Molecular Vision 2003; 9:277-287 <http://www.molvis.org/molvis/v9/a39/> Received 2 April 2003 | Accepted 26 June 2003 | Published 30 June 2003

Download Reprint

Interferon-γ regulation of the human mimecan promoter

Elena S. Tasheva, Gary W. Conrad
 
 

Division of Biology, Kansas State University, Manhattan, KS

Correspondence to: Elena S. Tasheva, Division of Biology, Ackert Hall, Kansas State University, Manhattan, KS, 66506-4901; Phone: (785) 532-6553; FAX: (785) 532-6653; email: est@ksu.edu

Abstract

Purpose: The human mimecan/osteoglycin promoter contains multiple interferon-stimulated response elements (ISRE) and interferon-γ-activation sites (GAS). ISRE and GAS motifs are present in a variety of interferon (IFN)-inducible genes. The purpose of this study was to investigate whether IFN-γ affects mimecan gene expression and, if so, to determine the cis-elements and transcription factors that mediate its action.

Methods: Electrophoretic mobility shift assay (EMSA) was used to investigate whether nuclear proteins from IFN-γ-treated cells bind to regions of the human mimecan promoter containing ISRE sites. Incubation of nuclear extracts with specific antibodies was used to identify transcription factors that bind to these sites. Transcriptional activity of the promoter was evaluated by transient transfections of human mimecan promoter/luciferase reporter constructs into corneal keratocytes and non-corneal cells. Co-transfection experiments were used to study the role of transcription factors that bind ISRE elements in the promoter and mediate the IFN-γ response. Expression of mRNAs was analyzed by reverse transcription-polymerase chain reaction.

Results: Using probes that correspond to two ISRE sites located in the first intron of the human mimecan gene, we detected specific DNA-protein complexes with nuclear extracts from IFN-γ-treated cells. Formation of DNA-protein complexes was abrogated by competition with unlabeled probe and one of the complexes was supershifted by the anti-interferon regulatory factor-1 (IRF-1) antibody. Interestingly, when probe that corresponds to a conserved E-box (CACATG) in the proximal promoter and nuclear extracts from IFN-γ-treated cells were used in EMSA, increased binding of upstream stimulatory factor-1 (USF-1) was observed. Co-transfection of a mimecan promoter construct that contained the entire first intron with IRF-1, or with both IRF-1 and USF-1 expression plasmids, suppressed luciferase activity of the promoter in corneal keratocytes and T-47D cells. In contrast, co-transfection experiments with IRF-2, or with both IRF-2 and USF-1, led to increased luciferase activity of the same promoter construct. RT-PCR analyses demonstrate that IFN-γ rapidly and transiently suppresses mimecan expression and induces IRF-1 and IRF-2 mRNAs in bovine corneal keratocytes.

Conclusions: IRF-1 binds to ISRE sites, located in the first intron of the human mimecan gene, and negatively regulates mimecan expression at the level of transcription. Consistent with these observations, an inverse correlation between the expression of mimecan and IRF-1 in bovine corneal keratocytes and non-corneal cells was demonstrated. IRF-2 positively regulates mimecan transcription in corneal keratocytes. Because the intact E-box in the proximal promoter was required for IRF-1 and IRF-2 effects on mimecan transcription, potential direct or indirect interactions between USF-1 and IRF-1 and IRF-2 are likely.

Introduction

The IFNs are a family of secreted glycoproteins that play roles in regulation of antiviral, antiproliferative, and immunomodulatory responses in vertebrate cells. Two types of IFNs have been described, the type I IFNs (α, β, ω, and τ) that are produced by leukocytes and fibroblasts, and the type II IFN (γ) that is produced by T-lymphocytes and natural killer cells in response to infections by viruses and bacteria, as well as in response to noninfectious agents [1,2]. Both types of IFNs exert their biological effects through the signal transduction pathways that initiate by binding of IFNs to their corresponding cell surface receptors; interferon α receptor (IFNAR) for all type I IFNs and interferon γ receptor (IFNGR) for type II IFN. IFN receptors have been detected in almost all mammalian cell types [3,4]. Binding of IFNs with their receptors leads to activation of the Janus kinases (Jaks), which in turn leads to tyrosine phosphorylation and activation of signal transducers and activators of transcription (Stat) proteins [5]. Activated Stat-1 proteins form dimers that translocate to the nucleus and selectively regulate the genes containing the GAS element in their promoters. Heterotrimers of Stat-1, Stat-2 and p48 (known as ISGF3γ and IRF9) also translocate to the nucleus and results in regulation of genes containing ISRE in their promoters [1,2]. The interferon regulatory factors also are induced by these pathways and form a group of secondary transcription factors that bind to ISRE sequences and affect transcription of different sets of genes, depending on the cell type and the nature of cellular stimuli [6]. In addition to the pathways described above, other signaling proteins and transcription factors have been shown to participate in the IFN-induced cellular responses. These include p38 mitogen-activated protein kinase (MAPK), transcription factors CBP/p300, USF1 and NF-κB, shown to act as co-activators, and the class II transactivator (CIITA) that can act as a suppressor of transcription [7-11]. IFNs are known to affect the expression of more than 50 cellular genes. Among IFN-regulatable genes are several members of the proteoglycan gene family, including aggrecan, perlecan and glypican, members of the small leucine-rich proteoglycan gene family, including decorin and biglycan, as well as other matrix proteins such as collagen α2(I) [12-14].

IFNs appear to play important roles in pathogenesis of many ocular diseases as evidenced by the following observations; (i) interferon α receptor 1-deficient mice are highly susceptible to viral infection [15], (ii) ectopic expression of IFN-γ in the lens of transgenic mice affects the growth of the whole eye, resulting in microphthalmia, microphakia, persistent hyperplastic primary vitreous, corneal vascularization, and arrest of retinal differentiation [16], (iii) IFN-γ contributes to corneal allograft rejection, as well as to development of herpetic stromal keratitis [17,18], and (iv) many interferon regulatory factors, including IFN consensus sequence-binding protein (ICSBP), which is thought to be expressed only in cells of macrophage and lymphocyte lineages, are constitutively expressed in the mouse lens [19,20]. Interferon regulatory factors not only affect the expression of IFN-regulated genes, but also play roles in chromatin remodeling, susceptibility to oncogeneic transformation, regulation of the cell cycle, and apoptosis [6]. Therefore, identifying genes that are regulated by IFNs and that are expressed in the eye may be of medical interest because it may suggest potential new therapeutic targets for therapy of diseases causing corneal opacity, corneal transplant rejections, aging and cancer.

Mimecan/osteoglycin is a small leucine-rich proteoglycan of the extracellular matrix that is transcribed and translated several-fold greater in the cornea than in other tissues. Importantly, in human and bovine corneas mimecan carries a long keratan sulfate chain attached to the protein core and thereby is converted to a keratan sulfate proteoglycan, whereas in other connective tissues mimecan is present as a non-sulfated matrix glycoprotein [21]. Similar to other members of the small leucine-rich proteoglycan gene family, mimecan plays a role in collagen fibrillogenesis, as illustrated by in vivo studies using mimecan-deficient mice [22,23]. Similar to other corneal keratan sulfate proteoglycans, mimecan may play roles in the acquisition and maintenance of corneal transparency [24-26]. It is also likely that corneal keratan sulfate proteoglycans contribute to the relative immunotolerance of the cornea [27]. Additional roles for mimecan in cellular growth control are likely, as judged by observations that; (i) the tumor suppressor protein p53 activates transcription of bovine and human mimecan genes [28,29], (ii) mimecan mRNA is absent or found in low levels in the majority of cancer cell lines and tumors [28], (iii) the level of mimecan mRNA is high in corneal keratocytes maintained in low serum or serum-free media, but rapidly decreases if these cells are grown in media containing serum [30], and (iv) growth factors, such as bFGF, modulate mimecan mRNA expression in corneal keratocytes and vascular smooth muscle cells [31,32].

The purpose of the present study was to investigate whether IFN-γ affects mimecan gene expression and, if so, to characterize the cis-elements mediating the IFN-γ action. The rationale for this study stems from; (i) the presence of multiple ISRE (consensus sequence, AGTTTCNNTTTCNY) and GAS (consensus sequence, TTNCNNNAA) binding motifs in the human mimecan promoter (Figure 1), (ii) the presence of similar sequences in the bovine mimecan promoter, and (iii) our previous work demonstrating that nuclear proteins from MG-63 cells bind to a region in the first intron of the human mimecan gene that spans ISRE-c [29]. In this report we show that IFN-γ transiently suppresses mimecan transcription in different cell types. We demonstrate that IRF-1 and IRF-2 are constitutively and inducibly expressed in bovine corneal keratocytes and that in these cells IRF-1 acts as a suppressor of mimecan transcription, whereas IRF-2 has the opposite effect. The intact E-box in the proximal promoter is required for IRF-1 and IRF-2 effects on mimecan transcription. Although interactions between USF-1 and IRF-1 and IRF-2 are not directly demonstrated in this study, potential direct or indirect interactions are likely.

Methods

Plasmids

Expression plasmid pcDNA3.1/hIRF1, containing human IRF1 cDNA driven by the promoter of the immediate-early gene of human cytomegalovirus (CMV), was obtained from Research Genetics (Huntsville, AL, catalog number H-K1000). Expression plasmid pCDM8/h-IRF-2, containing human IRF2 cDNA driven by the CMV promoter, was kindly provided by Dr. T. Taniguchi, Department of Immunology, Tokyo University, Tokyo, Japan. Expression plasmid pCX-USF1, containing USF-1 cDNA driven by the promoter of human cytomegalovirus (CMV), was kindly provided by Dr. Robert Roeder, Laboratory of Biochemistry and Molecular Biology, the Rockefeller University, New York, NY [33]. Expression plasmid A-USF, containing the USF dominant-negative mutant (A-USF) coding sequence, was kindly provided by Dr. Charles Vinson, Laboratory of Biochemistry, National Cancer Institute, Bethesda, MD [34,35]. The human mimecan promoter/luciferase reporter constructs, containing different lengths of the 5'-flanking region of the human mimecan gene including the first intron, have been described previously [29,36].

Cell culture, transfections, luciferase, and β-galactosidase assays

Primary bovine keratocytes were isolated and cultured as described [21,32]. MG-63 cells (human osteosarcoma cell line), T-47D cells (breast ductal carcinoma cell line), and U-937 cells (histiocytic lymphoma cell line) were obtained from the American Type Culture Collection (Manassas, VA). MG-63 cells were maintained in DMEM/F12, supplemented with 10% FBS and antibiotics, T-47D and U-937 cells were maintained in RPMI 1640 supplemented with 10% FBS and antibiotics. Transient transfection experiments were performed using FuGENE-6 transfection reagent (Roche Molecular Biochemicals, Indianapolis, IN) according to the manufacturer's protocol (12 μl reagent per 4 μg DNA). For each well, 2.5 μg of reporter gene and 250 ng of pSVβ-Gal plasmid (Clontech, Palo Alto, CA), to correct for transfection efficiency, were added. For co-transfection experiments, the indicated amounts of pcDNA3.1/hIRF1, pCDM8/h-IRF-2, or pCX-USF1 expression plasmids were used. When two plasmids were co-transfected and compared to data with only one transfected plasmid, an empty vector DNA was co-transfected in equal amounts to keep the transfection mix to the required final amount of DNA and to control for potential competition between transfected promoters for cellular transcription factors. Since all of the co-transfected plasmids used in this study contained the same promoter (the CMV promoter), an empty pcDNA3.1 vector was used in control co-transfection experiments. Transfections with the pGL3 basic plasmid were used for background determination and transfections with the pGL3 control plasmid were used as positive controls. Identical transfections were performed in duplicate and repeated at least three times. Luciferase and β-galactosidase assays were performed as previously described [28]. Transcriptional activity was expressed as relative luciferase activity after normalization for β-galactosidase activity. The results were presented as the means with standard error bars from three separate transfections performed in duplicate.

IFN treatment

Recombinant human IFN-γ was obtained from Roche Molecular Biochemicals (Indianapolis, IN). Recombinant bovine IFN-γ was kindly provided by Dr. T. Kellner, Biocor Animal Health (Omaha,NE). For each experiment, the medium was replaced with medium containing 0.1% FBS and 200 U/ml human IFN-γ, or 50 ng/ml bovine IFN-γ. Cells were incubated for the indicated periods and total RNA and nuclear extracts were prepared from both IFN-treated and untreated cells as controls. IFN-γ treatment of transiently transfected cells was performed as above 24 h after transfection. Luciferase and β-galactosidase activities were assayed 12 h after addition of IFN-γ.

Isolation of RNA and reverse transcription-polymerase chain reaction

Total RNA was isolated from the cells at indicated intervals before and after IFN-γ treatment using the Totally RNA Total RNA Isolation Kit (Ambion, Inc., Austin, TX). RNA (1 μg) was reverse-transcribed using the anchor primer oligonucleotide (dT)18, and Superscript II Reverse Transcriptase (Life Technologies, Inc., Githersburg, MD) for 50 min at 42 °C in buffer supplied by the manufacturer. The single-stranded cDNA products (2 μl) were used as templates in a 50 μl PCR amplification reaction that contained 10 mM TRIS, pH 9, 50 mM KCl, 0.1% Triton X-100 (v/v), 1.5 mM Mg₂Cl, Taq polymerase (5 U), 0.2 mM dNTPs and 100 ng of each GSP, and carried out for 20-24 cycles (95 °C/45 s, 60 °C/1 min, 72 °C/2 min) to remain in the linear phase of the logarithmic growth during PCR amplification. Resulting PCR products were analyzed by electrophoresis on 2% agarose gels, with DNA visualized by ethidium bromide staining. Polaroid photographs of gels were scanned and the intensity of the bands was estimated by densitometry using the Scion Image 1.60 program. The human mimecan cDNA (AF112465) was amplified using the following primers; Hm+2, 5'-tct cat tca ccc tcc cac ttg g-3' and Hm-1265, 5'-taa tgc gtg agt cct gct ggg ttg g-3'. For amplification of the glyceraldehyde-3-phosphate dehydrogenase (GAPDH), human control amplifier set from Clontech (Palo Alto,CA) was used (catalog number 5406-1). The human IRF-1 cDNA (NM_002198) was amplified using the following primers; hIRF1+187; 5'-cag agc caa aca tgc cca tca ctc gg-3' and hIRF1-1180; 5'-tgc tac ggt gca cag gga atg g-3'. The human IRF-2 cDNA (NM_002199) was amplified using the following primers; hIRF2+290; 5'-gat gca tgc ggc tag aca tgg gtg-3' and hIRF2-837; 5'-ggt aga gct cgc tca tgc tga ccg g-3'. The bovine mimecan cDNA (M37974) was amplified using the following primers; Bm+411; 5'-ttt agg atc cat taa aat gaa gac tct gca atc tac act tct cct g-3' and Bm-1120; 5'-gat agc tcg aga atg tat gac cct ata gg-3', designed to span the entire coding region of the gene and to contain BamH I and Xho I sites at their ends to facilitate the cloning and sequencing of amplified PCR fragments. The bovine IRF-1 cDNA (AJ490935) was amplified using the following primers; bovIRF1+11; 5'-aga tta att cca acc aaa tcc cag g-3', and bovIRF1-432; 5'-gga gct act gag tcc atc aga gaa gg-3. The bovine IRF-2 cDNA (AJ490936) was amplified using the following primers; bovIRF2 +301; 5'-tga gtc acc ttt ggg gct tag taa cgg-3', and bovIRF2-730; 5'-gtt gga cgt gac aaa ggt ggc cat gca gg-3'. The bovine GAPDH cDNA (U85042) was amplified using the following primers; Bgapdh+451; 5'-gtc atc cat gac cac ttt ggc atc gtg g-3' and Bgapdh-800; 5'-ttg aag tcg cag gag aca acc tgg-3'. All primers were synthezised by Integrated DNA Technologies Inc. (Coralville, IA).

Preparation of nuclear extracts and electrophoretic mobility shift assay

Nuclear extracts were prepared from MG-63 cells before and after treatment with IFN-γ using NE-PER, Nuclear and Cytoplasmic Extraction Reagents, from Pierce (Rockford, IL), according to the manufacturer's protocol. Nuclear extracts from U-937 and T-47D cells were obtained from Geneka Biotechnology Inc. (Montreal, Canada). Oligonucleotides for EMSA were synthesized by Integrated DNA Technologies,Inc., and annealed to generate double-stranded DNA probes. The following oligonucleotides were used in this study (sequence of the sense strand shown); ISRE-a, 5'-cac agt ttc aca ttc tgc a-3'; ISRE-b, 5'-ctc tgg ttt aag ttt cag tta cct c-3'; ISRE-c, 5'-act att ctt tca gtt ttc tta gct t-3'; E-box, 5'-ctg gca aag atc tct acc aca tgg aaa ctg-3'. Double-stranded DNA probes were 5' end-labeled with [γ³²P] ATP and T4 polynucleotide kinase. Nuclear extract (5 μg) was incubated with 25,000 cpm of labeled probe in a 20 μl volume of reaction buffer containing 50 mM NaCl, 10 mM Tris-HCl (pH 7.5), 1 mM MgCl₂, 0.5 mM EDTA, 0.5 mM DTT, 10% (v/v) glycerol and 2 μg poly(dI-dC)-poly(dI-dC, Amersham Biosciences, Piscataway, NJ). Binding reactions were performed at room temperature for 25 min and resolved on a 4% nondenaturating polyacrylamide gel in 50 mM Tris-HCl, 0.38 M glycine, 2 mM EDTA, pH 8.5 for 2 h at 100 V. The gels were dried and autoradiographed. For the supershift assay, 1 μg of each antibody directed against USF-1 (sc-229), c-Myc (sc-232), IRF-1 (sc-497), IRF-2 (sc-498) or Statp84/p91 (sc-346) was added 20 min after the nuclear extract had been incubated with the probe. This mixture then was incubated for an additional 45 min at 4 °C. All antibodies were obtained from Santa Cruz Biotechnology, Santa Cruz, CA.

Results

Mimecan, IRF-1, and IRF-2 mRNA levels are inversely correlated following IFN-γ treatment

To evaluate whether IFN-γ treatment would affect mimecan expression, total RNA was isolated from cultured primary bovine corneal keratocytes, MG-63, U-937 and T-47D cells before and after treatment with IFN-γ. MG-63, U-937 and T-47D cells were included in these experiments because they are among the few cancer cell lines that express mimecan (high levels in MG-63 and detectable levels in U-937 and T-47D cells) and because they respond to a variety of factors, including IFN-γ [37-39]. mRNA pools were analyzed by reverse transcription-polymerase chain reaction (RT-PCR). The amplifications were carried out for 20-24 cycles to remain in the linear phase of the logarithmic growth during PCR, thereby allowing semiquantitative assessment of the mRNA pools. In addition to mimecan, the levels of IRF-1, IRF-2, and GAPDH mRNAs also were evaluated. As shown in Figure 2A, the level of mimecan mRNA was decreased within 3 h of exposure to IFN-γ. This effect was reversible and at 24 h of exposure to IFN-γ the level of mimecan mRNA was restored and slightly increased. An inverse correlation was observed between the expressions of mimecan and those of IRF-1 and IRF-2 in IFN-γ-treated cells. In addition, the results (Figure 2A, 0 h) demonstrated that IRF-1 and IRF-2 are constitutively expressed in corneal keratocytes. As expected, transient induction of IRF-1 and IRF-2 mRNAs was observed after 3 h of treatment with IFN-γ with IRF-1 displaying an approximately 2.5-fold induction and IRF-2 increasing approximately 2-fold.

Similar results were obtained with MG-63, U-937 and T-47D cells (Figure 2B). The level of mimecan mRNA was decreased in these three cell types 3 h after IFN-γ treatment. Under the same conditions, the levels of IRF-1 and IRF-2 mRNA were increased in all three cell lines, thus confirming the lack of cytotoxic effect of IFN-γ in our assay. These results demonstrate that IFN-γ suppresses mimecan gene expression in the cell types described above. The results also demonstrate that mimecan, IRF-1, and IRF-2 mRNA levels are inversely correlated following IFN-γ treatment.

IRF-1 binds ISRE-b and ISRE-c in the human mimecan promoter

Using DNase I footprinting and EMSA, we previously demonstrated that nuclear proteins from untreated MG-63, but not from U-937 cells, as well as in vitro synthesized IRF-1, bind to a segment of the first intron of human mimecan gene that contains ISRE-c (Figure 3) [29]. In the present study, we examined whether IFN-γ could induce the binding of nuclear proteins to ISRE-a, -b, -c, and GAS sequences in the human mimecan gene. Nuclear extracts were prepared from MG-63, U-937 and T-47D cells treated with 200 U/ml IFN-γ for 2 or 4 h and EMSA was performed with labeled oligonucleotides spanning the sequences above. Nuclear extracts from these three cell lines formed complex bands with the labeled ISRE-a probe (Figure 3A, lanes 1, 6, and 11). However, addition of non-specific competitor poly (dI-dC)-poly (dI-dC) reduced the density of the complex bands in all three samples (Figure 3A, lanes 2, 7, and 12), thus demonstrating that binding of nuclear proteins to this ISRE is non-specific. Addition of anti-IRF-1, anti-IRF-2 and anti-Stat-1 antibodies did not lead to a supershift of the complexes (Figure 3A, lanes 3-5, 8-10, and 13-15) indicating that these complexes did not contain IRF-1, IRF-2 or Stat-1 proteins. DNase I footprinting using a probe that spans the GAS-b and -c elements and nuclear extracts from untreated and IFN-γ-treated cells did not show protected segments within this region (data not shown). In contrast, when EMSA was performed with labeled ISRE-b probe, specific complex bands were formed in all three nuclear extract samples (Figure 3B, top, lanes 1, 6, and 11). The addition of 50 times molar excess of the non-labeled ISRE-b probe reduced the density of the complex bands, thus demonstrating specificity of the binding proteins induced in these cells by IFN-γ (Figure 3B, top, lanes 2, 7, and 12). Addition of anti-IRF-1, anti IRF-2, and anti-Stat-1 antibodies did not lead to supershift of the compexes formed with MG-63 and U-937 extracts (Figure 3B, top, lanes 3-5 and 8-10). However, addition of anti-IRF-1 antibody led to supershift of one of the complexes formed with T-47D extract (Figure 3B, top, lane 11, the complex marked with a red arrow contains IRF-1; lane 13, supershifted complex is marked with a green arrow). These results demonstrate that the binding proteins induced in T-47D cells by IFN-γ contained IRF-1. Similar results were obtained with the ISRE-c probe. Nuclear extracts from U-937 and T-47D cells treated with IFN-γ formed specific complexes with a labeled ISRE-c probe (Figure 3B, bottom, lane 1 versus 6 and lane 11 versus 16). Addition of anti-IRF-1, -2, and anti-Stat-1 antibodies did not supershift the specific complex formed in U-937 extract (Figure 3B, bottom, lanes 8-10, red arrow). However, similar to EMSA results with ISRE-b probe, addition of anti-IRF-1 antibody led to supershift of one of the complexes formed with T-47D extract (Figure 3B, bottom, compare lanes 16 and 18). These results demonstrate that the binding proteins induced in T-47D cells by IFN-γ interact with ISRE-b and ISRE-c sequences and that at least one of the interacting proteins is IRF-1.

Transcription factors from IFN-γ-treated cells bind the E-box in the human mimecan promoter

We demonstrated previously that the E-box in the proximal promoter is important for transcription and is an element required for transcriptional activation of the human mimecan gene by p53 [36]. The E-box (CACRTG) is a DNA binding site for several transcription factors including the basic-helix-loop-helix-leucine zipper transcription factors, USF-1, USF-2, and their differentially spliced isoforms, the Myc/Max/Mad family of transcription factors, the Mitf (microphthalmia associated transcription factor), and Myo-D, myogenin [40-44]. The USF-1 was identified as a protein that binds to the E-box and supports transcription of mimecan in MG-63 cells [29]. In this study we sought to examine whether treatment of MG-63, U-937 and T-47D cells with IFN-γ could influence the binding of nuclear proteins to this E-box in the human mimecan promoter. Gel-shift analysis was performed using a double-stranded probe spanning the E-box motif (Figure 4). Consistent with our previous observations, nuclear extract from untreated MG-63 cells formed a complex band with labeled probe and this complex was supershifted by addition of anti-USF-1 antibody, but not by addition of anti-c-myc antibody (Figure 4, lanes 1-3). Nuclear extract from MG-63 cells, treated with IFN-γ for 2 h, formed a complex band with much higher density as compared to the band from untreated MG-63 cells. This complex band was competed by unlabeled probe, but could not be supershifted by addition of anti-USF-1 or anti-c-myc antibodies (Figure 4, lanes 4-8). Treatment with IFN-γ for 4 h led to reduced density of the complex (Figure 4, compare lanes 7 and 8 with lanes 9 and 10). These results demonstrate that a different complex is formed in IFN-γ-treated MG-63 cells. One possible explanation for these results is that additional proteins that bind to USF-1 may be contained in the dense band and may block access of USF-1 antibody to its target. Nuclear extracts from U-937 cells, treated with IFN-γ for 2 h, also formed a complex band that was competed by unlabeled probe but not by non-specific competitor poly (dI-dC)-poly (dI-dC, Figure 4, lanes 14-16). In contrast to results with MG-63, the complex formed with U-937 extract could be supershifted by addition of anti-USF-1 (Figure 4, lane 17). In addition, the density of supershifted complex was stronger in IFN-γ-treated cells (Figure 4, compare lanes 12 and 17, blue arrows) indicating that IFN-γ-treatment of U-937 cells led to increased binding of USF-1 to the E-box. Similar experiments performed with untreated and IFN-γ-treated T-47D cells demonstrated that IFN-γ also induced the binding of nuclear proteins to E-box and that one of the complexes induced by IFN-γ contained USF-1 in its molecule (Figure 4, lanes 19-25).

Together, the results presented so far demonstrate that; (i) nuclear proteins from all three cell lines bind to ISRE-b and ISRE-c elements, both of which are located in the first intron of the human mimecan gene, (ii) the proteins induced in T-47D cells by IFN-γ that bind to ISRE-b and ISRE-c sequences contain IRF-1, (iii) the proteins induced in U-937 and T-47D cells by IFN-γ that bind to the E-box contain USF-1, and (iv) treatment with IFN-γ appears to abolish binding of USF-1 to the E-box in MG-63 cells.

IRF-1 suppresses mimecan transcription in corneal keratocytes

The transcriptional repression of mimecan by IFN-γ was further investigated by transient cell transfection assay utilizing mimecan promoter/luciferase reporter gene constructs. Varying lengths of the mimecan promoter region linked to the luciferase gene were transfected into bovine corneal keratocytes and T-47D cells (Figure 5). The p-3694/-1146 construct contained 2.5 kb of the human mimecan promoter, including GAS-a, GAS-b, GAS-c, and ISRE-a sequences, as well as the E-box and the three Initiator elements located at the 5'-end of the first exon (Figure 1). The p-2316/-1146 construct contained 1.1 kb of the human mimecan promoter in which upstream GAS and ISRE sequences were deleted. The p-1314/-1146 construct was the minimal promoter region (168 bp) that was shown before to be functional, and contained only the E-box and the three Initiator elements [36]. The p-1341/-153 construct contained the entire first intron, in addition to the E-box and the three Initiator elements. Following transfection (24 h), to allow reporter expression, the cells were incubated in the presence or absence of IFN-γ for 12 h. In several independent experiments, only the p-1314/-153 construct showed transcriptional repression in the presence of IFN-γ in both cell types (Figure 5). These data are consistent with the results from EMSA (Figure 2) and demonstrate that the first intron of the human mimecan gene contains IFN-γ responsive elements.

To further investigate the potential role of IRF-1 as a regulator of mimecan expression, we cotransfected the IRF-1 expression plasmid together with p-1314/-1146 or p-1314/-153 into corneal keratocytes and T-47D cells. As shown in Figure 6, IRF-1 expression markedly diminished luciferase activity of the p-1314/-153 construct, but had little effect on luciferase activity of the p-1314/-1146 construct (in which the first intron is missing). The negative effect of IRF-1 on mimecan transcription was dose-dependent, since increased amounts of IRF-1 led to increased suppression of mimecan transcription. Interestingly, cotransfections with the IRF-2 expression plasmid had an opposite effect on the mimecan promoter and led to increased luciferase activity of the p-1314/-153 construct, although the results from EMSA did not show binding of IRF-2 to intronic ISRE-b and ISRE-c sequences. These results can be explained in two ways. First, IRF-2 also may bind to intronic ISRE sequences, but this binding was not detected in our system. Second, binding of IRF-2 to ISRE sequences may not be required for its effect on mimecan transcription. In support of the second conclusion, the luciferase activity of p-1314/-1146 cotransfected with IRF-2 was also increased, although not as dramatically as with the p-1314/-153 construct. Consistent with our previous observations, co-transfection of USF-1 led to activation of the human mimecan promoter [36]. To test for potential functional interactions between IRF-1 and IRF-2 with USF-1, cotransfections of human mimecan promoter constructs, together with expression vectors encoding the USF-1, or the dominant-negative mutant of USF, A-USF, were performed. In the A-USF, the basic HLH domain of USF-1 is replaced with an acidic sequence, which stabilizes the heterodimers between A-USF and USF, resulting in inhibition of USF DNA binding [34,35]. As shown in Figure 6, cotransfections of p-1314/-153 with both USF-1 (50 ng) and IRF-1 (1 μg) led to diminished luciferase activity of this promoter construct. Compared to luciferase activity of p-1314/-153 cotransfected with IRF-1 alone, these results indicate that USF-1 and IRF-1 may have opposite effects on mimecan transcription. Cotransfections of p-1314/-153 with both USF-1 (50 ng) and IRF-2 (1 μg) led to slightly increased luciferase activity, when compared to that of p-1314/-153 cotransfected only with USF-1, and decreased luciferase activity, when compared to that of p-1314/-153 cotransfected with IRF-2 (1 μg). Because co-transfections of USF-1 alone or IRF-2 alone led to increased promoter activities, these results can be explained with the involvement of an additional protein(s) that may bind both, USF-1 and IRF-2, and that may cause transcriptional repression of the mimecan promoter. Co-transfections of A-USF with either IRF-1 or IRF-2 abolished the effects of these factors when co-transfected alone, thereby supporting our conclusion that interactions between USF-1 and the two IRFs are likely.

To further confirm the roles of IRF-1 and IRF-2 on transcriptional regulation of human mimecan, bovine corneal keratocytes were transiently transfected with IRF-1 or IRF-2 expression plasmids. Transient introduction of IRF-1 in these cells, which had not been treated with IFN-γ, caused a moderate decrease in mimecan transcripts, whereas transient introduction of IRF-2 in these cells had the opposite effect (Figure 7). The modest effect is likely a result of the fact that transient transfection can introduce the transfected gene in only a portion of the cells and that transfection efficiency of primary cells is known to be low.

Discussion

In this study we investigated the expression of mimecan in bovine corneal keratocytes and other cells types in response to IFN-γ. We demonstrated that IRF-1 and IRF-2 are constitutively expressed in bovine corneal keratocytes and that after IFN-γ treatment the levels of both IRF-1 and IRF-2 mRNAs are upregulated by IFN-γ, with greater increases seen in the levels of IRF-1 mRNA. Notably, the levels of IRF-1 and IRF-2 mRNAs were inversely correlated with those of mimecan mRNA in all cell types used in this study.

Herein, we identify the cis-acting elements (two ISRE sites located in the first intron of the human mimecan gene) and transcription factors involved in this response. Using EMSA, with oligonucleotides spanning intronic ISRE-b and ISRE-c sequences, we identify specific DNA-protein complexes formed with nuclear extracts from IFN-γ-treated cell lines. As indicated by supershift experiments, IRF-1 was contained in one of the complexes formed with nuclear extract from T-47D cells, but not with nuclear extract from U-937 cells. Previously, we demonstrated that in MG-63 cells IRF-1 also binds to ISRE-c [29]. Therefore, at least in two cell types, IRF-1 binds to ISRE-c in the human mimecan promoter. The results obtained with nuclear extract from IFN-γ-treated U-937 cells can be explained either that proteins other than IRF-1 directly bind to ISRE-b and ISRE-c in the human mimecan promoter, or that additional proteins that bind IRF-1 are contained in the complexes so that they block access of anti-IRF-1, anti-IRF-2 and anti-Stat antibodies to their targets. Supportive of these explanations are the results from EMSA with T-47D extracts, showing that multiple DNA-protein complexes were formed.

Our results also indicate involvement of the E-box in IFN-γ-induced suppression of human mimecan trancription. By EMSA, binding of USF-1 to E-box oligonucleotide was detected in untreated MG-63, T-47D and U-937 cells. However, increased binding of USF-1 after IFN-γ treatment was detected in two of these cells types, T-47D and U-937. In contrast, IFN-γ treatment of MG-63 cells appeared to abolish the binding of USF-1 to its target. Instead, increased density of the complex formed with E-box oligonucleotide in these cells was observed. Antibody against c-myc, one of many other E-box-binding proteins, also had no effect on the DNA-protein complex in MG-63 cells. Although in this study we did not attempt to identify all known E-box-binding proteins, one explanation for these results could be that proteins other than USF-1 bind to the E-box in IFN-γ-treated MG-63 cells. Alternatively, proteins that bind to USF-1 may block access of anti-USF-1 antibody to its target. Involvement of the E box and intronic ISRE sites in transcriptional regulation of human mimecan by IFN-γ was further corroborated by transient transfection experiments. The expression of three constructs that contained only the 5'-upstream region of the human mimecan promoter was not affected by IFN-γ, demonstrating that the ISRE-a and GAS-a, -b, and -c elements play little role in mediating mimecan response to IFN-γ. In contrast, expression of the construct that contained the E-box plus the entire first intron was suppressed by IFN-γ in two cell types, corneal keratocytes and T-47D cells, demonstrating roles for the intronic ISRE and E-box elements. The effect of USF-1 and IRFs on mimecan gene transcription was evaluated by overexpression of each one of these transcription factors alone or in combination. We show here that overexpression of IRF-1 suppresses, but overexpression of IRF-2 activates, mimecan transcription. Importantly, cotransfection of A-USF abrogates the effects of IRF-1 and IRF-2. Based on these data, it can be concluded that mimecan transcription in response to IFN-γ is controlled, in part, by the presence and concentrations of USF-1 and IRFs and that functional interactions between these transcription factors appears to take place in the human mimecan promoter. In this view it is of interest to note that IRF-1 protein has been demonstrated to be very unstable (half-life about 30 min), whereas IRF-2 protein has been demonstrated to be stable (half-life about 8 h) [6]. The data presented in this study are consistent with these observations; the initial down-regulation of mimecan mRNA after treatment with IFN-gamma, mediated by IRF-1, is followed by up-regulation of mimecan mRNA, mediated by IRF-2. In addition, binding of IRF-2 to intronic ISRE sites may not be required for its effect on mimecan expression. USFs are ubiquitously expressed proteins known to interact with DNA as dimers, with USF-1-USF-2 heterodimer being the major one present in most cell types, less abundant USF-1 homodimers and very little USF-2 homodimers [45]. Interactions between USF-1 and other proteins have been described, including interactions between USF-1, IRF-1 and Stat1α in the type IV class II transactivator (CIITA) promoter in astrocytes, although the positions of E-box, ISRE and GAS regulatory elements in the CIITA promoter are different than those in the mimecan promoter [46,47]. Similar to mimecan, binding of IRF-1 and IRF-2 to an ISRE located in the first intron of the c-myb gene has been reported [48]. Furthermore, similar to mimecan, IRF-1 has been demonstrated to suppress transcription of the c-myb gene. Although in this report we do not address the exact molecular mechanism by which the described transcription factors may regulate mimecan transcription in response to IFN-γ, one possible explanation for the results presented in this study could be that binding of IRF-1 to intronic ISRE sites may cause pausing of RNA polymerase II and temporarily block transcription elongation of mimecan mRNA. As an attempt to overcome such a potential block in transcription, cells may respond with increased binding of USF-1 (and other transcription factors) to the E-box in the human mimecan promoter. A similar mode of transcriptional regulation has been shown to control expression of other genes, including c-myb, c-myc, c-fos, and adenosine deaminase [48-51]. Recently, recruitment of multiple interferon regulatory factors and histone acetyltransferase to the transcriptionally active IFN-γ promoters has been demonstrated [52]. Chromatin remodeling at the CIITA locus by IFN-γ also has been reported [53]. Although not directly addressed in this study, formation of chromatin-remodeling complexes at the mimecan promoter also may account for transcriptional regulation of this gene by IFN-γ. The roles of IRF-1 and IRF-2 on transcriptional regulation of human mimecan were further confirmed by transient transfections of bovine corneal keratocytes with IRF-1 or IRF-2 expression plasmids.

In conclusion, our study has several novel findings. First, it shows that mimecan expression in the cornea is suppressed by IFN-γ. Given the important role of corneal keratan sulfate proteoglycans in the maintenance of corneal transparency, as well as increased recognition of extracellular matrix proteins as important elements in lymphocyte (cells producing IFN-γ) positioning and effector functions in alloreactive responses, the results from these studies provide new information and may suggest new molecules as potential targets for immune intervention in corneal and organ transplantation. Second, our study identifies cis-regulatory elements and transcription factors that mediate the suppressive effect of IFN-γ on the human mimecan promoter in corneal keratocytes and other cells types. Third, this study shows that IRF-1 and IRF-2 are constitutively expressed in corneal keratocytes and therefore may have other, non-immunity-related functions in the cornea.

Acknowledgements

The authors thank Dr. R. Roeder for providing the human USF-1 expression plasmid, Dr. C. Vinson for providing the A-USF plasmid, Dr. Dr. T. Taniguchi for providing the IRF-2 expression plasmid, Dr. T. Kellner for providing recombinant bovine IFN-γ, and members of Dr. Terry Johnson's laboratory for the regular supply of bovine corneas. This work was supported by NIH Grant EY13395 to GWC and EST.

References

1. Stark GR, Kerr IM, Williams BR, Silverman RH, Schreiber RD. How cells respond to interferons. Annu Rev Biochem 1998; 67:227-64.

2. Taniguchi T, Takaoka A. The interferon-alpha/beta system in antiviral responses: a multimodal machinery of gene regulation by the IRF family of transcription factors. Curr Opin Immunol 2002; 14:111-6.

3. Darnell JE Jr, Kerr IM, Stark GR. Jak-STAT pathways and transcriptional activation in response to IFNs and other extracellular signaling proteins. Science 1994; 264:1415-21.

4. Schindler C, Darnell JE Jr. Transcriptional responses to polypeptide ligands: the JAK-STAT pathway. Annu Rev Biochem 1995; 64:621-51.

5. Darnell JE Jr. STATs and gene regulation. Science 1997; 277:1630-5.

6. Taniguchi T, Ogasawara K, Takaoka A, Tanaka N. IRF family of transcription factors as regulators of host defense. Annu Rev Immunol 2001; 19:623-55.

7. Goh KC, Haque SJ, Williams BR. p38 MAP kinase is required for STAT1 serine phosphorylation and transcriptional activation induced by interferons. EMBO J 1999; 18:5601-8.

8. Bhattacharya S, Eckner R, Grossman S, Oldread E, Arany Z, D'Andrea A, Livingston DM. Cooperation of Stat2 and p300/CBP in signalling induced by interferon-alpha. Nature 1996; 383:344-7.

9. Muhlethaler-Mottet A, Di Berardino W, Otten LA, Mach B. Activation of the MHC class II transactivator CIITA by interferon-gamma requires cooperative interaction between Stat1 and USF-1. Immunity 1998; 8:157-66.

10. Neish AS, Read MA, Thanos D, Pine R, Maniatis T, Collins T. Endothelial interferon regulatory factor 1 cooperates with NF-kappa B as a transcriptional activator of vascular cell adhesion molecule 1. Mol Cell Biol 1995; 15:2558-69.

11. Zhu XS, Ting JP. A 36-amino-acid region of CIITA is an effective inhibitor of CBP: novel mechanism of gamma interferon-mediated suppression of collagen alpha(2)(I) and other promoters. Mol Cell Biol 2001; 21:7078-88.

12. Sharma B, Iozzo RV. Transcriptional silencing of perlecan gene expression by interferon-gamma. J Biol Chem 1998; 273:4642-6.

13. Dodge GR, Diaz A, Sanz-Rodriguez C, Reginato AM, Jimenez SA. Effects of interferon-gamma and tumor necrosis factor alpha on the expression of the genes encoding aggrecan, biglycan, and decorin core proteins in cultured human chondrocytes. Arthritis Rheum 1998; 41:274-83.

14. Higashi K, Kouba DJ, Song YJ, Uitto J, Mauviel A. A proximal element within the human alpha 2(I) collagen (COL1A2) promoter, distinct from the tumor necrosis factor-alpha response element, mediates transcriptional repression by interferon-gamma. Matrix Biol 1998; 16:447-56.

15. Hwang SY, Hertzog PJ, Holland KA, Sumarsono SH, Tymms MJ, Hamilton JA, Whitty G, Bertoncello I, Kola I. A null mutation in the gene encoding a type I interferon receptor component eliminates antiproliferative and antiviral responses to interferons alpha and beta and alters macrophage responses. Proc Natl Acad Sci U S A 1995; 92:11284-8.

16. Egwuagu CE, Sztein J, Chan CC, Reid W, Mahdi R, Nussenblatt RB, Chepelinsky AB. Ectopic expression of gamma interferon in the eyes of transgenic mice induces ocular pathology and MHC class II gene expression. Invest Ophthalmol Vis Sci 1994; 35:332-41.

17. Babu JS, Kanangat S, Rouse BT. T cell cytokine mRNA expression during the course of the immunopathologic ocular disease herpetic stromal keratitis. J Immunol 1995; 154:4822-9.

18. King WJ, Comer RM, Hudde T, Larkin DF, George AJ. Cytokine and chemokine expression kinetics after corneal transplantation. Transplantation 2000; 70:1225-33.

19. Li W, Nagineni CN, Efiok B, Chepelinsky AB, Egwuagu CE. Interferon regulatory transcription factors are constitutively expressed and spatially regulated in the mouse lens. Dev Biol 1999; 210:44-55.

20. Li W, Nagineni CN, Ge H, Efiok B, Chepelinsky AB, Egwuagu CE. Interferon consensus sequence-binding protein is constitutively expressed and differentially regulated in the ocular lens. J Biol Chem 1999; 274:9686-91.

21. Funderburgh JL, Corpuz LM, Roth MR, Funderburgh ML, Tasheva ES, Conrad GW. Mimecan, the 25-kDa corneal keratan sulfate proteoglycan, is a product of the gene producing osteoglycin. J Biol Chem 1997; 272:28089-95.

22. Tasheva ES, Koester A, Paulsen AQ, Garrett AS, Boyle DL, Davidson HJ, Song M, Fox N, Conrad GW. Mimecan/osteoglycin-deficient mice have collagen fibril abnormalities. Mol Vis 2002; 8:407-15 <http://www.molvis.org/molvis/v8/a48/>.

23. Ameye L, Young MF. Mice deficient in small leucine-rich proteoglycans: novel in vivo models for osteoporosis, osteoarthritis, Ehlers-Danlos syndrome, muscular dystrophy, and corneal diseases. Glycobiology 2002; 12:107R-16R.

24. Klintworth GK, Smith CF. Abnormalities of proteoglycans and glycoproteins synthesized by corneal organ cultures derived from patients with macular corneal dystrophy. Lab Invest 1983; 48:603-12.

25. Hassell JR, Cintron C, Kublin C, Newsome DA. Proteoglycan changes during restoration of transparency in corneal scars. Arch Biochem Biophys 1983; 222:362-9.

26. Plaas AH, West LA, Thonar EJ, Karcioglu ZA, Smith CJ, Klintworth GK, Hascall VC. Altered fine structures of corneal and skeletal keratan sulfate and chondroitin/dermatan sulfate in macular corneal dystrophy. J Biol Chem 2001; 276:39788-96.

27. Funderburgh JL, Mitschler RR, Funderburgh ML, Roth MR, Chapes SK, Conrad GW. Macrophage receptors for lumican. A corneal keratan sulfate proteoglycan. Invest Ophthalmol Vis Sci 1997; 38:1159-67.

28. Tasheva ES, Maki CG, Conrad AH, Conrad GW. Transcriptional activation of bovine mimecan by p53 through an intronic DNA-binding site. Biochim Biophys Acta 2001; 1517:333-8.

29. Tasheva ES. Analysis of the promoter region of human mimecan gene. Biochim Biophys Acta 2002; 1575:123-9.

30. Beales MP, Funderburgh JL, Jester JV, Hassell JR. Proteoglycan synthesis by bovine keratocytes and corneal fibroblasts: maintenance of the keratocyte phenotype in culture. Invest Ophthalmol Vis Sci 1999; 40:1658-63.

31. Shanahan CM, Cary NR, Osbourn JK, Weissberg PL. Identification of osteoglycin as a component of the vascular matrix. Differential expression by vascular smooth muscle cells during neointima formation and in atherosclerotic plaques. Arterioscler Thromb Vasc Biol 1997; 17:2437-47.

32. Tasheva ES, Funderburgh ML, McReynolds J, Funderburgh JL, Conrad GW. The bovine mimecan gene. Molecular cloning and characterization of two major RNA transcripts generated by alternative use of two splice acceptor sites in the third exon. J Biol Chem 1999; 274:18693-701.

33. Du H, Roy AL, Roeder RG. Human transcription factor USF stimulates transcription through the initiator elements of the HIV-1 and the Ad-ML promoters. EMBO J 1993; 12:501-11.

34. Krylov D, Kasai K, Echlin DR, Taparowsky EJ, Arnheiter H, Vinson C. A general method to design dominant negatives to B-HLHZip proteins that abolish DNA binding. Proc Natl Acad Sci U S A 1997; 94:12274-9.

35. Qyang Y, Luo X, Lu T, Ismail PM, Krylov D, Vinson C, Sawadogo M. Cell-type-dependent activity of the ubiquitous transcription factor USF in cellular proliferation and transcriptional activation. Mol Cell Biol 1999; 19:1508-17.

36. Tasheva ES, Conrad GW. The UV responsive elements in the human mimecan promoter: a functional characterization. Mol Vis 2003; 9:1-9 <http://www.molvis.org/molvis/v9/a1/>.

37. Madisen L, Neubauer M, Plowman G, Rosen D, Segarini P, Dasch J, Thompson A, Ziman J, Bentz H, Purchio AF. Molecular cloning of a novel bone-forming compound: osteoinductive factor. DNA Cell Biol 1990; 9:303-9.

38. Ehinger M, Bergh G, Johnsson E, Baldetorp B, Olsson I, Gullberg U. p53-dependent and -independent differentiation of leukemic U-937 cells: relationship to cell cycle control. Exp Hematol 1998; 26:1043-52.

39. Goldstein D, Bushmeyer SM, Witt PL, Jordan VC, Borden EC. Effects of type I and II interferons on cultured human breast cells: interaction with estrogen receptors and with tamoxifen. Cancer Res 1989; 49:2698-702.

40. Sawadogo M, Van Dyke MW, Gregor PD, Roeder RG. Multiple forms of the human gene-specific transcription factor USF. I. Complete purification and identification of USF from HeLa cell nuclei. J Biol Chem 1988; 263:11985-93.

41. Sommer A, Bousset K, Kremmer E, Austen M, Luscher B. Identification and characterization of specific DNA-binding complexes containing members of the Myc/Max/Mad network of transcriptional regulators. J Biol Chem 1998; 273:6632-42.

42. Hodgkinson CA, Moore KJ, Nakayama A, Steingrimsson E, Copeland NG, Jenkins NA, Arnheiter H. Mutations at the mouse microphthalmia locus are associated with defects in a gene encoding a novel basic-helix-loop-helix-zipper protein. Cell 1993; 74:395-404.

43. Jane DT, Morvay LC, Koblinski J, Yan S, Saad FA, Sloane BF, Dufresne MJ. Evidence that E-box promoter elements and MyoD transcription factors play a role in the induction of cathepsin B gene expression during human myoblast differentiation. Biol Chem 2002; 383:1833-44.

44. Semov A, Marcotte R, Semova N, Ye X, Wang E. Microarray analysis of E-box binding-related gene expression in young and replicatively senescent human fibroblasts. Anal Biochem 2002; 302:38-51.

45. Sirito M, Lin Q, Maity T, Sawadogo M. Ubiquitous expression of the 43- and 44-kDa forms of transcription factor USF in mammalian cells. Nucleic Acids Res 1994; 22:427-33.

46. Yang XP, Freeman LA, Knapper CL, Amar MJ, Remaley A, Brewer HB Jr, Santamarina-Fojo S. The E-box motif in the proximal ABCA1 promoter mediates transcriptional repression of the ABCA1 gene. J Lipid Res 2002; 43:297-306.

47. Dong Y, Rohn WM, Benveniste EN. IFN-gamma regulation of the type IV class II transactivator promoter in astrocytes. J Immunol 1999; 162:4731-9.

48. Manzella L, Gualdi R, Perrotti D, Nicolaides NC, Girlando G, Giuffrida MA, Messina A, Calabretta B. The interferon regulatory factors 1 and 2 bind to a segment of the human c-myb first intron: possible role in the regulation of c-myb expression. Exp Cell Res 2000; 256:248-56.

49. Kerppola TK, Kane CM. Intrinsic sites of transcription termination and pausing in the c-myc gene. Mol Cell Biol 1988; 8:4389-94.

50. Fort P, Rech J, Vie A, Piechaczyk M, Bonnieu A, Jeanteur P, Blanchard JM. Regulation of c-fos gene expression in hamster fibroblasts: initiation and elongation of transcription and mRNA degradation. Nucleic Acids Res 1987; 15:5657-67.

51. Lattier DL, States JC, Hutton JJ, Wiginton DA. Cell type-specific transcriptional regulation of the human adenosine deaminase gene. Nucleic Acids Res 1989; 17:1061-76.

52. Au WC, Pitha PM. Recruitment of multiple interferon regulatory factors and histone acetyltransferase to the transcriptionally active interferon a promoters. J Biol Chem 2001; 276:41629-37.

53. Pattenden SG, Klose R, Karaskov E, Bremner R. Interferon-gamma-induced chromatin remodeling at the CIITA locus is BRG1 dependent. EMBO J 2002; 21:1978-86.