PDX-1 interaction and regulation of the Pancreatic Derived Factor (PANDER, FAM3B) promoter

Abstract

Pancreatic Derived Factor (PANDER) is a novel cytokine-like protein dominantly expressed within the endocrine pancreas. Our previous study demonstrated that the PANDER promoter was both tissue-specific and glucose-responsive. Surrounding the PANDER transcriptional start site are several putative A- and E-Box elements that may bind to the various pancreatic transcriptional factors of MafA, BETA2/NeuroD, and Pancreatic Duodenal Homeobox-1 (PDX-1). To characterize the transcriptional regulatory factors involved in PANDER gene expression, we performed co-transfection reporter gene analysis and demonstrated upregulation by all three transcription factors, with the greatest individual increase stemming from PDX-1. Potential binding of PDX-1 to A box (TAAT) regions of the PANDER promoter was demonstrated by chromatin immunoprecipitation (ChIP) and further corroborated by electrophoretic mobility shift assay (EMSA). Binding of PDX-1 to the A box regions was inhibited by mutagenized (TAGT) oligonucleotides. Site-directed mutagenesis of the three PDX-1 A box binding motifs revealed that A box sites 2 and 3 in combination were critical for maximal gene expression and deletion resulted in a 82% reduction in promoter activity. Furthermore, deletion of A box sites 2 and 3 completely diminished the glucose-responsiveness of the PANDER promoter. Our findings demonstrate that PANDER is a potential PDX-1 target gene and the A box sites within the promoter region are critical for basal and glucose-stimulated PANDER expression.

Figure 1. Effects of β-cell specific transcription factors on PANDER expression

Plasmids encoding the transcription factors MafA, PDX-1, BETA2/NeuroD, as well as one condition using all three, were co-transfected with the PANDER and phRL-TK luciferase reporter plasmids into NIH-3T3 cells. Twenty-four hours post-transfection, luciferase activity was measured and normalized to luciferase expression of the PANDER-luciferase plasmid co-transfected with pcDNA3.1 (empty vector). Data presented is from three experiments with each construct tested in triplicate, as % above ± S.E.M PANDER-luciferase plasmid co-transfected with pcDNA3.1. The asterisk indicates a P < 0.05 as determined by paired t-test with respect to luciferase activity from the PANDER-luciferase plasmid plus pcDNA3.1 empty vector. NS denotes not significant (P > 0.05) between PDX-1 and All (MafA, PDX-1, and BETA2/NeuroD co-transfected together) luciferase activity as determined by One-Way ANOVA.

Figure 2. Effects of site-directed mutagenesis of PDX-1 binding motif on PANDER promoter activity

(A) Structure of the mutagenized PANDER promoter luciferase constructs. A schematic diagram representing the PANDER promoter’s three A boxes, either wild type (i.e. black, 5′-TAAT-3′) or mutagenized (hatched bars, 5′-TAGT-3′). A plasmid encoding the −338 to +491 region of the PANDER gene, containing three A boxes (5′-TAAT-3v PDX-1 binding motifs) were subjected to site-directed mutagenesis to create seven mutant constructs, each containing a combination of one, two, or all three A boxes as shown above. Nucleotide locations are numbered and shown at the bottom of the diagram. TSS refers to transcriptional start site. (B) Luciferase activity of PANDER promoter luciferase constructs. Mutagenized PANDER promoter luciferase constructs (as diagramed in Figure 2A) were transiently co-transfected with Renilla luciferase reporter plasmid phRL-TK into β-HC9 cells in complete DMEM. Twenty-four hours following transfection, luciferase activity was detected by luminometry. WT denotes wild-type PANDER promoter luciferase construct that contains all 3 intact A box (TAAT) regions. Data are shown as mean ± S.E.M. percentage luciferase activity (firefly/renilla) relative to cells transfected with WT construct (open white bar). Luciferase activity above WT construct is shown with black filled bar whereas observed activity below WT construct is represented by back slashed bar. These data encompass two independent experiments with each construct transfected in triplicate. Single and double asterisk indicates P < 0.05 and 0.01 as determined by ANOVA, respectively, between mutagenized (shown according to construct number) and wild-type PANDER-luciferase plasmid (indicated by WT).

Figure 3. ChIP analysis of PDX-1 interaction with the PANDER promoter region

Cross-linked DNA isolated from murine insulinoma β-HC9 cells were subjected to ChIP analysis using either anti-PDX1 antibody, murine IgG antibody (negative control), or murine insulin antibody (positive control) followed by PCR amplification using primers specific for the PANDER promoter that either surround or encompass A box regions. (A) Schematic diagram of the PANDER promoter region indicating the approximate location of the A box sites (TAAT) and primers utilized for ChIP PCR amplification. A1, A2, and A3 refer to A-box sites. Arrows representing forward and reverse primers designate primer locations. (B) Representative PCR results following ChIP purification on a 2% ethidium bromide stained agarose gel. Lane 1, marker; Lane 2, no input DNA; Lane 3, mouse IgG immunoprecipitation (IP); Lane 4, blank; Lane 5, insulin IP; Lanes 6, 7, 8, PDX1 IP with PANDER specific amplification at various regions.

Figure 4. Binding of PDX-1 to the PANDER promoter as demonstrated by EMSA

Chemiluminescence of mobility-shifted DNA/protein complexes. 3′-Biotinylated oligonucleotides (above, DNA) imitating either the wild type (5′-TAAT-3′) A boxes (WT) and flanking regions within the PANDER promoter, or the site-directed mutagenesis-derived (5′-TAAT-3′) A boxes (MUT), were incubated with rat PDX-1 protein (PDX-1). The reaction mixtures were then electrophoresed in a non-denaturing gel and transferred and cross-linked to a nylon membrane, incubated in Streptavidin-Horseradish Peroxidase Conjugate (Pierce), and the resulting chemiluminescence was detected photographically. Evaluation of A boxes 1 (Lanes 1–3), 2 (Lanes 4–6), and 3 (Lanes 7–9) are indicated above the membrane. Lanes 1, 4, and 7 contain biotinylated WT oligonucleotides only; Lanes 2, 5, and 8 contain biotinylated WT oligonucleotides with rat PDX-1 protein (PDX-1), and Lanes 3, 6, and 9 contain MUT oligonucleotides with rat PDX-1 protein. Arrow indicates shifted oligonucleotide/protein complex.

Figure 5. Mutagenesis of A boxes impairs glucose-responsiveness of the PANDER promoter

(A) Mutagenized PANDER/luciferase plasmids are shown on the left side of the above figure and denoted by construct number as described previously (Figure 2A). A boxes are designated as A1, A2, and A3. Intact A boxes (5′-TAAT-3′) are represented by a black bar whereas mutagenized sites (5′-TAGT-3′) are shown by hatched bars. PANDER/luciferase constructs were transiently co-transfected with Renilla luciferase reporter plasmid phRL-TK into βTC3 cells in glucose-free DMEM. Four hours following transfection, the medium was replaced with DMEM containing incremental concentrations of D-glucose (i.e., 0 mM (white bar), 0.3 mM (vertical hatched bar), 5 mM (gray bar), 10 mM (black bar)). Twenty-four hours post-transfection, luciferase activity was detected by luminometry. Luciferase activity is shown as fold above respective construct at 0 mM glucose and normalized to renilla expression of the phRL-TK plasmid. All data are shown as mean ± S.E. from three independent experiments with each construct evaluated in triplicate. Asterik indicates a P value < 0.05 for the comparison of the PANDER/luciferase construct with that of the activity at 0 mM glucose according to ANOVA.