QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) All Versions of this Article: mol.106.033795v1 72/3/514    most recent Submit a response Alert me when this article is cited Alert me when eLetters are posted Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via Google Scholar Google Scholar Articles by Biggs, J. S. Articles by Yost, G. S. Search for Related Content PubMed PubMed Citation Articles by Biggs, J. S. Articles by Yost, G. S.

Transcription Factor Binding to a Putative Double E-Box Motif Represses CYP3A4 Expression in Human Lung Cells

Department of Pharmacology & Toxicology, University of Utah, Salt Lake City, Utah (J.S.B., J.W., N.S.C., G.S.Y.); Department of Pharmacology and Toxicology, University of Oulu, Oulu, Finland (J.H., P.U.); and University of Kuopio, Kuopio, Finland (H.R.)

Received December 27, 2006; accepted June 4, 2007

	Abstract

Top Abstract Materials and Methods Results Discussion References

 
Two vital enzymes of the CYP3A subfamily, CYP3A4 and CYP3A5, are differentially expressed in the human lung. However, the molecular mechanisms that regulate tissue-selective expression of the genes are poorly understood. The ability of the 5' upstream promoter region of these two genes to drive luciferase reporter activities in human lung A549 cells was dramatically different. The CYP3A5 promoter region activated luciferase gene expression by 10-fold over the promoterless construct, whereas the CYP3A4 promoter did not drive expression. Sequence comparisons of the promoters identified a 57-base pair insertion in the CYP3A4 promoter region (–71 to –127) that was absent in the CYP3A5 promoter. Deletion of the 57-bp motif from CYP3A4 or insertion into the CYP3A5 promoter, showed that this motif represses CYP3A4 expression in lung. EMSA analysis using nuclear extracts from either A549 cells or human lung tissues showed two specific protein/DNA complexes formed with the ³²P-labeled CYP3A4 57-bp oligonucleotide. EMSA analyses identified two E-box motifs as the minimal specific cis-elements. Supershift assays with antibodies directed against known double- or single-E-box binding factors (TAL1, EF1, E2A, HEB, etc.) failed to identify this factor as a previously characterized trans-acting double E-box binding protein. These results demonstrated that the 5'-upstream region of CYP3A4 contains an active putative double E-box repressor motif, not present in the 5'-upstream region of the CYP3A5 gene, that attenuates CYP3A4 expression in the human lung. We believe that this is the first documented case in which a cytochrome P450 gene is actively repressed in a tissue-specific manner.

The regulation of cytochrome P450 gene expression has received much attention in recent years, and the mechanisms responsible for CYP3A gene expression vary widely and are often complex (Guengerich, 1999; Quattrochi and Guzelian, 2001; Gibson et al., 2002; Goodwin et al., 2002b). To address the obvious importance of CYP3A in drug-drug interactions, a humanized mouse model of hepatic CYP3A4 regulation has been created (Zhang et al., 2003), and much work has been done in recent years to elucidate its regulatory mechanisms (Ding and Kaminsky, 2003; Schuetz, 2004; Xie et al., 2004). The 5'-flanking regions of the CYP3A subfamily of genes contain sequence motifs (i.e., cis-elements) that can regulate these genes in three generalized ways: 1) those that are involved in enhancing the enzyme's production [inducers: xenobiotic responsive enhancer module (XREM), pregnane X response element (PXRE), glucocorticoid receptor (GR), etc.] (Hukkanen et al., 2000, 2003; Goodwin et al., 2002b); 2) those that are involved in maintaining basal level expression (e.g., nuclear factor Y (NFY), nuclear factor B (NFB), constitutive androstane receptor (CAR), hepatocyte nuclear factor 3 (HNF3), etc.) (Iwano et al., 2001; Saito et al., 2001; Goodwin et al., 2002a; Bombail et al., 2004); and 3) those involved in turning genes off in the presence or absence of specific signals [e.g., CCAAT/enhancer-binding protein, -LIP, silencing mediator for retinoid and thyroid hormone receptors (SMRT)] (Chen and Evans, 1995; Chen and Li, 1998; Jover et al., 2002; Johnson et al., 2006). The combined actions of these and other cis-acting elements ensure that the proper amount of CYP3A protein is produced only at the precise time it is needed.

The CYP3A enzymes expressed in human lung are important determinants of pulmonary carcinogenesis caused by metabolism of several inhaled xenobiotic compounds (Piipari et al., 2000; Mollerup et al., 2001; Yeh et al., 2003). Other lung diseases cause significant morbidity and mortality, and specific P450 enzymes are at least partially responsible for these diseases (Ding and Kaminsky, 2003; Yeh et al., 2003). These P450 enzymes probably participate in the metabolism and bioactivation of polycyclic aromatic hydrocarbons and other procarcinogens present in combustion products, tobacco smoke, and ambient particulate matter (Nelson et al., 1996; Guengerich and Shimada, 1998). CYP3A4 and CYP3A5 are active in the metabolic detoxication of benzo[a]pyrene, but they are also partly responsible for the activation of benzo[a]pyrene-7,8-diol to carcinogenic diol epoxides that are capable of covalently binding to DNA. CYP3A4 and CYP3A5 exhibit differences in expression patterns, both within tissues and among individuals (Ding and Kaminsky, 2003). Transcripts of CYP3A4 have not been found in respiratory epithelial cell lines, such as the human A549 lung cell line and the immortalized bronchial epithelial BEAS-2B cell line. Nor have transcripts been detected in human bronchoalveolar macrophages or peripheral blood lymphocytes (Willey et al., 1996; Anttila et al., 1997; Hukkanen et al., 1997; Hukkanen et al., 2000; Piipari et al., 2000). In contrast, CYP3A5 is consistently expressed in all of these cell types. Moreover, tissue-specific expression of the CYP3A genes is thought to be a major factor influencing interindividual variation in both drug response and lung disease susceptibility (Piipari et al., 2000; Lamba et al., 2002). Thus, the dominant mechanisms that regulate CYP3A expression in the human lung must be established to understand interindividual and inter-racial susceptibility to inhaled xenobiotics.

	Materials and Methods

Top Abstract Materials and Methods Results Discussion References

 
Reagents. Plasmid/RNA isolation kits were purchased from QIAGEN (Valencia, CA). Bacterial artificial chromosomes containing human sequences for the CYP3A loci [AC005020, AC011904, and AF280107 [GenBank] (Gellner et al., 2001)] were purchased from Incyte Genomics (Palo Alto, CA). Oligonucleotides used in PCR amplification of genomic sequences and EMSA analyses were purchased from Integrated DNA Technologies (Coralville, IA). The pGL3-luciferase reporter-plasmid and dual-luciferase reporter assay system were purchased from Promega (Madison, WI). SDS-polyacrylamide gel electrophoresis reagents (37.1:1 acrylamide/bis-acrylamide, ammonium persulfate, and N,N,N',N'-tetra-methyl-ethylene-diamine) were purchased from Bio-Rad Laboratories (Hercules, CA). Human lung tissue was acquired from consenting donors through the Rocky Mountain Donor Services (Salt Lake City, UT). Antibodies against known E-box binding proteins were generous gifts from Dr. Frans van Roy (SIP1) and Dr. Cornelis Murre (E2A and E47) or purchased from Santa Cruz Biotechnology (Santa Cruz, CA). Platinum Pfx DNA polymerase, Platinum PCR Supermix High Fidelity, TOPO cloning kits, cell culture media, restriction enzymes, and all other molecular biology reagents were purchased from Invitrogen (Carlsbad, CA), BD Gentest (Woburn, MA), or Lonza Walkersville, Inc. (Walkersville, MD). Real-Time PCR reagents were purchased from Bioline (Randolph, MA) and Invitrogen. All sequence manipulations were conducted in silico using the Clone Manager Suite 7.0 (Scientific and Educational Software, Durham, NC) software package.

Cell Culture. Human adenocarcinoma A549 cells were obtained from American Type Culture Collection (Manassas, VA). A549 cells (p98) were seeded in 75-cm² flasks using cryopreserved aliquots (1 x 10⁶ cells; P87–90) and maintained for 48 h with Dulbecco's modified Eagle's medium/nutrient mixture with Ham's F12 medium, supplemented with 10% fetal bovine serum. For subculturing, cells were trypsin-disassociated and reseeded at 10% confluence in fresh media. For transfection assays, cells were trypsin-disassociated and reseeded in 96-well plates at a concentration of 1.0 x 10⁴ cells/well in 100 µl of the appropriate media. Normal Human Bronchial Epithelial (NHBE) cells were obtained from Lonza Walkersville and were grown in bronchial epithelial growth medium and cultured as above in flasks and dishes coated with collagen, fibronectin, and albumin. They were passaged no more than three times before transfection. Frozen primary hepatocytes were obtained from BD Gentest. They were thawed in a 75-cm² flask according to the supplier's instructions and then reseeded as described above for A549 cells into 96-well plates in hepatocyte growth medium.

Construction of Promoter-Luciferase Reporter Plasmids. Luciferase reporter constructs containing promoter regions from CYP3A5 (pGL3A5–218, pGL3A5–443, pGL3A5–647, pGL3A5–872, pGL3A5–1150, and pGL3A5–1365) and a fragment of the 5'-untranslated region of the gene (+31 base pairs) were adapted from previous work (Hukkanen et al., 2003). Chimeric CYP3A4-promoter- and CYP3A5-promoter-luciferase reporter plasmids were prepared by PCR amplification of the 5'-flanking regions of these genes from bacterial artificial chromosomes (i.e., AF_280107 for 3A4 clones; AC_005020 for 3A5 clones) using the primers listed in Tables 1 and 2. Both homology maps comparing the CYP3A4 and CYP3A5 promoters and stringent MATCH/TRANSFAC analyses were conducted to predetermine putative transcriptional binding motifs to guide primer design. Cloning primers introduced 5'-restriction sites for subsequent insertion into the multiple cloning site of the pGL3-Basic vector. For CYP3A4, a 5'-NheI site and a 3'-HindIII site were incorporated into the forward and reverse primers, respectively; for CYP3A5 constructs, a 5'-MluI site and a 3'-BglII site were used. pGL3-3A4 constructs encompassed regions surrounding the transcriptional start site (TSS) (i.e., from 3'-TSS + 118 to 5'-TSS, –35, –59, –72, –80, –150, –170, –222, –2795, and –13kb, respectively). A larger pGL3-3A5 5.3-kb plasmid was also created by PCR-amplification of a 2.3-kilobase pair region (bases 10,195 to 13,107) from a bacterial artificial chromosome containing CYP3A5 (AC_005020). Using the oligonucleotides Forward-3A5 5.3 kb-KpnI (5'-ggtaccATGCTCGTGTGCCTGATAAC-3'; sense, bases 10,195 to 10,216) and A5_P1_Reverse (5'-GCATTGCTTTGGGTAGTATGGAC-3'; antisense, bases 13,107 to 13,085), a product incorporating a KpnI restriction site at the 5'-end was created. Utilization of this new KpnI site, and a unique AflII restriction site (bases –2837 to –2842) in the CYP3A5 promoter, directed the proper insertion of this region into the original pGL3-3A5 3-kb plasmid, which was prepared separately. All plasmids were screened with restriction digestions using enzymes, which cut at least once in both the vector and the insert, and then sequenced by the University of Utah core sequencing facility.

Site-Directed Mutagenesis. Site-directed mutagenesis of the pGL3-CYP3A5–208 construct was performed with the QuikChange site-directed mutagenesis kit (Stratagene, La Jolla, CA) according to the manufacturer's instructions. The mutations were introduced using specific oligonucleotides (listed in Table 2). Correct assembly of the mutations was confirmed by sequencing.

PCR-Based Mutations of CYP3A4 and CYP3A5 Constructs. The chimeric reporter plasmids created above were mutated to introduce or knockout the 57-bp region for the CYP3A5 and CYP3A4 constructs, respectively. Primers used in these methods are listed in Tables 1 and 2. The CYP3A5-knock-in construct was made with forward and reverse primers with 3' ends that were complementary to the boundaries of homology within the CYP3A5 promoter, but whose 5' ends were complementary to the CYP3A4 57-bp region. To reduce errors in oligonucleotide syntheses, the 57-bp insertion was divided between the forward and reverse primers, and Platinum Pfx DNA polymerase (a proof-reading polymerase that immediately terminates at the end of a template leaving blunt ends) protocols were applied as the manufacturer suggested. CYP3A4 57-bp deletion mutants were generated by long-template PCR amplification of the entire vector excluding the targeted 57-bp region. To accomplish these deletions, the wild-type CYP3A4 constructs were subjected to a deletion protocol using Platinum PCR Supermix High Fidelity containing 20 ng of original template and 20 nM concentrations of each primer (3A4–57 bp-mut F and 3A4–57 bp-mut R) in a total volume of 50 µl and subjected to an initial melting step of 94°C for 2 min, 20 cycles of amplification (94°C for 30 s; 55°C for 90 s; 68°C for 8.5 min.), and a final capping step of 68°C for 3.5 min. Ten PCR reactions were pooled and the template strands were removed by digestion with DpnI. The mutated products were purified on a 0.7% agarose gel with an expected product size of 8 kb. Because the initial PCR reaction mixture contained Taq DNA polymerase, the resulting products also contained 3'-polyadenosine overhangs that were subsequently removed with mung bean nuclease (New England Biolabs, Ipswich, MA), gel-purified as before, and then self-ligated with T4-ligase in Rapid Ligation Buffer (Promega). Due to the lack of flexibility in selection of the priming site, this protocol consistently yielded a 64-bp knockout CYP3A4 construct with 7 bp more than the 57-bp region removed (the total region that was deleted was –71 to –132), but the 64-bp knockout still closely resembled the homologous region of the CYP3A5 gene. Multiple attempts to obtain a knockout construct with precisely 57 bp deleted were not successful. Similar attempts to restore this region to these constructs using the knockin technique described above were also unsuccessful. However, a similar knockout was created using the protocol described above with 3A4-SacII (Table 1) primers to introduce two SacII restriction sites flanking the 57-bp region. Creation of the CYP3A4-Spacer construct was accomplished using the 64-bp knockout as a template, using primers that harbored the additional spacer sequence at the 5' end such that they would be included in the final PCR product. To insure that there were no other mutations introduced into the pGL3-Basic vector during these exceptionally long polymerization events, all completed promoter constructs were subcloned back into the original pGL3-Basic vector, screened with restriction enzymes, and sequenced.

Transient Transfection and Luciferase Assay. Parallel luciferase reporter assays were conducted to compare the genetic differences between CYP3A4 and CYP3A5 transcriptional activity in cultured human lung A549 cells. Approximately 50-ml cultures of each plasmid were purified using the EndoFree Plasmid Maxi Kit (QIAGEN, Valencia, CA.), yielding 100–200 µg of DNA each. When confluence reached 70%, the A549 cells were transfected with 0.25 µg of reporter plasmid and 0.005 µg of Renilla reniformis luciferase plasmid (pRL-SV40) using FuGene 6 (Roche, Indianapolis, IN), according to the manufacturer's suggestions. Primary cell cultures (NHBE and hepatocytes) were transfected using Effectene reagent (QIAGEN), according to the manufacturer's suggestions. Cells were lysed for 36 h (A549) or 24 h (primary cultures) after the transfections, and the respective luciferase activities were determined using the dual-luciferase assay system (Promega). Firefly luciferase activities for the experimental constructs were normalized for transfection efficiency and cell loading using R. reniformis luciferase activity and total protein concentration, respectively. Data from these experiments were expressed as -fold luminescence over the activity of the promoterless pGL3-Basic reporter plasmid. The data for A549 transfections were presented as mean -fold luminescence (± S.D.) for three independent experiments performed in quadruplicate. NHBE and hepatocyte data were for a single experiment with nine replicates for each plasmid. Transfections of the mutated CYP3A5 –208 constructs to A549 cells were performed as described previously (Hukkanen et al., 2003).

Quantitative Real-Time PCR. Total RNA was purified from 10⁶ NHBE or A549 cells using TRIzol. One microgram of total RNA was used to synthesize first-strand cDNA using random hexomers and SuperScript II (Invitrogen), diluted 1:5 and then evaluated by quantitative real-time PCR using a Chromo-4 cycler (Bio-Rad Laboratories) and SYBR green (Invitrogen). The primers (a generous gift of Roger Gaedigk, University of Missouri, Kansas City, MO) were: for CYP3A4 detection, 5'-CTCTCATCCCAGACTTGGCCA-3' and 5'-ACAGGCTGTTGACCATCATAAAAG-3'; for CYP3A5 detection, 5'-GACCTCATCCCAAATTTGGCGG-3' and 5'-CAGGGAGTTGACCTTCATACGTT-3'; -actin was used as a housekeeping control gene (primers: 5'-GACAACGGCTCCGGCATGTGCA-3' and 5'-TGAGGATGCCTCTCTTGCTCTG-3'). We calculated the relative expression by using plasmid copy number standards.

Electrophoretic Mobility Shift Assay. Nuclear extracts from A549 or HepG2 cells and human lung tissues were prepared as described previously (Carr et al., 2003). EMSA was performed using the gel-shift assay system from Promega essentially as described by the manufacturer. Binding reaction mixtures were preincubated at room temperature for 10 min. The mixtures contained 4 µl of nuclear extract (4 µg for lung tissue and 6 µg for cell cultures), 0.005 to 0.01 pmol ³²P-labeled oligonucleotide probe, and 2 µlof5x binding buffer [50 mM Tris-HCl, pH 7.5, 250 mM NaCl, 5 mM MgCl₂, 2.5 mM EDTA, 2.5 mM dithiothreitol, 20% glycerol, and 0.25 mg/ml poly(dI-dC)·poly(dI-dC)] in a total volume of 10 µl. For competition experiments, a 100-fold molar excess of unlabeled double-stranded oligonucleotide was incubated for 15 min with nuclear extract before the addition of 1 µl of ³²P-labeled oligonucleotide probe (0.005–0.01 pmol). The mixtures were incubated for another 20 min at room temperature. Immediately before electrophoresis, gel loading dye (25 mM Tris-HCl, pH 7.5, 0.02% bromphenol blue, and 4% glycerol) was added to all binding reaction mixtures. The protein/DNA complexes and unbound probes were separated by nondenaturing gel electrophoresis using 4% polyacrylamide gels (1–8 V/length in centimeters) and detected by autoradiography.

Double-stranded oligonucleotides comprising the consensus binding sequences for AP1, NFB, Sp1, OCT1, and the 2F1-LSF (Carr et al., 2003) were included (175-fold molar excess) as nonspecific competitors in the gel shift assay system (Promega). Sequence-specific competitive oligonucleotides (1–11, Table 3) were generated by mutating five base pairs at a time, using the conversion of A>C and T>G; competitors M9 to M55 (Table 3) contain single base pair mutations corresponding to the distance 5'- to 3' within the 57-bp insertion; and E-box knockout competitors were created by converting the consensus sequence (CACCTG) to a sequence that is not known to bind any transcription factor (AAAAT). Competitive oligonucleotide probes were synthesized by either Integrated DNA Technologies (Coralville, IA) or the University of Utah Core Research Facilities. Sequences of the DNA probes used in competitive EMSA experiments are listed in Table 3.

View larger version (19K): [in this window] [in a new window]   Fig. 1. Transcriptional activity of pGL3-CYP3A4 and pGL3-CYP3A5 constructs transiently transfected into human lung A549 cells as measured using the Dual Luciferase reporter system (Roche). Expression levels were normalized to activity of the cotransfected R. reniformis luciferase vector, and are expressed as -fold of luciferase activity over the promoterless pGL3-Basic plasmid. For all experiments, pGL3-CYP3A5 constructs (A) expressed considerably higher luciferase activity than that of pGL3-CYP3A4 constructs (B) (50-fold versus 2-fold, respectively). Error bars represent S.D. of four replicate experiments run in quadruplicate.

	Results

Top Abstract Materials and Methods Results Discussion References

View larger version (20K): [in this window] [in a new window]   Fig. 2. Schematic representation of the first 150 base pairs of the CYP3A4 and CYP3A5 5'-upstream regions. The CYP3A4 sequence has a series of subtle differences (i.e., one or two base pairs) in the core binding sequences of the CCAAT-box, BTE-box, and the ER6 motif that have been previously shown (Iwano et al., 2001) to cooperate in driving basal expression of CYP3A5 in human hepatoma HepG2 cells. The most obvious difference between these two regions is a 57-base pair insertion (–71 to –127) in the CYP3A4 promoter, which includes another intact CCAAT-box (–119 to –124) [shown to constitute a portion of the functional C/EBP element (–121 to –130) for CYP3A4 regulation in HepG2 cells (Bombail et al., 2004)] and two E-box binding motifs, which are in phase, two full turns of the DNA double helix apart. Sizes and distances of salient features are not drawn to scale.

View larger version (34K): [in this window] [in a new window]   Fig. 3. Effect of site-directed mutations on CYP3A5 promoter transcriptional activity. A, the proximal –208 bp of the CYP3A5 promoter (H3A5) was compared with the corresponding region of the CYP3A5 pseudogene (H3A5P1). The differing nucleotides are numbered, and the transcription factor binding sites known to be important for CYP3A5 transcriptional activation in liver (Iwano et al., 2001) are indicated. B, mutated pGL3-CYP3A5 –208 constructs were transfected into A549 lung cells and the luciferase activities were compared with the luciferase activity produced by the wild-type (CYP3A5–208) construct. Luciferase activities were normalized to the activity of the cotransfected R. reniformis luciferase vector, and are expressed as -fold luciferase activity over the wild type. The data represent means ± S.D. of four separate replicates. The experiment was repeated three times with similar results.

Function of the CYP3A4 57-bp Insertion in the Repression of CYP3A4 Gene Expression. We examined the repressive ability of the CYP3A4 57-bp insertion in the human A549 cell line by creating a series of mutated CYP3A4 and CYP3A5 promoter luciferase constructs. Introduction of the CYP3A4 57-bp insertion into the analogous position of a CYP3A5 promoter construct (pGL3–3A5–3000) significantly reduced luciferase-driven expression of that construct in human A549 lung cells by roughly 50% (Fig. 4). Similar activity changes were observed for mutations in the smaller pGL3-3A5–433 and pGL3–3A5–862 luciferase constructs (Fig. 4). These data demonstrated that the CYP3A4 57-bp insertion has direct functional consequences on the low expression of CYP3A4 in human lung cells.

View larger version (24K): [in this window] [in a new window]   Fig. 4. Insertion of the CYP3A4 57-bp insertion into the concomitant region of the CYP3A5 promoter reduces expression in the human lung adenocarcinoma A549 cell line. Salient features of the CYP3A5/luciferase reporter constructs are depicted to the left of the bar graph with individual numbers representing relative distance from the published transcriptional start site for CYP3A5. Bar graphs represent the relative -fold activity over the promoterless pGL3-Basic vector. Data represent means ± S.D. from quadruplicate experiments that have been normalized for transfection efficiency and protein content.

View larger version (17K): [in this window] [in a new window]   Fig. 5. Transcriptional activity of pGL3–3A4 constructs, with and without the CYP3A4 57-bp region, transiently transfected into human lung A549 cells as measured using the Dual Luciferase reporter system (Roche). When the 57-bp region had been completely excised from the CYP3A4 promoter, no increase in luciferase activity was observed (64-bp Knockout and Spacer; p > 0.1). However, if the C/EBP binding motif (–121 to –130), shown to be essential for trans-activation in HepG2 liver cells (Rodríguez-Antona et al., 2003) was maintained (SacII construct), expression levels were increased by approximately 4-fold compared with the complete CYP3A4 construct (p < 0.05 significantly different compared with wild-type pGL3–3A4 expression levels). Expression levels were normalized to activity of the cotransfected R. reniformis luciferase vector and are expressed as -fold luciferase activity divided by the promoterless pGL3-Basic plasmid. Error bars represent the S.D. of four replicates that were normalized for transfection efficiency and protein content. The experiment was repeated twice with similar results.

View larger version (22K): [in this window] [in a new window]   Fig. 6. Transcriptional activity of wild-type pGL3–3A4 and pGL3–3A5 constructs, with and without the CYP3A4 57-bp region, transiently transfected into primary cell cultures. The largest (13 kb) CYP3A4 promoter (a kind gift from Dr. Tetsuya Kamataki, Sapporo, Japan), which presumably contained all viable transcriptional elements, did not enhance luciferase expression in NHBE cells. The SacII deletion construct (57-bp knockout) significantly (p < 0.05) increased luciferase activity over the intact CYP3A4 construct in NHBE cells, but this deletion did not alter promotion in primary hepatocytes, both of which showed activity above that of the control vector, pGL-3 Basic. Insertion of the CYP3A4 57-bp region (57 bp inserted) into the wild-type 3-kb pGL3–3A5 construct significantly decreased (p < 0.05) luciferase activity in transfected NHBE cells but not in primary hepatocytes. Data represent means ± S.D. from nine replicates that were normalized for transfection efficiency.

View larger version (6K): [in this window] [in a new window]   Fig. 7. Quantitative real-time PCR analysis of CYP3A4 and CYP3A5 transcripts in primary and immortalized human lung cells. Total RNA was extracted from 106 NHBE or A549 cells, and 1 µg of total RNA was used to synthesize first-strand cDNA and then analyzed by quantitative real-time PCR. Copy numbers were determined by extrapolation from a standard curve that was constructed with known amounts of CYP3A4 or CYP3A5 plasmid DNA, and standardized to actin levels. N.D., not detected (the mRNA amounts for CYP3A4 were below detectable levels in both lung cell types).

Analysis of Factor Binding to the CYP3A4 57-bp Region by EMSA. To demonstrate the specificity of nuclear factor(s) binding to the putative cis-element(s) identified by the luciferase experiments, EMSA experiments were conducted using ³²P-labeled double-stranded oligonucleotides corresponding to the wild-type sequence for the CYP3A4 57-bp insertion (–71 to –127), shown in Table 3, and nuclear extracts from A549 lung cells and whole-lung nuclear extracts. When radiolabeled wild-type CYP3A4 57-bp probe was incubated with nuclear extract from either source, two sequence-specific DNA-protein complexes were observed (Fig. 8; data from whole-lung extracts not shown). Binding was inhibited by the addition of a 100-fold molar excess of unlabeled wild-type oligonucleotide. In contrast, complex formation was not inhibited by the addition of 175-fold molar excess of unlabeled competitors harboring consensus binding sequences for AP1, NFB, Sp1, OCT1, and the CYP2F1-LSF1 motif (Table 3, Fig. 8). These results demonstrate the specific binding of nuclear factors within the human lung and human lung adenocarcinoma A549 cells to the CYP3A4 57-bp insertion. Furthermore, because consensus oligonucleotides harbor the highest affinity binding sites for their respective nuclear factors, these data also strongly suggest that the trans-acting element(s) is not AP1, NFB, Sp1, OCT1, or CYP2F1-LSF1.

View larger version (45K): [in this window] [in a new window]   Fig. 8. Competitive EMSA assays of serial mutations of the CYP3A4 57-bp promoter element. Putative E-boxes within the CYP3A4 57-bp insertion are capable of binding in a sequence-specific manner with human lung adenocarcinoma A549 cell nuclear proteins. The 32P-labeled CYP3A4 57-bp probe was incubated with nuclear extracts (6 µg of total protein) prepared from A549 lung cells with various competitors. When competitors were included in the incubations, they were added 10 min before the addition of the 32P-labeled CYP3A4 57-bp probe. Therefore, the ability to compete for, rather than displace, the 32P-labeled CYP3A4 57-bp probe for A549 nuclear proteins was assayed. The two major protein/DNA complexes are marked with arrows. Competitive oligonucleotides were created by mutating five consecutive base pairs at a time and are labeled with numbers above the lanes that correspond to the appropriate mutated sequences shown in Table 3. The CCAAT-box and the two putative E-box motifs are labeled above the 57-bp sequence, and the consensus sequences are underlined. The five base pairs that were mutated for each competitive incubation are identified above each lane and demarcated with vertical lines. Notations are as follows: probe, 32P-labeled CYP3A4 57-bp probe only; A549, 32P-labeled CYP3A4 57-bp probe incubated with A549 nuclear extracts; Cold, 100-fold molar excess of unlabeled CYP3A4 57 bp; and NFB, 175-fold molar excess of NFB consensus oligonucleotide.

View larger version (77K): [in this window] [in a new window]   Fig. 9. EMSA analysis of lung and liver cellular nuclear extract binding to the 57-bp motif. Nuclear extracts (6 µg) from A549 (lanes 1–3) or HepG2 (lanes 4–6) cells were mixed with 0.005 to 0.01 pmol of 32P-labeled 57-bp oligonucleotide probe, and the protein/DNA complexes and unbound probes were separated by nondenaturing gel electrophoresis using 4% polyacrylamide gels and detected by autoradiography as described under Materials and Methods. Lanes 2 and 5 are samples from incubations that contained a 100-fold excess of unlabeled specific 57-bp oligonucleotide, incubated for 15 min with nuclear extract before the addition of labeled probe. Lanes 3 and 6 are samples from incubations that contained a 100-fold excess of nonspecific oligonucleotide, incubated for 15 min with nuclear extract before the addition of labeled probe. Right-facing arrows point to the two major protein/DNA complexes of lung cell proteins with the 57-bp probe. Left-facing arrows point to the two major protein/DNA complexes of liver cell proteins with the 57-bp probe. Binding of liver nuclear extracts (lanes 4–6) to the probe was considerably weaker than lung extracts, because most of the radioactive probe stayed with the dye front at the bottom of the gel.

In an attempt to identify the trans-acting factor involved in the active repression of the CYP3A4 gene, supershift EMSA experiments were conducted using antibodies generated against the known double-E-box binding factors (SIP1, HEB, and EF1), as well as single E-box binding proteins (e.g., E2A, Myc, MyoD, and E47). None of the antibodies retarded the mobility of the protein/DNA complex (data not shown), indicating that these are not the trans-acting factors in question. Multiple attempts to purify and identify the lung nuclear factor from lung cells or tissues by DNA affinity chromatography were not successful. When combined, these data suggest that an uncharacterized human lung specific transcription factor(s) binds specifically to a double E-box motif within the CYP3A4 57-bp insertion and actively represses CYP3A4 expression in the human lung.

	Discussion

Top Abstract Materials and Methods Results Discussion References

 
Comparisons of the promoter regions of CYP3A4 and CYP3A5 show that they share more than 90% homology within the first 1 kb of upstream regulatory sequence and yet show striking differences in expression among individuals, races, developmental stages, tissues, and cell types. The regulation of the CYP3A cassette is also very complex with regard to its response to changes in physiological conditions.

Most studies on the CYP3A subfamily thus far have concentrated on the expression of these isoforms in the liver, where polymorphisms have been shown to be important. An intriguing ramification of previous studies (Kuehl et al., 2001) is that 75% of a population (i.e., those homozygous for CYP3A5*3) would express an extremely limited amount of CYP3A5 within their extrahepatic tissues, thus increasing the metabolic burden upon the CYP3A4 enzyme. Because CYP3A5 is probably the major CYP3A isoform expressed within human lung tissue (Anttila et al., 1997; Raunio et al., 1999), the overall metabolism of 3A substrates in lung cells would be drastically reduced or otherwise altered toward alternative metabolic pathways. Although procarcinogens metabolized by the CYP3A enzymes would not become bioactivated, drugs designed to target the lung might exhibit greater adverse side effects (e.g., long-term use of inhaled glucocorticoids might lead to superinfections) and carcinogens/toxins would exhibit extended half-lives. CYP3A individual genotypes have been correlated to lower incidence of lung and other types of cancer (Yeh et al., 2003; Keshava et al., 2004).

As expected in A549 cells, CYP3A5-Luc constructs exhibited more than 20-fold higher luciferase activity than CYP3A4-Luc constructs. This dramatic difference in basal expression among these constructs becomes quite apparent within nested deletions containing only the first 200 base pairs of the 5'-flanking region of these two genes (Fig. 1). Luciferase constructs that contained the 3A4 promoter gradually increased in activity as nested-deletions reached –80 but then dropped dramatically thereafter. In contrast, CYP3A5 constructs rapidly reached maximal levels with the first –208 base pairs of the promoter; a pattern of expression that closely matches previous studies in HepG2 liver cells (Iwano et al., 2001). The experiments with mutated CYP3A5 promoter constructs showed that the basic transcriptional element (BTE) and especially the CCAAT box are important for CYP3A transcriptional activation in A549 cells (Fig. 4), a finding that is surprisingly similar to the mechanisms of CYP3A transcriptional regulation in HepG2 cells.

We can conjecture about the genetic regulatory mechanisms in A549 lung cells by comparing the mechanisms governing expression of CYP3A4 and CYP3A5 in HepG2 cells. For instance, the basal regulation of CYP3A5 is governed primarily by the cooperative effects of NF-Y and specificity protein (Sp) family members binding to a CCAAT-box (–68 to –78) and the BTE (–46 to –67) in the proximal promoter of the CYP3A5 gene (Iwano et al., 2001). Analogous motifs are identifiable within the proximal promoter of the CYP3A4 gene. However, compared with the CYP3A5 promoter, the CYP3A4 CCAAT-box (–62 to –66) has a single mutation in the core binding motif, and the BTE (–36 to –57) also shows a difference in two consecutive nucleotides. Both perturbations confer markedly less expression activity than their respective CYP3A5 counterparts when tested in HepG2 cell lines (data not shown), and this is also the case in A549 cells (Fig. 1). In contrast, the transcriptional regulation of the CYP3A4 gene in HepG2 cell lines is controlled by a complex circuit of transcription factors that bind to motifs more distal to the related region of CYP3A5. Instead of being governed by the aforementioned regions, the basal expression of the CYP3A4 gene in HepG2 cell lines is only partially controlled by the binding of C/EBP and HNF-3 to a proximal ER-6 motif (–152 to –169) (Rodríguez-Antona et al., 2003); expression depends rather upon HNF-1, HNF-4, AP-1, and USF1 binding to a distal region (–10.9 to –11.4 kb), which was referred to as the "constitutive liver enhancer module of CYP3A4 (CLEM4)" (Matsumura et al., 2004). However, this is not the case in transient transfections of these same constructs in A549 cells, because the inclusion of the CLEM4 element did not increase expression (data not shown). What is important about these findings is that the underlying mechanisms responsible for the basal expression of these two genes are encoded within the sequences of the CYP3A4 and CYP3A5 promoters and that, although similar, the expression patterns within A549 lung cells are distinct from that of HepG2 liver cells.

In a comparison of the reporter activities of CYP3A4-Luc and CYP3A5-Luc in the human lung A549 cell line, dramatic differences in promoter-driven expression were observed within the first –150 and –208 base pairs for CYP3A4 and CYP3A5, respectively. The most obvious difference between these two promoter regions is a 57-base pair insertion in the CYP3A4 promoter (–71 to –127; Fig. 2). More importantly, when introduced into the same region of the CYP3A5 promoter, this insertion reduced the expression of CYP3A5-promoter constructs by roughly 50% (Fig. 4). It is important to note that the insertion of this 57-bp region disrupts the regulatory CCAAT-box that is vital to drive basal expression in HepG2 cells (Iwano et al., 2001) and in A549s (this study). However, this insertion also includes a CCAAT-box at the 5' end that evolutionarily conserves this motif almost perfectly, suggesting either that another protein is interacting with this introduced region or that distance from neighboring motifs is fundamental to maintaining this expression mechanism.

Two conclusions can be drawn from these observations: 1) differences between the proximal cis-elements of the CYP3A4 and CYP3A5 promoters reduce the expression of the CYP3A4 gene in human lung cells; and 2) control of the lung-specific differences in expression observed in previous studies (Kivistö et al., 1996; Anttila et al., 1997; Hukkanen et al., 2000, 2002) is a combination of these differences and the insertion of a repressor motif in the form of a 57-bp region containing two E-box motifs directly between the positive cis-elements and the transcriptional initiation site of the CYP3A4 gene. A schematic diagram illustrating the mechanism by which the CYP3A4 gene could be repressed through this region is depicted in Fig. 10A.

View larger version (30K): [in this window] [in a new window]   Fig. 10. A, schematic representation of the LSF2 mechanism of repressing CYP3A4 in the human lung. Actions of both the proximal (Rodríguez-Antona et al., 2003) and the distal (Matsumura et al., 2004) elements converge upon an ER-6 motif and a CCAAT-box to recruit the RNA Pol II transcriptional machinery to the CYP3A4 transcriptional start site. By binding to the 57-bp insertion located immediately downstream of this region (–71 to –128), LSF2 disrupts these events through one or a combination of actions: 1) altering the recognition of the CCAAT-box by these factors, 2) physically inhibiting the assembly of the complex, or 3) aiding the condensation of chromatin to mask the cis-elements. B, the sequence highlighting the protein/DNA interactions of the CYP3A4 promoter region. Highlighted are the 57-bp insertion (red); the HNF-3, ER-6, C/EBP, and TATA transcription factor binding sites (gray); the E-box sites (underlined); and the core nucleotides for binding of A549 nuclear extracts to the 57-bp probe from EMSA experiments (Fig. 8) or the core nucleotides for binding of HepG2 nuclear proteins (yellow). The transcription start site (+1) and various cloning sites are identified as well.

The codependence on both E-box motifs could also be associated with cooperative binding of multiple transcription factors to this region. Although the detailed interactions between these motifs have not been rigorously established, these data strongly suggest that the active repression of CYP3A4, and not the lack of a transcriptional activator, is the operative mechanism in the human lung controlling the differential expression of these two very important genes. We believe that this is the first case in which a cytochrome P4503A gene is actively "silenced" in a tissue-specific manner. We have termed this trans-repressor protein "lung-specific factor-2" (LSF2). Although the precise molecular mechanisms involved in LSF2 repression of CYP3A4 in the human lung remain unclear, we hypothesize that LSF2 actively represses the expression of the CYP3A4 gene in human lung cells.

The underlying mechanism could exhibit its effects through three different mechanisms (Johnson, 1995): 1) interfering directly with other cis-acting elements through competition for their cognitive cis-activation domain (Sekido et al., 1997), 2) physically blocking the guided assembly (or initiation) of the transcription machinery (Coumoul et al., 2002), or 3) having some chromatin remodeling capacity (Ringrose and Paro, 2004). A number of recent observations, both genetic and biochemical, suggest a different mechanism of P450 repression might be active in HepG2 liver cells. The PXR-SMRT mechanism of CYP3A4 repression in the absence of ligand (Johnson et al., 2006), involves both sequestration of trans-acting elements and direct recruitment of histone deacetylases involved in chromatin condensation. Supershift EMSA experiments using polyclonal goat antibodies raised against SMRT did not change the mobility of the A549 nuclear protein/DNA band. Therefore, it is highly unlikely that SMRT is the trans-element involved in the pulmonary repression mechanism.

Given that the 57-bp insertion is situated on the 3' end of the promoter, the simplest mechanism would be that LSF2 would hinder the formation of the initiation complex either through steric hindrance or through the recruitment of chromatin remodeling complexes. Active repression by LSF2 is an attractive mechanism because it can override a number of different signal cascades with the same protein, regardless of which additional transcription factors might be actively transcribed within the cell.

	Acknowledgements

 
We thank Cassandra Deering for her efforts in creating some of the luciferase-reporter constructs within these studies, Dr. Ronald Hines and Dr. Brian Carr for assistance with cell culture and luciferase assay techniques, and Dr. Tetsuya Kamataki (Hokkaido University, Sapporo, Japan) for providing the wild-type pGL3-CYP3A4-13 kb luciferase-reporter construct.

	Footnotes

 
This research was supported by the National Heart, Lung, and Blood Institute (grant HL60143), a National Institutes of Health Minority Graduate Research Assistance Supplement, and the Colgate-Palmolive/SOT Award for Student Research Training in Alternative Methods.

Article, publication date, and citation information can be found at http://molpharm.aspetjournals.org.

doi:10.1124/mol.106.033795.

ABBREVIATIONS: P450, cytochrome P450; PCR, polymerase chain reaction; EMSA, electrophoretic mobility shift assay; kb, kilobase pair(s); bp, base pair(s); NHBE, normal human bronchial epithelial cells; AP1, activator protein-1; NFB, nuclear factor B; C/EBP, CCAAT/enhancer-binding protein; ANOVA, analysis of variance; LSD, least significant difference; BTE, basic transcriptional element; NF-Y, nuclear factor Y; Sp1, specificity protein 1; LSF, lung-specific factor; A549, human lung adenocarcinoma A549 cells.

Address correspondence to: Dr. Garold S. Yost, Department of Pharmacology and Toxicology, University of Utah, 30 South 2000 East, Room 201, Salt Lake City, UT 84112-5820. E-mail: gyost{at}pharm.utah.edu

	References

Top Abstract Materials and Methods Results Discussion References

 
Anttila S, Hukkanen J, Hakkola J, Stjernvall T, Beaune P, Edwards RJ, Boobis AR, Pelkonen O, and Raunio H (1997) Expression and localization of CYP3A4 and CYP3A5 in human lung. Am J Respir Cell Mol Biol 16: 242–249.[Abstract]

Bombail V, Taylor K, Gibson GG, and Plant N (2004) Role of Sp1, C/EBPalpha, HNF3, and PXR in the basal- and xenobiotic-mediated regulation of the CYP3A4 gene. Drug Metab Dispos 32: 525–535.[Abstract/Free Full Text]

Burk O, Tegude H, Koch I, Hustert E, Wolbold R, Glaeser H, Klein K, Fromm MF, Nuessler AK, Neuhaus P, et al. (2002) Molecular mechanisms of polymorphic CYP3A7 expression in adult human liver and intestine. J Biol Chem 277: 24280–24288.[Abstract/Free Full Text]

Carr BA, Wan J, Hines RN, and Yost GS (2003) Characterization of the human lung CYP2F1 gene and identification of a novel lung-specific binding motif. J Biol Chem 278: 15473–15483.[Abstract/Free Full Text]

Chen JD and Evans RM (1995) A transcriptional co-repressor that interacts with nuclear hormone receptors. Nature 377: 454–457.[CrossRef][Medline]

Chen JD and Li H (1998) Coactivation and corepression in transcriptional regulation by steroid/nuclear hormone receptors. Crit Rev Eukaryot Gene Expr 8: 169–190.[Medline]

Coumoul X, Diry M, and Barouki R (2002) PXR-dependent induction of human CYP3A4 gene expression by organochlorine pesticides. Biochem Pharmacol 64: 1513–1519.[CrossRef][Medline]

Ding X and Kaminsky LS (2003) Human extrahepatic cytochromes P450: function in xenobiotic metabolism and tissue-selective chemical toxicity in the respiratory and gastrointestinal tracts. Annu Rev Pharmacol Toxicol 43: 149–173.[CrossRef][Medline]

Floyd MD, Gervasini G, Masica AL, Mayo G, George AL, Jr., Bhat K, Kim RB, and Wilkinson GR (2003) Genotype-phenotype associations for common CYP3A4 and CYP3A5 variants in the basal and induced metabolism of midazolam in European- and African-American men and women. Pharmacogenetics 13: 595–606.[CrossRef][Medline]

Gellner K, Eiselt R, Hustert E, Arnold H, Koch I, Haberl M, Deglmann CJ, Burk O, Buntefuss D, Escher S, et al. (2001) Genomic organization of the human CYP3A locus: identification of a new, inducible CYP3A gene. Pharmacogenetics 11: 111–121.[CrossRef][Medline]

Gibson GG, Plant NJ, Swales KE, Ayrton A, and El-Sankary W (2002) Receptor-dependent transcriptional activation of cytochrome P4503A genes: induction mechanisms, species differences and interindividual variation in man. Xenobiotica 32: 165–206.[CrossRef][Medline]

Goodwin B, Hodgson E, D'Costa DJ, Robertson GR, and Liddle C (2002a) Transcriptional regulation of the human CYP3A4 gene by the constitutive androstane receptor. Mol Pharmacol 62: 359–365.[Abstract/Free Full Text]

Goodwin B, Redinbo MR, and Kliewer SA (2002b) Regulation of cyp3a gene transcription by the pregnane x receptor. Annu Rev Pharmacol Toxicol 42: 1–23.[CrossRef][Medline]

Guengerich FP (1999) Cytochrome P-450 3A4: regulation and role in drug metabolism. Annu Rev Pharmacol Toxicol 39: 1–17.[CrossRef][Medline]

Guengerich FP and Shimada T (1998) Activation of procarcinogens by human cytochrome P450 enzymes. Mutat Res 400: 201–213.[Medline]

Hukkanen J, Hakkola J, Anttila S, Piipari R, Karjalainen A, Pelkonen O, and Raunio H (1997) Detection of mRNA encoding xenobiotic-metabolizing cytochrome P450s in human bronchoalveolar macrophages and peripheral blood lymphocytes. Mol Carcinog 20: 224–230.[CrossRef][Medline]

Hukkanen J, Lassila A, Paivarinta K, Valanne S, Sarpo S, Hakkola J, Pelkonen O, and Raunio H (2000) Induction and regulation of xenobiotic-metabolizing cytochrome P450s in the human A549 lung adenocarcinoma cell line. Am J Respir Cell Mol Biol 22: 360–366.[Abstract/Free Full Text]

Hukkanen J, Pelkonen O, Hakkola J, and Raunio H (2002) Expression and regulation of xenobiotic-metabolizing cytochrome P450 (CYP) enzymes in human lung. Crit Rev Toxicol 32: 391–411.[CrossRef][Medline]

Hukkanen J, Vaisanen T, Lassila A, Piipari R, Anttila S, Pelkonen O, Raunio H, and Hakkola J (2003) Regulation of CYP3A5 by glucocorticoids and cigarette smoke in human lung-derived cells. J Pharmacol Exp Ther 304: 745–752.[Abstract/Free Full Text]

Iwano S, Saito T, Takahashi Y, Fujita K, and Kamataki T (2001) Cooperative regulation of CYP3A5 gene transcription by NF-Y and Sp family members. Biochem Biophys Res Commun 286: 55–60.[CrossRef][Medline]

Johnson AD (1995) The price of repression. Cell 81: 655–658.[CrossRef][Medline]

Johnson DR, Li C-W, Chen L-Y, Ghosh JC, and Chen JD (2006) Regulation and binding of pregnane X receptor by nuclear receptor corepressor silencing mediator of retinoid and thyroid hormone receptors (SMRT) 10. Mol Pharmacol 69: 99–108.[Abstract/Free Full Text]

Jover R, Bort R, Gomez-Lechon MJ, and Castell JV (2002) Down-regulation of human CYP3A4 by the inflammatory signal interleukin-6: molecular mechanism and transcription factors involved. FASEB J 16: 1799–1801.[Abstract/Free Full Text]

Keshava C, McCanlies EC, and Weston A (2004) CYP3A4 polymorphisms—potential risk factors for breast and prostate cancer: a HuGE review. Am J Epidemiol 160: 825–841.[Abstract/Free Full Text]

Kivistö KT, Griese EU, Fritz P, Linder A, Hakkola J, Raunio H, Beaune P, and Kroemer HK (1996) Expression of cytochrome P 450 3A enzymes in human lung: a combined RT-PCR and immunohistochemical analysis of normal tissue and lung tumours. Naunyn Schmiedebergs Arch Pharmacol 353: 207–212.[Medline]

Koch I, Weil R, Wolbold R, Brockmoller J, Hustert E, Burk O, Nuessler A, Neuhaus P, Eichelbaum M, Zanger U, et al. (2002) Interindividual variability and tissue-specificity in the expression of cytochrome P450 3A mRNA. Drug Metab Dispos 30: 1108–1114.[Abstract/Free Full Text]

Kolars JC, Lown KS, Schmiedlin-Ren P, Ghosh M, Fang C, Wrighton SA, Merion RM, and Watkins PB (1994) CYP3A gene expression in human gut epithelium. Pharmacogenetics 4: 247–259.[Medline]

Kuehl P, Zhang J, Lin Y, Lamba J, Assem M, Schuetz J, Watkins PB, Daly A, Wrighton SA, Hall SD, et al. (2001) Sequence diversity in CYP3A promoters and characterization of the genetic basis of polymorphic CYP3A5 expression. Nat Genet 27: 383–391.[CrossRef][Medline]

Lamba JK, Lin YS, Schuetz EG, and Thummel KE (2002) Genetic contribution to variable human CYP3A-mediated metabolism. Adv Drug Deliv Rev 54: 1271–1294.[CrossRef][Medline]

Matsumura K, Saito T, Takahashi Y, Ozeki T, Kiyotani K, Fujieda M, Yamazaki H, Kunitoh H, and Kamataki T (2004) Identification of a novel polymorphic enhancer of the human CYP3A4 gene. Mol Pharmacol 65: 326–334.[Abstract/Free Full Text]

Mollerup S, Ovrebo S, and Haugen A (2001) Lung carcinogenesis: resveratrol modulates the expression of genes involved in the metabolism of PAH in human bronchial epithelial cells. Int J Cancer 92: 18–25.[CrossRef][Medline]

Nelson DR, Koymans L, Kamataki T, Stegeman JJ, Feyereisen R, Waxman DJ, Waterman MR, Gotoh O, Coon MJ, Estabrook RW, et al. (1996) P450 superfamily: update on new sequences, gene mapping, accession numbers and nomenclature. Pharmacogenetics 6: 1–42.[Medline]

Piipari R, Savela K, Nurminen T, Hukkanen J, Raunio H, Hakkola J, Mantyla T, Beaune P, Edwards RJ, Boobis AR, et a