Localization of regulatory elements mediating constitutive and cytokine-stimulated plasminogen gene expression.

The activity of plasmin, the major enzyme responsible for dissolving fibrin clots, is regulated by plasminogen activators, plasminogen activator inhibitors, alpha(2)-antiplasmin, and inflammatory mediators. Recent studies suggest that plasmin activity can be regulated also at the level of plasminogen gene expression. In this study, we characterized the murine plasminogen promoter and 5'-flanking region. The major transcription start site was identified at -83 bp relative to the ATG translational initiation codon. A series of 5'-flanking sequences up to 2400 bp upstream of the transcription initiation site were fused to the luciferase reporter gene and transfected into hepatocytic cells. A 106-bp 5'-flanking region of the murine plasminogen gene demonstrated sufficient functional promoter activity in plasminogen-expressing cells. IL-6 treatment stimulated luciferase activity driven by the 5'-flanking region and an intact consensus IL-6-responsive element at -791, was required for maximal stimulation by this cytokine. These results indicate the presence of regulatory elements in the 5'-flanking region of the murine plasminogen promoter that may regulate murine plasminogen gene expression and, hence, plasmin activity.

Plasminogen is the zymogen of the serine protease plasmin, which is the major enzyme responsible for degrading fibrin clots (1). Plasmin activity is regulated by the presence of plasminogen activators, plasminogen activator inhibitors, ␣ 2 -antiplasmin, and inflammatory mediators (2,3). Recent studies from our laboratory and others suggest that plasmin activity can be regulated also by the modulation of plasminogen gene expression (4 -7).
Plasminogen is synthesized primarily in the liver (8 -11) as an 810-amino acid residue polypeptide chain. Murine plasminogen contains two additional amino acid residues at positions 543 (Ser) and 587 (Gly). Like human plasminogen, murine plasminogen is converted to plasmin by cleavage of a single Arg-Val peptide bond by plasminogen activators (either tissue plasminogen activator or urokinase) (12). Murine plasmin is composed of a 562-amino acid heavy chain (derived from the amino terminus of plasminogen) that is disulfide-linked to a 231-amino acid light chain (derived from the carboxyl terminus of plasminogen). The catalytic domain contained in the 231amino acid light chain is 84% identical when comparing murine and human plasminogens (13).
Several reports suggest that plasminogen is an acute phase reactant (14 -18). In addition to its function in fibrinolysis, plasminogen participates in a variety of physiological processes including wound healing (19,20), vascular remodeling (21,22), and leukocyte migration (23). In recent years, the murine model has been used extensively to study both physiological and pathological processes associated with plasminogen deficiency. Pathobiological conditions associated with plasminogen deficiency that are observed in both plasminogen Ϫ/Ϫ mice and homozygous plasminogen-deficient humans include thrombotic disease (24 -26) and ligneous conjunctivitis (27)(28)(29).
Previous studies conducted in our laboratory demonstrated that plasminogen mRNA expression is increased in primary murine hepatocytes treated with interleukin 6 (IL-6). 1 Furthermore, mice injected with IL-6 exhibit increases in hepatic plasminogen mRNA and circulating plasminogen levels, compared with mice injected with saline (4). In the present study, we have sequenced the 5Ј-flanking region 2600 bp upstream of murine plasminogen exon I, delineated the transcriptional start site, and defined the minimal promoter region required for constitutive expression of the murine plasminogen gene in hepatocytic cells. Our studies demonstrate that a 106-bp fragment of the 5Ј-flanking region of the murine plasminogen gene is sufficient to confer transcription in plasminogenexpressing cell lines. We also show that IL-6 stimulates murine plasminogen gene expression and have identified cisacting elements in the plasminogen promoter that may provide a mechanism for IL-6-mediated functional regulation of the plasminogen gene in vivo.

Murine Plasminogen Promoter and 5Ј-Deletional Constructs-A
DASHII 129/SvJ murine genomic library was screened by in situ hybridization using a 32 P-labeled 580-bp EcoRI-NsiI fragment from murine plasminogen cDNA (13) containing the amino-terminal portion of the cDNA through the second kringle domain. Positive phage clones were isolated and screened by polymerase chain reaction (PCR) for the presence of exons I and II using the following exon-specific primer pairs: (a) mPLE1-5Ј (5Ј-CCGGTGCTGTTGGCCAGTCCC-3Ј) and mPLE1-3Ј (5Ј-CTGGTTTCAGAAGCAAGAGA-3Ј) corresponding to nucleotides 1-21 and 73-54 of the murine plasminogen cDNA and (b) mPLE2-5Ј (5Ј-GGGGACTCGCTGGATGGCTA-3Ј) and mPLE2-3Ј (5Ј-TTCACATT-TGGCCAAACAGT) corresponding to nucleotides 79 -98 and 186 -167 of the murine plasminogen cDNA (13). An 11.5-kb SacI genomic fragment of murine plasminogen (data not shown) was excised from purified phage DNA by restriction enzyme digestion with SacI (24). The fragment was inserted into the Bluescript II plasmid, and correct orientation was verified by using plasmid templates (30). The DNA sequence 2600 bp upstream from exon I was obtained using the dideoxy chain termination method and DNA Strider TM 1.2 software (Dr. C. Marck, Service de Biochemie et de Genetique Moleculaire, Batiment 142, Centre d'Etudes de Saclay, Gif-sur-Yvette Cedex, France). A 1052-bp Bst1107/MscI genomic fragment, containing the 5Ј-flanking region and 17 bp of exon I, was excised from the 2600-bp murine plasminogen DNA fragment and cloned into the SmaI site of the pUC19 plasmid. The DNA fragments were gel purified and subcloned into the KpnI/HindIII sites of the promoterless pGL2/Basic plasmid (Promega, Madison, WI) upstream of the luciferase reporter gene in both forward (pGL2/mPLPR) and reverse (pGL2/mPLPRЈ) orientations. The orientation of these constructs was verified by restriction digestion with EcoRI. The murine plasminogen promoter sequence was scanned for putative transcription factor binding sites using version 2.2 of the TRANSFAC 4.0 data base (31).
The deletional constructs were constructed via PCR amplifications using appropriate primers. The resulting fragments spanned 2400, 1712, 1064, 700, 500, 403, 250, and 106 bp upstream from the transcription initiation site. These fragments were cloned into the pGL2/ Basic plasmid using restriction sites in the linker region and sequenced using the dideoxy chain termination method. All of the constructs had the anticipated DNA sequences.
Human Plasminogen Promoter and 5Ј-Deletional Constructs-pGL2/ hPLPR consisted of a 1067-bp fragment (nucleotides Ϫ914 to ϩ154, relative to the transcription initiation site) of the human plasminogen promoter cloned into the pGL2/Basic plasmid in the forward orientation, upstream of the luciferase reporter gene. pGL2/hPLPRЈ had the same plasminogen nucleotide sequence cloned into the pGL2/Basic vector in the reverse orientation. The preparation of these constructs has been described previously (4). Human plasminogen 5Ј-deletional constructs were constructed via PCR amplifications using appropriate primers. The resulting fragments spanned 935, 710, 515, 290, 234, and 189 bp upstream of the ATG translational start site. These fragments were cloned into the pGL2/Basic plasmid using restriction sites in the linker region and sequenced by the dideoxy chain termination method. All of the constructs had the anticipated DNA sequences.
RNA Isolation and Primer Extension-Total RNA was harvested from livers of CB6F1 male 5-week-old mice using the guanidinium isothiocyanate procedure (32). For identification of the murine plasminogen transcription-initiation site, two oligonucleotide primers: 5Ј-GCACCTGGACAACTGTGTCC-3Ј, complementary to nucleotides Ϫ37 to Ϫ18, and 5Ј-CCTTATGGTCCATGTTGGGACTGGCC-3Ј, complementary to nucleotides Ϫ13 to ϩ13 of the cDNA sequence of murine plasminogen (13), were used. The methionine initiation (ATG) codon of the murine plasminogen gene was designated as nucleotide ϩ1 (Fig. 1). The primer phosphorylation reaction contained, in a total volume of 10 l, 10 pmol of primer, 50 mmol/liter Tris-HCl (pH 7.5), 10 mmol/liter MgCl 2 , 5 mmol/liter dithiothreitol, 0.1 mmol/liter spermidine, 30 Ci of [␥-32 P]ATP (3000 Ci/mmol (Amersham Biosciences)), and 10 units of T4 polynucleotide kinase (New England Biolabs, Beverly, MA). The mixture was incubated at 37°C for 10 min, and then heated to 90°C for 2 min to inactivate the T4 polynucleotide kinase. The primer-extension hybridization reaction contained the following in a final volume of 11 l: 10 -30 g of total liver mRNA, 1 ϫ 10 6 cpm oligonucleotide primer, 10 mmol/liter Tris-HCl (pH 8.3), 50 mmol/liter KCl, 10 mmol/liter MgCl 2 , 10 mmol/liter dithiothreitol, 1 mmol/liter each deoxynucleoside triphosphate, and 0.5 mmol/liter spermidine. The primer was annealed to the mRNA by heating to 65°C for 5 min and allowed to cool slowly to 22°C. Extension was carried out in a final volume of 20 l with the addition of 2 mmol/liter sodium pyrophosphate plus 1 unit of avian myeloblastosis virus reverse transcriptase (Promega) and incubated at 42°C for 30 min. The reaction was stopped by the addition of an equal volume of loading dye (10 mmol/liter EDTA, 0.1% xylene cyanol, and 0.1% bromphenol blue in 98% formamide). The extended samples were electrophoresed through a 7% polyacrylamide, 7 M urea sequencing gel (33).
To further study the cellular specificity of the murine plasminogen gene promoter responsiveness, we transfected the 106-bp mPLPR construct in the hepatic cell line, Hepa 1-6, and nonhepatic IMR-32 (neuroblastoma) and Caco-2 (colorectal adenocarcinoma) cell lines.
In all experiments, plasminogen promoter constructs were cotransfected with the Renilla luciferase reporter, pRL-TK (Promega) at a ratio of 50:1 for pGL2/experimental vector DNA to pRL-TK DNA. Cells were cultured for 24 -48 h at 37°C. Cell extracts were assayed for reporter gene activity 24 -48 h after addition of DNA to the cells, using the Dual Luciferase Reporter Assay System (Promega) and a Monolight 2001 luminometer (Analytical Luminescence Laboratory, San Diego, CA). The expression of the experimental reporter gene was normalized to the activity of the Renilla luciferase reporter gene and expressed as normalized -fold change in luciferase activity relative to the activity of the pGL2/Basic control plasmid.
Treatment of Hepa 1-6 Cells with Murine Interleukin 6 (mIL-6)-Hepa 1-6 cells transfected with the pGL2/mPLPR constructs were grown in 12-well cluster plates containing in each well (3.8 cm 2 ) 1.5 ml of DMEM supplemented with 10% FBS and 4 mmol/liter L-glutamine. At 75-80% confluence, cells were rinsed with phosphate-buffered saline and then grown in serum-free DMEM supplemented with 0.1% bovine serum albumin and 4 mmol/liter L-glutamine. For the murine IL-6 dose-response study, recombinant murine IL-6 (mIL-6; Sigma) was added in concentrations from 0 to 750 units/ml. Cell extracts were assayed for luciferase activity 48 h after addition of mIL-6 to the cells. For the mIL-6 time-course study, transfected Hepa 1-6 cells were incubated with 500 units/ml mIL-6. Cell extracts were then assayed for reporter gene activity at 0 (no treatment), 24, 48, and 72 h after addition of mIL-6 (500 units/ml) to the cells. As a positive control for IL-6 stimulation, cells were transfected with a pGL2/fibrinogen construct. pGL2/fibrinogen consisted of a 200-bp fragment of the ␤-chain of fibrinogen cloned into the pGL2/Basic plasmid in the correct orientation, upstream of the luciferase reporter gene. The effects of IL-6 stimulation on the serial 5Ј-deletional constructs of the murine and human plasminogen promoters were examined as well.
Site-directed Mutagenesis of the Interleukin-6-responsive Element (IL6-RE)-The pGL2-1.7kb and pGL2-2.4kb (mPLPR-2400/ mutIL6RE) mutated plasmids were constructed using the QuikChange site-directed mutagenesis kit according to the instructions from the manufacturer (Stratagene, La Jolla, CA). The site-specific mutated constructs were made using 25 ng of wild-type plasmid templates and the sense oligonucleotide 5Ј-CAAACGGACCTAAACACTGCACAGT-3Ј in combination with the antisense oligonucleotide 3Ј-GTTTGCCTG-GATTTGTGACGTGTCA-5Ј in a final volume of 50 l. Each construct was sequenced to confirm the incorporated mutation (Retrogen, San Diego, CA).
Statistical Analysis-All data are presented as means Ϯ S.E. of the mean. Statistical significance (p Ͻ 0.05) in all dual luciferase reporter assays was determined via one-way analysis of variance followed by the Student-Newman-Keuls post hoc test.

RESULTS
DNA Sequence of the 5Ј-Flanking Region of the Murine Plasminogen Gene-The sequence of the 5Ј-flanking region 2600 bp upstream of murine plasminogen exon I was determined (Fig.  1). The TRANSFAC data base was used to search for consensus transcription factor binding sites. Putative binding sites for the liver-enriched transcription factors: CCAAT/enhancer binding protein ␤ (C/EBP␤ or nuclear factor IL-6, NF-IL6), hepatic nuclear factor 1 (HNF-1) (34), and hepatic leukemia factor (HLF) were present in the 2600-bp murine plasminogen promoter sequence (Fig. 1). The sequence of the 2600-bp murine plasminogen gene promoter was aligned with the published sequence of the human plasminogen promoter. Comparison of 2600 bp of the 5Ј-flanking region of the murine gene with the corresponding sequence of the human plasminogen gene (35) showed 50% identity. In the 5Ј-flanking region spanning Ϫ250 bp relative to the ATG start site (designated as ϩ1), comparison between the murine and human plasminogen 5Ј-flanking region showed 70% identity. As shown in Fig. 2, within this region, a number of putative regulatory elements were conserved, notably for liver-specific transcription factors C/EBP␤, HNF-1, HLF, and ubiquitous factors activator protein 1 (AP-1) and nuclear factor B (NF-B). These data suggest the existence of a similar regulation pathway by binding factors in these two species.
Determination of the Transcriptional Start Site of the Murine Plasminogen Gene-The transcriptional start site of the murine plasminogen gene was determined by primer extension analysis. The end-labeled [␥-32 P]ATP primer, 5Ј-GCACCTG-GACAACTGTGTCC-3Ј, complementary to nucleotides Ϫ37 to Ϫ18, was used in extension reactions with total RNA from murine liver (the major site of plasminogen synthesis) and Nor-10 skeletal muscle cells (negative control). The major extension product was located 83 bp upstream from the ATG initiation codon (designated as ϩ1) and ended at a T residue, thus identifying Ϫ83 as the major transcription initiation site in the murine liver (Fig. 3). No bands were detected when the extension reaction was performed with RNA from the negative control, Nor-10 skeletal muscle cells (data not shown). The same start site was identified when a radiolabeled 26-oligonucleotide primer, 5Ј-CCTTATGGTCCATGTTGGGACTGGCC-3Ј, complementary to nucleotides Ϫ13 to ϩ13 was used (data not shown).
Functional Analysis of the Murine Plasminogen Promoter and 5Ј-Flanking Region in Hepatocytic Cells-The ability of the 1064-bp murine plasminogen promoter and 5Ј-flanking region to drive expression of a luciferase reporter gene was examined and directly compared with a human plasminogen promoter of similar length that we have characterized previously, pGL2/ hPLPR (4). To determine whether the murine plasminogen 5Ј-flanking region could confer liver-specific transcription, four cell lines (Hepa 1-6, Nor-10, Hep G2, and MCF-7) were transfected with each of six constructs: 1) 1064-bp pGL2/mPLPR, 2) 1064-bp pGL2/mPLPRЈ, 3) 1067-bp pGL2/hPLPR, 4) 1067-bp hPLPRЈ, 5) pGL2/Basic, and 6) pGL2/SV40. The dual luciferase reporter assay was employed for the quantitative measurement of plasminogen promoter activity. Hepa 1-6 cells transfected with the pGL2/mPLPR construct expressed luciferase activity that was 4.6-fold higher (p Ͻ 0.001) than cells transfected with the promoterless vector control, pGL2/Basic (Fig. 4A). No induction of luciferase activity relative to the activity of pGL2/ Basic was observed in Hepa 1-6 cells transfected with pGL2/ mPLPRЈ (Fig. 4A), a construct that is identical to pGL2/mPLPR except that the 1064-bp plasminogen 5Ј-flanking region is cloned in the reverse orientation. Luciferase expression driven by the positive control for transfection efficiency, pGL2/SV40, increased 123-fold (p Ͻ 0.05) in Hepa 1-6 cells (data not shown). To investigate cell specificity of the murine plasminogen promoter, we transfected cells of the Nor-10 murine skeletal cell line with the pGL2/mPLPR construct. (Plasminogen expression is not detectable in murine skeletal muscle (11).). As shown in Fig. 4B, there was no statistically significant difference in luciferase activity in Nor-10 murine skeletal muscle cells transfected with either pGL2/mPLPR or pGL2/mPLPRЈ when compared with cells transfected with pGL2/Basic. Luciferase expression driven by pGL2/SV40 increased 17-fold (p Ͻ 0.05) in Nor-10 cells when compared with cells transfected with the pGL2/Basic construct (data not shown).
We also examined the activity of the murine 5Ј-flanking region in human Hep G2 cells, a representative hepatoma line. Hep G2 cells transfected with the pGL2/mPLPR construct exhibited luciferase activity 32-fold greater than that of cells transfected with the pGL2/Basic construct (Fig. 4C, p Ͻ 0.001). Luciferase expression by cells transfected with pGL2/mPLPRЈ was not significantly increased compared with cells transfected with the pGL2/Basic construct (Fig. 4C). Hep G2 cells transfected with pGL2/SV40 provided a 504-fold stimulation (p Ͻ 0.05) of luciferase activity compared with pGL2/Basic (data not shown). As a control to examine cell specificity, MCF-7 (human breast carcinoma) cells were transfected with the pGL2/ mPLPR construct. As shown in Fig. 4D, there was no significant difference in luciferase activity in MCF-7 cells transfected with either pGL2/mPLPR or pGL2/mPLPRЈ when compared with cells transfected with pGL2/Basic (p Ͼ 0.05). Luciferase expression driven by pGL2/SV40 increased 104-fold (p Ͻ 0.05) in MCF-7 cells when compared with cells transfected with the pGL2/Basic construct (data not shown). Taken together, these results suggest that 1064 bp of the 5Ј-flanking region of the murine plasminogen gene are sufficient to confer liver-specific transcription.
When we compared the activities of the pGL2/mPLPR with a construct containing the proximal 1067 bp of the human plasminogen 5Ј-flanking region (pGL2/hPLPR), luciferase expression driven by the murine and human constructs differed by less than 2-fold in Hepa 1-6 cells (Fig. 4A). There was no statistical difference in the induction of luciferase activity between Hep G2 cells transfected with either the 1064-bp pGL2/ mPLPR or the 1067-bp pGL2/hPLPR (Fig. 4C). Hepa 1-6 and Hep G2 cells transiently transfected with the 1067-bp pGL2/ hPLPRЈ also showed no statistical difference in the induction of luciferase activity compared with the pGL2/Basic control (Fig.  4, A and C). The results indicate that the 5Ј-flanking regions of the murine and human plasminogen genes contain sequences that control the expression of these genes, and that constitutive promoter function is orientation-dependent. These data also suggest that the 1064-bp murine and 1067-bp human plasminogen promoter regions exhibit similar activities and, because the overall level of induction was much greater in Hep G2 cells than in Hepa 1-6 cells, that levels of transcription factors in the two cell lines, Hepa 1-6 and Hep G2, may differ. sequences immediately upstream of the ATG translational initiation site were constructed, and their abilities to drive luciferase expression were compared in both Hepa 1-6 and Nor-10 cells. The 5Ј-flanking constructs ranged in size from 106 to 2400 bp. The construct containing the first 106 bp of the 5Ј-flanking region of the murine plasminogen gene drove luciferase expression in Hepa 1-6 cells that was not statistically different from the luciferase expression driven by the 2400-bp construct (Fig. 5A). Thus, minimal promoter activity was contained in the first 106 bp upstream from the transcription initiation site. A deletion from Ϫ699 to Ϫ500 bp resulted in a 2-fold increase in luciferase activity, suggesting the presence of a repressor element in this region. Further deletion from Ϫ499 bp to Ϫ403 bp led to a 2-fold decrease in promoter activity, suggesting the presence of an enhancer element within this region. In negative controls, none of the constructs exhibited increased luciferase expression compared with the pGL2/Basic construct, when transfected into Nor-10 cells (data not shown). The results show that the 106-bp minimal mPLPR construct was sufficient to increase luciferase activity in Hepa 1-6 cells. The data also suggest that there are negative and positive cis-acting regulatory elements within 2400-bp of the 5Ј-flanking region of the murine plasminogen gene.
We also examined promoter activities of serial 5Ј-deletional human plasminogen promoter (hPLPR) constructs transfected into Hep 3B cells (Fig. 5B). A sequence consisting of the first 189 bp upstream of exon I of the human plasminogen gene, exhibited minimal promoter activity. A deletion from Ϫ709 bp to Ϫ515 bp resulted in a modest increase in luciferase activity, suggesting the presence of a repressor element within this region. Further deletion of sequences from Ϫ514 bp to Ϫ290 bp led to a 1.5-fold decrease in promoter activity suggesting the presence of an enhancer element within this region. Luciferase activity of MCF-7 cells transfected with the hPLPR deletional constructs did not significantly differ from the reporter gene activity of cells transfected with the pGL2/Basic construct (data not shown). These results suggest that the murine and human plasminogen 5Ј-flanking regions contain similar positive and negative cis-acting regulatory elements involved in liver-specific transcriptional activity of the plasminogen promoter.
To examine whether the 106-bp minimal promoter region of the murine plasminogen gene confers liver specificity, we transfected the 106-bp mPLPR construct into two nonhepatic plasminogen-expressing cell lines, IMR-32 (neuroblastoma) and Caco-2 (colorectal adenocarcinoma) (36) and compared luciferase expression with transfected Hepa 1-6 cells. As shown in Fig. 6, luciferase expression by cells transfected with the construct containing the first 106 bp (relative to the transcription initiation site) of the 5Ј-flanking region of the murine plasminogen gene was significantly increased (p Ͻ 0.05) compared with cells transfected with the pGL2/Basic construct in all three cell lines consistent with plasminogen expression by these cells. In addition, the minimal 106-bp mPLPR construct drove luciferase expression in each of the three plasminogenexpressing cell lines that was not statistically different from luciferase expression driven by the mPLPR-2400 construct (p Ͼ 0.05). These data suggest that a 106-bp fragment of the 5Јflanking region of the murine plasminogen gene is sufficient to direct transcription in plasminogen-expressing cells but sequences within this region do not confer liver specificity of plasminogen expression.
Stimulation of Plasminogen Promoter Activity in Hepa 1-6 Cells by Murine IL-6 -We have demonstrated previously that interleukin 6 increases plasminogen mRNA levels in primary murine hepatocytes (4). In addition, mice injected with IL-6 exhibit increases in hepatic plasminogen mRNA and circulating plasminogen levels compared with mice injected with saline (4). To examine whether the murine pGL2/mPLPR construct behaved as the endogenous gene, we tested whether cells transfected with the 1064-bp pGL2/mPLPR construct could respond to mIL-6 (murine IL-6). Hepa 1-6 cells were transfected with the 1064-bp pGL2/mPLPR, 1064-bp mPLPRЈ, or pGL2/Basic and grown in the presence of increasing concentrations of mIL-6 for 48 h. The maximal increase in luciferase activity expressed by cells transfected with the 1064-bp pGL2/mPLPR construct was achieved with 500 units/ml mIL-6 (2.2-fold) compared with untreated cells (Fig. 7A). As a positive control, Hepa 1-6 cells transfected with the pGL2/fibrinogen construct and incubated with 500 units/ml mIL-6 for 48 h exhibited an 2.3fold increase in luciferase activity compared with untreated cells (data not shown). The maximal concentration of 500 units/ml mIL-6 is similar to the concentration at which maximal stimulation of human plasminogen mRNA expression is achieved in primary murine hepatocytes with human IL-6 (4). In a separate set of experiments, a time-dependent increase in murine plasminogen promoter activity was also observed in Hepa 1-6 cells in response to mIL-6 treatment. As shown in minogen promoter behaves as the endogenous gene, with regard to the response to IL-6 treatment of the cells.
Experiments were then performed to localize the region(s) in the murine plasminogen promoter that mediate IL-6-dependent stimulation in Hepa 1-6 cells. The location of putative IL-6-responsive elements present in the murine plasminogen promoter region are depicted in Fig. 8A. Hepa 1-6 cells were transfected with the series of mPLPR 5Ј-deletional constructs and then incubated for 48 h in the presence of 500 units/ml mIL-6 prior to measuring luciferase activity. One region in the murine plasminogen gene appeared to predominantly regulate the increased gene expression in response to IL-6 treatment. The level of IL-6-dependent stimulation fell from 3.4-to 1.5-fold when the region from Ϫ1063 to Ϫ700 was deleted. This region contains a C/EBP␤ (NF-IL6) consensus sequence beginning at Ϫ791 bp (relative to the ATG codon). The 106-bp mPLPR construct, which contains two IL-6-responsive elements, positioned at 139 and 173 bp, respectively, did not significantly respond to IL-6 stimulation compared with the response of the promoterless vector (p Ͼ 0.05) (Fig. 8B). In addition, the pres-ence of 144 bp upstream of the 106-bp minimal promoter (250-bp construct containing an additional consensus IL-6-responsive element positioned at 206 bp) did not significantly alter IL-6 responsiveness compared with the promoterless vector. These results suggest that the three putative IL-6-responsive elements present in the region from Ϫ250 to Ϫ83 relative to the ATG codon may not play a major role in the IL-6 inducibility of the murine plasminogen gene in Hepa 1-6 cells.
To further investigate whether the IL6-RE motif at Ϫ791 plays a role in the induction of murine plasminogen gene expression by interleukin-6, Hepa 1-6 cells were transfected with either the wild-type (intact IL6-RE) 1712-bp mPLPR or 2400-bp mPLPR or mutant (containing a 3-bp mutated IL6-RE binding site) mPLPR-1712/mutIL6RE or mPLPR-2400/mutIL6RE constructs or the promoterless control vector, pGL2/Basic. The mutation within the putative IL6-RE binding motif is shown in Fig. 9A. IL-6-treated Hepa 1-6 cells transfected with either the mPLPR-1712/mutIL6RE or mPLPR-2400/mutIL6RE constructs did not exhibit increased luciferase activity compared with promoterless control vector, pGL2/ . Luciferase activities of the mPLPR and hPLPR constructs were compared with that of the promoterless control vector, pGL2/Basic. The values were calculated by dividing the amount of luciferase activity (normalized against the internal pRL-TK standard) of the murine plasminogen promoter or human plasminogen promoter by that of the pGL2/Basic control expressed by each cell line. The activity of the pGL2/Basic control is therefore 1 in each cell type tested. Results in panels A and C are given as mean Ϯ S.E. (n ϭ 5-12 transient transfections). *, p Ͻ 0.001, compared with the pGL2/Basic control. Basic (Fig. 9B). Under these conditions, IL-6-treated Hepa 1-6 cells transfected with either the wild-type mPLPR-1712 or mPLPR-2400 constructs exhibited significantly (p Ͻ 0.05) increased luciferase expression (1.8-and 2.2-fold, respectively) compared with cells transfected with the control vector alone (Fig. 9B) similar to the extent of stimulation at 24 h as shown in Fig. 7B. There was no statistical difference in the induction of luciferase activity between Hepa 1-6 cells transfected with either mPLPR-1712 or mPLPR-2400 constructs. These results suggest that the interleukin-6-responsive element positioned at Ϫ791 bp is essential for stimulation of plasminogen gene expression in response to IL-6.

DISCUSSION
An emerging area of research has demonstrated that the presence and regulation of plasminogen gene expression in various tissue and cell types plays a critical role in numerous physiologic and pathologic processes (19 -29). Therefore, the elucidation of the molecular mechanisms involved in the modulation of plasminogen gene expression requires the identification and characterization of the transcriptional regulatory regions of the plasminogen gene. In addition, the recent characterizations of mice deficient in plasminogen (24,25) provided an impetus for the study of the structure and function of the murine plasminogen promoter, to assess the applicability of the murine model to the human system. In the present study, we have determined the sequence 2.6 kb upstream from exon I of the murine plasminogen gene, identified the transcription initiation site, demonstrated cis-regulatory elements sufficient to direct tissue-specific regulation of the gene, and localized the minimal promoter region required by plasminogen-expressing cells. In addition, we demonstrated that expression of the murine 5Ј-flanking region was increased in response to IL-6 treatment, mimicking the function of the endogenous gene in vivo. Furthermore, we have localized a major region in the murine 5Ј-flanking sequence that is predominantly responsible for mediating the response to IL-6.
There is a distinct tissue-specific pattern of expression of the plasminogen gene with the liver being the predominant site of plasminogen synthesis (8 -11). We found that a 1064-bp murine plasminogen 5Ј-fragment cloned upstream of the luciferase reporter gene drove luciferase expression in the murine hepatoma cell line, Hepa 1-6, and the human hepatoblastoma cell lines, Hep G2 and Hep 3B. In controls, luciferase expression was not increased compared with the vector alone in cells that do not express plasminogen, murine Nor-10 skeletal muscle cells and human breast carcinoma MCF-7 cells. Furthermore, we demonstrated that the murine plasminogen minimal promoter was active in plasminogen-expressing cell lines IMR-32 and Caco-2 (36). Thus, the ability of the 5Ј-flanking region of the murine plasminogen promoter to drive luciferase expression is consistent with the known tissue expression of the plasminogen gene.
The promoter functions of the 1064-bp murine plasminogen 5Ј-fragment and a 1067-bp human plasminogen 5Ј-fragment (previously described from our laboratory (Ref. 4)) were similar. Compared with the promoterless vector, the luciferase activities driven by both murine and human 5Ј-flanking regions were both ϳ32-fold in human Hep G2 cells; ϳ6and ϳ13-fold, respectively, in human Hep 3B cells (data not shown); and ϳ5and ϳ9-fold, respectively, in murine Hepa 1-6 cells. Thus, interspecies promoter strengths were similar, although differences in the stimulating activities of the hepatocytic cells were observed. A single transcription start site for the murine plasminogen gene was identified 83 bp upstream of the ATG initiation codon. This result is similar to the utilization of a single transcription start site in the human plasminogen gene (35). Minimal murine plasminogen promoter activity was contained within the first 106 bp upstream of the transcription initiation site in Hepa 1-6 cells as well as the plasminogen-expressing cell lines Caco-2 and IMR-32 (36). Thus, this region is sufficient to direct plasminogen transcription in plasminogen-expressing cells. Sequence analysis also identified consensus sequences for the transcription factors AP-1 and NF-B within the 106-bp minimal promoter 5Ј-flanking region in both the murine and human plasminogen genes. These findings are consistent with other studies that have shown that these transcription factors play a significant role in both the basal and inducible transcription of a variety of genes associated with the acute phase response (37)(38)(39).
Alignment between the 5Ј-flanking regions spanning 250 bp from the ATG start site of the murine and human plasminogen genes showed a high overall degree of identity (70%). Within this region and upstream of Ϫ106, a number of putative regulatory elements were conserved, notably for liver-specific transcription factors C/EBP␤, HLF, and HNF-1 (34). Further studies are needed to determine whether binding of these transcription factors plays a role in the high level of plasminogen expression in hepatic versus nonhepatic tissues (11).
Results obtained with transfection studies using murine plasminogen promoter deletional constructs provided insight into the regions involved in plasminogen gene expression, and it is of interest to compare these results with those obtained with the human promoter. Deletional analysis revealed that the murine plasminogen promoter contains a negative (Ϫ699 to Ϫ500 bp) cis-regulatory element within 2400 bp of the 5Јflanking region of the murine plasminogen gene. Two possible candidates for transcriptional repressors are octamer factor 1 (Oct-1) and activator protein 1 (AP-1). Oct-1 and AP-1 are ubiquitously expressed and both can function as either activa- A, schematic representation of the murine plasminogen 5Ј-flanking region deletional constructs with putative IL-6-responsive elements illustrated. B, the murine plasminogen 5Ј-flanking region deletional constructs were transfected into Hepa 1-6 cells and assayed for luciferase activity following treatment with 500 units/ml IL-6 for 48 h (as described under "Experimental Procedures"). Results in panels A and B are expressed as mean Ϯ S.E. (relative to the unstimulated construct) of three independent experiments. *, p Ͻ 0.05, compared with -fold stimulation by IL-6 of cells transfected with the pGL2/Basic control. tors or repressors of transcription (40 -44). Similarly, removal of the homologous region on the human plasminogen gene (Ϫ709 to Ϫ515 bp) resulted in a modest increase in luciferase activity suggesting the presence of a repressor within this region as well (Fig. 5B). Similar results were obtained in another study, in which an increase in promoter activity in the human plasminogen gene was observed in Hep G2 cells upon deletion of sequences from Ϫ700 to Ϫ300 of the human plasminogen 5Ј-flanking region (35). Sequence analysis performed in the current study revealed the presence of putative Oct-1 and AP-1 consensus sites within the region from Ϫ709 to Ϫ515 bp in the human plasminogen promoter also.
Deletional analysis also revealed that the murine plasminogen gene contains positive (Ϫ499 to Ϫ403 bp) cis-regulatory elements within 2400 bp of the 5Ј-flanking region. One candidate for a transcriptional enhancer is C/EBP␤ (or NF-IL6 (Ref. 45)). A C/EBP␤ consensus sequence with forward (Ϫ496 bp to Ϫ488 bp) and reverse (Ϫ497 bp to Ϫ489 bp) orientations, relative to the ATG codon, was identified by TRANSFAC analysis in the murine plasminogen 5Ј-flanking region. C/EBP␤ has been shown to act as a transcriptional activator for several genes, including the acute phase protein, lipopolysaccharidebinding protein (43,46). The results from these deletion experiments point to the presence of both positive and negative regulatory elements in the 5Ј-flanking region that may additionally modulate the transcriptional regulation of the murine plasminogen gene. Our data suggest that similar elements that regulate constitutive expression of the plasminogen gene are present in the murine and human plasminogen 5Ј-flanking regions. Similar structure/function relationships observed for both the murine and human plasminogen promoter may suggest broad applicability of murine models to further investigate plasminogen transcriptional regulation and its potential role in human physiology and pathophysiology.
Plasminogen gene regulation in response to inflammatory mediators and cytokines has not been addressed in detail in the literature. However, several reports suggest that plasminogen behaves as an acute phase reactant (14 -18, 47, 48). The acute phase mediator, IL-6, is induced following induction of the acute phase response (49,50). We have shown previously that mice injected with IL-6 exhibit increases in hepatic plasminogen mRNA; consequently, circulating plasminogen levels are significantly higher in mice injected with IL-6, compared with mice injected with saline (4). Furthermore, primary murine hepatocytes treated with IL-6 also increase plasminogen mRNA expression (4).
Using reporter gene functional analysis, we characterized the IL-6-responsive elements of the murine plasminogen gene. A 1064-bp mPLPR construct conferred the strongest response to IL-6 stimulation in transfected murine hepatoma Hepa 1-6 cells. The experimental data correlate with the observations in the human plasminogen promoter wherein IL-6 stimulation results in a 4.5-fold increase in 1067-bp hPLPR expression in human hepatocarcinoma Hep 3B cells (4). The level of IL-6 induction of murine plasminogen gene expression was significantly decreased upon deletion of the region from Ϫ1063 to Ϫ700 bp, suggesting the presence of a functional IL-6-responsive element in this region. Mutation of the putative NF-IL6 consensus sequence positioned at Ϫ791 bp to Ϫ783 bp, relative to the ATG codon, abolished responsiveness of the murine plasminogen 5Ј-flanking region to IL-6, suggesting that the wild-type sequence is necessary for IL-6-stimulated plasminogen gene expression.
Recently we conducted a tissue survey for plasminogen mRNA expression in mice (11) and found that plasminogen mRNA is expressed broadly extrahepatically at low levels. Plasminogen mRNA is present in adrenal, kidney, brain, testis, heart, lung, uterus, spleen, thymus, and gut (11). The brain, testis, and the thymus cortex are separated from the circulation by anatomic barriers so that plasminogen synthesis within these tissues should provide an exclusive source of plasminogen. Two recent reports have demonstrated the regulation of plasminogen expression in such extrahepatic tissues not exposed to circulating plasminogen. Kainic acid stimulates plasminogen mRNA and protein levels in rodent hippocampal neurons (5,6), and interleukins-1␣ and -1␤ increase levels of plasminogen mRNA and protein in the cornea (7). Thus, regulation of the plasminogen gene may be particularly important at these sites of extrahepatic plasminogen synthesis. Analysis of regulatory elements within the murine plasminogen 5Јflanking region that modulate both extrahepatic constitutive plasminogen synthesis and stimulation of plasminogen synthesis by inflammatory mediators in both liver and in extrahepatic cells is a promising new area of investigation that should provide key insights into the physiologic and pathophysiologic functions of plasminogen.