INTRODUCTION

Top Abstract Introduction Methods Results Discussion References

A CHLORIDE CHANNEL, ClC-K1, found exclusively in the thin ascending limb of Henle's loop (tAL), is one of the nine currently known members of the mammalian ClC chloride channel family (1, 2, 11, 13, 23, 26, 27, 31, 32). ClC-1 is a skeletal muscle-specific chloride channel that accounts for the majority of membrane conductance of plasma membrane in skeletal muscle (26). Mutations of the ClCN1 gene have been reported in dominant and recessive types of myotonia (15). Among the ClC family, the only members other than ClC-1 that are expressed in a tissue-specific fashion are ClC-K1 (32) and ClC-K2 (1). ClC-K1 and ClC-K2 are very homologous to each other, with an amino acid identity of ~80% in rat K1 and K2 (1, 32) and ~90% in human K1 and K2 (14, 30). Recently, we determined chromosomal localization of human ClC-K1 and K2 genes (CLCNKA and CLCNKB, respectively) using the two-color fluorescence in situ hybridization (FISH) method (22). Both loci were found to be closely linked at 1p34-36 (22), suggesting that both genes arose from a common ancestor gene by gene duplication. Although they are very similar in structure, their intrarenal localizations are completely different: ClC-K1 is present in the thin ascending limb of Henle's loop (tAL) in the inner medulla (34), and ClC-K2 is present in the thick ascending limb of Henle's loop and collecting ducts (1). The tAL has the highest transepithelial chloride permeability among nephron segments (10), and in vitro perfusion studies (38) have suggested the presence of chloride channels within it. The functional characteristics of ClC-K1 expressed in Xenopus oocytes very closely matched those of chloride transport in tAL (34), and immunohistochemistry revealed that ClC-K1 is present in both the apical and basolateral plasma membranes of tAL (34). Based on these observations, ClC-K1 has been identified as a major chloride channel that mediates transepithelial chloride transport in tAL (34). Chloride transport in tAL constitutes a countercurrent system in the inner medulla and is thought to be important for the urinary concentration mechanisms (10). In fact, the expression of ClC-K1 is augmented in dehydrated rats (32), suggesting its involvement in urinary concentration mechanisms.

Little is known about the molecular mechanism(s) of tissue-specific expression of genes in the kidney. Recently, several cDNAs such as aquaporin-2 (6), V₂ vasopressin receptor, Na-K-Cl cotransporter (8), and Na-glucose cotransporter (SGLT2; see Ref. 12) were found to be expressed exclusively in the kidney. The 5'-flanking regions of the aquaporin-2 gene (33), V₂ vasopressin receptor gene (16), and Na-K-Cl cotransporter gene (9) have all been isolated, but no common sequence or mechanism(s) of cell-specific expression has been identified. We have examined the rat ClC-K1 gene as a model for kidney-specific gene expression. The 5'-flanking region of the ClC-K1 gene was isolated to commence examination of the molecular basis for kidney-specific expression of the rat ClC-K1 gene. The ~2.5-kb pair 5'-flanking region of the rat ClC-K1 gene was isolated, and its kidney cell-specific promoter activity was demonstrated. A regulatory element required for cell-specific and maximal promoter activity in the 5'-flanking region was further determined by deletion analysis and gel retardation assay. We found that a purine-rich element located ~40 bp upstream of the transcription start site was required for maximal promoter activity of the rat ClC-K1 gene, and the cell-specific promoter activity demonstrated in the ClC-K1-expressing cell line (IM cells) suggested this purine-rich element's involvement in the kidney-specific expression of ClC-K1 in vivo.

	METHODS

Top Abstract Introduction Methods Results Discussion References

Library screening. A rat genomic library in DASHII was purchased from Stratagene and was screened by filter hybridization with a ³²P-labeled rat ClC-K1 cDNA (32). Positive clones were plaque purified, and the genomic inserts were restiction mapped using Sal I, EcoR I, and Xba I. Restriction fragments were transferred to nylon membrane, and exon-containing fragments were identified by hybridization using various parts of ClC-K1 cDNA as ³²P-labeled probes.

Primer extension analysis. Primer extension analysis was performed using an antisense oligonucleotide (5'-ACGCAGTCCCACGAGTTCTTCCAT) that was complementary to the ClC-K1 cDNA [nucleotide (nt) 44-67; see Ref. 32]. This oligonucleotide was end labeled with [-³²P]ATP using polynucleotide kinase. Ten micrograms of poly(A)⁺ RNA from rat kidney were hybridized with 5 × 10⁵ counts/min (cpm) probe at 30°C for 3 h in 250 mM KCl, 10 mM tris(hydroxymethyl)aminomethane (Tris) · Cl, pH 8.0, and 1 mM EDTA, and the sample was reacted with 500 units of Superscript II reverse transcriptase (Life Technologies) in 75 mM KCl, 10 mM MgCl₂, 0.25 mM EDTA, 20 mM Tris · Cl, pH 8.0, 10 mM dithiothreitol, 0.25 mM dNTP, and 100 µg/ml actinomycin D. The sample was quenched with EDTA, extracted with phenol-chloroform, and precipitated with ethanol. Reaction products were analyzed on 6% polyacrylamide sequence gel.

5'-Rapid amplification of cDNA ends of rat ClC-K1 cDNA. 5'-Rapid amplification of cDNA ends (RACE)-Ready cDNA (Clontech, Palo Alto, CA) was used to clone more of the 5'-end portion of the rat ClC-K1 cDNA. Two ClC-K1 gene-specific primers [gene-specific primer 1, 5'-AACCTCCCCGGTATAGCCATTTGTG from nt 275-298 of rat ClC-K1 cDNA (32); gene-specific primer 2, 5'-TTCTGACACCTCGGCGGATCGGT from nt 121-143] were prepared, and the first polymerase chain reaction (PCR) was performed with gene-specific primer 1 and the anchor primer provided in the kit. Nested PCR was then performed using the first PCR product as a template with the gene-specific primer 2 and the anchor primer. PCR products were subcloned into pSPORT1 vector and sequenced.

Ribonuclease protection assay. To determine the transcriptional start site of the rat ClC-K1 gene in vivo, a 192-bp genomic fragment containing a putative transcription start site determined by primer extension analysis (from 125 to +64) was amplified by PCR and cloned into pSPORT1 vector at EcoR I and BamH I sites. Riboprobe in antisense orientation was synthesized using SP6 RNA polymerase and hybridized with 10 µg of poly(A)⁺ RNA from rat kidney [1 × 10⁵ cpm probe in hybridization solution (80% formamide, 40 mM piperazine-N,N'-bis(2-ethanesulfonic acid), pH 6.4, 400 mM sodium acetate, and 1 mM EDTA)]. After overnight incubation, the reaction was treated with ribonuclease (RNase) A (0.5 U/ml) and T1 (10 U/ml), precipitated, and analyzed by electrophoresis in a 10% denaturing polyacrylamide gel.

To determine the transcriptional start site of a reporter gene construct in transient transfection experiments, a 289-bp DNA fragment covering the portions of ClC-K1 promoter (from nt 128 to +64) and pGL2 basic (a 97-bp fragment from Bgl II site) was cloned into pSPORT1 vector. Riboprobe in antisense orientation was synthesized using T7 RNA polymerase and hybridized with 10 µg of total RNA from IM cells transfected with 20 µg of pLuc-128.

To measure the abundance of mouse ClC-K1 mRNA level in cultured cells, a 294-bp fragment encoding 98 amino acids from the first methionine of mouse ClC-K1 was cloned into pSPORT1 vector. Riboprobe in antisense orientation was synthesized using T7 RNA polymerase and hybridized with 10 µg of total RNA from IM cells and NIH-3T3 cells.

Cell culture. Inner medulla of the kidney from a transgenic mouse harboring temperature-sensitive simian virus 40 (SV40) large T antigen gene (a generous gift from Dr. M. Obinata, Tohoku University) was minced and treated with 1 mg/ml collagenase B (Boehringer Mannheim) for 30 min at 37°C, plated on standard culture dishes in RITC80-7 medium (Kyokuto Pharmaceutical Industrial, Tokyo, Japan) supplemented with 5% fetal bovine serum, 10 µg/ml transferrin, 1 µg/ml insulin, and 10 ng/ml epidermal growth factor, and cultured at 34°C. Each colony was cloned using a cloning cylinder, and the ClC-K1 mRNA abundance in each cell line was measured by RNase protection assay. The outer medullary collecting duct (OMCD) cells (5), kindly provided by Dr. H. Endoh (Kyorin University, Tokyo, Japan), were cultured under the same condition as used for the IM cells. NIH-3T3 cells and LLC-PK₁ cells were maintained in Dulbecco's modified Eagle's medium containing 10% fetal bovine serum.

Construction of reporter plasmids. After various 5'-flanking regions of rat ClC-K1 gene were synthesized by PCR using various 5'-primers containing Mlu I site and a common 3'-primer (ccagatctCAGTCCACCCTGCAGAGGTCCTGCGTGGCTG, nt +45 to +75) containing the Bgl II site, they were ligated into pGL2 basic vector at Mlu I and Bgl II sites. A pGL2 control containing a luciferase gene driven by the SV40 early region promoter/enhancer was used as positive control, and empty pGL2 basic was used as negative control. The sequences of 5'-flanking regions obtained by PCR were verified by sequencing. A reporter gene constuct containing the ~2.5-kb 5'-flanking region was generated as follows: a 5-kb Xba I-Kpn I (Fig. 1) fragment was subcloned into pSPORT1, and an Mlu I-BstX I fragment in this vector was then converted to a pLuc-495 construct cut with Mlu I and BstX I. 

View larger version (40K): [in this window] [in a new window]   Fig. 1.   Gene structure (A) and sequence (B) of the 5'-flanking region of rat ClC-K1 gene. A: filled and open boxes represent exons corresponding to the coding and noncoding region of rat ClC-K1 cDNA, respectively. Exon 2, shown as a hatched box, was identified by RACE analysis. B: +1 indicates the transcription start site. Arrows indicate the 5'-end of the luciferase constructs shown in Fig. 5. The purine-rich region is underlined. Dots indicate the 5'-ends of various RACE clones. HSF, heat shock factor.

Transient transfection and reporter gene assay. Luciferase reporter plasmids were introduced into cultured mammalian cells by electroporation. When cells plated on 150-mm plastic dishes reached 60-70% confluency, they were detached with 0.25% trypsin/EDTA, neutralized by complete medium, pelleted by brief centrifugation, resuspended in 500 µl of buffer (30.8 mM NaCl, 120.7 mM KCl, 1.46 mM KH₂PO₄, 8.1 mM Na₂HPO₄, 10 mM MgCl₂) containing 20 µg of pGL2 vectors and 5 µg of pSV--galactosidase vector (Promega), transferred to a cuvette (0.4 mm width), and electroporated at settings of 370 V and 960 µF. Ten minutes after electroporation, the cells were resuspended in prewarmed complete medium and seeded in a 60-mm dish. Forty-eight hours after transfection, the cells were harvested for luciferase and -galactosidase assay (Promega). Transfection efficiency was corrected using -galactosidase activity.

Electrophoretic mobility shift assay. Synthesized sense and antisense oligonucleotides were annealed and radiolabeled at the 5'-end with polynucleotide kinase and [-³²P]ATP. Nuclear extracts from IM and NIH-3T3 cells were prepared as described previously (3). After the nuclear extracts were incubated with 0.5 ng of radiolabeled DNA at room temperature for 30 min in a buffer containing 25 mM N-2-hydroxyethylpiperazine-N'-2-ethanesulfonic acid-KOH (pH 7.9), 90 mM KCl, 0.5 mM EDTA-NaOH (pH 8.0), 0.5 mM dithiothreitol, 0.5 mM phenylmethylsufonyl fluoride, 10% glycerol, 1 µg poly(dI-dC), and, if indicated, competitors, they were electrophoresed through 6% nondenaturing polyacrylamide gels (19:1 acrylamide-bisacrylamide) containing 10% glycerol in TGE buffer (50 mM Tris · HCl, pH 8.5, 380 mM glycin, and 2 mM EDTA-NaOH, pH 8.0) for 2 h at 4°C. Upon completion of electrophoresis, the gels were dried, and the protein-DNA complexes were visualized by autoradiography.

Statistical analysis. Multiple comparisons were performed using analysis of variance, and P < 0.01 was considered to be statistically significant.

	RESULTS

Top Abstract Introduction Methods Results Discussion References

Cloning of the rat ClC-K1 gene and identification of the transcription start site. A rat genomic library in DASHII vector was screened with rat ClC-K1 cDNA (32), and a clone (GK1-2) of ~18 kb was obtained. Southern blot analysis using various portions of rat ClC-K1 cDNA and subcloning of various restriction fragments and their partial sequencing revealed that GK1-2 contained sequences at the 5'-end of the rat ClC-K1 cDNA. These results suggested that GK1-2 was likely to contain the 5'-flanking promoter region of rat ClC-K1 gene (Fig. 1A).

The transcription initiation site was determined using primer extension analysis, ligation-anchored PCR, and RNase protection assay. The results of ligation-anchored PCR using 5'-RACE-Ready cDNA revealed the presence of more 5'-end sequences of ClC-K1 cDNA than was found in the original cDNA (Fig. 1B; see Ref. 32). Furthermore, the sequences newly identified by 5'-RACE were also observed just upstream of the 5'-end of the original cDNA sequence in the isolated genomic clone, clearly demonstrating that the genomic clone isolated in this study contained the 5'-flanking region of the rat ClC-K1 gene. As shown in Fig. 1B, the longest 5'-RACE clone contained 20-bp more of the 5'-untranslated region. As shown in Figs. 1A and 2A, cloning of 5'-RACE products also revealed the existence of another exon (85 bp) not present in the original cDNA, suggesting the existence of two types of ClC-K1 mRNAs.

View larger version (9K): [in this window] [in a new window]   View larger version (70K): [in this window] [in a new window]   View larger version (68K): [in this window] [in a new window]   Fig. 2.   A: schematic representation of transcriptional start site of ClC-K1 gene determined by primer extension analysis (B) and ribonuclease protection assay (C). Sequence ladders represent sequences unrelated to the ClC-K1 gene and were used only as size markers. D: transcriptional start site of a reporter gene (pLuc-128) in IM cells. Ribonuclease protection assay was performed using a radiolabeled riboprobe (289 bp) covering the native transcriptional start site. Lane 1, total RNA (20 µg) from untransfected IM cells. Endogenous mouse ClC-K1 message in IM cells was not detected with the rat ClC-K1 riboprobe. Lane 2, total RNA (20 µg) from IM cells transfected with pLuc-128 (20 µg). The length of the longest protected fragment was 160 bp, indicating that the transcriptional start site of pLuc-128 in IM cells was located at +2 in Fig. 1B.

The result of primer extension is shown in Fig. 2B. An end-labeled antisense oligonucleotide complementary to the 5'-untranslated region of rat ClC-K1 cDNA (nt 44-67; see Ref. 32) was annealed to poly(A)⁺ RNA from rat kidney and elongated with Moloney murine leukemia virus reverse transcriptase (SuperScript II; Life Technologies). A 108-bp major extended product and a 193-bp product were obtained, indicating that the 5'-end of the rat ClC-K1 gene was located 41 bp upstream of the 5'-end of the original cDNA. Because the difference of nucleotide length between long and short products matched the length of exon 2 (Figs. 1A and 2A), the longer product probably corresponded to the mRNA containing exon 2.

To confirm the initiation site, RNase protection assay was performed. A 192-bp riboprobe covering the initiation site was prepared and hybridized with poly(A)⁺ RNA from rat kidney. After digestion with RNase A and T1, the protected probe was analyzed by a denaturing polyacrylamide gel. As shown in Fig. 2C, a 64-bp protected fragment and several shorter fragments were detected. Because the 3'-end of the riboprobe corresponded to nt 23 of the original cDNA, the existence of a 64-bp protected fragment indicated that the transcription start site was present 41 bp upstream of the 5'-end of the original cDNA. Thus the transcription start site determined by the primer extension was confirmed by an independent method.

To check whether the transcription of a hybrid reporter gene was also initiated properly in the transient expression in IM cells, RNase protection assay was performed using the mRNA from pLuc-128-transfected IM cells. The length of the protected RNA probe was 160 nt (Fig. 2D, lane 2) after hybridization and RNase treatment, indicating that the transcriptional start site of a hybrid reporter gene was present 129 nt downstream of 128, i.e., at +2 in Fig. 1B. This is close enough to the native transcriptional start site.

Sequence of the proximal 5'-flanking region of the rat ClC-K1 gene. A fragment of GK1-2 clone from exon 1 to the Bgl II site was subcloned and sequenced (Fig. 1B). The transcription start site, the 5'-ends of various 5'-RACE products, and the 5'-ends of various reporter gene constructs are visible on Fig. 1. Although the 5'-flanking region lacked a TATA box and SpI site, it contained an AP-3 site at position 20, a consensus binding site for heat shock factor at 82, and several AP-2 sites (at 174, 225, 235, 290, and 316) and E-boxes (at 10, 174, 201, and 242). At positions 22, 426, and 446, the sequences matched the consensus binding site (TKNNGNAAK) for CCAAT/enhancer binding proteins. There was no adenosine 3',5'-cyclic monophosphate (cAMP)-responsive element, and none of the sequences were similar to osmotic-responsive elements (4, 21, 29). Purine-rich sequences were found at positions 40, 225, and 286. The sequence of the rat ClC-K1 promoter was aligned with the sequences of other cloned kidney-specific promoters of human aquaporin-2 (33), rat V₂ vasopressin receptor (16), and murine Na-K-Cl cotransporter (9), but no significant regions of homology were identified.

Cell-specific activity of the rat ClC-K1 promoter. To determine whether the 5'-flanking region of ClC-K1 gene had a functional promoter, reporter gene assay was performed using ClC-K1-expressing and -nonexpressing cells. Because none of the established kidney-derived cell lines, which included OMCD cells and LLC-PK₁ cells, expressed ClC-K1 (data not shown), tubular cell lines were established from transgenic mice harboring temperature-sensitive SV40 large T antigen gene as described previously (37). The established cell lines were screened for the expression of ClC-K1 message. Figure 3 shows the result of RNase protection assay performed to measure ClC-K1 message in NIH-3T3 fibroblasts and one of the established cell lines, i.e., IM cells. IM cells expressed ClC-K1 mRNA, which had an mRNA abundance (at 34°C) of ~ of that in the whole mouse kidney. In NIH-3T3 cells, no ClC-K1 message was detected. We picked the IM cells as ClC-K1-expressing recipient cells in reporter gene assay. The ~2.5-kb fragment of the 5'-flanking region cloned upstream to a promoterless luciferase reporter gene in pGL2 basic (pLuc-2500) was transfected into IM cells, OMCD cells, NIH-3T3 cells, and LLC-PK₁ cells using the electroporation method. Promoter activity was inferred from light output normalized for differences in transfection efficiency. Because there were no significant differences in the luciferase activities of a positive control vector, pGL2 control, in any of the cell lines (data not shown), the promoter activity of each construct in various cell lines was expressed as percent of pGL2 control to average the results of five experiments. Figure 4 shows the normalized light output after transfection. Transfection with pGL2 basic produced low levels of luciferase activity in all cell lines (343 ± 21 light units in IM, 289 ± 30 light units in OMCD, 326 ± 18 light units in NIH-3T3, and 350 ± 18 light units in LLC-PK₁ vs. 156 ± 12 in blank, mean ± SD, n = 3). As shown in Fig. 4, the luciferase activity resulting from transfection with pLuc-2500 plasmid into IM cells (~30% of pGL2 control) was much higher (P < 0.01) than the luciferase activities resulting from transfection into other cell lines (1-5% of pGL2 control). These results clearly demonstrated that the reporter gene assay using the ~2.5-kb 5'-flanking region of the ClC-K1 gene reflected the level of endogenous expression of ClC-K1 gene in each cell line, thus indicating that the ~2.5-kb 5'-flanking region of the ClC-K1 gene contained cis-acting regulatory elements for cell-specific expression of the ClC-K1 gene.

View larger version (40K): [in this window] [in a new window]   Fig. 3.   Ribonuclease protection assay for mouse ClC-K1 expression in mouse kidney and cultured cell lines; 10 µg of total RNAs were hybridized with cRNA probe (333 bp indicated by arrow only) for mouse ClC-K1. A 294-bp protected band, indicated by the arrow and asterisk, was present in the mouse kidney and IM cells but not in NIH-3T3 cells.

View larger version (10K): [in this window] [in a new window]   Fig. 4.   Promoter activities of the 2.5-kb 5'-flanking region of rat ClC-K1 gene in various cell lines. pLuc-2500 (20 µg) was cotransfected with pSV--galactocidase vector (5 µg) into various cell lines by the electroporation method. Promoter activity measured as luciferase activity was corrected by -galactosidase activity and expressed as % of pGL2 control, a positive control that is driven by simian virus 40 early promoter and enhancer. Data are expressed as means + SD (n = 5).

Deletion analysis of the 5'-flanking region of rat ClC-K1 gene promoter and identification of elements involved in cell-specific transcription. To identify the regulatory elements required for ClC-K1 gene promoter in IM cells, deletion analysis was performed. The plasmid constructs containing nested deletions of the 5'-flanking region were introduced into IM cells. As shown in Fig. 5, deletion from 2.5 to 128 kb resulted in a 1.3-fold increase of luciferase activity in IM cells. Further deletion to 51 did not affect luciferase activity in IM cells, but further deletion to 29 resulted in an ~90% reduction of luciferase activity. The promoter activity of the 29 construct in IM cells (3.6 ± 1.8% of pGL2b control, mean ± SD, n = 5) was significantly higher (P < 0.01) than the value for pGL2 basic (0.1 ± 0.02% of pGL2 control). Further deletion to +15 led to a complete loss of the minimal promoter activity. These results indicated that the 29-bp 5'-flanking region (from 0 to 29) had minimal promoter activity and that the element from 29 to 51 was required for the maximal activity of ClC-K1 promoter in IM cells. This 22-bp region was characterized by purine-rich sequence GGGGAGGGGGAGGGGAG.

View larger version (15K): [in this window] [in a new window]   Fig. 5.   Deletion analysis of ClC-K1 promoter in IM cells. Various 5'-flanking regions of rat ClC-K1 gene were ligated to luciferase reporter gene and transfected into IM cells with pSV--galactocidase vector (5 µg). Promoter activity measured as luciferase activity was corrected by -galactosidase activity and expressed as % of pGL2 control. Data are expressed as means + SD (n = 3).

To determine whether the IM cell-specific promoter activity in pLuc-2500 resulted from the 22-bp purine-rich element, the promoter activities of pLuc-29 and pLuc-51 in various cell lines were measured. As shown in Table 1, luciferase activities of pLuc-29 in various cell lines were low but significantly (P < 0.01) higher than those of pGL2 basic and pLuc+15, suggesting that the minimal promoter activity in pLuc-29 was present even in ClC-K1 nonexpressing cell lines. In contrast, the promoter activities of pLuc-51 in various cell lines showed a pattern of cell specificity similar to that observed in pLuc-2500. Although the promoter activity of pLuc-51 in each cell line was significantly (P < 0.01) higher than that of pLuc-29, the magnitude of increase was largest (~10-fold increase) in the IM cells. These results suggest that the region from 29 to 51 has a major role in the IM cell-specific promoter activity of the ClC-K1 gene.

Electrophoretic gel shift assay. To identify protein(s) binding to the 22-bp purine-rich element, electrophoretic mobility shift assay was performed using nuclear extracts from IM cells and NIH-3T3 cells. Figure 6 shows that nuclear extract from IM cells gave rise to four protein-DNA complexes (lane 2). All four bands were competed with a 100-fold excess of unlabeled double-strand probe (Fig. 6, lane 3, competitor 1). However, two lower bands could also be competed with 100-fold excess of the antisense single-strand DNA used for preparation of the double-strand DNA probe (lane 6, competitor 4), thus suggesting that both complexes were single-strand DNA-binding proteins. Furthermore, although incubation with a 100-fold excess of cold mutant probe (competitor 2) where four Gs in the wild-type probe were mutated to T (GGGGAGGGGGAGGGGAGGG to GAGGGGG) had no effect on the formation of the upper band, it did compete with the wild-type probe to form the lower band (Fig. 6). These results indicated that the protein(s) forming the upper band with the probe could be important for the ClC-K1 promoter. To confirm this, reporter gene assay was performed to measure enhancer activity of G to T mutant promoter, which could not form the upper band. As shown in Fig. 7, the mutant construct did not increase the minimal promoter activity (2.5 ± 0.5-fold increase, n = 3) as much as the wild-type construct (10 ± 1.5-fold, n = 3), confirming that the protein(s) forming the upper band was the most important for the promoter activity of the ClC-K1 gene. Subsequently, we tested the role of several G residues in forming the upper complex with nuclear extract from IM cells. Each mutant double-strand probe was end labeled, and gel mobility shift assay was performed. As shown in Fig. 8, the single mutation of G to T either at position 1 (lane 2), 2 (lane 3), or 3 (lane 4) abolished the formation of the upper band as well as the mutations at all four sites (lane 1). On the other hand, the mutation at position 4 (lane 5) had no effect on the formation of the upper band. Taken together, these results suggest that the sequence GGAGGGGGAGG between sites 1 and 3 may at least have some involvement in the formation of the DNA-protein complex.

View larger version (40K): [in this window] [in a new window]   Fig. 6.   Electrophoretic mobility shift assay detected a protein binding to the purine-rich sequence of ClC-K1 promoter in the nuclear extract from IM cells. Nuclear extract (3 µg) or bovine serum albumin (3 µg) was incubated with 32P-labeled purine-rich DNA probe in the presence or absence of various competitors. The competitor was 100 M excess of an unlabeled probe in lane 1, 100 M excess of a mutant probe (AGCCGAGGGGGTGTTG was converted into AGCCGAGGGGGTGTTG) in lane 2, 100 M excess of an unrelated probe in lane 3, and 100 M excess of single-strand (antisense strand) oligonucleotide in lane 4. Band marked with an arrow and asterisk is a specific DNA-protein complex. Band marked with an arrow only was competed with the mutant probe.

View larger version (9K): [in this window] [in a new window]   Fig. 7.   Effect of mutations in the purine-rich motif on its enhancer activity. To construct the pLuc-51 mutant, the mutations shown in Fig. 6 were introduced into pLuc-51. Promoter activities of pLuc-51 and pLuc-51 mutant in the IM cells were compared with the promoter activity of pLuc-29 and expressed as n-fold increases (means ± SD, n = 3).

View larger version (66K): [in this window] [in a new window]   Fig. 8.   Effects of various mutations in the purine-rich motif on the formation of the specific DNA-protein complex. Mutant probes are as follows: lane 1, AGCCGAGGGGGTGTTG; lane 2, AGCCGAGGGGGAGGGGAGGGTGTTG lane 3, AGCCGGGGAGGAGGGGAGGGTGTTG; lane 4, AGCCGGGGAGGGGGGGAGGGTGTTG; and lane 5, AGCCGGGGAGGGGGAGGGGGTGTTG. Nucleotides underlined are mutations introduced. Each probe was labeled with [-32P]ATP and incubated with 3 µg of nuclear extract from IM cells. Specific DNA-protein complex identified in Fig. 6 (bold arrow and asterisk) was only detected with probe 5.

Finally, we investigated whether or not the protein binding to the purine-rich element was specifically present in ClC-K1-expressing IM cells. As shown in Fig. 9, although nuclear extract from the NIH-3T3 cells could also form the upper band, the intensity of this band was much fainter than that formed by the nuclear extract from the IM cells.

View larger version (74K): [in this window] [in a new window]   Fig. 9.   Specific protein binding to the purine-rich motif was much more abundant in IM cells than in NIH-3T3 cells. Equal amounts (3 µg) of nuclear extract from IM and NIH-3T3 cells were incubated with the purine-rich probe. In lanes indicated by competitor (+), 100 M excess of an unlabeled probe was included in the reaction. Band indicated by an arrow represents a specific DNA-protein complex.

	DISCUSSION

Top Abstract Introduction Methods Results Discussion References

In this report, we describe the isolation and transcriptional analysis of the rat kidney-specific ClC-K1 chloride channel gene promoter. The 5'-flanking region of the rat ClC-K1 gene does not contain a TATA box, CAAT box, or any SpI sites. Promoters lacking a TATA box were first found in housekeeping genes and have more recently been found in various gene promoters involved in cell-specific transcription (16, 28). Among the cloned promoters of other kidney-specific genes, aquaporin-2 water channel and Na-K-Cl cotransporter promoter have a TATA box, but V₂ vasopressin receptor promoter lacks one. There were no extended regions of sequence similarity among these kidney-specific promoters, including ClC-K1 gene promoter. This suggests that each type of kidney cell is more likely to possess its own mechanism(s) of cell-specific expression rather than a common mechanism that governs all kidney-specific genes. In a promoter lacking a TATA box, GC-rich elements usually play an important role in transcriptional activation. In the ClC-K1 promoter, however, Sp1 sites and other GC-rich elements are not present. In contrast, several GA-rich clusters are found. Other recognition sequences of transcription factors found in the ClC-K1 promoter are AP-3, AP-2, and binding sites for heat shock factor and basic helix-loop-helix proteins (Fig. 1B). Because ClC-K1 expression is increased in dehydration in vivo (32), cAMP signal from vasopressin via V₂ receptor was speculated to enhance ClC-K1 gene transcription. Although there is no cAMP responsive element, several AP-2 sites known to regulate gene transcription in response to phorbol ester or cAMP are present in the ClC-K1 promoter (Fig. 1B). However, when we performed tests to determine the effects of these two agents on the ClC-K1 promoter activity in reporter gene assay, we could observe any (data not shown). The roles of other putative binding sites for transcription factors in the ClC-K1 promoter activity have not been examined.

The transcriptional start site of the rat ClC-K1 gene was determined by three independent methods in this study. Sequencing of several clones obtained by 5'-RACE revealed that some of them contained more 5'-end sequences of the gene compared with the initially isolated rat ClC-K1 cDNA (32). Furthermore, these sequences were found in the genomic clone isolated in this study just upstream of the 5'-end of the cDNA sequence, clearly identifying the genomic clone that we isolated as the ClC-K1 gene. The existence of another splice variant of ClC-K1 mRNA was also revealed by 5'-RACE. As shown in Figs. 1A and 2A, there is an 85-bp exon 2 that was not present in the original cDNA. Although primer extension analysis also confirmed the existence of this variant in the rat kidney, the physiological significance of this variant is not clear at present. The RACE clone that had the most 5'-end portion of the ClC-K1 gene had 20 more base pairs than the original cDNA. However, the primer extension experiment suggested that the real transcription start site was present 21 bp further upstream. Accordingly, we performed RNase protection assay to determine the exact transcriptional start site. Although several protected fragments were observed, the longest fragment was 64 bp, confirming that the transcriptional start site was 41 bp upstream of the 5'-end of the initial cDNA. The results of reporter gene assay were consistent with the result of the transcriptional start site. Although the construct (+15 to +75) not containing the transcriptional start site showed no significant promoter activity over pGL2 basic, the construct (29 to +75) containing the transcriptional start site showed minimal promoter activity, thus supporting the existence of the transcriptional start site in this region. RNase protection assay (Fig. 2D) also confirmed the proper initiation of transcription in reporter gene assay.

A 29-bp region from the transcription start site contains AP-3 consensus sequence (GTGGWWWG). AP-3 was shown to be an important element for SV40 promoter (17), but its binding protein has yet to be identified. As shown in Table 1 and Fig. 5, this region had minimal non-cell-specific promoter activity of the ClC-K1 gene. In contrast, the addition of purine-rich sequence to this minimal promoter enhanced the promoter activity in a cell-specific manner. As shown in Table 1, the minimal promoter activity was increased about two- to threefold by the purine-rich sequence in ClC-K1-nonexpressing cell lines. In contrast, in ClC-K1-expressing IM cells, the purine-rich sequence increased the minimal promoter activity ~10-fold. This enhancer activity was accompanied by the binding of specific protein to this element. As shown in Fig. 9, the specific protein binding to the purine-rich sequence was present much more abundantly in the IM cells than in the NIH-3T3 cells. Furthermore, the correlation between the enhancer activity and the binding of a specific protein was clearly demonstrated in Figs. 7 and 8 in which the mutant enhancer, which did not form the specific DNA-protein complex, did not show enhancer activity in IM cells. Thus the purine-rich motif and its binding protein have an essential role in the IM cell-specific activity of the ClC-K1 gene promoter. Because the level of endogenous expression of the ClC-K1 gene in the IM cells was much lower than that in vivo, the physiological relevance of these results to kidney-specific expression of ClC-K1 gene must be verified by further experiments utilizing transgenic mice. However, the cell-specific enhancer activity of purine-rich element observed only in the ClC-K1-expressing cell line could possibly be one of the important determinants of nephron-specific expression of the ClC-K1 gene in vivo.

Purine-rich sequences were found in the 5'-flanking region of several genes, especially in the neuron-specific gene promoters. In the promoter of -aminobutyric acid A receptor -subunit gene, the conserved purine-rich sequences were present, and the brain-specific protein (BSF1) binding to the AGAGGAGAGGGGAGAGGGGGGAG element was reported to be involved in the neuron-specific expression of the gene (19). In human neurofilament heavy gene promoter, the GGGAGGAGG element was identified to be important in brain-specific enhancement of transcription (25). Similar purine-rich elements were also observed in synapsin I gene promoter (AGAGAGGGGGAGGGGAAA; see Ref. 24) and Na-K ATPase ₃-subunit promoter (AGGGTGAAGGGGGAAGGGGGAG; see Ref. 20). The sequence found in the ClC-K1 gene promoter in this study is GGGGAGGGGGAGGGGAGGG, which is similar but not identical to these brain-specific enhancer elements. It is highly likely that a family of purine-rich DNA-binding proteins is present and involved in the tissue-specific expression of genes. It is also possible that these proteins are related to known transcription factors such as ETS family proteins and C₂H₂ zinc finger genes. ETS proteins recognize various sequences with the GGAA core sequence (36), and the expression of most of the known ETS proteins is cell specific (36). Some C₂H₂ zinc finger genes, for example, c-Krox and myeloid zinc finger gene 1 (MZF1), were shown to bind to GA-rich sequences (7,18). c-Krox was found to bind the GGGAGGG motif in the type I collagen gene promoter (7), and its predominant expression in skin raises the possibility that c-Krox is involved in the control of type I collagen gene expression in skin. MZF1, a transcriptional fator shown to bind AGTGGGGA and CGGGnGAGGGGGAA sequences (18), plays a key role in regulating hematopoiesis (18). It is of particular interest that a similar transcription factor, Kid-1, was isolated from rat kidney (35). Kid-1 contains 13 zinc fingers and posesses 45% homology within zinc finger domains with MZF1. Expression of mRNA for both Kid-1 and MZF1 is developmentally regulated and restricted to specific tissues. Thus a zinc finger gene similar to Kid-1 and MZF1 may bind to the purine-rich element in the ClC-K1 promoter. Isolation of a protein binding to this element will elucidate the mechanisms of kidney-specific expression of the ClC-K1 gene. Although the purine-rich sequence (GGGGAGGGGGAGGGGAGGG) determined in the ClC-K1 gene promoter is not exactly the same as purine-rich sequences found in the Na-K-Cl cotransporter, V₂ vasopressin receptor promoter, and aquaporin-2 water channel promoter, the Na-K-Cl cotransporter gene promoter contains GA repeats, and V₂ vasopressin receptor promoter and the aquaporin-2 water channel promoter contain several purine-rich regions such as AGGGAAAAGGAGGGGGGAAG at 110 (16) and GAGAAAGAGAG at +10 (33), respectively. If these sequences are found to be important in the cell-specific expression of each gene, similar transcription factors may be involved in the kidney-specific expression.

In conclusion, we have cloned the rat ClC-K1 gene promoter and observed that the promoter exhibits ClC-K1 expressing cell-specific activity. This cell-specific activity results from the purine-rich element ~40-bp upstream of the transcription start site.