c/BP␤ or with 35 S-labeled His-LacZ and diluted to 500 l with co-immunoprecipitaion buffer containing 20 mM Tris-HCl pH 7.5, 150 mM NaCl, 1 mM dithiothreitol, 5 g/ml aprotinin, 0.5 mM phenylmethylsulfonyl fluoride, and 0.1% Tween 20 (v/v). The samples were incubated at room temperature for 1 h with continuous rocking. 25 l of protein A-agarose beads (50% slurry, Upstate) and 2 g of His polyclonal antibody were then added and incubated at 4°C for overnight with continuous rocking. Samples were washed three times with TBS-Tween 0.05% and centrifuged. The pellets were suspended in 20 l of SDS-loading buffer, heated at 80°C for 5 min, and centrifuged briefly. The supernatants were loaded onto a 4ϳ20% PAGE gold precast minigel. After electrophoresis and fluorography, gels were dried and visualized on a phosphorimager.
In Vitro Phosphorylation of Egr1 by Activated ERK1/ERK2-The highly purified and already activated ERK1/ERK2 was purchased from Stratagene. An aliquot of Egr1 in vitro translation product was mixed on ice with 4 l of 10ϫ kinase reaction buffer, 1 l of the activated ERK, and 2 l of [␥-32 P]ATP in a reaction volume of 40 l. The kinase reaction was executed by incubation of the mixture at 30°C for 90 min. The standard ERK substrate PHAS-I was used in a separate reaction as a positive control. The phosphorylated products were examined by SDS-PAGE and phosphorimaging. After detection of phosphorylated Egr1, the phosphorylation reaction was repeated using unlabeled ATP. Electrophoretic Mobility Shift Assays (EMSA)-Oligonucleotide probes were annealed and end-labeled with T4 polynucleotide kinase in the presence of [␥-32 P]ATP. Each binding reaction was composed of 10 mM HEPES, pH 7.8, 2 mM MgCl 2 , 2 mM dithiothreitol, 80 mM NaCl, 10% glycerol, 1 g of poly (dI-dC), 1 g of bovine serum albumin, and an aliquot of Egr1 in vitro translation product without or with 6 g of HepG2 nuclear extract in a final volume of 40 l. The samples were incubated with 0.4 -0.5 ng of 32 P-labeled double-stranded synthetic oligonucleotide probe (40 -80 ϫ 10 3 cpm) for 10 min at room temperature. The reaction mixtures were loaded onto a 6% polyacrylamide gel and run in TGE buffer (50 mM Tris base, 400 mM glycine, 1.5 mM EDTA, pH 8.5) at 30 mA for 2.5-3 h at 4°C. Gels were dried and visualized on a phosphorimager. For supershift assays, antibody was incubated with samples for 30 min at room temperature prior to the addition of the probe.
Chromatin Immunoprecipitation (ChIP) Assays-HepG2 cells were untreated or treated with OM (50 ng/ml) for 1 h and thereafter were cross-linked with 0.37% formaldehyde at 37°C for 10 min. Total cell lysate was isolated, and the genomic DNA was sheared to sizes of between 200 and 600 bp by sonication. ChIP assays with rabbit antibodies to c/EBP␤, Egr1, CREB, ATF3, and with normal rabbit IgG as a negative control were performed according to the protocol of Upstate Biotechnology using aliquots of lysate obtained from 5 ϫ 10 6 cells. The immunocomplex was heated at 65°C for 4 h to revert the cross-linking between DNA and proteins. DNA was purified by repeated phenol/ chloroform extraction and ethanol precipitation. The purified DNA (designated as bound) was dissolved in 20 l of Tris buffer (10 mM Tris, pH 8.5). The DNA isolated using the same procedure with omission of the immunoprecipitation step was designated as the input DNA and was diluted 100ϫ prior to PCR. The bound and the input DNA were analyzed by PCR (31 cycles) with primers that amplify a 180-bp fragment of the human LDLR proximal promoter region from Ϫ124 to ϩ54, relative to the major transcription start site. The PCR conditions were 94°C for 5 min, 94°C for 30 s, 68°C for 30 s, 72°C for 30 s, and 72°C for 5 min. The 180-bp PCR product was visualized on a 2% agarose gel stained with ethidium bromide. The intensity of the PCR products was scanned with a BioRad Fluro-S MultiImager System and quantified by the Quantity One Program. Different amounts of template DNA were tested in the PCR reaction to ensure a linear range of DNA amplification.
Transient Transcription and Dual Luciferase Reporter Assays-HepG2 cells were seeded in 48-well plates at 9 ϫ 10 4 per well in 10% fetal bovine serum the day before transfection. A total of 100 ng of DNA per well was transfected into cells using the FuGENE 6 transfection reagent (Roche Applied Science, 1814443). The DNA ratios of the analyzing firefly luciferase reporter to expression vector and to Renilla luciferase reporter, pRL-SV40 were 65:25:10. For analysis of the LDLR promoter, transfected cells were incubated in medium containing 0.5% lipoprotein-depleted serum (LPDS)-containing cholesterol (10 g/ml cholesterol ϩ 1 g/ml 25-hydroxycholesterol). For analysis of pFR-Luc activity, transfected cells were incubated in medium containing 0.5% fetal bovine serum. 20 h after transfection, OM at the concentration of 50 ng/ml was added, and cells were lysed 4 h later to analyze luciferase activities using the Dual Luciferase assay system obtained from Promega.
Immunoprecipitation and Western Blot Analysis of Phosphorylated c/EBP␤-HepG2 cells seeded in 100-mm culture dishes were incubated in 0.5% fetal bovine serum overnight. OM was added to cells for 1 h, and cells were lysed in radioimmune precipitation assay buffer (150 mM NaCl, 50 mM Tris-HCl (pH 8.0), 1% Nonidet P-40, and 0.5% sodium deoxycholate) in the presence of protease inhibitors. 1 ml of cell lysate containing 500 g of protein was precleared with isotype-matching IgG for 1 h and then incubated with 2 g of rabbit anti-c/EBP␤ antibody at 4°C for 2 h. Protein A/G-Sepharose (30 l/sample) was then added to the lysate, and the mixture was further incubated overnight at 4°C. After washes, proteins bound to the beads were eluted with SDS sample buffer by boiling, separated by SDS-PAGE, and analyzed by Western blotting with monoclonal antibody that recognizes the phospho-Thr-Pro motif.
For detection of phosphorylated His-c/EBP␤ or His-c/EBP␤-T235A, HEK293 cells were untransfected or transfected with pcDNA4/His-c/ EBP␤-P1 and pcDNA4/His-c/EBP␤-T235A plasmids. Two days after transfection, cells were trypsinized, reseeded into two dishes, and cultured overnight in medium containing 0.5% fetal bovine serum. Transfected cells were stimulated with OM for 1 h and harvested for immunoprecipitation with anti-His antibody. The immunoprecipitation complexes were analyzed by Western blotting with the monoclonal antibody that recognizes the phospho-Thr-Pro motif.
For other Western blot analyses, cells were seeded in 60-mm dishes and were lysed in 100 l of whole cell lysis buffer (20 mM HEPES pH 7.9, 0.2% Nonidet P-40, 10% glycerol, 400 mM NaCl, and 0.1 mM EDTA) with protease inhibitor mixture (Roche Applied Science, 1836153). 50 g of protein from whole cell lysate per sample were used for Western blotting.

RESULTS
A Strong Interaction of Egr1 with SIRE DNA Requires Cooperation from other SIRE-binding Proteins-In order to determine whether Egr1 can bind to the SIRE motif by itself or it requires cooperation from other SIRE-binding proteins, we first produced Egr1 protein by using the TNT-coupled reticulocyte lysate system. The plasmid pCITEzif encoding Egr1 (8) or a control plasmid expressing LacZ was added to the rabbit reticulocyte lysate containing T7 polymerase and [ 35 S]methionine. After reactions, translation products were examined by SDS-PAGE and autoradiography. As shown in Fig. 1A, a major band with the molecular mass of 82 kDa, corresponding to the reported molecular mass of Egr1 protein on SDS-PAGE (80 -82 kDa) was detected in the reaction mixture containing the Egr1 plasmid (lane 1) and was not present in the sample without Egr1 DNA (lane 3). Some proteins with lower molecular mass in the reaction mixture of Egr1 were also shown, which might represent degraded translation products.
It has been previously reported that Egr1 contains a single ERK phosphorylation motif (PHSP, amino acids 24 -27) residing in the N-terminal transactivation domain (AD) (9). We were interested in determining whether ERK-induced phosphorylation of Egr1 affects the binding of Egr1 to the SIRE motif. To produce the phosphorylated Egr1, the highly purified and already activated ERK1/ERK2 was incubated with an aliquot of Egr1 in vitro translation product in the presence of [␥-32 P]ATP, and the kinase reaction was executed by incubation of the reaction mixture at 30°C for 90 min. The standard ERK substrate PHAS-I was used in a separate reaction as a positive control. The phosphorylated products were then examined by SDS-PAGE and phosphorimaging. Fig. 1B shows that a 32 Plabeled protein migrated as 82 kDa was only detected in the reaction containing the translation product of Egr1 (lane 3). After validation of these methods, we reproduced Egr1 as well as ERK-phosphorylated Egr1 protein using unlabeled methionine and ATP for the subsequent DNA binding assays.
To be able to compare the binding affinity of Egr1 to the SIRE sequence with its binding to the classic EBS, a 32 Plabeled SIRE probe and a 32 P-labeled Egr1 probe were incubated with equal amounts of the Egr1 translation product in the absence or presence of untreated HepG2 nuclear extract, which does not contain a detectable level of the Egr1 protein.
After a 15-min incubation at room temperature, reaction mixtures were separated by polyacrylamide gel electrophoresis, and results were visualized by a BioRad PhosphorImager. Fig.  1C shows that the binding of Egr1 to the labeled Egr1 probe was very strong and was not affected by the presence of HepG2 nuclear extract (compare lanes 1 and 2). In contrast, the binding of Egr1 to the labeled SIRE probe was weak by itself, however, a strong binding with an increase of 8-fold was detected in the presence of HepG2 nuclear extract (compare lanes  5 and 6). The Egr1-SIRE complex could be totally supershifted by anti-Egr1 antibodies (Fig. 1D, compare lanes 2 and 4), whereas no supershifted band was detected in the reaction only containing the nuclear extract of untreated HepG2 (Fig. 1D,  lane 3). The EMSA also revealed that phosphorylation of Egr1 by ERK has no effects on the interaction of Egr1 with Egr1 probe (Fig. 1C, lanes 3 and 4) or the SIRE probe (Fig. 1C, lanes  7 and 8). To further examine the effects of phosphorylation on the DNA binding activities of Egr1, HEK293 cells were transfected with pCMV-Egr1 or pCMV-Egr1-T26A. Two days after transfection, nuclear extracts were prepared from the transfected cells and the EMSA was performed with the nuclear extracts that express the wild-type Egr1 or the phosphorylation-deficient Egr1. The results showed that mutation of T26A on Egr1 did not diminish its DNA binding activity to EBS or SIRE (data not shown). Altogether, these experiments convincingly demonstrate that the binding of Egr1 with SIRE DNA is not affected by ERK-directed phosphorylation but it is strongly enhanced by other SIRE-binding proteins, possibly through protein to protein interaction.
Egr1 Interacts Specifically with c/EBP␤-To obtain insights into which SIRE-binding protein may interact with Egr1 at the SIRE motif, initially we examined the in vivo binding of c/EBP␤, CREB, ATF3, as well as Egr1 to the LDLR promoter in control and OM 1 h-stimulated HepG2 cells by ChIP assays. To determine whether c/EBP␤ functions as a Egr1 coactivator, we performed mammalian two-hybrid assays by using a series of constructs of GAL4-Egr1 fusion proteins (Fig. 3A) and the luciferase reporter pFR-Luc that contains 5 GAL4 binding sites. HepG2 cells were transiently transfected with pFR-Luc along with GAL4-Egr1 fusion constructs in the absence or presence of expression vectors for c/EBP␤ (pEF-NFIL-6) or CREB (RSV-CREB). The plasmid pRL-SV40 expressing the Renilla luciferase was also included as a control vector to normalize the variation of transfection efficiency. 20-h posttransfection, OM was added to cells for 4 h prior to cell lysis. Fig. 3B shows that the transactivating activities of the GAL4 fusion proteins containing the partial AD of Egr1 (pM-Egr1-1) and the whole AD of Egr1 (pM-Egr1-2) were greatly increased by coexpression of c/EBP␤ but not by coexpression of CREB, and OM stimulation further increased the stimulatory effect of c/EBP␤. The activity of the fusion protein pM-Egr1-3 that contains the AD and the repression domain (RD) of Egr1 (7, 10) was not increased by c/EBP␤, suggesting that the presence of the repression domain may interfere with the accessibility of c/EBP␤ to the Egr1 AD.
To further evaluate the specificity of the Egr1-c/EBP␤ interaction, we included the constructs of pBP-p53 and pAD-SV40T in the mammalian two-hybrid assays. As shown in Fig. 3C, the activity of GAL4-Egr1-2 is increased by pEF-NFIL-6 but was not increased by pAD-SV40T that expresses the vp16 AD and SV40 T antigen fusion protein. On the other hand, the activity of GAL4 and p53 fusion protein pBD-p53 was only increased by pAD-SV40T and was not increased by pEF-NFIL6. Importantly, OM has no stimulatory effect for the interaction of p53 and SV40 T antigen. These results confirm the specific inter-  1 and 2; 5 and 6). In lanes 3, 4, 7, and 8, the translated Egr1 was first phosphorylated by ERK and subsequently mixed with the probes. Equal amounts of Egr1 protein were used under all conditions. The data shown are representative of three separate assays. D, supershift assay with anti-Egr1 antibody. HepG2 nuclear extract was incubated with the 32 P-labeled SIRE probe without (lanes 1 and 3) or with 2 l of Egr1 (lanes 2 and 4) in the absence (lanes 1 and 2) or the presence of 2 l of anti-Egr1 antibody (lanes 3 and 4).

FIG. 2. ChIP analysis for Egr1 and c/EBP␤ association with the LDLR promoter in control and OM-treated HepG2 cells.
Antibodies to c/EBP␤, CREB, Egr1, and ATF3 were used in a ChIP analysis followed by PCR to amplify a 180-bp region surrounding the SIRE site from genomic DNA isolated from unstimulated and OM 1-h stimulated HepG2 cells. Normal rabbit IgG was included in the assay as a negative control for nonspecific binding. The PCR product was separately on a 2% agarose gel, stained with ethidium bromide, and quantified by a BioRad Fluro-S MultiImager system. Bound represents the DNA coimmunoprecipitated with antibody, while Input represents the starting material before immunoprecipitation. The data shown are representative of three separate ChIP assays. action between Egr1 and c/EBP␤ and further indicate a stimulatory role of OM in this interaction. So far our data have shown that a strong binding of Egr1 to the SIRE DNA requires the presence of other SIRE-binding proteins and that c/EBP␤ increases the transactivating activity of Egr1 containing only the N-terminal transcriptional activation domain. These results together suggest that Egr1 may be involved in the regulation of LDLR transcription through a new mechanism that does not require its DNA binding activity. To test this, we constructed an expression vector pCMV-Egr1-DN that encodes a fusion protein containing a FLAG tag, followed by a nuclear localization sequence (NLS) derived from the large T antigen, and the C terminus of Egr-1, including the entire DNA binding domain (DBD) of Egr1. We reasoned that this vector may compete the binding of Egr1 to EBS and thus act as a dominant-negative mutant to inhibit the Egr1 transactivating activity through the classic mode of Egr1, but it may not affect Egr1 activity on LDLR transcription if its DNA binding activity is not involved. Indeed, Fig. 4 shows that the OMinduced increase in the promoter activity of an Egr1 reporter pEBS 4 Luc was significantly decreased by cotransfection of pEBS 4 Luc with pCMV-Egr1-DN, whereas the stimulation of LDLR promoter activity by OM was not diminished at all by the expression of the Egr1 DBD, indicating that the Egr1 inhibitor blocked OM-induced binding of Egr1 to the EBS in pEBS 4 Luc promoter but it did not prevent OM-induced association of Egr1 with the LDLR promoter.
Next, we examined the effects of other members of the c/EBP and CREB family on LDLR transcription and the interaction with Egr1. The upper panel of Fig. 5A shows that among the transcription factors examined, LDLR promoter activities were increased by exogenous expression of c/EBP␤ and c/EBP␣, whereas c/EBP␦ and CREB/ATF had no effect. Interestingly, unlike c/EBP␤, c/EBP␣ did not increase the transactivating activity of Egr1 in the mammalian two-hybrid assay (Fig. 5A, lower  panel). We also analyzed the endogenous expressions of c/EBP␤, c/EBP␣, and c/EBP␦ in HepG2 cells by Western blot and found that c/EBP␤ protein was readily detected whereas expression levels of c/EBP␣ and c/EBP␦ were below the detection limit. However, expressions of these two proteins were seen after transfection of their expression vectors respectively in HepG2 cells (Fig. 5B). Together, these results demonstrating the absence of c/EBP␣ and c/EBP␦ protein expression in HepG2 cells and the lack of their interaction with Egr1 further indicate the specific roles of c/EBP␤ and Egr1 in the OM-mediated up-regulation of the LDLR transcription through the SIRE element. . The plasmid pUC18 was used to normalize total amount of DNA transfected into cells. PRL-SV40 was included in each transfection as an internal control. 20 h after transfection, cells were stimulated with OM for 4 h prior to cell lysis. The data shown are a representative of 4 -6 independent transfection assays. C, mammalian two-hybrid assay was conducted to examine the interaction between GAL4-Egr1 with c/EBP␤ or with SV40 T antigen and the interaction between GAL4-p53 with SV40 T antigen or with c/EBP␤. The data shown are a representative of four independent transfection assays.

An Intact c/EBP␤ Is Required for Its Interaction with Egr1-
While sequences of the DBD of the c/EBP family are highly homologous, the N-terminal transactivation domains among different members of this family are quite divergent (11)(12)(13). The fact that c/EBP␣ or c/EBP␦ could not interact with Egr1 suggests that the NH 2 -terminal AD of c/EBP␤ may play a critical role in the interaction with Egr1. In an attempt to define the domain requirement, we made four fusion constructs that express the AD of vp16 and the full-length or the truncated c/EBP␤ (Fig. 6A) and tested their abilities to activate the transactivating activity of Egr1. Fig. 6B shows that the Egr1 transactivating activity is increased 7.1-fold by the fusion protein containing a full-length c/EBP␤. Deletion of the C-terminal domains reduced the activity of c/EBP␤ to 2.3-fold whereas deletion of the N-terminal AD nearly abolished the activation. These data indicate that the N-terminal activation domain alone is not sufficient to enhance the Egr1-transactivating ability; an intact c/EBP␤ protein is needed. It is possible that maintaining the correct conformation of c/EBP␤ is important for its interaction with Egr1. This may also explain the relatively lower activity of the vp16 AD-c/EBP␤ fusion protein as compared with the wild-type c/EBP␤ (pEF-NFIL6), because of the attachment of the vp16 AD to the N terminus of the protein.
To determine whether Egr1 and c/EBP␤ can directly interact in the absence of other SIRE-binding proteins we performed immunoprecipitation assay using [ 35 S]methionine labeled in vitro translated products of Egr1 and a His-tagged c/EBP␤. In vitro translated His-LacZ was included as a negative control. Using rabbit anti-His antibody to immunoprecipitate, Egr1 was only detected in the immunoprecipitation complex of the reaction mixture containing Egr1 and His-c/EBP␤ (Fig. 6C,  lane 1) but not in the immunoprecipitation complex formed with the His-LacZ (Fig. 6C, lane 2). These results are in accordance with the findings obtained from in vivo DNA binding assays (Fig. 2) that detect the association of both Egr1 and c/EBP␤ with the LDLR promoter. We also produced the [ 35 S]methionine-labeled His-tagged N-terminal and C-terminal domains of c/EBP␤ (His-c/EBP␤-2, -3, and -4) and conducted similar immunoprecipitation assays. But we could not consistently detect a clear signal of labeled Egr1 with anti-His antibody, indicating that the interactions of Egr1 with the truncated c/EBP␤ proteins are very weak.
ERK-induced Phosphorylation of c/EBP␤ Enhances the Interaction of c/EBP␤ with Egr1-Previously we have shown that ERK is rapidly activated by OM (4) and blocking this activation totally abolishes the OM-induced Egr1 mRNA expression (3), the formation of the C2 complex with SIRE DNA (2), and the up-regulation of the LDLR expression (3,4). These prior studies combined with the results described above imply a functional role of ERK activation in the interaction of Egr1 with c/EBP␤. Both Egr1 and c/EBP␤ contain a single ERK phosphorylation site (9,13). The phosphorylation motif of Egr1 (PHSP, amino acids 24 -27) is located in the N-terminal AD that mediates the interaction with c/EBP␤. To determine whether the phosphorylation status is important for Egr1 to interact with c/EBP␤, we conducted the mammalian two-hybrid assays using the GAL4 fusion proteins containing the wild-type Egr1 (pM-Egr1-2) and the mutated Egr1 (pM-Egr1-2-S26A). In this assay, ERK activation is either induced by OM or by cotransfection with the plasmid MAPKK⌬N3-S218E-S222D that encodes a constitutively active form of MEK1 (14). Fig. 7 shows that the c/EBP␤-induced activation of Egr1 transactivation activity was not diminished by the Egr1 mutant. The c/EBP␤-induced increase in the luciferase activities were further increased by OM treatment in cells expressing either the Egr1 wild type or the phosphorylation-inactive mutant. Expression of the wild-type MEK1 did not produce an effect, however, expression of the constitutively active MEK1 caused a similar enhancement on the transactivating activities of both the wild-type and the mutant Egr1 in the presence of c/EBP␤ without OM treatment. These data demonstrate the nonessential role of Egr1 phosphorylation and point to a functional role of ERK-induced phosphorylation of c/EBP␤.
We further tested the c/EBP␤ mutant on the OM induction of LDLR promoter activity. The OM activity on the LDLR promoter was relatively reduced but not eliminated by cotransfection with the mutated c/EBP␤ as compared with the wild-type vector (Fig. 8B). The residual activity is likely caused by OMinduced Egr1 expression that can interact with unphosphorylated c/EBP␤, leading to increased LDLR promoter activity. In an attempt to demonstrate a direct effect of OM on induction of the ERK-directed phosphorylation of c/EBP␤, HepG2 cells were treated with OM for 1 h, and total cell lysates were harvested from control and OM-treated cells. The rabbit antic/EBP␤ antibody was used to immunoprecipitate c/EBP␤. The immunocomplexes were thereafter analyzed by Western blot to detect the phosphorylated c/EBP␤ using a monoclonal antibody that only recognizes the phospho-Thr-Pro motif of the ERK substrates. The membrane was subsequently probed with rabbit anti-c/EBP␤ antibody, and followed with anti-actin mAb. The results in Fig. 8C show that there is a low level of phosphorylated c/EBP␤ in control cells; OM treatment significantly increased the level of phosphorylated c/EBP␤ whereas the total c/EBP␤ protein level in HepG2 cells was not elevated after the short exposure to OM. To further verify the OM-induced phosphorylation site, 293 cells were transfected with pcDNA4/Hisc/EBP␤ or with pcDNA4/His-c/EBP␤-T235A. Two days posttransfection, cells were treated with OM for 1 h and harvested for immunoprecipitation with anti-His antibody. The immunocomplexes were analyzed by Western blotting using antibody that recognizes the phosphorylated His-tagged c/EBP␤. Fig. 8D shows that the phospho-c/EBP␤ was only detected in cells transfected with the wild-type His-c/EBP␤ (lane 4), and OM treatment significantly increased its abundance (lane 5). In contrast, this band was not present in untransfected cells (lane 1) and was not detected in cells transfected with the His-c/ EBP␤-T235 mutant (lanes 2 and 3). These results demonstrate that the OM-induced phosphorylation occurs at the ERK recognition site. All together, the results of Fig. 8 indicate that the ERK-induced phosphorylation of c/EBP␤ has a functional role in OM-mediated activation of the LDLR transcription. DISCUSSION Egr1, also known as ZIF268, NGFI-A, Knox-24, and TIS8, is an 80 -82-kDa nuclear phosphoprotein (5). Egr1 belongs to a group of genes termed the immediate early genes. A variety of environmental signals, including growth factors, hormones, neurotransmitters, and cytokines, induce Egr1 (9,(17)(18)(19)(20)(21). The prompt induction of Egr1 expression is poised to initiate genetic programs that would couple a stimulus with an appropriate cellular response. It is well known that Egr1 binds to a GC-rich motif (5Ј-GCG(T/G)GGGCG-3Ј) through its three zinc finger DNA binding domains and modulates transcriptions of a number of genes that participate in different cellular functions (8,(22)(23)(24)(25). However, in this study we provide compelling evidence from five different lines of investigation to demonstrate a different activation mechanism of Egr1 in the regulation of LDLR transcription via a protein-protein interaction with c/EBP␤, a transcription factor belonging to the basic leucine zipper transcription factor family.
First  8. The phosphorylation-defective c/EBP␤ mutant lacks the response to OM stimulation. A, mammalian two-hybrid assays were conducted by using pM-Egr1-2 with different amounts of the wild type or the Thr235A mutants of c/EBP␤. The asterisk sign indicates the statistically significant difference (p Ͻ 0.05) of the Egr1 transactivating activities between OM and control. B, HepG2 cells were transfected with pLDLR234Luc with different amounts of plasmid DNA encoding the wild-type c/EBP␤ (pEF-NFIL6) or the phosphorylation-defective c/EBP␤ mutant (pCMV-NFIL6-T235A). The OM-induced increases of luciferase activity are 4.9-and 3.7-fold in cells co-transfected with the wild-type c/EBP␤ and are 2.5-and 2.0-fold in cells co-transfected with the mutant c/EBP␤. C, detection of OM-induced phosphorylation of endogenous c/EBP␤. HepG2 cells were untreated or treated with OM for 1 h. Total cell lysates (500 g per sample) were incubated with 2 g of rabbit anti-c/EBP␤ antibodies and preceded with immunoprecipitation. The immunoprecipitation complexes were analyzed with immunoblotting with antibody recognizing only phosphorylated c/EBP␤. Aliquots of total cell lysates were probed with rabbit anti-c/EBP␤ antibodies to detect the expression levels of total c/EBP␤ protein. D, detection of OM-induced phosphorylation of transfected His-c/EBP␤. HEK293 cells were mocktransfected, transfected with the wild-type His-c/EBP␤, and transfected with the His-c/EBP␤-T235A mutant. Two days later, cells were trypsinized and reseeded into two dishes for each transfection. After an overnight incubation in medium with 0.5% fetal bovine serum, cells were treated with OM for 1 h. Immunoprecipitation with anti-His antibody was performed for each total cell lysate containing 1 mg of protein, and the immunocomplexes were analyzed by Western blotting to detect phosphorylated His-c/EBP␤. The anti-c/EBP␤ and anti-actin antibodies were used to analyze total lysates. The asterisk sign indicates the nonspecific band that is equally present in every immunoprecipitation sample. c/EBP␤ is the transcription factor that recruits Egr1 to the SIRE motif. Third, the mammalian two-hybrid assay using a series of constructs of GAL4-Egr1 fusion proteins demonstrate that among members of the c/EBP and CREB families that all have the ability to bind the SIRE DNA under in vitro conditions, only c/EBP␤ increases the transactivating activity of the GAL4-Egr1 fusion proteins that do not contain Egr1 DBD. Fourth, by using the approach of in vitro translation coupled with immunoprecipitation, we demonstrate a direct contact between Egr1 and c/EBP␤ in the absence of other SIRE-binding proteins. The last evidence to rule out the involvement of the Egr1 DNA binding activity in the OM-induced LDLR transcription is obtained by using a dominant-negative inhibitor of Egr1 (pCMV-Egr1-DN). The construct expressing the zinc finger domains of Egr1 inhibited the OM-induced activation of the pEBS 4 Luc luciferase reporter driven by 4 EBS repeats, whereas the inhibitor had no diminishing effect on OM-induced reporter driven by the LDLR promoter.
Although we have focused the study of Egr1-c/EBP␤ interaction on the regulation of LDLR transcription that does not have an EBS, it is possible that this interaction may also play a role in transcriptions of other Egr1-regulated genes. We have found that the pEBS 4 Luc reporter activity can be moderately increased by coexpression with c/EBP␤ but not with c/EBP␣ (data not shown). The question of whether endogenous cellular genes containing EBS can be modulated by c/EBP␤ through its interaction with Egr1 needs further investigation.
ERK activation is a critical event in the OM-stimulated LDLR expression and Egr1 expression. Egr1 and c/EBP␤ both contains a single ERK recognition motif. Previously we have postulated that the OM-induced ERK activation probably has dual effects on Egr1 in relation to LDLR transcription. ERK activates Egr1 promoter activity (13,21,27), resulting in prompt accumulation of Egr1 mRNA and protein in nucleus. Subsequently, the Egr1 protein could be further phosphorylated by ERK. These two combined events may account in part for the increased binding of Egr1 to the LDLR promoter. In this study, we show that in vitro phosphorylation of Egr1 with activated ERK did not increase Egr1 binding affinity to SIRE DNA or to EBS. Additionally, we found that the interaction between Egr1 and c/EBP␤ is stimulated by OM and by the constitutively activated MEK1 to similar degrees by using the wild type or mutated Egr1S26A mutant. These results together indicate that the phosphorylation site on Egr1 is nonessential for Egr1 transcriptional activity. In contrast to Egr1, the ERKdirected phosphorylation site on c/EBP␤ is likely functionally involved in the interaction between c/EBP␤ and Egr1, as we have shown that mutation of this site abolished the OM stimulatory effect on the Egr1 transactivating activity. Furthermore, transfection of pCMV-c/EBP␤-T235A lowered the effect of OM on induction of LDLR promoter activity. By utilizing a monoclonal antibody that only recognizes the Thr-235-phosphorylated c/EBP␤, we were able to demonstrate that OM treatment leads to an elevated phosphorylation level of c/EBP␤. Altogether, these data indicate the double roles of ERK in the OM-mediated up-regulation of LDLR expression. ERK activation leads to Egr1 expression and phosphorylation of c/EBP␤, both of which are important steps in the signaling cascade that ultimately changes the transcriptional rate of the LDLR in OM-treated HepG2 cells.
In conclusion, we have demonstrated a novel activation mechanism for Egr1 to regulate LDLR gene transcription that is different from its traditional act of DNA binding. Instead, Egr1 is recruited to the SIRE motif through its interaction with c/EBP␤. The interaction of Egr1 and c/EBP␤ is further strengthened by the ERK-induced phosphorylation of c/EBP␤, thereby leading to the formation of a transactivation complex at the SIRE site and resulting in a rapid increase of LDLR gene transcription.