Induction of the intronic enhancer of the human ciliary neurotrophic factor receptor (CNTFRalpha) gene by the TR4 orphan receptor. A member of steroid receptor superfamily.

A conserved hormone response element, CNTFR-DR1 (5′-GGG-3′), has been identified in the 5th intron of the α component of the ciliary neurotrophic factor receptor (CNTFRα) gene for the human TR4 orphan receptor (TR4). Electrophoretic mobility shift assay showed a specific binding with high affinity (Kd = 0.066 nM) between TR4 and the CNTFR-DR1. A reporter gene assay using chloramphenicol acetyltransferase demonstrated that the 5th intron of CNTFRα has an enhancer activity which could be induced by TR4 in a dose-dependent manner. Furthermore, our in situ hybridization data showed that abundant TR4 transcripts were detected in adult brain, in regions of cortical and hippocampal neurons, as well as in many developing neural structures, including brain, spinal cord, ganglia (sympathetic and sensory), and neuronal epithelia (retinal, otic, olfactory, and gustatory). The striking similarities in the expression patterns of TR4 and CNTFRα in the developing and postnatal nervous systems further support the potential role of TR4 in neurogenesis. Collectively, these data suggest that the human CNTFRα gene could represent the first identified neural-specific gene induced by TR4.

Members of the steroid/thyroid hormone receptor superfamily are transcription factors that can bind to specific DNA sequences called hormone response elements (HRE) 1 and thereby regulate the expression of their target genes (1). This superfamily includes receptors for steroid, thyroid, vitamin D 3 , and retinoids, and a large number of orphan receptors whose cognate ligands are still unknown (1). These members have been grouped into three categories according to their binding preference to HREs, the palindromic half-site AGAACA is pre-ferred for the binding by the androgen, glucocorticoid, mineralocorticoid, and progesterone receptors; the direct repeat with various spacing is preferentially recognized by the estrogen, thyroid, retinoic acid, retinoid X, vitamin D, and many orphan receptors; and the single half-site of AGGTCA preceded by two specific flanking nucleotides is favored by some members which bind as a monomer, such as the steroidogenic factor 1 (2), TR3 orphan receptor (TR3)/NGFI-B/nur77 (3), and the thyroid receptor (4). Some receptors, i.e. thyroid hormone receptor and chicken ovalbumin upstream promoter-transcription factors (COUP-TFs), are shown to be promiscuous for binding to different arrangements of the AGGTCA half-site (4,5).
The human and rat TR4 cDNAs were isolated from testis and hypothalamus by degenerative PCR cloning (6). The TR4 cDNA shows high homology in nucleotide sequence with the TR2 orphan receptor (TR2; another orphan receptor isolated from our laboratory (7)), suggesting that these two orphan receptors constitute an unique subfamily within the steroid receptor superfamily. TR4 mRNA are widely expressed in the adult rat brain (6). Within the supraoptic nucleus, TR4 is one of the most abundant steroid receptors expressed there with the order thyroid hormone receptor Ͼ COUP-TFII Ͼ TR4 ϭ COUP-TFI (8). Despite the high abundance of the TR4 transcripts in nervous tissues, the physiological function of TR4 in neurogenesis remains unclear.
To understand the role of TR4 in neurogenesis, we sought downstream regulatory events, namely target genes, which can be regulated by TR4. Thus, we examined the TR4 expression pattern during mouse embryogenesis to further confirm its participation in neuronal development, and then used this information to search for TR4 target genes involved in a neuronal-specific program. We then sought potential TR4 target genes based on the binding preference of TR4. In vitro binding assays suggested that TR4 is capable of binding to 5Ј-AGGTCA direct repeats with 1-6-base pair spacing (DR1-DR6). 2 We hypothesized that the ␣ component of ciliary neurotrophic factor receptor (CNTFR␣) with DR1 in its 5th intron may be a target gene for TR4.
Ciliary neurotrophic factor (CNTF) is a member of the cytokine superfamily. It utilizes a three-component receptor system consisting of an extracellular CNTF-binding protein, known as CNTFR␣ (9), as well as two signal transducing ␤ receptor subunits, gp130 and LIFR␤, which it shares with its cytokine relatives (10,11). CNTFR␣ is bound to the cell membrane by a glycosylphosphatidylinositol anchor and its major function is to convey ligand-specificity (9). However, upon certain stimula-tion, CNTFR␣ can also be released from its glycosylphosphatidylinositol anchor and mimics the effect of CNTF on target gene activation (12). The expression of CNTFR␣ is mostly restricted to nervous tissues and is detected in all neurons that have been shown to respond to CNTF (13)(14)(15)(16). Clinically, CNTF effectively protects against motor neuron degeneration in human Parkinson's disease (17). The trophic effect of CNTF is also observed on the denervated skeletal muscle (18). Consistent with the CNTF effect on motor neurons, the essential function of the CNTFR␣ gene has been proved in mice lacking CNTFR␣, which exhibit profound motor neuron deficits at birth (19). Despite the important function of CNTFR␣, the regulation of the CNTFR␣ gene remain almost completely unknown.
Because little is known about TR4 targets, we focused study here on a DR1 element present in CNTFR␣ which can be potentially regulated by TR4. Thus, the 5th intron of the CNTFR␣ gene (CNTFR-I5) was cloned to show the DR1 response element, where TR4 binds, is conserved and functional. The consequence for TR4 interaction with this DR1 in its natural gene context was determined using a reporter gene assay. Moreover, the expression pattern of TR4 during mouse embryogenesis was examined by in situ hybridization and then compared to that of CNTFR␣. Through these studies, our results suggest TR4 may function as an inducer in CNTFR␣ gene regulation and this event could happen in vivo. Thus, this study provides the first evidence for TR4 to interact with the neurocytokine signaling pathway during neurogenesis.

MATERIALS AND METHODS
Cloning of the 5th Intron of CNTFR␣ Gene-To clone the DNA fragment containing the 5th intron of the CNTFR␣ gene, polymerase chain reaction (PCR) was employed. Two oligonucleotides (E5: 5Ј-CAC-CTTCAATGTGACTGTGC-3Ј, E6: 5Ј-GTGCATGTAGCGAATGTGGC-3Ј) located in the coding regions flanking the 5th intron of human CNTFR␣ were synthesized for PCR amplification. The positions of E5 and E6 primers are 417-436 and 510 -491, respectively (20). Purified genomic DNA (1 g) isolated from several cell lines (human 1299 and Chinese hamster ovary) and from mouse tail were used as the templates. The following thermal cycling program was used to amplify the DNA fragment with the expected size around 270 base pairs: denaturation at 95°C for 1 min, annealing at 50°C for 1 min, and extension at 72°C for 1 min, for 35 cycles. The amplified DNA fragments (named as CNTFR-I5) from various template sources were gel purified and subcloned into pT7 blue vector (Novagen). Sequence analysis using the Sequenase kit (U. S. Biochemical Corp.) confirmed the identity of these fragments.
Reporter Constructions-To test for enhancer activity, pCAT-promoter vector (Promega), which contains an SV40 promoter upstream from the CAT gene, was used and the human CNTFR-I5 DNA fragments (from PCR) were inserted into this vector. Plasmids pCNTFRI5-CATe5ϩ and pCNTFRI5-CATe5Ϫ both contain the CNTFR-I5 insert at the BglII site of the pCAT-promoter vector with the orientation the same or opposite to CAT expression, respectively. Plasmid pCNTFRI5-CATe3ϩ inserted the CNTFR-I5 at the XbaI site of the pCAT-promoter vector with the orientation the same as CAT gene expression. The identities of these constructs were confirmed by DNA sequencing.
Coupled in Vitro Transcription and Translation-Plasmid pCMX-TR4 (21) containing the full-length human TR4 cDNA was in vitro transcribed and translated by the TNT system according to the manufacturer's instructions (Promega). The product of TR4 was analyzed in SDS-polyacrylamide gel electrophoresis and showed a major band with an expected molecular weight of 67.3 kilodaltons as described previously (21).
Electrophoretic Mobility Shift Assay (EMSA)-EMSA was carried out as described previously (22). Briefly, 0.05 l of TNT-expressed TR4 was included in each reaction. Double-stranded oligonucleotides corresponding to the human CNTFR-DR1 (5Ј-GCCCTGACCTCTGAC-CTCTC-3Ј) were end labeled by [␥-32 P]ATP. The mutated CNTFR-DR1 oligonucleotides (5Ј-GCCCTGAGCTCTGAGCTCTC-3Ј) contained two mutated nucleotides (letters underlined). For competition reactions, unlabeled double-stranded oligonucleotides were mixed with the la-beled probe prior to addition to the reactions. For antibody supershift analysis, 1 l of monoclonal number 29 anti-TR4 antibody 2 or 1 l of unrelated monoclonal antibody (IgG 2 isotype) was incubated with the reactions for 15 min at room temperature prior to loading on a 5% native gel.
Scatchard Analysis-The DNA-protein binding assay was performed as described previously with modifications (22). Briefly, 0.05 l of in vitro translated TR4 was incubated with various concentrations of the probe. Protein-DNA complexes were resolved on a 5% non-denaturing polyacrylamide gel (at 4°C) in 0.5 ϫ TBE. After exposing to a film to localize the free probe and protein-probe complex, the respective bands were excised and counted directly in a scintillation counter. The dissociation constant (K d ) value and B max were calculated using the Ebda program (Biosoft).
Cell Culture, Transfection, and CAT Assay-Chinese hamster ovary cells were cultured and transfected by the calcium phosphate coprecipitation procedure as described previously (22). To normalize the transfection efficiency, the pCMV␤ (Clontech) was co-transfected. Results were plotted as the mean Ϯ S.D. of at least three independent experiments of CAT expression normalized to ␤-galactosidase activity.
Cloning Mouse TR4 cDNA-To clone the full-length mouse TR4 cDNA, reverse-transcription followed by PCR (RT-PCR) was employed, following the standard protocol recommended by the manufacturer (Perkin-Elmer). Mouse testis total RNA (1 g) was used as a template. The N-and C-terminal mouse TR4 cDNA fragments were separately amplified using the primer sets TR4-16 and TR4-15 or TR4-3 and TR4-42, respectively. The primer sequences are as followings: TR4 -16 TR4 -42 (5Ј-CTATAGACTGACTCCGGTGATCTG-3Ј). The identity of PCR products was confirmed by sequencing and showed high homology to human counterparts. These two cDNA fragments were then ligated together to produce a full-length mouse TR4 cDNA.
Northern Blot Hybridization-Total RNAs from selected mouse tissues were isolated, electrophoresed, and transferred onto a nylon membrane as described previously (6). The blot was hybridized with the mouse mTR4N probe, washed, dried, exposed to PhosphorImager intensifying screen for 16 h. The image was scanned and printed.
In Situ Hybridization Analysis-Embryo collection, section preparation, and in situ hybridization were performed as described previously (22). Both the antisense and sense TR4 probes were included for each batch of experiments. Washes were performed with high stringency (2 ϫ SSC, 50% formamide at 65°C (1 ϫ SSC ϭ 0.15 M NaCl, 0.015 M sodium citrate, pH 7.0)) before and after RNase digestion (20 mg/ml for 30 min). Slides were dipped into Kodak NTB2 emulsion and exposed for 2 weeks. Subsequently, slides were developed in Kodak D19 developer, fixed, dehydrated, and mounted for dark field analysis. Some slides were stained with hematoxylin for light-field analysis.

RESULTS
Cloning of the 5th Intron of the CNTFR␣ Gene-Based on the genomic organization and DNA sequence published by Valenzuela et al. (20), the 5th intron of the human CNTFR␣ gene (CNTFR-I5) contains one perfect DR1 (5Ј-AGGTCACAGGTCA) and four copies of AGGTCA-like half-sites. It is especially amazing because so many AGGTCA-like sequences are simutaneously localized within a very small intron of around 176 base pair. We wondered whether some of these sites might be to various amounts of the probe was resolved by EMSA. Constant amounts of TR4 were incubated with varying concentrations of labeled CNTFR-DR1 probe, as indicated. The specific DNA-protein complex (arrowhead) and free probe were detected by autoradiography, excised from the gel, and counted directly in a scintillation counter. C, Scatchard plot of the result is shown. The ratio between specific DNA-protein binding and free DNA probe (bound/free) with respect to specific DNA-protein binding (bound, nM) was plotted. The dissociation constant (K d ) and B max values were generated from the Ebda program (Biosoft).
HREs regulated by steroid receptors. As shown in Fig. 1A, CNTFR-I5 is located in the middle of the coding sequence for the first fibronectin-like subdomain within the cytokine receptor-like domain of the CNTFR␣ gene ( Fig. 1A) (20). The presence and positions of introns in the cytokine receptor-like domain are conserved among members in the cytokine receptor superfamily (20), implying CNTFR-I5 may contribute a regulatory role through these response elements. To test this hypothesis, we cloned the CNTFR-I5 from human, hamster, and mouse by PCR (see "Materials and Methods"). Sequence comparisons between these three species showed that the DR1 and putative thyroid hormone receptor-binding site (TRE-like sequence) (4) were conserved but the TR3/NGFI-B/nur77-binding site (3) has one mismatch at its end (Fig. 1B). The other two AGGTCA-like sequences were not well conserved. The appearance of three different types of highly conserved HREs within a small intron suggests that CNTFR-I5 may be critical for CNTFR␣ gene regulation.
TR4 Binds Specifically to the CNTFR-DR1 with High Affinity-To examine whether CNTFR-DR1 could bind TR4, we carried out in vitro DNA binding experiments. Gel retardation assays were performed with in vitro translated TR4 using the 32 P-labeled CNTFR-DR1 oligonucleotide as a probe. As shown in Fig. 2A, a specific DNA-protein complex was formed in the presence of both probe and TR4 (lane 3, hollow arrow) but was absent in the reaction containing probe and the mock-translated control (lane 2). This TR4-CNTFR-DR1 complex could be essentially abolished by as low as a 10-fold molar excess of unlabeled CNTFR-DR1 oligonucleotide (lane 4), but remained intact in the presence of a 100-fold molar excess of "mutant" CNTFR-DR1 oligonucleotide (lane 6). Moreover, this retarded complex could be fully supershifted in the presence of the anti-TR4 monoclonal antibody (lane 8, solid arrow). As a negative control, an unrelated monoclonal antibody with the same IgG 2 isotype as that of TR4 showed no effect on the retarded complex (lane 7). Together, these data indicate that the CNTFR-DR1 is a specific binding site for TR4.
To determine the binding affinity of TR4 and CNTFR-DR1, we performed Scatchard analysis by EMSA. The typical EMSA pattern of protein-DNA complex formed between increasing amounts of CNTFR-DR1 probe (0.0039 -2 ng) and fixed amounts of TR4 was shown (Fig. 2B). The radioactivity of specific complex (bound) and unbound (free) probe were quantitated for the subsequent Scatchard plot analysis. The results are consistent with a single binding population for the specific DNA-protein complex with a dissociation constant (K d ) of 0.066 nM and B max of 0.089 nM (Fig. 2C). This binding affinity is about 15-45-fold higher than the K d range for steroid receptors and their HREs (4). At very low probe concentrations, the specific protein-DNA complex was still visible (Fig. 2B, lanes 2-4). That was further consistent with the calculated high affinity dissociation constant.
Enhancer Activity of CNTFR-I5 Induced by TR4 -To investigate whether TR4 could regulate the CNTFR␣ gene expression through interaction with CNTFR-I5, we carried out the CAT assay following co-transfection of expression vectors and CAT reporter constructs into Chinese hamster ovary cells. Three enhancer-reporter plasmids were created in the present study to test whether the enhancer activity of CNTFR-I5 is position-or orientation-dependent (Fig. 3A). As shown in  13 and 14 to lane 12). In contrast, this induction did not occur when the antisense TR4 expression vector was transfected ( lanes 5, 6, 10, 11, 15, and 16). Different orientations or posi-tions did not appear to affect the TR4-mediated transcriptional activity. TR4 does not significantly induce the CAT reporter in the absence of the DR1 (22). Taken together, these results suggest that TR4 might induce CNTFR␣ gene expression via the CNTFR-I5 enhancer.
The N-terminal Mouse TR4 Probe Is Highly Specific for Northern Blot and in Situ Hybridization Analyses-The entire coding region of mouse TR4 cDNA was cloned by RT-PCR (for details see "Materials and Methods") (Fig. 4A). Sequence comparison with other members of the steroid receptor superfamily indicated that the N-terminal mouse TR4 cDNA fragment (mTR4N) covered between primer TR4-16 and TR4-23 is the most divergent; the homology within this region is around 30% at nucleotide level between mouse TR4 and TR2. 3 To test the specificity of mTR4N, total RNA samples isolated from mouse FIG. 3. Functional analysis of the CNTFR-I5 gene fragment as regulated by TR4. A, construction of the enhancer reporter plasmids containing CNTFR-I5. Names assigned to each construct are as indicated. CNTFR-I5 DNA fragments were inserted into either upstream (e"5") or downstream (e"3") of the vector pCAT-promoter (pCAT-). Relative orientations of enhancer fragments with respect to the promoter are indicated by arrows; also indicated as "ϩ" or as "Ϫ" for the same or opposite direction to the promoter, respectively. B, dosage dependent induction of CAT activities by TR4 via CNTFR-I5. Various reporter constructs without (lanes 1, 2, 7, and 12) or with expression vector were transfected into Chinese hamster ovary cells. The reporter constructs for transfection were either pCATp (lane 1) or CNTFR-I5-CATe5ϩ (lanes 2-6) or CNTFR-I5-CATe5Ϫ (lanes 7-11) or CNTFR-I5-CAT3ϩ (lanes [12][13][14][15][16]. Expression plasmids co-transfected include: sense TR4 expression plasmid, 0.5 g (lanes 3, 8, 13), 3 g (lanes 4, 9,14); or antisense TR4 expression plasmid, 0.5 g (lanes 5, 10, 15), 3 g (lanes 6,11,16). Chloramphenicol conversion rates were calculated from Phos-phorImager quantifiable intensities. Fold induction was normalized relative to the CAT activity produced by the control plasmid pCATp without the co-transfection with TR4 expression plasmid. The data are average of at least three independent experiments with the error bars representing standard deviation.
testis and kidney were used as positive and negative controls, respectively. As shown in Fig. 4B, the mTR4N probe hybridized to two bands, 7.8 and 2.8 kilobase in size, in mouse testis (Fig.  4B, lane 1). The sizes of these two TR4-hybridized bands are different from those of TR2 transcripts, which are 2.4 (major band) and 2.9 kb (minor band) in size. No hybridization signal was detected in the kidney (Fig. 4B, lane 2). These data are consistent with previous observations (6) supporting that the mTR4N probe is specific and only hybridizes to TR4 transcripts. Expression of TR4 in Adult Brain-To test whether TR4 could potentially regulate CNTFR␣ gene expression in vivo, we examined TR4 expression by in situ hybridization. As shown in Fig. 5A, a coronal section through the rostral region of the forebrain demonstrated the presence of TR4 transcripts within the hippocampus, habenula nuclei, and thalamus as well as throughout the cortical layers. The most intense signals were found in the cortex, piriform cortex (Fig. 5, B and C), regions of Ammon's horn (Fig. 5A, CA1, CA2, and CA3), and dentate gyrus (Fig. 5, D and E). Consistent with previous results, the signal within the dentate gyrus was localized to the granule cells, which undergo active postnatal neurogenesis (6). No signal was detected with the sense mTR4N probe (data not shown).

Prominent Expression of TR4 in Neuronal Precursors during Embryonic Development Supports a Role for This Factor in
Neurogenesis--Ours and previous in situ studies suggested that TR4 expression is related to the neural proliferation status. To further pursue this possibility, we examined TR4 expression during embryogenesis. In situ hybridization analysis in mouse embryos revealed that TR4 expression was most prominent in the developing neural structures as well as in certain non-neuronal regions (Fig. 6, B-E). TR4 transcripts were detected as early as embryonic day 9 (E9) (Fig. 6B). This result was confirmed by RT-PCR with a cycle number at the early amplification phase (data not shown). Initially within the neural system of E9-E11, TR4 transcripts were present uniformly throughout the neural tube along both its rostrocaudal and dorsoventral axes (Fig. 6, B and C, ne). Subsequently, during E14-E16, expression became restricted to the ventricular zones of brain vesicles, where cells are rapidly proliferating (23) (Fig. 6, D-F, arrow and arrowhead). Sense mTR4N probe did not detect specific signal in E14 embryo (Fig. 6A). Intensive TR4 transcripts in E16 brain were also observed in the striatum (Fig. 6F, Sr), active dividing population of the cerebellar primordium (Fig. 6E, cp), and spinal cord (Fig. 6D, sp). The spinal motor neurons (Fig. 6D, m) were also positive for TR4 expression. This expression pattern is consistent with a potential role for TR4 in the proliferation and/or early differentiation of neuronal precursors within the central nervous system.
During the development of the peripheral nervous system, TR4 transcripts were also actively expressed within the ganglia and the neural epithelium of many sensory organs. Abundant TR4 transcripts were detected in the dorsal root (Fig. 6, C-E, dg), superior cervical (Fig. 6, G-H, c), sympathetic (Fig. 6H, s), and trigeminal (Fig. 6, G, V) ganglia. Its abundance within the dorsal root ganglia through E11-E16 suggests a role for TR4 through the progression of these neurons from precursors to postmitotic neurons (24). High levels of TR4 mRNA were also found within all targets of sensory innervation, including the neuronal epithelium of inner ear (Fig. 6, G, e), retina (Fig. 6I), nasal cavity (Fig. 6, J, oe), and tongue ( Fig. 6, J, t). The signal intensity of TR4 in the ganglia layer (neuron bodies) of the retina is stronger than the inner and outer nuclear layers (photoreceptors) (Fig. 6I, compare g to n). This pattern of TR4 expression implies that TR4 could have similar roles on the development and/or early differentiation of motor, sensory, sympathetic neurons as well as their precursors.

Striking Similarities in the Expression Patterns of TR4 and CNTFR␣ within the Developing and Postnatal
Nervous Systems-Widely overlapping expression domains of TR4 and CNTFR␣ (25) further strengthen our hypothesis that the regulation of CNTFR␣ by TR4 may happen in vivo. Both TR4 and CNTFR␣ transcripts were localized within the developing nervous system. In addition, the timing of initial expression are tightly correlated. CNTFR␣ expression is most notably in the neuroepithelial lining of the brain vesicles as well as in the developing dorsal root ganglia and spinal cord during E11-E15 of rat embryos with its transcripts restricted in the mitotically active populations (25). Our in situ data clearly show that TR4 have a similar expression pattern in E9-E14 mouse embryos (correlate to E10-E15 rat embryos) and our RT-PCR assay (not shown) show that TR4 transcripts are present as early as E9.
Several tissues expressing abundant TR4 transcripts are also known to respond to CNTF. Examples are the hippocampal (14), retinal (nuclear and ganglion layers) (15), and motor (neostriatum and spinal motor) neurons (16), as well as the sympathetic (superior cervical) and sensory (trigeminal and dorsal root) (25) ganglia. Since CNTFR␣ is essential for the motor neuron development (19), our in situ data showing abundant TR4 transcripts there indicate that TR4-mediated CNTFR␣ induction could be an important process for motor neuron survival. An interesting coincidence is that the neurotransmitter dopamine has been proposed as an activator for TR2 and, possibly, for TR4 (26). Assuming dopamine activates TR4 which may then induce CNTFR␣ expression, it is possible that the reduced dopamine level, in the case of Parkinson's disease, can affect the activation of TR4 and CNTFR␣. Thus, a higher CNTF concentration may be needed to maintain neuron integrity. This hypothesis reasonably explains the action of CNTF and proposes TR4 as a bridge between the dopamine and the CNTF signal transduction pathways.
Outside the developing nervous system, both TR4 and CNTFR␣ transcripts are detected in the skeletal muscle (18), liver (27), and in the neural crest-derived precursors invading the nasal region, second pharyngeal pouch, heart structures, and intestinal areas (25). Significant amounts of TR4 transcripts are also present in many proliferating cell populations where the expression of CNTFR␣ is unclear, such as in the pericartilage, the brown fat, and the developing organs (e.g. lung, kidney, and intestine). This widespread but specific expression pattern suggest TR4 may be involved in the normal development of the whole embryo. Gene knock-out experiments would probably be helpful to understand the TR4 function and the significance of TR4 to CNTFR␣ gene induction.
During the adult stage, TR4 and CNTFR␣ transcripts are co-expressed within neuron cell bodies of many brain regions, including cortex, hippocampus, and cerebellum. Outside the adult brain regions, whether both are co-expressed is unknown.
TR4 Induces Transcriptional Activation via a DR1 Located within CNTFR-I5-The DR1 present in CNTFR-I5 represent the first intronic inducible DR1 response element regulated by TR4. Using in vitro binding assay, a DR1 sequence has been shown to be the high affinity binding site for several steroid receptors, including RARs (28), RXRs (29), peroxisome proliferator-activated receptor (30), COUP-TFs (31), TR2 (22), and TR4. 2 For certain gene regulation to occur, these DR1-binding factors may compete with each other for the binding site depending on the affinity and the relative abundance. The relative affinity among these DR1 binding factors are unclear. However, our data show a very high affinity of TR4 binding to the CNTFR-DR1, indicating once TR4 is present, it could dominate over other DR1 binding factors for occupying the DR1 response element.
Upon binding to a DR1 present in the promoter, these DR1binding steroid receptors or their complexes may either activate or repress gene transcription in reporter gene assays. For example, RAR/RXR heterodimer mediates retinoic acid response on phosphoenolpyruvate carboxykinase gene induction (28). Peroxisome proliferator-activated receptor alone or complexed with RXR is able to activate the acyl-CoA oxidase gene expression (30). Although COUP-TFs were originally identified as required factors for ovalbumin gene transcription (32), in contrast, they repress retinoic acid induced transactivation on a DR1 reporter construct (33). Similar phenomena are also observed for TR2 (22) and TR4 2 on a DR1-mediated repression of the retinoic acid-induced CRBPII promoter activity. As opposed to the repression effect of TR4 on the CRBPII promoter, 2 our data demonstrate that TR4 could bind to a DR1 sequence and induce transcriptional activation (Fig. 2, 3). The effects of other DR1-binding factors on CNTFR-I5 have not been examined yet. These results indicate that through the binding to a similar DR1 sequence, whether the transcriptional activity is up or down-regulated may depend on the receptor, gene context, response element, and the assay system used. Other studies also have made similar observations (5).
TR4 may be able to interact with other DR1-binding proteins, such as RAR␣, RAR␤, RXR␥, COUP-TFI, COUP-TFII, and TR2, and exert combinational effects on CNTFR␣ regulation. Among these DR1-binding protein, RAR␣ (34), RAR␤ (34), and RXR␥ (35) are detected in the more differentiated neuronal and oral cavities (J). Ventricular zone was indicated as arrowheads (F). Abbreviations used are: c, superior cervical ganglion; cp, cerebellum primordium; cx, cerebral cortex; drg, dorsal root ganglia; e, otic epithelium; g, ganglion layer of retina; l, lens; m, motor neuron; ms, muscle; n, nuclear layer of retina; ne, neural epithelium; oe, olfactory epithelium; r, retina; s, sympathetic ganglia; sp, spinal cord; Sr, striatum; t, tongue epithelium; V, trigeminal ganglion. The size bars represent 1 and 100 nm for A-E and F-J, respectively. types, which is consistent with the roles of retinoids in neuron differentiation and their expression patterns do not match that of CNTFR␣. Since TR4, TR2, 3 COUP-TFI, and COUP-TFII (33) are localized mainly in the actively mitotic neuron populations which also express CNTFR␣, the chance for them to regulate CNTFR␣ transcription is much higher. Especially, COUP-TFII has been demonstrated to be involved in spinal motor neuron development in chicken (36). How to coordinate these DR1binding proteins for the regulation of CNTFR␣ will be an intriguing question to follow.
In addition, the presence of several 5Ј-AGGTCA-like halfsites within CNTFR-I5 indicates some steroid receptors which bind as a monomer may regulate the CNTFR␣ expression. The candidates include thyroid hormone receptor and TR3/NGFI-B/nur77. This idea is supported by their localization in adult brain (8,37). However, much lower protein-DNA affinity between these receptors and the half-site may not support these factors that would play major roles on CNTFR␣ regulation.
TR4 Acts as an Activator-In addition to the roles of TR4 in restricting other hormonal signaling pathways 2 and in inhibiting the SV40 promoter (21), our data here suggest that TR4 can also function as a transcriptional activator with high induction activity through CNTFR-I5. This TR4-mediated transcriptional activation is consistent as tested in Chinese hamster ovary or P19 cell lines (data not shown). In contrast, TR4 repressed the retinoic acid-induced transcriptional activity with a CAT reporter containing a core DR1 without the flanking sequences and no significant TR4-mediated induction is observed in the absence of retinoic acid. 2 These results suggest that the sequences flanking the DR1 may also influence the TR4-mediated gene regulation. Finally, this TR4 inducible system may provide a good assay system for searching the TR4 ligands and activators.
In summary, our data indicate that human CNTFR␣ is the first identified neuron-specific target gene regulated by TR4. Induction of CNTFR␣ gene expression can be mediated through a perfect DR1 response element located in its 5th intron. In addition, the expression patterns of both genes correlated well during the developmental and postnatal neurogenesis. These findings may extend our views to study the interactions between steroid receptors and cytokine/cytokine receptors.