The orphan receptor CAR-beta binds DNA as a heterodimer with the retinoid-X receptor and activates gene transcription in a constitutive manner. Here we show that, in contrast to the classical nuclear receptors, the constitutive activity of CAR-beta results from a ligand-independent recruitment of transcriptional co-activators. While searching for potential ligands of CAR-beta, we found that the steroids androstanol and androstenol inhibit the constitutive activity of CAR-beta. This effect is stereospecific: only 3alpha-hydroxy, 5alpha-reduced androstanes are active. These androstanes do not interfere with heterodimerization or DNA binding of CAR-beta; instead, they promote co-activator release from the ligand-binding domain. These androstane ligands are examples of naturally occurring inverse agonists that reverse transcriptional activation by nuclear receptors. CAR-beta (constitutive androstane receptor-beta), therefore, defines an unanticipated steroidal signalling pathway that functions in a manner opposite to that of the conventional nuclear receptor pathways.

We have identified and characterized a new orphan member of the nuclear hormone receptor superfamily, called MB67, which is predominantly expressed in liver. MB67 binds and transactivates the retinoic acid response elements that control expression of the retinoic acid receptor P2 and alcohol dehydrogenase 3 genes, both of which consist of a direct repeat hexamers related to the consensus AGGTCA, separated by 5 bp. MB67 binds these elements as a heterodimer with the 9-cis-retinoic acid receptor, RXR. However, MB67 does not bind or activate other retinoic acid response elements with alternative hexamer arrangements or any of several other wild-type and synthetic hormone response elements examined. The transactivation of retinoic acid response elements by MB67 is weaker than that conferred by the retinoic acid receptors but does not require the presence of all-trans retinoic acid, 9-cis-retinoic acid, or any exogenously added ligand. We propose that MB67 plays an important role in the complex network of proteins that govern response to retinoic acid and its metabolites.The nuclear hormone receptor superfamily is a large group of related transcription factors which includes members that bind a diverse array of ligands, including steroids, thyroid hormone (T3), all-trans retinoic acid (RA), 9-cis-retinoic acid (9-cis-RA), and vitamin D (reviewed in references 4,8,19,23,35). In mammals, approximately a dozen genes encode these conventional receptors. An even larger number of genes encode proteins known as orphan receptors (recently compiled in references 1 and 32). These orphans are structurally and functionally related to the conventional receptors but do not bind known ligands.The biological roles of the orphans are generally unknown. Some information is available for a limited number which were initially characterized prior to their identification as superfamily members. HNF-4, for example, was originally identified as a liver-specific basic transcription factor, with binding sites in several genes (59). Similarly, SF-I was identified as a factor associated with activation of expression of several steroidogenic enzymes (31). However, the majority of the orphans were isolated simply by cross-hybridization with probes derived from conventional receptors, and their functions remain largely undefined. By analogy with the conventional receptors, it has generally been thought that most of the orphans will eventually be demonstrated to be ligand-dependent transcriptional activators. However, recent results demonstrate that several, including 30,59), , and NGFI-B/Nur77 (13,48,70)

We have identified a new murine orphan member of the nuclear hormone receptor superfamily, termed mCAR, that is closely related to the previously described human orphan MB67, referred to here as hCAR. Like hCAR, mCAR expression is highest in liver. In addition to the most abundant mCAR1 isoform, the mCAR gene expresses a truncated mCAR2 variant that is missing the C-terminal portion of the ligand binding/dimerization domain. The mCAR gene has 8 introns, and this mCAR2 variant is generated by a splicing event that skips the 8th exon. mCAR1, like hCAR, binds as a heterodimer with the retinoid X receptor to the retinoic acid response element from the promoter of the retinoic acid receptor ␤2 isoform. Consistent with its lack of a critical heterodimerization interface, the mCAR2 variant does not bind this site. Both mCAR1 and hCAR are apparently constitutive transcriptional activators. This activity is dependent on the presence of the conserved C-terminal AF-2 transcriptional activation motif. As expected from its inability to bind DNA, the mCAR2 variant neither transactivates by itself nor inhibits transactivation by hCAR or mCAR1.The nuclear hormone receptor superfamily includes the receptors for a number of potent biological regulators, such as steroids, retinoids, and thyroid hormone. With the recent addition of nuclear prostaglandin receptors (1, 2), and an oxysterol receptor (3), there are now more than 15 genes in mammalian genomes that encode such conventional receptors. An even larger set of genes encodes the orphan receptors, which are related to the conventional receptors but do not have known ligands. Particularly since individual genes for superfamily members frequently encode more than one isoform as a consequence of either alternative promoter utilization or alternative mRNA splicing, the total number of proteins that belong to the nuclear receptor superfamily is large.

We describe an assay for determining the frequency and specificity of mutations occurring at hot spots within a population of DNA molecules. The procedure consists of (a) annealing the DNA population with a labeled oligonucleotide designed to prime DNA synthesis at the mutational hot spot; (b) DNA elongation in the presence of a single dideoxynucleoside triphosphate together with 1-3 deoxynucleoside triphosphates, and (c) quantitation of all limit elongation products by high-resolution gel electrophoresis followed by autoradiography and computing densitometry. Derivation of mutational frequency and specificity over a wide range of values is demonstrated for M13 viral DNA mixtures containing defined proportions of wild-type and mutant DNAs, as well as for M13 viral DNA populations obtained by transfection of DNA bearing a defined site-specific ethenocytosine lesion. The assay is shown to yield results similar to those obtained by laborious clone-by-clone sequencing of viral progeny. The method is not affected significantly by several tested variables and appears to be suitable for use as a quantitative assay for sequence microheterogeneity at defined positions within DNA populations. Application of the methodology demonstrates that ethenocytosine, an exocyclic DNA lesion induced by carcinogens such as vinyl chloride and urethane, is a highly efficient mutagenic lesion with a mutational specificity expected for noninstructive lesions.