Glucocorticoids inhibit glucose utilization and protein synthesis, and promote protein degradation

18 Glucocorticoids and FoxO3 exert similar metabolic effects in skeletal muscle. FoxO3 19 gene expression was increased by dexamethasone (Dex), a synthetic glucocorticoid, both in vitro 20 and in vivo. In C2C12 myotubes the increased expression is due to, at least in part, the elevated 21 rate of FoxO3 gene transcription. In the mouse FoxO3 gene, we identified three glucocorticoid 22 receptor (GR) binding regions (GBRs): one being upstream of the transcription start site, 23 17kbGBR; and two in introns, +45kbGBR and +71kbGBR. Together, these three GBRs contain 24 four 15-bp glucocorticoid response elements (GREs). Micrococcal nuclease (MNase) assay 25 revealed that Dex treatment increased the sensitivity to MNase in the GRE of +45kbGBR and 26 +71kbGBR upon 30 and 60 min Dex treatment, respectively. Conversely, Dex treatment did not 27 affect the chromatin structure near the -17kbGBR, in which the GRE is located in the linker 28 region. Dex treatment also increased histone H3 and/or H4 acetylation in genomic regions near 29 all three GBRs. Moreover, using chromatin conformation capture (3C) assay, we showed that 30 Dex treatment increased the interaction between the -17kbGBR and two genomic regions: one 31 located around +500 bp and the other around +73 kb. Finally, transcriptional coregulator, p300, 32 was recruited to all three GBRs upon Dex treatment. The reduction of p300 expression decreased 33 FoxO3 gene expression and Dex-stimulated interaction between distinct genomic regions of 34 FoxO3 gene identified by 3C. Overall, our results demonstrate that glucocorticoidactivated 35 FoxO3 gene transcription through multiple GREs, with chromatin structural change and DNA 36 looping. 37 38 INTRODUCTION 39 Glucocorticoids play a critical role in the regulation of skeletal muscle physiology. 40 Glucocorticoids inhibit glucose utilization and protein synthesis, and promote protein degradation 41 in skeletal muscle. Inhibiting glucose utilization preserves plasma glucose, the primary energy 42 source for brain, and promoting protein degradation and repressing protein synthesis produce free 43 amino acids, which can be used as the substrates for hepatic gluconeogenesis. These effects are 44 important metabolic adaptation for the survival of mammals during stress conditions, such as 45 fasting and starvation. However, chronic or excess glucocorticoid treatment can cause severe 46 metabolic disorders, such as muscle atrophy and insulin resistance in skeletal muscle (21, 32, 35, 47 39, 40). 48 Glucocorticoids exert their biological functions mainly through binding to an intracellular 49 receptor, the glucocorticoid receptor (GR). Upon binding to glucocorticoids, GR is recruited to 50 genomic glucocorticoid response elements (GRE) to regulate the transcription of its target genes. 51 These primary target genes then trigger glucocorticoid-regulated physiological responses. Our 52 previous chromatin immunoprecipitation sequencing (ChIPseq) experiments identified 4 GR 53 binding regions (GBR) in or near the mouse FoxO3 gene (22). The induction of FoxO3 mRNA 54 and protein levels by glucocorticoids has been shown previously both in vitro and in vivo (15, 29, 55 34, 50, 52). The FoxO3 gene encodes a transcription factor that plays a vital role in skeletal 56 muscle protein and glucose metabolism (9, 33). First, the transcription factor inhibits glucose 57 oxidation by activating the transcription of pyruvate dehydrogenase kinase 4 (PDK4) (23). 58 Interestingly, glucocorticoids also induce PDK4 gene transcription (23). In the human PDK4 gene 59 promoter, the binding of both FoxO3 and FoxO1, another member of the FoxO transcription 60 factor family, are necessary for the maximum level of glucocorticoid-induced PDK4 transcription 61 (23). Second, like glucocorticoids, the FoxO3 transcription factor stimulates the transcription of 62 genes that activate protein degradation, such as MuRF1 and atrogin-1; and genes that suppress 63 protein synthesis; such as Eif4ebp1 (37, 47). MuRF1 and atrogin-1 are muscle–specific ubiquitin 64 ligase that have been shown to mediate muscle atrophy caused by various conditions. 65 Constitutively active FoxO3 has been shown to be sufficient to induce muscle atrophy, whereas 66 dominant negative Foxo3 has been shown to prevent the muscle atrophy caused by either disuse 67 or glucocorticoids (37, 47). Furthermore, the dominant negative form of FoxO3 suppresses 68 glucocorticoid-induced atrogin-1 gene expression (37, 47). Recently, microRNA that decreases 69 FoxO3 gene expression has been shown to decrease glucocorticoid-induced expression of 70 atrogin-1 gene in C2C12 myotubes (15). Overall, these results strongly suggest that 71 glucocorticoid-regulated glucose and protein metabolism require the participation of FoxO3. 72 Because of the important role of FoxO3 in glucocorticoid action, we systematically 73 examine the mechanisms of glucocorticoid-regulated FoxO3 gene expression in this report. With 74 nuclear run-on assay, we examined whether the FoxO3 gene was transcriptionally regulated by 75 glucocorticoids. We characterized the four GBRs in the FoxO3 gene identified by ChIPseq, and 76 further identified 15-bp GREs that mediate the glucocorticoid response within these GBRs. The 77 acetylation status of histone H3 and H4 surrounding these GBRs was checked. Moreover, we 78 mapped the position of nucleosomes wrapped by these GBRs, and studied the effect of 79 glucocorticoid treatment on their chromatin structures. Finally, we used chromatin conformation 80 capture (3C) assay to test the potential interactions between GBRs and the genomic region near 81 the transcription start site (TSS). 82 83 EXPERIMENTAL PROCEDURES 84 Cell culture: Mouse C2C12 cells were purchased from the Cell and Tissue Culture Facility at the 85 University of California, Berkeley. They were maintained in Dulbecco’s modified Eagle’s 86 medium (DMEM; Mediatech) containing 10% fetal bovine serum (FBS; Tissue Culture 87 Biologicals) and incubated at 37°C with 5% CO2. Upon reaching 95~100 % confluence, C2C12 88 myoblasts were differentiated into myotubes with 2% horse serum (J.R. Scientific) in DMEM. 89 The C2C12 cells were maintained in 2% horse serum-containing DMEM, changed every 2 days, 90 until fully differentiated into myotubes, taking approximately 4-6 days. For cell culture 91 experiments, C2C12 myotubes were treated with various concentrations of Dex (Sigma), RU486 92 (Mifepristone, Sigma) or an equal volume (0.05% v/v of media) of vehicle control ethanol 93 (EtOH) or DMSO. 94 95 Animals: Male 8-week-old C57BL/6 mice were purchased from Charles River. Mice were 96 injected with 5 mg/kg Dex (Sigma) or PBS for 1 or 4 days. After the treatment, gastrocnemius 97 muscles were isolated from mice for gene expression analysis. For in vivo ChIP, male 8-week-old 98 C57BL/6 mice were injected with PBS or Dex at 11 am for 2 or 3 h. Gastrocnemius muscles were 99 then isolated for ChIP. Gastrocnemius muscle of 14 weeks old transgenic mice overexpressing 100 corticotropin-releasing hormone (CRH) (46) and wild type C57BL/6 mice were provided by the 101 laboratory of Charles Harris (Washington University, St. Louis, USA). In these CRH-Tg 102 transgenic mice, CRH gene was driven by mouse metallothioneine-1 promoter. Thus, CRH gene 103 is overexpressed in almost all tissues in these transgenic mice. The levels of corticosterone were 104 approximately 10 times higher than normal C57BL/6 mice (46). The Office of Laboratory Animal 105 Care at the University of California, Berkeley (Approval number R306-0111) approved all animal 106 experiments conducted in this paper. 107 108 Nuclear run-on: C2C12 myotubes were untreated or treated with 1 μM Dex for 30min, 1h, 2h or 109 4h. Cells were then washed once with PBS, and 3ml of lysis buffer (10 mM Tris-HCl pH 7.4, 3 110 mM MgCl2, 10 mM NaCl, 150 mM sucrose and 0.5% NP40) was added to each plate, followed 111 by incubation at 4C for 10-15 min. Cell lysate was collected and spun at 170x g at 4C for 5 min 112 to pellet the nuclei. Nuclei were washed once with lysis buffer w/o NP40, and resuspended in 113 freezing buffer (50 mM Tris-HCl at pH 8.3, 40% glycerol, 5 mM MgCl2 and 0.1 mM EDTA). 114 The total number of nuclei from each of the untreated and Dex-treated samples was counted, and 115 1x10 nuclei were used for in vitro transcription. Two aliquots from each sample were used, one 116 sample was incubated in 100 μl of 2x in vitro transcription buffer (200 mM KCl, 20 mM Tris117 HCl pH 8.0, 5 mM MgCl2, 4 mM dithiothreitol (DTT), 4 mM each of ATP, GTP and CTP, 200 118 mM sucrose and 20% glycerol) with 8 μl biotin-UTP (Roche, or equal amount from Epicentre), 119 and the other sample was incubated in 100 μl 2x in vitro transcription buffer with 8 μl UTP 120 (negative control) for 60 min at 30C. Then, 6 μl of 250 mM CaCl2 and 6 μl of RNAse-free 121 DNase (Roche) (10 U/ml) were added to stop the reactions. Total RNA was isolated using the 122 Nucleospin RNA II kit (Macherey-Nagel). 123 124 Dyna beads M-280 (Invitrogen) were washed twice in solution A (0.1 mM NaOH, 0.5 M NaCl) 125 for 5 min, once in solution B (0.1 M NaCl) for 5 min, and then resuspended in binding/wash 126 buffer (10 mM Tris-HCl pH 7.5, 1 mM EDTA and 2 M NaCl) with 1 μl (40 units) RNasin per 127 100 μl of beads. Then, 50 μl of beads were added to total RNA isolated, incubated at 42C for 30 128 min, followed by vigorous shaking on a shaker at room temperature for 3h. The beads were 129 precipitated with magnets and centrifugation, and supernatant was discarded. The beads were 130 then washed once for 15 min with 500 μl 15% formamide with 2x saline-sodium citrate (SSC) 131 buffer, twice for 5 min with 1 ml 2x SSC buffer, then resuspended in 30 μl RNaseand DNase132 free water. Finally, 10 μl of beads were used for each reverse transcription (RT) reaction prior to 133 real-time PCR (qPCR). These primers were used in qPCR: mFOXO3_runon_F, 134 ACTCCCGTCTTTTCCTCTCC; mFOXO3_runon_R, GGAAGTGATCTTGGCAGGTC; 135 mRPL19_cDNA_F, ATGGAGCACATCCACAAGC; and mRPL19_cDNA_R, 136 TCCTTGGTCTTAGACCTGCG. 137 138 RNA isolation and quantitative PCR: Total RNA was isolated from mouse gastrocnemius 139 muscles using TRI Reagent® RT (Molecular Research Center, Inc.). To synthesize randomly 140 primed cDNA, 0.5 μg of total RNA, 4 μl of 2.5 mM dNTP and 2 μl of 15 μM random primers 141 (New England Biolabs) were mixed at a volume of 16 μl, and incubated at 70°C for 10 min. 142 Then, a 4-μl cocktail containing 25 units of Moloney Murine Leukemia Virus (M-MuLV) 143 Reverse Transcriptase (New England Biolabs), 10 units of RNasin (Promega) and 2 μl of 10x 144 reaction buffer (New England Biolabs) was added, and samples were incubated at 42°C for 1h 145 and then at 95°C for 5 min. The cDNA was diluted and used to perform real-time quantitative 146 PCR (qPCR) using the EVA QPCR SuperMix Kit (Biochain), following manufacturer’s protocol. 147 qPCR was performed in either a 7900HT, 7500HT or StepOne PCR System (Applied 148 Biosystems) and analyzed with the ∆∆-Ct method, as supplied by the manufacturer (Applied 149 Biosystems). Rpl19 gene expression was used for internal normalization. Primer sequences for 150 qPCR are: Rpl19_cDNA_F, ATGGAGCACATCCACAAGC, Rpl19_cDNA_R, 151 TCCTTGGTCTTAGACCTGCG; FoxO3_cDNA_forward, TTCAACAGTACCGTGTTTGGAC; 152 and FoxO3_cDNA_reverse, AGTGTGACACGGAAGAGAAGGT. 153 154 Western blotting: RIPA buffer (10 mM Tris-HCl, pH 8.0, 1mM EDTA, 150 mM NaCl, 5% 155 glycerol, 0.1% sodium deoxycholate, 0.1% SDS, and 1% Triton X-100), supplemented with 156 protease inhibitors, was added to cell pellet. The mixture was gently rocked at 4°C for 1 h. The 157 supernatant was then collected as protein sample. NuPAGE Novex Bis-Tris mini gels 158 (Invitrogen) were used, following the manufaturer’s protocol, and proteins were transferred to 159 nitrocellulose membranes (Amersham) using semidry transfer (Bio-Rad) overnight. The next 160 day, membranes were blocked for 4 h at room temperature with 10% (wt/vol) nonfat milk in TBS 161 (50 mM Tris-base, 200 mM NaCl, pH 7.5). Membranes were then incubated in 5% milk in TBS 162 with appropriate primary antibody with gentle rocking overnight at 4°C. The following day, 163 membranes were washed with TBS plus 0.5% Tween-20 at pH 7.5 (TBST), and then incubated in 164 5% milk in TBS containing appropriate secondary antibody for at least 2 h at room temperature. 165 The membranes were then washed with TBST, and proteins were detected by chemiluminescence 166 (Western Lighting Plus-ECL, Perkin Elmer). For additional protein detection on the same 167 membrane, membranes were soaked in TBS overnight at 4°C, and stripped for 30 min in PBS 168 with 7 μl/ml β-mercaptoethanol, followed by 30min in PBS only, and 4 h in 10% milk in TBS 169 before re-probing with other primary antibodies. The following antibodies were used: FoxO3 (07170 702, Millipore), p300 (sc-584, Santa Cruz Biotechnology), β-actin (C4) mouse monoclonal IgG1 171 (sc-47778, Santa Cruz Biotechnology), GAPDH (abcam, ab9483), anti-rabbit IgG1-HRP (Cell 172 Signaling) and anti-goat IgG-HRP (sc-2768, Santa Cruz Biotechnology). Blots were scanned and 173 analyzed with Image J software (http://rsbweb.nih.gov/ij). β-actin or GAPDH was used as an 174 internal control. 175 176 Plasmids, transfection, and luciferase reporter assay: pGL4.10-E4TATA reporter plasmid was 177 generated by insertion of a 50-bp minimal E4 TATA promoter sequence (25) into the Bgl II to 178 Hind III sites of vector pGL4.10 to drive luciferase expression (1). Each chosen GBR fragment, 179 extending 100-150 bp upstream and downstream of the GBR, was amplified from genomic 180 C2C12 DNA (primer sequences are available in Supplemental Material S1), using the Expand 181 Long Template PCR System (Roche Applied Science) and cloned into the pGL4.10-E4TATA 182 vector with Kpn I/Xho I sites. We also replaced E4 TATA promoter region with the promoter 183 region of FoxO3 gene (-70 to +25 bp relative to its transcription start site, TSS). The QuikChange 184 Lightning mutagenesis kit (Stratagene) was used to make site-directed mutations per the 185 manufacturer’s instructions. Lipofectamine 2000 (Invitrogen) was used to transfect C2C12 186 myoblast according to the technical manual. Twenty-four h post-transfection, cells were treated 187 with either 1 μM Dex or control EtOH in differentiation media for 16-20 h. Cells were then 188 harvested and their luciferase activities were measured with the Dual-Luciferase Reporter Assay 189 kit (Promega) according to procedures in the technical manual. 190 191 Chromatin immunoprecipitation (ChIP): Fully differentiated C2C12 myotubes were treated 192 with 1 μM Dex or control EtOH for 1h, cross-linked in 2% formaldehyde for 3 min at 37°C. The 193 reactions were quenched with 0.125 M glycine. The cells were then washed with 1x PBS, and the 194 cells were scraped and lysed in cell lysis buffer (50 mM HEPES-KOH at pH 7.4, 1 mM EDTA, 195 150 mM NaCl, 10% glycerol, 0.5% Triton X-100), supplemented with protease inhibitor cocktails 196 (Calbiochem). The cell lysate was incubated for 1h at 4°C, and the crude nuclear extract was 197 collected by centrifugation at 600xg for 5 min at 4°C. The nuclei were resuspended in 1 mL of 198 ice-cold RIPA buffer (10 mM Tris-HCL at pH 8.0, 1 mM EDTA, 150 mM NaCl, 5% glycerol, 199 1% Triton X-100, 0.1% sodium deoxycholate, 0.1% SDS, supplemented with protease inhibitor). 200 The chromatin was fragmented with a Branson Sonifier 250 sonicator (13 min total, 20 sec pulse 201 at 35% power followed by 40 sec pause). To remove insoluble components, we centrifuged the 202 samples at 13,000 rpm for 15 min at 4°C and recovered the supernatant. Fifty μl of lysates were 203 retained and prepared as input. After proteinase K treatment and reverse cross-linking, DNA 204 fragments were purified using QIAquick PCR purification kit (Qiagen) as described at the end of 205 this paragraph. For GR ChIP, 1 μg of rabbit polyclonal anti-GR antibody (N499, provided by 206 Keith R. Yamamoto laboratory, UCSF) was added to the supernatant to immunoprecipitate GR207 bound chromatin at 4C overnight. For histone modification ChIP, the following antibodies were 208 used: anti-histone H3 (ab1791, abcam), anti-acetyl histone H3 (ab47915, abcam), anti-histone H4 209 (05-858, Millipore), anti-acetyl histone H4 (06-866, Millipore), CBP (sc-7300x, Santa Cruz 210 Biotechnology) and p300 (sc-584x, Santa Cruz Biotechnology). Normal rabbit IgG antibody (sc211 2027, Santa Cruz Biotechnology) was used as negative control for all ChIP. The next day, 100 μl 212 of 50% protein A/G (GR ChIP, Santa Cruz Biotechnology) or protein A (histone modification 213 ChIP, Upstate) bead slurry, containing 100 μg/ml salmon sperm DNA, was added into each 214 immunoprecipitation and nutated at 4C for 2h. The beads were then washed twice with RIPA 215 buffer, three times with RIPA buffer containing 500 mM NaCl, twice with LiCl buffer (20 mM 216 Tris at pH 8.0, 1 mM EDTA, 250 mM LiCl, 0.5% NP-40, 0.5% sodiumdeoxycholate) and one 217 time with RIPA buffer, all supplemented with protease inhibitor. After removing the remaining 218 wash buffer, 75 μl of proteinase K solution (TE pH 8.0, 0.7% SDS, 200 μg/ml proteinase K) was 219 added to each IP reaction, followed by incubation at 55°C for 3h and 65°C for overnight to 220 reverse formaldehyde cross-linking. ChIP DNA fragments were purified with QIAquick PCR 221 purification kit (Qiagen), eluting in 60 μl of Qiagen Elution Buffer. Primers used for ChIP are in 222 Supplemental Material S1. 223 For each ChIP experiment, the percentage of input was calculated by taking the ratio of 224 ChIP DNA fragment to input DNA fragment. Since input DNA fragments were isolate from 50 μl 225 whereas ChIP DNA fragments were purified from 950 μl lysate, the 19x difference was taking 226 into consideration for data calculation. 227 228 In vivo ChIP: The in vivo ChIP protocol was modified from two previous reports (41, 42). Mice 229 were injected intraperitoneally with Dex (5 mg/kg body weight, in 200 μl solution) or an equal 230 volume of PBS. Gastrocnemius muscles were excised and snap frozen in isopentane cooled by 231 liquid nitrogen and then crushed into a fine powder using a mortar and pestle. Samples were 232 suspended in a 1% formaldehyde solution in PBS and incubated at 37C for 10 minutes and 233 incubated for 5 minutes in 125mM of Glycine, followed by re-suspension in 50mM Tris pH8, 1% 234 SDS, 10mM EDTA, 1mM DTT, and protease inhibitors, then incubated at 10C for 10 min and 235 sonicated at 60% output for a total of 50 seconds. Cleared lysates were then diluted 3 fold in a 236 dilution buffer of 0.01% SDS, 1.1% Triton x-100, 1.2 mM EDTA, 16.7mM Tris pH 8, 167mM 237 NaCl, and protease inhibitor. 100ul of 50% protein A/G beads with 5ug of IgG antibody were 238 added and incubated for 1 h at 4 degrees C. Supernatants were collected and samples were 239 incubated with antibodies overnight at 4C. Beads were washed once with a low salt washing 240 buffer (150mM NaCl, 20 mM Tris pH 8.1, 0.1% SDS, 1% Triton X-100, and 2mM EDTA), a 241 high salt washing buffer (500mM NaCl, 20 mM Tris pH 8.1, 0.1% SDS, 1% Triton X-100, and 242 2mM EDTA), a LiCL washing buffer (250mM LiCl, 1% sodium deoxycholate, 1% Nonidet P-40, 243 1mM EDTA, and 10mM Tris pH 8.1), and twice with Tris-EDTA Buffer. Protein-chromatin 244 complexes were eluted with 10mM DTT, 1% SDS, and 100mM NaHCO3 for 1 hour and the 245 supernatant was then incubated overnight at 65C with 200mM NaCl. Proteins were digested 246 with Proteinase K for 3 hours at 55C and DNA fragments were isolated using QIAquick DNA 247 purification kit (QIAGEN, Valencia, CA). Real-time PCR were used to monitor the recruitment 248 of GR to various GBRs. 249 250 251 Micrococcal nuclease (MNase) assay: The protocol for MNase assay was previously described 252 (43). Briefly, C2C12 myotubes were treated with 1 μM Dex or an equal volume (0.05% v/v of 253 media) of vehicle control ethanol (EtOH) for 30 min or 60 min. Cells were cross-linked with 1% 254 formaldehyde for 3 min at 37C, and the reaction was quenched by the addition of glycine to a 255 final concentration of 0.125 M. Cells were then washed once with PBS and scraped in ice-cold 256 MNase NP-40 lysis buffer (10 mM Tris pH 7.4, 10 mM NaCl, 3 mM MgCl2, 0.5% NP-40, 0.15 257 mM spermine, 0.5 mM spermidine). After shaking for 3-5h at 4C in MNase lysis buffer, nuclei 258 were collected by centrifugation and washed in ice-cold MNase digestion buffer w/o CaCl2 (10 259 mM Tris at pH 7.4, 15 mM NaCl, 60 mM KCl, 0.15 mM spermine, 0.5 mM spermidine). 260 Samples were then resuspended in ice-cold MNase digestion buffer with CaCl2 (10 mM Tris at 261 pH 7.4, 15 mM NaCl, 60 mM KCl, 0.15 mM spermine, 0.5 mM spermidine, 1 mM CaCl2). 262 Nuclei were treated with 1 unit of MNase (Nuclease micrococcal from Staphylococcus aureus, 263 Sigma-Aldrich, N5386-200UN) for 60-90 min at 25C. Reactions were stopped by the addition of 264 80 μl MNase digestion buffer with CaCl2, 20 μl MNase stop buffer (100 mM EDTA, 10 mM 265 EGTA), 75 μg Proteinase K, and 20 μl 10% SDS, and then cells were incubated at 65C 266 overnight. Samples were run on 1.5% agarose gel, and single nucleosome-wrapped DNA (about 267 150 bp) was purified with Qiagen gel extraction kit. The concentration of the samples was 268 measured and diluted to 0.3 ng/μl for use in qPCR. The qPCR primers were designed to span 269 approximately 500-bp regions, covering the identified GREs in each of the GBRs. When there 270 were gaps between primers, the gaps were no more than 20 bps long. Primers used for MNase 271 assay are available in Supplemental Material S1. Notably, a titration curve using genomic DNA 272 was performed for each primer in individual MNase experiment to calculate for the absolute 273 DNA amount. 274 275 Lentiviral infection: Mouse C2C12 myoblasts were infected with p300 shRNA lentiviral particle 276 (sc-29432v, Santa Cruz Biotechnology) or control shRNA lentiviral particle (sc-108080, Santa 277 Cruz Biotechnology), and selected with 5μg/ml puromycin for several days. Sh-p300 and sh-scr 278 C2C12 myoblasts were then differentiated into myotubes. Three days after differentiation, 1 μΜ 279 Dex or EtOH were added to sh-p300 or sh-scr myotubes for 6 h, followed by Western blotting or 280 gene expression assay. 281 282 Chromatin conformation capture (3C): Two 15-cm plates of C2C12 myotubes were used: one 283 treated with 1 μM Dex, the other with control EtOH for 1 h. After treatment, cells were fixed in 284 2% formaldehyde and incubated at 37C for 3 min. Glycine was added to a final concentration of 285 0.125 M at room temperature for 5 min to quench the cross-linking. Cells were then washed with 286 ice-cold PBS and resuspended in 5 ml of ice-cold PBS. The cells were transferred through a 40 287 μm nylon cell strainer (BD Falcon 352340) and centrifuged at 320 x g (100 rpm) at 4C for 7 288 min. The pellet was then resuspended in 20 ml ice-cold lysis buffer (10 mM Tris-HCl at pH 8.0, 289 10 mM NaCl, 0.2% NP-40 and complete protease inhibitor) and rotated at 4C for at least 1h. 290 Samples were centrifuged at 15,000 to 18,000 rpm at room temperature for 5 min and 291 resuspended in 2 ml of 1.1x Bgl II restriction enzyme buffer, 60 μl of 10% SDS (final 292 concentration 0.3%), 200 μl of 20% Triton X-100 (final concentration 1.8%) and shaken at 37C 293 for 1h. Then, 1,600 units of Bgl II restriction enzyme were added to each sample and incubated at 294 37C overnight. The next day, 320 μl of 10% SDS (final concentration 1.3%) was added and 295 incubated at 65C for 30 min, followed by the addition of 1.5 ml of 20 % Triton X-100 (final 296 concentration 1%), 2.8 ml of 10x T4 DNA ligase buffer (NEB) and up to 28 ml of de-ionized 297 water, followed by incubation at 37C for 1h. Then, 1 μl of T4 DNA ligase (400 u) was added to 298 each sample, and samples were kept at 4C overnight. The next day, 60 μl of proteinase K (20 299 mg/ml, Fermentas) were added, and samples were incubated at 65C overnight. The next 300 morning, 120 μl of RNase A (10 mg/ml, Fermentas) were added, followed by 45 min incubation 301 at 37C. Four rounds of phenol/chloroform extraction were used to clear SDS, and aqueous phase 302 was recovered. Then, DNA was precipitated with 0.1 volumes of 3 M sodium acetate at pH 4.8 303 and 2 volumes of 100% ethanol in -80C overnight. Followed by centrifugation at 4000 rpm at 304 4C for 60 min, the DNA pellets were washed with 70% ethanol and spun at 4000 rpm at 4C for 305 10 min. The DNA pellets were dissolved in 500 μl of TE buffer, pH 8.0. Forward and reverse 306 primers for qPCR are designed to flank the junction of Bgl II restriction enzyme site, and paired 307 according to Fig. 6. Primers for 3C are listed in Supplemental Material S1. 308 309 Statistical Analysis: Error Bars represent Standard Error Mean (S.E.M.) for each group. Student’s 310 t test were used for all statistical analyses. 311 312 RESULTS 314 Glucocorticoids increase the expression of FoxO3 gene in vitro and in vivo 315 C2C12 myotubes were treated with vehicle control ethanol (EtOH) or dexamethasone 316 (Dex, 1 μM), a synthetic glucocorticoid, for 2, 6, 24 or 48 h. Except for 2 h treatment, Dex 317 markedly increased FoxO3 gene expression at all time points (Fig. 1A). We then used the shortest 318 yet effective treatment time (6 h) for the subsequent dose response analysis. As shown in Fig. 1B, 319 while treatment of 10 and 100nM of Dex significantly stimulated the FoxO3 gene expression, 1 320 μM of Dex provided the strongest induction. By treating C2C12 myotubes with a combination of 321 Dex and RU486, a GR antagonist, we found that RU486 co-treatment dose-dependently reduced 322 Dex-induced FoxO3 gene expression (Fig. 1C). These results demonstrated that Dex response on 323 FoxO3 gene expression in C2C12 myotubes is GR-dependent. 324 To study the glucocorticoid effect on FoxO3 gene expression in vivo, wild type (WT) 325 mice were injected with control PBS or Dex (5 mg/kg/day body weight) for 4 days. FoxO3 gene 326 expression was 2.5 fold higher in the gastrocnemius muscle from Dex-injected WT mice than that 327 from PBS-treated ones (Fig. 1D). Next, we utilized transgenic mice overexpressing corticotropin328 releasing hormone (CRH, CRH-Tg) (4, 14) to examine long-term glucocorticoid effects on 329 FoxO3 gene expression. These transgenic mice have chronically elevated levels of CRH, which 330 stimulates the secretion of adrenocorticotropin hormone (ACTH). ACTH further increases the 331 secretion of corticosterone to circulation. We found that FoxO3 gene expression was about 1.8 332 fold higher in gastrocnemius muscle from CRH-Tg mice than that from WT ones (Fig. 1E). 333 Overall, these experiments demonstrated that the expression of FoxO3 was induced by 334 glucocorticoids both in vitro and in vivo. 335 To confirm that Dex-upregulated FoxO3 mRNA levels were correlated with FoxO3 336 protein levels, C2C12 myotubes were treated with Dex or control EtOH for 24 h. Immunoblotting 337 revealed that Dex treatment resulted in 1.5 fold increase of FoxO3 protein levels (Fig. 1F). 338 Next, we performed nuclear run-on assay to test the induction of FoxO3 gene in 339 transcriptional level by glucocorticoids. Since transcriptional activation of the FoxO3 gene by 340 Dex should occur prior to the observation of its mRNA increase, we chose time points earlier 341 than 6 h for the nuclear run-on assay. Fig. 1G shows that FoxO3 transcription was significantly 342 induced by 0.5 h, 1 h, 2 h and 4 h of the treatment. These results indicate that the increased 343 FoxO3 gene expression was due to, at least in part, the induction of its transcription, as early as 344 0.5 h after Dex treatment. 345 Insulin antagonizes the effects of glucocorticoids on protein metabolism, as it activates 346 protein synthesis and reduces protein degradation (21, 40). To investigate whether insulin 347 represses glucocorticoid-induced FoxO3 gene expression, C2C12 myotubes were treated with 348 control EtOH, Dex, insulin or a combination of Dex and insulin for 6 h. As shown in Fig. 1H, 349 insulin reduces the basal expression of FoxO3 gene. The absolute level of FoxO3 induction by 350 Dex is also decreased by insulin; however, the fold change of Dex-induced FoxO3 gene 351 expression in the presence of insulin is similar to that in the absence of insulin (3.2 vs 3.2 fold, 352 represented as hash tags in Fig. 1H). These results suggest that insulin reduces the expression of 353 basal expression of FoxO3 gene rather than suppressing Dex-stimulated FoxO3 gene 354 transcription. 355 356 Identification of GBRs in the genomic region of FoxO3 gene 357 Previously, our ChIPseq identified 4 potential GBRs in or near FoxO3 genomic region 358 (22). These include the genomic region between -17455 to -17126 (relative to the TSS, referred to 359 as the -17kbGBR), between +45231 and +45317 (called the +45kbGBR), between +71380 and 360 +71565 (called the +71kbGBR) and between +98640 and +98777 (called the +93kbGBR). The 361 +45kbGBR and the +71kbGBR were located in introns, whereas the +93kbGBR was located in 362 the 3’ untranslated region. Using conventional ChIP we found that GR was recruited to the 363 17kbGBR, the +45kbGBR and the +71kbGBR, but not the +93kbGBR, upon 1h treatment of Dex 364 on C2C12 myotubes (Fig. 2A). We individually inserted each GBR upstream of the TATA box in 365 a luciferase reporter, pGL4.10-E4TATA and performed reporter assay. For C2C12 myoblasts 366 transfected with reporters containing the -17kbGBR, the +45kbGBR or the +71kbGBR, Dex367 treated cells gave a significantly higher luciferase activity than EtOH-treated ones (Fig. 2C, 2E, 368 2G). These results indicate that the -17kbGBR, the +45kbGBR and the +71kbGBR contain 369 functional GREs that confer glucocorticoid responses. 370 We also replaced the promoter region of pGL-17kbWT, pGL+45kbWT and 371 pGL+71kbWT with -70 to +25 (relative to TSS) region of FoxO3 gene. With FoxO3 promoter, 372 all three GBRs again conferred Dex response, though the levels of response were lower (Fig. 2H). 373 Next, we searched for sequences resembling the consensus GRE identified from our 374 ChIPseq (51), RGXACAnnnTGTXCY, in the -17kbGBR, the +45kbGBR and the +71kbGBR. 375 Based on the consensus sequences, nucleotide position 2, 4, 5, 6, 10, 11, 12 and 14 has to be a 376 specific nucleotide. We looked for sequences that contain at least 6 of these 8 nucleotides. We 377 mutated position 11 of this consensus GRE from a G to a C residue, or position 5, from C to G 378 (Fig. 2B, 2D, 2F). These residues have been previously shown to make direct contact with the GR 379 (28). In the -17kbGBR, two GRE-like sequences (GLSs) were found (Fig. 2B). Mutation of 380 GLS1 resulted in more than a 95% decrease in response to Dex, whereas mutation of GLS2 381 caused about 80% reduction of Dex response. Double mutation of GLS1 and 2 completely 382 abolished its response to Dex (Fig. 2C). These results suggested that both GLS1 and 2 are 383 required to confer glucocorticoid response for -17kbGBR, while GLS1 plays a more prominent 384 role. In this regard, the sequence of GLS1, but not the sequence of GLS2, is highly conserved in 385 human and rat FoxO3 gene (Table 1). 386 For the +45kbGBR, three GLSs were located (Fig. 2D). Mutation of GLS2 had no effect 387 on Dex response, whereas mutation of GLS3 gave a 57% decrease in response to Dex (Fig. 2E). 388 Furthermore, mutation of GLS1 completely eliminated its response to Dex (Fig. 2E). These 389 results indicated that GLS1 plays a primary role, and GLS3 plays an accessory role in mediating 390 glucocorticoid response. Notably, the GLS1 sequence is also highly conserved in human and rat 391 FoxO3 genes (Table 1). 392 For the +71kbGBR, four GLSs were found (Fig. 2F). Mutation of GLS2, 3 or 4 had no 393 effect on the Dex response (Fig. 2G). However, mutation in GLS1 completely removed its 394 response to Dex (Fig. 2G). Therefore, GLS1 alone conferred a complete glucocorticoid response 395 in the +71kbGBR, and its sequence is also conserved in human and rat FoxO3 genes (Table 1). 396 To test whether three GBRs identified from C2C12 myotubes are occupied by GR in 397 vivo, mice were injected with PBS or Dex at 11 am. Twoor three h post-injection their 398 gastrocnemius muscles were isolated for ChIP. We found that GR occupied +45kb GBR and 399 +71kb GBR in gastrocnemius muscle after 2 and 3 h PBS treatment (Fig. 2I and 2J, respectively). 400 This suggests that endogenous corticosterone were enough to induce GR recruitment to these two 401 GBRs. Treating mice with Dex for 2 and 3 h further enhanced GR recruitment on these two 402 GBRs in gastrocnemius muscle (Fig. 2I and 2J). On the contrary, GR was recruited to -17kb GBR 403 after 2 h PBS treatment (Fig. 2I and 2J), but the recruitment was not enhanced by Dex. Overall, 404 these results confirmed that GR is recruited to the three GBRs in vivo. 405 406 Glucocorticoids increase the level of acetylated histones in FoxO3 genomic region surrounding 407 GBRs 408 Histone hyperacetylation is highly associated with transcription activation. We monitored 409 the levels of acetylated histone H3 and H4 (AcH3 and AcH4, respectively) as well as total H3 and 410 H4 in FoxO3 genomic regions containing GRE and regions located upstream and downstream 411 from each GRE. For the -17kbGBR, the +45kbGBR and the +71kbGBR, there is no significant 412 change in the total H3 and H4 levels between control EtOH and Dex treatment (Fig. 3A, B, C). 413 All three genomic regions present significant levels of acetylated H3 and H4 before Dex 414 treatment (Fig. 3D, E, F). These observations suggest that those are likely regulatory regions that 415 play a role in controlling basal FoxO3 gene expression. Dex treatment markedly induced 416 hyperacetylation of H3 or H4 in genomic regions surrounding all three GREs (Fig. 3D, E, F), 417 though the status of histone acetylation was not affected on the GRE region of each GBR. The 418 ratio of acetylated histones versus total histones was calculated to more precisely reflect histone 419 hyperacetylation status. Dex treatment increased AcH4/H4 levels upstream of the +45kbGBR 420 (Fig. 3H), as well as downstream of all three GBRs (Fig. 3G-I). Dex treatment also elevated 421 AcH3/H3 level downstream of the -17kbGBR (Fig. 3G). Overall, these results indicate that 422 glucocorticoids increase the acetylation status of histones surrounding, but not within, each GRE. 423 424 Glucocorticoid treatment differentially induce chromatin structural changes in GBRs 425 Treatment of glucocorticoids has been shown to disrupt nucleosome assembly or change 426 the position of nucleosome in the genome. We used micrococcal nuclease (MNase) to map the 427 position of nucleosomes surrounding the three FoxO3 GBRs. 428 For the -17kbGBR, three nucleosomes were detected: -17550 to -17400, -17400 to 429 17250, and -17200 to -17050. Dex treatment did not affect the position of these three 430 nucleosomes, as their sensitivity to MNase was similar between EtOH-treated and Dex-treated 431 cells. Interestingly, the major GRE (-17231 to -17217) in the -17kbGBR is located in a linker 432 region between nucleosome 2 and 3 (Fig. 4A, 4B). 433 For the +45kbGBR, two nucleosomes were observed: +44850 to +45050 and +45250 to 434 +45450. These nucleosomes appeared to cover more than 146 bp of DNA, probably due to the 435 lack of overlapping primer sets in certain GC-rich regions. Nonetheless, the effect of 436 glucocorticoids on these nucleosomes is apparent. Dex treatment for 30 min markedly increased 437 the sensitivity to MNase of both nucleosomes (Fig. 4C). This increase of sensitivity, however, 438 was not seen in cells treated with Dex for 60 min (Fig. 4D). Thus, the density of chromatin 439 structure of these two nucleosomes in the +45kbGBR region was reduced upon 30 min Dex 440 treatment, and it was transient. 441 For the +71kbGBR region, three nucleosomes were detected: +71300 to +71420, +71420 442 to +71560, and +71560 to +71700. Dex treatment for 30 min did not significantly affect the 443 position or sensitivity to MNase of these nucleosomes (Fig. 4E). However, 60 min treatment 444 significantly increased the sensitivity to MNase of the 2nd nucleosome. In fact, this nucleosome 445 is un-recognizable in the 60 min Dex-treated sample (Fig. 4F). This result strongly suggests that 446 chromatin structure was loosened up in the region of +71420 to +71560. Interestingly, this region 447 harbors the highly conserved +71kb GLS1 sequence. 448 Overall, these results show that glucocorticoids differentially modulate the chromatin 449 structure of the three GBRs in the FoxO3 genomic region, suggesting that distinct mechanisms 450 are adapted by these GBRs to participate in GR-activated FoxO3 gene transcription. 451 452 Dex-induced FoxO3 gene transcription requires p300 453 One key hallmark for active GREs is the recruitment of transcriptional coactivators by 454 GR. We examined the recruitment of transcriptional coregulators, p300 and CBP, to FoxO3 455 GBRs upon 30 min Dex treatment. We found that p300, but not CBP, was recruited to all three 456 GBRs upon Dex treatment (Fig. 5A). Reducing p300 in C2C12 myotubes with shRNA 457 lentiviruses (Fig. 5B) markedly decreased the induction of FoxO3 gene and protein expression by 458 Dex (Fig. 5C and 5D). These results indicated that p300 directly participates in GR-activated 459 FoxO3 gene transcription. 460 461 The potential interactions between genomic regions near FoxO3 GBRs and 578 bp 462 downstream of TSS 463 All three FoxO3 GBRs are located far away from the TSS. Therefore, to stimulate FoxO3 464 transcription, they may need to interact with the genomic region near the TSS, where the basal 465 transcription machinery is located. We performed chromatin conformation capture (3C) to 466 examine the potential interaction between GBRs and TSS genomic regions. C2C12 myotubes 467 were treated with control EtOH or Dex for 30 min or 60 min, cross-linked, followed by nuclei 468 isolation. The samples were digested with restriction enzyme Bgl II, and then diluted for 469 intracellular re-ligation. With primers pairing between genomic regions near each GBR and TSS, 470 qPCR was used to assess the change in relative crosslinking frequencies between Dexand EtOH471 treated samples. We tested more than a hundred primer pairs, where most of them did not detect 472 any PCR products. Nevertheless, we found fourteen primer pairs that detected significant levels 473 of PCR products. Fig. 6A showed the location of these primers (the nucleotide number represent 474 the midpoint of each primer sequence) in the FoxO3 genomic region. Levels of these fourteen 475 ligated products were not affected by 30 min Dex treatment (data not shown). However, upon 60 476 min Dex treatment, the levels of two ligated products were significantly increased (Fig. 6B and 477 6C): between primer pair +73506 and -17648 (2.3 fold induction, primer 7), and primer pair 478 17648 and +578 (2.3 fold induction, primer 14). This finding suggests that Dex treatment 479 enhanced DNA looping in a time-dependent manner, and resulted in increased interaction 480 between the 3’ end of the +71kbGBR and the -17kbGBR, and between the -17kbGBR and the 481 genomic region approximately 578 bp downstream of TSS. 482 To test whether p300 is involved in Dex-induced DNA looping, we performed 3C in 483 EtOHor Dex-treated C2C12 myotubes that express scramble shRNA (control) or p300 shRNA. 484 Consistent with the results above, Dex treatment resulted in 1.8 fold induction between primer 485 pair +73506 and -17648, and 1.6 fold induction between primer pair -17648 and +578 in C2C12 486 myotubes that express scramble shRNA (Fig. 6D). In constrast, both inductions were abolished 487 with the presence of p300 shRNA (Fig. 6D). These results indicate that p300 is required for Dex488 induced interaction between these genomic elements in FoxO3 gene. 489 490 DISCUSSION 491 Glucocorticoids and FoxO3 play similar roles in the regulation of protein and glucose 492 metabolism in skeletal muscle. The induction of FoxO3 gene expression by glucocorticoids could 493 have important physiological and/or pathological implications. Here, we extensively studied the 494 mechanism of glucocorticoid-activated FoxO3 gene expression. We showed that glucocorticoids 495 activated the transcription of the FoxO3 gene in C2C12 myotubes. We identified three GBRs in 496 or near the FoxO3 gene, and the functional GREs within the GBRs. Interestingly, these three 497 GBRs are far away from FoxO3 TSS. Several lines of evidence suggest that they all play a role in 498 glucocorticoid-activated FoxO3 gene transcription. 499 First, all three GBRs confer glucocorticoid response when individually inserted into a 500 reporter plasmid. Among these GBRs, four functional GREs were identified from nine GRE-like 501 sequences. Intriguingly, the sequences of three out of the four GREs (one from each GBR) are 502 highly conserved in human and rat FoxO3 gene (Table 1), suggesting that these GREs in human 503 and rat may also play a role in glucocorticoid-regulated FoxO3 gene transcription. Alternatively, 504 these GREs are conserved due to selective pressure during evolution. Notably, previous studies 505 have shown that certain glucocorticoid-regulated genes, such as tyrosine aminotransferase and 506 dual specificity phosphatase 1, contain multiple GREs (7, 48). Moreover. ChIP sequencing results 507 revealed many glucocorticoid-regulated contain multiple GR binding regions in their genome (8, 508 22, 36, 51), which suggests that regulation of gene transcription by multiple GREs is likely a 509 common mechanism. 510 We also showed that GR was recruited to these GBRs in vivo. GR was recruited to +45kb 511 and +71kb GBR in gastrocnemius muscle of PBS-treated mice, which indicates a basal level of 512 GR occupancy at these GBRs. Notably, Dex further enhanced GR recruitment to both +45kb and 513 +71kb GBR. In contrast, GR occupancy on -17kb GBR was not enriched with Dex treatment, 514 which suggests that further GR recruitment to -17kb GBR likely required other signals or 515 transcription factors. 516 Second, we observed changes in chromatin structure surrounding two GBRs: +45kbGBR 517 and +71kbGBR. An increase in sensitivity to MNase was observed in the nucleosome that 518 contains the GRE in the +45kbGBR upon 30 min glucocorticoid treatment. However, there is no 519 significant change in chromatin structure of the +71kbGBR at the same time point. Instead, its 520 structure was disrupted upon 60 min glucocorticoid treatment. Furthermore, upon 30 min 521 glucocorticoid treatment, while histone H4 was hyperacetylated both immediately upstream and 522 downstream of both the +45kbGBR and downstream of -17kbGBR and +71kbGBR, H3 is 523 hyperacetylated downstream of -17kbGBR only. 524 The chromatin structure surrounding the -17kbGBR was not affected at either time point. 525 Interestingly, its GRE is located within the linker region between two nucleosomes. Thus, unlike 526 the GREs in the +45kbGBR and the +71kbGBR which are located within nucleosome, the GRE 527 in the -17kbGBR is already exposed for GR binding. It could explain the absence of chromatin 528 structural change in the -17kbGBR. In the mean time, increasing histone acetylation surrounding 529 GBRs would loosen up chromatin structure near GREs to allow other transcription factors to 530 associate with their respective binding sites to further assist transcriptional activation. 531 It is intriguing that these three GBRs respond to glucocorticoid treatment differently to 532 modify their chromatin structure. Previous studies have shown that the nucleotide sequence of a 533 GRE plays a central role in modulating GR function (24, 45), as distinct GRE sequences were 534 shown to induce different conformational change of GR (30). This is likely affecting the ability of 535 GR to associate with transcriptional coregulators and other DNA-binding transcription factors 536 (27, 45). We therefore hypothesize that distinct sets of transcriptional coregulators may be 537 recruited to FoxO3 GBRs in a time-dependent fashion. So far, p300 was the only transcriptional 538 coregulator that we found to be recruited to all three GBRs upon Dex treatment for 30 min. Dex539 induced FoxO3 gene transcription appears to require p300, as p300 reduction markedly decreased 540 the ability of Dex to stimulate FoxO3 gene expression. It is possible that p300, a histone 541 acetyltransferase, participates in the hyperacetylation of genomic regions at and/or surrounding 542 GREs shown in Fig 3. To understand the differential effects of Dex on the chromatin structure 543 change, one would need to identify transcription coregulators that are differentially recruited to 544 GBRs, possibly at distinct time points of Dex treatment, in future studies. 545 Third, in 3C experiments, we found that 60 min Dex treatment increases the interaction 546 between genomic region downstream of +71kbGBR and the -17kbGBR, and between the 547 17kbGBR and the genomic region around +578 bp. Based on these results, we contemplate a 548 model for FoxO3 genomic configuration upon GR activation (Fig. 7). In this model, the 549 17kbGBR is drawn to the +578 bp genomic region upon glucocorticoid treatment to be closer to 550 the TSS. If we consider these interactions in a three-dimensional geometry, the +71kbGBR could 551 also be close to the TSS through its communication with the -17kbGBR (Fig. 7). Likely, there is 552 no direct interaction between genomic regions around +73kb and +578bp, as we did not detect 553 any ligated PCR product with respective primers. Nonetheless, such spatial arrangements could 554 still play a role in transcription regulation (12, 31). There is no available restriction site that 555 allows us to study genomic regions closer to TSS than +578 region. Nevertheless, we speculate 556 that +578 region may contain regulatory elements required for maximal glucocorticoid response 557 on FoxO3 gene transcription. 558 It is likely that certain levels of DNA looping occur before the Dex treatment, as we had 559 positive qPCR results from 14 primer pairs in EtOH-treated C2C12 myotubes. These results are 560 not surprising, as DNA looping may be required to provide basal transcription of FoxO3. Nuclear 561 run-on first showed Dex-induced FoxO3 gene transcription at 30 min time point, whereas DNA 562 looping was augmented after 1 h Dex treatment in 3C analysis. This suggests that the basal level 563 of DNA looping is sufficient to support the initiation of FoxO3 transcription by Dex. Enhanced 564 DNA looping induced by Dex at 1 h time point may be required to sustain prolonged FoxO3 gene 565 transcription. Alternatively, we cannot exclude the possibility that DNA looping is increased 566 outside of tested genomic regions by 30 min of Dex treatment. We did not observe an increased 567 interaction between the +45kbGBR and genomic regions tested. However, it does not exclude the 568 possibility that the +45kbGBR is drawn to the TSS upon glucocorticoid treatment at an earlier or 569 later time point, or the possibility that the +45kbGBR may interact with genomic regions outside 570 the ones tested. 571 The induction of DNA looping by glucocorticoids has been shown in other cases (11, 13, 572 17). But the mechanism of glucocorticoid-induced DNA looping is not clear. One DNA binding 573 protein that can induce DNA looping is CTCF (10, 26, 38, 44). Cohesin has been shown to act 574 with CTCF to induce DNA looping (5, 6). Based on CTCF ChIP sequencing results from the 575 UCSC genome browser, there are multiple CTCF binding sites in the FoxO3 genomic region. It is 576 unclear whether these CTCF binding sites play a role in basal and Dex-induced DNA looping in 577 FoxO3 genomic region. In addition to CTCF, other transcription cofactor may participate in DNA 578 looping event in a glucocorticoid-dependent manner. It has been shown that Mediators form a 579 complex with cohesin to participate in the induction of DNA looping (18). SWI/SNF chromatin 580 remodeling complex has also been shown to induce chromatin looping (19). It is well established 581 that GR can recruit Mediator (2, 3) and SWI/SNF complexes (16, 49). So far we were not able to 582 detect the recruitment of components in Mediator and SWI/SNF to FoxO3 GBRs. However, p300 583 was recruited to all three GBRs. Recent studies suggest that histone hyperacetylation induced by 584 p300 contributes to DNA looping between locus control region and globin gene promoter (20). 585 Indeed, when we reduced p300 expression in C2C12 myotubes, the ability of glucocorticoids to 586 induce DNA looping within FoxO3 gene was markedly reduced. It is unclear how p300 exerts 587 this effect; nonetheless, these experiments established a role of p300 in this process. 588 Since the three GREs showed distinct regulatory characteristics, it is likely that they serve 589 different functions in the activation of FoxO3 gene transcription. We speculate that +45kbGBR is 590 involved in the initial step of transcriptional activation, as chromatin structure surrounding 591 +45kbGBR loosened up upon 30 min Dex treatment (Fig. 7). This initial step may include the 592 recruitment of proteins to initiate DNA looping, as we observed the interaction between 593 +71kbGBR and -17kbGBR upon 1 h Dex treatment, at which time point the chromatin structure 594 surrounding +45kbGBR has returned to the basal state. At this time point, chromatin structure 595 surrounding +71kbGBR has loosened up, which may in turn recruit additional GR and/or other 596 transcription factors, and interact with -17kbGBR to activate transcription (Fig. 7). To test the 597 role of each GRE, it would be necessary to delete or mutate each of them endogenously, and 598 examine the effects on transcriptional activation process. 599 In summary, in this report we have shown that glucocorticoids employ a novel 600 mechanism in regulating FoxO3 gene transcription, where three GBRs, with a total of four GREs, 601 are involved in the transcriptional activation process. The interaction between two GBRs and 602 genomic region approximately 578 bp downstream of TSS likely plays a role in Dex-induced 603 FoxO3 gene transcription. Further study of these transcriptional regulations not only is important 604 in understanding of glucocorticoid action in skeletal muscle, but also will provide a valuable 605 model for elucidating the complex mechanisms of GR-regulated gene transcription. 606 607 ACKNOWLEDGEMENTS 608 We thank Dr. Charlie Harris for providing tissues from CRH-Tg mice, Dr. Hei Sook Sul 609 for the comments on the manuscript, and Dr. Ofir Hakim in Gordon Hager’s laboratory for 610 suggestions on 3C experiments. 611 612 GRANTS 613 This study was supported by the NIH (R01DK083591) and the Muscular Dystrophy Association 614 (Research Grant 186068). T.K. was supported by the Dissertation Award Fellowship (18DT615 0010) from the University of California Tobacco-Related Diseases Research Program. 616 617 DISCLOSURES 618 No conflicts of interest, financial or otherwise, are declared by the authors. 619 620 621 AUTHOR CONTRIBUTIONS 622 T.K. and J.C.W. conception and design of research; T.K. performed experiments, 623 analyzed data, interpreted results, prepared figures and wrote the manuscript. P.H.L., T.C.C., 624 R.A.L., J.N., D.Z., C.L., A.C., Y.T., and E.C. performed experiments and analyzed data. J.C. 625 Wang reviewed the experiments, analyzed data, interpreted results and wrote the manuscript.626 REFERENCES6276281. Bolton EC, So AY, Chaivorapol C, Haqq CM, Li H, and Yamamoto KR. Celland629gene-specific regulation of primary target genes by the androgen receptor. Genes Dev 21:6302005-2017, 2007.6312. Chen W, and Roeder RG. The Mediator subunit MED1/TRAP220 is required for632optimal glucocorticoid receptor-mediated transcription activation. Nucleic Acids Res 35:6336161-6169, 2007.6343. Chen W, Rogatsky I, and Garabedian MJ. MED14 and MED1 differentially regulate635target-specific gene activation by the glucocorticoid receptor. Mol Endocrinol 20: 560-636572, 2006.6374. Coste SC, Murray SE, and Stenzel-Poore MP. Animal models of CRH excess and638CRH receptor deficiency display altered adaptations to stress. Peptides 22: 733-741,6392001.6405. Degner SC, Verma-Gaur J, Wong TP, Bossen C, Iverson GM, Torkamani A,641Vettermann C, Lin YC, Ju Z, Schulz D, Murre CS, Birshtein BK, Schork NJ,642Schlissel MS, Riblet R, Murre C, and Feeney AJ. CCCTC-binding factor (CTCF) and643cohesin influence the genomic architecture of the Igh locus and antisense transcription in644pro-B cells. Proc Natl Acad Sci U S A 108: 9566-9571, 2011.6456. Dorsett D. Cohesin: genomic insights into controlling gene transcription and646development. Curr Opin Genet Dev 21: 199-206, 2011.6477. Espinas ML, Roux J, Ghysdael J, Pictet R, and Grange T. Participation of Ets648transcription factors in the glucocorticoid response of the rat tyrosine aminotransferase649gene. Mol Cell Biol 14: 4116-4125, 1994.6508. Grontved L, John S, Baek S, Liu Y, Buckley JR, Vinson C, Aguilera G, and Hager651GL. C/EBP maintains chromatin accessibility in liver and facilitates glucocorticoid652receptor recruitment to steroid response elements. EMBO J 32: 1568-1583, 2013.6539. Gross DN, van den Heuvel AP, and Birnbaum MJ. The role of FoxO in the regulation654of metabolism. Oncogene 27: 2320-2336, 2008.65510. Guo Y, Monahan K, Wu H, Gertz J, Varley KE, Li W, Myers RM, Maniatis T, and656Wu Q. CTCF/cohesin-mediated DNA looping is required for protocadherin alpha657promoter choice. Proc Natl Acad Sci U S A 109: 21081-21086, 2012.65811. Hakim O, John S, Ling JQ, Biddie SC, Hoffman AR, and Hager GL. Glucocorticoid659receptor activation of the Ciz1-Lcn2 locus by long range interactions. J Biol Chem 284:6606048-6052, 2009.66112. Hakim O, Sung MH, and Hager GL. 3D shortcuts to gene regulation. Curr Opin Cell662Biol 22: 305-313, 2010.66313. Hakim O, Sung MH, Voss TC, Splinter E, John S, Sabo PJ, Thurman RE,664Stamatoyannopoulos JA, de Laat W, and Hager GL. Diverse gene reprogramming665events occur in the same spatial clusters of distal regulatory elements. Genome Res 21:666697-706, 2011.66714. Harris C, Roohk DJ, Fitch M, Boudignon BM, Halloran BP, and Hellerstein MK.668Large increases in adipose triacylglycerol flux in Cushingoid CRH-Tg mice are explained669by futile cycling. Am J Physiol Endocrinol Metab 304: E282-293, 2013.67015. Hudson MB, Rahnert JA, Zheng B, Woodworth-Hobbs ME, Franch HA, and Price671SR. miR-182 attenuates atrophy-related gene expression by targeting FoxO3 in skeletal672muscle. Am J Physiol Cell Physiol 307: C314-319, 2014.67316. John S, Sabo PJ, Johnson TA, Sung MH, Biddie SC, Lightman SL, Voss TC, Davis674SR, Meltzer PS, Stamatoyannopoulos JA, and Hager GL. Interaction of the675glucocorticoid receptor with the chromatin landscape. Mol Cell 29: 611-624, 2008.676 17. Kabotyanski EB, Rijnkels M, Freeman-Zadrowski C, Buser AC, Edwards DP, and677Rosen JM. Lactogenic hormonal induction of long distance interactions between beta-678casein gene regulatory elements. J Biol Chem 284: 22815-22824, 2009.67918. Kagey MH, Newman JJ, Bilodeau S, Zhan Y, Orlando DA, van Berkum NL,680Ebmeier CC, Goossens J, Rahl PB, Levine SS, Taatjes DJ, Dekker J, and Young681RA. Mediator and cohesin connect gene expression and chromatin architecture. Nature682467: 430-435, 2010.68319. Kim SI, Bresnick EH, and Bultman SJ. BRG1 directly regulates nucleosome structure684and chromatin looping of the alpha globin locus to activate transcription. Nucleic Acids685Res 37: 6019-6027, 2009.68620. Kim YW, and Kim A. Histone acetylation contributes to chromatin looping between the687locus control region and globin gene by influencing hypersensitive site formation.688Biochim Biophys Acta 1829: 963-969, 2013.68921. Kuo T, Harris CA, and Wang JC. Metabolic functions of glucocorticoid receptor in690skeletal muscle. Mol Cell Endocrinol 380: 79-88, 2013.69122. Kuo T, Lew MJ, Mayba O, Harris CA, Speed TP, and Wang JC. Genome-wide692analysis of glucocorticoid receptor-binding sites in myotubes identifies gene networks693modulating insulin signaling. Proc Natl Acad Sci U S A 109: 11160-11165, 2012.69423. Kwon HS, Huang B, Unterman TG, and Harris RA. Protein kinase B-alpha inhibits695human pyruvate dehydrogenase kinase-4 gene induction by dexamethasone through696inactivation of FOXO transcription factors. Diabetes 53: 899-910, 2004.69724. Lefstin JA, and Yamamoto KR. Allosteric effects of DNA on transcriptional regulators.698Nature 392: 885-888, 1998.69925. Lin YS, Carey MF, Ptashne M, and Green MR. GAL4 derivatives function alone and700synergistically with mammalian activators in vitro. Cell 54: 659-664, 1988.70126. Ling JQ, Li T, Hu JF, Vu TH, Chen HL, Qiu XW, Cherry AM, and Hoffman AR.702CTCF mediates interchromosomal colocalization between Igf2/H19 and Wsb1/Nf1.703Science 312: 269-272, 2006.70427. Lucas PC, and Granner DK. Hormone response domains in gene transcription. Annu705Rev Biochem 61: 1131-1173, 1992.70628. Luisi BF, Xu WX, Otwinowski Z, Freedman LP, Yamamoto KR, and Sigler PB.707Crystallographic analysis of the interaction of the glucocorticoid receptor with DNA.708Nature 352: 497-505, 1991.70929. Lutzner N, Kalbacher H, Krones-Herzig A, and Rosl F. FOXO3 is a glucocorticoid710receptor target and regulates LKB1 and its own expression based on cellular AMP levels711via a positive autoregulatory loop. PLoS One 7: e42166, 2012.71230. Meijsing SH, Pufall MA, So AY, Bates DL, Chen L, and Yamamoto KR. DNA713binding site sequence directs glucocorticoid receptor structure and activity. Science 324:714407-410, 2009.71531. Mercer TR, and Mattick JS. Understanding the regulatory and transcriptional716complexity of the genome through structure. Genome Res 23: 1081-1088, 2013.71732. Morgan SA, Sherlock M, Gathercole LL, Lavery GG, Lenaghan C, Bujalska IJ,718Laber D, Yu A, Convey G, Mayers R, Hegyi K, Sethi JK, Stewart PM, Smith DM,719and Tomlinson JW. 11beta-hydroxysteroid dehydrogenase type 1 regulates720glucocorticoid-induced insulin resistance in skeletal muscle. Diabetes 58: 2506-2515,7212009.72233. Nakae J, Oki M, and Cao Y. The FoxO transcription factors and metabolic regulation.723FEBS Lett 582: 54-67, 2008.72434. Nishimura M, Mikura M, Hirasaka K, Okumura Y, Nikawa T, Kawano Y,725Nakayama M, and Ikeda M. Effects of dimethyl sulphoxide and dexamethasone on726 mRNA expression of myogenesisand muscle proteolytic system-related genes in mouse727myoblastic C2C12 cells. J Biochem 144: 717-724, 2008.72835. Pivonello R, De Leo M, Vitale P, Cozzolino A, Simeoli C, De Martino MC,729Lombardi G, and Colao A. Pathophysiology of diabetes mellitus in Cushing's730syndrome. Neuroendocrinology 92 Suppl 1: 77-81, 2010.73136. Reddy TE, Pauli F, Sprouse RO, Neff NF, Newberry KM, Garabedian MJ, and732Myers RM. Genomic determination of the glucocorticoid response reveals unexpected733mechanisms of gene regulation. Genome Res 19: 2163-2171, 2009.73437. Sandri M, Sandri C, Gilbert A, Skurk C, Calabria E, Picard A, Walsh K, Schiaffino735S, Lecker SH, and Goldberg AL. Foxo transcription factors induce the atrophy-related736ubiquitin ligase atrogin-1 and cause skeletal muscle atrophy. Cell 117: 399-412, 2004.73738. Sanyal A, Lajoie BR, Jain G, and Dekker J. The long-range interaction landscape of738gene promoters. Nature 489: 109-113, 2012.73939. Schakman O, Gilson H, and Thissen JP. Mechanisms of glucocorticoid-induced740myopathy. J Endocrinol 197: 1-10, 2008.74140. Schakman O, Kalista S, Barbe C, Loumaye A, and Thissen JP. Glucocorticoid-742induced skeletal muscle atrophy. The international journal of biochemistry & cell biology74345: 2163-2172, 2013.74441. Shimizu N, Yoshikawa N, Ito N, Maruyama T, Suzuki Y, Takeda S, Nakae J, Tagata745Y, Nishitani S, Takehana K, Sano M, Fukuda K, Suematsu M, Morimoto C, and746Tanaka H. Crosstalk between glucocorticoid receptor and nutritional sensor mTOR in747skeletal muscle. Cell Metab 13: 170-182, 2011.74842. Shimizu N, Yoshikawa N, Wada T, Handa H, Sano M, Fukuda K, Suematsu M,749Sawai T, Morimoto C, and Tanaka H. Tissueand context-dependent modulation of750hormonal sensitivity of glucocorticoid-responsive genes by hexamethylene bisacetamide-751inducible protein 1. Mol Endocrinol 22: 2609-2623, 2008.75243. Shipp LE, Lee JV, Yu CY, Pufall M, Zhang P, Scott DK, and Wang JC.753Transcriptional regulation of human dual specificity protein phosphatase 1 (DUSP1) gene754by glucocorticoids. PLoS One 5: e13754, 2010.75544. Splinter E, Heath H, Kooren J, Palstra RJ, Klous P, Grosveld F, Galjart N, and de756Laat W. CTCF mediates long-range chromatin looping and local histone modification in757the beta-globin locus. Genes Dev 20: 2349-2354, 2006.75845. Starr DB, Matsui W, Thomas JR, and Yamamoto KR. Intracellular receptors use a759common mechanism to interpret signaling information at response elements. Genes Dev76010: 1271-1283, 1996.76146. Stenzel-Poore MP, Cameron VA, Vaughan J, Sawchenko PE, and Vale W.762Development of Cushing's syndrome in corticotropin-releasing factor transgenic mice.763Endocrinology 130: 3378-3386, 1992.76447. Stitt TN, Drujan D, Clarke BA, Panaro F, Timofeyva Y, Kline WO, Gonzalez M,765Yancopoulos GD, and Glass DJ. The IGF-1/PI3K/Akt pathway prevents expression of766muscle atrophy-induced ubiquitin ligases by inhibiting FOXO transcription factors. Mol767Cell 14: 395-403, 2004.76848. Tchen CR, Martins JR, Paktiawal N, Perelli R, Saklatvala J, and Clark AR.769Glucocorticoid regulation of mouse and human dual specificity phosphatase 1 (DUSP1)770genes: unusual cis-acting elements and unexpected evolutionary divergence. J Biol Chem771285: 2642-2652, 2010.77249. Trotter KW, and Archer TK. Nuclear receptors and chromatin remodeling machinery.773Mol Cell Endocrinol 265-266: 162-167, 2007.77450. Waddell DS, Baehr LM, van den Brandt J, Johnsen SA, Reichardt HM, Furlow JD,775and Bodine SC. The glucocorticoid receptor and FOXO1 synergistically activate the776 skeletal muscle atrophy-associated MuRF1 gene. Am J Physiol Endocrinol Metab 295:777E785-797, 2008.77851. Yu CY, Mayba O, Lee JV, Tran J, Harris C, Speed TP, and Wang JC. Genome-wide779analysis of glucocorticoid receptor binding regions in adipocytes reveal gene network780involved in triglyceride homeostasis. PLoS One 5: e15188, 2010.78152. Zheng B, Ohkawa S, Li H, Roberts-Wilson TK, and Price SR. FOXO3a mediates782signaling crosstalk that coordinates ubiquitin and atrogin-1/MAFbx expression during783glucocorticoid-induced skeletal muscle atrophy. Faseb J 24: 2660-2669, 2010.784785FIGURE LEGENDS786787Fig. 1. Induction of FoxO3 gene expression and protein by glucocorticoids in vitro and in788vivo.789(A) C2C12 myotubes were treated with 1 μM Dex or control EtOH for 2, 6, 24 or 48 h. n=6.790(B) C2C12 myotubes were treated with 10 nM, 100 nM, 1 μM and 5 μM Dex or control EtOH for7916h. n=6.792(C) C2C12 myotubes were treated with 50 nM Dex and with DMSO (vehicle control), or 25 nM,79350 nM, 100 nM, and 1 μM of RU486 for 6 h, n=5.794(D) Male C57BL/6 mice were injected with 5mg/kg/day of Dex or control PBS for 1 day (24 h)795or 4 days (96 h). Their gastrocnemius muscles were isolated. n=10.796(E) Gastrocnemius muscles were collected from wild type and CRH-Tg mice, and n=12. For A,797B, C and D, primers specific to the FoxO3 and the Rpl19 (internal control) genes were used in798qPCR for gene expressions analysis. Fold induction was calculated by normalizing to Rpl19, and799taking the ratio of Dex over EtOH.800(F) C2C12 myotubes were treated with 1 μM Dex or control EtOH for 24 h, and FoxO3 protein801levels were measured. n=3.802(G) Nuclear run-on assay for monitoring FoxO3 in in vitro transcription. C2C12 myotubes were803untreated or treated with 1 μM Dex for 0, 0.5, 1, 2 or 4 h. Primers specific to FoxO3 (across exon8041 and intron 1) and Rpl19 (internal control) were used in qPCR. n=4.805(H) FoxO3 gene expression after co-treatment of Dex and insulin. n=3. C2C12 myotubes were806treated with control EtOH, Dex, insulin (100 nM) or a combination of Dex and insulin for 6 h.807Fold induction is calculated by comparing treatment to EtOH. For all, error bars indicate SEM. *808indicates p < 0.05. # specifies 3.2-fold induction. *(rep) indicates statistical significant repression809compared to EtOH. ** indicates statistically significant difference in fold induction between 100810nM and 1 μM Dex treatment in B), and statistically significant difference in fold induction811between PBS, 50 nM Dex and 1 μM RU486 treatment in C).812813Fig. 2. The identification of the GREs in FoxO3 GBRs.814(A) ChIP to confirm GR recruitment to GBRs identified from ChIPseq. C2C12 myotubes were815treated with 1 μM Dex or control EtOH for 1 h. Primers flanking the -17kbGBR, the +45kbGBR,816the +71kbGBR, the +93kbGBR and Rpl19 (negative control) were used in qPCR. Error bars817represent the SEM of the fold enrichment from three independent experiments, and * indicates p818< 0.05.819(B) The -17kbGBR genomic region.820(C) Reporter assay for plasmid containing -17kbGBR.821(D) The -45kbGBR genomic region.822(E) Reporter assay for plasmid containing +45kbGBR.823(F) The +71kbGBR genomic region.824(G) Reporter assay for plasmid containing +71kbGBR.825For 2B, 2D and 2F, the location of each GBR, the sequence of consensus GRE, the sequence of826 GLSs, and each mutated nucleotide are shown. For 2C, 2E and 2G, reporter plasmids described in8272B, 2D and 2F were cotransfected with a human GR expression vector and a Renilla internal828control plasmid in C2C12 myoblasts, and their luciferase activities are shown. T test is calculated829by comparing each reporter plasmid to pGL4.10 E4TATA vector.830(H) FoxO3 neutral promoter region -70 to +25 bp, relative to its transcription start site, was831cloned into pGL4.10 reporter plasmid to give pGL4.10 FoxO3 promoter reporter. n ≥ 5. Fold832induction was calculated by taking the ratio of luciferase activity in Dex-treated samples over833EtOH-treated ones. Error bars represent the SEM, and * p < 0.001.834(I-J) ChIP to confirm GR recruitment to GBRs in gastrocnemius muscle after 2 h (I) and 3 h (J)835PBS or Dex treatment. Error bars represent the SEM of the fold enrichment from at least three836independent experiments, and * indicates p < 0.05.837838Fig. 3. Assessment of histone H3 and H4 acetylation near the GRE of the FoxO3 gene.839C2C12 myotubes were treated with 1 μM Dex or control EtOH for 30 min. ChIP was performed840with antibodies recognizing total histone H3 and H4 (H3 and H4, respectively) in A, B and C, and841with antibodies recognizing acetylated H3 and H4 (AcH3 and AcH4, respectively) in D, E and F.842The ratio of AcH3 or AcH4 over H3 or H4 is presented in G, H and I. For A, D and G) The -84317kbGBR primer marks FoxO3 genomic region of -17308 to -17184; the -17kb GBR 5’ primer, -84417678 to -17570; and the -17kb GBR 3’ primer, -16909 to -16786. For B, E and H) The845+45kbGBR primer defines FoxO3 genomic region of +45219 to +45302; the +45kb GBR 5’846primer, +44853 to +44929; and the +45kb GBR 3’ primer, +45631 to +45683. For C, F and I)847The +71kb GBR primer identifies FoxO3 genomic region of +71443 to +71555; the +71kb GBR8485’ primer, +71290 to +71381; and the +71kb GBR 3’ primer, +72129 to +72215. Error bars849represent SEM, n ≥ 6 and * p < 0.05.850851Fig. 4. Nucleosome mapping of the FoxO3 GBRs.852C2C12 myotubes were treated 1 μM Dex or control EtOH. Nucleosome positions were analyzed853with MNase digestion, followed by qPCR with primers spanning FoxO3 GBRs. For 4A and 4B,854nucleosome mappings between -17600 and -17100 of the FoxO3 gene are shown for 30 min and8551 h treatment, respectively. 11 primers were used to span this region. For 4C and 4D, nucleosome856mappings between +44800 and +45600 of the FoxO3 gene are shown for 30 min and 1 h857treatment, respectively. 9 primers were used to span this region. For 4E and 4F, nucleosome858mappings between +71300 and +71700 of the FoxO3 gene are shown for 30 min and 1 h,859respectively. 9 primers were used to span this region. For Fig. 4, the positions of nucleosomes are860drawn. Error bars represent SEM of the genomic DNA amount from at least five independent861experiments. * p < 0.05 for the indicated primer comparing Dex to EtOH-treated samples.862863Fig. 5. Acetyltransferase p300 is recruited to all three FoxO3 GBRs, and is involved in Dex864induced FoxO3 transcription.865(A) C2C12 myotubes were treated with 1 μM Dex or EtOH for 30 min. ChIP was performed with866antibodies recognizing IgG (control), CBP, p300 or GR, followed by qPCR with primers867spanning FoxO3 GBRs. Fold enrichment was calculated by taking the ratio Dex-treated samples868over EtOH-treated ones. Dashed line indicates EtOH-treated basal level of 1, and n=3.869(B) Western blotting shows that p300 is knocked down in C2C12 myotubes treated with shRNA870lentivial particle against p300 compared to that against control scramble (scr).871(C) FoxO3 gene expression in Dex or EtOH-treated myotubes infected with shRNA lentiviral872particle against p300 or control scr. Fold induction in represents the ratio of Dex-treated samples873to EtOH-treated ones. Error bar represents SEM, n=3 and *p < 0.05.874(D) FoxO3 protein expression in Dex or EtOH-treated myotubes infected with shRNA lentiviral875particle against p300 or control scr. One representative result was shown from two independent876 experiments. The intensity of bands were measured using Image J. The bar graph showed the877average of two independent experiments. Error bar represents SEM, *p < 0.05.878879Fig. 6. Chromatin Conformation Capture (3C) identified potential physical interaction880between genomic regions near GBRs and TSS of FoxO3 gene.881C2C12 myotubes were treated with 1 μM Dex or control EtOH for 1 h. DNA was digested with882Bgl II and re-ligated with DNA ligase intermolecularly, and qPCR was used to detect the amount883of ligated products obtained from individual primer set pairing Region 1 (-17648, -5014, -578),884Region 2 (+49809, +53118) and Region 3 (+70172, +73506, +75896). The nucleotide position885represents mid-point of each primer.886(A) A schematic diagram of Bgl II sites in Region 1, 2 and 3 of the FoxO3 gene.887(B) PCR amplification results of primer set 1-14. The data represented SEM of fold enrichment888(Dexto EtOH-treated samples) from 7 independent experiments. A bacterial artificial clone889(BAC, RP24-177H14) harboring the entire FoxO3 genomic region was used as control for890random chromatin interaction. The data shown here all have higher amount of ligated products in891Dexand EtOH-treated samples compared to PCR amplification from BAC. Ligated products892from primer set 7 and 14 are significantly enriched after Dex treatment, pairing +73506 to -17648893and -17648 to +578, respectively. * indicates p < 0.05.894(C) Circos plot interprets the results from B) and indicates relative crosslinking frequencies895observed between positions in Region 1, 2 and 3.896(D) 3C performed with C2C12 myotubes expressing scramble shRNA or p300 shRNA. Fold897enrichment was calculated by taking the ratio of Dexto EtOH-treated cells from two898independent experiments. * indicates p < 0.05.899900Fig. 7. Schematic Model of glucocorticoid-induced Foxo3 chromatin structure change901The chromatin structure of the FoxO3 gene upon 30 min and 1 h Dex treatment. The arrow902indicates the location of each conserved GRE. The black ovals indicate packed nucleosomes and903the white ovals indicate nucleosomes that are loosened or disrupted. Ac indicates hyperacetylated904histones. Before Dex treatment there are basal levels of DNA looping in FoxO3 gene. Dex905treatment for 1 h, however, induces DNA looping that increases the interaction between906+71kbGBR and -17kbGBR, and between genomic regions near -17kbGBR and TSS.907908 Mouse -17KbGBR GLS1 (-17231 to -17217) (chr 10, 42013789-42013775)AGGTCTTTCTGTTCC [6/8]Human (-16519 to -16533)(chr 6, 108971185-108971199) AGGTCTTTCTGTTCC [6/8]Rat (-12290 to -12276)(chr 20, 46275109-46275095)AGGGCATTCAGTTCC [6/8] Mouse +45kbGBR GLS1 (+45266 to +45252) (chr 10, 41951297-41951283)GTTACATCCTGGACT [6/8]Human (+61314 to +61328)(chr 6, 109049034-109049048) GGCACATTCTGTACT [8/8]Rat (+44026 to +44012)(chr 20, 46218771-46218757)GGTACGTCCTGTACT [7/8] Mouse +71kbGBR GLS1 (+71492 to +71478) (chr 10, 41925071-41925057)GGAATGGAATGTTCC [6/8]Human (+94403 to +94417)(chr 6, 109082123-109082137) GGAATGGAATGTTCT [6/8]Rat (+70364 to +70350)(chr 20, 46192453-46192439)GGAATGGAATGTTCC [6/8]Table 1. The Conservation of GRE in FoxO3 GBR *** The genomic location of mouse GRE and its conserved counterparts in human and rat genome are shown. Based on theconsensus GRE sequence identified from ChIP-seq, RGXACANNNTGTXCY, we looked for sequences that have at least 6identical nucleotides of 8 underlined nucleotides. The number of identical nucleotide (underlined) is shown in brackets. Mousesequences are based on mm9 assembly. Human sequences are based on hg18 assembly, and rat sequences are based on Baylor3.4/rn4 assembly. 00.511.522.533.5 2h 6h 24h 48hFoldinduction(Dex/EtOH)