Regulation of Metallothionein Transcription by the Metal-responsive Transcription Factor MTF-1

Every living organism must detoxify nonessential metals and carefully control the intracellular concentration of essential metals. Metallothioneins, which are small, cysteine-rich, metal-binding proteins, play an important role in these processes. In addition, the transcription of their cognate genes is activated in response to metal exposure. The zinc finger transcription factor MTF-1 plays a central role in the metal-inducible transcriptional activation of metallothionein and other genes involved in metal homeostasis and cellular stress response. Here we report that the phosphorylation of MTF-1 plays a critical role in its activation by zinc and cadmium. Inhibitor studies indicate that multiple kinases and signal transduction cascades, including those mediated by protein kinase C, tyrosine kinase, and casein kinase II, are essential for zinc- and cadmium-inducible transcriptional activation. In addition, calcium signaling is also involved in regulating metal-activated 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, 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 condition of protein kinase C inhibition, suggesting that specific dephosphorylation of this transcription factor contributes to its activation.

teins (1,2). The precise physiological function of MT has not been fully elucidated. However, proposed roles include (a) participation in maintaining the homeostasis of essential transition metals; (b) sequestration of toxic metals, such as cadmium and mercury; and (c) protection against intracellular oxidative damage (1)(2)(3)(4).
Metallothionein expression is primarily controlled at the level of transcription. Transcription can be induced by a variety of physiological agents and environmental stressors such as transition metals, glucocorticoids, cAMP, phorbol esters, alkylating agents, oxidizing agents, ultraviolet and ionizing radiation, and hypoxia (1,(5)(6)(7)(8)(9)(10)(11)(12). The activation of MT transcription by transition metals is mediated by regulatory elements, designated metal-responsive elements (MREs). MREs contain a 7-bp core sequence (TGCRCNC) and are present in multiple copies in the promoter/enhancer regions of almost all metalinducible MTs (13,14). Inducible transcription is mediated by a variety of other regulatory elements located in promoter/ enhancer regions of MT genes, which include glucocorticoid response elements, cAMP response elements, AP-1 elements, and antioxidant response elements (1,5,15,16).
Two models have been proposed to describe how the interaction among MREs, MTF-1, and zinc activates MT transcription. The first proposes that MTF-1 acts as a zinc sensor that cannot bind to DNA in the absence of metal (18, 24 -27). As cells accumulate zinc, the metal binds in the finger domains, causing a conformational change in the protein and subsequent binding to the MRE. When MT levels are sufficient, it chelates the metal from the zinc fingers; now, MTF-1 can no longer bind to the MREs and transcription ceases.
An alternative model hypothesizes the existence of a zincsensitive inhibitor, which complexes with MTF-1 rendering it inactive (28). As cellular zinc levels increase, the metal binds to the inhibitor releasing it from MTF-1, allowing the transcription factor to bind to the MRE to activate transcription. When MT levels are sufficient, the zinc is removed from the inhibitor, allowing it once again to bind to MTF-1 thereby inhibiting transcription.
Typically, zinc is used as the inducer when the interactions among MREs, MTF-1, and MT transcription are investigated.
There is a paucity of information regarding the molecular mechanism of MT transcriptional activation by other transition metals and environmental stressors. A combination of mutational and transcriptional analyses clearly show that cadmium activation of MT transcription is dependent on MTF-1 and MREs (18,29). MTF-1 binding to MREs and the activation of MT transcription is also induced by oxidative stress (25,30). 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 non-zinc stressors.
We propose a model in which the regulation of MT transcription, via the MTF-1/MRE interaction, is controlled by several signal transduction cascades that affect MTF-1 phosphorylation. It has been reported previously that the level of MTF-1 phosphorylation is modified following exposure to cadmium or zinc in vivo (31,32). The phosphorylation of MTF-1 may be affected by one or more metal-or stress-responsive signal transduction pathways. This model is based on several observations. First, protein motif analysis (PROSITE) (33) of MTF-1 indicates the presence of several evolutionarily conserved, potential phosphorylation sites. As many as 11 PKC sites, 13 casein kinase II sites, and 1 tyrosine kinase site are predicted in the four characterized MTF-1s from different species. Second, exposure of cells to activators of signal transduction cascades causes an increase in the steady-state level of MT mRNAs (5,7,12,34). Similarly, addition of signal transduction inhibitors attenuates or abolishes metal-inducible MT mRNA expression (35). Finally, many of the effectors that induce MT transcription, metals (zinc, cadmium, mercury, arsenic, chromium) as well as other environmental stressors (oxidative stress, radiation), modulate the activity of intracellular signal transduction cascades (36 -39).
In this report, we show that MTF-1 is phosphorylated in vivo on serine and tyrosine residues, and that the level of phosphorylation is affected by metal exposure and kinase inhibitors. In addition, several signal transduction cascades have been identified that modulate cadmium-and 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% fetal bovine serum for 16 h prior to transfection. These cells were then transfected using the calcium phosphate transfection method as previously described (40,41). COS-7 were grown to ϳ60% confluence in complete DMEM. The medium was then removed, and the cells were washed with Opti-MEM reduced serum medium (Invitrogen) immediately before transfection. Transient transfection was accomplished by lipofection using Lipofectin per manufacturer's instructions (Invitrogen).
Expression Plasmids and Fusion Genes-The plasmid encoding a vesicular stomatitis virus (VSV) G protein-tagged hMTF-1 fusion protein, designated phMTF-1-VSV, has been previously described (42). An expression vector encoding a myc-His 6 -tagged mMTF-1 fusion protein was constructed that contained a 2,007-bp cDNA fragment. This fragment includes 670 amino acids of mMTF-1; however, it is missing the five C-terminal amino acids ( 671 LQLPP 675 ). The mMTF-1 fusion protein contains a His 6 tag, which can be used for protein purification, and a Myc epitope, which can be used to identify the expressed protein by Western immunoblot analysis and for purification by immunoprecipitation.
To construct this expression vector, pc-mMTF-1, which contains the full-length mMTF-1 cDNA (18), was digested with NcoI and ApaI to yield a 2,007-bp cDNA fragment (1 bp 5Ј of the start codon and 19 bp 5Ј of the TAG stop codon). This fragment was inserted into the NcoI and ApaI sites of pGEM-5zf(ϩ) (Promega). The resulting plasmid, designated pG5-mMTF-1, was then digested with ApaI and SpeI to release the original fragment and an additional 18 bp from the multiple cloning site of pGEM-5zf(ϩ). This fragment was directionally cloned into the NheI (SpeI-compatible) and ApaI sites of pcDNA3.1-myc-his (Invitrogen). The final recombinant expression vector is designated pmMTF-1-myc-his.
Expression of both the hMTF-1-VSV and mMTF-1-myc-his mRNAs is under the control of the cytomegalovirus promoter, which allows high level expression in a variety of cells. Transfection of cells with either phMTF-1-VSV or pmMTF-1-myc-his expression plasmids results in the constitutive expression of the MTF-1 fusion proteins.
Two parameters were used to assess the biological activity of the mMTF-1 and hMTF-1 fusion proteins: the ability to activate MT transcription and to specifically bind to MREs. Co-transfection of either MTF-1 expression vector with Ϫ153CAT (see below) results in a significant increase in reporter gene expression in the absence of metals, compared with cells not transfected with pmMTF-1-myc-his or phMTF-1-VSV. This effect has been described previously in cells transfected with MTF-1 expression plasmids and has been attributed to the unusually high levels of MTF-1 (17). When cadmium was added to the culture medium, the level of reporter gene expression significantly increased (p Ͻ 0.05) compared with cells not exposed to cadmium.
Electrophoretic mobility shift assays demonstrated that the mMTF-1-myc-his and phMTF-1-VSV fusion proteins specifically bind to the MRE sequence. In vitro transcribed/translated fusion proteins bound to a 32 P-labeled MRE-containing oligonucleotide, and binding was successfully competed with the unlabeled MRE oligonucleotide. Furthermore, an oligonucleotide that contained a nonfunctional MRE sequence did not compete with the labeled probe (44). The results of the transfection and EMSA analyses with the mouse and human MTF-1 fusion proteins are similar to those previously reported for endogenous MTF-1 (17,19). They indicate that both hMTF-1-VSV and mMTF-1-myc-his are functionally equivalent to the non-fusion form of MTF-1.
A series of reporter plasmids were used to assess the levels of MT transcription. Metallothionein transcription in HEK293 and HeLa cells was determined from the amount of luciferase produced using mouse MT-I-Luc or 4xMREd-Luc reporter constructs. The construction of MT-I-Luc has been described previously (42). The 4xMREd-Luc reporter plasmid was created by excising the concatenated MREs from 4xMREd-OVEC (17) following digestion with XhoI and PvuII, which released a 111-bp fragment. The ends of the fragment were made blunt, and it was subsequently inserted into the SmaI site of pGL3-basic (Promega).
The level of MT transcription in COS-7 cells was established by measuring the level of chloramphenicol acetyltransferase produced using Ϫ42CAT, which contains the minimal mouse MT-I promoter (Ϫ42 to ϩ62); Ϫ153CAT, which contains the region of the mMT-1 gene from Ϫ153 to ϩ62; and MREdЈ 5 CAT, which contains five tandem copies of MREdЈ inserted upstream of TATA box in Ϫ42CAT. Complete descriptions of the MT-CAT reporter genes can be found in Ref. 43, by Dalton et al.
In Vivo 33 P and 32 P Labeling-HEK293 cells were transfected with 5 g of phMTF-1-VSV for 24 h. After transfection, the medium was replaced with phosphate-free DMEM (Invitrogen). After 3 h the medium was replaced with the phosphate-free DMEM containing [ 33 P]orthophosphate (0.1 mCi/ml). Either zinc (100 M ZnCl 2 ) or cadmium (60 M CdCl 2 ) was added to the 33 P-containing medium. Incubation proceeded for 3 h, after which nuclear and cytoplasmic extracts were prepared as previously described (17). The hMTF-1-VSV fusion protein was isolated from nuclear and cytoplasmic extracts by immunoprecipitation using anti-VSV antibody (Sigma). The amount of 33 P incorporated into hMTF-1-VSV was determined following SDS-PAGE by PhosphorImager analysis. COS-7 cells were transfected with 16 g of pmMTF-1-myc-his for ϳ6 h, at which time the medium was removed and replaced with complete DMEM. The cells were then allowed to recover for an additional 48 h. Cells were then washed with phosphate-free DMEM and then incubated in phosphate-free DMEM containing [ 32 P]orthophosphate (0.25 mCi/ml) for 30 min. Metals (50 M CdCl 2 or 100 M ZnCl 2 ) were then added to the culture medium, and the cells incubated an additional 1-4 h. Following this incubation, the mMTF-1-myc-his fusion protein was purified by either nickel affinity chromatography or immunoprecipitation. To determine the affects of inhibiting PKC activity on the level of mMTF-1 phosphorylation, the PKC inhibitor H-7 was applied to cells (100 M) 30 min prior to the addition of cadmium or zinc.
Nickel Affinity Chromatography-Cells were lysed in 50 mM Tris-HCl buffer, pH 8.0, containing 1% Triton X-100, 8 M urea, and 25 mM NaCl. The fusion protein was then purified using Ni-NTA affinity chromatography by first incubating the cell lysate with Ni-NTA resin (Qiagen) at 22°C for 30 min with constant shaking. At the end of this incubation, the Ni-NTA resin was collected and washed twice with 20 mM imidazole buffer. The mMTF-1-myc-his fusion protein was subsequently eluted with 500 mM imidazole.
Immunoprecipitation-Cell lysates were prepared in 50 mM Tris-HCl buffer, pH 8.0, containing 1% Triton X-100, 25 mM NaCl, 2 g/ml aprotinin, 100 g/ml phenylmethylsulfonyl fluoride, and 2 mg/ml pepstatin. Lysates were first pre-treated with a 50% (v/v) slurry of Protein G-agarose (Roche Biomedical) in lysis buffer for 3 h at 4°C with constant agitation. After removing the resin by centrifugation, a fresh aliquot of resin and anti-Myc monoclonal antibody (Invitrogen) was added to the treated lysate. The mixture was incubated for 24 h at 4°C, after which the resin was collected and washed three times with lysis buffer. The washed resin was then combined with SDS-PAGE sample buffer (40) and incubated for 6 min at 100°C, and the supernatant was collected.
mMTF-1-myc-his proteins isolated by affinity chromatography or immunoprecipitation were resolved by SDS-PAGE and subsequently transferred to polyvinylidene difluoride membranes (Millipore Corp). The identification and quantification of the 32 P-labeled mMTF-1-mychis fusion protein was accomplished by autoradiography and Phosphor-Imager analysis (Molecular Dynamics). The location of the fusion protein was determined by Western immunoblot analysis, using either anti-Myc or anti-poly-His 6 antibodies, and enhanced chemiluminescence. Proteins purified by immunoprecipitation were identified using an anti-mMTF-1 antibody.
Analysis of Phosphorylated Amino Acid Residues-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-1myc-his and exposed to metals as described above; however, [ 32 P]orthophosphate was omitted. The fusion protein was purified by immunoprecipitation, using anti-Myc antibodies, and resolved by SDS-PAGE. Identification of phosphorylated amino acid residues was accomplished following Western immunoblot analysis using anti-phosphothreonine, anti-phosphoserine, or anti-phosphotyrosine antibodies (Sigma). The level of mMTF-1 fusion protein expression was determined using anti-mMTF-1 antibody.
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 luciferaseand CAT-based reporter genes.
Luciferase Assay-HEK293 and HeLa were grown on six-well tissue culture dishes in DMEM supplemented with 5% fetal bovine serum for 16 h before transfection. Cells were transiently transfected with 2 g/well mMT-I-Luc or 4xMREd-Luc reporter plasmid, and 0.5 or 1 g/well pCMV-LacZ as a reference. Sixteen hours after transfection, cells were washed and then incubated for an additional 24 h in DMEM containing 5% fetal bovine serum. Following this incubation, the medium was replaced with DMEM containing 0.5% bovine serum albumin. After incubating for an additional 24 h under serum-free conditions, cells were treated with kinase inhibitors. Inhibitor treatments were initiated 30 min prior to the addition of cadmium (60 M CdCl 2 ) or zinc (100 M ZnCl 2 ). Cells were exposed to metals for 4 h in both the presence and absence of the inhibitors. Cell lysates were prepared and luciferase activities measured according to manufacturer's instructions (Promega). A description of the inhibitors and the concentrations used in the assays can be found in Table I.
Chloramphenicol Acetyltransferase Enzyme-linked Immunosorbent Assay-COS-7 cells were plated on to 24-well tissue culture dishes and incubated for 24 h. Cells were washed twice with phosphate-buffered saline and transfected with Ϫ42CAT, Ϫ153CAT, or MREdЈ 5 CAT reporter genes for 3-4 h. Cells were co-transfected with pSV-␤Gal (Promega), to control for transfection efficiency. Following this incubation, the transfection medium was replaced with complete DMEM, and cells were incubated for an additional 24 h.
To determine the effects of inhibiting signal transduction cascades on MTF-1/MRE-mediated MT transcription, chemical inhibitors (Table I) were added to the transfected cells in complete DMEM. Following a 0.5-1-h incubation, no metal, 50 M CdCl 2 , or 100 M ZnCl 2 was added, and cells incubated for an additional 3 h. Extracts were prepared, and CAT concentrations and ␤-galactosidase activities determined by sand-wich enzyme-linked immunosorbent assay using a CAT-ELISA kit (Roche Biotechnologies) and a ␤-galactosidase enzyme assay system (Promega), respectively. All assays were performed in triplicate and normalized to ␤-galactosidase activity.
Electrophoretic Mobility Shift Assays-HEK293 cells were incubated with 100 M ZnCl 2 or 60 M CdCl 2 for 3 h 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 containing core MRE consensus sequence with nuclear extracts as previously described (17). The sequences of the two strands in the MRE oligonucleotide are: cgagggagctctgcacacggcccgaaaagtg (plus strand) and tcgacacttttcgggccgtgtgcagagctccctcgagct (minus strand). The bases in bold designate core MRE sequence.

RESULTS
Phosphorylation of MTF-1 in Vivo-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 (Fig. 1A). Following a 3-h incubation with zinc or cadmium, 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 phospho-MTF-1 were still observed in the cytoplasmic fraction (Fig. 1A). These results confirm that MTF-1 is phosphorylated in both the presence and absence of added metals. In addition, they indicate that exposure to transition metals increases the level of phosphorylation.
To identify which types of amino acid residues are phosphorylated in vivo, nonradiolabeled, immunopurified mMTF-1myc-his was probed with anti-phosphoserine, anti-phosphothreonine, and anti-phosphotyrosine antibodies. Both antiphosphoserine and anti-phosphotyrosine antibodies crossreacted with mMTF-1-myc-his (Fig. 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 what was observed in the in vivo radiolabeling experiments, the protein is constitutively phosphorylated and the level of phosphorylation increases in response to metal exposure (Fig. 1).
Effects of Signal Transduction Cascade Activators and Inhibitors on MT Transcription-Protein motif analysis of MTF-1 indicates that it contains a variety of potential sites that may be phosphorylated via different signal transduction pathways. To determine which signal transduction pathway contributes to the regulation of MTF-1 activity, the effects of kinase inhibitors and activators on MRE-mediated MT transcription were investigated in HeLa, HEK293, and COS-7 cells.
Protein Kinase C-The effects of three PKC inhibitors, staurosporine, H-7, and BIM-I, on zinc-inducible MT transcription were investigated in HeLa cells. Exposure to staurosporine or H-7, in either the presence or absence of zinc, reduced mMT-I promoter activity to levels below that observed in cells not exposed to metal (Fig. 2). Treatment with BIM-I also significantly reduced the level of zinc-inducible mMT-1 transcription. Similar results were obtained with HeLa cells transfected with the 4xMREd reporter gene (Fig. 2). In these studies, however, treatment with staurosporine and H-7 did not reduce the level of reporter gene expression below that observed in cells not exposed to zinc. Thus, in these cells, PKC inhibitors did not affect the basal level of expression. The effect of H-7 on cadmium-and zinc-inducible MT transcription was also examined in COS-7 cells. Zinc-induced -153CAT and MREdЈ 5 CAT reporter gene activity were reduced by 77% (p Ͻ 0.001) and 83% (p Ͻ 0.001), respectively, in the presence of H-7 (Fig. 2). Treatment with H-7 reduced cadmiuminducible transcription to a similar extent, 86% (p Ͻ 0.001) and 74% (p Ͻ 0.005). The level of MT transcription in the presence of metal and H-7 was not significantly different from those of samples containing no metal, or cells transfected with the negative control plasmid -42CAT (p Յ 0.99).
The results from these studies indicate that the MTF-1/ MRE-mediated activation of MT transcription is regulated by PKC-responsive signal transduction pathways. These results are consistent with previous reports demonstrating that exposure of animals or cultured cells to PKC activators (TPA) increases the steady-state level of MT mRNA (7). Exposure of HeLa or COS-7 cells to the PKC activators (Ϫ)-indolactam or TPA, however, did not significantly increase the level of MREregulated MT transcription, even following prolonged exposure (Table II). This suggests that PKC activation is necessary but not sufficient for metal induction.
Casein Kinase II-Mammalian MTF-1s contain more than 11 predicted casein kinase II phosphorylation sites. The effects of the casein kinase II inhibitors DRB and heparin on metalinducible MT transcription were examined in HeLa, HEK293, and COS-7 cells. In HeLa cells, treatment with DRB significantly reduced the level of both cadmium-and zinc-inducible reporter gene expression at all of the concentrations tested (5-50 M) (Fig. 3). In HEK293 cells, however, DRB was an effective inhibitor of cadmium-inducible transcription at concentrations above 10 M, and of zinc-inducible transcription at 50 M (Fig. 3).
Treatment of COS-7 cells with heparin (2 g/ml) significantly inhibited cadmium-inducible reporter gene expression in cells transfected with -153CAT (Fig. 3). This concentration and cytoplasmic (lower panel) extracts were then prepared and hMTF-1 fusion proteins purified by immunoprecipitation. Immunocomplexes were resolved by SDS-PAGE, transferred onto nitrocellulose membranes, and visualized by PhosphorImager analysis. B, COS-7 cells were transfected with 16 g of pmMTF-1-myc-his and exposed to 100 M zinc for 4 h or no metal (NM). The fusion protein was purified by immunoprecipitation and resolved by SDS-PAGE. Identification of MTF-1-myc-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.

FIG. 2. Effect of PKC inhibitors on metal-inducible transcription.
HeLa cells were transfected with mMTI-Luc and CMV-Lac Z (upper panel) or 4xMREd-Luc and CMV-Lac Z (lower panel) and then serum-starved for 24 h. Cells were then exposed to the PKC inhibitors staurosporine, H-7, or BIM-I for 30 min prior to the addition of 100 M zinc. Following a 4-h incubation, cells were harvested and processed for luciferase and ␤-galactosidase assays. The level of reporter gene activity was also measured in cells exposed to inhibitors, but not to zinc (NM), or cells not exposed to metals and inhibitors (No Treatment). Table, COS-7 cells were transfected with Ϫ42CAT, Ϫ153CAT or MREdЈ 5 CAT, and pSV-␤Gal, for 3 h. After a 24-h recovery period, cells were pre-incubated with the PKC inhibitor H-7 for 30 min, and then cadmium (50 M) or zinc (100 M) was added. Following a 4-h incubation, cells were harvested and CAT concentrations and ␤-galactosidase activities determined. 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) were also determined. The data presented are the mean values from three independent experiments with the standard deviations. *, significant change in reporter gene expression (p Ͻ 0.05). of heparin, however, did not affect zinc-inducible expression. In addition, heparin did not affect cadmium-or zinc-inducible expression of the MREdЈ 5 CAT reporter. It may require higher concentrations of this inhibitor to affect MT transcription.
Treatment of cells with spermidine, a casein kinase II acti-vator, slightly activated reporter gene expression in COS-7 cells, in the absence of metals (Table II). The results from the inhibitor and activator studies using the three cell lines suggest that casein kinase II may modulate the activity of MTF-1.
Tyrosine Kinase-Inhibition of tyrosine kinase activity with herbimycin A (5 M) significantly reduced the level of cadmium-and zinc-inducible reporter gene activity in both HEK293 and HeLa cells (Fig. 4). 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 and Calmodulin-Calcium is required for the activation of many kinases; therefore, the effects of intracellular calcium chelators and calcium ionophores on MRE/MTF-1-mediated transcription were investigated. Exposure to the intracellular calcium chelator, BAPTA-AM, reduced the level of cadmium-and zinc-inducible 4xMREd promoter activity to levels below that observed in cells not exposed to metals (Fig. 5). This effect is not attributed to chelation of cadmium or zinc by BAPTA-AM, which is known to have very low affinity for these metals. Consistent with these results, exposure of COS-7 cells to the calcium ionophore A-23187 markedly increased the level of -153CAT and MREdЈ 5 CAT activity in the absence of added metals (Table II). These results suggest that calcium-mediated signal transduction pathways are involved in the activation of MTF-1.
Electrophoretic Mobility Shift Assay-Exposure to PKC, tyrosine kinase, or casein kinase II inhibitors decreases the level of MTF-1/MRE-mediated gene expression. To investigate the relation between kinase inhibition and the binding of MTF-1 to the MRE, EMSAs were performed. Pretreatment of HEK293 cells with kinase inhibitors did not significantly affect the cadmium-or zinc-inducible binding of MTF-1 to the MRE (Fig. 6). In contrast, the level of MTF-1-MRE complex formed increased in the presence of inhibitors. These results suggest that the regulation of MRE-mediated transcription by these kinases, or their related signal transduction pathways, is not mediated by inhibiting the DNA binding activity of MTF-1. This does not, however, exclude effects on cytoplasmic nuclear transport of MTF-1, or chromatin accessibility of template promoter targets.
Effect of Protein Kinase C Inhibition on MTF-1 Phosphorylation-We observed that the inhibition of PKC activity prevents both cadmium-and zinc-inducible MT transcription in the three cell lines examined. Therefore, the effect of inhibiting PKC activity on the phosphorylation of MTF-1 was investigated.
When PKC activity was inhibited in zinc-treated COS-7 cells with H-7, the amount of 32 P incorporated into mMTF-1-myc-his increased, compared with zinc-treated cells not exposed to H-7 (Fig. 7). Similarly, the amount of phosphorylation on the serine residues significantly increased for the zinc-treated cells exposed to H-7 (Fig. 7). Similar results were obtained when cells were exposed to cadmium and H-7 (results not shown). In control experiments, H-7 treatment alone caused an increase in the level of MTF-1 phosphorylation. It should be noted that treatment with H-7 did not affect the steady-state level of MTF-1 (Fig. 7).

FIG. 3. Effect of casein kinase II inhibition on metal-inducible transcription. HEK293 (upper panel) or HeLa (lower panel) cells were
transfected with 4xMREd-Luc and CMV-LacZ and then serum-starved for 24 h. Cells were exposed to various concentrations (0, 5, 10, and 50 M) of the casein kinase II inhibitor DRB for 30 min prior to the addition of either 100 M zinc or 60 M cadmium. Following a 4-h incubation, cells were harvested and processed for luciferase and ␤-galactosidase assays. The level of reporter gene activity was also measured in cells exposed to 50 M DRB but not metal (DRB Only), or cells not exposed to metals and inhibitor (No Treatment). Table, COS-7 cells were transfected with CAT reporter genes and pSV-␤Gal, for 3 h. After a 24-h recovery period, cells were pre-incubated with the casein kinase inhibitor, heparin for 30 min, and then cadmium (50 M) or zinc (100 M) was added. Following a 4 h incubation, cells were harvested, and CAT concentrations and ␤-galactosidase activities determined. 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) were also determined. The data presented are the mean values from three independent experiments with the standard deviations. *, significant change in reporter gene expression (p Ͻ 0.05).

DISCUSSION
It is well documented that metal-activated MT transcription is mediated primarily through the MRE-binding transcription factor MTF-1. Although this protein was discovered almost a decade ago, the mechanism by which metals activate transcription via MRE/MTF-1 interactions is still largely unresolved. Previous studies indicate that metal-inducible transcription requires several processes: nuclear translocation, DNA binding, and transcriptional activation. 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 (17). The binding of MTF-1 to MREs requires the occupancy of zinc finger 1 with zinc (24). These observations lead to the hypothesis that the biological activity of MTF-1 may be homologous to that of the yeast transcription factors ACE-1 and AMT-1, which function as copper sensors, binding metal to activate transcription (46). The sensor model of MTF-1, however, cannot explain the activation of MTF-1/MRE-mediated transcription by other metals or stressors. Occupancy of the zinc fingers by other transition metals even inhibits MTF-1binding to DNA in vitro (47). Thus, zinc is required for the formation of zinc fingers and DNA binding. DNA binding, how-ever, is not synonymous with transcriptional activation. This is supported by the observations that metal-inducible, MTF-1/ MRE-mediated transcription is completely blocked with PKC and casein kinase II inhibitors (Figs. 2 and 3), but these agents do not inhibit MTF-1 binding to DNA (Fig. 6).
Recent reports identified a second component in the activation of mammalian MT transcription: the metal-and stressinducible translocation of MTF-1 from the cytoplasm to the nucleus (42,48). Our studies demonstrate that, following metal exposure, the phosphorylated form of MTF-1 accumulates in the nucleus (Fig. 1). It is reasonable to hypothesize that modifications of specific residues could control nuclear translocation and transcriptional activation. It has been reported that, in response to IL-12 induction, STAT-4 is phosphorylated at both tyrosine and serine residues, which leads to its activation and nuclear translocation (49). The unique tyrosine kinase phosphorylation site in human MTF-1, Tyr 140 , is contiguous with the MTF-1 nuclear localization sequence 133 KRKEVKR 139

FIG. 4. Effect of tyrosine kinase inhibition on MT transcription.
HeLa (upper panel) or HEK293 (lower panel) cells were transfected with 4xMREd-Luc and CMV-LacZ and then serum-starved for 24 h. Cells were exposed to the tyrosine kinase inhibitor herbimycin for 30 min prior to the addition of 100 M zinc or 60 M cadmium. Following a 4-h incubation, cells were harvested and processed for luciferase and ␤-galactosidase assays. The level of reporter gene activity was also measured in cells that were not exposed to metal (NM), or cells not exposed to the inhibitors (No Treatment). The data presented are the mean values from three independent experiments with the standard deviations.

FIG. 5. Effect of intracellular calcium chelators on MT transcription.
HeLa cells were transfected with 4xMREd-Luc and CMV-LacZ and then serum-starved for 24 h. Cells were exposed to the intracellular calcium chelator BAPTA-AM (ϩ) for 30 min prior to the addition of 100 M zinc or 60 M cadmium. Cells were also exposed to metal, in the absence of BAPTA-AM (Ϫ). Following a 4-h incubation, cells were harvested and processed for luciferase and ␤-galactosidase assays. The data presented are the mean values from three independent experiments with the standard deviations, and are presented as the fold increase in reporter gene activity relative to that observed in cells that were not treated with metals or chelator (No Treatment).
FIG. 6. Effect of kinase inhibitors on the DNA binding activity of MTF-1. HEK293 cells were incubated with the indicated kinase inhibitors for 30 min prior to exposure to 100 M zinc or 60 M cadmium, or no metal (NM). Following a 3-h incubation, nuclear extracts were prepared and analyzed by EMSA 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 (No Inhibitor). The location of the MTF-1/MRE oligonucleotide complex is indicated by the arrow. (42). 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 proteins.
The binding of MTF-1 to DNA is highly correlated with transcriptional activation; however, a mechanistic link has not been described. Results presented in this report and previous studies indicate that DNA binding can be disparate from transcriptional activation (32,42). It has been clearly demonstrated that, under conditions where MT transcription is inhibited, DNA binding still occurs. Furthermore, the binding of MTF-1 to MREs increases. This increase may be caused by a change in the affinity of MTF-1 for DNA following a change in the level of phosphorylation. It may also be the result of an increase in concentration of nuclear-MTF-1 caused by either a higher level of nuclear transport or de novo synthesis of 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 metalinducible 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. 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 (Fig. 7). This is surprising at first glance, because H7, staurosporine, and BIM-I are potent inhibitors of metal-inducible MT transcription (Fig. 2). Although contradictory, these results are consistent with a model in which metal-activated transcription is controlled by a change in the phosphorylation pattern of MTF-1 that also involves dephosphorylation. A dephosphorylation activation mechanism is not without precedence. A similar mechanism is responsible for activating the binding of c-jun to the AP-1 promoter elements. The ability of c-Jun to bind DNA is a result of the activation of a phosphatase by PKC, which dephosphorylates residues adjacent to the DNA-binding domain, subsequently allowing AP-1 binding (50). Although c-Jun is dephosphorylated prior to DNA binding, it must remain phosphorylated at serine 63 and 73 to maintain proper function (51,52). A similar requirement could explain why serine residues in MTF-1 remain phosphorylated in the presence of zinc and cadmium.
In the dephosphorylation model, only transcriptional activation of MTF-1 is controlled by the removal of specific phosphorylated residues. The total level of MTF-1 phosphorylation may be greater following exposure to metals or stress, as shown in Fig. 1; however, specific residues are dephosphorylated to activate transcription. The dephosphorylation of MTF-1 is performed by a "MTF-1 phosphatase," whose activity may be controlled by upstream regulatory proteins.
The inhibitor studies indicate that metal-activated MT transcription via MTF-1/MRE interactions is controlled by signal transduction pathways that involve PKC, casein kinase II, tyrosine kinase, and calcium. The activation of the signal transduction pathways associated with these kinases by metals and other environmental stressors has been well documented (36,53,54). The activity of casein kinase II is also affected by environmental stressors. In response to oxidative stress, casein kinase II translocates to the nucleus, and the enzymatic activity of the kinase increases (54,55). The stress-inducible increase in heme oxygenase-1 transcription is dependent, in part, on the activity of tyrosine kinase (56).
The ability of PKC inhibitors and activators to affect the steady-state levels of MT mRNA, and the results presented in this report indicate that the MAP kinase signaling pathway directly regulates MT transcription, through MTF-1/MRE interactions. Transition metals and oxidative stress have been shown to modulate the activity of several members of the MAP kinase signaling cascade (57,58). Recently, c-Jun N-terminal kinase has been shown to contribute to the regulation of metalinducible MTF-1 transcriptional activation (32). In addition, p38 and extracellular signal-regulated kinase 1/2 are also involved in controlling metal-inducible transcription. 2 The observation that multiple signal transduction pathways contribute to the ability of MTF-1 to regulate MT transcription provides an explanation to the observation that deletion of various regions in MTF-1 affects its ability to activate transcription. However, no single domain could be assigned as the "metal-activation" domain (23). Because potential targets of the various signal transduction pathways are located throughout the protein, it is likely that changing the phosphorylation status of several of these residues would control metal-activated transcription.
Exposure to metals, organic chemicals, physical stress (heat, cold, radiation), intracellular damage, and physiologic agents (hormones, second messengers) causes an increase the steadystate level of MT mRNA (6, 59 -62). In addition, several of these non-metal stressors activate MT transcription via MREs (11,28,30,63). 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-1regulating signal transduction pathway could increase MT transcription. Furthermore, agents that activate parallel signal transduction pathways that converge with PKC, casein kinase II, and tyrosine kinase-regulated pathways will affect MT transcription.
Although originally isolated as the transcription factor that controls MT expression, MTF-1 has been shown to be essential A, Cells were exposed to 100 M zinc or no metal and [ 32 P]orthophosphate for 4 h 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, transferred onto a nitrocellulose membrane, and the amount of 32 P-labeled MTF-1 determined by PhosphorImage analysis. B, cells exposed to 100 M zinc or no metal for 4 h in the presence (H-7) or absence (NI) of H-7. The fusion protein was purified by immunoprecipitation and subsequently transferred to nitrocellulose membranes. The identification of MTF-1-myc-his (MTF-1), and phosphorylated serine (P-ser) was accomplished by Western immunoblot analysis using anti-mMTF-1 and antiphosphoserine antibodies, respectively. The arrowhead indicates the location of the MTF-1 fusion protein.
for normal liver development (64). Recently, several non-MT genes have been identified that contain MREs in their upstream regulatory region and the level of expression of which is affected by the loss of MTF-1 (65). Several of these genes have roles in mediating the stress response; however, for others, intracellular stress has not reported to affect their expression. These observations suggest that MTF-1 may be a more general transcription factor, and not limited to controlling stress response genes. The ability of several signal transduction pathways to affect MTF-1 activity may contribute to its activity in regulating liver development and non-stress response genes.
A detailed analysis of the specific residues in MTF-1 that are modified following exposure to zinc, cadmium or other stressors, and various kinase inhibitors will help to define how each signaling pathways affects the level of MTF-1 phosphorylation. The effects of individual and combinatorial site-specific mutations on the ability of MTF-1 to regulate MT transcription will be required to determine the precise mechanism by which metals and other stressors activate transcription via MTF-1/ MRE interactions and the functional relationship between the phosphorylation of specific residues and the in vivo activity of MTF-1.