Distinct functions are implicated for the GATA-4, -5, and -6 transcription factors in the regulation of intestine epithelial cell differentiation

Abstract

Based on conserved expression patterns, three members of the GATA family of transcriptional regulatory proteins, GATA-4, -5, and -6, are thought to be involved in the regulation of cardiogenesis and gut development. Functions for these factors are known in the heart, but relatively little is understood regarding their possible roles in the regulation of gut-specific gene expression. In this study, we analyze the expression and function of GATA-4, -5, and -6 using three separate but complementary vertebrate systems, and the results support a function for these proteins in regulating the terminal-differentiation program of intestinal epithelial cells. We show that xGATA-4, -5, and -6 can stimulate directly activity of the promoter for the intestinal fatty acid-binding protein (xIFABP) gene, which is a marker for differentiated enterocytes. This is the first direct demonstration of a target for GATA factors in the vertebrate intestinal epithelium. Transactivation by xGATA-4, -5, and -6 is mediated at least in part by a defined proximal IFABP promoter element. The expression patterns for cGATA-4, -5, and -6 are markedly distinct along the proximal-distal villus axis. Transcript levels for cGATA-4 increase along the axis toward the villus tip; likewise, cGATA-5 transcripts are largely restricted to the distal tip containing differentiated cells. In contrast, the pattern of cGATA-6 transcripts is complementary to cGATA-5, with highest levels detected in the region of proliferating progenitor cells. Undifferentiated and proliferating human HT-29 cells express hGATA-6 but not hGATA-4 or hGATA-5. Upon stimulation to differentiate, the transcript levels for hGATA-5 increase, and this occurs prior to increased transcription of the terminal differentiation marker intestinal alkaline phosphatase. At the same time, hGATA-6 steady-state transcript levels decline appreciably. All of the data are consistent with evolutionarily conserved but distinct roles for these factors in regulating the differentiation program of intestinal epithelium. Based on this data, we suggest that GATA-6 might function primarily within the proliferating progenitor population, while GATA-4 and GATA-5 function during differentiation to activate terminal-differentiation genes including IFABP.

FIG. 1

The xIFABP gene is transcribed in adult Xenopus gut specifically in the differentiating epithelium. Xenopus intestine was analyzed by whole-mount in situ hybridization. Sections of processed tissue are shown. (a) Villus architecture, with smooth-muscle (SM) layers on the left and the villus proximal-distal axis running left to right toward the tip (T). The dark strain indicates the pattern of xIFABP transcripts following development of the alkaline phosphatase reaction to detect the hybridized antisense xIFABP RNA probe. (b) Higher-magnification view of a villus section. Arrows in both panels indicate the positions along the villus axis that transcripts are first detected in the epithelium. The signal increases in intensity along the proximal-distal axis and is strongest at the distal tip (T). (c) Section from similarly processed tissue that was hybridized with a control sense-strand RNA probe.

FIG. 2

Structure of the xIFABP gene and sequence of the promoter region. (a) Genomic organization of exons along a 7-kb XbaI fragment. The positions of BglII and SacI restriction sites are also indicated. Solid boxes represent the four exons, and the arrow and arrowhead indicate the positions of the ATG translational initiation codon and stop codon, respectively. The positions of the exons, placement of the introns, and sizes of the introns separating the exons were determined by PCR and direct sequencing. (b) Sequence encompassing the exon/intron boundaries, relative to the xIFABP coding sequence. (c) Sequence of the xIFABP gene from −969 (around the upstream XbaI site) to +75, relative to the transcriptional start site (see Fig. 3). Small brackets (three located between −360 and −280 and one located around −42) indicate several sequences consistent with the GATA consensus binding sites. The larger bracket around −290 encompasses a conserved sequence shown in previous experiments to be important for regulating IFABP expression in transgenic mice (45). Note that the deviated GATA site in this sequence (TGATG) is in the reverse orientation. The vertical brace indicates the position at which upstream sequences were deleted in the Δ233 reporter. The long double-headed arrow indicates the sequence of the oligomer probe used in the gel mobility shift experiments (Fig. 5 and 8), and the small arrowheads point to the bases mutated in the pm reporter (Fig. 4) and in the mutant oligomers used in competition experiments (Fig. 5 and 8). The angled arrow indicates the transcriptional start site designated +1 (from the data in Fig. 3), and the solid triangle identifies the ATG translational initiation codon.

FIG. 2

Structure of the xIFABP gene and sequence of the promoter region. (a) Genomic organization of exons along a 7-kb XbaI fragment. The positions of BglII and SacI restriction sites are also indicated. Solid boxes represent the four exons, and the arrow and arrowhead indicate the positions of the ATG translational initiation codon and stop codon, respectively. The positions of the exons, placement of the introns, and sizes of the introns separating the exons were determined by PCR and direct sequencing. (b) Sequence encompassing the exon/intron boundaries, relative to the xIFABP coding sequence. (c) Sequence of the xIFABP gene from −969 (around the upstream XbaI site) to +75, relative to the transcriptional start site (see Fig. 3). Small brackets (three located between −360 and −280 and one located around −42) indicate several sequences consistent with the GATA consensus binding sites. The larger bracket around −290 encompasses a conserved sequence shown in previous experiments to be important for regulating IFABP expression in transgenic mice (45). Note that the deviated GATA site in this sequence (TGATG) is in the reverse orientation. The vertical brace indicates the position at which upstream sequences were deleted in the Δ233 reporter. The long double-headed arrow indicates the sequence of the oligomer probe used in the gel mobility shift experiments (Fig. 5 and 8), and the small arrowheads point to the bases mutated in the pm reporter (Fig. 4) and in the mutant oligomers used in competition experiments (Fig. 5 and 8). The angled arrow indicates the transcriptional start site designated +1 (from the data in Fig. 3), and the solid triangle identifies the ATG translational initiation codon.

FIG. 3

Primer extension analysis mapping of the xIFABP transcriptional start site. RNA from stage (st.) 62 tadpoles or stage 66 adult frogs was analyzed with a specific labeled primer. The reaction products were analyzed on a denaturing polyacrylamide gel alongside sequencing ladders made with the same primer and the genomic clone as a template. The arrow indicates the position of the single extension product.

FIG. 4

The xIFABP promoter is a direct target for transactivation by GATA factors. The graphs show the relative activity of a reporter gene regulated by the xIFABP promoter. (A) The basal activity of the full length (−1 to −969) promoter when cotransfected with the control (empty) expression vector was arbitrarily assigned a value of 1 (Vec). The relative activity of this promoter when cotransfected with expression vectors for xGATA-4 (G4), xGATA-5 (G5), or xGATA-6 (G6) is shown. The data were averaged from at least four independent experiments, and error bars indicate the standard deviations. In all cases, the transfection includes saturating amounts of the GATA factor expression vector, determined in preliminary transfections. (B) Similar transfections were performed with the xGATA-4 expression vector but with either the full-length reporter (WT) (presented as 100% activity) or reporters containing mutated promoters as indicated. The activity of a control GATA-dependent promoter (αD3 [14]) is shown for comparison. Note that the majority (but not all) of the reporter activity in the presence of GATA-4 is eliminated when the proximal GATA site is mutated (pm).

FIG. 5

GATA factors bind specifically to the xIFABP proximal promoter element. Gel mobility shift assays were performed by using as probe an oligomer containing sequences from −58 to −31 (Fig. 2c), including the proximal consensus GATA-binding site. Extracts were derived from cells transfected with the expression vector alone or the xGATA-4, xGATA-5, or xGATA-6 expression vector, as indicated. In addition to the probe and extract, incubation mixtures contained no competitor (lanes 1), competitor containing the WT promoter sequence (TE492/TE493) (lanes 2), competitor containing the pm mutated sequence that alters the GA of the consensus GATA-binding site (TE452/TE453) (lanes 3), or competitor containing a well-characterized GATA-binding site from the chicken α^D-globin promoter TE72/TE73 (lanes 4). The position of the major specific complex (arrow) and the position of the labeled probe (P) are indicated. The GATA-6 reactions consistently generate a less-abundant complex; it is not known if this is due to lower expression levels or binding affinity.

FIG. 6

GATA factors are differentially regulated along the villus axis within intestinal epithelium. (a) Intestine was isolated from 1-month-old chicks and processed by in situ whole-mount hybridization. Sections from tissues incubated with antisense RNA probes for cGATA-4 (upper panel), cGATA-5 (middle panel), or cGATA-6 (lower panel) are shown. The dark signal indicates the pattern of GATA factor transcripts, as in Fig. 1. The relative positions of the crypt (c) progenitor cells, the Paneth (p) cells, and the distal-tip (T) cells are indicated. Note that GATA-4 transcripts are distributed from the crypt to the tip, with increasing levels accumulating toward the distal tip. In contrast, GATA-5 transcripts are highly localized within differentiating cells of the tip, and GATA-6 transcript levels are highest in the less differentiated region located closer to the crypt zone. (b) Higher-magnification views of sections derived from tissue hybridized to probes for cGATA-5 (upper panel) or cGATA-6 (lower panel). Note that the transcript patterns are essentially complementary.

FIG. 6

GATA factors are differentially regulated along the villus axis within intestinal epithelium. (a) Intestine was isolated from 1-month-old chicks and processed by in situ whole-mount hybridization. Sections from tissues incubated with antisense RNA probes for cGATA-4 (upper panel), cGATA-5 (middle panel), or cGATA-6 (lower panel) are shown. The dark signal indicates the pattern of GATA factor transcripts, as in Fig. 1. The relative positions of the crypt (c) progenitor cells, the Paneth (p) cells, and the distal-tip (T) cells are indicated. Note that GATA-4 transcripts are distributed from the crypt to the tip, with increasing levels accumulating toward the distal tip. In contrast, GATA-5 transcripts are highly localized within differentiating cells of the tip, and GATA-6 transcript levels are highest in the less differentiated region located closer to the crypt zone. (b) Higher-magnification views of sections derived from tissue hybridized to probes for cGATA-5 (upper panel) or cGATA-6 (lower panel). Note that the transcript patterns are essentially complementary.

FIG. 7

Changes in GATA factor transcript levels correlate with induction of terminal differentiation in human gut epithelium cell lines. (a) RT-PCR analysis was used to measure the relative levels of RNA encoding human GATA-4, -5, and -6 (G4, G5, and G6, respectively) in samples derived from uninduced and proliferating CaCo-2, HT-29, or SW1417 cells. The S14 gene encodes a relatively abundant rRNA protein message that does not change significantly in different samples and is used as a positive control for the RT reactions. PCR products were labeled by including trace radiolabeled nucleotides in the reaction. The products were detected after gel electrophoresis by autoradiography. (b) HT-29 cells were either uninduced (lane 0) or induced with 5 mM sodium butyrate, and RNA was harvested at various times (in hours) to measure relative transcript levels by semiquantitative RT-PCR analysis. A representative autoradiograph following gel electrophoresis of the PCR products is shown. The IAP gene is a terminal-differentiation marker that is induced to high levels by 48 h. The hGATA-5 gene is an early target for activation by sodium butyrate, while the hGATA-6 transcript levels decline initially during differentiation before recovering at later time points. Similar kinetics were consistently noted in multiple experiments. (c) The same RNA as that used for the RT-PCR analysis in panel b was analyzed for hGATA-6 mRNA in a Northern blotting experiment. Total RNA was electrophoresed, blotted, and hybridized to an hGATA-6 cDNA probe. As shown in the upper panel, the RNA levels decrease prior to reaccumulating by 48 h, confirming the RT-PCR results. The lower panel shows the ethidium bromide-stained 28S rRNA, demonstrating equal RNA loading for each lane.

FIG. 8

Nuclear extracts derived from HT-29 cells contain binding activity that interacts specifically with the IFABP proximal GATA element, and the levels of this activity increase following sodium butyrate-induced differentiation. (a) The oligomer probe containing the xIFABP proximal promoter element (TE492/TE493, as in Fig. 5) was end labeled and incubated with nuclear extracts derived from uninduced (control) or induced cells cultured for 48 h in the presence of butyrate. Control cultures were incubated for 48 h in a mock induction, and equal amounts of total protein lysates were used in mobility shift assays. The positions of the free probe (P) and a complex formed with a nonspecific DNA-binding activity (ns) are indicated. The first lane contains the probe incubated in buffer alone. Lanes: 1, no specific competitor DNA; 2, 100-fold-excess unlabeled-probe competitor; 3, 100-fold-excess competitor DNA containing a specific mutation of the GATA cis element (TE452/TE453); 4, 100-fold-excess competitor containing the GATA-binding site from the chicken α^D-globin promoter (TE72/TE73). Note that the specific complex (large arrow), but not the nonspecific complex, is increased in abundance severalfold in extracts from induced cells. Lane xG4, extract from QT6 cells transfected with the xGATA-4 expression vector (as in Fig. 5) as a positive control for a specific complex (small arrow). (b) Gel mobility shift experiments were performed as for panel a with extracts derived from uninduced HT-29 cells (0 h) or cells induced for 24 or 48 h. In all cases, cells were cultured for 48 h following addition of butyrate to the 48-h sample; equal amounts of total nuclear proteins were incubated with the probe. Lanes: 1, no additional competitor; 2, 1 μl of preimmune rabbit serum; 3, 1 μl of immune serum derived from rabbits injected with synthetic peptides consistent with several regions of the hGATA-6 sequence. Note that the antibodies specifically inhibit or disrupt formation of the complex (arrow), while preimmune serum actually enhances binding. Although it is not known, the antibody is likely to interfere with binding in all GATA-4, -5, and -6 interactions in this experiment. P, free probe.