A. Materials
9. Reverse Transcriptase- Polymerase Chain Reaction (RT-PCR)
Total RNA was isolated using a RNAzolT M B and cDNA was synthesized using First-strand cDNA synthesis kit and 1 ㎍ of total RNA following the manufacture’s instructions. The PCR was conducted with 3 µl of the first-strand cDNA; 98 oC for 1 min, followed by 25 cycles for BETA2 and β-actin, and 30 cycles for SUR1 at 94 oC for 1 min, 55 oC for 30 sec, and 72 oC for 1 min, and finally 72 oC for 7 min. Primers were designed to recognize the separate exons to eliminate possible DNA contamination. The PCR primers for SUR1 were 5’-GCTCTTCATCACCTTCCCCATCCTC-3’ (forward) and 5’-CACAACCTGCG CTGGATCCTTACC-3’ (reverse); for BETA2 5’-CTCCGGGGTTATGAGAT CGTCAC-3’ (forward) and 5’-GATCTCTGACAGAGCCCA-3’ (reverse); for Pdx-1 5’-GGACACACAGCTCTACAAGGA-3’ (forward) and 5’-CATCACTGCCAG CTCCACCC-3’ (reverse); for β-actin 5’-CATGTTTGAGACCTTCAACACCCC-3’
(forward) and 5’-GCCATCTCCTGCTCGAAGTCTAG-3’ (reverse). The PCR products were analyzed on a 2 % agarose gel.
10. Nuclear extracts and immunoblotting
Nuclear extracts were prepared from transfected COS and 293T cells as
previously described by Attardi and Tjian.33 Briefly, cells were harvested 36 h after transfection and lysed in 25 mM Tris-Cl (pH 8.0), 2 mM MgCl2, 0.5 mM dithiothreitol (DTT), and 0.01% phenylmethylsulfonyl fluoride (PMSF) for 5 min at room temperatur e. Nonide P-40 was then added to a final concentration at 1700 ×g for 15 min, the resulting pellet was suspended in 10 mM Tris-Cl (pH 8.0), 400 mM LiCl, 0.5 mM DTT, and 0.01 % PMSF and kept on room temperature for 5 min.
After centrifugation at 12000 ×g for 2 min, the supernatant was harvested as nuclear extracts. Usually 30-60 ㎍ of cytosol and nuclear extracts were subjected to 12 % sodium dodecyl sulphate (SDS)-polyacrylamide gel electrophoresis. The proteins on the gel transferred electrophoretically onto the Immobilon-P Teflon menbrane and western analysis was performed with an specific antibodies. Finally, the membrane was incubated with horseradish peroxidase-conjugated anti- mouse IgG antibody and the proteins were visualized using an enhanced chemiluminescence (ECL) kit following the manufacture’s recommendation.
III. RESULTS
1. Transactivation of the SUR1 gene by BETA2/NeuroD
Previously, we defined that SUR1 is a target gene of BETA2/NeuroD. To investigate whether the endogenous BETA2/NeuroD confers high activity of pSUR-660CAT in HIT-T15 cells, we determined the effect of a dominant negative form of BETA/NeuroD, BETA2(1-233) (Fig. 4A). BETA2(1-233) functions as a dominant negative mutant because it contains the bHLH domain for heterodimerization with E47 but lacks transactivation domain. Coexpression of BETA2(1-233) reduced the promoter activity of the -660/-20 fragment in a dose dependent manner (Fig. 4B). Thus, transfection of 0.3 µg pCHA-BETA2(1-233) decreased the promoter activity of the -660/-20 fragment and the insulin promoter to about 20% and 35% of the reporter alone, respectively. Similarly an insulin reporter gene, pINSCAT448-, which contained the RIPE3 sequence was suppressed by BETA2(1-233). Under the same condition, BETA2(1-233) did not affect the
pSV2CAT. This result indicates that repression by BETA2(1-233) is specific to β-cell
specific genes and high promoter activity of –660/-20 of SUR promoter is due to BETA2/NeuroD-like factors in HIT-T15 cells.
BETA2
Fig. 4. Repression of the SUR1 promoter by a dominant negative mutant of BETA2/NeuroD.
A schematic diagram of full- length BETA2 and a truncated form BETA2(1-233) peptides were epitope tagged with hemagglutinin (HA) at the N-terminus.
bHLH, basic helix- loop-helix domain; AD, activation domain. B. Reporter genes were cotransfected with pcHA-233), an expression vector for BETA2(1-233). Although pSV2CAT lacking the E-box was not affected by BETA2(1-233), pINSCAT448- containing the E-box of the rat insulin II promoter was repressed by BETA2(1-233). Data are relative values with respect to the CAT activity of pSV2CAT and presented as averages ± standard errors of three independent experiments. *p<0.05, **p<0.01, both p values were estimated from t-test compared to the value of the reporter gene alone.
2. Determination of binding sites for BETA2/NeuroD
Four E-box-like sequences (E1- E4), which are potential sites for BETA2/NeuroD action, were found within the -660/-20 fragment. Since the -138/-20 fragment containing E4 was not synergistically activated by BETA2/NeuroD and E47, E4 alone is insufficient for tissue specific regulation. To determine which of the remaining E-box like sequences was necessary fo r BETA2/NeuroD binding, we carried out electrophoretic mobility shift assays (EMSA) using RIPE3 as a probe and the E-box sequences (E1-E3) of the SUR1 promoter as competitors (Fig. 5). As a source of BETA2/NeuroD, nuclear extracts were prepared from 293T cells transfected with expression vectors for BETA2(1-233) and E47. BETA2(1-233) has been shown to bind RIPE3 stronger than the full length BETA2 when expressed in 293T cells. Several complexes were detected and specific binding of the labeled RIPE3 probe was verified using an excess amount of unlabeled RIPE3 oligonucleotide (data not shown). Interestingly, only E3 was able to compete with RIPE3 for binding to BETA2/NeuroD (lane 6, 7). The binding was specific since the
same complexes disappeared by addition of an anti-NeuroD antibody (lane 8). This result indicates that the E3 box is necessary for binding of BETA2/NeuroD. To confirm the specificity of binding to E3, we carried out EMSA using E3 as a probe
Fig. 5. Determination of DNA binding ability of three putative E-boxes.
Nuclear extracts were prepared from 293T cells expressing a dominant negative form of BETA2/NeuroD, BETA2(1-233). EMSA was carried out with RIPE3, the E box of
consensus CANNTG
E1 GATCCACATCCAGCTGAGCCTAGA E2 GATCCCAGCGCACGTGCGCATTGA E3 GATCCAAGAGCAGCTGGAAGGGCA RIPE3a GATCTAGCCCCTCTGGCCATCTGCTGATCCG
the rat insulin II. Double stranded oligonucleotides containing E1, E2 or E3 of the SUR1 were used as competitors. Only E3 was able to compete with RIPE3 (solid triangle in lanes 6 and 7). The same complex disappeared in the presence of an anti-BETA2 antibody (lane 8).
E3 GATCCAAGAGCAGCTGGAAGGGCA E3m GATCCAAGAGTGGCTGGAAGGGCA
Fig. 6. Binding of BETA2/NeuroD to E3.
Nuclear extracts were prepared from 293T cells expressing a dominant negative form of BETA2/NeuroD, BETA2(1-233). A double stranded oligonucleotide containing E3 of the SUR1 promoter was used as a probe. Specific binding was confirmed by competition
with an excess amount of the unlabeled E3 (lanes 2 and 3) or E3m containing mutant in E3 (lanes 4 and 5) and RIPE3 (lanes 6 and 7). The specific complexes (solid triangle) disappeared by addition of an anti-BETA2 antibody (lane8).
(Fig. 6). Similar to RIPE3, BETA2/NeuroD bound to E3 specifically. E3 oligonucleotide (lane 2, 3) and RIPE3 (lane 6, 7) were able to compete with E3 of the SUR1 gene (closed triangle). In contrast, the E3m, a mutant oligonucleotide, did not compete with E3 (lane 4, 5). The specific complexes disappeared in the presence of an anti-NeuroD antibody (lane 8). This result indicates that BETA2/NeuroD can bind the E3 element at -141 bp of the SUR1 gene.
3. E3-mediated transactivation by BETA2/NeuroD
To determine whether the E3 box could confer transactivation of the SUR1 gene by BETA2/NeuroD, reporter genes, pSURE3(1+) and pSURE3m(1+), were constructed by ligating one copy of E3 or E3m to a heterologous promoter driving expression of luciferase in the pGL3-promoter vector, respectively (Fig. 7).
Coexpression of BETA2/NeuroD and E47 enhanced the luciferase expression by 3.8 fold for pSURE3(1+) while the pGL3-promoter was not affected at all. Mutation of E3 eliminated the stimulatory effect of BETA2/NeuroD (Fig. 8, left panel). To
confirm that E3 is the only functional E-box element in the -660/-20 fragment, we constructed pSUR-660E3m which contained a linker scanning mutation in E3 of the -660/-20 fragment. Mutation of E3 in the whole promoter context eliminated the
Fig. 7. Construction for E3 box and mutated E3 box of SUR1.
Reporter genes, pSURE3(1+), pSURE3m(1+), pSUR(-2432/-660), were constructed by ligating one copy of E3, E3m, or –2432/-660 to pGL3-promoter.
pSUR-660 and pSUR-660E3m were constructed using pGL2-basic (see the Materials
-660 -141 -20
and Methods for details).
Fig. 8. Mutation of E3 abolishes transactivation by BETA2/NeuroD.
Each of reporter genes (0.3 µg) was cotransfected with expression vectors for BETA2 and E47 into HIT-T15 cells. Note that mutation in E3 caused a loss of transactivation by BETA2/NeuroD and E47. Data are shown as averages ± standard
errors of three independent experiments with respect to the luciferase activity of pGL3 promoter or pGL2 basic. p values were estimated from t-test compared to the
BETA2/E47 - + - + - + - +
value of reporter genes alone (*p<0.05, **p<0.01).
stimulatory effect of BETA2/NeuroD (Fig. 8, right panel). Sequence analysis revealed nine more E-box- like sequences upstream of –660. To determine whether they might confer higher activity of pSUR-2432CAT than pSUR-660CAT in HIT-15 cells as shown in master’s thesis, we generated a reporter construct, pSUR-2432/-660 by ligating the -2432/-660 fragment to the pGL3-promoter. Coexpression of BETA2/NeuroD with or without E47 did not significantly enhance the -2432/-660 fragment (Fig. 8, left panel). Thus, E3 is essential for transactivation by BETA2/NeuroD.
4. Specificity of the E box-mediated transactivation by BETA2/NeuroD
In addition to BETA2/NeuroD, ngn3 is also a bHLH transcription factor expressed in pancreatic islet cells during their development. Thus, it is possible that ngn3 can also activate the SUR1 gene through the E3 element. We tested this possibility by coexpressing ngn3 and E47 in HIT-T15 cells. In HIT-T15 cells, the stimulatory effect of ngn3 was similar to that of BETA2/NeuroD (compare HIT-T15
cells in Fig. 9A and 9B). A reporter gene containing one copy of E3 increased 5.7 fold by coexpression of ngn3 and E47 (Fig. 10B). Similar result was obtained with a reporter gene containing the –660/-20 fragment. The stimulatory effect of ngn3 was
Fig. 9. Specificity of E3 for BETA2/NeuroD.
A, Reporter genes (0.3 µg) were cotransfected with expression vectors for
BETA2/NeuroD (0.1 µg) and E47 (0.05 µg) into HIT-T15 (hatched bar) or HeLa cells (filled bar). B, Reporter genes (0.3 µg) were cotransfected with expression vectors for ngn3 (0.1 µg) and E47 (0.05 µg) into HIT-T15 or HeLa cells. Data are shown as averages ± standard errors of three independent experiments with respect to
- + +
the luciferase activity of pSURE3(1+) or pSUR-660 in the absence of BETA2/NeuroD and E47. p values were estimated from t-test compared to the values of pSURE3 or pSUR-660Luc (*p<0.05, **p<0.01).
abrogated when E3 was mutated as shown with E3m (Fig. 9B) or –660E3m (data not shown). In contrast to HIT-T15 cells, transactivation by ngn3 was minimal compared to BETA2/NeuroD in HeLa cells (Fig. 9B). Coexpression of ngn3 and E47 could only activate the promoter activity of pSURE3 by 1.4 fold. In contrast, BETA2/NeuroD enhanced it by 3.1 fold under the same condition. Interestingly, both BETA2/NeuroD and ngn3 could not activate E3 in the homologous context of –660/-20 in HeLa cells. These results indicate that the E3 element is somehow specific for BETA2/NeuroD in HeLa cells. It has been shown that ngn3 is not detectable in mature β-cells and forced expression of ngn3 causes an increase in the level of BETA2/NeuroD in HIT-T15 cells.
Thus, it is possible that activation of E3 by ngn3 in HIT-T15 cells may be due to enhanced expression of BETA2/NeuroD, which is absent in HeLa cells. To test this possibility we investigated mRNA levels of SUR1 and BETA2/NeuroD before and after overexpression of ngn3. Stable cells were obtained by transfecting HIT-T15 and HeLa cells with the ngn3 expression vector. Total RNAs were prepared
from pools of G418-resistant cells and subjected to RT-PCR. Overexpression of ngn3 enhanced expression of both SUR1 and BETA2/NeuroD by 2.3 and 2.4 fold, respectively, in HIT-T15 cells (Fig. 10A and 10B). In contrast, ngn3 could not induce
Fig. 10. Specificity of E3 for BETA2/NeuroD.
RT-PCR products of SUR1 and BETA2/NeuroD with HIT and HeLa cells stably transformed with ngn3. A. BETA2/NeuroD and SUR1 products were detected as 620 bp and 117 bp fragments, respectively. Data present the most representative one from three independent experiments. B, The mRNA levels of SUR1 were
A
normalized to the β-actin mRNA and the RT-PCR data from three independent experiments were summarized as averages ± standard errors with respect to the value of untreated cells (*p<0.05, **p<0.01).
expression of SUR1 and BETA2/NeuroD in HeLa cells. This result suggests that ngn3 might activate the SUR1 promoter indirectly by inducing the expression of BETA2/NeuroD in HIT-T15 cells but not in HeLa cells.
5. Repression of BETA2/NeuroD promoter activity by AICAR in MIN cells.
AICAR is an adenosine analogue, which is taken up into cells and converted by adenosine kinase into a phosphorylated form, AICAR monophosphate (ZMP).
ZMP mimics the effects of AMP on both the allosteric activation and the phosphorylation of AMPK.34
To investigate the effects of glucose and AICAR on β-cells specific gene expression, islets were isolated from pancreas of Sprague-Dawley rats (200-250g male). The average number of islets isolated was ~300 per rat pancreas. The islets showed a regular spherical shape with will-defined smooth borders (Fig. 11). We carried out RT-PCR for the expression of BETA2/NeuroD, SUR1, and Pdx-1 to
investigate the effects of glucose and AICAR on mRNA levels of β-cell specific genes. Islets are incubated in high glucose condition, 30mM Glucose, at 6 h.
Increasing the glucose concentration from 5 to 30 mM significantly elevated the mRNA levels of BETA2/NeuroD in rat pancreatic islets (Fig. 12). Effect of
AICAR on SUR1 mRNA level showed similar results to BETA2/NeuroD (Fig. 12).
Interestingly, in islets treated with 400 µM AICAR for 6 h at 30 mM glucose, expression of BETA2/NeuroD and SUR1 decreased by 4.7 and 7.4 fold (Fig. 13). In contrast, AICAR did not affect the Pdx-1 level. The mRNA levels of each sample were normalized to the β-actin mRNA levels.
We also investigated mRNA levels of β-cell specific genes in MIN cells as a same condition with primary islets and similar results were obtained in MIN cells (Fig. 14). In MIN cells treated with 400 µM AICAR at 30 mM glucose, expression of
BETA2/NeuroD and SUR1 decreased by 3.1 and 2.2 fold, respectively. These results suggest that expression of BETA2/NeuroD and SUR1 were upregulated by a high level of glucose. Interestingly, the stimulatory effect of glucose was blocked in the presence of AICAR, but all genes of β-cells were not affected by AICAR.
To evaluate the effect of AICAR on BETA2/NeuroD gene transcription, we investigated BETA2/NeuroD promoter activity in MIN cells after transient
transfection of BETA2/NeuroD reporter gene with expression vectors, BETA2/NeuroD and ngn3, and treatment with 400 µM AICAR for 16 h before harvest. Coexpression of BETA2/NeuroD or ngn3 increased the promoter activity by
Fig. 11. Isolation and culture of rat pancreatic islets
Islets were isolated from pancreata of 200-250 g male Sprague-Dawley rats with collagenase digestion. Briefly, the common bile duct was cannulated and injected with 6 ml of cold M199 medium containing 1.5 mg/ml collagenase. The islets were collected and separated on Histopaque 1077 density gradient. The washed islets were picked individually under a dissecting microscope and cultured in RPMI medium with 11.1 mmol/l glucose, 10 % FBS, penicillin (100 U/ml), and
streptomycin (100 µg/ml) in a standard humidified culture condition of 5 % CO2 and 95 % air at 37 oC.
Fig. 12. Activation of BETA2/NeuroD and SUR1 by high glucose and repression by 400 µM AICAR, an AMPK activator.
Rat pancreatic islets were treated with 400 µM AICAR for 0-12h at 5mM or
30 mM glucose, and BETA2/NeuroD, SUR1, and Pdx-1 mRNA levels were examined by RT-PCR. BETA2/NeuroD, SUR1, and Pdx-1 products were detected as 390, 117, and 519 bp, respectively. The mRNA levels normalized to the β-actin
400 uM AICAR (h) 0 6 12 0 6 12 Low glucose High glucose
BETA2/NeuroD
SUR1
Pdx-1
β-Actin
mRNA. 12h after treatment with AICAR at 30 mM glucose, islets were suppressed BETA2/NeuroD and SUR1 mRNA level, but was without effect on Pdx-1 mRNA level. Data shown are the most representative of three independent experiments.
Fig. 13. Activation of BETA2/NeuroD and SUR1 by high glucose and repression by 400 mM AICAR, a AMPK activator.
The bar graphs represent quantitation of the results of RT-PCR in primary islets. The mRNA levels of BETA2 /NeuroD, SUR1, and Pdx-1 were normalized to the β-actin mRNA, and the RT-PCR data from three independent experiments were summarized as ± SE with respect of the value of untreated cells (*p<0.05, **p<0.01).
Fig. 14. Repression of the BETA2/NeuroD and SUR1 gene expression by glucose and AICAR.
MIN cells were treated with 400 µM AICAR for 12 h at 5 mM or 30 mM glucose, and RT-PCR were performed. BETA2/NeuroD and SUR1 products were detected as 390 bp and 117 bp fragments, respectively. The mRNA levels normalized
BETA2/NeuroD
SUR1
β-Actin AICAR - + - +
Glucose (mM) 5 5 30 30
Pdx-1
to the β-actin mRNA. BETA2/NeuroD and SUR1 were suppressed by 400 µM AICAR. Data shown are the most representative of three independent experiments.
Fig. 15. Repression of the BETA2/NeuroD and SUR1 gene expression by glucose and AICAR.
The bar graphs represent quantitation of the results of RT-PCR in MIN cells.
The mRNA levels of BETA2/NeuroD and SUR1were normalized to the β-actin mRNA, and the RT-PCR data from three independent experiments were summarized as ± SE with respect of the value of untreated cells.
1.7 and 2.3 fold 48 h after transfection. Importantly, when AICAR was treated, the promoter activity was decreased 4.2 and 5.7 fold, respectively (Fig. 16). MIN cells incubated at 30 mM glucose displayed greater BETA2/NeuroD promoter activity than cells maintained at 5 mM glucose (Data not shown). Addition of 400 µM AICAR completely repressed the promoter activity of the BETA2/NeuroD gene. As a positive control, we used a SUR1 reporter gene, pSURE3, which contained only activated E-box of SUR1 promoter region. In MIN cells at 30 mM glucose, expression of BETA2/NeuroD enhanced the promoter activity of the SURE3 reporter gene by 3.0 fold. After treated with 400 µM AICAR, the SUR1 promoter activity was decreased by 17.8 fold (Fig. 16). BETA2/NeuroD was originally isolated as a transactivator of the insulin gene. We also investigated a rat insulin II promoter activity in MIN cells after transient transfection of insulin reporter gene (RIP) and treatment with 200 and 400 µM AICAR for 16 h before harvest. In MIN cells treated
with 200 and 400 µM AICAR at 30 mM glucose, expression of insulin gene was
decreased by 2.4 fold (Fig. 17). These results suggest that expression of BET2/NeuroD- mediated genes, such as BETA2/NeuroD, SUR1, and insulin, were blocked in the presence of AICAR.
6. Binding of BETA2/NeuroD to E-box of the insulin promoter by AICAR
To verify the DNA binding ability of NeuroD by AICAR, we performed electrophoretic mobility shift assay using a double-stranded oligonucleotide containing the E-box of the rat insulin promoter element 3 (RIPE3). The RIPE3 site
contains the E-box, GCCATCTGC, which is conserved in all characterized mammalian insulin genes. Nuclear extracts were prepared from MIN cells incubated with or without AICAR for 12 h. Several complexes were detected, and specific binding of the labeled RIPE3 probes was verified using an excess amount of unlabeled RIPE3 oligonucleotides (Fig. 18, lane 2, 3). Interestingly, binding of NeuroD was decreased by AICAR (Fig. 18, lane 5). The specificity of the band for BETA2/NeuroD was shifted by BETA2/NeuroD antibody (Fig. 18, lane 6). Pdx-1 as well as BETA2/NeuroD is an important transcription factor for insulin gene. We also investigated DNA binding of Pdx-1 to A-box of the rat insulin promoter by AICAR.
The RIPA site contains the A-box, TAAT, which binds factors belonging to the homeodomain-containing protein family. Nuclear extracts were prepared from MIN cells incubated with or without AICAR for 12 h. Several complexes were detected,
Fig. 16. Repression of the BETA2/NeuroD and SUR1 promoter by 400 mM AICAR.
Reporter genes, pGL3-BETA2-2.2Luc and pGL3-SURE3Luc were cotransfected with pCMV-BETA2, an expression vector for BETA2/NeuroD and pCR3.1-ngn3, an expression vector for ngn3. To evaluate the effect of AICAR, MIN cells were treated with 400 M AICAR for 16h. Addition of AICAR completely repressed the promoter activity in MIN cells at high glucose.
Fig. 17. Repression of the rat insulin II promoter by AICAR.
Relative Luciferase Activity
2 0 4 0 6 0 8 0 100
AICAR (h) 0 200 400 0 200 400
Low Glucose High Glucose
Rat insulin II promoter (RIP), reporter gene, was transfected with AICAR in a dose-dependent for 16 h. Addition of AICAR completely repressed the promoter activity at high glucose.
and specific binding of the labeled RIPA probes was verified using an excess amount of unlabeled RIPA oligonucleotides (Fig. 19, lane 2, 3). DNA binding activity of Pdx-1 was not affected by AICAR (Fig. 19, lane 5). These complexes were supershifted when anti-BETA2 antibody was included in the incubation (Fig. 19, lane 6). These results suggest that the binding of BETA2/NeuroD to the E-box in the RIPE3 element was blocked by AICAR, whereas binding of Pdx-1 to the A-box in the RIPA element was not blocked by AICAR.
7. Expression of BETA2/NeuroD by AMPK activation.
We examined the changes of protein level of endogenous BETA2/NeuroD in MIN cells treated with 400 µM AICAR. Western analysis showed that the protein
We examined the changes of protein level of endogenous BETA2/NeuroD in MIN cells treated with 400 µM AICAR. Western analysis showed that the protein