Skip to main page content
Access keys NCBI Homepage MyNCBI Homepage Main Content Main Navigation
, 65 (6), 1081-1095.e5

ERK-Induced Activation of TCF Family of SRF Cofactors Initiates a Chromatin Modification Cascade Associated With Transcription

Affiliations

ERK-Induced Activation of TCF Family of SRF Cofactors Initiates a Chromatin Modification Cascade Associated With Transcription

Cyril Esnault et al. Mol Cell.

Abstract

We investigated the relationship among ERK signaling, histone modifications, and transcription factor activity, focusing on the ERK-regulated ternary complex factor family of SRF partner proteins. In MEFs, activation of ERK by TPA stimulation induced a common pattern of H3K9acS10ph, H4K16ac, H3K27ac, H3K9acK14ac, and H3K4me3 at hundreds of transcription start site (TSS) regions and remote regulatory sites. The magnitude of the increase in histone modification correlated well with changes in transcription. H3K9acS10ph preceded the other modifications. Most induced changes were TCF dependent, but TCF-independent TSSs exhibited the same hierarchy, indicating that it reflects gene activation per se. Studies with TCF Elk-1 mutants showed that TCF-dependent ERK-induced histone modifications required Elk-1 to be phosphorylated and competent to activate transcription. Analysis of direct TCF-SRF target genes and chromatin modifiers confirmed this and showed that H3S10ph required only Elk-1 phosphorylation. Induction of histone modifications following ERK stimulation is thus directed by transcription factor activation and transcription.

Keywords: ERK; Elk-1; H3 phosphorylation; SRF; chromatin; histone modification; immediate-early genes; ternary complex factor; transcription.

Figures

None
Figure 1
Figure 1
Hierarchy of TPA-Induced Histone Modifications at TSS Regions in MEFs Histone modifications at annotated TSSs were assessed by ChIP-seq using antibodies H3K9acS10ph, H4K16ac; H3K27ac, H3K9acK14ac, and H3K4me3. (A) The heatmap shows changes in ChIP-seq signal across at 2,364 TSS regions (−2 to +1 kb), where at least one modification shows a statistically significant change according to DESeq (among ∼12,000 with detectable ChIP-seq signal). Ranking is by average fold change. (B) Dendrogram plot for the five histone modifications (distance method “euclidean,” nesting method “single”). (C) Average difference in fold induction per compared pair of histone modifications (average of xi − yi, for xi > 1). (D) Metaprofiles for the 2,364 TSS regions, grouped by change in H3K9acS10ph (low, 1–1.5× [n = 1,093]; medium, 1.5–2× [n = 598]; high, ≥2× [n = 216]) and TSS regions exhibiting unchanged histone modifications (≥100 reads across each TSS; n = 5,546). ChIP-seq read counts extend 3 kb either side of the TSS. 150 bp on each side of the TSS, where low read counts presumably reflect nucleosome depletion, were excluded. Shaded areas denote difference between the TPA-induced and uninduced levels. (E) Differential profiles (TPA −0.3% FCS) per base at TSS regions displaying >2× increased H3K9acS10ph. For other TSS region classes, see Figure S2C. (F) Boxplot representation of average fold induction of precursor RNA (intronic RNA-seq reads) of the TSS-associated genes as a function of H3K9acS10ph fold change (middle line, median; top and bottom edges, 75th and 25th percentiles). For data summary, see Tables S1 and S2.
Figure 2
Figure 2
Most TPA-Induced Histone Modifications Are TCF Dependent Histone modifications at annotated TSSs were assessed in wild-type and TKO MEFs as in Figure 1. (A) Comparison of ChIP-seq signal in wild-type and TKO MEFs (colored and white boxes, respectively) at the 2,060 TCF-dependent TSS regions. Middle line, median; top and bottom edges, 75th and 25th percentiles; horizontal bars, 90th and 10th percentiles. Statistical significance by Wilcoxon matched-pairs signed rank test (∗∗∗∗p < 0.0001). (B) TCF dependence of TPA-induced changes in histone modifications. (C) Domain structure of Elk-1 (Buchwalter et al., 2004): in Elk-1nonA, all ERK sites in the activation domain (TAD) are substituted by alanine; Elk-1ΔFW lacks the FW motif required for Mediator recruitment (Balamotis et al., 2009). (D) Proportion of TSS regions where expression of wild-type Elk-1 restored regulated histone modifications in TKO MEFs. (E) Differential profiles, as in Figure 1E, of ChIP-seq signals across the 1,248 TSS regions where regulated histone modifications is restored by Elk-1. Left: profiles in wild-type (red) and TKO MEFs (black); right: differential profiles in TKO MEFs reconstituted with wild-type Elk-1 (red), Elk-1ΔFW (blue), Elk-1nonA (purple), or pMY vector (black) (Gualdrini et al., 2016). See Figure S4 for full data.
Figure 3
Figure 3
Signal-Induced Histone Modifications at TSSs Are Not Transcription Factor or Promoter Specific (A) Definition of SRF-linked TSSs exhibiting TCF-dependent TPA-induced histone modifications. TSSs are classified by the closest SRF ChIP-seq peak, with TSSs displaying Hi-C interaction with SRF shaded in red. “Direct” TCF-SRF controlled TSSs (n = 817) are defined as all those within 10 kb of an SRF site or that interact in Hi-C. (B) Hierarchy of TPA-induced histone modification changes at all TCF-dependent TSS regions (top) and direct TCF-dependent TSSs (bottom) displayed as in Figure 1B (see also Figure S2A). (C). Heatmap representation of the ChIP-seq signals at the 817 direct TCF-SRF controlled TSS regions, ranked by average fold change, compared with TPA-induced change in RNA synthesis. (D) Hierarchy of TPA-induced histone modification changes at the 304 TCF-independent TSS regions, with metaprofiles of the modifications, plotted as in Figure 1D, shown below. (E) Left: classification of the 2,404 DNase I HS sites showing significant changes in ChIP-seq signals. Gray, sites within 2 kb of a TSS; orange, remote sites with SRF ChIP-seq peaks; blue, others. Sequence motifs enriched at remote DNase I HS are shown at the right. (F) Differential metaprofiles of induced histone modifications at remote DNase I HS sites in wild-type MEFs; shading indicates change upon induction. (G) Differential metaprofiles of histone modification changes at the remote DNase I HS sites in wild-type (red line) and TKO (black lines) MEFs.
Figure 4
Figure 4
Ordered Histone Modification and Transcriptional Initiation Events at TPA-Induced Genes (A) Top: time course of H3K9acS10ph and H3K4me3 antibody ChIP-seq signal at TPA-inducible TSSs, ranked as in Figure 1A. Bottom: frequency distribution of fold change. (B) Time course metaprofiles (gray, resting; black, 5 min; blue, 15 min; red, 30 min) and differential profiles of H3K9acS10ph and H3K4me3 signal at TSSs where H3K9acS10ph is increased by 30 min TPA stimulation (n = 1,907). (C) Time course metaprofiles and differential profiles of H3S10ph, H3K9acS10ph, and H3K4me3 at remote SRF-positive and SRF-negative DNase I HS sites from Figure 3E, displayed as in (A). (D) Left: TCF direct target genes Egr1, Egr3, and Fosl1; red bars, SRF/TCF binding sites; black bars, PCR probe locations. Right: quantitative ChIP time courses of total H3, H3S10ph, H3K9acS10ph, H3K4me3, MED1, CDK9, and PolII normalized to the maximum detected signal (red dotted line) at the TSSs of TCF targets Egr1, Egr3, and Fosl1 and the TCF-independent TPA-induced TSSs Tpt1, Wsb1, and Spag9. Data are mean ± SEM; n = 4. See Figures S5A and S5B.
Figure 5
Figure 5
Function of the Elk-1 Transcriptional Activation Domain in Signal-Induced Histone Modifications at Egr1 (A) Quantitative ChIP at Egr1 with the indicated antibodies. Dashed lines, unstimulated cells; solid lines, TPA-stimulated cells; red, wild-type MEFs; black, TKO MEFs. Histone ChIP signals are normalized to total H3 (see Figure S5C). (B) Quantitative ChIP analysis of histone modification and transcriptional machinery recruitment at Egr1 in TKO MEFs reconstituted with wild-type Elk-1 (red), Elk-1FW (blue), Elk-1nonA (purple), or with the empty pMY vector control (black). (C) Signal-induced changes in H3S10ph and H3K9acS10ph at the 5′ flanking and transcribed sequences of Egr1, Egr2, Fos, and Ier2 in TKO MEFs reconstituted as in (B). Data are means ± SEM, n = 3. p < 0.05 by t test compared with TKO MEFs with empty vector. See also Figures S5D and S5E.
Figure 6
Figure 6
siRNA Screen Defines an Ordered Series of TCF-Dependent Chromatin Steps Required for Egr1 and Fos Activation (A) Left: siRNA screening strategy, with number of candidate hits remaining at each stage indicated. Right: summary of siRNA effects on transcription or ChIP signals at the 5′-flanking or TSS-proximal region of Egr1 (one-way ANOVA, corrected for multiple comparisons: ∗∗∗p < 0.001, ∗∗p < 0.01, p < 0.05; ns, not significant). (B) Quantitative ChIP at Egr1 with the indicated antibodies following depletion of MLL3 (red), CHD2 (blue), AURKB (purple), and KMT3C (green) or scrambled oligonucleotide control (black). Histone ChIP signals were normalized to H3. (C) Quantitative ChIP at Egr1, Egr3, and Fosl1 analyzed using 5′-flanking or TSS-proximal PCR probes (see Figure 4A) following depletion of MLL3 (red), CHD2 (blue), AURKB (purple), and KMT3C (green) or scrambled oligonucleotide control (black). Histone ChIP signals were normalized to H3. Data are mean ± SEM; n = 3.
Figure 7
Figure 7
Acute Inhibition of AURKB and KMT3C Inhibits IE Gene Transcription (A) Quantitative ChIP with the indicated antibodies at Egr1, Egr3, and Fosl1 analyzed using 5′-flanking or TSS-proximal PCR probes (see Figure 4A) upon TPA stimulation with 250 μM hesperadin (purple), 1 μM LLY-507 (green), or DMSO vehicle (black). Histone signals were normalized to H3. Dashed red line, TPA-induced level; solid yellow line, resting level in vehicle-treated cells. Data are mean ± SEM; n = 4. (B) Signal-induced transcription factor activation and histone modification. Dotted arrows indicate that inducible histone modifications depend on signal reception and recruitment of the transcription machinery by TCF Elk-1. (C) Signaling and activation of direct TCF-SRF target genes. Left: role of TCFs and chromatin modifiers. MLL3 and CHD2 are placed above AURKB, as they affect both basal and induced levels of H3K4me3. Right: sequence of events at the promoter. Box outlines indicate dependence on TCF functions; shading indicates dependence on chromatin modifiers. We did not assess the dependence of H4K16ac, H3K27ac, or H3K9acK14ac on CHD2 and MLL3. AURKB is an H3S10 kinase but may act indirectly, via a factor required both for S10ph and the other modifications. Similar results were obtained at Egr1, Egr3, and Fosl1, although Mediator recruitment was not detected on Fosl1 and was dependent on KMT3C at Egr3.

Similar articles

See all similar articles

Cited by 7 articles

See all "Cited by" articles

References

    1. Anders S., Huber W. Differential expression analysis for sequence count data. Genome Biol. 2010;11:R106. - PMC - PubMed
    1. Anest V., Hanson J.L., Cogswell P.C., Steinbrecher K.A., Strahl B.D., Baldwin A.S. A nucleosomal function for IkappaB kinase-alpha in NF-kappaB-dependent gene expression. Nature. 2003;423:659–663. - PubMed
    1. Aoki K., Kumagai Y., Sakurai A., Komatsu N., Fujita Y., Shionyu C., Matsuda M. Stochastic ERK activation induced by noise and cell-to-cell propagation regulates cell density-dependent proliferation. Mol. Cell. 2013;52:529–540. - PubMed
    1. Balamotis M.A., Pennella M.A., Stevens J.L., Wasylyk B., Belmont A.S., Berk A.J. Complexity in transcription control at the activation domain-mediator interface. Sci. Signal. 2009;2:ra20. - PMC - PubMed
    1. Benayoun B.A., Pollina E.A., Ucar D., Mahmoudi S., Karra K., Wong E.D., Devarajan K., Daugherty A.C., Kundaje A.B., Mancini E. H3K4me3 breadth is linked to cell identity and transcriptional consistency. Cell. 2014;158:673–688. - PMC - PubMed

MeSH terms

Feedback