The forkhead transcription factor Foxi1 directly activates the AE4 promoter

Abstract

Intercalated cells are highly specialized cells within the renal collecting duct epithelium and play an important role in systemic acid-base homoeostasis. Whereas type A intercalated cells secrete protons via an apically localized H+-ATPase, type B intercalated cells secrete HCO3-. Type B intercalated cells specifically express the HCO3-/Cl- exchanger AE4 (anion exchanger 4), encoded by Slc4a9. Mice with a targeted disruption of the gene for the forkhead transcription factor Foxi1 display renal tubular acidosis due to an intercalated cell-differentiation defect. Collecting duct cells in these mice are characterized by the absence of inter-calated cell markers including AE4. To test whether Slc4a9 is a direct target gene of Foxi1, an AE4 promoter construct was generated for a cell-based reporter gene assay. Co-transfection with the Foxi1 cDNA resulted in an approx. 100-fold activation of the AE4 promoter construct. By truncating the AE4 promoter at the 5'-end, we demonstrate that a fragment of approx. 462 bp upstream of the transcription start point is sufficient to mediate activation by Foxi1. Sequence analysis of this region revealed at least eight potential binding sites for Foxi1 in both sense and antisense orientation. Only one element was bound by recombinant Foxi1 protein in bandshift assays. Mutation of this site abolished both binding in bandshift assays and transcriptional activation by co-transfection of Foxi1 in the reporter gene assay. We thus identify the AE4 promoter as a direct target of Foxi1.

Figure 1. The AE4 promoter region

(A) Database search identified three regions within the 5′-region of the Ae4 gene that are conserved between mice and human indicating that these regions may be functionally important. The genomic regions promH1, promH2 and promH3, which include one, two or three conserved regions, were cloned into a luciferase expression vector. (B) Transcriptional activity of the promoter constructs in HEK-293T cells was determined by a luciferase-based reporter gene assay. Activity is shown as fold activation beyond basal activity of the pGL3 vector. The experiment was performed in triplicate and was repeated twice.

Figure 2. Transcriptional activities of the promoter constructs in HEK-293T cells determined by a luciferase-based reporter gene assay

Activities of the promoter constructs after co-transfection with Foxi1 are shown as fold activation compared with basal activity of the respective promoter construct, which was set as 1. Each experiment was performed in triplicate and was repeated at least twice. (A) pGL3 (1.0 μg of DNA) and pCMV (0.5 μg of DNA) vector served as a negative control. HEK-293T cells were transfected with 1.0 μg of promH1 and increasing amounts of Foxi1 cDNA as indicated. (B, C) Transfection of the AE4 promoter constructs promH1, promH2 and promH3 (1.0 μg of DNA each) in HEK-293T cells (B) and COS-7 cells (C) with pCMV (−) or with the pCMV-Foxi1 cDNA (0.5 μg) (+). All promoter constructs were strongly activated by co-transfection with Foxi1. Co-transfection of the pGL3 vector (1 μg) or the myoglobin promoter (1 μg of DNA) with Foxi1 did not result in significant luciferase activity.

Figure 3. Nucleotide sequence of the 5′-end of the AE4 gene (nt −462/−1)

Eight potential binding motifs in either sense or antisense direction were identified within the sequence as used for the promH3 construct, if stringency criteria were limited to a minimal core sequence TxTTT in the sense and antisense orientation. Motifs are highlighted by boldface letters. The oligonucleotides used for EMSA experiments are framed.

Figure 4. EMSA of recombinant Foxi1 with radioactively labelled oligonucleotides spanning putative Foxi1-binding sites

Labelled oligonucleotides were incubated with either bacterially expressed GST–Foxi1 (+) or bacterially expressed GST (−). Seven of these putative motifs did not show significant Foxi1 binding. A new band corresponding to Foxi1–DNA complexes was only formed with oligonucleotides −270/−243 (arrowhead).

Figure 5. Specificity of Foxi1 binding to the motif −270/−243

Whereas GST–Foxi1 binding to radioactively labelled oligonucleotide −270/−243 could be competed with increasing amounts of the unlabelled oligonucleotide (lane 3, 5-fold excess of unlabelled oligonucleotides; lane 4, 10-fold; lane 5, 50-fold; and lane 6, 100-fold), competition was not achieved with a mut-oligo, where the three central thymidines in positions −258/−256 had been replaced by guanines (mut-oligo −270/−243) (lane 8, 5-fold excess of unlabelled mut-oligo; lane 9, 10-fold; lane 10, 50-fold; and lane 11, 100-fold). The ³²P-labelled mut-oligo −270/−243 did not bind to GST–Foxi1 (lanes 12 and 13). Lane 1, labelled oligonucleotide −270/−243 in the presence of GST without Foxi1; lanes 2 and 7, labelled oligonucleotide −270/−243 in the presence of GST–Foxi1 without competitor.

Figure 6. The motif −270/−243 is a Foxi1-response element

Reporter plasmid promH3Δ lacking the identified binding motif −270/−243 and a mutant promH3 (promH3 mut), where the core sequence TTT in positions −258/−256 had been replaced by GGG, were analysed in the cell-based luciferase reporter assay. Cells transfected with the respective plasmids did not show induced luciferase activity in case of promH3Δ, and the mutated variant had low residual Foxi1-dependent transactivation activity (3-fold). Basal activity of the respective promoter construct without Foxi1 was defined as 1.