Identification of a cis-Regulatory Element for Δ12-Prostaglandin J2-induced Expression of the Rat Heme Oxygenase Gene

Abstract We recently reported that Δ12-prostaglandin (PG) J2 caused various cells to synthesize heme oxygenase, HO-1 (Koizumi, T., Negishi, M., and Ichikawa, A.(1992) Prostaglandins 43, 121-131). Here we examined the molecular mechanism underlying the Δ12-PGJ2-induced HO-1 synthesis. Δ12-PGJ2 markedly stimulated the promoter activity of the 5′-flanking region of the rat HO-1 gene from −810 to +101 in rat basophilic leukemia cells. From functional analysis of various deletion mutant genes we found that the Δ12-PGJ2-responsive element was localized in a region from −690 to −660, containing an E-box motif, which was essential for the Δ12-PGJ2-stimulated promoter activity. When the region containing the Δ12-PGJ2-responsive element was combined with a heterologous promoter, SV40 promoter, in the sense and antisense direction, the element showed an enhancer activity in response to Δ12-PGJ2. Gel mobility shift assays demonstrated that Δ12-PGJ2 specifically stimulated the binding of two nuclear proteins to the E-box motif of this region. These results indicate that Δ12-PGJ2 induces the expression of the rat HO-1 gene through nuclear protein binding to a specific element having an E-box motif.

Eicosanoids are oxygenated metabolites of arachidonic acid, and are regarded as modulators of cellular functions in various physiological and pathological processes (1). Eicosanoids are divided into two groups, conventional eicosanoids and cyclopentenone-type prostaglandins (PGs) 1 according to their mechanisms of action. Conventional eicosanoids, such as PGE 2 and PGD 2 , act on a cell surface receptor to exert their actions, and the molecular structures of their receptors have been revealed recently (2). Cyclopentenone PGs, such as ⌬ 12 -PGJ 2 and PGA 2 , have no cell surface receptor, but are actively transported into cells and accumulated in nuclei, where they act as potent inducers of cell growth inhibition and cell differentiation (3). The actions of cyclopentenone PGs are attributed to the synthesis of the various proteins induced by them, such as heat shock proteins (HSPs) (4,5), ␥-glutamylcysteine synthetase (6), collagen (7), gadd 153 (8), and heme oxygenase (9). In contrast to conventional eicosanoids, the molecular characterization of cyclopentenone PG actions has been hardly carried out.
Heme oxygenase is one of the most prominent proteins induced by ⌬ 12 -PGJ 2 , and it is a key enzyme in heme catabolism, oxidatively clearing heme to yield biliverdin, iron, and carbon monoxide (10). The biological functions of this enzyme are the production of biliverdin as a physiological antioxidant and the conservation of the iron (11). Furthermore, carbon monoxide produced on the enzymatic degradation of heme has been suggested to function as a neural messenger (12). Two isozymes of heme oxygenase, HO-1 and HO-2, have been identified (13). HO-2 is constitutively expressed, while HO-1 is drastically induced in response to a variety of stresses, including heavy metals, heat shock, and UV irradiation (14). We previously found that ⌬ 12 -PGJ 2 preferentially induced the synthesis of HO-1 in various cells involved in the reticuloendothelial system, in which active degradation of heme by HO-1 takes place during inflammation (9,15). In order to elucidate the mechanism underlying ⌬ 12 -PGJ 2 -induced protein synthesis, we examined the effect of ⌬ 12 -PGJ 2 on the promoter activity of the HO-1 gene. We report here that ⌬ 12 -PGJ 2 induces the expression of the rat HO-1 gene through nuclear protein binding to a specific ⌬ 12 -PGJ 2 -responsive element, located 660 base pairs upstream from the transcription initiation site.
Rat basophilic leukemia (RBL)-2H3 cells were obtained from the Japanese Cancer Research Resources Bank (Tokyo, Japan). The cells were cultured in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum, 4 mM glutamine, 0.2 mg/ml streptomycin, and 100 units/ml penicillin under humidified air containing 5% CO 2 at 37°C.
Northern Blots-Total RNA from RBL-2H3 cells was isolated by the acid guanidinium thiocyanate-phenol-chloroform method (16), and 5 g of each RNA was separated by electrophoresis on a 1.2% agarose gel, transferred to a nylon membrane (Hybond-N, Amersham Corp.), and hybridized with a 32 P-labeled fragment, corresponding to exon 3 of the rat HO-1 gene. The same filter was rehybridized with 32 P-labeled glyceraldehyde-3-phosphate dehydrogenase cDNA probe (Clontech). Hybridization was carried out at 68°C in 6 ϫ SSC, and the filter was washed at 68°C in 2 ϫ SSC. The filter was autoradiographed with x-ray film (Fuji RX). The radioactivity was determined with a Fuji BAS 2000 imaging analyzer (Fuji Film Co., Tokyo).
Construction of Plasmid DNA-The 5Ј-flanking region of the rat HO-1 gene (nucleotide residues Ϫ810 to ϩ101, relative to the transcription start site) was obtained by means of the polymerase chain reaction (PCR) from genomic DNA prepared from RBL-2H3 cells, as described previously (17). The PCR primers with restriction enzyme sites (BamHI for the forward primer and SalI for the reverse one) were designed according to the published sequence of the rat HO-1 gene (18). The BamHI/SalI fragment of the PCR product was first constructed into pBluescript II (Stratagene), and then the SacI/KpnI insert was constructed into pUC00CAT to generate plasmid RHO810, containing the 5Ј-flanking region of the HO-1 gene, upstream of the bacterial chloramphenicol acetyltransferase (CAT) reporter gene. The sequence of the cloned 5Ј-flanking region was confirmed by sequence analysis using the dideoxynucleotide chain termination method.
Successive deletion plasmids as to the 5Ј-flanking region, RHO600 (Ϫ600 to ϩ101), RHO320 (Ϫ320 to ϩ101), RHO270 (Ϫ270 to ϩ101), RHO-0 (ϩ1 to ϩ101), and RHO810 -600 (Ϫ810 to Ϫ600), were also generated by PCR using the respective forward primers with a BamHI site. To prepare a deletion plasmid without region (Ϫ600 to Ϫ270), two DNA fragments (Ϫ810 to Ϫ600, and Ϫ270 to ϩ101) were prepared by PCR and tandemly constructed into pUC00CAT. For the functional analysis using a heterologous promoter, the 5Ј-flanking region from Ϫ810 to Ϫ600 was inserted into pCAT promoter vector (Promega), which carries the SV40 promoter upstream from the CAT gene, and the DNA fragment from Ϫ810 to Ϫ600 was located upstream from the SV40 promoter, in the sense (pCAT-HOs) and antisense direction (pCAT-HOa).
Transfection and CAT Assay-RBL-2H3 cells were transiently transfected with the plasmid DNA constructs by the DEAE-dextran method (19). After the cells (4 ϫ 10 6 cells/assay) had been incubated for 30 min at 37°C in 0.5 ml of serum-free Dulbecco's modified Eagle's medium containing 8 g of plasmid DNA and 20 g of DEAE-dextran (Pharmacia Biotech Inc.), they were incubated for 10 min at 37°C with a dimethyl sulfoxide-hypertonic solution (28 mM Tris-HCl, pH 7.5, containing 0.4 M sucrose, 8% polyethylene glycol 4000, 84 mM NaCl, and 10% dimethyl sulfoxide). They were then cultured for 2 days in Dulbecco's modified Eagle's medium containing 10% fetal bovine serum.
Reactions were started by the addition of the test agents. After incubation for the indicated times, cellular extracts were prepared by four cycles of freezing and thawing. The CAT assay was performed as described previously (20). Cellular extracts containing equal amounts of protein were treated at 65°C for 10 min to inactivate deacetylase, and then incubated for 4 h at 37°C with [ 14 C]1-deoxychloramphenicol (50 nCi) and 0.1 mg/ml acetyl-CoA. The reaction mixtures were extracted with ethyl acetate, and then separated on a silica TLC plate (F1500, Schleicher & Schnell). After development of the TLC plate, the radioactivity was quantitated with a Fuji BAS 2000 imaging analyzer (Fuji Film, Co., Tokyo). CAT activity was normalized as to ␤-galactosidase activity due to co-transfected Rous sarcoma virus-␤-galactosidase, which was assayed as described previously (20).
Gel Mobility Shift Assay-A complementary pair of DNA fragments encoding region, Ϫ690 to Ϫ660, was synthesized with an ABI 391 DNA synthesizer (Applied Biosystems, Inc., CA), and then annealed as a probe. The fragment was radiolabeled at the 5Ј ends with [␥-32 P]ATP using T 4 polynucleotide kinase.
Nuclear extracts were prepared by detergent lysis method of Block et al. (21). The nuclear extracts (2 g) were incubated with 2 g of poly(dI-dC) and 1 ng of a 32 P-labeled probe (10,000 cpm) for 30 min at 30°C in 25 mM Hepes-NaOH (pH 7.9), containing 0.5 mM EDTA, 50 mM KCl, 10% glycerol, 0.5 mM dithiothreitol, and 0.5 mM phenylmethylsulfonyl fluoride. The reaction mixtures were electrophoresed on native 4% polyacrylamide gels at 4°C at 150 V for 2 h in 50 mM Tris-HCl (pH 8.5), containing 380 mM glycine and 2 mM EDTA. The gels were dried on Whatman 3MM paper and then autoradiographed with x-ray film (Fuji RX).
Point Mutation of the E-box Motif-A complementary pair of DNA fragments (Ϫ690 to Ϫ660) point-mutated at the E-box motif were synthesized with an ABI 391 DNA synthesizer. The sequences were as follows: 690F, 5Ј-ATAGATGTGTCAGAGCCCCGGGTCCTGAC-3Ј; 660R, 5Ј-AAGTCAGGACCCGGGGCTCTGACACATCT-3Ј. The mutated sites are underlined. They were annealed as the competitor for the gel mobility shift assay. A mutated plasmid as to the E-box motif was constructed by Kunkel's method (22), using RHO810 as a template and 690F as a mutated primer. The mutation in the plasmid was confirmed by sequence analysis.
UV Cross-linking-The 5Ј-end-labeled fragment from Ϫ690 to Ϫ660 was incubated with the nuclear extract under standard gel mobility shift assay condition. Following electrophoresis, the wet gel was irradiated for 20 min using a UV transilluminator (254 nm; 1780 microwatts/cm 2 ) at a distance of 12 cm of the UV source. Radioactive bands corresponding to protein-DNA complexes, I and II, were separately eluted by electroelution using Electro-elutor (Bio-Rad Model 422), and the eluate was subjected to SDS-12.5% PAGE, followed by autoradiography.

Induction of HO-1 mRNA and Stimulation of Promoter Activity of the 5Ј-Flanking Region of the HO-1 Gene by ⌬ 12 -PGJ 2 -
We examined the effect of ⌬ 12 -PGJ 2 on the expression of the HO-1 gene in RBL-2H3 cells by Northern blot analysis. As shown in Fig. 1A, after a 0.5-h lag time, ⌬ 12 -PGJ 2 gradually increased the mRNA level of HO-1. This stimulation of expression of HO-1 mRNA preceded the ⌬ 12 -PGJ 2 -induced HO-1 protein synthesis and HO-1 activity (9). ⌬ 12 -PGJ 2 concentration dependence caused marked accumulation of the mRNA, the maximum being reached at 10 M (Fig. 1B), and this concentration dependence was consistent with those of ⌬ 12 -PGJ 2induced HO-1 protein synthesis and HO-1 activity (9). On the other hand, the mRNA level of glyceraldehyde-3-phosphate dehydrogenase did not change.
To identify the cis-regulatory element for ⌬ 12 -PGJ 2 -induced expression of the HO-1 gene, we isolated the 5Ј-flanking region of the HO-1 gene by PCR amplification, which included the 810 base pairs upstream of the transcription initiation site, containing two known cis-acting elements, the heat shock element (HSE) and the metal-responsive element (MRE). To determine whether or not ⌬ 12 -PGJ 2 stimulates the promoter activity of the 5Ј-flanking region, we examined the effect of ⌬ 12 -PGJ 2 on the transient expression of the bacterial CAT reporter gene harboring the 5Ј-flanking region (Ϫ810 to ϩ101) in RBL-2H3 cells. As shown in Fig. 2, RHO810 containing the 5Ј-flanking region gave low but detectable CAT activity, and ⌬ 12 -PGJ 2 markedly stimulated it by about 5-fold, indicating that the region contains an element responsible for ⌬ 12 -PGJ 2 . Fig. 3 shows the time course and concentration dependence of the effect of ⌬ 12 -PGJ 2 on the promoter activity of RHO810. ⌬ 12 -PGJ 2 stimulated the promoter activity in a time-dependent manner, the time course being consistent with those of the ⌬ 12 -PGJ 2 -induced HO-1 protein synthesis and HO-1 activity (9). ⌬ 12 -PGJ 2 concentration-dependently stimulated it, the maximum being reached at 10 M (Fig. 3B). This concentration dependence was consistent with that of the mRNA induction by ⌬ 12 -PGJ 2 (Fig. 1B). We next examined the specificity of stim- ulation of the promoter activity for various PGs. As shown in Fig. 4A, ⌬ 12 -PGJ 2 and ⌬ 7 -PGA 1 markedly stimulated the activity, but the other PGs did not significantly stimulate it. HO-1 synthesis is known to be induced by other stimuli, such as heat shock, thiol-reactive agents, arsenite and diethylmaleate, and hemin, which is the substrate of heme oxygenase and a known strong inducer of HO-1 (23)(24)(25). Thus, we examined the effects of these agents on the promoter activity of RHO810. As shown in Fig. 4B, heat shock strongly stimulated the promoter activity, while hemin did not stimulate it at all. Arsenite and diethylmaleate only slightly stimulated it.
Identification of the ⌬ 12 -PGJ 2 -responsive Element-Heat shock and heavy metals have been shown to induce HO-1 mRNA through the respective elements, HSE and MRE, located in the 5Ј-flanking region (18). Previously, ⌬ 12 -PGJ 2 was shown to activate a heat shock factor (HSF), which binds to HSE upstream of the HSP gene promoter, and to induce the transcription of HSP (26,27). To identify the element responsible for the ⌬ 12 -PGJ 2 -stimulated promoter activity, we constructed a series of deletion plasmids as to the 5Ј-flanking region fused to the CAT gene, and then examined the effects of ⌬ 12 -PGJ 2 on the promoter activity of the mutant genes. As shown in Fig. 5, RHO600, containing MRE and HSE, or RHO320, containing only HSE, had completely lost the ⌬ 12 -PGJ 2 responsiveness, indicating that HSE or MRE is not the ⌬ 12 -PGJ 2 -responsive element. RHO270, without either HSE or MRE, showed basal promoter activity without ⌬ 12 -PGJ 2 responsiveness, but RHO-0, without the 5Ј-flanking region, lost basal promoter activity, indicating that the region from Ϫ270 to 0 is required for the basal promoter activity of HO-1 gene. Whereas the region (Ϫ810 to Ϫ600) itself did not show the basal promoter activity, this region (Ϫ810 to Ϫ600) combined with the region (Ϫ270 to ϩ101) regained the ⌬ 12 -PGJ 2 -stimulated promoter activity, suggesting that this region contains an enhancer-like element. These results indicate that ⌬ 12 -PGJ 2 stimulates the expression of the HO-1 gene through neither HSE nor MRE, but through an element located between Ϫ810 and Ϫ600. On the other hand, heat shock stimulated the promoter activity of either RHO600 or RHO320, but failed to stimulate the activity of the HSE-deleted mutant gene, RHO270 (data not shown).
Enhancer Function of the ⌬ 12 -PGJ 2 -responsive Element-To examine whether the region from Ϫ810 to Ϫ600 has an enhancer function, we constructed another fusion gene with a heterologous promoter, SV40 promoter (Fig. 6). ⌬ 12 -PGJ 2 stimulated the CAT activity by about 6-fold in the cells transfected with pCAT-HOs or pCAT-HOa containing the region from Ϫ810 to Ϫ600 in the sense or antisense orientation upstream from the SV40 promoter. This establishes that the ⌬ 12 -PGJ 2responsive element has an enhancer function.
Point Mutation of the E-box Motif in the ⌬ 12 -PGJ 2 -responsive Element-To identify more precisely the region responsible for the stimulation by ⌬ 12 -PGJ 2 and to detect nuclear protein binding to the region, we performed a gel mobility shift assay using various sized DNA fragments, from Ϫ810 to Ϫ600. We could not find potential binding sites for so far known transcription factors in the region from Ϫ810 to Ϫ600, but this region contained the consensus E-box motif, CANNTG (Ϫ673 to Ϫ668), for a large family of putative transcription factors, containing a basic helix-loop-helix domain (28). Fig. 7A shows the result of protein-DNA complex formation using a DNA fragment of the region from Ϫ690 to Ϫ660, containing the E-box motif. ⌬ 12 -PGJ 2 markedly induced two nuclear protein-DNA complexes with different mobilities (I and II). These complexes were displaced by double-stranded DNA fragments of Ϫ690 to Ϫ660, but not by either the single-stranded fragments or the E-box motif-mutated fragment, indicating that the nuclear pro- teins specifically recognize the E-box motif in this region. DNA fragments other than the region (Ϫ690 to Ϫ660) did not form complexes with nuclear proteins in response to ⌬ 12 -PGJ 2 (data not shown). To identify the nuclear proteins, which bind to the region from Ϫ690 to Ϫ660, the complexes of nuclear proteins and the Ϫ690 to Ϫ660 DNA fragment were photolabeled by UV irradiation. As shown in Fig. 7B, subsequent SDS-12.5% PAGE analysis of the photolabeled complexes showed radioactive bands with different molecular weight, 80,000 and 24,000, as DNA complex form, respectively.
We further examined the effect of ⌬ 12 -PGJ 2 on the promoter activity of RHO810 point-mutated at the E-box motif. As shown in Fig. 7C, this point mutation completely abolished the ⌬ 12 -PGJ 2 -induced stimulation of the promoter activity of RHO810. These results indicate that the region (Ϫ690 to Ϫ660) contains the ⌬ 12 -PGJ 2 -responsive element and that the E-box motif in this element is essential for this stimulation of the promoter activity. DISCUSSION We demonstrated here that ⌬ 12 -PGJ 2 drastically induced HO-1 mRNA through a ⌬ 12 -PGJ 2 -specific cis-regulatory element, containing a E-box motif, in RBL-2H3 cells. ⌬ 12 -PGJ 2 induces the syntheses of a variety of proteins (3). Among the ⌬ 12 -PGJ 2 -induced protein syntheses, the mechanism for the synthesis of HSP has been well characterized. A wide range of external stress stimuli, including heat shock, heavy metals, amino acid analogues, and oxidizing agents, drastically induce the expression of the HSP gene through activation of HSF, which binds to HSE located in the 5Ј-flanking sequence of the HSP gene (29). ⌬ 12 -PGJ 2 induces the expression of the HSP gene, and this induction is also mediated by HSF activation, as well as the above mentioned stimuli (26,27). We here showed that ⌬ 12 -PGJ 2 stimulated the promoter activity of the 5Ј-flanking region of the HO-1 gene and then induced HO-1 mRNA in RBL-2H3 cells (Figs. 1 and 2). Although the 5Ј-flanking region contains HSE, HSE is not necessary for the ⌬ 12 -PGJ 2 -stimulated promoter activity (Fig. 5), but this stimulation requires the specific region (Ϫ690 to Ϫ660) containing the E-box motif (Fig. 7), indicating that ⌬ 12 -PGJ 2 induces the gene expression through a specific element other than HSE. On the other hand, heat shock stimulated the promoter activity of RHO810, but this stimulation was completely abolished on removal of HSE, indicating that heat shock induces this gene expression through HSE. Furthermore, other stimuli, such as hemin and arsenite, did not stimulate the promoter activity of RHO810 or showed only very low stimulation (Fig. 4B). These findings demonstrate that this element specifically responds to ⌬ 12 -PGJ 2 activation. Thus, the element could be referred to as the ⌬ 12 -PGJ 2 -responsive element, and this is the first example of a cis-regulatory element showing a ⌬ 12 -PGJ 2 -specific response. Among various PGs, the stimulation of the promoter activity of HO-1 is specific for ⌬ 12 -PGJ 2 and ⌬ 7 -PGA 1 (Fig. 4A). The stimulation was not observed with A type cyclopentenone PGs as well as conventional PGs. Generally, the biological actions of PGA 1 and PGA 2 are much weaker than those of ⌬ 12 -PGJ 2 and ⌬ 7 -PGA 1 (3). A characteristic of cyclopentenone PGs, such as ⌬ 12 -PGJ 2 and PGA 2 , is that they contain ␣,␤-unsaturated ketones, which are very susceptible to nucleophilic addition reactions with thiols, and are essential for the actions of the PGs (30,31). PGA 1 or PGA 2 forms a monoconjugate with a thiol (32,33), but ⌬ 12 -PGJ 2 or ⌬ 7 -PGA 1 can form a bisconjugate with two thiols (30). Furthermore, the binding of PGA 1 to synthetic polymer-supported thiols as the model of thiol-containing proteins is reversible, but that of ⌬ 12 -PGJ 2 or ⌬ 7 -PGA 1 is irreversible (34). An irreversible bisconjugate with thiol groups of proteins may be required for the stimulation of the HO-1 promoter activity, or the stimulation by PGA 1 or PGA 2 might be marginal in the detection of the stimulation of the activity of CAT.
The 5Ј-flanking region of the HO-1 gene contains a number of DNA sequences of potential regulatory elements. The 5Ј-flanking region of the rat HO-1 gene up to position Ϫ600 contains several potential binding sites for different transcription factors: a transcription factor, Sp1, a positive regulator for the control of amino acid synthesis (GCN4), a heat shock transcription factor, and a metal-dependent transcription factor (18). The proximal promoter region within 149 base pairs of the upstream sequence of the mouse HO-1 gene contains several sequence elements for AP-1, AP-4, C/EBP, and c-Myc:Max/ FIG. 5. Functional analysis of various deletion mutant genes of RHO810. After cells (4 ϫ 10 6 cells) had been transiently transfected with RHO810, RHO600, RHO320, RHO270, RHO-0, RHO810 -600, or RHO810 without region (Ϫ600 to Ϫ270), (dRHO810), they were treated with (f) or without (Ⅺ) 10 M ⌬ 12 -PGJ 2 for 9 h. The CAT assay was performed for cellular extracts, as described under "Experimental Procedures." The values are expressed as fold of the untreated cell value of RHO810, and are the means for three independent experiments, which varied by less than 10%. The value for the acetylated chloramphenicol of the untreated cells was 0.295 Ϯ 0.043%.
FIG. 6. Functional analysis of the ⌬ 12 -PGJ 2 -responsive element using SV40 promoter. After cells (4 ϫ 10 6 cells) had been transiently transfected with pCAT, pCAT-HOs, or pCAT-HOa, they were treated with (f) or without (Ⅺ) 10 M ⌬ 12 -PGJ 2 for 9 h. The CAT assay was performed for cellular extracts, as described under "Experimental Procedures." The values are expressed as fold of the untreated cell value of pCAT, and are the means for three independent experiments, which varied by less than 10%. The value for the acetylated chloramphenicol of the untreated cells was 0.223 Ϯ 0.038%. USF, and is required for basal promoter activity (35). Several NFB and AP-2-like binding sites have been found in the 5Ј-flanking region of the human HO-1 gene up to position Ϫ500 (36). However, the more upstream region of rat HO-1 (Ϫ810 to Ϫ600), containing the ⌬ 12 -PGJ 2 -responsive element, has not yet been reported to contain potential binding sites for so far known transcription factors. Thus, the ⌬ 12 -PGJ 2 -responsive element is a novel cis-regulatory element and plays an important role in HO-1 gene expression. Recently, the heme oxygenase transcription factor, an essential transcription factor, was shown to interact with the cis-acting element (Ϫ51 to Ϫ35), located just upstream of the TATA box of the rat HO-1 gene, and this binding is essential for the basal expression of the HO-1 gene in rat glioma cells (37). In RBL-2H3 cells, the proximal 5Ј-flanking region (Ϫ270 to ϩ101) containing the heme oxygenase transcription factor binding element and TATA box showed the basal promoter activity. The proximal promoter region, containing the heme oxygenase transcription factor binding element and TATA box, is essential for the basal promoter activity. Whereas the region (Ϫ810 to Ϫ600) containing the ⌬ 12 -PGJ 2 -responsive element itself did not show the basal promoter activity, this region (Ϫ810 to Ϫ600), when combined with the proximal 5Ј-flanking region (Ϫ270 to ϩ101), enhanced the promoter activity of the region (Ϫ270 to ϩ101) in a ⌬ 12 -PGJ 2 -dependent manner (Fig. 5). Furthermore, this region (Ϫ810 to Ϫ600) enhanced activity of the exogenous promoter, SV40 promoter (Fig. 6). The ⌬ 12 -PGJ 2 -responsive element thus appears to act as an enhancer.
HO-1 mRNA was increased more than 30-fold by ⌬ 12 -PGJ 2 (Fig. 1), although the magnitude of stimulation of CAT activity was 5-fold (Fig. 3). Such a difference may simply represent that the fusion genes lack the additional element required for the maximal stimulation by ⌬ 12 -PGJ 2 . Alternatively, this may be due to a limitation of transient expression assays; namely, integration of the fusion genes into the genomic DNA is required for the maximal stimulation.
Using UV cross-linking, we identified two nuclear proteins, the ⌬ 12 -PGJ 2 -responsive factors, which specifically bind to the ⌬ 12 -PGJ 2 -responsive element, their apparent molecular weight being 80,000 and 24,000 as DNA complex form, indicating that two different proteins bind to the element in response to ⌬ 12 -PGJ 2 (Fig. 7B). The ⌬ 12 -PGJ 2 -responsive element contains an E-box motif, and this motif is essential for the ⌬ 12 -PGJ 2 -responsive factor binding to this element and for the ⌬ 12 -PGJ 2stimulated promoter activity (Fig. 7, A and C). The transcription factors which bind to the E-box motif contain a helix-loophelix domain and an adjacent basic amino acid region, and a variety of transcription factors, such as MyoD, c-myc, and Myf-5, belong to this family (38). In RBL-2H3 cells, the ⌬ 12 -PGJ 2 -responsive factors appear to belong to a family of basic helix-loop-helix type transcription factors.
Mast cells synthesize and subsequently release PGD 2 , which is converted into ⌬ 12 -PGJ 2 by serum albumin during the process of inflammation (39). The ⌬ 12 -PGJ 2 produced stimulates the transcription of the HO-1 gene in various cells, such as basophils, fibroblasts, and vascular endothelial cells, which are actively involved in the inflammatory response, and the reticuloendothelial system is the site at which the active degradation of heme by HO-1 may take place during inflammation. The induction of HO-1 by ⌬ 12 -PGJ 2 may play an important role in the fate of heme liberated during inflammation.
In summary, we here identified the ⌬ 12 -PGJ 2 -specific cisregulatory element responsible for the expression of the rat HO-1 gene. This study will contribute not only to understanding of the HO-1 gene expression mechanism but will also facilitate elucidation of the molecular mechanisms of cyclopentenone PG actions.