Cloning and Characterization of the Rat Lysyl Oxidase Gene Promoter

Lysyl oxidase (LO) stabilizes the extracellular matrix by cross-linking collagen and elastin. To assess the transcriptional regulation of LO, we cloned the 5′-flanking region with 3,979 bp of the rat LO gene. LO transcription started at multiple sites clustered at the region from –78 to –51 upstream of ATG. The downstream core promoter element functionally independent of the initiator predominantly activated the TATA-less LO gene. 5′ Deletion assays illustrated a sequence of 804 bp upstream of ATG sufficient for eliciting the maximal promoter activity and the region –709/–598 exhibiting strongly enhancing effects on the reporter gene expression in transiently transfected RFL6 cells. DNase I footprinting assays showed a protected pattern existing in the fragment –612/–580, which contains a nuclear factor I (NFI)-binding site at the region –594/–580 confirmed by electrophoretic mobility supershift assays. Mutations on this acting site decreased both NFI binding affinity in gel shift assays and stimulation of SV40 promoter activities in cells transfected with the NFI-binding site-SV40 promoter chimeric construct. Furthermore, at least two functional NFI-binding sites, including another one located at –147/–133, were identified in the LO promoter region –804/–1. Only NFI-A and NFI-B were expressed in rat lung fibroblasts, and their interaction with the LO gene was sensitively modulated by exogenous stimuli such as cigarette smoke condensate. In conclusion, the isolated rat LO gene promoter contains functionally independent initiator and downstream core promoter elements, and the conserved NFI-binding sites play a critical role in the LO gene activation.

enzyme catalyzes the initiation of cross-linking of collagen and elastin, major structural components of the extracellular matrix (ECM), by oxidizing peptidyl lysine residues within these proteins to peptidyl ␣-aminoadipic-␦-semialdehyde, leading to the formation of condensation products stabilizing polymeric collagen or elastin as insoluble fibers. Thus, LO plays a central role in ECM morphogenesis and tissue repair (1).
In addition to the major function in stabilizing the ECM, LO also exhibits other biological activities. As reported, expression of transfected LO cDNA inhibited Ha-ras-induced cell transformation indicating an anti-tumorigenic effect of LO (2). LO can oxidize lysine residues in various globular proteins other than collagen and elastin (1). Oxidation of basic fibroblast growth factor (bFGF) by LO blocks the proliferation of bFGFstimulated cells and highly tumorigenic bFGF autocrine-transformed cells (3). Purified mature bovine LO displays chemotactic activity for monocytes and vascular smooth muscle cells (4,5). LO and its oxidized substrates exist within the nuclei, potentially using histone H1 as a substrate and modulating the promoter activity of the collagen type III gene (6 -8). Increased LO activity is associated with fibrotic diseases such as lung and liver fibrosis and atherosclerosis (1), whereas decreased LO activity is associated with lathyritic agent-induced emphysema in animal models (9) and with disorders of copper metabolism like Menkes syndrome (10).
LO gene expression is regulated at the mRNA level in response to intra-or extracellular agents or conditions. LO transcription is up-regulated by transforming growth factor-␤1 and interleukin-1␤ (11,12) but down-regulated by bFGF and interferon-␥ (13,14). Steady-state LO mRNA levels were diminished in some cancer and transformed cell lines (1, 2) but elevated in invasive breast cancer cells (15). In addition, environmental agents can also act upon the LO gene resulting in its transcriptional modification. For example, our recent studies indicated that cigarette smoke condensate (CSC), the particulate phase of smoke, inhibited synthesis of nascent LO transcripts leading to reduced levels of LO mRNAs in treated rat fetal lung fibroblasts (16). Transcription of the LO gene is driven by proximal promoter sequences, which interact with ubiquitous transcription factors. A region of 1,865 bp of the rat LO promoter upstream of ATG was found to direct luciferase reporter gene expression. CSC suppressed luciferase activity in cells transiently transfected with the LO promoter-luciferase gene chimeric vector consistent with its effects on LO mRNA levels (16). Although several cis-elements have been characterized in the LO promoter region for human, mouse, and rat (17)(18)(19), the precise mechanisms for their activation remain to be understood. Because the rat LO gene is widely used as a model for studies on the expression and regulation of the LO gene, it is important and necessary to further define the rat LO gene promoter.
Nuclear factor I (NFI) was originally described as a factor required for the replication of adenovirus DNA and then shown as a transcription factor widely involved in the regulation of constitutive or inducible gene expression, including both transactivation and repression (20). NFI encoded by four different genes (nfi-A, nfi-B, nfi-C/CTF, and nfi-X) binds to the consensus sequence TTGGC(N5)GCCAA (N indicates any nucleotide) on duplex DNA as a dimer. Notably, NFI can also bind to the individual half-site, i.e. TTGGC or GCCAA, with a somewhat reduced affinity. The highly conserved N-terminal residues contain the DNA binding domain, whereas the proline-rich C-terminal residues constitute the transcriptional regulation domain (20). Four cysteine residues are conserved in all DNA binding domains of NFI isoforms, but only three cysteines are required for the DNA binding activity. The fourth cysteine nonessential for DNA binding makes NFI proteins sensitive to oxidative damage. Exposure of cells to H 2 O 2 inhibited the binding of NFI to its DNA consensus sequence (21). Mutation of the fourth cysteine induced NFI resistance to oxidative inactivation (22). The feature of oxidation sensitivity of NFI may play a critical role in the cellular response to oxidative stress (23). NFI has been reported to regulate the expression of a wide range of cellular and viral genes such as collagen (24), cytochrome P450 1A1 (21,25), metallothionein-I (26), mouse mammary tumor virus, etc. (27).
To further understand the control of LO gene transcription, we have cloned the 5Ј-flanking region of the rat LO gene, mapped its transcription start sites and core promoter elements, characterized the regulatory activities in transcription of its different subsegments, and demonstrated NFI as a critical transactivator interacting with the cognate cis-element in the promoter region for LO gene activation.

MATERIALS AND METHODS
Cell Culture-The rat fetal lung fibroblasts (RFL6) obtained from the ATCC were cultured in Dulbecco's modified Eagle's medium (DMEM) supplemented with 10% fetal bovine serum (FBS) at 37°C in a 5% CO 2 and 95% air incubator as described previously (28). Stock cultures were derived from the frozen cell line and passaged every 4 days. A total of six passages was used for experiments.
Primer Extension Analysis-To identify start sites of the LO transcription, we first carried out a primer extension assay as described (29). Total RNA was isolated from RFL6 cells with TRIzol reagent (Invitrogen). An antisense oligodeoxynucleotide, 5Ј-ATGATGCTCCCGGCTCGTCCCTGCT-3Ј (Integrated DNA Technologies, Coralville, IA), corresponding to positions from Ϫ23 to ϩ2 in the rat LO sequence using the first nucleotide preceding the ATG codon as Ϫ1, was labeled with [␥-32 P]ATP (PerkinElmer Life Sciences) by T4 polynucleotide kinase (New England Biolabs). Approximately 20,000 -40,000 cpm of labeled oligonucleotides were annealed with 20 g of total RNA in 5ϫ M-MLV first strand buffer (Invitrogen) and adjusted to a final volume of 19 l with diethyl pyrocarbonatetreated double distilled H 2 O. Samples were heated at 95°C for 1 min and subsequently incubated at 48°C for 45 min. Then 2 l of 0.1 M dithiothreitol, 8 l of 2.5 mM dNTP, and 40 units of M-MLV RT enzyme (Invitrogen) were added into the reaction mixture. After incubation at 37°C for 30 min, samples were mixed with 8 l of a formamide sequencing gel loading buffer, heated at 90°C for 1 min, and then chilled rapidly on ice. Aliquots of samples were analyzed on a 6% urea/acrylamide sequencing gel. The transcription start sites were identified by the sequencing ladders directly parallel to the run-off reverse transcripts (29).
5Ј-RLM-RACE-The assay for 5Ј-RNA ligase-mediated rapid amplification of cDNA ends (5Ј-RLM-RACE) was performed to confirm the rat LO transcription start sites with the First Choice TM RLM-RACE kit (Ambion, Austin, TX) following the manufacturer's instructions (30). Briefly, 10 g of total RNA extracted from RFL6 cells was treated with calf intestinal phosphatase to remove the free 5Ј-phosphate group. Tobacco acid pyrophosphatase was then used to specifically remove the cap structure from the full-length mRNA, leaving a 5Ј-monophosphate. Note that non-tobacco acid pyrophosphatase-treated RNAs were also included in experiments as negative controls. An RNA oligonucleotide adaptor was next ligated to the newly decapped mRNA by T4 RNA ligase. With the ligated RNA as a template, LO cDNA was synthesized by reverse transcription using M-MLV reverse transcriptase and (dT) 15 primers. The resulting cDNA was then amplified by nested PCR using Platinum Taq High Fidelity DNA polymerase (Invitrogen) as well as the rat LO gene primers (reverse) and the adaptor primers (forward) provided by the manufacturer. The gene-specific antisense inner primer 5Ј-GACTCTCGAGGGTTGTCACGCAG-CAGCAGAATGG-3Ј and the nested PCR outer primer 5Ј-CAGATGGGCTTGGAGTCCTC-3Ј were designed for the RLM-RACE assay based on the sequence of rat LO cDNA (16,31). The 5Ј-RLM-RACE PCR products were analyzed on agarose gels and cloned into the pBluescript II SK (Ϫ) vector for sequencing as described (30).
To evaluate functionalities of the initiator (Inr), the downstream core promoter element (DPE) and the NFI-binding sites in the LO gene activation, we also isolated two other LO promoter 5Ј deletions such as Ϫ80/Ϫ1 and Ϫ160/Ϫ1 containing Inr and DPE core promoter elements and NFI-binding sites, respectively. The reverse primer used for PCR amplification was the same as described above corresponding to the region from Ϫ23 to ϩ2 of the rat LO gene sequence. The forward primers for the LO promoter Ϫ80/Ϫ1 and Ϫ160/Ϫ1 were 5Ј-GATCGAGCTCTCCTTCGCGGGATCTGAGTC-3Ј and 5Ј-GATCGAGCTCCGGCCGCTCGCCCTTGGCAC-3Ј, respectively. LO promoter-luciferase reporter constructs were created with the pGL3-Basic vector. Using these newly and previously prepared LO promoter-reporter gene constructs compassing sequences Ϫ80/Ϫ1, Ϫ160/Ϫ1, and Ϫ804/Ϫ1 (see Fig.  3A and Fig. 8C) as DNA templates, mutations of the Inr, the DPE, or the NFI-binding sites were performed according to the QuikChange mutagenesis protocol (Stratagene, La Jolla, CA). Mutagenic primer pairs used for the PCR amplification were 5Ј-GATCTGAGTCCCTGTCTTGGTGTTTCTCCTAGC-CACGTCC-3Ј for the Inr mutagenesis, 5Ј-CGTCCCTCCC-CGAGAAGCCCCGAGCCGGGAGCATC-3Ј for the DPE mutagenesis, 5Ј-CTGCCGCTCGCCCTGAACACCAGTCC-CTGCGACC-3Ј for the NFI-binding site 1 mutagenesis, and 5Ј-CTTCATGCATATTTGAACTTGGGCCCATGGCCTG-GCTG-3Ј for the NFI-binding site 2 mutagenesis (complementary reverse primers not shown, mutated nucleotides labeled with underlines). Each mutation was verified by direct sequencing.
To further characterize the properties of specific regions of the LO promoter, synthetic oligonucleotides (Integrated DNA Technologies, Coralville, IA) of LO promoter fragments Ϫ709/ Ϫ676, Ϫ682/Ϫ641, Ϫ654/Ϫ609, and Ϫ612/Ϫ580, as shown in Fig. 5A, and the fragment Ϫ609/Ϫ573 containing the NFIbinding motif or various mutants as well as the NFI-binding consensus sequences, as shown in Fig. 6B, were annealed each to its complementary sequence, phosphorylated with T4 polynucleotide kinase, and ligated into the SmaI-treated pGL3-Promoter reporter vector (Promega, Madison, WI) upstream of the SV40 promoter sequence.
Each resulting recombinant construct and the pSV-␤-galactosidase plasmid (Promega, Madison, WI) or pRL-TK, an internal control for monitoring the transfection efficiency, were transiently cotransfected into RFL6 cells by Lipofectamine reagents (32). After a 24-h post-transfection, cells were growtharrested by incubation with 0.3% FBS/DMEM for an additional 24 h. In some experiments, cells were growth-arrested for 6 h and then exposed for 24 h to agents such as CSC at various doses in the FBS-free medium. Note that cells cotransfected with pGL3-Basic or pGL3-Promoter containing the luciferase gene without the LO promoter and ␤-galactosidase vectors were always included in any experiment to evaluate the background. Luciferase and ␤-galactosidase activities in cell lysates were measured by luminometry and spectrophotometry, respectively, as described by manufacturers. LO promoter-luciferase activities in transfected cells were normalized to the transfection efficiency as revealed by the ␤-galactosidase assay. Results are expressed as cpm/optical density of ␤-galactosidase. In experiments as shown in Fig. 3B, Fig. 8D, and Fig. 9B using the plasmid pRL-TK as an internal control, the LO promoterdirected expression of the firefly luciferase gene in the LO promoter-pGL3-Basic construct was normalized to Renilla luciferase activities derived from the pRL-TK vector and expressed as relative luciferase activities according to the instructions from the supplier (Promega, Madison, WI).
Nuclear Extract Preparation and EMSA-Nuclear extracts were prepared from rat RFL6 cells using nuclear and cytoplasmic extraction kit (Pierce). Protein concentrations were determined by the BCA protein assay reagents (Pierce). For the EMSA (33), synthetic oligonucleotides as shown in Fig. 6B and Fig. 8A were end-labeled with [␥-32 P]ATP (PerkinElmer Life Sciences) by T4 polynucleotide kinase (New England Biolabs) and annealed to their complements. A total volume of 20 l of reaction mixture containing 20 g of nuclear protein or bovine serum albumin (BSA), a negative control, 1 g of poly(dI-dC)⅐poly(dI-dC) (Sigma), and 10,000 -20,000 cpm of labeled probes was incubated for 20 min at room temperature. For competition experiments, unlabeled cold oligonucleotides such as the NFI-binding site wild type and various mutants, and NFIbinding site consensus sequences as shown in Fig. 6B at 100-fold molecular excess were added 10 min prior to addition of the radiolabeled probe. After the reaction, samples were subjected to native 4% PAGE and visualized by exposure of the dried gel to Kodak film. Supershift reactions were run as competition assays as described above with the exception that 2 g of a corresponding specific antibody against Oct1 (Biovision, Mountain View, CA), CdxA (Cemines, Goldon, CO), Pit-1A, or all NFI isoforms (Santa Cruz Biotechnology, Santa Cruz, CA) instead of cold probes was added, as shown in Fig. 6A and Fig.  8B. In the supershift combined with competition assay (Fig.  6D), competitors were added into the reaction mixture before addition of the antibody.
DNase I Footprinting-For the DNase I footprinting assay (34), a synthetic oligonucleotide of the LO promoter fragment Ϫ612/Ϫ580 was end-labeled with [␥-32 P]ATP (PerkinElmer Life Sciences) by T4 polynucleotide kinase. Twenty nmol of DNA was incubated for 30 min with 20 g of nuclear extract or BSA, a negative control. Samples were then treated with different concentrations of DNase I (1.0 -0.25 units; Sigma) for 5 min. Reactions were stopped by addition of DNase I stop solution (1% SDS, 200 mM NaCl, 20 mM EDTA, pH 8.0, and 40 g/ml tRNA). Reaction mixtures were phenol/chloroform-extracted and followed by ethanol precipitation. Resulting DNA samples were mixed with 5 l of loading buffer (95% (v/v) formamide, 10 mM EDTA, pH 8.0, 0.1% (w/v) bromphenol blue, 0.1% (w/v) xylene cyanol) and analyzed on 6% acrylamide gel in 1ϫ TBE buffer with a Sequi-Gen Sequencing Cell (Bio-Rad). Single strand DNA ladder (10/60) (Integrated DNA Technologies, Coralville, IA) labeled with [␥-32 P]ATP was electrophoresed alongside the digestion products. The gel was dried and exposed to Kodak film.
Reverse Transcription (RT)-PCR and Preparation of NFI Expression Vectors-To assess the expression of NFI isoforms in rat lung fibroblasts, RT-PCR was carried out as described (28). Briefly, approximately 500 ng of total RNA extracted from RFL6 cells by TRIzol were converted to cDNA, which was then amplified using the SuperScript One-Step RT-PCR with Platinum Taq kit (Invitrogen) under the following conditions: reverse transcription at 50°C for 30 min and predenaturation at 94°C for 2 min, denaturation at 94°C for 30 s, annealing at 60°C for 30 s, and extension at 72°C for 90 s in a total of 35 repetitive cycles. The final extension was performed at 72°C for 6 min. The forward (F) and reverse (R) primers used were as follows: 5Ј-GACTCAGCTGATGTATTCTTCGCTCTGTC-TCA-3Ј (F) and 5Ј-GACTCTCGAGTATCCCAGGTACCAG-GACTGTG-3Ј (R) for amplification of nfi-A cDNA; 5Ј-GACT-AAGCTTATGATGTATTCTCCCATCTGTC-3Ј (F) and 5Ј-GACTCTCGAGTTGCTTGTCTCCGCTTGAAGG-3Ј (R) for amplification of nfi-B cDNA; 5Ј-GACTAAGCTTATGTATT-CCTCCCCGCTCTGCC-3Ј (F) and 5Ј-GACTCTCGAGCCC-AGGTACCAGGACGGTCCTG-3Ј (R) for amplification of nfi-C cDNA; and 5Ј-GACTAAGCTTATGATTGGGGAGCA-GCAGAGAA-3Ј (F) and 5Ј-GACTCTCGAGAAAGTTGCTG-TCCCGGGATCC-3Ј (R) for amplification of nfi-X cDNA (restriction sites in primer pairs are labeled with underlines). PCR products were separated on agarose gels, stained with ethidium bromide, and visualized on a UV transilluminator. The mammalian NFI expression vectors pcDNA3.1-nfi-A and pcDNA3.1-nfi-B were, respectively, constructed by ligating the rat nfi-A and nfi-B coding sequences with the pcDNA3.1/V5-His-TOPO vector (Invitrogen) in their PvuII/XhoI sites for nfi-A and HindIII/XhoI sites for nfi-B. All expression vectors were sequenced to ensure fidelity and subjected to cotransfection into RFL6 cells with the LO promoter-reporter gene vector for assaying their capacities for the gene activation.
Chromatin Immunoprecipitation (ChIP) Assay-To determine cellular NFI binding to the LO promoter region, the ChIP assay was performed with the EpiQuik chromatin immunoprecipitation kit based on the protocol provided by the supplier (Epigentek Group Inc., Brooklyn, NY). Cellular components were cross-linked by incubation of control and CSC-treated cells at the same number (2 ϫ 10 6 ) with 1% formaldehyde at room temperature for 10 min. The cross-linking reaction was stopped by addition of glycine to a final concentration of 125 mM. Nuclei were extracted with a nuclear isolation buffer, resuspended in a nuclear lysis buffer with protease inhibitor mixture, and then sonicated to shear DNA to lengths between 200 and 1000 bp. After centrifugation, cell debris was discarded, and DNA-containing supernatants were diluted with the ChIP dilution buffer, and aliquots of samples were removed out as "input" DNA. Diluted DNA samples were transferred into the strip wells that were precoated with the antibody against rat NFI (Santa Cruz Biotechnology), RNA polymerase II (a positive control provided by the kit supplier), or nonspecific rat IgG (a negative control from Santa Cruz Biotechnology) and incubated at room temperature for 90 min with shaking. After successively washing wells with the washing buffer and finally with the TE buffer (10 mM Tris-HCl, pH 8.0, 1 mM EDTA), precipitated DNA-protein complexes as well as the input samples were treated with proteinase K (250 g/ml) in the DNA release buffer for 15 min and then incubated in the reverse buffer for 90 min at 65°C. The DNA samples were collected by the P-spin columns, washed with 70 and 90% ethanol successively, and then eluted with the elution buffer. Using purified DNA as a template, PCR was conducted under the following conditions: initial denaturation at 94°C for 2 min, 30 cycles each with denaturation at 94°C for 30 s, annealing at 55°C for 30 s, and extension at 72°C for 1 min, and final extension at 72°C for 5 min. Three forward (F) and reverse (R) primer pairs were used in PCR to characterize the NFI binding to the LO gene and the RNA polymerase II binding to the glyceraldehyde-3-phosphate dehydrogenase (GAPDH) gene including the following: pair 1, 5Ј-GGAAAGGGGAGAGGAGGAC-3Ј (F) and 5Ј-AGGAGG-GAGACCTCTTCGAG-3Ј (R) for amplification of the LO NFIbinding site 1 fragment; pair 2, 5Ј-CCAGGGCTGGTGAC-CTAATA-3Ј (F) and 5Ј-GACTTAATCTGGGCCGAACA-3Ј (R) for amplification of the LO NFI-binding site 2 fragment; and pair 3, 5Ј-GATGTTAGCGGGATCTCGCTCCTG-3Ј (F) and 5Ј-GTTCAACGGCACAGTCAAGGCTGAG-3Ј (R) for amplification of the RNA polymerase II binding region in the GAPDH promoter. PCR products were analyzed on a 2.2% agarose gel, stained with ethidium bromide, and visualized on a UV transilluminator. Densities of PCR-amplified gene fragments on the gel were measured with the 1D Scan software as described (28).
Statistical Analysis-Data were expressed as mean Ϯ S.D. of at least three independent experiments. Statistical differences between means were determined using one-way analysis of variance followed by Bonferroni's post hoc test or two-tailed Student's test when appropriate. A p value Ͻ 0.05 was considered significant.

RESULTS
Isolation of the 5Ј-Flanking Region of the Rat LO Gene-A genomic DNA fragment containing 3,979 bp compassing the 5Ј-flanking region of the rat LO gene upstream of the translation start codon ATG was isolated by PCR from RFL6 cells and restricted into the pGL3-Basic vector. This DNA fragment was shown below to drive the expression of the luciferase reporter gene indicating its promoter property. Sequence analysis revealed that the rat LO promoter region lacks canonical TATA and CAAT boxes. It contains putative binding sites for the following transcription factors: NFI, GR, SP1, Oct-1, TAF-1, C/EBP␥, etc. In addition, the cloned rat LO promoter also contains the metal-response element, the hypoxia-response element, the antioxidant-response element, etc. (Fig. 1). Comparative sequence alignment across species of rat, human, and mouse showed that the proximal promoter region was highly conserved, whereas the distal promoter region was more divergent among three species (Fig. 1). For example, regions Ϫ410/Ϫ1 (as shown here and below indicating the fragment between two nucleotide numbers using the first nucleotide preceding the ATG codon as Ϫ1), Ϫ804/Ϫ411, and Ϫ1,336/Ϫ805 displayed 75.4, 38.6, and 24.8% homology, respectively, in sequences of rat, human, and mouse ( Fig. 1). These analyses suggest that the proximal promoter region might play a fundamental role in the regulation of LO transcription in vivo.
The Transcription Start Sites of the Rat LO Gene-To identify start sites of the LO transcription, we first carried out a primer extension assay (29). Total cellular RNA isolated from RFL6 cells was incubated with the ␥-32 P-labeled LO antisense primer corresponding to positions from Ϫ23 to ϩ2 in the rat LO sequence using the first nucleotide preceding the ATG codon as Ϫ1 in the presence of reverse transcriptase. Following analysis of the reaction product ( Fig. 2A, lanes 5 and 6) on the sequencing gel, the transcription start sites were identified by the sequencing ladders ( Fig. 2A, lanes 1-4) directly parallel to the run-off reverse transcripts. As shown ( Fig. 2A), there was a conspicuous extension product parallel to the position of cytosine (labeled with underline) at the sequence CCCTG (arrow) in addition to several minor extension products (arrowheads) on the gel. Thus, the cytosine at the sequence of CCCTG is a major transcription start site which was then mapped at Ϫ61 bp upstream of ATG in the LO promoter region.
Furthermore, the transcription start sites of the LO gene were confirmed by 5Ј-RLM-RACE analysis (Fig. 2B) (30), which has proven to be a very sensitive and accurate method to identify full-length 5Ј ends of cDNAs by eliminating truncated messages from the amplification reactions. The advantage of this method over others is that only authentic capped 5Ј ends of mRNA are detected. After reverse transcription and nested PCR, a band of about 450 bp was revealed by the agarose gel (Fig. 2B). Notably, using identical procedures, non-tobacco acid pyrophosphatase-treated RNAs, a negative control, did not induce PCR products (data not shown). Resulting RT-PCR product was cloned into pBluescript II SK(Ϫ) vector. Sequencing of 16 clones showed that the initiations of LO transcripts were clustered at the region from Ϫ78 to Ϫ51 (Fig. 2C). In addition to two major start sites at the Ϫ51 adenosine and the Ϫ61 cytosine (31.3% of tested clones each, i.e. 5/16), other potential start sites for the LO transcription included Ϫ54 (6.3%), Ϫ55 (12.5%), Ϫ57 (6.3%), Ϫ60 (6.3%), and Ϫ78 (6.3%) upstream of the ATG. Notably, the transcription start site at the Ϫ51 adenosine overlaps with the sequence TCATTTT (labeled with underline) identical with the Inr element consensus sequence YYAN(T/A)YY (Y is pyrimidine; N is any nucleotide) (36,37). Furthermore, a sequence 5Ј-GGACG-3Ј is located at the region from Ϫ18 to Ϫ14 upstream of ATG consistent with the consensus sequence of the DPE, (A/G)G(A/T)(C/T)(G/ A/C) (37). Thus, a potential Inr core promoter element combined with the DPE was mapped in the LO promoter region (Fig. 1). As indicated above, the Ϫ61 cytosine as another major transcription start site has also been demonstrated by the primer extension assay. Thus, multiple transcription start sites exist in the rat LO gene.
Roles of the Inr and the DPE in the LO Promoter Activation-To evaluate roles of the Inr and the DPE in the LO gene activation, we introduced site-directed mutations at the Inr, the DPE, or both in the LO promoter-reporter gene construct spanning from Ϫ80 to Ϫ1 (Fig. 3A). LO core promoter activities were tested by transfection assays in RFL6 cells. As shown in Fig. 3B, mutations of the Inr consensus TCATTTT (Ϫ53/Ϫ47) to TCGGTGTT and the DPE consensus GGACG (Ϫ18/Ϫ14) to CCCCG (mutations are labeled with underlines) reduced the luciferase gene expression to 58 and 1.4% of the wild-type control, respectively, whereas double mutations decreased the promoter activity to 1.8% of the wild-type control. These results suggest that the Inr and the DPE did not work as one combined unit initiating the LO gene transcription because mutation of the Inr only partially inhibited the reporter gene expression. Furthermore, the DPE may represent a more important core promoter to activate the LO promoter in comparison to the Inr because its mutation essentially abolished the luciferase activity. Thus, the LO promoter contains functionally active Inr and DPE sequences, which act as independent core promoter modules for the LO gene transcription.
Deletion Analysis of the LO Promoter-To characterize the 5Ј-regulatory region of the LO gene, we further prepared a set of luciferase reporter gene constructs containing successive 5Ј deletions of the LO promoter (Fig. 4A). After transient transfection into RFL6 cells, the transcription activities derived from these constructs were tested. The background luciferase expression levels were evaluated by transfection of equal amounts of the pGL3-Basic reporter vector without the LO promoter insert. A series of systematic luciferase assays as shown in Fig. 4B indicated that a high level of the LO promoter activity was found in the construct PromϪ804 reaching 168% of the full-length control (PromϪ3,979). When the sequence length extended from Ϫ804 to Ϫ3,237, no significant alterations of luciferase activities were observed. Thus, critical control elements that regulate LO gene expression may not be pres-  (35). Gaps introduced to maximize similarity are indicated by dashes. Asterisks represent identical nucleotides. Translation start sites are labeled as ϩ1. Putative transcription factor-binding sites and cis-elements are labeled above the sequence. Core promoters Inr and DPE are labeled with boxes. AUGUST 31, 2007 • VOLUME 282 • NUMBER 35 ent in the region from Ϫ3,237 to Ϫ804. Inclusion of more nucleotides up to Ϫ3,979 from Ϫ3,237 resulted in a 60% decrease in luciferase activities relative to the full-length control, suggesting the presence of inhibitory cis-elements in this distal 5Ј-flanking region. More 5Ј deletions of nucleotides from the PromϪ804 induced reduction of luciferase activities such that the PromϪ709, Ϫ598, Ϫ410, and Ϫ277 decreased LO promoter activities to 109, 36, 22, and 10% of the full-length control, respectively. Note that there was one exception, i.e. the PromϪ517, which restored 17% promoter activity relative to that of the PromϪ598. Furthermore, deletions of the 5Ј-proximal regions upstream of ATG from the PromϪ1,865 totally abolished luciferase activities as shown by those in constructs PromϪ1,865/Ϫ804, PromϪ1,865/Ϫ911, and PromϪ1,865/ Ϫ1,335. It should be noted that the construct PromϪ804 significantly enhanced the expression of the luciferase gene reaching 75-fold of the pGL3-Basic control, whereas the construct PromϪ277 displayed only 2-fold of the basal level of luciferase activity in comparison to the pGL3-Basic. Thus, it appears that the major role of the fragment between Ϫ804 and Ϫ277 is transactivation of the LO gene. Taken together, these results suggest that there are at least four positive regulatory segments (i.e. Ϫ804/Ϫ709, Ϫ709/Ϫ598, Ϫ598/Ϫ410, and Ϫ410/Ϫ277) and two negative regulatory segments (i.e. Ϫ3,979/Ϫ3,237 and Ϫ598/Ϫ517) identified in the full-length, cloned rat LO promoter and that the sequence from Ϫ1 to Ϫ277 as a basic region and the sequence from Ϫ277 to Ϫ804 as a regulatory region both are essential for LO promoter functions.

Regulation of the Rat Lysyl Oxidase Gene Promoter
Screening cis-Elements in the Region from Ϫ709 to Ϫ598 bp of the Rat LO Promoter-As shown in deletion analysis (Fig. 4B), the upstream extension of the LO promoter from Ϫ598 to Ϫ709 markedly increased the pGL3-Basic luciferase activity reaching 3.2-fold of the former level, suggesting the critical role of this region in the regulation of the LO gene expression. To characterize cis-elements and their corresponding transcription factors in this region, we synthesized four oligonucleotides with overlap in their sequences, i.e. Ϫ709/Ϫ676, Ϫ682/Ϫ641, Ϫ654/Ϫ609, and Ϫ612/Ϫ580 (Fig. 5A), and we examined their functional activities. Each oligonucleotide was annealed to its complementary sequence and inserted into the pGL3-Promoter reporter vector upstream of the SV40 promoter sequence. Resulting constructs were transiently transfected into RFL6 cells and the expression of the reporter gene was determined. As shown (Fig. 5B), pSVϪ709/Ϫ676, pSVϪ682/ Ϫ641, pSVϪ654/Ϫ609, and pSVϪ612/Ϫ580 constructs all  1-4) were produced from PromϪ804 with the same oligonucleotide primer (RevϪ23/ϩ2). Notably, experiments shown here and below were repeated at least three times with reproducible results, and a representative one is presented. B, mapping of LO transcription start sites by RLM-RACE. Agarose gel electrophoresis of nested PCR products from the RLM-RACE procedure using decapped rat RFL6 total RNA ligated with an RNA oligonucleotide adaptor as a template. Molecular weight (MW) markers (base pairs) are indicated on the left. The major PCR product was marked by two arrows on the right. C, multiple transcription start sites clustered in the rat LO promoter region. Line A, numbers indicate the relative nucleotide positions upstream of ATG. Line B, asterisks show transcription start sites identified by cloning and sequencing the RLM-RACE products. Line C, indicates the number of clones initiated at the corresponding site labeled with an asterisk in a total of 16 sequenced clones, such that Ϫ61 and Ϫ51 have 5 clones each. exhibited enhanced levels of luciferase activities in transfected cells amounting to 186, 216, 168, and 179%, respectively, of the pGL3-Promoter basal level. These results suggest the presence of enhancer cis-elements in these LO promoter fragments, and thus the oligonucleotide Ϫ612/Ϫ580 as a subregion of the LO promoter was first screened by this laboratory for its cis-element property.
The DNase I footprinting assay was performed as described (34) in an effort to explore the binding sites of transcription factors in the subregion Ϫ612/Ϫ580. As shown (Fig. 5C), RFL6 nuclear extracts (lanes 5-8), but not BSA (lanes 1-4), an internal control, induced a protection pattern in this LO promoter subregion. Precise positioning of the protected area from Ϫ598 to Ϫ588 was achieved by comparison of the footprints with 32 P-labeled single strand DNA molecular weight ladders (M in Fig. 5C). Computational analysis (TESS, TF Search and Alibaba 2.1) showed that the protected region may be a potential binding site for several transcription factors such as Oct1, CdxA, Pit-1A, NFI, etc.
Identification of NFI Binding to the Subregion Ϫ612/Ϫ580 of the LO Promoter-To identify transcription factors that bind to the putative cis-element Ϫ612/Ϫ580 as described above, we conducted EMSA and supershift assays (33) with antibodies against Oct1, CdxA, Pit-1A, and NFI. As shown (Fig. 6A), only the NFI antibody, rather than other antibodies tested, yielded a supershifted band, indicating that the putative cis-element in this subregion is a NFI-binding site. Because the putative NFI-binding site (TTG-GCTTGGGCCCAT, Ϫ594/Ϫ580) locates at the 3Ј terminus of the fragment Ϫ612/Ϫ580, we further synthesized the oligonucleotides encompassing the sequence from Ϫ609 to Ϫ573 of the LO promoter containing the putative NFI-binding site or its mutants as shown (Fig. 6B). These oligonucleotides were used as competitors as described (38) in gel mobility shift (Fig. 6C) and supershift assays (Fig. 6D) in the presence of the RFL6 nuclear extract using 32 P-labeled NF1-binding site-wild type (NFI-BS-WT) as a probe. As expected (Fig. 6C), a 100-fold molar excess of cold NFI-BS-WT (lane 7) as well as NFI-binding site-high affinity consensus (NFI-BS-High) (lane 3) and NFI-binding site-consensus (NFI-BS-Con1 and NFI-BS-Con 2) (lanes 3 and 4) (38) oligonucleotides competed successfully with labeled probes to bind to the NFI protein. In contrast, NFI-binding site mutant oligonucleotides such as NFI-BS-Mut (Fig. 6C, lane 6) (38), NFI-BS-Mut-1, NFI-BS-Mut-2, NFI-BS-Mut-3, and NFI-BS-Mut-4 (lanes 8 -11) that were derived from the LO promoter sequence lost, at least in part, their competitive capacities to block the formation of the labeled DNA-protein complex. Moreover, similarly competitive effects elicited by NFI-BS-WT (lane 4) or mutant oligonucleotides (lanes 5-9) were also observed in the gel supershift assay (Fig. 6D). Together, these results strongly support the hypothesis that the NFIbinding site exists in the Ϫ612/Ϫ580 fragment of the rat LO promoter, and the transcription factor, NFI, specifically binds to this cis-element.

. Luciferase activities in RFL6 cells transfected with LO promoter-deletion constructs.
A, schematic representation of LO promoter-deletion constructs. B, relative luciferase activities of LO promoter-deletion constructs. LO promoter-deletion constructs were transiently transfected into RFL6 cells. Twenty four hours after transfection, cells were growth-arrested in 0.3% FBS/DMEM for an additional 24 h and then harvested for assaying luciferase activities. The plasmid pSV-␤-galactosidase was cotransfected with LO deletion constructs for data normalization. Data were expressed as % of the full-length promoter control (100% luciferase activity in cells transfected with PromϪ3,979 ϭ 3,500 cpm/optical density of ␤-galactosidase). All values represent the mean Ϯ S.D. of three experiments, each determined with triplicate dishes.

Functionalities of the LO Promoter-NFI-binding Site Ϫ594/ Ϫ580 and Its Response to CSC in Transfected RFL6 Cells-To
verify the biological role of the NFI cis-element in the gene expression, we cloned the subregion Ϫ609/Ϫ573 of the LO promoter containing the NFI-binding motif Ϫ594/Ϫ580 or various mutants as described above (Fig. 6B) into the pGL3-Promoter vector upstream of the SV40 promoter. The NFIbinding site-directed luciferase expression was examined in transfected RFL6 cells. As shown (Fig. 7A), cells transfected with the wild-type NFI-binding site plasmid (pSV-NFI-BS-WT) and with the positive control (pSV-NFI-BS-High) exhibited enhanced levels of luciferase activities reaching 257 and 182% of the pGL3-Promoter control, respectively. However, a point mutation in the first half of the conserved NFI-binding site (pSV-NFI-BS-Mut-1 and pSV-NFI-BS-Mut-2) decreased luciferase activities to 155 and 143% of the pGL3-Promoter control, respectively, whereas a point mutation in both halves of the NFI-binding site (pSV-NFI-BS-Mut-3 and pSV-NFI-BS-Mut-4) further reduced luciferase activities to 122 and 130% of the pGL3-Promoter control, respectively. These results indicated that the NFI-binding site in the LO promoter region Ϫ609/Ϫ573 was functionally active in transfected cells.
Because CSC is known to modulate LO transcription (16), we further examined the NFI-binding element in response to CSC. RFL6 cells were transfected with pSV-NFI-BS-WT chimeric constructs and exposed to CSC at indicated doses for 24 h. CSC effects on the SV40 promoter activated by the NFI-binding element was determined by assaying levels of luciferase activities and expressed as fold of the pGL3-Promoter control. As shown (Fig. 7B), 40, 60, and 80 g/ml of CSC inhibited NFI-binding site-directed luciferase activities by 34, 50, and 58%, respectively, consistent with the down-regulating effect of CSC on the LO mRNA level. Thus, the NFI interaction with its binding site in the LO promoter region is a critical molecular target for CSC insult. Synthesized LO promoter oligonucleotides were inserted into pGL3-Promoter vector upstream of the SV40 promoter. The constructs were then transiently transfected into RFL6 cells. Twenty four hours after transfection, cells were growth-arrested in 0.3% FBS/DMEM for an additional 24 h and then harvested for assaying luciferase activities. The plasmid pSV-␤-galactosidase was cotransfected into cells for data normalization. Data were expressed as % of the pGL3-Promoter control (100% luciferase activity in cells transfected with the pGL3-Promoter vector ϭ 2,540 cpm/optical density of ␤-galactosidase). All values represent the mean Ϯ S.D. of three experiments, each determined with triplicate dishes. **, p Ͻ 0.01 compared with the pGL3-Promoter basic control. C, in vitro DNase I footprinting of the LO promoter subregion Ϫ612/Ϫ580. A synthesized fragment of the LO promoter from Ϫ612 to Ϫ580 upstream of ATG was end-labeled with [␥-32 P]ATP using T4 polynucleotide kinase. The labeled fragment was then allowed to bind to BSA, an internal control, or RFL6 nuclear extract (NE) and digested with 2, 1, 0.5, or 0.25 units of DNase I for 2 min at 25°C. The material was then electrophoresed and visualized by autoradiography. 32 P-Labeled single strand DNA molecular weight ladders were used as a size marker (M). Lanes 1-4, partial digestion of the probe in the presence of BSA. Lanes 5-8, partial digestion of the probe in the presence of RFL6 nuclear extract.

Evaluation of Biological Activities of All Putative NFI-binding Sites in the LO Promoter Region
Ϫ804/Ϫ1-Computational analysis (Alibaba2.1) (39) showed that there are three putative NFI-binding sites located at Ϫ147/Ϫ133, Ϫ594/Ϫ580, and Ϫ676/Ϫ662 in the LO promoter region from Ϫ804 to Ϫ1, which yielded the maximal reporter gene expression (Fig. 4B).  (36). The LO promoter fragment Ϫ599/Ϫ573 is labeled with uppercase letters. The NFI-binding site Ϫ594/Ϫ580 is underlined, and the mutated nucleotides in the NFI-binding site are marked with a boldface letter. C, NFI binding to the subregion Ϫ609/Ϫ573 of the LO promoter determined by EMSA competition assays. 32 P-Labeled synthetic LO promoter oligonucleotide probe Ϫ609/Ϫ573 containing NFI-BS-WT and 5 g of nuclear extracts (NE) prepared from RFL6 cells or BSA, an internal control, were incubated in the reaction mixture in the absence or presence of 100-fold molecular excess of competitors (Comp) as indicated. After the reaction, samples were analyzed on the native 6% polyacrylamide gel followed by autoradiography. D, NFI binding to the subregion Ϫ609/Ϫ573 of the LO promoter determined by EMSA supershift competition assays. Experiments were performed as described in C, and competitors were added into the reaction mixture before addition of the antibody against NFI (Ab-NFI).
The NFI-binding site (Ϫ594/Ϫ580) containing LO promoter fragment (Ϫ609/Ϫ573) was demonstrated to enhance the SV40 promoter activity (Fig. 7A). To examine whether other putative NFI-binding sites are also capable of incorporation with NFI and regulation of the LO gene expression, we carried out electrophoretic mobility shift and supershift assays with synthetic oligonucleotides spanning LO promoter sequences Ϫ158/ Ϫ123, Ϫ604/Ϫ570, and Ϫ686/Ϫ652, respectively, as shown in Fig. 8A. After incubation with nuclear extracts from RFL6 cells, labeled oligonucleotides containing Ϫ147/Ϫ133, Ϫ594/Ϫ580, but not Ϫ676/Ϫ662, induced mobility shift bands in the absence of and supershift bands in the presence of the antibody against NFI in gel electrophoresis (Fig. 8B). These results indicated that in addition to the region Ϫ594/Ϫ580, the region of Ϫ147/Ϫ133 is another one of the functional NFI-binding sites in the LO promoter Ϫ804/Ϫ1.
To further confirm biological activities of these two NFIbinding sites, we used two LO promoter-reporter constructs containing NFI-binding sites Ϫ147/Ϫ133 and Ϫ594/Ϫ580 as templates. Mutations of these NFI-binding sites individually or in their combination were performed by PCR (Fig.  8C). The LO promoter-directed reporter gene expression was monitored by transfection assays in RFL6 cells. As shown in Fig. 8D, extension of the LO promoter in length from Ϫ160 to Ϫ804 enhanced the expression of the reporter gene by 18-fold. Mutations of NFI-binding sites Ϫ147/Ϫ133, Ϫ594/Ϫ580, and Ϫ147/Ϫ133 plus Ϫ594/Ϫ580 reduced LO promoter activities to 20, 58, and 42% of corresponding controls, respectively. These results provide strong evidence that there are at least two biologically active NFI-binding sites in the LO promoter region Ϫ804/Ϫ1, and the NFI binds to this cis-element enhancing the LO gene expression.

The Expression of NFI Isoforms and Their Effects on LO Promoter Activation in Rat Lung
Fibroblasts-RT-PCR was performed to identify the expression of NFI isoforms in rat lung fibroblasts. As shown in Fig. 9A, although four primer pairs were designed specifically and used for amplification of different NFI isoforms, only 1.5-kb nfi-A and 1.3-kb nfi-B cDNA bands were produced following RT-PCRs indicating these two, rather than all four, NFI isoforms expressed in rat lung fibroblasts. To further assess the role of major NFI isoforms in LO promoter activation, the LO promoter (from Ϫ804 to Ϫ1)-luciferase chimera was cotransfected into RFL6 cells with NFI expression constructs containing the full-length cDNA encoding the major NFI proteins, i.e. NFI-A and NFI-B. As indicated (Fig. 9B), NFI-A and NFI-B both enhanced LO promoter activities in transfected cells reaching 3.5-and 2.7-fold of the pcDNA3.1 control, respectively, for individual transfection, and 5.8-fold of the pcDNA3.1 control, 1.7-fold of the NFI-A  Fig. 6B were inserted upstream of the SV40 promoter in the pGL3-Promoter reporter vector. Resulting chimeric constructs were cotransfected with the plasmid pSV-␤-galactosidase into RFL6 cells. After a 24-h post-transfection, cells were growth-arrested in 0.3% FBS/DMEM for an additional 24 h. Luciferase and ␤-galactosidase activities in cell lysates were determined. NFI-BS-WT or mutant-modified SV40 promoter-directed luciferase activities were normalized to the transfection efficiency. Data were expressed as % of the pGL3-Promoter control (100% luciferase activity in cells transfected with the pGL3-Promoter vector ϭ 2,850 cpm/optical density of ␤-galactosidase). All values represent the mean Ϯ S.D. of three experiments, each determined with triplicate dishes. **, p Ͻ 0.01 compared with the pGL3-Promoter basic control. B, CSC effects on the activation by NFI-BS-WT of the SV40 promoter. RFL6 cells were cotransfected with pSV-NFI-BS-WT chimeric construct and the plasmid pSV-␤galactosidase. After a 24-h post-transfection, cells were growth-arrested for 6 h in 0.3% FBS/DMEM and incubated for 24 h in FBS-free media in the absence or presence of CSC at the indicated doses. In parallel, cells cotransfected with the pGL3-Promoter and pSV-␤-galactosidase vectors were exposed to CSC under same conditions to evaluate CSC effects on the SV40 promoter basal activity. After incubation, luciferase and ␤-galactosidase activities in cell lysates were determined. Luciferase activity for each group was normalized to transfection efficiency. CSC modification of NFI-BS-WT-directed SV40 promoter activity was expressed as fold of the basal level of the pGL3-Promoter activity in cells exposed to CSC at the same dosage. Data were expressed as % of control in cells without CSC treatment (100% ϭ the luciferase activity in cells transfected with pSV-NFI-BS-WT chimeric construct Ϭ the luciferase activity in cells transfected with the pGL3-Promoter vector ϭ 2.53-fold). All values represent the mean Ϯ S.D. of three experiments, each determined with triplicate dishes. *, p Ͻ 0.05; **, p Ͻ 0.01 compared with the pSV-NFI-BS-WT basic control.
control, and 2.1-fold of the NFI-B control, respectively, for both transfections in combination. Significantly, such enhancement of LO promoter activity by NFI isoforms was blocked in transfected cells incubated in the presence of 80 g/ml CSC (Fig. 9B). These results provide additional evidence for NFI activation of the LO gene.
Cellular NFI Binding to the Cognate cis-Element in the LO Promoter-To determine cellular NFI binding to the LO promoter region, the ChIP assay was carried out using an antibody against rat NFI. As demonstrated, there were at least two functionally active NFI-binding sites located at the regions Ϫ147/ Ϫ133 (site 1) and Ϫ594/Ϫ580 (site 2) in the LO promoter region Ϫ804/Ϫ1 that displayed the maximal promoter activity (Fig. 4B and Fig. 8). Thus, these two active NFI-binding sites were tested for their cellular protein binding by using specific primer pairs as described above. Because CSC inhibited LO promoter and NFI-binding site-directed SV40 promoter activity (Fig. 7B) (16), the NFI binding status of LO DNA was examined in cells treated with CSC at various doses, and its specificity was evaluated by comparison with that of RNA polymerase II binding to the GAPDH promoter under the same conditions. According to the designed primer pairs, we anticipated that LO promoter fragments Ϫ218/Ϫ77 and Ϫ635/Ϫ516, which con-  (labeled with underlines). B, evaluation by EMSA. 32 P-Labeled synthetic LO promoter fragments containing putative NFI-binding sites as shown in A and 5 g of nuclear extracts (NE) prepared from RFL6 cells or BSA, an internal control, were incubated in the reaction mixture in the absence or presence of the specific antibody against NFI. After the reaction, samples were analyzed on the native 6% polyacrylamide gel followed by autoradiography. Reaction products are shown in lanes 1-3 with the probe NFI-BS1 (Ϫ158/Ϫ123), lanes 4 -6 with the probe NFI-BS2 (Ϫ604/ 570), and lanes 7-9 with the probe NFI-BS3 (Ϫ686/Ϫ652). C, schematic representation of LO promoter-reporter chimeras with or without NFI-binding site mutations. The site-directed mutations of NFI-binding sites were performed with the QuikChange mutagenesis kit. The mutated site is labeled with shaded circle with ϫ, and the nonmutated site is marked with shaded circle. D, relative luciferase activities of LO promoter-reporter constructs. The LO promoter-reporter constructs each as shown in C, and the pRL-TK vector, an internal control, were transiently cotransfected into RFL6 cells. Twenty four hours after transfection, cells were growth-arrested in 0.3% FBS/DMEM for an additional 24 h and then harvested for assaying the reporter gene expression. Firefly luciferase activities elicited by the LO promoter with or without mutations of NFI-binding sites were normalized to Renilla luciferase activities derived from the pRL-TK vector. Data shown are the mean Ϯ S.D. of three experiments, each determined with triplicate dishes. *, p Ͻ 0.05; **, p Ͻ 0.01 relative to wild type controls. tain active NFI-binding sites 1 and 2, respectively, would result following immunoprecipitation with the NFI antibody and amplification by PCR. Indeed, gel analysis of PCR products confirmed our expectation. Approximately 140-and 120-bp DNA bands were observed on gels as the PCR amplified using primer pairs that encompass NFI-binding sites 1 and 2, respectively (Fig. 10, A and D). In contrast, no signal was detected in control experiments with a nonspecific antibody (Fig. 10, B and E). These results suggest that the assay conditions were appropriate and can be used to measure the relative levels of NFI binding to the LO gene in response to stimuli. As shown (Fig. 10, A and  D), CSC that inhibited LO promoter activity (16) markedly reduced NFI incorporation with the cis-elements at either binding site 1 (Fig. 10A) or binding site 2 (Fig. 10D) within the LO promoter region Ϫ804/Ϫ1 in a dose-dependent manner. Density measurements indicated that CSC at 40, 80, and 120 g/ml decreased NFI incorporation with site 1 to 54, 28, and 33% of the control, and with site 2 to 53, 42, and 41% of the control, respectively. These changes apparently were not because of the difference of DNA amounts that were initially used for assays. As noted, the same cell number (2 ϫ 10 6 ) for each group was used for DNA extraction, and there was no significant difference in the yields of PCR products among groups using input (before immunoprecipitation) DNA as a template (Fig. 10, C and F). More importantly, reduction of NFI binding to the LO promoter by CSC may be a specific, epigenetic response because identical treatment of cells with CSC did not change RNA polymerase II binding to the promoter region Ϫ148/Ϫ58 (upstream of ATG) of the GAPDH gene (Fig. 10G). These results reflect the status of cellular NFI binding to the LO gene, which is highly sensitive to modification of exogenous stimuli.

DISCUSSION
Regulation of LO transcription is a critical control point for the LO gene expression involved in physiology and pathology of the ECM (1). This study presents the data for the rat LO gene promoter by addressing three issues. First, it identified the transcription start sites of the rat LO gene. Second, it characterized functional Inr and DPE core elements and other active subregions of the rat LO gene promoter. Third, it demonstrated the NFI-binding site as an important cis-element for transactivation of the rat LO gene.
To our knowledge, few previous studies reported structural and functional characteristics of the LO promoter in mouse (18,19), rat, and human (17). Comparing LO promoter activities in c-Ha-ras-NIH-3T3 fibroblasts (RS485), 3T6-5 myofibroblastlike cells, and vascular smooth muscle cells indicated the main positive and negative cis-acting regions and binding sites for several putative transcription factors in the mouse LO promoter (40). Transcription factor AP2 binding to the control region of the mouse promoter was identified by the DNase I footprint assay (19). Analysis of over 13 kb of intervening sequence and 5Ј-flanking region of the human LO gene revealed a concentration of conserved consensus sequence elements within the first intron and 1-kb fragment immediately 5Ј of exon 1 (17). Human LO promoter activity in luciferase reporter constructs transfected into rat aortic smooth muscle cells was markedly elevated by serum deprivation (41). However, the detailed transcriptional regulation of the LO gene is still poorly understood. Although the rat LO cDNA was cloned in 1990 (42,43) and rat cell lines were widely used as a model for studies of LO biology, the rat LO gene promoter has not been well defined. Comparatively little is known about transcriptional control of the rat LO gene. Thus, we cloned and characterized the rat LO gene promoter in an effort to understand mechanisms for the LO transcriptional regulation.
The 5Ј-flanking region of the rat LO gene appears to contain no typical TATA or CAAT box. The TATA box as the first core promoter element identified in RNA polymerase II-transcribed genes is generally present 25-30 bp upstream of the transcription start site (36). Primer extension and 5Ј-RLM-RACE assays indicated multiple transcription start sites clustered in the rat LO promoter region from Ϫ78 to Ϫ51 relative to ATG. One major transcription start site is located at the Ϫ51 adenosine residue. The sequence from Ϫ53 to Ϫ46, i.e. 5Ј-TCATTTTT-3Ј, overlapping with the transcription start site Ϫ51 adenosine (labeled with underline) is fully homologous to the consensus sequence of the Inr element present in many TATA-less promoters (36,37). Notably, although alternative transcription start sites with predicted TATA boxes have been reported, sites of most putative TATA boxes are located far upstream of the proximal promoter excluding their role in the LO transcription initiation (18,19). Moreover, other transcription start sites located at Ϫ54, Ϫ55, Ϫ57, Ϫ60, Ϫ61, and Ϫ78 identified in this study do not have any similarities with those published, indicating complex regulation of the rat LO gene transcription.
Furthermore, we found that in addition to the Inr element mapped at Ϫ53/Ϫ46 of the LO gene promoter, the sequence from Ϫ18 to Ϫ14, i.e. 5Ј-GGACG-3Ј, upstream of ATG perfectly matches the consensus sequence of the DPE that is located ϳ30 bp after the adenosine residue in the Inr motif (37). Apparently, core promoter elements of the Inr in conjunction with the DPE may exist in the rat LO promoter region (Fig. 1). A variety of transcription factors has been shown to interact with the Inr element in a sequence-specific manner, including TFIID, TFII-I, and YY1. RNA-polymerase II binds to the Inr element mediating the Inr-dependent transcription (37). The DPE was originally identified as a downstream core promoter binding site for purified Drosophila TFIID (44). It was reported that TFIID bound to the Inr and DPE motifs cooperatively, as mutation of either the Inr or the DPE resulted in loss of TFIID binding to the core promoter (44,45). Thus, the Inr and the DPE coordinately function as a single core promoter unit for the RNA polymerase II-directed gene transcription (37). However, our studies indicated that the DPE in the rat LO gene promoter worked as an independent core promoter module playing a key role in the LO gene transcription. This conclusion is based on the observation that site-directed mutation of the DPE fully abolished the LO promoter activity, but the same treatment of the Inr only partially inhibited the LO promoter activity (Fig. 3, A and B). Functions of the DPE were well characterized in Drosophila (44,45) but rarely reported in mammalian genes (44,46). Therefore, discovery of the predominant and independent property of the DPE in comparison with the Inr in the TATA-less rat LO gene may be an unusual finding in mammalian promoters. As shown in Fig. 1, Inr and DPE core elements in the rat LO promoter are conserved in mouse and human.
To determine whether the 5Ј-flanking region of the LO gene is functionally active for the regulation of transcription, an ϳ4-kb genomic fragment spanning nucleotides from Ϫ3,979 to Ϫ1 was inserted in front of the luciferase gene in the reporter gene vector (PromϪ3,979). The promoter activity was assessed by transient transfection of the PromϪ3,979 construct into RFL6 cells. As shown in Fig. 4B, the PromϪ3,979 induced a 50-fold increase over the pGL3 basal level in reporter gene expression indicating its LO promoter property. A series of 5Ј deletions of the PromϪ3,979 indicated that the smallest construct PromϪ277 in which the core promoter element Inr-DPE is located retained promoter activity amounting to at least 2-fold of the pGL3-Basic control. The region from Ϫ3,979 to Ϫ278 contains four positive and two negative regulatory segments in our assay system. The maximal luciferase activity was found in the 804-bp region immediately upstream of ATG as shown in the construct PromϪ804. Similar results were observed in the mouse and human LO gene promoters. As reported, an 808-bp region and a 924-bp fragment before the ATG were required for the highest reporter gene expression in fibroblasts transfected with the mouse and human LO promoter constructs, respectively (17,18). Considering the high degree of homology in the proximal LO promoter sequences associated with the lack of the typical TATA box in rat, mouse, and human, these results suggest that the LO transcription control is likely to be regulated by similar elements in these species.
Progressive 5Ј deletion assays indicated the sequence from Ϫ1 to Ϫ277 as a basic region and the sequence from Ϫ278 to Ϫ804 as a regulatory region, and both are essential for the maximal expression of the LO promoter activity. Because the expansion of sequences from Ϫ598 to Ϫ709 greatly increased luciferase activity (Fig. 4B), it was necessary to characterize this region for its transactivation ability. Four synthetic oligonucleotides derived from the sequence Ϫ709/Ϫ598 were cloned into pGL3-Promoter vector upstream of the SV40 promoter, and the resulting chimeric constructs were transiently transfected into RFL6 cells. Results showed that all four synthetic LO promoter oligonucleotides enhanced levels of the SV40 promoter activity (Fig. 5B) consistent with the data obtained from the 5Ј deletion assays (Fig. 4B). Screening this region using footprint, gel shift, supershift, and competition assays demonstrated an NFI-binding site, i.e. 5Ј-TTGGCTTGGGCCCAT-3Ј, at the region of Ϫ594/Ϫ580, and NFI binding to this sequence enhanced SV40 promoter activity (Fig. 7A). Mutating as few as one nucleotide in the LO promoter-NFI-binding site at the region of Ϫ594/Ϫ580 successfully reduced its competition for NFI binding (Fig. 6, C and D) and inhibited luciferase activities in chimeric construct-transfected RFL6 cells (Fig. 7A). These results strongly support the hypothesis that NFI is one of the transcription enhancers that binds to the cis-element located at Ϫ594/Ϫ580, facilitating LO transcription.
The LO promoter region Ϫ804/Ϫ1 that yielded the maximal activity contains at least three putative NFI-binding sites at regions of Ϫ676/Ϫ662, Ϫ594/Ϫ580, and Ϫ147/Ϫ133. To gain insight into biological activities of all putative NFI-binding sites in the LO promoter, we performed electrophoretic mobility shift and supershift assays. Results showed that except the oligonucleotides containing the sequence of Ϫ676/Ϫ662, two oligonucleotides with the sequences Ϫ594/Ϫ580 and Ϫ147/Ϫ133 incorporated with NFI proteins probed by the specific antibody. Furthermore, site-directed mutations of the consensus TTGGC to TGAAC at the sequences Ϫ594/Ϫ580 and Ϫ147/ Ϫ133 individually or in their combination strongly inhibited LO promoter activities. In toto, these results are consistent with the enhancement of SV40 promoter activity by the LO NFIbinding site Ϫ594/Ϫ580 in chimeric construct-transfected cells as shown above and further support the conclusion that the NFI-binding site acts as an enhancer cis-element transactivating the LO gene.
The NFI family consists of four highly conserved genes whose protein products, i.e. NFI-A, NFI-B, NFI-C, and NFI-X, are able to form the homo-or the heterodimer for DNA binding. The N-terminal domain confers dimerization and DNA binding properties on NFI proteins. NFI dimerization is essential for its DNA binding activity (20). The RT-PCR assays indicated only NFI-A and NFI-B expressed in rat lung fibroblasts. Cellular NFI binding to the LO promoter region was further illustrated by the ChIP assay. More importantly, cells cotransfected with NFI-A and NFI-B expression vectors in combination displayed a greater transactivation activity (Fig. 9B) sug-gesting that the heterodimer of NFI-A and NFI-B may play a key role in maximal LO gene expression. NFI regulates the gene expression by multiple mechanisms (20). The C-terminal proline-rich domain is required for transcriptional modulation of an NFI site-containing promoter in Drosophila, yeast, and mammalian cells (20). A motif of the heptapeptide SPTSPSY, existing in the C-terminal proline-rich domain of NFI-C, is homologous to the C-terminal domain repeat YSPTSPS, present in RNA polymerase II, critical for transactivation of a gene. This domain was shown to interact directly with human TFIIB or yeast TBP, triggering the transcription machinery (47,48). NFI-A that was expressed in rat lung fibroblasts as determined in this study contains a similar heptapeptide sequence, SPTSPTY. Thus, NFI-A may participate in transcription initiation of the TATA-less LO gene, particularly as it binds to the cis-acting element Ϫ147/Ϫ133 near the cluster of transcription start sites (Ϫ78/Ϫ51). Consistent with this suggestion, NFI-A was shown to have a stronger capacity in comparison with NFI-B to activate the LO promoter (Fig. 9B). Histone H1 inhibited the gene transcription by binding to the consensus NFIbinding sites. NFI proteins were suggested to activate the gene transcription by the direct competition with the repressor histone H1 for the DNA sites (49). In addition, a number of studies have indicated that the C-terminal proline-rich domain of NFI proteins can interact with a variety of transactivators such as TAFII55, Sp1, YY1, etc. (50). Notably, the C-terminal domain of NFI proteins was also involved in the repression of the gene expression possibly by recruitment of corepressor proteins or by interaction with the basal transcriptional apparatus (20).
The NFI regulation of gene expression is sensitive to oxidation because its DNA binding domain contains a cysteine residue susceptible for oxidative damage (21)(22)(23). We tested this possibility and showed that the luciferase reporter gene expression in RFL6 cells transfected with LO NFI-binding site (Ϫ594/ Ϫ580)-SV40 promoter chimeric constructs was significantly inhibited by CSC in a dose-dependent manner (Fig. 7B). Luciferase activities in NFI-binding site-SV40 promoter chimeric construct-transfected cells treated with or without CSC were expressed relative to controls obtained from cells transfected with NFI element free pGL3-Promoter vector under the same conditions to ensure the inhibition mediated by the perturbation of the NFI and cis-element interaction by CSC. CSC inhibition of LO promoter activity was further confirmed by the cotransfection assay of NFI expression constructs with the LO promoter-reporter vector (Fig. 9B). Moreover, CSC blockage of cellular NFI binding to the LO gene was demonstrated by the ChIP assay (Fig. 10). These results indicate the high sensitivity of NFI interaction with its cognate cis-element of the LO gene in response to stimuli. As reported, each puff of cigarette smoke produces 10 14 -10 16 free radicals inducing oxidative damage to the lung (51). CSC contains at least 3,500 compounds, including oxidants, heavy metals, and carcinogens. CSC has been shown to inhibit LO transcription at multiple levels such as the LO promoter activity, the transcription initiation rate, and the steady-state RNA level (16). CSC induced elevated levels of cellular metallothionein and glutathione, markers for the oxidative stress, suggesting that oxidative damage may represent a key mechanism for CSC down-regulation of LO (28). In view of the molecular structural feature of the NFI-DNA binding domain, CSC components may oxidize N-terminal cysteine residues of NFI proteins, thus weakening its interaction with DNA inducing decreased cellular NFI binding to the LO gene ( Fig. 10) and inhibition of the reporter gene expression (Figs. 7B and 9B).
In sum, this study describes the cloning and characterization of the LO gene promoter in rat lung fibroblasts. The Inr core promoter element encompassed the major one of multiple transcription start sites mapped at Ϫ51 upstream of the ATG, whereas the DPE core promoter element worked predominantly and independently for the activation of the TATA-less LO gene. The maximal promoter activity was displayed within an 804-bp 5Ј-flanking region upstream of the exon 1 composed of a basic domain from Ϫ277 to Ϫ1 and a regulatory domain from Ϫ804 to Ϫ278. Two NFI-binding sites at the sequences Ϫ594/Ϫ580 and Ϫ147/Ϫ133 were determined, which acted as an enhancer element elevating the LO promoter activity. Two NFI isoforms, NFI-A and NFI-B, were expressed in rat lung fibroblasts, and NFI-A exhibited a stronger transactivation activity than NFI-B. NFI transactivation of the LO gene was strongly modulated in response to exogenous stimuli such as CSC. By pursuing these findings, we may elucidate mechanisms for controlling transcription of LO, a key enzyme for ECM cross-linking and remodeling.