Cooperative Interaction between Interferon ( IFN ) Stimulus Response Element and KB Sequence Motifs Controls IFNy and Lipopolysaccharide-stimulated Transcription from the Murine IP-10 Promoter

The transcriptional regulation of the murine IP-10 gene in lipopolysaccharide (LPS)  or  interferon (1FNy)-treated  macrophages  was  investigated by analysis of regions of the gene that flank the transcription start site. A series of sequence fragments were placed 5’ to the chloramphenicol acetyltransferase (CAT) reporter gene and ability to mediate transcription of CAT in response to IFNy or LPS treatment was studied following transient transfection in the macrophage-like cell line RAW 264.7. Analysis of larger constructs identified a potential negative regulatory site for IFNy response in the region between nucleotide positions -2002 and -930 and a positive regulator for LPS response in the region between bases -930 and -676. A 227-base fragment spanning positions -228 to -2 was the minimal sequence able to mediate LPSand IFNy-dependent transcription of CAT. Deletion of 24 bases, which included a highly conserved IFN stimulus response element (ISRE) from the -228 construct, abolished response to IFNy. A 33-base fragment containing the IP-10 ISRE was able to confer both IFNy and LPS sensitivity upon a heterologous promoter. The ability of LPS to stimulate CAT via the ISRE was apparently mediated by intermediate expression of endogenous IFNaIB. Elimination of bases -204 to -102 abolished sensitivity to LPS. This region contains two KB binding sites. Site-directed mutagenesis of key nucleotides in the ISRE and the two KB sites demonstrated that optimal response to IFNy required both the ISRE and one of the two KB sites, whereas optimal response to LPS required either both KB sites or one KB site and the ISRE. IFNy or LPS treatment induced sequencespecific binding activity for the ISRE and the two KB sites. These results indicate that the 230 nucleotides upstream from the transcription start site are important for transcriptional control of the IP-IO gene in response to IFNy and LPS. The three defined regulatory elements function in distinct fashion for each of the two stimuli; optimal response to either IFNy or LPS requires cooperation between at least two sites.

The transcriptional regulation of the murine IP-10 gene in lipopolysaccharide (LPS) or interferon (1FNy)-treated macrophages was investigated by analysis of regions of the gene that flank the transcription start site. A series of sequence fragments were placed 5' to the chloramphenicol acetyltransferase (CAT) reporter gene and ability to mediate transcription of CAT in response to IFNy or LPS treatment was studied following transient transfection in the macrophage-like cell line RAW 264.7. Analysis of larger constructs identified a potential negative regulatory site for IFNy response in the region between nucleotide positions -2002 and -930 and a positive regulator for LPS response in the region between bases -930 and -676. A 227-base fragment spanning positions -228 to -2 was the minimal sequence able to mediate LPSand IFNy-dependent transcription of CAT. Deletion of 24 bases, which included a highly conserved IFN stimulus response element (ISRE) from the -228 construct, abolished response to IFNy. A 33-base fragment containing the IP-10 ISRE was able to confer both IFNy and LPS sensitivity upon a heterologous promoter. The ability of LPS to stimulate CAT via the ISRE was apparently mediated by intermediate expression of endogenous IFNaIB. Elimination of bases -204 to -102 abolished sensitivity to LPS. This region contains two KB binding sites. Site-directed mutagenesis of key nucleotides in the ISRE and the two KB sites demonstrated that optimal response to IFNy required both the ISRE and one of the two KB sites, whereas optimal response to LPS required either both KB sites or one KB site and the ISRE. IFNy or LPS treatment induced sequencespecific binding activity for the ISRE and the two KB sites. These results indicate that the 230 nucleotides upstream from the transcription start site are important for transcriptional control of the IP-IO gene in response to IFNy and LPS. The three defined regulatory elements function in distinct fashion for each of the two stimuli; optimal response to either IFNy or LPS requires cooperation between at least two sites. The mononuclear phagocyte system participates in immunoregulation, orchestration of inflammation, and maintenance of homeostasis (1)(2)(3). Most of these functions are acquired following exposure of the cells to an almost bewildering array of extracellular stimuli (1)(2)(3). Among the most frequently encountered signals are bacterial cell wall products such as lipopolysaccharide (LPS)' and secretory products of lymphoid cells such as interferon y (IFNr) (4, 5). LPS and 1FN-y are both well known macrophage activating stimuli that act at least in part by the induction of new gene expression (4)(5)(6). These agents act by initiation of multiple intracellular signaling pathways, which culminate in altered gene-specific function in the nucleus (6)(7)(8).
IFNr and LPS have been shown to have overlapping activities with respect to the induction of certain gene products (5,6). Among these is the inflammatory protein 10-kDa (IP- 10) gene, originally identified in human U937 monocytic leukemia cells as an IFNy-inducible product (9) and subsequently in murine peritoneal macrophages on the basis of LPS sensitivity (10,11). Although the physiologic function(s) and significance of the IP-10 gene product have not been clarified, it has significant sequence homology with the super family of chemotactic factor genes recently designated the intercrine or small cytokine family (11)(12)(13). This superfamily is subdivided into two groups based upon the relative positioning of the first two cysteine residues in the sequence (CC uersus CXC). In the mouse the intercrine a family includes IP-10, MIP-2, and KC (10,11,(13)(14)(15)). The intercrine /3 family includes MIP-la and /3 and JE (13,14,16,17). Because of the diverse array of extracellular stimuli that can induce expression of IP-10 and the large number of cell types in which expression has been identified (18)(19)(20)(21), this gene represents a good model system for analysis of molecular mechanisms involved in the inducible expression of a single gene in response to distinct stimuli.
Many recent studies have focused on the identification and characterization of cis-acting sequence elements present in the region 5' of the transcription start sites of inducible genes and on trans-acting factors, which recognize and bind such sequences in stimulus-dependent fashion (22)(23)(24)(25)(26)(27)(28)(29). As a first step toward understanding the molecular mechanisms involved in regulating expression of the IP-10 gene, we have carried out a functional analysis of the 5"flanking region of the JP-10 gene using transient transfection of a reporter gene The abbreviations used are: LPS, lipopolysaccharide; IFN, interferon; CAT, chloramphenicol acetyltransferase; ISRE, IFN stimulus response element; DTT, dithiothreitol; PMSF, phenylmethylsulfonyl fluoride; EMSA, electrophoretic mobility shift assay; bp, base pair(s); PIPES, 1,4-piperazinediethanesulfonic acid NF, nuclear factor; IL-6, interleukin 6.

6678
Transcrlptional Regulation of IP-10 T C T T~  -1885   c -2002   c-Re1  lm/CK-l   -1884 T A T A G A T T C t T T T A T T T C A G T C A T T~C A T T M C T T T A T A~C A T~~T T~T A t T T T t T T A C A G A T C T G M T A T t T T~C T T~C C T A G C C A~T C C C  -1165  NF-IL6 c-Re1

-1164 A T G A C A T C T A T A C C A T A C T T G~A T M C C T Y i C T A T M t T C A~G C -T G G C C T G T T C T A t T C~G G G C T G C t T T C M G T A C~T T T A C T G G A T A C M T G T A T G T A T A C T A -1645 CK-2 -1644 C A G T~T t T T A G C A T -t T C T t T G T -~G T C C M T T T~C C C C T~T A f f i T G T T T~C A C C~G~C A T C~G A T M C T T C A G C C M G C~G A C A C -1525 I RF
He4t Shock

CK-1 -684 T f f i C T G C T C T G C A G A G T G C A T C C A C A T T~~C T~C C C T G~A C C A C A C A C T~C~G M G A C M T C~G~T C C C T G A~G M T C A G C A C f f i G G~C M T T A G~ -565 I RF OCT -564 ~T t C C T C T T G T M C t T G T~C A T T A~G T~C A C T T C C G G C t T C T t T T C T G M T G C C T G C T C T M C~T M G T T G T A T A G~C T G G C T~T T T f f i C A T G T G A T -445 CK-2 OCT -444 T~T T T T T T~G T T T A T T C C M G T A T T T T A T M T~G C C T A T t M G T~G C~T~C A G T G T~~C M C~T G T~C C A T G G T T A~C C T G A C T T A G A T A T C A G -325
-330

T A G G C T G G T T G M~C A C G~T A G M T~~G~C G M G M G G G M~G~~~G G G G G A G~-~G~G G G M G A G A G A G T C A G G A~f f i G C~ 396
397 MGAGTGGGATCG~~~TMGCCATGGATGCCTCCTCCT~GCCTGAGCCTMCCMTACTGTGAGCA~GCAT~TGCAGGA~TCGTM~GA~G~GCAGATCTCTCTTTACCATGAC~ 516

M G G G G T T A T C M W T A C T C A C G G M C C T G~T G T A T G T G T~T A C T A T T T M T G M C G A C T G T A~t T A G M T T C C T A G A T G T A T T T~T G T A T G C T T T~A T T G T A T A T G~ 2076
2077 AGMCTTCTGTCATCMGTATGAT~TffiGTAGTT~GTTTATTTTT-CCGTCCMTACCTTTTGTATTATGTMCATTC-GACMTGTACTGTATT~GTAGTAAGAG 2196 c an IFN stimulus response element (ISRE) and two structurally and functionally distinct KB elements, respectively.
Cell Culture-RAW 264.7 mouse macrophage cells were subcultured twice weekly in Dulbecco's modified Eagle's medium containing glutamine, penicillin, streptomycin, and 5% fetal bovine serum. For the preparation of total RNA, the cells were plated at a density of 5 X lo6 cells in 10-cm diameter dishes, cultured in the medium for 3 days, and then stimulated with appropriate stimuli as described in the text.
Northern Hybridization Analysis-Total cellular RNA was isolated by the guanidine isothiocyanate-cesium chloride method (30). Five micrograms of total RNA was separated on a 1% agarose, 2.2 M formaldehyde gel and subsequently blotted onto MAGNA nylon membrane with 20 X SSC by capillary transfer according to previously published methods (31). The RNA was cross-linked to the membrane by using UV cross-linker (Stratagene, La Jolla, CA). The blots were prehybridized for 8 to 12 h at 42 "C in 50% formamide, 1% SDS, 5 x SSC, 1 X Denhardt's solution (0.02% Ficoll, 0.02% bovine serum albumin, 0.02% polyvinylpyrrolidone), 0.25 mg/ml denatured salmon sperm DNA, and 50 mM sodium phosphate buffer (pH 6.5) and then hybridized with 1 X lo6 cpm/ml [~~-~~P]dCTP-radiolabeled IP-10 cDNA plasmid probe (specific activity 1 X 10Rcpm/pg DNA) prepared by nick translation at 42 "C for 16 to 24 h. After hybridization, blots were washed with 0.1% SDS, 2 X SSC for 30 min at room temperature followed by two washes at 55 "C. The blots were then exposed using XAR-5 x-ray film with screens at -70 "C.
Isolation of Genomic ZP-IO-A murine EMBL-3 genomic library was purchased from Clonetech (Palo Alto, CA). This library was prepared by partial MboI digestion of liver DNA from adult DBA/2J mouse and cloned into the BamHI site of EMBL-3. Approximately 1 X lo6 plaques were screened with a murine IP-10 cDNA probe (ll), and two positive clones were obtained. The isolated phage DNAs were digested with various restriction enzymes, subcloned into the pBluescript (Stratagene), and subjected to dideoxy sequence analysis as described below.
Nucleotide Sequence Analysis-Subfragments for sequence analysis were obtained by either restriction enzyme deletion or a nested deletion strategy using the Erase-a-Base kit (Promega). The nucleotide sequence of each of the subclones was determined by the dideoxy chain termination method on supercoiled templates using Sequenase DNA sequencing kits (U. S. Biochemical Corp.) and [cY-~*P]~ATP. Sequences were compiled using sequence analysis software (Mac-Vector, International Biotechnologies, Inc., New Haven, CT) and Gene Works (IntelliGenetics, Inc., Mountain View, CA) and compared with GenBank nucleic acid data base (GenBank release 71).
Preparation of Oligonucleotides-Oligonucleotides used in this study (Table I) were synthesized using an Applied Biosystems DNA synthesizer (model 381A). The crude oligonucleotides were purified on NAP-10 columns (Pharmacia LKB). Double-stranded oligonucleotides were prepared by annealing the complementary single strands. For the preparation of probe in the electrophoretic mobility shift assay (EMSA), double-stranded oligonucleotides were radiolabeled with the Klenow fragfnent of DNA pclymerase I and [&"P]dCTP in fill-in reaction for 5"protruding ends or with polynucleotide kinase and [y3'P]ATP for blunt ends.
Primer Extension Analysis-The position of the 5' terminus of murine IP-10 mRNA was determined by primer extension analysis using an end-labeled synthetic oligonucleotide corresponding to positions 69-85 of the murine cDNA. The radiolabeled oligonucleotide was annealed with 50 pg of total RNA from murine peritoneal macrophages stimulated with LPS (30 ng/ml) for 3 h prepared as described previously (19) in 400 mM KC1, 40 mM PIPES (pH 6.5), 80% formamide, and 1 mM EDTA for 16 h at 30 "C. The annealed primer was extended by reverse transcriptase (30 units, Boehringer Mannheim) in 50 mM Tris-C1 (pH 8.0), 6 r n~ MgC12, 10 mM DTT, 100 mM NaCl, 1 mM dNTPs, and 10 units of RNase inhibitor (Promega) for 90 min at 37 "C. The primer-extended hybrids were analyzed on a 6% polyacrylamide gel containing 7 M urea.
Construction of CAT Reporter Plasmids-A series of deletion mutants of the 5"flanking region of the IP-10 gene were constructed either by restriction endonuclease or exonuclease I11 digestion. In brief, the 5"flanking region of the IP-10 gene from -2004 to -2 (EcoRI to XhoI) was excised from a murine genomic IP-10 gene and subcloned into pBluescript (Stratagene). This vector (designated pBS IP2.0) was subsequently linearized with SmaI at the polylinker site of the vector and ligated with phosphorylated SalI linkers. After digestion with SalI, the 2.0-kilobase pair fragment was cloned into the SalI site of pCAT-Basic (Promega), which has no promoter or enhancer sequences (designated pCAT BO). The 5"deletions were site. Potential cis-acting elements in 5"flanking region are underlined. The consensus sequences of these cis-acting elements are cited from the following references: AP-1 (60), AP-2 (60), CK-1 (54), CK-2 (54), c-Re1 (59), KB (26) and -2 of the IP-10 genomic sequence were cloned into pCAT-Basic (pCAT BO). All CAT constructs (10 pg) were cotransfected with pSV-@ Galactosidase reference plasmid (5 pg) using DEAE-dextran. 24 h after transfection, the cells were stimulated with IFNy (100 units/ml) or LPS (100 ng/ml) for 18 h prior to asay for CAT activity as described under "Materials and Methods." After autoradiography, acetylated products were estimated by image analysis. A, a representative autoradiogram showing the results of CAT assays of extracts prepared from RAW 264.7 cells transfected with the reporter gene constructs illustrated in B. B, schematic representation of CAT constructs and normalized CAT activity. Potential cis-acting regulatory elements are also shown at the top of the figure. The CAT activity for each construct was normalized to @-galactosidase activity measured in the same sample. The relative CAT activity for each deletion mutants expressed as a percentage of that of pCAT B-243 stimulated with LPS. Values a t right indicate the -fold induction of CAT activity stimulated with IFNy or LPS uersw untreated but transfected cultures. The values presented are the mean from three independent experiments. constructed from the pBS IP2.0 using restriction sites (XbaI for the -930 deletion, PstI for the -676 deletion, EcoRV for the -330 deletion, BstXI for the -154 deletion, and AuaI for the -102 deletion).
The isolated fragments were ligated into the Sun site of pCAT BO. Other 5"deletions were constructed from the pCAT B-676 by exonuclease 111 digestion. All of the CAT constructs were verified by dideoxy sequencing analysis. These CAT constructs contain 19 bp of polylinker sequence (Sa& ChI, HindIII) at the 3' end of the inserted fragment.
Constructs containing the ISRE sequence in front of a heterologous promoter were made by ligating the synthetic ISRE oligonucleotide (Table I) into the BgnI site of the pCAT-Promoter (Promega) (which contains SV-40 promoter), or the BarnHI site of the pTK-CAT, which contains the Herpes virus thymidine kinase promoter (32).
Site-directed Mutagenesis-Site-specific mutants of the ISRE and the two KB binding sites of the murine IP-10 gene were created using a kit from Promega. A 242-bp HindIII fragment of pCAT B-243 (region between -2 and -243) was used as starting material. Positive clones were selected by ampicillin. Mutant clones were subjected to dideoxy sequence analysis to confirm the mutation sites and then the BO.
fragments were excised and recloned into the HindIII site of pCAT Transient Transfection-RAW 264.7 cells were seeded at a density of 3 X lo6 cells/lO-cm diameter dish 3 days prior to transfection. 10 pg of reporter CAT construct plasmid DNA and 5 pg of the plasmid pSV-@-galactosidase (containing the lac2 gene) were cotransfected by the DEAE-dextran method (Ref. 33; 300 pg/ml DEAE-dextran) for 30 min a t room temperature. The cells were incubated in Dulbecco's modified Eagle's medium containing 5% fetal bovine serum and 100 p~ chloroquine diphosphate for 90 min, washed with phosphatebuffered saline, replenished with fresh culture medium, and cultured for 24 h. 24 h after transfection, the cells were stimulated with LPS or IFNy for 18 h. CAT activity was assessed by determination of the conversion of ["C]chloramphenicol into acetylated forms detected by thin layer chromatography as described previously (34). The CAT activity for each construct was normalized to @-galactosidase activity (35) from the same sample. The relative CAT activity for each deletion mutant is expressed as a percentage of that of wild type pCAT B-243 stimulated with LPS. Preparation of Nuclear Extracts-Nuclear extracts were prepared by a modified method of Dignam et al. (36). RAW 264.7 cells were plated a t a density of 1 X 10' cells in 15-cm diameter dishes, stimulated, harvested, and resuspended in hypotonic buffer A (10 mM HEPES, pH 7.9, 10 mM KCI, 0.1 mM EDTA, 0.1 mM EGTA, 1 mM DTT, 0.5 mM PMSF) for 15 min on ice, then vortexed for 10 s with 0.6% Nonidet P-40. Nuclei were separated from cytosol by centrifugation at 12,000 X g for 60 s and were resuspended in buffer C (20 mM HEPES, pH 7.9, 25% glycerol, 0.4 M NaC1, 1 mM EDTA, 1 mM EGTA, 1 mM DTT, 0.5 mM PMSF) and shaken for 30 min a t 4 "C. Nuclear extracts were obtained by centrifugation a t 12,000 X g.

Genomic Sequence of the Murine IP-10 Gene-A clone
containing the IP-10 gene was obtained from a murine liver genomic library and subjected to dideoxy sequencing analysis. The complete nucleotide sequence of the IP-10 gene is shown in Fig. 1. Primer extension analysis revealed one potential transcriptional initiation site 65 bp upstream of ATG codon (data not shown). This start site (A) is consistent with the finding that RNA polymerase 11-dependent transcription initiates most frequently at adenine nucleotides (38). The predicted nucleotide length of the primary transcript is 2231 bases without a poly(A) tail. Alignment of the murine genomic IP-10 sequence with the human IP-10 genomic sequence indicates that the two genes are organized identically (39). Analysis of IP-10 mRNA Expression in RAW 264.7 Celh-We have previously reported that LPS, IFNr, and IFNB induce IP-10 transcription in thioglycollate-elicited peritoneal macrophages (19). To study the activity of the IP-10 promoter, various portions of the sequence flanking the transcription

TABLE I Nucleotide sequences of ISRE and KB oligonucleotides wed in this study
The numbers above the sequence refer to the distance from the transcriptional start site (Fig. 1). Lowercase letters represent the bases included for creating restriction sites. Boldface type indicates the substituted bases for mutation. Underlined sequences represent the consensus ISRE (28) and KB (26) region. R A or G; T T or C; W A or T; N any nucleotide.

-234
A. ISRE wild type -202 In RAW 264.7 cells, no constitutive expression of IP-10 mRNA was observed in unstimulated cells but was markedly enhanced in a dose-dependent manner in response to LPS or IFNy (Fig. 2). There was no cooperativity evident between LPS and IFNy; optimal doses of either agent resulted in maximal expression of IP-10 mRNA that could not be further elevated by addition of the second agent (not shown). If IFNy or LPS were provided at suboptimal doses, the combination was solely additive. These findings are consistent with previously published studies of IP-10 expression in primary peritoneal macrophages (19).

ISRE consensus
Functional Analysis of 5' cis-Regulatory Element of the IP- 10 Gene-A computer search revealed a large number of potential regulatory elements in the 5"flanking region of the murine IP-10 gene (Figs. 1 and 3B). Functional analysis of the IP-10 gene was carried out by cloning the fragment containing positions -2002 to -2 into a reporter plasmid (pCAT-BO) that lacks both enhancer and promoter sequences. This plasmid was transfected into the RAW 264.7 macrophage cell line by DEAE-dextran, and 24 h later the cells were stimulated with LPS or IFNy for 18 h and tested for CAT activity. Cultures transfected with a construct containing 2 kilobase pairs of the IP-10 5"flanking region exhibited a 9fold increase in CAT activity in response to LPS treatment but showed no significant response to IFNy (Fig. 3). Deletion of the region between -2002 and -930 resulted in constructs that showed a 3-fold induction of CAT activity in response to IFNy while remaining highly responsive to LPS. Removal of the region spanning nucleotides -930 to -676 reduced the sensitivity of the construct to LPS without affecting the sensitivity to IFNy. Constructs containing the region from -330 to -2 retained sensitivity to both stimuli. When the region from -330 to -243 was deleted both basal and inducible CAT activity increased. Removal of the region from -243 to -102 resulted in a loss of all response to LPS or IFNy. These results suggested the presence of a distal element between -2002 and -930, which negatively regulates response to IFNy, and an element between -930 and -676, which positively regulates response to LPS. The region between -243 and -102 is essential for LPS or IFNy inducibility. In order to more precisely localize the IFNy and/or LPS response sequence(s) in the region between -243 and -102, we constructed several more 5"deletion mutants and tested for the inducible CAT activity in response to LPS or IFNy. pCAT-243 and pCAT-228 gave high inducible CAT activities in response to both LPS and IFNy (Fig. 4). Deletion of 24 bases from pCAT-228 abolished response to 1FN-y. Both basal activity and LPS-inducible activity were also reduced in magnitude. These results indicate that an IFNy-response element exists in the region between -228 and -204. This region contains a highly conserved ISRE between -212 and -224 ( Fig. 1 and Table I). Deletion of nucleotides -204 to -154 had no effect on LPS-inducible CAT activity. However, sensitivity to LPS was lost when sequences between -154 and -102 were deleted. These results demonstrate that the region between -154 and -102 is necessary for LPS-inducibility.
This region contains a KB site at -113 to -104 ( Fig. 1 and Table I).
The ISRE Confers Inducibility by LPS or IFNy-We next tested whether the IP-10 ISRE sequence present at positions -212 to -224 could endow a heterologous promoter with sensitivity to IFNy or LPS. Plasmids were prepared containing one, two, or three copies of a 33-residue synthetic oligonucleotide corresponding to positions -234 to -202 (Table I) positioned just upstream of the thymidine kinase promoter (pTK-CAT plasmid). pTK-CAT is inactive without appropriate enhancer elements and was not inducible by either LPS or IFNy in RAW 264.7 cells (Fig. 5). When one copy of the ISRE sequence was placed in pTK-CAT, basal CAT activity was elevated and treatment with LPS and IFNy enhanced this 3.6-and 1.7-fold, respectively. pTK-CAT with two or three copies of the ISRE sequence enhanced CAT activity approximately 4-and %fold with LPS or IFNy. The increase in CAT activity was independent of the orientation of the ISRE sequence. Similar IFNy and LPS-inducible CAT activity was observed in constructs using the SV-40 promoter (data not shown).
LPS is known to stimulate expression of IFNa/@ by macrophages (1). Since the time period following stimulation and prior to CAT assay is long, the ability of LPS to stimulate FIG. 5. The IP-10 ISRE sequence confers LPS and IFN y inducibility on a heterologous promoter. A, one, two, or three copies of a 33-bp oligonucleotide containing the ISRE from the murine IP-10 gene (see Table I) were cloned immediately upstream of the enhancerless herpes virus thymidine kinase promoter CAT reporter plasmid (pTK 0) and transfected into RAW 264.7 cells as described in the legend to Fig. 3. The diagram a t left shows the structure of the CAT reporter constructs. The ISRE and TK promoter sequence are indicated by the filled arrow and hatched box, respectively. The number of copies and the orientation of ISRE sequence are indicated by the arrows. Values at right indicate the -fold induction of CAT activity in cells stimulated with 1FN-y or LPS relative to untreated cells. Similar results were obtained in three separate experiments. B, RAW 264.7 cells were transfected with the pTK CAT plasmid containing three copies of the IP-10 ISRE as described in the legend to Fig. 3. Following a 24-h rest the cells were either stimulated or not with LPS (100 ng/ml) in medium containing antiserum to IFNa/P (sufficient to neutralize 1300 antiviral units of IFNa/o) or an equivalent dilution of preimmune serum prior to assay of CAT activity. Similar results were obtained in two separate experiments.
CAT expression under control of the ISRE could result from the induced expression of IFNa/@. To test this possibility, RAW 264.7 cells were transfected with the pTK CAT plasmid containing three copies of the ISRE sequence and stimulated with LPS in the absence or presence of anti-serum to murine IFNa/@ or preimmune serum (Fig. 5B). Preimmune serum had no effect on the LPS-induced increase in CAT activity but anti-IFNa/@ serum markedly reduced induced expression without affecting the levels of basal or constitutive activity.

ISRE and KB Elements Cooperate in the IP-10 Promoter-
T o examine whether the ISRE and two KB sites cooperate in mediating response to IFNy or LPS, specific residues in these elements known to be critical for transacting factor recognition were changed by site-directed mutagenesis. Specific mutants utilized are shown in Fig. 6B. Three nucleotide mutations in the ISRE sequence of the -243 CAT construct abolished the IFNy-inducible CAT. This mutation also re-sulted in a reduced magnitude for basal and LPS-induced CAT activity although the ratio of induced to basal activity remained high (Fig. 6, pCAT Mu3). The mutation of either K B~ or K B~ sites individually diminished both the LPS-inducible and basal CAT activities; however, mutation of the K B~ site (pCAT Mul) more effectively reduced the response to LPS than did mutation of the K B~ site (pCAT Mu2). Similar quantitative modulation of IFNy-inducible CAT activity was reproducibly observed in transfection experiments utilizing plasmids containing an intact ISRE but mutated either in K B~ or KBZ. Interestingly, mutation of both KB sites (pCAT Mul-2), which retains the wild type ISRE sequence, abolished the IFNy-induced CAT activity. These data indicate that the KB-binding sites cooperate with the ISRE for efficient activation of IP-10 transcription by IFNy. The combination of the mutations in the ISRE and K B~ site (pCAT Mu2-3) resulted in a loss of LPS inducibility, suggesting the possibil- ity that sequences between -243 and -154 may have some negative effect on response to LPS, since deletion of residues -243 to -154, which contains the K B~ sequence, was sensitive to LPS (see Fig. 4, -154 CAT construct). No inducible activity was observed in the CAT constructs containing mutations in the ISRE and one or both KB sites.

IFNy and LPS Activate Nuclear Factors Binding the ISRE
and KB Sequences-In order to determine if IFNy and/or LPS could induce DNA-binding factor(s) that specifically recognize the IP-10 ISRE or KB sites, double-stranded synthetic oligonucleotides corresponding to the appropriate residues were prepared. An oligonucleotide spanning the sequence from -234 to -202 (Table I) was radiolabeled with [32P]dCTP and used in EMSA with nuclear extracts prepared from untreated LPS-or IFNy-treated RAW 264.7 cells (Fig. 7). Nuclear extracts from untreated cells displayed a major, rapidly migrating complex. When the cells were stimulated with IFNy or LPS, a more slowly migrating complex was observed (see arrow). This inducible binding activity was observed as early as 30 min after stimulation with IFNy but was seen in  (Table I).
Similar results were obtained in two separate experiments.
LPS-stimulated cells only following 3 h. Several additional minor binding complexes were observed that did not exhibit reproducible stimulus dependence. The binding specificity of the principle stimulus-dependent complex was demonstrated by competition studies with unlabeled oligonucleotides containing the intact ISRE sequence, the mutant ISRE, or unrelated sequences (not shown). Analysis of K B~ binding activity used a 30-mer corresponding to residues -98 to -123 while the oligonucleotide corresponding to the K B~ site was a 28-residue sequence containing residues -151 to -178 (Table I). Nuclear extracts from untreated cells formed two complexes with the K B~ sequence (Fig. 8). When cells were stimulated with IFNy for 2 h,. the levels of complex 2 moderately increased. A much more marked increase in the formation of both complexes 1 and 2 was observed in LPS-stimulated cells. Both IFNy-and LPSinducible complexes could be specifically competed with wild type K B~ oligonucleotide (Fig. 8 B ) while the mutant K B~ oligonucleotide (Table I) did not compete. Interestingly, when the K B~ sequence was used as a probe, complexes C1 and C2 were detectable at lower levels in untreated cells. IFNytreatment caused a modest increase in complex 2 and LPS treatment strongly induced the formation of complex C2 and an additional slowly migrating complex termed C3.

DISCUSSION
Expression of the IP-10 gene is inducible by diverse proinflammatory stimuli including LPS, IFNy, IFNaIP, PDGF, tumor necrosis factor a, and IL-1 in a variety cell types (19)(20)(21). 2 The primary goal of the work presented in this report was to elucidate how such different extracellular signals regulate the transcriptional activation of the IP-10 gene in the same cell population. Structural and functional characterization of the sequence region flanking the transcription start site of the murine IP-10 gene demonstrates that approximately -228 bases upstream from transcription start site are essential for IFNy inducibility and -154 bases upstream are requisite to LPS sensitivity. Although IFNy-and LPS-induced transcriptional activation of the IP-10 gene are both influenced by the ISRE and KB sequences, these elements function in distinct ways for each stimulus. These conclusions are based upon the following observations. 1) Deletion of the ISRE sequence from the -228 nucleotide fragment of IP-10 promoter construct abolished the sensitivity to IFNy without affecting sensitivity to LPS. 2) Deletion of both KB sequences from the -154 nucleotide fragment abolished sensitivity to LPS. 3) IFNy induced a nuclear factor that specifically recognized and bound an ISRE containing oligonucleotide within 30 min of stimulation while LPS-induced ISRE binding activity required a minimum stimulation time of 3 h. 4) Although both IFNy and LPS could drive transcription from constructs containing multiple copies of the IP-10 ISRE linked to a heterologous promoter, the response to LPS (but not IFNy) was blocked by including antibody to murine IFNaIP. These findings suggest that IFNy and LPS act via distinct (though perhaps overlapping) intracellular signaling pathways. This contention is supported by experimental findings demonstrating the possibility of additional regulatory elements distal to the 243 bp fragment which are functionally distinct. A negative element between -2002 and -930 suppresses response to IFNy without affecting response to LPS while a positive element between -930 and -630 enhances response to LPS but not to IFNy. These findings are consistent with many previous reports suggesting differences in the signaling responses of macrophages to IFNy or LPS (4,6,7,19).
Although the ISRE and the KB sequences, respectively, are essential for IFNy and LPS inducibility in the IP-10 gene promoter, one important characteristic of the transcriptional activation of the IP-10 gene induced by 1FN-y or LPS is the cooperative interaction between the ISRE and KB elements. The following observations indicate that the efficient transcriptional activation of the IP-10 gene requires the combined interaction of a t least two positive regulatory elements. First, LPS-induced transcriptional activation was completely abolished by the combination of mutation in one KB site (either K B~ or K B~) and in the ISRE or in both KB sites. Thus the ISRE and one KB site or two KB sites are necessary for LPS inducibility. Second, mutation of one KB site reduced or abolished the sensitivity to IFNy despite of the presence of an intact ISRE sequence. Thus the ISRE and a KB site are both necessary for IFNy inducibility. Since the ISRE sequence itself was able to confer IFNy inducibility to heterologous promoter, it is conceivable that another sequence motif between -243 and -154 may play some repressive role which could be overcome by an IFNy-inducible factor(s) binding to either of the KB sites. The possibility of a negative element in * Y. Ohmori   10 pg of nuclear extract protein was incubated with 0.5 ng of radiolabeled K B~ (lanes [1][2][3] or KB 2 (lanes 4-6) oligonucleotide (Table I)  the region between -154 and -243 is also supported by the finding that deletion of this region does not affect LPS sensitivity (Fig. 4), whereas mutation of the ISRE and one KB site abolish sensitivity (Fig. 6).
The 33-base sequence of the region between -234 and -202 contains a conserved ISRE sequence that confers responsiveness to IFNy and LPS in the context of a heterologous promoter. This result indicates that the IP-10 ISRE sequence functions as both IFNy and LPS responsive enhancer element. Work in multiple cultured cell systems has identified a diverse array of factors that can specifically interact with ISRE-like sequence elements (28,29,(40)(41)(42)(43)(44)(45)(46). For example, interferon-stimulated gene factor 3 is an oligomeric complex that may be activated following stimulation with type 1 IFN (28,41). Additional ISRE-binding proteins induced in response to IFN have been described (these include interferonstimulated gene factor 2 (41), interferon regulatory factor -1 (IRF-1, Ref. 42), IRF-2 (43), interferon consensus sequence-binding protein (44), and interferon binding factor 1 (45)). A number of additional ISRE binding activities, which are not responsive to IFN treatment, have also been identified (46). The relationship between the macrophage-derived complexes and those described previously in other cell types cannot be determined on the basis of the available data and will require analysis of the precise protein composition of the complexes themselves. The IFNy-inducible ISRE binding complex can be detected as early as 30 min after IFNy treatment and was independent of protein synthesis. This result suggests that induction of ISRE binding is a primary response to IFNr and involves the activation of preexisting cellular factor(s).
In contrast, the LPS-stimulated ISRE binding activity appeared only after 3 h of stimulation. Since LPS can directly induce IFNa//3 in macrophages, the involvement of the ISRE in mediating response to LPS may be indirect and depend upon the intermediate synthesis and secretion of IFNalj3. Indeed this possibility is supported by the finding that anti-