ATM, a DNA damage-inducible kinase, contributes to activation by high NaCl of the transcription factor TonEBP/OREBP

Abstract

High NaCl activates the transcription factor tonicity-responsive enhancer/osmotic response element-binding protein (TonEBP/OREBP), resulting in increased transcription of several protective genes, including the glycine betaine/γ-aminobutyric acid transporter (BGT1). High NaCl damages DNA, and DNA damage activates ataxia telangiectasia mutated (ATM) kinase through autophosphorylation on Ser-1981. TonEBP/OREBP contains ATM consensus phosphorylation sites at Ser-1197, Ser-1247, and Ser-1367. The present studies test whether ATM is involved in activation of TonEBP/OREBP by high NaCl. We find that raising osmolality from 300 to 500 mosmol/kg by adding NaCl activates ATM, as indicated by phosphorylation at Ser-1981. High urea and radiation also activate ATM, but they do not increase TonEBP/OREBP transcriptional activity like high NaCl does. Wortmannin, which inhibits ATM, reduces NaCl-induced TonEBP/OREBP transcriptional activation and BGT1 mRNA increase. Overexpression of wild-type TonEBP/OREBP increases ORE/TonE reporter activity much more than does overexpression of TonEBP/OREBP S1197A, S1247A, or S1367A. In AT cells (which express nonfunctional ATM), TonEBP/OREBP transcriptional and transactivating activity are further increased by expression of wild-type ATM but not of S1981A ATM. TonEBP/OREBP reciprocally coimmunoprecipitates with ATM kinase, demonstrating physical association. Additionally, antibody to ATM kinase supershifts TonEBP/OREBP bound to its cognate ORE/TonE DNA element. In AT cells, wortmannin further decreases high NaCl-induced increase in transcriptional activity, consistent with participation of signaling kinase(s) in addition to ATM. In conclusion, signaling via ATM is necessary for full activation of TonEBP/OREBP by high NaCl, but it is not sufficient.

In renal inner medullary interstitial fluid, NaCl concentration is variable, but normally it is very high (1), a necessary condition for concentration of the urine. Increase of NaCl concentration is hypertonic. It, like other hypertonic stresses, perturbs the function of cells by shrinking them and increasing the concentration of intracellular components, including inorganic ions. Despite this stress, the inner medullary cells survive and function. The nearly universal protective response to hypertonicity is accumulation of compatible organic osmolytes (2). Accumulation of organic osmolytes protects against hypertonicity by normalizing both cell volume and inorganic ion concentration (2). The renal compatible organic osmolytes include glycine betaine (betaine), myo-inositol (inositol), and sorbitol (3). Accumulation of these renal organic osmolytes is regulated by a transcription factor, tonicity-responsive enhancer/osmotic response element-binding protein (TonEBP/OREBP) (4, 5). TonEBP/OREBP initiates the accumulation of compatible organic osmolytes by increasing transcription of the betaine/γ-aminobutyric acid transporter (BGT1), the sodium-myo-inositol cotransporter (SMIT), and aldose reductase (4–6). BGT1 and SMIT transport betaine and inositol, respectively, into the cells, and aldose reductase catalyzes conversion of glucose to sorbitol.

Regulation of TonEBP/OREBP transcriptional activity is complex. Within 30 min of hypertonicity, TonEBP/OREBP becomes phosphorylated and translocates into the nucleus (4, 5, 7). Hours later, TonEBP/OREBP mRNA and protein abundance increase (4, 5). Also, hypertonicity increases transactivating activity of TonEBP/OREBP, associated with phosphorylation of its transactivating domain (8).

There is evidence that several protein kinases participate in activation of TonEBP/OREBP by hypertonicity, including p38 (9, 10), Fyn (10), and protein kinase A (PKAc) (11). However, there might be additional ones. In particular, the stress of high NaCl causes DNA double-strand breaks. Repair of these breaks correlates with cell survival and is repressed by 25 μM LY294002, an inhibitor of DNA-activated protein kinases (12). Responses to DNA damage include activation of a number of kinases. Ataxia telangiectasia mutated (ATM) kinase is activated early in the DNA damage response (13), apparently not directly by the DNA breaks but as a consequence of accompanying changes in chromatin structure (13). ATM phosphorylates a number of proteins, including p53, MDM2, CHK2, BRCA1, and NBS1, contributing to DNA repair, cell-cycle delay, and apoptosis (14–18). Because high NaCl causes DNA double-strand breaks, we hypothesized that ATM, activated by high NaCl, might also contribute to activation of TonEBP/OREBP. The purpose of the present studies was to test this hypothesis.

Experimental Procedures

Cell Culture. HEK293 cells (American Type Culture Collection) and AT cells (GM09607 B, in which ATM protein is truncated; Coriell Cell Repositories, Camden, NJ) were maintained according to instructions of the supplier. Osmolality of the basal medium was 300 mosmol/kg. Osmolality was increased by adding NaCl to 500 mosm/kg or urea to 600 mosm/kg. Wortmannin (Calbiochem) was solubilized with DMSO, and the same final concentration of DMSO (<0.25%) was added to controls. Some HEK293 cells were exposed to ionizing radiation (10 Gy, Gammacell 1000, MD Nordion, Ottawa) or UV radiation (15 J/m², UV Stratalinker, Stratagene).

Plasmids and Transfection. Human TonEBP/OREBP cDNA clone KIAA0827 was a gift from T. Nagase (Kazusa DNA Research Institute, Chiba, Japan), and wild-type or mutant (S1981A) Flag-ATM (15) were gifts from M. B. Kastan (St. Jude Children's Research Hospital, Memphis, TN). The ORE-X luciferase reporter construct contains two copies of human ORE-X (19) within a minimal IL-2 promoter (20) (hTonE-GL3, a gift from S. N. Ho, University of California, San Diego). The binary GAL4 reporter system has been described (8). In brief, plasmid pFR-Luc (Stratagene) contains the yeast GAL4-binding site (upstream activating sequence) upstream of a minimal promoter and the Photinus pyralis luciferase gene. Expression plasmid pFA-CMV (Stratagene) contains sequence coding for the yeast GAL4 DNA-binding domain (dbd) under control of a cytomegalovirus promoter. A fusion protein was generated by in-frame insertion of the sequence coding for amino acids 548–1,531 of KIAA0827 into pFA-CMV to generate GAL4dbd-548–1531. Sequence coding for TonEBP/OREBP amino acids 1–1,531 from KIAA0827 was cloned into pcDNA6 V5·His (Invitrogen) expression vector to generate 1-1531V5. Mutants (S1197A, S1247A, and S1367A) of 1-1531V5 were prepared by site-directed mutagenesis. All constructs were generated by using standard cloning procedures and verified by restriction enzyme digestion and DNA sequencing. DNA was transfected into cells by using Effectene according to the manufacturer's instructions (Qiagen, Valencia, CA).

Western Blot Analysis. Cells were lysed with mammalian protein extraction reagent (Pierce) according to the manufacturer's instructions, with added protease inhibitor mixture (Roche Diagnostics) and phosphatase inhibitor mixture (Sigma). Proteins were separated on 4–15% SDS-polyacrylamide gels and transferred to nitrocellulose membranes. Nonspecific binding was blocked with 5% nonfat milk powder and 0.1% Tween 20 in PBS. Membranes were incubated overnight at 4°C with anti-ATM1981S-P or anti-ATM (MAT3), gifts from M. B. Kastan. Blots were detected by enhanced chemiluminescence (Amersham Biosciences).

Electrophoretic Mobility-Shift Assay. Nuclear pellets from HEK293 cells treated 2 h at 500 mosmol/kg (NaCl added) were prepared by using nuclear and cytoplasmic extraction reagents (Pierce) according to the manufacturer's instructions. Nuclear pellets were resuspended in lysis buffer (21), centrifuged 10 min at 15,000 × g, and the supernatant was retained. Double-stranded ORE/TonE probe (–1,238 to –1,104 of the human aldose reductase gene) containing three OREs in native gene context was generated by annealing complementary 5′ biotinylated oligonucleotides (Integrated DNA Technologies, Coralville, IA). We preincubated 2 μg of nuclear extract with 0.6 μg of poly(dA-dT) in binding buffer (21) for 10 min at room temperature. Anti-TonEBP/OREBP (1 μg, NFAT5, Affinity Bioreagents, Neshanic Station, NJ) or anti-ATM (1 μl, 5× ATM D1611 antisera, gift from M. B. Kastan) was added to some reactions and incubated at 4°C for 1 h; 10 fmol of ORE/TonE with or without 100 fmol of nonbiotinylated ORE/TonE was added, and reactions were incubated for an additional 20 min. Total binding reaction volume was 20 μl. Reaction products were separated by electrophoresis on 0.7% SeaKem ME agarose gels in 0.5× Tris·borate/EDTA buffer at 4°C, transferred to nylon membranes, and crosslinked (UV Stratalinker). Blots were detected by using the LightShift chemiluminescent electrophoretic mobility-shift assay kit (Pierce) according to the manufacturer's instructions.

Real-Time PCR. Total RNA was isolated (RNeasy, Qiagen), and cDNA was prepared (TaqMan reverse transcription kit, Applied Biosystems), according to the manufacturer's instructions. PCR was performed on 8- and 80-ng cDNA samples in 20-μl reaction mixtures in triplicate (TaqMan PCR master mix, Applied Biosystems). Amplicons were detected with an ABI Prism 7900HT sequence detection system (Applied Biosystems). The primers for human BGT1 were 5′-CCCGAGGAGGGAGAGAAGTT-3′ and 5′-TCCATCTTGTTGGTCCATTGG-3′. The 6FAM-labeled probe was 5′-AAAGACGAGGACCAGGTGAAGGATCGG-3′. The detection system records the number of PCR cycles (Ct) required to produce an amount of product equal to a threshold value, which is a constant. From the Ct values we calculated the mRNA abundance in each experimental condition, relative to that of control cells at 300 mosmol/kg, as described (11).

Immunoprecipitation. At 24 h after HEK293 cells were transfected at 300 mosmol/kg with 1-1531V5, fresh medium at 300 or 500 mosmol/kg (NaCl added) was substituted. At 3 h later, cells were trypsinized and pelleted by centrifugation. Subsequent steps were at 4°C. The pellet from one 10-cm dish was extracted for 5 min with 1 ml of lysis buffer containing 50 mM Tris·HCl (pH 8.0), 150 mM NaCl, 1% Triton X-100, protease inhibitor mixture (Roche Diagnostics), and phosphatase inhibitor mixture (Sigma) and centrifuged at 15,000 × g for 10 min). Samples were precleared with 0.2 mg of protein A-agarose (Roche Diagnostics) and 12 μg of IgG-agarose (Santa Cruz Biotechnology) for 1 h and centrifuged. For immunoprecipitation of recombinant TonEBP/OREBP and any associated proteins, the precleared supernatants were incubated for 1 h with 4 μg of mouse anti-V5 (Invitrogen), then overnight with addition of 0.2 mg of protein A-agarose beads. The beads were resuspended in Laemmli sample buffer, and, after centrifugation, supernatant proteins were separated on a Tris·HCl/4–15% polyacrylamide gel (Bio-Rad). Proteins were transferred to a poly(vinylidene difluoride) membrane, which was then cut. One part of the membrane was incubated at 4°C overnight with mouse anti-TonEBP/OREBP (NFAT5, Affinity Bioreagents); the others were incubated with mouse anti-ATM (MAT3, gift from M. B. Kastan) or rabbit anti-PKAc (PKAcα, Santa Cruz Biotechnology). Blots were detected by ECL (Amersham Biosciences). As negative controls, mouse IgG-agarose conjugate (Santa Cruz Biotechnology) was substituted for anti-V5, or cells were transfected with empty pcDNA6-V5His vector. For immunoprecipitation of ATM and any associated proteins, the precleared supernatant was mixed with rabbit anti-ATM (ATM D1611, gift from M. B. Kastan). As a negative control, immunoprecipitation was performed with mouse IgG agarose conjugate (Santa Cruz Biotechnology).

Luciferase Assay. Cells grown in a six-well plate were transfected overnight with 0.2 μg of ORE-X luciferase reporter. Then, medium was substituted with the same osmolality (300 mosmol/kg) or elevated with NaCl to 500 mosmol/kg or with urea to 600 mosmol/kg. Luciferase activity was measured either 2 or 16 h later with the Luciferase Assay System (Promega).

Statistical Analysis. Data were compared by a one-way ANOVA followed by Dunnett's multiple comparison test for separation of significant means. Logarithmic transformation was applied first to data calculated as ratios. Results are expressed as mean ± SEM (n = 3). Differences were considered significant for P ≤ 0.05.

Results

ATM kinase is activated by intermolecular phosphorylation on serine 1981 (13). We used a phosphospecific antibody, ATM1981S-P, to measure ATM activation in HEK293 cells exposed to different osmotic stresses and DNA damaging agents (Fig. 1A). ATM1981S-P increases in cells exposed for 1 h to high NaCl or high urea, as well as ionizing or UV radiation. High urea (600 mosmol/kg) and UV (15 J/m²) have less effect than high NaCl (500 mosmol/kg) or ionizing radiation (10 Gy). The total amount of ATM is unaffected. Thus, the osmotic stress of high NaCl or urea activates ATM.

Fig. 1.

ATM is activated by different DNA damaging agents, including high NaCl, but only high NaCl increases transcriptional activity of TonEBP/OREBP. (A) Western blot analysis of HEK293 cells by using anti-ATM or phosphospecific anti-ATM 1981S-P. Cells were harvested 1 h after the indicated change. (Lower) Representative immunoblot. (Upper) Densitometric analysis normalized to control at 300 mosmol/kg. (B) ORE-X reporter activity in HEK293 cells was measured 2 h after the indicated change. Values are normalized to control cells at 300 mosmol/kg. Mean ± SEM (*, P ≤ 0.05, n ≥ 3).

We used a luciferase reporter system to measure TonEBP/OREBP transcriptional activity (Fig. 1B). Of the agents that activate ATM, only high salt increases TonEBP/OREBP transcriptional activity. Further, although low NaCl activates ATM (13), it reduces TonEBP/OREBP transcriptional activity (Fig. 1B). We conclude that high NaCl activates ATM, but activation of ATM is not, of itself, sufficient to activate TonEBP/OREBP.

We used wortmannin, an inhibitor of ATM activity, to determine whether ATM might be necessary for activation of TonEBP/OREBP by high NaCl. Raising osmolality from 300 to 500 mosmol/kg by adding NaCl for 16 h greatly increases TonEBP/OREBP transcriptional activity in HEK293 cells (Fig. 2A). A total of 1 or 20 μM wortmannin significantly inhibits the increase (Fig. 2 A). To determine whether the response to wortmannin seen in HEK293 cells solely is because of inhibition of ATM, we repeated the experiment in AT cells (Fig. 2B), which lack functional ATM. In AT cells, raising the osmolality from 300 to 500 mosmol/kg by adding NaCl increases TonEBP/OREBP transcriptional activity far less than in HEK293 cells, but wortmannin at 1 or 20 μM still significantly inhibits the response, indicating the involvement of kinases in addition to ATM. Activation of TonEBP/OREBP by hypertonicity increases transcription and mRNA abundance of BGT1 (22). Accordingly, raising osmolality from 300 to 500 mosmol/kg by adding NaCl for 16 h greatly increases BGT1 mRNA abundance in HEK293 cells (Fig. 2C). A total of 1 or 20 μM wortmannin significantly inhibits the increase (Fig. 2C). These results are consistent with a role for ATM in hypertonicity-induced activation of TonEBP/OREBP. However, because wortmannin also inhibits other phosphatidylinositol kinase kinases, including DNA-PK and phosphatidylinositol 3-kinase (23, 24), additional experiments were required to confirm the role of ATM.

Fig. 2.

Wortmannin inhibits TonEBP/OREBP transcriptional activity and BGT1 mRNA abundance when NaCl is high. HEK293 cells (A) or AT cells (B) were preincubated with wortmannin for 1 h at 300 mosmol/kg before increasing osmolality to 500 mosmol/kg by adding NaCl. ORE-X reporter activity was measured 16 h later. Values are normalized to control (DMSO) cells at 300 mosmol/kg. (C) HEK293 cells were preincubated with DMSO or wortmannin for 1 h at 300 mosmol/kg before increasing osmolality to 500 mosmol/kg by adding NaCl. BGT1 mRNA was measured 16 h later by real-time RT-PCR. Values are normalized to control cells at 300 mosmol/kg. Mean ± SEM (*, P ≤ 0.05, n ≥ 3).

Overexpression of TonEBP/OREBP increases ORE/TonE reporter activity at 300 mosmol/kg (20, 25), but the increase is small compared with the increase produced by hypertonicity. Cotransfection of wild-type TonEBP/OREBP expression plasmid increases ORE/TonE reporter activity at 300 mosmol/kg by 39-fold, which is much less than the 12,019-fold increase produced by adding NaCl to raise osmolality to 500 mosmol/kg (Fig. 3A). In contrast, when TonEBP/OREBP activation is already greatly increased at 500 mosmol/kg, overexpression of wild-type TonEBP/OREBP increases the reporter activity only by an additional 20% (Fig. 3A). TonEBP/OREBP contains three putative ATM phosphorylation sites (motif scanner, available at http://scansite.mit.edu/motifscanner/motifscan2.html), namely Ser-1197, Ser-1247, and Ser-1367. To determine the importance of these serines for TonEBP/OREBP transcriptional activity, we cotransfected HEK293 cells with the luciferase reporter and with wild-type TonEBP/OREBP or TonEBP/OREBP S1197A, S1247A, or S1367A. At 300 mosmol/kg, each of these mutants activates reporter activity significantly less than wild-type TonEBP/OREBP (Fig. 3B). We conclude that these serines are important for transcriptional activity of TonEBP/OREBP because they are phosphorylated by ATM or some other kinase. Also, although phosphorylation of these serines may be less at 300 mosmol/kg than at 500 mosmol/kg, it is still sufficient to contribute to the relatively small increase in transcriptional activity produced by overexpression of recombinant TonEBP/OREBP and to explain why recombinant TonEBP/OREBPs in which these serines are mutated have less transcriptional activity.

Fig. 3.

(A–C) Activation of ORE/TonE reporter activity by overexpression of wild-type or mutated TonEBP/OREBP. Cells were cotransfected at 300 mosmol/kg with ORE-X reporter and with empty vector, wild-type TonEBP/OREBP 1-1531V5, or 1-1531V5-S1197A, 1-1531V5-S1247A, or 1-1531V5-S1367A. At 24 h later, osmolality was either kept at 300 mosmol/kg or was increased to 500 mosmol/kg (NaCl added) for 16 h before measuring reporter activity. (A) HEK293 cells. Values are normalized to those with empty vector at 300 mosmol/kg. (B) HEK293 cells. Values are normalized to those with wild-type TonEBP/OREBP at 300 mosmol/kg (Left) or 500 mosmol/kg (Right). (C) AT cells. Values are normalized to cells transfected with wild-type TonEBP/OREBP at 300 mosmol/kg (Left) or 500 mosmol/kg (Right). (D) Activity of ATM contributes to ORE/TonE reporter activity. AT cells were cotransfected with ORE-X reporter and empty vector, wild-type, or S1981A ATM at 300 mosmol/kg. At 24 h later, osmolality was either kept the same or increased to 500 mosmol/kg (NaCl added) for 16 h before measuring luciferase reporter activity. Values are normalized to cells kept at 300 mosmol/kg after transfection with empty vector. (E) ATM increases TonEBP/OREBP transactivating activity. To measure TonEBP/OREBP transactivating activity, AT cells were cotransfected with the GAL4 upstream activating sequence reporter and GAL4dbd-548–1531 plus empty vector, wild-type ATM or ATM S1981A. At 24 h later, osmolality was either kept the same or increased to 500 mosmol/kg (NaCl added) for 16 h before measuring luciferase reporter activity. Values are normalized to cells kept at 300 mosmol/kg after transfection with empty vector. Mean ± SEM (*, P ≤ 0.05, n ≥ 3).

To distinguish whether it is ATM or some other kinase that is responsible for these effects, we repeated the experiments in cells that lack ATM activity (AT cells) (Fig. 3C). In AT cells, transfection of wild-type TonEBP/OREBP expression plasmid increases ORE/TonE reporter activity at 300 mosmol/kg by 14 ± 1.2-fold, and raising osmolality to 500 mosmol/kg by adding NaCl increases reporter activity by 137 ± 21-fold. Both increases are less than in HEK293 cells. Further, in AT cells at 300 mosmol/kg, overexpression of TonEBP/OREBP S1247A and S1367A increase reporter activity as much as overexpression of wild-type TonEBP/OREBP (Fig. 3C), in marked contrast to their lesser effect in HEK293 cells (Fig. 3B). We suggest that transcriptional activity of TonEBP/OREBP depends in part on phosphorylation of TonEBP/OREBP Ser-1247 and Ser-1367 by ATM or an ATM-dependent kinase. On the other hand, results with TonEBP/OREBP S1197A in HEK293 cells, which express ATM (Fig. 3B), are similar to the results in AT cells, which do not (Fig. 3C), implying involvement of one or more kinases other than ATM at that site.

To test further the role of ATM in activation of TonEBP/OREBP, we tested the effect on TonEBP/OREBP transcriptional activity of reconstituting AT cells with empty vector, wild-type ATM, or ATM S1981A, a mutation that prevents activation of ATM kinase activity (13). Wild-type ATM increases TonEBP/OREBP transcriptional activity both at 300 and 500 mosmol/kg, whereas ATMS1981A does not (Fig. 3D). We conclude that ATM regulates TonEBP/OREBP transcriptional activity.

We next determined the effect of ATM on transactivating activity of TonEBP/OREBP. TonEBP/OREBP contains a functional transactivation domain within amino acids 548–1,531 (8, 26). To measure transactivating activity, we cotransfected AT cells with the GAL4 upstream activating sequence reporter and GAL4dbd-548–1531 plus empty vector, wild-type ATM, or ATM S1981A. Reconstitution of the AT cells with wild-type ATM increases TonEBP/OREBP transactivating activity both at 300 and 500 mosmol/kg, but the inactive mutant, ATM S1981A, does not (Fig. 3E). We conclude that ATM regulates transactivating activity of TonEBP/OREBP.

Finally, to test whether ATM physically associates with TonEBP/OREBP, we immunoprecipitated proteins from HEK293 cells that were transfected with 1-1531V5, then immunoblotted with anti-TonEBP/OREBP or anti-ATM antibodies. Anti-V5 antibody coimmunoprecipitates TonEBP/OREBP and ATM from cells transfected with 1-1531V5 but not from cells transfected with empty vector both at 300 and 500 mosmol/kg (Fig. 4A Upper). Similarly, anti-ATM antibody coimmunoprecipitates TonEBP/OREBP and ATM both at 300 and 500 mosmol/kg (Fig. 4A Lower). The weaker TonEBP/OREBP band that coimmunoprecipitates from cells transfected with empty vector represents native TonEBP/OREBP, which evidently is less abundant than the overexpressed recombinant TonEBP/OREBP (Fig. 4A Lower). As a control, nonspecific IgG does not immunoprecipitate either TonEBP/OREBP or ATM (Fig. 4A). To determine whether ATM is a member of the protein complex associated with TonEBP/OREBP bound to ORE/TonE DNA elements, we performed electrophoretic mobility-shift assay binding reactions with and without anti-ATM. Addition of anti-ATM to the binding reaction results in a supershift (Fig. 4B). Addition of anti-TonEBP/OREBP results in a similar supershift (Fig. 4B). The mobility shift is specific. Only the TonEBP/OREBP-shifted band is eliminated by addition of 10-fold molar excess nonbiotinylated competitor in the absence of antibody (Fig. 4B, lane 5) or in the presence of anti-ATM (Fig. 4B, lane 6). We conclude that TonEBP/OREBP and ATM are physically associated in HEK293 cells and that this association is maintained in vitro when TonEBP/OREBP binds to its cognate DNA element, ORE/TonE.

Fig. 4.

TonEBP/OREBP, ATM, and PKAc are physically associated. (A) TonEBP/OREBP, ATM, and PKAc coimmunoprecipitate. HEK293 cells were transfected with 1–1531V5 or empty vector at 300 mosmol/kg. At 24 h later, osmolality was either kept the same or increased to 500 mosmol/kg (NaCl added) for 2 h. Cell extracts were immunoprecipitated with mouse IgG, mouse anti-V5 (Upper), or mouse anti-ATM antibody (Lower) and immunoblotted with anti-TonEBP/OREBP, ATM, or PKAc antibody (representative blots, n ≥ 3). (B) Anti-ATM supershifts TonEBP/OREBP complexed to the cognate ORE/TonE DNA element. Electrophoretic mobility-shift assay binding reactions were performed with nuclear extracts of HEK293 cells in 500 mosmol/kg medium (NaCl added, 2 h), by using an ORE/TonE probe containing three OREs in native gene context (lanes 1–6). Additions to reactions are anti-ATM, anti-TonEBP/OREBP, and/or 10× molar excess of nonbiotinylated probe (representative blot, n ≥ 3).

We previously found that PKAc is involved in osmotic regulation of transcriptional activity of TonEBP/OREBP and physically associates with TonEBP/OREBP (11). In agreement, PKAc coimmunoprecipitates both with TonEBP/OREBP (Fig. 4A Upper) and ATM (Fig. 4A Lower). Evidently, all are part of the same large transcriptional complex.

Discussion

High NaCl Damages DNA. Acute elevation of NaCl causes DNA double-strand breaks in cell culture (12), and the DNA breaks persist even after the cells are adapted to high NaCl (27). In a similar fashion, renal inner medullary cells in vivo are normally exposed to a variable, but always high, level of NaCl (1). In agreement with the results in cell culture, inner medullary cells in normal mice also exhibit numerous DNA breaks (27). Thus, both in cell culture and in vivo, high NaCl causes persistent DNA damage.

High NaCl Activates ATM. ATM protein kinase, mutations of which are associated with the human disease ataxia-telangiectasia, mediates responses to ionizing radiation-induced DNA damage in mammalian cells (13). ATM is held inactive in unirradiated cells as a dimer or higher-order multimer, with the kinase domain bound to a region surrounding serine 1981. Ionizing radiation induces rapid intermolecular autophosphorylation of serine 1981, resulting in dimer dissociation and initiation of ATM kinase activity. We now find that another DNA damaging agent, high NaCl, also activates ATM (Fig. 1 A). Many aspects of the DNA damage response are inhibited by high NaCl (28), but ATM evidently is exceptional in this regard. Thus, ionizing radiation causes phosphorylation of H2AX and assembly at sites of DNA damage of a complex that includes Mre11 and phosphorylated H2AX, yet this does not occur when DNA is damaged by high NaCl (28). A possible explanation for activation of ATM by high NaCl, independent of the other parts of the DNA damage response, is provided by the proposed special nature of the link between DNA damage and ATM activation. The introduction of even a few DNA double-strand breaks, by removing topological constraints on DNA loops, apparently causes a rapid change in some aspect of higher-order chromatin structure that initiates ATM activation (13). Evidence for this model is the rapidity of ATM activation and its occurrence at a distance from the DNA strand breaks. In support of the model, ATM is also activated in the absence of detectable DNA strand breaks when chromatin structure is altered by hypotonic swelling or by treatment with chloroquine or trichostatin. High NaCl may also perturb chromatin structure (29), which could explain its activation of ATM.

ATM Is Activated by High NaCl, Contributing to Activation of TonEBP/OREBP. In cells lacking functional ATM (AT cells), expression of wild-type ATM increases transcriptional (Fig. 3D) and transactivating activity (Fig. 3E) of TonEBP/OREBP at 500 mosmol/kg, but expression of inactive ATM (S1981A) does not. Thus, ATM contributes to activation of TonEBP/OREBP by high NaCl.

There are putative ATM phosphorylation sites in TonEBP/OREBP at amino acids Ser-1197, Ser-1247, and Ser-1367. To test whether these sites might be involved in activation of TonEBP/OREBP by ATM, we examined the effect of overexpression TonEBP/OREBP, either wild type or mutated at the sites. Overexpression of wild-type TonEBP/OREBP in HEK293 cells at 300 mosmol/kg increases transcriptional reporter activity (Fig. 3A), but TonEBP/OREBP S1197A, S1247A, and S1367A have much less effect (Fig. 3B), making it plausible that phosphorylation of TonEBP/OREBP by ATM at any of these sites might contribute to its activation. At 500 mosmol/kg TonEBP/OREBP transcriptional activity is already greatly increased by the high NaCl, and the additional effect of overexpressing wild-type TonEBP/OREBP is small (Fig. 3A), but the mutants still have slightly less effect than wild type (Fig. 3B).

Any transcriptional activity of the overexpressed TonEBP/OREBP that depends on ATM should be lacking in AT cells. In AT cells, transcriptional activity of TonEBP/OREBP S1247A equals that of wild-type TonEBP/OREBP (Fig. 3C), in marked contrast to its lesser effect in HEK293 (Fig. 3B) cells. The most likely explanation is that ATM contributes to activation of TonEBP/OREBP by phosphorylating it at Ser-1247.

On the other hand, in AT cells, overexpression of TonEBP/OREBP S1197A activates transcription of the ORE/TonE reporter less than overexpression of wild type (Fig. 3C), suggesting that a kinase other than ATM activates at this site. A likely candidate is DNA-PK, whose consensus phosphorylation motif is the same as for ATM (motif scanner).

Finally, the results with TonEBP/OREBP S1367A in AT cells are equivocal. At 300 mosmol/kg, overexpression of TonEBP/OREBP S1367A activates transcription of the ORE/TonE reporter as much as overexpression of wild-type TonEBP/OREBP (Fig. 3C), not less, as in HEK293 cells (Fig. 3B), consistent with phosphorylation at this site by ATM being important for activation of TonEBP/OREBP. However, at 500 mosmol/kg, even in AT cells, TonEBP/OREBP S1367A has slightly but significantly less transcriptional activity than wild type (Fig. 3C), consistent with a role for some additional kinase at this site.

Convergence of Multiple Pathways Is Necessary for High NaCl to Signal Activation of TonEBP/OREBP. ATM is activated not only by high NaCl but also by high urea, UV radiation and ionizing radiation (Fig. 1 A), and low NaCl (13). However, although high NaCl increases activity of a TonEBP/OREBP reporter, the other stresses do not (Fig. 1B) (8). Therefore, activation of ATM increases TonEBP/OREBP reporter activity but is not sufficient to induce it.

Other kinases also are activated by high NaCl and contribute to activation of TonEBP/OREBP. These include PKAc (11), p38 (9, 10), and Fyn (10). Like ATM, each of them is necessary for full activation of TonEBP/OREBP transcriptional activity by high NaCl, but inhibition of any of them, singly, does not fully prevent the activation.

TonEBP/OREBP is part of a large transcriptional complex (4). Based on mutual coimmunoprecipitation (Fig. 4A) and the ability of anti-ATM to supershift TonEBP/OREBP bound to its cognate DNA element ORE/TonE (Fig. 4B), ATM is part of the complex, as is PKAc (Fig. 4A) (11). It is not clear whether p38 or Fyn is also part of the complex or whether it participates by phosphorylating some other member of the complex or some upstream signaling molecule.

Further evidence that a multiplicity of hypertonicity-induced pathways signals to TonEBP/OREBP comes from the effects of wortmannin. Wortmannin inhibits ATM, which presumably contributes to its inhibition of high-NaCl-induced transcriptional activity of TonEBP/OREBP (Fig. 2 A). However, wortmannin also inhibits high NaCl-induced activation of the ORE/TonE reporter in AT cells (Fig. 2B), an effect that must involve some other kinase. Wortmannin generally inhibits phosphatidylinositol kinase family kinases, namely phosphatidylinositol 3-kinase and phosphatidylinositol kinase-related kinases, including ATM and DNA-PK (23, 24). Sensitivity to wortmannin differs among the phosphatidylinositol kinase family kinases. phosphatidylinositol 3-kinase is most sensitive, IC₅₀ being 3.0 nM wortmannin in CTLL-2 cells (24). In contrast, IC₅₀ for ATM and DNA-PK in A549 cells is 5.8 and 3.6 μM, respectively (23). We conclude that the signaling molecules in addition to ATM that are involved in hypertonicity-induced signaling to TonEBP/OREBP, probably include other phosphatidylinositol kinase family members.