Differential Activation of p38 Mitogen-activated Protein Kinase and Extracellular Signal-regulated Protein Kinases Confers Cadmium-induced HSP70 Expression in 9L Rat Brain Tumor Cells*

We have reported that treatment with CdCl2 at 40–100 μm induces the heat shock proteins (HSPs) in 9L rat brain tumor cells, during which the activation of heat shock factor (HSF) is essentially involved. By exploiting protein kinase inhibitors, we further analyzed the possible participation of specific protein kinases in the above processes. It was found that induction of HSP70 in cells treated with a high concentration of cadmium (i.e. 100 μm) is preceded by the phosphorylation and activation of p38 mitogen-activated protein kinase (p38MAPK), while that in cells treated with a low concentration (60 μm) is accompanied by the phosphorylation and activation of extracellular-regulated protein kinases 1 and 2 (ERK1/2). In 100 μm cadmium-treated cells, both HSP70 induction and HSF1 activation are eliminated in the presence of SB203580, a specific inhibitor of p38MAPK. By contrast, in 60 μm cadmium-treated cells, the processes are not affected by SB203580 but are significantly suppressed by PD98059, which indirectly inhibits ERK1/2 by acting on MAPK-ERK kinase. Taken together, we demonstrate that p38MAPK and ERK1/2 can be simultaneously or independently activated under different concentrations of cadmium and that the signaling pathways participate in the induction of HSP70 by acting on the inducible phosphorylation of HSF1. We thus provide the first evidence that both p38MAPKand ERK signaling pathways can differentially participate in the activation of HSF1, which leads to the induction of HSP70 by cadmium.


From the Department of Life Science, National Tsing Hua University, Hsinchu, Taiwan 30043, Republic of China
We have reported that treatment with CdCl 2 at 40 -100 M induces the heat shock proteins (HSPs) in 9L rat brain tumor cells, during which the activation of heat shock factor (HSF) is essentially involved. By exploiting protein kinase inhibitors, we further analyzed the possible participation of specific protein kinases in the above processes. It was found that induction of HSP70 in cells treated with a high concentration of cadmium (i.e. 100 M) is preceded by the phosphorylation and activation of p38 mitogen-activated protein kinase (p38 MAPK ), while that in cells treated with a low concentration (60 M) is accompanied by the phosphorylation and activation of extracellular-regulated protein kinases 1 and 2 (ERK1/2). In 100 M cadmium-treated cells, both HSP70 induction and HSF1 activation are eliminated in the presence of SB203580, a specific inhibitor of p38 MAPK . By contrast, in 60 M cadmium-treated cells, the processes are not affected by SB203580 but are significantly suppressed by PD98059, which indirectly inhibits ERK1/2 by acting on MAPK-ERK kinase. Taken together, we demonstrate that p38 MAPK and ERK1/2 can be simultaneously or independently activated under different concentrations of cadmium and that the signaling pathways participate in the induction of HSP70 by acting on the inducible phosphorylation of HSF1. We thus provide the first evidence that both p38 MAPK and ERK signaling pathways can differentially participate in the activation of HSF1, which leads to the induction of HSP70 by cadmium.

are induced in cells responding
to suboptimal growing conditions (1,2). Transactivation of the heat shock genes has been centered on the interaction of the HSFs with the HSEs, which are found in all heat shock genes (3)(4)(5). Although several members of HSF have been characterized, HSF1 has been demonstrated to be the most crucial transcription factor activated in mammalian cells responding to classical inducers of the heat shock response (5)(6)(7). Under normal growing conditions, HSF1 may exist as a latent monomeric form, with neither DNA binding nor transcription activity. Activation of HSF1 is through a multistep pathway, a nuclear localization and trimerization step by which HSF1 acquires the properties of a stable trimer, which correlates with the DNA binding activity, and a phosphorylation step by which HSF1 becomes transcriptionally active (5,8,9). On the other hand, HSF1 may also be constitutively phosphorylated, which represses its transcription activity (10,11). Thus, HSF1 may be constitutively or inducibly phosphorylated, and only inducible phosphorylation of this factor can confer the transcription activity. Several phosphorylation sites in HSF1 have been identified to contain the consensus sequence for proline-directed kinases including ERK1/2, p38 MAPK , and SAPKs, which can efficiently phosphorylate HSF1 in vitro (11,12). However, it is still unclear whether the MAPKs phosphorylate HSF1 in vivo.
MAPKs are proline-directed serine/threonine kinases and are themselves activated through phosphorylation on specific tyrosine and threonine residues in response to signaling pathways induced by mitogens or stress conditions. Several members of the MAPK family have been identified, including ERK1/2 (13,14), SAPKs (15)(16)(17), and p38 MAPK (18 -20). ERK1/2 are principally activated by growth factors (13,14), while both SAPKs and p38 MAPK are activated rather weakly by mitogens but very strongly by stress stimuli (16, 19 -24). Interestingly, these distinct MAPK cascades can be activated independently or simultaneously (25). Recently, protein kinase inhibitors with distinct substrate specificity have been employed to unravel the roles of p38 MAPK and ERK1/2 in cellular responses in a variety of experimental systems. SB203580, which belongs to a group of pyridinyl imidazole compounds that specifically bind to or near the ATP binding pocket of p38 MAPK , inhibits the kinase with an IC 50 of 0.6 M. However, it does not affect, even at 100 M, the activities of 12 other protein kinases, including ERK1/2 and SAPKs (26 -29). PD98059, on the other hand, is a flavonoid and a potent inhibitor of MAPK-ERK kinase. The drug is only weakly cytostatic and does not affect overall ERK contents at concentrations Ն10 M; however, exposure to PD98059 results in a rapid loss (Ͼ95%) of the dually phosphorylation (active) form of ERKs (IC 50 ϭ 1 M) (30 -32).
We previously reported that treatment of 9L RBT cells results in the induction of a battery of HSPs, including HSP70, and that HSF-HSE interaction is involved in this process (33). In this study, we have employed SB203580 and PD98059 to investigate whether activation of p38 MAPK and ERK1/2 are necessary for cadmium-induced HSP70 expression. We have focused on the effects of the protein kinase inhibitors on the induction of HSP70 as well as the activation of HSF1 in cadmium-treated 9L RBT cells. Herein we demonstrate that p38 MAPK and ERKs can be simultaneously or distinctly activated under different concentrations of cadmium and that the signaling pathways may participate in the induction of HSP70 by acting on the inducible phosphorylation of HSF1. We thus provide the first evidence that both p38 MAPK and ERKs may differentially participate in the activation of HSF1, which leads to the induction of the heat shock genes. The differential role of each of these two signaling pathways under stress is also discussed.
Cells and Drug Treatment-The 9L RBT cells (34) were maintained in Eagle's minimum essential medium plus 10% fetal calf serum supplemented with 100 units/ml penicillin G and 100 g/ml streptomycin in a 37°C incubator under 5% CO 2 and 95% air. Prior to each experiment, stock cells were plated in 25-cm 2 flasks or six-well plates at a density of 4Ϫ6 ϫ 10 4 cells/cm 2 . Exponentially growing cells at 80Ϫ90% confluency were used. Cadmium chloride was dissolved in water at a concentration of 100 mM and added to the culture medium to the desired concentrations for treatment. The cells were treated at 37°C for 2 h and allowed to recover for various durations as specified. For the studies concerning the effects of protein kinase inhibitors, the cells were preincubated with respective inhibitors, at various concentrations for 1 h, followed by treatment with cadmium in the presence of the inhibitors.
Metabolic Labeling and SDS-PAGE-De novo protein synthesis was revealed by [ 35 S]methionine labeling at a concentration of 20 Ci/ml. After various treatments, the cells were labeled for 1 h, washed in PBS (137 mM NaCl, 2.7 mM KCl, 4.3 mM Na 2 HPO 4 , 1.4 mM KH 2 PO 4 , pH 7.4), and lysed in sample buffer (35). Equal amounts of cell lysates were resolved by SDS-PAGE (35). The gels were then fixed, dried, and processed for autoradiography as described (33). Protein bands of interest were quantified by densitometric scanning (Molecular Dynamics, Inc., Sunnyvale, CA). The relative synthesis rate of HSP70 was presented as sum of pixel values of each band divided by that of actin in the same lane (internal control).
Nuclear Extract Preparation and Electrophoretic Mobility Shift Assay-Nuclear extracts were prepared from 9L RBT cells by a rapid fractionation protocol (36). Approximately 10 7 cells were trypsinized, collected by centrifugation at 100 ϫ g for 8 min at 4°C, washed once with PBS, and centrifuged as above. The pellet was resuspended in 0.5 ml of nuclear extraction buffer (10 mM Hepes-KOH, pH 7.9, 0.5% Triton X-100, 0.5 M sucrose, 0.1 mM EDTA, 10 mM KCl, 1.5 mM MgCl 2 , 1 mM dithiothreitol, 0.5 mM phenylmethylsulfonyl fluoride) and then homogenized by three strokes with a Dounce grinder. The samples were briefly centrifuged at 10,000 ϫ g, and the resulting pellet was lysed in 0.5 ml of nuclear extraction buffer supplemented with 0.5 M NaCl and 5% glycerol at 4°C for 30 min. The samples were then centrifuged at 14,000 ϫ g for 20 min, and the supernatant fractions were collected. The nuclear extracts obtained were dialyzed for 3-4 h against at least 50 volumes of dialysis buffer (10 mM Hepes-KOH, pH 7.9, 17% glycerol, 0.1 mM EDTA, 50 mM NaCl, 1 mM MgCl 2 , 1 mM dithiothreitol, 0.5 mM phenylmethylsulfonyl fluoride) and kept frozen at Ϫ70°C until use. HSE binding activities of the nuclear extracts were determined by electrophoretic mobility shift assay using double-stranded oligonucleotides as probes. The synthetic HSE oligonucleotide probe was prepared by annealing 5Ј-CTAACAGACCCGAAACTGCTGGAAGATTCTTGG-3Ј (corresponding to 152-184 of the genomic sequence of rat hsp70 (37)) with its complementary strand, followed by end labeling with [␥-32 P]ATP and polynucleotide kinase. The binding reaction mixture consisted of 2 g of nuclear extract and 1 ng of oligonucleotide (approximately 2 ϫ 10 4 cpm) in 15 l of binding buffer (10 mM Hepes-KOH, pH 7.9, 5 mM MgCl 2 , 4 mM Tris-HCl, 12% glycerol, 100 mM KCl, 1 mM EDTA, 1 mM dithiothreitol, 300 g/ml bovine serum albumin, and 20 g of poly(dI-dC)). The reaction was allowed to proceed at room temperature for 30 min. The samples were subsequently resolved on 6% nondenaturing polyacrylamide gels in TBE buffer (50 mM Tris, 50 mM boric acid, 1 mM EDTA, pH 7.0). After electrophoresis, the gels were dried and processed for autoradiography.
Immunoblot Analysis-Whole cell extracts were fractionated by SDS-PAGE and then transferred to nitrocellulose membrane in TBE buffer by using a semidry transfer apparatus according to the manufacture's protocols (OWL scientific, Woburn, MA). After blocking with 3% nonfat milk in TTBS (0.5% Tween 20, 20 mM Tris-HCl, pH 7.4, 0.5 M NaCl) for 90 min and washing once with TTBS, the membranes were incubated with a 1:1000 dilution of respective antibodies against phospho-p38 MAPK , p38 MAPK , phospho-ERK1/2, ERK1/2, phospho-c-Jun, and c-Jun at room temperature for 12 h. Subsequently, the membranes were washed three times with TTBS, for 10 min each time, and incubated with a 1:5,000 dilution of horseradish peroxidase-conjugated antimouse or anti-rabbit secondary antibodies for 60 min. After washing three times for 5 min each time with TTBS, the immune complexes were detected by the ECL system and quantified by densitometric scanning.
Activity Assays for p38 MAPK and ERK1/2-Cells were stressed by 0, 60, 100 M CdCl 2 for 2 h and then lysed in kinase lysis buffer (20 mM Tris-HCl, pH 7.5, 1 mM EDTA, 1% Triton X-100, 10 mM ␤-glycerophosphate, 50 mM sodium pyrophosphate, 1 mM Na 3 VO 4 , 1 mM benzamidine, 0.1% ␤-mercaptoethanol) for 10 min on ice. Insoluble material was removed by centrifugation (10,000 ϫ g, 20 min, 4°C). The cell lysate was mixed with 5 l (200 ng) of glutathione S-transferase-MAPKAPK-2 or MBP, 5 l of [␥-32 P]ATP, and 2 l of magnesium/ATP mixture and then incubated for 15 min at 30°C with agitation. The solution was finally mixed with an equal volume of 2ϫ sample buffer for SDS-PAGE. After electrophoresis, the gels were dried and processed for autoradiography. 32 P i Labeling and Immunoprecipitation-For the determination of HSF1 phosphorylation, cells were incubated with 1 mCi of 32 P i in 1 ml of phosphate-free Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum for 1 h. After treatment, cells were rapidly chilled on ice, washed twice with ice-cold PBS, and then lysed in lysis buffer (20 mM Hepes-KOH, pH 7.9, 5 mM EDTA, 10 mM EGTA, 5 mM NaF, 0.1 g/ml microcystin LR, 10% glycerol, 1 mM dithiothreitol, 0.4 M KCl, 0.4% Nonidet P-40, and protease inhibitors (5 g/ml leupeptin, 5 g/ml aprotinin, 5 g/ml pepstatin, 1 mM benzamidine, and 50 g/ml phenylmethylsulfonyl fluoride)) for 10 min on ice. Insoluble material was removed by centrifugation (10,000 ϫ g, 20 min, 4°C). The protein content of the cell lysate was determined by Bradford assay (Bio-Rad), and an equal amount of protein was used in each reaction. After preclearing by incubation with 1 g of protein G-Sepharose (Amersham Pharmacia Biotech) for 2 h at 4°C followed by centrifugation to remove pellets, the cell lysates were incubated with a 1:5,000 dilution of anti-HSF antibodies (Affinity Bio-reagents Inc., Golden, CO) for 12 h. Immune complexes were further precipitated with 1 g of protein G-Sepharose, washed five times with lysis buffer, and finally mixed with an equal volume of 2ϫ sample buffer for SDS-PAGE. After electrophoresis, the gels were dried and processed for autoradiography.
Flow Cytometry-Flow cytometry was performed on a fluorescenceactivated cell sorter, FACScan (Becton-Dickinson, San Jose, CA), using CellQuest and Modfit LT software. For cell cycle analysis, aliquots of 1 ϫ 10 6 cells were fixed in 70% ethanol on ice for 2 h and collected by centrifugation (1,000 ϫ g, 10 min, 4°C). The pellets were resuspended in 1 ml of PBS buffer containing 20 g/ml RNase A and incubated at room temperature for 30 min. RNase was removed by centrifugation (1,000 ϫ g, 10 min, 4°C), and the fixed cells were resuspended in 1 ml of PBS followed by incubation with 100 l of 10 g/l propidium iodide (PI) for 30 min. The labeled cells were analyzed with FACScan according to the manufacturer's instructions.
Preparation of Genomic DNA from 9L RBT Cells-The DNA extraction procedure was carried out at 4°C unless otherwise specified. Approximately 10 8 cells treated with or without CdCl 2 were collected by centrifugation (1,000 ϫ g, 5 min) and resuspended in 10 ml of PBS buffer, and the pellet was collected by centrifugation (1,000 ϫ g, 5 min). The cells were resuspended in 3.6 ml of proteinase K buffer (10 mM Tris-Cl, pH 7.4, 10 mM EDTA, 150 mM NaCl, 0.4% SDS), and 0.4 ml of 10 mg/ml proteinase K was added. The mixture was incubated at 65°C for 15 min and 37°C overnight. The solution was extracted three times with an equal volume of phenol/chloroform (1:1) and finally one time with an equal volume of chloroform followed by centrifugation (2,500 ϫ g, 10 min). One-tenth volume of 3 M sodium acetate, pH 7.0, and 2.5 volumes of ethanol were both added to the aqueous layer with gentle mix to precipitate the genomic DNA. The DNA pellet was rinsed with 5 ml of 80% ethanol and recollected by centrifugation (2,500 ϫ g, 5 min). To dissolve the precipitated DNA, 4.5 ml of proteinase K buffer without SDS was added, and the solution was incubated at room temperature overnight. Subsequently, 25 l of 10 mg/ml RNase A was added, and the solution was further incubated at 37°C for 30 min, extracted with 5 ml of phenol/chloroform (1:1), and separated by centrifugation (2,500 ϫ g, 10 min). The genomic DNA was precipitated by the addition of 9 ml of ethanol, rinsed with 80% ethanol, and collected by centrifugation (2,500 ϫ g, 10 min). The isolated DNAs were loaded and separated on a 1% agarose gel in 1ϫ TAE buffer and stained with ethidium bromide for examination with a UV transilluminator.

Effects of Protein Kinase Inhibitors on Cadmium-induced HSP70
Expression in 9L RBT Cells-As a first step to elucidate the possible involvement of protein kinases in the induction of HSP70 under cadmium treatment, several protein kinase inhibitors, including staurosporine, SB203580, bisindolylmaleimide (BIM), and H89, which inhibit general kinases, p38 MAPK , protein kinase C, and protein kinase A, respectively, were employed. Cells were respectively incubated with the above inhibitors for 1 h, treated with 100 M CdCl 2 for 2 h along with each inhibitor, and metabolically labeled with [ 35 S]methionine for 1 h. De novo protein synthesis was monitored by autoradiography (Fig. 1A). Staurosporine appeared to inhibit the induction of HSP70, indicating that certain protein kinase(s) are involved in the process (Fig. 1, lane 3). Furthermore, SB203580, but not BIM or H89, was found to be able to significantly suppress the induction of HSP70 in cadmium-treated cells (Fig. 1, lane 4). The results indicated that the p38 MAPK , but not the protein kinase A or protein kinase C, signaling pathway is involved in this process.
Effects of Protein Kinase Inhibitors on the HSE Binding Activities of Nuclear Proteins Extracted from Cadmium-treated 9L RBT Cells-We have previously demonstrated the formation of HSE⅐HSF complexes during the induction of HSPs in cadmium-treated 9L RBT cells (33). Herein we further analyzed the effects of the protein kinase inhibitors on the HSE binding activities of the nuclear proteins extracted from both the treated and untreated cells. As shown in Fig. 2, two slowly migrating complexes, designated as complexes I and II, were detected (Fig. 2, lane 2). In the untreated cells, staurosporine could completely abolish the formation of complex I (Fig. 2, lane  3), while SB203580, BIM, and H89 could only reduce the binding intensity but did not affect the binding patterns (Fig. 2,  lanes 4 -6). Enhanced binding activity toward HSE was found in the cadmium-treated cells (Fig. 2, lane 7); however, both staurosporine and SB203580 eliminated the formation of complex I (Fig. 2, lanes 8 and 9), which could still be detected in cells preincubated with BIM or H89 (Fig. 2, lanes 10 and 11). These observations supported the notion that there are constitutive as well as inducible phosphorylation processes of HSFs and indicated that complex I may represent the presence of the inducible phosphorylation form. Moreover, the effects of the protein kinase inhibitors on the formation of complex I (inducible phosphorylation form) closely correspond to those on the induction of HSP70 shown in Fig. 1. Taken together, the data indicated that in cells treated with 100 M CdCl 2 for 2 h, the p38 MAPK signaling pathway is involved in the cadmium-induced activation of HSF, which in turn confers the induction of HSP70.
Correlation of HSP70 Induction and p38 MAPK Activation in Cadmium-treated 9L RBT Cells-To study the relationship between HSP70 induction and p38 MAPK activation, the levels of p38 MAPK , tyrosine phosphorylation of p38 MAPK , and HSP70 were determined by immunoblotting analysis. The anisomycintreated cell lysates, supplied by the manufacturer, were also used as additional controls to assure the authentic detection of p38 MAPK and phospho-p38 MAPK . The results clearly demonstrated that Tyr 182 of p38 MAPK underwent phosphorylation in cells treated with 100 M CdCl 2 (Fig. 3, lane 5) despite no observable changes in the levels of p38 MAPK with or without cadmium treatment. Furthermore, it was found that phosphorylation of p38 MAPK was completely abolished and inductive accumulation of HSP70 was significantly suppressed in cells preincubated with SB203580 (Fig. 3, lane 6).
The correlation between p38 MAPK activation and HSP70 induction was further examined in cells treated with 100 M CdCl 2 for 2 h followed by different recovery durations (Fig. 4). The synthesis of HSP70 was evaluated by metabolic labeling followed by autoradiography, while the alterations in the phosphorylation of p38 MAPK and total expression of p38 MAPK were determined by immunoblotting using respective antibodies. Fig. 4A showed that synthesis of HSP70 was induced immediately after treatment, reached its maximum after 2 h, and diminished after 8 h of recovery. Consistently, p38 MAPK remained unphosphorylated in the untreated cells, but the phosphorylated form of p38 MAPK appeared immediately after treatment. The protein kinase was found to be phosphorylated until 4 h but completely dephosphorylated after 8 h of recovery (Fig.  4B). Quantitative analysis of the data showed that phosphorylation of p38 MAPK appeared to precede the induction HSP70 (Fig. 4C). These experiments clearly demonstrated that treatment with cadmium would result in phosphorylation of p38 MAPK , which subsequently led to HSP70 induction in the treated 9L RBT cells.
Concentration-dependent Activation of p38 MAPK and ERK1/2 in Cadmium-treated 9L RBT Cells-The correlation between the HSP70 induction and MAPK activation was further investigated in cells treated with different concentrations of cadmium for 2 h followed by 2 h of recovery. Induction of HSP70 was determined by metabolic labeling, and levels of p38 MAPK and phospho-p38 MAPK were determined by immunoblotting as in the previous experiments. The results clearly revealed that HSP70 was induced under all concentrations being applied (Fig. 5A). Surprisingly, phosphorylation of p38 MAPK could only be detected in cells treated with 80 and 100 M CdCl 2 (Fig. 5B, lanes 4 and 5), indicating that the kinase could only be activated by high concentrations of cadmium and

FIG. 2. Effects of protein kinase inhibitors on the binding activities of HSF to the HSE derived from the promoter of hsp70 in cadmium-treated 9L RBT cells.
Cells were incubated with the protein kinase inhibitors and treated with CdCl 2 as specified in the legend to Fig. 1, and nuclear extracts were prepared from cells recovered for 2 h after treatment. Synthetic oligonucleotides corresponding to the HSE of rat hsp70 were annealed, end-labeled, and mixed with the nuclear extracts. The DNA-protein complexes were resolved by 6% nondenaturing PAGE and visualized by autoradiography. Complexes I and II are marked at the right. Similar results were obtained in three independent experiments.

FIG. 4. Induction of HSP70 in cadmium-treated 9L RBT cells is
preceded by the activation of p38 MAPK . Cells were treated with 100 M cadmium for 2 h and allowed to recover for up to 8 h. At different intervals as indicated, the cells were labeled with [ 35 S]methionine for 1 h and then lysed. The cell lysates were resolved by SDS-PAGE followed by autoradiography (A). Additionally, the cell lysates were analyzed by immunoblotting using anti-phospho-p38 MAPK and anti-p38 MAPK as the primary antibodies (B). The relative synthesis rate of HSP70 was determined as described in the legend of Fig. 1, while the relative levels of p38 MAPK phosphorylation were presented as the sum of pixel values of each phospho-p38 MAPK band divided by that of p38 MAPK in the same lane (C). Data represent the means Ϯ S.D. of three independent experiments. thus may be responsible for HSP70 induction under these conditions. In an attempt to investigate whether another class of MAPK pathways is involved in HSP70 induction upon cadmium treatment, we have examined the roles of ERKs, which have been reported to be involved in the in vitro phosphorylation of HSF1. By exploiting antibodies against ERK1/2 and phospho-ERK1/2, we found that phosphorylation of the ERKs was initially induced by 40 M CdCl 2 , peaked at 60 M CdCl 2 , and subsided at higher concentrations. Phosphorylation of ERK1/2 became barely detectable in cells treated with 100 M CdCl 2 (Fig. 5C). On the other hand, when antibodies against the substrate of SAPKs (c-Jun) and phospho-c-Jun were employed, no change in protein level and phosphorylation level was detected in cells treated with up to 100 M CdCl 2 (data not shown). Taken together, our results showed that both p38 MAPK -and ERK1/2-mediated pathways can be simulta-neously or independently activated in cadmium-treated cells, depending on the concentration used, whereas the SAPK signaling pathway appeared not to be involved in this process. Concentration-dependent differential involvement of the p38 MAPK and ERKs in cadmium-induced HSP70 was further substantiated by the employment of specific inhibitors of these kinases, PD98059 and SB203580. As shown in Fig. 6, preincubation with PD98059 abolished HSP70 induction in cells treated with 60 M CdCl 2 but not in those treated with 100 M CdCl 2 (Fig. 6A, lanes 3 and 6). Conversely, SB203580 effected abolishment of HSP70 induction upon CdCl 2 treatment at 100 M but not at 60 M (Fig. 6A, lanes 4 and 7). Furthermore, induction of HSP70 by 60 M CdCl 2 was strictly associated with the activation of ERK1/2, since both HSP70 induction and ERK1/2 phosphorylation were simultaneously eliminated by PD98059 but not by SB203580 in these samples (Fig. 6B, lanes  2-4). In contrast, the induction of HSP70 by 100 M CdCl 2 was strictly confined to the activation of p38 MAPK , since both processes in these cells were simultaneously inhibited by SB203580 but not by PD98059 (Fig. 6B, lanes 5-7). In addition to examining the phosphorylation of p38 MAPK and ERK1/2 themselves, the activation of the protein kinases was assessed by the phosphorylation of their respective downstream effectors, MAPKAPK-2 and MBP. It was found that MAPKAPK-2 was phosphorylated in the presence of 100 M CdCl 2 (Fig. 7A,  lane 3), while MBP was phosphorylated in the presence of 60 M CdCl 2 (Fig. 7B, lane 2). Furthermore, each phosphorylation process could be blocked by its specific inhibitor (Fig. 7, A, lane  4, and B, lane 3). Altogether, we provided concrete evidence that there is differential involvement of p38 MAPK and ERK1/2 in cadmium-induced synthesis of HSP70.  4 and 7). The cells were then allowed to recover for 2 h and labeled with [ 35 S]methionine for 1 h before they were lysed. The cell lysates were processed for autoradiography and immunoblotting as described. Shown are the autoradiogram (A) and the immunoblots obtained by using antibodies against phospho-p38 MAPK and phospho-ERK1/2 (B). Similar results were obtained in three independent experiments.
Phosphorylation of HSF1 in Cadmium-treated 9L RBT Cells-We subsequently measured the phosphorylation level of HSF1 in the cells under cadmium treatment to investigate whether the processes of differential activation of p38 MAPK and ERK1/2 were authentically coupled to the inductive phosphorylation of HSF1. As in the previous experiments, cells were prelabeled by [ 32 P]orthophosphate and then treated with 60 or 100 M CdCl 2 with or without preincubations of the protein kinase inhibitors. After treatment, HSF1 was immunoprecipitated from the cell lysates, and the immunocomplexes were analyzed by SDS-PAGE followed by autoradiography. We found that HSF1 was phosphorylated in cells treated with both 60 and 100 M CdCl 2 (Fig. 8). In accordance with the results shown in Fig. 6, preincubation with PD98059 remarkably eliminated the HSF1 phosphorylation in cells treated with 60 but not 100 M CdCl 2 . In contrast, the phosphorylation of HSF1 under 100 M CdCl 2 could only be abolished in cells preincubated with SB203580. These data further substantiated the notion that phosphorylation of HSFs and HSP70 induction in cells treated with different concentrations of cadmium are mediated through p38 MAPK and ERK1/2 differentially.
Mitogenic and Apoptotic Effects Induced in Cadmium-treated 9L RBT Cells-In order to investigate whether 60 and 100 M CdCl 2 , respectively, cause the mitogenic and apoptotic effect in 9L RBT cells, changes in DNA synthesis and fragmentation in each condition have been monitored by flow cytometry employing PI staining and agarose gel electrophoresis as shown in Fig.  9. In our experiments, CdCl 2 at a concentration of 60 or 100 M was used for induction, and the cells were allowed to execute their stress response in 6 or 12 h. As shown in Fig. 9A, the percentage of cells arrested in G 2 /M phase was 13.6%, which drastically increased to 30.8% once the cells were treated with 60 M CdCl 2 and 6 h of recovery. The increase indicated vigorous mitogenesis in the treated cells. On the other hand, for cells treated with 100 M CdCl 2 and 6 h of recovery, the percentage increased to only 21.1% accompanied by the appearance of some apoptotic cells. The apoptotic cells became the major component as the cells treated with 100 M CdCl 2 were allowed to recover for 12 h. Meanwhile, the fraction arrested in G 2 /M phase decreased to only 2.3%. The genomic DNAs derived from cells with or without CdCl 2 treatment were further examined on a 1% agarose gel as shown in Fig. 9B. It was clear that no additional DNA fragmentation occurred in cells treated with 60 M CdCl 2 and 6 h of recovery as compared with the control, but severe DNA fragmentation was observed in cells treated with 100 M CdCl 2 . It was also evident that as the post-treatment time extended, more smaller fragments could be detected. The results were in agreement with what was derived with the PI assay, i.e. low and high concentration of cadmium chloride applied in the system indeed induces different response in the treated cells. We have proved that 60 and 100 M CdCl 2 individually cause the mitogenic and apoptotic effect, respectively. DISCUSSION We have demonstrated the differential activation of ERK1/2 and p38 MAPK upon CdCl 2 treatment in 9L RBT cells. Both of these two signal cascades lead to the phosphorylation and activation of HSF1, which in turn confers HSP70 induction upon treatment with CdCl 2 . Under low cadmium concentration (60 M), ERK1/2 were activated and then HSF1 was phosphorylated. On the other hand, p38 MAPK was activated and respon- sible for phosphorylation of HSF1 in cells treated with a high concentration (100 M) of cadmium. ERK1/2 and p38 MAPK are subgroups of the MAPK family that play key roles in transducing extracellular signals to the nucleus (25,38). In general, ERK-and p38 MAPK -mediated signaling represent independent pathways, with distinct upstream activators and downstream targets. The ERK1/2 pathway consists of a protein kinase cascade linking growth and differentiation signals with transcription in the nucleus. Growth factor receptors and tyrosine kinases activate Ras, which in turn activates c-Raf, MAPK-ERK kinase 1/2, and thus ERK1/2. Activated ERK1/2 then translocates to the nucleus and activates transcription by phosphorylation of transcription factors such as Elk1 and Stat (14,39). On the other hand, p38 MAPK has been shown to be activated by MAPK kinases 3 and 6 (40,41). Subsequently, activated p38 MAPK phosphorylates and activates MAPKAPK-2 (20) and ATF-2 (21,42). We have demonstrated that ERK1/2 are activated in cells treated with 60 M CdCl 2 , while p38 MAPK is activated in cells treated with 100 M CdCl 2 . Moreover, both ERK1/2 and p38 MAPK are found to be phosphorylated by 80 M CdCl 2 , indicating that the kinase cascades can be simultaneously or separately activated in cadmium-treated cells. However, the SAPK signaling pathway appeared to be uninvolved in the cadmium-activated HSF1, since the processes could be completely abolished by the protein kinase inhibitors PD98059 and SB203580. Since the distinct ERK and p38 MAPK signaling pathways principally represent mitogenic response and stress response, respectively, it is conceivable that there are concentration-dependent responses to cadmium with different biological consequences. Our experimental data proved that activation of ERK1/2 in cells upon treatment with a low concentration of cadmium causes a mitogenic response representing an adaptation route for a future down-regulation stress response elicited by cadmium. By contrast, activation of p38 MAPK upon treatment with a high concentration of cadmium is associated with induction of an apoptotic response (43,44). However, whether p38 MAPK and ERK1/2 are directly involved in the apoptotic and mitogenic processes still remains to be investigated.
As mentioned previously, it has been shown that HSF1 is a suitable in vitro substrate for proline-directed kinases including ERK1/2 (11,12). However, two-dimensional PAGE analysis showed that the pattern of phosphorylation in vitro varies from that of in vivo with only one phosphopeptide in common (11). Moreover, phosphorylation of HSF1 by ERK1/2 is likely to result in a suppression of the transactivation activity (11,13,45). In in vivo studies, both of the ERK and p38 MAPK signaling pathways have been linked to the regulation of HSF1. For instance, it has been shown that overexpression of c-Raf and thus up-regulation of the Raf/MAPK-ERK kinase/ERK pathway would lead to enhanced phosphorylation of HSF1 (10,45) and that p38 MAPK is activated in heat-shocked and osmotically stressed cells in which heat shock genes are induced (18,46). In the present studies, the respective activations of ERK1/2 and p38 MAPK by low and high concentrations of cadmium are strictly coupled to the inducible phosphorylation of HSF1, as revealed by exploiting PD98059 and SB203580 (Fig. 8). Our findings indicate that activation of ERK1/2 or p38 MAPK could be directly linked to phosphorylation and activation of HSF1, suggesting that HSF1 can serve as a downstream effector for both ERK1/2 and p38 MAPK signaling pathways.
Our electrophoretic mobility shift assay and 32 P i labeling experiments revealed that cadmium treatment induces HSE binding activity and phosphorylation of HSF1. However, the constitutive binding of HSF1 to HSE could be abolished only by staurosporine and not by SB203580, and it is also shown that neither SB203580 nor PD98059 could affect the constitutively phosphorylated form of HSF1. It appeared that the constitutive and inducible phosphorylation forms of HSF1 were regulated by multiple mechanisms and that distinct kinases were involved in different stages. We have demonstrated the involvement of ERK1/2 and p38 MAPK pathways in the inducible phosphorylation and thus transcription activity of HSF1; however, the candidates responsible for constitutive phosphorylation of HSF1 and the detailed regulatory mechanism for HSF1 phosphorylation need to be elucidated.
In conclusion, we have demonstrated that the p38 MAPK and ERK1/2 signaling pathways are differentially activated by cadmium at different concentrations and that the activation of these MAPKs will converge to the phosphorylation and activation of HSF1, which in turn transactivates the heat-shock genes. However, the upstream activation mechanisms still await further elucidation.