|Home | About | Journals | Submit | Contact Us | Français|
Previous work from this laboratory demonstrated that arachidonic acid activates c-jun NH2-terminal kinase (JNK) through oxidative intermediates in a Ca2+-independent manner (Cui X and Douglas JG. Arachidonic acid activates c-jun N-terminal kinase through NADPH oxidase in rabbit proximal tubular epithelial cells. Proc Natl Acad Sci USA 94: 3771–3776, 1997.). We now report that JNK can also be activated via a Ca2+-dependent mechanism by agents that increase the cytosolic Ca2+ concentration (Ca2+ ionophore A23187, Ca2+-ATPase inhibitor thapsigargin) or deplete intracellular Ca2+ stores [intracellular Ca2+ chelator 1,2-bis(2-aminophenoxy)ethane-N,N,N′,N′-tetraacetic acid (BAPTA)-AM]. The activation of JNK by BAPTA-AM occurs despite a decrease in cytosolic Ca2+ concentration as detected by the indicator dye fura 2, but appears to be related to Ca2+ metabolism, because modification of BAPTA with two methyl groups increases not only the chelation affinity for Ca2+, but also the potency for JNK activation. BAPTA-AM stimulates Ca2+ influx across the plasma membrane, and the resulting local Ca2+ increases are probably involved in activation of JNK because Ca2+ influx inhibitors (SKF-96365, nifedipine) and lowering of the free extracellular Ca2+ concentration with EGTA reduce the BAPTA-induced JNK activation.
The c-jun NH2-terminal kinase (JNK) is a member of the mitogen-activated protein kinase (MAPK) superfamily. We recently reported for rabbit proximal tubule cells that this enzyme is activated by arachidonic acid in a Ca2+-independent manner, even though cytosolic Ca2+ concentrations ([Ca2+]i) increased when arachidonic acid metabolites levels were raised (3, 18). The conclusion that Ca2+ was not an important player in the activation of JNK by arachidonic acid was based on the observations that the presence of intracellular and extracellular Ca2+ chelators did not change the arachidonic acid-dependent activation. However, in many other cell types, JNK is activated in a Ca2+-dependent manner (15, 21, 27, 31, 38). Therefore, the current experiments were designed to investigate the relationship between [Ca2+]i and JNK activation in rabbit proximal tubule cells. Interestingly, we observed that the intracellular Ca2+ chelator 1,2-bis(2-aminophenoxy)ethane-N,N,N′,N′-tetraacetic acid (BAPTA)-AM activated JNK, similar to known elevators of [Ca2+]i, such as the Ca2+ ionophore A-23187 and the Ca2+ pump inhibitor thapsigargin. BAPTA-AM activation of JNK is also largely Ca2+ dependent, although in a paradoxical way because it results from chelation of cytosolic Ca2+. This chelation causes store depletion and activation of “capacitative Ca2+ influx” (24, 32, 34) and, therefore, local increases of [Ca2+]i.
A23187, EGTA, thapsigargin, purified rabbit IgG, and anti-rabbit IgG agarose beads were purchased from Sigma Chemical (St. Louis, MO). Cell culture medium and additives were from GIBCO BRL (Gaithersburg, MD). BAPTA-AM, dimethyl BAPTA-AM, and fura 2-AM were from Molecular Probes (Eugene, OR). [γ-32P]ATP and [3H]arachidonic acid were obtained from DuPont-NEN (Boston, MA). was from ICN Biomedicals (Aurora, OH). Rabbit anti-JNK1(FL) polyclonal antibody, which cross reacts with all three isoforms of JNK, recombinant activating factor-2 (ATF-2), and horseradish peroxidase (HRP)-conjugated goat anti-rabbit antibody were purchased from Santa Cruz Biotechnology (Santa Cruz, CA). Rabbit polyclonal anti-phospho JNK, anti-phospho ATF-2, and anti-phospho c-jun antibodies were from New England Biolabs (Beverly, MA).
Proximal tubule epithelial cells were isolated from New Zealand White rabbits as previous described (18). They were maintained in modified DMEM:F-12 (1:1) media supplemented with 5% fetal calf serum (FCS), 5 μg/ml insulin, 5 μg/ml transferrin, 0.5 μM hydrocortisone, 350 μg/ml l-glutamine, 100 U/ml penicillin, and 100 μg/ml streptomycin. Subconfluent monolayers of first-passage cells were employed for experiments.
Cells were serum deprived for 16–18 h before any experiment. The kinase assay was conducted as previously described (3). In brief, after an experimental treatment, cells were washed twice with ice-cold Dulbecco's PBS and were lysed on ice by adding 0.3 ml of lysis buffer [50 mM Tris (pH 7.2), 1% (vol/vol) Triton X-100, 1 mM Na3VO4, 1 mM EGTA, 0.2 mM phenylmethylsulfonyl fluoride, 25 μg/ml leupeptin, and 10 μg/ml aprotinin]. The samples were centrifuged at 14,000 g for 10 min. Protein content in the supernatants was determined by the BCA Protein Assay according to the manufacturer's instructions (Pierce, Rockford, IL). Two hundred micrograms of lysate protein in a total volume of 800 μl was precleared with 1 μg of nonimmune rabbit IgG and 30 μl of goat anti-rabbit IgG agarose beads on a rotating plate for 1.5 h at 4°C. After centrifugation at 10,000 g for 10 min, the supernatant was mixed with 1 μg of anti-JNK1(FL) polyclonal antibody and 25 μl of goat anti-rabbit IgG agarose beads on a rotating plate overnight at 4°C. The beads were pelleted and washed twice with lysis buffer and once with the kinase assay buffer [20 mM HEPES, pH 7.6, 20 mM MgCl2, 25 mM β-glycerol phosphate, 0.1 mM Na3VO4, and 2 mM dithiothreitol]. The kinase assay was carried out at 30°C for 15 min in 30 μl of assay buffer containing 0.5 μg of ATF-2, 20 μM ATP, and 2 μCi of [γ-32P]ATP. The reaction was terminated by addition of Laemmli's sample buffer followed by boiling for 5 min. The samples were resolved by 10% SDS-PAGE followed by staining with Coomassie brilliant blue to check for protein loading. The gel was dried, and the incorporation of 32P was visualized by autoradiography. Gel slices of the 69-kDa ATF-2 bands were also cut out, and the radioactivity was measured by liquid scintillation counting.
Cell lysates were prepared as described above, following treatments indicated in the figure legends. Fifteen micrograms of total cell lysate protein were subjected to SDS-PAGE and then transferred to a polyvinylidene difluoride membrane by electroblotting at 200 mA for 1.5 h. The membrane was incubated overnight at 4°C with 5% nonfat milk dissolved in Tris-saline buffer containing 0.1% (vol/vol) Tween 20 (TTBS), followed by three washes with TTBS. The membrane was then incubated with primary antibodies overnight at 4°C, followed by six washes and another incubation with 1:2,000 dilution of HRP-conjugated goat anti-rabbit antibody at room temperature for 1 h. After extensive washes, the immunoreactive proteins were detected by enhanced chemiluminescence (Amersham Pharmacia Biotech, Piscataway, NJ). The exposure autoradiograph is analyzed by the OS-Scan Image Analysis System to obtain the densitometry data.
[Ca2+]i was measured as previously described (13). Briefly, cells were grown on collagen-coated glass coverslips and loaded with the fluorescent Ca2+ indicator dye fura 2. The coverslip was then mounted in a Sykes-Moore chamber (Bellco, Vineland, NJ) and constantly perfused with Ringer buffer. Under the microscope, cells were alternately illuminated through epifluorescent optics with light at 340 and 380 nm. Light emitted >510 nm was captured by an intensified charge-coupled device camera. The video images were digitized and analyzed with Image-1/FL software (Universal Imaging, Media, PA). Images at the two wavelengths were each acquired for one-half second, and then the ratio image (340/380) was calculated with acquisition of ratio images every 5 s. Free [Ca2+]i was derived from the following equation (9): [Ca2+]i = Kd × β × (R − Rmin)/(Rmax − R) where Kd = 224 nM, β is the ratio of fluorescence at 380 nm with low (10 mM EGTA) and high (1.2 mM Ca2+) free Ca2+, R is the ratio of fluorescence with excitation at 340 and 380 nm, and Rmin and Rmax are the fluorescence ratios with low and high free Ca2+. Experiments were carried out at 37°C.
Cells were grown in 24-well culture plates to subconfluence. Serum was removed from the culture medium 1 day before experiments. To measure Ca2+ uptake, the old medium was completely removed and cells incubated at room temperature (25°C) with new medium containing 45Ca2+ (1 μCi/well, 6.7 Ci/mol CaCl2) in the presence or absence of BAPTA-AM or dimethyl BAPTA-AM for 1–20 min, respectively. To terminate 45Ca2+ uptake, the cells were washed three times with ice-cold sucrose solution. The cells were then lysed with 0.1 N NaOH/1% SDS, and the 45Ca2+ counts remaining in each well were determined by scintillation counting.
Cells were subcultured into 12-well plates to subconfluence and labeled with 0.5 μCi/ml per well [3H]arachidonic acid for 3 h before treatment. The wells were washed three times with medium containing 5% fatty acid-free FCS to remove free, labeled arachidonic acid. After treatment of cells, the medium was removed and the released arachidonic acid was determined by scintillation counting. Each data point is the average from at least three wells.
For fura 2 experiments, fluorescence data from at least 12 (n) cells were converted to [Ca2+]i and shown as the mean in the figures and given as mean ± SE in RESULTS. 45Ca2+ uptake data are expressed as percentage of control without treatment, whereby n represents the number of wells. JNK activity is quantitated by counts per minute from kinase assays or by densitometry units from immunoblots (see Fig. 4B) and is expressed relative to that in unstimulated cells. Data are given as means ± SE of n assays. The effects of putative inhibitors for JNK were quantified by setting the stimulated portion of JNK activity (total stimulated activity minus baseline activity) to 100% and calculating the decrease brought about by putative inhibitors. Thus 100% inhibition indicates that the experimental drug completely suppressed any effect of the stimulatory agent present simultaneously. Significance between differently treated groups was determined by unpaired two-tail Student's t-test.
To delineate the relationship between changes in [Ca2+]i and JNK activation, a variety of manipulations were employed to change [Ca2+]i. The Ca2+ ionophore A-23187 (1 μM) induced a significant increment in [Ca2+]i from a basal level of 87 ± 18 nM to a peak of 1,353 ± 576 nM (n = 12) within 5 min, which was followed by a slow return to about 258 nM by 13 min (a level that was 3-fold higher than the basal one). Although significant JNK activation could be observed as early as 5 min, maximum activation of 443 ± 55% (n = 3, P < 0.01, relative to basal level) occurred at 15 min (Fig. 1A). Similarly, thapsigargin (1 μM) induced a transient Ca2+ elevation with a peak value of 1,153 ± 382 nM (n = 12), followed by return to a plateau concentration that was twofold higher than the basal one. This treatment also activated JNK with a maximum activation of 462 ± 87% (n = 6, P < 0.01) at 15 min (Fig. 1B).
BAPTA-AM is widely used as an agent that is converted intracellularly to the Ca2+ chelator BAPTA. Interestingly, BAPTA-AM activated JNK in renal proximal tubule epithelial cells. As shown in a time series experiment in Fig. 2A, JNK activity was induced by 5 min and peaked at 30 min (290 ± 31%, n = 16, P < 0.01); the activity returned to the basal level after 2 h. To assess whether BAPTA-AM indeed chelated cytosolic Ca2+, [Ca2+]i was monitored continuously by fura 2 fluorescence. BAPTA-AM decreased [Ca2+]i from basal 101 ± 10 nM (n = 12) to 66 ± 7 nM (n = 12) at 5 min (a 35% decrease) when cells were bathed in a medium containing normal [Ca2+], i.e., 1.05 mM. [Ca2+]i eventually normalized at ~1 h during continuous exposure to BAPTA-AM.
To evaluate JNK activation in intact cells, the cell lysate was probed for the phosphorylated (activated) form of JNK and its downstream signaling products, i.e., phospho-ATF-2 and phospho-c-jun (Fig. 2B). Equal sample loading was verified by stripping the membrane and immunoblotting for total JNK protein with anti JNK(FL) antibody (data not shown). BAPTA-AM clearly activated JNK by these criteria (Fig. 2B).
Although BAPTA-AM is a Ca2+ chelator, the activation of JNK is not necessarily related to this property. To resolve this question, we compared the effects of BAPTA-AM and dimethyl BAPTA-AM, reasoning that, if the BAPTA-AM effect was related to Ca2+ chelation, then dimethyl BAPTA-AM should be more potent because it has a higher affinity for Ca2+ (Kd: 40 nM vs. 160 nM, Fig. 3A). Although both analogs gave a similar time course of JNK activation, dimethyl BAPTA-AM activation of JNK was either more intensive at the same concentration, or was effective at a lower concentration than that of BAPTA-AM as determined by both vitro kinase assay (Fig. 3B) and immunoblot for the phosphorylated forms of JNK (Fig. 3C). Therefore, it is likely that JNK activation by BAPTA-AM is somehow related to cytosolic Ca2+ chelation or cellular Ca2+ metabolism.
Although it is possible that JNK is activated by both cytosolic [Ca2+] increases or decreases, both essentially representing stress situations, paradoxical activation of a number of Ca2+-dependent processes by chelation of cytosolic Ca2+ has been shown before, e.g., for exocytosis in chromaffin cells and Ca2+-dependent inhibition of adenyl cyclase (6, 7). The paradox is explained by depletion of endoplasmatic Ca2+ stores that, in turn, results in activation of plasma membrane Ca2+ channels (usually termed activation of “Ca2+ release-activated channels” or “capacitative Ca2+ entry”) and therefore highly localized increases in [Ca2+]. Localized [Ca2+] increases around channel openings would not have been detected with the fura 2 method employed in this study because of insufficient temporal and spacial resolution (about 5 s and 5 μm, respectively).
To assess whether BAPTA-AM activates JNK by a Ca2+-dependent mechanism, similar to A-23187 and thapsigargin, we measured to what extent JNK activation is dependent on Ca2+ influx. Cells were stimulated with BAPTA-AM or thapsigargin in the absence or presence of Ca2+ influx inhibitors or an extracellular Ca2+ chelator. As shown in Fig. 4A, SKF-96365, a general Ca2+ influx inhibitor (19), and nifedipine, an L-type Ca2+ channel blocker, significantly reduced BAPTA-AM-induced JNK activation by 62% and 99.5%, respectively. Both drugs together brought BAPTA-AM-induced JNK activation to below the basal level. In comparison, JNK activation induced by thapsigargin was inhibited by SKF-96365 and nifedipine by 87% and 90%, respectively. Furthermore, chelation of extracellular Ca2+ with 3 mM EGTA decreased both BAPTA-AM- and thapsigargin-induced JNK activation by 62% (n = 3) and 94% (n = 3), respectively (Fig. 4B). These results indicate that localized Ca2+ influx and [Ca2+] increases, and not bulk cytosolic [Ca2+], represent a major pathway for JNK activation in the case of BAPTA-AM and thapsigargin.
To verify that BAPTA-AM and dimethyl BAPTA-AM also increase Ca2+ influx across the plasma membrane of proximal tubule cells, as predicted by the capacitative Ca2+ entry mechanism, cellular 45Ca2+ uptake was measured. As shown in Fig. 5A, exposure to BAPTA-AM (10 μM) increased 45Ca2+ uptake significantly at 20 min. Dimethyl BAPTA-AM caused an even higher Ca2+ influx that was significant at both 1 min and 20 min. The greater efficacy of dimethyl BAPTA-AM compared with BAPTA-AM is again in line with its greater chelation affinity for Ca2+, greater store depletion, and hence greater activation of capacitative Ca2+ entry. The dimethyl BAPTA-AM-induced 45Ca2+ uptake was blocked by SKF-96365 and nifedipine by 89% and 66%, respectively (Fig. 5B), indicating that specific Ca2+ entry pathways had been activated.
Our previous data (3, 18) demonstrated that arachidonic acid induced JNK activation by a Ca2+-independent mechanism. To investigate whether BAPTA-AM and thapsigargin involve arachidonic acid as a mediator in the activation of JNK, we measured the release of arachidonic acid after exposure to BAPTA-AM and thapsigargin. As shown in Fig. 6, whereas thapsigargin significantly increased free arachidonic acid, BAPTA-AM had only a moderate, statistically nonsignificant effect.
JNK is an important signal molecule associated with mitogenesis and apoptosis and is activated especially during stress-induced injury (2, 14, 23, 36, 37). Therefore, a detailed understanding of its activation mechanisms is of considerable interest. In many cell types, JNK activation has been correlated with elevated [Ca2+]i and is thus considered to be “Ca2+ dependent” (15, 21, 27, 31, 38). However, no consensus has emerged from many different types of experiments that elevation of [Ca2+]i is the direct cause of JNK activation. In addition, the Ca2+ ionophore A-23187 has been found to be insufficient to activate JNK in several cell types (8, 15, 31).
The complexity of intracellular Ca2+ metabolism in regulating cellular function has become obvious with the increased resolution provided by modern electrophysiological and fluorescence-imaging technology (33). It is now recognized that many cellular functions are normally regulated by localized and not global cytosolic increases of [Ca2+]i (30). Local, short-lived [Ca2+]i increases can be sufficient to act as signal transducers. For example, Lau et al. (16) have demonstrated by photolysis of caged Ca2+ that a local transient [Ca2+] elevation is sufficient to induce new filopodia in neuronal growth cones. Another important aspect of intracellular Ca2+ metabolism is the phenomenon of Ca2+-induced Ca2+ release that serves to amplify localized increases of [Ca2+]i so that relatively high [Ca2+] concentrations can be achieved for short periods of time (1, 4, 10, 25, 26). For example, in excitable cells, such as cardiac myocytes and neurons, Ca2+-induced Ca2+ release from the internal stores is triggered by a transient microdomain of high [Ca2+] beneath an open L-type Ca2+ channel (17, 35).
On this background, the paradoxical, Ca2+-influx-dependent activation of JNK by BAPTA-AM and dimethyl BAPTA can be explained in rabbit proximal tubule cells. BAPTA-AM activates “capacitative Ca2+ influx” in this cell type (that could be inhibited by either SKF-96365 or nifedipine, see 45Ca2+ data) and other epithelial cells, measured usually as electrical Ca2+ current (ICRAC, I = current; CRAC = Ca2+ release-activated channel) (34). Although the phenomenon of capacitative Ca2+ entry is well established for many cell types, including epithelial cells, the signal transduction pathway leading from store depletion to plasma membrane Ca2+ channel activation is not (24, 32). Paradoxical activation of Ca2+-dependent processes through store depletion and local [Ca2+] increases would be expected to be highly variable depending on cell type and on the kinetic and binding properties of the cytosolic Ca2+ chelator employed, because rates and extent of store depletion, magnitude of Ca2+ stores, efficiency of signal transduction, or extent of activatable Ca2+ channels in the plasma membrane may be different. Thus variability of Ca2+-dependent JNK activation with cell type is consistent with indirect activation through capacitative Ca2+ entry. In quantitative terms, a minimum of 62% of JNK activation by BAPTA-AM is explained with Ca2+ entry based on the inhibition data by chelation of extracellular (62%) or the SKF inhibitor (62%). The mechanism that results in the residual JNK activation by BAPTA-AM after chelation of extracellular Ca2+ is not understood, however, it is also nifedipine sensitive.
Although there have been indications that BAPTA-AM may induce JNK activation in other cell types, either no specific data were presented or the effect was suggested as being nonspecific (27, 38). Interestingly, the thapsigargin data suggest that JNK activation in this case also occurs largely through the capacitative Ca2+ entry pathway, although store depletion occurs through a different mechanism, i.e., inhibition of the Ca2+-ATPase in the endoplasmatic reticulum. Differences in cytosolic [Ca2+] between BAPTA-AM and thapsigargin can actually explain differences between the two agents in the duration of JNK activation. A comparison of Fig. 1B and Fig. 2A indicates a shorter duration of JNK activation by thapsigargin compared with BAPTA. High [Ca2+]i with thapsigargin has been shown to inactivate ICRAC (11, 22, 29) and would thus explain a shorter duration of JNK activation. Our finding of prolongation of JNK activation by BAPTA-AM is analogous to several electrophysiology observations demonstrating that inclusion of BAPTA in the patch pipette attenuates the inactivation of ICRAC that normally occurs in the high [Ca2+]i situation with thapsigargin alone (11, 28–30). The nature of the channels activated by store depletion is not known at the moment. Although nifedipine is known as an inhibitor of voltage-dependent L-type channels, it should be recognized that the concentration of 10 μM used in the studies is high and could block other types of channels.
Previous observations from our laboratory demonstrated that arachidonic acid and the cytochrome P-450 products mediate the Ca2+-dependent activation of p42 and p44 MAPK via the tyrosine kinase-Shc/Grb2/Sos-Ras pathway (5, 12). JNK activation appears to be more complex with at least a Ca2+-dependent and a Ca2+-independent step. These steps may be in the same pathway, although parallel pathways of activation cannot be excluded. Cytosolic phospholipase A2 has been shown to be activated by thapsigargin via either increased [Ca2+]i or local [Ca2+] (20). BAPTA-AM activation of JNK could also involve arachidonic acid release (Fig. 6), even though our experiments could not show statistical significance for the increase. To clarify this point, another study is required.
We thank Nnennaya Nkemere and Pearl Whitley for technical assistance involving 45Ca2+ uptake experiments, cell isolation, and tissue culture. We thank Drs. Meredith Bond, Andrea Romani, and Carlos Obejero-Paz for instructive discussions. We also thank Drs. George Dubiyak and Clark Distelhorst for a critical review of the manuscript.
This work was supported by American Heart Association Mid-America Research Consortium Beginning Grant-in-Aid Award 9806215, National Heart, Lung, and Blood Institute Grants HL-41618 and HL-07714, and National Institute of Diabetic, Digestive, and Kidney Diseases Grant DK-27651.