CRTR-1, a developmentally regulated transcriptional repressor related to the CP2 family of transcription factors.

CP2-related proteins comprise a family of DNA-binding transcription factors that are generally activators of transcription and expressed ubiquitously. We reported a differential display polymerase chain reaction fragment, Psc2, which was expressed in a regulated fashion in mouse pluripotent cells in vitro and in vivo. Here, we report further characterization of the Psc2 cDNA and function. The Psc2 cDNA contained an open reading frame homologous to CP2 family proteins. Regions implicated in DNA binding and oligomeric complex formation, but not transcription activation, were conserved. Psc2 expression in vivo during embryogenesis and in the adult mouse demonstrated tight spatial and temporal regulation, with the highest levels of expression in the epithelial lining of distal convoluted tubules in embryonic and adult kidneys. Functional analysis demonstrated that PSC2 repressed transcription 2.5-15-fold when bound to a heterologous promoter in ES, 293T, and COS-1 cells. The N-terminal 52 amino acids of PSC2 were shown to be necessary and sufficient for this activity and did not share obvious homology with reported repressor motifs. These results represent the first report of a CP2 family member that is expressed in a developmentally regulated fashion in vivo and that acts as a direct repressor of transcription. Accordingly, the protein has been named CP2-Related Transcriptional Repressor-1 (CRTR-1).

Consistent with the ability of mouse CP2 to activate transcription, human CP2/LBP-1c has been shown to activate transcription from an SV40 promoter (6) and from cellular promoters such as those directing expression of the serum amyloid A3 gene (11). The ability to activate transcription is conserved among other family members; LBP-1b activates transcription from the Ϫ155/Ϫ131 region of the human P450scc promoter (9), LBP-1a activates transcription from a human immunodeficiency virus, type I promoter (8), and chicken CP2 activates transcription from the ␣A-crystallin gene promoter (10).
Members of the CP2 family of transcription factors bind a consensus DNA sequence consisting of a direct bipartite repeat sequence, CNRG-N 6 -CNRG (3,10). Binding sites for this family of proteins have been described in the viral and cellular promoters described above; binding sites for CP2/LBP-1c have been described in the ␥-fibrinogen (12,3), synthase kinase-3␤ (13), and ␥-globin (5) promoters; binding sites for NF2d9 have been described in the Cyp 2d-9 (steroid 16␣-hydroxylase) promoter (4); and binding sites for LBP-9 have been described in the human P450scc promoter (9). Amino acids 63-270 of CP2 share sequence similarity with the region required for DNA binding in the Drosophila melanogaster transcription factor grainyhead (grh) (2). This region is highly conserved within other CP2 family members and appears to be important for DNA binding because LBP-1d, which is translated from an alternatively spliced form of CP2/LBP-1c and lacks amino acids 189 -239, is unable to bind the LBP-1c DNA binding sequence (7,8). N-and C-terminal truncation studies have defined the minimum DNA binding region of LBP-1c between amino acids 65 and 383 (14).
Mammalian members of the CP2 family are generally expressed ubiquitously (4,5,10,18). Whereas LBP-9 expression in cultured cell lines suggests some regulation of expression (9), the expression of this gene has not been mapped in vivo. Using differential display PCR 1 analysis we identified three novel genes that exhibit regulated expression during pluripotent cell differentiation. 2 Expression of these genes was temporally regulated during conversion of ES cells to EPL cells, an in vitro system that recapitulates conversion of inner cell mass to primitive ectoderm in vivo (19), and in the pluripotent cells of the pregastrulation mouse embryo. In this paper we report further analysis of one of these genes, denoted Psc2, which was expressed in pluripotent cells in vivo at 3.5 and 4.5 days post coitum (d.p.c.) and down-regulated around 4.75 d.p.c.. We demonstrate that the Psc2 cDNA encodes a novel mouse member of the CP2 family, which differs from the known members in two respects. Firstly, expression of this gene is tightly regulated in vivo in both temporal and spatial fashion, with the strongest expression detected in the epithelial lining of distal convoluted tubules (DCTs) in the embryonic and adult kidney. Secondly, the protein exhibits a novel transcriptional repression activity, localized in the N terminus of the protein, when tethered to a heterologous promoter. Accordingly, we have renamed the gene CP2-Related Transcriptional Repressor-1 (CRTR-1).

MATERIALS AND METHODS
cDNA Isolation and Sequencing-CRTR-1 cDNA clones were isolated from a ZAP II ES cDNA library derived from D3 ES cell RNA (CLONTECH Inc.). Library screening was carried out as described previously (20) using random primed [␣-32 P]dATP-labeled (Geneworks Ltd., Adelaide, South Australia, Australia) DNA probes (Gigaprime kit, Geneworks Ltd.). Clones q1, 1.2, 6A, 8.2.1, and 8B were isolated by successive screening of the library with a 736-bp fragment isolated by EcoRI digestion of the cloned CRTR-1 differential display PCR fragment, 2 a 132-bp fragment isolated by EcoRI/AccI digestion of cDNA clone q1, a 190-bp fragment isolated by EcoRI/NcoI digestion of cDNA clone 1.2, a 384-bp fragment isolated by EcoRI digestion of cDNA clone 6A, and a 500-bp fragment isolated by EcoRI digestion of cDNA clone 8.2.1, respectively (see Fig. 1A). Third round duplicate positive plaques from each library screen were isolated, grown to high titer, and excised by the ZAP excision process into pBluescript-SK ϩ vector as described in the manufacturer's instructions. CRTR-1 cDNA fragments were isolated by restriction digestion, subcloned into pBluescript-KS ϩ , and sequenced by automated DNA sequencing (PE Biosystems) using the universal sequencing primer and reverse sequencing primer. CRTR-1 nucleotide and open reading frame sequences were analyzed using DNASIS version 2.0 computer software (Hitachi Software Engineering Co.). CRTR-1 nucleotide and amino acid sequence comparisons were carried out using the GenBank TM data base and BLAST software (21). Protein sequence alignments were performed using the ClustalW and Boxshade software (22).
RNA Isolation, Riboprobe Synthesis, and Ribonuclease Protection Analysis-Cytoplasmic RNA was isolated from D3 ES and EPL (19) cells as described previously (29). Mouse embryos from 10.5 to 17.5 d.p.c., 16.5-d.p.c. embryonic tissues, and adult mouse tissues were isolated and homogenized, and total RNA was isolated using the guanidinium isothiocyanate method (30).
CRTR-1 antisense riboprobes for use in RNase protections were synthesized as described previously (31) by transcription of HincII linearized cDNA clone 1.2 (see Fig. 1A) with T3 RNA polymerase (Roche Molecular Biochemicals). Ribonuclease protection assays were performed on 10 g of total RNA as previously described (31,33) except that hybridizations were for 14 h at 45°C. To reduce overexposure of the loading control, low specific activity mGAP probes were synthesized using 40 Ci of [␣-32 P]rUTP in the reaction, whereas CRTR-1 probes were synthesized using 125 Ci of [␣-32 P]rUTP. 37,000 counts/min of mGAP probe and 150,000 counts/min of all other probes were added to each hybridization. [␣-32 P]rUTP was obtained from Geneworks Ltd.
Digoxygenin-labeled riboprobes for wholemount in situ hybridization were generated from the 736-bp CRTR-1 differential display PCR fragment cloned into pBluescript-KS ϩ2 (see Fig. 1A). Sense and antisense transcripts were generated as described (32) by BamHI digestion and transcription with T7 RNA polymerase and XhoI digestion and transcription with T3 RNA polymerase, respectively. 33 P-labeled sense and antisense CRTR-1 probes for in situ hybridizations were generated as described elsewhere (33) by transcription of pCRTR-1-1.2.8 linearized with BamHI and XhoI and transcribed with T3 and T7 RNA polymerase, respectively. [␣-33 P]rUTP was obtained from Geneworks Ltd.
Wholemount in Situ Hybridization and in Situ Hybridization-16.5d.p.c. kidneys were isolated from BALB/c mouse embryos. Wholemount in situ hybridization was carried out as described elsewhere (32) except that 16.5-d.p.c. embryonic kidneys were prewashed for 20 min in radioimmune precipitation buffer three times before post-fixing, prehybridized, hybridized, and washed at 65°C. Probed embryonic kidneys were dehydrated at room temperature for 10 min in 100% methanol, 15 min in isopropanol, and 2 ϫ 15 min in Histo-Clear (National Diagnostics) before embedding in paraffin wax. 7-m serial sections were cut using a Leica microtome, floated on water at 45°C, placed onto silanized microscope slides, de-paraffinized in Histo-Clear, and re-hydrated through a methanol series; then the sections were counterstained with methyl green and mounted with DePex and a coverslip. Sections were viewed using a Nikon Eclipse TE300 inverted microscope and photographed with Ektachrome 100 ASA slide film (Kodak).
Adult kidneys were dissected from BALB/c female mice and fixed as above. 7-m serial sections were cut as described for wholemount in situ hybridization. Radiolabeled in situ hybridization was carried out as described elsewhere (34) with the following modifications. Sections were heated to 55°C for 30 min before de-paraffinization in Histo-Clear. Sections were prehybridized at 52°C for 1 h and washed twice in 2ϫ SSPE for 2 min prior to overnight hybridization with riboprobes. Post-hybridization washes were carried out prior to RNase digestion as follows: 50% formamide, 2ϫ SSPE, 0.1% SDS, 10 mM ␤-mercaptoethanol at 52°C for 5 min; 50% formamide, 2ϫ SSPE, 10 mM ␤-mercaptoethanol at 52°C for 5 min; 50% formamide, 2ϫ SSPE, 10 mM ␤-mercaptoethanol at 60°C for 10 min; and twice in 2ϫ SSPE at room temperature for 5 min. Slides were air-dried and warmed to 37°C. Sections were counterstained with hematoxylin, mounted with DePex and a coverslip, viewed using light and dark field condensers on a Zeiss Axioplan microscope, and photographed with Ektachrome 160T ASA slide film (Kodak).
Cell Culture and Transfections-ES and EPL cell culture was carried out as described previously (19). D3 ES cells, COS-1 cells, and 293T cells were maintained as described previously (35).

CRTR-1 cDNA Isolation and Sequence-Based on the known expression in pluripotent cells in vitro and in vivo, CRTR-1
cDNA clones were isolated from a D3 ES cell ZAP II cDNA library (CLONTECH Inc; (20)) using successively more 5Ј CRTR-1-specific probes (Fig. 1A). Clones were confirmed to be CRTR-1-specific using Southern analysis (data not shown), sequence data, and expression analysis ( Fig. 1B and data not shown). Ribonuclease protection using riboprobes generated from the CRTR-1 cDNA clone 1.2 demonstrated rapid downregulation of CRTR-1 expression upon differentiation of ES cells to EPL cells (Fig. 1, B and C), consistent with the pattern described previously for the differential display PCR product using Northern blot and in situ hybridization. 2 Both strands of CRTR-1 cDNA fragments were sequenced to generate the CRTR-1 cDNA sequence (GenBank TM accession number AF311309). The 9405-bp cDNA contained a poly(A) tail 25 bp downstream of a consensus polyadenylation signal (AATAAA), consistent with the typical positioning of the polyadenylation signal 10 -30 bp upstream of the poly(A) tail (39).
A 1446-bp open reading frame extended from nucleotide 92 to nucleotide 1537 and was followed by a long 3Ј untranslated region of 7868 bp. The CRTR-1 protein predicted from this compiled sequence is 481 amino acids long, with a predicted molecular mass of 54,702 daltons. No other significant reading frames were identified.
CRTR-1 Sequence Analysis-Comparison of the predicted CRTR-1 amino acid sequence with entries in protein sequence data bases revealed considerable similarity to a group of proteins related to the mouse transcription factor CP2 (2) ( Table I). Fig. 2 shows a multiple sequence alignment of reported CP2 family members with CRTR-1. Included in Table I and the multiple sequence alignment are the mouse family members CP2 (2) and NF2d9 (4); human family members LBP-1a, LBP-1b, LBP-1c, LBP-1d (8), and LBP-9 (9); and the DNA binding domain of the D. melanogaster protein GRH (40).
Conservation between CRTR-1 and the other mammalian proteins was extensive (Fig. 2) and extended across the CRTR-1 sequence, with the exception of amino acids 1-47 and 381-401, which were conserved only with LBP-9. Furthermore, the 51-amino acid deletion at position 189 specific to LBP-1d and the 37-amino acid insertion (amino acids 274 -312) specific to LBP-1b were not found in CRTR-1. Similarity to GRH was confined to amino acids 632-865, shown to be sufficient for DNA binding to elements in the Dopa decarboxylase (Ddc) promoter (41). The failure of proteins containing deletions within this region (LBP-1d (7, 8) and chicken CP2 (10)) or truncated N-terminally past amino acid 65 or C-terminally past amino acid 383 (LBP-1c (14)) to bind DNA is consistent A, schematic figure outlining the cDNA library screening strategy employed to isolate CRTR-1 cDNAs. The CRTR-1 transcript, represented by the shaded box, was predicted to be 9400 bp in length by Northern analysis (see Footnote 2). cDNA clones, and the probes used for their isolation, are indicated together with the restriction sites used for their excision. RI, EcoRI; A, AccI; N, NcoI; dd, differential display; kb, kilobase(s). B, ribonuclease protection analysis of CRTR-1 expression using 10 g of total RNA isolated from ES and EPL cells cultured in 50% MEDII in the presence of leukemia inhibitory factor for 2, 4, 6, and 8 days. An mGAP-specific antisense riboprobe was used as a loading control. C, quantitation of ribonuclease protection analysis used to confirm isolated CRTR-1 cDNAs as represented in B. The expression of CRTR-1 was normalized against the mGAP loading control. The standard mean error of three independent ribonuclease protection assays using three different CRTR-1-specific antisense riboprobes is shown.
with the identification of this region as a DNA binding sequence that is conserved in the CRTR-1 protein.
Truncation studies have localized an oligomerization domain within LBP-1c to amino acids 266 -403 (14). This region was well conserved within CRTR-1, suggesting a potential for formation of homo-and hetero-oligomeric protein complexes. Within the equivalent regions of CP2 (amino acids 398 -425) and LBP-1c/LBP-1d are located a glutamine/proline repeat and a polyglutamine repeat, respectively, which have been predicted to form a transcriptional activation domain (2) but are not conserved in CRTR-1 or LBP-9.
LBP-9 (9) is the CP2 family member that shows the greatest level of similarity to CRTR-1 (Table I; Fig. 2). Whereas there was considerable conservation of amino acid sequences between these proteins, similarity was restricted to the open reading frame and did not extend into the reported, incomplete 3Ј untranslated region of LBP-9.
Regulated Expression of CRTR-1 during Mouse Development-Expression of CRTR-1 has been shown to be specifically regulated in pluripotent cell populations in vitro and in vivo. 2 CRTR-1 expression during later mouse development was investigated by ribonuclease protection analysis using total embryonic RNA isolated from 10.5-17.5 d.p.c. embryos, tissuespecific total RNA samples isolated from 16.5-d.p.c. embryos, and tissue-specific total RNA samples isolated from adult mice. CRTR-1 expression was not detected in total RNA isolated from 12.5-and 13.5-d.p.c. embryos and was expressed at low levels in 10.5-and 11.5-d.p.c. embryos (Fig. 3A). CRTR-1 expression was highest between 14.5 and 17.5 d.p.c. Of the 16.5-d.p.c. embryonic tissues analyzed (Fig. 3B), CRTR-1 was not detected in 16.5-d.p.c. embryonic brain. Low levels of CRTR-1 expression were detected in 16.5-d.p.c. embryonic intestine, limb, lung, and skin, with highest expression in 16.5 d.p.c. embryonic kidney. Levels of CRTR-1 expression observed in 16.5-d.p.c. embryonic kidney were comparable with levels of expression observed in ES cells. Of the tissue-specific total RNA samples isolated from adult mice, CRTR-1 was not detected in brain, heart, liver, and spleen (Fig. 3C). CRTR-1 was expressed at low levels in lung, mesenteric lymph nodes, muscle, ovary, and thymus; at elevated levels in placenta, testis, and small intestine; and at high levels in adult kidney and stomach, which expressed CRTR-1 at levels 7-and 1.5-fold greater than ES cells, respectively. Expression of CRTR-1 was therefore specifically regulated in a temporal and spatial fashion both during embryogenesis and in the adult mouse.
Expression of CRTR-1 in Embryonic and Adult Kidney Is Restricted to the Distal Convoluted Tubules-Cellular localization of CRTR-1 expression was investigated in embryonic and adult kidneys. Wholemount in situ hybridization analysis was carried out using kidneys isolated from 16.5-d.p.c. embryos where CRTR-1 expression was demonstrated to be highest (Fig. 3B). Embryonic kidneys were probed with CRTR-1-specific sense and antisense digoxygenin-labeled riboprobes prior to embedding, sectioning, and counterstaining. Kidney sections probed with CRTR-1 sense control probe showed no specific staining (Fig. 4A). Kidney sections probed with CRTR-1 antisense probe showed specific staining representing CRTR-1 expression in the epithelial monolayer lining a subset of tubules in the embryonic kidney cortex (Fig. 4, B and C). CRTR-1expressing tubules were identified as DCTs because they were located adjacent to glomeruli, consistent with the location of DCTs within the kidney cortex (42). Furthermore, only a small proportion of the tubules present in any cortical section expressed CRTR-1, consistent with the greater relative representation of proximal convoluted tubules in this region of the kidney (43)(44)(45)(46). Finally, the morphology of CRTR-1-expressing tubules was clear and open, consistent with the morphology of DCTs but distinct from that of proximal convoluted tubules, in which the epithelium forms a brush border consisting of microvilli that project into the lumen of the tubule (45,46). CRTR-1 expression was not detected in proximal convoluted tubules, glomeruli, or kidney vasculature (Fig. 4, B and C).
CRTR-1-expressing cells in the adult mouse kidney were determined by radiolabeled in situ hybridization to kidney sections because the greater volume of the adult kidney precludes the use of wholemount in situ hybridization. Adult kidneys were sectioned and probed with CRTR-1-specific, [␣-33 P]labeled sense and antisense riboprobes. Hybridization was not detected using CRTR-1 sense control probe (Fig. 4, D and E). Adult kidney sections probed with CRTR-1 antisense probe showed specific localization of CRTR-1 expression to the epithelial monolayer lining a subset of tubules in the adult kidney cortex (Fig. 4, F and G). Consistent with the expression in embryonic kidneys, these tubules were identified as DCTs. CRTR-1 transcripts were not detected in the proximal convoluted tubules, glomeruli, or kidney vasculature (Fig. 4G). This analysis demonstrates that expression of CRTR-1 is spatially regulated in at least two distinct sites, the pluripotent cells of the developing mouse embryo 2 and the epithelial cells lining the embryonic and adult kidney distal convoluted tubules. (1,3) and in vivo (3, 8 -11) transcription assays. The ability of CRTR-1 to act as a transcriptional regulator could not be investigated using target gene expression because the DNA binding sequence for this protein is unknown. The transcriptional activity of CRTR-1 was therefore assessed as a fusion protein with amino acids 1-174 of the Gal4 DNA binding domain (DBD) (28) in the plasmid pGalO⅐CRTR-1. pTK-MH100x4-LUC (36), which contains a luciferase gene regulated by the thymidine kinase (TK) promoter and four upstream tandem copies of the Gal4 binding site, was used as a reporter.

CRTR-1 Acts as a Transcriptional Repressor in a Variety of Cell Types-Members of the CP2 family have been reported to act as transcriptional activators in both in vitro
Plasmids were transfected into COS-1 cells, and levels of luciferase activity were analyzed in cell extracts 36 h posttransfection. Cotransfection of the Gal4-DBD, pGalO (23)(24)(25)(26)(27), with pTK-MH100x4-LUC did not alter the reproducible levels of luciferase activity (Fig. 5A, column 1) compared with pTK-MH100x4-LUC alone (Fig. 5A, column 2). Cotransfection of pGalO⅐CRTR-1 with pTK-MH100x4-LUC resulted in a 10 -15fold reduction in luciferase activity (Fig. 5A, column 3). CRTR-1-mediated transcriptional repression was also demonstrated in 293T (Fig. 5B) and ES cells (Fig. 5C), where expression of the Gal4-DBD-CRTR-1 fusion protein reduced luciferase expression 2.5-and 3.5-fold, respectively. This transcriptional repression was specific for the reporter plasmid pTK-MH100x4-LUC and not a result of general transcriptional toxicity of the Gal4-  DBD-CRTR-1 fusion protein, because luciferase activity within these assays was normalized against expression of Renilla luciferase, expressed from pRLTK under the control of the constitutive thymidine kinase promoter. Furthermore, expression of luciferase from pHRE-Luc (37,38), in which expression of luciferase is controlled by the SV40 promoter and three upstream copies of the hypoxia-inducible factor response element, was not altered by cotransfection with pGalO⅐CRTR-1 (Fig. 5B,  columns 4 and 5). These results demonstrate that CRTR-1 acts as a transcriptional repressor in a variety of cell types including ES cells and 293T cells, representative of in vivo expression.
The Ability of CRTR-1 to Repress Transcription Resides in an N-terminal Repression Domain-The N-terminal 40 amino acids of CP2 have been shown to contain the CP2 transcriptional activation domain. 3 This region of CRTR-1 was not conserved with members of the CP2 family reported to act as transcriptional activators but was closely related to LBP-9, which can antagonize LBP-1b-mediated transcriptional activation (9).

CRTR-1 Is a Novel Mouse Member of the CP2 Family of Transcription Factors-The
CRTR-1 open reading frame was closely related to a group of proteins including the mouse transcription factor CP2 (Table I and Fig. 2) (2), the founding member of an expanding group of highly conserved proteins implicated in transcriptional control. Amino acid conservation across the CP2 open reading frame suggested conservation of functionally important regions of the CRTR-1 protein. In particular, a potential DNA binding domain, distinct from structurally characterized DNA binding domains, was identified between CRTR-1 amino acids 45 and 260, consistent with conservation of this region with the DNA binding domain of the D. melanogaster protein GRH (41) and deletion mapping of the LBP-1c DNA binding domain (14). Furthermore, a region implicated by deletion mapping in homo-oligomerization of LBP-1c (14) was highly conserved with CRTR-1 residues 261-386, suggesting that this protein is likely to support the formation of protein complexes. Whereas CRTR-1 shared greatest identity (88%) with the recently reported human protein LBP-9 (9), conservation was restricted to the open reading frame and did not extend into the 3Ј untranslated region. This suggests 3 S. Jane, personal communication. either that the reported CRTR-1 and LBP-9 cDNAs are derived from alternative splicing of a homologous gene in mice and humans or that the proteins are not products of homologous genes. Sequence analysis therefore identified CRTR-1 as a novel mouse member of the CP2 family, with potential roles in transcriptional control and the formation of protein complexes.
CRTR-1 Is a Novel Transcriptional Repressor-CRTR-1 was shown to repress transcription when bound at a heterologous promoter. Whereas the extent of repression varied from 2.5-to 15-fold in different cell types (Fig. 5), conservation of this activity in different cell lines suggests that these results are indicative of normal CRTR-1 activity. This is supported by the fact that 293T and ES cells, in which repression was demonstrated, are representative of in vivo expression sites in kidney and pluripotent cells, respectively. The transcriptional repression activity of CRTR-1 was found to be localized to the Nterminal 52 amino acids, a region that does not show strong homology to other members of the CP2 family, with the exception of LBP-9, which has been shown to antagonize LBP-1bmediated transcriptional activation by an unknown mechanism (9). The results presented here demonstrate that the N-terminal 52 amino acids of CRTR-1 are both necessary and sufficient for CRTR-1-mediated transcriptional repression when recruited to the promoter by DNA binding and that the observed transcriptional repression was not the result of steric hindrance caused by Gal4-DBD-CRTR-1 fusion proteins.
Activity as a transcriptional repressor distinguishes CRTR-1 from most other members of the CP2 family, which have been reported to act as transcriptional activators (1, 3, 8 -11). This is consistent with the lack of amino acid conservation at the N terminus, which contains the activation domain in CP2, 4 and with the lack of polyglutamine-and glutamine/proline-rich sequences suggested as activation domains in the LBP-1c and CP2 sequences, respectively (2). LBP-9, identified as a sequence-specific binding protein on the Ϫ155/Ϫ131 region of the P450scc promoter, also exhibits unusual transcriptional activity. Whereas LBP-1b, which also binds this sequence, activated transcription of a linked reporter gene 21-fold in JEG-3 cells, LBP-9 did not activate transcription in the same system (9). Transfection of cells with increasing amounts of LBP-9 suppressed the LBP-1b-mediated reporter activation to basal lev-els. The mechanism of inhibition was not resolved and could result from direct repression of transcription, steric exclusion of LBP-1b from the DNA binding site, or displacement of LBP-1b from the promoter by formation of complexes with LBP-9. By contrast, CRTR-1 is the first reported CP2 family member that represses transcription directly from a heterologous promoter. Conservation of the 52-amino acid region of CRTR-1, responsible for transcriptional repression, with the equivalent region of LBP-9 may provide a mechanistic explanation for the suppression of LBP-1b-mediated transcription activation by LBP-9. In particular, if heteromeric complexes including LBP-9 can be localized at the P450scc promoter, this protein and possibly CRTR-1 may be capable of acting as a dominant repressor of promoters that are activated by CP2 family proteins. Resolution of this possibility requires identification of the CRTR-1 DNA binding sequence and binding partners.
Transcriptional repression can be mediated through several different mechanisms such as interference with assembly of the transcriptional machinery (47) or recruitment of corepressors including histone deacetylases (48). Protein sequence motifs present in DNA-binding transcriptional repressors that mediate interaction with corepressor proteins include PXDLS in the ikaros protein (49 -51), WRPW in Hairy-related bHLH proteins (52,53), and a Gly/Arg-rich sequence present in the transcription factor YY1 (54). Furthermore, a histone deacetylase-independent mechanism of transcriptional repression has been described for methyl-CpG-binding protein 2 that is dependent on the presence of a conserved 30-amino acid sequence that contains two clusters of basic amino acids (55). These motifs could not be identified within the N-terminal 52 amino acids of CRTR-1, suggesting a novel mechanism of transcriptional repression for this protein. Conservation of repressor activity in cell lines of diverse origin and properties such as ES cells, 293T cells, and COS-1 cells suggests that factors required for CRTR-1-mediated repression are widely expressed.
Expression of CRTR-1 Is Spatially and Temporally Regulated during Mouse Development-Expression of CRTR-1 was shown to be spatially and temporally regulated, both during embryogenesis and in the adult mouse. In vitro, CRTR-1 was expressed in ES cells and rapidly down-regulated upon differ- entiation to EPL cells (Fig. 1, B and C). An equivalent expression pattern has been described in vivo where CRTR-1 expression in pluripotent cells of 3.5-d.p.c. mouse embryos is downregulated at around 4.75 d.p.c. 2 Re-expression of CRTR-1 in the embryo from 10.5 to 12.5 d.p.c. and at higher levels from 14.5 to 17.5 d.p.c. was demonstrated by ribonuclease protection analysis (Fig. 3A). This analysis also indicated differential CRTR-1 expression in various tissues. For example, in the 16.5-d.p.c. embryo (Fig. 3B), CRTR-1 expression was greatest in the kidney, at low levels in the intestine, limb, lung, and skin, and not detected in the brain, whereas in the adult mouse, CRTR-1 (Fig. 3C) was expressed at highest levels within adult mouse kidney, at moderate levels in the stomach, testis, placenta, and small intestine, at low levels in lung, mesenteric lymph nodes, muscle, ovaries, and thymus, and not detected in brain, heart, liver, and spleen. Detailed investigation of CRTR-1 expression in 16.5-d.p.c. embryonic and adult mouse kidney (Fig. 4) demonstrated restriction of expression to the epithelium of distal convoluted tubules, identified by morphological and histological criteria.
Tight spatial and temporal localization of CRTR-1 expression in vitro and in vivo distinguishes CRTR-1 from most other CP2 family members that are reported to be expressed ubiquitously (4,5,10,18). Variable expression of LBP-9 as detected by reverse transcriptase-PCR in cultured cell lines may also be indicative of regulated expression. This analysis (9) was limited to cell lines of placental (JEG-3), adrenal (NCI-H295A), cervical (HeLa), hepatic (HepG2), and kidney (COS-1) origin and human adrenal tissue. LBP-9 expression was detected at highest levels in JEG-3 cells, at lower levels in COS-1, HepG2 and HeLa cells, and was not detected in NCI-H295A cells or human adrenal tissue. CRTR-1 was expressed in embryonic and adult kidney and placenta, but expression was not detected in adult liver. Whereas direct parallels cannot be drawn between expression in vivo and potentially deregulated expression in cell lines in vitro, the differential sites of CRTR-1 and LBP-9 expression support the suggestion, based on sequence conservation, that these genes may not represent homologues.
The sites of CRTR-1 expression in vivo suggest at least two functions for CRTR-1 in the mouse: in pluripotent cells during early mouse development and in the development and function of kidney DCTs. Whereas both pluripotent cells and the DCT lining are epithelial in origin, other epithelial cells including the lining of kidney proximal convoluted tubules did not express detectable CRTR-1, excluding a general role for the CRTR-1 protein in cells of this type. It is of interest that both demonstrated sites of CRTR-1 expression in vivo are associated with cavitation, within the egg cylinder and DCTs, respectively. A general model for cavitation, based on integrated action of extracellular diffusible "death" signals and matrixlocalized survival signals, has been hypothesized, based on mechanistic investigation of proamniotic cavity formation in pluripotent cell populations (56). CRTR-1 expression could potentially be responsive to signals of this nature (57).
DCTs form part of the nephron, the basic filtration unit of the kidney, and become functional at around 16.0 d.p.c. (58). These tubules control blood pH through regulated ion channels that direct the reabsorption of Na ϩ and HCO 3 Ϫ ions from kidney filtrates and the secretion of K ϩ and H ϩ ions (59,60). This process is regulated by the signaling molecule aldosterone, a ligand for the mineralcorticoid receptor that is associated with regulation of genes required for Na ϩ and H ϩ exchange (61,62). By RNase protection, CRTR-1 expression in the embryonic and adult kidney was at similar or higher levels to that in ES cells. This points to an extremely high level of expression in kidney DTCs, which comprise only a small proportion of the cells within the kidney, suggesting important CRTR-1 function at this location. DCTs arise from the metanephric mesenchyme, which is located near the cortical periphery after 13 d.p.c. Although CRTR-1 expression in the embryo may be associated with induction of these tubules during development, continued high level expression of the gene at later stages of embryogenesis and in the adult is suggestive of a role for CRTR-1 in DCT function and physiology. Resolution of the functional relevance of CRTR-1 expression at different locations and developmental stages awaits functional investigation.