Renal tubule-specific transcription and chromosomal localization of rat thiazide-sensitive Na-Cl cotransporter gene.

The molecular mechanism underlying the renal expression localization of the thiazide-sensitive Na-Cl cotransporter (TSC) gene was studied. The TSC gene was localized to chromosome 19p12-14. In cultured cells, tissue-specific transcription activity of the 5'-flanking region of the rat rTSC gene (5'FL/rTSC) was demonstrated, and the major promoter region was located between position -580 and -141. To further examine the tissue-specific transcription, transgenic rats harboring the 5'FL/rTSC fused upstream of the LacZ gene were generated. Immunohistochemical analysis clearly showed that LacZ gene expression was co-localized to distal convoluted tubules (DCT) with TSC, indicating that the 5'FL/rTSC regulates the renal tubule-specific TSC expression. Because a transcription factor, HFH-3 (hepatocyte nuclear factor-3/folk head homologue-3), had also been localized to DCT, a possible role of the putative cis-acting element (HFH-3/rTSC, -400/-387 position) for HFH-3 binding in the tissue-specific transcription was examined. Deletion and mutation analyses suggested that transcription of the HFH-3/rTSC was actually responsive to HFH-3, and electrophoretic mobility shift assay showed a direct binding of in vitro synthesized HFH-3 to the HFH-3/rTSC. In conclusion, the rTSC gene is localized to rat chromosome 19p12--24. The transcription regulatory region of the TSC gene confers DCT-specific gene expression. DCT-specific transcription factor HFH-3 may be involved in the renal tubule-specific transcription of TSC gene.

tant molecule for reabsorption of NaCl in the kidney and a target of thiazide diuretics. TSC cDNAs were cloned, and the TSC function was characterized (1)(2)(3). Mutations possibly leading to loss-of-function in the human TSC gene have been shown to cause Gitelman's syndrome, which is characterized by dehydration, hypokalemic metabolic alkalosis, hypomagnesemia, and hypocalciuria (4,5). Recently, TSC-deficient mice were generated, and their phenotype was shown to bear a good resemblance to Gitelman's syndrome (6). Thus, TSC plays an important role not only in NaCl metabolism but in acid-base balance and in the metabolism of other electrolytes such as potassium, magnesium, and calcium.
TSC mRNA expression has been shown to be localized to the distal convoluted tubule (DCT) in the kidney by in situ hybridization histochemistry (7,8) as well as reverse transcription and polymerase chain reaction (RT-PCR) with microdissected nephron segments (9). Immunohistochemistry using a specific antibody against TSC has also shown that immunoreactive TSC is localized to DCT (10,11). However, no study to uncover the molecular mechanism underlying the renal tubule-specific TSC expression has yet been performed in any species.
Recently, a transcription factor, hepatocyte nuclear factor-3/ folk head homologue-3 (HFH-3), was isolated and partly characterized (12). HFH-3 belongs to an HFH/winged helix transcription factor family and is identical to FREAC-6 (13). Members of the family have been shown to be involved in tissue-or cell-specific gene expression, as well as cellular differentiation during embryonic development (14,15). Interestingly, HFH-3 mRNA was reported to be localized to the epithelium of the DCT, where TSC is also localized. We therefore hypothesized that HFH-3 might be involved in the TSC gene expression.
In the present study, we studied the tissue-specific expression of the TSC gene in terms of gene transcription by an in vivo as well as in vitro experiments. The structure of the 5Ј-flanking region of the rat TSC gene (5ЈFL/rTSC) was revealed, and the rTSC gene was mapped to chromosome 19p12-14 by FISH. The transcription function was shown to be tissue-specific in cultured cells. Moreover, we generated transgenic rats harboring the 5ЈFL/rTSC fused upstream of the LacZ gene to examine the tissue-specific transcription of 5ЈFL/rTSC. Immunohistochemical analysis clearly demonstrated that LacZ expression was localized to DCT in the transgenic rat kidney, indicating that the 5ЈFL/rTSC regulates the DCT-specific TSC localization in vivo. Furthermore, we identified a functional HFH-3 binding site in the major promoter region of 5ЈFL/rTSC, which was actually bound by in vitro synthesized HFH-3. HFH-3 may be involved at least in part in the DCTspecific expression of TSC. * This study was supported in part by grants-in-aid from the Ministry of Education, Science, and Culture, Japan and by the Takeda Foundation for Metabolic Disorder Research, Japan. 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 U.S.C. Section 1734 solely to indicate this fact.
The nucleotide sequence(s) reported in this paper has been submitted to the GenBank TM /EBI Data Bank with accession number(s) AB024534.

MATERIALS AND METHODS
Southern Blot Analysis-To use as a probe in Southern blot analysis of rTSC gene, an rTSC cDNA fragment was first cloned by a PCR-based method with a rat kidney cDNA library (rat kidney Marathon-ready cDNA, CLONTECH) as a template using a pair of primers (sense, 5Ј-CCC GGA GCC ATA ATG GCA GAG CTA CCC G-3Ј (from position 1 to 28 in previously reported rTSC cDNA (2)); antisense, 5Ј-CTT CTC CAG CTG GAG AGC GTG AGT TCC G-3Ј (from position 3,070 to 3,097 in rTSC transcript)) under the following PCR condition: 95°C, 30 s and 70°C, 5 min for 25 cycles. The PCR product was then subcloned into pCR-TOPO (Invitrogen) by a TA cloning method. Sequencing was performed to confirm the complete sequence identity with the previously reported rTSC cDNA (2). The BalI digested fragment of the rTSC cDNA (892 bp) was labeled with [␣-32 P]dCTP by a random primer method (BcaBEST labeling kit, TaKaRa) and used as a probe. Rat genomic DNA was extracted from liver as described previously (16). Genomic DNA was then digested with restriction enzymes BamHI, EcoRI, or HindIII. Digested DNA was then resolved on a 0.9% agarose gel and transferred to a nylon membrane (HyBond N ϩ , Amersham Pharmacia Biotech). Hybridization was performed in Rapid-Hyb Buffer (Amersham Pharmacia Biotech) at 65°C for 2 h with a radiolabeled probe (25 ng). The membrane was then incubated with 2ϫ SSC, 0.1% SDS at room temperature for 20 min, 0.1ϫ SSC, 0.1% SDS at 65°C for 15 min, and 0.1ϫ SSC, 0.1% SDS at 65°C for 15 min. The blot was exposed to an x-ray film (Kodak) for 72 h.
Cloning of the 5Ј-Flanking Region of rTSC Gene (5ЈFL/rTSC)-The 5ЈFL/rTSC was cloned by a PCR-based method using the Genome Walker Kit (CLONTECH) for rat gene cloning. Briefly, the first PCR was carried out based on the manufacturer's instruction with the genespecific primer (GSP)-1, 5Ј-ATG GGT CAA ATG GCT GGG CTG GCT ATT G-3Ј (designed to hybridize with the region from position 120 to 147 in rTSC cDNA (2)), and the adapter primer (AP)-1, 5Ј-GTA ATA CGA CTC ACT ATA GGG C-3Ј (supplied in the kit). Nested PCR was then performed with the GSP-2, 5Ј-TGC ACA GAG CAT CGC CTG GCA TCT CTG T-3Ј (designed to anneal to the region from position 31 to 58 in rTSC cDNA (2)), and the AP-2, 5Ј-ACT ATA GGG CAC GCG TGG T-3Ј, using the first PCR product as a template. The resultant PCR product was subcloned into pCR-TOPO (Invitrogen), and sequenced in both directions. Putative transcription factor binding sites in 5ЈFL/ rTSC were identified by the TRANSFAC 3.4 data base using Transcription Element Search Software (TESS) (18).
Primer Extension Method-Primer extension was performed using the primer extension system-AMV reverse transcriptase (Promega) by a modification of the previously reported method (19). Total RNA was extracted from rat renal cortex by the guanidinium thiocyanate-cesium chloride centrifugation method (16). An oligonucleotide (5Ј-ACT GCA CAG AGC ATC GCC TGG CAT CTC TGT CAC GGG TAG C-3Ј) designed to anneal to the region from position 21 to 60 in rTSC cDNA (2) was labeled with IRD700 (Aloka) at the 5Ј-end and used as a primer. Annealing was carried out with 1 pmol of the primer and 25 g of total RNA at 50°C for 8 h, and the extension reaction was performed at 42°C for 2 h. The extended product was resolved on a 0.8% polyacrylamide gel. To determine the length of the DNA product, the sequence ladder of pBluescript SK(ϩ) (Invitrogen) was also resolved in the same gel. The fluorescent DNA was analyzed by automated DNA analyzer (LI-COR dNA analyzer 4200).
Rapid Amplification of the 5Ј-end of cDNA (5Ј-RACE)-5Ј-RACE was performed with rat kidney Marathon-ready cDNA (CLONTECH). PCR was carried out according to the manufacturer's instruction with the adapter primer, 5Ј-GCC ACA CCC ACG GCA TTG GCA AAG GCG A-3Ј (supplied in the kit), and an rTSC-specific primer, 5Ј-CCA TCC TAA TAC GAC TCA CTA TAG GGC-3Ј (from position 693 to 720 in rTSC transcript), under the following conditions: 95°C, 30 s; 68°C, 2 min, for 30 cycles. The PCR product was then subcloned into pCR-TOPO (Invitrogen) by a TA cloning method. The 5Ј-end of the rTSC cDNA was determined by sequencing.
Cell Culture-HEK293 (a human embryonic kidney epitheloid cell line), A10 (a rat vascular smooth muscle cell line), and HepG2 (a human hepatocyte cell line) cells were cultured in Dulbecco's modified Eagle's medium containing 10% fetal bovine serum.
Electrophoretic Mobility Shift Assay (EMSA)-An expression construct to synthesize a fusion protein of the HFH-3 DNA binding domain and glutathione S-transferase (GST) (12) was a generous gift from Dr. Robert H. Costa (University of Illinois at Chicago). GST-HFH-3 fusion protein was isolated from Escherichia coli cultures and purified by glutathione-Sepharose 4B (Amersham Pharmacia Biotech) according to the manufacturer's instruction. The purity of GST-HFH-3 was confirmed by SDS-polyacrylamide gel electrophoresis followed by Coomassie staining (data not shown). The double strand DNA of HFH-3/rTSC was radiolabeled and used as a probe. EMSA was performed with the probe and in vitro synthesized GST-HFH-3 as described previously (23,24). A double strand DNA containing the consensus HFH-3 binding was created by annealing with oligonucleotides 5Ј-AGC TGC ACG TTC GTT GTT TAT GTA CCG AGC G-3Ј (sense) and 5Ј-TCG ACG CTC GGT ACA TAA ACA ACG AAC GTG C-3Ј (antisense), in which underlining indicates the consensus HFH-3 binding sequence, and was used as a competitor. mHFH-3/rTSC oligonucleotides (containing the mutated HFH-3 binding site) were also used as competitor.
Generation of Transgenic Rats Bearing 5ЈFL/rTSC-LacZ-The transgene used herein consisted of the 2,110-bp rTSC promoter fused to the LacZ gene. A fragment encoding the rTSC promoter region (Ϫ2,110 to ϩ45) was excised from rTSC/Luc with KpnI and subcloned to the KpnI site of pSV-␤-galactosidase control vector (Promega). The KpnI site of this plasmid was located 87-bp upstream of the LacZ coding region, and the frame for the intact LacZ gene expression was confirmed by sequencing. The transgene fragment excised by digestion with SalI was purified using the QIAquick gel extraction kit (Qiagen).
Transgenic rats were generated using pronuclear microinjection as the standard procedure. The microinjection was performed by YS New Technology Laboratory (Tochigi, Japan). Briefly, the transgene fragment was microinjected to the pronuclei of fertilized single-cell oocytes obtained from Harlan Sprague-Dawley rats. Embryos that survived microinjection were transferred into the oviduct of pseudo-pregnant Wistar rats. Transgenic rats were selected with amplification of LacZ by PCR from genomic DNA extracted from tails. The number of integrated copies of the transgene was estimated by Southern blotting with the transgene fragment as a probe.
Immunohistochemistry-The transgenic and nontransgenic rats (16week) were anesthetized with diethyl ether, and perfused with saline through cannulae inserted into the left ventricles. The kidneys were removed and fixed for 24 h with 20% formaldehyde and mounted onto paraffin blocks. Rabbit anti-␤-galactosidase polyclonal antibody (Biogenesis) was used at a dilution of 1:200. Rabbit anti-TSC antibody, a gift from Dr. Steven C. Hebert (11), was used at a dilution of 1:250. Staining was performed by the avidin-biotinylated peroxidase complex method using an RTU Vectastain universal avidin-biotin complex kit (ABC, Vector) and a VIP substrate kit (Vector). Sections were counter-stained with hematoxylin.
Data Analysis-p Ͻ 0.05 by Student's t test was considered statistically significant.

Southern Blot Analysis and Chromosome Mapping of rTSC
Gene-A single band was observed in either the BamHI or HindIII digest of rat genomic DNA (ϳ6.5 and ϳ4.2 kilobase pairs in length, respectively), suggesting that rTSC is a single copy gene (Fig. 1). Additionally, two bands in the EcoRI digest suggested the presence of EcoRI site in the intron between TSC exons complementary to the probe. To determine the chromosomal localization of rTSC gene, we next performed FISH using three different DNA fragments of rTSC gene as probes. After hybridization, one hundred metaphases were analyzed by recording the number and position of fluorescent spots on chromosomes. Twin fluorescent spots were identified in eighty three metaphases at chromosome 19, and single fluorescent spot at the same chromosome in twenty metaphases. A typical metaphase showing the presence of twin fluorescent spots at chromosome 19 is shown in Fig. 2A. The position of fluorescent spots were shown in Fig. 2B, and these spots were clustered at 19p12-14 (p Ͻ 0.01).
Structure of 5ЈFL/rTSC-A 2.1-kilobase pair fragment of 5ЈFL/rTSC was isolated and sequenced (DDBJ/EMBL/Gen-Bank TM data bases with accession number AB024534). 5ЈFL/ rTSC (Fig. 3)  Transcription Initiation Site of rTSC Gene-The transcription initiation site of the rTSC gene was determined by both the primer extension method and 5Ј-RACE. Total RNA extracted from rat renal cortex was reverse transcribed using the 5Ј-end labeled DNA primer (designed to be a complement to the region from position 21 to 60 of rTSC cDNA). As shown in Fig.  4A, the longest extended product had 67 bases in length, indicating that the transcription initiation site of the rTSC gene is located 8 bases upstream of the 5Ј-end of the cloned cDNA (2) (Fig. 4C). Moreover, we performed 5Ј-RACE to confirm further the transcription initiation site. A PCR product 720 bp in length was subcloned, and the 5Ј-end was determined by sequencing. As shown in Fig. 4B, the 5Ј-end was completely identical to the transcription initiation site determined by the primer extension method (Fig. 4C).
Transcription Activity of 5ЈFL/rTSC-Luciferase reporter gene assay was performed to examine the transcription activity of 5ЈFL/rTSC. HEK293 cells (in which endogenous TSC mRNA expression was detected) were transfected with either luciferase expression vector alone (control) or rTSC/Luc. Luciferase expression with rTSC/Luc was significantly higher than with control ( Fig. 5A; 26.2 Ϯ 1.8-fold compared with control) in transfected HEK293 cells, suggesting that the 5ЈFL/rTSC is transcriptionally active. We next conducted the deletion analysis of 5ЈFL/rTSC. As shown in Fig. 5B, transfection of the luciferase expression vector containing the full-length (ϳ2.1 kilobase pairs) 5ЈFL/rTSC (-2,093 rTSC/Luc) showed the highest luciferase expression, and the transcription activity of 5ЈFL/ rTSC was dependent on the fragment length. Luciferase activity was high in cells transfected with the rTSC/Luc containing deletion fragments from position Ϫ2,093 to Ϫ580. However, further deletion of the promoter to position Ϫ141 resulted in a marked decrease in luciferase activity.
Tissue-specific Transcription of rTSC Gene-Endogenous expression of TSC mRNA was examined in HEK293, HepG2, or A10 cells by the RT-PCR method. An expected length of PCR product (609 bp) was detected with the total RNA from HEK293 cells (Fig. 6A), and its sequence identity to rTSC was confirmed, whereas no PCR products for rTSC were detected in the other cells. As a positive control, GAPDH mRNA was detected in either HEK293, HepG2, or A10 cells by RT-PCR (Fig.  6A). The PCR product of rTSC mRNA was also detected with the total RNA from renal cortex by the same RT-PCR protocol. TSC was thus indicated to be expressed in HEK293 cells but not in either HepG2 or A10 cells. We therefore used these cell lines to examine the tissue-specific transcription activity of 5ЈFL/rTSC. As shown in Fig. 6B, the transcription activity of 5ЈFL/rTSC estimated by luciferase expression was pronounced in HEK293 cells, whereas it was modest in both HepG2 and A10 cells. It was thus suggested that the 5ЈFL/rTSC is able to regulate the tissue-specific gene transcription.
Establishment of Transgenic Rats Bearing 5ЈFL/TSC-LacZ-227 microinjected oocytes were transplanted into eight Wistar rats. 58 rats were born, seven of which had the transgene. The copy number of the integrated transgene was estimated as three to six copies by Southern blotting. Every transgenic rat grew normally in appearance. The transgene was inherited in the Mendelian fashion.
Expression Co-localization of LacZ with TSC-The transgenic rat harboring six copies of the transgene was subjected to immunohistochemistry. In the cortical region of the transgenic rat kidney, immunoreactive ␤-galactosidase was observed at cortical distal tubules (Fig. 7A), whereas no stain was observed in the control rat (Fig. 7B). Using the sequential sections from the transgenic rat kidney, we observed co-localization of immunoreactive ␤-galactosidase (arrows, Fig. 7, C and E) and TSC (arrowheads, panels D and F). No immunostaining was observed in the medullar region (Fig. 7G).
A Role of HFH-3 on Transcription Activity of 5ЈFL/rTSC-To examine the effect of HFH-3 overexpression on the transcription activity of 5ЈFL/rTSC, hHFH-3 cDNA was isolated, and an expression vector carrying the cDNA (pcHFH-3) was synthesized. In vitro translated HFH-3 with this construct was resolved in SDS-polyacrylamide gel electrophoresis, and we identified a new band with the expected size corresponding to HFH-3 protein (data not shown). As shown in Fig. 8, luciferase expression in HepG2 cells transfected with Ϫ2,093 rTSC/ Luc was significantly higher by cotransfection with pcHFH-3 than with mock (2.28 Ϯ 0.14-fold expression compared with mock; p Ͻ 0.01). This enhancement of transcription with pcHFH-3 was also observed in transfection with Ϫ1,139 rTSC/Luc and Ϫ580 rTSC/Luc, whereas it was not observed with Ϫ141 rTSC/Luc. Thus, the HFH-3-responsive element was suggested to be present between position Ϫ580 and Ϫ141, and we identified an element homologous to HFH-3 binding consensus sequence (DBD TRT TTR YDT D) at position Ϫ393 (TCC TTT TTG TTA TA).
To evaluate the transcription responsiveness of HFH-3 to putative HFH-3 binding site, we mutated this site and examined the effect of HFH-3 overexpression on its transcription activity. In HEK293 cells expressing HFH-3 mRNA (data not shown), transcription activity of Ϫ580 (mHFH-3) rTSC/Luc was reduced markedly (Fig. 9, open column) compared with that of Ϫ580 rTSC/Luc (Fig. 9, hatched column).
Direct Binding of HFH-3 to HFH-3/rTSC in EMSA-EMSA was performed to examine a direct binding between HFH-3 and HFH-3/rTSC. As shown in Fig. 11, in vitro synthesized HFH-3 protein reacted with the radiolabeled HFH-3/rTSC and formed a protein-DNA complex (Lane 1). Using unlabeled HFH-3/rTSC as a cold competitor, the formation of protein-DNA complex was significantly inhibited (lane 2). Moreover, the complex disappeared with an excess of oligonucleotides containing the HFH-3 binding consensus sequence as a cold competitor (lane 3). However, the protein-DNA complex was kept unchanged with an excess of oligonucleotides of mHFH-3/rTSC as a competitor (lane 4). These results indicate that HFH-3/rTSC is actually bound by HFH-3 and the binding is sequence-specific. DISCUSSION The rTSC gene was shown to be localized to 19p12-14 by FISH. In either human or mouse, the TSC gene has also been mapped to chromosome 16q13 (3,4,25) or 8 (26), respectively. These results suggest the segmental chromosome similarity in their regions.
Genes of the Na-K-Cl cotransporter (NKCC2) (2), sodium phosphate cotransporter (NPT2) (27), chloride channel CLC-K1 (28), and kidney-specific cadherin (Ksp-cadherin) (29) have been shown to be expressed specifically in the kidney. The transcription regulatory regions of those genes have already been cloned (30 -32), and the tissue-specific transcription mechanisms have been partly examined. Particularly, the kidney-specific CLC-K1 gene transcription has recently been shown to depend on an interaction between a transcription factor, MAZ, and a transcription repressor, KKLF (33). In the present study, we analyzed the transcriptional function of the rTSC gene in terms of tissue-specific gene expression.
The 5ЈFL/rTSC contains two putative glucocorticoid-responsive elements (Fig. 3). It has been reported that glucocorticoids and mineralocorticoids stimulate the thiazide-sensitive sodium transport at DCT, inducing an increase in the number of thiazide diuretic binding sites (34,35) or in TSC protein expression (36). The glucocorticoid-responsive elements may possibly be involved in the steroid-induced TSC expression at the gene transcription level. A putative cAMP-responsive element was also identified, implying that TSC transcription could be regulated by intracellular cAMP formation by an agent such as calcitonin, which has been shown to increase the number of renal thiazide diuretic binding sites (37). Moreover, a couple of putative C/EBP␤ binding sites were identified. Because C/EBP␤ is expressed in kidney and induced by interleukin-6 mediating an inflammatory response (38), TSC gene transcription may possibly be affected by inflammation via interleukin-6.
Luciferase reporter gene analysis was performed with the chimeric reporter expression vector containing 5ЈFL/rTSC (rTSC/Luc). When rTSC/Luc was transfected into HEK293 cells (which express TSC mRNA), marked luciferase expression was observed (Fig. 5A). Additionally, the transcription activity is dependent upon the length of 5ЈFL/rTSC, indicating that the 5ЈFL/rTSC has a significant transcription function. Deletion analysis also showed a marked decrease (75%) in the transcription activity between position Ϫ580 and Ϫ141 (Fig. 5B), suggesting that a major promoter is located between these positions.
Tissue-specific transcription of TSC gene was focused in the present study. As shown in Fig. 6A, transcription activity of 5ЈFL/rTSC was pronounced in HEK293 cells expressing TSC mRNA, whereas it was suppressed in either HepG2 or A10 cells lacking TSC mRNA expression (Fig. 6), suggesting that the 5ЈFL/rTSC plays a crucial role in the tissue-specific TSC gene transcription. To further study the tissue-specific transcription of TSC in kidney, we generated the transgenic rats with the 5ЈFL/rTSC fused upstream of the LacZ gene as a transgene. As shown in Fig. 7, immunoreactive ␤-galactosidase was present at the same tubular region as TSC, probably at DCT in the renal cortical region, and neither was detected in the other renal tissues. Thus, this in vivo gene transcription model has clearly demonstrated that DCT-specific TSC gene expression is dependent on the 5ЈFL/rTSC, and it is suggested that the region contains some important elements responsible for the DCT-specific transcription.
HFH-3 belongs to the HFH/winged helix transcription factor family (12), and its family members are involved in tissuespecific gene expression and cell differentiation during embryonic development (14,15). HFH-3 mRNA is localized to DCT in kidney (12), but the target gene of HFH-3 has not yet been identified. We hypothesized that HFH-3 might be involved in the renal tubule-specific gene transcription of TSC, because a putative HFH-3 binding site was identified in the 5ЈFL/rTSC. We first examined the effect of HFH-3 overexpression on the transcription activity of 5ЈFL/rTSC in the transfected HEK293 cells. As shown in Fig. 7, the stimulation was significantly observed with Ϫ2,093, Ϫ1,139, and Ϫ580 rTSC/Luc but not with Ϫ141 rTSC/Luc. Consistent with this result, there is a putative HFH-3 binding site (HFH-3/rTSC) at the Ϫ393 position (Fig. 3), 5Ј-TCC TTT TTG TTA T-3Ј, which bears a homology to the previously reported HFH-3 binding sequence (DBD TRT TTR YDT D) (12) by 10 of 13 nucleotides (77%). To examine the transcription activity of HFH-3/rTSC, the HFH-3/rTSC in Ϫ580 rTSC/Luc was mutated, and its transcription function was examined in HEK293 cells (expressing both TSC and HFH-3 mRNAs). By the mutation, we observed a marked loss of transcription function. It is therefore suggested that the HFH-3/rTSC plays an important role in TSC gene transcription in HEK293 cells. We next focused on the transcriptional function of HFH-3/rTSC by examining a functional interaction between HFH-3 and HFH-3/rTSC in HepG2 cells. HFH-3 overexpression activated the transcription of HFH-3/rTSC-tk-Luc in a dose-dependent manner (Fig. 8), whereas the transcription stimulation was not observed in transfected cells with the mHFH-3/rTSC. These results indicate that HFH-3 can stimu- In the cortical region of transgenic rat kidney, immunoreactive ␤-galactosidase is observed at distal tubules (A), whereas no stains are observed in a control rat (B). In the sequential sections from the transgenic rat, co-localization of immunoreactive ␤-galactosidase (arrows, C and E) and TSC (arrowheads, D and F) is observed at the cortical distal tubules. In the medullar region, no immunoreactive ␤-galactosidase (G) is detected.
late the transcription through an interaction with HFH-3/ rTSC. However, the magnitude of the transcription stimulation (ϳ2-3-fold that of mock) was not comparable with the transcription activity of Ϫ2,029 rTSC/Luc in HEK293 cells. Some other elements in the 5ЈFL/rTSC may also be involved in the gene transcription to enhance the HFH-3 transactivation.
EMSA was performed to examine an interaction between HFH-3 and HFH-3/rTSC. In EMSA, a protein-DNA complex was formed with the radiolabeled HFH-3/rTSC and in vitro synthesized HFH-3. The complex formation was inhibited with an excess of the HFH-3 consensus sequence oligonucleotides but not with an excess of mHFH-3/rTSC. HFH-3 therefore can actually bind the HFH-3/rTSC and may lead to transactivation of TSC gene. Putative target genes for HFH-3 have been suggested to be the mineralocorticoid receptor, Na/H exchanger NHE3, Na/K-ATPase ␣2 subunit, and Na/K-ATPase ␤2 subunit (12), although any interaction with HFH-3 has not been proven even in these the transcription activity of the rTSC gene genes. We have demonstrated here that HFH-3 can stimulate through the HFH-3/rTSC via an interaction with HFH-3. This is the first report in which a target gene for HFH-3 has been identified.