Structure and Expression of the Orphan Nuclear ReceptorSHP Gene*

To determine the organization of the orphan nuclear receptor SHP gene (Seol, W., Choi, H.-S., and Moore, D.D. (1996) Science 272, 1336–1339), genomic clones were isolated from human and mouse genomic libraries. TheSHP gene was composed of two exons interrupted by a single intron spanning approximately 1.8 kilobases in human and 1.2 kilobases in mouse. Genomic Southern blot analysis and fluorescence in situ hybridization of human metaphase chromosomes indicated that the SHP gene is located at the human chromosome 1p36.1 subband. The 5′-flanking regions of human and mouse SHPgenes were highly conserved, showing 77% homology in the region of approximately 600 nucleotides upstream from the transcription start site. Primer extension analysis was carried out to determine the transcription start site of human SHP to 32 nucleotides downstream of a potential TATA box. The human SHP gene was specifically expressed in fetal liver, fetal adrenal gland, adult spleen, and adult small intestine. As expected from this expression pattern, the activity of the mouse SHP promoter measured by transient transfection was significantly higher in the adrenal-derived Y1 cells than HeLa cells.

The nuclear receptor superfamily is a group of transcription factors regulated by small hydrophobic hormones such as retinoic acid, thyroid hormone, and steroids and also includes a large number of related proteins that do not have known ligands, referred to as orphan nuclear receptors (for reviews see Refs. 1, 2). The nuclear receptors directly regulate transcription by binding to specific DNA sequences named hormone response elements, generally located in promoters of target genes. The nuclear hormone receptors share a common domain structure. The central DNA binding domain (DBD) 1 includes two zinc binding modules, which consist of a series of invariant cysteine residues. A conserved helical region termed the P box within the DBD (3) makes base-specific contacts and serves as one of the main criteria used for classification of the nuclear receptor superfamily. The C-terminal ligand binding domain (LBD) binds to the cognate ligands. This domain also contains dimerization and transcriptional activation functions. A less well conserved hinge domain that separates DBD and the ligand binding domain has been thought to serve merely as a flexible linker. However, recent results demonstrate that it is also involved with transcriptional repression, at least for a subset of receptors (4). In addition, it was also shown to contain nuclear localization signals (1,2). A quite variable N-terminal domain includes a transcriptional activation function with some receptors.
Although ligands have not been identified for orphan nuclear receptors, a variety of results indicate that they have important functions. The simplest is that knockout mutations of these orphans in mice frequently have shown much more dramatic defects than similar mutations of the conventional receptor genes (5-7). We have recently reported an unusual orphan member of the nuclear receptors that contains a ligand binding domain but lacks the conserved DBD (8). This orphan receptor interacts, both in vitro and in the yeast two-hybrid system with several conventional and orphan members of the receptor superfamily, including retinoid receptors (RAR and RXR), thyroid hormone receptor, and the orphan receptor CAR. In mammalian cells, this receptor specifically inhibited transactivation by the superfamily members with which it interacted, suggesting that it functions as a negative regulator of receptor-dependent signaling pathways. On the basis of its small size and its ability to interact with several superfamily members, it was named SHP (small heterodimer partner) (8). The ability of SHP to interact with multiple superfamily members was independently confirmed by isolation of its rat homologue on the basis of its interaction with the PPAR (peroxisome proliferator-activated receptor) (9).
The ability of SHP to interact with a range of receptors with different affinities suggests that the level of SHP gene expression could play a pivotal role in regulation of receptor-dependent signaling pathways. Thus, information regarding the SHP promoter and its activity is crucial to understanding the function of this novel orphan receptor. In this report, we characterized the structure of the human and mouse SHP genes, the location of the human gene, the expression pattern of SHP in various tissues, and the basal SHP promoter.

EXPERIMENTAL PROCEDURES
Isolation of SHP Genes from Genomic Phage Libraries-Human and mouse genomic phage lambda libraries (CLONTECH, CA) were screened using the SHP cDNA as a probe. One human and three mouse SHP clones were independently isolated and confirmed by Southern blot analysis. Boundaries between exons and introns were determined by DNA sequencing and polymerase chain reaction analysis. The 5Јflanking region of the SHP gene was analyzed by restriction mapping, DNA sequencing, and polymerase chain reaction analysis. Sequence homology between human and mouse genes was analyzed by using the GCG program. The nucleotide sequence(s) reported in this paper has been submitted to the GenBank TM /EBI Data Bank with accession number(s) AF044315 (mouse) and AF044316 (human).
Primer Extension Analysis-The transcription start site of the SHP gene was determined by primer extension assay with an end-labeled oligonucleotide primer complementary to positions ϩ34 to ϩ69 of the human SHP coding sequence. Total RNA was prepared from HepG2 and JEG3 cells using guanidinium isothiocyanate and CsCl density gradient centrifugation (10). The isolated total RNAs were further purified using the Poly(A)-Tract mRNA isolation Kit (Promega, WI). The primer extension assay was basically performed as described previously (10). Briefly, 32 P-labeled primer was mixed with 0.5 g of poly(A) ϩ RNA in 20 l of hybridization buffer and denatured at 65°C for 90 min. The mixture was slowly cooled down to room temperature. After the addition of 30 l of reverse transcription buffer (75 mM Tris-HCl, pH 8.3, 37.5 mM KCl, 4.5 mM MgCl 2 ), 2 l of 0.1 M dithiothreitol, 1 l of 10 mM dNTPs, and 0.5 l of Superscript II reverse transcriptase (Life Technologies, Inc.), the reaction was incubated at 42°C for 60 min and terminated by the addition of 1 l of 0.5 M EDTA. The extended transcripts were recovered by ethanol precipitation after RNA digestion and phenol extraction and resolved on a 6% denaturing polyacrylamide gel. The start site was determined by comparison to sequencing reactions carried out on the genomic clone using the same primer in parallel.
Northern Blot Analysis-A human adult tissue blot was purchased from CLONTECH (Palo Alto, CA), 13-week-old human fetus tissue RNA was isolated, and Northern blot analysis was carried out as described previously (11).
Construction of Promoter-Luciferase Fusion Vectors-A fragment containing 2080 base pairs of 5Ј-flanking sequences from the mouse SHP gene was generated by polymerase chain reaction using the genomic clone as template and used to replace the thymidine kinase (TK) promoter fragment in the pTKLuc reporter plasmid (12) to generate pSHP2080Luc. The 5Ј-deletion constructs were prepared by digestion with BamHI and restriction enzymes as indicated in Fig. 6 followed by Klenow fill-in and T4 ligation.
Cell Culture and Transient Transfection-Y1 and HeLa cells were maintained in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum and antibiotics. The day before transfection, cells were split into 12-well plates. On the next day, the medium was changed, and 4 h later, cells were transfected in triplicate using calcium phosphate co-precipitation as described previously (12). For each well, equal molar amounts (0.5 g of pSHP2080Luc, with others adjusted accordingly) of the pSHPLuc constructs was co-transfected with 0.5 g of pTKGH as an internal control. The medium was replaced 18 h after transfection, and cells were further grown for 36 h in culture medium. The medium was collected for measuring growth hormone activity, and the cells were harvested in lysis buffer (Promega, WI) according to the manufacturer's protocol. GH and luciferase activities were determined as described previously (11).
Genomic Southern Blot Analysis-Monochromosomal somatic cell hybrid blot was obtained from Biosis Laboratories (New Haven, CN). A zoo blot containing EcoRI-digested genomic DNAs from nine eukaryotic species was purchased from CLONTECH (Palo Alto, CA). The blots were hybridized with human SHP cDNA as a probe and washed with 2ϫ SSC (1ϫ SSC ϭ 0.15 M NaCl and 0.015 M sodium citrate), 0.1% SDS at room temperature for 30 min and 0.1ϫSSC, 0.1% SDS for 30 min at 65°C.
Fluorescence in Situ Hybridization-Phytohemoagglutinin (PHA)stimulated lymphocyte was synchronized with methotrexate, and human male metachromosomes were hybridized with a biotin-labeled SHP genomic DNA probe (13). Bromodeoxyuridine release technique was applied for the delineation of R-or G-band. Briefly, SHP probe was labeled with biotin dUTP by nick translation, and the labeled probe was combined with human Cot-1 DNA and hybridized to normal metaphase chromosomes in a solution containing 50% formamide, 10% dextran sulfate, and 2ϫ SSC. Specific hybridization signals were detected by incubating slides in fluorescein-conjugated avidin. The slides were counter-stained with propidium iodide and analyzed with a Zeiss fluorescent microscope. A total of 80 metaphase cells were examined, with 65 exhibiting specific signals. Arrows indicate the band specifically detected in human chromosome 1. B, a Southern blot containing 4 g of EcoRI-digested genomic DNAs from different species was hybridized with the human SHP cDNA as a probe as described (11). kb, kilobases.

RESULTS
Structure of the SHP Gene-To investigate the genomic organization of the SHP gene, we isolated three genomic clones for the murine SHP gene and one for the human SHP gene, respectively, from genomic phage lambda libraries by using SHP cDNA (8) as a probe. The genomic structures of human and mouse SHP genes were characterized by Southern blotting and polymerase chain reaction analysis, and sequences of the intron-exon boundaries and part of the coding regions were determined. As indicated in Fig. 1, A and B, both the human and mouse SHP genes consist of two exons interrupted by one intron, and the sequence of exon-intron boundaries conform to the consensus splicing signals. No evidence for the existence of the DBD of nuclear hormone receptors was found with neither the human nor the mouse genomic clones (data not shown). The intron in the human and mouse SHP gene lies between the first and second nucleotides of the codon for aspartic acid 181. This is quite reminiscent of the structure of the gene encoding the closest relative of SHP, DAX-1, which also consists of two exons with an intron located exactly at the same relative position (14).
Chromosomal Location of the SHP Gene-To determine the chromosomal location of the SHP gene, a monochromosomal somatic cell hybrid Southern blot was hybridized with human SHP cDNA as a probe. The monochromosomal somatic cell hybrid panel consists of 24 hybrid cell lines, each carrying 1 or 2 human chromosomes within the context of mouse or hamster background, as shown in Fig. 2A. Somatic cell hybrid 0A1AR that contained human chromosome 1 showed a 3.8-kilobase band hybridized to the human SHP cDNA probe. The identical band also showed up in control lanes of human and mixtures of human-mouse genomes. These results indicate that the SHP gene is located in human chromosome 1. To further locate the SHP gene on the human chromosome 1, fluorescence in situ hybridization was carried out (Fig. 3). To confirm the precise location of the gene, the fluorescent signals were visualized directly on R-banded metaphase spreads, and the identical spreads were separately stained with Giemsa (data not shown). These analyses demonstrated that the SHP gene resides in a single locus on human chromosome 1 at position 1p36.1. A genomic Southern blot containing EcoRI-digested DNAs from various species was hybridized with human SHP cDNA as a probe (Fig. 2B). Two genomic EcoRI fragments (approximately 7 and 2.8 kilobases) from human DNA, as well as one or two bands from various species were hybridized to the probe, indicating that the SHP genes are well conserved throughout different species (Fig. 2B). Since the human SHP genomic clone has three EcoRI sites in its intron, those two bands are likely to originate from an identical SHP genome. This result is consistent with an idea that the SHP gene is a single-copy gene.
Expression of the SHP Gene-Northern blot analysis was FIG. 4. Expression of SHP mRNA. A, a Northern blot containing 2 g of poly(A) ϩ mRNA from the indicated human tissues was hybridized with a human SHP probe as described previously (11), and the identical blot was rehybridized with TRIP3 gene (25) as a control. B, a Northern blot containing 25 g of total RNA from a 12-week-old human fetus was hybridized with a human SHP probe as previously described (11). The identical blot was rehybridized with ␤-actin cDNA as a control. kb, kilobases.
used to identify the expression pattern of the SHP gene in various human tissues. SHP mRNA was detected in spleen and small intestine among human adult tissues examined (Fig. 4A). In addition, SHP gene expression was also detected in fetal liver and fetal adrenal gland (Fig. 4B). These results confirm and extend previous results with the mouse (8) and rat (9) genes, which indicated a relatively tissue specific pattern of expression. To identify the transcription start site for the SHP gene, primer extension analysis was used with HepG2 mRNA as a template (Fig. 5A). The apparent start site of transcription identified by these studies lies 32 nucleotides downstream from a consensus TATA box. This TATA motif is present within both the human and murine sequences, which are relatively well conserved over the 600 base pairs of 5Ј-flanking region (Fig. 5B) and match the 5Ј-end of the longest cDNA clone isolated previously (8). To confirm that the 5Ј-flanking sequences of the SHP gene can confer promoter activity, an approximately 2-kilobase fragment of the mouse sequences was inserted into a luciferase reporter construct. As expected from the SHP ex-pression in the adrenal gland, this construct showed significantly higher basal activity in the adrenal-derived Y1 cells than HeLa cells (Fig. 6). Finally, a series of 5Ј-deletion mutants were constructed to locate sequences required for the promoter activity in Y1 cells. As indicated in Fig. 6, full activity was observed with rather short fragment from Ϫ139 to ϩ22, with a further deletion to Ϫ68 showing significantly decreased expression. Thus, these results map the minimal SHP sequences necessary for the promoter activity in adrenal cells to approximately Ϫ139 to ϩ22. The specific transcription factors involved in this activity remain unclear. DISCUSSION Here, we characterized the structure and expression of the SHP gene. As it was the case with DAX-1 (14), the SHP gene consists of two exons with a single intron located near the C terminus. Since this position lies within a relatively well conserved region of the ligand binding domain, it will be interesting to examine whether other receptor genes also have an analogous intron. Accordingly, several receptors, including CAR (15), vitamin D receptor (16), and thyroid hormone receptor (17), have been shown to contain exactly analogous insertions in the conserved aspartate residue. In addition, the steroid receptors GR (glucocorticoid receptor), MR (mineralocorticoid receptor), AR (androgen receptor) and PR (progesterone receptor) were shown to share an insertion in what appears to be an analogous position (18). However, this intron is not present in every receptor genes known (e.g. the Drosophila orphan E75 (19)), as expected from the generally observed intron loss over evolutionary time spans. However, the quite extensive divergence of the putative ligand binding domains of SHP and DAX-1 from those of other nuclear hormone receptor members (8) suggest that this intron must have been present in the earliest members of the nuclear receptor superfamily. Since superfamily members are found in species ranging from Caenorhabditis elegans to mammals, this intron must be quite ancient.
Based on the results of genomic Southern blot and fluorescence in situ hybridization analysis, we concluded that the SHP gene is located at a single locus on the human chromosome one at position 1p36.1. At least two interesting functions map to this region: the tumor suppressor gene BCDS2, associated with ductal carcinoma of the breast (20), and an apparent tumor suppressor associated with cutaneous malignant melanomadysplastic nevus (21). Other studies indicate a loss of heterozygosity in this region in various tumors (e.g. Ref. 22). However, neither its postulated functions nor its limited pattern of expression suggest that SHP is a strong candidate for a tumor suppressor.
In addition to the previous results suggesting the SHP expression in liver (8), we observed specific expressions in small intestine, spleen, ovary, adrenal gland, and testis ( Fig. 4 and data not shown). In particular, the expression in adrenal gland, ovary, and testis overlaps with that of the SHP relative, DAX-1, which also lacks the conventional DBD. A recent report suggests that DAX-1 inhibits transactivations by the orphan receptor SF-1, which is also expressed in these tissues (23). Intriguingly, preliminary results suggest that SHP shows a functional synergy with SF-1. 2 Thus, it will be interesting to compare the developmental patterns of SHP expression with those of DAX-1 and SF-1 (24).
In agreement with previous results with the rat SHP gene (9), the primer extension analysis reported here indicate that SHP transcription is initiated 32 nucleotides downstream of a consensus TATA box. The transient transfection results indicate that a relatively short fragment encompassing this region confers promoter activity in Y1 cells but not in HeLa cells. The molecular basis for this difference in expression, which is consistent with the expression of SHP transcripts in adrenal gland, remains unknown. However, another potential linkage to SF-1 is provided from the presence of several consensus SF-1 binding sites in both the human and mouse promoter regions. Although this would be consistent with the expressions of SF-1 in Y1 cells, most of these sites lie outside of the minimal region required for basal promoter activity in these cells. Further work will be necessary to determine the role of SF-1 and other factors in the SHP expression.