Figure 1.
System for the analysis of chromatin changes during glucocorticoid treatment.
(A) Experimental schematic. (B) Example of PAGE of chromatin fragments used to isolated mononucleosomal fragments (as well as altosomal fragments analyzed in Fig. S15). Complete gels and quantitation are shown in Fig. S1. (C) RT-PCR for mRNA levels of CD44 (control), GEM (Dex repressed) and SGK1 (Dex activated) endogenous genes. Results show the signal at the indicated time with Dex divided by the signal without Dex. (D) Western blot versus BRG1, using anti-actin as a control, shows transfection with BRG1 specific siRNA reduces BRG1 protein levels. (E) MMTV luciferase transactivation by Dex, as measured by luminometer, is attenuated in BRG1 siRNA transfected cells.
Figure 2.
Hormone addition results in chromatin changes at individual GR-regulated genes.
UL3 cells were transfected with BRG1-specific or control siRNA oligos, and treated, after 71 hours with 10 nM Dex (or ethanol vehicle) for 1 hr. Alternatively, untransfected cells were treated with Dex for 4 hrs. (A) Quantile normalized nucleosome/bare ratios for the MMTV LTR promoter region from the integrated MMTV-luciferase construct for No Dex (blue line), Dex 1 hr (red line) or Dex 4 hrs (green line). Blue bars: approximate locations of nucleosome boundaries from indirect end labeling experiments [19], [48]. Blue & green ovals: nucleosome positions for No Dex or Dex 4 hrs. Arrows: primer sets used to detect Nucs B, D & F for Fig. 2B. We calculate that a nucleosome/bare ratio of 1.0 corresponds to ∼50% nucleosome occupancy, while a value of 2.1 corresponds to 100% occupancy (e.g. a nucleosome covering a given location on all gene copies). Because a nucleosomal fragment must cover the full length of each oligo to give a strong signal, well positioned nucleosomes are represented by signal peaks greater than 1.0 and spanning only about 9 oligos (∼90 bp), an effect which also increases the apparent width of nucleosome-free linker regions. For more details, see Additional Methods, in Text S1. (B) MNase footprint PCR results for MMTV, using the primer sets indicated in (A). The PCR signal from mononucleosomal fragments at the indicated positions and times after Dex addition was normalized by dividing by the NoDex signal, and the results from both pairs of primers corresponding to each position averaged. Bars indicate standard error of the mean from 6 PCR reactions. In addition to the progressive decrease at Nuc D at 20 mins, 1 hr and especially 4 hrs, moderately decreased signal at Nucs D and F was also seen at 4 hrs (consistent, perhaps, with the somewhat lower occupancy, especially for Nuc D, seen in Fig. 2A). (C) As for (A), but showing the nucleosome profile for the Dex-repressed endogenous MYC promoter. Ovals indicate nucleosome positions mapped using indirect-end labeling in human HL60 cells [51], [52], [53]. The light blue line and squares shows array mapping results from A375 cells ([6], GEO # GSE6385). Arrows show the minor (P1) and major (P2) MYC promoters. (D) As for (A), but showing results for the Dex-induced endogenous HSD11B2 promoter. Grey arrows indicate sites where nucleosome occupancy increases or decreases after 1 hr of dex treatment. (E) Comparison of No Dex and Dex 1 hr experimental results from two separate experiments. Conditions for the second experiment were identical, except that cells were not transfected with control siRNA oligos before hormone treatment. Arrows indicate sites of statistically significant nucleosome occupancy changes. We found that almost all locations that met this criterion for experiment #1 (e.g. dark blue and red curves at arrows) showed an occupancy difference in the same direction in experiment #2 (e.g. blue-green and orange curves at arrows). (F) No Dex and Dex 1 hr treatments in BRG1 knock down cells (dark blue and dark red lines). The data from control transfected cells (from (D)) is shown as stippled lines.
Figure 3.
Average nucleosome occupancy rises transiently +Dex surrounding TSSes of GR activated and GR repressed genes.
(A) & (B) Average nucleosome density relative to the transcription start sites of the four Dex repressed genes on the array: GEM, PLK2, POMC & MYC. (C) & (D) Average nucleosome density vs. TSSes for nine of the GR-activated endogenous genes on the array: HSD11B2, TSC22D3, PCK1, ZBTB16, SDPR, SGK1, SLC19A2, CYP3A4 & SRGN. MMTV was excluded from this analysis to avoid contributions of non-MMTV vector and luciferase sequences upstream of ∼−1200 and downstream of ∼+100). (A) & (C) Comparison of UL3 cells treated with No Dex, Dex 1 hr and Dex 4 hr, as indicated. (B) & (D) Comparison of No Dex and Dex 1 hr conditions in BRG1 knock down and siRNA control cells.
Figure 4.
1 hr Dex treatment increases nucleosome occupancy at GR-unregulated promoters.
(A) & (B) As per Fig. 3, but showing average normalized nucl/bare ratios for all GR-independent genes on the array (CCNA1, CCNB2, CCND1, CCNE1, CD44, CDK1, CDKN1A, CSF1, E2F1, GAPDH, UGT1A6 & UGT1A8). (C) –Dex or +Dex 1 hr results from biological replicate samples (dotted lines) and average from both samples (solid lines).
Figure 5.
Proximal GRBSes are associated with a lesser increase in nucleosome occupancy after Dex.
Nucleosome density relative to TSSes was plotted for either: (A) all genes that contain a GR-binding site within 500 bp upstream of the TSS (top panel, SDPR, SLC19A2, SRGN, GEM & POMC – MMTV was excluded because of the presence of luciferase and vector sequences, and lack of knowledge of the insertion site), or (B) all genes that lack a promoter proximal GR-binding site, but which have a GR-binding site >1 kb upstream of the TSS (bottom panel, HSD11B2, SGK1, TSC22D3 & PLK2).
Figure 6.
Genomewide increase in promoter nucleosome occupancy upon activation of human CD4+ T cells.
Sequence reads from fifty Solexa lanes for resting versus activated human CD4+ T cells [5] were mapped to the human genome, and relative nucleosome frequency versus TSSes determined for genes at least 2-fold induced or at least 2-fold repressed by activation (A), or for genes whose expression changed less than 1.5-fold as a result of activation (B).
Figure 7.
HL60 differentiation to granulocytes is associated with increased promoter nucleosome occupancy.
Average quantile-normalized nucleosome occupancy for control HL60 cells versus DMSO-differentiated HL60 cells relative to (A) the TSSes of all genes on the array that were significantly up- or down-regulated by 5-day DMSO treatment (up: MYC, CCNE1, CCND1, CCNB2, CD44 & ZBTB16, down: GAPDH & SGK1, [59]), or (B) all genes on the array that were not regulated by DMSO treatment (showing expression levels well above background and less than a 1.3-fold change with DMSO treatment: CDK1, E2F1, CSF1, SRGN & TSC22D3, [59]). A similar effect was seen for both up- and downregulated genes, when analyzed separately (data not shown). (C) Comparison of average nucleosome densities for control conditions (–Dex for UL3s, –DMSO for HL60s) and after stimulation (+Dex 1 hr for UL3s, +DMSO 5 days for HL60s). Plots showing that increased occupancy +DMSO is only seen for promoter regions (at any given nucl/bare ratio in –DMSO samples) are shown in Fig. S3C & D.