The JI
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     
 

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Saito, H.
Right arrow Articles by Kishimoto, T.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Saito, H.
Right arrow Articles by Kishimoto, T.
The Journal of Immunology, 2000, 164: 5833-5843.
Copyright © 2000 by The American Association of Immunologists

IFN Regulatory Factor-1-Mediated Transcriptional Activation of Mouse STAT-Induced STAT Inhibitor-1 Gene Promoter by IFN-{gamma}1

Hiroshi Saito2,*, Yoshiaki Morita*, Minoru Fujimoto*, Masashi Narazaki*, Tetsuji Naka* and Tadamitsu Kishimoto{dagger}

* Department of Molecular Medicine, Osaka University Medical School, and {dagger} Osaka University, Suita City, Osaka, Japan


   Abstract
Top
 Abstract
Introduction
 Materials and Methods
 Results
 Discussion
 References
 
STAT-induced STAT inhibitor-1 (SSI-1), also referred to as suppressor of cytokine signaling-1 and JAK-binding protein, is a member of a new family, the members of which are negative regulators of cytokine signals. SSI-1 is induced by various cytokines; however, the transcriptional mechanism of the SSI-1 gene is not fully understood. Here, we showed that transcription of the mouse SSI-1 gene was initiated from six adjoining sites accompanying three GC boxes and a single GC box-like element near them, but not from the TATA box or an initiator sequence. We also showed that IFN-{gamma} induced SSI-1 mRNA more strongly than IL-6 in NIH-3T3 fibroblasts and that this IFN-{gamma} effect was mediated by Stat1. To determine the signal pathway downstream of Stat1, transcriptional activities of several mutant promoters were examined. The region mediating stimulatory effect of IFN-{gamma} to the gene transcription was localized to the -88/-60 region containing three tandem GAAA units, named variant IFN-{gamma}-responsive element (VIRE), while four IFN-{gamma} activation site (GAS)-like elements located far upstream were not related to the IFN-{gamma} response. Gel-shift assays revealed that IFN-{gamma} induced IFN regulatory factor-1 (IRF-1) binding to VIRE, but not that of IRF-2 or three components of ISGF3. Furthermore, forced expression of IRF-1 mimicked and that of IRF-2 inhibited the stimulatory effect of IFN-{gamma} on SSI-1 gene transcription. Finally, mouse embryonal fibroblasts lacking IRF-1 showed impaired SSI-1 mRNA induction by IFN-{gamma}. These results demonstrated that IRF-1, which is induced by activation of Stat1, mediated transcriptional activation of the SSI-1 gene by IFN-{gamma} via VIRE.


   Introduction
Top
 Abstract
 Introduction
Materials and Methods
 Results
 Discussion
 References
 
Avariety of cytokines function through binding to members of the cytokine receptor superfamily, subsequently activating the Janus family of protein tyrosine kinases (JAKs)3 (1, 2, 3). Although the mechanisms of signal transduction and activation of cytokine-inducible genes were well established, the mechanisms of attenuation of their signals have not yet been fully determined. We have cloned a protein, STAT-induced STAT inhibitor-1 (SSI-1) (4), also referred to as suppressor of cytokine signaling-1 and JAK binding protein (5, 6), which was involved in attenuation of cytokine signals. Eight SSI-1-related proteins are now known. All the proteins possess a central SH2 domain and a highly conserved domain at the C terminus (7); thus, those are recognized as a family. Recent identification of the new family of proteins (4, 5, 6, 7, 8, 9), induced by various cytokines and involved in negative regulation of their own signals, brought new aspects for autoregulation of cytokine signals.

SSI-1 mRNA is induced by various cytokines, such as IL-4 and IL-6, leukemia inhibitory factor (LIF), and G-CSF in hemopoietic cell lines. It has been shown that SSI-1 mRNA induction by IL-6 or LIF is mediated by Stat3. SSI-1 was shown to be capable of suppressing the signal transduction of IL-6 or LIF by decreasing the tyrosine phosphorylation status of a receptor glycoprotein component (gp130) and Stat3 in murine myeloid leukemia cells (M1 cells) (4). Further studies showed that SSI-1 could bind to JAKs and inhibit their tyrosine phosphorylation and their activation, resulting in suppression of IL-6 signaling. Structurally, SSI-1 family proteins share two homologous domains, an SH2 domain and a C-terminal conserved domain, which we have called the SC-motif, also referred to as suppressor of cytokine signaling-1 box or the CH domain (10, 11). It has also been demonstrated that three district domains of SSI-1, the pre-SH2 domain, the SH2 domain, and the SC-motif, are involved in the function of SSI-1 (10). On the other hand, the mechanisms of induction of SSI-1 mRNA by cytokines have not fully been examined. Isolation and characterization of the promoter region of the mouse SSI-1 gene are necessary to clarify the mechanisms of cytokine-induced expression of SSI-1.

Besides hemopoietic cells, fibroblasts are also targets of various cytokines (12, 13, 14). IFN-{gamma} plays a role, particularly in the regulation of fibrosis: inhibition of collagen synthesis, decreased fibrous tissue formation, and retardation of wound healing (15). It was shown that IFN-{gamma} induced SSI-1 mRNA in a fibroblast cell line (16) and that SSI-1 inhibited IFN-{gamma} signaling (17), suggesting the presence of autoregulatory mechanisms of IFN-{gamma} signals mediated by SSI-1. Signals of IFN-{gamma} are transduced via two kinds of consensus sequences for the IFN-{gamma} response. One is the IFN-{gamma} activation site (GAS) element, a binding site for the Stat1 homodimer. The other is IRF-E/IFN-stimulated regulatory element (ISRE), a binding site for IRFs or ISGF3 (18). It has not yet been determined whether either element is involved in the induction of SSI-1 promoter by IFN-{gamma}.

In this study we have cloned the upstream region of the mouse SSI-1 gene and characterized it. We determined the structure of the promoter region of this gene and found a variant IFN-{gamma}-responsive element (VIRE) that was not identical with the consensus sequences reported previously (19). Furthermore, we demonstrated that IRF-1 mediated transcriptional activation of the mouse SSI-1 gene by IFN-{gamma} via VIRE.


   Materials and Methods
Top
 Abstract
 Introduction
 Materials and Methods
Results
 Discussion
 References
 
Cell culture

The mouse fibroblast cell line, NIH-3T3, and the mouse preadipocyte cell line, NIH-3T3L1, were obtained from American Type Culture Collection (Manassas, VA), and the mouse calvaria-derived osteoblastic cell line, MC3T3E1, was obtained from RIKEN Cell Bank (Tsukuba, Japan). Stat1- and IRF-1-deficient mouse embryonal fibroblasts were gifts from Drs. R. D. Schreiber and T. Taniguchi, respectively. NIH-3T3 cells and MC3T3E1 cells were maintained in DMEM (low glucose) and {alpha}-MEM, respectively, supplemented with 10% FCS. Embryonal fibroblasts and NIH-3T3L1 cells were maintained in DMEM (high glucose) supplemented with 10% FCS.

RNA preparation and Northern blot analysis

NIH-3T3, MC3T3E1, and NIH-3T3L1 cells were plated in each adequate medium supplemented with 2% FCS at densities of 1 x 106, 6 x 105, and 3 x 105 cells/100-mm tissue culture plate, respectively. On the following day, cells were washed twice with PBS and incubated for 28 h in serum-free medium (adequate medium/Ham’s F-12, 1/1). Thereafter, various cytokines or growth factors were added, and cells were harvested after various time intervals. Total RNA was extracted from each cell using RNAzol B (Tel-Test, Friendswood, TX). Ten-microgram aliquots were subjected to agarose gel electrophoresis and transferred to a membrane (Hybond-N+, Amersham, Arlington Heights, IL), which was hybridized with radiolabeled mouse SSI-1 cDNA probe at 65°C using Rapid-hyb buffer (Amersham).

S1 nuclease protection assay

S1 nuclease protection assay was principally performed using an S1 assay kit according to the manufacturer’s instructions (Promega, Madison, WI). A single-stranded antisense DNA probe corresponding to the -23/+67 region of the sequence upstream of the mouse SSI-1 gene was prepared with a DNA synthesizer and purified by PAGE (Nisshinseiphun, Tokyo, Japan). An aliquot of the probe, end labeled with [{gamma}-32P]ATP, was allowed to hybridize with 3 µg of poly(A) RNA prepared from IFN-{gamma}-stimulated NIH-3T3 cells or yeast transfer RNA at 55°C overnight.

Constructions of reporter plasmids

Genomic DNA including the entire coding region of the mouse SSI-1 gene was obtained by screening a mouse genomic library (Stratagene, La Jolla, CA). pSI1 and pSI3 were constructed by inserting SpeI-PvuII and XbaI-PvuII luciferase reporter plasmids (pGL3, firefly luciferase reporter vector, Promega). For construction of pSI2, a SmaI-SacI fragment was isolated from pSI1 and blunted by T4 polymerase (Toyobo, Tokyo, Japan), and then ligated into the SmaI site of pGL3. The PstI-BglII fragment of p00tk CAT (20) containing two GC boxes, a CAAT box, and a TATA box (-201/+17) of herpes simplex virus thymidine kinase promoter was ligated into blunted HindIII-BglII sites of pRL2 and SI4, yielding in pTK and pSI4, respectively. Hybridized oligonucleotides (identical with a gel-shift probe and cold competitors) were ligated into the SmaI site of pTK, yielding in pTKSGM2, pTKSGM3, and pTKSGM4. Two sets of ISRE sequence (see Fig. 5GoC) were inserted into the SmaI site of pTK, yielding pIRED. pSI5 and pSI7 were made by inserting fragments, prepared by PCR amplification using pSI3 as templates, into the SmaI site of pGL3. Constructs with mutations at GAAA units (pSG1–4, change to CTTT) or GC boxes (pGCM1–6) were prepared by the PCR-based mutagenesis method (21). Sequences of all constructs prepared though PCR amplification were verified by nucleotide sequencing (ABI PRISM 310 Genetic Analyzer, Perkin-Elmer, Norwalk, CT).



View larger version (27K):
[in this window]
[in a new window]
  		FIGURE 5. Effects of forced expression of IRF-1 and -2 on transcriptional activities of the mouse SSI-1 gene promoter. One hundred nanograms of pACT vector, pACT1(IRF-1), or pACT2(IRF-2) was cotransfected with 750 ng of various reporter plasmids and 15 ng of pRLSV40 luciferase control plasmid into NIH-3T3 cells. A, After 20 h of serum starvation, vehicle ({square}) or 20 ng/ml of IFN-{gamma} ({blacksquare}) was added and further incubated for 22 h. B, pACT vector ({square}), pACT1 ({blacksquare}), or pACT2 () was cotransfected with reporters and luciferase control plasmids. After 20 h of serum starvation, vehicle was added and further incubated for 22 h. Relative luciferase activity was defined as the ratio of the activities of reporter and control plasmids. Each column and bar represent the mean ± SE relative luciferase activities in three independent experiments. Bars for SE were compressed in several columns.

 
Transient transfection assays

NIH-3T3 cells were plated on 48-well dishes at a density of 1.5 x 104 cells/well. On the following day, a mixture of 0.75 µg of reporter plasmid, 0.015 µg of pRLSV40 control plasmid, 4.5 µl of SuperFect transfection reagent (Qiagen, Valencia, CA), and 50 µl of DMEM (low glucose) was incubated for 10 min at 25°C, and then combined with 250 µl of DMEM with 2% FCS. The resultant 300 µl of the mixture was added to each well, and cells were incubated for 3 h at 37°C. Additionally, cells were incubated for 28 h in serum-free medium, then various cytokines were added to the cells. After 22 h, cells were lysed with 65 µl of lysis buffer (Promega). The luciferase activity of each lysate was measured using Dual-Luciferase Reporter Assay System (Promega). For cotransfection experiments, 0.1 µg of pACT (mammalian expression plasmid bearing ß-actin promoter), pACT1 (human IRF-1 expression plasmid), or pACT2 (human IRF-2 expression plasmid) was added to the transfection mixtures.

Nuclear extracts preparation and gel-shift assays

Nuclear extracts were prepared as described previously (22) with minor modifications. NIH-3T3 cells, serum-starved for 28 h, were stimulated with 200 ng/ml of IFN-{gamma} (Calbiochem, La Jolla, CA) and harvested at various times, then collected and washed twice with ice-cold PBS. Sucrose was omitted from buffer A, and the NaCl concentration was changed to 0.4 M in buffer B. At concentration steps, Centricon-30 filters (Amicon, Beverly, MA) were used.

Sense and antisense strands of oligonucleotides corresponding to the -88/-60 region of the mouse SSI-1 gene promoter were allowed to hybridize and were end labeled with [{gamma}-32P]ATP (Amersham; 6000 Ci/mmol). Nuclear extracts were incubated with a 32P-labeled probe in 15 µl of binding buffer (12 mM HEPES (pH 7.9), 12% glycerol, 60 mM NaCl, 1.5 mM MgCl2, and 2 mM DTT) with 2 µg of poly(dI-dC) (Sigma) with or without competitor oligonucleotides. Following a 30-min incubation at 25°C, the reaction mixtures were subjected to electrophoreses in a 6% polyacrylamide (acrylamide/bisacrylamide, 29/1) gel in 0.7x TBE buffer (63 mM Tris base, 63 mM boric acid, and 0.35 mM EDTA, pH 8.3). Sequences of the sense strands of oligonucleotides for preparing the probe and cold competitors were shown in Fig. 4GoC. The sequence of GC competitor was 5'-CCAGGCAGGCCCCGCCCACAGGTCC-3'. All Abs used for supershift assays were purchased from Santa Cruz Biotechnology (Santa Cruz, CA).



View larger version (42K):
[in this window]
[in a new window]
  		FIGURE 4. Detection of complexes formed on the three tandem GAAA units (VIRE) in the mouse SSI-1 gene promoter. Nuclear extracts were prepared from NIH-3T3 cells before or after stimulation with IFN-{gamma} (200 ng/ml) and were subjected to gel-shift assays using the double-stranded oligonucleotides corresponding to the -88/-60 region as a probe. For competition assays, a 75-fold molar excess of nonradiolabeled oligonucleotides was added to reaction mixtures before addition of the radiolabeled probe. WT, M2, M3, M4, and M5 represent wild-type and M2, M3, M4, and M5 mutant competitors of IFRE, respectively. GC and ISRE represent GC nonrelated and canonical ISRE sequence competitors, respectively (A). For supershift assays, 2 µg of each Ab was added to the reaction mixtures before addition of the radiolabeled probe (B). Arrows, complexes A and B, and supershifted band. C, Sequence alignment of the competitors used for gel-shift assays. Three GAAA elements corresponding to G2, G3, and G4 elements in the mouse SSI-1 promoter (see Fig. 2GoB) are shown in boldface. G2/3 and G3/4 elements are shown by thin and thick underlines, respectively. The core binding sequence of IRF-1 in ISRE is double underlined. The ISRE sequence is that present in the promoter of human guanylate binding protein gene (23 ). The consensus sequence of IRF-1 binding (IRF-E) was referred (18 ). GAAA units present in ISRE and IRF-E are also shown in boldface.

 

   Results
Top
 Abstract
 Introduction
 Materials and Methods
 Results
Discussion
 References
 
Induction of SSI-1 mRNA expression by IFN-{gamma}

We first examined the effects of various cytokines and growth factors on induction of SSI-1 mRNA expression in a mouse fibroblast cell line, NIH-3T3, by Northern blot analyses. Overnight serum-starved NIH-3T3 cells exhibited no detectable expression of SSI-1 mRNA (Fig. 1GoA, lane 1). Of the factors tested, IFN-{gamma} induced high levels of expression of SSI-1, IL-6 and TNF-{alpha} induced low levels of expression, and IL-2, IL-4, and TGF-ß did not induce expression (Fig. 1GoA). Fig. 1GoB shows the time course of SSI-1 mRNA induction in NIH-3T3 cells by IFN-{gamma}. The mRNA level was maximal within 3 h after stimulation, declined thereafter, and then maintained a considerable level up to 22 h. We next examined IFN-{gamma}-induced expression of SSI-1 in other fibroblast cell lines and M1 leukemic cells. In M1 cells, SSI-1 mRNA has been shown to be induced by IL-6 (4) and was also induced by IFN-{gamma}, although the amount of mRNA induction appeared to be less than that in NIH-3T3 cells (Fig. 1GoC, lanes 1, 2, 7, and 8). NIH-3T3L1 and MC3T3E1 cell lines are derived from NIH-3T3 fibroblasts and exhibit characteristics of preadipocytes and osteoblasts, respectively. In addition to NIH-3T3 cells, these two fibroblast-derived cells responded to IFN-{gamma} and expressed SSI-1 mRNA, suggesting that the responsiveness to IFN-{gamma} is common to fibroblast-derived cells rather than specific to one cell line (Fig. 1GoC).



View larger version (44K):
[in this window]
[in a new window]
  		FIGURE 1. Effects of various cytokines and growth factors on mouse SSI-1 mRNA expression. Total RNA (10 µg/lane) was allowed to hybridize to the mouse SSI-1 (A–D) or IRF-1 cDNA probe (D). The upper panels and the middle panel in D show the results autoradiographed for 24 h. The lower panels show ethidium bromide staining of the gels. Positions of 28S and 18S ribosomal RNAs are shown at right. Arrows indicate the position of the mouse SSI-1 mRNA. A, NIH-3T3 cells were stimulated with 200 ng/ml of various factors for 3 h. B, NIH-3T3 cells were stimulated by 200 ng/ml of IFN-{gamma} for the indicated periods. C, Various cell lines were stimulated by 200 ng/ml of IFN-{gamma} for 1.5 h. D, Embryonal fibroblasts prepared from Stat1-deficient mice and their siblings as wild-type controls were stimulated with 200 ng/ml of IFN-{gamma} for 3 h.

 
As previously shown that SSI-1 mRNA was induced by IL-6 in a Stat3-dependent manner in M1 cells and that activation of Stat1 was a main pathway of IFN-{gamma} signaling, we attempted to verify whether Stat1 mediates this SSI-1 mRNA induction by IFN-{gamma}. Responsiveness to IFN-{gamma} of embryonal fibroblasts prepared from Stat1-deficient mice was analyzed. As shown in Fig. 1GoD, Stat1-/- fibroblasts did not respond to IFN-{gamma} stimulation in terms of SSI-1 mRNA induction, and concordantly, IRF-1 mRNA induction was diminished, while Stat1+/+ fibroblasts responded to IFN-{gamma} as expected. These results indicated that IFN-{gamma} induced SSI-1 via a Stat1-dependent pathway.

Determination of transcription initiation sites of the mouse SSI-1 gene

Cloning of the genomic DNA of the SSI-1 gene revealed that the entire coding region was confined to a single exon (data not shown). To detect a putative exon(s) consisting of a noncoding region(s) located upstream of the translation initiation codon, RT-PCR was performed using an antisense primer including the region just downstream of the ATG codon and several sense primers randomly selected from the region far upstream from the genomic sequences. As a result, an intron 512 bp in length was detected, the sequence of which was present in genomic DNA but not in cDNA derived from an RNA of NIH-3T3 cells, and the sequences of boundaries between this sequence of 512 bp and those presented in the cDNA coincided with the consensus sequences of the exon-intron boundary (Fig. 2GoA). Subsequently, S1 nuclease protection assays were performed using the genomic sequence just upstream of the intron as a probe. A nuclease-free control showed that the probe used here was intact (Fig. 2GoB, lane 5). The reaction using poly(A) RNA prepared from IFN-{gamma}-stimulated NIH-3T3 cells revealed six adjoining protected bands, while the reaction using yeast transfer RNA revealed no band (Fig. 2GoA, lanes 6 and 7). Sequencing of the region upstream of the SSI-1 gene revealed that four GAS-like elements are localized between -645 to -443 and four GAAA units between -94 to -66. In addition, the TATA box or initiator sequence was not found around the transcription initiation sites. Fig. 2GoC showed a genomic structure around the mouse SSI-1 gene. Sequence data in GenBank (AB000677) indicated that the mouse SSI-1 gene is located ~3 kb downstream of the mouse transition protein-2 (TP2) gene, and our present results showed that the mouse SSI-1 gene was composed of two exons. These results also indicated that the promoter of the mouse SSI-1 gene belongs to the TATA-less promoter and that it has six adjoining transcription initiation sites.



View larger version (65K):
[in this window]
[in a new window]
  		FIGURE 2. Structure of the promoter region of the mouse SSI-1 gene. A, The upstream nucleotide sequence of the mouse SSI-1 gene is depicted. The nucleotide sequence corresponding to the exons is shown in uppercase letters. The translation initiation codon (ATG) is underlined. Transcription initiation sites are shown by filled circles, and the position of the most 5' site of those is defined as +1. Consensus sequences of GC box (GGGCGG), GC box-like element (GGGTGG), and GAS element (TTC(N)3–4GAA) are indicated by boldface letters. Four GAAA units are indicated by boxes and named G1–G4. B, S1 nuclease protection assay was performed using a single-stranded antisense nucleotide corresponding to the -23/+67 region as a probe. The sequence ladder of corresponding region is shown at the left (lanes 1–4). End-labeled probes were allowed to hybridize with 3 µg of poly(A) RNA prepared from IFN-{gamma}-stimulated NIH-3T3 cells (lanes 5 and 6) or yeast transfer RNA (lane 7), then incubated at 37°C for 30 min in the presence (lanes 6 and 7) or the absence (lane 5) of S1 nuclease. C, The genomic structure of the mouse SSI-1 gene is depicted. The mouse TP2 gene is located just upstream of the SSI-1 gene. Open boxes, exons; bent arrow, a transcription initiation site. The locations of various upstream regions cloned into pSI1, -2, -3, and -4 luciferase constructs are also shown.

 
Effects of various cytokines and growth factors on mouse SSI-1 gene promoter activity

We prepared a reporter construct bearing the -2177/+122 region of the mouse SSI-1 promoter just upstream of the luciferase gene (pSI1) and subjected it to transient assays in mouse NIH-3T3 fibroblasts in the presence or the absence of various factors (Table IGo). Of the factors tested, IFN-{gamma} induced promoter activity to the greatest extent at both physiological (10 ng/ml) and pharmacological (200 ng/ml) doses. IL-6 and TNF-{alpha} also induced activity at pharmacological doses, but not to a marked extent at physiological doses. The effects of IL-2, IL-4, and TGF-ß on the activity were marginal. These results were consistent with those for SSI-1 mRNA expression levels; IFN-{gamma} was the strongest inducer of the factors tested, and IL-6 and TNF-{alpha} were the next strongest. To localize the regions responding to IL-6 or TNF-{alpha}, we analyzed the effects of IL-6 and TNF-{alpha} on the transcriptional activity of the promoter using another reporter construct, pSI3, containing the -105/+122 region of the SSI-1 gene promoter. NIH-3T3 cells were stimulated with or without 200 ng/ml of factors, and promoter activities of pSI3 were evaluated with the ratio of luciferase activities of reporter and control plasmids. Results were as follows: 0.333 ± 0.023 (no factor), 0.637 ± 0.040 (200 ng/ml of IL-6), and 0.590 ± 0.030 (200 ng/ml of TNF-{alpha}; n = 3). These results showed that the -105/+122 region of the SSI-1 gene promoter still retained responsiveness to IL-6 and TNF-{alpha}.


View this table:
[in this window]
[in a new window]
  		Table I. Effects of various factors for transcriptional activity of mouse SSI-1 promoter1

 
Detection of the regions responsible for IFN-{gamma} responsiveness in the mouse SSI-1 gene promoter

As Stat1 was essential for the induction of SSI-1 by IFN-{gamma} shown in Fig. 1GoD, we next attempted to determine the region(s) involved in the induction of transcription of the SSI-1 gene by IFN-{gamma} stimulation. Various mutants of the promoter were subjected to transient transfection assays in the presence or the absence of IFN-{gamma} in NIH-3T3 cells. pSI1 (-2777/+122) exhibited ~25-fold basal transcriptional activity compared with herpes simplex virus thymidine kinase (TK) promoter and a 2.6-fold increase in response to IFN-{gamma}, while TK promoter exhibited no response to IFN-{gamma}. Deleted constructs, pSI2 (-835/+122) and pSI3 (-105/+122), exhibited almost the same basal activities and IFN-{gamma} responsiveness as pSI1, whereas pSI7 (-67/+122) and pSI11 (-50/+122) exhibited no IFN-{gamma} responsiveness in addition to decreased basal activities (Fig. 3GoA). In addition, pSI4 (-2777/-106) including four GAS-like elements, exhibited basal activity only 2% of pSI12 and no response to IFN-{gamma} stimulation. To rule out the possibility that the unresponsiveness of the -2777/-106 region to IFN-{gamma} may be due to lack of basal promoter activity, the region was ligated to TK promoter and subjected to transient transfection assays. Subsequent construct pTKSI4 showed almost same basal activity as pTK and no response to IFN-{gamma} (Fig. 3GoB). The downstream deletion construct, pSI12 (-105/+18), exhibited a 4.9-fold increase in transcriptional activity when stimulated by IFN-{gamma}, demonstrating no contribution of the +19/+122 region to IFN-{gamma} responsiveness. These results suggested that IFN-{gamma} responsiveness was localized around the -105/-67 region and that four GAS-like elements were not involved in IFN-{gamma} responsiveness.



View larger version (36K):
[in this window]
[in a new window]
  		FIGURE 3. Functional analysis of the mouse SSI-1 gene promoter in NIH-3T3 cells. Structures of various luciferase reporter plasmids are depicted on the left of each panel (A, B, and C). Bent arrows, Transcription initiation sites. Open oval and closed circle, GAS-like sequence and the GAAA unit, respectively. Open square, GC box or GC box-like element (GC1). TK, herpes simplex viral thymidine kinase gene promoter. Various reporter constructs were transfected with pRLSV40 control plasmid in NIH-3T3 cells and serum starved for 28 h, then incubated with vehicle alone (open bar) or with 200 ng/ml of IFN-{gamma} (filled bar) for 22 h. Relative luciferase activity is defined as the ratio of the activities of reporter and control plasmids. The values are the mean of values from three independent experiments, and SE is shown by a thin line on the bar graph on the right of each panel.

 
Visual search revealed that the -105/-60 region included four GAAA units (tentatively named G1, G2, G3, and G4 units from upstream). Because IRF-E/ISRE, one of IFN-{gamma}-responsive elements, contained two GAAA units, we attempted to verify the contribution of each GAAA unit to IFN-{gamma} responsiveness by testing luciferase activities of reporter plasmids with each element disrupted. pSG2, pSG3, and pSG4 (G2, G3, and G4 units disrupted, respectively) lost IFN-{gamma} responsiveness, while pSG1 (G1 unit disrupted) exhibited nearly the same response as the wild type (pSI12; Fig. 3GoA), suggesting no contribution of the G1 unit to IFN-{gamma} responsiveness. To further assess the importance of the GAAA units in IFN-{gamma} responsiveness, wild-type and various mutants of the -88/-60 region were inserted upstream of a heterologous promoter (TK promoter), and IFN-{gamma} responsiveness was determined (Fig. 3GoB). pTKSGWT (including G2, G3, and G4 units) exhibited 3.6-fold higher luciferase activity in nonstimulated conditions than pTK (TK promoter alone) and responded to IFN-{gamma} stimulation with 2.6-fold increased luciferase activity. In contrast, pTKSGM2, pTKSGM3, and pTKSGM4 (mutations in G2, G3, and G4 units, respectively) each had no response to IFN-{gamma}. Therefore, the -88/-60 region, including the three tandem GAAA units (G2, G3, and G4), was suggested to be an IFN-{gamma}-responsive element, and each GAAA unit was indispensable for responsiveness to IFN-{gamma}. The element was tentatively named VIRE because its sequence was not identical with canonical IRF-E/ISRE.

We also analyzed regions other than GAAA units important for promoter activity. Because the -105/+18 region contained two GC boxes (GGGCGG) and single GC box-like element (GGGTGG), tentatively named GC1, GC2, and GC3 in order from upstream, we attempted to determine their contributions to the promoter activity by introducing mutations into these boxes. The second and third guanine residues of GC2 were replaced by thymidine residues to eliminate binding of the transcription factor SP1 to the box, yielding pGCM2. It exhibited decreased luciferase activity in nonstimulated conditions (46% that of pSI3), but not decreased IFN-{gamma} responsiveness (Fig. 3GoC). To confirm the contribution of GC2 to promoter activity, another mutation was introduced in the box (GGGCGG->GTTGTT). The resultant construct, pGCM4, also yielded essentially the same results as pGCM2. In addition, combined disruption of GC1 and GC2, or of GC1, GC2, and GC3 (pGCM5 and pGCM6, respectively), further reduced promoter activity in nonstimulated conditions (23% and 13% that of pSI3, respectively) without impairing IFN-{gamma} responsiveness (Fig. 3GoC). These results indicated that these three boxes were all involved in the transcriptional activity of the SSI-1 promoter in nonstimulated conditions, but not in IFN-{gamma} responsiveness.

Detection by gel-shift assays of complexes formed on the VIRE

Having established the importance of the -88/-60 region (VIRE) for the induction of SSI-1 by IFN-{gamma}, we next examined nuclear proteins that bind to the region. Nuclear extracts of NIH-3T3 cells were prepared before or after IFN-{gamma} stimulation and subjected to gel-shift assays using the -88/-60 region (VIRE) as a probe. Two complexes (A and B) were predominant in the reaction using extracts of IFN-{gamma}-stimulated cells (Fig. 4GoA, lane 2). Complex B was present in the reaction using extracts of nonstimulated cells and was not affected by IFN-{gamma} treatment, whereas complex A was detected only after the stimulation of IFN-{gamma} (Fig. 4GoA, lanes 1 and 2). The formation of either complex was sequence specific, because the wild-type competitor competed well with both complexes, but the nonrelated GC competitor did not (Fig. 4GoA, lanes 3 and 8). To verify the sequence requirements of the two complexes, competition assays were performed. The sequences used for the assays were shown in Fig. 4GoC. It was shown that complex B was well competed by M4 mutant competitor, but not by M2 and M5 mutant competitors, and that complex A was well competed by M2, M4, and M5 competitors but not by M3 competitor. Supershift assays revealed that complex A was shifted by addition of anti-IRF-1 Ab, but not by anti-IRF-2 or Abs for components of the ISGF3 complex (p48, Stat1, and Stat2), while complex B was not affected by any of the Abs tested here (Fig. 4GoB). In brief, at least two complexes were associated to VIRE: complex A, including IRF-1, was induced by IFN-{gamma} stimulation; complex B was present before the stimulation and was not related to IRFs or ISGF3. In addition, it was also shown that only two tandem GAAA units (the G2/3 or G3/4 element; see Fig. 4GoC) were sufficient for complex A formation in vitro.

IRF-1 enhanced transcriptional activity of the mouse SSI-1 promoter via the VIRE

To examine the effects of IRF-1 for induction of the SSI-1 gene directly, we expressed IRF-1 proteins with various reporter plasmids in NIH-3T3 cells (Fig. 5Go). Cotransfection of the IRF-1 expression plasmid increased the luciferase activity of pSI12 by ~4-fold over the value in mock-transfected controls, and IFN-{gamma} treatment produced a further increase in luciferase activity (Fig. 5GoA). In contrast, cotransfection of IRF-2 expression plasmid suppressed IFN-{gamma}-induced enhancement of the luciferase activity of pSI12. Because gel-shift assays showed that only two tandem GAAA units were sufficient for complex A formation, we tested mutated constructs, pSG2 and pSG4, both of which contained only two GAAA units. pSG2 and pSG4 showed no response to forced expression of IRF-1. pIRED is a luciferase construct bearing a TK promoter and two tandem repeats of canonical ISRE sequence (see Fig. 4GoC) that was present in the promoter of the human guanylate binding protein gene and mediated transcriptional enhancement of IRF-1 (23, 24). Forced expression of IRF-1 enhanced the luciferase activity of pIRED, and further enhancement by IFN-{gamma} was not observed, suggesting that forced expression of IRF-1 completely mimicked the effects of IFN-{gamma} when canonical ISRE was used. To further define the responsiveness of the VIRE to IFNs, TK promoter constructs bearing wild-type or mutant VIRE were used as a reporter (Fig. 5GoB). pTKSGWT showed a 2.5-fold increase in luciferase activity by forced expression of IRF-1, while pTKSGM2, pTKSGM3, and pTKSGM4 did not show any increase in luciferase activity by IRF-1 expression. These results suggested that IRF-1 induced transcription of the SSI-1 gene through VIRE and that IRF-2 blocked IFN-{gamma}-induced transcriptional enhancement of the SSI-1 gene. It was also suggested that all three GAAA units were required for VIRE to respond to IRF-1.

Role of each GAAA unit in IRF-1-mediated transcriptional activation of the mouse SSI-1 promoter

According to the results obtained to date, two tandem GAAA units could bind to IRF-1, whereas three GAAA units were required for IFN-{gamma}- and IRF-1-mediated induction. To dissolve the discrepancy, we first compared the degree of complex A formation on wild-type and M2 mutant VIRE. Fig. 6GoA showed decreased complex A formation when M2 mutant was used as a probe, suggesting lower affinity for IRF-1 of M2 mutant than wild type VIRE. We next analyzed the affinities of wild-type and mutant VIREs for IRF-1 using increasing molar excess of competitors (Fig. 6GoB), and band intensities were quantified by densitometric scanning (Fig. 6GoC). M2 and M4 showed lower affinities for IRF-1 than wild type. Requirement for 50% binding inhibition was estimated at ~6-, 15-, and 20-fold molar excesses of WT, M4, and M2 competitors, respectively, suggesting that two tandem GAAA units (G2 and G3, or G3 and G4) can bind to IRF-1 but with lower affinity than three tandem GAAA units. In addition, even a 75-fold molar excess of M7 (G2 and G4 units disrupted) as well as M3 (G3 unit disrupted) did not show any decrease in IRF-1 binding, suggesting that the G3 unit was required, but not sufficient, for IRF-1 binding. To examine the effects of this different affinity of each element for IRF-1 binding in vivo, transcriptional activities of several reporters were tested by cotransfection of increasing amounts of IRF-1 expression plasmid (pAct1; Fig. 6GoD). Luciferase activity of pSI12 (bearing wild-type VIRE) was increased by cotransfection of pAct1, reaching a maximum by 200 ng of pAct1. pSG2 showed lower basal activity than pSI12, and the effect of pAct1 up to 100 ng was minimal, whereas >200 ng of pAct1 increased pSG2 activity. Both pSG3 and pSG4 showed lower basal activity and apparently no response to pAct1. DNA binding and transcriptional activity of wild-type and mutant VIRE were summarized in Fig. 6GoF based on the results in Figs. 3GoA, 5A, and 6, B–E. In brief, G3 and G4 units were indispensable for IRF-1-mediated transcriptional activation, and G2 was also important for that, but the effect of disruption of the G2 unit was partially overcome by the larger amount of IRF-1.



View larger version (47K):
[in this window]
[in a new window]
  		FIGURE 6. Role of each GAAA unit in IRF-1-mediated transcriptional activation of the mouse SSI-1 promoter. A, Either wild-type (lane 1) or M2 mutant (lane 2) VIRE was end labeled, mixed with 2 µg of nuclear extracts prepared from IFN-{gamma}-stimulated NIH-3T3 cells, and subjected to gel-shift assays. B, Competition assays were performed using wild-type VIRE as a probe and 2 µg of extracts from IFN-{gamma}-stimulated cells. Ten-, 25-, or 75-fold molar excesses of several mutant competitors were added to reaction mixtures. A representative of the results is shown. C, Band intensities of complex A in the competition assays were quantified by densitometric scanning with the FMBIO analysis program. The percent DNA binding was expressed for the intensity of complex A without any competitors as 100% and is shown in log scale. Each point represents the mean of results obtained from two independent experiments. D, Six hundred nanograms of several reporters were cotransfected with 0, 10, 30, 100, 200, or 400 ng of pAct1, and luciferase activities were measured. Total amounts of plasmids transfected per well were adjusted to 1 µg with addition of pActC. The mean and SE were calculated from results of three independent experiments. E, Seven hundred fifty nanograms of several reporters were transfected into NIH-3T3 cells. Cells were treated with either vehicle ({blacksquare}) or 20 ng/ml of IFN-{gamma} ({square}) and subjected to luciferase assays. F, An open box and an open oval represent a putative complex B binding site and a GAAA unit, respectively.

 
In addition, the dose-dependent increase in IRF-1-mediated transcriptional activity (Fig. 6GoD) and the coincidence of decreased basal activity with impairment of IRF-1-mediated induction (Fig. 6GoF) suggested that the basal activity of VIRE was also mediated by IRF-1. As the presence of IRF-1 in nonstimulated nuclear extract was obscure (Fig. 4GoA, lane 1), larger amounts of extracts were subjected to gel-shift assays (Fig. 7GoA), and components of the retarded band were identified by anti-IRF-1 and -IRF-2 Abs. From this experiment, it was clear that IRF-1 binding activity was already present in nonstimulated nuclear extracts. In addition, cotransfection with >200 ng of pAct2 suppressed the basal activity of pSI12 but not that of pSG3 (Fig. 7GoB). This observation indicated that the basal activity was suppressed by IRF-2 in a VIRE-dependent manner. These findings suggested that both basal and induced VIRE-mediated transcriptional activities of SSI-1 promoter depend on the amount of IRF-1 binding activity.



View larger version (31K):
[in this window]
[in a new window]
  		FIGURE 7. Detection of IRF-1 in nonstimulated NIH-3T3 cells and suppression of the basal activity of the SSI-1 promoter by IRF-2. A, Either 0 (lane 1), 2 (lane 2), or 10 µg (lanes 3–5) of nuclear extracts obtained from nonstimulated HIN3T3 cells after 20-h serum starvation was subjected to gel-shift assays using wild-type VIRE as a probe. For supershift assays, 2 µg of anti-IRF-1 (lane 4) or anti-IRF-2 (lane 5) Ab was added to each reaction mixture before addition of the probe. B, Six hundred nanograms of pSI12 and pSG3 were cotransfected with 0, 100, 200, or 400 ng of pAct2, and luciferase activities were measured. The percent luciferase activities were calculated by adjusting the activities of the samples with 0 ng of pAct2 as 100%. Each circle and bar represent the mean ± SE obtained from four independent experiments.

 
To analyze relation of complex B to transcriptional activity of VIRE, another mutant, M6, was prepared in which all three GAAA units were intact. M6 mutant did not compete for complex B formation (Fig. 6GoB, lanes 5–7), indicating almost no affinity of M6 mutant for complex B. However, the affinity for IRF-1 binding (Fig. 6GoC), basal activity and IRF-1-mediated transcriptional activation (Fig. 6GoD), and response to IFN-{gamma} (Fig. 6GoE) of M6 mutant VIRE were indistinguishable from those of wild-type VIRE. These findings suggested that complex B did not contribute to basal or induced SSI-1 promoter activity.

SSI-1 mRNA inducibility by IFN-{gamma} in IRF-1-deficient fibroblasts

To further confirm the involvement of IRF-1 in SSI-1 mRNA induction by IFN-{gamma}, IRF-1-deficient embryonal fibroblasts were treated with IFN-{gamma}, and SSI-1 mRNA induction was assessed. As shown in Fig. 8Go, SSI-1 mRNA induction by IFN-{gamma} in IRF-1-/- fibroblasts was ~40% of that in IRF-1+/+ wild-type fibroblasts. As shown in the middle column of Fig. 8GoA, IRF-1 mRNA of a slightly small length was detected in the sample from IRF-1-/- fibroblast, which reflected the replacement of exon 2 by the neo resistance gene and indicated that the IRF-1 gene was disrupted (25). These results clearly demonstrated that IRF-1 was involved in the induction of SSI-1 mRNA by IFN-{gamma}.



View larger version (29K):
[in this window]
[in a new window]
  		FIGURE 8. SSI-1 mRNA inducibility by IFN-{gamma} in IRF-1-deficient fibroblasts. A, Nuclear extracts of embryonal fibroblasts were prepared from IRF-1-deficient mice and their siblings as wild-type controls, and NIH-3T3 cells were stimulated with 200 ng/ml of IFN-{gamma} for 3 h. Then, total RNAs were prepared and hybridized to the mouse SSI-1 (upper panel) or IRF-1 (middle panel) cDNA probe. The lower panel shows ethidium bromide staining of the gel. B, Northern blot analyses were repeated three times with different preparations of RNA samples from IRF-1-/- and IRF-1+/+ cells and the same preparation of the RNA samples from NIH-3T3 cells as a control. Band intensity was quantified by densitometric scanning with the FMBIO analysis program and is shown as the intensity relative to that of the IFN-{gamma}-treated NIH-3T3 sample. Horizontal bars are the SE of the results of three experiments.

 
Synergistic effects of TNF-{alpha} and IFN-{gamma} on SSI-1 mRNA induction

Because synergy between TNF-{alpha} and IFN-{gamma} in transcriptional activation of several genes including IRF-1 was evident, the synergistic effect of those two in SSI-1 gene transcription was analyzed. Addition of TNF-{alpha} or IFN-{gamma} alone induced SSI-1 mRNA, and the combination of TNF-{alpha} and IFN-{gamma} further induced it. The synergistic effect roughly coincided with that on IFR-1 mRNA induction (Fig. 9GoA). Transient transfection assays using pSI12 as a reporter revealed that TNF-{alpha} and IFN-{gamma} also showed synergy on transcriptional activation of the mouse SSI-1 gene promoter (Fig. 9GoB). Because pSI12 did not possess any NF-{kappa}B binding sites, a direct effect of TNF-{alpha} on the SSI-1 promoter is unlikely. It is speculated that the synergy between TNF-{alpha} and IFN-{gamma} on SSI-1 promoter activation is an indirect effect via IRF-1 induction.



View larger version (31K):
[in this window]
[in a new window]
  		FIGURE 9. Synergistic effect of TNF-{alpha} and IFN-{gamma} on SSI-1 mRNA induction. A, NIH-3T3 cells were stimulated by TNF-{alpha} (200 ng/ml) and/or IFN-{gamma} (20 ng/ml) for 3 h. Then, total RNAs were prepared and hybridized to the mouse SSI-1 (upper panel) or IRF-1 (middle panel) cDNA probe. The lower panel shows ethidium bromide staining of the gel. B, The pSI12 luciferase construct bearing the -105/+18 region of the mouse SSI-1 gene was transfected into NIH-3T3 cells and serum starved for 28 h, then incubated with 200 ng/ml of TNF-{alpha} and/or 20 ng/ml of IFN-{gamma}. After 22 h, cells were harvested and subjected to luciferase assays.

 

   Discussion
Top
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
References
 
In analyses of the region upstream of the mouse SSI-1 gene, we identified six adjoining transcription initiation sites by S1 nuclease protection assay and a 512-bp intron by RT-PCR. These findings indicate that the mouse SSI-1 gene is composed of at least two exons: exon 1 was comprised of a 5' noncoding sequence, and exon 2 contained the entire coding region. Our primer extension analyses in several different conditions did not yield clear results because of the presence of extremely high GC-rich sequences in the 5' end of mRNA (data not shown), leaving the possibility that another promoter is present far upstream from the one we identified. The mouse SSI-1 gene is located just downstream of the TP2 gene. pSI4, one of the luciferase constructs, contains a part of genomic DNA, the 5' end of which is just downstream of the 3' end of TP2 gene (26), implying that pSI4 contained the entire upstream region of the SSI-1 gene except downstream of -105 (see Fig. 2GoC). In transient transfection assays, the transcriptional activity of pSI4 was marginal (~2% of pSI12 in nonstimulated conditions; Fig. 3GoA). In addition, the -2777/-106 region affected TK promoter activity in neither nonstimulated nor IFN-{gamma}-stimulated conditions (Fig. 3GoB). These observations make the presence of another promoter within the region (-2777/-106) unlikely.

SSI-1 mRNA was induced by IFN-{gamma} stimulation in the serum-starved mouse fibroblast cell line NIH-3T3, and this induction by IFN-{gamma} was mediated by Stat1, because IFN-{gamma} treatment of Stat1-deficient fibroblasts exhibited no induction of SSI-1 mRNA. IFN-{gamma} stimulation mediated by Stat1 induces the transcription of various genes via two kinds of consensus sequences on their promoters (18). One is the GAS sequence, a binding site for Stat1 homodimer. The other is IRF-E/ISRE, a binding site for IRFs or ISGF3. Sequence analyses showed analogous sequences of both were presented on the SSI-1 promoter.

Four GAS-like sequences (TTC(N)3–4GAA) were found between -645 and -443 of the promoter (see Fig. 2GoB), raising the possibility that IFN-{gamma}-induced activation of the promoter may be mediated directly by Stat1 homodimer through the GAS-like sequences of the promoter. Results of deletion analyses (Fig. 3GoA) showed that luciferase activities of pSI3 (lacking all four GAS-like sequences) were almost the same as those of pSI1 and pSI2 in both nonstimulated and IFN-{gamma}-stimulated conditions. In addition, pSI4 and pSI4TK, containing all four GAS-like sequences, exhibited no response to IFN-{gamma} (Fig. 4Go). These findings indicated no contribution of these GAS-like sequences to the IFN-{gamma} responsiveness of the promoter.

Deletion analyses localized a region responsible for the IFN-{gamma} response, and introduction of mutations within this region (-105/-60) demonstrated that all three GAAA units (G2, G3, and G4 units) were required for the promoter to respond to IFN-{gamma} (Fig. 2GoB). The finding was further confirmed by introduction of the -88/-60 region (tentatively named VIRE) into a heterologous promoter (Fig. 3GoB). The sequence of VIRE in the mouse SSI-1 promoter resembled that of IRF-E/ISRE. The difference between this VIRE and IRF-E/ISRE is that the sequence of the former contains three GAAA units, while IRF-E/ISRE contains only two GAAA units. Gel shift assays detected two complexes that bound to the VIRE in the mouse SSI-1 promoter. Complex B was present both before and after IFN-{gamma} stimulation with no apparent change in intensity, whereas complex A was induced by IFN-{gamma} and was shown to be composed of IRF-1. Neither complex was affected by Abs for three components of ISGF3, suggesting that ISGF3 bears no relation to VIRE-mediated induction of mouse SSI-1 gene promoter by IFN-{gamma} (Fig. 4Go). Cotransfection experiments demonstrated that forced expression of IRF-1 mimicked IFN-{gamma} stimulation, while IRF-2 diminished induction of the SSI-1 promoter by IFN-{gamma}. It was also shown that this transcriptional activation by IRF-1 was abolished by introduction of mutation into VIRE (Fig. 5Go). Furthermore, IRF-1-deficient fibroblasts showed impaired response to IFN-{gamma} in terms of the SSI-1 mRNA induction (Fig. 6Go). These findings demonstrated that IFN-{gamma} induced IRF-1, which bound to VIRE and thereby activated the mouse SSI-1 gene promoter.

Recently, several VIRE-like elements were reported. Elements containing complete three tandem GAAA units with 2-base spacers and those with one base mismatch are listed in Table IIGo. The VIRE-like element, present in the promoter region of bovine IFN-stimulated gene 17 (bISG17), was suggested to bind to IRF-1 and transduce signals of IFNs to the bISG17 gene (27). TAP1 and low molecular mass polypeptide 2 genes share a common promoter in which a VIRE-like element is present. The VIRE-like element was shown to bind to IRF-1 and to play a crucial role in up-regulation of TAP1 and low molecular mass polypeptide 2 genes by IFN-{gamma} (28). However, exact sequence requirements for IRF-1 binding and for transcriptional activation of the genes were not examined in both reports. Concerning human IFN-stimulated gene 15 (hISG15), it was shown that IRF3 homodimer, induced by virus-stimulated phosphorylation, bound to a VIRE-like element in the hISG15 promoter and activated transcription of the gene (29). These three genes were all IFN inducible, suggesting the importance of VIRE in response to IFNs.


View this table:
[in this window]
[in a new window]
  		Table II. VIREs of several genes

 
In the present study we have analyzed sequence requirements both for IRF-1 binding and for transcriptional activation by IRF-1 or IFN-{gamma}. Analyses of the crystal structure of IRF-DNA complexes have established the recognition sequence of IRF-1 and IRF-2 as AANNGAAA (32, 33). Because the recognition sequence spanned two GAAA units, the G2/G3 element or G3/G4 element (see Fig. 3GoC) was raised as a candidate for the IRF-1 binding site within VIRE. Competition gel-shift assays revealed that the G3 unit was required for IRF-1 binding (Fig. 6Go, B and C). However, M7 mutant (only G3 unit was intact) could not bind to IRF-1, indicating that only the G3 unit is not sufficient for IRF-1 binding and that an additional element is required. Because both M2 (with an intact G2/G3 element) and M4 (with an intact G3/G4 element) mutants could bind to IRF-1 in vitro, we could not determine which site was predominantly responsible for IRF-1 binding. Cotransfection experiments demonstrated that M3 and M4 mutant VIREs did not show any response to a large extent of forced expression of IRF-1, indicating that both G3 and G4 units were crucial for IRF-1-mediated induction of the SSI-1 promoter in vivo. By contrast, M2 mutant VIRE could respond to overexpression of IRF-1, while the responsiveness was much lower than that of wild-type VIRE. This indicated that VIRE could transduce IRF-1 stimuli without the G2 unit with much less efficiency than wild-type VIRE in vivo. These findings suggested that the G3 and G4 units (or G3/G4 element) were the predominant binding site for IRF-1, and a crucial cis-element for IRF-1 mediated transcriptional activation of the SSI-1 gene.

In view of the findings that disruption of the G2 unit elicited lower binding affinity for IRF-1 and lesser efficiency of IRF-1-mediated transcriptional activation, it was suggested that the role of the G2 unit was to increase the affinity of the G3/G4 element for IRF-1, thereby increasing the effect of IRF-1 on transcriptional activation. Homodimerization of IRF-2 DNA binding domains on the sequence 5'-ACAAGTGAAAGTGAAA-3' was reported (33) (underlining indicates the nucleotides contacted with two IRF-2 DNA binding domains). The first binding of the IRF-2 DNA binding domain was shown to cause enhancement of the second binding, and the reason for this cooperativity was suggested to be that the first binding introduced DNA structural distortions, thereby increasing the affinity of the DNA for the second binding. Although homodimerization of IRF-1 was not demonstrated in the report above, it is possible that homodimerization may occur on the VIRE of the SSI-1 promoter and that the homodimerization may be required for full activity of the VIRE. This hypothesis can explain the enhancing effect of the G2 unit for IRF-1 binding. However, we could not detect homodimerization of IRF-1 in the gel-shift assays using increasing amounts of recombinant IRF-1 proteins (data not shown). Ternary complex formation and coordination of IRFs and another transcriptional factor were reported. The Ig light chain gene enhancer E{lambda}2–4 contains two domains, {lambda}A and {lambda}B, that are essential for enhancer activity. The {lambda}B element consists of two juxtaposed, but distinct, transcription factor-binding sites, PU.1 and IRF-4. Binding of IRF-4 to the element was marginal without PU.1, whereas with PU.1, IRF-4 binding was enhanced, and both factors functioned as mutually dependent transcriptional activators of the composite element (34). In this view, an unknown factor(s) may bind to the G2 unit and enhance IRF-1 binding to the G3/G4 element, thereby eliciting transcriptional activation of the SSI-1 gene. Complex B is a candidate for the unknown factor that cooperatively functions with IRF-1. To test the possibility, we prepared a new mutant VIRE (M6) and a corresponding reporter construct (pSG6) bearing the three complete GAAA units but having no affinity for complex B. M6 mutant showed almost the same affinity for IRF-1 as wild-type VIRE (Fig. 6GoC), and pSG6 responded to forced expression of IRF-1 to almost the same extent as pSI12 (Fig. 6GoD). These findings made cooperativity between the complex B and IRF-1 unlikely.

We have demonstrated the involvement of IRF-1 in transcriptional activation of the SSI-1 gene by IFN-{gamma}; however, IRF-1-deficient fibroblasts still responded to IFN-{gamma} in terms of SSI-1 mRNA induction to even a lesser extent than wild-type fibroblasts. It is possible to speculate that IRFs other than IRF-1 may transduce IFN-{gamma} stimuli, especially in IRF-1-deficient fibroblasts. Alternatively, an additional mediator(s) of the Stat1-dependent pathway other than IRF-1 and ISGF3 may be present and enhanced in IRF-1-deficient circumstance.

Because the mouse SSI-1 promoter had neither a TATA box nor initiator consensus sequence in an adequate location, it is reasonable to speculate that the GC box(es) may confer basal transcriptional activity. Although we were able to demonstrate the contributions of GC1, GC2, and GC3 to the activity, pGCM6 (with all three boxes disrupted) still had some activity (Fig. 3GoC). Thus, we could not rule out the presence of an element(s) required for the basal transcription of the promoter other than these three boxes.

In addition to IFN-{gamma}, the mouse SSI-1 promoter was induced by IL-6 and TNF-{alpha}, while the fold induction by the latter two was less than that by IFN-{gamma} (Table IGo). The shorter construct, pSI3, also resulted in ~2-fold induction of transcription, indicating that the region responsible for IL-6- or TNF-{alpha}-induced activation was present within the -105/+122 region of the promoter. Induction of IRF-1 by IL-6 stimulation is well established in various cells, such as M1 cells, human breast carcinoma T47D cells, and murine hybridoma B9 cells. In T47D cells and B9 cells, IL-6 was a less potent inducer of IRF-1 mRNA than IFN-{gamma} (35, 36). TNF-{alpha} also induced transcriptional activation of the IRF-1 gene promoter through the proximal {kappa}B site, and the strength of this transcriptional activation was ~15% that of IFN-{gamma} (37). Thus, transcriptional induction of the SSI-1 promoter by IL-6 or TNF-{alpha} may be mediated by IRF-1, and the strength of induction may depend on the degree of induction of IRF-1 by IL-6 or TNF-{alpha}. Concerning TNF-{alpha}, synergistic effects of TNF-{alpha} and IFN-{gamma} on SSI-1 mRNA induction were evident, as seen for IRF-1 mRNA induction, which further emphasizes the involvement of IRF-1 in TNF-{alpha}-mediated induction of the SSI-1 mRNA. In addition, we previously showed that IL-6 induced SSI-1 mRNA and that this induction was blocked by forced expression of a dominant negative form of Stat3 protein in M1 cells (4). It was also found that induction of IRF-1 mRNA by IL-6 was mediated through Stat3 (38). Thus, it is possible to speculate that IL-6-induced SSI-1 mRNA expression is mediated through IRF-1, which was induced by binding of activated Stat3 to the GAS sequence on the IRF-1 promoter. The present study showed that IL-4 did not induce the SSI-1 mRNA in NIH-3T3 cells, although we previously reported that IL-4 induced SSI-1 mRNA in IL-4-dependent CT4S cells (4). Chen et al. reported that IL-4 signal transduction was impaired in NIH-3T3 cells (39). Thus, this discrepancy may be due to the differences in responsiveness to IL-4 between cell lines.


   Acknowledgments
 
We thank Drs. Tadatsugu Taniguchi and Naoki Hata for providing IRF-1 and IRF-2 expression plasmids, and Drs. R. D. Schreiber and Tadatsugu Taniguchi for providing Stat1-deficient and IRF-1-deficient fibroblasts, respectively. We also thank Dr. Tomoshige Matsumoto for his technical assistance, and Keiko Tsujii and Mariko Tagami for their excellent secretarial assistance in preparing the manuscript.


   Footnotes
 
1 This work was supported by a grant-in-aid from the Ministry of Education, Science, and Culture, Japan. Back

2 Address correspondence and reprint requests to Dr. Hiroshi Saito, Department of Molecular Medicine, Osaka University Medical School, 2-2 Yamada-oka, Suita City, Osaka 565-0871, Japan. Back

3 Abbreviations used in this paper: JAK, Janus family of protein tyrosine kinases; SSI-1, STAT-induced STAT inhibitor-1; LIF, leukemia inhibitory factor; SC-motif, SSI COOH-terminal motif; IRF-1, IFN regulatory factor-1; IRF-E, consensus sequence of IRF-1 binding; VIRE, variant IFN-{gamma}-responsive element; TK, thymidine kinase; TP2, transition protein-2; IPTG, isopropyl-ß-D-thiogalactopyranoside; GAS, IFN-{gamma} activation site; ISRE, IFN-stimulated regulatory element. Back

Received for publication October 4, 1999. Accepted for publication March 22, 2000.


   References
Top
 Abstract
 Introduction
 Materials and Methods
 Results
 Discussion
 References
 

  1. Kishimoto, T., S. Akira, M. Narazaki, T. Taga. 1995. Interleukin-6 family of cytokines and gp130. Blood 86:1243.[Free Full Text]
  2. Ihle, J. N.. 1995. Cytokine receptor signaling. Nature 377:591.[Medline]
  3. Jr Darnell, J. E., I. M. Kerr, G. R. Stark. 1994. Jak-STAT pathways and transcriptional activation in response to IFNs and other extracellular signaling proteins. Science 264:1415.[Abstract/Free Full Text]
  4. Naka, T., M. Narazaki, M. Hirata, T. Matsumoto, S. Minamoto, A. Aono, N. Nishimoto, T. Kajita, T. Taga, K. Yoshizaki, et al 1997. Structure and function of a new STAT-induced STAT inhibitor. Nature 387:924.[Medline]
  5. Starr, R., T. A. Willson, E. M. Viney, L. J. Murray, J. R. Rayner, B. J. Jenkins, T. J. Gonda, W. S. Alexander, D. Metcalf, N. A. Nicola, et al 1997. A family of cytokine-inducible inhibitors of signaling. Nature 387:917.[Medline]
  6. Endo, T. A., M. Masuhara, M. Yokouchi, R. Suzuki, K. Mitsui, H. Sakamoto, M. Ohtsubo, H. Misawa, Y. Kanekura, A. Yoshimura. 1997. A new protein containing an SH2 domain that inhibits JAK kinases. Nature 387:921.[Medline]
  7. Naka, T., M. Fujimoto, T. Kishimoto. 1999. Negative regulation of cytokine signaling: STAT-induced STAT inhibitor. TIBS 24:394.
  8. Minamoto, S., K. Ikegame, K. Ueno, M. Narazaki, T. Naka, H. Yamamoto, T. Matsumoto, H. Saito, S. Hosoe, T. Kishimoto. 1997. Cloning and functional analysis of new members of STAT-induced STAT inhibitor (SSI) family: SSI-2 and SSI-3. Biochem. Biophys. Res. Commun. 237:79.[Medline]
  9. Masuhara, M., H. Sakamoto, A. Matsumoto, R. Suzuki, H. Yasukawa, K. Mitsui, T. Wakioka, S. Tanimura, A. Sasaki, H. Misawa, et al 1997. Cloning and characterization of novel CIS family genes. Biochem. Biophys. Res. Commun. 239:439.[Medline]
  10. Narazaki, M., T. Matsumoto, M. Fujimoto, Y. Morita, H. Saito, T. Kajita, K. Yoshizaki, T. Naka, T. Kishimoto. 1998. Three distinct domains of SSI-1/SOCS-1/JAB protein are required for its suppression of interleukin 6 signaling. Proc. Natl. Acad. Sci. USA 95:13130.[Abstract/Free Full Text]
  11. Hilton, D. J., R. T. Richardson, W. S. Alexander, E. M. Viney, T. A. Wilson, N. S. Sprigg, R. Starr, S. E. Nicholson, D. Metcalf, N. A. Nicola. 1998. Twenty proteins containing a C-terminal SOCS box form five structural classes. Proc. Natl. Acad. Sci. USA 95:114.[Abstract/Free Full Text]
  12. Chomarat, P., J. Banachereau. 1997. An update on interleukin-4 and its receptor. Eur. Cytokine Netw. 8:333.[Medline]
  13. Harada, H., M. Kitagara, N. Tanaka, H. Yamamoto, K. Harada, M. Ishihara, T. Taniguchi. 1993. Anti-oncogenic and oncogenic potentials of interferon regulatory factors-1 and -2. Science 259:971.[Abstract]
  14. Tremblay, G. M., M. Jordana, J. Gauldie, and B. Särnstrand. 1995. Fibroblasts as effector cells in fibrosis. In Lung Biology in Health and Disease, Vol. 80. S. H. Phan and R. S. Thrall, eds. Marcel Dekker, New York, p. 541.
  15. Raghu, G., M. Kinsella. 1993. Cytokine effects on extracellular matrix. , ed. In Lung Biology in Health and Disease Vol. 61:491. Marcel Dekker, New York.
  16. Sakamoto, H., H. Yasukawa, M. Masuhara, S. Tanimura, A. Sasaki, K. Yuge, M. Ohtsubo, A. Ohtsuka, T. Fujita, T. Ohta, et al 1998. A Janus kinase inhibitor, JAB, is interferon-{gamma}-inducible gene and confers resistance to interferons. Blood 92:1668.[Abstract/Free Full Text]
  17. Adams, T. E., J. A. Hansen, R. Starr, N. A. Nicola, D. J. Hilton, N. Billestrap. 1998. Growth hormone preferentially induces the rapid, transient expression of SOCS-3, a navel inhibitor of cytokine receptor signaling. J. Biol. Chem. 273:1285.[Abstract/Free Full Text]
  18. Taniguchi, T., M. S. Lamphier, N. Tanaka. 1997. IRF-1: the transcription factor linking the interferon response and oncogenesis. Biochim. Biophys. Acta 1333:M9.[Medline]
  19. Tanaka, N., T. Kawakami, T. Taniguchi. 1993. Recognition DNA sequences of interferon regulatory factor (IRF-1) and IFR-2, regulators of cell growth and the interferon system. Mol. Cell. Biol. 13:4531.[Abstract/Free Full Text]
  20. Saito, H., H. Kouhara, T. Harada, S. Miyake, H. Sugiyama, T. Kishimoto, B. Sato. 1994. Mapping of a transcription element critical for expression of the fibroblast growth factor receptor 1 gene. Biochem. Biophys. Res. Commun. 198:1020.[Medline]
  21. Higuchi, R., B. Krummel, R. K. Saiki. 1988. A general method of in vitro preparation and specific mutagenesis of DNA fragments: study of protein and DNA interactions. Nucleic Acids Res. 16:7351.[Abstract/Free Full Text]
  22. Saito, H., T. Oka. 1996. Hormonally regulated double- and single-stranded DNA-binding complexes involved in mouse ß-casein gene transcription. J. Biol. Chem. 271:8911.[Abstract/Free Full Text]
  23. Wang, I. M., J. C. G. Blanco, S. Y. Tsai, M. J. Tsai, K. Ozato. 1996. Interferon regulatory factors and TFIIB cooperatively regulate interferon-responsive promoter activity in vivo and in vitro. Mol. Cell. Biol. 16:6313.[Abstract]
  24. Lew, D. J., T. Decker, I. Strehlow, J. E. Darnell. 1991. Overlapping elements in the guanylate-binding protein gene promoter mediate transcriptional induction by {alpha} and {gamma} interferons. Mol. Cell. Biol. 11:182.[Abstract/Free Full Text]
  25. Tanaka, N., M. Ishihara, M. Kitagawa, H. Harada, T. Kimura, T. Matsuyama, M. S. Lamphier, S. Aizawa, T. W. Mak, T. Taniguchi. 1994. Cellular commitment to oncogene-induced transformation or apoptosis is dependent on the transcription factor IRF-1. Cell 77:829.[Medline]