Skip to main page content
Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
. 2017 Sep 5;20(10):2313-2327.
doi: 10.1016/j.celrep.2017.08.030.

The SET1 Complex Selects Actively Transcribed Target Genes via Multivalent Interaction With CpG Island Chromatin

Affiliations
Free PMC article

The SET1 Complex Selects Actively Transcribed Target Genes via Multivalent Interaction With CpG Island Chromatin

David A Brown et al. Cell Rep. .
Free PMC article

Abstract

Chromatin modifications and the promoter-associated epigenome are important for the regulation of gene expression. However, the mechanisms by which chromatin-modifying complexes are targeted to the appropriate gene promoters in vertebrates and how they influence gene expression have remained poorly defined. Here, using a combination of live-cell imaging and functional genomics, we discover that the vertebrate SET1 complex is targeted to actively transcribed gene promoters through CFP1, which engages in a form of multivalent chromatin reading that involves recognition of non-methylated DNA and histone H3 lysine 4 trimethylation (H3K4me3). CFP1 defines SET1 complex occupancy on chromatin, and its multivalent interactions are required for the SET1 complex to place H3K4me3. In the absence of CFP1, gene expression is perturbed, suggesting that normal targeting and function of the SET1 complex are central to creating an appropriately functioning vertebrate promoter-associated epigenome.

Keywords: CpG island; DNA methylation; H3K4me3; SET1; chromatin; epigenetics; histone; histone methylation; multivalent; transcription.

Figures

None
Figure 1
Figure 1
The Cellular Dynamics of CFP1 Are Governed by Its Chromatin-Binding Domains and Association with the SET1A Complex (A) A schematic illustrating the GFP-tagged versions of CFP1 stably expressed in mouse C127 cells and used in the FRAP studies in (B). These include GFP alone (GFP), wild-type CFP1 (WT), CFP1 with a mutated CXXC domain (CXXC), CFP1 with a mutated PHD domain (PHD), and CFP1 with combined CXXC and PHD mutations (P/C). (B) Biexponential fits describing the recovery of fluorescence intensity over time for each of the proteins described in (A). Fits were calculated using post-bleach fluorescence intensity recovery data collected at 8 frames per second from >42 cells across biological triplicates. (C) A boxplot indicating the half time of recovery (t1/2) in seconds for the FRAP curves in (B). Boxes show interquartile range (IQR) and whiskers extend by 1.5 × IQR. The p values indicating statistically significant differences are indicated above the boxplot. The p value denotes statistical significance using a Student’s t test. (D) A schematic illustrating the GFP-tagged versions of CFP1 and SET1A used in the FRAP studies in (E). This includes GFP alone (GFP), wild-type CFP1 (WT), CFP1 with combined PHD and CXXC mutations (P/C), CFP1 with a mutated SET1 interaction domain (SID), CFP1 with combined PHD/CXXC/SID mutations (P/C/S), and wild-type SET1A. (E) Biexponential fits describing the recovery of fluorescence intensity over time for the proteins described in (D). CFP1 and SET1 fits were calculated using post-bleach fluorescence intensity recovery data collected at 13 frames per second from >28 cells across biological triplicates, whereas the GFP fit was limited to 8 cells, in which the nuclear border was clearly defined. (F) A boxplot indicating the half time of recovery (t1/2) in seconds for the FRAP curves in (E). Boxes show IQR and whiskers extend by 1.5 × IQR. The p values indicating statistically significant differences are indicated above the boxplot. The p value denotes statistical significance using a Student’s t test.
Figure 2
Figure 2
Binding of CFP1 to Chromatin Relies on the CXXC and PHD Domains but Not Interaction with SET1 (A) A genomic snapshot of the NMI-associated Orai1 gene promoter showing the signal from ATAC, Bio-CAP, H3K4me3, RNA Pol II, RNA, and CFP1 sequencing experiments. (B) Heatmaps of the sequencing signals in (A) ranked by Bio-CAP signal over all NMIs. (C) A Venn diagram illustrating the overlap between CFP1 peaks and NMIs. (D) Bar graph illustrating the percentage of CFP1-bound and -unbound NMIs that overlap with transcription start sites (TSSs). (E) Boxplots illustrating the enrichment of H3K4me3 at TSS-associated NMIs that are bound (CFP1+) or unbound (CFP1−) by CFP1. Boxes show IQR and whiskers extend by 1.5 × IQR. The p value denotes statistical significance calculated by a Wilcoxon signed rank test. (F) As in (E), illustrating enrichment of RNA PolII. (G) A genomic snapshot of the NMI-associated Orai1 gene promoter showing the GFP ChIP-seq signal for GFP, GFP-CFP1, and mutated forms of GFP-CFP1. (H) Metaplot and heatmap analysis of GFP ChIP-seq signal for GFP, GFP-CFP1, and mutated forms of GFP-CFP1 over all NMIs. (I) Heatmap analysis of GFP ChIP-seq signal over all NMIs for the same cell lines as in (H).
Figure 3
Figure 3
Multivalent Binding to Non-methylated DNA and H3K4me3 Determines the Occupancy of CFP1 on Chromatin In Vivo (A) Heatmap analysis of Bio-CAP and ChIP-seq signal over all NMIs for the indicated endogenous or GFP fusion proteins. The intensity scale for GFP-CFP1 and GFP-CFP1-PHD mutant are indicated below. (B) The relative enrichment of the features heatmapped at NMIs in (A) was plotted across H3K4me3 enrichment percentiles. (C) 1H,15N heteronuclear single quantum coherence titration experiments using the CFP1 PHD domain and an H3K4me3 peptide. (D) As in (C) for an unmodified H3K4me0 peptide. (E–G) Measurements using intrinsic fluorescence spectroscopy for the PHD domain binding to (E) H3K4me3, (F) H3K4me2, and (G) H3K4me1. (H) A table illustrating the quantitative measurements of binding affinities for the CFP1 PHD domain bound to H3K4 substrates, as determined by intrinsic fluorescence spectroscopy and NMR spectroscopy.
Figure 4
Figure 4
CFP1 Is the Central Determinant in SET1A Occupancy on Chromatin (A) A schematic illustrating the Cfp1fl/fl mouse ESC model, in which the addition of tamoxifen (TAM) leads to deletion of CFP1. (B) Western blot analysis of CFP1 following a time course of TAM treatment. (C) A genomic snapshot illustrating CFP1 ChIP-seq signal at the Ube2j2 gene in untreated (upper panel) and tamoxifen-treated cells (lower panel). (D) Metaplot analysis of CFP1 ChIP-seq signal at CFP1-bound NMI-associated TSSs in untreated (solid line) and tamoxifen-treated cells (dashed line). p values denote statistical significant calculated by a Wilcoxon signed rank test comparing read counts across the represented interval in UNT versus TAM. (E) A genomic snapshot illustrating SET1A ChIP-seq signal at the at Ube2j2 gene in untreated (upper panel) and tamoxifen-treated Cfp1fl/fl ESCs (lower panel). (F) Metaplot analysis of T7-SET1A ChIP-seq signal at CFP1-bound NMI-associated TSSs in untreated (solid blue line) and tamoxifen-treated Cfp1fl/fl ESCs (dashed blue line). The solid black line illustrates ChIP-seq signal for the T7 antibody in an untagged cell line (Mock). p values denote statistical significance calculated by a Wilcoxon signed rank test comparing read counts across the represented interval in UNT versus TAM. (G) A scatterplot of the SET1A and CFP1 ChIP-seq signal at TSSs. R value indicates Spearman rank correlation. Genes right of the dashed line correspond to CFP1-bound (CFP1+) sites. (H) A scatterplot of the log2-fold change in SET1A ChIP-seq signal compared to CFP1 ChIP-seq signal at TSSs. R value indicates Spearman rank correlation. Genes right of the dotted line correspond to CFP1-bound (CFP1+) sites. (I) Genomic snapshots illustrating a gene where SET1A ChIP-seq signal is lost following removal of CFP1 (left panel) and a gene that is more highly transcribed and retains some SET1A (right panel) following removal of CFP1. (J) Boxplots illustrating the log2-fold change in SET1A ChIP-seq signal (left panel) and expression level based on 4SU-RNA-seq (right panel) of the top 10% of genes that lose or retain SET1A. Boxes show IQR and whiskers extend by 1.5 × IQR. The p value denotes statistical significance calculated by the Mann-Whitney test.
Figure 5
Figure 5
CFP1 Exploits Multivalent Interactions with CpG Island Chromatin to Shape H3K4me3 (A) Western blot analysis of bulk H3K4me in untreated (UNT) and TAM-treated Cfp1fl/fl mouse ESCs. (B) Genomic snapshot illustrating H3K4me3 ChIP-seq signal at the Ap3d1 NMI, which is bound by CFP1 (left panel) and the Smoc2 NMI that is not bound by CFP1 (right panel). The H3K4me3 ChIP-seq signal without tamoxifen treatment is represented in blue (UNT) and the signal with tamoxifen treatment is represented in black (TAM) in the overlay. (C) H3K4me3 signal around CFP1+ (left) and CFP1− (right) NMI-associated TSSs in untreated (solid line) and tamoxifen-treated cells (dashed line). (D) A scatterplot illustrating that highly (left panel) but not lowly (right panel) expressed genes lose H3K4me3 at their TSS following tamoxifen treatment. The scatterplots correspond to non-divergent genes with an H3K4me3 peak overlapping their TSS, and a gene was considered to be lowly expressed if it had less than −2.5 log2 FPKM nuclear RNA seq signal over the gene body. (E) Genomic snapshot illustrating H3K4me3 ChIP-seq signal at the Hexim1 NMI in UNT and TAM-treated cells. The UNT and TAM-treated samples are colored blue and black in the overlay, respectively. The upper panel corresponds to the parental Cfp1fl/fl line, with the lower panels corresponding to the parental line rescued with at GFP, GFP-CFP1, GFP-P/C, or GFP-SID transgenes. (F) Metaplot and boxplot analysis of the H3K4me3 signal at TSSs as described in (C) for cell lines described in (E) without and with tamoxifen treatment. (G) Boxplot analysis of H3K4me3 signal at CFP1+ NMI-associated TSSs +/- 1 kb. Boxes show IQR and whiskers extend by 1.5 × IQR.
Figure 6
Figure 6
Loss of CFP1 Leads to Widespread Effects on Gene Expression (A) An MA plot showing log2-fold change in the nuclear RNA-seq signal of CFP1-bound NMI-associated genes in UNT and TAM-treated Cfp1fl/fl cells. Red and green points depict significantly downregulated (1,108) and upregulated genes (584) that change in expression by more than 1.5-fold. (B) Boxplots indicating the expression of genes that are downregulated (red), not significantly (N.S.) changing (gray), or upregulated (green). (C–E) Mean distribution of CFP1 ChIP-seq signal (C), CpG density (D), and SET1A ChIP-seq signal (E) around TSSs of downregulated (red) and upregulated (green) genes. p values denote statistical significance calculated by Mann-Whitney test comparing ChIP-seq read counts across a 200-bp interval flanking the TSS in downregulated versus upregulated genes. (F) Correlation density plot of changes in gene expression (nucRNA-seq) and H3K4me3 at TSSs of CFP1-bound NMI genes. Only genes whose TSSs overlap an H3K4me3 peak and do not have a divergent TSS within 2 kb were considered. R value indicates Spearman rank correlation.

Similar articles

See all similar articles

Cited by 17 articles

See all "Cited by" articles

References

    1. Andreu-Vieyra C.V., Chen R., Agno J.E., Glaser S., Anastassiadis K., Stewart A.F., Matzuk M.M. MLL2 is required in oocytes for bulk histone 3 lysine 4 trimethylation and transcriptional silencing. PLoS Biol. 2010;8 - PMC - PubMed
    1. Ardehali M.B., Mei A., Zobeck K.L., Caron M., Lis J.T., Kusch T. Drosophila Set1 is the major histone H3 lysine 4 trimethyltransferase with role in transcription. EMBO J. 2011;30:2817–2828. - PMC - PubMed
    1. Austenaa L.M., Barozzi I., Simonatto M., Masella S., Della Chiara G., Ghisletti S., Curina A., de Wit E., Bouwman B.A., de Pretis S. Transcription of mammalian cis-regulatory elements is restrained by actively enforced early termination. Mol. Cell. 2015;60:460–474. - PubMed
    1. Bernstein B.E., Kamal M., Lindblad-Toh K., Bekiranov S., Bailey D.K., Huebert D.J., McMahon S., Karlsson E.K., Kulbokas E.J., 3rd, Gingeras T.R. Genomic maps and comparative analysis of histone modifications in human and mouse. Cell. 2005;120:169–181. - PubMed
    1. Birke M., Schreiner S., García-Cuéllar M.P., Mahr K., Titgemeyer F., Slany R.K. The MT domain of the proto-oncoprotein MLL binds to CpG-containing DNA and discriminates against methylation. Nucleic Acids Res. 2002;30:958–965. - PMC - PubMed

LinkOut - more resources

Feedback