- Email: [email protected]
Site A of the MCP-1 distal regulatory region functions as a transcriptional modulator through the transcription factor NF1
Molecular Immunology 37 (2000) 623 – 632 www.elsevier.com/locate/molimm
Site A of the MCP-1 distal regulatory region functions as a transcriptional modulator through the transcription factor NF1 Shantha N. Kumar, Jeremy M. Boss * Department of Microbiology and Immunology, Emory Uni6ersity School of Medicine, 1510 Clifton Road, Atlanta, GA 30322, USA Received 24 July 2000; accepted 14 November 2000
Abstract The monocyte chemoattractant protein-1 (MCP-1) functions to recruit monocytes and macrophages to areas of inflammation and is a prototypic chemokine subjected to coordinate regulation by immunomodulatory agents. TNF mediated regulation of MCP-1 occurs through a distal regulatory region located 2.5 kb upstream of the transcriptional start site. Within this region are two NF-kB motifs that are each critical for function. Site A, located within the distal regulatory region and upstream of the kB elements is required for maximal induction by TNF. However, unlike the kB elements and other MCP-1 regulatory elements, Site A is constitutively occupied by factors in vivo. To better understand the nature of Site A function, this report identified a Site A binding protein and provides a functional analysis of the element in driving transcription. The results showed that the transcription factor NF1/CTF binds to Site A both in vitro and in vivo. While Site A has no transcriptional activity on its own, it was found to augment the transcriptional activity of a GAL4-VP16 reporter system in an orientation and position independent manner. Because NF1 is known to interact with factors that modify nucleosomes, these results suggest a unique role for Site A in regulating MCP-1 expression. © 2001 Elsevier Science Ltd. All rights reserved. Keywords: MCP-1; Chemokine; Gene expression; NF-1
1. Introduction Monocyte chemoattractant protein-1 (MCP-1) is a proinflammatory CC chemokine secreted by fibroblasts, endothelial cells, vascular smooth muscle cells, monocytes, T cells, and other cell types that mediate the influx of cells to sites of inflammation (Baggiolini and Dahinden, 1994; Howard et al., 1996; Ransohoff et al., 1996; Baggiolini et al., 1997). MCP-1 expression has been observed in a large number of tissues during disease progression, including atherosclerosis (Schwartz et al., 1991; Nelken et al., 1991; Yla-Herttuala et al., 1991), arthritis (Koch et al., 1995; Taylor et al., 2000), and cancer (Mantovani et al., 1993; Negus et al., 1998; Kopydlowski et al., 1999). In these cases, the influx of macrophages into these tissues has been suggested to exacerbate the disease. Thus, the expression of MCP-1, which is likely to be critical for fighting infectious * Corresponding author. Tel.: +1-404-7275973; fax: + 1-4047271719. E-mail address: [email protected] (J.M. Boss).
disease, must be tightly regulated. Regulation of MCP1 expression has been shown to occur at the transcriptional level by stimulatory agents like TNF (Ping et al., 1996; Jones et al., 1997), IFN-g (Satriano et al., 1993; Zhou et al., 1998), PDGF (Cochran et al., 1983; Rollins et al., 1988; Freter et al., 1992), and stress factors (Shyy et al., 1995; Glabinski et al., 1996). In contrast, retinoic acid, glucocorticoids, and estrogen have been shown to down modulate MCP-1 expression (Kawahara et al., 1991; Brach et al., 1992; Hanazawa et al., 1994; Frazier-Jessen and Kovacs, 1995). Because macrophages secrete TNF, and TNF induces MCP-1, one can envision a positive feedback loop generated by both of these cytokines for the rapid infiltration of macrophages to an area of inflammation. The transcriptional regulation of MCP-1 is controlled by two complex elements (Fig. 1A). The proximal regulatory region is sufficient for PDGF and IFN-g induced expression and is necessary for TNF-induced expression. The proximal regulatory region consists of three subelements: GC box, Site B, and kB-3 (Ping et al., 1996). The GC box, which binds Sp1 in vivo and in
0161-5890/01/$ - see front matter © 2001 Elsevier Science Ltd. All rights reserved. PII: S 0 1 6 1 - 5 8 9 0 ( 0 0 ) 0 0 0 9 7 - 3
624
S.N. Kumar, J.M. Boss / Molecular Immunology 37 (2000) 623–632
vitro is critical to all aspects of MCP-1 regulation (Ping et al., 1999a, 2000). While Site B and the kB-3 box are functional in vivo (Ping et al., 1996, 1999b), the factors that bind these elements have not been defined. In the human MCP-1 gene, kB-3 is required for IFN-g induction and interacts with STAT1 (Satriano et al., 1993; Zhou et al., 1998). The distal regulatory region contains four subelements (Ping et al., 1996). Two NF-kB binding sites, termed kB-1 and kB-2 surround a DMS hypersensitive region (HS). Both the kB sites and the HS region are required for MCP-1 regulation by TNF (Ping et al., 1999b). The kB sites bind p65 and at least one of these sites binds a p65/50 heterodimer (Ueda et al., 1994, 1997; Ping et al., 1999b). The protein binding to the HS is not known. The spacing between the two kB sites is not strict and there does not appear to be a requirement for HMG-I(Y), a minor groove architectural protein
Fig. 1. EMSAs of nuclear extracts from mouse liver using Site A oligo as a probe display a specific banding pattern. (A) Schematic map of the murine MCP-1 gene’s regulatory elements. (B) Site A sequence (shaded region) and schematic of DNA competitors SA-1, SA-2 and SA-3. (C) Liver nuclear extracts were tested to detect protein complexes binding to Site A in an EMSA in the presence of various specific and non-specific competitors. Lane 1, no extract; lane 2, no competitor; lanes 3 –5, poly(dIdC) competitor DNA (concentrations); lanes 6 – 8 contain SA-1, 2, and 3 competitor DNAs; lane 9 contains Site A competitor; and lane 10 contains a non-specific competitor. All competitor DNAs were used at 50 ng/assay. Bands a and b were used in subsequent methylation interference analysis.
found to be important for other TNF induced genes (Ping et al., 1999b). Site A, which lies just upstream of the kB sites, is required for maximal expression in fibroblasts treated with TNF. Deletion of Site A from reporter constructions containing the upstream kB sites resulted in a two-thirds loss in activity, suggesting an important role for Site A (Ping et al., 1996). The Site A sequence, which consists of 41 bp is highly conserved between murine and human MCP-1 genes. Site A is unique in the MCP-1 gene with regard to protein occupancy in vivo (Ping et al., 1996). In vivo genomic footprinting studies (IVGF) performed on the MCP-1 gene in both control and TNF-treated murine fibroblasts found that Site A sequences were resistant or hypersensitive to methylation by dimethylsulfate (DMS), suggesting that proteins were constitutively bound to the site. This constitutive occupancy was unexpected as in control cells no other regions of protein occupancy were observed between Site A and the start of transcription, which is 2.5 kb downstream of Site A. This included the Sp1 binding site as well as kB-3 and Site B, sites that were capable of binding factors in vitro. Upon TNF treatment, Site A showed no changes in its IVGF pattern, while all the sites described above showed distinct and clear changes in their IVGF pattern, indicating transcription factor binding and assembly. These results suggested that Site A may play an important and unique role in MCP-1 regulation by TNF. In this report we have undertaken a biochemical approach to identify Site A binding factors. Nuclear extracts prepared from mouse liver and fibroblasts revealed a distinct pattern of binding to Site A. This banding pattern was analyzed further by in vitro methylation interference assays, which identified a site that was similar to a NF1 transcription factor binding site. Further analyses showed that the in vitro binding patterns were most likely due to NF1 binding. In vivo chromatin immunoprecipitation (ChIP) assays were used with aNF1 antisera to show that NF1 could be found on Site A in fibroblasts. Additionally, we analyzed the ability of Site A to stimulate/interfere with the transactivation of an acidic activator protein. The results showed that Site A had no intrinsic transcriptional activity but that it could augment the activity of other transcriptional activators in an orientation and position independent manner. Together these results describe a role for NF-1 and Site A as a modulator of transcriptional activation.
2. Materials and methods
2.1. Cell lines and tissue culture NIH3T3 fibroblasts were purchased from ATCC and
S.N. Kumar, J.M. Boss / Molecular Immunology 37 (2000) 623–632
grown in Dulbecco’s modified Eagle’s media supplemented with 10% bovine calf serum (Hyclone, Logan, UT), penicillin (50 U/ml), streptomycin (50 mg/ml) and L-glutamine (1 mM) (Life Sciences). Cells were grown at 37°C in an atmosphere of 8% CO2.
2.2. Plasmids The plasmid pDRGY-CAT served as the initial subcloning reporter vector (Riley and Boss, 1993). pDRGY-CAT contains a single Gal4 DNA binding site 5%-CGGAAGACTCTCTCCTCCG upstream of the Y box element and the minimal promoter region (−56– + 3) of the human HLA-DRA gene. These sequences drive expression of the chloramphenicol acetyltransferase (CAT) gene. pDRG-CAT, which lacks the Y box, was derived from pDRGY-CAT by double digestion with XbaI and SalI, filling in the ends and ligating the plasmid ends together (Riley and Boss, 1993). pDRG-CAT is inactive in B cells, fibroblasts, and T cells. In current studies pDRGY-CAT was also digested with XbaI and SalI, removing the Y box, the ends were filled in by the Klenow fragment of DNA polymerase and a double stranded Site A oligonucleotide 5%-AGAACTGCTTGGCTGCAGGCCCAGCATCTGGAGCTCACATT was blunt-end ligated with the ends to obtain plasmids pDRGSA-CAT and pDRGRSACAT, which have Site A in forward and reverse orientations, respectively. To make the plasmid pDRSAG-CAT, the plasmid pDRG-CAT was cut with HindIII, filled-in with the Klenow fragment of DNA polymerase and ligated with Site A oligonucleotide. Plasmid pDRSAGSA-CAT was created by cutting pDRGSA-CAT with HindIII, filling-in the ends as above and religating with Site A oligo upstream to the Gal4 site. All constructs were verified for integrity by DNA sequencing.
2.3. Nuclear extracts and electrophoretic mobility shift assays (EMSA) Nuclear extracts were prepared from NIH3T3 fibroblasts or mouse liver using the methods of Shapiro et al. (1988) and as described by Ping et al. (1996). For extracts from TNF treated cells, 500 U/ml of recombinant human TNF (Genzyme) was added for 30 min to NIH3T3 fibroblasts that were 90 – 95% confluent, and nuclear extracts prepared as above. EMSAs were performed using 6 mg of crude liver or fibroblast extract as outlined by Ping et al. (1996). Binding reactions were carried out on ice in 15 mM Hepes, pH 7.9, 50 mM KCl, 10% glycerol, 2 mM DTT, 0.12 mM EDTA, 5 mM MgCl2, 0.5 mg poly(dIdC):poly(dIdC), 0.2 mg salmon sperm DNA, 5 mg BSA, 0.125% NP-40, and 0.5 mM PMSF. The Site A probe, shown in Fig. 1B was end-labeled with [g --32P]ATP and purified on a non-de-
625
naturing acrylamide gel. The indicated competitor DNAs (50 ng per reaction) were annealed and incubated with the extracts for 5 min on ice prior to addition of the probe, which was incubated for an additional 30 min period. Consensus and mutant NF-1 oligos were purchased from Santa Cruz Biotechnology. For antibody supershift assays, the antiserum (1–2 ml) was added to the reactions for 30 min after incubation with the probe. Pre-immune and anti-CTF antisera were kindly provided by Dr Naoko Tanese (New York University, NY). The protein–DNA complexes were resolved on a 5% polyacrylamide gel (29:1 acrylamide:bis), fixed, dried and autoradiographed.
2.4. Methylation interference assays Both strands of an extended sequence spanning Site A (5%-TCAGATTCTCCGGCCCATGAGAGAACTGCTTGGCTGCAGGCCCAGCATCTGGAGC-3%) were synthesized and used for in vitro methylation interference analysis. Single-stranded oligos were labeled with [g-32P]ATP and T4 polynucleotide kinase, and after inactivating the enzyme by heating at 80°C for 3 min, they were annealed to their complementary unlabeled strands. Duplex DNAs were gel-purified. Methylation interference analysis was performed as described previously (Hasegawa et al., 1991). In these assays, the standard EMSA binding reaction was scaled up approximately 10-fold, using 5× 106 cpm of probe and the DNA was methylated prior to the reaction by treatment with DMS.
2.5. Western and Southwestern blots Western blotting was performed as described by Towbin et al. (1979), using 50–100 mg of the indicated nuclear extract. a-NF1 antiserum against the conserved amino-terminal region of NF1 (Santa Cruz Technology) was used at a dilution of 1:1000 for 1 h at room temperature. HRP-conjugated a-rabbit IgG (Sigma, St. Louis, MO), diluted 1:3000, was used as the secondary antiserum. Blots were developed using enhanced chemiluminescence according to the manufacturer’s directions (ECL systems, Amersham Life Sciences, Arlington Heights, IL). Southwestern (DNA-protein) blotting was performed as described by Gao et al. (1996). Nuclear proteins (30 – 50 mg) from mouse liver or fibroblasts were denatured in the presence of b-mercaptoethanol, separated by 10% SDS-PAGE, and electroblotted onto nitrocellulose membranes (Micron Separations, MA) in transfer buffer (25 mM Tris–HCl, pH 8.3, 192 mM glycine, 20% methanol, 1 mM EDTA). Proteins were renatured by step-wise incubation in 6 M guanidine hydrochloride containing Z buffer (5 mM Hepes, pH7.5, 8% glycerol,
626
S.N. Kumar, J.M. Boss / Molecular Immunology 37 (2000) 623–632
50 mM NaCl, 12.5 mM MgCl2, 10 mM ZnSO4, and 0.01% Nonidet P-40), followed by three consecutive dilutions of the solution with Z buffer at 4°C over a period of 8 h. Non-specific sites on the membrane were then blocked with Z buffer containing 3% nonfat dried milk and 3 mg poly(dI:dC) – poly(dI:dC) for 30 min at 25°C. After a brief incubation in Z buffer containing 0.25% nonfat dried milk, 500 000 cpm/ml [g-32P]ATP end-labeled Site A probe was incubated with the membrane for 16 h at 4°C. Blots were washed three times in Z buffer and analyzed using a phosphorimager (Molecular Dynamics).
2.6. Chromatin immunoprecipitation (ChIP) assay ChIP assays were performed as outlined by Moreno et al. (1999) with the following modifications. NIH 3T3 fibroblasts grown to about 70% confluency were used instead of B lymphocytes. Chromatin-immune complexes were formed with the indicated antiserum (ca. 5 mg) and collected by incubating for 1 h at 4°C with Protein A-sepharose beads instead of para-magnetic beads. Following extensive washing as described, the purified DNA was dissolved in 30 ml water from which 3 ml were used as a template for PCR. The primers used for PCR amplification span a 121 bp region that encodes Site A and were 5%-GCTCAGACTAGGCCTTTGTTGAGT and 5%-GCAGTATTGGAAGTTCCCAGACCC.
2.7. Transient transfection assays Transient transfections were performed in NIH3T3 cells by electroporation of approximately 5× 106 cells/ assay at 300 V and 980 mF as described previously (Ping et al., 1996), with addition of 1 mg of a luciferase expression vector to control for transfection efficiency. Transfections were done in triplicate, using 20 mg of reporter construct and where applicable 20 mg of the chimeric Gal4-VP16 activation vector. Cells were harvested 36 h post transfection and assayed for expression of the CAT reporter by ELISA as described by the manufacturer (Boehringer-Mannheim, Indianapolis, IN). Luciferase activity was assayed using a luminometer and the luciferase assay system of Promega Biotech, Madison, WI.
3. Results
3.1. The 5 % proximal region of Site A binds se6eral protein complexes Using in vivo genomic footprinting analysis to probe
the regulatory mechanism controlling MCP-1 induction by TNF, only the distal regulatory region designated as Site A was found to be occupied in a constitutive manner regardless of the expression status of the MCP1 gene (Ping et al., 1996; Klein et al., 1997). Upon TNF induction, which resulted in the occupancy of three additional distal regulatory region and three proximal regulatory region sites, Site A remained unchanged (Fig. 1A). Deletion of Site A from the expression vectors resulted in a substantial decrease in TNF mediated expression (ca. 50–18-fold) (Ping et al., 1996). Thus, while Site A was important for full expression, its presence was not required for induction. Together these data suggested Site A may function as a modulator of gene expression. To begin to test this hypothesis and understand how Site A may function, the protein(s) that bind to Site A were characterized. Because Site A was constitutively occupied in vivo, nuclear extracts were prepared from murine liver and murine fibroblast cultures and analyzed for their binding activity by EMSA. Gel-shift assays were performed using labeled Site A DNA (Fig. 1B) as a probe to determine the pattern of nuclear proteins binding to elements within Site A. Nuclear extracts from mouse liver produced approximately four bands, with the upper two bands exhibiting the strongest activity (Fig. 1C). The binding was shown to be specific as it was competed for by the addition of unlabelled Site A (Fig. 1C, lane 9) but not by a nonspecific competitor DNA (Fig. 1C, lane 10). Additionally, increasing amounts of poly(dIdC):poly(dIdC), a general non-specific competitor did not alter the pattern. To locate the site of binding on the Site A probe, DNAs spanning three different sections of Site A (Fig. 1B) were used as competitors in EMSAs (Fig. 1C). The SA-1 competitor DNA, but not SA-2 or SA-3, was able to compete for binding activity, suggesting that the Site A factor was binding to the 5% half of the Site A probe. Extracts prepared from both murine liver and fibroblasts were found to produce similar EMSA banding complexes; however, band b appeared to be stronger in the fibroblast extracts (Fig. 2). Nuclear extracts isolated from TNF-treated fibroblasts produced identical patterns to the non-treated fibroblast extracts presented (data not shown). Because all of the bands that formed were equally competed for by the specific competitors, this suggested that the factors that bind Site A may exist in several isoforms or multimeric combinations within the cell. It was, therefore, likely that the fibroblast extract has more of the band b isoform than of the band a isoform.
3.2. The 5 % half of Site A interacts with NF1 /CTF The 5% half of Site A contains a sequence that dis-
S.N. Kumar, J.M. Boss / Molecular Immunology 37 (2000) 623–632
Fig. 2. NF1 binds to Site A. Nuclear extracts prepared from mouse liver and NIH3T3 fibroblasts were used in EMSAs using Site A oligo as probe. Competitor DNAs (50 ng/assay) are: NS, nonspecific DNA; Site A, unlabelled probe DNA; NF1, a commercial consensus NF1 binding site; and NF1 mut, a commercial NF1 site with a mutation that blocks binding. Antibody mediated supershift assays were carried out by the addition of antibody 20 min after the reaction was initiated. Antibodies used were to NF1/CTF and Sp1 as indicated. The supershifted bands are indicated by an asterisk. No competitor or antibody was added to lanes 2 and 10. Lane 1 had no extract. The two major bands (a and b) discussed in the text are indicated.
plays homology to a NF1/CTF consensus DNA binding site (Fig. 3) and shows a 14/20 bp match with a site in the promoter of the aspartate aminotransferase promoter that is thought to bind nuclear factor-1/CCAAT transcription factor (NF1/CTF) (Gronostajski et al., 1985; Garlatti et al., 1996). NF1/CTF proteins are encoded by four genes NF1-A, NF1-B, NF1-C and NF1-X. In some cases multiple isoforms generated by alternative RNA splicing can be found (Kruse and Sippel, 1994). NF1 sites are common to many genes and play varying roles in DNA replication and
Fig. 3. Alignment of NF1/CTF consensus sequences. Sequences were aligned manually. The references for the aspartate amino transferase gene (Garlatti et al., 1996), adenovirus gene (Gronostajski et al., 1985), and p53 gene (Lee et al., 1998) are provided.
627
Fig. 4. Identification of Site A binding protein by methylation interference and Southwestern analysis. Bands a and b shown in Fig. 1C and Fig. 2 were analyzed by methylation interference using probes labeled on the coding or non-coding strand and compared to unmethylated free probe (F). Arrows denote the guanines that inhibit protein binding by methylation interference. The sequence of the region analyzed with Site A shaded shows the positions of the methylation sensitive sites in this assay (stars) and the positions obtained from previous IVGF studies (Ping et al., 1996) (open arrows show protected sites while shaded arrows show hypersensitive sites).
transcriptional activation (Jones et al., 1987). To determine if this site is indeed the binding site of the protein/DNA complexes, in vitro methylation interference assays were performed using DMS-treated Site A DNA with extracts from mouse liver. In this experiment, the EMSA reaction was scaled up 10-fold and the complexes were resolved by a preparative EMSA gel. Bands a and b and the unbound probe were excised and the DNA purified, cleaved with piperidine, and analyzed on a DNA sequencing gel. Compared to the unbound lanes (Fig. 4) strong methylation interference is observed at G-2437 and G-2436 of the coding strand and G-2435 of the non-coding strand. Weaker interference was observed at G-2429 on the coding strand. The patterns for both bands a and b were similar. A summary and comparison of the methylation interference data with the IVGF (Ping et al., 1996), which is an in vivo methylation protection experiment, is shown in Fig. 4. The data showed that overlapping patterns were generated with the two techniques on the 5% half of Site A. This also indicates that we did not detect the factor that binds to the 3% half of Site A. Importantly, the 5% binding site overlaps the NF1 site. Demonstration that the observed patterns were due to the NF1 site was investigated further using three different in vitro approaches. The first was to determine if a canonical NF1 site but not a mutant NF1 site could compete for factor binding. As shown in Fig. 2, lanes 5
628
S.N. Kumar, J.M. Boss / Molecular Immunology 37 (2000) 623–632
and 6, a consensus NF1 oligo could compete for all binding activity, but the mutant NF1 DNA could not. The second approach investigated whether antibodies to NF1 could alter the migration of the complexes in the EMSA. Indeed, the addition of antibodies reactive to the C-terminal half of NF1 to the DNA-binding reaction resulted in a supershift in both the liver and fibroblast extracts. The supershifted band appears in the wells and can be resolved using a lower crosslinked gel (69:1 acrylamide:bis); however, in such gels the resolution of the bands a and b become diffuse (data not shown). The third in vitro approach compared the patterns derived from a Western blot analysis of the nuclear extracts for NF1 protein with those of a Southwestern blot, in which the same material was analyzed for binding of a Site A DNA probe. In this experiment nuclear extracts prepared from liver, NIH3T3 fibroblasts, and NIH3T3 fibroblasts treated with TNF were analyzed on the same gel. The resulting Western blot showed a predominant band at 32 – 34 kDa in all three extracts, which corresponds to the size reported by Gao et al. (1996) and Adams et al. (1995) for rat NF1 (Fig. 5). Additional bands appeared in the liver extract at higher molecular weights and are either crossreactive proteins or isoforms of NF1. The Southwestern blot showed a single band in all three extracts that aligns precisely with the NF1 bands in the Western blot.
Fig. 6. Chromatin immunoprecipitation assay shows NF1 protein binds to Site A in vivo. A 2.5% agarose gel of the ChIP assay performed to determine if the NF1/CTF protein bound to Site A in the MCP-1 promoter in vivo is shown. NIH3T3 fibroblasts were treated with formaldehyde to crosslink their DNA binding proteins to DNA. Chromatin was isolated and immunoprecipitated as described in Section 2 using either no antibody (lane 3), pre-immune antisera (lane 2) or anti-CTF antisera (lane 5) from Dr Tanese (New York University, NY). Positive and negative controls for the assay included genomic DNA (lane 1) or no DNA (lane 4), respectively. M, molecular weight markers.
3.3. NF1 binds to Site A in 6i6o
Fig. 5. Western and Southwestern analysis of Site A protein show overlapping patterns. SDS-PAGE of nuclear extracts prepared from mouse liver (L) or NIH3T3 fibroblast cultures treated with (F +) or without (F − ) TNF were blotted onto nitrocellulose membranes. The blot was divided in half. One half was probed with antibody to NF1 which identifies a major 32–34 kDa protein in all three extracts. The other half was probed with radiolabeled Site A as described.
To determine if NF1 associated with Site A in vivo, the chromatin immunoprecipitation assay recently described for analyses of histone acetylation and other DNA-specific transcription factor binding was employed (Dedon et al., 1991; Li et al., 1999; Moreno et al., 1999). In this assay, fibroblasts are fixed in formaldehyde, chromatin is isolated, and sheared by sonication. Anti-NF1 antibodies or preimmune antiserum were used to immunoprecipitate the crosslinked protein-DNA complexes. The crosslinks of the precipitated complexes were reversed and the DNA that was associated with the complex was used as a template for PCR. Primers spanning Site A were used to PCR amplify the DNA in the precipitated complexes. A strong amplified product of 121 bp spanning Site A was obtained using antisera to NF1 (Fig. 6). No band was observed in the pre-immune sera lane. These data suggest that NF1 binds to Site A of the MCP-1 gene in vivo.
S.N. Kumar, J.M. Boss / Molecular Immunology 37 (2000) 623–632
3.4. Site A functions as a position and orientation-independent modulator of transcription To begin to understand the role of Site A in mediating MCP-1 transcription, Site A was cloned into an enhancerless CAT reporter construct. This reporter construct has the TATA promoter sequences from the MHC class II HLA-DRA gene as well as a yeast Gal4 DNA binding site but does not contain the MHC class II regulatory elements. Regardless of Site A’s orientation, proximity to the TATA element, or whether there were two copies versus one copy of Site A, no expression was detected in these transient transfection assays, suggesting that Site A was incapable of directing transcription on its own (Fig. 7). To determine if Site A could augment the activity of another DNA binding protein, advantage was taken of the Gal4 DNA binding site in the construction, which could bind Gal4-activator fusion proteins. Transient cotransfections with the Gal4 DNA binding domain alone and the reporter plasmid did not show an increase in activity of the reporter plasmid. Cotransfection of Gal4-VP16, a fusion that contains the transactivation region of the herpes simplex VP16 protein, showed a greater than 50-fold increase in activity over just the DNA binding domain vector (pSG424). Interestingly, the presence of Site A in any of the orientations described above led to a significant 1.6–2.5 fold increase in activity, the highest level occurring when the Gal4 site was flanked by Site A elements. Thus, while Site A did not display
629
enhancer-like activity by functioning independently of other elements, its presence was able to modulate the activity of a strong transactivator, GAL4-VP16. This activity is consistent with its role in the context of the MCP-1 gene, where it is responsible for a 2–3 fold increase in the level of expression from an NF-kB dependent regulatory element.
4. Discussion The focus of this study was to identify the factors binding to Site A of the MCP-1 gene and characterize the general activity of this region in modulating transcription. Previous studies have shown that TNF induction of MCP-1 involves the assembly of a multi-component complex in the distal regulatory region, containing NF-kB p65/p50 and p65/p65 heterodimers and the appearance of a strong DMS-hypersensitive site in the region (Ueda et al., 1994; Ping et al., 1996; Ueda et al., 1997; Ping et al., 1999b). In the current analysis, both in vitro and in vivo evidence has identified the factor binding to the 5% half of Site A as belonging to the NFI/CTF family of transcription factors. Using nuclear extracts prepared from both mouse liver and established fibroblast cell lines, a number of protein/DNA complexes were found. In liver cells the top two complexes produced identical methylation interference patterns, suggesting that the complexes were
Fig. 7. Site A facilitates transcriptional activation but does not activate on it’s own. Transient transfection analysis of chimeric Gal4–Site A constructs were carried out in NIH3T3 fibroblasts. A single Gal4 site and one or more Site A elements were introduced into a minimal pDR-CAT promoter/reporter construct, as indicated in the schematic, to assess the ability of Site A to function as a transcriptional activator. Thirty-six hours after transfection, the cultures were harvested and the level of the CAT protein measured by ELISA. The ELISA OD was then normalized to the expression of a cotransfected luciferase reporter construction. An average of three experiments are shown with the S.E. of the mean indicated.
630
S.N. Kumar, J.M. Boss / Molecular Immunology 37 (2000) 623–632
related. NF1/CTF family proteins are encoded by at least four independent genes (Inoue et al., 1990; Kruse and Sippel, 1994). Depending on the species the names of the genes differ. In the mouse, four isoforms have been identified as NF1-A, NF1-B, NF1-C, and NF1-X (Chaudhry et al., 1997). NF1 isoforms share extensive homology in their N-terminal domains, which is responsible for dimerization and DNA binding (Kruse and Sippel, 1994). The C-terminal regions of these proteins is more divergent and may be responsible for protein-protein interactions and transcriptional modulation. The C-terminal region contains a proline-rich transactivation domain that shares homology with the C-terminal domain (CTD) of the large subunit of RNA polymerase II (Wendler et al., 1994). These isoforms may also be expressed in a tissue specific manner during development (Chaudhry et al., 1997). Thus, the multiple DNA/protein complexes observed may be a result of different isoforms binding to the Site A probes in vitro. Fibroblasts displayed a similar pattern in the EMSA; however, the intensity of each of the protein/ DNA complexes was not as even as the liver cell extract. We interpret this to imply that one of the isoforms is expressed at a higher level in fibroblasts as compared to liver cells. It is interesting to note however, that the Southwestern blot analysis revealed only a single band. This may be due to the relative abundance of one isoform over another and the relatively low sensitivity of the Southwestern blot assay as compared to the EMSA. Depending on the isoform and promoter context, NF1/CTF family members may have different roles with respect to gene expression. In one system using JEG-3 cells, an embryonic choriocarcinoma cell line, the transactivation potential of the four isoforms on the MMTV promoter was NF1-B\NF1-X \ NF1-C\ NF1-A (Chaudhry et al., 1997). Consistent with the above results, another study, comparing the NF1-B and NF1-C splice variants found that NF1-B isoforms had a greater transactivation potential on a large number of minimal promoters than the NF1-C isoforms (Osada et al., 1999). Interestingly, a repression domain in NF1-A consistently showed lower levels of activity when cotransfected with the reporter plasmids (Osada et al., 1999). Similarly, the rat homolog of NF1-A, NF1-L, was found to function in the negative regulation of the rat peripherin gene (Adams et al., 1995). Thus, NF1 isoforms may have distinct functions in regulating gene expression. Given the data that show that deletion of Site A results in a decrease in expression of the MCP-1 gene and the fact that Site A augments the stimulatory ability of the acidic activator in the current system, it is likely that the NF1 family member that binds Site A is one which provides transcriptional activation signals. NF1 proteins have been found to function through a variety of mechanisms, although all of these involve
protein-protein interactions. NF1/CTF proteins can facilitate the assembly of a functional transcriptional initiation complex (Dusserre and Mermod, 1992) and associate with components of basal transcriptional machinery, including TFIIB (Kim and Roeder, 1994), TATA-binding protein (Xiao et al., 1994) and TBP-associated factors (Dusserre and Mermod, 1992). NF1 is also known to interact with other sequence specific transcription factors such as ATF2 (Alonso et al., 1996), thyroid transcription factor (Ortiz et al., 1999) or mediate hormone-induced gene expression by co-operating or competing with nuclear receptors of insulin (Cooke and Lane, 1999), glucocorticoids (Garlatti et al., 1996), androgens (Darne et al., 1998), and progestins (Bruggemeier et al., 1990). We found in our previous studies that Site A exerted a modulatory effect on NF-kB dependent transcription of the MCP-1 gene (Ping et al., 1996). This effect was substantial as the removal of Site A resulted in a two-thirds reduction of activity. The enhancing effect of NF1 on NF-kB-mediated transcription could be due to a direct synergistic effect, such as the stabilization of binding of factors in that region. However, in vitro EMSAs to determine if NF1 and NF-kB could bind cooperatively to the distal regulatory element (Site A/ kB sites) of MCP-1 did not show cooperative binding of the two factors (data not shown). A more intriguing role may be in the recruitment of coactivators to the region. The coactivator CBP, which interacts with many transcription factors including CREB, NF-kB, and nuclear receptors, also can interact with NF1 (Leahy et al., 1999). NF1 binds to CBP at it’s CREBbinding site (Leahy et al., 1999). CBP, which exhibits intrinsic histone acetyltransferase activity, is thought to help remodel chromatin to allow transcription factor access. The binding of NF1 to Site A in a constitutive manner may be related directly to its interactions with CBP. In such a scenario, NF1 and CBP allow the chromatin to remain open. This open structure would be important for the subsequent recruitment of NF-kB in a manner that allows the MCP-1 gene to respond quickly to external stimuli such as TNF. Contacts of NF1 with histone H3 (Alevizopoulos et al., 1995) and nucleosomal cores suggest interaction with chromatin components that may play a role in transcriptional control as in the case of the mouse mammary tumor virus (MMTV) promoter (Beato and Chavez, 1997). It has been shown that NF1/CTF-1 can counteract the repression mediated by histone H1 (Dusserre and Mermod, 1992). Thus, independent of CBP, NF1 binding may promote chromatin accessibility to a region. Alternatively, the binding of NF1 may stabilize nucleosome positioning in a region. Thus, although NF1 binding to its site shows variability that is dependent on nucleosome positioning (Blomquist et al., 1999), once bound it may control repositioning within a region.
S.N. Kumar, J.M. Boss / Molecular Immunology 37 (2000) 623–632
Not all NF1 sites are likely to act the same. As described above, depending on the gene and available isoforms, NF1 can have divergent effects. With regard to TNF induction, we noted an NF1 site in the TNF response region of the manganous superoxide dismutase gene (MnSOD) (Jones et al., 1997). This TNF response region was located in an intron several kb downstream of the start of transcription. The interesting difference between the MCP-1 NF1 site and the MnSOD site is that the MnSOD site was only occupied following TNF induction. Thus, the context of the NF1 site must be extremely important for its function, even though it is of remarkable coincidence that it was also in close proximity to a functional NF-kB motif. Due to its constitutive occupancy, Site A represents a novel region within the MCP-1 flanking DNA. Its role as modulator of transcriptional activators is now established and is most likely to occur through the action of NF1. Its role in chromatin accessibility in this region has yet to be investigated but may shed light on how regulatory elements several kb upstream of the transcriptional start site can communicate and direct transcription of their downstream promoters. Acknowledgements We thank members of the laboratory for critique and suggestions of this work. We also thank Dr D. Ping for his contributions to the beginning of this project. We acknowledge Dr N. Tanese of New York University for kindly providing pre-immune and anti-CTF antisera. We also thank Yvonne DeBellotte for secretarial assistance with preparation of the manuscript. This work was supported by the National Cancer Institute of the NIH CA74271. References Adams, A.D., Choate, D.M., Thompson, M.A., 1995. NF1-L is the DNA-binding component of the protein complex at the peripherin negative regulatory element. J. Biol. Chem. 270, 6975 – 6983 (Published erratum appears in J. Biol. Chem. 18 August 1995 270 (33), 19668). Alevizopoulos, A., Dusserre, Y., Tsai-Pflugfelder, M., von der Weid, T., Wahli, W., Mermod, N., 1995. A proline-rich TGF-b-responsive transcriptional activator interacts with histone H3. Genes Dev. 9, 3051 – 3066. Alonso, C.R., Pesce, C.G., Kornblihtt, A.R., 1996. The CCAATbinding proteins CP1 and NF-1 cooperate with ATF-2 in the transcription of the fibronectin gene. J. Biol. Chem. 271 (36), 22271 – 22279. Baggiolini, M., Dahinden, C.A., 1994. CC chemokines in allergic inflammation. Immunol. Today 15, 127–133. Baggiolini, M., Dewald, B., Moser, B., 1997. Human chemokines: an update. Annu. Rev. Immunol. 15, 675–705. Beato, M., Chavez, S., 1997. Nucleosome-mediated synergism between transcription factors on the mouse mammary tumor virus promoter. Proc. Natl. Acad. Sci. USA 94 (7), 2885–2890.
631
Blomquist, P., Belikov, S., Wrange, O., 1999. Increase nuclear factor 1 binding to its nucleosomal site mediated by sequence-dependent DNA structure. Nucleic Acids Res. 27 (2), 517 – 525. Brach, M.A., Gruss, H-J., Riedel, D., Asano, Y., Vos, S.D., Herrmann, F., 1992. Effect of antiinflammatory agents on synthesis of MCP-1/JE transcripts by human blood monocytes. Mol. Pharmacol. 42, 63 – 68. Bruggemeier, U., Rogge, L., Winnacker, E.L., Beato, M., 1990. Nuclear factor I acts as a transcription factor on the MMTV promoter but competes with steroid hormone receptors for DNA binding. EMBO J. 9 (7), 2233 – 2239. Chaudhry, A.Z., Lyons, G.E., Gronostajski, R.M., 1997. Expression patterns of the four nuclear factor I genes during mouse embryogenesis indicate a potential role in development. Dev. Dyn. 208, 313 – 325. Cochran, B.H., Reffel, A.C., Stiles, C.D., 1983. Molecular cloning of gene sequences regulated by platelet-derived growth factor. Cell 33, 939 – 947. Cooke, D.W., Lane, M.D., 1999. Transcription factor NF1 mediates repression of the GLUT4 promoter by cyclic-amp. Biochem Biophys. Res. Commun. 260, 600 – 604. Darne, C., Veyssiere, G., Jean, C., 1998. Phorbol ester causes ligandindependent activation of the androgen receptor. Eur. J. Biochem. 256, 541 – 549. Dedon, P.C., Soults, J.A., Allis, C.D., Gorovsky, M.A., 1991. A simplified formaldehyde fixation and immunoprecipitation technique for studying protein – DNA interactions. Anal. Biochem. 197, 83 – 90. Dusserre, Y., Mermod, N., 1992. Purified cofactors and histone H1 mediate transcriptional regulation by CTF/NF-1. Mol. Cell. Biol. 12 (11), 5228 – 5237. Frazier-Jessen, M.R., Kovacs, E.J., 1995. Estrogen modulation of JE/monocyte chemoattractant protein-1 mRNA expression in murine macrophages. J. Immunol. 154, 1838 – 1845. Freter, R.R., Irminger, J.-C., Porter, J.A., Jones, S.D., Stiles, C.D., 1992. A novel 7-nucleotide motif located in 3% untranslated sequences of the immediate-early gene set mediates platelet-derived growth factor induction of the JE gene. Mol. Cell. Biol. 12, 5288 – 5300. Gao, B., Jiang, L., Kuno, G., 1996. Transcriptional regulation of a1b adrenergic receptors (a1b AR) by nuclear factor 1 (NF1): a decline in the concentration of NF1 correlates with the downregulation of a1b AR gene expression in regnerating liver. Mol. Cell. Biol. 16 (11), 5997 – 6008. Garlatti, M., Aggerbeck, M., Bouguet, J., Barouki, R., 1996. Contribution of a nuclear factor 1 binding site to the glucocorticoid regulation of the cytosolic aspartate aminotransferase gene promoter. J. Biol. Chem. 271 (51), 32629 – 32634. Glabinski, A.R., Balasingam, V., Tani, M., et al., 1996. Chemokine monocyte chemoattractant protein-1 is expressed by astrocytes after mechanical injury to the brain. J. Immunol. 156, 4363–4368. Gronostajski, R.M., Adhya, S., Nagata, K., Guggenheimer, R.A., Hurwitz, J., 1985. Site-specific DNA binding of nuclear factor I: analyses of cellular binding sites. Mol. Cell. Biol. 5 (5), 964–971. Hanazawa, S., Takeshita, A., Kitano, S., 1994. Retinoic acid suppression of c-fos gene inhibits expression of tumor necrosis factor-ainduced monocyte chemoattractant JE/MCP-1 in clonal osteoblastic MC3T3-E1 cells. J. Biol. Chem. 269, 21379–21384. Hasegawa, S.L., Sloan, J.H., Reith, W., Mach, B., Boss, J.M., 1991. Regulatory factor-X binding to mutant HLA-DRA promoter sequences. Nucleic Acids Res. 19, 1243 – 1249. Howard, O.M.Z., Ben-Baruch, A., Oppenheim, J.J., 1996. Chemokines: progress toward identifying molecular targets for therapeutic agents. Trends Biotechnol. 14, 46 – 51. Inoue, T., Tamura, T., Furuichi, T., Mikoshiba, K., 1990. Isolation of complementary DNAs encoding a cerebellum-enriched nuclear factor I family that activates transcription from the mouse myelin basic protein promoter. J. Biol. Chem. 265, 19065 – 19070.
632
S.N. Kumar, J.M. Boss / Molecular Immunology 37 (2000) 623–632
Jones, K.A., Kadonaga, J.T., Rosenfeld, P.J., Kelly, T.J., Tjian, R., 1987. A cellular DNA-binding protein that activates eukaryotic transcription and DNA replication. Cell 48, 79–89. Jones, P.L., Ping, D., Boss, J.M., 1997. TNF-a and IL-1b regulate the murine manganese superoxide dismutase gene through a complex intronic enhancer. Mol. Cell Biol. 17, 6970–6981. Kawahara, R.S., Deng, Z-W., Deuel, T.F., 1991. Glucocorticoids inhibits the transcriptional induction of JE, a platelet-derived growth factor-inducible gene. J. Biol. Chem. 266, 13261–13266. Kim, T.K., Roeder, R.G., 1994. CTD-like sequences are important for transcriptional activation by the proline-rich activation domain of CTF1. Nucleic Acids Res. 22 (2), 251. Klein, S.A., Dobmeyer, J.M., Dobmeyer, T.S., et al., 1997. Demonstration of the Th1 to Th2 cytokine shift during the course of HIV-1 infection using cytoplasmic cytokine detection on single cell level by flow cytometry. AIDS 11 (9), 1111–1118. Koch, A.E., Kunkel, S.L., Strieter, R.M., 1995. Cytokines in rheumatoid arthritis. J. Invest. Med. 43, 28–38. Kopydlowski, K.M., Salkowski, C.A., Cody, M.J., et al., 1999. Regulation of macrophage chemokine expression by lipopolysaccharide in vitro and in vivo. J. Immunol. 163, 1537–1544. Kruse, U., Sippel, A.E., 1994. The genes for transcription factor nuclear factor I give rise to corresponding splice variants between vertebrate species. J. Mol. Biol. 238, 860–865. Leahy, P., Crawford, D.R., Grossman, G., Gronostajski, R.M., Hanson, R.W., 1999. CREB binding protein coordinates the function of multiple transcription factors including nuclear factor I to regulate phosphoenolpyruvate carboxykinase (GTP) gene transcription. J. Biol. Chem. 274 (13), 8813–8822. Lee, M., Song, H., Park, S., Park, J., 1998. Transcription of the rat p53 gene is mediated by factor binding to two recognition motifs of NF1-like protein. Biol. Chem. 379, 1333–1340. Li, X.Y., Virbasius, A., Zhu, X., Green, M.R., 1999. Enhancement of TBP binding by activators and general transcription factors. Nature 399, 605 – 609. Mantovani, A., Sozzani, S., Bottazzi, B., et al., 1993. Monocyte chemotactic protein-1 (MCP-1): signal transduction and involvement in the regulation of macrophage traffic in normal and neoplastic tissues. Adv. Exp. Med. Biol. 351, 47–54. Moreno, C.S., Beresford, G., Louis-Plence, P., Morris, A.C., Boss, J.M., 1999. CREB regulates MHC class II expression in a CIITAdependent manner. Immunity 10, 143–151. Negus, R.P.M., Turner, L., Burke, F., Balkwill, F.R., 1998. Hypoxia down-regulates MCP-1 expression: implications for macrophage distribution in tumors. J. Leukocyte Biol. 63, 758–765. Nelken, N.A., Coughlin, S.R., Gordon, D., Wilcox, J.N., 1991. Monocyte chemoattractant protein-1 in human artheromatous plaques. J. Clin. Invest. 88, 1121–1127. Ortiz, L., Aza-Blanc, P., Zannini, M., Cato, A.C.B., Santisteban, P., 1999. The interaction between the forehead thyroid transcription factorTTF-2 and the constitutive factor CTF/NF-1 is required for efficient hormonal regulation of the thyroperoxidase gene transcription. J. Biol. Chem. 274 (21), 15213–15221. Osada, S., Matsubara, T., Daimon, S., et al., 1999. Expression, DNA-binding specifity and transcriptional regulation of nuclear factor 1 family proteins from rat. Biochem. J. 342, 189–198. Ping, D., Jones, P.L., Boss, J.M., 1996. TNF regulates the in vivo occupancy of both distal and proximal regulatory regions of the murine monocyte chemoattractant protein-1 (MCP-1/JE) gene. Immunity 4, 455 – 469. Ping, D., Boekhoudt, G., Boss, J.M., 1999. Trans-retinoic acid blocks PDGF-induced expression of the murine monocyte chemoattractant-1 gene by blocking the assembly of a promoter proximal Sp1 binding site. J. Biol. Chem. 274 (in press).
.
Ping, D., Boekhoudt, G.H., Rogers, E.M., Boss, J.M., 1999b. NF-kB p65 mediates the assembly and activation of the tumor necrosis factor responsive element of the murine monocyte chemoattractant-1 gene. J. Immunol. 162, 727 – 734. Ping, D., Boekhoudt, G.H., Zhang, F., et al., 2000. Sp1 binding is critical for promoter assembly and activation of the MCP-1 gene by tumor necrosis factor. J. Biol. Chem. 275, 1708 – 1714. Ransohoff, R.M., Glabinski, A., Tani, M., 1996. Chemokines in immune-mediated inflammation of the central nervous system. Cytokine Growth Factor Rev. 7, 35 – 46. Riley, J.L., Boss, J.M., 1993. Class II MHC transcriptional mutants are defective in higher order complex formation. J. Immunol. 151, 6942 – 6953. Rollins, B.J., Morrison, E.D., Stiles, C.D., 1988. Cloning and expression of JE, a gene inducible by platelet-derived growth factor and whose product has cytokine-like properties. Proc. Natl. Acad. Sci. USA 85, 3738 – 3742. Satriano, J.A., Hora, K., Shan, Z., Stanley, E.R., Mori, T., Schlondroff, D., 1993. Regulation of monocyte chemoattractant protein-1 and macrophage colony-stimulating factor-1 by IFNgamma, tumor necrosis factor-alpha, IgG aggregates, and cAMP in mouse mesangial cells. J. Immunol. 150, 1971 – 1978. Schwartz, C.J., Valente, A.J., Sprague, E.A., Kelley, J.L., Nerem, R.M., 1991. The pathogenesis of atherosclerosis: an overview. Clin. Cardiol. 14, I-1 – I-16. Shapiro, D.J., Sharp, P.A., Wahli, W.W., Keller, M.J., 1988. A high-efficiency HeLa cell nuclear transcription extract. DNA 7, 47 – 55. Shyy, J.Y.-J., Li, Y.-S., Lin, M.-C., et al., 1995. Multiple cis-elements mediate shear stress-induced gene expression. J. Biomech. 28, 1451 – 1457. Taylor, P.C., Peters, A.M., Paleolog, E., et al., 2000. Reduction of chemokine levels and leukocyte traffic to joints by tumor necrosis factor alpha blockade in patients with rheumatoid arthritis. Arthritis Rheum. 43 (1), 38 – 47. Towbin, H., Staehelin, T., Gordon, J., 1979. Electrophoretic transfer of proteins from polyacrylamide gels to nitrocellulose sheets: procedure and some applications. Proc. Natl. Acad. Sci. USA 76, 4350 – 4354. Ueda, A., Okuda, K., Ohno, S., et al., 1994. NF-kB and Sp1 regulate transcription of the human monocyte chemoattractant protein-1 gene. J. Immunol. 153, 2052 – 2063. Ueda, A., Ishigatsubo, Y., Okubo, T., Yoshimura, T., 1997. Transcriptional regulation of the human monocyte chemoattractant protein-1 gene. J. Biol. Chem. 272, 31092 – 31099. Wendler, W., Altmann, H., Ludwig-Winnacker, E., 1994. Transcriptional activation of NFI/CTF1 depends on a sequence motif strongly related to the carboxyterminal domain of RNA polymerase II. Nucleic Acids Res. 22, 2601 – 2603. Xiao, H., Lis, J.T., Greenblatt, J., Friesen, J.D., 1994. The upstream activator CTF/NF1 and RNA polymerase II share a common element involved in transcriptional activation. Nucleic Acids Res. 22 (11), 1966 – 1973. Yla-Herttuala, S., Lipton, B.A., Rosenfeld, M.E., et al., 1991. Expression of monocyte chemoattractant protein 1 in macrophagerich areas of human and rabbit atherosclerotic lesions. Proc. Natl. Acad. Sci. USA 88, 5252 – 5256. Zhou, Z.H., Chartuvedi, L., Han, P., et al., 1998. Interferon-gamma induction of the human monocyte chemoattractant protein (hMCP)-1 gene in astrocytoma cells: functional interaction between a gamma activated site (GAS) and a GC-rich element. J. Immunol. 160, 2908 – 2916.
Site A of the MCP-1 distal regulatory region functions as a transcriptional modulator through the transcription factor NF1 Shantha N. Kumar, Jeremy M. Boss * Department of Microbiology and Immunology, Emory Uni6ersity School of Medicine, 1510 Clifton Road, Atlanta, GA 30322, USA Received 24 July 2000; accepted 14 November 2000
Abstract The monocyte chemoattractant protein-1 (MCP-1) functions to recruit monocytes and macrophages to areas of inflammation and is a prototypic chemokine subjected to coordinate regulation by immunomodulatory agents. TNF mediated regulation of MCP-1 occurs through a distal regulatory region located 2.5 kb upstream of the transcriptional start site. Within this region are two NF-kB motifs that are each critical for function. Site A, located within the distal regulatory region and upstream of the kB elements is required for maximal induction by TNF. However, unlike the kB elements and other MCP-1 regulatory elements, Site A is constitutively occupied by factors in vivo. To better understand the nature of Site A function, this report identified a Site A binding protein and provides a functional analysis of the element in driving transcription. The results showed that the transcription factor NF1/CTF binds to Site A both in vitro and in vivo. While Site A has no transcriptional activity on its own, it was found to augment the transcriptional activity of a GAL4-VP16 reporter system in an orientation and position independent manner. Because NF1 is known to interact with factors that modify nucleosomes, these results suggest a unique role for Site A in regulating MCP-1 expression. © 2001 Elsevier Science Ltd. All rights reserved. Keywords: MCP-1; Chemokine; Gene expression; NF-1
1. Introduction Monocyte chemoattractant protein-1 (MCP-1) is a proinflammatory CC chemokine secreted by fibroblasts, endothelial cells, vascular smooth muscle cells, monocytes, T cells, and other cell types that mediate the influx of cells to sites of inflammation (Baggiolini and Dahinden, 1994; Howard et al., 1996; Ransohoff et al., 1996; Baggiolini et al., 1997). MCP-1 expression has been observed in a large number of tissues during disease progression, including atherosclerosis (Schwartz et al., 1991; Nelken et al., 1991; Yla-Herttuala et al., 1991), arthritis (Koch et al., 1995; Taylor et al., 2000), and cancer (Mantovani et al., 1993; Negus et al., 1998; Kopydlowski et al., 1999). In these cases, the influx of macrophages into these tissues has been suggested to exacerbate the disease. Thus, the expression of MCP-1, which is likely to be critical for fighting infectious * Corresponding author. Tel.: +1-404-7275973; fax: + 1-4047271719. E-mail address: [email protected] (J.M. Boss).
disease, must be tightly regulated. Regulation of MCP1 expression has been shown to occur at the transcriptional level by stimulatory agents like TNF (Ping et al., 1996; Jones et al., 1997), IFN-g (Satriano et al., 1993; Zhou et al., 1998), PDGF (Cochran et al., 1983; Rollins et al., 1988; Freter et al., 1992), and stress factors (Shyy et al., 1995; Glabinski et al., 1996). In contrast, retinoic acid, glucocorticoids, and estrogen have been shown to down modulate MCP-1 expression (Kawahara et al., 1991; Brach et al., 1992; Hanazawa et al., 1994; Frazier-Jessen and Kovacs, 1995). Because macrophages secrete TNF, and TNF induces MCP-1, one can envision a positive feedback loop generated by both of these cytokines for the rapid infiltration of macrophages to an area of inflammation. The transcriptional regulation of MCP-1 is controlled by two complex elements (Fig. 1A). The proximal regulatory region is sufficient for PDGF and IFN-g induced expression and is necessary for TNF-induced expression. The proximal regulatory region consists of three subelements: GC box, Site B, and kB-3 (Ping et al., 1996). The GC box, which binds Sp1 in vivo and in
0161-5890/01/$ - see front matter © 2001 Elsevier Science Ltd. All rights reserved. PII: S 0 1 6 1 - 5 8 9 0 ( 0 0 ) 0 0 0 9 7 - 3
624
S.N. Kumar, J.M. Boss / Molecular Immunology 37 (2000) 623–632
vitro is critical to all aspects of MCP-1 regulation (Ping et al., 1999a, 2000). While Site B and the kB-3 box are functional in vivo (Ping et al., 1996, 1999b), the factors that bind these elements have not been defined. In the human MCP-1 gene, kB-3 is required for IFN-g induction and interacts with STAT1 (Satriano et al., 1993; Zhou et al., 1998). The distal regulatory region contains four subelements (Ping et al., 1996). Two NF-kB binding sites, termed kB-1 and kB-2 surround a DMS hypersensitive region (HS). Both the kB sites and the HS region are required for MCP-1 regulation by TNF (Ping et al., 1999b). The kB sites bind p65 and at least one of these sites binds a p65/50 heterodimer (Ueda et al., 1994, 1997; Ping et al., 1999b). The protein binding to the HS is not known. The spacing between the two kB sites is not strict and there does not appear to be a requirement for HMG-I(Y), a minor groove architectural protein
Fig. 1. EMSAs of nuclear extracts from mouse liver using Site A oligo as a probe display a specific banding pattern. (A) Schematic map of the murine MCP-1 gene’s regulatory elements. (B) Site A sequence (shaded region) and schematic of DNA competitors SA-1, SA-2 and SA-3. (C) Liver nuclear extracts were tested to detect protein complexes binding to Site A in an EMSA in the presence of various specific and non-specific competitors. Lane 1, no extract; lane 2, no competitor; lanes 3 –5, poly(dIdC) competitor DNA (concentrations); lanes 6 – 8 contain SA-1, 2, and 3 competitor DNAs; lane 9 contains Site A competitor; and lane 10 contains a non-specific competitor. All competitor DNAs were used at 50 ng/assay. Bands a and b were used in subsequent methylation interference analysis.
found to be important for other TNF induced genes (Ping et al., 1999b). Site A, which lies just upstream of the kB sites, is required for maximal expression in fibroblasts treated with TNF. Deletion of Site A from reporter constructions containing the upstream kB sites resulted in a two-thirds loss in activity, suggesting an important role for Site A (Ping et al., 1996). The Site A sequence, which consists of 41 bp is highly conserved between murine and human MCP-1 genes. Site A is unique in the MCP-1 gene with regard to protein occupancy in vivo (Ping et al., 1996). In vivo genomic footprinting studies (IVGF) performed on the MCP-1 gene in both control and TNF-treated murine fibroblasts found that Site A sequences were resistant or hypersensitive to methylation by dimethylsulfate (DMS), suggesting that proteins were constitutively bound to the site. This constitutive occupancy was unexpected as in control cells no other regions of protein occupancy were observed between Site A and the start of transcription, which is 2.5 kb downstream of Site A. This included the Sp1 binding site as well as kB-3 and Site B, sites that were capable of binding factors in vitro. Upon TNF treatment, Site A showed no changes in its IVGF pattern, while all the sites described above showed distinct and clear changes in their IVGF pattern, indicating transcription factor binding and assembly. These results suggested that Site A may play an important and unique role in MCP-1 regulation by TNF. In this report we have undertaken a biochemical approach to identify Site A binding factors. Nuclear extracts prepared from mouse liver and fibroblasts revealed a distinct pattern of binding to Site A. This banding pattern was analyzed further by in vitro methylation interference assays, which identified a site that was similar to a NF1 transcription factor binding site. Further analyses showed that the in vitro binding patterns were most likely due to NF1 binding. In vivo chromatin immunoprecipitation (ChIP) assays were used with aNF1 antisera to show that NF1 could be found on Site A in fibroblasts. Additionally, we analyzed the ability of Site A to stimulate/interfere with the transactivation of an acidic activator protein. The results showed that Site A had no intrinsic transcriptional activity but that it could augment the activity of other transcriptional activators in an orientation and position independent manner. Together these results describe a role for NF-1 and Site A as a modulator of transcriptional activation.
2. Materials and methods
2.1. Cell lines and tissue culture NIH3T3 fibroblasts were purchased from ATCC and
S.N. Kumar, J.M. Boss / Molecular Immunology 37 (2000) 623–632
grown in Dulbecco’s modified Eagle’s media supplemented with 10% bovine calf serum (Hyclone, Logan, UT), penicillin (50 U/ml), streptomycin (50 mg/ml) and L-glutamine (1 mM) (Life Sciences). Cells were grown at 37°C in an atmosphere of 8% CO2.
2.2. Plasmids The plasmid pDRGY-CAT served as the initial subcloning reporter vector (Riley and Boss, 1993). pDRGY-CAT contains a single Gal4 DNA binding site 5%-CGGAAGACTCTCTCCTCCG upstream of the Y box element and the minimal promoter region (−56– + 3) of the human HLA-DRA gene. These sequences drive expression of the chloramphenicol acetyltransferase (CAT) gene. pDRG-CAT, which lacks the Y box, was derived from pDRGY-CAT by double digestion with XbaI and SalI, filling in the ends and ligating the plasmid ends together (Riley and Boss, 1993). pDRG-CAT is inactive in B cells, fibroblasts, and T cells. In current studies pDRGY-CAT was also digested with XbaI and SalI, removing the Y box, the ends were filled in by the Klenow fragment of DNA polymerase and a double stranded Site A oligonucleotide 5%-AGAACTGCTTGGCTGCAGGCCCAGCATCTGGAGCTCACATT was blunt-end ligated with the ends to obtain plasmids pDRGSA-CAT and pDRGRSACAT, which have Site A in forward and reverse orientations, respectively. To make the plasmid pDRSAG-CAT, the plasmid pDRG-CAT was cut with HindIII, filled-in with the Klenow fragment of DNA polymerase and ligated with Site A oligonucleotide. Plasmid pDRSAGSA-CAT was created by cutting pDRGSA-CAT with HindIII, filling-in the ends as above and religating with Site A oligo upstream to the Gal4 site. All constructs were verified for integrity by DNA sequencing.
2.3. Nuclear extracts and electrophoretic mobility shift assays (EMSA) Nuclear extracts were prepared from NIH3T3 fibroblasts or mouse liver using the methods of Shapiro et al. (1988) and as described by Ping et al. (1996). For extracts from TNF treated cells, 500 U/ml of recombinant human TNF (Genzyme) was added for 30 min to NIH3T3 fibroblasts that were 90 – 95% confluent, and nuclear extracts prepared as above. EMSAs were performed using 6 mg of crude liver or fibroblast extract as outlined by Ping et al. (1996). Binding reactions were carried out on ice in 15 mM Hepes, pH 7.9, 50 mM KCl, 10% glycerol, 2 mM DTT, 0.12 mM EDTA, 5 mM MgCl2, 0.5 mg poly(dIdC):poly(dIdC), 0.2 mg salmon sperm DNA, 5 mg BSA, 0.125% NP-40, and 0.5 mM PMSF. The Site A probe, shown in Fig. 1B was end-labeled with [g --32P]ATP and purified on a non-de-
625
naturing acrylamide gel. The indicated competitor DNAs (50 ng per reaction) were annealed and incubated with the extracts for 5 min on ice prior to addition of the probe, which was incubated for an additional 30 min period. Consensus and mutant NF-1 oligos were purchased from Santa Cruz Biotechnology. For antibody supershift assays, the antiserum (1–2 ml) was added to the reactions for 30 min after incubation with the probe. Pre-immune and anti-CTF antisera were kindly provided by Dr Naoko Tanese (New York University, NY). The protein–DNA complexes were resolved on a 5% polyacrylamide gel (29:1 acrylamide:bis), fixed, dried and autoradiographed.
2.4. Methylation interference assays Both strands of an extended sequence spanning Site A (5%-TCAGATTCTCCGGCCCATGAGAGAACTGCTTGGCTGCAGGCCCAGCATCTGGAGC-3%) were synthesized and used for in vitro methylation interference analysis. Single-stranded oligos were labeled with [g-32P]ATP and T4 polynucleotide kinase, and after inactivating the enzyme by heating at 80°C for 3 min, they were annealed to their complementary unlabeled strands. Duplex DNAs were gel-purified. Methylation interference analysis was performed as described previously (Hasegawa et al., 1991). In these assays, the standard EMSA binding reaction was scaled up approximately 10-fold, using 5× 106 cpm of probe and the DNA was methylated prior to the reaction by treatment with DMS.
2.5. Western and Southwestern blots Western blotting was performed as described by Towbin et al. (1979), using 50–100 mg of the indicated nuclear extract. a-NF1 antiserum against the conserved amino-terminal region of NF1 (Santa Cruz Technology) was used at a dilution of 1:1000 for 1 h at room temperature. HRP-conjugated a-rabbit IgG (Sigma, St. Louis, MO), diluted 1:3000, was used as the secondary antiserum. Blots were developed using enhanced chemiluminescence according to the manufacturer’s directions (ECL systems, Amersham Life Sciences, Arlington Heights, IL). Southwestern (DNA-protein) blotting was performed as described by Gao et al. (1996). Nuclear proteins (30 – 50 mg) from mouse liver or fibroblasts were denatured in the presence of b-mercaptoethanol, separated by 10% SDS-PAGE, and electroblotted onto nitrocellulose membranes (Micron Separations, MA) in transfer buffer (25 mM Tris–HCl, pH 8.3, 192 mM glycine, 20% methanol, 1 mM EDTA). Proteins were renatured by step-wise incubation in 6 M guanidine hydrochloride containing Z buffer (5 mM Hepes, pH7.5, 8% glycerol,
626
S.N. Kumar, J.M. Boss / Molecular Immunology 37 (2000) 623–632
50 mM NaCl, 12.5 mM MgCl2, 10 mM ZnSO4, and 0.01% Nonidet P-40), followed by three consecutive dilutions of the solution with Z buffer at 4°C over a period of 8 h. Non-specific sites on the membrane were then blocked with Z buffer containing 3% nonfat dried milk and 3 mg poly(dI:dC) – poly(dI:dC) for 30 min at 25°C. After a brief incubation in Z buffer containing 0.25% nonfat dried milk, 500 000 cpm/ml [g-32P]ATP end-labeled Site A probe was incubated with the membrane for 16 h at 4°C. Blots were washed three times in Z buffer and analyzed using a phosphorimager (Molecular Dynamics).
2.6. Chromatin immunoprecipitation (ChIP) assay ChIP assays were performed as outlined by Moreno et al. (1999) with the following modifications. NIH 3T3 fibroblasts grown to about 70% confluency were used instead of B lymphocytes. Chromatin-immune complexes were formed with the indicated antiserum (ca. 5 mg) and collected by incubating for 1 h at 4°C with Protein A-sepharose beads instead of para-magnetic beads. Following extensive washing as described, the purified DNA was dissolved in 30 ml water from which 3 ml were used as a template for PCR. The primers used for PCR amplification span a 121 bp region that encodes Site A and were 5%-GCTCAGACTAGGCCTTTGTTGAGT and 5%-GCAGTATTGGAAGTTCCCAGACCC.
2.7. Transient transfection assays Transient transfections were performed in NIH3T3 cells by electroporation of approximately 5× 106 cells/ assay at 300 V and 980 mF as described previously (Ping et al., 1996), with addition of 1 mg of a luciferase expression vector to control for transfection efficiency. Transfections were done in triplicate, using 20 mg of reporter construct and where applicable 20 mg of the chimeric Gal4-VP16 activation vector. Cells were harvested 36 h post transfection and assayed for expression of the CAT reporter by ELISA as described by the manufacturer (Boehringer-Mannheim, Indianapolis, IN). Luciferase activity was assayed using a luminometer and the luciferase assay system of Promega Biotech, Madison, WI.
3. Results
3.1. The 5 % proximal region of Site A binds se6eral protein complexes Using in vivo genomic footprinting analysis to probe
the regulatory mechanism controlling MCP-1 induction by TNF, only the distal regulatory region designated as Site A was found to be occupied in a constitutive manner regardless of the expression status of the MCP1 gene (Ping et al., 1996; Klein et al., 1997). Upon TNF induction, which resulted in the occupancy of three additional distal regulatory region and three proximal regulatory region sites, Site A remained unchanged (Fig. 1A). Deletion of Site A from the expression vectors resulted in a substantial decrease in TNF mediated expression (ca. 50–18-fold) (Ping et al., 1996). Thus, while Site A was important for full expression, its presence was not required for induction. Together these data suggested Site A may function as a modulator of gene expression. To begin to test this hypothesis and understand how Site A may function, the protein(s) that bind to Site A were characterized. Because Site A was constitutively occupied in vivo, nuclear extracts were prepared from murine liver and murine fibroblast cultures and analyzed for their binding activity by EMSA. Gel-shift assays were performed using labeled Site A DNA (Fig. 1B) as a probe to determine the pattern of nuclear proteins binding to elements within Site A. Nuclear extracts from mouse liver produced approximately four bands, with the upper two bands exhibiting the strongest activity (Fig. 1C). The binding was shown to be specific as it was competed for by the addition of unlabelled Site A (Fig. 1C, lane 9) but not by a nonspecific competitor DNA (Fig. 1C, lane 10). Additionally, increasing amounts of poly(dIdC):poly(dIdC), a general non-specific competitor did not alter the pattern. To locate the site of binding on the Site A probe, DNAs spanning three different sections of Site A (Fig. 1B) were used as competitors in EMSAs (Fig. 1C). The SA-1 competitor DNA, but not SA-2 or SA-3, was able to compete for binding activity, suggesting that the Site A factor was binding to the 5% half of the Site A probe. Extracts prepared from both murine liver and fibroblasts were found to produce similar EMSA banding complexes; however, band b appeared to be stronger in the fibroblast extracts (Fig. 2). Nuclear extracts isolated from TNF-treated fibroblasts produced identical patterns to the non-treated fibroblast extracts presented (data not shown). Because all of the bands that formed were equally competed for by the specific competitors, this suggested that the factors that bind Site A may exist in several isoforms or multimeric combinations within the cell. It was, therefore, likely that the fibroblast extract has more of the band b isoform than of the band a isoform.
3.2. The 5 % half of Site A interacts with NF1 /CTF The 5% half of Site A contains a sequence that dis-
S.N. Kumar, J.M. Boss / Molecular Immunology 37 (2000) 623–632
Fig. 2. NF1 binds to Site A. Nuclear extracts prepared from mouse liver and NIH3T3 fibroblasts were used in EMSAs using Site A oligo as probe. Competitor DNAs (50 ng/assay) are: NS, nonspecific DNA; Site A, unlabelled probe DNA; NF1, a commercial consensus NF1 binding site; and NF1 mut, a commercial NF1 site with a mutation that blocks binding. Antibody mediated supershift assays were carried out by the addition of antibody 20 min after the reaction was initiated. Antibodies used were to NF1/CTF and Sp1 as indicated. The supershifted bands are indicated by an asterisk. No competitor or antibody was added to lanes 2 and 10. Lane 1 had no extract. The two major bands (a and b) discussed in the text are indicated.
plays homology to a NF1/CTF consensus DNA binding site (Fig. 3) and shows a 14/20 bp match with a site in the promoter of the aspartate aminotransferase promoter that is thought to bind nuclear factor-1/CCAAT transcription factor (NF1/CTF) (Gronostajski et al., 1985; Garlatti et al., 1996). NF1/CTF proteins are encoded by four genes NF1-A, NF1-B, NF1-C and NF1-X. In some cases multiple isoforms generated by alternative RNA splicing can be found (Kruse and Sippel, 1994). NF1 sites are common to many genes and play varying roles in DNA replication and
Fig. 3. Alignment of NF1/CTF consensus sequences. Sequences were aligned manually. The references for the aspartate amino transferase gene (Garlatti et al., 1996), adenovirus gene (Gronostajski et al., 1985), and p53 gene (Lee et al., 1998) are provided.
627
Fig. 4. Identification of Site A binding protein by methylation interference and Southwestern analysis. Bands a and b shown in Fig. 1C and Fig. 2 were analyzed by methylation interference using probes labeled on the coding or non-coding strand and compared to unmethylated free probe (F). Arrows denote the guanines that inhibit protein binding by methylation interference. The sequence of the region analyzed with Site A shaded shows the positions of the methylation sensitive sites in this assay (stars) and the positions obtained from previous IVGF studies (Ping et al., 1996) (open arrows show protected sites while shaded arrows show hypersensitive sites).
transcriptional activation (Jones et al., 1987). To determine if this site is indeed the binding site of the protein/DNA complexes, in vitro methylation interference assays were performed using DMS-treated Site A DNA with extracts from mouse liver. In this experiment, the EMSA reaction was scaled up 10-fold and the complexes were resolved by a preparative EMSA gel. Bands a and b and the unbound probe were excised and the DNA purified, cleaved with piperidine, and analyzed on a DNA sequencing gel. Compared to the unbound lanes (Fig. 4) strong methylation interference is observed at G-2437 and G-2436 of the coding strand and G-2435 of the non-coding strand. Weaker interference was observed at G-2429 on the coding strand. The patterns for both bands a and b were similar. A summary and comparison of the methylation interference data with the IVGF (Ping et al., 1996), which is an in vivo methylation protection experiment, is shown in Fig. 4. The data showed that overlapping patterns were generated with the two techniques on the 5% half of Site A. This also indicates that we did not detect the factor that binds to the 3% half of Site A. Importantly, the 5% binding site overlaps the NF1 site. Demonstration that the observed patterns were due to the NF1 site was investigated further using three different in vitro approaches. The first was to determine if a canonical NF1 site but not a mutant NF1 site could compete for factor binding. As shown in Fig. 2, lanes 5
628
S.N. Kumar, J.M. Boss / Molecular Immunology 37 (2000) 623–632
and 6, a consensus NF1 oligo could compete for all binding activity, but the mutant NF1 DNA could not. The second approach investigated whether antibodies to NF1 could alter the migration of the complexes in the EMSA. Indeed, the addition of antibodies reactive to the C-terminal half of NF1 to the DNA-binding reaction resulted in a supershift in both the liver and fibroblast extracts. The supershifted band appears in the wells and can be resolved using a lower crosslinked gel (69:1 acrylamide:bis); however, in such gels the resolution of the bands a and b become diffuse (data not shown). The third in vitro approach compared the patterns derived from a Western blot analysis of the nuclear extracts for NF1 protein with those of a Southwestern blot, in which the same material was analyzed for binding of a Site A DNA probe. In this experiment nuclear extracts prepared from liver, NIH3T3 fibroblasts, and NIH3T3 fibroblasts treated with TNF were analyzed on the same gel. The resulting Western blot showed a predominant band at 32 – 34 kDa in all three extracts, which corresponds to the size reported by Gao et al. (1996) and Adams et al. (1995) for rat NF1 (Fig. 5). Additional bands appeared in the liver extract at higher molecular weights and are either crossreactive proteins or isoforms of NF1. The Southwestern blot showed a single band in all three extracts that aligns precisely with the NF1 bands in the Western blot.
Fig. 6. Chromatin immunoprecipitation assay shows NF1 protein binds to Site A in vivo. A 2.5% agarose gel of the ChIP assay performed to determine if the NF1/CTF protein bound to Site A in the MCP-1 promoter in vivo is shown. NIH3T3 fibroblasts were treated with formaldehyde to crosslink their DNA binding proteins to DNA. Chromatin was isolated and immunoprecipitated as described in Section 2 using either no antibody (lane 3), pre-immune antisera (lane 2) or anti-CTF antisera (lane 5) from Dr Tanese (New York University, NY). Positive and negative controls for the assay included genomic DNA (lane 1) or no DNA (lane 4), respectively. M, molecular weight markers.
3.3. NF1 binds to Site A in 6i6o
Fig. 5. Western and Southwestern analysis of Site A protein show overlapping patterns. SDS-PAGE of nuclear extracts prepared from mouse liver (L) or NIH3T3 fibroblast cultures treated with (F +) or without (F − ) TNF were blotted onto nitrocellulose membranes. The blot was divided in half. One half was probed with antibody to NF1 which identifies a major 32–34 kDa protein in all three extracts. The other half was probed with radiolabeled Site A as described.
To determine if NF1 associated with Site A in vivo, the chromatin immunoprecipitation assay recently described for analyses of histone acetylation and other DNA-specific transcription factor binding was employed (Dedon et al., 1991; Li et al., 1999; Moreno et al., 1999). In this assay, fibroblasts are fixed in formaldehyde, chromatin is isolated, and sheared by sonication. Anti-NF1 antibodies or preimmune antiserum were used to immunoprecipitate the crosslinked protein-DNA complexes. The crosslinks of the precipitated complexes were reversed and the DNA that was associated with the complex was used as a template for PCR. Primers spanning Site A were used to PCR amplify the DNA in the precipitated complexes. A strong amplified product of 121 bp spanning Site A was obtained using antisera to NF1 (Fig. 6). No band was observed in the pre-immune sera lane. These data suggest that NF1 binds to Site A of the MCP-1 gene in vivo.
S.N. Kumar, J.M. Boss / Molecular Immunology 37 (2000) 623–632
3.4. Site A functions as a position and orientation-independent modulator of transcription To begin to understand the role of Site A in mediating MCP-1 transcription, Site A was cloned into an enhancerless CAT reporter construct. This reporter construct has the TATA promoter sequences from the MHC class II HLA-DRA gene as well as a yeast Gal4 DNA binding site but does not contain the MHC class II regulatory elements. Regardless of Site A’s orientation, proximity to the TATA element, or whether there were two copies versus one copy of Site A, no expression was detected in these transient transfection assays, suggesting that Site A was incapable of directing transcription on its own (Fig. 7). To determine if Site A could augment the activity of another DNA binding protein, advantage was taken of the Gal4 DNA binding site in the construction, which could bind Gal4-activator fusion proteins. Transient cotransfections with the Gal4 DNA binding domain alone and the reporter plasmid did not show an increase in activity of the reporter plasmid. Cotransfection of Gal4-VP16, a fusion that contains the transactivation region of the herpes simplex VP16 protein, showed a greater than 50-fold increase in activity over just the DNA binding domain vector (pSG424). Interestingly, the presence of Site A in any of the orientations described above led to a significant 1.6–2.5 fold increase in activity, the highest level occurring when the Gal4 site was flanked by Site A elements. Thus, while Site A did not display
629
enhancer-like activity by functioning independently of other elements, its presence was able to modulate the activity of a strong transactivator, GAL4-VP16. This activity is consistent with its role in the context of the MCP-1 gene, where it is responsible for a 2–3 fold increase in the level of expression from an NF-kB dependent regulatory element.
4. Discussion The focus of this study was to identify the factors binding to Site A of the MCP-1 gene and characterize the general activity of this region in modulating transcription. Previous studies have shown that TNF induction of MCP-1 involves the assembly of a multi-component complex in the distal regulatory region, containing NF-kB p65/p50 and p65/p65 heterodimers and the appearance of a strong DMS-hypersensitive site in the region (Ueda et al., 1994; Ping et al., 1996; Ueda et al., 1997; Ping et al., 1999b). In the current analysis, both in vitro and in vivo evidence has identified the factor binding to the 5% half of Site A as belonging to the NFI/CTF family of transcription factors. Using nuclear extracts prepared from both mouse liver and established fibroblast cell lines, a number of protein/DNA complexes were found. In liver cells the top two complexes produced identical methylation interference patterns, suggesting that the complexes were
Fig. 7. Site A facilitates transcriptional activation but does not activate on it’s own. Transient transfection analysis of chimeric Gal4–Site A constructs were carried out in NIH3T3 fibroblasts. A single Gal4 site and one or more Site A elements were introduced into a minimal pDR-CAT promoter/reporter construct, as indicated in the schematic, to assess the ability of Site A to function as a transcriptional activator. Thirty-six hours after transfection, the cultures were harvested and the level of the CAT protein measured by ELISA. The ELISA OD was then normalized to the expression of a cotransfected luciferase reporter construction. An average of three experiments are shown with the S.E. of the mean indicated.
630
S.N. Kumar, J.M. Boss / Molecular Immunology 37 (2000) 623–632
related. NF1/CTF family proteins are encoded by at least four independent genes (Inoue et al., 1990; Kruse and Sippel, 1994). Depending on the species the names of the genes differ. In the mouse, four isoforms have been identified as NF1-A, NF1-B, NF1-C, and NF1-X (Chaudhry et al., 1997). NF1 isoforms share extensive homology in their N-terminal domains, which is responsible for dimerization and DNA binding (Kruse and Sippel, 1994). The C-terminal regions of these proteins is more divergent and may be responsible for protein-protein interactions and transcriptional modulation. The C-terminal region contains a proline-rich transactivation domain that shares homology with the C-terminal domain (CTD) of the large subunit of RNA polymerase II (Wendler et al., 1994). These isoforms may also be expressed in a tissue specific manner during development (Chaudhry et al., 1997). Thus, the multiple DNA/protein complexes observed may be a result of different isoforms binding to the Site A probes in vitro. Fibroblasts displayed a similar pattern in the EMSA; however, the intensity of each of the protein/ DNA complexes was not as even as the liver cell extract. We interpret this to imply that one of the isoforms is expressed at a higher level in fibroblasts as compared to liver cells. It is interesting to note however, that the Southwestern blot analysis revealed only a single band. This may be due to the relative abundance of one isoform over another and the relatively low sensitivity of the Southwestern blot assay as compared to the EMSA. Depending on the isoform and promoter context, NF1/CTF family members may have different roles with respect to gene expression. In one system using JEG-3 cells, an embryonic choriocarcinoma cell line, the transactivation potential of the four isoforms on the MMTV promoter was NF1-B\NF1-X \ NF1-C\ NF1-A (Chaudhry et al., 1997). Consistent with the above results, another study, comparing the NF1-B and NF1-C splice variants found that NF1-B isoforms had a greater transactivation potential on a large number of minimal promoters than the NF1-C isoforms (Osada et al., 1999). Interestingly, a repression domain in NF1-A consistently showed lower levels of activity when cotransfected with the reporter plasmids (Osada et al., 1999). Similarly, the rat homolog of NF1-A, NF1-L, was found to function in the negative regulation of the rat peripherin gene (Adams et al., 1995). Thus, NF1 isoforms may have distinct functions in regulating gene expression. Given the data that show that deletion of Site A results in a decrease in expression of the MCP-1 gene and the fact that Site A augments the stimulatory ability of the acidic activator in the current system, it is likely that the NF1 family member that binds Site A is one which provides transcriptional activation signals. NF1 proteins have been found to function through a variety of mechanisms, although all of these involve
protein-protein interactions. NF1/CTF proteins can facilitate the assembly of a functional transcriptional initiation complex (Dusserre and Mermod, 1992) and associate with components of basal transcriptional machinery, including TFIIB (Kim and Roeder, 1994), TATA-binding protein (Xiao et al., 1994) and TBP-associated factors (Dusserre and Mermod, 1992). NF1 is also known to interact with other sequence specific transcription factors such as ATF2 (Alonso et al., 1996), thyroid transcription factor (Ortiz et al., 1999) or mediate hormone-induced gene expression by co-operating or competing with nuclear receptors of insulin (Cooke and Lane, 1999), glucocorticoids (Garlatti et al., 1996), androgens (Darne et al., 1998), and progestins (Bruggemeier et al., 1990). We found in our previous studies that Site A exerted a modulatory effect on NF-kB dependent transcription of the MCP-1 gene (Ping et al., 1996). This effect was substantial as the removal of Site A resulted in a two-thirds reduction of activity. The enhancing effect of NF1 on NF-kB-mediated transcription could be due to a direct synergistic effect, such as the stabilization of binding of factors in that region. However, in vitro EMSAs to determine if NF1 and NF-kB could bind cooperatively to the distal regulatory element (Site A/ kB sites) of MCP-1 did not show cooperative binding of the two factors (data not shown). A more intriguing role may be in the recruitment of coactivators to the region. The coactivator CBP, which interacts with many transcription factors including CREB, NF-kB, and nuclear receptors, also can interact with NF1 (Leahy et al., 1999). NF1 binds to CBP at it’s CREBbinding site (Leahy et al., 1999). CBP, which exhibits intrinsic histone acetyltransferase activity, is thought to help remodel chromatin to allow transcription factor access. The binding of NF1 to Site A in a constitutive manner may be related directly to its interactions with CBP. In such a scenario, NF1 and CBP allow the chromatin to remain open. This open structure would be important for the subsequent recruitment of NF-kB in a manner that allows the MCP-1 gene to respond quickly to external stimuli such as TNF. Contacts of NF1 with histone H3 (Alevizopoulos et al., 1995) and nucleosomal cores suggest interaction with chromatin components that may play a role in transcriptional control as in the case of the mouse mammary tumor virus (MMTV) promoter (Beato and Chavez, 1997). It has been shown that NF1/CTF-1 can counteract the repression mediated by histone H1 (Dusserre and Mermod, 1992). Thus, independent of CBP, NF1 binding may promote chromatin accessibility to a region. Alternatively, the binding of NF1 may stabilize nucleosome positioning in a region. Thus, although NF1 binding to its site shows variability that is dependent on nucleosome positioning (Blomquist et al., 1999), once bound it may control repositioning within a region.
S.N. Kumar, J.M. Boss / Molecular Immunology 37 (2000) 623–632
Not all NF1 sites are likely to act the same. As described above, depending on the gene and available isoforms, NF1 can have divergent effects. With regard to TNF induction, we noted an NF1 site in the TNF response region of the manganous superoxide dismutase gene (MnSOD) (Jones et al., 1997). This TNF response region was located in an intron several kb downstream of the start of transcription. The interesting difference between the MCP-1 NF1 site and the MnSOD site is that the MnSOD site was only occupied following TNF induction. Thus, the context of the NF1 site must be extremely important for its function, even though it is of remarkable coincidence that it was also in close proximity to a functional NF-kB motif. Due to its constitutive occupancy, Site A represents a novel region within the MCP-1 flanking DNA. Its role as modulator of transcriptional activators is now established and is most likely to occur through the action of NF1. Its role in chromatin accessibility in this region has yet to be investigated but may shed light on how regulatory elements several kb upstream of the transcriptional start site can communicate and direct transcription of their downstream promoters. Acknowledgements We thank members of the laboratory for critique and suggestions of this work. We also thank Dr D. Ping for his contributions to the beginning of this project. We acknowledge Dr N. Tanese of New York University for kindly providing pre-immune and anti-CTF antisera. We also thank Yvonne DeBellotte for secretarial assistance with preparation of the manuscript. This work was supported by the National Cancer Institute of the NIH CA74271. References Adams, A.D., Choate, D.M., Thompson, M.A., 1995. NF1-L is the DNA-binding component of the protein complex at the peripherin negative regulatory element. J. Biol. Chem. 270, 6975 – 6983 (Published erratum appears in J. Biol. Chem. 18 August 1995 270 (33), 19668). Alevizopoulos, A., Dusserre, Y., Tsai-Pflugfelder, M., von der Weid, T., Wahli, W., Mermod, N., 1995. A proline-rich TGF-b-responsive transcriptional activator interacts with histone H3. Genes Dev. 9, 3051 – 3066. Alonso, C.R., Pesce, C.G., Kornblihtt, A.R., 1996. The CCAATbinding proteins CP1 and NF-1 cooperate with ATF-2 in the transcription of the fibronectin gene. J. Biol. Chem. 271 (36), 22271 – 22279. Baggiolini, M., Dahinden, C.A., 1994. CC chemokines in allergic inflammation. Immunol. Today 15, 127–133. Baggiolini, M., Dewald, B., Moser, B., 1997. Human chemokines: an update. Annu. Rev. Immunol. 15, 675–705. Beato, M., Chavez, S., 1997. Nucleosome-mediated synergism between transcription factors on the mouse mammary tumor virus promoter. Proc. Natl. Acad. Sci. USA 94 (7), 2885–2890.
631
Blomquist, P., Belikov, S., Wrange, O., 1999. Increase nuclear factor 1 binding to its nucleosomal site mediated by sequence-dependent DNA structure. Nucleic Acids Res. 27 (2), 517 – 525. Brach, M.A., Gruss, H-J., Riedel, D., Asano, Y., Vos, S.D., Herrmann, F., 1992. Effect of antiinflammatory agents on synthesis of MCP-1/JE transcripts by human blood monocytes. Mol. Pharmacol. 42, 63 – 68. Bruggemeier, U., Rogge, L., Winnacker, E.L., Beato, M., 1990. Nuclear factor I acts as a transcription factor on the MMTV promoter but competes with steroid hormone receptors for DNA binding. EMBO J. 9 (7), 2233 – 2239. Chaudhry, A.Z., Lyons, G.E., Gronostajski, R.M., 1997. Expression patterns of the four nuclear factor I genes during mouse embryogenesis indicate a potential role in development. Dev. Dyn. 208, 313 – 325. Cochran, B.H., Reffel, A.C., Stiles, C.D., 1983. Molecular cloning of gene sequences regulated by platelet-derived growth factor. Cell 33, 939 – 947. Cooke, D.W., Lane, M.D., 1999. Transcription factor NF1 mediates repression of the GLUT4 promoter by cyclic-amp. Biochem Biophys. Res. Commun. 260, 600 – 604. Darne, C., Veyssiere, G., Jean, C., 1998. Phorbol ester causes ligandindependent activation of the androgen receptor. Eur. J. Biochem. 256, 541 – 549. Dedon, P.C., Soults, J.A., Allis, C.D., Gorovsky, M.A., 1991. A simplified formaldehyde fixation and immunoprecipitation technique for studying protein – DNA interactions. Anal. Biochem. 197, 83 – 90. Dusserre, Y., Mermod, N., 1992. Purified cofactors and histone H1 mediate transcriptional regulation by CTF/NF-1. Mol. Cell. Biol. 12 (11), 5228 – 5237. Frazier-Jessen, M.R., Kovacs, E.J., 1995. Estrogen modulation of JE/monocyte chemoattractant protein-1 mRNA expression in murine macrophages. J. Immunol. 154, 1838 – 1845. Freter, R.R., Irminger, J.-C., Porter, J.A., Jones, S.D., Stiles, C.D., 1992. A novel 7-nucleotide motif located in 3% untranslated sequences of the immediate-early gene set mediates platelet-derived growth factor induction of the JE gene. Mol. Cell. Biol. 12, 5288 – 5300. Gao, B., Jiang, L., Kuno, G., 1996. Transcriptional regulation of a1b adrenergic receptors (a1b AR) by nuclear factor 1 (NF1): a decline in the concentration of NF1 correlates with the downregulation of a1b AR gene expression in regnerating liver. Mol. Cell. Biol. 16 (11), 5997 – 6008. Garlatti, M., Aggerbeck, M., Bouguet, J., Barouki, R., 1996. Contribution of a nuclear factor 1 binding site to the glucocorticoid regulation of the cytosolic aspartate aminotransferase gene promoter. J. Biol. Chem. 271 (51), 32629 – 32634. Glabinski, A.R., Balasingam, V., Tani, M., et al., 1996. Chemokine monocyte chemoattractant protein-1 is expressed by astrocytes after mechanical injury to the brain. J. Immunol. 156, 4363–4368. Gronostajski, R.M., Adhya, S., Nagata, K., Guggenheimer, R.A., Hurwitz, J., 1985. Site-specific DNA binding of nuclear factor I: analyses of cellular binding sites. Mol. Cell. Biol. 5 (5), 964–971. Hanazawa, S., Takeshita, A., Kitano, S., 1994. Retinoic acid suppression of c-fos gene inhibits expression of tumor necrosis factor-ainduced monocyte chemoattractant JE/MCP-1 in clonal osteoblastic MC3T3-E1 cells. J. Biol. Chem. 269, 21379–21384. Hasegawa, S.L., Sloan, J.H., Reith, W., Mach, B., Boss, J.M., 1991. Regulatory factor-X binding to mutant HLA-DRA promoter sequences. Nucleic Acids Res. 19, 1243 – 1249. Howard, O.M.Z., Ben-Baruch, A., Oppenheim, J.J., 1996. Chemokines: progress toward identifying molecular targets for therapeutic agents. Trends Biotechnol. 14, 46 – 51. Inoue, T., Tamura, T., Furuichi, T., Mikoshiba, K., 1990. Isolation of complementary DNAs encoding a cerebellum-enriched nuclear factor I family that activates transcription from the mouse myelin basic protein promoter. J. Biol. Chem. 265, 19065 – 19070.
632
S.N. Kumar, J.M. Boss / Molecular Immunology 37 (2000) 623–632
Jones, K.A., Kadonaga, J.T., Rosenfeld, P.J., Kelly, T.J., Tjian, R., 1987. A cellular DNA-binding protein that activates eukaryotic transcription and DNA replication. Cell 48, 79–89. Jones, P.L., Ping, D., Boss, J.M., 1997. TNF-a and IL-1b regulate the murine manganese superoxide dismutase gene through a complex intronic enhancer. Mol. Cell Biol. 17, 6970–6981. Kawahara, R.S., Deng, Z-W., Deuel, T.F., 1991. Glucocorticoids inhibits the transcriptional induction of JE, a platelet-derived growth factor-inducible gene. J. Biol. Chem. 266, 13261–13266. Kim, T.K., Roeder, R.G., 1994. CTD-like sequences are important for transcriptional activation by the proline-rich activation domain of CTF1. Nucleic Acids Res. 22 (2), 251. Klein, S.A., Dobmeyer, J.M., Dobmeyer, T.S., et al., 1997. Demonstration of the Th1 to Th2 cytokine shift during the course of HIV-1 infection using cytoplasmic cytokine detection on single cell level by flow cytometry. AIDS 11 (9), 1111–1118. Koch, A.E., Kunkel, S.L., Strieter, R.M., 1995. Cytokines in rheumatoid arthritis. J. Invest. Med. 43, 28–38. Kopydlowski, K.M., Salkowski, C.A., Cody, M.J., et al., 1999. Regulation of macrophage chemokine expression by lipopolysaccharide in vitro and in vivo. J. Immunol. 163, 1537–1544. Kruse, U., Sippel, A.E., 1994. The genes for transcription factor nuclear factor I give rise to corresponding splice variants between vertebrate species. J. Mol. Biol. 238, 860–865. Leahy, P., Crawford, D.R., Grossman, G., Gronostajski, R.M., Hanson, R.W., 1999. CREB binding protein coordinates the function of multiple transcription factors including nuclear factor I to regulate phosphoenolpyruvate carboxykinase (GTP) gene transcription. J. Biol. Chem. 274 (13), 8813–8822. Lee, M., Song, H., Park, S., Park, J., 1998. Transcription of the rat p53 gene is mediated by factor binding to two recognition motifs of NF1-like protein. Biol. Chem. 379, 1333–1340. Li, X.Y., Virbasius, A., Zhu, X., Green, M.R., 1999. Enhancement of TBP binding by activators and general transcription factors. Nature 399, 605 – 609. Mantovani, A., Sozzani, S., Bottazzi, B., et al., 1993. Monocyte chemotactic protein-1 (MCP-1): signal transduction and involvement in the regulation of macrophage traffic in normal and neoplastic tissues. Adv. Exp. Med. Biol. 351, 47–54. Moreno, C.S., Beresford, G., Louis-Plence, P., Morris, A.C., Boss, J.M., 1999. CREB regulates MHC class II expression in a CIITAdependent manner. Immunity 10, 143–151. Negus, R.P.M., Turner, L., Burke, F., Balkwill, F.R., 1998. Hypoxia down-regulates MCP-1 expression: implications for macrophage distribution in tumors. J. Leukocyte Biol. 63, 758–765. Nelken, N.A., Coughlin, S.R., Gordon, D., Wilcox, J.N., 1991. Monocyte chemoattractant protein-1 in human artheromatous plaques. J. Clin. Invest. 88, 1121–1127. Ortiz, L., Aza-Blanc, P., Zannini, M., Cato, A.C.B., Santisteban, P., 1999. The interaction between the forehead thyroid transcription factorTTF-2 and the constitutive factor CTF/NF-1 is required for efficient hormonal regulation of the thyroperoxidase gene transcription. J. Biol. Chem. 274 (21), 15213–15221. Osada, S., Matsubara, T., Daimon, S., et al., 1999. Expression, DNA-binding specifity and transcriptional regulation of nuclear factor 1 family proteins from rat. Biochem. J. 342, 189–198. Ping, D., Jones, P.L., Boss, J.M., 1996. TNF regulates the in vivo occupancy of both distal and proximal regulatory regions of the murine monocyte chemoattractant protein-1 (MCP-1/JE) gene. Immunity 4, 455 – 469. Ping, D., Boekhoudt, G., Boss, J.M., 1999. Trans-retinoic acid blocks PDGF-induced expression of the murine monocyte chemoattractant-1 gene by blocking the assembly of a promoter proximal Sp1 binding site. J. Biol. Chem. 274 (in press).
.
Ping, D., Boekhoudt, G.H., Rogers, E.M., Boss, J.M., 1999b. NF-kB p65 mediates the assembly and activation of the tumor necrosis factor responsive element of the murine monocyte chemoattractant-1 gene. J. Immunol. 162, 727 – 734. Ping, D., Boekhoudt, G.H., Zhang, F., et al., 2000. Sp1 binding is critical for promoter assembly and activation of the MCP-1 gene by tumor necrosis factor. J. Biol. Chem. 275, 1708 – 1714. Ransohoff, R.M., Glabinski, A., Tani, M., 1996. Chemokines in immune-mediated inflammation of the central nervous system. Cytokine Growth Factor Rev. 7, 35 – 46. Riley, J.L., Boss, J.M., 1993. Class II MHC transcriptional mutants are defective in higher order complex formation. J. Immunol. 151, 6942 – 6953. Rollins, B.J., Morrison, E.D., Stiles, C.D., 1988. Cloning and expression of JE, a gene inducible by platelet-derived growth factor and whose product has cytokine-like properties. Proc. Natl. Acad. Sci. USA 85, 3738 – 3742. Satriano, J.A., Hora, K., Shan, Z., Stanley, E.R., Mori, T., Schlondroff, D., 1993. Regulation of monocyte chemoattractant protein-1 and macrophage colony-stimulating factor-1 by IFNgamma, tumor necrosis factor-alpha, IgG aggregates, and cAMP in mouse mesangial cells. J. Immunol. 150, 1971 – 1978. Schwartz, C.J., Valente, A.J., Sprague, E.A., Kelley, J.L., Nerem, R.M., 1991. The pathogenesis of atherosclerosis: an overview. Clin. Cardiol. 14, I-1 – I-16. Shapiro, D.J., Sharp, P.A., Wahli, W.W., Keller, M.J., 1988. A high-efficiency HeLa cell nuclear transcription extract. DNA 7, 47 – 55. Shyy, J.Y.-J., Li, Y.-S., Lin, M.-C., et al., 1995. Multiple cis-elements mediate shear stress-induced gene expression. J. Biomech. 28, 1451 – 1457. Taylor, P.C., Peters, A.M., Paleolog, E., et al., 2000. Reduction of chemokine levels and leukocyte traffic to joints by tumor necrosis factor alpha blockade in patients with rheumatoid arthritis. Arthritis Rheum. 43 (1), 38 – 47. Towbin, H., Staehelin, T., Gordon, J., 1979. Electrophoretic transfer of proteins from polyacrylamide gels to nitrocellulose sheets: procedure and some applications. Proc. Natl. Acad. Sci. USA 76, 4350 – 4354. Ueda, A., Okuda, K., Ohno, S., et al., 1994. NF-kB and Sp1 regulate transcription of the human monocyte chemoattractant protein-1 gene. J. Immunol. 153, 2052 – 2063. Ueda, A., Ishigatsubo, Y., Okubo, T., Yoshimura, T., 1997. Transcriptional regulation of the human monocyte chemoattractant protein-1 gene. J. Biol. Chem. 272, 31092 – 31099. Wendler, W., Altmann, H., Ludwig-Winnacker, E., 1994. Transcriptional activation of NFI/CTF1 depends on a sequence motif strongly related to the carboxyterminal domain of RNA polymerase II. Nucleic Acids Res. 22, 2601 – 2603. Xiao, H., Lis, J.T., Greenblatt, J., Friesen, J.D., 1994. The upstream activator CTF/NF1 and RNA polymerase II share a common element involved in transcriptional activation. Nucleic Acids Res. 22 (11), 1966 – 1973. Yla-Herttuala, S., Lipton, B.A., Rosenfeld, M.E., et al., 1991. Expression of monocyte chemoattractant protein 1 in macrophagerich areas of human and rabbit atherosclerotic lesions. Proc. Natl. Acad. Sci. USA 88, 5252 – 5256. Zhou, Z.H., Chartuvedi, L., Han, P., et al., 1998. Interferon-gamma induction of the human monocyte chemoattractant protein (hMCP)-1 gene in astrocytoma cells: functional interaction between a gamma activated site (GAS) and a GC-rich element. J. Immunol. 160, 2908 – 2916.