Research ArticleBIOCHEMISTRY

Condensin II is anchored by TFIIIC and H3K4me3 in the mammalian genome and supports the expression of active dense gene clusters

See allHide authors and affiliations

Science Advances  21 Jun 2017:
Vol. 3, no. 6, e1700191
DOI: 10.1126/sciadv.1700191
  • Fig. 1 Condensin II and TFIIIC complexes associate in mammalian cells.

    (A and B) Subunits of the TFIIIC complex (blue) were pulled down using anti-condensin (SMC2 and SMC4) antibodies (Ab) in nuclear extracts from HEK293T cells for co-IP experiments, followed by immunoblotting. IgG, immunoglobulin G. (C) All subunits of the condensin II complex (magenta) were pulled down using anti–TFIIIC-220 antibodies in co-IP, followed by immunoblotting with the indicated antibody. (D and E) Subunits of the TFIIIC complex were pulled down from nuclear extracts made from cell lines that stably expressed Flag-SMC2 or Flag-NCAPH2, followed by immunoblotting with the indicated antibody. (F) All subunits of the condensin II complex were pulled down with anti-FLAG antibody from a cell line stably expressing Flag-TFIIIC-110, followed by immunoblotting with the indicated antibody. (G) Subunits of the TFIIIC complex were pulled down from nuclear extracts derived from a cell line that stably expressed BirA*-NCAPH2 in a BioID assay. (H) BioID assays with nuclear extracts from a cell line that stably expressed BirA*-TFIIIC-110 had all subunits of the TFIIIC complex. Each experiment was performed in triplicate.

  • Fig. 2 Definition of CTS in the mouse genome.

    (A) NCAPH2 and NCAPD3 were colocalized with TFIIIC binding sites in mESCs based on ChIP-seq data from the present study (NCAPH2) and from Dowen et al. (NCAPD3) (33). (B and C) Knockdown of TFIIIC-90 caused a significant reduction of NCAPH2 binding to tRNA genes. However, knockdown of NCAPH2 did not alter the binding of TFIIIC-220 to tRNA genes. (D) Colocalization of TFIIIC-220 and NCAPH2 formed the basis for the definition of CTS (10,043 sites; 64.7% of total high-confidence TFIIIC sites) and CFTS (4921 sites; 32.9% of total TFIIIC sites). Pol III, RNA polymerase III. (E and F) Knockdown of TFIIIC-90 caused a significant reduction of NCAPH2 binding to CTS, whereas knockdown of NCAPH2 did not change the binding of TFIIIC to CTS. (G) Genome-wide annotation chart of CTS in mESCs. (H) Gene Ontology (GO) chart shows the localization of CTS at promoters of housekeeping genes. FDR, false discovery rate. (I) TFIIIC-220 has higher peak intensity at CTS than at CFTS. 5′UTR, 5′ untranslated region.

  • Fig. 3 Identification of CTS in the human genome.

    (A) NCAPH2 and NCAPD3 were colocalized with TFIIIC binding sites in human HEK293 cells based on ChIP-seq data from the present study. (B) Colocalization of TFIIIC-220 and NCAPH2 formed the basis for the definition of CTS (11,169 sites; 70.7% of total high-confidence TFIIIC sites) and CFTS (4627 sites; 29.3% of total TFIIIC sites). (C) Genome-wide annotation chart of CTS in HEK293 cells. (D) GO chart shows the localization of CTS at promoters of housekeeping genes.

  • Fig. 4 The condensin II complex interacts with H3K4me3.

    (A) H3K4me3 and H3K27ac, but not H3K9me3 or H3K27me3, colocalized with CTS. (B and C) Pull-down assays were performed by mixing nuclear extracts from HEK293T cells with biotinylated histone peptides (HP) modified as indicated. Condensin II and TFIIIC-220 were pulled down with H3K4me3, but not H3K27ac, histone peptides. (D) Pull-down assays were performed by mixing full-length proteins of condensin II generated by in vitro translation with biotinylated histone peptides with and without H3K4me3 modification. NCAPD3 and SMC2 interacted with the H3K4me3 histone peptide. (E and F) Co-IP with anti-SMC2 and anti-NCAPD3 antibodies pulled down chromatin with H3K4me3, but not H3K27ac. (G) HEAT repeat domains from NCAPD3 or NCAPG2 and the C terminus of NCAPD3 were identified and cloned out as described (4). Pull-down assays were performed by mixing HEAT repeat domains from NCAPD3 or NCAPG2 or the C terminus of NCAPD3 with biotinylated H3K4me3 histone peptides. HEAT repeat domains from NCAPD3 interacted with the H3K4me3 peptide. Three independent experiments were performed for each assay.

  • Fig. 5 Condensin II and TFIIIC are colocalized at TAD boundaries.

    (A) CTS, but not CFTS, were enriched at TAD boundaries in mESCs. (B) Twenty-eight percent of CTS were within 50 kb of a TAD boundary in mESCs, in contrast to 13% TAD boundary association of random control peaks by chance (P = 2.1 × 10−16, χ2 test). (C) CTS, but not CFTS, were enriched at TAD boundaries in human HEK293 cells. (D) Representative image from a genome browser showing tracks of in situ Hi-C and ChIP-seq of NCAPH2, TFIIIC-220, and H3K4me3. (E) Fifty-five percent of CTS were within 50 kb of a TAD boundary in mESCs, in contrast to 41% TAD boundary association of random control peaks by chance (P = 2.11 ×10−8, χ2 test). (F) The total number of genes that are significantly down-regulated upon condensin knockdown is 1516 in mESCs. Eighty-five percent (1293) of genes that were significantly down-regulated upon NCAPH2 knockdown were within 50 kb of a TAD boundary. (G) A total of 2415 genes are significantly down-regulated upon condensin knockdown in HEK293 cells. Sixty-five percent (1570) of genes that were significantly down-regulated upon NCAPH2 knockdown were within 50 kb of a TAD boundary.

  • Fig. 6 Condensin II supports interaction and expression of histone gene clusters.

    (A) Contact profiles of control and NCAPH2 knockdown mESCs at histone gene loci measured by 4C-seq. Knockdown of NCAPH2 reduced interactions between two histone gene loci (1.86 ± 0.25–fold change, P = 7.81 × 10−8). The green dotted line indicates the anchor point. (B) Knockdown of NCAPH2 by short hairpin RNA (shRNA) in mESCs significantly reduced the expression of histone genes in the loci shown in (A). Error bars, SEM. (C) Knockdown of NCAPH2 in HEK293 cells by siRNA significantly reduced the expression of histone genes. Error bars, SEM. (D) The size of the histone clusters in NCAPH2 knockdown mESCs is significantly smaller than the sizes of control cells, indicating that the formation of the clusters is disrupted in NCAPH2 knockdown cells. Fluorescent staining was performed with an antibody to NPAT, a protein that labels the histone clusters (31). Error bars, SD. FWHM, full width at half maximum. (E) Working model for how condensin II and TFIIIC complexes tether dense active promoters in the mammalian genome: Condensin II and TFIIIC complexes (CTS) are colocalized at densely clustered active promoters at TAD boundaries. Condensin II may support the expression of the genes within the boundaries. Some TADs associated with the lamina are less transcriptionally active. The lower right inset shows interaction between CTS that are associated with transcriptionally active promoters. The upper right inset depicts details of the CTS, including condensin II (H, NCAPH2; D, NCAPD3; G, NCAPG2), the TFIIIC complex, the B box–like sequence, and the H3K4me3 histone modification. Condensin II may be recruited to the chromatin, in part, by the HEAT domain of NCAPD3 binding to the H3K4me3 histone modification and, in part, by the interaction with TFIIIC.

Supplementary Materials

  • Supplementary material for this article is available at http://advances.sciencemag.org/cgi/content/full/3/6/e1700191/DC1

    fig. S1. TFIIIC is not associated with CTCF and cohesin in mammalian cells, whereas the association of condensin II and TFIIIC complexes is stable.

    fig. S2. Knockdown of NCAPH2 did not affect the binding of TFIIIC-220 to the genome.

    fig. S3. CTS are preferentially localized at the promoter/TSS of highly expressed gene.

    fig. S4. CTS are correlated with the promoters of active genes.

    fig. S5. Binding of architectural proteins, active promoter, and enhancer markers at CTS but not at CFTS.

    fig. S6. CTS have higher CpG islands and GC content than CFTS.

    fig. S7. Down-regulation of highly transcribed genes upon condensin II knockdown in mESC and HEK293.

    fig. S8. Histone peptide assays and IP experiments.

    fig. S9. Effect of reduced H3K4me3 on the binding of NCAPH2 and TFIIIC-220 to CTS.

    fig. S10. Characterization of CTS at TAD boundaries in mESCs.

    fig. S11. Characterization of CTS at TAD boundaries in HEK293 cells.

    fig. S12. CTS are not correlated with LAD or LOCK domains in mESCs.

    fig. S13. Clustered CTS are associated with stronger TAD boundaries in mESCs.

    fig. S14. Interactions between two histone gene loci on chr13.

    fig. S15. Chromatin interactions between two highly expressed gene loci were reduced upon NCAPH2 knockdown.

    fig. S16. Formation of histone clusters may be disrupted in NCAPH2 knockdown mESCs.

    table S1. Public ChIP-seq data sets analyzed.

  • Supplementary Materials

    This PDF file includes:

    • fig. S1. TFIIIC is not associated with CTCF and cohesin in mammalian cells, whereas the association of condensin II and TFIIIC complexes is stable.
    • fig. S2. Knockdown of NCAPH2 did not affect the binding of TFIIIC-220 to the genome.
    • fig. S3. CTS are preferentially localized at the promoter/TSS of highly expressed gene.
    • fig. S4. CTS are correlated with the promoters of active genes.
    • fig. S5. Binding of architectural proteins, active promoter, and enhancer markers at CTS but not at CFTS.
    • fig. S6. CTS have higher CpG islands and GC content than CFTS.
    • fig. S7. Down-regulation of highly transcribed genes upon condensin II knockdown in mESC and HEK293.
    • fig. S8. Histone peptide assays and IP experiments.
    • fig. S9. Effect of reduced H3K4me3 on the binding of NCAPH2 and TFIIIC-220 to CTS.
    • fig. S10. Characterization of CTS at TAD boundaries in mESCs.
    • fig. S11. Characterization of CTS at TAD boundaries in HEK293 cells.
    • fig. S12. CTS are not correlated with LAD or LOCK domains in mESCs.
    • fig. S13. Clustered CTS are associated with stronger TAD boundaries in mESCs.
    • fig. S14. Interactions between two histone gene loci on chr13.
    • fig. S15. Chromatin interactions between two highly expressed gene loci were reduced upon NCAPH2 knockdown.
    • fig. S16. Formation of histone clusters may be disrupted in NCAPH2 knockdown mESCs.
    • table S1. Public ChIP-seq data sets analyzed.

    Download PDF

    Files in this Data Supplement:

Navigate This Article