The N-CoR complex enables chromatin remodeler SNF2H to enhance repression by thyroid hormone receptor

Abstract

Unliganded thyroid hormone receptor (TR) actively represses transcription via the nuclear receptor corepressor (N-CoR)/histone deacetylase 3 (HDAC3) complex. Although transcriptional activation by liganded receptors involves chromatin remodeling, the role of ATP-dependent remodeling in receptor-mediated repression is unknown. Here we report that SNF2H, the mammalian ISWI chromatin remodeling ATPase, is critical for repression of a genomically integrated, TR-regulated reporter gene. N-CoR and HDAC3 are both required for recruitment of SNF2H to the repressed gene. SNF2H does not interact directly with the N-CoR/HDAC3 complex, but binds to unacetylated histone H4 tails, suggesting that deacetylase activity of the corepressor complex is critical to SNF2H function. Indeed, HDAC3 as well as SNF2H are required for nucleosomal organization on the TR target gene. Consistent with these findings, reduction of SNF2H induces expression of an endogenous TR-regulated gene, dio1, in liver cells. Thus, although not apparent from studies of transiently transfected reporter genes, gene repression by TR involves the targeting of chromatin remodeling factors to repressed genes by the HDAC activity of nuclear receptor corepressors.

Figure 1

GalTR/UAS-Lucif model system requires N-CoR for inhibiting transcription in unliganded state. (A) T3-dependent luciferase activity of a 293T cell line with stably integrated (UASx5)-sv40 promoter-luciferase reporter gene and Gal4-TRβ fusion receptor (GalTR/UAS-Lucif). Cells were incubated with T3 (final concentration 10 nM) for 24 h (B) ChIP analysis at the integrated gene in GalTR/UAS-Lucif cells in the absence or presence of T3 (10 nM, 24 h). Samples were analyzed by real-time PCR, normalized to the 36B4 gene, and plotted as fold change relative to no ligand (−T3). ^*P<0.05 versus no ligand. (C) Immunoblot following transfection of cells with siRNA for N-CoR. (D) Transcription assay following treatment with siRNA for N-CoR in control UAS-Lucif cells (lack receptor) and GalTR/UAS-Lucif cells. Luciferase activity was normalized to β-gal expression and then plotted. ^*P<0.05 versus siRNA control.

Figure 2

SNF2H mediates repression of a stably integrated TR regulated gene. (A) Immunoblot following transfection of cells with siRNA for SNF2H. (B) Transcription assay following treatment with siRNA for SNF2H in control UAS-Lucif cells (lack receptor) and GalTR/UAS-Lucif cells. Luciferase activity was normalized to β-gal expression and then plotted. ^*P<0.05 versus siRNA control. (C, D) Transcription assay in transient reporter system. The (UASx5)-sv40 promoter-luciferase reporter was transiently transfected along with Gal4-DBD or Gal4-TR into 293T cells treated with siRNA for (C) SNF2H or (D) N-CoR. Luciferase activity was normalized to β-gal expression and plotted as fold repression relative to Gal4-DBD.

Figure 3

N-CoR is required for SNF2H recruitment. ChIP analysis at the integrated gene in GalTR/UAS-Lucif cells following treatment with control or N-CoR siRNA. Samples were analyzed by real-time PCR, normalized to the 36B4 gene, and plotted as fold change relative to control siRNA treatment. ^*P<0.05 versus siRNA control (n=3).

Figure 4

SNF2H does not bind N-CoR/HDAC3 complex. (A) 293T cells were transfected with pCMX-Flag-N-CoR and whole-cell lysates were immunoprecipitated with anti-Flag agarose beads. Bound proteins were separated by SDS–PAGE and immunoblotted with Flag, SNF2H or HDAC3 antibodies. Input is 1%. (B) Immunoprecipitation of 293T whole-cell lysates was performed with normal rabbit IgG or rabbit SNF2H antibody and followed by immunoblot analysis. Input is 3%.

Figure 5

SNF2H preferentially binds unacetylated histone H4 tails. (A) Immunoprecipitation of 293T whole-cell lysates was performed with normal rabbit IgG or rabbit SNF2H antibody and followed by immunoblot analysis for SNF2H or histone H4. Input is 3%. (B) Recombinant Flag-SNF2H was incubated with bead-bound biotinylated unacetylated histone H4 peptide (H4 peptide) or tetra-acetylated H4 peptide (Ac-H4 peptide). Beads were washed and bound protein was separated by SDS–PAGE and immunoblotted. Input is 10%.

Figure 6

HDAC3 is required for SNF2H recruitment. (A) Immunoblot following transfection of cells with siRNA for HDAC3. (B) ChIP analysis at the integrated gene in GalTR/UAS-Lucif cells following treatment with control or HDAC3 siRNA. Samples were analyzed by real-time PCR, normalized to the 36B4 gene, and plotted as fold change relative to control siRNA treatment. ^*P<0.05 versus siRNA control (n=3).

Figure 7

SNF2H and HDAC3 regulate chromatin architecture of a TR-repressed gene. (A) Nuclei from GalTR/UAS-Lucif cells were collected following treatment with control and (B) SNF2H siRNA or (C) HDAC3 siRNA and digested with MNase. After DNA purification, complete digestion to mononucleosome size was confirmed by agarose gel electrophoresis. Samples were analyzed by real-time PCR using primers that amplified a 90 bp region in the sv40 promoter (black arrows) or a 120 bp region containing the luciferase start site (white arrows). Results were normalized to the 36B4 gene and plotted as fold change relative to control siRNA treatment. ^**P<0.01 versus siRNA control (n=3).

Figure 8

SNF2H represses endogenous dio1 gene and recruitment requires HDAC activity. (A) Immunoblot following electroporation of HepG2 cells with siRNA for SNF2H. (B) Reverse transcriptase analysis of dio1 and 36B4 mRNA expression in HepG2 cells following treatment with control or SNF2H siRNA. Samples were analyzed by real-time PCR, normalized to Gapdh, and plotted as fold change relative to control siRNA treatment. ^*P<0.05 versus siRNA control (n=3). (C) Immunoblot following incubation of HepG2 cells with vehicle (water) or sodium butyrate (5 mM, 5 h). (D) ChIP analysis at the hdio1 promoter in HepG2 cells following treatment with sodium butyrate (5 mM, 5 h). Samples were analyzed by real-time PCR, normalized to the 36B4 gene, and plotted as fold change relative to vehicle treatment. ^*P<0.05 versus vehicle (n=3).

Figure 9

The N-CoR/HDAC3 corepressor couples histone deacetylation to SNF2H-dependent chromatin remodeling in NR repression. The N-CoR/HDAC3 complex associates with unliganded NR and deacetylates local histones, resulting in active repression of the gene. Deacetylation of histone tails also permits stable recruitment of the ATP-dependent chromatin remodeler SNF2H, which can further modify chromatin, possibly by sliding nucleosomes as depicted in the model.