BETA2 Activates Transcription From the Upstream Glucokinase Gene Promoter in Islet β-Cells and Gut Endocrine Cells

Rat βGK promoter sequences from −402 to +14 were cloned into the luciferase reporter vector pSVOAPL2L (13). Mutant 1 (mt1; −221 AAGGTG −216) and 2 (mt2; CATGTG) were generated in this construct using the Quick-Change mutagenesis kit (Stratagene); the mutated nucleotide is in italics, and the essential bHLH protein binding sequences are underlined. The minimal rat βGK E-box-driven construct contains three tandem copies of rat βGK −227 to −210 sequence cloned directly upstream of the TATA-box in the luciferase reporter vector pTATALuc (22). Each construct was verified by DNA sequencing (Baylor College of Medicine, DNA Sequencing Core Facility). Plasmid DNAs were purified using the Qiagen Maxi-Prep column (Qiagen). BETA2, E47, and upstream stimulatory factor (USF) were expressed in transfected cells from CMV-BETA2 (23), pCR3.1-E47 (22), psvUSF1 (24), and psvUSF2 (24), respectively.

Monolayer cultures of β (hamster HIT-T15 M2.2.2, mouse MIN6 and βTC-3), baby hamster kidney (BHK), and STC-1 enteroendocrine cells were grown in Dulbecco’s modified Eagle’s Medium supplemented with 10% fetal bovine serum (FBS), 50 μg/ml streptomycin, and 50 μg/ml penicillin. RIN β-cells were maintained in RPMI media with 10% FBS, 50 μg/ml streptomycin, and 50 μg/ml penicillin. MIN6 β-cells were maintained as described previously (25). Cells were plated at a density of 4 × 10⁴ cells per well in 12-well tissue culture dishes the day before transfection. Fugene reagent (5 μl; Roche Biochemicals) was used to transfect 500 ng of test DNA and 10 ng of pRL-TK DNA per well. Renilla luciferase (pRL-TK) was used as a recovery marker. Extracts were prepared 40 h after transfection and assayed for firefly and Renilla luciferase using the dual luciferase assay kit (Promega). The firefly luciferase activity from the test reporter construct was normalized to the Renilla luciferase activity from the cotransfected internal control plasmid. Each experiment was carried out at least three independent times.

Nuclear extracts were prepared from HIT T-15 and STC-1 cells using the methods described by Shelton et al. (10). Gel shift reactions (20 μl) were incubated at 24°C for 20 min in the presence of 5 μg nuclear extract, 120 mmol/l KCl, 1 μg poly(dI-dC), 100 ng sheared salmon sperm DNA, 12.5 mmol/l HEPES (pH 7.8), 2.5 mmol/l dithiothreitol, 0.1 mmol/l EDTA, 5% glycerol, and 50,000 cpms radiolabeled double-stranded oligonucleotide probe (600 cpm/nmol). The TNT-coupled reticulocyte lysate system (Promega) was used to in vitro transcribe/translate BETA2 (23) and E47 (22); 100 ng of poly(dI-dC) was used in place of salmon sperm DNA in these binding reactions. The NH₂ terminus-specific BETA2/NeuroD and E47 antisera (Santa Cruz Biotechnology) were preincubated with extract protein for 20 min before adding the probe. The DNA probe was end-labeled using [γ-³²P]ATP and T4 polynucleotide kinase. The completed binding assays were electrophoresed on a 6% acrylamide gel (from ICN) buffered with 250 mmol/l Tris (pH 7.6), 1.9 mmol/l glycine, and 10 mmol/l EDTA at 4°C (160 V for 3.5 h) before drying and autoradiography. The probe and competitor sequences were: rat βGK −231/−198, −231 GGGAACTGAGCAGGTGGTAATGTCTACCA −198; rat βGK mt1, −231 GGGAACTGAGAAGGTGGTAATGTCTACCA −198; rat βGK mt2, −231 GGGAACTGAGCATGTGGTAATGTCTACCA −198; rat βGK E-box-binding mutant (mtE), −231 GGGAACTGAGACTGGTGTAATGTCTACCA −198; rat insulin II, −104 CCCCTCTGGCCATCTGCTGATCC −86; U (USF consensus binding site), CACCCGGTCACGTGGCCTAGACC (Santa Cruz Biotechnology). The mutated nucleotides are in italics, and the bHLH protein binding sequences are underlined. The methylation interference analysis was performed with a rat βGK (−231/−198 bp) probe as described previously (10). The free and bound complex bands were excised, cleaved, and resolved on 20% acrylamide gels.

MIN6 cells (∼0.5–1.0 × 108) were exposed to 1% formaldehyde in Dulbecco’s modified Eagle’s medium for 5 min at 23°C, and then glycine was added to 125 mmol/l. Cells were washed with cold PBS, pelleted, and incubated in 0.6 ml of 1% SDS, 10 mmol/l EDTA, 50 mmol/l Tris-HCl (pH 8.1), and 1 mmol/l phenylmethylsulfonyl fluoride (PMSF) for 10 min on ice. The lysed samples were transferred to tubes containing 250-mg glass beads (≤10⁶ mm diameter) and treated at setting 4 with a Virsonic 100 sonicator (Virtis) for 12 10-s pulses at 4°C to fragment the chromatin. Then, 100-μl aliquots were incubated with 0.9 ml dilution buffer (0.01% SDS, 1.1% Triton X-100, 1.2 mmol/l EDTA, 16.7 mmol/l Tris-HCl, pH 8.1, 167 mmol/l NaCl, 1 mmol/l PMSF, and 0.1% protease inhibitor cocktail for mammalian cells; Sigma) and 60 μl of 50% protein G-Sepharose (blocked with 1 mg/ml BSA and 1 mg/ml sonicated salmon sperm DNA) for 1 h at 4°C. After removal of the Sepharose beads by centrifugation, 5 μg anti-BETA2 (NeuroD1) polyclonal IgG (Santa Cruz Biotechnology), 10 μg normal goat polyclonal IgG (Santa Cruz Biotechnology), or no antibody was added to the supernatant, and the reaction was incubated for 1 h at 4°C. Next, 60 μl of 50% protein G-Sepharose was added, and the reaction was incubated at 4°C for 3 h. The Sepharose beads were collected by centrifugation and washed extensively. The formaldehyde-induced cross-linking was reversed by incubating the beads overnight at 65°C with 0.5 ml elution buffer (1% SDS and 0.1 mol/l NaHCO₃) and 20 μl of 5 mol/l NaCl. The immunoprecipitated DNA was purified by standard methods and resuspended in 105 μl of nuclease-free water (Promega). PCR analysis was performed using Ready-to-Go PCR beads (Amersham Pharmacia Biotech) with 15 pmol of each primer and 10 μl of immunoprecipitated DNA per reaction. The cycling parameters were 1 cycle of 95°C (2 min) and 28 cycles of 95°C (30 s), 61°C (30 s), and 72°C (30 s). The primers used for amplification were: mouse βGK, −256 GTGATAGGCACCAAGGCACTGAC and −1 CGGTGCTTCTGTTCCAACCAGG; and phosphoenolpyruvate carboxykinase (PEPCK), 5′-GAGTGACACCTCACAGCTGTGG and 5′- GGCAGGCCTTTGGATCATAGCC. The correctness of the PCR products was confirmed by sequencing. Amplified products were electrophoresed through a 1.4% agarose gel in Tris-acetate-EDTA buffer and visualized by ethidium bromide staining.

The region spanning the UPE and Pal control sites between −280 to +14 bp of the upstream rat βGK promoter has been shown to be sufficient to drive β-cell-selective expression (10,13). To understand the role of the conserved E-box element at −221 to −216 bp in βGK expression, single-point mutants were constructed that should either disrupt (mt1) or have little effect (mt2) on the binding of the bHLH family of proteins that regulate E-box activity (Fig. 1B) (26). Their effect on activation was compared with the Pal-1 site mutant (−170/−161) and the wild-type −402/+14 luciferase construct in transfected βGK-expressing islet β-cell (HIT T-15, MIN6, βTC3, and RIN) and enteroendocrine (STC-1) cell lines. The expression pattern of the mutant constructs was essentially the same between lines. The results showed that the predicted dysfunctional −221 bp mutation within the bHLH factor contact site reduced βGK promoter-driven activity by 55, 45, 49, 48, and 35% in HIT T-15, MIN6, βTC3, RIN, and STC-1 cells, respectively, whereas the nonessential site mutation at −219 bp had little or no effect (Fig. 1C) in the tested cell lines. The loss in E-box activity was comparable to the Pal-1 site mutant in HIT-15 (Fig. 1C) and STC-1 (data not shown) cells. These results strongly suggested that bHLH factor(s) binding to the conserved E-box element at −221 to −216 bp was involved in βGK promoter activity within the islet and intestine.

The insulin E-box element is activated by a β-cell-enriched activator complex composed of BETA2 and E47 that contributes in cell-selective transcription (23,27). To characterize the factor(s) that bind to the βGK E-box element, gel mobility shift assays were performed with HIT T-15 nuclear extracts and probes to both the βGK and insulin E-box. Two major protein-DNA complexes (termed a and b) were shared between these elements (Fig. 2A). Competition assays performed with the unlabeled wild type and a consensus mtE indicated that each complex contained a bHLH binding factor(s) (Fig. 2A). The a complex was only detected in βGK-expressing HIT T-15 and STC-1 nuclear extracts, whereas the b complex was also found in nuclear extracts prepared from nonexpressing cells (data not shown).

The bHLH factor complexes that bind in gel shift assays to the insulin E-box contain the cell-enriched BETA2 transcription factor (i.e. as BETA2:E47) and the ubiquitously distributed USF (27,28). To determine whether these proteins were involved in binding to the βGK E-box, a super-shift assay was performed with antibodies against BETA2, E47, and the two isoforms of USF (i.e. USF1 and USF2). The addition of antibodies to BETA2 or E47 to the binding reactions specifically disrupted the a complex, and a higher super-shifted complex was detected in each lane (Fig. 2B). In contrast, the preimmune serum had no effect on complex formation. In addition, antibodies raised against USF1, but not USF2, selectively disrupted the b complex (compare U1 with U2 in Fig. 2C). An unlabeled USF E-box consensus binding site competitor also specifically eliminated b complex binding (Fig. 2C). These same complexes were detected in gel shift assays performed with MIN6 and βTC-3 nuclear extracts (data not shown). These results demonstrated that the two E-box element complexes contained BETA2:E47 (a complex) and USF-1 (b complex).

Methylation interference analysis was performed on the BETA2:E47 complex to identify the nucleotide contact sites within the βGK E-box element. The core CA and TG nucleotides were found to be the primary contact sites, although variant nucleotides were also contacted (Fig. 3A). A similar in vitro binding pattern for BETA2:E47 has also been found with the E-box motifs of rat insulin I (29), rat insulin II (27), pro-opiomelanocortin (POMC) (30), and secretin (31) genes. Interestingly, both the −221 and −219 bp were strong contact sites (Fig. 3B), although only the −221-bp mutation (i.e. mt1) within the invariant bHLH factor binding core affected E-box element activity in transfection (Fig. 1C) and competition assays (Fig. 3C).

To determine whether BETA2:E47 or USF1 could directly activate βGK E-box-mediated activation, we analyzed the effect of each upon expression of pβGKEboxTATALuc, a reporter plasmid that contains three copies of the βGK E-box inserted directly upstream of the promoter region in the pTATALuc reporter gene. pβGKEboxTATALuc was transfected into BHK cells either alone or in the presence of the BETA2, E47, USF1, or USF2 expression plasmids. βGK E-box-mediated expression was stimulated by E47 alone but more effectively by E47 and BETA2 (Fig. 4A). These factors did not affect the activity of the insert-less vector pTATALuc (Fig. 4B). Furthermore, activation was not observed with BETA2, USF1, or USF2 alone (Fig. 4). Because BETA2 and E47 only appear to be found together in cells (23), these results strongly suggest the βGK E-box activity is mediated by this cell-enriched activator complex.

The ChIP assay is a powerful tool for analyzing transcription factor binding in vivo (32,33). To directly address whether BETA2 binds to the upstream promoter of the endogenous βGK gene, formaldehyde cross-linked chromatin from βGK-expressing MIN6 β-cells was treated with BETA2 polyclonal IgG or goat polyclonal IgG. The DNA brought down by this treatment was PCR amplified with βGK and PEPCK promoter-specific primers. The antibody to BETA2 was capable of immunoprecipitating upstream βGK promoter sequences, whereas goat polyclonal IgG or the control without IgG could not (Fig. 5). However, the BETA2 IgG did not immunoprecipitate transcription control sequences from the PEPCK gene, which is not transcribed in β-cells. These results demonstrate that BETA2 occupies the βGK promoter region of the endogenous βGK gene in β-cells. The same conclusion was drawn in ChIP experiments performed in βTC-3 cells (data not shown). Together with the gel shift and transfection data described above, these results demonstrate that BETA2 regulates βGK promoter-mediated transcription.

βGK plays a critical role in adult islet β-cell function by catalyzing the rate-limiting step of glucose-induced insulin release (34). βGK mRNA transcription in the islet, gut, pituitary, and brain is controlled by sequences in the upstream promoter region (10,11,13). Experiments performed in transgenic mice have shown that tissue-specific expression is mediated by cis-acting elements found between −280 and +14 bp (3). Here, we show that a bHLH protein complex containing BETA2 and E47 regulates activation in islet β-cell and gut enteroendocrine cells from the conserved E-box element at −221/−216 bp.

Our initial strategy involved testing within a GK gene promoter construct spanning −402 to +14 bp, the effect of site-directed mutants within the potential E-box element at −221 to −216 bp. Under these circumstances, a mutant that blocked bHLH protein binding reduced activation by ∼50% in HIT T-15, βTC-3, MIN6, and RIN β-cell lines, whereas no effect was observed in a mutant that did not influence binding. In contrast, a previous study found that there was only an 18% decrease in activation in a 5′-flanking deletion mutant that removed this E-box region (10). Most likely, sequences within the vector minimized the control features of the GK E-box in this study.

Both BETA2:E47 and USF-1 were shown to bind specifically to the βGK E-box probe in gel shift assays (Fig. 2). To test their role in activation, the ability of BETA2, E47, USF-1, and USF-2 to stimulate βGK E-box-driven reporter expression was analyzed in BHK cells. E47 and BETA2 stimulated activity, in contrast to either of the USF proteins (Fig. 4). Because the USF proteins independently bind to their DNA regulatory target (35), their inability to activate indicates that they are not involved in βGK E-box control in vivo. Most significantly, BETA2 was shown to specifically bind to a region spanning the E-box within the endogenous βGK promoter, using the ChIP assay (Fig. 5). These data strongly suggest that BETA2:E47 specifically binds to and activates the upstream βGK promoter in the islet and gut. BETA2:E47 has also been shown to control insulin (23), glucagon (islet β-cells) (36), secretin (gut enteroendocrine cells) (31,37), and POMC (pituitary corticotropes) (38) gene expression in βGK-producing cell types.

In addition to its role as a transcriptional activator, BETA2 plays a significant role in the development of the pancreas, brain, and other gut endocrine cells. Targeted deletion of BETA2 in mice, for instance, results in compromised formation of islet cells (23), hippocampal and cerebellar neurons (39), and both secretin- and cholecystokinin-expressing enteroendocrine cells (37). Although βGK expression has not been fully analyzed in BETA2-null mice, the overlapping expression pattern of BETA2 with βGK-producing cells in affected tissues implies a potential role of BETA2 in controlling βGK gene transcription.

Because the E-box, Pal, and UPE elements appear to be maintained within the upstream transcription unit of all mammalian βGK genes (17), the conserved sequences within this region are likely to contain the binding sites for the other key factors required in transcriptional activation. Indeed, our experimental results suggest that selective transcription from the βGK promoter results from BETA2 cooperating with other factor(s) having a more cell-restricted expression pattern. It is likely that control is mediated through functional interactions with the PDX-1 homeodomain protein that binds to the three A+T-rich UPE elements of βGK (10,40). PDX-1 contributes in transcription of other β-cell-enriched genes, including insulin (15,41), islet amyloid polypeptide (42–44), and GLUT2 (16). In the insulin gene, activation by BETA2:E47 and PDX-1 is mediated by the p300/CBP (CREB binding protein) coactivator (45), a mechanism that also may be utilized in the βGK promoter.

Dysfunctional mutations in the coding region of GK are linked to a form of diabetes, termed maturity-onset diabetes of the young-type 2 (MODY-2) (46). Although GK was the first MODY locus discovered, the genetic basis for several other MODY subtypes were identified as transcription factors, including PDX-1 (47). Heterozygous mutations in PDX-1 in mice results in glucose intolerance (48). The inability to provide insulin in sufficient amounts to meet the body’s needs is also found in a form of type 2 diabetes caused by mutations in BETA2 (49). Our results indicate that conditions that effect the level or function of a factor involved in GK gene expression, like BETA2 and/or PDX-1, could also compromise β-cell function.

This work was supported by the Veterans Administration (Research Career Development and MERIT Review Award to J.M.M.) and the National Institutes of Health (NIH RO1 DK55091 to R.S.) and partially by the Vanderbilt University Diabetes Research and Training Center Molecular Biology Core Laboratory (Public Health Service grant P60 DK20593 from the National Institutes of Health) and an National Institutes of Health grant to M.-J.T.

We are grateful to Lisa Kelly and Dr. Hsiang-Po Huang for expert technical assistance and to Dr. Michele Sawadogo (University of Texas M.D. Anderson Cancer Center) for providing the USF expression vectors.