- Email: [email protected]
Induction of the Heme Oxygenase-1 Gene by Metalloporphyrins
Archives of Biochemistry and Biophysics Vol. 380, No. 2, August 15, pp. 219 –227, 2000 doi:10.1006/abbi.2000.1921, available online at http://www.idealibrary.com on
Induction of the Heme Oxygenase-1 Gene by Metalloporphyrins Ying Shan, Joyce Pepe, Tze Hong Lu, Kimberly K. Elbirt, Richard W. Lambrecht, and Herbert L. Bonkovsky 1 Department of Medicine and Department of Biochemistry and Molecular Biology, University of Massachusetts Medical School and The Center for Study of Disorders of Iron & Porphyrin Metabolism, UMass Memorial Health Care, Worcester, Massachusetts 01655
Received December 30, 1999, and in revised form May 5, 2000
Induction of expression of heme oxygenase-1 (HO-1) has been studied in primary cultures of chick embryo liver cells and in the LMH line of avian hepatoma cells. Cells were transiently transfected with selected constructs containing portions of the 5ⴕ-untranslated (promoter) region of the HO-1 gene linked to luciferase as reporter gene. LMH cells that had been stably transfected with selected wild type or mutant constructs were also studied. Metalloporphyrins, especially Fe protoporphyrin (heme) and Co protoporphyrin strongly induced luciferase expression in both types of transfected cells. Low concentrations of Zn mesoporphyrin, an inhibitor of HO activity, exerted a synergistic effect on heme-, but not Co protoporphyrin-dependent induction. The antioxidant and OSH donor N-acetyl cysteine had little effect on the metalloporphyrin-dependent inductions of HO-1, in contrast to its marked inhibitory effect on the sodium arsenite-dependent induction of the HO-1 gene. Deletional analysis showed that the key element(s) required for the metalloporphyrin-dependent induction of HO-1 is located between ⴚ3.6 and ⴚ5.6 kb upstream of the transcription starting point. Data from electrophoretic mobility shift and site-directed mutagenesis experiments excluded a role for consensus AP-1 binding elements at ⴚ1576, ⴚ3647, or ⴚ4578 in the inductions produced by heme or Co protoporphyrin. © 2000 Academic Press
Key Words: cobalt protoporphyrin; enhancer; gene expression; heme; heme oxygenase-1; oxidative stress; promoter.
1 To whom correspondence should be addressed at University of Massachusetts Memorial Health Center, Department of Medicine, Room S6-737, 55 Lake Avenue, North, Worcester, MA 01655. Fax: (508) 856-3981. E-mail: [email protected].
0003-9861/00 $35.00 Copyright © 2000 by Academic Press All rights of reproduction in any form reserved.
Heme oxygenases (HO) (EC 1.14.99.3) are ubiquitous enzymes that catalyze the breakdown of heme (iron-protoporphyrin IX) by the specific oxidative cleavage of the ␣-methene bridge. The products of the reaction are biliverdin, iron, and carbon monoxide (CO); the carbon atom of the CO formed is derived exclusively from the ␣-methene carbon of the heme macrocycle. Two major forms of HO, called HO-1 and HO-2, have been described. HO-1 is detectable in many tissues and can be induced to high levels, mainly by increases in gene transcription rates. Heme and other metalloporphyrins, especially Coand Mn-protoporphyrin (MnPP), are inducers of HO-1, as are several transition metals (e.g., Cd, Co, Fe), arsenicals (e.g., sodium, arsenite, phenylarsine oxide), agents that deplete intracellular levels of GSH and other thiols, and physical forces that increase oxidative or other stress on cells (e.g., UV light, heat shock). Indeed another name for HO-1 is heat shock protein (hsp) 32, and HO-1 is now used widely as a harbinger of cellular stress responses (for reviews, see Refs. 1 and 2). HO-2 is an homologous protein of ⬃36 kDa molecular mass, which is also widely distributed in many tissues in the body, is present in highest concentration in selected areas of the brain and testes of mammals, and is essentially uninducible. HO-3 has a lesser degree of homology to the other forms and much lower specific activity; its principal role may actually be to bind or transport heme within cells. The importance of both HO-1 and HO-2 in heme and iron metabolism has been established by recent descriptions of abnormal phenotypes in mice with targeted gene disruptions of one or the other form of HO (3–5). In addition, a Japanese child with developmental delay, seizures, hemolytic anemia, and renal dysfunction was recently reported to have markedly reduced levels of HO-1, 219
220
SHAN ET AL.
due to compound heterozygous defects in the HO-1 genes inherited from his parents (6). The mechanisms whereby chemical and physical perturbations increase expression of the HO-1 gene are starting to be elucidated. For example, recent work in our laboratory showed that increases in the major mitogen-activated protein (MAP) kinases, c-Jun N-terminal kinase (Jnk), extracellular regulated kinase (Erk), and p38, underlie the induction produced by sodium arsenite, cadmium, or heat shock. An important role for consensus AP-1 binding elements, several of which are found in the 5⬘-untranslated region (UTR) of the HO-1 gene, was established by site-directed mutagenesis and electrophoretic mobility shift assays (EMSA) (7). In these studies, heme did not produce any increase in MAP kinase activities, although it too induced both the endogenous HO-1 gene and the HO-1 5⬘-UTR-Luc reporter gene constructs studied. Others have provided evidence that other kinases (e.g., protein kinases A and C) may also play a role in HO-1 induction (8, 9). The functions and mechanisms of action of inducers of HO-1 are dependent upon many factors, including the species of organism and the organs and cell-types studied. For example, although high concentrations (50 –100 M) of heme may increase endothelial cell injury in some rat or human organs, previous work from our laboratory in normal avian hepatocytes in primary culture (10) or the LMH line of avian hepatocellular carcinoma cells (7) indicated that heme and other metalloporphyrins induce HO-1 by nontoxic, nonstress dependent pathways. Chick embryo liver cells (CELC) provide a simple, cheap, robust, and relevant system for studying heme and porphyrin metabolism, including the regulation of HO-1 expression. CELCs maintain normal levels and inducibility of both 5-aminolevulinate synthase, the rate-controlling enzyme of heme synthesis, and of HO, the rate-controlling enzyme of heme breakdown. In addition, the kinetics of the heme synthetic pathway in CELC are closely similar to that of normal human liver, a property not shared by rodent liver models (11, 12). We have also found that LMH cells provide a continuous cell line that retains the desirable properties of normal levels and regulation of the rate-controlling steps of heme metabolism (13, 14). Among the potential additional advantages of LMH cells are that they are a single, homogeneous cell type, that unlimited numbers can be obtained, and that stably-transfected lines can be established. In this paper, we report studies that characterize the metalloporphyrin-dependent induction of HO-1 in both CELC and LMH cells, and that begin to delineate portions of the 5⬘-UTR of the HO-1 gene that are central to the inductive response.
MATERIALS AND METHODS Materials. All tissue culture dishes were from Corning Glass Inc. (Corning, NY). Culture flasks were from Falcon, VWR Scientific (Bridgeport, NJ). Ferric (Fe 3⫹)-protoporphyrin IX 䡠 Cl (heme), chromium (Cr 3⫹)-mesoporphyrin IX 䡠 Cl, manganese (Mn 3⫹)-protoporphyrin IX 䡠 Cl, cobalt (Co 3⫹)-protoporphyrin IX 䡠 Cl, tin (Sn ⫹4)-mesoporphyrin IX 䡠 2Cl, and zinc (Zn 2⫹)-mesoporphyrin IX were from Porphyrin Products (Logan, UT). Chloroform, isopropyl alcohol, acrylamide, and bis-acrylamide were from Fisher (Pittsburgh, PA). Waymouth’s MB 752/1 media, Williams’ E media, geneticin, gelatin, fetal bovine serum, and calf serum were from Gibco BRL, (Grand Island, NY). -Mercaptoethanol, bovine serum albumin (BSA), EDTA, EGTA, formaldehyde (37% v/v), formamide, glycylglycine, penicillin/streptomycin, phenyl methyl sulfonyl fluoride, piperacillin, sodium arsenite, sodium dodecyl sulfate, trypsin, dithiothreitol (DTT), 3,3⬘,5-triodo-L-thyronine (T-3), spermine, and o-nitrophenyl-D-galactopyranoside were from Sigma (St. Louis, MO). Dexamethasone was from Gensia Pharmaceuticals (Irvine, CA). DNA Maxiprep kits and gel extraction kits were from Qiagen (Santa Clarita, CA). All 32P-radionucleotides were from NEN Life Science Products (Boston, MA). LMH cells were a generous gift from D. L. Williams, Department of Pharmacological Sciences, SUNY (Stony Brook, NY). The pPGK- gal plasmid was a gift from P. Dobner (Department of Molecular Genetics and Microbiology, University of Massachusetts Medical School, Worcester, MA). The primers GLprimer2 and RVprimer3, Wizard plasmid DNA preparation kits, pGL3 Promoter plasmid, and Luciferase Assay Reagent were purchased from Promega (Madison, WI). DNA sequencing was performed by the sequencing facility of the Dana Farber Cancer Institute (Boston, MA) or by the Nucleic Acids Facility, University of Massachusetts Medical School. All chemicals were of the highest purity available. Cell cultures and treatments. Primary chick liver cell cultures were prepared with minor modifications of a previously described method (15). In brief, 17-day-old Barred Rock chick embryos were killed by decapitation, and their livers were immediately removed and placed in Hanks’ balanced salt solution. Livers were chopped under sterile conditions and then transferred to a 125-mL sterile flask and incubated for a total of 20 min in a 37°C with trypsin and DNAse. The cell suspension was then washed with in a buffer containing ammonium chloride to lyse the erythrocytes and resuspended repeatedly by shaking with Williams’ E medium until the total desired volume of medium (10 mL/liver) had washed over the cells. The cells (3 mL/dish) were then plated on 6-well culture dishes. LMH cells were maintained in Waymouth’s MB 752/1 medium, supplemented with 100 U/mL penicillin, 100 g/mL streptomycin, and 10% (v/v) fetal bovine serum. The cells were routinely passaged twice a week (7). For some experiments, 1 g/mL piperacillin was added to the culture medium. All metalloporphyrins were dissolved in Me 2SO; sodium arsenite was prepared as described (13, 16) and stored at ⫺20°C until used. Plasmid construction. The chick HO-1 gene expression plasmids used to transfect CELC and LMH cells were constructed as follows. The construction of plasmids pcHO7.1-Luc, pcHO5.6-Luc, pcHO4.6Luc, pcHO3.6-Luc, pcHO2.5-Luc, pcHO1.6-Luc, pcHO0.6-Luc, and pcHOTATA-Luc has been described previously (7). The pcHOPB-Luc reporter plasmid was constructed by cloning 913 base pairs (⫺4.6 to ⫺3.6 kbp) of the chick HO-1 proximal promoter into the pGL3 with promoter vector. Plasmid pcHO7.1-Luc was partially digested with PmlI, the ends were filled in using the Klenow fragment, redigested with BglII, and then ligated to SmaI/BglII sites upstream of the luciferase reporter gene. The plasmids pcHOPN-Luc and pcHONBLuc were constructed by cloning the PmlI/Nhel fragment (⫺4.6 to ⫺4.1 kbp) or the NheI/BglII fragment (⫺4.1 to ⫺3.6 kbp) of pcHO4.6Luc, respectively, into the pGL3 promoter vector. The orientation of the inserts was verified by sequencing using the commercial primers GLprimer2 and RVprimer3 from Promega.
HO-1 INDUCTION BY METALLOPORPHYRINS Transfections. Transient and stable transfections were carried out by the calcium phosphate precipitation technique (16, 17). For transient transfections, about 40 h after initial plating of the cell cultures, the medium was carefully removed and 2 mL of fresh Williams’ E medium [containing dexamethasone, T-3, glutamine, and 10% (v/v) calf serum] was added. Two hours later, transfections were performed for 4 –5 h by the calcium phosphate method. Transfections were ended by removing the transfection medium and gently washing the cells twice with a buffer containing 0.01 M sodium phosphate (pH 7.4) and 0.15 M NaCl (PBS) to remove the extracellular DNA. Then fresh Williams’ E medium (with dexamethasone, T-3, and glutamine, but without serum) was added. Typically, some of the plates received treatments at this time. For stable transfections, LMH cells were plated into 6-cm tissue culture dishes (5 ⫻ 10 5/6-cm plate) the day before transfection. Cells were fed with 5 mL of Waymouth’s medium supplemented with 10% fetal bovine serum 2 h prior to transfections. LMH cells were exposed to the DNA-CaPO 4 precipitate for 16 h. The DNA mixture was composed of 10 g of the chick HO-1 expression plasmid and 2 g of pSV2neo. The precipitate was removed, and the cells were cultured in Waymouth’s complete medium for 24 h the transfected cells were then split (1:15) and plated on 10-cm dishes in Waymouth’s complete medium. Geneticin (G418 sulfate) was added 24 h later to a final concentration of 100 g/mL, and resistant colonies were selected over a 2–3 week period. Resistant colonies were transferred to 48well dishes and propagated in the presence of G418. Colonies containing luciferase activity were seeded in 6-well dishes at a density of 6 ⫻ 10 5 cells/well. Forty hours after seeding, cells were incubated at 37°C in 2 mL of serum-free medium. After 2 h, some of the plates received treatment. Assessment of reporter gene activity. Reporter gene expression and activation was assessed by quantitation of luciferase activity, normalized to -galactosidase activity, and protein content. For luciferase activities, transfected cells were washed twice with PBS, and harvested by scraping in 250 L of glycylglycine harvest buffer (25 mM glycylglycine, pH 7.8, 15 mM magnesium sulfate, 4 mM EGTA, 1 mM DTT). Cells were lysed by three freeze–thaw cycles (3 min in liquid nitrogen followed by 3 min at 37°C), followed by a 10-min centrifugation at 14,000g at 4°C. The supernatant was retained, and 15-L aliquots of cell lysate were used for each assay. Luciferase activity measurements were carried out using a Monolight 2010 luminometer (Analytical Luminescence Laboratories, Ann Arbor, MI). Relative light units were recorded. For -galactosidase activities, 15 L of cell lysate was added to 200 L of Z Buffer (60 mM Na 2HPO 4, 40 mM NaH 2PO 4, 10 mM potassium chloride, 1 mM magnesium sulfate, 50 mM -mercaptoethanol, adjusted to pH 7.0) and 100 L of 5 mg/mL o-nitrophenyl-D-galactoside (ONPG) dissolved in 0.1 M potassium phosphate, pH 7.0. The samples were mixed and incubated at 37°C for 1.5–2 h. The absorbance at 420 nm was measured (18). Protein concentrations were determined from absorbance at 562 nm measured by the bicinchoninic acid method on a Spectronic GENESYS 2 spectrophotometer, using BSA as standard. Site-directed mutagenesis of the chick HO-1 reporter gene construct. Site-directed mutagenesis of pcHO5.6-Luc was carried out using Quick Change mutagenesis kits from Stratagene (La Jolla, CA) following the manufacturer’s instructions. pcHO5.6-Luc was used as template, and the AP-1 elements located ⫺1576, ⫺3647, and ⫺4578 bp from the transcription start site in pcHO5.6-Luc were mutated. The oligonucleotides 5⬘GCAGA GCAAG ACAGGAAAAG CATGGCTTCG TCAGGCTGGG AGCGCTGAG3⬘ and 5⬘CTCAGCGCTC CCAGCCTGAC GAAGCCATGC TTTTCCTGTC TTGCTCTGC3⬘ were used as primers for the mutagenesis reaction. After the mutagenesis reactions were completed, the reaction products were introduced into XL1-Blue competent cells (Stratagene). The mutants were confirmed by DNA sequencing (UMass Nucleic Acid Facility).
221
DNA/protein interaction analysis. Extracts of nuclear proteins from cultured cells were prepared as described previously (17, 19). In a typical electrophoretic mobility shift assay (EMSA), 6 g of the nuclear protein were incubated at 37°C for 30 min with 32P-labeled 24-mer oligonucleotide probe (20,000 cpm), containing the AP-1 consensus sequence (5⬘-AGCATGGCTGAGTCAGGCTGGGAG-3⬘) [the consensus sequence is underlined]. The DNA-binding reaction was carried out in a reaction buffer containing 10 mM Tris–HCl pH 7.5, 1 mM MgCl 2, 1 mM EDTA, 1 mM DTT, 50 mM NaCl, 1.5 g dI– dC, 3 g BSA, and 12% (v/v) glycerol, either with or without competitors. The competitors used were unlabeled AP-1 consensus oligonucleotides or mutant AP-1 oligonucleotides. Proteins in the reaction mixture were then separated by polyacrylamide gel electrophoresis (PAGE) (4.5% polyacrylamide gel run at 100 V for 1 h). The gel was dried and scanned by PhosphorImager (17). Statistical analysis of data. Experiments were repeated two to four times; except for gel shift assays, every experiment included at least triplicate samples for each treatment group. Representative results from single experiments are presented. Statistical analyses were performed with JMP 3.0.2 software (SAS Institute, Cary, NC). For normally distributed data, differences in mean values were assessed by analysis-of-variance techniques (ANOVA), with the Tukey–Kramer correction for multiple pair-wise comparisons, or Dunnett’s test versus a control. For experiments with non-normally distributed data, the Wilcoxon/Kruskal–Wallis (rank-sum) test was used. P values ⬍0.05 were considered significant.
RESULTS
Induction of Transfected Heme Oxygenase-1 Promoter Luciferase Reporter Constructs by Metalloporphyrins Some metalloporphyrins, especially heme (FePP) and CoPP, gave significant induction of HO-1 promoter/reporter constructs in both CELCs and LMH cells (Figs. 1A and B). CoPP had the greatest effect on reporter gene expression with a 5-fold or 42-fold increase in normalized luciferase activity in CELCs or LMH cells, respectively. Lesser degrees of induction were observed with this construct after exposure to heme (4.2-fold or 16-fold in CELC or LMH cells). CrMP and MnPP produced slight increases (1.8-fold in CELC but not in LMH cells); however, no induction was observed for SnMP or ZnMP in either CELCs or LMH cells. Optimization of Heme- or CoPP-Mediated Induction of pcHO-Luc Further analysis was performed on the heme- and CoPP-mediated increases in reporter gene expression in cells transfected with either pcHO5.6-Luc or pcHO7.1-Luc. Dose–response and time course experiments established that the most effective dose for heme and CoPP were 10 M in transiently transfected CELCs or in stably transfected LMH cells. Peak levels of induction occurred at 10 –15 h for heme, after which the luciferase activities fell. For CoPP, peak induction occurred at 15 h (Fig. 2). Therefore, these conditions were used to further investigate the mechanism of heme and CoPP induction of HO-1 gene transcription.
222
SHAN ET AL.
an inhibitor of HO (20). We therefore added low concentrations (10 –200 nM) of ZnMP alone or with CoPP or heme (5 or 10 M) to transfected CELC or LMH cells. As shown in Fig. 3, ZnMP alone had no effect on luciferase activities nor did it affect induction produced by CoPP. In contrast, ZnMP produced the expected synergism of induction by heme, to levels similar to those produced by CoPP. Effect of N-Acetylcysteine (NAC) on the Metalloporphyrin-Dependent Induction of HO-1 Promoter-Luc Reporter Activity Effects of pretreatment with NAC on normalized luciferase activity from heme-, CoPP-, and arsenitetreated LMH cells are shown in Fig. 4. NAC did not significantly affect the induction by CoPP or heme, but markedly abrogated arsenite-mediated induction of the reporter gene activity. Deletion Analysis of the Chick HO-1 Gene Promoter/ Enhancer
FIG. 1. Effects of metalloporphyrins on reporter gene activities. (A) Effects of metalloporphyrins on pcHO5.6-Luc activity in transiently transfected CELCs. Transient transfections were carried out by the calcium phosphate method. Following transfections, cells were washed twice with PBS and 2 mL of serum-free medium was added, after which some cells were treated with 1 L/mL Me 2SO or with 10 M cobalt protoporphyrin (CoPP), chromium mesoporphyrin (CrMP), heme, manganese protoporphyrin (MnPP), tin mesoporphyrin (SnMP), or zinc mesoporphyrin (ZnMP). After 15 h, cells were harvested and luciferase assays and -galactosidase assays were carried out, as described under Materials and Methods. Luciferase activities were normalized to -galactosidase activities and protein content. (B) Effects of metalloporphyrins on pcHO7.1-Luc activity in stably transfected LMH cells. Monolayers of LMH cells, stably transfected with plasmid pcHO7.1-Luc, were treated with 1 L/mL Me 2SO or with 10 M CoPP, CrMP, heme, MnPP, SnMP, or ZnMP in serum-free medium. After 15 h, cells were harvested and lysed, and assays were performed on 15 L of lysate. Luciferase activities were normalized to protein content. Data are presented as means ⫾ SE, n ⫽ 12 for Me 2SO, CoPP, and heme; n ⫽ 3 for the other treatments. a indicates an increase over no treatment, P ⬍ 0.05; b indicates an increase over heme treatment, P ⬍ 0.05.
A Low Concentration of ZnMP Synergizes the HemeDependent Induction of the Reporter Gene We postulated that the difference in efficacy of upregulation of the luciferase reporter gene activity by CoPP vs heme is because heme can be catabolized by HO, whereas CoPP cannot. Therefore, we reasoned that, if we added an inhibitor of HO, that itself does not up-regulate the HO-1 gene, with heme, we might attain activities of luciferase similar to those produced by CoPP alone. We previously showed that ZnMP is such
In an effort to identify the metalloporphyrin-responsive regulatory elements within the upstream 5⬘-flanking sequences of the chick HO-1 gene, we performed deletional analysis of the entire distal upstream 7.1-kb region. Increasing lengths of the 7.1 kb upstream region (Fig. 5), including the promoter and distal enhancers, were linked to a luciferase reporter gene, and transiently or stably transfected into CELC or LMH cells and assayed for luciferase activity in response to metalloporphyrins. We observed that the key elements required for the metalloporphyrin-dependent induction of the luciferase reporter are located between ⫺3.6 and ⫺5.6 kb of the 5⬘-UTR of the HO-1 gene (Fig. 5). To further localize the promoter/enhancer activity, three subfragments (PB, PN, and NB) were cloned into the vector pGL3-promoter, and initially analyzed by transient expression assays in CELCs. Two of the three subfragments, PB and NB, increased the expression of luciferase activity (3- to 4.6-fold) by heme or CoPP treatment (Fig. 5), indicating that the predominant element mediating up-regulation of the reporter is located ⫺4.1 to ⫺3.6 kb upstream of the transcription starting point. AP-1 Elements Play Little Role in the Induction of HO-1 Gene Expression by Heme and CoPP Other recent studies from our laboratory (2, 7, and Bonkovsky et al., unpublished) have established that consensus AP-1 binding elements play a key role in induction of the chick HO-1 gene produced by sodium arsenite, a prototypical “stress response.” Therefore, we investigated whether such AP-1 binding elements are also important for induction of HO-1 gene expres-
HO-1 INDUCTION BY METALLOPORPHYRINS
223
FIG. 2. Dose and time dependence of induction of pcHO5.6-Luc and pcHO7.1-Luc by heme and CoPP in CELC or LMH cells. (A, B) CELCs were transfected with pcHO5.6-Luc plasmid DNA using the calcium phosphate method as described under Materials and Methods. Following transfection, cells were treated with Me 2SO only (white bars), or with 0, 5, 10, or 30 M heme (dark gray bars), or CoPP (light gray bars) for 15 h (A) or 1 L/mL Me 2SO (E), 10 M heme (Œ), or 10 M CoPP (■), for increasing lengths of time (B). Cells were harvested, and luciferase assays and -galactosidase assays were carried out as described under Materials and Methods. Luciferase activities were normalized to -galactosidase activities and protein content. (C, D) LMH cells that had been stably transfected with pcHO7.1-Luc were grown as monolayers. Cells were treated with 0, 1, 2, or 3 L/mL Me 2SO (white bars), 0, 5, 10, or 30 M heme (dark gray bars), or CoPP (light gray bars) for 15 h (C) or with 1 L/mL Me 2SO (E), 10 M heme (Œ), or 10 M CoPP (■) for increasing lengths of time (D). Cells were harvested, and luciferase assays and -galactosidase assays were carried out as described under Materials and Methods. Luciferase activities were normalized to protein content. Data are presented as means ⫾ SE, n ⫽ 3. a indicates an increase over no treatment, P ⬍ 0.05; b indicates an increase over heme treatment, P ⬍ 0.05. Error bars are not shown when less than the size of symbol.
sion in response to heme and CoPP. For those studies, we chose the reporter construct pcHO5.6-Luc, which contains three consensus AP-1 binding elements located at ⫺1576, ⫺3647, and ⫺4578 base pairs upstream of the transcription starting point (2, 7, and Bonkovsky et al., unpublished). The role of these AP-1 sites in transcriptional activation was studied by making site-directed mutations in 2 out of 7 base pairs. As illustrated in Fig. 6A, five separate mutants with single, double, or triple AP-1 site mutants were made and tested for their ability to be induced by treatment with heme, CoPP, or arsenite (as positive control). In mutants 2–5, the inductions by arsenite were significantly reduced, whereas the inductions by heme and CoPP were not significantly affected (except for mutant 2⬘s response to CoPP [Fig. 6B]). Consistent with these results, in EMSA analyses, we found that heme or CoPP pretreatment produced little, if any, increase in the binding of nuclear proteins to an oligonucleotide containing the AP-1 consensus binding element, whereas pretreatment with sodium arsenite (the positive control) produced a sizable increase in binding (Fig. 7). Specificity of these results was estab-
lished by showing that a 50-fold excess of unlabeled nucleotide successfully completed for binding to these proteins (compare lanes 6 and 3 of Fig. 7), whereas a 50-fold excess of unlabeled nucleotide with a mutated AP-1 element did not (lane 7, Fig. 7). We therefore conclude that these AP-1 elements play little, if any, role in the heme- or CoPP-dependent induction of the chick HO-1 gene. DISCUSSION
Our major findings may be summarized as follows: (i) some metalloporphyrins, especially heme (FePP) and CoPP, induce expression of HO-1 promoter-Luc reporter genes in both CELC and LMH cells (Figs. 1 and 2); (ii) low concentrations of ZnMP, a potent, virtually irreversible inhibitor of HO activity, synergizes the heme-, but not the CoPP-dependent induction of the reporter gene (Fig. 3); (iii) unlike its effects on sodium arsenite-dependent induction, NAC has little, or no, inhibitory effect on the metalloporphyrin-dependent induction of HO-1 promoter-Luc reporter activity (Fig. 4); (iv) the key elements required for the metal-
224
SHAN ET AL.
FIG. 3. Effects of ZnMP on the induction of reporter gene activities by heme and CoPP. (A, B) CELCs were transfected with pcHO5.6-Luc plasmid DNA using the calcium phosphate method as described under Materials and Methods. Following transfection, cells were treated with 1 L/mL Me 2SO (E), 5 M heme (‚), or 10 M heme (Œ) and with increasing concentrations of ZnMP for 15 h (A), or 1 L/mL Me 2SO (E), 5 M CoPP (䊐), or 10 M CoPP (■) with increasing concentrations of ZnMP for 15 h (B). Cells were harvested and lysed, and assays were performed on 15 L of lysate. Luciferase activities were normalized to -galactosidase activities and protein content. (C, D) LMH cells that had been stably transfected with pcHO7.1-Luc were grown as monolayers. After 48 h, cells were treated with 1 L/mL Me 2SO (E), 5 M heme (‚), or 10 M heme (Œ) and with increasing concentrations of ZnMP for 15 h (C), or 1 L/mL Me 2SO (E), 5 M CoPP (䊐), or 10 M CoPP (■) with increasing concentrations of ZnMP for 15 h (D). Cells were harvested and lysed. Assays were performed on 15 L of lysate. Luciferase activities were normalized to protein content. Data are presented as means ⫾ SE, n ⫽ 3. a indicates an increase over no ZnMP treatment, P ⬍ 0.05. Error bars are not shown when less than the size of symbol.
loporphyrin-dependent induction of the reporter gene are located between ⫺3.6 and ⫺5.6 kb of the 5⬘-UTR of the HO-1 gene (Figs. 5 and 6); and (v) data from both EMSA and site-directed mutagenesis experiments indicate that the consensus AP-1 binding elements at ⫺1576, ⫺3647, and ⫺4578, alone or in concert, play little, if any, role in the inductions produced by heme or CoPP (Figs. 6 and 7). The parallelism between the inductions of the luciferase reporter in the present experiments (Fig. 1A) and the inductions of the endogenous HO-1 gene previously reported by our laboratory (20) supports the relevance of our current constructs and work to the general question, “what controls expression of HO-1?” The high degree of inducibility of HO-1, and the myriad chemical and physical factors (most of which produce oxidative or other “stress”) capable of producing such induction, emphasize the importance of HO-1 as part of the cellular system of defense against such stress. Thus, it is not surprising that the 5⬘-UTR of the HO-1 gene contains numerous consensus AP-1 binding sites, some of which have already been shown to be essential for the responses observed after exposure of cells to such in-
FIG. 4. Lack of effect of N-acetylcysteine (NAC) on induction of reporter gene activities by heme and CoPP. Transient transfections of CELCs were carried out by the calcium phosphate method as described under Materials and Methods. Following transfections, some cells were pretreated with 0, 5, 10, or 20 mM NAC for 1 h and then treated with 1 L/mL Me 2SO (E), 15 M sodium arsenite (F), 10 M heme (Œ), or 10 M CoPP (■) for 15 h in serum-free medium. Cells were harvested and lysed. Assays were performed on 15 L of lysate. Luciferase activities were normalized to -galactosidase activities and protein content. Data are presented as means ⫾ SE, n ⫽ 3. a indicates an increase over Ars treatment without NAC treatment, P ⬍ 0.001. NAC had no significant effect on heme or CoPP induction of luciferase activity.
HO-1 INDUCTION BY METALLOPORPHYRINS
225
FIG. 5. Deletion analysis of the upstream region of the chick HO-1 gene in CELC and LMH cells after heme or CoPP treatment. A partial restriction endonuclease map of the chick HO-1 gene is presented. Recognition sites for restriction endonucleases, MluI, PmlI (P), NheI (N), and BglII (B) are shown. The portion of the 5⬘-flanking region cloned in to the luciferase expression vector and the fragments (pcHO4.6 –3.6, 4.6 – 4.1, 4.1–3.6) are appropriately positioned below the restriction endonuclease map. Transient or stable transfections were carried out as described under Materials and Methods, after which some cells were treated with 10 M heme or 10 M CoPP for 15 h. Cells were harvested and lysed, and assays were performed on 15 L of lysate. Luciferase activities were normalized to -galactosidase activities and protein content (CELCs), or normalized to protein content (LMH cells). Data are presented as means ⫾ SE, n ⫽ 3. a indicates an increase over Me 2SO, P ⬍ 0.05.
ducers as sodium arsenite, H 2O 2, or UV light (2, 7, 21). A protective effect of HO-1 induction may be mediated by its ability to decrease intracellular levels of heme and/or hemoproteins (i.e., to diminish potential toxicity of heme per se) or by its ability to increase the products of the HO-1 reaction, especially biliverdin and bilirubin, which are potent physiological anti-oxidants (22) and CO, which is a physiologic vasodilator and neurotransmitter (1, 23–25). Indeed, CO has many properties and functions closely akin to NO (nitric oxide), a compound widely recognized for its myriad signaling functions in neuromuscular physiology (25, 26). Clearly, high concentrations of heme in cells may increase oxidative stress. However, it is far from clear whether heme at relatively low concentrations (1–10 M) produces very much oxidative stress in hepatocytes, which require substantial levels of heme synthesis and hemoproteins for normal function and which have high levels of GSH and other antioxidants. Indeed, these low concentrations of heme have been reported to induce HO-1 to high levels in normal hepatocytes without producing detectable increases in oxidative stress (10, 27). These results led us to suggest that heme and other metalloporphyrins are capable of inducing HO-1 gene expression by nonstress pathway(s). This conclusion is supported by the effects of CoPP, which is not a substrate for HO-1 and which is not able to catalyze the formation of active oxygen species within cells, but is the most efficacious metal-
loporphyrin inducer of HO-1 identified thus far (Figs. 1 and 2; 20). This induction seems clearly to be due to the macrocycle and not to free cobalt (Fig. 1B). In order for heme to reach the same degree of efficacy of induction of HO-1 as CoPP, it is necessary to inhibit endogenous HO-1 activity (e.g., by addition of ZnMP, see Fig. 3), so that higher intracellular levels of heme will persist long enough for maximal HO-1 induction to occur (20, 27, and Fig. 3). A mechanism of HO-1 induction by metalloporphyrins, that is independent of MAP kinase-stress pathways, is further supported by the observation that NAC is unable to abrogate the inductive effects of CoPP (Fig. 4). The partial abrogation of heme-dependent induction by NAC that we observed in some experiments suggests that heme may function by both MAP kinase-stress and by nonstress pathways. Recent work in murine hepatoma cells posited a key role for complexes of CoPP and rabbit hemopexin and hemopexin receptors in induction of HO-1 and other effects of CoPP on liver cells (28). However, studies in several other cell types, including CELCs, and in intact rodents have failed to support a key role for hemopexin or hemopexin receptors in mediating effects of metalloporphyrins on hepatocytes or other cells (29 –34). When hemopexin has been carefully prepared and purified, to avoid desialylation, it has consistently been found to retard and limit the uptake of heme into cells, including CELC (29, 30,
226
SHAN ET AL.
33). Furthermore, CoPP alone, without hemopexin, has been shown to be taken up by CELC and normal rat liver, to be incorporated into hemoproteins (35, 36), and to induce HO-1 (Figs. 1 and 2). Although hemopexin does bind heme with high avidity, its concentration is low and it is rapidly depleted whenever appreciable levels of heme (or other metalloporphyrins) occur in plasma [e.g., in intravascular hemolysis or in heme therapy of the acute porphyrias, tyrosinemia, or lead poisoning (37, 38)]. Then hemealbumin complexes form, and the heme of such complexes is taken up rapidly and selectively by hepatocytes where it is able to replete the regulatory heme pool, repress induction of 5-aminolevulinate synFIG. 7. Lack of effect of heme on nuclear protein binding to AP-1 consensus element. CELCs were treated with 15 M sodium arsenite (as positive control), 10 M heme or 10 M CoPP for 2 h, after which cells were harvested and nuclear proteins were isolated as described under Materials and Methods. A 6-g amount of the nuclear proteins was incubated with or without a 50-fold excess of unlabeled AP-1 or mutated AP-1 oligonucleotide at room temperature for 5 min, after which the ␥- 32P-labeled oligonucleotide containing the AP-1 consensus element (underlined) (5⬘-AGCATGGCTGAGTCAGGCTGGGAG3⬘) was added with 30 min. further incubation at 37°C. The oligonucleotide bound to the nuclear proteins was electrophoretically separated from the unbound form on a 4.5% polyacrylamide gel. The gel was dried and scanned by a PhosphorImager. The lane assignments are as follows: (1) no protein; (2) extracts from untreated cells; (3) extracts from cells treated with Ars; (4) extracts from cells treated with heme; (5) extracts from cells treated with CoPP; (6) extracts from cells treated with Ars plus a 50-fold excess of unlabeled oligonucleotide; (7) extracts from cells treated with Ars plus a 50-fold excess of unlabeled mutated oligonucleotide.
FIG. 6. Mutational analysis of the AP-1-binding elements of pcHO5.6-Luc in CELC: Effects of sodium arsenite, heme, and CoPP. (A) Site-directed mutants of putative AP-1 elements of plasmid pcHO5.6-Luc, corresponding to the diagrammed fragments. (B) Transient transfections were carried out by the calcium phosphate method as described under Materials and Methods. Following transfections, some cells were treated with 15 M sodium arsenite (Ars), 10 M heme, or 10 M CoPP for 15 h in serum-free medium. Cells were harvested and lysed, and assays were performed on 15 L of lysate. Luciferase activities were normalized to -galactosidase activities and protein content. Data are presented as means ⫾ SE, n ⫽ 3. a indicates an increase over wild type (wt), P ⬍ 0.05.
thase, increase levels and activities of holo-cytochromes P450 and tryptophan pyrrolase, and induce HO-1 (2, 39, 40). In summary, metalloporphyrins (especially heme and CoPP) induce HO-1 by a mechanism fundamentally different from that of sodium arsenite and other “stress” inducers. In the chick, the key element(s) required for the metalloporphyrin-mediated induction of HO-1 are located ⫺3.6 to ⫺5.6 kb upstream of the transcription starting point and are different from the AP-1 binding elements that play key roles in stressmediated inductions. Efforts to pinpoint the locations of the metalloporphyrin responsive elements and to identify their cognate binding proteins are ongoing in our laboratory. ACKNOWLEDGMENTS We thank D. L. Williams (SUNY, Stony Brook) for providing a starter supply of LMH cells and O. Gildemeister for helpful discussions. We thank J. Blackwood for help with typing the manuscript and P. LeClair for help with statistical analyses and preparation of final figures. Supported by a grant from NIH (DK 38825 to H.L.B.) and by the Research and Education Fund of the Division of Digestive Disease & Nutrition, UMass Memorial Health Care. The opinions
HO-1 INDUCTION BY METALLOPORPHYRINS expressed in this paper are those of the authors; they do not necessarily reflect the official views of NIH, the USPHS, the University of Massachusetts, or UMass Memorial Health Care.
REFERENCES 1. Maines, M. D. (1988) FASEB J. 2, 2557–2568. 2. Elbirt, K. K., and Bonkovsky, H. L. (1999) Proc. Assoc. Am. Phys. 111, 438 – 447. 3. Poss, K. D., and Tonegawa, S. (1997) Proc. Natl. Acad. Sci. USA 94, 10919 –10924. 4. Poss, K. D., and Tonegawa, S. (1997) Proc. Natl. Acad. Sci. USA 94, 10925–10930. 5. Burnett, A. L., Johns, D. G., Kriegsfeld, L. J., Klein, S. L., Calvin, D. C., Demas, G. E., Schramm, L. P., Nelson, R. J., Snyder, S. H., and Poss, K. D. (1998) Nat. Med. 4, 84 – 87. 6. Yachie, A., Niida, Y., Wada, T., Igarashi, N., Kaneda, H., Toma, T., Ohta, K., Kasahara, Y., and Koizumi, S. (1999) J. Clin. Invest. 103, 129 –135. 7. Elbirt, K. K., Whitmarsh, A. J., Davis, R. J., and Bonkovsky, H. L. (1998) J. Biol. Chem. 273, 8922– 8931. 8. Immenschuh, S., Kietzmann, T., Hinke, V., Wiederhold, M., Katz, N., and Muller-Eberhard, U. (1998) Mol. Pharmacol. 53, 483– 491. 9. Masuya, Y., Hioki, K., Tokunaga, R., and Taketani, S. (1998) J. Biochem. 124, 628 – 633. 10. Cable, E. E., Gildemeister, O. S., Pepe, J. A., Lambrecht, R. W., and Bonkovsky, H. L. (1997) Mol. Cell. Biochem. 169, 13–20. 11. Bonkovsky, H. L., Healey, J. F., Sinclair, P. R., and Sinclair, J. F. (1985) Biochem. J. 227, 893–901. 12. Healey, J. F., Bonkowsky, H. L., Sinclair, P. R., and Sinclair, J. F. (1981) Biochem. J. 198, 595– 604. 13. Gabis, K. K., Gildemeister, O. S., Pepe, J. A., Lambrecht, R. W., and Bonkovsky, H. L. (1996) Biochim. Biophys. Acta, Gen. Subj. 1290, 113–120. 14. Kolluri, S., Elbirt, K. K., and Bonkovsky, H. L. (1999) Biochim. Biophys. Acta 1472, 658 – 667. 15. Lincoln, B. C., Aw, T. Y., and Bonkovsky, H. L. (1989) Biochim. Biophys. Acta 992, 49 –58. 16. Lu, T. H., Pepe, J. A., Gildemeister, O. S., Tyrrell, R. M., and Bonkovsky, H. L. (1997) Biochim. Biophys. Acta, Gene Struct. Expression 1352, 293–302. 17. Lu, T. H., Lambrecht, R. W., Pepe, J., Shan, Y., Kim, T., and Bonkovsky, H. L. (1998) Gene 207, 177–186. 18. Sambrook, J., Fritsch, E. F., and Maniatis, T. (1989). Molecular Cloning: A Laboratory Manual Cold Spring Harbor Laboratory Press, Plainview, NY. 19. Attardi, L. D., and Tjian, R. (1993) Genes Dev. 7, 1341–1353.
227
20. Cable, E. E., Pepe, J. A., Karamitsios, N. C., Lambrecht, R. W., and Bonkovsky, H. L. (1994) J. Clin. Invest. 94, 649 – 654. 21. Tyrrell, R. M., and Basu-Modak, S. (1994) Methods Enzymol. 234, 224 –235. 22. Stocker, R., Yamamoto, Y., McDonagh, A. F., Glazer, A. N., and Ames, B. N. (1999) Science 235, 1043–1046. 23. Marks, G. S. (1994) Cell. Mol. Biol. 40, 863– 870. 24. Maines, M. D. (1996) Methods Enzymol. 268, 473– 488. 25. Maines, M. D. (1997) Annu. Rev. Pharmacol. Toxicol. 37, 517– 554. 26. Snyder, S. H., Jaffrey, S. R., and Zakhary, R. (1998) Brain Res. Brain Res. Rev. 26, 167–175. 27. Russo, S. M., Pepe, J. A., Cable, E. E., Lambrecht, R. W., and Bonkovsky, H. L. (1994) Eur. J. Clin. Invest. 24, 406 – 415. 28. Eskew, J. D., Vanacore, R. M., Sung, L., Morales, P. J., and Smith, A. (1999) J. Biol. Chem. 274, 638 – 648. 29. Sinclair, P. R., Bement, W. J., Gorman, N., Liem, H. H., Wolkoff, A. W., and Muller-Eberhard, U. (1988) Biochem. J. 256, 159 – 165. 30. Sinclair, P. R., Bement, W. J., Healey, J. F., Gorman, N., Sinclair, J. F., Bonkovsky, H. L., Liem, H. H., and Muller-Eberhard, U. (1994) Hepatology 20, 741–746. 31. Potter, D., Chroneos, Z. C., Baynes, J. W., Sinclair, P. R., Gorman, N., Liem, H. H., Muller-Eberhard, U., and Thorpe, S. R. (1993) Arch. Biochem. Biophys. 300, 98 –104. 32. Taketani, S., Immenschuh, S., Go, S., Sinclair, P. R., Stockert, R. J., Liem, H. H., and Muller-Eberhard, U. (1998) Hepatology 27, 808 – 814. 33. Bonkovsky, H. L., Gildemeister, O. S., Pepe, J. A., Lambrecht, R. W., and Hrkal, Z. (1996) Eur. J. Clin. Chem. Clin. Biochem. 34, 931–932. 34. Noyer, C. M., Immenschuh, S., Liem, H. H., Muller-Eberhard, U., and Wolkoff, A. W. (1998) Hepatology 28, 150 –155. 35. Sinclair, J. F., Sinclair, P. R., Healey, J. F., Smith, E. L., and Bonkowsky, H. L. (1982) Biochem. J. 204, 103–109. 36. Cable, E. E., Cable, J. W., and Bonkovsky, H. L. (1993) Hepatology 18, 119 –127. 37. Muller-Eberhard, U., and Fraig, M. (1993) Am. J. Hematol. 42, 59 – 62. 38. Muller-Eberhard, U., and Liem, H. H. (1974) in Structure and function of plasma proteins (Allison, A. C., Ed.), pp. 35–53, Plenum Press, London. 39. Bloomer, J. R., and Bonkovsky, H. L. (1989) DM, Dis.-Mon. 35, 1–54. 40. Bonkovsky, H. L. (1990) in “Hepatology: A Textbook of Liver Disease” (Zakim, D., and Boyer, T. D., Eds.), pp. 378 – 424, Saunders, Philadelphia.
Induction of the Heme Oxygenase-1 Gene by Metalloporphyrins Ying Shan, Joyce Pepe, Tze Hong Lu, Kimberly K. Elbirt, Richard W. Lambrecht, and Herbert L. Bonkovsky 1 Department of Medicine and Department of Biochemistry and Molecular Biology, University of Massachusetts Medical School and The Center for Study of Disorders of Iron & Porphyrin Metabolism, UMass Memorial Health Care, Worcester, Massachusetts 01655
Received December 30, 1999, and in revised form May 5, 2000
Induction of expression of heme oxygenase-1 (HO-1) has been studied in primary cultures of chick embryo liver cells and in the LMH line of avian hepatoma cells. Cells were transiently transfected with selected constructs containing portions of the 5ⴕ-untranslated (promoter) region of the HO-1 gene linked to luciferase as reporter gene. LMH cells that had been stably transfected with selected wild type or mutant constructs were also studied. Metalloporphyrins, especially Fe protoporphyrin (heme) and Co protoporphyrin strongly induced luciferase expression in both types of transfected cells. Low concentrations of Zn mesoporphyrin, an inhibitor of HO activity, exerted a synergistic effect on heme-, but not Co protoporphyrin-dependent induction. The antioxidant and OSH donor N-acetyl cysteine had little effect on the metalloporphyrin-dependent inductions of HO-1, in contrast to its marked inhibitory effect on the sodium arsenite-dependent induction of the HO-1 gene. Deletional analysis showed that the key element(s) required for the metalloporphyrin-dependent induction of HO-1 is located between ⴚ3.6 and ⴚ5.6 kb upstream of the transcription starting point. Data from electrophoretic mobility shift and site-directed mutagenesis experiments excluded a role for consensus AP-1 binding elements at ⴚ1576, ⴚ3647, or ⴚ4578 in the inductions produced by heme or Co protoporphyrin. © 2000 Academic Press
Key Words: cobalt protoporphyrin; enhancer; gene expression; heme; heme oxygenase-1; oxidative stress; promoter.
1 To whom correspondence should be addressed at University of Massachusetts Memorial Health Center, Department of Medicine, Room S6-737, 55 Lake Avenue, North, Worcester, MA 01655. Fax: (508) 856-3981. E-mail: [email protected].
0003-9861/00 $35.00 Copyright © 2000 by Academic Press All rights of reproduction in any form reserved.
Heme oxygenases (HO) (EC 1.14.99.3) are ubiquitous enzymes that catalyze the breakdown of heme (iron-protoporphyrin IX) by the specific oxidative cleavage of the ␣-methene bridge. The products of the reaction are biliverdin, iron, and carbon monoxide (CO); the carbon atom of the CO formed is derived exclusively from the ␣-methene carbon of the heme macrocycle. Two major forms of HO, called HO-1 and HO-2, have been described. HO-1 is detectable in many tissues and can be induced to high levels, mainly by increases in gene transcription rates. Heme and other metalloporphyrins, especially Coand Mn-protoporphyrin (MnPP), are inducers of HO-1, as are several transition metals (e.g., Cd, Co, Fe), arsenicals (e.g., sodium, arsenite, phenylarsine oxide), agents that deplete intracellular levels of GSH and other thiols, and physical forces that increase oxidative or other stress on cells (e.g., UV light, heat shock). Indeed another name for HO-1 is heat shock protein (hsp) 32, and HO-1 is now used widely as a harbinger of cellular stress responses (for reviews, see Refs. 1 and 2). HO-2 is an homologous protein of ⬃36 kDa molecular mass, which is also widely distributed in many tissues in the body, is present in highest concentration in selected areas of the brain and testes of mammals, and is essentially uninducible. HO-3 has a lesser degree of homology to the other forms and much lower specific activity; its principal role may actually be to bind or transport heme within cells. The importance of both HO-1 and HO-2 in heme and iron metabolism has been established by recent descriptions of abnormal phenotypes in mice with targeted gene disruptions of one or the other form of HO (3–5). In addition, a Japanese child with developmental delay, seizures, hemolytic anemia, and renal dysfunction was recently reported to have markedly reduced levels of HO-1, 219
220
SHAN ET AL.
due to compound heterozygous defects in the HO-1 genes inherited from his parents (6). The mechanisms whereby chemical and physical perturbations increase expression of the HO-1 gene are starting to be elucidated. For example, recent work in our laboratory showed that increases in the major mitogen-activated protein (MAP) kinases, c-Jun N-terminal kinase (Jnk), extracellular regulated kinase (Erk), and p38, underlie the induction produced by sodium arsenite, cadmium, or heat shock. An important role for consensus AP-1 binding elements, several of which are found in the 5⬘-untranslated region (UTR) of the HO-1 gene, was established by site-directed mutagenesis and electrophoretic mobility shift assays (EMSA) (7). In these studies, heme did not produce any increase in MAP kinase activities, although it too induced both the endogenous HO-1 gene and the HO-1 5⬘-UTR-Luc reporter gene constructs studied. Others have provided evidence that other kinases (e.g., protein kinases A and C) may also play a role in HO-1 induction (8, 9). The functions and mechanisms of action of inducers of HO-1 are dependent upon many factors, including the species of organism and the organs and cell-types studied. For example, although high concentrations (50 –100 M) of heme may increase endothelial cell injury in some rat or human organs, previous work from our laboratory in normal avian hepatocytes in primary culture (10) or the LMH line of avian hepatocellular carcinoma cells (7) indicated that heme and other metalloporphyrins induce HO-1 by nontoxic, nonstress dependent pathways. Chick embryo liver cells (CELC) provide a simple, cheap, robust, and relevant system for studying heme and porphyrin metabolism, including the regulation of HO-1 expression. CELCs maintain normal levels and inducibility of both 5-aminolevulinate synthase, the rate-controlling enzyme of heme synthesis, and of HO, the rate-controlling enzyme of heme breakdown. In addition, the kinetics of the heme synthetic pathway in CELC are closely similar to that of normal human liver, a property not shared by rodent liver models (11, 12). We have also found that LMH cells provide a continuous cell line that retains the desirable properties of normal levels and regulation of the rate-controlling steps of heme metabolism (13, 14). Among the potential additional advantages of LMH cells are that they are a single, homogeneous cell type, that unlimited numbers can be obtained, and that stably-transfected lines can be established. In this paper, we report studies that characterize the metalloporphyrin-dependent induction of HO-1 in both CELC and LMH cells, and that begin to delineate portions of the 5⬘-UTR of the HO-1 gene that are central to the inductive response.
MATERIALS AND METHODS Materials. All tissue culture dishes were from Corning Glass Inc. (Corning, NY). Culture flasks were from Falcon, VWR Scientific (Bridgeport, NJ). Ferric (Fe 3⫹)-protoporphyrin IX 䡠 Cl (heme), chromium (Cr 3⫹)-mesoporphyrin IX 䡠 Cl, manganese (Mn 3⫹)-protoporphyrin IX 䡠 Cl, cobalt (Co 3⫹)-protoporphyrin IX 䡠 Cl, tin (Sn ⫹4)-mesoporphyrin IX 䡠 2Cl, and zinc (Zn 2⫹)-mesoporphyrin IX were from Porphyrin Products (Logan, UT). Chloroform, isopropyl alcohol, acrylamide, and bis-acrylamide were from Fisher (Pittsburgh, PA). Waymouth’s MB 752/1 media, Williams’ E media, geneticin, gelatin, fetal bovine serum, and calf serum were from Gibco BRL, (Grand Island, NY). -Mercaptoethanol, bovine serum albumin (BSA), EDTA, EGTA, formaldehyde (37% v/v), formamide, glycylglycine, penicillin/streptomycin, phenyl methyl sulfonyl fluoride, piperacillin, sodium arsenite, sodium dodecyl sulfate, trypsin, dithiothreitol (DTT), 3,3⬘,5-triodo-L-thyronine (T-3), spermine, and o-nitrophenyl-D-galactopyranoside were from Sigma (St. Louis, MO). Dexamethasone was from Gensia Pharmaceuticals (Irvine, CA). DNA Maxiprep kits and gel extraction kits were from Qiagen (Santa Clarita, CA). All 32P-radionucleotides were from NEN Life Science Products (Boston, MA). LMH cells were a generous gift from D. L. Williams, Department of Pharmacological Sciences, SUNY (Stony Brook, NY). The pPGK- gal plasmid was a gift from P. Dobner (Department of Molecular Genetics and Microbiology, University of Massachusetts Medical School, Worcester, MA). The primers GLprimer2 and RVprimer3, Wizard plasmid DNA preparation kits, pGL3 Promoter plasmid, and Luciferase Assay Reagent were purchased from Promega (Madison, WI). DNA sequencing was performed by the sequencing facility of the Dana Farber Cancer Institute (Boston, MA) or by the Nucleic Acids Facility, University of Massachusetts Medical School. All chemicals were of the highest purity available. Cell cultures and treatments. Primary chick liver cell cultures were prepared with minor modifications of a previously described method (15). In brief, 17-day-old Barred Rock chick embryos were killed by decapitation, and their livers were immediately removed and placed in Hanks’ balanced salt solution. Livers were chopped under sterile conditions and then transferred to a 125-mL sterile flask and incubated for a total of 20 min in a 37°C with trypsin and DNAse. The cell suspension was then washed with in a buffer containing ammonium chloride to lyse the erythrocytes and resuspended repeatedly by shaking with Williams’ E medium until the total desired volume of medium (10 mL/liver) had washed over the cells. The cells (3 mL/dish) were then plated on 6-well culture dishes. LMH cells were maintained in Waymouth’s MB 752/1 medium, supplemented with 100 U/mL penicillin, 100 g/mL streptomycin, and 10% (v/v) fetal bovine serum. The cells were routinely passaged twice a week (7). For some experiments, 1 g/mL piperacillin was added to the culture medium. All metalloporphyrins were dissolved in Me 2SO; sodium arsenite was prepared as described (13, 16) and stored at ⫺20°C until used. Plasmid construction. The chick HO-1 gene expression plasmids used to transfect CELC and LMH cells were constructed as follows. The construction of plasmids pcHO7.1-Luc, pcHO5.6-Luc, pcHO4.6Luc, pcHO3.6-Luc, pcHO2.5-Luc, pcHO1.6-Luc, pcHO0.6-Luc, and pcHOTATA-Luc has been described previously (7). The pcHOPB-Luc reporter plasmid was constructed by cloning 913 base pairs (⫺4.6 to ⫺3.6 kbp) of the chick HO-1 proximal promoter into the pGL3 with promoter vector. Plasmid pcHO7.1-Luc was partially digested with PmlI, the ends were filled in using the Klenow fragment, redigested with BglII, and then ligated to SmaI/BglII sites upstream of the luciferase reporter gene. The plasmids pcHOPN-Luc and pcHONBLuc were constructed by cloning the PmlI/Nhel fragment (⫺4.6 to ⫺4.1 kbp) or the NheI/BglII fragment (⫺4.1 to ⫺3.6 kbp) of pcHO4.6Luc, respectively, into the pGL3 promoter vector. The orientation of the inserts was verified by sequencing using the commercial primers GLprimer2 and RVprimer3 from Promega.
HO-1 INDUCTION BY METALLOPORPHYRINS Transfections. Transient and stable transfections were carried out by the calcium phosphate precipitation technique (16, 17). For transient transfections, about 40 h after initial plating of the cell cultures, the medium was carefully removed and 2 mL of fresh Williams’ E medium [containing dexamethasone, T-3, glutamine, and 10% (v/v) calf serum] was added. Two hours later, transfections were performed for 4 –5 h by the calcium phosphate method. Transfections were ended by removing the transfection medium and gently washing the cells twice with a buffer containing 0.01 M sodium phosphate (pH 7.4) and 0.15 M NaCl (PBS) to remove the extracellular DNA. Then fresh Williams’ E medium (with dexamethasone, T-3, and glutamine, but without serum) was added. Typically, some of the plates received treatments at this time. For stable transfections, LMH cells were plated into 6-cm tissue culture dishes (5 ⫻ 10 5/6-cm plate) the day before transfection. Cells were fed with 5 mL of Waymouth’s medium supplemented with 10% fetal bovine serum 2 h prior to transfections. LMH cells were exposed to the DNA-CaPO 4 precipitate for 16 h. The DNA mixture was composed of 10 g of the chick HO-1 expression plasmid and 2 g of pSV2neo. The precipitate was removed, and the cells were cultured in Waymouth’s complete medium for 24 h the transfected cells were then split (1:15) and plated on 10-cm dishes in Waymouth’s complete medium. Geneticin (G418 sulfate) was added 24 h later to a final concentration of 100 g/mL, and resistant colonies were selected over a 2–3 week period. Resistant colonies were transferred to 48well dishes and propagated in the presence of G418. Colonies containing luciferase activity were seeded in 6-well dishes at a density of 6 ⫻ 10 5 cells/well. Forty hours after seeding, cells were incubated at 37°C in 2 mL of serum-free medium. After 2 h, some of the plates received treatment. Assessment of reporter gene activity. Reporter gene expression and activation was assessed by quantitation of luciferase activity, normalized to -galactosidase activity, and protein content. For luciferase activities, transfected cells were washed twice with PBS, and harvested by scraping in 250 L of glycylglycine harvest buffer (25 mM glycylglycine, pH 7.8, 15 mM magnesium sulfate, 4 mM EGTA, 1 mM DTT). Cells were lysed by three freeze–thaw cycles (3 min in liquid nitrogen followed by 3 min at 37°C), followed by a 10-min centrifugation at 14,000g at 4°C. The supernatant was retained, and 15-L aliquots of cell lysate were used for each assay. Luciferase activity measurements were carried out using a Monolight 2010 luminometer (Analytical Luminescence Laboratories, Ann Arbor, MI). Relative light units were recorded. For -galactosidase activities, 15 L of cell lysate was added to 200 L of Z Buffer (60 mM Na 2HPO 4, 40 mM NaH 2PO 4, 10 mM potassium chloride, 1 mM magnesium sulfate, 50 mM -mercaptoethanol, adjusted to pH 7.0) and 100 L of 5 mg/mL o-nitrophenyl-D-galactoside (ONPG) dissolved in 0.1 M potassium phosphate, pH 7.0. The samples were mixed and incubated at 37°C for 1.5–2 h. The absorbance at 420 nm was measured (18). Protein concentrations were determined from absorbance at 562 nm measured by the bicinchoninic acid method on a Spectronic GENESYS 2 spectrophotometer, using BSA as standard. Site-directed mutagenesis of the chick HO-1 reporter gene construct. Site-directed mutagenesis of pcHO5.6-Luc was carried out using Quick Change mutagenesis kits from Stratagene (La Jolla, CA) following the manufacturer’s instructions. pcHO5.6-Luc was used as template, and the AP-1 elements located ⫺1576, ⫺3647, and ⫺4578 bp from the transcription start site in pcHO5.6-Luc were mutated. The oligonucleotides 5⬘GCAGA GCAAG ACAGGAAAAG CATGGCTTCG TCAGGCTGGG AGCGCTGAG3⬘ and 5⬘CTCAGCGCTC CCAGCCTGAC GAAGCCATGC TTTTCCTGTC TTGCTCTGC3⬘ were used as primers for the mutagenesis reaction. After the mutagenesis reactions were completed, the reaction products were introduced into XL1-Blue competent cells (Stratagene). The mutants were confirmed by DNA sequencing (UMass Nucleic Acid Facility).
221
DNA/protein interaction analysis. Extracts of nuclear proteins from cultured cells were prepared as described previously (17, 19). In a typical electrophoretic mobility shift assay (EMSA), 6 g of the nuclear protein were incubated at 37°C for 30 min with 32P-labeled 24-mer oligonucleotide probe (20,000 cpm), containing the AP-1 consensus sequence (5⬘-AGCATGGCTGAGTCAGGCTGGGAG-3⬘) [the consensus sequence is underlined]. The DNA-binding reaction was carried out in a reaction buffer containing 10 mM Tris–HCl pH 7.5, 1 mM MgCl 2, 1 mM EDTA, 1 mM DTT, 50 mM NaCl, 1.5 g dI– dC, 3 g BSA, and 12% (v/v) glycerol, either with or without competitors. The competitors used were unlabeled AP-1 consensus oligonucleotides or mutant AP-1 oligonucleotides. Proteins in the reaction mixture were then separated by polyacrylamide gel electrophoresis (PAGE) (4.5% polyacrylamide gel run at 100 V for 1 h). The gel was dried and scanned by PhosphorImager (17). Statistical analysis of data. Experiments were repeated two to four times; except for gel shift assays, every experiment included at least triplicate samples for each treatment group. Representative results from single experiments are presented. Statistical analyses were performed with JMP 3.0.2 software (SAS Institute, Cary, NC). For normally distributed data, differences in mean values were assessed by analysis-of-variance techniques (ANOVA), with the Tukey–Kramer correction for multiple pair-wise comparisons, or Dunnett’s test versus a control. For experiments with non-normally distributed data, the Wilcoxon/Kruskal–Wallis (rank-sum) test was used. P values ⬍0.05 were considered significant.
RESULTS
Induction of Transfected Heme Oxygenase-1 Promoter Luciferase Reporter Constructs by Metalloporphyrins Some metalloporphyrins, especially heme (FePP) and CoPP, gave significant induction of HO-1 promoter/reporter constructs in both CELCs and LMH cells (Figs. 1A and B). CoPP had the greatest effect on reporter gene expression with a 5-fold or 42-fold increase in normalized luciferase activity in CELCs or LMH cells, respectively. Lesser degrees of induction were observed with this construct after exposure to heme (4.2-fold or 16-fold in CELC or LMH cells). CrMP and MnPP produced slight increases (1.8-fold in CELC but not in LMH cells); however, no induction was observed for SnMP or ZnMP in either CELCs or LMH cells. Optimization of Heme- or CoPP-Mediated Induction of pcHO-Luc Further analysis was performed on the heme- and CoPP-mediated increases in reporter gene expression in cells transfected with either pcHO5.6-Luc or pcHO7.1-Luc. Dose–response and time course experiments established that the most effective dose for heme and CoPP were 10 M in transiently transfected CELCs or in stably transfected LMH cells. Peak levels of induction occurred at 10 –15 h for heme, after which the luciferase activities fell. For CoPP, peak induction occurred at 15 h (Fig. 2). Therefore, these conditions were used to further investigate the mechanism of heme and CoPP induction of HO-1 gene transcription.
222
SHAN ET AL.
an inhibitor of HO (20). We therefore added low concentrations (10 –200 nM) of ZnMP alone or with CoPP or heme (5 or 10 M) to transfected CELC or LMH cells. As shown in Fig. 3, ZnMP alone had no effect on luciferase activities nor did it affect induction produced by CoPP. In contrast, ZnMP produced the expected synergism of induction by heme, to levels similar to those produced by CoPP. Effect of N-Acetylcysteine (NAC) on the Metalloporphyrin-Dependent Induction of HO-1 Promoter-Luc Reporter Activity Effects of pretreatment with NAC on normalized luciferase activity from heme-, CoPP-, and arsenitetreated LMH cells are shown in Fig. 4. NAC did not significantly affect the induction by CoPP or heme, but markedly abrogated arsenite-mediated induction of the reporter gene activity. Deletion Analysis of the Chick HO-1 Gene Promoter/ Enhancer
FIG. 1. Effects of metalloporphyrins on reporter gene activities. (A) Effects of metalloporphyrins on pcHO5.6-Luc activity in transiently transfected CELCs. Transient transfections were carried out by the calcium phosphate method. Following transfections, cells were washed twice with PBS and 2 mL of serum-free medium was added, after which some cells were treated with 1 L/mL Me 2SO or with 10 M cobalt protoporphyrin (CoPP), chromium mesoporphyrin (CrMP), heme, manganese protoporphyrin (MnPP), tin mesoporphyrin (SnMP), or zinc mesoporphyrin (ZnMP). After 15 h, cells were harvested and luciferase assays and -galactosidase assays were carried out, as described under Materials and Methods. Luciferase activities were normalized to -galactosidase activities and protein content. (B) Effects of metalloporphyrins on pcHO7.1-Luc activity in stably transfected LMH cells. Monolayers of LMH cells, stably transfected with plasmid pcHO7.1-Luc, were treated with 1 L/mL Me 2SO or with 10 M CoPP, CrMP, heme, MnPP, SnMP, or ZnMP in serum-free medium. After 15 h, cells were harvested and lysed, and assays were performed on 15 L of lysate. Luciferase activities were normalized to protein content. Data are presented as means ⫾ SE, n ⫽ 12 for Me 2SO, CoPP, and heme; n ⫽ 3 for the other treatments. a indicates an increase over no treatment, P ⬍ 0.05; b indicates an increase over heme treatment, P ⬍ 0.05.
A Low Concentration of ZnMP Synergizes the HemeDependent Induction of the Reporter Gene We postulated that the difference in efficacy of upregulation of the luciferase reporter gene activity by CoPP vs heme is because heme can be catabolized by HO, whereas CoPP cannot. Therefore, we reasoned that, if we added an inhibitor of HO, that itself does not up-regulate the HO-1 gene, with heme, we might attain activities of luciferase similar to those produced by CoPP alone. We previously showed that ZnMP is such
In an effort to identify the metalloporphyrin-responsive regulatory elements within the upstream 5⬘-flanking sequences of the chick HO-1 gene, we performed deletional analysis of the entire distal upstream 7.1-kb region. Increasing lengths of the 7.1 kb upstream region (Fig. 5), including the promoter and distal enhancers, were linked to a luciferase reporter gene, and transiently or stably transfected into CELC or LMH cells and assayed for luciferase activity in response to metalloporphyrins. We observed that the key elements required for the metalloporphyrin-dependent induction of the luciferase reporter are located between ⫺3.6 and ⫺5.6 kb of the 5⬘-UTR of the HO-1 gene (Fig. 5). To further localize the promoter/enhancer activity, three subfragments (PB, PN, and NB) were cloned into the vector pGL3-promoter, and initially analyzed by transient expression assays in CELCs. Two of the three subfragments, PB and NB, increased the expression of luciferase activity (3- to 4.6-fold) by heme or CoPP treatment (Fig. 5), indicating that the predominant element mediating up-regulation of the reporter is located ⫺4.1 to ⫺3.6 kb upstream of the transcription starting point. AP-1 Elements Play Little Role in the Induction of HO-1 Gene Expression by Heme and CoPP Other recent studies from our laboratory (2, 7, and Bonkovsky et al., unpublished) have established that consensus AP-1 binding elements play a key role in induction of the chick HO-1 gene produced by sodium arsenite, a prototypical “stress response.” Therefore, we investigated whether such AP-1 binding elements are also important for induction of HO-1 gene expres-
HO-1 INDUCTION BY METALLOPORPHYRINS
223
FIG. 2. Dose and time dependence of induction of pcHO5.6-Luc and pcHO7.1-Luc by heme and CoPP in CELC or LMH cells. (A, B) CELCs were transfected with pcHO5.6-Luc plasmid DNA using the calcium phosphate method as described under Materials and Methods. Following transfection, cells were treated with Me 2SO only (white bars), or with 0, 5, 10, or 30 M heme (dark gray bars), or CoPP (light gray bars) for 15 h (A) or 1 L/mL Me 2SO (E), 10 M heme (Œ), or 10 M CoPP (■), for increasing lengths of time (B). Cells were harvested, and luciferase assays and -galactosidase assays were carried out as described under Materials and Methods. Luciferase activities were normalized to -galactosidase activities and protein content. (C, D) LMH cells that had been stably transfected with pcHO7.1-Luc were grown as monolayers. Cells were treated with 0, 1, 2, or 3 L/mL Me 2SO (white bars), 0, 5, 10, or 30 M heme (dark gray bars), or CoPP (light gray bars) for 15 h (C) or with 1 L/mL Me 2SO (E), 10 M heme (Œ), or 10 M CoPP (■) for increasing lengths of time (D). Cells were harvested, and luciferase assays and -galactosidase assays were carried out as described under Materials and Methods. Luciferase activities were normalized to protein content. Data are presented as means ⫾ SE, n ⫽ 3. a indicates an increase over no treatment, P ⬍ 0.05; b indicates an increase over heme treatment, P ⬍ 0.05. Error bars are not shown when less than the size of symbol.
sion in response to heme and CoPP. For those studies, we chose the reporter construct pcHO5.6-Luc, which contains three consensus AP-1 binding elements located at ⫺1576, ⫺3647, and ⫺4578 base pairs upstream of the transcription starting point (2, 7, and Bonkovsky et al., unpublished). The role of these AP-1 sites in transcriptional activation was studied by making site-directed mutations in 2 out of 7 base pairs. As illustrated in Fig. 6A, five separate mutants with single, double, or triple AP-1 site mutants were made and tested for their ability to be induced by treatment with heme, CoPP, or arsenite (as positive control). In mutants 2–5, the inductions by arsenite were significantly reduced, whereas the inductions by heme and CoPP were not significantly affected (except for mutant 2⬘s response to CoPP [Fig. 6B]). Consistent with these results, in EMSA analyses, we found that heme or CoPP pretreatment produced little, if any, increase in the binding of nuclear proteins to an oligonucleotide containing the AP-1 consensus binding element, whereas pretreatment with sodium arsenite (the positive control) produced a sizable increase in binding (Fig. 7). Specificity of these results was estab-
lished by showing that a 50-fold excess of unlabeled nucleotide successfully completed for binding to these proteins (compare lanes 6 and 3 of Fig. 7), whereas a 50-fold excess of unlabeled nucleotide with a mutated AP-1 element did not (lane 7, Fig. 7). We therefore conclude that these AP-1 elements play little, if any, role in the heme- or CoPP-dependent induction of the chick HO-1 gene. DISCUSSION
Our major findings may be summarized as follows: (i) some metalloporphyrins, especially heme (FePP) and CoPP, induce expression of HO-1 promoter-Luc reporter genes in both CELC and LMH cells (Figs. 1 and 2); (ii) low concentrations of ZnMP, a potent, virtually irreversible inhibitor of HO activity, synergizes the heme-, but not the CoPP-dependent induction of the reporter gene (Fig. 3); (iii) unlike its effects on sodium arsenite-dependent induction, NAC has little, or no, inhibitory effect on the metalloporphyrin-dependent induction of HO-1 promoter-Luc reporter activity (Fig. 4); (iv) the key elements required for the metal-
224
SHAN ET AL.
FIG. 3. Effects of ZnMP on the induction of reporter gene activities by heme and CoPP. (A, B) CELCs were transfected with pcHO5.6-Luc plasmid DNA using the calcium phosphate method as described under Materials and Methods. Following transfection, cells were treated with 1 L/mL Me 2SO (E), 5 M heme (‚), or 10 M heme (Œ) and with increasing concentrations of ZnMP for 15 h (A), or 1 L/mL Me 2SO (E), 5 M CoPP (䊐), or 10 M CoPP (■) with increasing concentrations of ZnMP for 15 h (B). Cells were harvested and lysed, and assays were performed on 15 L of lysate. Luciferase activities were normalized to -galactosidase activities and protein content. (C, D) LMH cells that had been stably transfected with pcHO7.1-Luc were grown as monolayers. After 48 h, cells were treated with 1 L/mL Me 2SO (E), 5 M heme (‚), or 10 M heme (Œ) and with increasing concentrations of ZnMP for 15 h (C), or 1 L/mL Me 2SO (E), 5 M CoPP (䊐), or 10 M CoPP (■) with increasing concentrations of ZnMP for 15 h (D). Cells were harvested and lysed. Assays were performed on 15 L of lysate. Luciferase activities were normalized to protein content. Data are presented as means ⫾ SE, n ⫽ 3. a indicates an increase over no ZnMP treatment, P ⬍ 0.05. Error bars are not shown when less than the size of symbol.
loporphyrin-dependent induction of the reporter gene are located between ⫺3.6 and ⫺5.6 kb of the 5⬘-UTR of the HO-1 gene (Figs. 5 and 6); and (v) data from both EMSA and site-directed mutagenesis experiments indicate that the consensus AP-1 binding elements at ⫺1576, ⫺3647, and ⫺4578, alone or in concert, play little, if any, role in the inductions produced by heme or CoPP (Figs. 6 and 7). The parallelism between the inductions of the luciferase reporter in the present experiments (Fig. 1A) and the inductions of the endogenous HO-1 gene previously reported by our laboratory (20) supports the relevance of our current constructs and work to the general question, “what controls expression of HO-1?” The high degree of inducibility of HO-1, and the myriad chemical and physical factors (most of which produce oxidative or other “stress”) capable of producing such induction, emphasize the importance of HO-1 as part of the cellular system of defense against such stress. Thus, it is not surprising that the 5⬘-UTR of the HO-1 gene contains numerous consensus AP-1 binding sites, some of which have already been shown to be essential for the responses observed after exposure of cells to such in-
FIG. 4. Lack of effect of N-acetylcysteine (NAC) on induction of reporter gene activities by heme and CoPP. Transient transfections of CELCs were carried out by the calcium phosphate method as described under Materials and Methods. Following transfections, some cells were pretreated with 0, 5, 10, or 20 mM NAC for 1 h and then treated with 1 L/mL Me 2SO (E), 15 M sodium arsenite (F), 10 M heme (Œ), or 10 M CoPP (■) for 15 h in serum-free medium. Cells were harvested and lysed. Assays were performed on 15 L of lysate. Luciferase activities were normalized to -galactosidase activities and protein content. Data are presented as means ⫾ SE, n ⫽ 3. a indicates an increase over Ars treatment without NAC treatment, P ⬍ 0.001. NAC had no significant effect on heme or CoPP induction of luciferase activity.
HO-1 INDUCTION BY METALLOPORPHYRINS
225
FIG. 5. Deletion analysis of the upstream region of the chick HO-1 gene in CELC and LMH cells after heme or CoPP treatment. A partial restriction endonuclease map of the chick HO-1 gene is presented. Recognition sites for restriction endonucleases, MluI, PmlI (P), NheI (N), and BglII (B) are shown. The portion of the 5⬘-flanking region cloned in to the luciferase expression vector and the fragments (pcHO4.6 –3.6, 4.6 – 4.1, 4.1–3.6) are appropriately positioned below the restriction endonuclease map. Transient or stable transfections were carried out as described under Materials and Methods, after which some cells were treated with 10 M heme or 10 M CoPP for 15 h. Cells were harvested and lysed, and assays were performed on 15 L of lysate. Luciferase activities were normalized to -galactosidase activities and protein content (CELCs), or normalized to protein content (LMH cells). Data are presented as means ⫾ SE, n ⫽ 3. a indicates an increase over Me 2SO, P ⬍ 0.05.
ducers as sodium arsenite, H 2O 2, or UV light (2, 7, 21). A protective effect of HO-1 induction may be mediated by its ability to decrease intracellular levels of heme and/or hemoproteins (i.e., to diminish potential toxicity of heme per se) or by its ability to increase the products of the HO-1 reaction, especially biliverdin and bilirubin, which are potent physiological anti-oxidants (22) and CO, which is a physiologic vasodilator and neurotransmitter (1, 23–25). Indeed, CO has many properties and functions closely akin to NO (nitric oxide), a compound widely recognized for its myriad signaling functions in neuromuscular physiology (25, 26). Clearly, high concentrations of heme in cells may increase oxidative stress. However, it is far from clear whether heme at relatively low concentrations (1–10 M) produces very much oxidative stress in hepatocytes, which require substantial levels of heme synthesis and hemoproteins for normal function and which have high levels of GSH and other antioxidants. Indeed, these low concentrations of heme have been reported to induce HO-1 to high levels in normal hepatocytes without producing detectable increases in oxidative stress (10, 27). These results led us to suggest that heme and other metalloporphyrins are capable of inducing HO-1 gene expression by nonstress pathway(s). This conclusion is supported by the effects of CoPP, which is not a substrate for HO-1 and which is not able to catalyze the formation of active oxygen species within cells, but is the most efficacious metal-
loporphyrin inducer of HO-1 identified thus far (Figs. 1 and 2; 20). This induction seems clearly to be due to the macrocycle and not to free cobalt (Fig. 1B). In order for heme to reach the same degree of efficacy of induction of HO-1 as CoPP, it is necessary to inhibit endogenous HO-1 activity (e.g., by addition of ZnMP, see Fig. 3), so that higher intracellular levels of heme will persist long enough for maximal HO-1 induction to occur (20, 27, and Fig. 3). A mechanism of HO-1 induction by metalloporphyrins, that is independent of MAP kinase-stress pathways, is further supported by the observation that NAC is unable to abrogate the inductive effects of CoPP (Fig. 4). The partial abrogation of heme-dependent induction by NAC that we observed in some experiments suggests that heme may function by both MAP kinase-stress and by nonstress pathways. Recent work in murine hepatoma cells posited a key role for complexes of CoPP and rabbit hemopexin and hemopexin receptors in induction of HO-1 and other effects of CoPP on liver cells (28). However, studies in several other cell types, including CELCs, and in intact rodents have failed to support a key role for hemopexin or hemopexin receptors in mediating effects of metalloporphyrins on hepatocytes or other cells (29 –34). When hemopexin has been carefully prepared and purified, to avoid desialylation, it has consistently been found to retard and limit the uptake of heme into cells, including CELC (29, 30,
226
SHAN ET AL.
33). Furthermore, CoPP alone, without hemopexin, has been shown to be taken up by CELC and normal rat liver, to be incorporated into hemoproteins (35, 36), and to induce HO-1 (Figs. 1 and 2). Although hemopexin does bind heme with high avidity, its concentration is low and it is rapidly depleted whenever appreciable levels of heme (or other metalloporphyrins) occur in plasma [e.g., in intravascular hemolysis or in heme therapy of the acute porphyrias, tyrosinemia, or lead poisoning (37, 38)]. Then hemealbumin complexes form, and the heme of such complexes is taken up rapidly and selectively by hepatocytes where it is able to replete the regulatory heme pool, repress induction of 5-aminolevulinate synFIG. 7. Lack of effect of heme on nuclear protein binding to AP-1 consensus element. CELCs were treated with 15 M sodium arsenite (as positive control), 10 M heme or 10 M CoPP for 2 h, after which cells were harvested and nuclear proteins were isolated as described under Materials and Methods. A 6-g amount of the nuclear proteins was incubated with or without a 50-fold excess of unlabeled AP-1 or mutated AP-1 oligonucleotide at room temperature for 5 min, after which the ␥- 32P-labeled oligonucleotide containing the AP-1 consensus element (underlined) (5⬘-AGCATGGCTGAGTCAGGCTGGGAG3⬘) was added with 30 min. further incubation at 37°C. The oligonucleotide bound to the nuclear proteins was electrophoretically separated from the unbound form on a 4.5% polyacrylamide gel. The gel was dried and scanned by a PhosphorImager. The lane assignments are as follows: (1) no protein; (2) extracts from untreated cells; (3) extracts from cells treated with Ars; (4) extracts from cells treated with heme; (5) extracts from cells treated with CoPP; (6) extracts from cells treated with Ars plus a 50-fold excess of unlabeled oligonucleotide; (7) extracts from cells treated with Ars plus a 50-fold excess of unlabeled mutated oligonucleotide.
FIG. 6. Mutational analysis of the AP-1-binding elements of pcHO5.6-Luc in CELC: Effects of sodium arsenite, heme, and CoPP. (A) Site-directed mutants of putative AP-1 elements of plasmid pcHO5.6-Luc, corresponding to the diagrammed fragments. (B) Transient transfections were carried out by the calcium phosphate method as described under Materials and Methods. Following transfections, some cells were treated with 15 M sodium arsenite (Ars), 10 M heme, or 10 M CoPP for 15 h in serum-free medium. Cells were harvested and lysed, and assays were performed on 15 L of lysate. Luciferase activities were normalized to -galactosidase activities and protein content. Data are presented as means ⫾ SE, n ⫽ 3. a indicates an increase over wild type (wt), P ⬍ 0.05.
thase, increase levels and activities of holo-cytochromes P450 and tryptophan pyrrolase, and induce HO-1 (2, 39, 40). In summary, metalloporphyrins (especially heme and CoPP) induce HO-1 by a mechanism fundamentally different from that of sodium arsenite and other “stress” inducers. In the chick, the key element(s) required for the metalloporphyrin-mediated induction of HO-1 are located ⫺3.6 to ⫺5.6 kb upstream of the transcription starting point and are different from the AP-1 binding elements that play key roles in stressmediated inductions. Efforts to pinpoint the locations of the metalloporphyrin responsive elements and to identify their cognate binding proteins are ongoing in our laboratory. ACKNOWLEDGMENTS We thank D. L. Williams (SUNY, Stony Brook) for providing a starter supply of LMH cells and O. Gildemeister for helpful discussions. We thank J. Blackwood for help with typing the manuscript and P. LeClair for help with statistical analyses and preparation of final figures. Supported by a grant from NIH (DK 38825 to H.L.B.) and by the Research and Education Fund of the Division of Digestive Disease & Nutrition, UMass Memorial Health Care. The opinions
HO-1 INDUCTION BY METALLOPORPHYRINS expressed in this paper are those of the authors; they do not necessarily reflect the official views of NIH, the USPHS, the University of Massachusetts, or UMass Memorial Health Care.
REFERENCES 1. Maines, M. D. (1988) FASEB J. 2, 2557–2568. 2. Elbirt, K. K., and Bonkovsky, H. L. (1999) Proc. Assoc. Am. Phys. 111, 438 – 447. 3. Poss, K. D., and Tonegawa, S. (1997) Proc. Natl. Acad. Sci. USA 94, 10919 –10924. 4. Poss, K. D., and Tonegawa, S. (1997) Proc. Natl. Acad. Sci. USA 94, 10925–10930. 5. Burnett, A. L., Johns, D. G., Kriegsfeld, L. J., Klein, S. L., Calvin, D. C., Demas, G. E., Schramm, L. P., Nelson, R. J., Snyder, S. H., and Poss, K. D. (1998) Nat. Med. 4, 84 – 87. 6. Yachie, A., Niida, Y., Wada, T., Igarashi, N., Kaneda, H., Toma, T., Ohta, K., Kasahara, Y., and Koizumi, S. (1999) J. Clin. Invest. 103, 129 –135. 7. Elbirt, K. K., Whitmarsh, A. J., Davis, R. J., and Bonkovsky, H. L. (1998) J. Biol. Chem. 273, 8922– 8931. 8. Immenschuh, S., Kietzmann, T., Hinke, V., Wiederhold, M., Katz, N., and Muller-Eberhard, U. (1998) Mol. Pharmacol. 53, 483– 491. 9. Masuya, Y., Hioki, K., Tokunaga, R., and Taketani, S. (1998) J. Biochem. 124, 628 – 633. 10. Cable, E. E., Gildemeister, O. S., Pepe, J. A., Lambrecht, R. W., and Bonkovsky, H. L. (1997) Mol. Cell. Biochem. 169, 13–20. 11. Bonkovsky, H. L., Healey, J. F., Sinclair, P. R., and Sinclair, J. F. (1985) Biochem. J. 227, 893–901. 12. Healey, J. F., Bonkowsky, H. L., Sinclair, P. R., and Sinclair, J. F. (1981) Biochem. J. 198, 595– 604. 13. Gabis, K. K., Gildemeister, O. S., Pepe, J. A., Lambrecht, R. W., and Bonkovsky, H. L. (1996) Biochim. Biophys. Acta, Gen. Subj. 1290, 113–120. 14. Kolluri, S., Elbirt, K. K., and Bonkovsky, H. L. (1999) Biochim. Biophys. Acta 1472, 658 – 667. 15. Lincoln, B. C., Aw, T. Y., and Bonkovsky, H. L. (1989) Biochim. Biophys. Acta 992, 49 –58. 16. Lu, T. H., Pepe, J. A., Gildemeister, O. S., Tyrrell, R. M., and Bonkovsky, H. L. (1997) Biochim. Biophys. Acta, Gene Struct. Expression 1352, 293–302. 17. Lu, T. H., Lambrecht, R. W., Pepe, J., Shan, Y., Kim, T., and Bonkovsky, H. L. (1998) Gene 207, 177–186. 18. Sambrook, J., Fritsch, E. F., and Maniatis, T. (1989). Molecular Cloning: A Laboratory Manual Cold Spring Harbor Laboratory Press, Plainview, NY. 19. Attardi, L. D., and Tjian, R. (1993) Genes Dev. 7, 1341–1353.
227
20. Cable, E. E., Pepe, J. A., Karamitsios, N. C., Lambrecht, R. W., and Bonkovsky, H. L. (1994) J. Clin. Invest. 94, 649 – 654. 21. Tyrrell, R. M., and Basu-Modak, S. (1994) Methods Enzymol. 234, 224 –235. 22. Stocker, R., Yamamoto, Y., McDonagh, A. F., Glazer, A. N., and Ames, B. N. (1999) Science 235, 1043–1046. 23. Marks, G. S. (1994) Cell. Mol. Biol. 40, 863– 870. 24. Maines, M. D. (1996) Methods Enzymol. 268, 473– 488. 25. Maines, M. D. (1997) Annu. Rev. Pharmacol. Toxicol. 37, 517– 554. 26. Snyder, S. H., Jaffrey, S. R., and Zakhary, R. (1998) Brain Res. Brain Res. Rev. 26, 167–175. 27. Russo, S. M., Pepe, J. A., Cable, E. E., Lambrecht, R. W., and Bonkovsky, H. L. (1994) Eur. J. Clin. Invest. 24, 406 – 415. 28. Eskew, J. D., Vanacore, R. M., Sung, L., Morales, P. J., and Smith, A. (1999) J. Biol. Chem. 274, 638 – 648. 29. Sinclair, P. R., Bement, W. J., Gorman, N., Liem, H. H., Wolkoff, A. W., and Muller-Eberhard, U. (1988) Biochem. J. 256, 159 – 165. 30. Sinclair, P. R., Bement, W. J., Healey, J. F., Gorman, N., Sinclair, J. F., Bonkovsky, H. L., Liem, H. H., and Muller-Eberhard, U. (1994) Hepatology 20, 741–746. 31. Potter, D., Chroneos, Z. C., Baynes, J. W., Sinclair, P. R., Gorman, N., Liem, H. H., Muller-Eberhard, U., and Thorpe, S. R. (1993) Arch. Biochem. Biophys. 300, 98 –104. 32. Taketani, S., Immenschuh, S., Go, S., Sinclair, P. R., Stockert, R. J., Liem, H. H., and Muller-Eberhard, U. (1998) Hepatology 27, 808 – 814. 33. Bonkovsky, H. L., Gildemeister, O. S., Pepe, J. A., Lambrecht, R. W., and Hrkal, Z. (1996) Eur. J. Clin. Chem. Clin. Biochem. 34, 931–932. 34. Noyer, C. M., Immenschuh, S., Liem, H. H., Muller-Eberhard, U., and Wolkoff, A. W. (1998) Hepatology 28, 150 –155. 35. Sinclair, J. F., Sinclair, P. R., Healey, J. F., Smith, E. L., and Bonkowsky, H. L. (1982) Biochem. J. 204, 103–109. 36. Cable, E. E., Cable, J. W., and Bonkovsky, H. L. (1993) Hepatology 18, 119 –127. 37. Muller-Eberhard, U., and Fraig, M. (1993) Am. J. Hematol. 42, 59 – 62. 38. Muller-Eberhard, U., and Liem, H. H. (1974) in Structure and function of plasma proteins (Allison, A. C., Ed.), pp. 35–53, Plenum Press, London. 39. Bloomer, J. R., and Bonkovsky, H. L. (1989) DM, Dis.-Mon. 35, 1–54. 40. Bonkovsky, H. L. (1990) in “Hepatology: A Textbook of Liver Disease” (Zakim, D., and Boyer, T. D., Eds.), pp. 378 – 424, Saunders, Philadelphia.