Role of a Distal Enhancer Containing a Functional NF- (cid:107) B-binding Site in Lipopolysaccharide-induced Expression of a Novel (cid:97) 1 -Antitrypsin Gene*

(cid:97) 1 -Antitrypsin ( (cid:97) 1 -AT) is one of the major proteinase inhibitors in serum. Its primary physiological function is to inhibit neutrophil elastase activity in lung, but it also inhibits other serine proteases including trypsin, chymotrypsin, thrombin, and cathepsin. We have previ- ously reported a novel (cid:97) 1 -AT, S-2 isoform, from rabbit that is induced up to 100-fold in the liver during acute inflammatory condition (Ray, B. K., Gao, X., and Ray, A. (1994) J. Biol. Chem. 269, 22080–22086). Here, we present evidence that the expression of this (cid:97) 1-AT S-2 gene is also induced in lipopolysaccharide (LPS)-treated pe- ripheral blood monocytes. From the cloned genomic DNA, we have identified a distal LPS-responsive en- hancer located between (cid:50) 2438 and (cid:50) 1990 base pairs upstream of the transcription start site. In vitro DNA- binding studies demonstrated an interaction of an LPS-inducible NF- (cid:107) B-like nuclear factor with a (cid:107) B-element present in this enhancer region. Antibodies against p65 and p50 subunits of NF- (cid:107) B supershifted the DNA-protein complex. A mutation of the NF- (cid:107) B-binding element virtually abolished the LPS-responsive induction of the chimeric promoter in monocytic cells. Furthermore, overexpression of NF- (cid:107) B induced the wild-type promoter activity. Taken together, these results demon- strated that during LPS-mediated inflammation, NF- Rel family of transcription factors play a crucial role in the transcriptional induction of the inflammation responsive 1 -AT gene.

␣ 1 -Antitrypsin (␣ 1 -AT) is one of the major proteinase inhibitors in serum. Its primary physiological function is to inhibit neutrophil elastase activity in lung, but it also inhibits other serine proteases including trypsin, chymotrypsin, thrombin, and cathepsin. We have previously reported a novel ␣ 1 -AT, S-2 isoform, from rabbit that is induced up to 100-fold in the liver during acute inflammatory condition (Ray, B. K., Gao, X., and Ray, A. (1994) J. Biol. Chem. 269, 22080 -22086). Here, we present evidence that the expression of this ␣1-AT S-2 gene is also induced in lipopolysaccharide (LPS)-treated peripheral blood monocytes. From the cloned genomic DNA, we have identified a distal LPS-responsive enhancer located between ؊2438 and ؊1990 base pairs upstream of the transcription start site. In vitro DNAbinding studies demonstrated an interaction of an LPSinducible NF-B-like nuclear factor with a B-element present in this enhancer region. Antibodies against p65 and p50 subunits of NF-B supershifted the DNA-protein complex. A mutation of the NF-B-binding element virtually abolished the LPS-responsive induction of the chimeric promoter in monocytic cells. Furthermore, overexpression of NF-B induced the wild-type promoter activity. Taken together, these results demonstrated that during LPS-mediated inflammation, NF-B/ Rel family of transcription factors play a crucial role in the transcriptional induction of the inflammation responsive ␣ 1 -AT gene.
␣ 1 -Antitrypsin (␣ 1 -AT), 1 also known as ␣ 1 proteinase inhibitor, is one of the major protease inhibitors in plasma. ␣ 1 -AT has a broad range of activities (Laurell and Jeppsson, 1975) but mainly protects the elastic fibers in lung alveoli from excessive digestion by neutrophil elastase (Olsen et al., 1975;Gadek et al., 1980). The importance of this particular function was first proposed by Laurell and Eriksson (1963), after observing that genetically ␣ 1 -AT-deficient individuals often develop either a degenerative lung disease early in life (Eriksson, 1964) or a liver disease (Sharp et al., 1969). Genetic ␣ 1 -AT deficiency seems to affect 1 in 2000 Europeans and is manifested by emphysema in adults and liver disease in children. The genetic variants associated with the deficiency of ␣ 1 -AT are termed as P i s , P i z and P i null . Individuals homozygous for the P i s allele exhibit a deficiency level of 60%; those homozygous for the P i z allele exhibit 10 -15% deficiency; and P i null individuals show almost no detectable amount of ␣ 1 -AT in their serum (Gitlin and Gitlin, 1975;Allen et al., 1974;Muensch et al., 1986). Increased expression of ␣ 1 -AT also have clinical importance in the case of individuals carrying P i zz ␣ 1 -AT mutant allele. A glutamate to lysine substitution at position 342 (Nukiwa et al., 1987;Sifers et al., 1988) results in aggregation of the mutant ␣ 1 -AT protein within the rough endoplasmic reticulum (Lomas et al., 1992) of hepatocytes and forms insoluble intracellular inclusions which lead to the hepatocellular damage in these patients. There is now strong evidence that the liver disease of the P i zz homozygotes is a direct consequence of ␣ 1 -AT accumulation and degradation in the hepatocyte (Carlson et al., 1988). Since inflammatory condition increases the synthesis of ␣ 1 -AT protein severalfold, prevention of inflammation and pyrexia could considerably lower the accumulation of the defective protein in the liver of a homozygous P i zz patient. ␣ 1 -AT is primarily synthesized in the liver (Laurell and Jeppsson, 1975), to a lesser extent in bronchoalveolar and breast milk macrophages, and in blood monocytes (Perlmutter et al., 1985). It is also reportedly expressed at lower levels in submandibular glands (Chao et al., 1990), renal tubular epithelial cells, intestinal epithelial cells, nonparietal cells of gastric mucosa, and pancreatic islet cells (Carlson et al., 1988;Kelsey et al., 1987;Koopman et al., 1989). The plasma level of ␣ 1 -AT in a variety of species is normally 2-4 mg/ml, but levels increase 3-4-fold under inflammatory conditions, pregnancy, and after administration of synthetic androgen danazol (Laurell and Rannevik, 1979).
Due to the clinical importance, molecular basis of the regulation of ␣ 1 -AT gene expression has been intensively studied. The minimal promoter element required for liver specific basal expression of human ␣ 1 -AT gene was reported to be confined within 261 nucleotides from the transcription start site (Cilliberto et al., 1985;DeSimone et al., 1987). But the mouse ␣ 1 -AT gene which is structurally similar to the human gene requires 500 bases upstream of the transcription start site for its maximal basal expression (Grayson et al., 1988). The difference between the enhancers of these two genes is very surprising. Also, the presence of negative elements that are active in liver cells is reported for both human (DeSimone and Cortese, 1989) and mouse ␣ 1 -AT genes (Montgomery et al., 1990). Despite numerous reports on ␣ 1 -AT promoter and its liver-specific expression, very little is known regarding the mechanisms or regulatory elements that control inducible expression of ␣ 1 -AT during inflammatory condition. It is only known that, in liver, ␣ 1 -AT is induced by IL-6, whereas in monocytes and macrophages it is induced by both IL-6 and bacterial lipopolysaccha-ride (Perlmutter et al., 1989;Barbey-Morrel et al., 1987). During IL-6-induced expression, human ␣ 1 -AT promoter utilizes a distal transcriptional initiation site located about 2 kb upstream of the hepatocyte-specific transcription start site (Hafeez et al., 1992). A lack of reports on the mechanisms of the inducible expression of ␣ 1 -AT is possibly due to its moderate level of induction which makes it less conducive for such regulation studies. Our recent studies have identified a highly inducible isoform of ␣ 1 -AT in rabbit whose expression is increased about 100-fold in response to inflammatory signals . Rabbit genome contains multiple ␣ 1 -AT genes that are significantly different in terms of their expression under inflammatory condition. In response to inflammation, the mRNA level of the two members of this family, F and S-1, increases nominally (1.5-fold), whereas expression of the S-2 isoform increases about 100-fold. Thus S-2 isoform becomes a major component of ␣ 1 -AT in the rabbit under inflammatory conditions . We have therefore initiated a study to understand the regulation of the inducible expression of the rabbit gene as a prelude to the establishment of a rabbit model system to study the pathology associated with the overexpression and deficiency of ␣ 1 -AT.
To identify the elements required for transcriptional induction of the rabbit ␣ 1 -AT S-2 gene, we performed a detailed analysis of the S-2 promoter region. In this report, we show that a distal region located about 2.5 kb upstream of the transcription start site confers the LPS inducibility to the S-2 promoter. We have identified one NF-B-like-binding element (Sen and Baltimore, 1986) at this region. Mutation of the NF-B site virtually eliminates LPS responsiveness in a reporter construct containing the rabbit S-2 upstream promoter. Further characterization showed that several members of the NF-B family can bind to this NF-B enhancer motif. Also, cotransfection of NF-B induces expression of the reporter CAT construct. These results provide strong evidence that NF-B might be involved as a major regulator in the induction of ␣ 1 -AT gene brought about by LPS-mediated inflammation.

MATERIALS AND METHODS
Preparation of Rabbit Monocyte Cells-Rabbit peripheral blood monocytes were isolated from citrated blood obtained from normal healthy male New Zealand white rabbits. The citrated blood was centrifuged at 300 ϫ g for 20 min and 800 ϫ g for another 20 min to sediment the cells. The buffy coat layer was removed and washed three times by centrifugation using RPMI 1640 medium supplemented with L-glutamine and gentamycin (50 g/ml). Washed mononuclear cells were resuspended in RPMI 1640 medium plus 5% heat-inactivated serum and allowed to adhere at 37°C for 2 h. Next, the nonadherent cells were removed by vigorous washing with warm RPMI 1640 medium. The adherent cells were cultured in RPMI 1640 supplemented with 5% heat-inactivated serum at 37°C, 5% CO 2 . For LPS stimulation, the cells were treated with 10 g of LPS/ml of medium at 37°C, 5% CO 2 for different times as indicated in the text and figure legends.
RNA Isolation and Northern Blot Analysis-Total RNA was isolated from uninduced and 5-h LPS-induced blood monocyte cells as described by Chomczynski and Sacchi (1987). Fifty micrograms of total RNA were fractionated on a 1% agarose gel containing 2.2 M formaldehyde and transferred onto a nylon membrane. Hybridization was carried out using a 32 P-labeled rabbit ␣ 1 -AT S-2 cDNA as probe. S1 Nuclease Protection Assay-Ten micrograms of total RNA isolated from uninduced and LPS-induced monocytes were incubated with a molar excess of 5Ј-end-labeled antisense S-2-specific oligonucleotide in 40 l of hybridization buffer (20 mM Tris-HCl, pH 7.4, 400 mM NaCl, 1 mM EDTA, 10 mM dithiothreitol, 80% formamide) at 75°C for 15 min (Sambrook et al., 1989). The sequence of the S-2 antisense oligonucleotide is 5Ј-GGGCATAGAATAGAGTAC-3Ј, as described previously . Next, the tubes were transferred quickly to a water bath set at 50°C and incubated overnight. Annealed RNA and oligonucleotide was precipitated with ethanol and 0.1 volume of 3 M sodium acetate, pH 5.5. The pellet was washed twice with 80% ethanol and vacuum dried. The dried pellet was dissolved in 10 l of S1 nuclease buffer (300 mM NaCl, 60 mM sodium acetate, pH 4.5, 3.4 mM ZnCl 2 , 2 g of carrier DNA) and incubated at 15°C in the presence of 1.2 units of S1 nuclease. The optimum concentration of S1 nuclease and time of incubation were determined by initial pilot experiments. The reaction was stopped by adding 10 mM EDTA and 5 l of loading buffer containing 8 M urea. After heating at 75°C for 5 min, the reaction mixture was electrophoresed in a 10% polyacrylamide, 8 M urea sequencing gel.
Isolation of Rabbit ␣ 1 -AT S-2 Genomic Clone-A rabbit liver EMBL3 genomic library (Ray and Ray, 1991) was screened by the hybridization technique using a radiolabeled probe consisting of the cDNA sequence of rabbit ␣ 1 -AT S-2 protein. Five independent positive clones were identified using this probe. In order to determine which of these five clones contains the S-2-specific genomic sequence, first DNA fragments carrying the reactive center regions of all five clones were identified by hybridization analysis and were subsequently cloned. The DNA sequence of S-2 isoform in this region is unique and distinguishable from the other isoforms . From sequence analysis using a primer downstream of the reactive center region, the right genomic clone containing the S-2 gene sequence was identified. Next, a restriction map of the S-2 genomic clone was generated by single and double digests with various restriction enzymes and Southern hybridization.
DNA Sequence Analysis-Regions of phage DNA spanning the ␣ 1 -AT gene were subcloned and sequenced by dideoxynucleotide chain termination method (Sanger et al., 1977) using a Sequenase sequencing kit (U. S. Biochemical Corp.).
Nuclear Extracts, Oligonucleotides, and Electromobility Shift Assay (EMSA)-Nuclear extracts were prepared from uninduced and 4-h LPS (10 g/ml)-induced blood monocytes by following the method of Dignam et al. (1983). Protein concentrations were measured by the method of Bradford (1976). DNA-binding assays were carried out following a standard protocol described earlier . The labeling of DNA probe, containing sequences from Ϫ2295 to Ϫ2250 of ␣ 1 -AT S-2 gene, was performed by filling in the overhangs at the termini by Klenow enzyme and [␣-32 P]dATP. In some binding assays, competitor oligonucleotides were added in the reaction mixture prior to the addition of the radiolabeled probe. For antibody interaction studies, a mixture of anti-p50 and anti-p65 NF-B proteins were added prior to the addition of the probe. The DNA-protein complexes were resolved in 6% native polyacrylamide gels.
A palindromic NF-B oligonucleotide sequence used as a competitor in EMSA is 5Ј-GATCCATGGGGAATTCCCCATG-3Ј. For self-annealing, the oligonucleotide was heated at 95°C for 2 min in 50 mM Tris, pH 7.4, 60 mM NaCl, 1 mM EDTA and allowed to cool slowly to room temperature for 2-3 h. A mutant NF-B oligonucleotide was also used as a competitor in EMSA, and its sequence is 5Ј-GATCCATCTCGAAT-TCGAGATG-3Ј. Underlined bases represent mutated sequences.
Plasmid Constructs, Cell Cultures, and Transfections-pAT-3.5 CAT plasmid was prepared by ligating the 5Ј-flanking and exon 1 sequence (from positions Ϫ3.5 kb to ϩ70) of the rabbit ␣ 1 -AT S-2 gene into a promoterless plasmid vector pBLCAT3 (Luckow and Schutz, 1987). Several deletion constructs containing the ␣ 1 -AT S-2 gene promoter sequences from Ϫ2438 to ϩ70, Ϫ1990 to ϩ70, and Ϫ619 to ϩ70 were also cloned into a pBLCAT3 vector. A plasmid with a mutated NF-Bbinding site, located between bp Ϫ2287 and Ϫ2276, was created by site-directed mutagenesis (using a kit from Promega) of the wild-type ␣ 1 -AT S-2 (Ϫ2438/ϩ70) CAT plasmid. The sequence of the mutated NF-B element is 5Ј-GAGCTCCTTTCGAG-3Ј, where underlined bases represent mutated sequences. The resulting constructs were analyzed by DNA sequencing to determine the orientation and authenticity of the insert DNAs.
The BNL liver cell (BNL CL.2) used in the transfection assays (ATCC TIB73) is a normal embryonic liver cell line. These cells were cultured in Dulbecco's modified Eagle medium containing a high level of glucose (4.5 g/liter) and supplemented with 10% fetal bovine serum. Ten g of reporter plasmid were used in each transfection assay, with 2 g of pSV-␤-gal plasmid (Promega) as a control for measuring transfection efficiency. In cotransfection experiments, various amounts of CMV-p65 and CMV-p50 plasmid DNAs were added (indicated in figure legends) together with reporter plasmid DNAs and carrier plasmid DNA so that the total amount of DNA in each transfection assay remained constant. Transfection of liver cells was carried out by the calcium phosphate method (Graham and Van der Eb, 1973) with minor modifications as described in . Monocytes used in transfection assays were prepared from rabbit peripheral blood as described above. Monocytes were transfected using the DEAE-dextran method (Sambrook et al., 1989). For LPS-stimulation of transfected monocytes, cells were incubated in the presence of 10 g/ml Escherichia coli LPS (Sigma). As a measure of monitoring transfection efficiency, cells were cotransfected with pSV-␤ galactosidase (Promega) plasmid. Chloramphenicol acetyltransferase (CAT) was assayed (Sambrook et al., 1989) using ␤-galactosidase-equivalent amounts of cell extracts which were heated at 60°C for 10 min to inactivate endogenous acetylase. All transfections were repeated at least three times.

LPS Stimulates S-2 Gene Expression in Blood Monocytes-
Earlier studies showed that the ␣ 1 -AT S-2 gene is highly induced in acute phase liver . To examine, if expression of S-2 gene is also induced in monocyte cells, we used Northern blot analysis of total cellular RNA isolated from peripheral blood monocytes following 5 h of LPS stimulation. Results in Fig. 1A demonstrate that LPS mediates an increase in ␣ 1 -AT gene expression in peripheral blood monocytes. To determine specifically the expression of the S-2 isoform of ␣ 1 -AT genes, we performed S1 nuclease analysis (Fig. 1B). An antisense 18-mer oligonucleotide specific to the reactive center region of S-2 isoform was 32 P-labeled at the 5Ј end by [␥-32 P]ATP and T 4 polynucleotide kinase. Molar excess of this oligonucleotide was hybridized with total cellular RNA prepared from normal and LPS-induced blood monocytes. After hybridization, the RNA-oligonucleotide duplex was briefly digested with S1 nuclease and electrophoresed in a sequencing polyacrylamide gel in which the protected duplex fragment migrates slower than the digested oligonucleotide probe. It is evident that expression of the S-2 gene is significantly induced in monocyte cells in response to LPS treatment (compare lanes 3 and 4). The control reaction using yeast tRNA (lane 1) indicated absence of any nonspecific hybridization demonstrating the specificity of this assay.
Organization of the Rabbit ␣ 1 -Antitrypsin S-2 Gene-To understand the mechanism of induction of this protein, we undertook the cloning and characterization of the rabbit ␣ 1 -AT S-2 gene. Using a rabbit ␣ 1 -AT cDNA as probe for screening of a EMBL3 rabbit genomic library, five independent hybridizing clones were detected. Previous studies  indicated that in rabbit at least three isoforms of ␣ 1 -AT gene exist. In order to know which of these five genomic clones code for the highly inducible S-2 isoform, we used the following strategy. Although the overall cDNA sequences of the three ␣ 1 -AT isoforms are 96% homologous, a striking mismatch is present at the reactive center region . Thus, we first sought to identify the DNA fragments carrying the reactive center region of the five genomic clones. This was accomplished by hybridization of the BamHI fragments of the genomic clones using a short DNA fragment containing sequences around the reactive center region as a probe. Next, these BamHI fragments were subcloned and sequenced for further characterization. On the basis of the sequence similarity at the reactive center region, one -phage clone was positively identified to be carrying the gene coding the S-2 isoform of the ␣ 1 -AT protein. This phage clone contains an insert of about 19 kb in size. Southern blot hybridization and DNA sequence analyses with exon-specific oligonucleotides indicated that the rabbit S-2 gene is composed of five exons separated by four introns. These results are summarized in Fig. 2. The arrangements of exons and introns in the S-2 gene are very similar to those in the human and mouse ␣ 1 -AT genes (Long et al., 1984;Krauter et al., 1986). Inspection of the sequences at splice junctions of the S-2 gene showed that they all contain the correct conserved sequence at the 5Ј-splice donor and at the 3Ј-splice acceptor sites (Sharp, 1987).
The 5Ј-Flanking Region of ␣ 1 -AT S-2 Gene Contains Potential Regulatory Elements-The sequences of the 5Ј-flanking promoter region of the S-2 gene is presented in Fig. 3. The transcription start site of this gene is known from a previous primer extension analysis . The sequence of the first 47 nucleotides of the 51-nucleotide 5Ј-untranslated region ob-FIG. 1. LPS stimulation of ␣ 1 -AT S-2 gene expression in monocytes. A, Northern analysis of the total ␣ 1 -AT mRNA isolated from rabbit peripheral blood monocytes grown in the absence (lane 1) or in the presence (lane 2) of LPS for 5 h. Total RNA samples were fractionated in a 1% agarose gel containing 2.2 M formaldehyde according to the method described by  and hybridized to a 32 P-labeled ␣ 1 -AT cDNA probe. As a control, the membrane was reprobed with an actin cDNA probe that reveals the qualitative and quantitative estimation of the two samples. B, result of an S1 nuclease assay (described under "Materials and Methods") that determines the level of ␣ 1 -AT S-2 transcripts in the mRNA preparations from monocytes grown in the absence (lane 3) or presence of LPS (lane 4). As a negative control yeast transfer RNA (lane 1) was used. Lane 2 contains untreated 32 P-labeled S-2 isoform-specific antisense oligonucleotide only. The S 1 -protected fragments were separated on a 10% polyacrylamide, 8 M urea sequencing gel and visualized by autoradiography.
FIG. 2. Physical map of the ␣ 1 -AT S-2 gene of rabbit. A genomic clone, ␣ 1 -AT12, containing a 19-kb insert encompass the S-2 gene. Its identity as S-2 isoform-specific was determined by Southern hybridization using S-2 gene-specific probe. Location of the gene within a 10-kb region spanning BamHI and SalI restriction sites was determined by Southern hybridization analysis with the ␣ 1 -AT S-2 cDNA probe , and a restriction map was generated. The relative position and size of the exons, identified by Roman numerals, were determined by DNA sequence analysis of the exons and the surrounding intronic regions. Transcription start site is indicated by an arrow.
tained from S-2 cDNA matched completely with the exon 1 sequence of S-2 genomic DNA. The remaining 4 nucleotides of the untranslated region and the translation initiator ATG codon are located in exon 2 (not shown). In addition to the TATA box for the binding of transcription factor TFIID, several other transcription factor binding sites, including those for HNF1/LF-B1 (Frain et al., 1989;Costa et al., 1989), HNF3 (Grayson et al., 1988), HNF2/LF-A1 (Rangan and Das, 1990), and C/EBP (Johnson et al., 1987;Costa et al., 1989;DeSimone et al., 1987) are located. Several Sp1 binding GC boxes 5Ј-GGGCGG-3Ј (Kadanoga et al., 1988;Gidoni et al., 1984) are present in direct or inverse orientation at the proximal promoter (within Ϫ570 bp) region. At the distal region about 2400 bp upstream of the transcription start site, a potential NF-B (Sen and Baltimore, 1986)-binding site is identified. Also, several APRF binding motifs (Wegenka et al., 1993) are seen at the distal region. The sequence of the 5Ј-flanking region and exon I of the rabbit S-2 gene was compared and aligned with the human (Long et al., 1984) sequence (Fig. 4). It is evident that a high level of homology (ϳ67%) exists between these genes at the proximal region (sequences up to Ϫ200 from ϩ1). However, further upstream of this region, at sequences between Ϫ550 and Ϫ200, the overall homology drops to less than 55%. Interestingly, the sequences identified as the DNA-binding elements of HNF1/LF-B1, HNF2/LF-A1, C/EBP, and HNF3 transcription factors, located within the proximal region, are highly conserved between the species, indicating the importance of these elements and the cognate transcription factors in controlling ␣ 1 -AT gene expression. The divergence at the distal promoter region is quite intriguing.
The 5Ј-Flanking Region of the S-2 Gene Has Inducible Promoter Activity in Liver and Monocytic Cells-A 3.5-kb DNA fragment containing the upstream 5Ј-flanking region and the first exon, from nucleotides Ϫ3551 to ϩ70 was cloned in the right orientation into the plasmid vector pBLCAT3 (Luckow and Schutz, 1987). This vector does not contain a functional promoter and is entirely dependent upon the functional promoter activity of the ligated heterologous gene sequence. The recombinant reporter gene (pAT-3.5 CAT) was transiently transfected into BNL CL.2 liver and rabbit peripheral blood monocyte cells to analyze its promoter activity. As a control, we also used the parent plasmid pBLCAT3. Transfected liver cells were grown in the absence and presence of 25% conditioned medium, a source of active cytokines obtained from LPS-induced monocyte cells (Ray et al., 1993). Transfected monocytes were induced using 10 g/ml Escherichia coli lipopolysaccharide (Sigma) in the incubation medium. Forty-eight hours after transfection, cells were harvested, and CAT activity was measured. As an internal control for transfection efficiency, cells were cotransfected with a second plasmid pSV-␤ galactosidase (Promega). Results presented in Fig. 5 show that pAT-3.5CAT is expressed at a low level in both BNL liver and monocyte cells. However, the promoter activity is highly induced when BNL cells are treated with conditioned medium. A similar level of induction was seen in LPS-treated (10 g/ml) monocytes. These data suggest that the DNA sequences present in the 5Ј-flanking region (Ϫ3551/ϩ70) of the rabbit ␣ 1 -AT S-2 gene contain elements necessary for induction of reporter gene expression in both liver and monocyte cells under inflammatory condition.
Identification of a Distal LPS-responsive Enhancer Region-To further define the regulatory regions required for LPS-mediated ␣ 1 -AT gene induction, a series of deletion constructs containing variable lengths of upstream promoter sequences were made (Fig. 6). These plasmids were transiently transfected in rabbit monocyte cells, grown for 24 h with or without LPS, and CAT activity was measured. Results presented in Fig. 6 show that the construct containing a 3.5-kb upstream promoter sequence is highly responsive to LPS stimulation. Deletion of sequences up to Ϫ2438 has little effect on the LPS responsiveness of the ␣ 1 -AT S-2 promoter. However, removal of sequences up to Ϫ1990 bp resulted in a severe drop in the promoter inducibility by LPS. Further deletion of up-stream sequence up to Ϫ619 bp caused a modest increase in the basal transcription which is independent of LPS. This is possibly due to the removal of a negative element which might be located upstream of the Ϫ619-bp position. Identity of this putative element is yet to be determined. Together, these data demonstrated that an LPS-dependent transcriptional enhancer element is located between the PstI sites at position Ϫ2438 and Ϫ1990.
The LPS-responsive Region of ␣ 1 -AT S-2 Promoter Contains a Functional NF-B-binding Site-To probe further into the nature of the regulatory elements present in the LPS-responsive region of the ␣ 1 -AT S-2 promoter, the sequence of this region was examined. As shown in Fig. 3, a NF-B-binding element is seen to be present in this region. To test whether this element is an active binding site for transcription factor NF-B, we performed EMSA. Nuclear extracts prepared from unstimulated and LPS-stimulated monocyte cells (4 h) were analyzed by electromobility shift assay using a probe spanning the S-2 promoter from Ϫ2295 to Ϫ2250 which contains an NF-B element. Two DNA-protein complexes were seen only with induced nuclear extract (Fig. 7A, lane 2). The complexes were competed by molar excess of an oligonucleotide containing NF-B binding element (Fig. 7B, lane 2). Inability of a mutant oligonucleotide to compete for the complex formation (Fig. 7B,  lane 3) indicates specificity of the NF-B binding. To further verify that the factor binding to the S-2 promoter is indeed a member of the NF-B family, we used specific antibodies to NF-B proteins in EMSA. Anti-p65 and anti-p50 antibody effectively supershifted the DNA-protein complexes (Fig. 8). Activation and participation of p50 subunit of NF-B is evident from the results in lanes 2 and 3. Likewise, involvement of p65 subunit of NF-B is evident from the results in lanes 4 and 5. These data indicated that both the p50 and p65 subunit of the NF-B family can interact with the B element of S-2 gene, located at position Ϫ2287 and Ϫ2276 in different combinations to form the complexes A and B.
Mutation of the NF-B Site in S-2 Promoter Diminishes LPS FIG. 4. Comparison of the promoter region of rabbit ␣ 1 -AT S-2 gene and human ␣ 1 -AT gene. S-2 gene sequences from ϩ47 to Ϫ964 were used for comparison. Symbols: q, identical nucleotides; -, absence of nucleotides; R, rabbit, and H, human sequence. The position of the transcription start site is indicated by an arrow.
FIG. 5. Functional analysis of the promoter region of ␣ 1 -AT S-2 gene. A DNA fragment containing sequences from a BamHI (Ϫ3551) to an ApaI (ϩ70) of the ␣ 1 -AT S-2 gene was subcloned in right orientation into plasmid vector pBLCAT3. The resultant plasmid, pAT-3.5 CAT, was used in transfection of liver and monocyte cells to assess the inducibility of the ␣ 1 -AT S-2 gene promoter DNA. BNL CL.2 liver and peripheral blood monocyte cells were transfected separately with pBLCAT3 (as a control) and pAT-3.5 CAT plasmid DNA (10 g of each) in the presence and absence of 25% conditioned medium or LPS (10 g/ml). Details of the transfection and CAT assay are described under "Materials and Methods." CAT activity was measured as radioactivity (counts/min) in [ 14 C]chloramphenicol acetate produced by an equivalent amount of transfected cell extracts. Transfections were normalized to the ␤-galactosidase control activity. Plotted values represent an average CAT activity relative to that of uninduced vector plasmid pBLCAT3 from three separate transfection assays, and the range represents the maximum values obtained in an individual assay.
Responsiveness of ␣ 1 -AT S-2 Promoter-To determine whether the interaction of NF-B is indeed responsible for the LPSmediated transcriptional induction of the ␣ 1 -AT gene, we selectively mutagenized the NF-B binding element of the S-2 promoter, and a mutated chimeric gene was constructed that contained this altered NF-B-binding element. The sequence was changed from 5Ј-GGGGCTTTCCCC-3Ј to 5Ј-GCTCCTTTC-GAG-3Ј. This mutated construct designated as pAT2.4mt NF-B CAT reporter gene was then transfected into monocytes, and its response to LPS was compared with the reporter gene containing a wild-type sequence (Fig. 9). Mutation of the NF-B element virtually eliminated the induction of the mutated reporter gene, while the wild-type promoter, in the presence of LPS, induced CAT activity more than 15-fold over background. Thus, the NF-B-binding site appears to be a necessary regulatory element for activation of the ␣ 1 -AT gene by LPS.

Overexpression of NF-B Induces ␣ 1 -AT Gene Expression-
For further verification of the role of NF-B in potentiating transcription of the ␣ 1 -AT S-2 gene, NF-B expression vectors were cotransfected into blood monocytes together with wildtype pAT2.4 CAT or pAT2.4 mt NF-B CAT reporter plasmids. Overexpression of NF-B resulted in strong transactivation of the ␣ 1 -AT promoter in a manner which was dependent on the integrity of the identified NF-B-binding element (Fig. 10). Both p50 and p65 subunits of NF-B induced transcription in a dose-dependent manner. These results clearly demonstrate the involvement of NF-B in the transcriptional induction of the ␣ 1 -AT S-2 gene.

DISCUSSION
Extremely high inducibility of the ␣ 1 -AT S-2 gene under inflammatory conditions has prompted a detailed analysis of the regulatory elements of this gene. ␣ 1 -AT is one of the major proteinase inhibitors in serum with a broad range of activities but mainly protects the elastic fibers in lungs from excessive digestion by neutrophil elastase. During the host response to inflammation/injury expression of ␣ 1 -AT is induced. However, the molecular mechanisms of this induction process is still FIG. 6. Analysis of the LPS responsive promoter of the ␣ 1 -AT S-2 gene. Specific sequences of the ␣ 1 -AT S-2 gene (horizontal bars) inserted upstream of the CAT gene of pBLCAT3 vector are aligned with a schematic locating specific restriction enzyme sites. Peripheral blood monocytes were transiently transfected with the constructs indicated (10 g of DNAs were used for each plasmid). After transfection, cells were maintained in the medium in the absence of stimulation or treated with 10 g/ml of LPS for an additional 24 h. The histogram represents uninduced (solid bars) and LPS-induced (hatched bars) CAT activity. Details of transfection and the CAT activity assay are described under "Materials and Methods." Relative CAT activity was determined as described in Fig. 5.   FIG. 7. Distal promoter element of ␣ 1 -AT S-2 gene interacts with the LPS-induced nuclear factors. EMSAs were performed with 32 P-labeled ␣ 1 -AT S-2 promoter DNA containing a NF-B element (Ϫ2295 to Ϫ2250). A, nuclear extracts (10 g of protein), prepared from normal untreated (lane 1) and 4-h LPS-treated (lane 2) monocytes, were incubated with the 32 P-labeled probe following the method described under "Materials and Methods," and the resulting complexes were resolved in a 6% native polyacrylamide gel. The positions of two DNAprotein complexes A and B formed by the factors in the LPS-induced nuclear extract are shown. B, prior to the addition of 32 P-labeled probe, the LPS-induced nuclear extract was incubated with competitor oligonucleotides containing either the wild-type (wt) (lane 2) or mutant (mt) (lane 3) sequence for the NF-B element. The sequences of these two oligonucleotides are described under "Materials and Methods." A control assay containing no competitor oligonucleotide is shown in lane 1.

FIG. 8. Identification of NF-B/Rel family members that interact with the distal promoter element of the ␣ 1 -AT S-2 gene.
EMSA was performed using 32 P-labeled promoter DNA (Ϫ2295 to Ϫ2250) and nuclear extract (10 g of protein) from LPS-treated rabbit blood monocytes. Prior to the addition of the 32 P-labeled probe, nuclear extract was incubated with 0.5 and 1 l of a 10-fold diluted anti-p50 antisera (lanes 2 and 3, respectively) or anti-p65 antisera (lanes 4 and 5, respectively). These antisera have been described earlier (Ray et al., 1995). As a control, a nonspecific antisera was also used (lane 6). The preincubation was done at 4°C for 45 min and then incubated with the 32 P-labeled probe. Lane 1 contains no antisera. The DNA-protein complexes were resolved in a 6% native polyacrylamide gel.
unclear. This report presents the first clear evidence for a regulatory element that controls the inducible expression of an inflammation responsive ␣ 1 -AT gene. The major findings of the present study include identification of a distal LPS-responsive promoter region that contains a NF-B-binding element. This element binds with NF-B, and its mutation abolishes the LPS-mediated inducibility of ␣ 1 -AT promoter construct in monocytic cells. Overexpression of NF-B in monocytes induced the ␣ 1 -AT expression. The data taken together strongly suggest that LPS-mediated ␣ 1 -AT gene induction is controlled by NF-B. Involvement of NF-B in the ␣ 1 -AT gene induction process has not been reported earlier.
What is the significance of the high level induction of the ␣ 1 -AT S-2 gene under inflammatory conditions? A logical explanation may be to down-regulate the proteolytic cascade associated with the cell death pathway activated during inflammation. Interleukin 1-␤ produced and activated in response to inflammation can activate many cellular events causing cell death. IL-1␤-converting enzyme (ICE), a mammalian proteinase processes pro-IL-1␤ to the active form mature IL-1␤ by proteolytic cleavage at aspartic residues (Black et al., 1988). ICE is also homologous to the ced-3 gene of Caenorhabditis elegans that is essential for apoptosis (Ellis et al., 1991). Recently it has been shown that ICE can be inhibited by a novel cysteine proteinase inhibitor CrmA originally identified in cowpox virus (Ray et al., 1992). Expression of crmA can prevent cell death mediated by ICE as well as quelling other IL-1-mediated inflammatory response. Overexpression of crmA can also prevent cell death mediated by TNF receptor associated death domain protein (Hsu et al., 1995), Fas associated death domain protein (Chinnaiyan et al., 1995), and a death protease Yama, a mammalian homolog of ced-3 . In view of high inducibility of the ␣ 1 -AT S-2 gene under inflammatory conditions, it is possible that a rabbit genome has evolved and maintained this highly inducible proteinase inhibitor to maintain cellular homeostasis and cope with the influx of proteinases brought in by infectious agents upon wounding or infection. It is known that parasites such as Schistosoma mansoni (McKerrow et al., 1985) and bacteria such as Pseudomonas (Werb et al., 1982) synthesize serine proteinases that act as virulence factors to increase the efficiency of infection.
In an effort to elucidate the mechanism of transcriptional induction, we undertook the cloning and characterization, as well as sequencing, of the immediate upstream promoter region of the rabbit ␣ 1 -AT S-2 gene. This gene spans about a 10-kb region and consists of five exons. A translation initiator ATG codon is present at the beginning of the second exon, while the first exon codes for the 5Ј-untranslated region. The proximal promoter region, up to 200 bp upstream of the transcription start site (Fig. 4) of S-2 gene, shows a high degree of homology with the proximal region of human and mouse ␣ 1 -AT genes. This region contains a number of major transcription factorbinding elements including HNF1/LF-B1, HNF2/LF-A1, HNF3, and C/EBP which are seen to be highly conserved. Such a high conservation of several transcription factor binding sequences argues for their importance in controlling the basal expression of ␣ 1 -AT across the species. These elements, however, are not sufficient for the inflammation-induced expression of the S-2 gene. Additional regulatory elements are crucial for its inducibility. This is evident from the detailed promoter analyses of this gene (Fig. 6).
Functional studies show that promoter of the S-2 isoform is highly inducible both in liver and monocyte cells in response to cytokines present in the conditioned medium or LPS (Fig. 5). Transient transfection of monocyte cells (Fig. 6) using a series of CAT gene reporter constructs, demonstrates that upstream sequences located between Ϫ2438 and Ϫ1990 (a 0.5-kb PstI fragment) are important for the induction response mediated by LPS. Analysis of the LPS-responsive region of S-2 promoter revealed the presence of a NF-B-binding element in this region. Drastic reduction of LPS responsiveness, when the NF- FIG. 9. Site-directed mutagenesis of the NF-B element abolishes LPS inducibility. Peripheral blood monocytes were transiently transfected with 10 g of CAT reporter plasmids. The pAT-2.4 CAT plasmid contained wild-type S-2 promoter (Ϫ2438/ϩ70) DNA, and the pAT-2.4mtNF-B CAT plasmid contained the same DNA selectively mutated at the NF-B-binding element. The sequence of the wild-type (wt) and mutant (mt) NF-B elements are indicated. The underlined bases represent mutated bases. After transfection, one set of transfected cells were stimulated with 10 g/ml LPS for 24 h. Relative CAT activity was determined as described in Fig. 5.   FIG. 10. Overexpression of NF-B potentiates ␣ 1 -AT S-2 promoter activity. Two CAT reporter plasmids (10 g each), pAT-2.4 CAT and pAT-2.4mtNF-B CAT, carrying either wild-type S-2 promoter DNA (Ϫ2438/ϩ70) or the same DNA selectively mutated at the NF-B site, respectively, were cotransfected in blood monocyte cells with increasing amounts (0, 1, 2, 3, 4, and 5 g) of the plasmids expressing either NF-B (p65) or NF-B (p50) proteins. Relative CAT activity was determined as described in Fig. 5. B-binding element is altered, clearly suggests that NF-B plays a crucial role for efficient induction of the ␣ 1 -AT S-2 gene by LPS in monocytic cells. The abilities of (i) inducible NF-Blike factors to bind to the ␣ 1 -AT promoter, (ii) anti-p65 and anti-p50 antibodies to supershift the LPS-inducible DNA-protein complex to the ␣ 1 -AT promoter, and (iii) the overexpression of p65 and p50 subunits of NF-B to activate ␣ 1 -AT gene expression further attest to the role of NF-B as a regulator of ␣ 1 -AT gene induction. NF-B is a family of pleiotropic inducible transcription factors originally identified as nuclear factors that bind to the B enhancer motif of immunoglobulin light chain (Sen and Baltimore, 1986). The transcription factor is activated by a number of extracellular signals and cytokines including IL-1 and tumor necrosis factor. Recent numerous studies have shown that NF-B regulates a wide variety of genes including those of IL-1␤, IL-6, IL-8, IL-2, and serum amyloid A (Bensi et al., 1990;Libermann and Baltimore, 1990;Stein and Baldwin, 1993;Hoyos et al., 1989;Ray et al., 1995). The involvement of NF-B in the transcriptional induction of ␣ 1 -AT broadens the range for this factor to yet another gene that is induced in response to inflammatory stress.
In this report, we have laid the foundation for more detailed analyses of the regulation of the inducible expression of the ␣ 1 -AT gene in liver and monocyte/macrophages during the inflammatory response. Increased expression of ␣ 1 -AT also poses a serious health problem in individuals carrying the P i zz ␣ 1 -AT mutant allele. During inflammation, synthesis of ␣ 1 -AT in these patients is increased, causing greater intracellular accumulation of the defective protein (Barbey-Morrel et al., 1987). Understanding the molecular mechanisms of inducible expression of ␣ 1 -AT will allow for development of therapies to intervene the increased synthesis and accumulation of the defective protein in the liver of a homozygous P i zz individual. Further work along this line is currently in progress.