A Novel Distal Enhancer Confers Chorionic Expression on the Human Renin Gene*

Renin catalyzes the rate-limiting step of the renin-angiotensin system, which regulates blood pressure and electrolyte homeostasis. To determine cell-specific human renin gene control elements, the transcriptional activity of promoter regions up to position −8876 was studied in renin-expressing cells. A positive regulatory region conferring ∼57-fold higher transcriptional activity to the human renin gene promoter in chorionic cells was identified between nucleotides −5777 and −5552. It had the orientation-independent activity typical of classical enhancers. It also conferred ∼59-fold higher transcriptional levels from the heterologous simian virus 40 (SV40) promoter in chorionic cells and ∼6-fold higher transcriptional levels in Calu-6 and As4.1 cells, whereas no effect was measured in non-renin-expressing cells. DNase I footprinting showed that this enhancer contains three binding sites for chorionic cell nuclear extracts. Functional analysis suggested that the activity of the enhancer is regulated by differential mechanisms in the three renin-expressing cells involving a complex arrangement of AP-1 motifs binding cell-specific members of the basic leucine zipper family of transcription factors. Thus, our results demonstrate that this enhancer plays a key role in the expression of the human renin gene in the chorion and may also be involved in its regulated expression in other tissues.

Renin, a key enzyme of the renin-angiotensin system regulates blood pressure and electrolyte homeostasis. The juxtaglomerular (JG) 1 cells of the kidney are the principal site of renin synthesis, but renin is also synthesized in a variety of nonrenal tissues. Chorionic tissue is one of the main extrarenal sites at which a local renin synthesis has been demonstrated at the protein (1,2) and the gene levels (3). Therefore, cultured chorionic cells (4) have been a model of choice for studying the regulation of the human renin gene transcription. Efforts to dissect human renin gene transcriptional regulation have focused mainly on the proximal promoter. Duncan et al. (5) demonstrated by transient transfections in chorionic cells that the first 584 bp of the human renin 5Ј-flanking region are able to direct chloramphenicol acetyltransferase expression. We then studied two elements, the cyclic AMP responsive element (Ϫ229 to Ϫ220) and a sequence similar to that of homeodomain-containing transcription factor binding site, located from nucleotides Ϫ79 to Ϫ69, which are responsible for basal renin gene promoter activity in chorionic cells (6,7). We have also shown that renal and chorionic tissues contain similar nuclear binding proteins recognizing the same regions in the human renin proximal promoter (8).
More distal human renin promoter regions up to nucleotide Ϫ2824 have been studied but no cis-acting element functionally important for promoter activity was identified (7). These distal regions were less efficient at directing reporter gene expression, suggesting that there may be a discrete silencer between nucleotides Ϫ2824 and Ϫ892. Transgenic animal studies have also failed to produce a coherent description of the human renin gene control elements in the coding, 3-kb 5Ј-flanking region and 1.2-kb 3Ј-flanking region (9 -11), demonstrating the need to look for more distal specific cis-regulating elements. Recently, Yan et al. (12) showed that a 220-bp region located 12 kb upstream from the transcription start site of the human renin gene gave up to 47-fold higher transcription rates when associated with the proximal promoter and used to transfect mouse As4.1 cells.
Differential regulation of the proximal promoter has been reported in human renin-producing chorionic cells (7) and Calu-6 cells (13,14). Therefore, the differential regulation of transcription may be involved in cell-specific expression of the human renin gene in the chorion. We isolated a genomic clone containing the human renin gene and its surrounding sequences. The functional activity of the promoter until position Ϫ8876 was determined by transient transfections of human chorionic cells, Calu-6 cells, and mouse As4.1 cells. An enhancer element was discovered far upstream from the transcription start site (Ϫ5777 to Ϫ5552). It conferred an ϳ59-fold activation of transcription in chorionic cells and gave smaller but significant increases in Calu-6 cells (ϳ6-fold) and As4.1 cells (ϳ6.5-fold). When used with a heterologous promoter, enhancer transcriptional activity was selectively increased in renin-producing cells. DNase I footprinting experiments detected at least three specific DNA-protein interactions with chorionic cell nuclear extracts, the functional activity of which was assayed in the three cell types either by reconstitution experiments with oligonucleotides encoding each protected sequence or by site-directed mutagenesis. This is the first report of a human renin gene enhancer with a chorionic cell-marked effect, suggesting its involvement in the regulation of the transcription of the renin gene in the chorion and that it may be responsible for the expression of the renin gene in different tissues. * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 1 The abbreviations used are: JG, juxtaglomerular; SV40, simian virus 40; AP-1, activator protein 1; NF-E2, nuclear factor erythroid 2; HEL, human erythroid leukemia; bZIP, basic leucine zipper; bp, base pair(s); kb, kilobase(s); PCR, polymerase chain reaction.
Cell Culture and Transfection Analysis-Primary cultures of chorionic cells (4) and Calu-6 cells, derived from a pulmonary carcinoma (ATCC HTB-56) (14), were cultured as described previously. As4.1 cells (ATCC CRL2193), a mouse renin-expressing cell line obtained from a kidney tumor of a transgenic mouse harboring a mouse ren-2 5Ј-flanking region/SV40 antigene, were cultured as described by Sigmund et al. (15). Cells were transiently transfected with DNA using FuGENE TM 6 transfection reagent (Boehringer Mannheim). Chorionic, Calu-6, As4.1, and COS-7 cells (ATCC CRL 1651) were plated in 12-well plates. For chorionic cells, 3.4 l of FuGENE TM 6 plus 1.5 g of renin reporter construct and 200 ng of pCH110 (16) (as an internal control for transfection efficiency) were used. For Calu-6 and COS-7 cells, 200 ng of renin reporter construct and 30 ng of pCH110 were used with 0.92 or 0.46 l of FuGENE TM 6, respectively. For As4.1 cells, we used 4.6 l of FuGENE TM 6 plus 2 g of renin reporter construct and 300 ng of pCH110. The cells were incubated overnight with DNA, and the medium was then replaced by fresh medium for 24 h. Transfected cells were washed three times with cold 1ϫ phosphate-buffered saline and lysed by adding 200 l of reporter gene assay lysis buffer (Boehringer Mannheim). A 50-l aliquot to which 50 l of luciferase reporter buffer (Promega) was added in a microplate scintillation and luminescence counter (Top Count, Packard) to measure luciferase activity. Calu-6 cells lysate was incubated 1 h at 50°C to destroy high endogenous ␤-galactosidase activity, and each cell type ␤-galactosidase activity was estimated by measuring the luminescence of a 50-l aliquot using ␤-galactosidase reporter gene assay (Boehringer Mannheim). All results are given as the mean of at least three independent transfection experiments Ϯ S.E.
Preparation of Nuclear Extracts, DNase I Footprinting, and Electromobility Shift Assays-Human chorionic cell and COS-7 cell nuclear extracts were prepared as described previously (17). The radioactive (Ϫ5777 to Ϫ5552) fragment used for DNase I footprinting of the renin gene promoter was obtained by PCR amplification using the following primers: 5Ј-TGCAGGTGCCAGAACATTTCTC-3Ј and 5Ј-AATCCCCGC-CCTGATACCAACG-3Ј hybridizing to plasmid pGL 3 -promoterϩ(Ϫ5777 to Ϫ5552). These primers were labeled with [ 32 ␥P]dATP by T4 DNA polynucleotide kinase before PCR amplification. DNase I footprint analysis was performed as described previously (6) with 30 g of chorionic cell or COS-7 cell nuclear extracts and 1/100 and 1/200 dilutions of DNase I (10 IU/l) (Boehringer Mannheim). The renin gene doublestranded oligonucleotides used for gel shift assays were as described in the plasmid constructs section. The AP-1 and NF-E2 oligonucleotides were as described by Andrews et al. (18): 5Ј-TGGGGAACCTGTTCT-GAGTCACTGGAG-3Ј and 5Ј-TGGGGAACCTGTGCTGAGTCACTG-GAG-3Ј. Consensus binding sites are underlined. These doublestranded oligonucleotides were end-labeled with [ 32 ␥P]dATP by T4 DNA polynucleotide kinase. Each binding reaction was performed with 7 g of human chorionic cell nuclear extracts incubated with 25,000 cpm end-labeled oligonucleotides for 30 min at 4°C in 20 l of buffer containing 6 mM MgCl 2 , 0.8 mM EDTA pH 8.0, 60 mM KCl, 30 mM Hepes pH 7.6, 5% Ficoll 400, 1 mM DTT, 175 mM NaCl, 10 mM Hepes, pH 7.9, 12.5% glycerol, 3 g of poly(dI-dC), and 1 g of sonicated salmon sperm DNA. When performing electromobility supershift assays, antisera directed against c-Jun, c-Fos, JunB, fra-2, kindly provided by Dr. Lallemand (CNRS, URA 1644, France), MafG/K, and p45NFE-2, kindly provided by Dr. Deveaux (INSERM U 91) were added to the reaction mixture and then incubated for 1 h at 4°C before electrophoresis. The mixture was subjected to electrophoresis in nondenaturing conditions in 5% polyacrylamide gels in 22 mM Tris borate/0.5 mM EDTA. For competition experiments, unlabeled oligonucleotides were used at a 50to 200-fold molar excess.
Site-directed Mutagenesis by Recombinant PCR-To destroy the NF-E2 and AP-1 binding sites in footprint C, two PCR reactions, one using sense and GL 2 primer (Promega) and the other using reverse and RV 3 primer (Promega) were performed (mutations are indicated in italics and underlined). These two PCR products were mixed and subjected to a new PCR reaction with GL 2 and RV 3 . The overlap containing the mutated sites allowed efficient recombinant PCR of the whole fragment which was inserted into pGL 3 -promoter vector and fully sequenced. The following oligonucleotides were used: NF-E2mut sense, GAAAATGAGAtgTCATTACT; NF-E2mut reverse, AGTAATGAcaTCT-CATTTTCC; C/AP-1mut sense, TCATTACTtgCCTGGCTACTTC; and C/AP-1mut reverse, AAGTAGCCAGGcaAGTAATGA. Double mutants were obtained by the same strategy.
To destroy the AP-1 sequence of the B footprint, B/AP-1mut reverse (GGAGGCTTAGCATGACcaACTCCATTTTG) and RV 3 were used to perform a PCR using pGL 3 -promoter-enhancer. The PCR product was digested by BlpI and SacI and inserted into pGL 3 -promoter vector. To create AB(892)-pGL 3 -promoter-enhancer, a PCR reaction using BlpI reverse primer (AGTGGAGGGAGGCTTAGCATG) and RV 3 was performed using pGL 3 -892. PCR product was digested by BlpI and SacI and ligated into pGL 3 -promoter-enhancer. The remaining digested vector was blunt-ended and self-ligated to create ⌬AB-pGL 3 -promoterenhancer.

Identification of an Upstream Positive Regulatory Region-
Transcriptional activity of the human renin gene distal promoter was performed by transient transfections of various regions, up to position Ϫ8876, inserted upstream from a promoterless luciferase gene construct in human renin-expressing chorionic cells. The activity measured was compared with that in human Calu-6 cells and in mouse As4.1 cells (Fig.  1). When the renin gene 5Ј-flanking region was lengthened from Ϫ892 to Ϫ4713, promoter activity, expressed as a percentage of pGL 3 -892 activity, decreased in human cells whereas it increased progressively in mouse As4.1 cells. When the promoter region was lengthened from Ϫ4713 to Ϫ5777, the resulting plasmid pGL 3 -5777 had a very high level of transcriptional activity in chorionic cells (4412 Ϯ 489) versus (77 Ϯ 6) for pGL 3 -4713. These results suggest that there is an enhancer between Ϫ5777 bp and Ϫ4713 bp that confers ϳ57-fold higher transcription rates in human chorionic cells. We also studied more distal fragments to Ϫ8876 and found that the (Ϫ7476 to Ϫ5777) region contained cis-acting regions that were func-tional only in As4.1 cells. The (Ϫ8876 to Ϫ7476) region had no transcriptional activity in any cell type. The highest level of transcription of the pGL 3 ϩ(Ϫ7476 to Ϫ5777) construct in As4.1 cells was consistent and reproducible but was not investigated further because no such stimulation was observed in human cells.
To determine more precisely the location of the (Ϫ5777 to Ϫ4713) enhancer, this fragment was inserted upstream of the first 892 bp of the human renin gene promoter driving the luciferase reporter gene, creating pGL 3 -892ϩ(Ϫ5777 to Ϫ4713) plasmid. The (Ϫ5777 to Ϫ4713) fragment conferred an ϳ39-fold higher activity to the pGL 3 -892 plasmid in chorionic cells compared with an ϳ6-fold higher activity in Calu-6 cells and ϳ3fold higher activity in As4.1 cells. Enhancer activity was observed in each cell type when the (Ϫ5777 to Ϫ4713) fragment was inserted in the sense or antisense orientation. Its activity was sometimes higher in the antisense than in the native orientation, demonstrating the orientation-independent classical enhancer activity of this fragment. Serial deletion mutant fragment of this 1-kb region were produced, and it was found that the more distal 225-bp region of the (Ϫ5777 to Ϫ5552) fragment had full enhancer activity with 56-fold higher transcriptional activity in chorionic cells compared with a 5-fold and an 8-fold higher transcriptional activity in Calu-6 cells and As4.1 cells, respectively (Fig. 1).
Cell-specificity Analysis-To test whether the activity of the human renin enhancer was confined to renin-expressing cells, the 225-bp enhancer fragment activity was studied using the heterologous SV40 promoter to drive luciferase gene expression. When it was cloned upstream from the SV40 promoter (Table I), its effect was similar to that observed with the native promoter with a 59-fold increase in reporter gene activity in chorionic cells, 6-fold in Calu-6 cells, and 6.5-fold in As4.1 cells.
Therefore, the activity of the enhancer was cell-specific and promoter-independent. There was no difference in the activity of the pGL 3 -promoter and pGL 3 -promoterϩ(Ϫ5777 to Ϫ5552) reporter plasmids in COS-7 cells (Table I), suggesting that the effect of the enhancer on the transcription of the human renin gene may be specific to renin-expressing cells.
Binding of Chorionic Cell Nuclear Extracts to the Human Renin Enhancer-To further determine the sequences involved in enhancer activity, an in vitro DNase I footprinting assay was performed using chorionic cell nuclear extracts. Three regions were protected by nuclear proteins (Ϫ5766 to Ϫ5745), (Ϫ5743 to Ϫ5719), and (Ϫ5651 to Ϫ5619), named A, B, and C, respectively (Fig. 2, panel I). These three footprints were specific to chorionic cells because DNase I footprinting experiments performed with COS-7 cell nuclear extracts showed only one footprint in the B region, slightly shorter than B and named BЈ (Fig. 2, panel III). To detect consensus matches in protected nucleotide sequences, the enhancer sequence was subjected to a screen of consensus matches against a library of transcription factor binding sites (19). This analysis allowed us to identify a putative binding site for NF-E2 (18) and AP-1 (20,21) in footprint C and a putative binding site for AP-1 in footprints B (Fig. 2, panel II).
To analyze the specificity of the DNA-protein interactions, we performed gel shift assays with oligonucleotides spanning the three protected sequences and chorionic cell nuclear extracts. No specific interactions were detected with an oligonucleotide corresponding to footprint A (data not shown). A weak interaction was detected with the oligonucleotide containing the footprint B (Ϫ5747 to Ϫ5719), which was, however, specific because unlabeled homologous oligonucleotide competed for binding (Fig. 3). This interaction was not competed by a 100fold molar excess of oligonucleotides corresponding to foot- prints A (Ϫ5766 to Ϫ5743) or C (Ϫ5656 to Ϫ5615) but was competed by a 50-to 100-fold molar excess of the oligonucleotide containing footprint AB (Ϫ5766 to Ϫ5719). It was not competed by a 100-fold molar excess of oligonucleotides con-taining consensus recognition sites for AP-1 or NF-E2. A specific interaction was observed with an oligonucleotide corresponding to footprint C containing both NF-E2 and AP-1 consensus binding sites (Fig. 4). It was competed by unlabeled oligonucleotide containing footprint C but not by a 100-fold molar excess of oligonucleotides corresponding to footprints A, B, or AB. In addition, both AP-1 and NF-E2 oligonucleotides competed for binding to the footprint C sequence.
To examine further the interactions at this motif, gel shift assays were performed with labeled oligonucleotides corresponding to bona fide NF-E2 and AP-1 binding sites. Specific interactions also occurred because bindings were competed specifically by molar excesses of both unlabeled AP-1 and NF-E2 oligonucleotides and by the footprint C sequence (Fig. 5). To discriminate between an AP-1 and/or NF-E2 binding activity to the NF-E2 motif, gel shift experiments were conducted with Human Erythroid Leukemia (HEL) cell nuclear extracts and labeled NF-E2 probe, showing two shifted complexes as described by Deveaux et al. (22) (data not shown). Only the upper AP-1 complex was competed by the footprint C sequence, whereas the lower NF-E2 complex was not competed (data not shown), demonstrating that the interaction with the NF-E2 binding site of the footprint C motif is related to AP-1. Supershift experiments were performed to characterize further the proteins present in chorionic cell nuclear extracts that bind to motif C. Experiments conducted with antisera directed against c-Fos, c-Jun, JunB, fra-2, MafG/K (23), and p45NF-E2 (18) showed no supershift (data not shown). Positive controls were performed with the AP-1 probe and chorionic cell nuclear ex-tracts for antisera directed against c-Fos, c-Jun, JunB, and fra-2 and with HEL cell nuclear extracts binding to the NF-E2 probe for antisera directed against MafG/K and p45NF-E2 (data not shown). These data suggest that tissue-specific transcription factors expressed in the chorion, but different from typical members of the basic leucine zipper (bZIP) family, are involved in regulating the activity of the enhancer.
Enhancer Functional Analysis-We addressed the relative importance of each binding site by reconstitution experiments where one copy of each footprint motif was inserted upstream from the SV40 promoter in the pGL 3 -promoter vector. These constructs were transiently transfected into the three reninexpressing cell types and COS-7 cells. None of these fragments had any effect on SV40 promoter activity in COS-7 cells (Table  I). In chorionic cells, the activity of pGL 3 -promoterϩA, pGL 3 -promoterϩB, pGL 3 -promoterϩC, pGL 3 -promoterϩAB, and pGL 3 -promoterϩABC constructs were respectively 2-, 2.5-, 3-, 4-, and 3-fold greater than that of pGL 3 -promoter vector. As described in Table I, none of the footprint motifs alone or in combination was responsible for the enhancer activity in any of the cell types studied (chorionic cells, Calu-6 cells, and As4.1 cells). Therefore, to study each footprint motif in its native context, point mutations were performed in each of the identified regions, i.e. the AP-1 motif in the B region (B/AP-1mut), the AP-1 motif in the C region (C/AP-1mut), and the AP-1 part   FIG. 3. Gel mobility shift analysis with human chorionic cell nuclear extracts and the human renin gene promoter (؊5747 to ؊5719). A doublestranded oligonucleotide containing the 5Ј-CTCGAGATCTGGGGGTCAGAGGC-AAAATGGAGTCAGTCA-3Ј region of the human renin gene promoter corresponding to footprint B (Ϫ5747 to Ϫ5719) was used as a probe. Competition experiments were performed with a 50-to 100fold molar excess with homologous DNA or with the (Ϫ5656 to Ϫ5615) oligonucleotide (footprint C), the (Ϫ5766 to Ϫ5743) oligonucleotide (footprint A), the (Ϫ5766 to Ϫ5719) oligonucleotide (footprint AB), or the AP-1 or the NF-E2 binding sites oligonucleotides. Specific DNA⅐protein complexes are indicated by an arrow. n.s., nonspecific interactions. of the NF-E2 motif in the C region (NF-E2mut) in pGL 3 -promoter enhancer. In chorionic cells, mutations of the C/AP-1 motif or the NF-E2 motif in footprint C are responsible for the loss of 85 and 82% of the enhancer activity (Fig. 6). When both mutations were performed in the same fragment (C/AP-1mut ϩ NF-E2mut), enhancer activity was abolished. In addition, mutation in the AP-1 motif of the B region had not effect alone but enhanced the effects of mutations in the C region. In reninproducing Calu-6 and As4.1 cells, a roughly similar pattern could be observed except that the smaller enhancer activity was never abolished in any single or double mutants. As expected, any of the mutations had no effect on pGL 3 -promotervector activity in COS-7 cells (data not shown).
Sequence Analysis-To detect putative secondary structures and sequence repeats, the enhancer region was further analyzed. An optimal alignment between the (Ϫ5777 to Ϫ5552) region encompassing the footprint sites of the enhancer and the (Ϫ892 to Ϫ717) region showed that footprints A and B are very well conserved in the proximal promoter with 91 and 100% identity, respectively. In contrast, the footprint C sequence is only 42.4% conserved with the (Ϫ763 to Ϫ731) region (Fig. 7). To determine whether these conserved sequences are functionally interchangeable, the AB(892) region was inserted instead of the AB(5777) within the enhancer sequence. A 2-fold loss of activity was observed in chorionic cells, which is comparable with the effect of deleting the AB sequence. In other reninexpressing cell types, either the replacement or the deletion of the AB sequence had no effect. DISCUSSION Transgenic animal studies (9 -11) did not lead to the identification of cis-acting elements directing cell-specific expression of the human renin gene. Functional analysis of the human renin gene proximal promoter identified positive and negative regulatory elements in chorionic cells (6) and Calu-6 cells (14) and a transcriptional silencer specific to renin-producing cells in the first intron (14,24). From these studies, it has been suggested that there may be cis-acting regions conferring differential and specific renal and extra-renal human renin gene expression located far upstream from the transcription start site, in the distal promoter. Recently, Yan et al. (12) identified an enhancer in the 5Ј-flanking region of the human gene that is very similar to that of the enhancer-mediating mouse ren-1c gene expression in As4.1 cells (25).
The aim of the present study was to identify the regulatory elements in the distal regions of the human renin gene promoter responsible for renin expression in the chorion. A 225-bp enhancer located between bases Ϫ5777 and Ϫ5552 of the human renin gene was shown to activate transcription although to different degrees in the three renin-expressing cell models FIG. 4. Gel mobility shift analysis with human chorionic cell nuclear extracts and the human renin gene promoter (؊5656 to ؊5615). A doublestranded oligonucleotide containing the 5Ј-CTCGAGATCTGGATTGGGGAAAA-TGAGACATCATTACTCACCTGGCT-AC-3Ј region of the human renin promoter corresponding to footprint C (Ϫ5656 to Ϫ5615) was used as a probe. Competition experiments were performed with a 50-to 200-fold molar excess of homologous DNA or with a 100-fold molar excess of the (Ϫ5747 to Ϫ5719) oligonucleotide (footprint B), the (Ϫ5766 to Ϫ5743) oligonucleotide (footprint A), the (Ϫ5766 to Ϫ5719) oligonucleotide (footprint AB), or the AP-1 and NF-E2 binding site oligonucleotides. Specific DNA⅐ protein complexes are indicated by an arrow.
described until now. The degree of activation mediated by the enhancer was 57-fold stimulation in chorionic cells and was weaker in Calu-6 cells (ϳ5-fold) and As4.1 cells (ϳ8-fold). In human cells, regulatory elements were strictly restricted to this 225-bp fragment because no cis-acting regions were detected either downstream, between bases Ϫ5552 and Ϫ2824, or upstream to bases Ϫ8876. Species differences may be responsible for the differences in the activity of the (Ϫ7476 to Ϫ5777) fragment, which activated transcription in mouse As4.1 cells but not in human cells. The enhancer activity of the 225-bp fragment was promoter-independent because it was roughly similar either with the SV40 promoter or the proximal renin promoter. The (Ϫ5777 to Ϫ5552) region acted typically as a classical enhancer in each renin-expressing cell type in both sense and antisense orientations.
DNase I footprinting assays showed that three regions within the 225-bp fragment were protected by chorionic cell nuclear extracts, whereas only one is protected by non-reninproducing COS-7 cell nuclear extracts. A functional analysis of each footprint motif showed that the overall level of stimulation can be accounted for by the addition of the functional effects of each of the protected regions in Calu-6 and As4.1 cells but cannot account for the final 59-fold stimulation of the transcription in chorionic cells. There were no consensus binding sites for known transcription factors in footprint A, but footprints B and C contained a motif similar to the AP-1 binding site that interacts with bZIP factors of the Jun and Fos families of transcription factors (20,21). Specific interactions with the footprint B motif were only competed by homologous DNA but not by oligonucleotides spanning consensus AP-1 binding sites, showing that the AP-1 consensus binding site of the footprint B motif is not involved. In addition to its characterized AP-1 binding site, footprint C also contains a consensus binding site for NF-E2, a member of the AP-1 superfamily (18). Transcription factor NF-E2 is a heterodimeric protein comprising the erythroid-and megakaryocytic-restricted p45 NF-E2 and the small bZIP p18 (26), also known as MafK present in most, if not all tissues (27) or the recently described MafG (23). Gel shift assays with anti-p45 NF-E2 and anti-MafG/K antisera did not show, either by supershift or binding inhibition, any related binding activity to the footprint C motif. Gel shift assays with antisera directed against c-Fos, c-Jun, JunB, and fra-2 did not block or supershift the motif C⅐chorionic cell nuclear extracts complex. These results strongly suggest that chorionic nucleoproteins binding to the footprint C motif are related but distinct from well characterized AP-1 family members.
The strong enhancer activity observed in chorionic cells seems to be conferred by footprint C since point mutations in the NF-E2 or the bona fide AP-1 binding site strongly reduced enhancer activity. In addition, enhancer activity was disrupted by both mutations in chorionic cells. Nevertheless, reconstitution experiments showed that footprint C alone cannot confer the enhancer activity. Contacts between proteins bound at nonadjacent sites on a short DNA fragment may be facilitated by distortion of the DNA structure that enhance the recruitment of proteins to their DNA targets (28). The transcription factors binding to this region in chorionic cells may affect the architecture assembly and function of the regulatory nucleoprotein complex. Alternatively, this human renin enhancer activity may result from the cooperative binding of transcription factors to their respective cognate sites. A chorionic-specific AP-1-related bZip transcription factor or a chorionic cellrestricted intermediary factor recruited by the binding of transcription factors may be responsible for the 59-fold enhancer activity in renin-producing chorionic cells.
In summary, this study presents a more complete picture than was previously available of the overall spatial arrangement of regulatory sequences responsible for the human renin gene expression. Moreover, although different levels of enhancer activity were observed in all three renin-producing cell models, interactions with tissue-specific trans-acting factors binding to this enhancer may be responsible for the differential transcriptional regulation of the human renin gene in extrarenal and renal tissue. This is the first case that such a sequence has been shown to be efficient and active in chorionic renin-producing cells. Isolation of the transcription factors involved in the regulation of this enhancer should provide impor- FIG. 7. a, sequence alignment of conserved residues between the (Ϫ5777 to Ϫ5605) and the (Ϫ892 to Ϫ717) region. The three DNase I-protected regions, A, B and C, are boxed. Conserved residues are indicated by stacked dots. Broken lines indicate gaps introduced for optimum alignment. b, functional activity of the conserved AB region. Footprints AB of the Ϫ5777 region were either deleted or exchanged by the AB region of the Ϫ892 region in the pGL 3 -promoter-enhancer plasmid context. Each value is the mean of at least three independent transfection experiments performed in triplicate wells. Error bars represent S.E. tant new insights into the molecular events responsible for the differential levels of renin gene expression in the chorion and other tissues.