A Direct Interaction between the Aryl Hydrocarbon Receptor and Retinoblastoma Protein

The aryl hydrocarbon receptor (AhR) is a ligand-activated transcription factor in eukaryotic cells that alters gene expression in response to the environmental contaminant 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD). In 5L hepatoma cells, TCDD induces a G1 cell cycle arrest through a mechanism that involves the AhR. The retinoblastoma tumor suppressor protein (pRb) controls cell cycle progression through G1 in addition to promoting differentiation. We examined whether the human AhR or its dimerization partner, the AhR nuclear translocator, interacts with pRb as a basis of the TCDD-induced cell cycle arrest.In vivo and in vitro assays reveal a direct interaction between pRb and the AhR but not the AhR nuclear translocator protein. Binding between the AhR and pRb occurs through two distinct regions in the AhR. A high affinity site lies within the N-terminal 364 amino acids of the AhR, whereas a lower affinity binding region colocalizes with the glutamine-rich transactivation domain of the receptor. AhR ligand binding is not required for the pRb interaction per se, although immunoprecipitation experiments in 5L cells reveal that pRb associates preferentially with the liganded AhR, consistent with a requirement for ligand-induced nuclear translocation. These observations provide a mechanistic insight into AhR-mediated cell cycle arrest and a new perspective on TCDD-induced toxicity.

2,3,7,8-Tetrachlorodibenzo-p-dioxin (TCDD) 1 represents a class of environmental contaminants that include the polychlorinated biphenyls. TCDD triggers a variety of responses in animals including liver toxicity, immune suppression, hyperplasia, and developmental and reproductive toxicity. In humans, TCDD exposure causes chloracne, but concern also extends to TCDD-induced tumor promotion, birth defects, and immunotoxicity (1)(2)(3). Studies in aryl hydrocarbon receptor (AhR) knock out mice indicate that most TCDD-induced toxic-ity is mediated by the AhR (4,5). The AhR is a ligand-activated transcription factor that acts in concert with the AhR nuclear translocator (Arnt) (6). Target gene expression involves binding of the AhR-Arnt complex to an enhancer sequence called a dioxin or xenobiotic response element (7). Cloning and characterization of the AhR and Arnt proteins revealed that both contain a basic helix-loop-helix (bHLH) and PAS homology domain (8 -10). The basic region confers AhR DNA binding and also contains the nuclear localization signal (3,11), whereas the HLH motif and PAS domain both mediate protein dimerization (12,13). Furthermore, the PAS region and AhR ligandbinding domain overlap (10). The C terminus of each protein contains a transactivation domain (TAD). Although both TADs are functionally competent (e.g. in yeast expression systems), in vivo studies indicate that the TAD of AhR predominates, at least in TCDD-induced CYP1A1 gene expression (14,15).
A new perspective on AhR function and TCDD action comes from two recent reports describing a relationship between AhR activity and the cell cycle (16,17). Ma and Whitlock (16) examined differences in the growth rates of wild-type (Hepa1) and AhR-defective mouse cell lines and determined that the AhR influences G 1 cell cycle progression. AhR-defective cells that contain only 10% of wild-type AhR levels exhibit a prolonged transition through the G 1 phase, but when transfected with an AhR expression construct, these cells grew with normal doubling times. Weiss and co-workers (17), using wild-type (5L) and AhR-defective (BP8) rat hepatoma cells, observed a TCDDdependent G 1 arrest only in the 5L cells. Furthermore, expression of the AhR in the BP8 cells ordinarily refractory to TCDDinduced growth arrest reconstituted the inhibitory phenotype, suggesting direct involvement by the AhR in the G 1 arrest. Progression through G 1 phase of the cell cycle is regulated by pRb. pRb is a 110-kDa nuclear phosphoprotein that undergoes cyclic phosphorylation and dephosphorylation during the cell cycle (18,19). The hypophosphorylated, active form of pRb triggers cells to arrest in G 1 , whereas progression beyond the G 1 checkpoint and entry into S phase occur following pRb inactivation by cyclin-dependent kinase (Cdk)-mediated hyperphosphorylation (20). Mechanistically, pRb is believed to function as a repressor protein in G 1 arrest by binding to the transcription factor E2F, preventing E2F-mediated transcription of genes required for S phase. The adenovirus E1A oncoprotein relieves E2F repression by binding to and sequestering pRb (21). It is noteworthy that E1A also suppresses drug induction of rat CYP1A1 through a mechanism involving the xenobiotic response element (22). It is conceivable that E1A disrupts a functional interaction between the AhR-Arnt complex and pRb necessary for AhR-mediated gene expression. Similarly, the TCDD-inducible G 1 cell cycle arrest in 5L cells may involve an interaction between pRb and the AhR-Arnt complex.
These observations prompted us to examine whether the * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18  AhR or Arnt protein is capable of interacting with pRb. Immunocoprecipitation results detected an interaction between the AhR-Arnt complex and pRb. Using the yeast two-hybrid system and the GST fusion protein pull-down assay, we show a direct interaction between pRb and the AhR. In contrast, the Arnt protein does not interact with pRb. The pRb-AhR interaction involves primarily a sequence(s) located within the N-terminal 364 amino acids of the AhR, although a C-terminal 83-amino acid region encompassing the glutamine-rich domain is also involved. The data also suggest that the pRb-AhR interaction per se is not contingent upon ligand binding. Collectively, the data demonstrate that the AhR and pRb directly interact in vivo and in vitro.

EXPERIMENTAL PROCEDURES
Materials-Restriction endonucleases and other DNA modifying enzymes (T4 DNA ligase, calf intenstinal alkaline phosphatase) were purchased from Life Technologies Inc. and New England Biolabs (Beverly, MA). Reduced glutathione and glutathione-agarose were from Sigma. The MATCHMAKER LexA two-hybrid system used and yeast culture media were from CLONTECH. The Taq and KlenTaq DNA polymerases were obtained from Qiagen and Sigma, respectively. The pRb antibody (G3-245) was from Pierce. TCDD was from the National Cancer Institute Chemcal Carcinogen Reference Standard Repository. Western-Star and Galacton-Star kits were purchased from Tropix (Bedford, MA). Radioactive compounds were acquired from Amersham Pharmacia Biotech. The TNT-coupled transcription and translation system was from Promega (Madison, WI). Protein G-coupled Sepharose resin and custom synthesized oligonucleotides were from Life Technolgies Inc.
Yeast Plasmid Construction and the Two-hybrid Assay-All AhR and Arnt LexA fusion proteins were constructed using pLexA vector (MATCHMAKER system, CLONTECH). This vector encodes a 202amino acid LexA repressor DNA-binding domain under the control of the yeast P ADH1 promoter followed by a multiple cloning site. The human AhR and Arnt constructs were derived from the plasmid phuAHR (Ref. 23; provided by Dr. C. Bradfield, University of Wisconsin, Madison, WI) and plasmid pBM5NeoM1-1 (Ref. 6; provided by Dr. O. Hankinson, UCLA, Los Angeles, CA). Full-length and truncated coding regions of the AhR were PCR amplified using the primer hAhR-F with the appropriate reverse primer. PCR products were subcloned into the NotI site of pLexA and designated pYAhRFL, pYAhR1-672, pYAhR1-589 and pYAhR1-528. The full-length and truncated human Arnt coding regions were PCR amplified using the primer hArnt-F with either hArnt-R (full length) or hArnt694-R (truncated). Arnt PCR products were subcloned into the BamHI site of pLexA to create pYArntFL and pYArnt1-694. The human pRb construct encoding amino acids 374 -928 was generated by PCR amplification of p2206TRB (Ref. 24; provided by Dr. R. Weinberg, MIT, Cambridge, MA) using the primers Rb374-F and Rb928-R. The PCR product was subcloned into the XhoI site of pB42AD generating pB42ADpRb, comprising an in-frame fusion between pRb and a vector-encoded 88-residue activation domain (B42). 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 used two reporters (LEU2 and lacZ) under the control of LexA operators.
Quantitation of ␤-Galactosidase Activity-Quantitation of ␤-galactosidase activity was performed on liquid cultures using the chemiluminescent reporter assay Galacton-light™ (Tropix) in accordance with the manufacturer's protocol. Briefly, yeast transformants were grown on selective plates for 3-5 days, after which 3-5 colonies were used to inoculate 5 ml of minimum dropout glucose medium and incubated at 30°C overnight. Subsequently, 0.5 ml was used to inoculate 5 ml of selection medium containing galactose and grown in log phase for 5 h. A 1.5-ml aliquot was removed and washed once with 1.5 ml of Z buffer (60 mM Na 2 HPO 4 , 40 mM NaH 2 PO 4 , 10 mM KCl, 1 mM MgSO 4 , 35 mM ␤-mercaptoethanol) and resuspended in 300 l of Z buffer. Cells were permeabilized by freeze-thawing twice in liquid nitrogen for 1 min, after which triplicate assays were performed using 20 l of the cell lysate with 70 l of the Galacton™ substrate in a 96-well microplate and incubated at room temperature for an hour. ␤-Galactosidase activity presented as relative light units was determined in a MLX Microtiter® plate luminometer (Dynex Technologies Inc., Chantilly, VA) using parameters suggested by the manufacturer.
Preparation of GST Fusion Proteins-The bacterial expression vector pGEX-KG (25) was used to generate GST fusion proteins with the human AhR (amino acids 1-364 and 537-774) and human Arnt (amino acids 510 -780). PCR amplified products using plasmid phuAHR and the primer sets GST-hAhR-F/GST-hAhR364-R and GST-hAhR537-F/G-ST-hAhR774-R were directionally subcloned into XbaI/HindIII-cut and EcoRI/HindIII-cut vector, respectively, to give pGEXAhR1-364 and pGEXAhR537-774. An Arnt-encoding cDNA was generated by PCR using plasmid pBM5NeoM1-1 and primers GST-hArnt510-F and GST-hArnt780-R followed by directional cloning into EcoRI/HindIII-cut pGEX-KG vector generating pGEXArnt510 -780. GST fusion proteins were expressed in Escherichia coli following induction by 1 mM isopropyl-thio-␤-D-galactopyranoside for 3 h. Fusion proteins were recovered from cells resuspended in NETN buffer (150 mM NaCl, 5 mM EDTA, 50 mM Tris, pH 8.0, 0.1% Nonidet P-40) containing protease inhibitors, following two passages through the French press and clarification of the lysate at 10,000 ϫ g for 10 min. Fusion proteins were purified using glutathione-agarose beads.
In Vitro Binding Assay-A cDNA encoding full-length human pRb was generated by PCR amplification of p2206TRB using the primers pCR3Rb-F and pCR3Rb928-R and cloned into pCR3.1 (Invitrogen) by TA cloning. In vitro transcription and translation of pRb was performed in the TNT T7-coupled rabbit reticulocyte lysate system in the presence of [ 35 S]methionine according to the protocol provided by the manufacture. Reactions were carried out at 30°C for 90 min. Binding assays were carried out on ice for 2 h using 10 l of the translated pRb incubated with glutathione-agarose beads containing 5 g of GST or GST fusion protein in 1 ml of NETN buffer containing protease inhibitors. Beads were washed six times in NETN buffer and separated by 7.5% SDS-PAGE and detected by autoradiography using a GS-525 MolecularImager® system (Bio-Rad).
Cell Culture-Wild-type rat hepatoma 5L cells and the AhR-defective BP8 variants were grown in 100-mm plates as monolayers in Dulbecco's modified Eagle's medium containing 10% fetal bovine serum, 100 units/ml penicillin, and 100 g/ml streptomycin in 5% CO 2 atmosphere at 37°C.

RESULTS
The TCDD-inducible G 1 arrest in 5L cells may bespeak an interaction between pRb and the AhR-Arnt complex and is supported by immunocoprecipitation experiments (Fig. 1). Immunoprecipitates from control or TCDD-treated 5L cell lysates using an antibody against human pRb were analyzed by Western blotting for the presence of AhR, Arnt protein, and pRb. A TCDD-inducible coprecipitation of the AhR and Arnt protein is revealed by the enrichment of these proteins without a corresponding increase in pRb levels (compare lanes 2 and 3). Digital imaging and quantitation of the bands reveal that AhR and Arnt levels are increased 2.5-and 2.2-fold, respectively, after normalization for pRb. Hence, the data suggest that the AhR-Arnt complex does indeed interact with pRb upon TCDD activation of the AhR. Furthermore, the result also reveals that the AhR-Arnt complex is capable of interacting with the hypophosphorylated, active form of pRb, because the inactive hyperphosphorylated form of pRb detectable in the 5L cell lysate (lane 1, ppRb) is not a component of the immunoprecipitates. Precipitation of pRb, AhR, or Arnt was not observed with nonimmune IgG (data not shown).
To examine whether the AhR or Arnt protein directly contacts pRb, we used the yeast two-hybrid assay. The assay uses two reporter genes (LEU2 and lacZ) under the control of multiple LexA operators, detectable as growth of the leucine auxotrophs and blue colony formation on leucine-deficient X-gal plates. Fusion constructs linking the full-length human AhR or Arnt protein to the DNA-binding domain of the LexA repressor induced both reporters in the absence of the plasmid pB42ADpRb (Fig. 2, C, AhRFL, and G, ArntFL), indicating that the TADs contained in the full-length AhR and Arnt are both transcriptionally competent. This agrees with past observations (14). For the purpose of the yeast two-hybrid assay, however, this domain needed to be removed. Because the AhR TAD is complex, comprising several regions each capable of transactivation (27), the AhR C terminus was progressively deleted in a series of LexA fusion constructs to remove the entire TAD (AhR1-672, 1-589, and 1-528). In contrast, the TAD of Arnt is confined primarily to the last 34 amino acids (27). The Arnt1-694 construct removes the TAD as well as the glutamine-rich region. When transformed into cells in the absence of pB42ADpRb, none of the AhR and Arnt truncated constructs induced ␤-galactosidase expression nor grew on leucine-deficient plates (Fig. 2, ϪpRb). Cotransformation with pB42ADpRb stimulated ␤-galactosidase expression and growth on selective medium in cells transformed with the truncated AhR proteins (Fig. 2, ϩpRb). Restoration of reporter expression by the pLexA bait and pB42AD prey constructs is interpreted as evidence for a direct pRb-AhR interaction. Specificity for this interaction is suggested by the failure of pRb to interact with Arnt1-694.
The AhR-pRb interaction and the impact of ligand binding were examined quantitatively by the yeast two-hybrid assay using liquid cultures (Fig. 3). Background ␤-galactosidase activity in cells (i.e. transformed with the reporter plasmid p8op-lacZ) was not increased in cells cotransformed with the pB42ADpRb construct, indicating that pRb alone is incapable of transcriptional activation. The full-length AhR in the absence of pRb triggered a 200-fold increase in ␤-galactosidase activity (Fig. 3, compare columns 1 and 5). Cotransformation of pRb fails to induce AhRFL-driven ␤-galactosidase activity further (compare columns 5 and 7), suggesting that the pRb interaction with AhRFL is probably masked by the potent TAD of the receptor. In contrast, the truncated AhR constructs were almost devoid of inherent AhR transactivation activity inducing at most a 2-3-fold increase in reporter expression (Fig. 3,  compare column 1 with columns 9, 13, and 17). Interactions between pRb and the truncated AhR proteins stimulated a 10 -100-fold increase in ␤-galactosidase activity, being most pronounced with the AhR1-672 construct (compare columns 9 and 11). The Ϸ20-fold difference in pRb-dependent ␤-galacto-  sidase activity between AhR1-672 and AhR1-589 suggests that pRb may be interacting, at least in part, with this 83amino acid AhR region between residues 589 and 672 (compare columns 11 and 15). Analysis of the Arnt constructs reveals that although the full-length protein stimulates a 6000-fold increase of ␤-galactosidase activity (Fig. 3, compare columns 1  and 21), the truncated Arnt1-694 construct expresses only residual transactivation activity (4-fold above background). Significantly, neither protein confers enhanced ␤-galactosidase activity in the presence of pRb, consistent with the failure of Arnt to bind pRb in the plate assay (Fig. 2). The decrease in ArntFL-induced ␤-galactosidase activity in the presence of pRb construct is thought to be due to the activation domain encoded by pB42ADpRb interfering with the activity of the TAD of Arnt. Control Western blot experiments on yeast cell extracts using antibodies against LexA and pRb reveal that expression of the LexA fusion constructs and pB42ADpRb are comparable among the transformants (data not shown).
In yeast cells devoid of pRb, the AhR ligand ␤NF stimulated a modest (2-3-fold) increase in ␤-galactosidase activity in the cells expressing the AhRFL, AhR1-672, or AhR1-589 fusion protein but not AhR1-528. Notably, no synergism in reporter expression was detected in ␤NF-treated cells cotransformed with pB42ADpRb, suggesting that ligand binding is neither a prerequisite for nor a facilitator of the AhR-pRb interaction. These results are reconcilable with the TCDD dependence observed in the 5L cell immunoprecipitates (Fig. 1) if ligand is required for AhR nuclear translocation and colocalization with the Arnt protein and pRb but not for complex formation. ␤NF actually appears to suppress the pRb interaction with the shorter AhR constructs (i.e. AhR1-589 compare columns 15 and 16 ; AhR1-528, compare columns 19 and 20). It is conceiv-able that a ligand-induced conformational change in these severely truncated proteins lacking the region between residues 589 and 672 renders them less able to bind pRb.
Lavender et al. (28) recently showed that HBP1, a high mobility group-box transcription factor, binds pRb through two distinct pRb-binding sites. High affinity binding involved the LXCXE consensus pRb-binding motif in HBP1, while a separate, low affinity binding site colocalized with the HBP1 activation domain. In contrast, binding between pRb and MyoD occurs through the bHLH domain of MyoD (29). In addition to a TAD and bHLH domain, visual inspection of the amino acid of the AhR sequence reveals an LXCXE motif located (amino acids 331-335 in the human AhR) within the PAS domain. Given the existence of multiple putative pRb-binding motifs in the AhR, GST fusion pull-down studies were performed (Fig. 4). This in vitro assay also provides a convenient method to identify potential pRb interactions within the AhR and Arnt protein TADs masked in the yeast two-hybrid assay. GST fusion proteins coupled to glutathione agarose were used as affinity reagents to bind full-length 35 S-radiolabeled pRb generated by in vitro transcription and translation (Fig. 4, Input). Retention of radiolabeled pRb was analyzed by SDS-PAGE and autoradiography. Both the GST fragment alone and the GST-hArnt (aa 510 -780) protein bound pRb at near background levels defined as binding to naive beads. Because the fragment used encompasses the Arnt protein TAD, the data confirm that this domain does not bind pRb. In contrast, GST-hAhR (aa1-364) protein bound pRb almost 20-fold above background. GST-hAhR (aa537-774) encompassing the glutamine-rich component of the TAD of the receptor also bound pRb, although with a lower affinity (i.e. 5.2-fold) than the N-terminal region. The results clearly show that pRb interacts specifically with both N-and C-terminal regions of AhR, placing one site in the N-terminal 364 amino acids containing both the bHLH and the LXCXE motif. The second site lies between residues 537 and 774. In fact, we can narrow this down further to an 83-amino acid glutamine-rich region between residues 589 and 672, because deleting this region resulted in a Ϸ20-fold decrease in ␤-galactosidase reporter expression (Fig. 3). DISCUSSION This paper presents evidence for a direct interaction between the AhR and pRb and seems to involve at least two distinct regions within the AhR. The GST-AhR pull-down data place one site in the N terminus of the AhR encompassing the first 364 amino acids and a second site within residues 537-774. The Ϸ20-fold decrease in ␤-galactosidase activity in the twohybrid assay following removal of residues 589 -672 suggests that the latter site lies within this 83-amino acid region. The N-terminal region contains both a bHLH and LXCXE motif, whereas the C-terminal sequence harbors a glutamine-rich TAD. Studies on the high mobility group-box transcription factor, HBP1, revealed that its LXCXE motif and activation domain functioned as high and low affinity pRb-binding sites, respectively (28). This resembles the Ϸ4-fold enhanced binding of pRb to GST-hAhR (aa1-364) containing a LXCXE motif relative to GST-hAhR (aa537-774) harboring the TAD (Fig. 4). Because the MyoD-pRb interaction involves the bHLH domain in MyoD (29), we are currently investigating whether the AhR-pRb interaction occurs through the bHLH or LXCXE motif. However, the very existence of the LXCXE motif in the AhR and the absence of pRb binding by the bHLH domain in Arnt suggests that the AhR-pRb interaction occurs through the LX-CXE motif.
We hypothesize that a mechanism involving the AhR-pRb interaction is responsible for the TCDD-inducible G 1 arrest in 5L cells. A role for the AhR in cell cycle arrest is supported by several observations. For example, benzo[a]pyrene induces an AhR-dependent G 1 arrest in murine Swiss 3T3 cells (30). In MCF7 human breast epithelial cells, TCDD causes a significant decrease in cell number when compared with the untreated counterparts over a 72-h period (31). Whether this is due to a G 1 arrest is currently unknown, but MCF7 cells exposed to indole-3-carbinol at concentrations 4-fold above the K d for binding to the AhR arrest in G 1 phase concomitant with Cdk6 down-regulation (32,33). TCDD induces the Cdk inhibitors p INK4a , p21 Waf1 , p27 Kip1 , as well as pRb in a dose-dependent manner in mice (34), consistent with conditions inducing G 1 arrest. Also, ras activation in NIH3T3 cells up-regulates cyclin D1 levels resulting in pRb hyperphosphorylation (35). This is noteworthy because activated ras down-regulates AhR function in MCF10A cells (36), suggesting that ras may indirectly suppress AhR activity by affecting pRb phosphorylation. This implies that AhR function may rely on an active pRb and predicts that AhR activity is cell cycle-dependent. Consistent with this idea, nocodazole-treated Hepa1 cells arrested in G 2 /M no longer induce CYP1A1 expression following TCDD exposure (37).
From a functional standpoint, pRb is generally viewed as a transcriptional repressor (18,19). pRb binds to the E2F activation domain actively suppressing E2F-regulated transcription of S phase-specific genes involved in the G 1 to S phase transition. This involves pRb-mediated recruitment of the histone deacetylase, HDAC1, believed to modify chromatin into a transcriptionally inactive form (38). Whether AhR-mediated G 1 arrest requires pRb transcriptional repression or activation is currently unclear. E1A suppression of CYP1A1 induction is consistent with the latter, because E1A functionally inactivates pRb by sequestration (21,28). Interpreting the E1A effect on CYP1A1 expression is complicated, however, by the recent observation that the coactivator p300/CBP interacts with the TAD in Arnt (22). E1A also binds p300/CBP, raising the pros-pect that E1A inhibits Arnt transactivation by interfering with p300/CBP function. Given that induction of CYP1A1 requires the AhR TAD (15), coupled with our results that pRb interacts with the AhR TAD, we speculate that E1A inhibits CYP1A1 induction primarily by disrupting the AhR-pRb interaction. This scenario imputes that in the context of CYP1A1 expression, pRb functions as a transcriptional activator rather than a repressor. Positive regulation by pRb is known in a few cases, most notably in glucocorticoid receptor-mediated transcriptional activation (39). In glucocorticoid receptor-mediated transcription, pRb acts in concert with hBrm, a homolog of the Saccharomyces cerevisiae SWI2/SNF2 protein, which functions in nucleosome disruption (40). Because nucleosome disruption also occurs during AhR-mediated induction of CYP1A1 (15), the prospect that a similar mechanism is involved must be considered.
Protein-DNA cross-linking studies (41) and purification (42) of the rat liver AhR DNA-binding complex detected a 110-kDa protein comprising part of the AhR-DNA complex that is distinct from Arnt. Given that hypophosphorylated pRb is 110 kDa, we speculate this species may in fact be pRb, suggesting that pRb is an integral component of the AhR complex. DNA binding experiments using purified, baculovirus-expressed AhR and Arnt revealed that AhR DNA binding could be reconstituted only following addition of "a heat-sensitive factor(s) present in soluble extracts from a variety of cell types," including mammalian, insect, and plant sources (43). The broad distribution of pRb or functional homologs (44,45) raises the possibility that these varied cell extracts promoted AhR-Arnt DNA binding by virtue of an pRb activity in the extracts. Because DNA binding by the AhR and Arnt alone may occur under some in vitro conditions, however (46), the precise role for pRb in AhR DNA binding requires further study. Attempts to demonstrate the presence of pRb in the DNA-bound AhR complex from rat liver using anti-pRb antibodies failed to supershift or disrupt the complex in a gel mobility shift assay (data not shown). This may reflect occlusion of the epitope during DNA binding. E box-binding complexes containing MyoD similarly failed to supershift with antibodies against pRb, even though MyoD and pRb bind to one another, and E box binding by MyoD-containing heterodimers is stabilized by pRb (29).
From a toxicological perspective, the AhR-pRb interaction and TCDD-induced G 1 arrest in 5L cells appear at odds with the notion that TCDD is a tumor promoter. Tumor formation is observed in animals treated with relatively high doses of TCDD, yet epidemiological evidence from a population in Seveso, Italy, exposed to low dioxin levels correlates with a significant reduction in total tumors (47). Hence, a thorough examination of the role played by the AhR in regulating cell cycle progression should clarify some of the confusion surrounding the involvement of TCDD in tumor promotion. In contrast, the toxic effects of TCDD on tissue differentiation are reconcilable with the role pRb plays in development and differentiation (18,19). It is conceivable that persistent TCDD signaling through the AhR may disrupt normal pRb-mediated differentiation processes. Hence, the realization that the AhR and pRb interact is pivotal in our efforts aimed at deciphering the mechanism of TCDD toxicity.