Activation of gene expression by metal-responsive signal transduction pathways.

Metallothioneins are small, cysteine-rich, metal-binding proteins that play important roles in maintaining intracellular metal homeostasis and in transition metal detoxification. MTF-1 (metal transcription factor-1) plays a central role in regulating the metal-inducible, transcriptional activation of metallothionein. Here we report that the phosphorylation of MTF-1 plays a critical role in the activation of MTF-1/metal-responsive element-mediated transcription. Inhibitor studies indicate that signal transduction cascades, including those mediated by protein kinase C, tyrosine kinase, and casein kinase II, are essential for zinc- and cadmium-inducible transcription. In addition, calcium signaling is also involved in regulating transcription. In contrast, cAMP-dependent protein kinase may not be directly involved in the metal response. Contrary to what has been reported for other transcription factors, the inhibition of transcriptional activation does not impair the binding of MTF-1 to DNA, suggesting that phosphorylation is not regulating DNA binding. Elevated phosphorylation of MTF-1 is observed under conditions of protein kinase C inhibition, suggesting that dephosphorylation of this transcription factor mediates its activation.

Metallothioneins (MTs) are a family of evolutionarily conserved, low molecularweight, cysteine-rich, metal-binding proteins (1,2). The precise physiological role of MT has not been fully elucidated. However, proposed roles include a) participation in maintaining the homeostasis of essential transition metals, b) detoxification of nonessential metals, and c) protection against intracellular oxidative stress (1-3).
Two models have been proposed to describe how the interaction among MREs, MTF-1, and zinc activates MT transcription (21,22). Although the current models can account for the regulation of MT transcription by zinc, they do not adequately explain the control of MTF-1 activity by nonzinc stressors.
There is a scarcity of information regarding the molecular mechanism of MT transcriptional activation by other transition metals and environmental stressors.
We propose a model in which the regulation of MT transcription, via the MTF-1/MRE interaction, is controlled by multiple signal transduction cascades that affect MTF-1 phosphorylation. This model is based on several observations. First, MTF-1 contains several evolutionarily conserved, potential phosphorylation sites: protein kinase C (PKC), casein kinase II, and tyrosine kinase. Second, exposure of cells to activators of signal transduction cascades causes an increase in the steady-state level of MT mRNAs (4,6,11,23). Similarly, the addition of signal transduction inhibitors attenuates or abolishes metal-inducible MT mRNA expression (24). Finally, many of the effectors that induce MT transcription (metals as well as other environmental stressors) modulate the activity of intracellular signal transduction cascades (25)(26)(27)(28).
In this report, we show that MTF-1 is phosphorylated on serine and tyrosine residues. In addition, several signal transduction cascades have been identified that modulate cadmiumand zinc-inducible MTF-1/MRE-mediated activation of MT transcription.
For transient transfection studies, HEK293 and HeLa cells were grown in DMEM supplemented with 5% FBS before transfection. These cells were then transfected using the calcium phosphate transfection method (29,30). Transient transfection of COS-7 cells was accomplished by lipofection, using Lipofectin per manufacturer instructions (Life Technologies, Carlsbad, CA, USA).

Expression Plasmids and Fusion Genes
Plasmids encoding either a vesicular stomatitis virus (VSV) G protein-tagged human MTF-1 (hMTF-1) fusion protein, designated phMTF-1-VSV, or a myc-his 6 -tagged mouse MTF-1 (mMTF-1) fusion protein, designated pmMTF-1-myc-his, were engineered. The expressed fusion proteins contain either a VSV or a myc epitope that can be used a) to identify the expressed protein by Western immunoblot analysis and b) for purification by immunoprecipitation. Expression of both the hMTF-1-VSV and mMTF-1-myc-his mRNAs is under the control of the cytomegalovirus promoter, which allows for highlevel expression in a variety of cell types.
A series of reporter plasmids were used to assess the levels of MT transcription. MT transcription in HEK293 and HeLa cells was determined from the amount of luciferase produced using mMT-I-Luc or 4xMREd-Luc reporter constructs (29). The level of MT transcription in COS-7 cells was determined by measuring the level of chloramphenicol acetyltransferase (CAT) produced using -42CAT, -153CAT, and MRE-d´5CAT reporter genes (31). VOLUME 110 | SUPPLEMENT 5 | OCTOBER 2002 • Environmental Health Perspectives In Vivo 33 P and 32 P Labeling HEK293 cells were transfected with phMTF-1-VSV. After transfection the medium was replaced with phosphate-free DMEM containing 33 P-orthophosphate. Either zinc (100 µM ZnCl 2 ) or cadmium (60 µM CdCl 2 ) was added to the 33 P-containing medium. Incubation proceeded for 3 hr, after which nuclear and cytoplasmic extracts were prepared (14). The hMTF-1-VSV fusion protein was isolated by immunoprecipitation, and the amount of 33 P incorporated into hMTF-1-VSV was determined after SDS-PAGE by PhosphorImager analysis. COS-7 cells were transfected with pmMTF-1-myc-his, after which they were washed with phosphate-free DMEM and then incubated in phosphate-free DMEM containing 32 P-orthophosphate. Metals (50 µM CdCl 2 or 100 µM ZnCl 2 ) were added, and the cells were incubated an additional 4 hr. After this incubation, the mMTF-1-myc-his fusion protein was purified by immunoprecipitation.
Amino acid residues that are phosphorylated in MTF-1, and the effects of metal exposure on their level of phosphorylation were determined by Western immunoblot analysis. COS-7 cells were transfected with pmMTF-1-myc-his and exposed to metals. The fusion protein was purified by immunoprecipitation and resolved by SDS-PAGE.
Phosphorylated amino acid residues were identified after Western immunoblot analysis using anti-phosphothreonine, anti-phosphoserine, or anti-phosphotyrosine antibodies.

Signal Transduction Cascade Inhibition
The effects of exposing HEK293, HeLa, and COS-7 cells to kinase inhibitors on the levels of MRE-mediated MT transcription were determined using luciferase-and CAT-based reporter genes.
To measure luciferase activity, HEK293 and HeLa cells were transiently transfected with mMT-I-Luc or 4xMREd-Luc reporter plasmids. After incubating under serum-free conditions, cells were treated with kinase inhibitors (Table 1). Inhibitor treatments were initiated 30 min before the addition of cadmium (60 µM CdCl 2 ) or zinc (100 µM ZnCl 2 ). Cells were exposed to metals for 4 hr in both the presence and absence of the inhibitors. Cell lysates were prepared and luciferase activity measured according to manufacturer instructions.
CAT concentrations were measured using a CAT-enzyme linked immunosorbent assay. COS-7 cells grown in complete DMEM were transfected with -42CAT, -153CAT, or MRE-d´5CAT reporter genes. Chemical inhibitors were added to the transfected cells in complete DMEM. After a 0.5-to 1-hr incubation, either no metal, 50 µM CdCl 2 , or 100 µM ZnCl 2 , was added, and cells were incubated for an additional 3 hr. Extracts were prepared, and CAT concentrations and β-galactosidase activities were determined.

Electrophoretic Mobility Shift Assays
HEK293 cells were incubated with 100 µM ZnCl 2 or 60 µM CdCl 2 for 3 hr in the presence or absence of kinase inhibitors. Binding reactions were performed by incubating 2-5 fmol of a 32 P-end-labeled, double-stranded oligonucleotide that contains the core MRE consensus sequence with nuclear extracts as previously described (14).

Phosphorylation of MTF-1
Radiolabeled proteins corresponding to the MTF-1 fusion proteins were isolated from transfected HEK293 and COS-7 cells.
In vivo labeling of both the hMTF-1 and mMTF-1 fusion proteins showed that MTF-1 is constitutively phosphorylated in the absence of added metal. The phosphorylated form of MTF-1 is located primarily in the cytoplasm of nonexposed cells (Figure 1). After zinc or cadmium exposure, the level of MTF-1 phosphorylation increased. This increase was primarily observed as a higher level of phosphorylation in the nuclear form of MTF-1. However, significant levels of phosphorylated MTF-1 were still observed in the cytoplasmic fraction ( Figure 1).
To identify which types of amino acid residues are phosphorylated in vivo, immunopurified mMTF-1-myc-his was treated with anti-phospho antibodies. Both anti-phosphoserine and anti-phosphotyrosine antibodies cross-reacted with mMTF-1-myc-his ( Figure  1); however, the binding of anti-phosphothreonine antibodies was not observed (data not shown). These results indicate that MTF-1 is phosphorylated at multiple sites: serine and tyrosine residues. Similar to that observed in the in vivo radiolabeling experiments, the protein is constitutively phosphorylated, and the level of phosphorylation increases in response to metal exposure ( Figure 1).

Effects of Signal Transduction Cascade Activators and Inhibitors on MT Transcription
The effects of the PKC inhibitors, staurosporine, H-7, and BIM-I on zinc-and cadmium-inducible MT transcription were investigated. Exposure to these inhibitors reduced mMT-I-based and MRE-based promoter activity to levels below that observed in cells not exposed to metal (Figure 2, Table 1).
Treatment of HEK293 cells with the casein kinase II inhibitor DRB significantly Metals Toxicity • Adams et al.  reduced the level of cadmium-and zincinducible reporter gene expression ( Table 1). Treatment of COS-7 cells with heparin (2 µg/mL) significantly inhibited cadmiuminducible reporter gene expression. This concentration of heparin, however, did not affect zinc-inducible expression ( Figure 2). Consistent with observation, treatment of cells with spermidine, a casein kinase II activator, slightly activated reporter gene expression in COS-7 cells in the absence of metal ( Table 1). Inhibition of tyrosine kinase activity with herbimycin A significantly reduced the level of cadmium-and zinc-inducible reporter gene activity (Table 1). These results suggest that phosphorylation of the tyrosine residue in MTF-1 is involved in the activation of MRE/MTF-1-regulated MT transcription.
Calcium is required for the activation of many kinases; therefore, the effects of intracellular calcium chelators and calcium ionophores on MTF-1/MRE-mediated transcription were investigated. Exposure to the intracellular calcium chelator BAPTA-AM reduced the level of cadmium-and zincinducible promoter activity (Table 1). Consistent with these results, exposure of COS-7 cells to the calcium ionophore A-23187 increased the level of promoter activity in the absence of added metals (Table  1). These results suggest that calcium-mediated signal transduction pathways are involved in the activation of MTF-1.
Inhibition of cAMP-dependent protein kinase (PKA) with Rp-cAMP significantly reduced the levels of zinc-inducible reporter gene expression in HeLa and HEK293 cells (Table 1). However, this change in expression can be attributed to an inhibition of basal activity, suggesting that PKA may not directly mediate metal-inducible transcription. Treatment of COS-7 cells with the PKA inhibitor HA1004 reduced cadmiuminducible reporter gene expression by a small yet significant degree. Zinc-inducible transcription was not affected (Figure 2). Exposure to a PKA activator, dibutyryl-cAMP, did not significantly affect the MT gene expression ( Table 1). The PKA inhibitor and activator data suggest that signaling pathways regulated by PKA may not be directly involved in the regulation of MTF-1 activity.

Electrophoretic Mobility Shift Assay
Exposure to PKC, tyrosine kinase, or casein kinase II inhibitors decreases the level of MTF-1/MRE-mediated gene expression. In contrast, treatment of HEK293 cells with kinase inhibitors did not significantly affect the cadmium-or zinc-inducible binding of MTF-1 to the MRE (Figure 3). In addition, the level of MTF-1-MRE complex formed increased in the presence of inhibitors. These results suggest that the regulation of MREmediated transcription by these kinases, or their related signal transduction pathways is not mediated by inhibiting the DNA binding activity of MTF-1.

Effect of PKC Inhibition on MTF-1 Phosphorylation
The inhibition of PKC activity prevents both cadmium-and zinc-inducible MT transcription. When PKC activity was inhibited in zinc-treated COS-7 cells, the amount of 32 P incorporated into mMTF-1-myc-his increased, compared with zinc-treated cells not exposed to H-7 (Figure 4). Similarly, the amount of phosphorylation on the serine residues significantly increased for both zincand cadmium-treated cells exposed to H-7 (results not shown). It should be noted that treatment with H-7 did not affect the steadystate level of MTF-1.
Metals Toxicity • Regulation of metal-responsive gene expression Nuclear and cytoplasmic extracts were then prepared and hMTF-1 fusion proteins purified by immunoprecipitation. (B) COS-7 cells were transfected with 16 µg of pmMTF-1-myc-his and exposed to 100 µM zinc (Zn) for 4 hr or no metal (NM). The fusion protein was purified by immunoprecipitation and resolved by SDS-PAGE. Identification of MTF-1myc-his (MTF-1) and phosphorylated serine (P-ser) or tyrosine (P-tyr) residues was accomplished by Western immunoblot analysis using anti-mMTF-1, anti-phosphoserine, and anti-phosphotyrosine antibodies, respectively. After a 24-hr recovery period, cells were treated with kinase inhibitors and metals. Reporter gene activity in the presence of metal, but in the absence of inhibitors (NI), and in cells not exposed to metals or inhibitors (NM) was also determined. The data presented are the mean values from three independent experiments with the standard deviations.  Figure 3. Effect of kinase inhibitors on the DNA binding activity of MTF-1. HEK293 cells were incubated with the indicated kinase inhibitors for 30 min before exposure to 100 µM zinc or 60 µM cadmium, or no metal (NM). After a 3-hr incubation, nuclear extracts were prepared and analyzed using a 32 P-labeled MRE-containing oligonucleotide. Nuclear extracts were also prepared from cells that were exposed to cadmium or zinc but not inhibitors.

Discussion
It is well documented that metal-activated MT transcription is mediated primarily through the MRE-binding transcription factor MTF-1.
Previous studies indicate that metal-inducible transcription requires several processes: nuclear translocation, DNA binding, and transcriptional activation of MTF-1. Here we report that activation of MT transcription via the MRE is dependent on interactions with several signal transduction pathways and a change in the level of MTF-1 phosphorylation. The initial examination of MTF-1 activity demonstrated that transcription depends on the binding of this factor to the MRE (14). DNA binding, however, is not synonymous with transcriptional activation. This is supported by the observations that metalinducible, MTF-1/MRE-mediated transcription is significantly blocked with PKC and casein kinase II inhibitors ( Figure 2, Table 1), but these agents do not inhibit MTF-1 binding to DNA (Figure 4).
Recent reports identified a second component in the activation of mammalian MT transcription: the metal-and stress-inducible translocation of MTF-1 from the cytoplasm to the nucleus (29,32). Our studies demonstrate that after metal exposure, the phosphorylated form of MTF-1 accumulates in the nucleus (Figure 1). It is reasonable to hypothesize that modifications of specific residues could control nuclear translocation and transcriptional activation. The unique tyrosine kinase phosphorylation site in human MTF-1, tyr 140 , is contiguous with the MTF-1 nuclear localization sequence 133 KRKEVKR 139 (29). Phosphorylation of this residue may contribute to the regulation of MTF-1 nuclear transport. It should be noted that the location of the unique, putative tyrosine phosphorylation site that is contiguous with nuclear localization sequence is conserved in vertebrate MTF-1.
The observations that a) MTF-1 is phosphorylated, b) its level of phosphorylation increases in response to metal exposure, and c) several kinase inhibitors block or attenuate metal-inducible MT transcription support a model in which transcriptional activation via the MTF-1/MRE interaction is controlled by several signal transduction pathways that ultimately affect the level of MTF-1 phosphorylation. The activation of the signal transduction pathways that involve PKC, casein kinase II, tyrosine kinase, and calcium by metals and other environmental stressors has been well documented (25,(33)(34)(35). A codicil to this hypothesis is that the interaction among several signaling pathways may "fine-tune" the basal and inducible activity of MTF-1.
In the presence of the PKC inhibitor H-7, the level of MTF-1 phosphorylation increases ( Figure 4); however, H-7, staurosporine, and BIM-I are potent inhibitors of metal-inducible MT transcription ( Figure 2, Table 1). Although contradictory, these results are consistent with a model in which metal-activated transcription is controlled by the dephosphorylation of MTF-1. A similar mechanism is responsible for activating the binding of c-Jun to the AP-1 promoter elements. (36). In the dephosphorylation model, only transcriptional activation of MTF-1 is controlled by the dephosphorylation of specific phospho residues. The total level of MTF-1 phosphorylation may be greater after exposure to metals or stress, as shown in Figure 1; however, specific residues are dephosphorylated to activate transcription. The dephosphorylation of MTF-1 is performed by a yet unidentified "MTF-1 phosphatase," whose activity may be controlled by upstream regulatory proteins.
Exposure to metals, organic chemicals, physical stress, intracellular damage, and physiologic agents causes an increase the steady-state level of MT mRNA (5,(37)(38)(39)(40). In addition, several of these nonmetal stressors activate MT transcription via MREs (10,22,41,42). The observation that multiple signal transduction pathways ultimately converge at MTF-1 to activate MT transcription provides a mechanistic link between exposure to structurally unrelated stressors and the activation of MT transcription. Potentially, any stressor that can activate an MTF-1-regulating signal transduction pathway could increase MT transcription. Furthermore, agents that activate parallel signal transduction pathways that converge with pathways regulated by PKC, casein kinase II, and tyrosine kinase will affect MT transcription.
In the future a detailed analysis of the specific residues in MTF-1 that are modified after exposure to metals or other stressors, and various kinase inhibitors, will help to define how each signaling pathway affects the level of MTF-1 phosphorylation. The effects of individual and combinatorial mutations on the ability of MTF-1 to regulate MT transcription will help to determine the precise mechanism by which metals and other stressors activate transcription via MTF-1/MRE interactions. Figure 4. Effect of PKC inhibition on the level of MTF-1 phosphorylation. COS-7 cells were transfected with 16 µg of pmMTF-1-myc-his for 24 hr. Cells were exposed to 100 µM zinc and 32 Porthophosphate for 4 hr in the presence (H-7) or absence (NI) of the PKC inhibitor H-7. The fusion protein was purified by immunoprecipitation, immunocomplexes were then resolved by SDS-PAGE and transferred onto a nitrocellulose membrane, and the amount of 32 P-labeled MTF-1 was determined by PhosphorImage analysis (Molecular Dynamics, Sunnyvale, CA, USA). The arrowhead indicates the location of the MTF-1 fusion protein.