Abstract
The aryl hydrocarbon receptor (AhR) belongs to the basic helix-loop-helix/periodicity/AhR nuclear translocator/simple-minded (Per-Arnt-Sim) family of transcription factors that regulate critical functions during development and tissue homeostasis. Within this family, the AhR is the only member conditionally activated in response to ligand binding, typified by 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD). We recently demonstrated that the AhR interacts with the retinoblastoma protein (pRb). This report presents evidence that a LXCXE motif in the AhR protein confers pRb binding, which is necessary for maximal TCDD induced G1 arrest in rat 5L hepatoma cells. The data support a mechanism whereby pRb seems to regulate G1 cell cycle progression distinct from the direct repression of E2F-mediated transcription. Furthermore, the results indicate that the AhR-pRb interaction regulates TCDD induction of CYP1A1, suggesting that pRb may be a general AhR coactivator.
The aryl hydrocarbon Receptor (AhR) belongs to a growing family of transcription factors characterized by a basic helix-loop-helix (bHLH) DNA-binding domain and a PAS homology domain involved in protein dimerization (Schmidt and Bradfield, 1996). Interest in the AhR dates back to the mid-1970s because of its involvement in 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) toxicity (Poland and Knutson, 1982). Among the various bHLH/PAS members, the AhR is the only protein conditionally activated in response to ligand binding. AhR ligands include the halogenated aromatic hydrocarbons typified by TCDD, polycyclic aromatic hydrocarbon (e.g., polychlorinated biphenyls), and related nonphysiological xenobiotics (e.g., heterocyclic amines). Several natural compounds are known to bind the AhR (Bjeldanes et al., 1991; Sinal and Bend, 1997; Schaldach et al., 1999), but whether they constitute physiologically relevant endogenous ligands requires further study. The structural diversity among the ligands, however, attests to the promiscuous nature of AhR ligand binding.
The unliganded AhR is a cytosolic protein bound to the chaperone 90-kDa heat shock protein 90 (hsp90) and an immunophilin-like molecule (reviewed by Whitlock, 1999). Upon ligand binding, the AhR translocates into the nucleus concomitant with dissociation from hsp90 and binds to regulatory cis elements in partnership with the Arnt protein, itself a member of the bHLH/PAS family. TCDD-induced AhR/Arnt DNA binding results in transcriptional regulation of target genes, including several genes encoding for the drug metabolizing enzymesCYP1A1, 1A2, 1B1 and glutathioneS-transferase Ya (Poland and Knutson, 1982; Whitlock, 1999). Two lines of evidence implicate the AhR in physiological processes germane to normal development and tissue homeostasis. First, TCDD exposure in animals induces teratogenesis, immunosuppression, reproductive defects, and tumor promotion. Second, AhR-null mice are resistant to TCDD toxicity, displaying instead certain abnormal phenotypes, including liver defects, a hyperproliferative phenotype in organs such as the stomach and vasculature, and decreases in circulating T and B cells (Schmidt et al., 1996; Gonzalez and Fernandez-Salguero, 1998). These observations suggest that the AhR influences normal cell proliferation and differentiation. Studies on growth rates of AhR-positive and -negative mouse and rat hepatoma cell lines revealed that the AhR modulates G1 cell cycle progression (Ma and Whitlock, 1996; Weiss et al., 1996). The pronounced TCDD-induced G1 arrest in the rat 5L hepatoma cell line involves an AhR-mediated induction of the cyclin-dependent kinase inhibitor, p27Kip1(Kolluri et al., 1999). Likewise, TCDD is known to inhibit DNA synthesis in primary hepatocytes and hepatocyte proliferation in vivo after partial hepatectomy (Bauman et al., 1995; Hushka and Greenlee, 1995).
We recently showed that the AhR and Retinoblastoma tumor suppressor protein (pRb) interact directly (Ge and Elferink, 1998). Mechanistically, hypophosphorylated pRb regulates transition through the G1/S restriction point by binding to and repressing several members of the E2F transcription factor family, thus preventing expression of genes required for S phase (Dyson, 1998). pRb and the related “pocket” proteins p107 and p130, contain a conserved A/B pocket domain that interacts with proteins harboring an LXCXE motif. These include the viral oncoproteins E1A and the simian virus 40 large T-antigen. Binding of pRb to E2F requires additional C-terminal residues composing the “large A/B pocket” (Qin et al., 1992). Hyperphosphorylation of pRb by the sequential action of the cyclin-dependent kinases 4/6 and 2 (CDKs) respectively, inactivates pRb culminating in entry into S phase (Sherr and Roberts, 1999). D-type cyclins control CDK4/6 activity and cyclins A and E regulate CDK2. In turn, CDK inhibitors comprise the INK4 proteins (p15, p16, p18, and p19) that bind to and inhibit the cyclin D-associated CDKs, and the more promiscuous Cip/Kip inhibitors (p21Cip1, p27Kip1, and p57Kip2), which suppress most cyclin-CDK complexes, particularly cyclin E-CDK2 and cyclin A-CDK2 (Sherr and Roberts, 1999).
The documented importance of pRb in controlling E2F-mediated transcription notwithstanding, pRb interacts with many other proteins, raising the possibility that its action as a tumor suppressor depends in part on these interactions. Previously, we speculated that the LXCXE motif in the AhR contributes to the AhR-pRb interaction (Ge and Elferink, 1998). Yeast two-hybrid experiments confirm that this motif is indeed involved in the AhR-pRb interaction, which in subsequent functional studies is shown to be necessary for maximal TCDD-induced G1 arrest in rat 5L hepatoma cells.CYP1A1 expression studies suggest that the AhR-pRb interaction is required for maximal AhR transcriptional activity implying that pRb may be an AhR coactivator. The ramifications of these observations are considered in the context of cell cycle control.
Experimental Procedures
Materials.
Restriction endonucleases and other DNA modifying enzymes (T4 DNA ligase, calf intestinal alkaline phosphatase) were purchased from Life Technologies Inc. (Gaithersburg, MD) and New England Biolabs (Beverly, MA). The MATCHMAKER LexA Two-Hybrid System used and yeast culture media were from CLONTECH (Palo Alto, CA). TheTaq and KlenTaq DNA polymerases were obtained from Qiagen (Santa Clarita, CA) and Sigma (St. Louis, MO), respectively. The pRb antibody (G3-245) was from Pierce (Rockford, IL). TCDD was from the National Cancer Institute Chemical Carcinogen Reference Standard Repository. [3H]TCDD was from the Cambridge Isotope Laboratories (Andover, MA). Western-Star and Galacton-Star kits were purchased from Tropix (Bedford, MA). Radioactive compounds were acquired from Amersham Pharmacia Biotech (Piscataway, NJ). Protein-G coupled Sepharose resin, custom synthesized oligonucleotides, LipofectAMINE Plus and G418 were from Life Technologies Inc. The pTRE and pTet-Off vectors were from CLONTECH and the hygromycin B was from Sigma. The hydroxylapatite was obtained from Bio-Rad (Richmond, CA).
Oligonucleotides.
Sequences of the oligonucleotides used are presented in Table 1. The underlined nucleotide doublets denote the mutated bases introduced to generate the alanine-substituted constructs.
Yeast Plasmid Construction and the Two-Hybrid Assay.
All AhR LexA-fusion proteins were constructed using pLexA vector (MATCHMAKER system) and were derived from a series of yeast ‘bait’ constructs described elsewhere (Ge and Elferink, 1998). Creation of the bHLH deletion constructs took advantage of two conveniently locatedEcoRI sites, one in the pLexA vector polylinker and the second at position 346 in the human AhR coding sequence.EcoRI digestion of pYAhRFL, pYAhR1–672, pYAhR1–589 and pYAhR1–528, followed by religation of the arms recovered the reading frame and removed a 381-bp fragment encoding the bHLH domain of the human AhR, to yield the constructs pYAhR117–853, pYAhR117–672, pYAhR117–589 and pYAhR117–528. Generation of the alanine-substituted constructs involved subcloning the NotI inserts from pYAhRFL, pYAhR1–672, pYAhR1–589, and pYAhR1–528 into a modified pLexA vector engineered to remove the vector's EcoRI andBamHI sites by EcoRI/BamHI digestion, end-filling with Klenow and religation. Site-directed mutagenesis involved a two-stage PCR protocol using KlenTaq polymerase in all the PCR intended for DNA cloning. The first round of PCR used the primer pairs oligonucleotides 3:2, 5:2, 7:2, 4:1, 6:1, and 8:1 as six individual PCR reactions (25 cycles). The second round of PCR combined DNA from reactions 3:2 and 4:1; 5:2 and 6:1; 7:2 and 8:1, and used oligonucleotides 1 and 2 to amplify the PCR products, resulting in 1052-bp PCR products that were digested with EcoRI andBamHI, generating 925-bp fragments encoding the L331A, C333A, or E335A alanine substitutions. These fragments were used to replace the corresponding DNA fragment in pYAhRFL, pYAhR1–672, and pYAhR1–589 to generate the series of alanine substituted constructs. The pRb prey construct (pB42ADpRb) encodes amino acids 374 to 928 of the human pRb (Ge and Elferink, 1998). Constructs were checked by DNA sequencing using the dideoxynucleotide chain termination method. Yeast cells (leucine auxotrophic strain EGY48) were transformed using the LiAc method as described in the MATCHMAKER manual (CLONTECH). The two-hybrid assay uses two reporters (LEU2 andlacZ), under the control of LexA operators.
Expression Plasmids.
A full-length, wild-type rat AhR cDNA was generated from a cDNA clone (Elferink and Whitlock, 1994) by PCR amplification using oligonucleotides 13 and 14, and subcloned into theXbaI site of pTRE (CLONTECH) to yield pTRErAhR. Generation of the E333A mutation involved PCR amplification of pTRErAhR with primer pairs, oligonucleotides 9:12 and 10:11 (25 cycles), followed by a second round of PCR with oligonucleotides s11 and 12. The 654-bp PCR product was digested with Eco47III and SpeI, and the resultant 533 fragment cloned into pTRErAhR replacing theEco47III/SpeI fragment encompassing the LXCXE motif generating pTRErAhRE333A. Successful cloning was confirmed by sequencing. Synthesis of the recombinant adenoviral clone involved subcloning the rAhR sequence from pTRErAhR into the XbaI site of the pAdTrack-CMV shuttle vector (He et al., 1998). Recombination and analysis of the recombinants was performed as described by He et al. (1998). Viral stocks were prepared by recovering the viruses from infected human embryonic kidney 293 packaging cells, purified by CsCl banding and frozen at −80°C as single-use aliquots. The virus AdrAhRFL expresses the full-length rat AhR and the green fluorescent protein (GFP) from separate but identical CMV promoters. Viral titers were determined as GFP expression forming units/ml on 293 cells. We found that the number of expression forming units in BP8 cells is ≈200-fold lower than that of 293 cells, thus requiring a higher multiplicity of infection for AhR expression in BP8 cells.
Cell Culture, Transfections and Infections.
Wild-type rat hepatoma 5L cells and AhR-defective BP8 variants were grown as monolayers in Dulbecco's modified Eagle's medium containing 10% fetal bovine serum, 100 U/ml penicillin, and 100 μg/ml streptomycin in 5% CO2 atmosphere at 37°C. All transfections were performed using LipofectAMINE Plus (Life Technologies) in accordance with the manufacturer's recommendations. For the stable transfections BP8 cells were transfected with the plasmid pTet-Off (CLONTECH) and grown in medium containing 500 μg/ml G418 for 14 days. A stable transfectant expressing the Tet repressor was clone purified before transfection with pTRErAhR or pTRErAhRE333A, and pTK-Hyg at a 10:1 ratio, respectively. Cells resistant to G418 and hygromycin B (100 μg/ml) selection were clone purified and screened for inducible AhR expression after removal of 1 μg/ml doxycycline. The BP8-WT and BP8-E333A lines used in this study express at a physiological level, the wild-type and alanine-substituted AhRs, respectively. Although the reason remains unclear, repeated attempts to isolate inducible cell lines proved unsuccessful. The lines are maintained in medium containing 500 μg/ml G418 and 100 μg/ml hygromycin B. Cells infected with the virus AdrAhRFL were seeded at 106 cells per 100-mm dish and cultured overnight before infection at a multiplicity of infection of 100:1. TCDD treatment was initiated 24 to 36 h after infection.
Flow Cytometry and Microscopy.
Cells were trypsinized, washed twice in PBS containing 1 g/l glucose, and 5 mM EDTA, fixed in ice-cold 70% ethanol at 3 × 106 cells/ml, and stored at 4°C for at least 18 h. Cells were stained with 50 μg/ml propidium iodide and 1 mg/ml RNase A for 30 min in the dark at room temperature in the PBS buffer. DNA content analyses were performed on a Becton Dickinson FACSCalibur cytometer using CellQuest and ModFit software. For microscopy, cells were grown on chambered glass slides (Nunc, Naperville, CT) and infected with AdrAhRFL. Cells were fixed in 4% (w/v) paraformaldehyde for 10 min, washed three times in PBS, and imaged on a Nikon Microphot-SA fluorescence microscope fitted with a GFP filter and SPOT CCD camera. These studies were performed in the NIEHS Center for Molecular and Cellular Toxicology Imaging and Cytometry Facility Core at Wayne State University.
Immunoprecipitation and Western Blots.
Subconfluent cultures were harvested by scraping on ice in 1 ml per 100-mm plate of buffer composed of 25 mM HEPES, pH 7.6, 1.5 mM EDTA, 5 mM dithiothreitol, and 10% glycerol containing 1 mM sodium vanadate, 1 mM NaF, 1 mM phenylmethylsulfonyl fluoride, 1 μg/ml leupeptin, 1 μg/ml aprotinin, and 1 μg/ml pepstatin A). Cells were lysed by Dounce homogenization (20 strokes) and the cell debris was removed by centrifugation at 18,000g for 30 min at 4°C. Protein from 200 × 103 cells was treated with Me2SO or 10 nM TCDD for 2 h/20°C before immunoprecipitation with a primary antibody (IPn Ab) against pRb (anti-Rb) or the AhR (anti-AhR) for 4 h on ice, followed by precipitation with Protein-G resin. Beads were washed four times in radioimmunoprecipitation assay buffer, and the bound proteins were fractionated by 7.5% SDS/polyacrylamide gel electrophoresis, transferred to polyvinylidene difluoride membrane, blocked with 4% bovine lacto transfer optimizer in Tris-buffered saline containing 0.1% Tween 20. Filters were probed with the anti-pRb antibody and antibodies against mouse AhR and Arnt (kindly provided by Dr. R. Pollenz (Medical University of South Carolina, Charleston, SC) for 4 h at room temperature, followed by an alkaline phosphatase-conjugated secondary antibody for 1 h. Detection was by chemiluminescence using Western-Star and imaging on a GS-525 Molecular Imager.
RT-PCR.
Total RNA was extracted from the cells using the method of Chomczynski and Sacchi (1987). First strand cDNA was generated from 1 μg of total RNA using a dT primer and Superscript II reverse transcriptase. PCR (30 cycles) using Taq polymerase was performed using oligonucleotide 15 and 16 (rat CYP1A1) and 17 and 18 (rat GAPDH) in the same reaction tube. The PCR products were fractionated on a 1.2% agarose gel and visualized by ethidium bromide staining. Images were captured digitally using a gel documentation system (Alpha Innotech, San Leandro, CA) and the bands quantitated using Molecular Analyst software (Bio-Rad).
Hydroxylapatite Binding Assay.
Cytosol from the 5L and BP8 stable transfectants was prepared as described previously (Reiners et al., 1997). Cytosol (2 mg/ml, 0.1 ml) was incubated with 1% (v/v) Me2SO or increasing concentrations of [3H]TCDD (100 pM–10 μM, 40 Ci/mmol) in the absence or presence of 100-fold excess 2,3,7,8-tetrachlorodibenzofuran for 2 h at 20°C. [3H]TCDD binding to hydroxylapatite followed the conditions described by Gasiewicz and Neal (1982). Specific [3H]TCDD binding was determined by subtracting the scintillation counts for nonspecific binding ([3H]TCDD+2,3,7,8-tetrachlorodibenzofuran) from total [3H]TCDD binding for each ligand concentration.
Electrophoretic Mobility Shift Assay.
5L, BP8, BP8-WT, and BP8-E333A cells were grown to 80 to 90% confluence. The cytosolic fraction was prepared and the mobility shift assay performed as described in detail by Reiners et al. (1997). The complementary oligonucleotides 5′-GATCCGGCTCTTCTCACGCAACTCCGAGCTCA-3′ and 5′-GATCTGAGCTCGGAGTTGCGTGAGAAGAGCCG-3′ contain an AhR DNA binding site (underlined) were annealed and used as the DNA probe in the assay.
Results
We recently showed that the human AhR makes direct contact with the pRb through two distinct domains in the AhR (Ge and Elferink, 1998). One site is located within the N-terminal 364 amino acids of the AhR, whereas the second is confined to a glutamine (Q)-rich region located in the receptor's C-terminal transactivation domain (TAD). Within the N-terminal region are two potential pRb binding sequences: one is the bHLH domain and the second is a LXCXE motif. Both elements in other proteins are known to bind pRb (Gu et al., 1993; Lee et al., 1998). Using the yeast two-hybrid system, we first examined whether the AhR bHLH domain binds pRb (Fig. 1). We generated a series of “bait” constructs fusing the LexA repressor DNA binding domain to the AhR missing the bHLH domain (Fig. 1A). In addition, the constructs contain various C-terminal deletions because previous studies showed that the full-length AhR possesses a potent TAD within this region capable of driving yeast reporter expression in the absence of a “prey” construct (Ge and Elferink, 1998). Removal of the TAD was necessary to detect a two-hybrid interaction. Except for the yeast transformed with the AhR117–853 bait construct (containing the entire AhR TAD), the absence of pB42ADpRb prey construct prevented the leucine auxotrophic cells from growing on the leucine deficient X-gal plates (Fig. 1B, -pRb). Growth of cells containing the C-terminal truncated AhR constructs required cotransformation with pB42ADpRb (Fig.1B, +pRb). Quantitative β-galactosidase activity measuring LacZ reporter gene expression corroborates the plate assay detectingLEU reporter expression (results not shown). Hence, the evidence shows that the bHLH domain is not required for the AhR-pRb interaction.
Next, we examined whether the LXCXE motif is responsible for the physical interaction with pRb (Fig. 2). We introduced alanine substitutions by site-directed mutagenesis at each of the three conserved residues in the human AhR LXCXE motif (Fig.2A; L331A, C333A, and E335A). Figure 2B shows the growth properties of the full-length AhR and two C-terminal truncation constructs (1–672 and 1–589) on leucine deficient X-gal plates in the absence or presence of pRb. Consistent with previous results examining growth on leucine deficient plates (Ge and Elferink, 1998), cells expressing the truncated wild-type AhR (WT) grow only in the presence of pB42ADpRb (Fig. 2B, +pRb). In contrast, coexpression of pRb fails to support cell growth of cells containing the mutated AhR constructs implying that the AhR-pRb interaction requires each of the conserved amino acids within the LXCXE motif. Our observations agree with the well-documented role for the LXCXE motif as a consensus pRb binding sequence (Weinberg, 1995). Quantitative β-galactosidase activity measurements support the plate assay results (results not shown). It is worth noting that the alanine substitutions in the full-length AhR (AhRFL) do not disrupt TAD activity, suggesting that the receptor's C-terminal region is not noticeably affected by the mutations.
To explore the functional significance of the AhR-pRb interaction we initiated studies in the 5L (AhR positive) and BP8 (AhR negative) rat hepatoma cell lines (Weiss et al., 1996). Asynchronous 5L cells exposed to 10 nM TCDD for 24 h undergo a pronounced G1 arrest, marked by a shift from 48.8 ± 2.5% to 84.1 ± 3.3% of the cells in G1(Fig. 3). Unlike the 5L cells, the BP8 clonal variant, selected for its resistance to benzpyrene toxicity, fails to arrest after TCDD treatment. Pharmacological evidence for the role of the AhR in the TCDD-induced G1 arrest comes from a dose-response study (Fig.4). The EC50concentration for G1 arrest is 30 to 100 pM and is maximal by 300 pM, closely approximating the dose-response forCYP1A1 induction (Nebert, 1989). Furthermore, 1 μM 3′methoxy-4′-nitroflavone, a high-affinity AhR antagonist that competes with TCDD for AhR binding and prevents receptor activation (Henry et al., 1999) completely suppresses the G1 arrest induced by 150 pM TCDD. This result implies that the G1 arrest requires AhR activation to a transcriptionally competent form. Using a recombinant adenovirus expression system (He et al., 1998) to express the wild-type rat AhR in BP8 cells, we obtained genetic evidence for the AhR's role in TCDD-induced G1 arrest (Fig.5). Approximately 20 to 30% of the BP8 cells were infected with the adenovirus 24 h before TCDD treatment as monitored by GFP expression (Fig. 5A). Infected cultures were treated with 10 nM TCDD for 24 h and sorted by flow cytometry using GFP as a marker (Fig. 5B). Noninfected (BP8-GFP) and infected (BP8+GFP) cells were collected and analyzed for DNA content by flow cytometry using propidium iodide (Fig. 5C). DNA content reveals that 86.8% of the AhR expressing cells (BP8+GFP) are in G1 whereas only 65.1% of the uninfected cells (BP8-GFP) from the same culture were in G1. This result agrees with the observations reported by Weiss et al. (1996). Control experiments using an adenovirus expressing LacZ confirm that the TCDD-induced G1 arrest is specific for AhR expression (data not shown).
Given the AhR's involvement in TCDD-induced G1arrest in 5L cells and the well-documented role of pRb in blocking G1 cell cycle progression (Dyson, 1998), we asked whether the AhR-pRb interaction is necessary for the TCDD-induced G1 arrest. We generated stably transfected BP8 cell lines, one expressing the wild-type rat AhR constitutively (BP8-WT), and the second line expressing constitutively a mutated AhR harboring an alanine substitution at position 333 in the rat receptor's LXCXE motif (BP8-E333A). This change corresponds to the E335A mutation in the human AhR examined in the yeast two-hybrid system (see Fig. 2). Western blotting confirms expression of the AhR proteins in the BP8-WT and BP8-E333A lines at levels ≈30% and ≈100% of that in 5L cells, respectively (Fig. 6A). TCDD induced G1 arrest in the 5L, BP8, and stable transfectant lines was examined by flow cytometry (Fig. 6B). The results show that TCDD induced a pronounced G1arrest in the BP8-WT line increasing the G1population by 27% from 56 to 83%, comparable with that seen with the 5L cells. In contrast, TCDD shifts the percentage of BP8-E333A cells in G1 by only 9% from 54 to 63%. This represents a statistically significant decline in the TCDD inducible G1 arrest (**p < 0.01), although the modest response seen in the BP8-E333A cells is siginificant (*p < 0.05). This residual G1arrest may reflect pRb binding to the Q-rich region in the AhR (Ge and Elferink, 1998), or represent AhR activity that is independent of pRb. Notwithstanding, the evidence suggests that the AhR mediated G1 arrest in 5L cells is at least partially dependent upon the AhR-pRb interaction through the LXCXE motif.
Coimmunoprecipitation experiments were performed to confirm that the AhR's E333A point mutation in BP8-E333A cells indeed disrupts the AhR-pRb interaction (Fig. 7). Because AhR expression in the BP8-WT cells is too low to reliably detect the AhR in immunoprecipitates, we infected BP8 cells with the AdrAhRFL adenovirus to obtain AhR expression comparable with that in BP8-E333A cells. We and others have previously documented AhR-pRb coprecipitation in 5L cells (Ge and Elferink, 1998) and MCF-7 cells (Puga et al., 2000). We show here that an anti-AhR antibody can coprecipitate pRb and the Arnt protein in a TCDD-dependent manner from the virally infected BP8 cells (compare lanes 3 and 4). In contrast, although the mutant AhR in BP8-E333A cells binds the Arnt protein, it failed to associate with pRb. Hence, in keeping with the yeast two-hybrid observations, the E333A substitution in the LXCXE motif disrupts the AhR-pRb interaction in the hepatoma cells. Because Arnt binds to the mutant AhR, the conformation of the receptor's HLH domain (amino acids 28-80) seems to be unaltered by the alanine substitution at position 333.
The LXCXE motif lies within the AhR ligand-binding domain (Burbach et al., 1992). Hence, it is formally possible that the diminished G1 arrest response in the BP8-E333A cells is caused by a mutation-induced conformational change in the AhR affecting ligand binding and receptor activation to a DNA binding form. Saturation-binding analyses were performed by incubating cytosols from 5L, BP8-WT, and BP8-E333A cells with [3H]TCDD over a range of concentrations, and specific binding at each concentration determined using the hydroxylapatite assay (Gasiewicz and Neal, 1982). The overlapping saturation binding curves obtained with AhR from the various cell lines indicates that TCDD binding affinities for the wild-type and mutant receptor are essentially identical (Fig.8). Therefore, ligand binding does not account for the functional difference between the wild-type and mutant AhR. Because ligand binding affinity is critically dependent upon hsp90 binding to this region of the AhR (Coumailleau et al., 1995), these data also infer that the E333A mutation does not disrupt the AhR-hsp90 interaction. Hence, this result provides compelling evidence that the E333A mutation—located within the ligand-binding domain—does not impart a conformational change upon the receptor protein. Ligand binding triggers AhR activation to a DNA binding form. Thus we also examined AhR DNA binding in vitro using the electrophoretic mobility shift assay (EMSA) (Fig. 9). EMSA detected the formation of TCDD-inducible AhR-DNA complexes in extracts from each of the cell lines expressing the AhR (Fig. 9, arrow). The evidence shows that the mutated AhR in BP8-E333A cells retains the capacity to bind DNA. However, comparing AhR protein levels (Fig. 6A) with the DNA binding properties (Fig. 10) in each cell line suggests that pRb may be contributing to AhR DNA complex formation. Because the E333A mutation specifically disrupts the pRb interaction, the identical migratory behavior of the DNA complex obtained with the wild-type or mutant AhR suggests that pRb is ordinarily not part of the in vitro complex detected by the EMSA. This agrees with our inability to “supershift” the AhR-DNA complex immunologically with an antibody against pRb (data not shown). Failure to detect protein components in the EMSA that contribute to DNA binding is not uncommon.Nguyen et al. (1999) could not detect the coactivator ERAP 140 in the AhR-DNA complex by EMSA, despite showing that it contributed to AhR DNA binding and function. Likewise, Gu et al. (1993), studying the MyoD-pRb interaction, failed to detect pRb in the MyoD-DNA complex in the EMSA, and estrogen receptor DNA binding and function is enhanced by cyclin D1 without forming a detectable ternary complex in the EMSA (Zwijsen et al., 1997). These reports illustrate that physiologically significant interactions in vivo are not necessarily recapitulated by the EMSA in vitro.
Kolluri et al. (1999) recently showed that the TCDD induced G1 arrest depends on AhR-mediated induction of p27Kip1. The evidence presented here indicates that the AhR-pRb interaction is required for maximal TCDD induced G1 arrest. This suggests that pRb binding may be needed for full AhR transcriptional activity. Because TCDD inducibleCYP1A1 expression is a hallmark of AhR transcriptional activity, we examined CYP1A1 expression in the various cell lines (Fig. 10). Total RNA was isolated from the 5L, BP8, BP8-WT, and BP8-E333A cell lines treated with vehicle or 10 nM TCDD for 24 h, and subjected to RT-PCR using primers specific for CYP1A1 and GAPDH (as an internal PCR control for quantitative purposes). The CYP1A1 expression profile in BP8-WT cells is nearly identical to that detected in 5L cells, whereas the induction response in BP8-E333A cells is significantly weaker (*p < 0.05) being only about 40% of that detected in the cells expressing a wild-type AhR. This resembles the difference in the G1 arrest response (Fig. 6B, 27% in the BP8-WT versus 9% in the BP8-E333A cells). Given that the AhR in BP8-WT and BP8-E333A cells exhibits a comparable level of DNA binding (Fig.9), the difference in AhR-mediated transcriptional activity implies that pRb may be functioning as a coactivator.
Discussion
This article provides evidence that maximal TCDD induced G1 arrest and CYP1A1 expression in rat 5L and BP8-WT cells relies on the AhR-pRb interaction involving the receptor's LXCXE motif. Furthermore, the evidence suggests that pRb functions as an AhR coactivator. Recently, Lu and Danielsen (1998)demonstrated that pRb binding to the androgen receptor is essential for its transcriptional activity, confirming that pRb can function as a transcriptional coactivator. Overexpression of pRb also potentiates glucocorticoid receptor transcriptional activity in concert with hBrm, a homolog of the Saccharomyces cerevisiae SWI2/SNF2 protein involved in chromatin remodeling (Singh et al., 1995). Because hBrm binds to pRb through a LXCXE motif however, pRb potentiation of AhR activity probably involves a mechanism distinct from that used by the glucocorticoid receptor. Instead, AhR coactivation may rely upon the interaction of pRb with TAFII250, an integral component of the TFIID basal-transcription complex (Brehm and Kouzarides, 1999). It is noteworthy that drug-inducible expression of a CAT reporter construct under the control of the rat CYP1A1promoter could be repressed by cotransfection of a construct expressing the E1A viral oncoprotein (Sogawa et al., 1989). Futhermore, the effect of E1A was on XRE-mediated enhancer activity, suggesting that AhR-driven trans-activation was repressed by E1A. Because E1A contains a LXCXE motif and functions as an oncoprotein by sequestering pRb (Whyte et al., 1989), E1A's repression of theCYP1A1 promoter may involve a mechanism denying the AhR access to pRb.
We previously showed that the AhR-pRb interaction occurs through two distinct AhR domains (Ge and Elferink, 1998). Yeast two-hybrid experiments (Fig. 2) and the coprecipitation results with the hepatoma cells (Fig. 7) confirm that the LXCXE motif makes up one of the AhR-pRb interaction domains. Binding to the LXCXE motif is negatively regulated by phosphorylation at pRb residues Thr 821 and Thr 826 (Zarkowska and Mittnacht, 1997), suggesting that pRb binding to the AhR is restricted to the hypophosphorylated, active form of pRb. Coprecipitation data presented here and elsewhere supports this interpretation (Fig. 7; Ge and Elferink, 1998; Puga et al., 2000). Given that hypophosphorylated pRb is limited to the G0 and G1 phases of the cell cycle, it predicts that pRb-dependent AhR activity will be cell-cycle dependent. Efforts are currently underway to examine this prediction. The second pRb-binding domain is confined to an 83-amino-acid, Q-rich region within the TAD (Ge and Elferink, 1998). In vivo studies indicate that in the native chromosomal setting, TCDD-induced CYP1A1 gene expression is heavily dependent upon the receptor's TAD (Ko et al., 1996). In this context, the residual CYP1A1 expression (Fig. 10) and G1 arrest (Fig. 6) seen in BP8-E333A cells may reflect pRb binding through the Q-rich domain that is undetectable under in vitro condition (Fig. 7) but nevertheless persists in vivo. Alternatively, this activity may be independent of pRb, relying instead on a direct AhR interaction with components of the transcriptional machinery such as transcription factor IIB (Swanson and Yang, 1998) or another coactivator such as RIP140, which was recently shown to bind the AhR Q-rich region (Kumar et al., 1999).
Kolluri et al. (1999) demonstrated that the TCDD-induced cell cycle arrest in 5L cells involves AhR-mediated induction of the CDK inhibitor p27Kip1, although the precise induction mechanism remains unclear. p27Kip1 is a potent inhibitor of cyclinE-CDK2 and cyclinA-CDK2 activity and the amount of “free” p27Kip1 responsible for inhibiting CDK2 activity is tightly regulated (Sherr and Roberts, 1999). By suppressing CDK2 activity, pRb remains active as an inhibitor of E2F function and cells are prevented from entering S phase. We propose a model (Fig.11) wherein AhR-mediated G1 arrest results from p27Kip1 induction preventing pRb phosphorylation, thereby keeping E2F repressed. Because our data suggest that hypophosphorylated pRb can function as an AhR coactivator, we envision establishment of a positive feedback loop sustaining AhR transcriptional activity—conditional upon the presence of an AhR agonist. In contrast, transition through the G1/S checkpoint relies on CDK4/6 and CDK2-mediated pRb hyperphosphorylation to derepress E2F-regulated gene expression. In turn, E2F facilitates its own transcriptional activity by controlling expression of the cyclin E gene (Ohtani et al., 1995), thereby increasing cyclin E-CDK2 activity and hastening pRb hyperphosphorylation. Hence, E2F activity also establishes a positive feedback mechanism driving entry into S phase. Collectively, the opposing actions of p27Kip1 and cyclin E-CDK2 function as a “binary switching mechanism” wherein G1/S phase transition seems to require not only the emancipation of E2F transcriptional activity, but also AhR inactivation to terminate synthesis of p27Kip1. As a regulatory component common to both pathways, pRb hyperphosphorylation meets both endpoints simultaneously. Hyperphosphorylation of pRb may also explain why oncogenic ras suppresses AhR activity in MCF-10A human breast cancer cells (Reiners et al., 1997). Activation ofras and the downstream mitogen-activated protein kinase cascade in response to mitogens induces cyclin D1 expression (Peeper et al., 1997), triggering the activation of cyclin D-CDK4/6. CDK4/6 directly phosphorylates pRb but also sequesters p27Kip1 away from the cyclin E-CDK2 complex, thus increasing CDK2 activity and further pRb phosphorylation (Sherr and Roberts, 1999). Hence, the inhibitory effect of ras on AhR activity may be due to pRb hyperphosphorylation.
An alternative mechanism for how the AhR might contribute to cell cycle arrest involves the formation of a ternary complex between E2F, pRb, and the AhR that represses E2F regulated gene expression. Puga et al. (2000) recently proposed such a mechanism based on the expression of E2F-driven reporter constructs. Repression of E2F activity by pRb has been shown to involve recruitment of histone deacetylase 1 by binding with the LXCXE motif in the deacetylase (Brehm et al., 1998), promoting histone deacetylation and formation of a transcriptionally less active chromatin conformation. However, the existence of the E2F/pRb/AhR ternary complex has not been demonstrated, nor has the AhR been shown to possess histone deacetylase activity. How, then, does TCDD suppress the E2F-driven reporter expression in mouse Hepa 1 cells (Puga et al., 2000)? In our hands, TCDD induces a rapid (4 h) and sustained (at least 48 h) increase in p27Kip1 protein in the Hepa 1 cells (C. J. Elferink and A. Levine, unpublished observations). Moreover, the increase requires a functional AhR and Arnt protein. Discerning between a p27Kip1mediated mechanism and one involving a ternary complex will require further study, but it is worth noting that the published data favoring the ternary complex is entirely compatible with the model outlined in Fig. 11.
Evidence indicates that AhR ligand binding is necessary for nuclear translocation, dimerization with Arnt, and maintenance of a transcriptionally active AhR complex (Lees and Whitelaw, 1999). Although our study dealt with the effects of TCDD on cell cycle progression, the impact of AhR function in normal cell cycle progression in vivo cannot be fully gauged until a physiological (endogenous) ligand is identified. It will be interesting to see if the ligand is cell-cycle regulated.
Acknowledgments
We thank Drs. He and Vogelstein for providing the reagents necessary to develop the adenoviral expression system. We are also grateful to Dr. Pollenz for the AhR and Arnt antibodies.
Footnotes
- Received May 25, 2000.
- Accepted December 1, 2000.
-
Send reprint requests to: Cornelis J. Elferink, Ph.D., Institute of Chemical Toxicology, 2727 Second Ave., Room 4000, MCHT, Wayne State University, Detroit, MI 48201. E-mail:cornelis_elferink{at}wayne.edu
-
↵1 Current address: Laboratory of Genetics, National Cancer Institute, National Institutes of Health, Bethesda, Maryland.
-
Supported by the National Institute of Environmental Health Sciences (NIEHS) Grant R29-ES07800 and, in part, by NIEHS Center Grant ES06639.
Abbreviations
- AhR
- aryl hydrocarbon receptor
- bHLH
- basic helix-loop-helix
- TCDD
- 2,3,7,8-tetrachlorodibenzo-p-dioxin
- PAS
- Per-Arnt-Sim (periodicity/aryl hydrocarbon receptor nuclear translocator/simple-minded)
- hsp90
- 90-kDa heat shock protein
- Arnt
- aryl hydrocarbon receptor nuclear translocator
- pRb
- retinoblastoma protein
- CDK
- cyclin-dependent kinase
- bp
- base pair(s)
- PCR
- polymerase chain reaction
- GFP
- green fluorescent protein
- WT
- wild-type
- RT
- reverse transcription (or transcriptase)
- Me2SO
- dimethyl sulfoxide
- TAD
- transactivation domain
- EMSA
- electrophoretic mobility shift assay
- The American Society for Pharmacology and Experimental Therapeutics