THE JOURNAL OF BIOLOGICAL CHEMISTRY Q 1993 by The American Society for Biochemistry and Molecular Biology, Inc.

Vol. 268, No. 10, Issue of April 5, pp. 7365-7371,1393 Printed in U.S.A.

Regulation of Heme Oxygenase and Metallothionein Gene Expression by the Heme Analogs, Cobalt-, and Tin-Protoporphyrin*

(Received for publication, July 9, 1992)

Ann Smith$#, Jawed AlamT, Pablo V. EscribaSII, and William T. Morgan$ From the $Division of Molecular Biology and Biochemistry, School of Biological Sciences, University of Missouri-Kansas City, Kansas City, Missouri 64110 and the TDepartment of Molecular Genetics, Alton Ochsner Medical Foundation, New Orleans, Louisiana 70121

Two heme analogs, cobalt- and tin-protoporphyrin (CoPP and SnPP, respectively) have been used to probe the heme-hemopexin interaction, hemopexin receptor binding, and the mechanism of regulation of heme ox- ygenase (HO) and metallothionein-1 (MT-1) gene expression by hemopexin. Both CoPP and SnPP are HO inhibitors and hemopexin binds SnPP (Morgan, W. T., Alam, J., Deaciuc, V., Muster, P., Tatum, F. M., and Smith, A. (1988) J. Biol. Chern. 263,8226-8231) and CoPP. The association of CoPP with hemopexin produces characteristic changes in the absorbance spectrum of CoPP and quenches the intrinsic fluores- cence of hemopexin. Binding of CoPP is tight (& ca. 3 X 10” M) although of lower affinity than heme itself (& < p ~ ) ; and CoPP binding, like heme, produces conformational changes in hemopexin shown by an increase in the molar ellipticity at 233 nm and affords protection from proteolysis of the hinge region be- tween the two structural domains of hemopexin. The coordination of the central cobalt atom is predicted to be similar to that of heme and to involve His66 and Hisl2, of rabbit hemopexin. Furthermore, CoPP-hem- opexin, like SnPP-hemopexin, binds to the hemopexin receptor as shown by competitive inhibition studies with radioactive heme-hemopexin.

The effect of free heme analogs and their hemopexin complexes on HO and MT gene regulation was inves- tigated and compared with the extent of induction by heme and heme-hemopexin. Free CoPP is a more effec- tive inducer of HO steady state mRNA levels than free heme and produces a &fold increase within 1 h com- pared to only a 2-fold increase with heme, but free SnPP (up to 10 PM) produces no detectable increase in HO mRNA. In contrast, by 3 h heme-hemopexin and SnPP-hemopexin increase HO mRNA levels 11- and 6-fold, respectively; but the CoPP-hemopexin complex causes no detectable change in HO mRNA levels.

The complexes of hemopexin with heme or either of the two heme analogs are effective inducers of metal- lothionein (MT) mRNA. Induction of MT mRNA by heme-hemopexin is rapid, increasing 4-fold within 1 h and 14-fold by 3-4 h. Strikingly, an even more rapid and slightly more extensive induction of MT mRNA is

* This work supported by National Institutes of Health Grant DK- 37463 (to A. S.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

To whom correspondence should be addressed: Div. of Molecular Biology and Biochemistry, School of Biological Sciences, University of Missouri, Kansas City, MO 64110. Tel.: 816-235-2579; Fax: 816-

)I Present address: Dept. of Biologia, Universidad Islas Baleares, 235-5158,

Cta Valldemossa Km. 7.5, 07071 Palma de Mallorca, Spain.

seen in response to either CoPP- or SnPP-hemopexin complexes, with MT mRNA rising %fold within 1 h. In contrast, free heme and the free analogs are far less effective inducers, increasing MT mRNA levels and in vitro transcription rates only 3-4-fold and declining after 2-3 h.

The uptake of SnPP by Hepa cells is facilitated by hemopexin but interestingly, binding to hemopexin significantly retards the uptake of CoPP. Both free CoPP and SnPP alone are accumulated by cells at 37 “C. In conclusion, since CoPP-hemopexin binds to the hemopexin receptor yet CoPP uptake is minimal, and since MT mRNA levels are induced, while those of HO are not, occupation of the hemopexin receptor alone appears to be sufficient to influence intracellular events, such as a signaling pathway, important for MT gene transcription.

Heme oxygenase (HO; EC 1.14.99.3),’ the rate-limiting enzyme in heme catabolism producing iron and biliverdin IXa , is rapidly induced in many cell types by incubation with exogenous heme (1, 2) or by heme or hemoglobin injected intravenously in rats (3-5). HO and amino-levulinate syn- thase are considered to play key roles in the control of intracellular heme turnover needed for efficient heme-protein synthesis. Induction of HO occurs in a variety of tissues, particularly liver, spleen, and kidney in vivo (6). Of interest here is the fact that hemopexin-mediated heme transport is an important factor in HO regulation since the hemopexin receptor provides a means to direct heme intracellularly uia a defined route (7) and since there is no evidence for free heme in the circulating plasma (8). Moreover, heme-hemopexin has been shown to rapidly cause the induction of HO mRNA by activation of transcription of the HO gene in mouse hepatoma and human promyelocytic HL-60 cells (9), and the extent of induction appears to be directly related to intracellular heme concentration (8).

Heme-hemopexin not only causes a rapid stimulation of the expression of HO (9) but also increases the steady state level of metallothionein (MT) mRNA in a time- and dose- dependent manner (10). Comparison of the regulation of HO and MT genes in response to free heme or to heme-hemopexin points to significant differences in the mechanism of gene regulation by heme (10). For example, in contrast to HO regulation, heme-hemopexin is a far more effective inducer of M T isozyme 1 than free heme which readily accumulates in

‘The abbreviations used are: HO, heme oxygenase; heme, iron- protoporphyrin; CoPP, cobalt-protoporphyrin IX; SnPP, tin-proto- porphyrin IX, MT, metallothionein; DMEM, Dulbecco’s modified Eagle’s medium.

7365

7366 Heme Analog-mediated Gene Regulation

cells in uitro. This suggests that, for increased accumulation of MT-1 transcripts, an interaction between the hemopexin complex and its receptor and/or targetting of heme via the hemopexin receptor pathway is necessary. The concomitant regulation of HO and MT in response to heme-hemopexin may be related to the proposed role for MT as an intracellular antioxidant capable of trapping hydroxyl and superoxide rad- icals. Surface events, including binding to the hemopexin receptor followed by heme release with subsequent heme transport and enhanced intracellular levels of heme and iron moving through the cell, may increase the levels of radicals generated during normal cellular metabolism. Alternatively, or in addition, MT induction by heme-hemopexin may be due to a need to sequester intracellular zinc which would compete with iron and occupy sites on regulatory proteins.

Moreover, inhibition of protein synthesis by cycloheximide augments the heme-hemopexin-mediated accumulation of MT-1 mRNA but abrogates induction of HO gene expression (10). This suggests that transcriptional regulatory activity by heme or metals may require decreased concentrations or inactivation of a labile repressor in the case of MT as well as synthesis or cycloheximide-sensitive processing of additional inducer-specific trans-acting factor(s) for both MT and HO.

Understanding the regulation of expression of HO is com- plicated by the fact that regulatory responses are not re- stricted to heme. Heavy metals including cobalt and cadmium also increase HO enzymic activity in rats (6, 11-13) and HO mRNA in hepatoma (14) and other cultured cell lines (15). This regulation of the HO gene is considered to involve two2 metal-regulatory elements (MRE (16), similar in heptad core (TGCA/TCNC) to the five MREs (TGCPuCNC) present in the MT gene (17, 18). Cd2+ not only activates HO transcrip- tion but also acts to stabilize mRNA simultaneously, e.g. increasing HO mRNA half-life from 5.3 h to 6.2 h in mouse hepatoma cells (31). In addition, some specificity exists in the response of HO and MT MRE'S to metals since the levels of extracellular zinc, which potently induce MT (19), do not affect HO mRNA levels (10). Further, Cd2+ is a far more effective inducer of HO mRNA than Co2+ on a molar basis in mouse hepatoma cells (9).

Specific heme-responsive elements (HRE) and labile, heme- inducible trans-acting factor(s) (10) have been proposed to be involved in HO gene transcriptional activation after heme transport by the hemopexin receptor (10). Due to similarities in structure and chemistry, heme-analogs, like CoPP and tin- protoporphyrin (SnPP), may bind to the same regulatory sites as heme itself. Since metals other than iron, e.g. Co2+, can substitute as a substrate for ferrochelatase in vitro (24) and CoPP is a competitive inhibitor of HO (23), we investigated whether CoPP also regulates HO gene expression and whether some of the observed effects of metals on HO expression (25) could be due to intracellular formation of heme analogs.

We have previously shown that HO mRNA is rapidly increased when the heme analog SnPP bound to hemopexin is transported into hepatoma cells via the hemopexin receptor (20). The high affinity interaction of SnPP-hemopexin with the hemopexin receptor suggests that SnPP is transported by the same pathway used by heme to HO. However, SnPP-like CoPP is a competitive inhibitor of HO (X) , not a substrate. The induction of HO mRNA in response to SnPP-hemopexin (20) provides a mechanism whereby HO protein levels are induced in vivo (22), yet the HO enzymic activity is inhibited (23).

Here, Co-PP was used to probe the heme-binding site of HPX, and both CoPP and SnPP were used to define further

J. Alam and A. Smith, manuscript submitted.

the elements involved in the heme-mediated regulation of expression of the HO and MT genes. A preliminary report of some of this work has been presented (45).

EXPERIMENTAL PROCEDURES

Materials-All enzymes were purchased from Bethesda Research Laboratories. Zeta-Probe nylon membrane was obtained from Bio- Rad; the random priming DNA labeling kit, from Pharmacia LKB Biotechnoloy Inc.; and [a3'P]dCTP, [c~-~'P]UTP, and [T-~*P]ATP, from Du Pont-New England Nuclear. Protein synthesis inhibitors were from Sigma; and mesoheme (iron-mesoporphyrin IX), cobalt-, tin- and iron-protoporphyrin IX (heme), from Porphyrin Products (Logan, UT). All other chemicals were reagent grade. Mesoheme is used here rather than the less stable protoheme, and mesoheme- hemopexin is chemically and biologically equivalent to protoheme- hemopexin (26, 27). Except where specifically indicated, mesoheme is referred to as heme.

Isolation and Characterization of Hemopexin, Its Heme and Heme Analog Complexes-Hemopexin was isolated from rabbit serum and mesoheme- and protoheme-hemopexin complexes were prepared, pu- rified, and characterized as previously described (27,28). Stock solu- tions of CoPP (-1 mM) were made fresh in 15 mM sodium phosphate buffer, pH 7.4, and the concentration was determined in 0.1 M NaOHpyridine:H20 (3:1017, v/v) using an extinction coefficient ( E = M" cm") of 1.8 X lo5 at 424 nm (29). To minimize dimerization, solutions of SnPP or mesoheme in dimethyl sulfoxide were prepared fresh each day, and their concentrations were determined using an E at 406 nm of 1.64 X 10' in 0.5% pyridine containing 1 drop of NH,OH/ 100 ml for SnPP3 and 1.7 X lo5 at 394 nm in dimethyl sulfoxide for mesoheme. Complexes of hemopexin with mesoheme, CoPP, or SnPP were prepared by mixing 1.1 equivalent of tetrapyrrole with 1 equiv- alent of protein. Unbound ligand was removed by dialysis at 4 "C. Full saturation of the hemopexin complexes was confirmed, and concentration of stock solutions was determined by absorbance spec- troscopy using published procedures (27,28). Estimations of binding affinity were assessed from the addition of 0.2 WM increments of CoPP to a 2 FM solution of rabbit hemopexin in PBS, pH 7.4, and analyzing the binding data using ENZFITTER (Biosoft, Elsevier).

Cell Culture-The growth, maintenance, and incubation conditions of human myeloid leukemia HL-60 cells (706) and of Hepa cells (99) have been described. Additional details are given in the figure legends. To investigate the transport of CoPP and SnPP, mouse Hepa cells

bated for up to 2 h at 37 "C in HEPES-buffered DMEM, pH 7.4, in exponential growth were rinsed free of culture medium and incu-

containing either CoPP, SnPP, CoPP-hemopexin, or SnPP-hemo- pexin (10 p ~ ) . To stop transport, the experimental medium was removed by aspiration, and the cells rinsed in ice-cold Dulbecco's phosphate-buffered saline (3 X 2 ml) followed by extraction of SnPP into 1 N perchloric acidmethanol (5050, v/v) using a procedure based on Anderson et al. (30). CoPP was extracted by the addition to each well of 1.0 ml of 1 N perchloric acid:methanol mixed with dimethyl sulfoxide (17% v/v). The cell extracts were then centrifuged for 5 min in an Eppendorf microcentrifuge, and the amount of heme analog in the supernatant was estimated using fluorescence or absorbance spectroscopy for SnPP and CoPP, respectively. Standard curves for SnPP or CoPP determinations were made by adding aliquots of fresh stock solutions of the heme analogs into cell extracts isolated from control cells incubated in Hepes-buffered DMEM. Emission spectra of SnPP were recorded in cell extracts from 500-700 nm after exci- tation at 405 nm (IO-mm slit width) using a Perkin-Elmer 650-40 spectrofluorimeter. CoPP concentrations in cell extracts were deter- mined from the absorbance spectrum recorded from 340-450 nm using a concentration range of 0.05-0.35 FM.

Plasmids, Preparation, and Analysis of RNA-The mouse MT-1 gene (from nucleotide residue -750 to +1240) cloned into the vector pBX322 was kindly provided by Dr. R. Palmiter (University of Washington, Seattle, WA).

Total (31) and cytoplasmic RNA (32) were isolated, and Northern and RNA dot blot analyses were carried out as described (31). To enable specific measurement of MT-1 mRNA by quantitative primer

prepared by labeling a heptadeca-nucleotide (5"CGGTGGAGCAG- extension analysis (33), the mouse MT-1-specific DNA probe was

GAGCAG-3') complementary to residues 85-101 of the MT-1 gene (34) at the 5'-end with [T-~~PIATP using T4 polynucleotide kinase.

B. Burnham, personal communication.

Heme Analog-mediated Gene Regulation 7367

After hybridization to alkali-denatured pBX322/MT-l DNA and extension using the Klenow fragment of Escherichia coli DNA polym- erase I and unlabeled deoxynucleoside triphosphates, the extended product was digested with the restriction endonuclease RsaI, resolved on a polyacrylamide/urea gel, and the 75-base-long, single-stranded probe was subsequently eluted. The mouse HO probe was an end- labeled dodeca-nucleotide (5’-TCGGGCTGTGGACGCTCCAT-3’) complementary to the sense strand of the HO gene (35) from residues +129 to +148. This was hybridized to denatured plasmid DNA containing the appropriate genomic sequence and extended using the Klenow fragment. The extended product was digested with XhoI, and the 79-base-long, single-stranded probe was isolated as described above. Autoradiographic signals were quantitated using a Biomed laser scanning densitometer, Model SLR 2D/lD.

Cell Binding Studies-The minimal deviation hepatoma cells (mouse Hepa cells) derived from the mouse solid tumor BW 7756 were grown in Dulbecco’s modified Earle’s medium containing 2% fetal calf serum as previously described (38). Binding of ‘261-labeled complexes was measured by adding 1-ml aliquots of Hepes-buffered DMEM, pH 7.4, containing the protein and either heme-hemopexin or CoPP-hemopexin as competitive inhibitor to 1-1.5 X 10‘ cells. After 3 h at 4 “C the cells were washed and harvested, and the amount of radioactive heme-’261-hemopexin specifically bound determined using previously published procedures (38, 39). Additional details are given in the figure legends.

Regulation of Heme Oxygenase and Metallothionein-Measurement of cellular HO and MT mRNA levels was carried out by Northern analysis using probes and experimental conditions previously de- scribed in detail (10, 31). Nuclei preparations (36), nuclear run on assays, and isolation of labeled RNA (37) with hybridization and washing were as described previously (31). In general, all induction experiments were repeated at least twice. For the nuclear run-ons, after exposure of the filter to the x-ray film, the “dot” samples were excised from the filter, and the amount of radioactive probe hybridized was quantitated using liquid scintillation radiometry. The -fold in- creases were calculated for each set of data by normalizing with respect to the time zero dpm for each as described previously (31), and the extent of variation in induction was always less than 10% of the mean for separate experiments.

RESULTS

Characterization of the Interaction of Cobaltprotoporphyrin IX with Hemopexin: Binding and Conformational Changes- Hemopexin binds the heme analog, CoPP, as shown by char- acteristic changes in the absorbance spectrum of CoPP upon incubation with hemopexin (Fig. 1, A and B). There is an increase in absorbance and a red shift of the maximum wavelength from 419 to 426 nm. The heme-iron is normally hexa-coordinated, interacting with the 4 tetrapyrrole nitro- gens and 2 histidine residues, HisM and His’27 of the protein4 (40), and a similar coordination is predicted for Co. The affinity of hemopexin for CoPP was determined to have a Kd of - 0.3 p~ from measurements of change in absorbance upon adding CoPP to hemopexin (Fig. 1C). Similar results ( K d of 0.45 p ~ ) were obtained by analysis of the quenching of the intrinsic tryptophan fluorescence of hemopexin upon binding increasing amounts of CoPP (Fig. 1D).

The binding of CoPP produces conformational changes in hemopexin similar to those produced by heme. These include an increase in the unusual positive molar ellipticity at 233 nm in the circular dichroism spectrum of hemopexin, albeit some- what smaller than seen with heme (see Fig. 2 A ) , and protec- tion from proteolysis of the hinge region between the two structural domains of the protein (27,41) upon ligand binding (Fig. 2B).

Comparison of the Affinity of CoPP-Hemopexin and Heme- Hemopexin Complexes with the Hemopexin Receptor-The conformational changes in hemopexin produced by heme binding are necessary for tight interaction with its specific

‘ W. T. Morgan, P. Muster, F. M. Tatum, S.-M. Kao, J. Alam, and A. Smith, manuscript submitted.

receptor (26). To address whether CoPP-hemopexin would bind to the hemopexin receptor, competitive binding studies were carried out in mouse Hepa cells (Fig. 3). Both COW- hemopexin and heme-hemopexin produce similar amounts of specific binding of heme-’251-hemopexin. Thus, CoPP-hemo- pexin binds to the hemopexin receptor with approximately the same affinity as heme-hemopexin complexes.

Regulation of Heme Oxygenase Expression by Heme Analogs and Their Complexes with Hemopexin: Comparison with That by Heme and Heme-Hemopexin-The extent of induction of HO mRNA was determined over 4 h. Within 1 h of incubation of Hepa cells with free heme or CoPP (10 p ~ ) HO mRNA is induced and steadily increases for 3-4 h (Fig. 4A). Interest- ingly, CoPP is a more effective inducer than heme, the natural substrate for HO, producing almost twice the amount of mRNA. In contrast, SnPP produced no detectable induction of HO under the same experimental conditions.

Heme-hemopexin was an effective inducer of HO mRNA producing an 11-fold increase by 3 h, and SnPP-hemopexin, unlike free SnPP, also caused induction, but somewhat less effectively, increasing HO mRNA 6-fold (Fig. 4A). Notably, CoPP, when presented to the cells complexed with hemo- pexin, was without effect on HO mRNA levels, unlike free CoPP.

Regulation of Metallothionein Gene Expression by Heme Analogs and Their Complexes with Hemopexin: Comparison with Heme and Heme-Hemopexin-Heme-hemopexin (10 PM) increased MT mRNA approximately 4-fold by 1 h in mouse Hepa cells, and the levels continued to increase for 3-4 h up to 14-fold (Fig. 4B). An even more rapid induction of MT reaching maximum levels (13-16-fold) within 2 h was ob- served in response to both the CoPP- and SnPP-hemopexin complexes. In contrast, free heme increased MT mRNA levels only 2-3-fold within 4 h (Fig. 5). Free CoPP caused a %fold increase by 1 h, but the mRNA levels decreased to control amounts by 4 h (Figs. 4B and 5). The time course of induction by free SnPP was similar to that of CoPP, but the effect was maintained for 4 h.

Nuclear run-on analysis (Fig. 6) showed that CoPP acti- vated transcription of both HO and MT genes. The effect on HO was extensive, but only a small, dose-related effect on MT was apparent. MT gene transcription, like that of HO (data not shown), was increased by SnPP. No effect was seen on the actin, L7, and PEPCK control genes.

Comparison of the Extent of Uptake of Heme Analogs by Mouse Hepa Cells Incubated with Free Analogs or Their Hem- opexin Complexes-Because free CoPP is an effective inducer of HO, the lack of induction of HO by CoPP-hemopexin suggested that CoPP was not being transported into these cells even though the CoPP-hemopexin complex binds to the receptor (vide supra). The amount of CoPP taken up by Hepa cells was determined after incubating the cells with either CoPP or CoPP-hemopexin at 37 “C. As shown in Fig. 7A, cells accumulated -83 and 150 pmol of CoPP by 1 and 2 h after incubation with 10 WM CoPP, but no detectable uptake occurred when cells were incubated with CoPP-hemopexin. When Hepa cells were incubated with 10 p~ SnPP or SnPP- hemopexin for 1 or 2 h, binding to hemopexin almost doubled the amount of SnPP transported (Fig. 7B). Intracellular concentrations after 1- and 2-h incubation were 0.65 and 0.83, and 1.15 and 1.65 pmol/well with SnPP and SnPP-hemo- pexin, respectively. These results show that hemopexin facil- itates the uptake of SnPP into cells, but in contrast, prevents or significantly retards the uptake of CoPP.

7368 Heme Analog-mediated Gene Regulation

0.2 0 0

0 n 0

n 4

0.1 .. - CI

0.0

2.0

W 0 C 0

1.5

2 1.0

0

a v)

0.5

0.0 I 250 300 350 400 450 500 550 600 650

Wavelength (nrn)

C

0 e

0.4

0.3

0.2

0.1

0.0

ow4 0 . o l 0.0 " " " " ' 0.2 0.4 0 . 6 0.5

[CoPP 1x1. pM [CoPP 1x1. pM

0.6

0.4 0

L - I - 0.2

FIG. 1. Spectroscopic characterization of the CoPP-hemopexin complex. A and B, absorbance spectra of CoPP and CoPP- hemopexin. Absorbance spectra of free CoPP (dotted l i n e ) and the CoPP-hemopexin complexes (solid line) in 10 mM sodium phosphate buffer, pH 7.4, are shown. The visible spectra are shown in the righthand panel B (with an expanded scale). The concentration of CoPP and complex was 12 pM. C and D, analysis of the binding of CoPP by hemopexin using absorbance and fluorescence spectroscopy. A titration of the binding of hemopexin (2 p ~ ) in 10 mM sodium phosphate buffer, pH 7.4, containing 0.15 M NaCl by the addition of 0.2 PM increments of CoPP is presented in C. A nonlinear regression analysis yielded an apparent K d of 0.3 p~ & 0.01. A titration of the quenching of the intrinsic fluorescence of hemopexin upon addition of increasing amounts of CoPP is shown in D. Nonlinear regression analysis yielded an apparent Kd of 0.46 p~ & 0.05. Additional experimental details are given under "Experimental Procedures."

DISCUSSION

The heme analogs CoPP and SnPP were used to probe the heme-hemopexin interaction, receptor binding, and the ef- fects of hemopexin on the regulation of expression of the HO and MT genes. These analogs offer unique properties for this purpose. They are both strong competitive inhibitors of heme oxygenase; they bind to hemopexin (the SnPP interaction with hemopexin and effects on HO expression were described previously (20)), and the analog-hemopexin complexes bind to the hemopexin receptor. Furthermore, CoPP but not SnPP can be formed enzymically by ferrochelatase. Here, SnPP was used to contrast with CoPP and to extend the results for MT gene regulation (Ref. 10 and see below).

Hemopexin has been shown here to bind CoPP similarly to heme, probably via 2 histidine residues and undergoes the conformational changes associated with heme binding (111) necessary for receptor recognition (26, 41, 42). Additional evidence for similarities between heme and CoPP binding have come from differential scanning calorimetry studies, where the melting temperature (T,) for heme-hemopexin is 67 "C and for CoPP-hemopexin is 69 " C 6 CoPP is bound by hemopexin with similar affinity to that previously determined for cobalt-meso- and cobalt-deutero-porphyrin IX (Kd of 1 x

M (43)). In addition, competitive inhibition studies dem- onstrated that CoPP-hemopexin complexes bind to the hem-

' M.-L. Wu and W. T. Morgan, manuscript submitted.

opexin receptor on mouse Hepa cells with approximately the same affinity as heme-hemopexin.

CoPP is shown here to be a far more effective inducer of HO mRNA than heme and than cobalt (31). Furthermore, the lack of induction of heme oxygenase by CoPP-hemopexin (in contrast to free CoPP) suggests that, although CoPP- hemopexin binds to the hemopexin receptor like heme-hem- opexin, subsequent events differ. While CoPP levels increased intracellularly after incubation of hepatoma cells with free CoPP, no detectable uptake occurred with CoPP-hemopexin. This suggests that CoPP bound by hemopexin it is not readily transported into cells (see below).

These observations help to explain some discrepancies in the literature concerning the effects of CoPP on heme oxy- genase gene expression in vitro compared with its effects in intact rats. Hepatic HO enzymic activity increases in rats injected subcutaneously with either CoPP or SnPP (23). The effect of CoPP on HO was greater than SnPP, although the hepatic accumulation of CoPP was significantly less. (Hepatic cobalt and tin, presumably as CoPP and SnPP, respectively, were 4-6 and 20-40 pg/g liver, respectively, at 4-8-h postin- jection.) In these studies the rats received the heme analogs by subcutaneous injection, which makes determination of the circulating form presented to hepatocytes difficult. However, it is likely that these two heme-analogs, like heme, are bound to hemopexin, albumin, and (if present at high enough con- centrations for sufficient time) to lipoproteins. CoPP would

Heme Analog-mediated Gene Regulation 7369

A.

7

- opo-hernopexin "". heme-hemopex in

.......... CoPP-hernopexin

I 220 230 240 250 260

Wavelength (nm) 0.

1 2 3 4 5 6 7 8 9 1 0 1 1 1 2 1 3

FIG. 2. Conformational changes in hemopexin associated with binding CoPP. A , circular dichroism spectra of CoPP-hemo- pexin complexes. Circular dichroism spectra of apo-hemopexin (solid line), heme-hemopexin (dashed line), and CoPP-hemopexin (dotted l i n e ) were recorded at 25 "C in 100 mM sodium phosphate buffer, pH 7.4. The hemopexin concentration was 14.5 p M in each case. Data are presented as molar ellipticity, deg.cm2.dmol". B, protection of the hinge region of hemopexin from proteolysis by CoPP. Hemopexin was treated with plasmin (1:50 w/w ratio in phosphate-buffered saline a t 25 "C) in the presence and absence of ligands, electrophoresed on a sodium dodecyl sulfate-polyacrylamide gel and stained with Coo- massie Blue. Shown are: apo-hemopexin after 0, 30, 60, and 90 min exposure to plasmin (lanes 1-4, respectively) and similar time courses for heme-hemopexin and CoPP-hemopexin in lanes 5-8 and lanes 9- 12, respectively. Lune 13 contained molecular weight standards. Ap- proximately 20 pg of protein were applied to each lane.

be able to enter the liver after dissociation from albumin or lipoproteins and rapidly activate the HO gene, whereas CoPP bound to hemopexin would not. In fact, binding of CoPP by hemopexin i n uiuo would delay or prevent hepatic uptake. The observed induction i n vivo is not due to free cobalt, because CoPP, unlike heme, is not a substrate for HO. How- ever, transcriptional regulation of the HO gene (25) after cellular uptake of Co2+ is probably due to combined effects of the metal itself uia the two metal responsive elements on the HO promoter as well as from the effects of CoPP formed intracellularly via ferrochelatase on the heme-responsive ele- ment.

The induction of HO mRNA by free CoPP is far more rapid and extensive than previously seen in these hepatoma cells with the metal Co2+ (50 p~ (31)). This contrasts with the conclusions of Lin et al. (25), who reported that after subcu- taneous injections in rats Co2+ was a more effective inducer of HO than CoPP, but who, as pointed out above, overlooked

I I I

T - c 300 c Q

0 L n m E 200

2 S 3 0

.a 100 - 0

LC E

I I I I 0 2000 4000 6000 aooo

Metalloporphyrin-HPX, uM

FIG. 3. Competit ive inhibition of binding of heme-'Z61-hem- opexin to mouse hepatoma cells by CoPP-hemopexin. Binding studies were carried out using published procedures (16) described under "Experimental Procedures." Briefly, mouse Hepa cells were rinsed in ice-cold HEPES-buffered DMEM ( 3 X 2 ml), followed by addition of radioactive ligand (70 nM heme-'2sII-hemopexin) dissolved in binding buffer in the presence or absence of nonradioactive heme- hemopexin (solid circles) or CoPP-hemopexin complexes (triangles).

A. B. 0 1 2 3 4 0 1 2 3 4

moo e 0 Heme 0 . 0 0 0 0 Heme-HPX .. 0 0. CoPP-HPX

e 0 COPP

0 I 0 0 SnPP.HPX 0 0 c SnPP

FIG. 4. Regulation of HO and MT expression by heme, CoPP, SnPP, a n d b y their complexes with hemopexin. A , time course of the induction of heme oxygenase mRNA levels. Mouse Hepa cells (-2 X lo6 cells/well) were incubated in 2 ml of serum-free DMEM containing heme, heme-analogs, or their hemopexin com- plexes (all a t 10 p ~ ) for 0, 1, 2, 3, or 4 h as indicated. Total RNA (5 pg) was applied to nylon membranes and hybridized to the HO probe, and the bound probe was detected by autoradiography. The data presented are from one individual experiment carried out twice. R, time course of induction of M T mRNA levels. Mouse Hepa cells (-2 X lo6 cells/well) were incubated in 2 ml of serum-free DMEM containing heme, heme-analogs, or their complexes with hemopexin (all a t 10 qM) for 0, 1, 2, 3, or 4 h as indicated. Total RNA (5 pg) was applied to nylon membranes and hybridized MT-1 probe detected by autoradiography. The data presented are from one individual exper- iment which was repeated twice.

the binding of CoPP to hemopexin and other circulating plasma proteins. The induction of HO by CoPP shown here in hepatoma cells appears to be due to targeting of the heme analog to sites involved in normal regulation by heme, possi- bly complexed to a protein (IO), that binds to a cis heme- responsive element in the HO promoter.

Regulation of MT gene expression by heme and heme analogs presented to cells via the hemopexin receptor appears to be controlled in an even more complex manner than HO. Clearly, M T expression does not respond as extensively to changes in intracellular heme or heme analog levels as the HO gene does, because nonprotein-bound heme, CoPP, and SnPP are far less effective inducers of MT than their respec- tive hemopexin complexes. Nevertheless, MT gene transcrip- tion is increased by heme and heme analogs providing evi-

7370 Heme Analog-mediated Gene Regulation

Heme Oxygenase rnRNA Metallothionein mRNA

0 1 2 5 4 - 2 40 1 CoPP-Hernopexih D

g 10 0: 2ol

0- 0 1 2 3 4

10 r

::MI 1 40 CoPP

10

0 0 1 2 3 4

l 0 I SnPP

5

0 1 2 5 4 0 1 2 3 4

Hours Hours

2o [Heme-Hemopexin 2o [ Heme

15

10

5

0 IdJ 5 0 0 1 2 5 4 2 o 0 1 2 3 4

2o r CoPP-Hemopexln [ COPP

15

10

0 5 : L b 5

0 1 2 3 4 0 1 2 3 4 0

2o r SnPP-Hemopexin 2o r S ~ P P

15

10

0 5 ldm 5

0 1 2 3 4 0 1 2 3 4 0

Hours Hours

FIG. 5. Comparison of the extent of induction of HO and MT mRNA when heme, CoPP, and SnPP are delivered to cells via the hemopexin receptor. The experimental conditions are the same as described in the legend to Fig. 4. The open and hatched areas show the extent of HO (LHS panels) or M T (RHS panels) mRNA induction after cells were incubated with either heme (upper panels), CoPP (middle panels), or SnPP (lower panels) free or bound to hemopexin, respectively. The relative amounts of steady state HO and MT mRNA were determined by quantitation of the amount of bound probe (data shown in Fig. 4, A and B ) using liquid scintillation counting and the extent of induction normalized to the time zero sample in each set as described under “Experimental Procedures.” The data presented in this figure are the average values of two individual experiments.

puc9

HO 0. B-Actin 0 0

L7

MT 6 0 PEPCK 0 0 d,

FIG. 6. Stimulation of MT-1 and HO gene transcription in isolated nuclei from Hepa cells incubated with CoPP. Mouse Hepa cells (-4 X lo7 cells/flask) were incubated in 20 ml of serum- free Dulbecco’s modified Eagle’s medium containing either 5 or 10 p~ CoPP or buffer alone for 2 h at 37 “C. In vitro transcription reactions were carried out and plasmid DNA (5 pg) blotted onto Zeta- Probe nylon membranes. Each filter was then incubated with labeled transcript (-5 X lo6 cpm) for 72 h followed by autoradiography. The results from one of two independent nuclear run-on experiments are shown. PEPCK, phosphoenolpyruvate carboxykinase; L7, ribosomal protein L7.

dence that gene expression is regulated in part via a heme- responsive element in the M T promoter. M T induction with- out uptake of CoPP from CoPP-hemopexin also demonstrates that MT regulation takes place in response to events which occur upon binding of hemopexin to its receptor independ- ently of metalloporphyrin uptake. This may involve PKC- mediated phosphorylation and lead to intracellular signaling,

resulting in an activation of M T gene transcription. An ob- vious target for phosphorylation would be the transcription factors and their associated factors involved in MT gene regulation. Along these lines, we have shown that phorbol esters double the rate of endocytosis of hemopexin via acti- vation of protein kinase C (39).

In sum, regulation of genes in response to heme-hemopexin can be divided into at least two groups: those which respond directly to increases in intracellular heme levels (e.g. the HO and hemopexin genes) and those which respond to surface events including hemopexin receptor occupancy ( eg . the mouse M T I and 116 genes). Additional levels of complexity also exist since increased HO gene transcription due to heme requires protein synthesis or some other cycloheximide-sen- sitive processing for activation of trans-acting factors, while the MT (10) and hemopexin’ (44) genes rapidly respond to heme-hemopexin by a mechanism independent of protein synthesis and are “superinduced by heme when protein syn- thesis is inhibited. In addition, recent work’ has shown that histone H3 and H4 mRNA levels are regulated by heme and heme-hemopexin, providing evidence for additional regula- tory consequences to cells of heme and hemopexin as well as suggesting a role for heme in the regulation of cell growth. In work to be reported elsewhere: transient transfection assays have established that a t least 600 base pairs 5’ to the TATA box of the mouse MT-1 promoter are required for activation of gene transcription in response to heme-hemopexin, and

P. V. Escriba and A. Smith, unpublished observations. Y. Ren and A. Smith, manuscript in preparation.

* N. Hooper, J. Alam, and A. Smith, manuscript in preparation. Y. Ren, R. Palmiter, and A. Smith, manuscript in preparation.

Heme Analog-mediated Gene Regulation 7371

0.02

m 0 c 0

M 2 0.01

0.00

A I I

B

350 400 450 500 550 600 650

Wavelength (nm) Wavelength (nrn)

FIG. 7 . Uptake of free and hemopexin-bound, CoPP, and SnPP by mouse hepatoma cells. A , cells were incubated with either free CoPP or CoPP-hemopexin at 10 PM in HEPES-buffered DMEM for 1 or 2 h at 37 “C as described in detail under “Experimental Procedures.” After extraction, the amount of CoPP was determined by comparing the absorbance spectrum of the cell extracts shown here with a standard curve for CoPP over the concentration range 0.05-0.35 PM. Shown are: free CoPP 1 h (fine dotted line) and 2 h (medium dotted line); CoPP- hemopexin 1 h (dashed line) and 2 h (solid line). B , cells were incubated with either free SnPP or SnPP-hemopexin as described above for CoPP in A . After extraction, the amount of SnPP taken up was determined by comparing the fluorescence spectra of cell extracts with a SnPP standard. Shown are: free SnPP 1 h (fine dotted line) and 2 h (medium dotted line); SnPP-hemopexin 1 h (dashed line) and 2 h (solid l i n e ) .

there is no role for the metal-responsive elements. In these assays, MT gene expression is more sensitive on a molar basis to heme-hemopexin than to zinc, which is without effect in serum-free medium at 10 and 50-80 WM is required for equiv- alent levels of induction. Current efforts are directed to fur- ther delineate the mechanisms whereby heme and heme- hemopexin exert their regulatory effects on mammalian gene expression.

Acknowledgements-We gratefully acknowledge the help of Dr. S. Farooqui in carrying out the binding experiments with CoPP-hemo- pexin, of M.-L. Wu in the circular dichroism studies and of V. Deaciuc and J. Finlay for their technical contributions.

REFERENCES 1. Shibahara, S., Yoshida, T., and Kikuchi, G. (1978) Arch. Biochem. Eiophys.

2. Shibahara, S., Yoshida, T., and Kikuchi, G. (1979) Arch. Biochem. Eiophys.

3. Tenhunen, R., Marver, H. S., and Schmid, R. (1970) J. Lab. Clin. Med. 75,

4. Pimstone, N. R., Engel, P:, Tenhunen, R., Seitz, P. T., Marver, H. S., and

5. Pimstone, N. R., Tenhunen, R., Seitz, P. T., Marver, H. S., and Schmid,

6. Ishizawa, S., Yoshida, T., and Klkuchi, G. (1983) J . BioE. Chem. 258,4220-

8. Smith, A. (1990) in Biosynthesis of Heme and Chlorophylls (Dailey, H. A,, 7. Smith, A., and Hunt, R. C. (1990) Eur, J. Cell Biol. 63,234-245

9. Alam, J., and Smith, A. (1989) J. Biol. Chem. 264, 17637-17640

11. Becker, R. C. (1991) Cardlology 78, 257-266 10. Alam, J., and Smith, A. (1992) J. Biol. Chem. 267, 16379-16384

12. Arizono, K., Okanari, E., Ueno, K., and Ariyoshi, T. (1991) J. Enuiron. Sei.

13. Maines, M. D., and Kappas, A. (1976) Bwchem. J. 1 6 4 , 125-131 14. Taketani, S., Kohno, H., Yoshinaga, T., and Tokunaga, R. (1988) Bkhern.

15. Caltahiano, M. M., Koestler, T. P., Poste, G., and Greig, R. G. (1986) J.

16. Muller, R. M., Taguchi, H., and Shibahara, S. (1987) J. Bid. Chern. 262,

188,243-250

197,607-617

410-421

Schmid, R. (1971) J. Clm. Inuest. 60,2042-2050

R. (1971) J. Exp. Med. 133, 1264-1281

4225

Jr., ed) pp. 435-489, McGraw-Hill Inc., New York

Health A 26A, 941-951

Int. 17,665-672

Biol. Chem. 261, 13381-13386

6795-6802

17. Stuart, G. W., Searle, P. F., and Palmiter, R. D. (1985) Nature 317, 828-

18. Culotta, V. C., and Hamer, D. H. (1989) Mol. CeU. Biol. 9, 1376-1380 19. Durnam, D. M., and Palmiter, R. D. (1981) J. B~ol. Chem. 266,5712-5716 20. Morgan, W. T., Alam, J., Deaciuc, V., Muster, P., Tatum, F., and Smith,

21. Drummond, G. S., and Kappas, A. (1981) Proc. Natl. Acad. Sci. U. S. A.

22. Drummond, G. S., and Kappas, A. (1981) Proc. Natl. Acad. Sci. U. S. A.

23. Sardana, M. K., and Kappas, A. (1987) Proc. Natl. Acad. Sci. U. S. A. 84,

24. Jones, M. S., and Jones, 0. T. G. (1969) Biochem. J. 113,507-514 25. Lin, J. H.-C., Villalon, P., Martasek, P., and Abraham, N. G. (1990) Eur.

26. Smith, A., and Morgan, W. T. (1979) Biochem. J. 182,47-54 27. Smith, A,, and Morgan, W. T. (1984) J. Biol. Chern. 259, 12049-12053 28. Smith, A. (1985) Biochem. J. 231,663-669 29. Yonetani, T., Yamamoto, H., and Woodrow, G. V., I11 (1974) J. Biol. Chem.

30. Anderson, K. M. (1984) Pharmacol. Exp. Ther. 228,327-333 31. Alam, J., Shibahara, S., and Smith, A. (1989) J. Bzol. Chem. 264, 6371-

32. Greenberg, M. E., and Ziff, E. B. (1984) Nature 311,433-438 33. Kingston, R. E. (1987) in Current Protocols in Mokculor Biology (Ausubel,

F. M., Brent, R., Kingston, R. E., Moore, D. D., Seidman, J. G., Smith, J. A., and Struhl, K., e&) pp. 8.1, John Wiley & Sons, Inc., New York

34. Glanville, N., Durnam, D. M., and Palmiter, R. D. (1981) Nature 292, 267-269

83 1

A. (1988) J. Biol. Chem. 263, 8226-8231

78,6466-6470

78,6466-6470

2464-2468

J. Biochem. 192,577-582

249,682-690

6375

35. Kageyama, H., Hiwasa, T., Tokunaga, K., and Sakiyama, S. (1988) Cancer Res. 48,4795-4798

36. Marzluff, W. F., and Huang, R. C. C. (1984) in Tramcription and Tram- lotion: A Practical A roach (Hames, B. D., and Higgins, s. J., eds) pp.

37. Groudine, M., Peretz, M., and Weintraub, H. (1981) Mol. Cell. Biol. 1 , 281- 89-129, IRL Press, E s h . , D.C.

38. Smith, A., and Ledford, B. E. (1988) Biochem. J. 256,941-950 288

39. Smith, A., Farooqui, S. F., and Morgan, W. T. (1991) Biochem. J. 276, 41 7-42.5

40. Morgan, W. T., Muster, P., Tatum, F., Kao, S. M., Alam, J., and Smith, A.

41. Smith, A., Tatum, F. M., Muster, P., Burch, M. K., and Morgan, W. T. (1993) J. Biol. Chem. 268,6256-6262

42. Morgan, W. T., Muster, P., Tatum, F. M., McConnell, J., Conway T. P., (1988) J . Biol. Chem. 263,5224-5229

43. Morgan, &. T. (1975) Ann. N . Y. Acad. Sei. 244,624-650 Hensle , P I and Smith, A. (1988) J. Biol. Chern. 263,8220-8225‘

44. Ren, Y., and Smith, A. (1992) FASEB J. 6 , A71 (Abstr. 407) 45. “A Century of Porphyria” Symposium, Glasgow, Scotland, 1989 (1991)

” . ”.

FASEB J. 6, 71 (ahstr.)