Chromatin-bound mitogen-activated protein kinases transmit dynamic signals in transcription complexes inβ-cells
- Michael C. Lawrence*,
- Kathleen McGlynn*,
- Chunli Shao*,
- Lingling Duan*,
- Bashoo Naziruddin†,
- Marlon F. Levy†, and
- Melanie H. Cobb*,‡
- *Department of Pharmacology, University of Texas Southwestern Medical Center at Dallas, Dallas, TX 75390; and
- †The cGMP Islet Cell Processing Laboratory, Islet Cell Transplant Program, Baylor University Medical Center, Dallas, TX 75246
-
Contributed by Melanie H. Cobb, July 8, 2008 (received for review June 13, 2008)
Abstract
MAPK pathways regulate transcription through phosphorylation of transcription factors and other DNA-binding proteins. In pancreaticβ-cells, ERK1/2 are required for transcription of the insulin gene and several other genes in response to glucose. We show that binding of glucose-sensitive transcription activators and repressors to the insulin gene promoter depends on ERK1/2 activity. We also find that glucose and NGF stimulate the binding of ERK1/2 to the insulin gene and other promoters. An ERK1/2 cascade module, including MEK1/2 and Rsk, are found in complexes bound to these promoters. These findings imply that MAPK-containing signaling complexes are positioned on sensitive promoters with their protein substrates to modulate transcription in situ in response to incoming signals.
The MAPKs ERK1/2 are required for insulin gene transcription induced by acute exposure of pancreaticβ-cells to increased glucose (1, 2). Pancreaticβ-cells produce, store, and release insulin in response to physiological demand by sensing changes in the circulating glucose concentration, normally maintained at ≈5–10 mM. Other nutrients, hormones, neurotransmitters, and paracrine factors have a modulatory role, enhancing or suppressing the effects of glucose. Glucose entry and metabolism increases the ATP/ADP ratio. Allosteric regulation by nucleotides causes closure of the ATP-dependent potassium channel, depolarization of the membrane, calcium influx through voltage-dependent channels, and release of calcium from intracellular stores (3). These changes lead to the activation of signaling pathways, which triggerβ-cells to release insulin to promote glucose removal from the circulation restoring a resting concentration close to 5 mM. Acute replenishment of insulin stores occurs by translation of preexisting insulin mRNA and proinsulin processing. The long-term maintenance of insulin production results from glucose- and nutrient-induced insulin gene transcription (4).
Glucose-induced calcium influx initiates ERK1/2 activation in a manner that depends on the calcium, calmodulin-dependent phosphatase calcineurin (5). The classical kinase cascade, consisting of Raf and the MAP/ERK kinases MEK1/2, is used. ERK1/2 are activated to a low but detectable extent inβ-cells in low (2–3 mM) and basal glucose (≈5 mM) and are further activated as the glucose concentration increases up to 8–10 mM (2). Thus, ERK1/2 are regulated by glucose throughout the circulating concentration range. Glucagon like peptide (GLP)-1 and related hormones further increase insulin secretion and ERK1/2 activity (5, 6).
Several factors bind to the insulin gene promoter to enhance transcription in response to glucose (4). PDX-1 and Beta2 (NeuroD1) synergistically activate insulin gene transcription; mutations in either can lead to maturity-onset diabetes of the young (MODY) (7). MafA also contributes to glucose-responsiveness (8–10) and forms a complex with nuclear factor of activated T cells (NFAT), which is activated by calcium signaling. Gene disruption experiments revealed key roles for NFAT inβ-cell function, as was suggested by its impact on the insulin gene promoter (11, 12). Insulin gene transcription is blocked by calcineurin inhibitors, which prevent the activation of both NFAT and ERK1/2 by glucose and other events that depolarizeβ-cells (2, 5). Negative regulators of insulin promoter activity include Jun and the CCAAT/enhancer binding proteinβ(C/EBP-β), which are increased inβ-cells during prolonged exposure to low and high glucose, respectively (13, 14). Several of these key factors, Beta2, PDX-1, MafA, NFAT, and C/EBP-β, are ERK1/2 substrates (1, 2, 15–18). MafA, NFAT, and C/EBP-βassociate with the insulin gene promoter in an ERK1/2-dependent manner (2). Transactivating activities of Beta2 and PDX-1 depend at least partly on ERK1/2 activity (1).
Among the best characterized ERK1/2-sensitive elements is the serum response element (SRE) in the c-fos promoter (19). Elk1 and other ternary complex factors (TCFs) are ERK1/2 substrates that bind the serum response factor and target it to the SRE (19–21). ERK1/2 also phosphorylate and, in some cases, also bind to general transcription factor and mediator complex subunits (22–25). In exploring the mechanisms through which ERK1/2 exert actions on transcription inβ-cells, we found that the kinases themselves are components of chromatin-bound transcription factor complexes on both the insulin and c-fos genes.
Results
ERK1/2 Regulate Binding of Transcription Factors to the Insulin Gene Promoter.
To identify mechanisms by which ERK1/2 regulate insulin gene transcription, we used ChIP to determine whether ERK1/2 could affect binding of factors to a region of the promoter from -247 to -2 bp, encompassing the glucose-sensitive A2–E1 region (Fig. 1A). Experiments were performed in Min6 and INS-1 culturedβ-cells and in human islets as indicated in the figures. We compared ChIP from cells exposed acutely to stimulatory glucose or depolarized with K+ (Fig. 1B). We used 10-fold serial dilutions of precipitated DNA in PCR to determine input thresholds for detecting relative changes in enrichment of associated factors in response to these two cell treatments for later comparison to samples from unstimulated and inhibitor-treated cells (Fig. 1C). At a 1:100 dilution, glucose and K+ differentially affected binding of factors to the promoter, indicating that these changes reflect relative affinities of factor binding to the promoter that can be used to compare the binding of a given factor with the promoter under different conditions (Fig. 1B). The K+-induced binding of these factors was seen at 5 min but not at 10 min or later (data not shown), corresponding to the transient effect of K+ on ERK1/2 activity. Although transient, K+-induced depolarization was sufficient to produce a 2- to 3-fold increase in insulin promoter activity detected 4 h after treatment and reduced by MEK1/2 or calcineurin inhibitors (Fig. 1E). Glucose caused a ≈10-fold or greater stimulation of the insulin gene promoter, which was also reduced by inhibitors of MEK1/2 or calcineurin (Fig. 1C).
Differential effects of glucose and K+ depolarization on promoters in islets andβ-cells. (A) Schematic representation of the relative location of primers used to amplify the insulin gene promoter in ChIP assays and the sites of factor binding. (B) ChIP assay on the insulin promoter in Min6 cells showing effects of glucose and K+ depolarization on enrichment of transcription factor association detected at 1-, 10-, and 100-fold dilutions of ChIP DNA by PCR. (C) Effects of inhibitors on glucose- and K+- induced transcription factor association. (D) Effects of acute glucose (30 min) and K+ (5 min) stimulation in cells that were cultured in 5.5 mM (Basal) or 25 mM glucose for three days (long-term). (E) Assay of insulin promoter reporter activity in response to glucose and K+ stimulation (4 h) in Min6 cells that were cultured in either basal or long-term high glucose. NS, no stimulation; G and Glc, glucose; U, U0126; FK, FK520; R, rapamycin.
Under the conditions described above, the glucose-induced binding of PDX-1, Beta2, and MafA was blocked by U0126 and the calcineurin inhibitor (Fig. 1C), as shown for MafA in ref. 2. Rapamycin, which blocks the mTorC1 pathway, did not prevent binding of any of these factors to the insulin gene promoter. The K+-dependent interactions were also blocked by the MEK1/2 and calcineurin inhibitors, showing that stimuli that activate ERK1/2 with different kinetics send distinct ERK1/2-dependent signals to this gene (6).
Effect of Glucose Outside the Normal Range on Transcription Factor Binding to the Insulin Gene Promoter.
ERK1/2 inhibit insulin gene transcription in cells chronically exposed to high glucose (≥11 mM for 24 h or more) as might occur in poorly controlled diabetes (2). Incubation in high glucose for days increases the amount of C/EBP-β, which suppresses insulin gene transcription (13). In cells exposed to high glucose for three days, placed in low glucose for 2 h, and then stimulated with glucose for 30 min, C/EBP-βwas bound to the insulin gene promoter in an ERK1/2-dependent manner, but other factors were no longer detected (Fig. 1D). In contrast, the K+ response was unchanged; both NFAT and MafA were on the promoter and C/EBP-βwas not detected, suggesting that the stronger activation of NFAT interferes with C/EBP-βbinding (2). Glucose inhibited reporter activity in cells exposed to long-term high glucose in a manner reversed by U0126 or FK520 (Fig. 1E). The reduced promoter activity under these conditions reflects the inhibitory effect of C/EBP-βon transcription (2).
ERK1/2 Are Bound to Chromatin on Gene Promoters.
A key question is how ERK1/2 coordinate binding of transcription factors to multiple promoters and regulate their functions. We considered the idea that ERK1/2 may associate with chromatin-bound transcription factor complexes to access components directly on promoters, as reported for Hog1 in yeast, for example (26). To test this idea, we used both glucose and NGF, another ligand that activates ERK1/2 by a different mechanism inβ-cells (Fig. 2B) (6, 27) as a measure of specificity. We examined both the insulin gene promoter and the fos gene promoter, because fos is a well characterized target for the effects of ERK1/2 on transcription. ERK1/2 stimulate fos transcription through the ternary complex factor Elk-1 and the transcription factor TFII-I and indirectly through phosphorylation of the serum response factor by the ERK1/2-activated kinase Rsk (19, 24). Glucose stimulation of Min6β-cells caused a doubling of Fos mRNA (Fig. 2A). Similar results were observed in INS-1 cells. As in other systems, inhibition of ERK1/2 caused a severalfold reduction in Fos mRNA.
Signal-transduced gene regulation and enrichment of ERK1/2 with promoter DNA in islets andβ-cells. (A) Quantitative PCR of fos gene expression in Min6 cells exposed to 5.5 or 25 mM glucose in the presence of U0126. (B) Diagram of differential regulation of ERK1/2 by glucose and NGF. (C) ChIP analysis of insulin and fos gene promoters in INS-1 cells in response to glucose, GLP-1, and NGF without or with FK520. (D) ChIP analysis of the insulin promoter in INS-1 cells in response to NGF in the presence of FK520 and U0126 (Upper); pERK1/2 immunoblot (Lower). (E) ChIP assay time course of the fos gene promoter in response to glucose and K+ in INS-1 cells without or with FK520. (F) Schematic representation of the relative location of primers used to amplify the fos gene promoter in ChIP assays and the contained elements.
We found ERK1/2 bound to both promoters in INS-1 cells and in human islets after exposure to glucose or glucose plus GLP-1 (Figs. 2 C and E and 3A, C, and D). Blocking activation of ERK1/2 by using the MEK1/2 inhibitors U0126 or PD98059 or the calcineurin inhibitor prevented the glucose-induced association of ERK1/2 with the promoters, just as these drugs blocked promoter binding of the ERK1/2-dependent transcription factors. NGF activated ERK1/2 independently of calcineurin inβ-cells and stimulated association of ERK1/2 with both promoters. The MEK1/2 inhibitor but not the calcineurin inhibitor blocked NGF-induced binding of ERK1/2 to the promoters (Figs. 2 C and D and 3 A and C).
Stimulus-dependent association of MAPK signaling components with insulin and fos gene promoter DNA. (A) ChIP analysis of B-raf, MEK1, and ERK2 association with the fos promoter in INS-1 cells in response to glucose, GLP-1, and NGF in the presence of U (U0126), PD (PD098059), and FK (FK520). (B) Immunoprecipitation of Raf1 and B-Raf from glucose-stimulated INS-1 lysates. (C) MEK1/2 and ERK1/2 association with insulin and fos promoters in human islets in response to glucose, GLP-1, and NGF in the presence of U (U0126) and FK (FK506). (D) ERK1/2, RSK2, NFAT, and calcineurin A association with the insulin promoter in human islets in response to glucose and GLP-1 in the presence of U (U0126) and FK (FK506). Rabbit IgG was used as an antibody control in ChIPs.
The association of ERK1/2 with the fos gene promoter was detected within 5 min of exposure of INS-1 cells to 11 mM glucose, just as increased ERK1/2 activity was observed and persisted for at least 30 min (Fig. 2E) (1). ERK1/2 activation after depolarization with K+ is more transient than with glucose (5), which paralleled ChIP findings in INS-1 cells showing ERK1/2 binding at 5 min but not at longer times (Fig. 2E). FK520 blocked both glucose- and K+-induced promoter association of ERK1/2.
We probed for the potential association of upstream kinase cascade components on these promoters, including MEK1/2 and B-Raf. B-Raf is activated by glucose upstream in the ERK1/2 cascade inβ-cells (L.D. and M.H.C., unpublished data) and MEK1/2 are both activated in these cells (data not shown) (1). B-Raf was not found on the Fos (Fig. 3A) or insulin (data not shown) gene promoters, despite the fact that the B-Raf antibody immunoprecipitates the protein (Fig. 3B). In contrast, MEKs were identified on the insulin and fos gene promoters in glucose- and NGF-stimulated cells and in human islets, as deduced by ChIP with isoform-specific antibodies (Fig. 3 A and C). MEKs were lost from the promoter complexes in cells or islets exposed to U0126. As with ERK1/2 binding, glucose-induced but not NGF-induced MEK binding was blocked by FK506. ERKs were detected by using either an ERK1/2 antibody or an ERK2-selective antibody.
We tested the possible chromatin association of the ERK1/2-stimulated protein kinase Rsk2. Rsk2 phosphorylates DNA binding proteins, including the cAMP response element binding protein, NFAT, and histones may contribute to ERK1/2-dependent gene transcription. ChIP with antibodies to Rsk showed that it was associated with the insulin gene promoter (Fig. 3D); its promoter complex binding was blocked by MEK1/2 or calcineurin inhibitors. Thus, Rsk2 is bound to the promoter in a manner that parallels and depends on activation of ERK1/2. Calcineurin regulates both ERK1/2 and NFAT function in these cells. Thus, we also probed for its presence on the promoter (Fig. 3D). Calcineurin, like Rsk2, was present on the promoter in cells stimulated with glucose with or without GLP-1, and binding was reduced by either MEK1/2 or calcineurin inhibitors. As observed above, glucose did not cause tight binding of NFAT to the insulin gene promoter; however, the addition of GLP-1 with glucose was sufficient to induce NFAT binding (Fig. 3D).
ERK1/2 Are Bound to the Same Promoters as Their Transcription Factor Substrates.
To determine whether ERK1/2 were bound simultaneously with transcription factor substrates to promoters, sequential ChIP was used. Material immunoprecipitated from islets treated with glucose or glucose plus GLP-1 with anti-ERK1/2 was subjected to a second immunoprecipitation to determine whether transcription factors were present in ERK1/2 complexes (Fig. 4A). Antibodies to Beta2, MafA, and PDX-1, but not NFAT or C/EBP-β, also precipitated the insulin gene promoter first pulled down with anti-ERK1/2 from glucose-treated cells. Similarly, in the presence of glucose plus GLP-1, sequential ChIP experiments with antibodies to ERK1/2, Beta2, MafA, PDX-1, and NFAT revealed insulin gene promoter DNA, but antibodies to C/EBP-βwere again ineffective. These results are consistent with data of Fig. 1B, showing the glucose-sensitive association of these factors with the insulin gene promoter, and support the hypothesis that these factors are in the same complexes with ERK1/2 (Fig. 4A).
Cooccupancy of ERK1/2 with transcription factors on the insulin gene promoter and requirement for MAPK signaling components. (A) Sequential ChIP assay for cooccupancy of ERK1/2 with transcription factors on the insulin gene promoter in human islets in response to glucose and GLP-1. (B) ChIP analysis of effects of transfected MAPK signaling components on ERK1/2 and MEK1/2 association with insulin gene promoter DNA. Rabbit IgG and goat IgG were used as antibody controls in ChIPs.
Promoter Binding of the Kinases is Induced by Active MEK1 and Inhibited by Kinase-Inactive ERK2.
As an independent means of determining if ERK1/2 activity is required for promoter binding, we coexpressed 410 bp of the insulin gene promoter along with constructs encoding active and inactive kinases in INS-1 cells to test their effects on promoter binding. The inactive mutant K52R ERK2 is commonly used to inhibit ERK1/2 activation in cells by competition for MEK1/2; K52R ERK2 is phosphorylated in cells, even though it has little or no kinase activity. Expression of K52R ERK2 with the insulin gene promoter fragment blocked binding not only of ERK1/2 but also of MEK1/2 to the promoter (Fig. 4B), supporting the conclusion that ERK1/2 kinase activity is required for promoter binding.
We have shown that activation of ERK1/2 by activated mutants of proteins in the signaling cascade is sufficient to stimulate insulin gene promoter activity (1). Thus, we examined the effect of expressing constitutively active MEK1 (MEK1R4F) or MEK1-ERK2 fusion proteins, ERK2 activated as a consequence of fusion to MEK1, on promoter interactions. Both active MEK1 and the fusion proteins were sufficient to induce binding of ERK2 and MEK1 to the promoter (Fig. 4B). MEK2 bound in the presence of glucose but was not found with the promoter if any of the MEK1 proteins was overexpressed, suggesting that MEK1 is saturating binding sites.
Discussion
We show that glucose and NGF stimulate the binding of ERK1/2 to the insulin gene and other promoter complexes. Their association can be driven by expression of active forms of cascade components and blocked by a kinase-inactive ERK2 mutant and by MEK1/2 inhibitors. Kinase-dead ERK2 is highly phosphorylated in cells when expressed; thus, ERK2 phosphorylation is not sufficient for promoter binding in the absence of its activity. Kinase binding, like binding of the transcription factors themselves, requires kinase activity. The yeast p38 MAPK homolog has been found on numerous promoters in response to osmotic stress (26). Protein kinases, notably cyclinA-Cdk2, have also been identified in promoter complexes in mammalian cells (26, 28). Here, we find an ERK1/2 cascade module, including MEK1/2 and Rsk, on signal-sensitive promoters. Our work suggests that MAPK-containing signaling complexes are bound to responsive promoters along with their protein substrates to respond to incoming signals in situ to modulate transcription. From a position directly in protein complexes bound to chromatin, the kinase module may be regulated and phosphorylate transcription factors not otherwise accessible, coactivators and modifying enzymes, and RNA polymerase subunits themselves.
Regardless of the extent of glucose exposure or the transcriptional output, ERK1/2 control all of the glucose-sensitive interactions that we examined, and ERK1/2 are themselves present in promoter complexes in high glucose, whether acute or long-term. Yet, the combination of ERK1/2-dependent factors associated with the insulin gene promoter varies depending on the recent history of glucose exposure of the cells, as does the transcriptional output.
At least a portion of the acute and long-term glucose-induced changes in insulin gene transcription are mediated through ERK1/2 effects on factor binding to the promoter (1, 2). Many of the transcription factors that bind the promoter in an ERK1/2-sensitive manner are reported ERK1/2 substrates, although there is little in vivo data. DNA binding of a Beta2/E47 heterodimer depends on prior phosphorylation with ERK2. Transactivation assays with Beta2 and PDX-1 suggest that consensus ERK1/2 sites phosphorylated in vitro are required for maximum activity. A consensus ERK1/2 phosphorylation site in C/EBP-βwas shown to be required for Ras-stimulated interaction with serum response factor (16). Interestingly, calcineurin is also promoter bound in an ERK1/2 activity-dependent manner. This association is unlikely to be due to interaction with NFAT, because calcineurin was detected on the promoter even when NFAT was not. Interactions with Rsk or other factors may account for the presence of calcineurin in DNA-bound complexes.
ERK1/2 may also have indirect effects on proteins present in promoter complexes, for example, through downstream protein kinases such as Rsk, which may phosphorylate NFAT (29). A second indirect effect may arise from synergistic interactions of transcription factors on the insulin gene promoter. Phosphorylation of one factor may promote the association of tightly bound partners that themselves may not be substrates (30).
Changes in the transcription factor composition occur in cells in long-term high glucose, including increases in C/EBP-βand Myc, and decreases in PDX-1 (2, 13, 31–33). Myc has been shown to compete with Beta2 for binding to E-boxes but lacks the transactivating activity or the capacity to bind to p300. These events may account for the loss of glucose-induced Beta2 binding to the insulin promoter in high glucose.
Distinct stimuli lead to the recruitment of different factors to the insulin gene promoter in a manner differentially sensitive to the glucose state. In contrast to glucose, depolarization induced by the addition of extracellular K+ causes the same factor binding pattern regardless of exposure to glucose and preferentially recruits NFAT. This stimulus-specific response may arise from the difference in calcium signaling kinetics. The rapid but transient increase in calcium may lead to a more potent, if short-lived, activation of NFAT. The kinetics and amplitude of changes in intracellular free calcium have been suggested to lead to varied signaling outputs (34). C/EBP-βhas been suggested to compete with a MafA-NFAT complex for DNA binding. A strong NFAT signal in depolarized cells may enhance promoter binding of an NFAT-MafA complex despite the accumulation of C/EBP-βin long-term high glucose. Phosphorylation of promoter bound complexes may occur at different rates due to differences in localization or substrate access, resulting in different amounts of binding-competent species in the vicinity of the promoter. Other potential contributors to differences in response to glucose and depolarization include the activation of calcium-independent pathways by glucose, introducing an additional layer of regulation on promoter binding.
In conclusion, we find that signaling cascade components, including protein kinases and phosphatase, reside in acutely regulated promoter complexes. The protein kinases examined to date all appear on promoters when they are active and are no longer captured on chromatin if their activity is blocked. Association can be induced without ligands by expression of activated kinase mutants. Because the transcription factors also bind in a parallel manner, dependent on activity of the activating kinase, it seems most likely that the kinases bind along with their transcription factor substrates. The fact that multiple kinases from a kinase cascade are found in the complexes suggests that the more downstream kinases may be inactivated and reactivated in situ in the vicinity of or loosely bound to DNA. These findings suggest a dynamic regulatory pattern, in which MAPKs, phosphatases, and other protein kinase families participate along with transcription factors, polymerases, and chromatin remodeling enzymes to acutely control transcription activation and repression.
Experimental Procedures
Cell Culture.
Cells were either maintained in basal glucose (5.5 mM) or long-term high glucose [11 mM, INS-1 (35); 16 mM, human islets; 25 mM, Min6 (36)] for three days or longer before treatments. These conditions were selected from those often reported in the literature for each cell system. Culture in high glucose is the standard culture condition forβ-cell lines. For treatments, cells were placed in 2 mM glucose for 2 h, unless otherwise indicated, before stimulating for 5–30 min with high glucose (2, 37). FK506, FK520, and rapamycin were used at a concentration of 100 nM, and U0126 at 5 or 10 mM, unless otherwise indicated.
Isolation of Human Pancreatic Islets.
Islets were isolated from human cadaveric pancreata according to methods described in ref. 38 at the cGMP Islet Cell Processing Laboratory, Baylor Regional Transplant Institute. After procurement, pancreata were preserved by using a combination of University of Wisconsin solution and oxygenated perfluorodecalin (two-layer method) before islet isolation (39, 40). The pancreatic duct was cannulated and initially perfused with cold collagenase enzyme NB1 (1.4 mg/ml) combined with neutral protease (SERVA Electrophoresis). The pancreatic tissue was then digested by using a Ricordi Chamber with enhanced mechanical agitation at 37°C. Islets were purified from pancreatic tissue by a continuous density gradient of high-osmolality Ficoll- or iodixanol- based solutions by using a COBE2991cell processor (Gambro Laboratories). The purity and quantity of islets were determined by staining with diphenylthiocarbazone. The viability was determined by staining with fluorescein diacetate and propidium iodide. Islets were cultured in RPMI or KRBH media in 5.5 or 16 mM glucose according to experimental design.
DNA Constructs.
Luciferase promoter-reporters pGL3-rInsI containing -410 to + 1 of the rat I insulin promoter and expression vectors encoding pCMV5-MEK1 R4F, pCMV5-Myc-ERK2-MEK1, pCMV5-Myc-ERK2-MEK1(LA), and pCMV5-ERK2 K52R were as described (2, 37, 41, 42).
Single and Sequential Chromatin Immunoprecipitation.
Chromatin from INS-1, Min6, or human pancreatic islets was cross-linked with 1% formaldehyde and sonicated as described in ref. 37. Antibodies immobilized on protein A-Sepharose beads were used to immunoprecipitate factors cross-linked to DNA. The ERK1/2, ERK2, MEK1, and MEK2 antibodies were previously described (X837, A249, A2227, A2228) (43, 44). The following antibodies were purchased from Santa Cruz Biotechnology: NFATc3 (F-1), c-Maf (M-153), C/EBP-β(C-19), PDX-1 (N-18), NeuroD1 (N-19), Fos (4), Jun (H-79), RSK2 (C-19), B-Raf (F-7), calcineurin. DNA-protein complexes were eluted in 100 mM NaHCO3 buffer containing 1% SDS. For sequential ChIP, eluates were diluted 10-fold to reduce SDS concentration, and multiple samples were pooled and scaled-up 18-fold to provide sufficient material for a second round of immunoprecipitations with a different antibody. cross-links were reversed with 0.2 M NaCl at 65°C for 5 h, and the DNA was purified by phenol-chloroform extraction. We used 10-fold serial dilutions of precipitated DNA to determine thresholds for detecting changes in factor association with DNA by PCR by using the following primer sets: 5′-CTGGGAAATGAGGTGGAAAA-3′and 5′-AGGAGGGGTAGGTAGGCAGA-3′(rInsI -329 to -90); 5′-AACTGGTTCATCAGGCCATC-3′and 5′-ACTGGGTCCCCACTACCTTT-3′(mInsII -247 to -2); 5′-GAGGAAGAGGTGCTGACGAC-3′and 5′-CCATCTCCCCTACCTGTCAA-3′(hIns -194 to -41); 5′-TCCCTCCCTCCTTTACACAG-3′and 5′-AGCGGAACAGAGAAACTGGA-3′(rFos -328 to -63); and 5′-GTTGAGCCCGTGACGTTTAC-3′to 5′-TAGGACATCTGCGTCAGCAG-3′(hFos -463 to -288).
Real-Time Quantitative PCR.
Total RNA was isolated from cells by using TRI reagent (Ambion). cDNA was synthesized with reverse transcriptase by using a High-Capacity cDNA Archive (ABI) with random hexamers. TaqMan probe-based PCR was performed by using the ABI 7500 DNA Sequence Detection System with standard fluorescent chemistries and thermal cycling conditions specified by the manufacturer: 50°C for 2 min, 95°C for 10 min for one cycle and an additional 40 cycles at 95°C for 15 sec, ramp to 60°C for 1 min. TaqMan Gene Expression Assays were obtained from Applied Biosystems; 18S rRNA was used as an internal expression control and was amplified with primer and probe sequences from Ribosomal RNA Control Reagents (Applied Biosystems).
Transfections and Reporter Assays.
Min6 and INS-1 cells were cultured for two days in basal or high glucose before transfection with reporter and expression plasmids by using Lipofectamine2000 (Invitrogen) in antibiotic-free medium overnight. On the third day, glucose was removed for 12 h before stimulation with glucose. Cells were harvested in Reporter Lysis Buffer (Promega) supplemented with 200 mM b-glycerophosphate and 2 mM Na3VO4. Reporter assays were performed by using Dual Luciferase Reporter Assay System (Promega) by using a TD-20/20 bioluminometer (Turner Designs) as described in refs. 2 and 37.
Immunoblot Analysis.
Cells were harvested in lysis buffer (50 mM Hepes, pH 7.5/150 mM NaCl/1.5 mM MgCl2/1 mM EGTA/10% glycerol/0.2 mM Na3VO4/50 mM b-glycerophosphate/100 mM NaF/0.1% Nonidet P-40). Protein samples were resolved by electrophoresis on 10% polyacrylamide gels in SDS and immunoblotted by using anti-pERK1/2 antibody (Sigma) and horseradish peroxidase-conjugated anti-mouse secondary antibody (2, 37).
Statistical Analyses.
Quantitative data are expressed as means ± SD determined from at least three independent experiments, unless otherwise stated. Statistical significance was calculated by one-tailed unpaired Student's t test.
Acknowledgments
We thank Michael White (Department of Cell Biology), Elliott Ross (Department of Pharmacology), Eric Wauson, Kyle Wedin, Arif Jivan, and Charles Heise of the Cobb laboratory for comments about the data and manuscript, and Dionne Ware for administrative assistance. This work was supported by the National Institutes of Health Grants DK55310 and DK34128 (to M.H.C.). During the initial stages of this work M.L. was supported by American Diabetes Foundation and William K. Warren Medical Research Institute Mentor-based Postdoctoral Fellowship. C.S. and L.D. completed this work in partial fulfillment of the requirements for the Ph.D.
Footnotes
- ‡To whom correspondence should be addressed. E-mail: melanie.cobb{at}utsouthwestern.edu
-
Author contributions: M.C.L. and M.H.C. designed research; M.C.L., K.M., C.S., and L.D. performed research; B.N. and M.F.L. contributed new reagents/analytic tools; M.C.L. analyzed data; and M.C.L. and M.H.C. wrote the paper.
-
The authors declare no conflict of interest.
-
This article is a PNAS Direct Submission.
- © 2008 by The National Academy of Sciences of the USA
