INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

THE KIDNEY MEDULLA is the only tissue in mammals that is hypertonic under physiological conditions, due to the operation of the urinary concentrating mechanism. The degree of hypertonicity in the renal medulla fluctuates widely depending on the hydration status of the animal. In dehydrated mammals, osmolarity of the renal medulla reaches >1,000 mosM in humans and 3,000 mosM in rats. When exposed to hypertonicity for more than several hours, cells in the renal medulla accumulate small organic molecules such as betaine, myo-inositol, taurine, sorbitol, and glycerophosphorylcholine (6, 23). These compounds are termed "compatible osmolytes" because, in contrast to electrolytes, they do not perturb the function of macromolecules at high intracellular concentrations (33). Accumulation of compatible osmolytes lowers the intracellular concentration of electrolytes toward the isotonic level and thereby protects cells from the stress of hypertonicity (33). Prevention of the accumulation of compatible osmolytes results in necrosis in the renal medulla (29) and inhibition of growth and ultimately death in cultured cells (9, 27, 32).

Some compatible osmolytes are accumulated by specific Na⁺-coupled transporters: the Na⁺- and Cl-coupled betaine/-aminobutyric acid transporter (BGT1) for betaine (31), the Na⁺- and Cl-coupled taurine transporter (26), and the Na⁺-coupled myo-inositol transporter (SMIT) (14). Sorbitol is synthesized from glucose in a reaction catalyzed by aldose reductase (AR) (2). Transcription of all these genes is stimulated by hypertonicity (reviewed in Ref. 12), leading to increased abundance of mRNA (15, 16), increased activity of transporter or enzyme (3, 17), and, finally, accumulation of compatible osmolytes (23). Hypertonicity-induced stimulation of transcription plays the critical role in adaptation of the renal cells to hypertonicity.

We identified a regulatory sequence element named tonicity-responsive enhancer (TonE) within a 13-bp region in the promoter region of the BGT1 gene (24). TonE mediates the transcriptional stimulation in response to hypertonicity. Subsequently, TonE-like regulatory sequences have been found in the 5'-flanking region of the AR gene of several species (4, 5, 10), suggesting that a common mechanism regulates these genes. However, little is known about how mammalian cells recognize hypertonicity and how the signal is conveyed to TonE.

Studies in yeast revealed a signal transduction pathway linking the sensing of hypertonicity to the transcriptional response (reviewed in Ref. 19). Signals from two tonicity sensors in the membrane converge on Pbs2, a homolog of mitogen-activated protein (MAP) kinase kinase. When yeast cells are switched to hypertonic media, Pbs2 and its downstream kinase, Hog1, a MAP kinase homolog, are immediately activated, resulting in the stimulation of transcription of GPD1 (1) in an as yet undetermined way. GPD1 encodes glycerol-3-phosphate dehydrogenase, the rate-limiting enzyme for synthesis of glycerol which is the predominant compatible osmolyte of yeast. In mammalian cells, including those in the kidney medulla, hypertonicity stimulates three cascades of MAP kinase homologs: ERKs, JNK/SAP kinase, and p38. However, preventing activation of these cascades does not interfere with the transcriptional stimulation of BGT1 or AR (11, 13), indicating that MAP kinase homologs are not involved in signaling to TonE.

In this study, we rigorously characterize the nucleotide sequence of TonE and the TonE binding proteins (TonEBPs) that specifically interact with TonE. In vivo footprinting and other data presented here strongly support the idea that TonEBPs are transcription factors that stimulate transcription in response to hypertonicity. Characteristics of TonEBPs revealed in this study provide fundamental information for cloning TonEBPs, whose regulation will in turn provide clues to the hypertonicity signaling pathways in mammals.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Cell culture. Madin-Darby canine kidney (MDCK) cells were maintained in a defined medium (30). Cells were grown to a confluent monolayer before they were treated with hypertonicity and other agents. Medium was made hypertonic by adding 200 mM raffinose. For treatment with inhibitors (Fig. 3B), cells grown in isotonic medium were incubated with the inhibitor for 30 min, and then cells were switched to hypertonic medium containing the inhibitor for an additional 6 h.

Preparation of nuclear extracts and electrophoretic mobility shift assay. MDCK cells cultured in isotonic or hypertonic medium were chilled to 4°C, and nuclear extracts were prepared as described (24). To prepare probes for electrophoretic mobility shift assay (EMSA), single-stranded oligonucleotides shown in Figs. 1 and 2 were synthesized and purified (Genetic Core Facility, Johns Hopkins University). To obtain a double-stranded probe, 200 pmol of each complementary oligonucleotide were annealed in 100 µl containing (in mM) 150 NaCl, 10 MgCl₂, and 50 Tris hydrochloride, pH 7.9. An aliquot of nuclear extract (4 µg protein) was incubated initially for 10 min at 25°C in 20 µl containing (in mM) 20 HEPES (pH 7.9), 100 KCl, 0.1 EDTA, 10% glycerol, 1 mM dithiothreitol, 1.5 µg poly(dA-dT), and 5 mM MgCl₂. The mixture was then incubated for an additional 20 min after 10 fmol of ³²P-labeled probe was added. The reaction was electrophoresed on a 4% polyacrylamide gel (79:1, acrylamide-bisacrylamide) in a buffer containing (in mM) 45 Tris, 45 borate, and 1 EDTA. The EMSA gels were dried and radioactivity was visualized and quantitated using a PhosphorImager and the computer program ImageQuant (Molecular Dynamics, Sunnyvale, CA).

View larger version (94K): [in this window] [in a new window]   Fig. 1.   Binding of various tonicity-responsive enhancer (TonE) sequences to nuclear extracts from MDCK cells cultured in hypertonic medium. A: sequence of cTonE, rTonE, and hTonE. TonE sequence in each probe is underlined. cTonE is 67/45 region of the canine BGT1 gene plus overhanging sequences compatible with Spe I site (in lower case letters). rTonE and hTonE were made by replacing the TonE portion (62/51 or 62/52) of cTonE with TonE sequences from rabbit (5) and humans (21), respectively. B: autoradiograms of electrophoretic mobility shift assay (EMSA) gels, using 32P-labeled cTonE (left) or hTonE (right). Each lane was loaded with a binding reaction containing 4 µg of nuclear extract from hypertonic MDCK cells along with 0.5 nM probe alone or with 15 or 25 nM (left to right) unlabeled cTonE, rTonE, or hTonE, as indicated, top. Bands corresponding to the slowly migrating bands described previously (24), and free probes are marked by solid and open arrows, respectively.

View larger version (27K): [in this window] [in a new window]   View larger version (13K): [in this window] [in a new window]   Fig. 2.   Relationship between binding to the slowly migrating bands and luciferase induction. A: binding to the slowly migrating bands and induction of luciferase in response to hypertonicity by hTonE, 11 mutants of hTonE where a single base pair was mutated in each mutant as indicated, and cTonE. Binding of each double-stranded oligonucleotide to the slowly migrating bands was measured as %competition of 32P-hTonE (0.5 nM) binding in the presence of the oligonucleotide at 25 nM, using EMSA as shown in Fig. 1. Means of two determinations from a single nuclear extract prepared from MDCK cells cultured in hypertonic medium for 24 h are shown. Variability between the duplicate determinations was <15%. Two tandem copies of each oligonucleotide were inserted in front of the SV40 promoter and a luciferase reporter gene using a commercial vector, pGL2-promoter (Promega). Luciferase activity was measured in MDCK cells transfected with each reporter construct and then cultured in hypertonic and isotonic medium. Multiple-fold induction of luciferase by hypertonicity (activity in cells cultured in hypertonic medium/activity in cells cultured in isotonic medium) is presented as means ± SE (n = 3-10). Line denotes 1-fold induction (no induction) of luciferase. See text for futher details. B: correlation of luciferase induction and binding to the slowly migrating bands (data from A).

Construction of reporter genes. To generate a reporter gene construct, a double-stranded TonE nucleotide was phosphorylated and ligated into the Spe I site of pBluescript II SK(+) (Stratagene, La Jolla, CA). Clones containing two copies of the oligonucleotide were picked and sequenced for verification. The "2 TonE" fragment was moved into the Kpn I/Sac I site upstream of the SV40 promoter and the Photinus luciferase gene, using a commercial vector, pGL2-promoter (Promega, Madison, WI). pRL-CMV (Promega) containing the Renilla luciferase gene under the strong promoter of cytomegalovirus served to assess transfection efficiency. The "-actin control" construct containing the Photinus luciferase gene under the control of the promoter of the -actin gene (24) was used as a reference (see below).

Transfection and analysis of reporter gene expression. A half million MDCK cells were seeded onto a 60-mm tissue culture dish. The next day, they were transfected with 2 µg of a 2 TonE reporter construct along with 0.05 µg of pRL-CMV using DEAE-dextran (24). For transfection of the -actin control plasmid, 0.01 µg was used along with 0.05 µg of pRL-CMV. Transfected cells were maintained in isotonic medium for 24 h and then switched to hypertonic medium or maintained in isotonic medium for another 24 h before analysis. Activity of the Photinus and Renilla luciferase in extracts of the transfected cells was determined using a commercial kit, Dual-Luciferase Reporter Assay System (Promega). For each sample, activity of the Photinus luciferase is divided by the activity of the Renilla luciferase to correct for transfection efficiency. The corrected Photinus luciferase activity of experimental constructs was expressed relative to that of the -actin control plasmid in isotonic or hypertonic conditions, as described (24). In each experiment, all transfections were performed in duplicate. Expression of Photinus luciferase driven by the SV40 promoter alone (pGL2 promoter by itself ) was not affected by hypertonicity (0.98 ± 0.15-fold induction by hypertonicity; means ± SE, n = 7).

RNA isolation and Northern analysis. RNA was isolated from MDCK cells using Trizol Reagent (Life Technologies, Gaithersburg, MD), and 10 µg of each sample was electrophoresed on a 1% agarose gel containing 2.2 M formaldehyde. In all samples, equal loading of RNA was confirmed by visual inspection of the ethidium staining. RNA was transferred to a nitrocellulose membrane (Schleicher & Schuell, Keene, NH) and hybridized to random-primed BGT1 cDNA (31) in 6× standard sodium citrate (SSC) (1× SSC contains 150 mM NaCl and 15 mM trisodium citrate), 0.5% SDS, 5× Denhardt's solution, and 100 µg/ml of herring sperm DNA at 65°C. Membranes were washed for 30 min at 65°C in 0.5× SSC and 0.1% SDS. Radioactivity was detected and quantified as described for the EMSA gels.

In vivo footprinting. To methylate G residues of the genomic DNA in vivo, MDCK cells in isotonic or hypertonic medium were incubated in the same medium containing 0.1% dimethyl sulfate for 2 min at room temperature. Cells were washed quickly to remove dimethyl sulfate, and DNA was isolated. As a control to detect all the G residues, DNA isolated from MDCK cells was treated with 0.1% dimethyl sulfate for 2 min in vitro. The methylated DNA was converted to the single-stranded form and cleaved at sites immediately 3' to the methylated G residues by treatment with piperidine at 90°C. The cleaved G residues were detected using ligation-mediated polymerase chain reaction (PCR), as described in Ref. 18. To amplify the "sense" strand of the promoter region of the BGT1 gene, the cleaved DNA was annealed to an anti-sense primer IVF-1 [TGTATGGGGCAGGGTGCAGC: complementary to +67/+86 region where numbers represent positions of nucleotides in the BGT1 gene and its 5' flanking region. Nucleotides are numbered positively starting from the transcription start site (+1) to the downstream direction, and they are numbered negatively in the upstream direction starting from the nucleotide immediately upstream of the transcription start site (1) (see below)], which was then extended using Vent DNA polymerase (New England Biolabs, Beverly, MA). A staggered double-stranded linker (18) was ligated, and 18 cycles of PCR were performed using a nested primer IVF-2 (CCAGAGCAGTAGGATGGGAGCCACCAA: complementary to +32/+58 region) and the linker primer. Two additional rounds of PCR were performed using IVF-2 end labeled with ³²P, and the reaction was electrophoresed on a sequencing gel to visualize the PCR products. Radioactivity was visualized and quantified by PhosphorImager, as described for EMSA (see above). The "anti-sense" strand of the promoter region was amplified and visualized in the same way using two primers: IVF-3 (GGTCTGTCTGACGGTAAACTTG, corresponding to 214/193 region) and IVF-4 (CTAGATAGGCTCCTTGAGGTTTGCTC, corresponding to 184/159 region).

Ultraviolet cross-linking. Nuclear extract (30 µg of protein) was initially incubated for 10 min at 25°C with 50 fmol of ³²P-labeled human TonE (³²P-hTonE) or canine TonE (³²P-cTonE) (see Fig. 1) and 2.5 µg of poly(dA-dT) in 50 µl containing the same buffer used for EMSA analysis (see above) and then chilled on ice for 10 min. Samples on ice were irradiated with ultraviolet (UV) for 60 min using the Stratalinker apparatus (Stratagene). Samples were boiled for 3 min in Laemmli buffer and then electrophoresed on an SDS-polyacrylamide gel. Radioactivity of dried gels was detected using a PhosphorImager.

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

Identification of TonEBPs and delineation of TonE sequence. We have previously located a TonE sequence of the canine BGT1 gene within 13 bp in the 5' flanking region, i.e., 62/50 region (24). Using a 35-bp probe corresponding to 69/35 region of the gene, we detected many nuclear proteins that bound to the probe, including those migrating very slowly in EMSA gels. Although the intensity of the (very) slowly migrating bands was much weaker than that of other bands, the binding of the probe to the slowly migrating bands, but not to others, was specifically competed by oligonucleotides containing the 62/50 sequence but not by oligonucleotides with mutations in the 62/50 region (24). To investigate the slowly migrating bands further, we prepared a shorter probe of 23 bp, which contains the TonE region with 5-bp flanking sequences on both sides (cTonE in Fig. 1; equivalent to 67/45 region of the BGT1 gene). Unlike the 69/35 probe described above, cTonE bound efficiently to the slowly migrating bands while it bound to other proteins with much less efficiency (Fig. 1B, left, far left lane). cTonE binding in the slowly migrating bands displayed the same specificity as the 69/35 probe, i.e., the profile of competition by wild-type and mutant oligonucleotides described (24) was the same for both probes (not shown). The slowly migrating bands appeared as two bands when the EMSA gel was electrophoresed for a longer time (Fig. 3A).

View larger version (31K): [in this window] [in a new window]   View larger version (20K): [in this window] [in a new window]   Fig. 3.   Regulation of TonEBPs in MDCK cells. A: time course of TonE binding protein (TonEBP) activation and increase in the abundance of BGT1 mRNA in response to hypertonicity. Top: 3.5- and 2.9-kb BGT1 mRNA bands in a Northern analysis (duplicate samples at each time point) and TonEBP bands (slowly migrating bands as in Fig. 1) of nuclear extracts of MDCK cells switched to hypertonic medium for the period of time shown. Bottom: relative abundance of mRNA or relative activity of TonEBPs (abundance or activity in cells cultured in hypertonic medium divided by abundance or activity in cells cultured in isotonic medium at each time point) is presented as means ± SE (n = 3 or 4). B: effects of inhibitors on hypertonicity-induced increase in BGT1 mRNA abundance and TonEBP activity. MDCK cells were incubated with actinomycin D (Act D, 1 µg/ml), cycloheximide (CHX, 3 µg/ml), and 2-aminopurine (2-AP, 3 mM) for 30 min before their medium tonicity was increased by adding 200 mM raffinose. After 6 h, abundance of BGT1 mRNA and activity of TonEBPs relative to cells treated with control (Cont) isotonic medium without inhibitor (indicated by dashed line) were measured as in A (means ± SE, n = 3).

After we reported the TonE sequence, other investigators reported tonicity-responsive enhancers with sequences similar to TonE, mostly from the promoter region of the AR gene of several species (4, 5, 10, 21). To test whether these enhancers are functionally related to TonE, we prepared rat TonE (rTonE) and hTonE by replacing appropriate portions of cTonE with the 12-bp enhancer from the rabbit AR gene (5) and the 11-bp enhancer from an unknown human gene (10, 21), respectively (Fig. 1A). hTonE was identified from a human genomic clone based on its TonE activity, but the origin of the portion of the clone containing hTonE is not known (21). At any rate, both rTonE (not shown) and hTonE (Fig. 1B, right) formed the slowly migrating bands. The binding by any one of the three probes (cTonE, rTonE, and hTonE) to the slowly migrating bands was competed by all the TonEs, indicating that these three elements interact with the same nuclear proteins, presumably transcription factors (see below for more). Relative affinity of the probes for the slowly migrating bands based on the degree of competition was hTonE > rTonE > cTonE.

To pursue the relationship between the slowly migrating bands and the enhancer activity of TonE further and also to identify the key nucleotides of TonE, 11 hTonE mutants were generated by changing a single base pair, a purine to a pyrimidine and vice versa, in each mutant as shown in Fig. 2A. hTonE was chosen as the founding wild type because it has the highest affinity to the slowly migrating bands (Fig. 1) and very high enhancer activity (see below), making it easier to detect reduction in affinity/enhancer activity of mutants, compared with using cTonE or rTonE as a wild-type TonE. Nucleotides at the 8th and 10th position of TonE were not mutated because they are variable among TonE sequences already reported (see Table 1). We determined the ability of all hTonE mutants, hTonE, and cTonE to bind to the slowly migrating bands and to induce transcription in response to hypertonicity, i.e., tonicity-responsive enhancer (TonE) activity, as described below.

To quantify binding (or affinity) of each sequence to the slowly migrating bands, percent reduction (or competition) of ³²P-hTonE (0.5 nM) binding to the slowly migrating bands in the presence of 25 nM of the test sequence was determined from the EMSA gels using ImageQuant software, as described in MATERIALS AND METHODS. hTonE competed 96%, while cTonE competed 61% of ³²P-hTonE binding to the slowly migrating bands. Mutant hTonEs showed a wide range of competition, from 0 to 95%, as shown in Fig. 2A.

To quantify enhancer activity, two tandem copies of each of the wild-type and mutant sequences were inserted in front of the SV40 promoter and the Photinus luciferase reporter gene, using a commercial vector, pGL2-promoter. These reporter constructs were individually transfected into MDCK cells, and the expression of luciferase under isotonic and hypertonic culture conditions was determined. Multiple-fold induction of luciferase in response to hypertonicity (i.e., TonE activity) was calculated by dividing the luciferase activity in cells in hypertonic medium by the luciferase activity in cells in isotonic medium. As shown in Fig. 2A, hTonE induced the reporter gene expression 21-fold, whereas cTonE induced 2.2-fold in response to hypertonicity.

TonE activity of the tandem cTonE (67/45 of the BGT1 gene; see Fig. 1) (above, 2.2-fold) is much lower than that of tandem 69/50 region of the BGT1 gene, which was over 12-fold in previous studies (24). We measured the enhancer activity of the tandem 69/50 region along with the tandem cTonE and confirmed that the tandem 69/50 construct induces over 12-fold. Arrangement of the sequences (head to head, tail to tail, head to tail, and tail to head) did not affect the degree of induction (data not shown). Because the 69/50 oligonucleotide was dimerized with Cla I overhang sequence (TCGA), whereas cTonE was dimerized with Spe I overhang sequence (CTAG) and the same commercial vector (pGL2-promoter) was used for both constructs, differences in sequences around the TonE (62/52 region, see below) must have contributed to the differences in luciferase induction.

Mutant hTonEs that displayed relative binding activity <50, i.e., hTonE-m2, hTonE-m3, hTonE-m4, hTonE-m5, hTonE-m8, and hTonE-m9 (see Table 1 for nomenclature), did not induce luciferase expression by hypertonicity. hTonE-m1 displayed binding of 55 and luciferase induction of 1.79 ± 0.28 (mean ± SE, n = 7), whereas hTonE-m6 displayed a moderate luciferase induction of 7.06, although its binding of 87 was rather high. The rest, hTonE-m7, hTonE-m10 (C to A in the 12th position), and hTonE-m11 (G to T in the 13th position), were as active as hTonE in binding and in luciferase induction. The absence of effects of mutations in the 12th and 13th position (hTonE-m10 and hTonE-m11) indicates that these residues are not part of TonE. Other TonE sequences are 11 bp long as well (4, 5, 10, 21) (see also Table 1). We conclude that TonE of the BGT1 is 11 bp long, located in 62/52 region. Based on all the TonE sequences in this study and those reported by others, we derived a consensus sequence of TonE: YGGAAnnnYnY (Y is C or T; n is any nucleotide).

Data in Fig. 2A were replotted in Fig. 2B to look for a correlation between binding to the slowly migrating bands and induction of luciferase by hypertonicity. The induction of luciferase increased dramatically as binding exceeded 50. The steepness of the curve may be due to cooperativity in the binding of specific transcription factors to the tandem TonE sites on the reporter constructs, as a result of protein-protein interactions, as in many other regulatory elements and their corresponding transcription factors (25). The relationship between the binding to the slowly migrating bands and the induction of transcription strongly supports the idea that the slowly migrating bands represent the transcription factors specifically interacting with TonE. Accordingly, the proteins in the slowly migrating bands are named TonE binding proteins (TonEBPs).

Regulation of TonEBPs by hypertonicity. Nuclear extracts prepared from hypertonic MDCK cells contain considerably more TonEBP activity (TonEBP activity refers to DNA binding activity, as determined by the radioactivity of the slowly migrating bands in EMSA gels) than those prepared from isotonic cells (24), indicating that hypertonicity increases the activity of TonEBPs. To investigate the temporal relationship of TonEBP activation and transcription of the BGT1 gene, we determined the time course of changes in the activity of TonEBPs and BGT1 mRNA abundance after MDCK cells were switched to hypertonic medium. The mRNA abundance was taken as an indicator of transcription, because we have previously shown that changes in the abundance of BGT1 mRNA closely reflect changes in transcription of the BGT1 gene in this model (28). As shown in Fig. 3A, the activity of TonEBPs increased steadily, reaching ninefold isotonic level at 18 h and then declined to a lower level, a profile that is quite similar to the time course of transcription of the BGT1 gene as determined by nuclear run-on assays (28). The abundance of BGT1 mRNA rose and reached >10-fold isotonic level at 18 h, similar to previous results (28). The data add further evidence that TonEBPs mediate transcriptional stimulation of the BGT1 gene.

To further investigate the relationship between TonEBPs and hypertonicity-induced transcription of BGT1 and also to explore the signaling pathway to activation of TonEBPs, we examined the effect of an inhibitor of transcription (actinomycin D), an inhibitor of protein synthesis (cycloheximide), and an inhibitor of serine/threonine protein kinases (2-aminopurine) (7). To minimize toxic side effects, MDCK cells were treated with these agents at the lowest effective concentrations in reducing the induction of mRNA in response to hypertonicity and only for the initial 6 h after medium tonicity was raised. We did not detect any obvious toxicity by these inhibitors, based on apprearance of cells and yield of protein and RNA. In previous studies, we determined that the increase in mRNA can be reliably detected at 6 h (28). Actinomycin D, as anticipated, prevented the increase in BGT1 mRNA abundance. It did not, however, prevent activation of TonEBP, suggesting that transcription is not required for the initial activation of TonEBPs (Fig. 3B). Cycloheximide, on the other hand, decreased both BGT1 mRNA abundance and TonEBP activity below the isotonic control level. Protein synthesis appears necessary for activation of TonEBPs. The reduction in transcription of the BGT1 gene (as indicated by the decrease in mRNA abundance) by cycloheximide may be due the decreased activity of TonEBPs. 2-Aminopurine prevented the increase in BGT1 mRNA abundance but not the activation of TonEBPs, i.e., DNA binding. It is possible that activation of transcription (transactivation) by TonEBPs involves serine/threonine phosphorylation, as in many transcription factors (8).

The observation that the hypertonicity-induced increase in TonEBP activity is temporally correlated with the transcription of the BGT1 gene (Fig. 3A) implies that binding of TonEBPs to the TonE site in vivo should also correlate with the time course of TonEBP activity seen in EMSA gels. To test this directly, we performed in vivo footprinting analysis of MDCK cells (Fig. 4). The primers used for this analysis enabled us to examine 160/+70 region of the BGT1 gene. With the exception of the G residue at 44 position (Fig. 4A) and the complementary G residue at 40 position (Fig. 4B), all the bands seen in control (in vitro methylated) lanes were also seen in experimental (especially isotonic samples) with similar intensity. In other words, most of the promoter area was quite accessible for methylation by dimethyl sulfate, indicating that the promoter/regulatory region is available to receive the signal of hypertonicity in the form of TonEBP binding. The second and third G residues of the TonE site of the BGT1 gene were protected from methylation (Fig. 4A) in a temporal correlation with the abundance of TonEBPs (Fig. 3A). The G residue complementary to the C residue in the last (11th) position of TonE was also protected by hypertonicity with a similar time course (Fig. 4B). Because mutations in these residues affect binding to TonEBPs (Fig. 4C), the protection of these residues from methylation is likely due to binding of TonEBPs.

View larger version (47K): [in this window] [in a new window]   View larger version (45K): [in this window] [in a new window]   View larger version (15K): [in this window] [in a new window]   Fig. 4.   Changes in the occupancy of the TonE site of the BGT1 gene in response to hypertonicity. Site occupancy was determined using the in vivo footprinting technique, in which the efficiency of methylation of G residues by dimethyl sulfate in vivo was measured. MDCK cells cultured in hypertonic medium (H) for the times shown and their matched isotonic controls (I) were treated with dimethyl sulfate before isolation of their DNA. DNA was cleaved with piperidine, and quantitative PCR was performed to detect methylated G residues. 32P-labeled primers were used for the last two rounds of PCR to visualize amplified products in sequencing gels as shown. To visualize all the G residues, dimethyl sulfate treatment was performed in vitro after the DNA was isolated from MDCK cells (control). A, left: PCR products of the "sense" strand. Sequence of the TonE site (62/52 region) is shown at left, and its corresponding region in the gel is demarcated. To obtain calibrated signal of the GG residues at 61/60 position, radioactivity of the bands representing these GG residues (arrowhead, right) was divided by radioactivity of bands representing reference G residues (3 doublets at 71/70, 64/63, and 47/46 positions) marked by  on the right; , G at 44 position. Changes in the calibrated signal of the 61/60 bands in hypertonic cells relative to that of isotonic control at each time point are shown in A, right. Means ± SE are shown (n = 4). B, left: PCR products of the "anti-sense" strand. TonE region is indicated as in A. To obtain calibrated signal of the G residue at 52 position (complementary to the bold C at left, marked by arrowhead), radioactivity of the band corresponding to this G residue was divided by radioactivity of bands representing reference G residues complementary to 73, 67, and 49 positions (); , G residue complementary to the C at 40 position. B, right: changes in calibrated signal of the complementary 52 G residue in hypertonic cells, compared with that of isotonic control at each time point. G residue complementary to the C residue at 53 position was poorly methylated in all samples, including control. C: summary of changes in methylation by hypertonicity. Double-stranded sequence of BGT1 TonE is shown. Nucleotides whose mutation did not change activity (Fig. 2) are shown in non-boldface letters.  G residues whose methylation was decreased by hypertonicity; arrowhead shows G residue whose methylation was increased by hypertonicity (see text).

On the other hand, methylation of the G residue in the 8th position of the sense strand increased by hypertonicity with a time course similar to the TonEBP activity (27% increase at 6 h, 82% increase at 18 h, and 46% increase at 48 h, n = 4; the signal was calibrated to the reference residues as described in Fig. 4A). Because this residue can be changed without affecting binding to TonEBP (Table 1), it is likely that binding of TonEBPs creates a conformational change in TonE sequence, rendering the 8th residue more accessible to dimethyl sulfate. At any rate, all the changes in the methylation described above were localized and limited to the TonE region, supporting the idea that the TonE site is the major point of control in the response to hypertonicity.

Biochemical characterization of TonEBPs. To characterize DNA binding components of TonEBPs, radiolabeled hTonE or cTonE was covalently cross-linked to TonEBPs by UV irradiation. ³²P-TonE was incubated with nuclear extracts and then irradiated with short-wavelength UV light. The mixture was denatured in SDS sample buffer and resolved on an SDS polyacrylamide gel. The same results were obtained for hTonE and cTonE (Fig. 5): there was a large polypeptide of 200 kDa, whose abundance is much greater in hypertonic MDCK cells. We have not detected any other band whose intensity increased consistently in the hypertonic cells compared with the isotonic cells. Labeling of the 200-kDa polypeptide was specifically blocked by 100 nM unlabeled hTonE or cTonE but not by 100 nM hTonE-m8, which does not bind TonEBP (see Fig. 2A), indicating that this protein is a DNA binding component of TonEBPs.

View larger version (69K): [in this window] [in a new window]   Fig. 5.   UV cross-linking of TonE to nuclear extracts. Nuclear extract isolated from MDCK cells cultured in isotonic (I) or hypertonic (H) medium for 24 h was incubated with 1 nM of 32P-hTonE (A) or 32P-cTonE (B) and irradiated with UV light. In some experiments, 100 nM of unlabeled hTonE, hTonE-m8, or cTonE was added to the incubation mixture as a competitor as indicated, top. Reaction mixtures were then boiled in SDS sample buffer and size fractionated by electrophoresis in 8% (A) or 6% (B) polyacrylamide gel. B: 3rd lane was loaded with 2.5 times equivalent of other lanes to detect reduction in signal intensity of the 200-kDa band. Radioactivity of the dried gel was visualized by PhosphorImager. Mobility of size markers are shown, right.

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

TonE may be a general enhancer of hypertonicity. The data presented here indicate that the TonE sequences of the BGT1 and AR genes are functionally identical. The sites of expression for these genes are, however, rather distinct in the renal medulla; AR mRNA expression is limited to the inner medulla, i.e., the thin limb of the loop of Henle and the inner medullary collecting duct, whereas BGT1 mRNA is abundantly expressed in the thick ascending limb (outer medulla) and the inner medullary collecting duct (16). Although the regulation by hypertonicity is mediated by the TonE sequences, other factors, such as tissue-specific transcription factors must determine the tissue distribution of expression. At this time, it is not known whether the TonE sequences are involved in the regulation of other genes whose transcription is also stimulated by hypertonicity such as SMIT (14), the taurine transporter (26) and CD9 (22). It would not be surprising to find that TonEs are involved in their regulation.

Based on the consensus sequence presented in Table 1, one would expect to find an average of one copy of TonE per 2 kb of genomic DNA sequence. Because only a few genes have been shown to be stimulated by hypertonicity, there may be requirements in addition to a copy of the TonE sequence in the promoter region for transcriptional stimulation. One possibility is that more than one TonE clustered within a short range may be required. In the human AR gene, three TonEs are present within 130 bp (located ~1.1 kb upstream of the promoter), and all of them are required for full regulation of the promoter (10). When the TonE of BGT1 is placed in front of the promoter of the SMIT gene, one copy or two tandem copies are not sufficient to stimulate the promoter in response to hypertonicity (20). On the other hand, four tandem copies of the BGT1 TonE stimulate the SMIT promoter over eightfold in response to hypertonicity (20), indicating that a synergistic action of multiple TonEs is required for significant stimulation. Although one TonE at 62/52 plays the major role in the regulation of BGT1, there is an unidentified element(s) in 185/70 region that is required for full regulation of the promoter (24). It is not known whether this unidentified element is related to TonE.

TonEBPs are transcription factors and targets of hypertonicity signaling pathways. Based on a tight correlation between binding (or affinity) and stimulation of transcription (Fig. 2B) in a panel of mutant TonE sequences, we characterized TonEBPs that are likely to be transcription factors that interact with the RNA polymerase complex at the promoter. This is further supported by the observation that in vivo methylation of the G residues that are critical for TonEBP binding in vitro (Fig. 2A) is specifically protected in a temporal correlation with the TonEBP activity (Fig. 4). Activation of TonEBPs appears to be the critical step in that it correlates well with the occupancy of the TonE site and transcription. TonEBPs should be the target of the signaling pathways by which hypertonicity stimulates transcription of specific genes whose products lead to accumulation of compatible osmolytes. Alternatively, TonEBPs themselves may serve as tonicity sensors.

Studies using inhibitors provide basic clues to the signaling pathways for activation of TonEBPs. It appears that transcription is not required but translation is necessary for stimulation of TonEBPs (Fig. 3B). Interestingly, inhibition of serine/threonine phosphorylation interferes with transactivation but not the activation of binding of TonEBPs, suggesting that there are two different levels of regulating TonEBPs: activation of binding to TonE and activation of the ability of TonEBP to increase transcription (transactivation). Although TonEBPs are seen as two close bands in EMSA gels, UV cross-linking reveals only one DNA binding polypeptide of 200 kDa. This 200-kDa polypeptide may exist in two forms seen in the EMSA gels possibly as a component of multi-subunit complexes.