J Biol Chem, Vol. 274, Issue 30, 21095-21103, July 23, 1999

The Cyclic AMP Response Element Modulator Family Regulates the Insulin Gene Transcription by Interacting with Transcription Factor IID^*

Akari Inada, Yoshimichi Someya, Yuichiro Yamada, Yu Ihara, Akira Kubota, Nobuhiro Ban, Rie Watanabe, Kinsuke Tsuda, and Yutaka Seino^§

From the Department of Metabolism and Clinical Nutrition, Graduate School of Medicine, Kyoto University, Kyoto, Japan 606-8507 and the  Faculty of Integrated Human Studies, Kyoto University, Kyoto, Japan 606-8501

	ABSTRACT

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

We analyzed a mechanism of transcriptional regulation of the human insulin gene by cyclic AMP response element modulator (CREM) through four cyclic AMP response elements (CREs). We isolated two novel CREM isoforms (CREMQ1 and CREMQ2), which lack one of the glutamine-rich domains, Q1 and Q2 respectively, and six known isoforms (CREM, CREM, inducible cyclic AMP early repressor (ICER) I, ICER I, CREM-17X, and CREM-17) from rat pancreatic islets and the RINm5F pancreatic -cell line. CREM isoforms functioned as efficient transcriptional activators or repressors to modulate insulin promoter activity by binding to all of the insulin CREs. The binding activity of repressors is higher than that of activators and suppressed not only basal activity but also activator-induced activities. Furthermore, CREM activator interacted directly with the transcription factor IID components hTAF_II130 and TATA box-binding protein (TBP). These results suggest that the activation of the insulin gene transcription by CREM activator is mediated by not only direct binding to the CREs but also by recruiting transcription factor IID to the insulin promoter via its interaction with hTAF_II130 and TBP. On the other hand, the CREM repressor ICER competitively interrupts the binding of the activators to CREs and does not interact with either TBP or hTAF_II130; therefore, it might fail to stabilize the basal transcriptional machinery and repress transactivation.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Insulin gene transcription is regulated through cis-acting elements in response to stimulation, such as glucose concentration or increased cAMP level (1). One of the cis-acting elements, cAMP response element (CRE),¹ was first identified as an inducible enhancer of genes that can be transcribed in response to increased cAMP levels (2, 3).

In a previous study, we identified four functional CREs in the human insulin gene and demonstrated that each of these CREs has different sequences from the others and also from the consensus CRE motif (TGACGTCA) (4). Since a remarkable reduction of insulin gene transcription is observed when CREs are mutated (4), transcription factors that bind to this regulatory element might play an important role in the regulation of insulin gene transcription. However, which transcription factors are primarily involved in the regulation of insulin gene transcription through CREs is unclear.

Transcription factors such as CREB, CRE-BP1, and ATF1, which belong to the ATF/CREB family, bind to CRE (5-8). CRE modulator (CREM) is a member of this family, and the CREM gene generates both transcriptional activators (CREM, -, -1, and -2) and repressors (CREM, CREM, CREM, S-CREM, CREM-17X, and CREM-17) by alternative splicing (9-14). In addition, use of an intronic promoter (P2) generates inducible cAMP early repressors (ICERs) (15, 16). This property is unique to CREM within this family of transcription factors. CREM and CREM contain the phosphorylation domain (P-box) and two glutamine-rich domains (Q1 and Q2), which may be essential for gene activation and function as transcriptional activators. The CREM repressors, CREM, -, and -, contain the P-box but lack both Q1 and Q2. Thus, by a shuffling of exons, various combinations of functional domains are generated, and a large number of functionally different CREM isoforms are produced (12). These isoforms have been well studied in the hypothalamic-pituitary-gonadal axis, where they play key physiological and developmental roles by regulating gene transcription through CREs (17-22). Since the human insulin gene also possesses CREs, CREM could be a candidate for promoting insulin gene transcription.

Recent studies have demonstrated direct interactions of site-specific transcriptional factors and TFIID (23-25), a multiprotein complex consisting of the TATA box-binding protein (TBP) and at least eight TBP-associated factors (TAFs) (26-28), and revealed that it enhances the rate of promoter binding and stabilization of TFIID-promoter complexes (29).

Here we have analyzed the mechanism of transcriptional regulation of the insulin gene by CREM. We identified two novel CREM isoforms, CREMQ1 and CREMQ2, and six known isoforms in pancreatic islets and the pancreatic -cell line. We demonstrated that these isoforms regulated insulin gene transcription by binding to CREs and that CREM and its phosphorylation were involved in glucose-induced insulin promoter activity. We showed the direct interaction between CREM isoforms and the TFIID component, hTAF_II130 or TBP, suggesting a mechanism by which CREM isoforms might regulate insulin gene transcription.

	EXPERIMENTAL PROCEDURES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Immunohistochemistry-- Wistar rat pancreas embedded in paraffin was cut into serial sections at 3.5 µm. For immunocytochemistry, the avidin-biotin complex method with alkaline-phosphatase was used as described previously (30) with a slight modification. Briefly, normal goat serum (diluted to 1:75; DAKO, Kyoto, Japan) for the inhibition of nonspecific binding of secondary antibody, the anti-CREM polyclonal antibody (diluted to 1:500; Upstate Biotechnology Inc., Lake Placid, NY), or the anti-insulin polyclonal antibody (diluted to 1:500; DAKO), biotin-labeled goat anti-rabbit IgG serum (diluted to 1:300; DAKO), and avidin-biotin-alkaline phosphatase complex (diluted to 1:100; DAKO) were sequentially applied. Staining was visualized by alkaline phosphatase substrate (red) (Vector Laboratories, Burlingame, CA). A section of rat testis was used as a positive control for CREM. Anti-CREM polyclonal antibody was raised against recombinant mouse CREM expressed in Escherichia coli.

Isolation of Pancreatic Islets and RNA-- For each experiment, about 5000 islets were isolated from the pancreases of 12 Wistar rats by collagenase digestion (31), followed by purification on Ficoll gradients. Total RNA was isolated from freshly isolated islets by the guanidium thiocyanate-cesium chloride method (32) and were used as a template for cDNA synthesis (Superscript reverse transcriptase (Life Technologies, Inc.).

Reverse Transcriptase-Polymerase Chain Reaction Analysis and Plasmid Constructs-- Rat pancreatic islet cDNA and the pancreatic -cell line RINm5F (33) cDNA were amplified by the polymerase chain reaction using rat CREM-specific oligonucleotides. To exclude any amplification product derived from genomic DNA that could contaminate the RNA preparation, total RNA without reverse transcription was amplified as a negative control. The sequences of the 5' primers for the full-length of CREM and ICER were as follows: primer a, 5'-CCGTATGACCATGGAAACAG-3' (nt 5-24, Ref. 34); primer b, 5'-ATGGCTGTAACTGGAGATGA-3' (nt 1-20, Ref. 15), respectively. The sequence of the 3' primer (primer c) was 5'-CAGGTCCAAGTCAAACACAG-3' (nt 1054-1035). Exon-specific primers were set to determine the expression of each isoform. The sequences of the isoform-specific 5' primers for CREM were as follows: primer d, 5'-TCTAGCTCAGGTTTCTGTAG-3' (nt 113-132); primer e, 5'-TCTAGCTCAGGTAGCAACAA-3' (nt 113-122, nt 270-279). Sequences of the 3' primers were as follows: primer f, 5'-GTCTCCTCATCTTGAACAAC-3' (nt 706-687); primer g, 5'-GTCTCCTCATTGTATTGCCC-3' (nt 706-697, nt 507-498). Primers a, b, c, d, e, f, and g correspond to Fig. 1B. Each DNA fragment was subcloned into the SmaI site of pBluescript II SK(+), and the sequences were confirmed by the dideoxynucleotide chain termination procedure. The full-length sequences of the CREM isoforms, CREM, CREMQ1, CREMQ2, CREM, ICER I, and ICER I, were subcloned into the cytomegalovirus promoter-driven expression plasmid. Mutation of CREM serine 117 to alanine, S117ACREM, was constructed by oligonucleotide-directed site mutagenesis using an in vitro mutagenesis system (Promega, Madison, WI) and the following nucleotide primer: 5'-TCACGAAGACCCGCATATAGAAAA-3' (boldface denotes the mutation). The sequences of the activation domains were as follows: primer h, 5'-GGTAGCAACAATTGCAGAGA-3' (nt 270-289), primer i, 5'-GATTGCTATAGCCCAAGGTGG-3' (nt 508-527); primer j, 5'- TTAGTATTGCCCCGTGCTAGTC-3' (nt 507-488); primer k, 5'-TTATTGAACAACAACCTGGCT-3' (nt 696-677). Primers h, i, j, and k correspond to Fig. 3. DNA fragments of P, Q2, PQ2, and S117APQ2 (mutation of CREM serine 117 to alanine) were subcloned in frame to the 3'-end of the GAL4 DNA-binding domain in plasmid pCG4. The control pCG4 was generated by introducing the coding sequence for the GAL4 DNA-binding domain (amino acids 1-147) from pGBT9 (CLONTECH, Palo Alto, CA), downstream of the cytomegalovirus promoter in plasmid pCMV6c.

Cell Culture and Transfection-- Plasmids for the 5xGAL4-TATA-luciferase reporter gene were kindly provided by Dr. R. A. Maurer (Oregon Health Sciences University, Portland, OR). HIT-T15 cells (hamster pancreatic -cell line; Ref. 35) were cultured in RPMI 1640 medium supplemented with 10% fetal calf serum, 100 units/ml penicillin, and 100 µg/ml streptomycin in an atmosphere of 5% CO₂ at 37 °C. Transient transfections in HIT-T15 cells were performed by the calcium phosphate precipitation technique. Briefly, the media were switched to Dulbecco's modified Eagle's medium containing 10% fetal calf serum 4 h before transfection. The cells were transfected with a mixture of 12 µg of the luciferase reporter plasmid, pHI-luc, which contains human insulin gene promoter (nt 339 to +112; Ref. 4) or 5xGAL4-TATA-luciferase reporter plasmid, which contains 5xGAL4 DNA-binding sites, 0-4 µg of each full-length CREM expression plasmid or part-length CREM expression plasmid in which GAL4 DNA-binding domain fused, and 4 µg of internal control plasmid p-act--gal (36). For CREB and CREB expression plasmid, human CREB and CREB cDNA were amplified from human liver cDNA and were subcloned into the cytomegalovirus promoter-driven expression plasmid. The sequence of the 5' primer was 5'-ATGACCATGGAATCTGGAGC-3' (nt 144-163; Ref. 37), and the sequence of the 3' primer was 5'-TTAATCTGATTTGTGGCAGT-3' (nt 1169-1150). Expression plasmids of CREB, CREB, and CRE-BP1 (7) were co-transfected. To analyze promoter activity stimulated by glucose, cells were treated with various concentration of glucose. Forty-eight hours after transfection, the cells were harvested, and cell extracts were prepared for luciferase assays and -galactosidase assays. -Galactosidase assays were performed for internal control. All transfection experiments were repeated more than three times.

Expression of Recombinant CREM Isoforms and Electrophoretic Mobility Shift Assays-- Electrophoretic mobility shift assays were carried out to measure binding activity of CREM, CREMQ2, and ICER I to the human insulin CREs with double-stranded oligonucleotide probes. For production of glutathione S-transferase (GST) fusion proteins, the DNA fragments of CREM, CREMQ2, and ICER I were subcloned into pGEX4T-3 (Amersham Pharmacia Biotech) and transformed into E. coli strains M15. Following induction of protein expression with isopropyl thio--D-galactoside, the GST fusion proteins were purified according to the manufacturer's instructions. To check the size of products, the proteins were visualized by Coomassie Blue staining following fractionation by SDS-PAGE. Cy5-labeled DNA probes and 0-1.5 µg/µl proteins were incubated for 30 min at 25 °C. Electrophoreses were performed at 15 °C, at 38 milliamps constant current, and the binding activity was analyzed using an ALF DNA sequencer. Sense strands are shown (Table I): CRE1 (nt 220 to 191), 5'-TAAGACTCTAATGACCCGCTGGTCCTGAGG-3'; CRE2 (nt 188 to 164), 5'-GAGGTGCTGACGACAAGGAGATCT-3'; CRE3 (nt +13 to +32), 5'-CAGGCTGCATCAGAAGAGGC-3'; and CRE4 (nt +57 to +72), 5'-GGCCTTTGCGTCAGGTGGGC-3'.

Western Blot Assays-- HIT-T15 cells were harvested and were suspended in radioimmune precipitation buffer (50 mM Tris-HCl (pH 7.4), 150 mM NaCl, 1% Triton X-100, 1% sodium deoxycholate, 0.1% SDS, 22 mM EDTA, 1% Trasilol). Cell lysis was sonicated in radioimmune precipitation buffer containing 0.5 mM phenylmethylsulfonyl fluoride. After removal of insoluble debris by centrifugation, supernatants were loaded on 12.5% SDS-PAGE and transferred to nitrocellulose membranes (Millipore Corp., Bedford, MA). Membranes were blocked with 5% skim milk in phosphate-buffered saline and then incubated with anti TBP-antibody (diluted to 1:250; Santa Cruz Biotechnology, Inc., Santa Cruz, CA) for 2 h at 37 °C. Membranes were incubated with horseradish peroxidase-linked anti-rabbit IgG antibody (Amersham Pharmacia Biotech) for 30 min at 37 °C. Each of these steps was followed by three washes for 10 min in 5% Tween 20/phosphate-buffered saline. Detection was performed as described in the ECL protocol (Amersham Pharmacia Biotech).

In Vitro Transcription and Translation-- The plasmid encoding human TAF_II130 (hTAF_II130) cDNA was a generous gift of Dr. N. Tanese (New York University). Human TBP (hTBP) cDNA was amplified from human liver cDNA. The sequence of the 5' primer was 5'-ATGGATCAGAACAACAGCCT-3' (nt 242-261; Ref. 38), and the sequence of the 3' primer was 5'-TTACGTCGTCTTCCTGAATC-3' (nt 1261-1242). hTBP and hTAF_II130 cDNA were subcloned into the SmaI site of pGEM 3Z and the EcoRI site of pBluescript II SK(+), respectively. hTBP and hTAF_II130 proteins were synthesized in a coupled in vitro transcription-translation system in the presence of [³⁵S]methionine (Promega). The radioactive products were visualized by separation on a 12.5% SDS-PAGE and autoradiography.

In Vitro Protein Interaction Assays-- GST pull-down assays were performed as described previously (39, 40). Each assay contained GST fusion protein immobilized on glutathione-Sepharose beads (Amersham Pharmacia Biotech). Briefly, HIT-T15 cell extract or [³⁵S]methionine-labeled in vitro translated hTBP or hTAF_II130 was incubated with GST, GST-CREM, GST-CREMQ2, GST-CREM, GST-ICER I in binding buffer (20 mM Tris-HCl (pH 7.6), 0.1 mM EDTA, 20% glycerol, 300 mM KCl, 0.1% Nonidet P-40, 0.5 mM phenylmethylsulfonyl fluoride, 2 mM MgCl₂, 5 mM 2-mercaptoethanol) for 1 h at 4 °C. After incubation, the beads were washed five times with 1 ml of binding buffer. The bound proteins were resolved by 12.5% SDS-PAGE and visualized by autoradiography or by an anti-TBP antibody as described above.

	RESULTS

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

Expression of the CREM Isoforms in Rat Pancreatic Islets-- To determine if CREM is expressed in pancreatic islets, immunohistochemical studies of 8-week-old rat pancreatic islet sections were performed using anti-CREM antibody. To ascertain the specificity of the CREM antibody, a testis section was used as a positive control (data are not shown). In agreement with previous data (9, 18, 41), the expression of CREM protein was found to be restricted to round spermatides, mostly at stage VII-VIII of spermatogenesis, when the spermatozoa begin to be released into the lumen. We used two continuous rat pancreatic islet sections, one for the anti-insulin antibody (Fig. 1A, left panel) and the other for the anti-CREM antibody (Fig. 1A, right panel). Pancreatic  cells were stained red with the anti-insulin antibody. The CREM proteins were detected both in the nuclei and cytoplasm of the pancreatic islets, mainly in the pancreatic  cells. To identify which isoforms were present in the pancreatic islet cells and the pancreatic -cell line, we performed reverse transcriptase-polymerase chain reaction with rat pancreatic islet cDNA and the RINm5F -cell line cDNA using primers a-g (Fig. 1B). We isolated eight CREM isoforms, of which two are novel isoforms and the other six are known isoforms. A schematic presentation of two novel isoforms (CREMQ1 and CREMQ2), five repressors (CREM, ICER I, ICER I, CREM-17X, and CREM-17), and an activator (CREM) is shown in Fig. 1B. CREMQ1 and CREMQ2 lack one of the glutamine-rich domains (Q1 and Q2, respectively) and contain an exon Ia (DNA binding domain (DBD) I). These structures are similar to those of CREM2 and -1 (9, 12), which also lack one of the glutamine-rich domains (Q1 and Q2, respectively), and contain an exon Ib (DBD II).

View larger version (76K): [in this window] [in a new window]   Fig. 1.   Expression of the CREM isoforms in rat pancreatic islets. A, immunohistochemical localization of CREM protein in pancreatic islets of 8-week-old rats using anti-CREM antibody (dilted to 1:500). Immunocytochemical staining was performed by the avidin-biotin complex method. To ascertain the specificity of the CREM antibody, a testis section was used as a positive control. In agreement with previous data (9, 18, 41), the expression of CREM protein was found to be restricted to round spermatides (data are not shown). Two continuous rat pancreatic islet sections, one for the anti-insulin antibody (Fig. 1A, left panel) and the other for the anti-CREM antibody (Fig. 1A, right panel). Left panel, pancreatic  cells were stained red with the anti-insulin antibody (magnification, × 400). Right panel, pancreatic islets were stained red with the anti-CREM antibody. The CREM proteins were detected both in nuclei and cytoplasm of the pancreatic islets, mainly in the pancreatic  cells (magnification, × 400). B, schematic presentation of the structure of the CREM isoforms is shown. Functional domains are two glutamine-rich domains (Q1, Q2), the phosphorylation domain (P-BOX), and the DNA-binding domain (DBD I). Two novel isoforms, CREMQ1 and CREMQ2, lacked Q1 and Q2, respectively, and contained DBD I. Other isoforms are the known activator (CREM) and the repressors (CREM, ICER I, ICER I, CREM-17X, and CREM-17) (9-14). ICER is transcribed from an alternative intronic promoter. The arrows, a-g, represent the positions of synthetic oligonucleotide primers used to detect isoforms.

Effect of CREM Isoforms on Insulin Promoter Activity-- Previously, we have reported that CREM-17X and CREM-17 functioned as strong repressors (13). To analyze the function of the other CREM isoforms in insulin gene transcription, the isolated isoforms were subcloned into the cytomegalovirus promoter-driven expression plasmid vector and were transfected into HIT-T15 cells with the luciferase reporter plasmid containing the human insulin promoter (pHI-luc). CREM, CREMQ1, and CREMQ2 activated insulin promoter 9-, 8-, and 2-fold, respectively (Fig. 2A). On the other hand, CREM, ICER I, and ICER I repressed the activity to about 30, 25, and 25% of controls, respectively. We then tested the abilities of the repressors, CREM and ICER, to repress not only basal transcriptional activity but also activator-induced transcriptional activity. CREM activator induced transcriptional activity, whereas CREM and ICER I blocked this effect to about 30-50% of controls (Fig. 2, B and C). Moreover, CREM and ICER I also repressed CREB-, CREB-, and CRE-BP1-induced transcriptional activity to about 30% of controls.

View larger version (26K): [in this window] [in a new window]   Fig. 2.   Effect of the CREM isoforms on insulin gene transcription. A, to determine if the CREM isoforms stimulate the insulin promoter, the isolated isoforms were subcloned into the cytomegalovirus promoter-driven expression plasmid and transfected into HIT-T15 cells with the luciferase reporter plasmid containing the human insulin gene promoter (pHI-luc). B and C, to determine if CREM and ICER I repress not only basal insulin promoter activity but also activator-induced insulin promoter activity, the same amount of CREM or ICER I expression plasmid was co-expressed with the expression plasmids of CREM, CREMQ1, CREMQ2, CREB, CREB, and CRE-BP1. Closed bars represent co-transfection with CREM (B) or ICER I (C). Transfections were repeated more than three times; bars represent the mean ± S.E. *, p < 0.05; **, p < 0.01.

Transcriptional Activity of the Activation Domain-- To investigate the function of the phosphorylation (P) and glutamine-rich (Q) domains in the transcriptional activity of CREM, the P domain alone, the Q2 domain alone, and the combination of P and Q2 domains (PQ2) were fused in frame to the 3'-end of the GAL4 DNA-binding domain in plasmid pCG4. To further examine if the phosphorylation is involved in CREM activation, we mutated the serine residue to alanine at position 117 in the P domain (S117APQ2), which is the target for the phosphorylation (9, 42, 43). Schematic presentation of GAL4-fused expression plasmids is shown (Fig. 3, left). HIT-T15 cells were transfected with the GAL4-fused expression plasmid, and the luciferase reporter plasmid containing 5xGAL4-binding sites (5xGAL4-TATA-luciferase reporter plasmid). Luciferase activity increased about 6- and 13-fold in the presence of Q2 domain alone or PQ2 domains, respectively, but showed no change in the absence of Q2 domain (Fig. 3, right). The Ser to Ala mutation substantially decreased PQ2 domain activity.

View larger version (32K): [in this window] [in a new window]   Fig. 3.   Transcriptional activity of the activation domain. To investigate the function of the P domain and Q2 domain, P domain alone, Q2 domain alone, and the combination of P and Q2 domain (PQ2) were fused in-frame to the 3'-end of the GAL4 DNA-binding domain (black box) in plasmid pCG4. To further examine if the phosphorylation is involved, Ser117 was mutated by Ala (S117APQ2). The arrows, h-k, represent the positions of the synthetic oligonucleotide primers used. Exons E and F correspond to the phosphorylation (P) domain, and exon G corresponds to the glutamine-rich (Q) domain, Q2. The asterisk in exon E represents a serine residue at position 117 (Ser117), which is the target for the phosphorylation. HIT-T15 cells were transfected with the GAL4-fused expression plasmid and the luciferase reporter plasmid containing the 5xGAL4-binding sites (5xGAL4-TATA-luciferase reporter plasmid). C, control. Transfections were repeated more than three times; bars represent the mean ± S.E. *, p < 0.05; **, p < 0.01.

Effect of Glucose Stimulation on the Insulin Gene Transcription-- Glucose stimulation has been reported to increase insulin mRNA (44, 45). We investigated whether CREM and CREB are involved in glucose-induced insulin promoter activity. To determine the involvement of the phosphorylation in CREM, Ser¹¹⁷ in the P domain of CREM was mutated by Ala (S117ACREM). CREM, CREB, and S117ACREM were transfected into HIT-T15 cells with pHI-luc, and the cells were treated with various concentrations of glucose. Interestingly, insulin promoter activity was increased remarkably by glucose stimulation in a dose-dependent manner in the presence of CREM (Fig. 4A). However, in the presence of S117ACREM and CREB, no significant increase in insulin promoter activity was observed by glucose stimulation, suggesting that CREM but not CREB is involved in glucose-induced insulin promoter and that the increase is dependent on Ser¹¹⁷. We next examined if glucose stimulation induces PQ2 domain activation; the PQ2 and S117APQ2 domains were transfected with 5xGAL4-TATA-luciferase reporter plasmid, and the cells were treated with various concentrations of glucose. The activity of the PQ2 domain increased in a dose-dependent manner, but the S117APQ2 activity did not (Fig. 4B). Again, this increase is dependent on Ser¹¹⁷, suggesting that glucose might induce the activation of CREM by phosphorylation.

View larger version (14K): [in this window] [in a new window]   Fig. 4.   Effect of glucose stimulation on insulin promoter. A, to determine if CREM and CREB are involved in glucose-induced insulin promoter activity, HIT-T15 cells were transfected with pHI-luc and the expression plasmid of CREM, S117ACREM (Ser117 was mutated by Ala) and CREB. Cells were treated with various concentrations of glucose. In the presence of CREM, a significant increase in a dose-dependent manner in insulin promoter activity was observed by glucose stimulation (p < 0.05 versus 0 mM glucose). But in the presence of S117ACREM and CREB, no significant increase in insulin promoter activity was observed by glucose, suggesting that CREM but not CREB is involved in glucose-induced insulin promoter and that the increase is dependent on Ser117, since mutating this residue to alanine leads to a strong decrease. B, to determine if glucose stimulation induces PQ2 domain activity, GAL4-fused PQ2 and S117APQ2 domain were transfected with 5xGAL4-TATA-luciferase reporter plasmid, and the cells were treated with various concentrations of glucose. In the presence of PQ2, a significant increase in luciferase activity was observed in a dose-dependent manner by glucose stimulation (p < 0.01 versus 0 mM glucose), but not in the presence of S117APQ2 domain. Transfections were repeated more than three times; bars represent the mean ± S.E. *, p < 0.05; **, p < 0.01.

Binding Activity of the CREM Isoforms to Insulin Gene CREs-- We have reported four CRE-like sequences in the human insulin gene that are somewhat different from each other and from the consensus CRE motif (4). To determine whether CREM isoforms bind to these CRE-like sequences of the human insulin gene, we performed electrophoretic mobility shift assays using bacterially produced CREM, CREMQ2, and ICER I proteins and Cy5-labeled human insulin gene CRE probes (Table I). The binding activity was analyzed using an ALF DNA sequencer. The binding of bacterially produced CREM, CREMQ2, and ICER I proteins to each CRE (CRE1 to 4) were detected as the peaks of fluorescence derived from Cy5-labeled probes (Fig. 5, A, B, and C). We next determined if ICER I competes with CREM or CREMQ2 for CRE3. As shown in Fig. 5D, ICER I competed with CREM in a dose-dependent manner to bind to CRE3 (lanes 1-4). Similar results were obtained using CREMQ2 (Fig. 5E). When the same amounts of ICER I and CREM were added (lane 5), CRE3 of the insulin gene bound to ICER I more than CREM. On the other hand, the binding activity of ICER I to CRE3 was the same as that of CREMQ2.

View larger version (31K): [in this window] [in a new window]   Fig. 5.   Binding activity of the CREM isoforms to human insulin gene CREs. To determine if the CREM isoforms bind to the CRE-like sequences of the human insulin gene, electrophoretic mobility shift assays were performed using Cy5-labeled human insulin gene CRE probes and bacterially produced CREM, CREMQ2, and ICER I proteins. The proteins were incubated with Cy5-labeled probes and electrophoresed. The binding activity was analyzed using an ALF DNA sequencer. The binding of CREM (A), CREMQ2 (B), or ICER I (C) to each CRE (CRE1 to 4) was detected as the peaks of fluorescence derived from Cy5-labeled probes. The amounts of proteins are noted as arbitrary units. When the relative amounts of ICER I were increased (lanes 1-4), ICER I competed with CREM (D) or CREMQ2 (E) in a dose-dependent manner to bind to CRE3. The same amounts of ICER I and CREM (D) or CREMQ2 (E) were added (lane 5).

Interaction of CREM with TFIID in Vitro-- To determine whether CREM might interact directly with the TFIID components TBP and hTAF_II130 we performed in vitro interaction assays. The GST-fused CREM isoforms GST-CREM, GST-CREMQ2, GST-CREM, and GST-ICER I were bacterially produced. [³⁵S]methionine-labeled hTBP and TAF_II130 were synthesized in a coupled in vitro transcription-translation system using rabbit reticulocyte lysate in the presence of [³⁵S]methionine. Immobilized GST-CREM isoforms were incubated with [³⁵S]methionine-labeled hTBP or hTAF_II130. Specifically bound protein was resolved by SDS-PAGE and visualized by autoradiography. As shown in Fig. 6A, hTBP bound to CREM, CREMQ2, and CREM (lanes 3-5). However, hTBP did not bind only to ICER I (lane 6). To confirm the interaction between TBP and the CREM isoforms, we prepared cell extracts from HIT-T15 cells and examined the interaction of intracellular TBP and bacterially produced CREM isoforms. Specifically bound proteins were subjected to Western blotting with a polyclonal TBP antibody (Fig. 6B). hTBP bound to CREM and CREMQ2 (lanes 3 and 4) but did not bind to ICER I (lane 5). We next investigated the interaction of hTAF_II130 and the CREM isoforms. hTAF_II130 bound only to CREM and CREMQ2 (lanes 3 and 4). The binding of hTAF_II130 to CREMQ2 was very weak. Moreover, hTAF_II130 did not bind to either CREM or ICER I (lanes 5 and 6) (Fig. 6C).

View larger version (28K): [in this window] [in a new window]   Fig. 6.   Interaction of CREM with TFIID detected in vitro. To determine if CREM interacts directly with the TFIID components TBP and TAFII130, in vitro interaction assays were performed. GST fusion CREM isoforms, GST-CREM, GST-CREMQ2, GST-CREM, and GST-ICER I, were bacterially produced. The plasmids containing the cDNA encoding human TBP (hTBP) and human TAFII130 (hTAFII130) were used to produce in vitro translated and [35S]methionine-labeled proteins. A and C, the GST and GST-CREM isoform beads were incubated with [35S]methionine-labeled proteins. After incubation, the beads were washed five times with the binding buffer. The bound protein was resolved by 12.5% SDS-PAGE and visualized by autoradiography. The arrows indicate the specific protein bands. The input samples contained one-sixth of the amounts used for incubation. A, hTBP bound to CREM, CREMQ2, and CREM (lanes 3-5) but not to ICER I (lane 6). The lack of binding to control GST demonstrates the specificity of the binding (lane 2). All of the isoforms that bound to hTBP commonly possess the phosphorylation domain, suggesting that CREM-hTBP interaction might be mediated by the phosphorylation domain. B, to confirm the interaction between TBP and CREM isoforms, cell extracts from HIT-T15 cells were prepared, and intracellular TBP and CREM isoforms were examined. Specifically bound proteins were subjected to Western blotting with a polyclonal anti-TBP antibody. hTBP bound to CREM and CREMQ2 (lanes 3 and 4) but not to ICER I (lane 5). C, the interaction of hTAFII130 and CREM isoforms was investigated. hTAFII130 bound to CREM and very weakly to CREMQ2 (lanes 3 and 4) but not to CREM or ICER I (lanes 5 and 6), suggesting that CREM-hTAFII130 interaction might be mediated by both of the glutamine-rich domains.

	DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

The insulin gene has multiple cis-acting elements, and its transcription is thought to be regulated by the interaction of various positive-acting and negative-acting factors (1). In addition, multiple factors such as general polymerase II transcription factors are involved in transcription by interacting with each other and forming the initiation complex (19-25). In the present study, the mechanism of transcriptional regulation of the human insulin gene by CREM through CREs was examined.

First, we demonstrated by reverse transcriptase-polymerase chain reaction and by immunohistochemistry that CREM is expressed in pancreatic -cells at the levels of mRNA and protein. We isolated two novel CREM isoforms (CREMQ1 and CREMQ2) and the six known isoforms (CREM, CREM, ICER I, ICER I, CREM-17X, and CREM-17) from rat pancreatic islets and the RINm5F pancreatic -cell line. It has been reported that all positively acting isoforms contain at least one of the glutamine (Q)-rich domains, Q1 and Q2 (12). Considering that CREMQ1 and CREMQ2 also contain one Q-rich domain and that their structures are very similar to that of the CREM1 and 2 activators (9), CREMQ1 and CREMQ2 were supposed to function as activators. As expected, CREMQ1 and CREMQ2 were found in the present study to function as activators of insulin promoter. In addition, we showed that CREM functions as an efficient activator, while CREM, ICER I, and ICER I act as strong repressors of insulin promoter. In the previous study, we have shown that CREM-17X and CREM-17 function as efficient repressors of insulin promoter (13).

We also found by immunohistochemistry that CREM proteins are located both in the nuclei and cytoplasm of pancreatic islets. A nine-amino acid sequence (RRKKKEYVK) defined as the nuclear translocation signal of CREB (46) is also conserved in the DBD of CREM isoforms (14). A previous study detected a translation product of the plasmid encoding CREB cDNA lacking the putative nuclear translocation signal in the cytoplasm when transfected in COS-1 cells (46), suggesting that the CREM isoforms containing nuclear translocation signals may occur in the nuclei and that the isoforms that lack the nuclear translocation signal occur in the cytoplasm.

In a previous study, we have identified four CRE-like sequences (CRE1 to 4) in the human insulin gene (4). These CREs are the characteristic sequences, because the consensus CRE sequence is TGACGTCA, but the sequences of CRE1 and CRE2 show the variation at the 3'-end, while CRE3 and CRE4 show it at the 5'-end. Therefore, the examination for the binding of CREM to these CRE-like sequences was required. In the current study, we demonstrated that the CREM activators, CREM and CREMQ2, bind to each CRE. This result suggested that CREM activator modulates insulin promoter activity by binding directly to CREs. In addition, we demonstrated that the CREM repressor, ICER I, also binds to each CRE and that the binding activity of repressor to insulin CRE is higher than that of activators, suggesting that CREM repressor suppresses insulin promoter activity by competitively interrupting the binding of the activators to CREs.

We compared the ability of the members of the CREB/ATF family to activate the insulin promoter. We found in the present study that CREB could not highly activate insulin promoter, compared with CREM. Therefore, it is possible that among the members of the CREB/ATF family, CREM plays a pivotal role in modulating insulin promoter activity in pancreatic  cells. We further compared the ability of CREM and CREB to activate insulin promoter in the presence of glucose. It is well known that glucose increases insulin promoter activity, but the mechanism is unclear (44, 45). Here we showed for the first time that CREM but not CREB is involved in glucose-induced insulin promoter activity. It is interesting that CREM and CREB show a difference in the ability to promote insulin promoter not only in the absence but also in the presence of glucose, despite their extensive homology. This result suggests that CREM plays a role different from that of CREB in promoting insulin promoter in pancreatic  cells. Since Ser¹¹⁷ of CREM has been shown to be the target for phosphorylation by protein kinase A, protein kinase C, calmodulin kinase, and p34^cdc2 and such phosphorylation increases CREM activity (9, 42, 43), we mutated Ser¹¹⁷ to analyze the involvement of this residue in glucose-induced insulin promoter activity. We found by mutation analysis that the phosphorylation of CREM at serine 117 is necessary for glucose-induced insulin promoter activity. Taken together, it is suggested that glucose stimulation might activate an intracellular signal transduction to enhance the phosphorylation of Ser¹¹⁷, which increases CREM trans-regulatory function to promote insulin promoter activity.

It has been reported that TFIID components, TAFs, interact with each other (39, 47) and that certain TAFs can function as co-activators by direct interaction with site-specific transcription factors (24, 26, 48, 49); therefore, it was of interest to determine if CREM directly interacts with TFIID. In this study, we found that CREM activators interact with both TBP and hTAF_II130. Previously, hTAF_II130 and its Drosophila homologues dTAF_II100 and dTAF_II110 have been shown to intreract with CREB or Sp1 directly through the glutamine-rich domain (Q2) of CREB or Sp1 (24, 25, 48-51). Considering that all CREM activators possess one or two glutamine-rich domains (Q1, Q2) (9), Q1 and Q2 seem to be required for CREM to interact with hTAF_II130 and to function as an activator; conversely, the interaction of Q1 and Q2 with hTAF_II130 may enable CREM to function as an activator. Our results suggested that CREM-hTAF_II130 interaction and CREM-TBP interaction might be mediated by the glutamine-rich domains (Q1, Q2) and by the phosphorylation domain of CREM, respectively. Since it has been reported that the interaction of TFIID with site-specific transcriptional factors enhances the rate of binding of TFIID to the promoter or stabilizes TFIID-promoter complexes (29), it is possible that an interaction of CREM with TBP and hTAF_II130 may enhance and stabilize the binding of TFIID to the insulin promoter. In other words, CREM activators could increase the levels of insulin gene transcription by both binding directly to the CREs and recruiting TFIID to the insulin promoter via its interaction with hTBP and hTAF_II130 (Fig. 7). On the other hand, there were no CREM repressors that interact with both hTAF_II130 and TBP, suggesting that CREM repressor might fail to stabilize TFIID-promoter complexes. Therefore, it is suggested that CREM repressor might suppress insulin promoter activity by competitively interrupting the binding of the activators to CREs and, further, by failing to stabilize the basal transcriptional machinery. Since the only interactions of CREM with hTBP and hTAF_II130 were examined in this study, further studies of the interactions between CREM and the individual components of the TFIID complex such as hTAF_II250 (52), hTAF_II100 (25, 39), and hTAF_II80 (47) will be required to reveal the molecular mechanisms of the the insulin gene transcription in more detail.

View larger version (25K): [in this window] [in a new window]   Fig. 7.   Activation and repression of the human insulin gene transcription by CREM and ICER. Left, CREM activates insulin promoter not only by direct binding to CREs but also by recruiting TFIID to the insulin promoter via its interaction with hTBP and hTAFII130. Right, ICER competitively interrupts the binding of the CREM activators to CREs and also does not interact with either TBP or hTAFII130, which might fail to stabilize the basal transcriptional machinery and thus repress trans-activation.

In conclusion, we found in this study that CREM activators and repressors are expressed in pancreatic  cells and that they play important roles in modulating the insulin gene transcription in response to glucose. An important point from our study is that the activation of insulin gene transcription by CREM activator is mediated by not only direct binding to the CREs but also by recruiting TFIID to the insulin promoter via its interaction with hTBP and hTAF_II130. Furthermore, the CREM repressor competitively interrupts the binding of the activators to CREs and does not interact with both TBP and hTAF_II130; thereby, it might fail to stabilize the basal transcriptional machinery and repress transactivation.

	ACKNOWLEDGEMENTS

We thank Dr. N. Tanese (New York University) for kindly providing human TAF_II130 plasmid and Dr. R. A. Maurer (Oregon Health Sciences University, Portland, OR) for 5xGAL4-TATA-luciferase reporter plasmid. We acknowledge Dr. S. Ishii (Tsukuba Life Science Center, RIKEN, Tsukuba, Japan) for critical reading of the manuscript and Dr. S. Toyokuni (Department of Pathology and Biology of Diseases, Kyoto University, Kyoto, Japan) for comments on immunohistochemistry.

	FOOTNOTES

^* This study was supported in part by Grants-in-Aid for Scientific Research 09470219 and for Creative Basic Research NP10NPO201 from the Ministry of Education, Science, Sports and Culture, Japan; by a grant for Diabetic Research from Tsumura and Co., Japan; and by grants for the "Research for the Future" Program from the Japan Society for the Promotion of Science (JSPS-RFTF97I00201).The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

^§ To whom correspondence should be addressed: Dept. of Metabolism and Clinical Nutrition, Graduate School of Medicine, Kyoto University, 54 Shogoin-Kawahara-cho, Sakyo-ku, Kyoto 606-8507, Japan. Tel.: 81-75-751-3560; Fax: 81-75-771-6601; E-mail: inada@metab.kuhp.kyoto-u.ac.jp.

	ABBREVIATIONS

The abbreviations used are: CRE, cyclic AMP response element; CREM, cyclic AMP response element modulator; CREB, cyclic AMP response element-binding protein; ICER, inducible cyclic AMP early repressor; nt, nucleotide(s); TBP, TATA box-binding protein; hTBP, human TBP; TAF, TBP-associated factor; hTAF, human TAF; GST, glutathione S-transferase; PAGE, polyacrylamide gel electrophoresis.

	REFERENCES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES