Transcriptional Regulation of the Rat Renin Gene by Regulatory Elements in Intron I

Library Screening

A Sprague-Dawley (SD) genomic library (Stratagene) was screened by the use of standard methods¹⁷ and an EcoRI/XbaI fragment of the rat renin gene comprising 1.2 kb 5′- flanking region, exon I, and 316 bp intron I as a radiolabeled probe. Four clones (r1, r4.1, r4.2, and r8) could be isolated and mapped by restriction analysis (Figure 1⇓). A 9.7 kb SacI fragment from clone r8 containing 3.9 kb 5′-flanking region, exon I, and the entire intron I was then subcloned to generate the basic renin gene vector.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1.

Restriction map of the newly isolated rat renin genomic clones r1, r4.1, r4.2, and r8. Top, originally isolated sequence (RATRENAA¹⁸). Exons (E1 to E9) are shown by open boxes interrupted by introns I1 to I8. The probe used for the isolation was a 1.6 kb EcoRI/XbaI fragment containing exon 1 indicated by a bar. Bg indicates BglII; E, EcoRI; X, XbaI.

Site-Directed Mutagenesis of the 2 Translation Start Codons

PCR-directed site-specific mutagenesis of the 2 translation start codons ATG I and ATG II (at position +33 bp and +48 bp with respect to the transcription start site¹⁸ was performed by the use of a 4 primer system with a pair of overlapping internal primers. The total length of the sequence remained unchanged. The primers were designed as follows: 5′ primer rR12 [5′ TCT AGA GAC ACG CAT AAT CAG TGC 3′] and 3′ primer rR13 [5′ CAG AGA GGC GGC CGC CTC CCG CCG ATC TGG TCC 3′] for mutagenesis of start codons ATG I/II and 5′ primer rR14 [5′ CGG GAG GCG GCC GCC TCT CTG GG 3′] and 3′ primer rR15 [5′ CCT GCA AAA CTC TAG AGG GGC CG 3′] for mutagenesis of start codon ATG II (bold letters indicate mutated residues). In addition to the mutagenesis of ATG II, a new NotI restriction site was introduced in the sequence. PCR products were subcloned in a T-vector system (Promega), sequenced, combined at the NotI restriction site, and integrated as a 1.8 kb XbaI fragment in the basic renin gene vector, replacing the former genomic part with functional ATG start codons.

Reporter Gene Constructs

Two reporter genes, β-galactosidase from E. coli (lacZ) and firefly luciferase, were used in parallel. The basic lacZ vectors contained a 121 bp nuclear translocation signal (NTS) fused in frame to the lacZ gene and either a 850 bp intron and polyadenylation signal (pA) cassette from SV40 or a splice acceptor site (SAS) from the c-myc gene and a pA cassette of SV40 in a pGEM4 plasmid backbone (Promega). The basic luciferase vector, pΔLux, was generated as described.¹⁹ For constructs containing introns, pΔLux was modified by adding the SAS of the c-myc gene upstream of the luciferase gene.

Reporter gene constructs with parts of the renin gene or the cytomegalovirus immediate early gene (CMV) promoter were cloned by standard methods. Restriction enzymes or Bal31 nuclease¹⁷ was applied for the shortening of the intron-I fragments.

Cell Culture and Transfection Experiments

Two cell lines, 293 (human embryonic kidney cells²⁰ ) and L8 (rat skeletal muscle myoblasts²¹ ) were used in the present study. Both were cultured under optimal conditions and transfected with 5 μg of the respective plasmids by the CaPO₄ coprecipitation method.¹⁹ As an internal control, 2 μg of pUHC13-1²² (for lacZ plasmids) or pCH110 (Pharmacia, for luciferase plasmids) containing the lacZ gene under control of the SV40 early gene promoter were added to the precipitation mixture. Both reporter gene activities were measured in each cell extract 40 hours after transfection as described.¹⁹ The lacZ activity measured was normalized to the luciferase activity of the same extract if the promoter to test was linked to lacZ and vice versa. The values were statistically analyzed by use of ANOVA.

Preparation of Nuclear Extracts and Gel-Retardation Experiments

293 and L8 cells were grown to 75% confluence, harvested after a short trypsin treatment, and pelleted at low speed. Nuclear extracts were prepared as described by Dignam et al,²³ except that HEPES was replaced by Tris-HCl, pH 7.9. Protein content was quantified according to Bradford²⁴ by the use of the BioRad protein assay system. Salt content was determined by conductivity measurement. Gel-retardation experiments were performed as follows: In 20 μL total reaction volume shift buffer (20 mmol/L HEPES, pH 8.4, 60 mmol/L KCl, 4% Ficoll) was mixed with 2 mmol/L DTT, 2 μg bovine serum albumin, 2 μg poly(dI-dC) heteropolymer, 2.5 to 7 μg nuclear extract, and 15 000 to 25 000 cpm end-labeled probe. After 30 minutes incubation at 30°C, samples were loaded to a 4% to 6% (39:1) polyacrylamide gel in 0.25× TBE (25 mmol/L Tris-HCl, 25 mmol/L boric acid, 0.05 mmol/L EDTA) and electrophoresed at 15 V/cm. Gels were dried and exposed to Kodak XAR-5 X-ray film.

DE 52 Purification and Heparin-Sepharose Chromatography of Nuclear Extracts

Purification and chromatographic separation was carried out at 4°C essentially as described by Scheidereit et al.²⁵ A DEAE cellulose column (DE 52, Whatman, 1 mL packed column volume [cv] per 4 to 10 mg protein load) was equilibrated in buffer C (20 mmol/L Tris-HCl, pH 7.9, 25% glycerol, 0.42 M NaCl, 1.5 mmol/L MgCl₂, 0.2 mmol/L EDTA, 0.5 mmol/L DTT, and 0.5 mmol/L PMSF). Crude nuclear extracts were adjusted to 0.42 M NaCl, passed over the column, eluted with 1 additional cv buffer BC 500 (20 mmol/L Tris-HCl, pH 7.9, 20% glycerol, 0.5 M KCl, 0.2 mmol/L EDTA, 0.5 mmol/L DTT, and 0.5 mmol/L PMSF) and dialysed against 60× volume buffer BC 100 (buffer BC with 0.1 M KCl). The purified material was loaded onto a heparin-sepharose column (1 mL cv per 3.5 mg protein load, equilibrated with 6 cv BC 100) at 2 cv per hour, washed with 2 cv BC 100 at 2 cv per hour and eluted with a linear gradient (5 cv) from 0.1 to 0.5 M KCl, again at 2 cv per hour. In a final step, the column was washed with 0.3 cv of BC 1000 (buffer BC with 1 M KCl) to elute remaining proteins. The gradient was collected in 32 fractions and protein and salt content was analyzed as described above.

Methylation-Interference Assay

For the methylation-interference assay, probes were labeled by a fill-in reaction of restriction sites by the use of exo-Klenow polymerase (Stratagene) and [³²P]dGTP or [³²P]dATP. Partial methylation of the probes was subsequently performed essentially as described,¹⁷ by the use of 2×10⁷ cpm probe and 1 μg poly(dI-dC) as carrier for the methylation reaction at 25°C and with 1.5 M sodium acetate/1 M 2-mercaptoethanol for reaction termination. After 2 ethanol precipitations and rinsing, the partially methylated DNA was used for gel-retardation assays as described above except that the gels were not dried before X-ray film exposure. All visible retarded complexes as well as the unbound DNA fraction were cut out from the gel, purified, and cleaved in 100 μL 10% piperidine at 90°C for 30 minutes. After precipitation and rinsing, the DNA was loaded onto an 8% polyacrylamide-sequencing gel and visualized by autoradiography.

Cloning of the Rat Renin Gene

Previously,¹⁸ 1.2 kb of the rat renin promoter region had been cloned. For the present study, a longer 5′-flanking region was considered to be necessary to include as many regulatory elements of the promoter region as possible. By screening of a rat genomic library (Stratagene), 4 clones (r1, r4.1, r4.2, and r8) could be isolated, comprising 16.5 kb 5′-flanking region, all exons and introns and 5.8 kb 3′-flanking region (Figure 1⇑). A comparison between these clones revealed a BglII restriction fragment length polymorphism probably due to different amounts of the tandem repeats present at this site in intron I of the rat renin gene.^{18 26}

Transcriptional Effects of Intron I

Conversion of the coding exon I of the rat renin gene into a noncoding one by PCR-directed site-specific mutagenesis of the 2 translation start sites at position +33 and +48, respectively, served as a starting point for the subsequent construction of 30 different reporter gene constructs, 16 containing the 3.9 kb renin 5′-flanking region and 14 containing the 760 bp CMV promoter element. For each promoter fragment, experiments with 2 reporter genes (lacZ and luciferase) were performed yielding essentially identical results.

All constructs were transfected into 2 different cell types, the human embryonic kidney cell line 293 not expressing renin² and the rat skeletal myoblast cell line L8 containing renin transcripts detectable by RT-PCR (not shown). Schematic representations of all constructs used in the transfection experiments and their transcriptional activity are shown in Figure 2⇓ (renin promoter) and Figure 3⇓ (CMV promoter).

• Download figure
• Open in new tab
• Download powerpoint

Figure 2.

Transcriptional activity of renin promoter/renin intron constructs. Constructs containing 3.9 kb of the rat renin promoter and exon 1 (rR1) and the indicated segments of intron I (rRIΔ1 to 11, rRI 1) were transfected into 293 (A) or L8 (B) cells and reporter gene activities were measured after 40 hours. The data were normalized to a cotransfected construct carrying another reporter gene under the control of an ubiquitously active promoter and are given as percentage of the activity of rR1 (set as 100%). Mean±SE of at least 3 independent experiments performed in triplicate are shown. Constructs carrying the renin promoter (as1) and the intron I (asI 1) in antisense orientation served as negative controls. * indicates P<0.05 versus the next shorter construct; #, P<0.05 versus rR1.

• Download figure
• Open in new tab
• Download powerpoint

Figure 3.

Expression of CMV promoter/renin intron constructs. Constructs containing 760 bp of the CMV promoter (C1) and renin exon 1 and the indicated segments of intron I (CIΔ1 to 11, CI 1) were transfected into 293 (A) or L8 (B) cells and reporter gene activities were measured after 40 hours. The data were normalized to a cotransfected construct carrying another reporter gene under the control of an ubiquitously active promoter and are given as percentage of the activity of C1 (set as 100%). Mean±SE of at least 3 independent experiments performed in triplicate are shown. The reporter gene activities for all intron-bearing constructs was significantly different than the 1 of C1. * indicates P<0.05 versus the next shorter construct.

In 293 cells, addition of parts of the rat renin intron I to reporter gene constructs driven by the homologous renin 5′-flanking region resulted in a down-regulation of reporter gene activity (Figure 2A⇑). Integration of the first 65 bp of intron I (rRIΔ1) led to a significant 2-fold decrease of lacZ or luciferase activity from 100±6.7% to 51.5±2.5%. Elongation of intron I to 117 bp (rRIΔ3) and further to 275 bp (rRIΔ5) resulted in additional reductions of reporter gene activities to 17.2±1.9% and 9.0±2.8%, respectively. Although longer sequences of intron I (rRIΔ7,9,10,11) showed only insignificant additional changes in reporter gene activity, inclusion of almost the entire intron I, except for the last 40 bp at the 3′ end, (rRI1) led to a further significant reduction of the transcription level to 2.7±0.9%. This is exactly in the range of negative control constructs bearing the renin sequences in antisense orientation (as1 and asI1).

Transfection of the same constructs into L8 cells revealed a completely opposite pattern of reporter gene activity (Figure 2B⇑). Integration of the first 65 bp of intron I (rRIΔ1) increased lacZ or luciferase activity 3-fold from 100±2.0% to 296±12%. Elongation of intron I to 117 bp (rRIΔ3) decreased transcription activity by a factor of 2.3 to 131.5±6.5%. Variations between 117 and 410 bp length of intron I did not display any significant changes. Further elongation of intron I to 538 bp (rRIΔ9) and 750 bp (rRIΔ11), however, resulted first in a significant decrease in reporter gene activity to 79.6±2.0%, followed by an increase to 178.5±3.5%. Inclusion of the full-length intron (rRI1) significantly decreased reporter gene activity to 77.9±4.2%, compared with the intron-less construct (rR1).

Transfection experiments with renin intron constructs controlled by the heterologous CMV promoter in 293 cells displayed a similar but even stronger inhibitory influence of intron I on transcriptional activity as seen in the corresponding renin constructs (Figure 3A⇑). Integration of the first 65 bp of intron I (CIΔ1) decreased the lacZ or luciferase activity 5-fold to 16.9±1.5% (instead of 2-fold in the case of construct rRIΔ1). Elongating intron I to 103 bp caused a reduction of reporter gene activity to 6.3±0.4%. However, further elongation elicited no additional significant effect except for the addition of the full-length intron I lowering the transcription level to 1.1±0.3%.

The transcriptional activities of the CMV promoter/renin intron constructs transfected into L8 cells were basically comparable to the ones in 293 cells, but contrasted to the results obtained with the corresponding renin promoter/renin intron constructs in L8 cells (Figure 3B⇑). Integration of the first 65 bp of intron I (CIΔ1) decreased reporter gene activity by a factor of 3.2 to 30.8±5.2%, an activity level that was also found in intron deletion constructs with a length of intron I between 103 and 432 bp. Further elongation of intron I to 538 bp (CIΔ9), 750 bp (CIΔ11), and 5.3 kb (CI1) resulted again in the wavy pattern of reporter gene activity seen with renin promoter/renin intron constructs with activity levels of 9.4±1.6%, 28.5±2.8%, and 9.5±2.1%, respectively. Thus, the addition of the full-length renin intron reduced transcriptional activity 10-fold for the CMV promoter but only by 22% for the renin promoter in L8 cells.

DNA-Protein Interactions in Intron I

The sequence causing the most obvious changes in reporter gene activity, ie, the coding part of exon I (79 bp) and the first part of intron I (65 bp) was further analyzed for possible DNA-protein interactions. Nuclear extracts of both cell lines were fractionated on heparin-sepharose columns and the fractions were analyzed by gel-retardation assays. As probes, different end-labeled fragments were used containing either the exon I-intron I sequence mentioned above (+45 to +189 bp) or a part of it.

The elution profile of the heparin-sepharose column using 293 nuclear extracts is depicted in Figure 4A⇓, a typical gel-retardation experiment with gradient-separated extracts and a fragment covering the first 65 bp of intron I (+124 to +189 bp) as probe is shown in Figure 4B⇓. The protein-DNA complexes separated by the shift experiment were named B1 to B18 with regard to their degree of retardation and their affiliation to low salt (B1-B8) or high salt (B9-B18) portions of the gradient. Although most complexes were common to both 293 and L8 nuclear extracts, B4, represented by faint (Figure 4B⇓) to moderate (data not shown) signals, was exclusively restricted to 293 extracts. On the other hand, a complex named B15′, presenting a weak and variable signal, could only be observed in L8 extracts (data not shown).

• Download figure
• Open in new tab
• Download powerpoint

Figure 4.

Protein-DNA complexes in the first 65 nucleotides of rat renin intron I. Nuclear extracts from 293 cells were fractionated by a 100 to 700 mmol/L KCl gradient on a heparin-sepharose column and the fractions were analyzed for salt and protein content by conductivity measurements (□) and the Bradford method (○), respectively (A), and for DNA-binding activity by a gel-retardation assay (B) with 1 μL of each fraction and the first 65 nucleotides of intron I endlabelled with ³²P. Triangles mark the complexes used for the methylation-interference assay (see Figure 5⇓). C indicates no protein; load, 1 μL of 293 nuclear extract before heparin-sepharose chromatography.

To determine the sequence specificity of the different binding activities methylation-interference assays were performed. The first 65 bp of intron I were labeled at the 5′-end and randomly methylated at an average of 1 guanine (G) or (to a lesser extent) adenine (A) residue per probe molecule. After incubation with the heparin-sepharose fractions 14, 20, 24, and 30 derived from 293 nuclear extract, protein-DNA complexes (B1-B18) were separated by polyacrylamide-gel electrophoresis, purified from the gel, cleaved at methylated residues and subsequently analyzed on a denaturing sequencing gel as outlined in Methods (Figure 5⇓). The methylated probe cleaved only by piperidine without prior incubation with nuclear extracts and separation by electrophoresis as well as the free probe served as controls. Both displayed an identical G/A sequence pattern indicating that the reaction procedures have no effect per se on the methylated probe. Analyzing the sequence pattern of all complexes only B4 showed significant differences in the signal intensity of G and A residues (Figure 5⇓). The signal intensities of the residues at positions +162 (G), +163 (A), +166 (G), and +167 (G) decreased markedly whereas that of +168 (A) increased significantly indicating that methylation at these nucleotide positions interfered with sequence-specific protein binding. Analysis of the surrounding sequence profile of the putative binding region revealed a dodecamer sequence (positions +159 to +171) with 8 out of 12 nucleotide positions arranged in a palindrome-like configuration. Neither B12 (Figure 5⇓) nor any other protein-DNA complex in the gel-retardation experiments exhibited sequence-specific binding in the methylation-interference assay (data not shown).

• Download figure
• Open in new tab
• Download powerpoint

Figure 5.

Sequence-specific DNA-binding in rat renin intron I. Protein-DNA complexes were formed with the first 65 nucleotides of intron I partially methylated and endlabelled with ³²P and 293 nuclear extracts. After polyacrylamide gel electrophoresis, complexes B4 and B12 as well as free probes from 2 lanes (see Figure 4⇑) were eluted from the gel, cleaved at the methylated residues and analyzed on a sequencing gel in comparison to the untreated probe. Only complex B4 showed selective binding to DNA that is unmethylated at the positions indicated by the white letters. The pseudo-palindromic sequence surrounding these positions is indicated by arrows. Positions are numbered relative to the transcriptional start site.