Characterization of DNA complexes formed by the nuclear receptor constitutive androstane receptor

The nuclear receptor constitutive androstane receptor (CAR) acts as a xenobiotic sensor and regulates the expression of enzymes, such as several cytochromes P450s and the UDP-glucuronosyltransferase (UGT) type 1A1. CAR binds as a heterodimer with the retinoid X receptor (RXR) to specific DNA sites, called response elements (REs). Clusters of CAR REs, referred to as phenobarbital response enhancer modules (PBREMs), have been identified in several CAR target genes. In this study we confirm that REs formed by direct repeats of two AGTTCA hexamers with 4 spacing nucleotides are optimal for the binding of CAR-RXR heterodimers. In addition, we found that the heterodimers also form complexes on everted repeat-type arrangements with 8 spacing nucleotides. We also observed that CAR is able to bind DNA as a monomer and to interact in this form with different coregulators even in the presence of RXR. Systematic variation of the nucleotides 5'-flanking to both AGTTCA hexamers showed that the dinucleotide sequence modulates the DNA complex formation of CAR monomers and CAR-RXR heterodimer by a factor of up to 20. The highest preference was found for the sequence AG and lowest for CC. The increased DNA affinity of CAR is mediated by the positively charged arginines 90 and 91 located in the carboxyl-terminal extension of the DNA-binding domain of the receptor. Furthermore, we show that one of the three CAR REs of the human UGT1A1 PBREM is exclusively bound by CAR monomers and this is regulated by ligands that bind to this nuclear receptor. This points to a physiological role for CAR monomers. Therefore, both CAR-RXR heterodimers and CAR monomers can contribute to the gene activating function of PBREMs in CAR target genes.     

The Nuclear Orphan Receptor CAR-Retinoid X Receptor Heterodimer Activates the Phenobarbital-Responsive Enhancer Module of the CYP2B Gene

PBREM, the phenobarbital-responsive enhancer module of the cytochrome P-450 Cyp2b10 gene, contains two potential nuclear receptor binding sites, NR1 and NR2. Consistent with the finding that anti-retinoid X receptor (RXR) could supershift the NR1-nuclear protein complex, DNA affinity chromatography with NR1 oligonucleotides enriched the nuclear orphan receptor RXR from the hepatic nuclear extracts of phenobarbital-treated mice. In addition to RXR, the nuclear orphan receptor CAR was present in the same enriched fraction. In the phenobarbital-treated mice, the binding of both CAR and RXR was rapidly increased before the induction of CYP2B10 mRNA. In vitro-translated CAR bound to NR1, but only in the presence of similarly prepared RXR. PBREM was synergistically activated by transfection of CAR and RXR in HepG2 and HEK293 cells when the NR1 site was functional. A CAR-RXR heterodimer has thus been characterized as a trans-acting factor for the phenobarbital-inducible Cyp2b10 gene.     

Repression of CAR-mediated transactivation of CYP2B genes by the orphan nuclear receptor, short heterodimer partner (SHP)

The induction of CYP2B gene expression by phenobarbital (PB) is mediated by the translocation of the constitutive androstane receptor (CAR) from the cytoplasm to the nucleus. The CAR/RXR heterodimer binds to two DR-4 sites in a complex phenobarbital responsive unit (PBRU) in the CYP2B gene. The short heterodimer partner (SHP), an orphan nuclear receptor that lacks a conventional DNA binding domain, was initially identified by its interaction with CAR. We have examined the role of SHP in CAR-mediated transactivation of the CYP2B gene. Coexpression of SHP inhibited the transactivation of the CYP2B gene by CAR in cultured hepatoma cells and the p160 coactivator GRIP1 reversed the inhibition. The interaction of CAR with SHP was confirmed by GST pulldown experiments. SHP did not block the binding of either CAR/RXR to the PBRU or binding of GRIP1 to the CAR/RXR complex in gel mobility shift assays, but slightly increased CAR/RXR binding and slightly altered the mobility of the CAR/RXR/GRIP1 complex, suggesting an interaction of SHP with these complexes. The presence of SHP in the complexes, however, could not be detected in an antibody supershift assay. Recombinant corepressors mSin3A, SMRT, and HDAC1, but not NCoR1, interacted with GST-SHP but each of these corepressors in liver nuclear extracts bound to GST-SHP. SMRT and NCoR1 inhibited CAR-mediated activation independent of SHP, but mSin3A and HDAC1 had little effect alone, and were additive with SHP. These studies demonstrate that SHP does not inhibit CAR-mediated trans-activation by interfering with DNA binding or by competition with GRIP1. Instead, SHP may either inhibit recruitment of other coactivators by GRIP1 or actively recruit corepressors directly to the CAR/RXR/PBRU complex.     

HX531, a retinoid X receptor antagonist, inhibited the 9-cis retinoic acid-induced binding with steroid receptor coactivator-1 as detected by surface plasmon resonance

HX531 is a retinoid X receptor (RXR) antagonist that inhibits 9-cis retinoic acid-induced neutrophilic differentiation of HL-60 cells. In order to elucidate the inhibitory mechanism of HX531, we have developed a novel ligand sensor assay for RXR in which the receptor-coactivator interaction is directly monitored using surface plasmon resonance (SPR) biosensor technology. A 20-mer peptide from steroid receptor coactivator-1 (SRC-1), containing nuclear receptor interaction motif LXXLL was immobilized on the surface of a BIAcore sensor chip. Injection of human recombinant RXR with or without 9-cis retinoic acid resulted in ligand-dependent interaction with the SRC-1 peptide. Kinetic analysis revealed dissociation constants (KD) of 9-cis RA-preincubated RXR to SRC-1 was 5.92 x 10(-8)M. Using this technique, we found that 1 microM HX531 reduced the ka value of liganded-RXR with SRC-1, suggesting that HX531 reduced the affinity of RXR to SRC-1. This SPR assay system was applied to obtain quantitative kinetic data of RXR ligand binding to the SRC-1 peptide and the alteration of these data by antagonists.     

RAR antagonists diminish the level of DNA binding by the RAR/RXR heterodimer

A purified RAR/RXR-DeltaAB heterodimer was obtained by production of His-tagged RAR and untagged RXR in Escherichia coli, followed by combined purification on a Ni(2+) affinity column using excess RXR extract, and finally a gel filtration chromatography step to isolate a pure heterodimer. The purified heterodimer preparation bound 9-cisRA at a level of 0.85-0.95 mol of binding sites per mole of protein monomer. Titration of a 26 kDa fluorescent labeled fragment of the SRC-1 coactivator protein with the purified heterodimer in the presence of the agonist 9-cisRA yielded a binding affinity near 300 nM, whereas no binding was observed in the absence of agonist. Binding of the purified heterodimer to a DR5 target was identical in the absence of ligand and in the presence of 9-cisRA. Competition by unlabeled specific and nonspecific DNA allowed us to demonstrate that the binding curve was bimodal. The first phase of binding was highly specific and of high affinity. This phase also exhibited a high degree of cooperativity in the binding profile. Nonspecific DNA efficiently competes for the second phase. Thus, the first phase of binding likely corresponds to the formation of the specific heterodimer complex in which heterodimerization is energetically coupled to DNA binding. While agonist binding had no effect on the apparent affinity of the heterodimer for DR5, a series of antagonists significantly destabilized the heterodimer-DR5 complex, either through a direct decrease in the affinity of the protein for the DNA or through destabilization of the heterodimer itself. Impeding the interaction between the heterodimer and DNA appears as an additional mechanism of antagonist action of varying efficiency, depending upon the chemical structure of the antagonist.     

Discrete roles for peroxisome proliferator-activated receptor gamma and retinoid X receptor in recruiting nuclear receptor coactivators

Peroxisome proliferator-activated receptor gamma (PPARgamma) plays a major role in adipogenesis. PPARgamma binds to DNA as a heterodimer with retinoid X receptor (RXR), and PPARgamma-RXR can be activated by ligands specific for either receptor; the presence of both ligands can result in a cooperative effect on the transactivation of target genes. How these ligands mediate transactivation, however, remains unclear. PPARgamma is known to interact with both the p160/SRC-1 family of coactivators and the distinct, multisubunit coactivator complex called DRIP. A single DRIP subunit, DRIP205 (TRAP220, PBP), binds directly to PPARgamma. Here we report that PPARgamma and RXR selectively interacted with DRIP205 and p160 proteins in a ligand-dependent manner. At physiological concentrations, RXR-specific ligands only induced p160 binding to RXR, and PPARgamma-specific ligands exclusively recruited DRIP205 but not p160 coactivators to PPARgamma. This selectivity was not observed in interaction assays off DNA, implying that the specificity of coactivator binding in response to ligand is strongly influenced by the allosteric effects of DNA-bound heterodimers. These coactivator-selective effects were also observed in transient-transfection assays in the presence of overexpressed p160 or DRIP coactivators. The results suggest that the cooperative effects of PPARgamma- and RXR-specific ligands may occur at the level of selective coactivator recruitment.     

Analysis of multiple nuclear receptor binding sites for CAR/RXR in the phenobarbital responsive unit of CYP2B2

The phenobarbital (PB) responsive enhancers in CYP2B genes contain a core of two direct repeat-4 nuclear receptor binding sites, NR-1 and NR-2, which flank an NF-1 site and appear to be most important for PB responsiveness. Additional sequences outside the core are required for maximal PB responsiveness, including a third direct repeat-4 site, NR-3. The PB response is mediated by constitutive androstane receptor (CAR) which binds as a CAR/RXR heterodimer to the NR sites. To determine the relative importance of the third NR site, each of the NR sites was mutated individually and in all combinations in the rat PB responsive unit (PBRU). Mutation of NR-3 resulted in similar effects on transactivation of the PBRU by CAR in HepG2 cells as did mutations of NR-1 and NR-2. The recruitment of GRIP1/SRC-2 by CAR/RXR to the PBRU assessed by gel shift assays was cooperatively enhanced if more than one NR site in the PBRU was occupied by CAR/RXR. NR-3 in combination with NR-1 or NR-2 was equal to NR-1 and NR-2 in mediating this cooperative recruitment. Recruitment of SRC-1 and GRIP1/SRC-2 was similar for all NR sites, while some selectivity of NR-1 for SRC-3 was observed. SRC-3 also exhibited CAR-independent activation of the PBRU in HepG2 cells. Micrococcal nuclease mapping of nucleosomes revealed that the NR-1/NR-2 core of the PBRU is present in a nucleosome while NR-3 is present in the linker adjacent to the nucleosome. In the linear sequence NR-3 is further from NR-1 than NR-2 is, but in a nucleosomal structure, NR-3 is well positioned for cooperative recruitment of GRIP1/SRC-2 by CAR/RXR that is bound to NR-3 and either NR-1 or NR-2, while NR-1 and NR-2 are on opposite sides of the nucleosome separated by the histone core. These results demonstrate that NR-3 is functionally similar to NR-1 and NR-2 in CAR transactivation of the PBRU in vitro and suggest that NR-3 may have a greater role in a chromatin context in vivo than is apparent from transient transfection studies.