Dominant-Negative c-Jun NH2-terminal Kinase 2 Sensitizes Renal Inner Medullary Collecting Duct Cells to Hypertonicity-induced Lethality Independent of Organic Osmolyte Transport*

The c-Jun NH2-terminal protein kinases (JNKs), as well as the extracellular signal-regulated protein kinases (ERKs) and p38 mitogen-activated protein kinase, are activated in renal cells in response to extracellular hypertonicity. To determine whether activation of JNKs by hypertonicity is isoform-specific, renal inner medullary collecting duct cells were stably transfected with cDNA’s encoding hemagglutinin (HA)-tagged JNK1 and JNK2 isoforms, and the expressed kinases were immunoprecipitated with an anti-HA antibody. Whereas both recombinant kinases were equivalently expressed, only immunoprecipitates from the HA-JNK2 cells displayed hypertonicity-inducible JNK activity. Furthermore, expression of dominant-negative JNK2 (HA-JNK2-APF) in stable clones inhibited hypertonicity-induced JNK activation by 40–70%, whereas expression of dominant-negative JNK1 (HA-JNK1-APF) had no significant inhibitory effect. Independent HA-JNK2-APF (but not HA-JNK1-APF) clones displayed greatly reduced viability relative to neomycin controls after 16 h of exposure to 600 mosm/kg hypertonic medium with percent survival of 20.5 ± 2.7 and 31.5 ± 7.3 for two independent HA-JNK2-APF clones compared with 80.1 ± 1.0 for neomycin controls (p < 0.001, n = 5, mean ± S.E.). However, neither JNK mutant blocked either regulatory volume increase or hypertonicity-induced enhancement of uptake of inositol, an organic osmolyte putatively involved in long term adaptation to hypertonicity. These results define JNK2 as the primary hypertonicity-activated JNK isoform in IMCD-3 cells and demonstrate its central importance in cellular survival in a hypertonic environment by a mechanism independent of acute regulatory volume increase as well as regulation of organic osmolyte uptake.

The cells of the inner medulla of the mammalian nephron are uniquely exposed to large fluctuations in extracellular tonicity due to the changes that occur during diuretic and antidiuretic states. One means of adaptation to a hypertonic environment that has been well described is the intracellular accumulation of "non-perturbing" osmolytes that occurs either by uptake via sodium-coupled transporters (1) or by the generation of sorbitol through the action of aldose reductase on glucose (2). This process involves increased transcription of transporter genes (3,4) mediated by the osmotic response element that resides in the promoter region of these genes (5). The signaling pathways that impinge upon the osmotic response element remain undefined.
Transcriptional regulation is often mediated by mitogenactivated protein (MAP) 1 kinase pathways (6), which are stimulated by diverse extracellular signals including hypertonicity (7,8). In this regard, cells of renal origin display osmotic activation of multiple members of the MAP kinase family, including the extracellular signal-regulated kinases (ERKs), c-Jun NH 2 -terminal kinases (JNKs), and p38 MAP kinases (9 -11). However, we have recently shown that induction of osmolyte transport by hypertonicity is not significantly impacted by pharmacologic inhibition of the ERK pathways (11). The marked activation of the JNKs and more modest activation of p38 MAP kinase (11), which is the mammalian counterpart of the osmoregulated HOG-1 in yeast (12), highlights the JNK pathway as being of significance with regard to osmoregulation in the kidney. However, the existence of three distinct JNK genes with several splice variants produced from each gene hampers the clear dissection of the role of JNK in osmoregulation in eukaryotic cells, since different isoforms may play different physiological roles. For example, in small cell lung cancer cells both JNK1 and JNK2 are activated by exposure to UV light but only the JNK1 isoform appears to regulate UV-induced apoptosis (13). Additionally, whereas JNK activation has been linked to induction of apoptosis in some cell types (14), its physiological role in the cellular response to osmotic challenge has not yet been elucidated. The present study was therefore undertaken to define the JNK isoforms regulated by hypertonicity in renal inner medullary collecting duct cells and to determine whether such activation plays a role in cellular adaptation to a hypertonic environment.

EXPERIMENTAL PROCEDURES
Materials-Cell culture media and serum were from Life Technologies, Inc. Recombinant GST-c-Jun(1-79) and ATF-2NT(1-254) were expressed in Escherichia coli and purified using glutathione-agarose (Sigma) and Ni ϩ -nitrilotriacetic acid-agarose (Qiagen, Studio City, CA), respectively, as described previously (15). Anti-JNK antisera were from Santa Cruz Biotechnology. Radioisotopes were from NEN Life Science Products. The osmolarity of all solutions used was checked with an Advanced Instruments Model 3MO Micro-Osmometer.
Cell Culture-The established inner medullary collecting duct cell line IMCD-3 is an immortalized line generously provided by Dr. Steve Gullans (Boston, MA) (16). The cells were routinely propagated in a 1:1 mixture of Dulbecco's modified Eagle's medium (DMEM) and Ham's F-12 nutrient mixture, supplemented with 10% fetal bovine serum, 100 U/ml penicillin, and 100 g/ml streptomycin. * This work was supported by National Institutes of Health Grants DK-19928 and GM-48826. 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.
Generation of JNK1 and JNK2 Stable Transfectants-The HA-JNK1, HA-JNK1-APF, HA-JNK2, and HA-JNK2-APF cDNAs (17)(18)(19) were ligated between the HindIII and HpaI sites of the retroviral vector pLNCX (20) and packaged into replication-defective retrovirus using 293T cells and the retrovirus component expression plasmids SV-Ϫ -A-MLV and SV-Ϫ -env Ϫ -MLV as described previously (21,22). IMCD-3 cells were cultured for 24 h in virus-containing conditioned medium that had been filtered through a 0.45 filter and supplemented with 8 g/ml polybrene. Positive infectants were selected in 500 g/ml G418, cloned, and further characterized by Western blots using anti-HA antisera, as well as by kinase activity assay (see below). Experiments were performed on cells that had been passaged less than 6 times.
Assay of JNK Kinase Activity by Immunoprecipitation-To determine activity of transfected HA-JNK's, IMCD-3 cells in 100-mm tissue culture dishes were incubated at 37°C for 10 min in DMEM:F-12 medium alone or medium supplemented with 150 mM NaCl (600 mosM final). They were then washed three times in isosmotic phosphatebuffered saline, and lysed in 0.2-0.5 ml of lysis buffer (50 mM ␤-glycerophosphate (pH 7.2), 0.5% Triton X-100, 0.1 mM sodium vanadate, 2 mM MgCl 2 , 1 mM EGTA, 1 mM dithiothreitol, 2 g/ml leupeptin, 4 g/ml aprotinin). The lysate was centrifuged at 4°C for 10 min (10,000 ϫ g) to remove nuclei and cell debris, and the supernatants were adjusted to 100 -200 g of protein in 0.5 ml, to which was added 5 l of mouse monoclonal antiserum directed against the influenza HA epitope (Boehringer Mannheim), and 100 l of 10% protein G-Sepharose (Pharmacia). After 2 h of rocking incubation at 4°C, the immunoprecipitates were washed three times in lysis buffer and resuspended in 40 l of 50 mM ␤-glycerophosphate (pH 7.2), 0.1 mM sodium vanadate, 20 mM MgCl 2 , 200 M [␥-32 P]ATP (5000 cpm/pmol), and 100 g/ml recombinant NH 2 -terminal domain of ATF-2 (ATF-2NT) (6) for 20 min at 30°C. The reaction was stopped by the addition of SDS sample buffer, and the lysates were heated in a boiling water bath for 5 min and subsequently subjected to SDS-polyacrylamide gel electrophoresis on a 10% polyacrylamide gel and followed by autoradiography. The bands corresponding to phosphorylated ATF-2NT were excised and counted in a liquid scintillation counter.
Assessment of kinase inhibition by HA-JNK-APF constructs was carried out similarly, except that 100 l of 10% GST-c-Jun-agarose beads were used in place of anti-HA antiserum and ATF-2 peptide, and the final ATP concentration was 20 M.
Assay by Mono Q Fast Performance Liquid Chromatography of Dominant-Negative JNK Inhibition of Endogenous JNK Activity-JNK activity was also assessed following fractionation on Mono Q fast performance liquid chromatography as described previously (11). Portions of cell lysates prepared as described above (0.5 ml, 1.0 -1.5 mg of protein) were applied to a Pharmacia HR5/5 Mono Q anion exchange column equilibrated in 50 mM ␤-glycerophosphate (pH 7.2), 0.1 mM sodium vanadate, 1 mM EGTA, and 1 mM dithiothreitol, and eluted with a 30-ml gradient of 0 -600 mM NaCl in the same buffer. Fractions (1 ml) were collected, and 20-l aliquots were mixed with 20 ml of 50 mM ␤-glycerophosphate (pH 7.2), 0.1 mM sodium vanadate, 20 mM MgCl 2 , 200 mM [␥-32 P]ATP (5,000 counts/min Ϫ1 pmol Ϫ1 ), 50 mg/ml IP-20 (TTYADFIAS-GRTGRRNAIHD), and 100 g/ml recombinant ATF-2NT. The IP-20, a protein kinase A inhibitor, is routinely included in all protein kinase assays. The kinase reactions were incubated for 30 min at 30°C and terminated with 10 l of SDS sample buffer, and ATF-2NT phosphorylation was assessed following SDS-polyacrylamide gel electrophoresis by liquid scintillation counting of the excised ATF-2NT bands.
Determination of Hypertonicity-induced Cell Lethality-Subconfluent IMCD-3 cells on 12-well tissue culture plates were treated for 16 h with 1 ml of DMEM:F-12 medium or medium to which sufficient NaCl had been added to bring final osmolality to 400, 500, or 600 mosM/kg. Supernatants were collected, and remaining adherent cells were lifted in Hank's balanced salt solution containing 5 mg/ml porcine trypsin and an appropriate amount of NaCl so that the trypsin solution was isotonic with the treatment medium. Cell suspensions were washed once with the isotonic Hank's balanced salt solution and mixed 1:1 (v/v) with 0.4% trypan blue immediately before counting on a hemocytometer.
Assay of Inositol Uptake-Inositol uptake was measured as in Veis et al. (23). Briefly, IMCD-3 cells were treated for 16 h with inositol-free DMEM (Life Technologies, Inc.) supplemented with 1% bovine serum albumin, fraction V (U. S. Biochemical Corp.), after which 5 l of [ 3 H]inositol (specific activity 20 Ci/mmol) was added into 0.5 ml of medium for 2 h. The cells were then vigorously washed three times in isotonic phosphate-buffered saline and lysed in 100 l lysis buffer. The lysate was centrifuged at 10,000 ϫ g, and separate aliquots of the supernatant were either counted in a liquid scintillation counter or analyzed for protein concentration.
Measurement of Mean Cell Volume-Cell volume regulation was studied by observing the mean cell volume at various times after hypertonic treatment. Approximately 10 million cells were trypsinized, resuspended in DMEM to inactivate trypsin, centrifuged for 1 min at 2,000 ϫ g, and resuspended in 4 ml of buffer E, consisting of 10 mM HEPES (pH 7.3), 140 mM NaCl, 4 mM KCl, 0.5 mM CaCl 2 , 2 mM MgCl 2 , 1 mM KH 2 PO 4 , and 300 mM glucose and having an osmolarity of approximately 300 mosM. After 30 min of equilibration, sufficient NaCl was added to bring the osmolarity to 600 mosM/kg, and the mean cell volume of 10,000 cells per time point was measured using a Coulter Multisizer, Coulter Sample Stand II, and Multisizer AccuComp version 1.19 software, utilizing an aperture tube diameter of 100 m. Changes in cell volume over time are expressed as relative volume by normalizing to mean cell volume measurements taken on cells before addition of hypertonic NaCl.

RESULTS AND DISCUSSION
Selective Osmotic Regulation of JNK2-To ascertain whether the activation of JNK by hypertonicity is isoform specific, we prepared IMCD-3 cell clones stably expressing HAtagged JNK1 and JNK2. The expression of both HA-tagged JNK isoforms was confirmed by Western blot (Fig. 1A). However, whereas neither the neomycin control (devoid of HAtagged JNK) nor HA-JNK1 was stimulated by hypertonicity, the HA-JNK2 showed a consistent activation in each of three experiments; one is shown in Fig. 1B. A similar specificity was noted for activation of HA-JNK2 and not HA-JNK1 by UV light. 2 Inhibition of JNK2 Signaling Sensitizes IMCD-3 Cells to Hypertonicity-induced Lethality-To characterize the role of 2 P. Wojtaszek and T. Berl, unpublished observations.

FIG. 1. JNK2 but not JNK1 is activated in IMCD-3 cells by hypertonicity. A, IMCD-3 cells expressing epitope-tagged JNK1 and
JNK2 were generated by retrovirus-mediated gene transfer. Cell lysates were subjected to SDS-polyacrylamide gel electrophoresis and Western blot analysis with 12CA5 antibody. B, kinase activity in 12CA5 immunoprecipitates from cells expressing epitope-tagged JNK1 or JNK2 treated for 10 min in either 300 or 600 mosM/kg medium was measured using ATF-2NT peptide as substrate (see "Experimental Procedures"). Representative data from three individual experiments is shown. WT, wild type. JNK2 in response to osmotic stress, we developed stably transfected IMCD-3 cell clones expressing nonphosphorylatable mutants of JNK1 (JNK1-APF) and JNK2 (JNK2-APF). In these mutants, the phosphorylation site, Thr-Pro-Tyr (TPY), is altered to Ala-Pro-Phe (APF), rendering the expressed kinase incapable of being phosphorylated, and hence incapable of being activated (18,24). These constructs are predicted to behave as competitive inhibitors of activation of cellular JNK1 and JNK2, respectively, by competing for binding of upstream MAP kinase kinases which normally phosphorylate and thereby activate the JNKs. Fig. 2A shows the expression of these constructs in two clones each of HA-JNK1-APF and HA-JNK2-APF.To demonstrate whether these constructs inhibit hypertonicity-stimulated JNK activity, GST-c-Jun was employed in JNK activity assays of lysates of cells exposed to hypertonicity. Utilization of this substrate measures activity of all JNK isoforms.
Activation of GST-c-Jun phosphorylation in HA-JNK1-APF clones was not significantly different from that of the neomycin controls (Fig. 2B), which is the predicted result if JNK1 is not osmotically regulated, as the results in Fig. 1 suggest. In contrast, HA-JNK2-APF clones consistently displayed a significant inhibition of hypertonicity-induced GST-c-Jun phosphorylation. Two independent stable HA-JNK2-APF clones displayed mean inhibitions of 42 Ϯ 15% and 48 Ϯ 17%, respectively, as illustrated in Fig. 2B (n ϭ 3, mean Ϯ S.E.).
It was important to confirm that the apparent inhibition of JNK activity observed by immunoprecipitation truly reflected inhibition of endogenous JNK activity and was not simply the result of an excessive amount of dominant-negative JNK saturating the GST-c-Jun beads used in the JNK binding assay. Therefore, lysates from cells subject to hypertonicity were also fractionated by Mono Q fast performance anion exchange liquid chromatography, and the JNK activity in the fractions was measured. Consistent with the percent inhibition by JNK2-APF observed in the GST-c-Jun binding assay, Fig. 2C shows a 50% inhibition of hypertonicity-stimulated ATF-2NT phosphorylation in the JNK2-APF clone 9, but not in the neomycin control or the JNK1-APF clone 5.
These experiments thus support the above observation that JNK2 is responsible for the majority of hypertonicity-induced JNK activity. If JNK1 activation played a significant role in this response, it would be expected that HA-JNK1-APF would at least partially block phosphorylation of GST-c-Jun in lysates from hypertonically exposed cells. However, only HA-JNK2-APF expression resulted in inhibition of hypertonicity-stimulated phosphorylation of GST-c-Jun, suggesting that JNK2 is the primary source of hypertonicity-stimulated JNK activity. Interestingly, in yeast lacking HOG1 MAP kinase, which is required for growth on hypertonic medium, expression of JNK1 but not JNK2 rescues hypertonic growth (19). This suggests that functions of different JNK isoforms may not be dictated by amino acid sequence similarity per se but rather by another factor such as differential affinity for specialized upstream MAP kinase kinases or downstream targets.
To determine if JNKs play a protective role in cellular adaptation to hypertonicity, we investigated the possibility that the presence of the dominant-negative JNK could impact on cell survival after osmotic challenge. Fig. 3 summarizes the survival data in which we employed two distinct stable clones each of HA-JNK1-APF and HA-JNK2-APF and assayed cell survival after hypertonic challenge utilizing trypan blue exclusion as- say. Both JNK2-APF clones displayed decreased survival relative to neomycin controls and to JNK1-APF clones, which were not significantly different from controls. After 16 h at 500 mosM/kg, the two HA-JNK2-APF clones displayed percent survival values of 62.8 Ϯ 3.1 and 68.2 Ϯ 1.9 compared with 82.6 Ϯ 2.8 for neomycin controls (p Ͻ 0.01, n ϭ 5, mean Ϯ S.E.), and at 600 mosM/kg, 20.5 Ϯ 2.7 and 31.5 Ϯ 7.3 compared with 80.1 Ϯ 1.0 for neomycin controls (p Ͻ .001, n ϭ 5, mean Ϯ S.E.). It is also of note that inhibition of the ERK pathway with PD098059 or of the p38 MAP kinase pathway with SB203580 did not alter the ability of cells to survive osmotic challenge. 2 Finally, staining with acridine orange:ethidium bromide revealed that the process leading to cell death was clearly necrosis and not apoptosis. 3 The adaptive response to hypertonicity involves increases in cellular content of organic osmolyte transporters leading to enhancement in the uptake of solute (23). We tested whether dominant-negative JNK2 may confer sensitivity to hypertonic challenge by interfering with hypertonicity-stimulated organic osmolyte uptake. However, HA-JNK2-APF stable transfectants displayed initial rates of inositol uptake after 16 h in a hypertonic environment at a level comparable to that of neomycin controls (Fig. 4). These studies were done at 500 mosM/kg in view of the sensitivity of the transfected cells to prolonged incubations at higher tonicity (see above and Fig. 3). Furthermore, inositol uptake measurements were normalized per milligram of cell protein, and the culture dishes were vigorously washed three times with isotonic buffer before lysis to avoid interference by protein from nonviable cells in the protein determination assay. These results do not entirely rule out the role for JNK activation in initiating the transcription of sodium myo-inositol transporter genes, as the residual JNK activity (Fig. 2, B and C) could be sufficient to initiate the process. However, since the degree of inhibition of cellular JNK2 activity in these mutants is sufficient to increase cell lethality after hypertonic challenge both at 500 and 600 mosM/kg, it would be predicted that if a primary function of JNK2 was to activate inositol uptake, some decrement in inositol uptake would be observable in these experiments.
In addition to these physiological measurements, Kultz et al. (25) have demonstrated that dominant-negative mutants of JNK1 and of p38 MAP kinase fail to block activation of an osmotic response element reporter gene construct when transiently transfected into renal PAP-HT25 cells. Our experiments complement these observations, since failure to observe transcriptional activation of the osmotic response element did not exclude the possibility that JNKs act at a site distal to transcription of transporter genes to modulate osmolyte uptake. By examining the final step in sodium myo-inositol trans-

FIG. 3. Expression of JNK2-APF but not JNK1-APF sensitizes IMCD-3 cells to hypertonicity-induced lethality.
Subconfluent cultures of stable transfectants were treated for 16 h in control medium (300 mosM/kg) or medium supplemented with sufficient NaCl for final osmolarities of 400, 500, and 600 mosM/kg. Non-trypan blue-staining cells were then counted. The graph shows percent survival and mean Ϯ S.E. for n ϭ 5 at each osmolarity. Statistical analysis by analysis of variance and Tukey-Kramer multiple comparisons test indicates that neither JNK1-APF clone is significantly different from controls, but both JNK2-APF clones are significantly different from controls at 500 mosM (p Ͻ 0.01) and at 600 mosM (p Ͻ 0.001).
FIG. 4. Expression of neither JNK1-APF nor JNK2-APF blocks hypertonicity-stimulated increase in inositol uptake. Subconfluent cultures were incubated for 16 h in inositol-free control medium (300 mosM/kg) or hypertonic medium (500 mosM/kg) after which [ 3 H]inositol was added to the medium for 2 h to measure initial rates of uptake, as in Veis et al. (23). The cells were then thoroughly washed in phosphate-buffered saline and lysed in lysis buffer. The lysate was centrifuged at 10,000 ϫ g, and the supernatant was counted in a liquid scintillation counter. Data shown are mean Ϯ S.E. (n ϭ 3).
FIG. 5. Expression of JNK2-APF does not inhibit regulatory volume increase after a 600 mosM/kg challenge. Cells were gently harvested and subjected to hypertonic challenge over a 30 min period as described under "Experimental Procedures." Mean cell volume of 10,000 cells per data point was monitored using a Coulter Multisizer, and data are expressed as percent of mean cell volume of samples before additional NaCl was added to bring the osmolarity to 600 mosM/kg. port, the present experiments suggest that the JNK pathway is not involved in regulation at any step in the process.
Another characteristic of cell adaptation to hypertonicity is the regulatory volume increase, by which a cell recovers from osmotic shrinkage and regains the volume it displayed before hypertonic treatment. To determine if the JNK2 pathway might contribute to this process, control and JNK-APF cells were subjected to 600 mosM/kg, and their cell volume was monitored with a Coulter Multisizer. Fig. 5 shows that both neomycin and JNK2-APF cells regained their original volume by 20 min, and that the JNK2-APF cells were not significantly different from controls in either time course or extent of regulatory volume increase. Similar data were obtained with the JNK1-APF clones, as well as with neomycin, JNK1-APF, and JNK2-APF cells treated with the ERK inhibitor PD098059 or the p38 kinase inhibitor SB203580. 2 These data suggest that the mechanism by which JNK2-APF sensitizes cells to killing by hypertonicity does not involve the cell's ability to carry out regulatory volume increase, and in addition, that neither the JNK, ERK, or p38 pathways play a significant role in this process.
Whereas these experiments address primarily volume-regulatory mechanisms by which a cell responds to hypertonic stress, it is likely that other systems also play a role in survival of hypertonicity. For example, several groups (26,27) have described members of the heat shock protein 70 superfamily that are up-regulated by exposure to hypertonic NaCl. The hypothesis that these heat shock proteins carry out an osmoprotective function, and that the JNK2 pathway may be involved in their regulation, is presently under investigation. Alternatively, JNK2 may activate as yet unidentified factors that stabilize the cellular machinery to the intracellular ionic changes that occur during osmoregulation.
In summary, we have for the first time determined that osmotic activation of JNK activity in renal inner medullary collecting duct cells is isoform-specific, with the major regulated form identified as JNK2. We have also dissociated JNK2 from the adaptive response that recruits sodium myo-inositol transporters and from the ability to carry out regulatory volume increase, but have defined a crucial role for JNK2 in the survival of inner medullary collecting duct cells in a hypertonic environment.