Role of Proximal Promoter Elements in Regulation of Renin Gene Transcription*

Mouse As4.1 cells, obtained after transgene-targeted oncogenesis to induce neoplasia in renal renin-express-ing cells, express high levels of renin mRNA from the endogenous Ren-1 c gene. We have used these cells to characterize the role of the Ren-1 c proximal promoter ( (cid:49) 6 to (cid:50) 117) in the regulation of renin gene transcription. It was found that 4.1 kilobases (kb) of Ren-1 c 5 (cid:42) flanking sequence, in combination with the proximal promoter, are required for strong activation ( (cid:59) 2 orders of magnitude over the basal level of the promoter alone) of the chloramphenicol acetyltransferase reporter in transfection assays. Within the 4.1-kb fragment, a 241- base pair region was identified that retains full activity in an orientation-independent manner in combination with the promoter. The resulting transcripts initiate at the normal renin start site. Electrophoretic mobility shift assays identified a sequence at approximately position (cid:50) 60 in the promoter region that binds nuclear proteins specific for renin-expressing As4.1 cells. Muta- tions in this sequence, which disrupt binding of nuclear protein(s), completely abolish activation of transcrip- tion by the 4.1-kb fragment. Activation of transcription by the 241-base pair enhancer was still observed, al- though it was diminished in magnitude (60-fold over the mutated promoter alone). We present

Renin, a protease involved in the regulation of mammalian blood pressure and electrolyte balance, is expressed in a complex developmental and tissue-specific pattern. Studies of mouse and rat fetuses have revealed that, early in development, renin expression is abundant in smooth muscle cells in the intrarenal arteries (1)(2)(3). However, as renal development proceeds, there is a progressive restriction of renin expression to smaller arteries and arterioles until, ultimately, it is limited to a small group of specialized cells at the most distal region of the afferent arteriole. These cells are known as juxtaglomerular cells and represent a rare cell type in the mature kidney (Ͻ0.1% of the total kidney cell population) (4). Under normal physiological conditions, the juxtaglomerular cells of the kidney are the principal source of active renin in the circulation (5,6). Interestingly, the developmental restriction of renin-expressing cells in the kidney is not a terminal event. Vascular smooth muscle cells appear to retain the ability to express renin and are actually recruited to a renin-expressing phenotype in response to a variety of physiological or pharmacological cues (7,8). Renin mRNA is also expressed in several extrarenal tissues that do not share an obvious embryological origin with the renin-expressing cells of the kidney (see Ref. 9 for review).
In mice, some strains contain a duplicated copy of the renin gene. This second locus is designated Ren-2 and is physically linked to the Ren-1 gene common to all mice (10 -12). Although Ren-1 and Ren-2 are expressed at approximately equal levels in the kidney, expression differences distinguish the two loci in extrarenal tissues (13)(14)(15). Expression differences are also observed at extrarenal sites for alleles of the Ren-1 locus, and by genetic and transgenic analyses, these differences appear to be regulated by sequences closely linked to the structural gene (16,17).
The lack of a suitable culture system for transfection analysis has hampered efforts to define regulatory regions controlling renin gene expression. However, evidence that important regulatory sequences may lie several kilobases distal to the renin gene has been suggested from the study of transgenic mice harboring chimeric genes containing renin 5Ј-flanking sequence linked to the coding region of SV40 T antigen. T antigen constructs containing either 0.45 or 2.5 kb 1 of Ren-1 or Ren-2 upstream sequence, respectively, demonstrated inappropriate expression in transgenic mice (18). However, transgenes containing more extensive 5Ј-flanking sequence (4.6 kb) demonstrated the correct tissue-and cell-specific expression of renin throughout development and in young adults (3,19) and an appropriate response to physiological perturbation with captopril (20).
To determine cis-acting regulatory regions controlling expression of the renin genes, cell lines that do not express renin were previously and widely used as a model system (40 -43). Data obtained with these cell lines suggested the involvement of proximal promoter sequences in renin transcriptional regulation. The fact that these cell lines do not express their endogenous chromosomal renin gene has raised caveats and limited interpretation.
In this work, we used the clonal cell line As4.1 (ATCC CRL2193), which maintains expression of its endogenous renin gene over long-term culture and was established from a kidney tumor of a Ren-2 5Ј-flanking sequence/SV40 T antigen transgenic mouse (21). The As4.1 cell line demonstrates several features that are characteristic of kidney juxtaglomerular cells, including high level expression of renin mRNA and the ability to store and secrete active renin (21). These cells represent a potentially valuable resource for the elucidation of cis-acting elements and trans-acting factors involved in the regulation of renin gene expression. In this report, we present our experiments using the As4.1 cell line, which support the hypothesis that regulatory sequences essential for cell-specific expression of renin are present in both the proximal promoter and distal 5Ј-flanking sequences.

EXPERIMENTAL PROCEDURES
Materials-Restriction enzymes and DNA-modifying enzymes were purchased from Life Technologies, Inc., Promega, and Boehringer Mannheim. [ 14 C]Chloramphenicol was obtained from Amersham Corp. Silica gel plates for thin layer chromatography were obtained from J. T. Baker Inc. Deoxyribonucleotides and n-butyryl-CoA were from Sigma. NotI linkers and the oligonucleotide 18-Bam were obtained from New England Biolabs Inc. and the Biopolymer Core Facility at the Roswell Park Cancer Institute, respectively. The phagemid vector pTz18R was obtained from Pharmacia Biotech Inc. Cell culture media, serum, and supplements were from Life Technologies, Inc. and Sigma. T7 RNA polymerase and the vectors pCAT-Basic, pUC19, and pGEM-7Zf Ϫ were purchased from Promega. The RNase protection assay kit was purchased from Ambion Inc. Oligonucleotides used for electrophoretic mobility shift assays (EMSAs) were obtained from DNA International or the Biopolymer Core Facility at the Roswell Park Cancer Institute.
Cell Lines-The renin-expressing As4.1 cell line was previously established in this laboratory from a kidney tumor of a transgenic mouse harboring a Ren-2 5Ј-flanking sequence/SV40 T antigen hybrid transgene (21). This line has been submitted to the American Type Culture Collection (ATCC CRL2193). The endogenous renin gene in the As4.1 cell line is Ren-1 c . The mouse Ltk Ϫ cell line (obtained from Dr. R. G. Hughes, Roswell Park Cancer Institute) was originally derived from the subcutaneous connective tissue of a C3H mouse (22). The Ltk Ϫ cell line also has an endogenous Ren-1 c gene on chromosome 1, but does not express renin (23). SV-T2 is a mouse embryonic cell line (SV40 virustransformed BALB/c/3T3 embryonic cells) obtained from the American Type Culture Collection (CCL163.1).
Plasmid Constructions-Plasmid pCAT-NotI was made by removing the BamHI site (located 3Ј of the CAT gene) of pCAT-Basic and converting it to a NotI site by use of a linker. Plasmid pϪ117CAT was constructed by inserting a renin promoter fragment (ϩ6 to Ϫ117); numbering relative to the major transcription start site) into the XbaI site adjacent to the CAT coding sequences of pCAT-NotI. The renin promoter fragment described above was derived from a DBA/2J Ren-2 plasmid (24). The BamHI site located at position Ϫ117 of the Ren-2 gene is missing in Ren-1 due to a single base polymorphism at position Ϫ115 of the renin 5Ј-flanking sequence. To join sequences upstream of the Ren-1 c gene with the renin promoter-CAT construct, a BamHI site was created at position Ϫ117 in the 5Ј-flanking region of Ren-1 c by sitedirected mutagenesis (26). Briefly, a 5.3-kb BamHI fragment containing the sequences from ϩ1.2 to Ϫ4.1 kb of the Ren-1 c gene was isolated from a previously described clone ( BALB-1) (25), ligated into the BamHI site of phagemid pTz18R, and subjected to site-directed mutagenesis. The 18-Bam primer (5Ј-GGCCAAGGATCCAGCTCC-3Ј) was used in oligonucleotide-directed mutagenesis of the plasmid pTz18R single-stranded form to cause a C 3 T substitution at position Ϫ115 of Ren-1 c and thus create a BamHI site at position Ϫ117. The modified construct was identified by restriction digest analysis and later sequenced to confirm the specific base change.
Plasmid R1C-4.1CAT was constructed by inserting the 5Ј-flanking sequence (Ϫ117 to Ϫ4100) of the BALB/c Ren-1 c gene into a BamHI site present at position Ϫ117 of the renin promoter in pϪ117CAT. Plasmids containing different lengths of Ren-1 c 5Ј-flanking sequence were constructed using restriction sites present on the 4.1-kb fragment: PstI (Ϫ1.8 kb), SphI (Ϫ2.625 kb), AccI (Ϫ2.866 kb), and PstI (Ϫ3.1 kb). Plasmid RiboϪ117/CAT was constructed by subcloning a 385-bp BamHI-EcoRI fragment of pϪ117CAT (position Ϫ117 in the renin promoter to the EcoRI site at position ϩ270 within the CAT gene) into pGEM-7zf Ϫ . The sequence of the 4.1-kb fragment of Ren-1 c flanking region has been submitted to the Genome Sequence Data Bank under accession number L78789.
Plasmid DNA was prepared by the Triton lysis method (27) and two sequential cesium chloride/ethidium bromide equilibrium centrifugation steps. A minimum of two different plasmid preparations have been tested for each construct.
Site-directed Mutagenesis of the Renin Gene Promoter Region (ϩ6 to Ϫ117)-Polymerase chain reaction was used to mutate the renin proximal promoter. The ϩ6/Ϫ117 promoter fragment within the pϪ117CAT plasmid was used as a template. The oligonucleotide primers used were as follows: A, 5Ј-GGCCCTGGGGTAccgAcTCAGAGCAG-3Ј (in the renin promoter sequence with mutations (lower-case) within the A/T-rich region forming a unique KpnI restriction site (underlined)); B, 5Ј-GACGGGCTCTAGACTGTGTAGCCCAGTCCCCCT-3Ј (containing sequence from the 3Ј-end of the renin promoter and overlapping pCAT-Basic sequence with an XbaI restriction site (underlined)); C, 5Ј-CTGCTCTGAgTcggTACCCCAGGGCC-3Ј (sequence complementary to oligonucleotide primer A); and D, 5Ј-GACGGCAAGCTTGCATGCCTG-CAGGTCGACTC-3Ј (pCAT-Basic sequence upstream of the renin promoter, with the PstI restriction site (underlined)). The polymerase chain reaction product amplified with oligonucleotide primers A and B was treated with KpnI and PstI. The polymerase chain reaction product amplified with oligonucleotide primers C and D was treated with KpnI and XbaI. Both fragments were subcloned into the pCAT-Basic plasmid treated with PstI and XbaI, resulting in a cloned renin promoter fragment with four point mutations within the A/T-rich region. The construct with mutated sequence was identified by restriction analysis and later sequenced to confirm the specific base change.
Cell Culture and Transient Transfections-As4.1 cells were propagated in Dulbecco's modified Eagle's medium containing 10% fetal bovine serum. As4.1 cells were transfected by electroporation (28) using a Cell Porator and electroporation chambers (0.4-cm electrode spacing; Life Technologies, Inc.). 72 h prior to transfection, As4.1 cells were subcultured into duplicate 75-cm 2 flasks (2 ϫ 10 6 cells/flask) containing 20.0 ml of Opti-MEM reduced-serum medium supplemented with 2% fetal bovine serum and 1 mg/ml ALBUMAX II. The switch from Dulbecco's modified Eagle's medium ϩ 10% fetal bovine serum to the supplemented reduced-serum medium is optimal for cell attachment post-electroporation and expression of transfected renin-CAT constructs in As4.1 cells. Conditions for electroporation were as follows. Cells (2 ϫ 10 7 ) were resuspended in 1.0 ml of 1 ϫ HeBS (25 mM Hepes, pH 7.05, 140 mM NaCl, 5 mM KCl, 0.75 mM Na 2 HPO 4 , 6 mM glucose) with a total plasmid DNA concentration of ϳ25-50 g/ml plus 250 g/ml sonicated salmon sperm DNA added as carrier. The cells were then exposed to a single electric impulse of 300 V at a capacitance setting of 1180 microfarads and subsequently transferred to 25-cm 2 flasks containing 8.0 ml of supplemented reduced-serum medium plus penicillin (100 units/ml) and streptomycin (100 g/ml).
To correct CAT values in each experiment for transfection efficiency, cells were cotransfected with 3 g of plasmid containing Rous sarcoma virus promoter driving ␤-galactosidase (RSV-␤-gal). Promoterless CAT plasmid (pCAT-Basic) was used to determine the background of CAT activity in these assays.
Chloramphenicol Acetyltransferase Assay-Cells were harvested 48 -72 h post-transfection and resuspended in 125 l of 0.25 M Tris, pH 8.0. The extract was subjected to three freeze/thaw cycles, and prior to CAT assays, extracts were heated to 60°C for 10 min. The protein concentration was determined by the method of Bradford (29) using bovine albumin as a protein standard. CAT activity was determined as described by Gorman et al. (30), except that n-butyryl-CoA (31) was substituted for acetyl-CoA in the assay. Equal amounts of protein ( ␤-Galactosidase Activity Assay-␤-Galactosidase activity was determined by using the Galacto-Light Plus TM chemiluminescent reporter assay according to the manufacturer's instructions (Tropix Inc., Bedford, MA). Chemiluminescence was measured on a Monolight 2010 TM luminometer (Analytical Luminescence Laboratory, San Diego, CA).

Results are expressed as light units/microgram of cell lysate.
RNase Protection Assay-Plasmid RiboϪ117/CAT was transcribed in vitro with SP6 RNA polymerase essentially as described by the manufacturer of the expression vector (Ambion, Inc.), using 50 Ci of [␣-32 P]GTP (3000 Ci/mmol) and 1 g of plasmid (linearized with BamHI). Full-length probe was isolated from a 7 M urea, 5% acrylamide sequencing gel and eluted overnight at 37°C in 0.5 M NH 4 OAc, 1 mM EDTA, 0.2% SDS. Purified probe (1 ϫ 10 5 cpm) was hybridized to 30 g of RNA isolated from transfected or nontransfected As4.1 cells and analyzed according to the manufacturer's instructions using the RNase protection assay kit (Ambion Inc.). Samples were then electrophoresed on 7 M urea, 5% acrylamide sequencing gels.
Nuclear Extract Isolation-Crude nuclear extracts from all cell lines used were prepared by the method of Gorski et al. (32). The concentration of proteins in resulting extracts was 2-3 mg/ml. Extracts were stored as aliquots at Ϫ70°C and thawed only once.
EMSA-The oligonucleotides used in this assay were annealed and labeled with [␥-32 P]ATP using T4 polynucleotide kinase, and doublestranded forms were purified from a 20% nondenaturing polyacrylamide gel. Nuclear extracts were preincubated for 15 min at room temperature in a 10-l reaction mixture containing 12 mM Hepes, pH 7.9, 60 mM KCl, 4 mM MgCl 2 , 1 mM EDTA, 1 mM dithiothreitol, 12% glycerol, 0.5 mM phenylmethylsulfonyl fluoride, and 1 g of poly(dI-dC). The concentration of nuclear extract was 2 g/10 l in each individual reaction mixture. Unlabeled competitor was included in some of the binding reactions as indicated. ϳ0.1 ng of radiolabeled double-stranded oligonucleotides (20,000 dpm) were added to each 10-l reaction mixture, and the incubation was continued for 30 min at room temperature. 2 l of loading dye (50% glycerol, 0.01% bromphenol blue, 1 mM Hepes, pH 7.9) were then added to each sample. The incubation mixture was loaded on a 5% polyacrylamide gel (19:1 acrylamide/bisacrylamide) in 1 ϫ TBE buffer (90 mM Tris borate, pH 8.0, 20 mM EDTA). Running buffer was 0.5 ϫ TBE buffer. Gels were run at ϩ4°C, after which they were transferred to Whatman No. 3MM paper, covered with Saran Wrap, and exposed to Kodak X-AR film for 5-15 h at Ϫ70°C with intensifying screens.

RESULTS
Three sites within the Promoter Region (ϩ6 to Ϫ117) of Ren-1 c Bind Nuclear Proteins from As4.1 Cells-To analyze the renin gene sequences involved in the regulation of transcrip-tion in As4.1 cells, we first divided the mouse renin gene 5Ј-flanking sequence into two regions: (a) the proximal promoter, a 123-bp fragment (ϩ6 to Ϫ117); and (b) a more distal segment extending out to Ϫ4.1 kb. Within the proximal promoter, mouse Ren-1 and Ren-2 genes have Ͼ99% sequence identity and ϳ80% identity with the rat and human renin genes (51,52). More distal of position Ϫ117, interspecific homology between renin genes decreases significantly. Farther upstream are repetitive sequences that are unique to the mouse genes and that extend some 500 -800 bases, depending upon the particular mouse locus in question, before homology again resumes with rat or human renin 5Ј-flanking sequences.
To identify sequences in this promoter fragment that are binding sites for nuclear proteins that may play a role in transcriptional regulation of the renin gene, we constructed 10 double-stranded oligonucleotides, 26 bp in length and overlapping each other by 13 bp, covering the entire renin proximal promoter with some additional sequence (ϩ12 to Ϫ130), as shown in Fig. 1A. These oligonucleotides were used as probes in EMSA with nuclear proteins isolated from As4.1 cells (Fig. 1B). 3 out of 10 oligonucleotides spanning the renin promoter were able to bind nuclear proteins from As4.1 cells: oligonucleotides 2, 5, and 9, covering positions Ϫ90, Ϫ60, and Ϫ20, respectively. Judging by the mobility of these complexes, all three interactions resulted from binding of different proteins (Fig. 1B).
Oligonucleotides that exhibited binding with nuclear extracts from renin-expressing As4.1 cells were tested by EMSA for binding to nuclear extracts from two non-renin-expressing cell lines, Ltk Ϫ and SV-T2. SV-T2 cells express SV40 T antigen, which is also present in As4.1 cells, due to expression of the transgene. The SV-T2 cell line was used to control for the possibility that T antigen or some other gene product induced by T antigen binds to the renin promoter sequences. Complexes of nuclear proteins from As4.1 cells with oligonucleotides 2 (position Ϫ90) and 9 (position Ϫ20) appeared to be nonspecific for renin-expressing cells since the same complexes were observed with nuclear extracts from control cell lines (data not shown). Complexes formed with oligonucleotide 5, which produced a very strong signal in EMSA (implying relative abundance of proteins binding to that sequence), are specific for As4.1 cells (Fig. 1C). Formation of these complexes with nuclear proteins from As4.1 cells could be effectively competed with unlabeled oligonucleotide 5 (Fig. 1C), suggesting that this interaction is sequence-specific. Nuclear proteins from SV-T2 and Ltk Ϫ cells did not form a major complex with oligonucleotide 5, as was observed with As4.1 cells (indicated by arrow in Fig. 1C).
Sequence Specificity of Binding of Nuclear Proteins from As4.1 Cells to the Ϫ60 Region of the Mouse Ren-1 c Gene-In Fig.  2, a series of EMSAs were performed to determine the minimal sequence within oligonucleotide 5 necessary for binding of cellspecific nuclear proteins from As4.1 cells.
First, the two oligonucleotides overlapping oligonucleotide 5 (oligonucleotides 4 and 6) (see Figs. 1A and 2A) were tested in competition assays (Fig. 2B) to further delineate core binding sequence. Neither sequence was able to compete complex formation with oligonucleotide 5. These results are consistent with findings shown in Fig. 1B, where these oligonucleotides, when used as probes, did not form complexes with nuclear proteins from As4.1 cells.
Second, a sequence with slightly different coordinates was chosen for further EMSA because it covers the region with the highest percent identity among mouse, rat, and human renin gene promoters (51,52) and includes a large portion of oligonucleotide 5 sequence. This sequence, designated 1 c (Fig. 2A), formed complexes with the same mobility as oligonucleotide 5 and was an effective competitor for complexes formed with oligonucleotide 5 (Fig. 2B). This suggests that the first 7 base pairs on the 5Ј-end of oligonucleotide 5, which are not present in 1 c ( Fig. 2A), are not involved in the formation of the complex.
Third, an oligonucleotide with sequence derived from the human renin gene analogous to oligonucleotide 1 c (HR) ( Fig.  2A) successfully competed oligonucleotide 5 interaction (Fig.  2B) and formed the same complex when used as a probe (data not shown). Although the human renin sequence in this region (HR) ( Fig. 2A) is highly homologous to the mouse sequence, some differences are located 1) in the sequence preceding the A/T-rich region (TGGGG for mouse and CAGGG for human renin) and 2) downstream of the A/T-rich region (AAG for Ren-1, GAG for Ren-2, and GGG for human renin). Competition studies presented in Fig. 2B with human renin gene (HR) and in Fig. 2C with Ren-2 (2 d ) sequences indicate that these differences do not affect protein binding to this region.
The A/T-rich sequence on 1 c appears to be central for complex formation with nuclear proteins from As4.1 cells. Surprisingly, oligonucleotide 6, which contains the intact A/T-rich sequence, did not form any complexes in EMSA or compete with binding to oligonucleotide 5. Because A/T-rich sequence is located on the end of oligonucleotide 6, we considered the possibility that the stability of the double-stranded form was altered, which in turn could affect complex formation. Alternatively, it is possible that sequence just upstream of the A/T-rich region is required for protein binding. To test these possibilities, we mutated oligonucleotide 6 (6M) by the addition of three deoxycytidine residues to its 5Ј-end immediately preceding the A/Trich sequence (Fig. 2A). This resulted in strong binding of nuclear proteins to oligonucleotide 6M, and the resulting complex had the same mobility as oligonucleotide 5 (Fig. 2B). Also, in contrast to oligonucleotide 6, which was an inefficient competitor for oligonucleotide 5 (Fig. 2B), the mutated sequence 6M was capable of competing the oligonucleotide 5 complex.
The effect of sequence addition to oligonucleotide 6 is most likely stabilization of the oligonucleotide in its double-stranded form, although we cannot rule out the possibility that a short nucleotide stretch (without specific sequence requirement) upstream of the A/T-rich region is all that is required for stable complex formation.
Fourth, to delineate the sequence on 1 c responsible for protein binding, six 1 c mutants (M1 to M6) were used in EMSA as competitors for 1 c complex formation ( Fig. 2A). Mutants M3 and M6, with altered A/T-rich sequence (TAATAAAT), lost their competition ability (Fig. 2C), while mutation of the two nucleotides immediately following the A/T box (mutant M2)  Fig. 1A) as well as mutated oligonucleotide 6 (6M) are also shown. B, oligonucleotides 5 and 6M were used as probes in EMSA without competitor or competed with oligonucleotides related to 1 c , as indicated. C, EMSAs were performed in which formation of the major 1 c complexes was competed with that of related 1 c sequences, as indicated. A 50-fold excess of competitor over the probe was used. Control samples (Ϫ lane) did not contain competitor. resulted in significantly decreased competition (Fig. 2C). Mutations M1, M2, and M5, which do not affect the A/T-rich sequence, still efficiently competed the 1 c binding.

High Level Cell-specific Transcription of the Reporter Gene in As4.1 Cells Requires the Presence of Both the Ren-1 c 5Ј-Flanking Region and Proximal
Promoter-Constructs were developed by fusing Ren-1 c 5Ј-flanking fragments upstream of the basal renin promoter-CAT construct. These were used to address minimal sequence requirements for initiation of reporter gene transcription in As4.1 cells in transient transfection assays. When the 123-bp renin promoter in either orientation in the CAT construct was transfected into As4.1 cells, there was no significant increase in CAT transcription compared with a promoterless CAT construct.
In an effort to find elements that activate transcription of the reporter gene, we attached various lengths of Ren-1 c 5Ј-flanking fragment (Ϫ117 to Ϫ4100) to the renin proximal promoter construct and assayed for induction of CAT activity after transfection into As4.1 cells. The addition of 5Ј-flanking sequence from positions Ϫ117 to Ϫ1800 (Fig. 3, construct d) and from positions Ϫ117 to Ϫ2625 (construct e) to the 123-bp renin promoter did not result in significant increase in CAT activity. However, the activity of the renin promoter was induced Ͼ70fold in the As4.1 cells when fragments Ϫ117 to Ϫ2866 (Fig. 3, construct f), Ϫ117 to Ϫ3100 (construct g), and Ϫ117 to Ϫ4100 (construct h) of Ren-1 c 5Ј-flanking sequence were added to the renin promoter. High level activation was also observed in constructs containing truncated forms of the 5Ј-flanking sequence: fragments Ϫ2625 to Ϫ4100 (Fig. 3, construct i) and Ϫ2625 to Ϫ3100 (construct k). Since constructs with fragments Ϫ2866 to Ϫ4.1 (Fig. 3, construct j) and Ϫ2866 to Ϫ3100 (construct l) did not show significant activity in this assay, it was concluded that the active sequence is positioned on a 241-bp fragment, from positions Ϫ2625 to Ϫ2866. In combination with the renin promoter, this fragment retained full activity (Fig. 3,  construct m), in an orientation-independent manner (construct n), typical of classical enhancers. Furthermore, constructs containing this enhancer in either orientation correctly initiated transcription from the renin promoter when analyzed by the RNase protection assay (see Fig. 5).
To examine the functional role of the Ϫ60 region in the renin promoter, the A/T-rich sequence was mutated to match mutant M6 (5Ј-GGCCCTGGGGTAccgAcTCAGAGCAG-3Ј), which did not compete for proteins binding with oligonucleotide 1 c (5Ј-GGCCCTGGGGTAATAAATCAGAGCAG-3Ј) in EMSA (Fig.  2C). When 4.1 kb of Ren-1 c 5Ј-flanking sequence were placed in front of the mutated promoter, no induction of CAT transcription above basal levels was observed (Fig. 4, construct c versus  d). With the 241-bp enhancer and mutated renin promoter (Fig.  4, construct f), activation was ϳ4-fold lower than with the control (enhancer with wild-type promoter) (construct e), but was still significantly higher than with the mutated promoter alone (ϳ60-fold) (compare constructs f and b).
Thus, sequence between positions Ϫ2625 and Ϫ2866 of the Ren-1 c 5Ј-flanking sequence is capable of activating transcription of the renin promoter in the absence of protein binding at the Ϫ60 region, but for the maximal induction of renin transcription, both sequences are required. On the other hand, activity of the 4.1-kb fragment is completely dependent on the presence of a functional A/T-rich region within the renin promoter.
Transfected Ren-CAT Constructs Correctly Initiate Transcription in As4.1 Cells-Total RNAs isolated from nontransfected and transfected As4.1 cells were analyzed by the RNase protection assay as described under "Experimental Procedures." The mRNA start site was determined using a labeled 432-nucleotide antisense cRNA probe (Fig. 5A). A schematic diagram of the transfected renin-CAT constructs analyzed by the RNase protection assay is given in Fig. 5B. Constructs with the upstream region (Ϫ117 to Ϫ4100) present in both the forward (3) and reverse (4) orientations relative to the promoter were used. The expected 276-nucleotide RNase protection product for correctly initiated transcripts is depicted below the map. RNA prepared from As4.1 cells transfected with the 4.1-kb 5Ј-flanking region in either orientation yielded protected fragments of the predicted size (276 nucleotides) for mRNA initiating at the major renin transcription start site (Fig. 1A). The initiation site used by transfected Ren-CAT constructs is the same one found for endogenous Ren-1 and Ren-2 genederived renal transcripts (25). DISCUSSION Much work has been aimed at determining the cis-acting regulatory regions controlling cell-specific expression of the renin genes. Analyses of renin constructs in transgenic mice (3,18,19,38,39) and after transfection into cell lines that lack expression of their endogenous renin genes (40 -43) have been difficult to correlate, and therefore, no clear picture has emerged.
In previous experiments with transgenic mice carrying a Ren-2 genomic transgene (including 2.5 kb of 5Ј-flanking sequence, exon/intron region, and 3Ј-flanking sequence), correct tissue-and cell-specific expression was reported (39). Additional studies by Sola et al. (18), however, showed that the 2.5 kb of the Ren-2 5Ј-flanking region were incapable of directing appropriate expression on their own when linked to SV40 T antigen in transgenic mice. In contrast, we have demonstrated correct tissue-and cell-specific expression of both a Ren-2 genomic transgene and a Ren-2/SV40 large T antigen fusion construct using larger amounts of 5Ј-flanking sequence (5.3 and 4.6 kb, respectively) (3,19,38). Taken together, these results suggest that important tissue-and cell-specific regulatory elements may lie upstream of Ϫ2.5 kb in the mouse renin 5Ј-flanking region and that there may possibly be some redundancy of information.
To further define the role of the renin 5Ј-flanking sequences controlling expression that were first analyzed in transgenic studies, we have used a kidney-derived cell line (As4.1) that expresses renin mRNA and protein over long-term culture. To facilitate analysis, we initially divided the 5Ј-region upstream of the cap site into a proximal promoter region extending from positions ϩ6 to Ϫ117, which included the cap site and the TATA box as well as some additional sequences, and a more distal segment extending to Ϫ4.1 kb of the Ren-1 c gene. Ϫ117 bp was chosen as a truncation point of the proximal region because this coordinate lies near the boundary of the first breakdown in homology evident between renin genes of different species (51).
To delineate the proximal promoter region more precisely, we have scanned the relevant region for identifiable protein-DNA interactions by EMSA. In this region, EMSA detects three sequence-specific interactions near positions Ϫ90, Ϫ60, and Ϫ20. The Ϫ90 site is a weak interaction that is not cell-specific; the Ϫ60 site is a strong cell-specific interaction encompassing the A/T-rich region; and the Ϫ20 site near the cap site is again not cell-specific. The mouse Ren-1 c Ϫ60 region was found to bind nuclear proteins from mouse kidney (43), and all three sites were detected by footprint analysis using renin-expressing human chorionic primary cells (33). Therefore, we examined a series of mutations directed at defining more precisely the critical sequence requirements for binding at the Ϫ60 site. Direct EMSA and competition assays suggest that the minimal site includes N (1-3) TAATAAATCA.
Sun et al. (35,36) and Gilbert et al. (37) detected binding of nuclear proteins from the pituitary cell line GC to a region on the human renin promoter that is analogous to the mouse Ϫ60 region. The authors suggested the possibility that transcription factor related to Pit-1/GH-1 (56) could be responsible for binding to the Ϫ60 region in the human renin gene. Although that sequence was originally identified as a Pit-1/GH-1-binding site, the sequence as found in the mouse has only partial similarity to Pit-1/GH-1 recognition sequence. Significantly, GC cells (which do not express their endogenous renin gene), used for transfection experiments by Sun et al. (35,36) and Gilbert et al. (37), do express Pit-1/GH-1. Since the recognition site existing in the renin promoter shares some homology with a Pit-1/GH-1 site, transcription of the renin gene may be artificially elevated in these studies. Double-stranded oligonucleotide containing the human Pit-1/GH-1 sequence did not form complexes with nuclear proteins from As4.1 cells, nor was it able to compete formation of the major cell-specific 1 c complex (data not shown). Furthermore, no mRNA for Pit-1/GH-1 is detected in As4.1 cells by Northern blot analysis at moderate stringency. 2 There are also other classes of factors, besides the POU homeodomain protein family (44), of which Pit-1/GH-1 is a member, that bind A/T-rich sites, such as high mobility group/high mobility group box proteins (45,46), serum-responsive factors (47), and factors related to serum-responsive factors (myocyte enhancer factors) (48) as well as dispersed paired-like homeodomain proteins, e.g. MHox (49). Thus, while the transcription factor affecting regulation of the Ren-1 c gene by binding to the A/T-rich site in As4.1 cells may still belong to the POU homeodomain class family, the precise identity of this factor is, at present, unknown.
In our transfection assays using the As4.1 cell line, we were not able to detect significant increases in reporter gene transcription with the 123-bp promoter (ϩ6 to Ϫ117), alone or in combination with up to 2.5 kb of 5Ј-flanking sequence. Furthermore, this report has confirmed what was first suggested with Ren/Tag transgenes, that a regulatory element located between residues Ϫ2600 and Ϫ4100 of Ren-1 c functions as an enhancer. When both 4.1 kb of the 5Ј-flanking sequence and the promoter region of Ren-1 c were included in renin-CAT reporter gene constructs, the renin promoter was activated ϳ2 orders of magnitude in renin-expressing As4.1 cells. Further analysis of the 4.1-kb flanking region resulted in delimitation of the Ϫ2625/Ϫ2866 region, which retains full activity. This sequence acts as an enhancer in that it promotes high levels of correctly 2 T. A. Black, unpublished data.

FIG. 5.
Mapping of the transcription start sites for transfected renin-CAT constructs in As4.1 cells. A, 30 g of total RNAs isolated from nontransfected and transfected As4.1 cells were analyzed by the RNase protection assay as described under "Experimental Procedures." RNA markers were generated by in vitro transcription of linearized templates of known sizes using SP6 RNA polymerase. Minus-RNase and tRNA controls are shown in the first two lanes. B, shown is a schematic diagram of the transfected renin-CAT constructs analyzed by the RNase protection assay. Constructs with the upstream region (Ϫ118 to Ϫ4100) in both the forward (3) and reverse (4) orientations relative to the promoter were examined (fourth and fifth lanes, respectively). The expected 276-nucleotide (nt) RNase protection product for correctly initiated transcripts is depicted below the map. initiated transcription independent of its orientation relative to the renin promoter.
Mutations within the Ϫ60 region of the promoter, which eliminate nuclear protein(s) binding to this sequence, completely abolish activation of reporter gene transcription by the 4.1-kb 5Ј-flanking sequence. Activation by the enhancer region was reduced by the same promoter mutations, but remained significant (60-fold compared with the promoter alone). Data presented here are consistent with the presence of negative regulatory element(s) in the Ren-1 c 5Ј-flanking region between positions Ϫ117 and Ϫ2600. The presence of potential negative regulatory elements upstream of the mouse Ren-1 d gene has been reported by others previously (44,50,(53)(54)(55). The suggested position of these putative element(s) lies in a region between the enhancer we have identified and the proximal promoter, consistent with our findings.
To account for all the data presented, we currently favor a model in which a complex interplay of cell-specific and general factors is associated with generation of the cell specificity of renal renin expression. In this model, a specific factor from renin-expressing cells is required to bind at the crucial proximal promoter site to relieve what are possibly ubiquitously functional, negative influence(s) interacting farther upstream. This relief, or abrogation, of inhibition permits an enhancer function associated with the most distal elements to boost expression from the basal transcription machinery only in renin-expressing cells. This model has the potential to explain not only the relatively exquisite tissue specificity of renin expression, but also certain key features of renal renin expression, namely, the plasticity of expression in renal vascular smooth muscle found as a function of development or under physiological and pathophysiological perturbation.
In conclusion, these results, derived by transient transfection analysis of our renin-expressing cell line, are consistent with previous findings from transgenic mice. These data provide compelling evidence that sequences located both distal and proximal to the promoter are involved in the tissue-and cellspecific expression of the mouse renin genes.