INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

Cytochrome Cyp1B1 gene expression in rodents and humans is inducible both through activation of the AhR¹ by TCDD and by a cAMP-mediated pathway (1-7). This work has also shown that CYP1B1 is one of the largest known P450 cytochromes (543 amino acids) that is translated from an exceptionally large 5.2-kb mRNA, most of which is contributed by a 3.1-kb 3'-untranslated region. Nevertheless, the structure of the human CYP1B1 gene is uniquely compact compared with other mammalian cytochrome P450 genes as evidenced by the presence of only three exons, the first of which is untranslated (8).

In addition to these unusual sequence characteristics, Cyp1B1 exhibits an exceptional pattern of tissue expression. Constitutive CYP1B1 and chemically induced transcription linked to the AhR are seen in stromal fibroblasts isolated from rodent embryos (1, 9, 10) and from endocrine-regulated tissues such as mammary gland (11), uterus, and prostate, where Cyp1A1 is very poorly expressed (1, 4, 7). Cyp1B1 is also expressed constitutively in rodent steroidogenic tissues and is stimulated by hormones that typically elevate cAMP levels in these cells (4, 6). For many rodent tissues (lung, liver, and kidney), there is essentially no constitutive CYP1B1, and elevated expression is only seen after administration of AhR agonists. However, this induction through the AhR is 30-40 times less effective than for Cyp1A1 (1, 4). For rat mammary cells, cell culture conditions greatly affect Cyp1B1 expression. In culture, Cyp1B1 is preferentially expressed in stromal fibroblasts and partly in epithelia, whereas Cyp1A1 is seen predominantly in epithelia (11); however, Cyp1B1 is expressed constitutively in ductal epithelia in vivo.² For human cells, CYP1B1 is again expressed in fibroblasts in preference to CYP1A1 and is coexpressed with CYP1A1 in selected epithelial cell types. Skin and mammary cells coexpress CYP1A1 and CYP1B1 following TCDD induction, but only Cyp1B1 is seen constitutively in these cells (11, 12).

The AhR mediates gene transcription by forming a nuclear heterodimer with the related helix-loop-helix protein AhR nuclear translocator (Arnt) (13-15). The formation of this complex is stimulated by binding of agonists such as TCDD and polycyclic aromatic hydrocarbons to the AhR, which then dissociates from the associated cytosolic partners such as HSP90 (16). The mechanism by which the AhR·Arnt complex stimulates transcription has been extensively dissected through analysis of the Cyp1A1 5'-flanking region (17). The AhR·Arnt heterodimer preferentially selects the core T_nGCGTG sequence from random combinations of oligonucleotide heptamers, but less strictly requires the starting T in the sequence context provided by certain genes (18). Most AhR-responsive genes contain multiple copies of such xenobiotic-responsive elements (XREs), and typical of enhancer elements, the response is increased by repetition of the elements and is independent of the orientation (19-21). In vivo DNA footprinting of the Cyp1A1 upstream region suggests that activation of an upstream XRE may open up the downstream region for binding by other nuclear regulatory complexes (13, 22). The enhancer activity of XREs in the Cyp1A1 promoter is increased by the presence of adjacent GC-rich elements (23). Thus, an additional set of proteins may work in cooperation with the AhR·Arnt complex to transfer a signal to the transcriptional machinery (13, 17). The AhR protein sequence has also been dissected by mutagenesis in transfection assays to establish the presence of trans-acting domains in the C-terminal part of the protein (24). The participation of labile repressor proteins has been indicated by cycloheximide treatment, which produces a stimulatory effect on Cyp1A1 transcription. This effect is attributed to removal of labile repressor proteins that bind close to certain XREs (25). Distinct elements also provide additional negative trans-acting effects that may be specific to the cell type or growth state of the cells (26, 27).

The results presented here show an extensive analysis of the mouse Cyp1B1 gene, including its 5'-regulatory region and transcription start site in embryo fibroblasts. This work establishes that several of the features of the Cyp1A1 5'-flanking region are repeated in the upstream region of the mouse Cyp1B1 gene, particularly a similar enhancer region that directs AhR-mediated transcriptional activation. This report also describes several important differences in the regulation of basal transcription that may account for the different expression pattern seen for Cyp1B1 mRNA. Most notably, we demonstrate for the first time a mechanism for involvement of the AhR in basal expression of a gene in addition to TCDD induction and a novel protein complex formed at a pair of enhancer XREs.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Isolation of Cyp1B1 Genomic Clones-- A 129SV mouse liver genomic library contained within the Lambda Fix II phage (Stratagene) was screened using a SmaI restriction fragment of the mouse cDNA (1). This 1028-base pair probe encompasses 175 bp upstream of the initiation ATG codon that encodes the first methionine and 853 bp downstream. Two genomic clones (clones 8 and 10) were enriched to homogeneity by repeated rescreening at lower densities, and -DNA was purified at a high multiplicity of infection according to standard protocols (28). SalI restriction digestion of clone 8 (14.5 kb) resulted in one restriction fragment of 9.5 kb and another fragment of 5 kb. Clone 10 revealed an ~9-kb restriction fragment and another 3-kb fragment.

DNA Sequencing-- Double-stranded DNA sequencing of the 5'-flanking region, exon 1, and intron 1 was carried out according to the dideoxy termination method (29) using the double-stranded template PGEM24.7 and Sequenase Version 2.0 (U. S. Biochemical Corp.). The universal T7 and PX1-3 5'-GCT GCG ATG AAG CGT GGT oligomers were synthesized, and end labeling was carried out using T4 kinase and [-³²P]ATP. Based on the sequence obtained from these first reactions, synthetic primers were designed to "walk" along this template from both the 5'-upstream and 3'-downstream regions. Exon 2, intron 2, and exon 3 sequences were determined by the Applied Biosystems PRISM dye terminator cycle sequencing method as described by the manufacturer. The first two primers used for the sequences were designed according to the Cyp1B1 cDNA sequences. The DNA sequence was analyzed using the Genetics Computer Group sequence analysis software program (Version 7).

Primer Extension-- C3H10T1/2 cells were cultured in Dulbecco's modified Eagle's/F-12 medium and 10% fetal bovine serum containing penicillin (50 units/ml). Cells were treated with TCDD (1 nM) for 2 h and harvested in phosphate-buffered saline. Preparation of poly(A)⁺-enriched RNA was carried out according to Bradley et al. (30). Primer extension was carried out according to standard protocols (28). 3'-Downstream [-³²P]ATP-labeled oligonucleotides (5'-GAC CTA GAC ACC TGA GGC CCG CTG CTT TAG-3', 5'-AGC GGG ACC TTA GGG-3' and 5'-CTG CGC GCT GGA GCA AAG CTC AAC CAG GAG-3'; 5 × 10⁶ cpm) were used as primers. The same labeled oligonucleotides were used as primers for double-stranded sequencing of the Cyp1B1 genomic DNA.

Construction of 5'- and 3'-Deletions Linked to the Luciferase Reporter Gene-- A fragment of Cyp1B1 genomic DNA containing exon 1 and its 5'-flanking region was amplified by PCR using T7 and PX1-3 as the primers. The PCR products were digested with SacI and treated with T4 DNA polymerase prior to subcloning into the SacI and SmaI sites of the luciferase reporter vector pGL3Basic (Promega, Madison, WI). With this procedure, plasmids p1075/+371, p1075/+150, p1075/159, and p1075/207 were generated. p1075/+371 contains all of exon 1 (371 bp) and 1075 bp of 5'-flanking region. The DNA fragments with which p1075/+267, p1075/+22, and p1075/90 were constructed were generated using 5'-upstream primer T7 and 3'-downstream primers that had a NheI restriction site linker. The DNA fragments used to generate constructs p821/+124, p432/+124, and p210/+124 were prepared using a 5'-upstream primer that had a KpnI linker and 3'-downstream primers that had an NheI linker. The native orientation of all constructs was verified by DNA sequence analyses. A 265-bp DNA fragment (821 to 1075) was generated by PCR using 5'-upstream primer T7 and a 3'-downstream primer (5'-CAA CGG TAC CGC CAA CAA ACG GTT GGG TTG-3') that had a KpnI restriction site linker. The PCR products were digested by KpnI and then subcloned into the KpnI sites of plasmid p210/+124. Both the native and reverse orientations of the 265-bp DNA were inserted into p210/+124. p210N260 was the native orientation, and p210R260 was the reverse orientation.

Transfection and Luciferase and -Galactosidase Assays-- C3H10T1/2 and Hepa-1 cells and AhR-deficient embryo fibroblast cell lines were cultured in Dulbecco's modified Eagle's/F-12 medium and 5% fetal bovine serum. The AhR-deficient mouse primary embryo fibroblasts (kindly provided by Dr. P. Fernandez-Salguero, National Institutes of Health, Bethesda, MD) were used to generate AhR-deficient embryo fibroblast cell lines according to the method described by Reznikoff et al. (31). All plasmids containing constructs were transfected by the calcium phosphate coprecipitation method (32). The AhR-encoding cDNA plasmid (pmuAhR) was kindly provided by Dr. Christopher Bradfield (University of Wisconsin, Madison, WI). Approximately 24 h prior to transfection, cells were seeded at 5.5 × 10⁵ in 60-mm dishes. 300 µl of transfection buffer contained 6.5 µg of reporter recombinant plasmid and 1.5 µg of internal reference plasmid pCH110, in which the -galactosidase gene is driven by the SV40 promoter. 5 h following transfection, the cells were treated with 15% glycerol for 7 min, and then fresh medium containing 7% fetal bovine serum was supplied again. TCDD (1 nM) was added to the induction group 24 h prior to harvesting. Luciferase activity was determined using the luciferase assay system (Promega) according to the manufacturer's instructions and immediately measured in a luminometer. The -galactosidase assay was carried out in a total volume of 250 µl of assay buffer containing 0.12 M Na₂HPO₄, 0.08 M NaH₂PO₄, 0.02 M KCl, 0.002 M MgCl₂, 0.1 M -mercaptoethanol, 50 µg of o-nitrophenyl--galactoside, and 100 µg of cell extract.

Mobility Shift Assay-- Nuclear extracts were prepared from control and TCDD-induced (1 nM/2 h) C3H10T1/2 cells, and mobility shift DNA-binding protein assays were carried out as described (33). The oligonucleotides were labeled with [-³²P]ATP using T4 polynucleotide kinase. 10 µg of the nuclear protein and 10,000 cpm of probe were used in each reaction in the experiments. Competition experiments were conducted by co-incubation with 10, 50, and 200-fold excesses of unlabeled competitors. The sequence of Cyp1A1 DXE1 is GATCTACGGCTCCCCTCCCCCAGCTAGCG (positions 1109 to 1085). The rest of the oligonucleotide sequences are shown in Table I.

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

Mouse Cyp1B1 Gene Structure and Identification of the Transcription Start Site-- Restriction endonuclease analysis of a 14.5-kb mouse genomic clone has established a map of the exon/intron distribution and their representation within this genomic clone. Hybridization of these restriction fragments with probes specific for exons 2 and 3 has defined the positions of these exons. A complete DNA sequence analysis was carried out using the genomic clone as the template to encompass all sequences found in the 4.9-kb cDNA, 2.967-kb intron sequences, and 1.1 kb of upstream sequence. Comparison of the sequence obtained from the genomic clone with the cDNA sequence established that the gene includes sequences identical to those found in the cDNA and thus has allowed us to determine the mouse Cyp1B1 gene structure. A diagram summarizing the mouse Cyp1B1 gene structure is presented in Fig. 1A. The mouse Cyp1B1 gene exhibits three exons. Introns 1 and 2 are composed of 0.376 and 2.591 kb, respectively. Exon 1 is composed of 0.371 kb. The open reading frame (1.63 kb) is encoded by exon 2 (starting at the second nucleotide; 1.042 kb) and 16% of exon 3 (3.78 kb). The intron 1/exon 2 and intron 2/exon 3 junction sites obey the GT/AG rule (Fig. 1B), characteristic of exon/intron junction sites (34).

View larger version (27K): [in this window] [in a new window]   Fig. 1.   A and B, mouse Cyp1B1 gene structure. A schematic representation of the mouse Cyp1B1 gene structure is presented in A. The map shows three exons and two introns, including their respective restriction sites. The locations of the translation initiation site and stop codon are indicated in exons 2 and 3, respectively. The exon/intron junction site sequences of the Cyp1B1 gene are presented in B. 5'-Flanking and intronic sequences are uppercase, whereas exonic sequences are lowercase. C, identification of the transcription start site. Primer extension was carried out using avian myeloblastosis virus reverse transcriptase. A 3'-downstream end-labeled oligonucleotide (5'-GAC CTA GAC ACC TGA GGC CCG CTG CTT TAG-3') was used as the primer for extension from C3H10T1/2 mRNA. The same primer was used for double-stranded DNA sequencing of the mouse genomic clone. Lanes 1-4 represent the DNA sequencing ladder, and lane 5 shows the primer extension products. The arrows denote the major CYP1B1 extension product consisting of two bands.

Location of Putative cis-Acting Elements-- Fig. 2 shows the 5'-flanking region of Cyp1B1; 26 bp upstream of the start site, there is a TATA box-like sequence (TTAAAA), and 13 bp farther upstream, there is a steroidogenic factor-1 (SF-1) motif (TCCAGT). This transcription factor is selectively expressed in steroidogenic cells and interacts with shared promoter elements in several steroidogenic P450 genes to increase expression in response to cAMP elevation (35). Two SP-1 sites, implicated in the constitutive expression of P450 genes (21) and separated by an XRE, are located 65-87 nucleotides upstream of the initiation start site. The experiments described below show that this 90-base (1 to 90) region is necessary to confer full basal promoter activity to the Cyp1B1 gene. DNA sequencing also revealed 11 putative XREs (T_nGCGTG or GCGTG), five of which are located between 820 and 1075. We will demonstrate later that this 265-bp segment is the primary region for TCDD-mediated induction of the Cyp1B1 gene. Two other XREs are located in the proximal promoter region and immediately on the 3'-side of the transcription start site, respectively. The remaining four XREs are located in the region between the proximal promoter and the TCDD enhancer region. The 5'-flanking region also contains multiple GC-rich sequences in proximity to XREs (not highlighted in Fig. 2) that resemble DXEs, shown to be critical for TCDD induction (36). In addition, we found five E-box elements, implicated in the binding of helix-loop-helix Arnt nuclear regulatory factor (37), and two further SF-1 motifs. One E-box, one SF-1 site, and four DXEs are located in the TCDD enhancer region. The noncoding exon 1 exhibited two E-box elements and one XRE.

View larger version (83K): [in this window] [in a new window]   Fig. 2.   Nucleotide sequence of the mouse Cyp1B1 5'-flanking region including exon 1 and intron 1. Putative XRE, SP-1, SF-1, and E-box elements are underlined. The TATA-like element is double-underlined. The arrows indicate the major (doublet) and secondary start sites. The negative numbers represent the 5'-flanking region counting negatively from the major transcription start site. The shaded box represents the enhancer region, and white box represents the proximal promoter.

Upstream cis-Acting Regulatory Elements Involved in TCDD Induction-- To identify cis-acting elements involved in regulation of the mouse Cyp1B1 gene, a set of Cyp1B1-luciferase reporter plasmids were transfected into two mouse cell lines (C3H10T1/2 and Hepa-1) that differ in their expression of Cyp1B1. C3H10T1/2 cells are mouse stromal fibroblasts that predominantly express Cyp1B1, and Hepa-1 cells are mouse hepatoma cells that preferentially express Cyp1A1. 3'-End deletion of 578 nucleotides (from +371 to 207) did not significantly change the TCDD-induced activity of luciferase in either cell line (Fig. 3A). By contrast, a 5'-deletion of 265 bp (from 1075 to 821) completely removed TCDD induction in both cell lines (Fig. 3).⁴ The data suggest that TCDD induction is primarily dependent on positive regulatory elements present between 821 and 1075, where five XREs are localized. Other more proximal XREs play, at most, a secondary role in TCDD induction. Further 5'-deletion of 611 nucleotides (from 821 to 210) did not result in any change in TCDD-induced luciferase activity (Fig. 3B). Although induction factors varied up to 3-fold between separate cultures, analysis of transfections in transfection experiments with different batches of cells showed that the most active Cyp1B1-luciferase plasmid (p1075/+150) was equally inducible in both C3H10T1/2 (10.0 ± 3.0-fold) and Hepa-1 cells (9.34 ± 3.4-fold) (Fig. 3C).

View larger version (28K): [in this window] [in a new window]   Fig. 3.   TCDD induction of luciferase activity in C3H10T1/2 and Hepa-1 cells following transfections of Cyp1B1-luciferase deletion constructs. Systematic 3'- and 5'-deletions of the mouse Cyp1B1 cis-regulatory region were generated, and C3H10T1/2 and Hepa-1 cells were transfected with each construct. The transfection efficiency was normalized by cotransfection with a -galactosidase-expressing vector. The numbers represent -fold induction measured from triplicate assays in single experiments, representative of multiple experiments. The S.D. in the induction factor was calculated from the induced activities divided by a mean basal activity and therefore reflects variability in the induced activity. A, 3'-deletions of the mouse Cyp1B1 flanking region including exon 1; B, 5'-deletions of the same segment (SS, transcription start site); C, mean ± S.D. of induction factors for six different experiments.

Upstream cis-Acting Regulatory Elements Involved in Constitutive Expression-- Basal promoter activities (Fig. 4, A and B) were increased in both cell lines ~3-fold by deletion of the exon 1 sequence +267 to +150, indicating the presence of negative regulation through effects on either promoter activity or mRNA stability. Deletion of sequence +22 to 90 removed 85% of the basal promoter activity (Fig. 4, A and B), whereas TCDD induction was retained (Fig. 3A). This is fully consistent with the primer extension and DNA sequence analyses, which both showed that the sequence between +22 and 90 contains the major start site and key elements of the mouse Cyp1B1 promoter region. Further deletion to 207 had little additional effect on the low basal promoter activity (Fig. 4, A and B). These results indicate that the 110-bp sequence from +20 to 90 is essential for constitutive expression. The significant promoter activity remaining after deletions to 207, particularly when enhanced 9-fold by TCDD induction, indicates a second weak promoter region. This is consistent with the second upstream start site indicated by the primer extension experiments.

View larger version (40K): [in this window] [in a new window]   Fig. 4.   Constitutive transcriptional activities produced by 3'- and 5'-Cyp1B1-luciferase deletion constructs in C3H10T1/2 and Hepa-1 cells. 3'- and 5'-deletions of the mouse Cyp1B1 cis-regulatory region were generated, and C3H10T1/2 and Hepa-1 cells were transfected with each construct. Luciferase activity was assayed 48 h following transfection in the absence of an inducer. The 5'- and 3'-numbers are designated relative to the transcription start site (, upstream; +, downstream). A and B show luciferase activities of 3'-deletion constructs. The 3'-ends of the Cyp1B1 fragments are indicated (1075 is the 5'-end). C and D show luciferase activities obtained from 5'-deletion constructs. The 5'-ends of the Cyp1B1 fragments are indicated (+124 is the 3'-end). Luciferase activity was normalized to -galactosidase activity. Relative luciferase activities are calculated as the ratio of each activity to the control vector activity (defined as 10). The data represent the mean ± S.E. of triplicate measurements from a single experiment (representative of three). *, the 3'-end of this fragment is at +150. Deletion from +150 to +22 did not affect luciferase activity.

1075

Effect of the AhR on Cyp1B1 Promoter Activity-- Using embryos obtained from previously characterized AhR^/ mice (38), we have generated C57B6 mouse embryo fibroblast cell lines that we have shown by immunoblotting to be completely deficient in AhR.⁵ Expression of Cyp1B1 in primary mouse embryo fibroblasts and mouse embryo fibroblast lines developed from wild-type C57B6 mice was essentially the same as in C3H10T1/2 cells that were derived in a similar manner from C3H mice (39). Mouse embryo fibroblasts derived from AhR^/ C57B6 mice were used to define the role of the AhR in the basal expression of CYP1B1. Constitutive CYP1B1 protein expression is undetectable in all the cultures of AhR-deficient fibroblast cell lines (data not shown), whereas expression is normal in equivalent lines generated from heterozygous littermates that carry a single AhR allele that expresses AhR at wild-type levels.⁵ Fig. 5 shows that both basal expression and TCDD induction of Cyp1B1 promoter activity are lost in the AhR-deficient cells. When the cDNA encoding the AhR is cotransfected with Cyp1B1-luciferase into the deficient cells, the uninduced luciferase activity is increased 10-fold compared with transfection with p1075/+371 Cyp1B1-luciferase alone. This relative activity is comparable to that seen with C3H10T1/2 cells, which normally express the AhR. When this cotransfection was accompanied by TCDD stimulation, there was an increase in luciferase activity of only 25% compared with uninduced activity. The same result was obtained in two separate transfection experiments. This increase in induction compares with 8-fold induction by TCDD when the p1075/+371 Cyp1B1-luciferase construct is transfected into C3H10T1/2 cells.

View larger version (18K): [in this window] [in a new window]   Fig. 5.   Basal and TCDD induction of the Cyp1B1-luciferase construct is dependent on the AhR. An AhR-deficient mouse embryo fibroblast cell line was cotransfected with the Cyp1B1-luciferase construct (p1075/+371) and either plasmid pSV.Sport1 or pSV.Sport1 containing the AhR-expressing cDNA (pmuAhR). As experimental controls, a parallel pair of cotransfections were performed with the pGL3Basic vector (without the Cyp1B1 insert). Following transfection, cells were left untreated (Control) or induced with TCDD for 24 h and subsequently assayed for luciferase activity. The -galactosidase expression vector pCH110 was also cotransfected. Luciferase activities was normalized to -galactosidase activity. The data represent the mean ± range of duplicate measurements. Similar results were obtained from one additional experiment.

Binding of Nuclear Proteins to XREs-- The binding of nuclear protein, particularly the AhR·Arnt complex to XREs found in the enhancer domain, has been analyzed by gel mobility shift assays. Four 30-base oligonucleotides were tested containing XRE1/XRE2 (oligonucleotide A), XRE3 (oligonucleotide B), XRE4 (oligonucleotide C), and XRE5 (oligonucleotide D), respectively (Table I). Each oligonucleotide was tested in gel mobility shift assays with nuclear extracts from control and TCDD-induced C3H10T1/2 cells. These extracts were also tested with a 30-base oligonucleotide containing rat Cyp1A1 XRE1 (oligonucleotide 1A1). This is located in an equivalent 5'-flanking region of the Cyp1A1 gene and has been used in many other studies (18, 36). Mouse embryo fibroblast cells, however, contain 5-10 times less AhR and Arnt (40) than hepatoma cells such as Hepa-1, which have been used for these previous studies with the Cyp1A1 XRE.

View larger version (55K): [in this window] [in a new window]   Fig. 6.   Three enhancer Cyp1B1 XREs form complexes, two of which are distinct from AhR·Arnt complexes. A and B, nuclear extracts from C3H10T1/2 cells. Nuclear extracts from untreated () and TCDD-treated (+) C3H10T1/2 cells were analyzed by gel mobility shift assays with 30-mer oligonucleotides containing Cyp1B1 XREs (see Table I). A, competition between Cyp1B1 XREs and excess Cyp1A1 XRE1 for C3H10T1/2 nuclear protein. TCDD-induced extracts were mixed with the indicated oligonucleotides and 0-, 10-, 50-, and 200-fold molar excesses of the competitor. B, cross-competition between Cyp1B1 XREs (oligonucleotides A and C) for TCDD-induced C3H10T1/2 nuclear protein. Cyp1A1 DXE1 was tested as a control competitor at 10-, 50-, and 200-fold molar excesses. C, nuclear extracts from AhR-deficient embryo fibroblasts compared with extracts from congenic wild-type fibroblasts. Nuclear extracts from untreated () and TCDD-treated (+) C3H10T1/2 cells (wild type (Wt)) and two independent AhR-deficient embryo fibroblast lines, B6EF/B (AhB) and B6EF/F (AhF), were analyzed by gel mobility shift assays with 32P-labeled probes: oligonucleotides C (XRE4), D (XRE5), and A (XRE1/XRE2). The arrows indicate the positions of the AhR·Arnt and alternative XRE complexes.

View larger version (46K): [in this window] [in a new window]   Fig. 7.   A, alignment of the mouse and human CYP1B1 enhancers. The DNA sequences for equivalent mouse and human CYP1B1 (8) enhancer regions as defined by these deletion experiments were analyzed using the BESTFIT option of the Genetics Computer Group sequence analysis software program. Conserved elements are double-underlined. The top sequence represents mouse enhancer segments. B, comparison of the mouse Cyp1B1 enhancer region and the equivalent rat Cyp1A1 5'-flanking region. White boxes represent XREs, and shaded boxes represent auxiliary DXEs as defined by Robertson et al. (36).

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

This characterization of the mouse Cyp1B1 gene establishes that the Cyp1B1 gene is radically different in its structure and regulation from the Cyp1A1 gene, even though they share 40% identity in their protein sequences. Although much more selectively expressed in cells than Cyp1A1, Cyp1B1 is often constitutively expressed (1, 4). Here we show that Cyp1B1 differs from Cyp1A1 through a key contribution of the AhR to this constitutive expression. There are several points of close similarity to human CYP1B1, notably in the gene structure and in features of an AhR-responsive enhancer region (Fig. 7A) (8). This is clearly defined here as critical for both basal and TCDD-induced expression. The mouse Cyp1B1 proximal promoter shares features with human CYP1B1, but also shows several striking differences.

The mouse Cyp1B1 gene, like the human CYP1B1 gene (8), has three closely spaced exons compared with the seven exons of the CYP1A1 gene (42), which is more typical of mammalian P450 genes (43, 44). Previous work has shown that Cyp1B1 is also exceptional among mammalian P450 genes in the length of the translated region (543 versus 520 amino acids for Cyp1A1) and the 3'-untranslated region (3.15 versus 1 kb). In spite of these differences, Cyp1B1 shares with Cyp1A1 inducibility by TCDD and other polycyclic aromatic hydrocarbons through the AhR. This work demonstrates the presence of an upstream enhancer domain that is similar to that in Cyp1A1 in location, size, and number of XREs (five). This provides strong support to the idea that both induction processes are governed by the same mechanism. In spite of this similarity, Cyp1B1 exhibits different cell selectivity in expression compared with Cyp1A1 and, unlike Cyp1A1, commonly exhibits substantial constitutive expression. The analysis of promoter activity presented here rigorously establishes for the first time that the AhR is required for basal expression by acting through the same enhancer domain. Thus, the similar enhancer domains in Cyp1B1 and Cyp1A1 function differently under basal conditions. We have also shown complex formation distinct from AhR·Arnt heterodimers with two XREs in the Cyp1B1 enhancer. These unusual TCDD-insensitive complexes may explain high basal expression of Cyp1B1. Transient transfections with 5'-flanking constructs of Cyp1B1 and Cyp1A1 reproduce only some of the cell selectivity of TCDD-induced expression seen with the genes. This is consistent with the idea that chromatin structure and modification are important elements in TCDD action and this selectivity (13).

The double start site for the Cyp1B1 gene detected by primer extension is conserved in a variety of rodent cells, such as rat mammary fibroblasts and rat adrenal glands.⁷ The gene sequence upstream of the start site does not contain a TATA box element, but exhibits a TATA-like sequence (TTAAAA) similar to what has been reported for human CYP1B1 (8). The initiation site is flanked by XREs (GCGTG) on the upstream and downstream sides, suggesting that the AhR and Arnt may be directly involved in the transcription initiation process. The upstream XRE separates a pair of SP-1-like sequences immediately adjacent to the TATA-like sequence. Recent analysis of the human CYP1B1 promoter established the importance of these conserved SP-1- and TATA-like sequences to basal promoter activity (45). Work from different laboratories has shown that SP-1 elements, in cooperation with an XRE, are capable of producing an enhancement of transcriptional activation in a synergistic fashion (21, 46). This proximal promoter region is also notable for a pair of SF-1-binding elements that have been associated with cAMP-mediated induction of hydroxylases involved in steroidogenesis (35, 47-49). Since Cyp1B1 expression is under hormonal control mediated by cAMP in endocrine-regulated tissues (4-6), these putative SF-1 elements (AGGTCA) may provide binding sites for this transcription factor. However, the mouse and human sequences within the proximal promoter regions are quite different. For example, the human CYP1B1 promoter sequence lacks the XREs that flank the mouse Cyp1B1 transcription start site and the SF-1 element. An SF-1-binding motif in the proximal promoter regions of Cyp11A1, Cyp21, and CYP19 are crucial for up-regulation in response to cAMP (50-52).

A second low efficiency start site located 332 nucleotides upstream of the first site by primer extension repeats the upstream 10-base sequence AGAGGGTTGG. This secondary site lacks the upstream SP-1- or TATA-like sequences, but when linked to luciferase (1075 to 207), retains 5-10% of the maximum basal activity without affecting TCDD induction. Multiple start sites have been reported for genes containing TATA-less promoters such as the housekeeping aspartate aminotransferase gene (53). Initiation of transcription from different start sites could be tissue-specific and may contribute to tissue-specific gene expression (54-56).

Cyp1B1-luciferase reporter analysis reveal that the same 265-bp sequence (810 to 1075) is critical for regulation of basal activity and TCDD stimulation. This enhancer contains a cluster of five XREs, which are completely conserved in the human CYP1B1 enhancer (Fig. 7A). Deletion of this region not only removed nearly all TCDD induction, but decreased basal expression 10-fold. The mouse Cyp1B1 enhancer region shows substantial similarity to the key upstream enhancer of rat Cyp1A1 (as illustrated in Fig. 7B), which is also located at ~1 kb upstream of the start site (821 to 1110) (57). This region in the rat Cyp1A1 gene contains five XREs and six GC-rich elements (DXEs), which have been linked to a capacity to modify induction mediated by the AhR (17, 36). Alignment of the Cyp1B1 and Cyp1A1 enhancer regions showed a 47-bp region (971 to 1018) that is over 70% identical to a similarly placed sequence in rat Cyp1A1 (940 to 987). Two DXEs (DXE4 and DXE5) that are located in this 47-bp region in rat Cyp1A1 bind distinct nuclear factors that are different from the AhR, Arnt, and SP-1.⁸ The mouse Cyp1B1 enhancer region has four GC-rich elements that resemble DXEs. Since the Cyp1B1 enhancer region has similar sequence characteristics, the mechanism of the regulation of Cyp1B1 response to TCDD may be very similar to that described for Cyp1A1. However, Cyp1A1 is rarely expressed at significant constitutive levels, indicating that the respective enhancers respond differently to low levels of constitutive nuclear AhR. This mouse Cyp1B1 enhancer also contains one SF-1 site (position 907), which is conserved in the corresponding human CYP1B1 enhancer and may contribute to cAMP responsiveness.

Gel mobility shift assays show that AhR·Arnt heterodimers bind to one site in the Cyp1B1 enhancer region (XRE5) in a manner analogous to Cyp1A1 XRE1, including TCDD stimulation of binding. XRE5 is also flanked on the 3'-side by a GC-rich sequence that is similar to DXE1, which is located adjacent to XRE2 in the Cyp1A1 enhancer. Surprisingly, the remaining three XRE-containing oligonucleotides formed different complexes. XRE1/XRE2 (oligonucleotide A) and XRE4 (oligonucleotide C) formed complexes of the same size as the AhR·Arnt complex of XRE5. They are clearly distinct complexes since they are retained in AhR-deficient embryo fibroblasts and Arnt-deficient Hepa-1 cells. In addition, these complexes were formed to a higher extent, were insensitive to TCDD, were not blocked by anti-AhR antibodies, and were only weakly competitive with XRE5 or Cyp1A1 XRE1. These two anomalous XREs bound the same protein, as evidenced by very high mutual competition but weak competition with XRE5 and Cyp1A1 XRE1. XRE1/XRE2 and XRE4 share a 12-base core sequence (GCGGCGCACGCA) that is identical in 10 out of 12 positions in XRE5. Interestingly, even though XRE4 and XRE5 bind distinct complexes, they also share an identical 11-base sequence that contains the XRE. The GG pair from the consensus 12-base sequence that is replaced by TC in XRE5 may determine which complex binds to this DNA element. This alternative complex does not depend on the XRE sequence, and indeed, binding is enhanced by substitution in this element (Table I, XRE4/MutA). The oligonucleotide containing XRE3 formed a much smaller complex that did not involve either AhR·Arnt heterodimers or the XRE sequence. Mutational analysis shows that the E-box sequence, but not the adjacent XRE, is critical to this complex and that even when this is substituted, no binding occurs to the XRE.

These anomalous complexes that form in the enhancer region may well contribute to the basal activity of Cyp1B1. These complexes suggest that the sensitivity of basal promoter activity to the AhR is due to a large increase in an additional AhR-stimulated protein. Similar complexes were observed with nuclear extracts from Hepa-1 cells. Significantly, we have found that TCDD induction is retained after deletion of a sequence containing XRE3, XRE4, and XRE5 from a rat Cyp1B1-luciferase promoter construct.⁹ It seems that when an extended promoter sequence is present, the AhR exhibits strong TCDD-responsive enhancer activity through the XRE1/XRE2 element alone. This suggests that the blocking secondary structures form less readily in the full enhancer than in the 30-mers. Consistent with enhancer characteristics, this region mediates TCDD induction effectively in either native or reverse orientations when directly linked to the proximal promoter region. Interestingly, the native orientation appears to be much more effective in stimulating basal promoter activity. This difference could reflect a sensitivity of the enhancer orientation to the lower amount of nuclear AhR under basal conditions or be caused by a fundamental difference between the basal and TCDD-induced mechanisms for the enhancer elements. The second interpretation is consistent with the novel complexes described above for the Cyp1B1 enhancer region that are independent of TCDD.

Although this enhancer region functions equally well in mediating TCDD induction in C3H10T1/2 fibroblasts and Hepa-1 cells, basal Cyp1B1 promoter activity is 10-fold more effective in C3H10T1/2 fibroblasts. This is far lower than the differences observed for in vivo expression of the genes in these cells (>100-fold). However, like expression of the genes, we again see selectivity in basal transcription rather than induction, which is also retained for Cyp1B1 in Hepa1 cells, although with greatly diminished total activities. The major difference seems to reside with a 10-fold greater effectiveness of the basal activity of the upstream enhancer region in C3H10T1/2 cells.

Evidence has been previously presented that TCDD activation of the AhR causes an initial cooperative binding to the enhancer domain that then causes an opening up of cis-acting elements more proximal to the transcription start site (13, 22). The low cell selectivity shown by the transiently transfected Cyp1B1 and Cyp1A1 promoter constructs between C3H10T1/2 cells and Hepa-1 cells indicates either that cis-acting elements farther upstream are critical determinants of cell-specific expression or that there is cell-selective exposure of the distinct Cyp1B1 or Cyp1A1 chromatin regions. Selective expression of Cyp1A1 and Cyp1A2 was only modeled by luciferase promoter constructs when they were fully integrated into the genome after stable transfections (58).

Deletion analysis also revealed two inhibitory domains in the Cyp1B1 upstream region that each decreased basal promoter activity without affecting TCDD induction. The inhibitory region located between 210 and 432 also contained the minor start site (332), two XREs, an E-box, and a putative SP-1 site (18-base G-rich sequence, G₆TG₆TG₄). A negative cis-acting regulatory element has been identified for the Cyp1A1 gene in rat epidermal cells (27). This is associated with differentiation at high passage numbers concomitant with down-regulation of Cyp1A1 and Cyp1B1 (27). A sequence similar to half of this element is found in the mouse Cyp1B1 region immediately downstream of the enhancer element in a position equivalent to the Cyp1A1 element (27). This cis-acting element is distinct from the negative regulatory element described for the CYP1A1 5'-flanking region by Hines and co-workers (59, 60). The second inhibitory domain is located in exon 1 (+150 to +269). Since exon 1 will be included in the luciferase transcript, this element may either inhibit transcription or affect the stability or translation efficiency of the transcript.

Recent work with AhR-deficient mice (AhR^/) has shown that Cyp1A2 and Ugt106 basal expression is lost compared with heterozygous AhR^+/ littermates that express the AhR (38). In studies to be published elsewhere, we have developed AhR^/ fibroblast cell lines from these mice⁵ and have shown that these cells do not express either basal or TCDD-induced Cyp1B1, consistent with involvement of the AhR in basal expression. We have also seen that there is no basal or TCDD-induced Cyp1B1-luciferase activity unless the cells are cotransfected with a vector encoding the AhR. The response to TCDD is small probably because activities produced by the receptor alone are remarkably high. Basal activities are comparable to basal activities in C3H10T1/2 cells in spite of low levels of transfected AhR. Similar high basal activities have been seen for Cyp1A1 promoter constructs in AhR-deficient hepatoma cells (61). Hyper-responsiveness has also been seen after cotransfection of steroid receptors with reporter constructs (62). This probably reflects the diminished association of nucleosomes with the transiently transfected promoters.

This first dissection of the upstream region of mouse Cyp1B1 has revealed some strong similarity to Cyp1A1, sufficient to suggest a common mechanism of AhR action in transcriptional activation of these genes. We have also provided clear evidence for the role of the AhR in constitutive and TCDD-induced activation of Cyp1B1. It remains to be determined whether different sequences in the 265-bp upstream enhancer play a role in the basal and induced activities. Cyp1B1 enhancer activity is also balanced against two domains that confer inhibitory effects on transcription. Recent human genetic studies have shown that a deficiency of human CYP1B1 leads to congenital glaucoma (63). This confirms previous suggestions based on the selective expression in steroid-responsive and steroidogenic tissues that CYP1B1 produces a physiological oxygenase product that has regulatory functions (4). It will be important to identify factors such as those binding to the enhancer XREs that may contribute to constitutive regulation of this upstream region.