Address correspondence and reprint requests to Dr. V. Matre at Department of Biochemistry, University of Oslo, P. O. Box 1041 Blindern, 0316 Oslo, Norway.
Abstract : The thyrotropin-releasing hormone (TRH) receptor (TRHR) is widely distributed throughout the central and peripheral nervous systems. In addition to its role in controlling the synthesis and secretion of thyroid-stimulating hormone and prolactin from the anterior pituitary, TRH is believed to act as a neurotransmitter as well as a neuromodulator. We have isolated genomic λ and P1-derived artificial chromosome clones encoding the human TRHR. The gene was found to be 35 kb with three exons and two introns. A 541-bp intron 1 (-629 to -89 relative to the translation start site) is conserved between human and mouse. A large intron 2 of 31 kb disrupts the open reading frame (starting in position +790) in the sequence encoding the supposed junction between the third intracellular loop and the putative sixth transmembrane domain. A similar intron was found in chimpanzee and sheep but not in rat and mouse. Promoter analysis of upstream regions demonstrated cell type-specific reporter activation, and sequencing of 2.5 kb of the promoter revealed putative cis-acting regulatory elements for several transcription factors that may contribute to the regulation of the TRHR gene expression. Functional analysis of potential response elements for the anterior pituitary-specific transcription factor Pit-1 revealed cell type-specific binding that was competed out with a Pit-1 response element from the GH gene promoter.
The thyrotropin-releasing hormone (TRH) receptor (TRHR) belongs to the superfamily of G protein-coupled, seven-transmembrane domain receptors Strader et al., (1995). In response to binding of ligand, these receptors signal the activation of a variety of cellular processes. The ligand of TRHR is TRH, a small neuropeptide with the sequence pGlu-His-Pro-NH2. The existence of a family of thyrotropin-related peptides occurring in both the animal and the plant kingdoms indicates that thyrotropin-related peptides have a wide phylogenetic distribution and probably have acquired diverse functions during evolution (Jackson and Reichlin, 1977 ; Lackey, 1992). In the hypothalamohypophyseal portal system in mammals, TRH is synthesized and released from the hypothalamus and transported via the portal vascular system to the anterior pituitary (Vale et al., 1977 ; O'Leary and O'Connor, 1995 ; Gershengorn and Osman, 1996). TRH plays an essential role in controlling the synthesis and secretion of thyroid-stimulating hormone and prolactin (PRL) from the anterior pituitary. TRH and its receptor are widely distributed throughout the central and peripheral nervous systems as well as in extraneural tissues (Sharif, 1985 ; Calza et al., 1992 ; Kaji et al., 1993 ; Zabavnik et al., 1993 ; Satoh et al., 1993a ; Fliers et al., 1994 ; Fukusumi et al., 1995). The signal pathways elicited by TRHR have been studied in PRL-producing pituitary adenoma cells (GH cells) where TRH has been found to exert a bifunctional effect by activating phospholipase C probably via the GTP-binding protein Gq/11 (Aragay et al., 1992 ; Hsieh and Martin, 1992) in addition to coupling to adenylyl cyclase via Gαs or a Gαs-like protein (Paulssen et al., 1991, 1992).
We report here the cloning and organization of the gene encoding the human TRHR. A prominent feature of the gene is a large intron of 31 kb interrupting the open reading frame (ORF) in the encoded junction between the third intracellular loop and the putative sixth transmembrane domain. We have also detected a similar intron in chimpanzee and sheep. An additional intron was uncovered in the 5′-untranslated region (5′-UTR). We detected promoter activity within a fragment of 770 bp upstream of the putative translation start site and analyzed cell type specificity for the TRHR gene promoter activity. Sequencing of 2.5 kb of promoter region DNA allowed the identification of elements that may contribute to the regulation of the TRHR gene expression. Two potential recognition elements for the Pit-1 transcription factor were further tested in competitive DNA-binding experiments with extracts from pituitary adenoma cell lines.
MATERIALS AND METHODS
Genomic library screening
Human placenta HL 1067j genomic λ library was screened with probes representing different parts of the human TRHR cDNA. The “5′-cDNA probe” spans the region from -12 (HindIII) to +234 (StuI) of the human TRHR cDNA, whereas the “3′-cDNA probe” spans from +862 (HincII) to +1,405 (BglII) (Matre et al., 1993). The libraries were plated at a density of 30,000 plaques/plate. Approximately 0.7-1 × 106 plaques were screened. The library was screened using standard methods (Sambrook et al., 1989). Positive plaques were purified to homogeneity by three rounds of screening before DNA was isolated using the Qiagen Lambda Kit.
Isolation of P1-derived artificial chromosome (PAC) clones harboring human TRHR gene
A central fragment of the cDNA [position +528 (EcoRV) to +1,105 (EcoRV)] encoding the human TRHR (Matre et al., 1993) was used to screen a human genomic PAC library (Ioannou et al., 1994) using the screening services of Genome-Systems (St. Louis, MO, U.S.A.). The library used was the Pieter de Jong (Roswell Cancer Institute) human PAC library derived from male leukocytes. The screening resulted in two clones, pGS8292 (with clone address PAC-18-F19 and GS control no. 8292) and pGS8293 (with clone address PAC-241-N11 and GS control no. 8293). Plasmids were prepared from small cultures (3 ml) of these clones using a modified alkaline purification protocol (Sambrook et al., 1989 ; Ioannou and Jong, 1996).
Restriction analysis of genomic DNA
Cloned genomic DNA was digested with restriction enzymes and electrophoresed on 0.8% agarose gels (λ DNA) or on 0.5% SeaKem Gold agarose gels (PAC DNA). The PAC clones were also analyzed by pulsed field gel electrophoresis (Bio-Rad Chef Mapper) in 1% agarose gels. The separation was 2-200 kb on 6-V/cm gradient (switch 0.1-17.33 s). Run time was 10.5 h.
Southern blotting and hybridization
The DNAs in the gels were transferred to Hybond nylon membranes by passive diffusion and immobilized (Sambrook et al., 1989). The hybridizations with the 5′-cDNA probe and the 3′-cDNA probe were performed according to Church and Gilbert (1984). When the 3′-UTR oligo probe (sequence : 5′-TCTAAGTCAGCCAAAATCCTGGTA-3′) was used, the hybridization temperature was reduced to 45°C and the washing conditions adjusted accordingly.
DNA sequence analysis
DNA fragments from restriction enzyme-digested genomic clones were subcloned into the pBluescript II SK(+) (Stratagene). Manual DNA-sequencing reactions were performed using the Sequenase kit (U.S. Biochemical). Automatic sequencing was performed on an ALFexpress system (Pharmacia) and on an ABI 373 sequencer. Sequence analysis was performed using the Wisconsin Sequence Analysis Package Program (version 8.1 ; Genetics Computer Group, Madison, WI, U.S.A.).
PCR and RT-PCR
PCR analysis was performed using standard procedures (Sambrook et al., 1989) employing Vent DNA polymerase (New England Biolabs) and appropriate primers. Human CNS mRNA was extracted from tissues obtained through the Netherlands Brain Bank and isolated by the method of Chirgwin et al. (1979). RT-PCR was performed following the protocol of Superscript Preamplification for first-strand cDNA synthesis (BRL). PCR was carried out according to the protocol of the GeneAmp DNA Amplification Reagent Kit (Perkin-Elmer) with minor modifications using 2 μl of the RT reaction and Vent DNA polymerase (New England Biolabs). The primers used were F1 with the sequence 5′-TTGGAAAGGGCTGTGAGGGTTTGGA-3′, F2 with the sequence 5′-GGGGAAATGTCATCTCAAAGGCTTGCTATTTT-3′, and R1 with the sequence 5′-CTATGACAGTCGACTTTAGAAGCTTAAAGTTTCTCGG-3′. For chimpanzee sequences, DNA extracted from two animals was kindly supplied by the Institute of Zoology, Zoological Society of London (U.K.). DNA from two sheep was extracted from tissue using standard methods. Primers were designed to match a human coding sequence upstream of intron 2 (nucleotide 376-394) and after ~30 bp into the human intron sequence. Primer sequences were 5′-GCAATCTGTCACCCCATC-3′ and 5′-AACTCTATCCCACATTTCCTCTAT-3′.
Analysis of promoter activity of TRHR upstream sequences by transfection of reporter constructs in pituitary adenoma cells
A fragment of the cloned human TRHR gene spanning the region from -1,080 to -1 relative to the initiator ATG was amplified by PCR using as forward primer an oligo with the sequence 5′-ctatagttgcatgc TCTAGACCCCTGGCTTATGA-3′ and as reverse primer an oligo with the sequence 5′-ctatgacagtcga CTTTAGAAGCTTAAAGTTTCTCGG-3′ (lowercase letters indicate sequences not present in the TRHR gene). The resulting 1.1-kb PCR product was digested and inserted into the pCAT basic vector (Promega) between the SphI and SalI sites, resulting in the plasmid pCAT-TRHR1.1. Two smaller plasmids were constructed using appropriate restriction sites ; pCATTRHR0.77 incorporating the region from -765 to -1 relative to the initiator ATG and pCAT-TRHR0.49 containing the region from -493 to -12. The pCAT-TRHR plasmids were transfected into the pituitary adenoma cell line GH4C1 and GH12C1 and also in the African green monkey kidney cell line COS-1. Six-centimeter plates (21 cm2) containing ~5 × 105 cells were transfected with 5 μg of reporter DNA using the liposomal transfection reagent DOTAP (Boehringer-Mannheim) as described by the manufacturer. A pSV-β-galactosidase expression vector (Promega) was included in these transfections for normalization purposes. Seventy-two hours after transfection, chloramphenicol acetyltransferase (CAT) activity was determined as described (Gorman et al., 1982). To estimate CAT activity, 30-60 μl of cell extract (typically containing 25-50 μg of protein) was added to 125 μl of reaction mix containing 250 mM Tris-HCl (pH 7.8), 0.5 mM acetyl-CoA, and 0.6 mCi of [14C]chloramphenicol. The mixture was incubated at 37°C for 2 h, after which another aliquot of acetyl-CoA was added. The final mixture was then incubated at 37°C overnight. Then [14C] chloramphenicol was extracted with 1 ml of ethyl acetate and processed as described (Gorman et al., 1982). The plates were exposed to a storage PhosphorImager screen overnight and scanned in a Molecular Dynamics PhosphorImager SI. β-Galactosidase activity was determined following the protocol of the β-galactosidase enzyme assay (TB 097 ; Promega) using 10-50 μl of cell extract.
Electrophoretic mobility shift assay (EMSA)
DNA binding was monitored by EMSA using whole-cell extracts prepared as described by Oelgeschläger et al. (1995). Two duplex oligonucleotide probes were derived from sequences in the TRHR gene promoter : Pit-1A : 5′-ctaagaggtG-TATTCATatttatttttc-3′ from position -421 to -394 and Pit-1B : 5′-ctgttaATGAATAtgtactatgac-3′ from position -121 to -98 (putative Pit-1-binding sites shown in uppercase letters). Labeled duplex oligonucleotides were obtained by end-labeling one DNA strand using polynucleotide kinase and [γ-32P] ATP. After annealing to the complementary strand, labeled duplex oligos were purified by polyacrylamide gel electrophoresis. Complex formation and gel analysis were performed as described (Oelgeschläger et al., 1995).
Isolation of λ clones containing human TRHR gene
To obtain λ clones containing the gene encoding the human TRHR, a human genomic DNA library was screened twice with two different cDNA fragments (Matre et al., 1993) : one encoding the first 78 amino acids (5′-cDNA probe) and the other encoding the last 111 amino acids of the human TRHR (3′-cDNA probe). Analysis of selected clones from each screen (λ10 and λ 12, respectively) resulted in the deduced map shown in Fig. 1B. Using labeled probes from the far downstream region of λ10 or from the far upstream region of λ12, we were not able to identify any overlapping regions in the two λ clones, despite an insert in the λ10 clone extending 7.3 kb downstream of the ATG initiator codon and an insert in the λ 12 clone extending 8.5 kb upstream of the translation stop codon. Sequencing verified that the ORF was interrupted by an intron starting in position +790, as previously suggested (Matre et al., 1993). Hence, the TRHR ORF is divided between two exons separated by an intron of >14.6 kb. These exons turned out to be exon 2 and exon 3 (see below).
Since the λ10 clone contained 9.3 kb of DNA upstream of the initiator ATG, it most probably incorporates the promoter of the gene. A detailed restriction map of this clone was derived by Southern blot analysis (Fig. 1A). Based on this map, several subclones were isolated and sequenced. A site for the rare-cutting restriction enzyme SacII identified in position -819 bp was taken as a point of reference in the subsequent long-range mapping of the gene (Fig. 1A-C).
Isolation of PAC clones containing human TRHR gene
To further examine the size of the large intron that interrupts the coding region, we decided to screen a human PAC library containing inserts in the 100 to 200-kb range. Two PAC clones positive for the 5′-cDNA probe were isolated (designated pGS8292 and pGS8293). One of these was found by colony hybridization also to contain sequences complementary to the 3′-cDNA probe (pGS8292). The pGS8292 clone would consequently span the entire central intron that interrupts the coding region of TRHR. With this clone in hand, a long-range mapping of the human TRHR gene was undertaken.
Characterization of PAC clone
The PAC clone pGS8292 was mapped by digestion with rare-cutting restriction enzymes and analyzed by pulsed field gel electrophoresis combined with Southern blotting (Fig. 2). A map was constructed on the basis of restriction fragments obtained with NotI, SalI, SacII, PacI, and SmaI (Fig. 1C). The insert was excised from the pGS8292 plasmid by NotI and its size estimated to be 174 kb, hybridizing with both the 5′-cDNA probe and the 3′-cDNA probe (Fig. 2). The insert contained a single SacII site 0.8 kb upstream of the translation start site that facilitated the localization of exon 1.
The positioning of an SmaI site 74 kb into the insert was critical for estimating the intron size. To precisely locate the end of the ORF, we took advantage of a PacI site identified by sequencing 665 bp downstream of the translation stop codon (PacI at 61 kb in Fig. 1C). An oligonucleotide probe was made matching the sequence just downstream of this PacI site in the 3′-UTR region. This probe hybridized to a 17-kb PacI fragment in the PAC clone that was reduced in size to 13 kb in a double digest with PacI + SmaI (Fig. 2). This places the PacI site 13 kb upstream of the SmaI site and by deduction also locates the stop codon in the ORF 13.7 kb (13 + 0.665) upstream of the SmaI site, as illustrated in Fig. 1C. From this we conclude that the large central intron (intron 2 ; see below) in the human TRHR gene has a size of 31 kb and that the size of the entire gene is 35 kb.
Detection of intron 2 of TRHR in chimpanzee and sheep genomic DNA
The mapped intron 2 seemed at first to be a unique feature of the human TRHR as no intron was found in the ORF of the mouse and rat TRHR genes (de la Pena et al., 1992b ; Duthie et al., 1993 ; Satoh et al., 1993b). To extend the evolutionary perspective of our finding, we set out to determine whether the large intron 2 of the TRHR gene was present in chimpanzee and sheep genomic DNA. Primers were designed to generate a PCR product spanning the human splice site between exon 2 and intron 2. PCR products were obtained from both animal genomic DNAs having the expected length of 464 bp. The PCR products were sequenced and aligned as shown in Fig. 3. They display high homology, particularly in the exon 2 region prior to the indicated start of the intron. In the last row, the mouse cDNA sequence (Straub et al., 1990) is also shown to illustrate the major change in homology observed after the start of intron 2 in the human gene. No PCR products could be detected with primers designed to span across the entire intron 2 in chimpanzee and sheep genomic DNA. The presence of sequences equivalent to exon 3 of human TRHR in chimpanzee and sheep genomic DNA was shown by sequencing PCR amplification products generated with primers, both of which were complementary to human exon 3 (R. P. Murray-McIntosh, Z. Velickovic, and D. Penny, unpublished results). It is concluded that sequences equivalent to human exon 2 and 3, and at least the 5′ end of intron 2, are present in the genomic DNA of chimpanzee and sheep.
Presence of intron 1 common among different species revealed by analysis of exon-intron organization of 5′-UTR
Due to limited cDNA sequence data from the 5′-UTR, we tried an alternative approach to provide information on the exon-intron organization in this region of the gene. A sequence comparison of rat cDNA (Sellar et al., 1993) and human genomic DNA using the Compare program of the Wisconsin Sequence Analysis Program (Genetics Computer Group) revealed a high degree of conservation between the human gene and the rat cDNA in the ′5-UTR. A dotplot revealed an abrupt interruption at position -89 by a stretch of 541 bp of DNA present only in the human genomic sequence. A closer inspection of the aligned sequences revealed consensus splicing signals at both ends of this extra sequence, as illustrated in Fig. 4A. This strongly suggests the presence of an intron of 541 bp in the 5′-UTR of the human gene.
Experimental verification of this conclusion was obtained by PCR. Employing two sets of PCR primers (F1 + R1 and F2 + R1, location given in Fig. 4A). we compared the PCR products obtained using as template either reverse-transcribed human hypothalamic mRNA or genomic TRHR DNA. The genomic DNA gave PCR products of the expected size of 795 bp for the F1 + R1 set (lane 2) and 580 bp for the F2 + R1 set (lane 4). Both fragments hybridized with a TRHR promoter-specific probe spanning the entire upstream region (from -1 to -1,080), as illustrated in Fig. 4B, as well as with an intron-specific probe (results not shown). On the other hand, the RT-PCR product obtained with the F1 + R1 primer set was reduced to 250 bp, consistent with the loss of a 541-bp spliced-out intron (Fig. 4B, lane 3). This PCR product hybridized with the promoter-specific probe (Fig. 4B, lane 3) but gave no signal with an intron 1-specific probe (results not shown). The intron-specific primer set F2 + R1 gave no visible PCR product after amplification of reverse-transcribed mRNA (Fig. 4B, lane 5). These observations demonstrate the presence in the 5′-UTR of an intron of 541 bp, as first expected.
Analysis of cell type-specific promoter activity
To analyze the putative promoter region of the TRHR gene, a total of 2.5 kb of λ10 DNA upstream of the initiator ATG was sequenced, as shown in Fig. 5. This sequence contained five TATA sequences (underlined in Fig. 5), two of which (-1,588 TATAGAT and -1,458 TATAAAG) also conform well with the extended consensus sequence TATAWAW (Gannon et al., 1979). However, no convincing candidate transcription start site was identified using the Proscan II program (Prestridge, 1995). The previously isolated human TRHR cDNAs were too short to indicate a likely transcription initiation region. We therefore attempted to determine the transcription start sites using either rapid amplification of cDNA ends or primer extension procedures, but a single start was not detected.
The exact location of the human 5′-UTR intron 1 compared with the length of the rat cDNA implicated a putative transcription start site located ≥880 bp upstream of the translation initiation site. No TATA-related sequences, however, were found close to this region.
To analyze directly whether the upstream region had promoter activity, three CAT reporter plasmids were constructed with reporter gene fused to different segments of the putative TRHR promoter region. The plasmid pCAT-TRHR1.1 had its start point at position -1,080, upstream of the expected translation start site. pCAT-TRHR0.77 started at position -765 in exon 1, and pCAT-TRHR0.49 had its start point in the first intron at position -493 upstream of expected ATG. The pCAT-TRHR plasmids were transfected into the pituitary adenoma cell lines (GH4C1 and GH1C1) and COS-1 cells. Promoter activity was measured as CAT activity normalized to β-galactosidase activity derived from a co-transfected reference plasmid. As shown in Fig. 6, significant CAT reporter activity was obtained in the pituitary but not in COS-1 cells with the two largest upstream segments tested, demonstrating that the analyzed regions indeed contained a functional promoter. No activity was found with pCAT-TRHR0.49, indicating that no transcription is initiated in the intron downstream of position -493 and that the TATA box in -439 TATACTT is probably not involved in transcription initiation. The fact that pCAT-TRHR1.1 had the highest reporter activity lends support to the presumed transcription start site at -880 and suggests that the TATA boxes at -1,458 and -1,588 are not necessary for promoter activity. Interestingly, a considerable reporter activity was obtained with pCAT-TRHR0.77, suggesting that additional transcription start sites have to be present downstream of the -880 region. The promoter activity was indeed cell specific, present only in pituitary adenoma cells but not in COS-1 (African green monkey kidney) cells.
We analyzed the entire sequence upstream of the initiator ATG for potential regulatory cis-elements that may act as binding sites for different transcription factors. Comparison of 2.5 kb of the 5′ flanking sequence with a transcription factor database revealed putative cis-acting regulatory elements for transcription factors like Pit-1, C/EBP, Ap-1, GR, and Myb, as illustrated in Fig. 7.
To further examine cell type-determining response elements, two potential recognition sites for the Pit-1 transcription factor were tested in an EMSA using whole-cell extracts from GH4C1 and COS-1 cells (Fig. 8). Two oligos were labeled containing the Pit-1 site A, from -412 to -405, and Pit-1B, from -115 to -108 (sequences listed in Materials and Methods). The EMSA analysis revealed two shifted bands for each of the Pit-1 oligos when incubated with cell extracts from GH4C1 cells, but only very weak shifts could be seen with cell extracts from COS-1 cells. Stronger binding was obtained with the Pit-1B probe than with Pit-1A. Upon incubation with a 50-fold excess of an unrelated oligo, the shifted Pit-1 bands remained unaffected, whereas the weak shifts observed with COS-1 extracts disappeared. Upon incubation with a 50-fold excess of an unrelated oligo, harboring a well characterized consensus Pit-1 site from the human GH promoter (Theill et al., 1989), the slowest Pit-1 band was severely affected, strongly indicating that the Pit-1 transcription factor is involved in this complex. Also, the second shifted band was reduced in intensity upon competition with a Pit-1 consensus oligo, but clearly less than the first one, suggesting that this shift might be nonspecific or involve another factor with some affinity for Pit-1 sites. These data are consistent with the involvement of pituitary-specific Pit-1 factor in binding to intronic cis-elements in the TRHR promoter contributing to the cell type-specific expression of TRHR.
We have isolated genomic clones corresponding to the entire amino acid coding sequence of the human TRHR, determined the exon-intron organization of the gene, and analyzed basal and cell type-specific promoter activity. We have also identified putative binding sites for transcriptional regulatory factors in the promoter region and analyzed two putative Pit-1 response elements in more detail.
A prominent feature of the gene is the presence of an ORF-disrupting intron in position + 790 (codon 263), as already suggested previously from comparison of TRHR cDNAs (Matre et al., 1993). This intron was too large to be contained in any λ clones but was identified in a PAC clone with a 174-kb insert that both harbored the TRHR-encoding exons and spanned the entire ORF-disrupting intron. Mapping of this clone enabled us to estimate the intron size to 31 kb and calculate the size of the entire gene to 35 kb harboring three exons. Originally, it was believed that most G protein-coupled receptors (GPCRs) were encoded by single exons. However, accumulating sequence data from cloned GPCRs have now revealed several examples of receptors encoded by multiple exons. No consistent pattern of intron positions with respect to the structure of the receptors is apparent. The rat luteinizing hormone receptor (11 exons) has all introns located in the region encoding the amino-terminal domain (Koo et al., 1991). Examples of GPCRs with introns disrupting the sequence encoding the transmembrane domains are the family of human opsins (five-six exons) (Nathans et al., 1986), human tachykinin receptors (five exons) (Graham et al., 1991), and human bombesin receptors (three exons) (Corjay et al., 1991). Like TRHR, these have one of the introns interrupting the third intracellular loop, but in a different position. Certain dopamine receptors also have introns in the amino acid coding region (five-seven exons) (Sokoloff et al., 1990 ; Van Tol et al., 1991). In these receptors, the third intracellular loop is interrupted in several positions by three introns. The bovine ETB endothelin receptor (seven exons) resembles the TRHR in having one of its introns located in a position precisely at the end of the third intracellular loop (Mizuno et al., 1992).
Comparison of the genomic nucleotide sequences of human, mouse, and rat TRHR genes with those of the corresponding cDNAs reveals that the large intron occurs only in the human TRHR. In rodents, a single exon apparently encodes the entire ORF except for the very C-terminal amino acids (de la Pena et al., 1992b ; Duthie et al., 1993 ; Satoh et al., 1993b). The recent access to more cloned TRHR genes now suggests a pattern where the absence of the ORF-disrupting intron in the rat and mouse genes appears to be the exception rather than the rule. Our analysis of chimpanzee and sheep genomic DNA provided evidence for an intron in these TRHR genes located precisely as the human ORF-disrupting intron. The recent cloning of the bovine TRHR gene also revealed a large intron in this position (Takata et al., 1998). Furthermore, the sequence of some isoforms of TRHR in chicken also suggests that the avian TRHR gene is interrupted by an intron at a homologous position (Sun et al., 1998). Ox, sheep, chimpanzee, and humans evolved more recently than rat and mouse (Penny and Hasegawa, 1997). In light of the chicken data, it is tempting to hypothesize that the large intron was lost during development of the rodent branch and not inserted at a later stage in evolution.
The location of the ORF-disrupting intron in TRHR is intriguing because it interrupts the reading frame precisely in the sequence corresponding to the expected junction between the third intracellular loop and the putative sixth transmembrane domain. In general, the third intracellular loop is of critical importance for the function of GPCRs, being instrumental in coupling to G proteins (Strader et al., 1995) and containing several putative sites for phosphorylation (Lefkowitz, 1993). Furthermore, independent expression of the third intracellular loop may even act as a receptor-specific antagonist (Luttrell et al., 1993 ; Hawes et al., 1994). The existence of an intron at this position therefore raises the possibility of a truncated form of the TRHR with potentially distinct functions. Although truncated human isoforms remain to be identified, the chicken TRHR gene appears to encode several isoforms, some of which have truncations in the end of the third intracellular loop and a sequence consistent with termination in an intron. So far, no function was found for these isoforms (Sun et al., 1998).
Evidence for an additional intron 1 in the 5′-UTR was also found. First, by direct comparison of the rat cDNA sequence with the human genomic sequence, it was found that the homology was interrupted by an insert in the human sequence of a 541-bp region flanked by consensus splice signals. Second, a comparison of PCR products from genomic clones with RT-PCR products obtained with mRNA from healthy hypothalamic human tissue demonstrated the presence of this intron.
Several aspects of the human TRHR promoter were analyzed. Localization of putative transcription start sites was attempted by direct sequence analysis because of low expression levels of the receptor in CNS tissue that rendered mRNA-based identification difficult. Promoter activity was demonstrated in a 1.1-kb fragment upstream of initiator ATG harboring a suggested start site at about —890 (Iwasaki et al., 1996). These observations exclude three putative TATA boxes (located at -1,588, -1,458, and -1,393) as important for basal promoter activity. Also, the -439 TATA box candidate can probably be excluded as no CAT activity was demonstrated with the intron 1-based fragment of 0.49 kb. The basal promoter activity might therefore be directed by a TATA-less type of promoter or other remaining TATA box candidate that might be functional. Interestingly, we also found substantial promoter activity in GH4C1 cells using a reporter construct with start point in position -765, implying that start sites in addition to the -880 region must be operative. Transfecting different cell lines revealed a clear cell type-specific promoter activity with both active reporter constructs. This indicates that cis-elements determining cell type-specific activation must be located downstream of position -765.
We identified several putative response elements for different transcriptional activators in a region >2 kb upstream of the translation start site. Some of these are located in the first intron, a phenomenon seen in many regulated genes. The presence of glucocorticoid response elements is consistent with reports demonstrating up-regulation of TRHR mRNA in response to dexamethasone (Yang and Tashjian, 1993a). The AP-1 sites might be relevant for the 12-O-tetradecanoylphorbol 13-acetate response and the CRE (cyclic AMP-responsive element) sites relevant for the adenylyl cyclase coupling. It is clear that TRHR is regulated by many different signals responding to, e.g., estradiol (Kimura et al., 1994), 1,25-dihydroxyvitamin D3 (Atley et al., 1995), epidermal growth factor (Monden et al., 1995), thyroid hormone (Hollenberg et al., 1995 ; Schomburg and Bauer, 1995), and TRH itself (Yang and Tashjian, 1993b). The presence of a putative binding site for the pituitary-specific Pit-1 factor in the POU domain family is particularly interesting in relation to expression of the TRHR gene in distinct cell types. We therefore analyzed two Pit-1 candidate sites further and found that both bound to factors present in extracts from GH4C1 cells, but not in COS-1 cells, and that the binding was competed for by an oligo encoding a well characterized Pit-1 site in the human growth hormone promoter. We suggest, therefore, that binding of the Pit-1 transcription factor to these sequences in the intron sites of the TRHR gene contributes to the cell-specific expression of the receptor.
In conclusion, we have mapped the human TRHR gene in great detail and included a size determination of a large ORE-disrupting intron. The human exon-intron stucture is distinct from mouse and rat and may contribute to the understanding of the evolution of TRHR genes. The present study revealed several functional sites in the TRHR promoter and will provide a starting point for the analysis of how the expression of the TRHR gene is regulated through defined regulatory elements located partly in the first intron.