Differential regulation of the glucose-6-phosphatase TATA box by intestine-specific homeodomain proteins CDX1 and CDX2

Abstract

Glucose-6-phosphatase (Glc6Pase), the last enzyme of gluconeogenesis, is only expressed in the liver, kidney and small intestine. The expression of the Glc6Pase gene exhibits marked specificities in the three tissues in various situations, but the molecular basis of the tissue specificity is not known. The presence of a consensus binding site of CDX proteins in the minimal Glc6Pase gene promoter has led us to consider the hypothesis that these intestine-specific CDX factors could be involved in the Glc6Pase-specific expression in the small intestine. We first show that the Glc6Pase promoter is active in both hepatic HepG2 and intestinal CaCo2 cells. Using gel shift mobility assay, mutagenesis and competition experiments, we show that both CDX1 and CDX2 can bind the minimal promoter, but only CDX1 can transactivate it. Consistently, intestinal IEC6 cells stably overexpressing CDX1 exhibit induced expression of the Glc6Pase protein. We demonstrate that a TATAAAA sequence, located in position -31/-25 relating to the transcription start site, exhibits separable functions in the preinitiation of transcription and the transactivation by CDX1. Disruption of this site dramatically suppresses both basal transcription and the CDX1 effect. The latter may be restored by inserting a couple of CDX- binding sites in opposite orientation similar to that found in the sucrase-isomaltase promoter. We also report that the specific stimulatory effect of CDX1 on the Glc6Pase TATA-box, compared to CDX2, is related to the fact that CDX1, but not CDX2, can interact with the TATA-binding protein. Together, these data strongly suggest that CDX proteins could play a crucial role in the specific expression of the Glc6Pase gene in the small intestine. They also suggest that CDX transactivation might be essential for intestine gene expression, irrespective of the presence of a functional TATA box.

Figure 1

Glc6Pase promoter activity in Caco-2 and HepG2 cells. Caco-2 cells (black bars) and HepG2 (white bars) were transiently transfected with each luciferase reporter plasmid containing different fragments of the Glc6Pase promoter (1 µg) together with pCMV-RL plasmid (2 ng) as a control for correction for transfection efficiency. LUC activity was determined 48 h after transfection and was normalized relative to the level of RL activity. The transcriptional activity of each construct is expressed relative to the LUC activity of pGL2‘basic’ and is the mean ± S.E.M. of at least three independent experiments performed in duplicate.

Figure 2

Transactivation of the Glc6Pase promoter by CDX1 and CDX2 in HepG2 cells. HepG2 cells were transiently transfected with the 5′-end deleted Glc6Pase promoter fragments fused to a LUC reporter gene in pGL2‘basic’ (1 µg), pCMV-RL (2 ng, used as internal control), without or with pCDX1-S (100 ng) or pCDX2-S (100 ng). LUC activities were determined 48 h after transfection and were normalized relatively to the level of RL activities. The transactivation level of each construct by CDX1 (black bars) and CDX2 (grey bars) is expressed as fold of induction over the basal condition (without CDX) relatively of pGL2‘basic’. Results are the mean ± S.E.M. of at least three independent experiments performed in duplicate. * and **, significantly different from without CDX, P < 0.05 and P < 0.01, respectively.

Figure 3

The overexpression of CDX1 in IEC6 cells induces the Glc6Pase expression. The protein expression was analysed by western blotting on whole protein extracts from IEC6-CAT cells or IEC6-CDX1 cells. Blots were revealed with anti-CDX1 (upper panel), anti-Glc6Pase (middle panel) or anti-actin (lower panel).

Figure 4

Specific binding of CDX1 to the TATA box of the Glc6Pase promoter and interaction with the TATA-binding protein. (A) Gel shift assays were carried out with 5 (lanes 1) or 15 µg (lanes 2 and 3) of whole cell extracts from HeLa cells, differentiated Caco-2 cells, CDX1-expressing HeLa cells, or CDX2-expressing HeLa cells and double-strand radiolabelled oligonucleotide matching the Glc6Pase TATAAAA sequence. Control was performed without protein (no extract). Competition experiments were performed in the presence of 100-fold excess of unlabelled Glc6Pase TATAAAA oligonucleotide (lanes 3). ‘1’ and ‘2’ on the left refer to DNA–protein complexes. The arrowhead above 2 indicates the presence of a non-reproducible complex. (B) Gel shift assays were carried out with 15 µg of whole cell extract from HeLa cells and double-strand radiolabelled oligonucleotide matching the Glc6Pase TATAAAA sequence. Competition experiments were performed with 100-fold excess of unlabelled oligonucleotides (Glc6Pase TATAAAA and consTATA box). Supershift experiment was performed with anti-TBP antibody, and indicated by an asterisk. (C) CDX1 and TBP interaction. HTC116 cells were co-transfected with vehicle alone (pCB6 vector), HA-CDX1 expressing vector (pCB6-HA-CDX1) or HA-CDX2 expressing vector (pCB6-HA-CDX2) and TBP expressing vector (pXJ41-hTBP plasmid). Immunoprecipitation (IP) was performed with anti-HA antibody and revealed by western blotting (WB) with anti-HA (upper panel) or anti-TBP (lower panel).

Figure 5

Specific transactivation and binding competition of CDX proteins on the Glc6Pase promoter. (A) Amino acid sequence of the homeodomains of the murine and human CDX1 and CDX2 proteins. Uppercase letters indicate the differences between these proteins. (B) HepG2 cells were transiently transfected with the –35/+60B construct (1 µg) together with pCMV-RL (2 ng, used as internal control), without or with 100 ng of pCDX1-S or pCDX2-S, or pCDX1HD2, or pCDX2HD1. (C) HepG2 cells were transiently transfected with the –35/+60B construct (1 µg) together with pCMV-RL (2 ng), in the presence of pCDX1-S (100 ng) and pCDX2-S (1–200 ng, black circles) or pCDX2-HD1 (1–200 ng, black triangles), or pCDX1-HD2 (1–200 ng, black squares) expression vectors. LUC activities were determined 48 h after transfection and were normalized with regards to the level of RL activities. The transactivation rate is expressed as fold of induction over the basal condition. Results are the mean ± S.E.M. of at least three independent experiments performed in duplicate.*, significantly different from without CDX, P < 0.01.

Figure 6

Mutation of the TATA box alters the CDX1 transactivation of the Glc6Pase promoter. (A) Mutations (M1 and M2) introduced in the TATAAAA region of the Glc6Pase promoter. Wild type (WT) TATA-box site is underlined. Mutated bases are in lowercase and bold. (B and C) HepG2 cells were transiently transfected with the –80/+60B and –55/+60B wild type (WT) or mutated constructs (1 µg) together with pCMV-RL (2 ng, used as internal control), without or with 100 ng of pCDX1-S. LUC activities were determined 48 h after transfection and were normalized with regards to the level of RL activities. (B) Basal transcription activity of the wild type and mutated forms of the –80/+60B constructs. The transcriptional activity of each construct is expressed relative to the LUC activity of pGL2‘basic’ and is the mean ± S.E.M. of at least three independent experiments performed in duplicate. **, significantly different from pGL2basic activity, P < 0.01. (C) Transactivation level by CDX1, expressed as fold of induction over the basal condition (without CDX1). Results are the mean ± S.E.M. of at least three independent experiments performed in duplicate. **, significantly different from without pCDX1-S, P < 0.01.

Figure 7

Insertion of CDX-BS to overcome the lack of functional TATA box. (A) The mutations introduced to replace the TATA box of the Glc6Pase promoter in the –80/+60B TATAM2 construct. Wild-type (WT) TATA box site is in bold uppercase. Mutated bases are in lowercase and bold. The mutants contain one CDX-BS (T-CBS-S, T-CBS-AS) or two CDX-BS (T-CBS-AS+S and T-CBS-AS+AS) instead of the TATA-box. Arrows indicate the CDX-BS in sense and antisense orientation. Sense orientation corresponds to the CDX-BS formed by the TATA-box. (B) Relative transactivation level by CDX1 of the mutants compared to the wild type construct. HepG2 cells were transiently transfected with the –80/+60B wild type (WT) or mutated constructs (1 µg) together with pCMV-RL (2 ng, used as internal control), without or with 100 ng of pCDX1-S. LUC activities were determined 48 h after transfection and were normalized with regards to the level of RL activities. Transactivation level by CDX1 was expressed as fold of induction over the basal condition (without CDX1) and compared with that obtained with the WT construct (100%). Results are the mean ± S.E.M. of at least three independent experiments performed in duplicate. **, significantly different from WT construct transactivation, P < 0.01.