2,3,7,8-Tetrachlorodibenzo-p-dioxin-inducible, Ah Receptor-mediated Bending of Enhancer DNA*

The environmental contaminant 2,3,7&tetrachlo- rodibenzo-p-dioxin produces its biological effects by binding to an intracellular protein, the Ah receptor. The liganded receptor activates transcription by bind- ing to a specific recognition motif within a dioxin-responsive enhancer upstream of the target CYPlAl gene. Here, we have used gel retardation to analyze the interaction between the liganded Ah receptor and five circularly permuted DNA fragments that contain a receptor recognition motif. Our findings indicate that the binding of the liganded receptor to its recognition motif bends the DNA at (or near) the site of the protein- DNA interaction. This observation implies that Ah receptor-induced DNA distortion may contribute to the activation of CYPlAl transcription by 2,3,7,8-te- trachlorodibenzo-p-dioxin. 2,3,7,8-Tetrachlorodibenzo-p-dioxin (TCDD,’ dioxin) is the prototype for a class of halogenated aromatic hydrocarbons that are widespread environmental contaminants, which pose a potential to health

The environmental contaminant 2,3,7&tetrachlorodibenzo-p-dioxin produces its biological effects by binding to an intracellular protein, the Ah receptor. 2,3,7,8-Tetrachlorodibenzo-p-dioxin (TCDD,' dioxin) is the prototype for a class of halogenated aromatic hydrocarbons that are widespread environmental contaminants, which pose a potential hazard to human health (l-4). In animals, TCDD elicits a variety of morphological, biochemical, immunologic, reproductive, and neoplastic effects (1, 2). Ligand-binding studies reveal that TCDD-responsive cells contain an intracellular protein, which binds the dioxin saturably and with high affinity and is known as the Ah receptor (5,6). Receptordefective tissues fail to respond to TCDD, a finding which implicates the protein in the mechanism of dioxin action (1, 7-9). We currently envision that TCDD produces its diverse effects by altering gene expression in Ah receptor-dependent fashion (1, 10). The receptor has not been purified, and many of its properties remain unknown.
The induction of aryl hydrocarbon hydroxylase activity constitutes a useful model response for analyzing the mechanism of TCDD action. The cytochrome P-450IAl enzyme catalyzes hydroxylase activity and is responsible for the oxygenation of polycyclic aromatic hydrocarbons, such as the environmental carcinogen, benzo(a)pyrene yses using enhancer DNA and nuclear extracts from uninduced and TCDD-induced cells reveal that the TCDD-receptor complex interacts with a specific "core" recognition motif which is present in multiple copies in the DNA upstream of the CYPlAl gene (19,20). Furthermore, the liganded receptor binds in a 1:l ratio to its recognition motif (20

RESULTS
Previous analyses in other systems reveal that a bent DNA fragment migrates more slowly during electrophoresis than unbent DNA of identical size and base composition (22). The degree of' bending influences the electrophoretic mobility of the DNA fragment; the larger the bend, the slower the mobility (26). The position of the bend along the DNA fragment also affects its electrophoretic behavior: the more centrally located the bend, the slower the mobility (22,27,28). DNA bending may be intrinsic (i.e. a function of the specific nucleat ide sequence) or dynamic (i.e. produced by the binding of a protein).
To determine whether the binding of the liganded Ah receptor to its recognition motif produces a bend in the DNA, we generat.ed a series of five "circularly permuted" DNA fragment.s using the strategy outlined in Fig. 1. The DNA fragments are identical in size and base composition, contain a single recognition motif for the liganded Ah receptor, and differ only in the position of the recognition motif along the DNA. The five fragments were end labeled with "'P and were incubated with nuclear extracts from uninduced or TCDDinduced mouse hepatoma cells; the resulting protein-DNA complexes were analyzed by gel retardation and autoradiography. The autoradiogram reveals a TCDD-inducible band (designated R, in Fig. 2)  ,"I'-1,nbeled I)ermutation f'ragments were incuhated with nuclear extracts from either uninduced (-'1'('1)1)) or TCDL-induced (1 nM recognition motif for the liganded Ah receptor is most centrally located (the sample designated RV) has a slower mobility than complexes in which the recognition motif is located toward an end of the DNA (the samples designated RI and R). These findings suggest that the bend in the DNA occurs in the vicinity of the recognition motif. As a control, we note that the mobility of a second, constitutive protein-DNA complex (designated RL in Fig. 2) varies only slightly (and not in a pattern indicative of DNA bending) among the five DNA fragments. The results of competition studies imply that the constitutive band represents a specific protein-DNA interaction (18-20). Together, these observations indicate that the variation observed for the TCDD-inducible band is not simply a nonspecific effect due to the binding of any protein to the DNA. Finally, short autoradiographic exposures reveal that the mobilities of the five protein-free DNA fragments (designated F in Fig. 2) are identical, demonstrating the absence of intrinsic DNA bending (data not shown).
To document that the liganded Ah receptor contributes to the bending of the DNA, we performed modified gel retardation experiments using nuclear extracts prepared from cells exposed to ["'1]2-iodo-7,8-dibromodibenzo-p-dioxin, a high affinity agonist for the Ah receptor (29). In these studies, the liganded receptor protein contained the radioactive label, and the five permuted DNA fragments were unlabeled. Under these conditions, the gel retardation analyses detect only the protein-DNA complexes that contain the liganded Ah receptor. We have shown previously that the mobility of the protein-DNA complex observed when the protein is labeled is identical to the mobility of the complex ohserved when the DNA is labeled (19). The results shown in Fig. 3  muted DNA fragments. The mobilities of the complexes vary with the position of the recognition motif, again implying that the protein-DNA interaction bends the DNA. The mobility pattern for the five protein-DNA complexes is similar to the pattern observed when the DNA fragments are labeled (compare the patterns designated R, in Figs. 2 and 3). Thus, our findings indicate that the liganded Ah receptor is a component of a nucleoprotein complex containing bent DNA. Our results do not exclude the possibility that the nucleoprotein complex also includes other proteins that contribute to DNA bending, although there is no reason to favor this idea. The issue of whether the liganded Ah receptor can by itself bend DNA requires receptor purification before it can be addressed. To determine the location of the DNA bend, we plotted the mobilities of the five TCDD-inducible protein-DNA complexes as a function of the position of the five restriction sites used to generate the permuted DNA fragments. This procedure allows us to estimate directly the position of the bend with respect to the position of the recognition motif. The results (Fig. 4) indicate that. within experimental error (five to seven nucleotides), the DNA bend coincides with the recognition motif. We interpret these observations to mean that the binding ofthe liganded Ah receptor to its DNA recognition motif induces a bend in the DNA at the receptor-binding site.

DISCUSSION
Others have shown that protein-induced DNA bending occurs in several prokaryotic and eukaryotic systems (22,27,28,30) and that the phenomenon may contribute to the processes of recombination (31), replication (32), and transcription (22,33). The studies reported here imply that the binding of the liganded Ah receptor to its DNA recognition motif produces a bend in the DNA. We envision that the bending may stabilize the receptor-DNA complex by increasing the contacts between the protein and the nucleic acid. This idea is consistent with our previous observations that nucleotides beyond the core recognition motif appear to strengthen the interaction between the liganded receptor and DNA (19). Other experiments, analogous to those described above but utilizing permuted DNA fragments containing just the core motif, have been inconclusive, due to the decreased strength of the protein-DNA interaction, combined with the background noise associated with the use of crude nuclear extracts. Thus, we do not yet know whether the binding of the liganded receptor to just the core motif is sufficient to induce DNA bending or whether nucleotides that flank the cole are also required. Purification of the Ah receptor may facilitate such studies.
The functional significance of our observation is unknown. The fact that the liganded receptor can bend DNA in vitro suggests that the receptor-enhancer interaction may produce a change in DNA structure in uiuo. For example, the energy expended in DNA bending may facilitate base-pair opening, which should, in turn, increase the flexibility of the DNA in the vicinity of the bend (34). If increased DNA flexibility were to occur in uiuo, we envision that it could facilitate the formation of other protein-DNA and protein-protein contacts required for the initiation of transcription. We note that the DNA upstream of the CYPlAl gene contains six copies of the recognition motif for the liganded Ah receptor (20). Thus, the potential for TCDD-inducible, receptor-mediated DNA distortion in UI,Y~ appears substantial. These observations suggest that the analysis of the liganded receptor's effects on DNA structure in the intact cell (i.e. in chromatin) may he a productive area for future research.