A Novel Colonic Repressor Element Regulates Intestinal Gene Expression by Interacting with Cux/CDP

Abstract

Intestinal gene regulation involves mechanisms that direct temporal expression along the vertical and horizontal axes of the alimentary tract. Sucrase-isomaltase (SI), the product of an enterocyte-specific gene, exhibits a complex pattern of expression. Generation of transgenic mice with a mutated SI transgene showed involvement of an overlapping CDP (CCAAT displacement protein)-GATA element in colonic repression of SI throughout postnatal intestinal development. We define this element as CRESIP (colon-repressive element of the SI promoter). Cux/CDP interacts with SI and represses SI promoter activity in a CRESIP-dependent manner. Cux/CDP homozygous mutant mice displayed increased expression of SI mRNA during early postnatal development. Our results demonstrate that an intestinal gene can be repressed in the distal gut and identify Cux/CDP as a regulator of this repression during development.

Morphogenesis of the digestive tract is the result of several steps through development beginning with gross morphogenesis of the organ and then cytodifferentiation of the epithelium and induction of intestine-specific genes in the establishment of the functional adult epithelium (40). In the mouse small intestine, cytodifferentiation occurs between embryonic days 14 and 15 with the transition from a stratified epithelium to a columnar epithelium with nascent villi. Small intestine villi lengthen and crypts form through the first two postnatal weeks. The early development of the mouse colon shows transient similarities to that of the small intestine, with the formation of villus structures. This temporary colonic architecture regresses early during development to form the adult colonic mucosa (39). Early developmental similarities between the small intestine and colon in humans have also been described (21).

The sucrase-isomaltase (SI) gene represents a useful model to elucidate the mechanisms of intestinal development. This gene encodes an enzyme that is expressed in the brush border of mature enterocytes in a complex developmental pattern (19, 20, 42). A low level of SI gene expression is first detectable in the mouse embryonic intestine, and this expression remains stable through the first 2 weeks of life after birth. A similar low and transient level of expression is also detected in the colon during early postnatal life (42). Between days 16 and 17, there is a dramatic induction of SI expression in enterocytes located at the crypt-villus junction in the small intestine (42). Similar transient expression of SI, as well as other intestinal brush border enzymes, has also been observed in the human fetal colon (48). It has been suggested that the mechanisms directing transient SI expression in the colon are recapitulated in the process of human colonic neoplasia, with the reexpression of SI in the majority of colonic adenomatous polyps and adenocarcinomas (4, 9, 45). A comparable recurrence of SI expression was also found in rodent colon tissues subjected to oncogene-mediated transformations (29). Therefore, the delineation of the molecular mechanisms involved in intestinal SI gene regulation during development could also provide insights into the molecular alterations that occur during colonic neoplasia.

An evolutionarily conserved SI promoter contains three major positive regulatory elements, SIF1 (SI footprint 1), SIF2, and SIF3 (41). Caudal-related homeodomain proteins (Cdx1 and Cdx2) interact with the SIF1 element and induce gene transcription in vitro (36, 38). The SIF2 and SIF3 elements of the human SI promoter interact with hepatocyte nuclear factor 1 (HNF-1) proteins to regulate transcription (47). The SIF3 element is essential to support SI promoter activity in the mouse intestinal epithelium (7). Similar to the pattern for the endogenous gene, this short SI promoter directs transcription to mouse enterocytes in developmental and differentiation-dependent patterns and fails to support expression in colonocytes. Taken together, these studies confirm the importance of complementary activators for regulating SI gene expression. By extension, the presence of significant amounts of HNF-1 (32) and Cdx proteins (33) in the colonic epithelium supports the hypothesis that transcriptional repression rather than lack of activation is involved in the colonic regulation of SI expression.

To identify regulatory elements that may repress SI expression during postnatal development, we performed a computer-based analysis of the short SI promoter (46). We identified a region between nucleotides −73 and −64 which predicted an overlapping consensus site for both the CDP (CCAAT displacement protein)/Cut repressor and GATA activators of transcription. Murine Cux/CDP and human CDP/Cut are transcriptional repressors closely related to the cut protein of Drosophila melanogaster (26, 43). CDP homologues have in common a unique homeodomain and three similar regions, designated cut repeats (CR), which function as DNA binding domains (2, 13). Several reports, mainly based on gene expression studies, suggest that CDP homologues function as transcriptional repressors by direct competition with transcriptional activators (1, 16, 18, 43) or by active repression by interacting with the HDAC1 protein to promote histone deacetylation (17, 18). Although different roles for the Cux/CDP protein have been proposed in a variety of distinct cell lineages, the potential role for this protein in the intestine still remains unexplored.

In addition to a Cux/CDP binding element, a GATA consensus site overlaps within this element. GATA transcription factors are characterized by a highly conserved DNA binding domain consisting of two zinc fingers of the motif Cys-X₂-Cys-X₁₇-Cys-X₂-Cys (22). GATA-4, -5, and -6 are expressed in various mesoderm- and endoderm-derived tissues such as heart, liver, lung, gonad, and gut, where they play critical roles in regulating tissue-specific gene expression (22).

The main objective of this study was to evaluate the functional role of the CDP-GATA site in the regulation of SI promoter activity during intestinal development. We report that Cux/CDP and GATA-4 proteins are capable of interacting with this overlapping site, named CRESIP (colon-repressive element of the SI promoter). We present evidence that the novel CRESIP is crucial for colonic repression of the SI promoter and that Cux/CDP acts through this site to repress SI.

MATERIALS AND METHODS

Plasmid construction and mutagenesis.

The −201 to +54mSI-hGH construct was described elsewhere (42). Point mutagenesis of the −201 to +54mSI-hGH plasmid was performed with the Transformer site-directed mutagenesis kit (Clontech Laboratories Inc., Palo Alto, Calif.). Oligonucleotide GATm (5′-GAATATTAAACATTTCGAGGCTTGTGAAAG-3′) was designed to create point mutations (underlined) within CRESIP. Integrity of the mutant construct was confirmed by sequence analysis. The coding sequences of various portions of the mouse Cux/CDP protein mapping with the different CR and homeodomain (HD) regions were amplified by PCR by using as a template the pRcCMV/Cux expression vector (kindly provided by J.-F. Brunet, INSERM-CNRS, Marseille-Luminy, France). For PCR, the following paired primers were used for amplification: CR1, CR1up (1373 to 1393; 5′-CCAGCCAATCAGAAAGTGCTG-3′) and CR1down (1642 to 1623; 5′-TGCTACGGAGTGCCAGGATG-3′); CR2, CR2up (2554 to 2574; 5′-AGCAGTACGAGGTCTACATG-3′) and CR2down (2792 to 2772; 5′-TGGCCCAGCTCTCCATTCAG-3′); CR3, CR3up (3094 to 3113; 5′-CCCTCAGCATCCAAGAATTA-3′) and CR3down (3355 to 3336; 5′-CCATCAGCTTCTCCACATTG-3′); HD, HDup (3440 to 3461; 5′-TGGGTATTGACTATAGCCAAG-3′) and HDdown (3694 to 3673; 5′-GGCTTCCAGCTTGAATCTCC-3′); CR3/HD, CR3up (3074 to 3093; 5′-TGCCTCTCTCTGGACACTCAG-3′) and HDdown (3719 to 3700; 5′-GAGTCGCTGGCACCAGCCTG-3′).

PCR amplification was carried out with Pfu polymerase (Invitrogen, Carlsbad, Calif.) for 35 cycles with denaturation at 95°C, annealing of the primers at 55°C, and reaction extension at 72°C, each for 1 min. The respective fragments of amplified CR1 (269 bp), CR2 (238 bp), CR3 (261 bp), HD (254 bp), and CR3/HD (645 bp) were subcloned into the bacterial expression vector pGEX-4T-3 (Amersham Pharmacia Biotech, Piscataway, N.J.). Integrity of subcloned PCR products was confirmed by sequence analysis. The Cux/CDP mutant glutathione S-transferase (GST) fusion protein was created by subcloning an EcoRI-XhoI fragment of the Cux/CDP cDNA that encodes amino acids 132 to 1067 into the pGEX-4T-3 vector. The Cux/CDPmut expression vector was created by a partial digestion of the pRcCMV/Cux vector with XhoI, which resulted in the release of the 3′ coding region, which encodes the HD and the repression domain of the Cux/CDP protein (18).

Preparation of bacterial fusion proteins.

The pGEX-4T-3/Cux domain constructs were transformed in the BL21 (DE3) pLysS strain (Promega Biotech, Madison, Wis.), and the expression of GST fusion proteins was induced with 0.1 mM IPTG (isopropyl-β-d-thiogalactopyranoside) for 3 h. The fusion proteins were then purified with a glutathione-Sepharose 4B column (Amersham Pharmacia Biotech) according to the manufacturer's recommendations. The purification of the GST-Cux fusion proteins was confirmed by Coomassie staining of sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) gel and by immunoblotting with an antibody to GST (Amersham Pharmacia Biotech).

Transgenic and knockout mice.

The promoter/reporter −201 to +54 MUT mSI-hGH construct was released by digestion with XbaI and SphI and purified. Transgenic mice were produced by the Transgenic Core Facility at the University of Pennsylvania. The DNA construct was injected into the male pronuclei of fertilized eggs and implanted into pseudopregnant females by standard methodology. DNA from tail biopsy samples of the resulting mice was extracted with the QIAamp tissue kit (Qiagen Inc., Valencia, Calif.). The presence of the transgene in mouse genomic DNA was determined by PCR and Southern analysis as described previously (19, 20). Transgene founders of the BGSJL/F1 strain (Jackson Laboratory, Bar Harbor, Maine) were bred with normal CD1 mice (Charles River), and offspring were analyzed for the transgene by PCR. The generation of Cux/CDP ΔHD mice was described elsewhere (34). C57BL/6J mice heterozygous for the targeted allele were subsequently bred with normal CD1 mice. Homozygous Cux/CDP ΔHD mice were identified by PCR, and their identity was confirmed on the basis of their typically small size.

RNA analysis.

RNA was extracted from multiple tissues using a CsCl gradient method as previously described (19). RNase protection assays were performed using the RPA II kit (Ambion, Austin, Tex.) according to manufacturer's recommendations. Riboprobes for the detection of mouse SI (mSI), hGH, and m36B4 mRNA were prepared as previously described (7).

In situ hybridization.

Intestinal tissues were fixed in 4% paraformaldehyde and embedded in paraffin in a Swiss roll orientation such that the entire length of the small or large intestinal tract could be identified on single sections. In situ hybridization was performed exactly as described previously (33). Sense and antisense hGH ³³P-labeled riboprobes were synthesized by digesting pGEM-hGH with EcoRI (sense) and XbaI (antisense) and using, respectively, SP6 and T7 RNA polymerases. Following hybridization and washes, slides were dehydrated through increasing concentrations of ethanol, air dried, and exposed on a BIOMAX MR film (Eastman Kodak Company, Rochester, N.Y.) for 24 h. Slides were then dipped in liquid emulsion (Kodak; NTB2, diluted 1:1 in water) and exposed in light-tight boxes at 4°C. The slides were developed after 2 weeks of exposure and stained with hematoxylin and eosin (H&E). Essentially the same procedure was used for the nonisotopic digoxigenin in situ hybridization except that the probes were immunologically detected with an alkaline phosphatase-conjugated antidigoxigenin antibody according to the manufacturer's recommendations (Roche Molecular Biochemicals, Indianapolis, Ind.).

Isolation of nuclear proteins from adult intestinal epithelium.

Nuclear proteins were isolated from the intestinal epithelium of adult mice by an adaptation of a previously described method that used human intestine (27). Briefly, mice were sacrificed and the intestine was separated into sections of proximal jejunum, ileum, and proximal colon. Each section was opened longitudinally and rinsed with cold phosphate-buffered saline (PBS). The sections were further cut in 5-mm-long pieces and incubated in 5 ml of cold MatriSperse (Becton Dickinson, Franklin Lakes, N.J.) in 15-ml tubes at 4°C for 18 to 24 h. The epithelial layer was dissociated by gentle manual shaking. The epithelial suspension was collected, centrifuged, and washed with cold PBS. Nuclear proteins were then isolated from the epithelial cell pellet as described previously (41).

Western blot analysis.

Twenty micrograms of nuclear protein extracts was analyzed by 4 to 12% N,N-methylenebisacrylamide-Tris or 3 to 8% Tris-acetate NuPAGE (Invitrogen) electrophoresis and transferred to an Immobilon-P membrane. Western blotting was then performed exactly as described previously (7). The following antibodies were used: a CDP affinity-purified guinea pig polyclonal antibody raised against the full protein (26) and kindly provided by E. Neufeld and a Cdx2 affinity-purified rabbit polyclonal antibody raised against a peptide mapping to the N terminus of murine Cdx2 (31).

EMSA.

Electrophoretic mobility shift assays (EMSA) were performed essentially as described previously (36) with some modifications. The reactions were performed in 20 μl of binding buffer (10 mM Tris-HCl [pH 7.5], 50 mM KCl, 5 mM MgCl₂, 1 mM EDTA, 5% glycerol, 0.2 mM dithiothreitol) (35) containing 5 μg of nuclear extracts or 50 ng of GST fusion proteins, 1 μg of poly(dI-dC) (50 ng for GST fusion proteins), and 20,000 cpm of ³²P-labeled DNA probe for 30 min. Complexes were then separated on a 4% polyacrylamide gel, dried, and exposed on a phosphorimager screen. The following double-stranded oligonucleotides were used as DNA probes: WT-CRESIP (5′-AAACATTGATAGGCTTGTGA-3′) and MUT-CRESIP (5′-AAACATTTCGAGGCTTGTG A-3′).

Immunohistochemistry.

Immunohistochemistry was performed exactly as described previously (33). After blocking with protein-blocking agent (Coulter-Immunotech, Miami, Fla.), the slides were incubated overnight at 4°C with anti-goat CDP (SC-6327) at a concentration of 0.8 μg/ml. The CDP primary antibody was visualized with biotinylated anti-goat secondary antibody and an avidin-biotin detection system, according to the protocol provided by Vector Laboratories (Burlingame, Calif.). The slides were developed with DAB (3,3′-diaminobenzidine tetrahydrochloride; Vector Laboratories). The tissue was lightly counterstained with hematoxylin and mounted with Permount.

Cell culture and transient transfections.

Caco-2 and Cos-7 cell lines were obtained from the American Type Culture Collection (Manassas, Va.). Cells were maintained in Dulbecco's modified Eagle medium (DMEM) supplemented with 4.5 g of d-glucose/liter, 25 mM HEPES, 10% fetal bovine serum, 50 U of penicillin/ml, and 50 μg of streptomycin/ml in 5% CO₂. Transfections were performed with Lipofectamine (Invitrogen) according to manufacturer's recommendations. Caco-2 cells at 10 to 20% confluence were incubated with a constant amount of 1.5 μg of total DNA and 2.5 μl of Lipofectamine/ml of OPTI-MEM for 18 h. The medium was then changed to DMEM complete medium containing 10% fetal bovine serum. Luciferase activity was determined 72 h after the transfection with the luciferase assay kit (Promega Biotech). Each experiment was repeated at least three times in triplicate. The pSV-β-galactosidase expression vector was cotransfected in each experiment as a measure of transfection efficiency, and the results were reported as light units per unit of β-galactosidase. Cos-7 cells were incubated with a total of 9 μg of expression vectors and 15 μl of Lipofectamine/6 ml of OPTI-MEM for 18 h. Nuclear proteins were extracted 48 h after the transfection as described previously (36).

RESULTS

CRESIP functions as a colonic repressor element for SI transcription during murine postnatal development.

We have previously identified a number of elements that promote transcription of the SI promoter in intestinal epithelial cells (36, 41). To seek other regulatory elements located within a short evolutionarily conserved SI promoter, we analyzed this promoter by using the TRANSFAC transcription factor database website (46). This analysis identified two putative overlapping sites localized between nucleotides −73 to −64 of the mouse SI promoter that were highly predicted to interact with both the CDP/Cut and GATA proteins (Fig. 1A). To assess a functional role of this region on SI gene regulation in the mouse intestinal epithelium, we designed a transgenic construct containing nucleotides −201 to +54 of the mSI gene and introduced a GAT-to-TCG (MUT) mutation in the core of this site (Fig. 1A). The SI promoter was linked to the hGH reporter gene (Fig. 1B and C). Three founders (founders 5, 8, and 28) derived from construct −201 to +54 MUT mSI-hGH (SI MUT) were found to have the transgene integrated into their genome (data not shown). The effect of the SI promoter mutation on hGH expression in these transgenic lines was compared to that on hGH expression of a previously characterized transgenic line that had integrated the wild-type −201 to +54 mSI-hGH (SI WT) transgene construct (42). Total RNA isolated from different portions of the adult intestine was analyzed by an RNase protection assay. Each RNA sample was incubated with probes for hGH and mSI in the same hybridization solution in order to correlate hGH mRNA levels with the endogenous SI mRNA levels. The SI WT transgene line showed high-level expression of hGH in the small intestine (Fig. 1B), as previously reported (42). The profile of expression of hGH mRNA in the small intestine in transgene lines carrying the SI MUT construct was similar to that in lines carrying the SI WT construct (Fig. 1C). However, a high level of hGH expression was detected in the cecum and proximal colon in each of the transgene lines harboring the mutant construct in contrast to the level of hGH in lines harboring the SI WT construct (Fig. 1C). The distal colon consistently had a weak level of hGH expression among the different lines analyzed (Fig. 1C). The pattern of transgene expression in the mutant lines was then further characterized during postnatal development by in situ hybridization. The expression of hGH in founder 5 and founder 8 lines was examined at multiple time points throughout postnatal intestinal development. The small and large intestines were mounted in Swiss roll configuration (see Materials and Methods), and sections were hybridized with either an antisense or sense ³³P-labeled hGH riboprobe. Exposure of the hybridized sections on an autoradiographic film revealed the level of transgene expression along the intestine from the proximal to the distal extremity and during different stages of development (Fig. 2). hGH expression was not detected before postnatal day 17 in the small intestine (Fig. 2A), as previously observed for the SI WT construct and endogenous SI mRNA (42) (data not shown). Expression of hGH in the SI MUT transgenic mice was strongly detected in the colon at postnatal day 4, and this level of expression was maintained throughout postnatal development (Fig. 2B). The proximal-to-distal gradient of transgene expression in the colon was observed at each time point during development. No signal was detected in the most distal portion of the colon (Fig. 2B). Serial sections hybridized with the sense riboprobe did not show a signal after exposure on film. The developmental pattern of transgene expression in the founder 5 line was identical to that in the founder 8 line (data not shown).

FIG. 1.

Functional importance of the common GATA and Cux/CDP sites for SI promoter activity in the mouse intestine. (A) Distribution and comparison of the GATA and Cux/CDP consensus sites within the mouse SI promoter. The two putative overlapping sites are aligned to consensus sites. The single nucleotide difference between the CDP consensus site and the SI promoter sequence is boxed. The GAT-to-TCG mutation is also indicated. (B) RNase protection analysis of hGH transgene expression. Total RNA (5 μg) isolated from intestinal tissues of a transgenic line with the genomic integrated −201 to +54 WT mSI-hGH construct (42) was used as a reference control in an RNase protection assay for the simultaneous detection of hGH and mSI mRNA. (C) Total RNA was isolated from intestinal tissues of three different adult founder (Fo) lines obtained with the −201 to +54 MUT mSI-hGH construct and was analyzed as described for panel B. Du, duodenum; PJ, proximal jejunum; DJ, distal jejunum; IL, ileum; Ce, cecum; PC, proximal colon; DC, distal colon.

FIG. 2.

Profile of transgene expression along the horizontal axis of the intestine during mouse postnatal (pn) development. Small intestines (A) and large intestines (B) of SI MUT transgenic mice were disposed and fixed in a Swiss roll configuration, and 5-μm-thick sections were subjected to a radioactive in situ hybridization with an antisense hGH probe (hGH) and exposed on a BIOMAX MR film for 24 h. Serial sections were counterstained with H&E to visualize the entire length of each intestinal segment. The proximal (prox.) and distal (dist.) intestinal extremities are indicated for each section.

The expression pattern of the transgene along the vertical axis of the intestine was further characterized. In the small intestines of the mice carrying the SI MUT construct, hGH mRNA was not detected before postnatal day 17 and the observed expression in enterocytes located at the crypt-villus junction and lower villus was restricted (Fig. 3A). In adult mice, crypts displayed no or very low levels of hGH mRNA expression. However, there was a marked increase in expression in the cytoplasm of the enterocytes localized at the crypt-villus junction and villus (Fig. 3B to D). This pattern of expression was similar to the one in SI WT mice (Fig. 3E). There was, however, a striking difference between colonic expression of the transgene in the SI MUT mice and in SI WT mice. The proximal colons of the SI MUT mice showed a gradient of transgene expression, with a dramatic increase in expression in the cytoplasm of colonocytes near the surface epithelium and very low levels in the bottom of the crypts (Fig. 4A to C). This pattern was found consistently at each time point during postnatal development. This was in contrast to results for the SI WT mice, in which no detectable transgene expression in the colon was observed (Fig. 4D). Therefore, the mutation of the element between nucleotides −73 and +64 that has consensus sites for both Cux/CDP and GATA does not affect the pattern of SI transcription during postnatal development of the small intestine. However, this element is crucial in the repression of colonic SI transcription during mouse postnatal and adult life. We name this element CRESIP.

FIG. 3.

Profile of transgene expression along the vertical axis of the small intestine during mouse postnatal development. Small intestine sections from postnatal day 17 (A) and adult (B to E) transgenic mice were subjected to radioactive (A, B, and E) or nonisotopic (C and D) in situ hybridization with an antisense hGH probe (AS hGH). Sections were counterstained with H&E when indicated. Black arrow, cytoplasm-restricted staining; red arrow, empty nucleus.

FIG. 4.

Profile of transgene expression along the vertical axis of the proximal colon during mouse postnatal development. Sections of proximal colons from postnatal (pn) days 4 and 17 and adult SI MUT (A to C) or SI WT (D) transgenic mice were subjected to radioactive (A and D) or nonisotopic (B and C) in situ hybridization with an antisense hGH probe (AS hGH). Sections were counterstained with H&E when indicated. Black arrow, cytoplasm-restricted staining; red arrow, nucleus.

Cux/CDP and GATA-4 proteins interact with CRESIP.

We examined whether Cux/CDP and GATA proteins could interact with CRESIP in vitro. EMSA was performed using either a wild-type CRESIP (WT-CRESIP)- or GAT-to-TCG-mutated CRESIP (MUT-CRESIP)-labeled probe that spans nucleotides −76 to −57 of the SI gene promoter (Fig. 1A). A full-length CDP-GST fusion protein was ineffective in interacting with CRESIP, corroborating the observation that the full-length CDP is incapable of a stable DNA binding interaction in vitro (data not shown) (24). Since the Cux/CDP transcription factor displays a complex protein structure that contains four different DNA-interacting domains (Fig. 5A, top), we investigated which of the specific DNA-interacting domains were capable of interacting with CRESIP. Comparable amounts of CR1-, CR2-, CR3-, HD-, and CR3/HD-GST fusion proteins were calibrated on Coomassie blue-stained SDS-PAGE gels (Fig. 5A, middle) and used in EMSA experiments. For all the fusion proteins tested, only the region that comprised both the CR3 and HD domains specifically interacted with the WT-CRESIP-labeled probe (Fig. 5A, bottom). No retarded complex was formed when the MUT-CRESIP-labeled probe was used under the same conditions (Fig. 5A, bottom). To test whether the HD domain was indeed essential for the interaction of the Cux/CDP protein with this site, we generated a Cux/CDP-GST fusion protein that lacked both the HD and C-terminal regions (Fig. 5B, top). Incubation of the fusion protein with the WT-CRESIP probe produced a specific retarded complex that was competed with the addition of an excess of unlabeled WT-CRESIP oligonucleotides but not with MUT-CRESIP oligonucleotides (Fig. 5B, bottom). No retarded complex was formed when the MUT-CRESIP-labeled probe was used under the same conditions. This indicated that the CR3/HD region binds to CRESIP but that the HD is dispensable for the Cux/CDP protein to interact with the SI promoter.

FIG. 5.

Characterization of Cux/CDP and GATA interaction with CRESIP. (A) GST fusion proteins for each different CR and HD domain of the Cux/CDP protein were generated (top) and calibrated by SDS-PAGE and Coomassie staining (middle). Arrow, partial cleavage of the GST tag. GST fusion proteins (GST-fus.; 50 ng ) were used for EMSA with labeled WT-CRESIP or MUT-CRESIP oligonucleotides (bottom). Competitions (Comp) were performed with a 1,000-fold molar excess of the WT-CRESIP or MUT-CRESIP unlabeled oligonucleotides. (B) A GST-Cux/CDP mutant (mut) protein that lacked portions of the N-terminal HD- and C-terminal regions was constructed (top). Western blot analysis was performed with a rabbit anti-GST polyclonal antibody on a GST affinity-purified bacterial extract that overexpressed the GST-Cux/CDPmut protein (middle). The GST-Cux/CDPmut fusion protein was used for EMSA and was competed with a 1,000-fold molar excess of the WT-CRESIP or MUT-CRESIP unlabeled oligonucleotides (bottom). (C) Nuclear extracts of Cos-7 cells (5 μg) transfected with an empty (ctl) or GATA-4 (G4) expression vector were used for each binding reaction. Competitions (Comp) were performed with a 100-fold molar excess of WT-CRESIP or MUT-CRESIP unlabeled oligonucleotides.

We next investigated whether GATA proteins could interact with CRESIP. Cos-7 cells were transfected with a GATA-4 expression vector, and nuclear extract from these cells was used for EMSA. A specific retarded complex was formed with the WT-CRESIP probe (Fig. 5C). This complex was competed with the addition of an excess of unlabeled WT-CRESIP oligonucleotides, while the addition of MUT-CRESIP oligonucleotides failed to compete (Fig. 5C). No complex was found when the MUT-CRESIP probe was incubated with GATA-4-containing nuclear extracts (Fig. 5C). In similar conditions, GATA-5 and GATA-6 did not efficiently interact with CRESIP (data not shown). Therefore, both Cux/CDP and GATA-4 proteins interact specifically with CRESIP.

Cux/CDP functions as a transcriptional repressor of the SI promoter.

We tested whether Cux/CDP and GATA-4 could functionally regulate SI gene promoter activity in intestinal cells. We first explored whether the Cux/CDP repressor could influence the transcriptional activity of the SI promoter. Cotransfection experiments using the Caco-2 cell line, which produces endogenous levels of SI after reaching confluence, were performed (3, 28). Early preconfluent Caco-2 cells do not have the potential to support SI promoter activity (7). We thus designed an experiment where the SI promoter could be activated with a constant amount of Cdx2, and then we tested the effect of Cux/CDP on activation of the SI promoter by comparing activation in the presence of Cux/CDP and Cdx2 to that in the presence of Cdx2 alone. Cotransfection of preconfluent Caco-2 cells with the Cdx2 expression vector alone produced strong activation of the SI promoter/luciferase construct (Fig. 6A). The addition of a Cux/CDP expression vector in the cotransfection assay resulted in a twofold reduction of the Cdx2-dependent activation of the SI promoter/luciferase construct in Caco-2 cells. The repressive properties of CDP have been shown to rely on the C-terminal region of the protein (18). Although our results demonstrated that a Cux/CDP recombinant protein that lacked both the C-terminal and HD domains was still capable of interacting with CRESIP (Fig. 5B), we tested whether this protein was functionally effective to repress SI promoter activity in Caco-2 cells. The Cux/CDPmut expression vector, designed to produce a truncated Cux/CDP protein lacking both the HD and C-terminal regions, was cotransfected with the Cdx2 expression vector and the SI promoter/luciferase construct. No significant reduction of the Cdx2-dependent activation of the SI promoter/luciferase construct was found when the Cux/CDPmut construct was used in this cotransfection assay (Fig. 6A). Point mutations of CRESIP that abolish the binding of the Cux/CDP protein (Fig. 5A) enhanced the Cdx2-dependent induction of the SI promoter and eliminated the repressive effect of Cux/CDP (Fig. 6A). Western analysis confirmed that both the Cux/CDP and the Cux/CDPmut proteins were synthesized when cotransfected in preconfluent Caco-2 cells and that exogenous Cdx2 and endogenous CDP/Cut protein production was not influenced under this condition (Fig. 6B). The exogenous Cux/CDP protein was shorter than the endogenous human CDP protein because of the use of an N-terminally truncated mouse CDP construct (43). The detection of two distinct migrating bands for each of the Cux/CDP proteins was intriguing and could be related to the susceptibility of CDP protein to specific proteolysis during cell cycle progression (24). Interestingly, no significant repressive effect of Cux/CDP was observed when cotransfected Caco-2 cells reached confluence (data not shown). Addition of either GATA-4, -5, or -6 expression vectors with the SI promoter/luciferase construct did not positively regulate the activity of the SI promoter in cotransfection experiments (data not shown). These observations suggest that the Cux/CDP protein can repress the SI promoter through CRESIP and that GATA proteins alone do not affect SI promoter activity.

FIG. 6.

Repressive effect of Cux/CDP on SI transcriptional activity in Caco-2 cells. (A) Caco-2 cells were transfected at low confluence (10 to 20%) with Lipofectamine and with 300 ng of either −201 to +54 mSI-pGL2basic (WT-SI/pGL2) or −201 + 54 GATmut mSI-pGL2basic (MUT-SI/pGL2) reporter vectors, 100 ng of simian virus 40-β-galactactosidase (βgal), and different combinations of pRC/CMV-Cdx2 (25 ng), pRC/CMV-Cux/CDP (600 ng), and pRC/CMV-Cux/CDPmut (600 ng) expression vectors, as indicated. The pRC/CMV plasmid was used as an empty control vector to calibrate the various amounts of expression vectors used in each condition. Results obtained in triplicate were reported as fold differences (means ± standard deviations) from results for transfection with the reporter construct alone and are representative of three independent experiments. (B) Western blot analysis with CDP (26) and Cdx2 antibodies was performed on protein extracts from Caco-2 cells cotransfected as described for panel A. The molecular mass for each protein is indicated. CDP/Cut, endogenous human CDP; Cux/CDP, transfected mouse CDP.

Cux/CDP displays a complex pattern of expression along the horizontal and vertical axes of the mouse intestine.

We next evaluated the cellular distribution of the Cux/CDP protein in the intestine by immunohistochemistry. The Cux/CDP protein was strongly detected in the nuclei of crypt cells (Fig. 7A) and more predominantly detected in the cytoplasm of enterocytes in the villus of the jejunum (Fig. 7A). Cux/CDP was also detected in cellular components located in the lamina propria (Fig. 7A). The detection of Cux/CDP in nuclei was uniform along the vertical axis of the ileum (Fig. 7B). The Cux/CDP protein was prominent in both the cytoplasm and nuclei of colonic epithelial cells, with a gradient of expression along the vertical axis and with strong expression near the surface epithelium (Fig. 7C). The specificity of the staining was confirmed by peptide-blocking experiments (data not shown). We next performed Western blot analysis on nuclear extracts prepared from isolated intestinal epithelial cells from different segments of the adult mouse intestine. A gradient of expression for the Cux/CDP protein was observed along the proximal-to-distal axis of the intestine. The Cux/CDP protein was weakly detected in the nuclei of enterocytes of the proximal jejunum but became strongly expressed in the proximal colon as demonstrated by immunoblots (Fig. 7D). YY1 protein expression was monitored as a control for protein integrity and remained relatively stable (Fig. 7D).

FIG. 7.

Expression profile of Cux/CDP protein along the vertical and horizontal axes of adult mouse intestine. (A to C) Immunohistochemistry was performed with a CDP polyclonal antibody on paraffin sections of adult mouse jejunum (A), ileum (B), and proximal colon (C). The CDP signal is brown (DAB). Black arrows, nuclear staining; red arrows cytoplasmic staining; black arrowheads, cellular components of the lamina propria. Magnification, ×170. (D) Western blot analysis was performed with CDP and YY1 antibodies on similar amounts of nuclear protein extracted from isolated adult mouse intestinal epithelial cells. PJ, proximal jejunum; IL, ileum; PC, proximal colon.

Cux/CDP is involved in SI gene repression during early postnatal intestinal development.

The mutation of CRESIP revealed the critical importance of this site for maintaining repression of the SI gene promoter during postnatal development in transgenic mice. Moreover, the Cux/CDP protein was able to repress SI promoter activity in transfection studies. Therefore, we further investigated the role of Cux/CDP in the regulation of SI expression during intestinal development with the use of a homozygous Cux/CDP mutant mouse model (34). As previously described, homozygous (CDP^ΔHD/ΔHD) mutant mice displayed partial lethality at or shortly after birth (34). The mutant mice had a reduced stature and gained little weight in comparison to the heterozygous and control mice (34). We first confirmed the level of Cux/CDP protein in the intestinal epithelia of the homozygous (CDP^ΔHD/ΔHD) mice. Western analysis was performed with nuclear extracts prepared from isolated colonocytes of mutant and control mice. The wild-type Cux/CDP protein (180 to 190 kDa) was detected in protein extracts from control heterozygous mice but was absent from homozygous mutant mice (Fig. 8A). As previously observed in thymic nuclear extracts (34), a shorter and less-intense band (150 to 160 kDa), which corresponded to the Cux/CDP^ΔHD protein, was produced in both the heterozygous and homozygous mutant mice (Fig. 8A). Levels of Cdx2 protein expression in the mutant and control animals were comparable, confirming the integrity of the nuclear proteins (Fig. 8A). We next assessed whether intestinal SI gene expression was affected in the few surviving Cux/CDP^ΔHD/ΔHD mice. Total RNA was isolated from the entire small intestines or the colons of mice sacrificed at postnatal days 4, 7, and 16 and analyzed by RNase protection assay. Each RNA sample was incubated with a probe for either mSI or m36B4 as a control for RNA integrity. SI mRNA expression was induced at postnatal day 4 in the small intestines of Cux/CDP^ΔHD/ΔHD mice, whereas levels of SI expression in the colons of both mutant and control animals were similarly low (Fig. 8B). At postnatal day 7, SI mRNA levels were induced in both the small intestines and the colons of the mutant Cux/CDP^ΔHD/ΔHD mice (Fig. 8B). The SI mRNA level was modestly increased in the small intestines of mutant mice compared to that in small intestines of control mice at postnatal day 16, a period when intestinal SI gene transcription is dramatically induced in the small intestine epithelium (42). A high level of SI mRNA was still detected in the colons of Cux/CDP^ΔHD/ΔHD mice (Fig. 8B, day 16). Following the transition from suckling to weaning, SI mRNA was no longer expressed in the colons of the very few surviving mutant mice (data not shown). These results confirm that the Cux/CDP protein acts as a repressor of SI gene transcription within an early window of intestinal postnatal development.

FIG. 8.

SI mRNA expression is altered in CDP^ΔHD/ΔHD mutant mice. (A) Western blot analysis was performed with CDP (26) and Cdx2 antibodies on similar amounts of colonic protein extracts. Arrows, Cux/CDP wild-type (wt) and truncated mutant protein (mut), respectively. (B) Total RNA (5 μg) isolated from intestinal tissues of CDP control (+/ΔHD) and mutant (ΔHD/ΔHD) mice was analyzed by an RNase protection assay for mSI and m36B4 mRNA. Sm. Int., small intestine.

DISCUSSION

The SI gene represents a useful model to study intestinal mechanisms that coordinate gene activation and repression. SI expression is tightly regulated during postnatal development and becomes strongly induced in the small intestine at the transition from suckling to weaning (40). SI reexpression in the colonic epithelium is also observed under certain circumstances such as colonic neoplasia (4, 9, 45). Our study has undertaken the functional characterization of a novel colonic repressor element of the SI promoter (CRESIP) in the intestinal epithelium. CRESIP interacts with both Cux/CDP and GATA-4 proteins. Cux/CDP functionally represses the SI promoter in transfection studies and Cux/CDP mutant mice express induced levels of SI mRNA during early postnatal intestinal development. These findings demonstrate that the SI gene promoter is actively repressed in the intestine and identify Cux/CDP as a repressor of SI.

Our laboratory has focused on transcription factors that appear to be important in the positive regulation of SI, such as HNF-1α (7) and Cdx2 (37). Although HNF-1α and Cdx2 are expressed along the entire tract of the adult intestine, there are two possible explanations for the lack of SI expression in the normal colonic mucosa: either other uncharacterized activators of SI are not expressed or unknown repressors of SI are active in colonocytes. The determination that a mutation of CRESIP results in strong SI promoter activity in the colonic epithelium suggests that repression is most likely the mechanism involved in the regulation of SI colonic expression. One may hypothesize that the release of repression mechanisms can promote both HNF-1α and Cdx2 proteins to interact with the SI promoter and lead to the establishment of a stable activator complex that can support a high level of SI gene expression. This concept correlates well with the decreasing gradient of transgene expression observed along the proximal-to-distal colons of the SI MUT transgenic mice (Fig. 2B). Indeed, the Cdx2 protein is subjected to a similar pattern of decreasing expression along the horizontal axis of the colonic epithelium at early stages during mouse postnatal development (33). Although we present some functional evidence that Cux/CDP competes with the Cdx2 activation of the SI promoter via CRESIP, the specific and detailed mechanisms that underline the Cux/CDP action on the mouse SI gene promoter in the intestinal epithelium context will require additional work.

The Cux/CDP homozygous mutant mice showed an important role for the intact Cux/CDP protein in the repression of SI gene expression in the intestinal epithelium in vivo. However, an important issue is raised since the pattern of SI derepression is also found in the early postnatal development stage of the small intestine. In addition, colonic SI gene expression becomes fully repressed following the transition from suckling to weaning in the rare surviving null mice. These observations illustrate important discordance with the characterization of CRESIP in SI MUT transgenic mice. Indeed, the SI promoter does not support detectable hGH reporter expression before the transition from suckling to weaning in the small intestine and is strongly activated in the colon throughout adult life. We propose several explanations to account for these differences. First, it is possible that other putative silencing elements localized outside the short SI promoter participate with CRESIP in SI gene repression during intestinal development. Second, we cannot exclude the possibility that the GATA site is important in promoting expression of SI during early development, when Cux/CDP-dependent transcriptional repression is prevented. Third, the production of the truncated Cux/CDP protein in the intestines of the mutant mice could still be effective in repressing the SI gene in the in vivo context. Indeed, the CR domains of the CDP protein have recently been demonstrated to be sufficient by themselves for the displacement properties of the CDP/Cut protein (23). The Cux/CDP mutant protein is able to interact in vitro with CRESIP but inefficiently represses the Cdx2-dependent activation of the SI promoter in cotransfection experiments with Caco-2. Whether this situation is reflected in the in vivo context is hard to predict mainly because of the possible complexity involved in the establishment of the chromatin-modifying activities that specifically control SI gene transcription. Finally, it is plausible that other unidentified Cux/CDP-related proteins could partially compensate for the Cux/CDP mutation later during adult life. This hypothesis is supported by the fact that some mutant mice are able to survive for several months. The compensation hypothesis has also been proposed for mice where the coding sequences for C-terminal CR3 and HD domains have been replaced with an in-frame lacZ gene because of the lack of morphological alterations in various tissues that normally express Cux/CDP (12). Another member of the Cux/CDP family, Cux2, displays functions similar to those of Cux/CDP but is restricted to the neural tissue (30). We thus hypothesize that the Cux/CDP protein is involved in intestinal repression of the SI gene via CRESIP but that other compensatory proteins, especially during adult life, are involved in this complex process.

Several studies suggest a role for CDP in the control of cellular proliferation and/or differentiation (25). Genetic studies of flies concluded that the Cut homologue functions as a determinant of cell type specification (5, 6, 15). In addition, it has been suggested that CDP/Cut serves as a cell cycle-dependent transcriptional factor in proliferating cells (8). CDP/Cut has also been reported to form a complex with retinoblastoma protein-related protein p107 and cyclin A (44) and to control the expression of several genes involved in the control of cellular proliferation such as the p21 (8), thymidine kinase (16), and histone genes (11) and c-myc (10). Interestingly, the c-Myc oncogene is an important target of the adenomatous polyposis coli pathway and is overexpressed in colorectal cancer (14). Since the CRESIP mutation leads to derepression of the SI gene promoter during early postnatal intestinal development in transgenic mice, it is tempting to hypothesize that this element and, by extension, the Cux/CDP protein are involved in SI derepression during the early progression of colonic neoplasia. Whether this protein belongs to specific pathways that regulate the proliferation versus differentiation of intestinal epithelial cells remains to be investigated. Interestingly, a reduction of the protein level of human homologue CDP/Cut is observed in various human adenoma-carcinoma cell lines (F. Boudreau and P. G. Traber, unpublished results).

In conclusion, we have identified a novel DNA element that is involved in the in vivo colonic epithelial repression of the SI gene. Based on transfection and gene deletion studies, we propose that the Cux/CDP transcriptional repressor is partly involved in this repression by acting via CRESIP. One challenge is to clarify the putative role of the CDP protein in the progression of colonic neoplasia and to identify other transcriptional factors that interact with CRESIP.