Article Images (0)
×![]()
featured_play_listSave article to folder
X
Create new folder
Full Text
PDF
Table of Contents
Abstract
EP400 Deposits H3.3 Into Promoters And Enhancers During Gene Activation
Abstract
Introduction
Results
Discussion
Experimental Procedures
Protein Purification
In Vitro Transcription And Immobilized Template (IT) Assays
ChIP Analysis In The Tet-VP16 U2OS Cell Line
SiRNA Knockdown Of EP400
Chromatin Immunoprecipitation, Library Preparation And Analysis
RNA Extraction, MRNA-Seq Library Preparation And Analysis
In Vitro Histone Exchange Assays
Acknowledgments
Footnotes
Abstract
Gene activation in metazoans is accompanied by the presence of histone variants H2AZ and H3.3 within promoters and enhancers. It is not known however what protein deposits H3.3 into chromatin or whether variant chromatin plays a direct role in gene activation. Here we show that chromatin containing acetylated H2AZ and H3.3 stimulates transcription in vitro. Analysis of the Pol II pre-initiation complex on immobilized chromatin templates revealed that the E1A Binding Protein p400 (EP400) was bound preferentially to and required for transcription stimulation by acetylated double-variant chromatin. EP400 also stimulated H2AZ/H3.3 deposition into promoters and enhancers and influenced transcription in vivo at a step downstream of the Mediator complex. EP400 efficiently exchanged recombinant histones H2A and H3.1 with H2AZ and H3.3, respectively, in a chromatin- and ATP-stimulated manner in vitro. Our data reveal that EP400 deposits H3.3 into chromatin alongside H2AZ and contributes to gene regulation after PIC assembly.
EP400 Deposits H3.3 into Promoters and Enhancers During Gene Activation
Mol Cell . Author manuscript; available in PMC 2017 Jan 7. Published in final edited form as: Mol Cell. 2016 Jan 7; 61(1): 27–38. Published online 2015 Dec 6. doi: [ 10.1016/j.molcel.2015.10.039 ] PMCID: PMC4707986 NIHMSID: NIHMS735541 PMID: 26669263 Suman K. Pradhan , 1 Trent Su , 1 Linda Yen , 2 Karine Jacquet , 3 Chengyang Huang , 1 Jacques Cote , 3 Siavash K. Kurdistani , 1, 2 and Michael F. Carey 1, 2 Suman K. Pradhan 1 Department of Biological Chemistry, David Geffen School of Medicine, UCLA, 351A Biomedical Sciences Research Building, 615 Charles E Young Drive South, Los Angeles, CA 90095-1737 Find articles by Suman K. Pradhan Trent Su 1 Department of Biological Chemistry, David Geffen School of Medicine, UCLA, 351A Biomedical Sciences Research Building, 615 Charles E Young Drive South, Los Angeles, CA 90095-1737 Find articles by Trent Su Linda Yen 2 The Molecular Biology Institute, UCLA, Paul D. Boyer Hall, 611 Charles E. Young Drive South, Los Angeles, California 90095-1570 Find articles by Linda Yen Karine Jacquet 3 Laval University Cancer Research Center, CHU de Quebec Research Center-Oncology, Hôtel-Dieu de Québec, 9 McMahon Street, Quebec City (Quebec), CANADA G1R 2J6 Find articles by Karine Jacquet Chengyang Huang 1 Department of Biological Chemistry, David Geffen School of Medicine, UCLA, 351A Biomedical Sciences Research Building, 615 Charles E Young Drive South, Los Angeles, CA 90095-1737 Find articles by Chengyang Huang Jacques Cote 3 Laval University Cancer Research Center, CHU de Quebec Research Center-Oncology, Hôtel-Dieu de Québec, 9 McMahon Street, Quebec City (Quebec), CANADA G1R 2J6 Find articles by Jacques Cote Siavash K. Kurdistani 1 Department of Biological Chemistry, David Geffen School of Medicine, UCLA, 351A Biomedical Sciences Research Building, 615 Charles E Young Drive South, Los Angeles, CA 90095-1737 2 The Molecular Biology Institute, UCLA, Paul D. Boyer Hall, 611 Charles E. Young Drive South, Los Angeles, California 90095-1570 Find articles by Siavash K. Kurdistani Michael F. Carey 1 Department of Biological Chemistry, David Geffen School of Medicine, UCLA, 351A Biomedical Sciences Research Building, 615 Charles E Young Drive South, Los Angeles, CA 90095-1737 2 The Molecular Biology Institute, UCLA, Paul D. Boyer Hall, 611 Charles E. Young Drive South, Los Angeles, California 90095-1570 Find articles by Michael F. Carey Author information Copyright and License information Disclaimer 1 Department of Biological Chemistry, David Geffen School of Medicine, UCLA, 351A Biomedical Sciences Research Building, 615 Charles E Young Drive South, Los Angeles, CA 90095-1737 2 The Molecular Biology Institute, UCLA, Paul D. Boyer Hall, 611 Charles E. Young Drive South, Los Angeles, California 90095-1570 3 Laval University Cancer Research Center, CHU de Quebec Research Center-Oncology, Hôtel-Dieu de Québec, 9 McMahon Street, Quebec City (Quebec), CANADA G1R 2J6 Correspondence: ude.alcu.tendem@yeracm , 310-206-7859 Copyright notice Publisher's Disclaimer The publisher's final edited version of this article is available at Mol Cell See other articles in PMC that cite the published article.
Abstract
Gene activation in metazoans is accompanied by the presence of histone variants H2AZ and H3.3 within promoters and enhancers. It is not known however what protein deposits H3.3 into chromatin or whether variant chromatin plays a direct role in gene activation. Here we show that chromatin containing acetylated H2AZ and H3.3 stimulates transcription in vitro. Analysis of the Pol II pre-initiation complex on immobilized chromatin templates revealed that the E1A Binding Protein p400 (EP400) was bound preferentially to and required for transcription stimulation by acetylated double-variant chromatin. EP400 also stimulated H2AZ/H3.3 deposition into promoters and enhancers and influenced transcription in vivo at a step downstream of the Mediator complex. EP400 efficiently exchanged recombinant histones H2A and H3.1 with H2AZ and H3.3, respectively, in a chromatin- and ATP-stimulated manner in vitro. Our data reveal that EP400 deposits H3.3 into chromatin alongside H2AZ and contributes to gene regulation after PIC assembly. Keywords: EP400, H3.3, H2AZ, RNA Polymerase II, Preinitiation Complex (PIC), Histone Exchange
Keywords: EP400, H3.3, H2AZ, RNA Polymerase II, Preinitiation Complex (PIC), Histone Exchange
Introduction
Chromatin poses a structural barrier to the RNA Polymerase II (Pol II) transcriptional machinery. The process of gene activation leads to recruitment of chromatin modifying and remodeling complexes that coordinate Pol II transcription with the remodeling and eviction of nucleosomes [reviewed in ( Li et al., 2007 )]. In metazoans, the canonical histones at gene promoters and enhancers are replaced by specific histone variants, H2AZ and H3.3, which correlate with active regulatory elements and transcribed genes ( Ahmad and Henikoff, 2002 ; Chen et al., 2014 ; Chen et al., 2013 ; Mito et al., 2005 ). Although the amino acid sequences of histone variants differ from the canonical histones, their role in activated transcription is unclear. Studies from the Felsenfeld and Henikoff groups, and others, suggest a positive correlation between occupancy of variant chromatin at the promoter and transcriptional activity ( Hardy et al., 2009 ; Jin et al., 2009 ; Millar et al., 2006 ; Talbert and Henikoff, 2010 ). Moreover, activation of transcription triggers H3.3 deposition ( Schwartz and Ahmad, 2005 ). These correlations raised the important questions of whether variant chromatin facilitates transcription or accompanies it, and how variant chromatin is targeted to promoters and enhancers in mammalian cells. Our current understanding is that H2AZ and H3.3 are enriched at the promoter and enhancer regions of active genes although research indicates they also play roles in genomic stability and DNA repair ( Adam et al., 2013 ; Ray-Gallet et al., 2011 ; Xu et al., 2012 ). H3.3 distributes throughout the coding regions of genes ( Goldberg et al., 2010 ). H3.3 is also present at the promoters of developmentally regulated genes, which are typically silenced or weakly transcribed in embryonic stem cells ( Banaszynski et al., 2013 ). H3.3 is expressed throughout the cell cycle and it has been argued that Chd1 and 2 deposit it into chromatin ( Konev et al., 2007 ; Siggens et al., 2015 ). H3.3 deposition is facilitated by the histone chaperones Hira and Atrx/Daxx ( Banaszynski et al., 2013 ; Goldberg et al., 2010 ; Szenker et al., 2011 ). H2AZ deposition, performed by SWR1 in yeast and SRCAP and EP400 in human cells, plays a role in gene regulation. ( Gevry et al., 2007 ; Mizuguchi et al., 2004 ; Ruhl et al., 2006 ), while the histone chaperone ANP32E facilitates H2AZ removal ( Obri et al., 2014 ). Despite our growing knowledge of variant chromatin across the genome, mechanistic studies that define its function beyond correlations are lacking. Additionally, given that H2AZ and H3.3 appear to be inserted coordinately into nucleosomes during transcription, the question of how they are linked to each other remains unclear as different proteins apparently mediate their assembly. The coordination of transcription with histone modification and remodeling is critical to gene regulation because specific histone modifications recruit distinct effector proteins that alter the chromatin landscape to facilitate different stages of Pol II function ( Li et al., 2007 ). To understand this issue, we have been employing a GAL4-VP16-responsive in vitro transcription system coupled with immobilized template assays to capture and identify PIC composition under various conditions ( Johnson et al., 2002 ). We identified a comprehensive list of factors recruited in a GAL4-VP16-responsive model promoter system by using multidimensional protein identification technology (MuDPIT) validated by immunoblotting ( Chen et al., 2012 ; Lin et al., 2011 ). These included TFIID and the general transcription factors or GTFs (TFIIB, TFIIE, TFIIF and TFIIH), Mediator complex, Pol II and its elongation complexes (the Cdk9-containing P-TEFb and PAF) together with the chromatin modification factors p300, SAGA, Tip60 and Set1 complex, and the ATP-dependent chromatin remodeling proteins Chd1, Ino80 and EP400. Among these, the factors segregate into two categories; those whose recruitment depends on Mediator, including GTFs and TFIID, and those that are activator-dependent but Mediator-independent including SAGA and Tip60-EP400 ( Chen et al., 2012 ). Indeed, SAGA does not participate in PIC assembly but acts afterwards to allow chromatin transcription in vitro. EP400 is a SWR1-class ATP-dependent chromatin remodeling protein originally identified as an interacting partner for the adenovirus E1A protein required for viral replication and cellular transformation ( Fuchs et al., 2001 ). EP400 is a subunit of the Tip60-EP400 complex comprising ~16 subunits, including the Tip60 histone acetyltransferase (HAT), the PI3K family protein kinase homologue TRRAP and Brd8 ( Cai et al., 2003 ; Doyon and Cote, 2004 ). The EP400 H2AZ exchange potential has been reported to play a role in regulated gene expression [reviewed in ( Venkatesh and Workman, 2015 )] and DNA repair ( Xu et al., 2012 ). Tip60-EP400 also regulates embryonic stem cell identity ( Fazzio et al., 2008 ) and cell cycle ( Chan et al., 2005 ; Tyteca et al., 2006 ). EP400 is associated with gene promoters ( Fazzio et al., 2008 ). Although EP400 is primarily considered a subunit of the Tip60-EP400 complex, it also resides within another complex that lacks Tip60 and is involved in repression ( Fuchs et al., 2001 ; Park et al., 2010 ). To understand the role of histone variants in transcription, we recreated GAL4-VP16-stimulated transcription by acetylated chromatin bearing H2AZ and H3.3 in vitro. We captured PICs on variant chromatin and demonstrated they were enriched in EP400. We validated our biochemical data in a Dox-inducible Tet-VP16 dependent reporter system and genome-wide in U2OS cells. Knockdown of EP400 led to a reduction in transcription and deposition of both H2AZ and H3.3. Histone exchange experiments revealed that EP400 could exchange canonical histones with both H2AZ and H3.3 in vitro. Our data suggest that EP400 mechanistically links variant histone deposition to gene transcription.
Results
Variant Chromatin Stimulates Transcription In Vitro To understand how variant chromatin influenced transcription, we generated chromatinized promoters in vitro bearing H2AZ-H2B dimers, H3.3-H4 tetramers or both and identified conditions where they stimulated transcription. We refer to the chromatin as double variant or single variant depending on whether it contains both H3.3 and H2AZ or each alone. We took into consideration that transcriptionally active chromatin is heavily acetylated in vivo. In proteomic analyses, H3.3 is enriched in H3K9 and H3K18 acetylation relative to H3.1 in Kc cells ( McKittrick et al., 2004 ). To provide sufficient levels of histone acetylation, we pre-acetylated the chromatin with recombinant Gcn5 and Esa1, the catalytic subunits of yeast SAGA (acetylates H3 and H2B) and NuA4 (acetylates H4 and H2A), respectively, the latter being the homologue of human Tip60 ( Doyon and Cote, 2004 ; Grant et al., 1997 ). In the absence of HATs, GAL4-VP16 stimulated transcription to similar levels on canonical and variant chromatin ( Figure 1A , lanes 2, 5, 8 and 11). However, after histone acetylation, the double variant chromatin displayed a highly reproducible stimulation of transcription versus canonical ( Figure 1A ; lane 12 vs. 3). Because canonical chromatin has been widely used in and is highly active for in vitro transcription, we were expecting a transcription stimulation rather than absolute dependence on variant chromatin. Importantly, the stimulation was significantly weaker on single variant chromatin bearing either H2AZ (lane 6) or H3.3 (lane 9) with the combination (lane 12) eliciting the greatest effect. These data show that double variant chromatin directly stimulates transcription upon histone acetylation. The ability to recreate transcription stimulation by double variant chromatin provided an assay to identify factors that mediate the effect. fig ft0 fig mode=article f1 fig/graphic|fig/alternatives/graphic mode="anchored" m1 Open in a separate window Figure 1 caption a7 caption a8 Variant Chromatin Stimulates Transcription In Vitro in an EP400 Dependent Manner (A) In vitro transcription. End-biotinylated G5E4T was assembled into chromatin using canonical, single or double variant octamers as indicated and immobilized on paramagnetic beads. Chromatin was acetylated by Gcn5 and Esa1 (HAT) as indicated and incubated with nucleotides (NTPs) in HeLa nuclear extract +/− GAL4-VP16 (Activator). A phosphorimage of the primer-extension gel is shown and graphed using mean/SD spectral units of 3 replicates. Double asterisk indicates p value <0.01 by Student’s t-test; n.s. not significant. (B) Immobilized template (IT) capture of PICs. Immobilized canonical or double variant chromatin was as in (A) minus NTPs. Purified PICs were immunoblotted with antibodies to indicated proteins. (C) Quantitation of select proteins from immunoblot in (B) of acetylated canonical or double variant chromatin. (D) Levels of H3 acetylation measured for canonical and double variant chromatin using pan-H3Ac antibody: B (Bound) and S (Supernatant/Wash) fractions. (E) EP400 immunoblots of mock (M) and EP400-depleted (EP400Δ) extracts over a 9-fold titration range. (F) IT analysis of select PIC components in Mock and EP400Δ extracts. (G) Silver stained gel of EP400 preparation. (H) In vitro transcription of acetylated chromatin in Mock or EP400Δ extracts. Graphed as in (A) for 3 replicates but +/− recombinant EP400 as indicated. Student’s t-test p value <0.01 indicated by **. EP400 is a Variant Chromatin Specific Transcription Factor In Vitro To identify a specific factor that was either recruited or enriched on acetylated variant chromatin, we employed the immobilized template assay to capture an activator-responsive PIC from HeLa nuclear extract ( Johnson et al., 2002 ). We employed immunoblotting of representative subunits of all the PIC components identified by our previous MuDPIT study to determine if any of them were enriched on double variant chromatin ( Figure 1B ) ( Chen et al., 2012 ; Lin et al., 2011 ). Without chromatin acetylation, GAL4-VP16 stimulated recruitment of the factors identified by MuDPIT including general transcription factors (GTFs), TFIID, Mediator complex, Pol II and others at nearly identical levels on templates bearing canonical or double variant chromatin ( Figure 1B ). In contrast, EP400, a component of the Tip60 HAT complex, was enriched by an average of 2.5-fold ( Figure 1B ) on acetylated double variant versus acetylated canonical chromatin (graphed in Figure 1C ). Although Brd8, another Tip60 complex subunit, was enriched on acetylated double variant chromatin, oddly the Tip60 HAT subunit was not ( Figure 1B ). We conclude that EP400 enrichment within a PIC is stimulated by acetylated double variant chromatin. Histone acetylation did not typically alter the amounts of histones bound to the immobilized chromatin templates. The amounts of chromatin-bound H3.3 and H3.1, along with the amounts of H2B, used as a measure of dimers with H2AZ or H2A, were not greatly affected by acetylation ( Figure 1B , bottom panels). Furthermore, acetylation did not cause H3.3 or H3.1 bound (B) to the immobilized template to dissociate into the supernatant and wash (S) as measured by a pan-acetyl H3 antibody ( Figure 1D ). Additionally, H3.3 and H3.1 were acetylated at similar levels. We conclude that neither the transcriptional stimulatory effect of the acetylated double variant chromatin nor the enhanced binding of EP400 is due to differences in acetylation or instability of acetylated chromatin. To investigate EP400’s role in transcription, we immunodepleted it (EP400Δ) from HeLa extracts ( Figure 1E ) and examined if its loss affected PIC assembly or transcription of acetylated double variant versus canonical chromatin. Although EP400Δ did not affect recruitment of representative transcription factors to the PIC ( Figure 1F ), we observed a reproducible and significant decrease in transcription from variant chromatin. Addition of recombinant EP400 ( Figure 1G ) restored the transcriptional stimulation from double variant chromatin in the depleted extracts with little effect on canonical chromatin ( Figure 1H ). The data in Figure panels 1F and H suggests that EP400 acts independent of PIC assembly to facilitate transcription on acetylated double variant chromatin. EP400 Affects Transcription In Vivo The in vitro data prompted us to investigate whether EP400 contributes to transcriptional activation in a cell-based system. We employed a U2OS cell line bearing >10 integrated copies of the doxycyline (Dox)-inducible Tet-VP16-responsive luciferase reporter gene ( Black et al., 2006 ; Lin et al., 2011 ) ( Figure 2A ). Time-course experiments showed that reporter gene mRNA accumulated in a gradual manner within an 8-hour window ( Figure 2B ). Locus specific ChIP experiments at the promoter of the reporter gene ( Figure 2C ) detected reproducible enrichment of Tet-VP16, Pol II, Mediator and EP400. We conclude that EP400 is recruited to the promoter in a VP16-responsive manner in vivo similar to its recruitment in vitro. fig ft0 fig mode=article f1 fig/graphic|fig/alternatives/graphic mode="anchored" m1 Open in a separate window Figure 2 caption a7 caption a8 EP400 is Necessary for Transcription in a Model Cell Based Reporter System (A) Schematic of U2OS Tet-On VP16 reporter system. (B) Quantitation of F-Luc reverse transcriptase-PCR products at indicated time points (hours post Dox-induction: h.p.i) in untreated, mock and EP400 siRNA-KD cells (KD) 72-h post-transfection of siRNA. (C) ChIP-qPCR of promoter at the indicated time points with antibodies against VP16, Pol II, MED1 (Mediator), EP400 and H3.3. Bar graph indicates enrichment at time points relative to input DNA. (D) EP400 Immunoblot of Mock vs. siRNA KD normalized via GAPDH levels. (E) Immunoblot analysis of H2AZ, H3.3, H3 and H4 for crosslinked chromatin fractions or whole cell extracts. The numbers indicate the intensity of signal in KD normalized to 1 for mock (M); representative blots are shown. (F) ChIP of Pol II, Mediator and H3.3 in Mock versus EP400 KD cells time post Dox. Student’s t-test p value <0.01 is **; not significant is indicated by n.s. We next examined whether EP400 is important for reporter gene activation in vivo. EP400 was knocked down (KD) by siRNA for 72 h and gene activation was compared with Mock siRNA-treated cells. Immunoblot analysis revealed a ~4-fold EP400 KD at low (1X) and high (3x) siRNA amounts as reported ( Mattera et al., 2010 ). Under KD conditions, we observed a 6-fold decrease in reporter gene expression as compared to mock-treated cells ( Figure 2B ). Because EP400 KD is known to affect H2AZ incorporation in vivo ( Gevry et al., 2007 ; Gevry et al., 2009 ; Lee et al., 2012 ), we analyzed chromatin-associated levels of H2AZ alongside H3, H2B and H3.3. H2AZ levels decreased in chromatin from EP400 siRNA treated cells but strikingly, we observed a greater decrease in H3.3 ( Figure 2E ). By contrast, the levels of H2AZ and H3.3 in whole cell extract were relatively unaffected ( Figure 2E ). The decrease of H3.3 was unexpected. We therefore measured H3.3 accumulation at the reporter gene promoter and found that it was deposited as a time-dependent function of gene activation ( Figure 2C ). However, under EP400 KD conditions, histone variant H3.3 deposition was significantly compromised along with Pol II recruitment, whereas the binding of Tet-VP16 (data not shown) and Mediator remained unchanged ( Figure 2F ). These data suggest that EP400 is necessary for transcription on variant chromatin and may contribute to H3.3 deposition. This observation is important because little is known about EP400’s role in gene regulation. To determine whether our observation was a nuance of our reporter gene or a global phenomenon, we next asked whether the model template results were paralleled genome-wide. EP400 is Found at the Promoter and Coding Regions Genome-wide and Correlates with H3.3 and Gene Expression To perform the genome-wide analysis, we carried out gene expression measurements (RNA-Seq) and chromatin immunoprecipitation (ChIP-Seq) in the same U2OS cell line as the model reporter gene. Pol II, Mediator, EP400 and the histone variants H3.3 and H2AZ were compared at the promoter under Mock and EP400 KD conditions. To gauge the effectiveness of the parallel ChIP experiments, we generated heatmaps sorted from high to low EP400 binding by p values of the enrichment. The heatmaps show that under Mock conditions, the rank enrichment of EP400 correlates roughly with that of Pol II, Mediator, H3.3 and H2AZ ( Figure 3A ). Importantly, the enrichment of promoter-bound Mediator, Pol II, EP400, H2AZ and H3.3 appears somewhat proportional to the rank abundance of mRNA. Pol II and Mediator are concentrated near the promoter. H3.3 and H2AZ peaked in a bifurcated manner around the promoter, but H3.3 extended outward into the gene body, as did EP400 ( Figures 3A and B ). The extension of H3.3 into gene bodies has been noted previously ( Goldberg et al., 2010 ; Jin et al., 2009 ). Indeed, metagene analysis with genes binned by mRNA levels corresponding to the top (C1), middle (C2), and bottom (C3) 10% of genes demonstrated that H3.3 and EP400 enrichment in the promoter and gene body rank-correlated with gene expression ( Figure 3C ). fig ft0 fig mode=article f1 fig/graphic|fig/alternatives/graphic mode="anchored" m1 Open in a separate window Figure 3 caption a7 caption a8 Genome-wide Analysis of EP400 and Variant Chromatin at Promoters (A) Heatmaps of EP400, Pol II, Mediator (MED26), H3.3 and H2AZ at U2OS promoters (3 kb flanking TSSs). P values for enrichment are plotted and ranked by EP400. The mRNA heatmaps plot FPKM of Mock alongside fold change (FC) of log 2 KD/Mock FPKM (green is downregulated, red is upregulated and black is no change). P value of differences by Wilcoxon Rank-Sum is 9E-19. (B) Representative genome browser view. (C) Metagene analyses of (A) clustered by gene expression. The colored lines indicate enrichment for each protein of top (red;C1), middle (green;C2), and bottom (purple;C3) 10% of expressed genes; black line (C4) is average for all genes. (D) Fold change in enrichment upon EP400 KD for ChIP experiments in (A) plotted as log2 (KD/Mock) on EP400-bound gene promoters by significant peaks (TSS ± 3kb). Green indicates reduced binding of indicated protein. Total H3 (H3) ChIP data are added to this plot but not shown in A. (E) Boxplots of -log 2 poisson p values of H3.3, total H3, H2AZ, Pol II, and MED26 ChIP signals in M and KD conditions. Error by Wilcoxon Rank-Sum Test is shown. Remarkably, EP400 KD elicited a significant and reproducible negative effect on H3.3 enrichment ( Figure 3A , compare Mock with KD). The data are further supported by heatmaps of fold change in Figure 3D , ranked on a promoter-by-promoter basis by EP400 abundance (significant peaks). The columns show a decrease in abundance of the indicated proteins upon EP400 KD. EP400 KD elicited the most evident change on H3.3 enrichment. The KD also decreased Pol II and H2AZ enrichment with less change in MED26. Total H3 levels at promoters, as measured by pan-H3 antibody ChIP, decreased far less than H3.3. These data suggest but do not prove the possibility that the loss of H3.3 is being compensated for by increased retention of H3.1. Note from the boxplots of Figure 3E that the decreased enrichment of all proteins under KD conditions is significant by Wilcoxon Rank-Sum tests. However, the p values ranking from lowest to highest are H3.3, Pol II, H2AZ and total H3 with the highest p value being MED26. The KD elicited effects on gene expression for many genes as illustrated in the heat map showing the log2 ratio of KD/Mock gene expression ( Figure 3A ; green is downregulated and red is upregulated); we will discuss these data further below. In conclusion, our genome-wide analysis in the presence of EP400 siRNA is largely consistent with our model reporter results showing that decreased EP400 levels lead to a decrease in H3.3 deposition but have a much smaller effect on Mediator. As predicted from the model reporter data, EP400 KD impacted gene expression ( Figure 4A ). Notably, among the top ~10,000 genes ranked by an FPKM greater than 1, 6110 genes are affected ≥1.5 fold upon KD; 4024 genes are downregulated and 2086 are upregulated ( Figure 4A ). The boxplot analysis in Figure 4B demonstrates that in Mock cells, the genes downregulated upon KD express ~10-fold more mRNA (by median) than the upregulated ones, suggesting that EP400’s predominant role in the cell is as a positive transcriptional regulator. The upregulated genes might be due to repression by EP400, which is observed in some contexts ( Fazzio et al., 2008 ; Tyteca et al., 2006 ). Also, H3.3 is found on weakly transcribed genes bound by PRC2 ( Banaszynski et al., 2013 ). The heatmaps in Figure 4C directly show the effect of KD on downregulated (green) and upregulated (red) genes by comparing side-by-side the FPKM in Mock and KD. fig ft0 fig mode=article f1 fig/graphic|fig/alternatives/graphic mode="anchored" m1 Open in a separate window Figure 4 caption a7 caption a8 Analysis of EP400 Knockdown on Variant Chromatin Gene Expression (A) Venn diagram of differentially expressed genes in EP400 KD versus Mock. RNAs from the top 9964 genes (by >1 FPKM value) analyzed for changes ≥1.5 fold. 4024 were down regulated (green) and 2086 (red) were upregulated in EP400 KD conditions. (B) Boxplot representing downregulated (green) and upregulated (red) genes by FPKM in M and KD conditions using a log 2 scale. (C) Heatmaps showing the downregulated (green) and upregulated (red) genes in M and KD by FPKM and ranked by fold change for clarity. (D) Gene Ontology analysis of affected genes by category. P values were corrected for multiple hypothesis testing using Benjamini correction. Gene Ontology analysis identified translation, RNA processing, cell cycle, chromatin organization and mitosis as the top hits for the functional enrichment of EP400-bound genes (p<10 −10 ). Such a result might be expected for a highly proliferative cancer cell line and is consistent with EP400 binding to the viral oncoprotein E1A during cellular transformation ( Fuchs et al., 2001 ) and the EP400-mediated cell cycle effects ( Chan et al., 2005 ; Tyteca et al., 2006 ) ( Figure 4D ). We conclude the EP400 is required for full expression of a subset of highly transcribed genes but is also necessary for repression of fewer genes with the caveat that some effects may be indirect, i.e., cases where EP400 might positively or negatively regulate a specific transcriptional repressor. EP400 is Enriched at Enhancers and Regulates H3.3 Occupancy Tip60-EP400 participates in H2AZ exchange at enhancers ( Gevry et al., 2009 ). We therefore analyzed EP400 distribution at enhancers identified by the presence of both the H3K4me1 and H3K18ac histone modifications under mock conditions. H3K18 is acetylated by the p300 and CBP HATs ( Jin et al., 2011 ). We restricted our analysis to distal intergenic enhancers located ≥3 kb upstream or downstream of an annotated TSS or TTS. When the enrichment p values are sorted by the relative abundance of H3K4me1, EP400 correlates with the presence of H3K18ac, H3.3 and Pol II ( Figure 5A ) consistent with the browser track ( Figure 5B ). fig ft0 fig mode=article f1 fig/graphic|fig/alternatives/graphic mode="anchored" m1 Open in a separate window Figure 5 caption a7 caption a8 Analysis of EP400 Knockdown on Enhancer Occupancy and Gene Expression (A) Heatmaps of H3K4me1, H3K18ac, EP400, H3.3, H2AZ and Pol II at active distal intergenic U2OS enhancers (−/+5 kb from center of H3K4me1 peaks) in KD and/or Mock cells. Active enhancers were scored by the H3K4me1 and H3K18ac marks, excluding +3 kb upstream of TSS or downstream of TTS, and sorted by H3K4me1 enrichment in Mock and EP400 siRNA KD cells. (B) Browser track of putative enhancers. Those showing markers consistent with active enhancers are indicated by arrows below the tracks, i.e., identifiable peaks of Pol II, Mediator, EP400, H3.3, H2AZ, H3K18ac and H3K4me1. (C) Fold change in enrichment of H3.3, H2AZ and Pol II as in Fig. 3D for all EP400-bound enhancers by significant peaks, ranked by H3K4me1 abundance. Wilcoxon Rank-Sum p values for change in H3.3 and H2AZ are both <2E-308 (saturated), and for Pol II is 2E-148. (D) Boxplots of -log 2 Poisson p-values of H3.3, H2AZ and Pol II in M and KD conditions for top 10% enhancers with p values. (E) Nearest neighbor effects by GREAT analysis. Average FPKM was plotted for nearest enhancer-proximal genes for Top, Middle and Bottom 10% of putative enhancers as sorted by H3K4me1 enrichment. The fold-change ( Figure 5C ) and boxplot analyses ( Figure 5D ) of the relationships reveal a strong decrease in H3.3 levels upon EP400 KD. The effect of EP400 KD on H2AZ levels was also evident by the enrichment heatmaps in (A), fold-change heatmaps in (C), and boxplots in (D), consistent with a previous report ( Gevry et al., 2009 ). Pol II occupancy also decreased at enhancers upon EP400 KD although its overall enrichment at enhancers was lower than promoters (note scales used Figure 5A vs. vs.3A 3A ). Genomic Regions Enrichment of Annotation Tool (GREAT) was used to identify the nearest neighboring genes to enhancers. The mRNA FPKM values from the enhancer-neighboring genes decrease upon EP400 KD cells ( Figure 5E ). Collectively, our data establish a link between EP400 and H3.3 occupancy at both promoters and enhancers and suggest that EP400 bound at enhancers might contribute to full levels of expression from the nearest neighboring genes. However, without establishing a direct enhancer-gene relationship, the data are simply correlative. H2AZ and H3.3 Exchange Activity of EP400 In Vitro The data from the genome-wide study raised the possibility that EP400 is necessary for H3.3 deposition during transcription. It is difficult to address in vivo whether H3.3 deposition is a direct effect of EP400 or an indirect effect due to other EP400-affected transcription steps that, in turn, influence histone deposition. However, others have shown that EP400 mediates H2AZ exchange in standard in vitro assays ( Gevry et al., 2007 ). In order for an H2AZ-H3.3 double variant nucleosome to be deposited during gene activation, it would have to replace an entire canonical nucleosome. We imagined two ways that a nucleosome could be replaced. In one model, the presence of a canonical nucleosome would stimulate the swapping out with a variant nucleosome. In the other, the canonical nucleosome would be evicted by an alternate process, which would then create space for the assembly of the histone variants on naked DNA. The relevant in vitro predictions of the two models are that histone variants would be deposited onto a template in either a chromatin or DNA stimulated manner. To test these predictions directly, we incubated magnetic bead-immobilized templates, bearing either canonical chromatin or naked DNA, with free variant H2AZ-H2B dimers, H3.3-H4 tetramers or double variant octamers in the presence of recombinant EP400 plus or minus the addition of ATP ( Figure 6 ). As previously noted ( Gevry et al., 2007 ), EP400 co-purifies with ATP and the enzyme Apyrase was necessary to reduce exchange activity in the absence of added ATP. Although a modest amount of background binding by the histone variants was to be expected, EP400 stimulated the reproducible deposition of both H2AZ-containing histone dimers ( Figure 6A ) and H3.3 tetramers ( Figure 6B ) onto canonical nucleosome-bound beads in an ATP-stimulated manner. ATPγS did not support the exchange. Importantly, we detected clear H2A and H3.1 eviction into the supernatant (Sup) in parallel to the deposition of H2AZ and H3.3 ( Figure 6A and B ). In contrast, EP400 deposited variant chromatin less efficiently on a naked DNA template versus a canonical chromatin template ( Figure 6C ). This result is in agreement with the idea that EP400 functions via interaction with chromatin ( Ruhf et al., 2001 ). Although EP400 could perform the reverse exchange of H2A and H3.1 into variant chromatin, its activity was substantially lower in side-by-side reactions than exchange of histone variants into canonical chromatin ( Figure 6D , compare panel I for H2AZ with panel II for H2A; Figure 6E , compare panel I for H3.3 with panel II for H3.1). In contrast, EP400 efficiently exchanged a His-tagged H3.3 onto Flag-tagged H3.3-containing chromatin ( Figure 6E , panel III). We conclude that assembly of variant chromatin is stimulated by the presence of chromatin on the template and proceeds in a direction that favors deposition of histone variants in place of canonical ones. Our data confirm previous reports that EP400 is an H2AZ exchange factor and importantly, extend that functionality to an H3.3 exchange factor and provide a plausible explanation for how H3.3 is deposited during transcription. fig ft0 fig mode=article f1 fig/graphic|fig/alternatives/graphic mode="anchored" m1 Open in a separate window Figure 6 caption a7 caption a8 Histone Exchange by EP400 In Vitro 300 ng immobilized canonical chromatin was incubated as indicated with Apyrase-treated EP400 and 300 ng H3.3-H4 tetramers or H2AZ-H2B dimers +/−1 mM ATP. Chromatin was captured, washed and subjected to immunoblotting for H2AZ (A) or H3 (B). FLAG-H3.3 is distinguished from H3.1 by mobility and detected using a pan H3 antibody; H2B on beads as a control. (C) Alternatively, 150 ng of immobilized naked DNA or 300 ng canonical chromatin was incubated with EP400 and 300 ng of histone variant octamers as above and FLAG-immunoblotted for H3.3. Graphs of (A–C) represent ≥3 independent experiments. Input and conditions for (D) and (E) were as above. (D) I. shows the H2AZ exchange into bead-bound H2A-containing chromatin; II. shows reverse reaction of H2A deposition into H2AZ chromatin. (E). I. shows H3.3 exchange into H3.1 chromatin; II. shows the reverse reaction of H3.1 deposition into H3.3 chromatin; III. shows exchange of His-H3.3 into Flag-H3.3 chromatin.
Variant Chromatin Stimulates Transcription In Vitro
To understand how variant chromatin influenced transcription, we generated chromatinized promoters in vitro bearing H2AZ-H2B dimers, H3.3-H4 tetramers or both and identified conditions where they stimulated transcription. We refer to the chromatin as double variant or single variant depending on whether it contains both H3.3 and H2AZ or each alone. We took into consideration that transcriptionally active chromatin is heavily acetylated in vivo. In proteomic analyses, H3.3 is enriched in H3K9 and H3K18 acetylation relative to H3.1 in Kc cells ( McKittrick et al., 2004 ). To provide sufficient levels of histone acetylation, we pre-acetylated the chromatin with recombinant Gcn5 and Esa1, the catalytic subunits of yeast SAGA (acetylates H3 and H2B) and NuA4 (acetylates H4 and H2A), respectively, the latter being the homologue of human Tip60 ( Doyon and Cote, 2004 ; Grant et al., 1997 ). In the absence of HATs, GAL4-VP16 stimulated transcription to similar levels on canonical and variant chromatin ( Figure 1A , lanes 2, 5, 8 and 11). However, after histone acetylation, the double variant chromatin displayed a highly reproducible stimulation of transcription versus canonical ( Figure 1A ; lane 12 vs. 3). Because canonical chromatin has been widely used in and is highly active for in vitro transcription, we were expecting a transcription stimulation rather than absolute dependence on variant chromatin. Importantly, the stimulation was significantly weaker on single variant chromatin bearing either H2AZ (lane 6) or H3.3 (lane 9) with the combination (lane 12) eliciting the greatest effect. These data show that double variant chromatin directly stimulates transcription upon histone acetylation. The ability to recreate transcription stimulation by double variant chromatin provided an assay to identify factors that mediate the effect. fig ft0 fig mode=article f1 fig/graphic|fig/alternatives/graphic mode="anchored" m1 Open in a separate window Figure 1 caption a7 caption a8 in an EP400 Dependent Manner (A) In vitro transcription. End-biotinylated G5E4T was assembled into chromatin using canonical, single or double variant octamers as indicated and immobilized on paramagnetic beads. Chromatin was acetylated by Gcn5 and Esa1 (HAT) as indicated and incubated with nucleotides (NTPs) in HeLa nuclear extract +/− GAL4-VP16 (Activator). A phosphorimage of the primer-extension gel is shown and graphed using mean/SD spectral units of 3 replicates. Double asterisk indicates p value <0.01 by Student’s t-test; n.s. not significant. (B) Immobilized template (IT) capture of PICs. Immobilized canonical or double variant chromatin was as in (A) minus NTPs. Purified PICs were immunoblotted with antibodies to indicated proteins. (C) Quantitation of select proteins from immunoblot in (B) of acetylated canonical or double variant chromatin. (D) Levels of H3 acetylation measured for canonical and double variant chromatin using pan-H3Ac antibody: B (Bound) and S (Supernatant/Wash) fractions. (E) EP400 immunoblots of mock (M) and EP400-depleted (EP400Δ) extracts over a 9-fold titration range. (F) IT analysis of select PIC components in Mock and EP400Δ extracts. (G) Silver stained gel of EP400 preparation. (H) In vitro transcription of acetylated chromatin in Mock or EP400Δ extracts. Graphed as in (A) for 3 replicates but +/− recombinant EP400 as indicated. Student’s t-test p value <0.01 indicated by **.
EP400 is a Variant Chromatin Specific Transcription Factor In Vitro
To identify a specific factor that was either recruited or enriched on acetylated variant chromatin, we employed the immobilized template assay to capture an activator-responsive PIC from HeLa nuclear extract ( Johnson et al., 2002 ). We employed immunoblotting of representative subunits of all the PIC components identified by our previous MuDPIT study to determine if any of them were enriched on double variant chromatin ( Figure 1B ) ( Chen et al., 2012 ; Lin et al., 2011 ). Without chromatin acetylation, GAL4-VP16 stimulated recruitment of the factors identified by MuDPIT including general transcription factors (GTFs), TFIID, Mediator complex, Pol II and others at nearly identical levels on templates bearing canonical or double variant chromatin ( Figure 1B ). In contrast, EP400, a component of the Tip60 HAT complex, was enriched by an average of 2.5-fold ( Figure 1B ) on acetylated double variant versus acetylated canonical chromatin (graphed in Figure 1C ). Although Brd8, another Tip60 complex subunit, was enriched on acetylated double variant chromatin, oddly the Tip60 HAT subunit was not ( Figure 1B ). We conclude that EP400 enrichment within a PIC is stimulated by acetylated double variant chromatin. Histone acetylation did not typically alter the amounts of histones bound to the immobilized chromatin templates. The amounts of chromatin-bound H3.3 and H3.1, along with the amounts of H2B, used as a measure of dimers with H2AZ or H2A, were not greatly affected by acetylation ( Figure 1B , bottom panels). Furthermore, acetylation did not cause H3.3 or H3.1 bound (B) to the immobilized template to dissociate into the supernatant and wash (S) as measured by a pan-acetyl H3 antibody ( Figure 1D ). Additionally, H3.3 and H3.1 were acetylated at similar levels. We conclude that neither the transcriptional stimulatory effect of the acetylated double variant chromatin nor the enhanced binding of EP400 is due to differences in acetylation or instability of acetylated chromatin. To investigate EP400’s role in transcription, we immunodepleted it (EP400Δ) from HeLa extracts ( Figure 1E ) and examined if its loss affected PIC assembly or transcription of acetylated double variant versus canonical chromatin. Although EP400Δ did not affect recruitment of representative transcription factors to the PIC ( Figure 1F ), we observed a reproducible and significant decrease in transcription from variant chromatin. Addition of recombinant EP400 ( Figure 1G ) restored the transcriptional stimulation from double variant chromatin in the depleted extracts with little effect on canonical chromatin ( Figure 1H ). The data in Figure panels 1F and H suggests that EP400 acts independent of PIC assembly to facilitate transcription on acetylated double variant chromatin.
EP400 Affects Transcription In Vivo
The in vitro data prompted us to investigate whether EP400 contributes to transcriptional activation in a cell-based system. We employed a U2OS cell line bearing >10 integrated copies of the doxycyline (Dox)-inducible Tet-VP16-responsive luciferase reporter gene ( Black et al., 2006 ; Lin et al., 2011 ) ( Figure 2A ). Time-course experiments showed that reporter gene mRNA accumulated in a gradual manner within an 8-hour window ( Figure 2B ). Locus specific ChIP experiments at the promoter of the reporter gene ( Figure 2C ) detected reproducible enrichment of Tet-VP16, Pol II, Mediator and EP400. We conclude that EP400 is recruited to the promoter in a VP16-responsive manner in vivo similar to its recruitment in vitro. fig ft0 fig mode=article f1 fig/graphic|fig/alternatives/graphic mode="anchored" m1 Open in a separate window Figure 2 caption a7 caption a8 EP400 is Necessary for Transcription in a Model Cell Based Reporter System (A) Schematic of U2OS Tet-On VP16 reporter system. (B) Quantitation of F-Luc reverse transcriptase-PCR products at indicated time points (hours post Dox-induction: h.p.i) in untreated, mock and EP400 siRNA-KD cells (KD) 72-h post-transfection of siRNA. (C) ChIP-qPCR of promoter at the indicated time points with antibodies against VP16, Pol II, MED1 (Mediator), EP400 and H3.3. Bar graph indicates enrichment at time points relative to input DNA. (D) EP400 Immunoblot of Mock vs. siRNA KD normalized via GAPDH levels. (E) Immunoblot analysis of H2AZ, H3.3, H3 and H4 for crosslinked chromatin fractions or whole cell extracts. The numbers indicate the intensity of signal in KD normalized to 1 for mock (M); representative blots are shown. (F) ChIP of Pol II, Mediator and H3.3 in Mock versus EP400 KD cells time post Dox. Student’s t-test p value <0.01 is **; not significant is indicated by n.s. We next examined whether EP400 is important for reporter gene activation in vivo. EP400 was knocked down (KD) by siRNA for 72 h and gene activation was compared with Mock siRNA-treated cells. Immunoblot analysis revealed a ~4-fold EP400 KD at low (1X) and high (3x) siRNA amounts as reported ( Mattera et al., 2010 ). Under KD conditions, we observed a 6-fold decrease in reporter gene expression as compared to mock-treated cells ( Figure 2B ). Because EP400 KD is known to affect H2AZ incorporation in vivo ( Gevry et al., 2007 ; Gevry et al., 2009 ; Lee et al., 2012 ), we analyzed chromatin-associated levels of H2AZ alongside H3, H2B and H3.3. H2AZ levels decreased in chromatin from EP400 siRNA treated cells but strikingly, we observed a greater decrease in H3.3 ( Figure 2E ). By contrast, the levels of H2AZ and H3.3 in whole cell extract were relatively unaffected ( Figure 2E ). The decrease of H3.3 was unexpected. We therefore measured H3.3 accumulation at the reporter gene promoter and found that it was deposited as a time-dependent function of gene activation ( Figure 2C ). However, under EP400 KD conditions, histone variant H3.3 deposition was significantly compromised along with Pol II recruitment, whereas the binding of Tet-VP16 (data not shown) and Mediator remained unchanged ( Figure 2F ). These data suggest that EP400 is necessary for transcription on variant chromatin and may contribute to H3.3 deposition. This observation is important because little is known about EP400’s role in gene regulation. To determine whether our observation was a nuance of our reporter gene or a global phenomenon, we next asked whether the model template results were paralleled genome-wide.
EP400 is Found at the Promoter and Coding Regions Genome-wide and Correlates with H3.3 and Gene Expression
To perform the genome-wide analysis, we carried out gene expression measurements (RNA-Seq) and chromatin immunoprecipitation (ChIP-Seq) in the same U2OS cell line as the model reporter gene. Pol II, Mediator, EP400 and the histone variants H3.3 and H2AZ were compared at the promoter under Mock and EP400 KD conditions. To gauge the effectiveness of the parallel ChIP experiments, we generated heatmaps sorted from high to low EP400 binding by p values of the enrichment. The heatmaps show that under Mock conditions, the rank enrichment of EP400 correlates roughly with that of Pol II, Mediator, H3.3 and H2AZ ( Figure 3A ). Importantly, the enrichment of promoter-bound Mediator, Pol II, EP400, H2AZ and H3.3 appears somewhat proportional to the rank abundance of mRNA. Pol II and Mediator are concentrated near the promoter. H3.3 and H2AZ peaked in a bifurcated manner around the promoter, but H3.3 extended outward into the gene body, as did EP400 ( Figures 3A and B ). The extension of H3.3 into gene bodies has been noted previously ( Goldberg et al., 2010 ; Jin et al., 2009 ). Indeed, metagene analysis with genes binned by mRNA levels corresponding to the top (C1), middle (C2), and bottom (C3) 10% of genes demonstrated that H3.3 and EP400 enrichment in the promoter and gene body rank-correlated with gene expression ( Figure 3C ). fig ft0 fig mode=article f1 fig/graphic|fig/alternatives/graphic mode="anchored" m1 Open in a separate window Figure 3 caption a7 caption a8 Genome-wide Analysis of EP400 and Variant Chromatin at Promoters (A) Heatmaps of EP400, Pol II, Mediator (MED26), H3.3 and H2AZ at U2OS promoters (3 kb flanking TSSs). P values for enrichment are plotted and ranked by EP400. The mRNA heatmaps plot FPKM of Mock alongside fold change (FC) of log 2 KD/Mock FPKM (green is downregulated, red is upregulated and black is no change). P value of differences by Wilcoxon Rank-Sum is 9E-19. (B) Representative genome browser view. (C) Metagene analyses of (A) clustered by gene expression. The colored lines indicate enrichment for each protein of top (red;C1), middle (green;C2), and bottom (purple;C3) 10% of expressed genes; black line (C4) is average for all genes. (D) Fold change in enrichment upon EP400 KD for ChIP experiments in (A) plotted as log2 (KD/Mock) on EP400-bound gene promoters by significant peaks (TSS ± 3kb). Green indicates reduced binding of indicated protein. Total H3 (H3) ChIP data are added to this plot but not shown in A. (E) Boxplots of -log 2 poisson p values of H3.3, total H3, H2AZ, Pol II, and MED26 ChIP signals in M and KD conditions. Error by Wilcoxon Rank-Sum Test is shown. Remarkably, EP400 KD elicited a significant and reproducible negative effect on H3.3 enrichment ( Figure 3A , compare Mock with KD). The data are further supported by heatmaps of fold change in Figure 3D , ranked on a promoter-by-promoter basis by EP400 abundance (significant peaks). The columns show a decrease in abundance of the indicated proteins upon EP400 KD. EP400 KD elicited the most evident change on H3.3 enrichment. The KD also decreased Pol II and H2AZ enrichment with less change in MED26. Total H3 levels at promoters, as measured by pan-H3 antibody ChIP, decreased far less than H3.3. These data suggest but do not prove the possibility that the loss of H3.3 is being compensated for by increased retention of H3.1. Note from the boxplots of Figure 3E that the decreased enrichment of all proteins under KD conditions is significant by Wilcoxon Rank-Sum tests. However, the p values ranking from lowest to highest are H3.3, Pol II, H2AZ and total H3 with the highest p value being MED26. The KD elicited effects on gene expression for many genes as illustrated in the heat map showing the log2 ratio of KD/Mock gene expression ( Figure 3A ; green is downregulated and red is upregulated); we will discuss these data further below. In conclusion, our genome-wide analysis in the presence of EP400 siRNA is largely consistent with our model reporter results showing that decreased EP400 levels lead to a decrease in H3.3 deposition but have a much smaller effect on Mediator. As predicted from the model reporter data, EP400 KD impacted gene expression ( Figure 4A ). Notably, among the top ~10,000 genes ranked by an FPKM greater than 1, 6110 genes are affected ≥1.5 fold upon KD; 4024 genes are downregulated and 2086 are upregulated ( Figure 4A ). The boxplot analysis in Figure 4B demonstrates that in Mock cells, the genes downregulated upon KD express ~10-fold more mRNA (by median) than the upregulated ones, suggesting that EP400’s predominant role in the cell is as a positive transcriptional regulator. The upregulated genes might be due to repression by EP400, which is observed in some contexts ( Fazzio et al., 2008 ; Tyteca et al., 2006 ). Also, H3.3 is found on weakly transcribed genes bound by PRC2 ( Banaszynski et al., 2013 ). The heatmaps in Figure 4C directly show the effect of KD on downregulated (green) and upregulated (red) genes by comparing side-by-side the FPKM in Mock and KD. fig ft0 fig mode=article f1 fig/graphic|fig/alternatives/graphic mode="anchored" m1 Open in a separate window Figure 4 caption a7 caption a8 Analysis of EP400 Knockdown on Variant Chromatin Gene Expression (A) Venn diagram of differentially expressed genes in EP400 KD versus Mock. RNAs from the top 9964 genes (by >1 FPKM value) analyzed for changes ≥1.5 fold. 4024 were down regulated (green) and 2086 (red) were upregulated in EP400 KD conditions. (B) Boxplot representing downregulated (green) and upregulated (red) genes by FPKM in M and KD conditions using a log 2 scale. (C) Heatmaps showing the downregulated (green) and upregulated (red) genes in M and KD by FPKM and ranked by fold change for clarity. (D) Gene Ontology analysis of affected genes by category. P values were corrected for multiple hypothesis testing using Benjamini correction. Gene Ontology analysis identified translation, RNA processing, cell cycle, chromatin organization and mitosis as the top hits for the functional enrichment of EP400-bound genes (p<10 −10 ). Such a result might be expected for a highly proliferative cancer cell line and is consistent with EP400 binding to the viral oncoprotein E1A during cellular transformation ( Fuchs et al., 2001 ) and the EP400-mediated cell cycle effects ( Chan et al., 2005 ; Tyteca et al., 2006 ) ( Figure 4D ). We conclude the EP400 is required for full expression of a subset of highly transcribed genes but is also necessary for repression of fewer genes with the caveat that some effects may be indirect, i.e., cases where EP400 might positively or negatively regulate a specific transcriptional repressor.
EP400 is Enriched at Enhancers and Regulates H3.3 Occupancy
Tip60-EP400 participates in H2AZ exchange at enhancers ( Gevry et al., 2009 ). We therefore analyzed EP400 distribution at enhancers identified by the presence of both the H3K4me1 and H3K18ac histone modifications under mock conditions. H3K18 is acetylated by the p300 and CBP HATs ( Jin et al., 2011 ). We restricted our analysis to distal intergenic enhancers located ≥3 kb upstream or downstream of an annotated TSS or TTS. When the enrichment p values are sorted by the relative abundance of H3K4me1, EP400 correlates with the presence of H3K18ac, H3.3 and Pol II ( Figure 5A ) consistent with the browser track ( Figure 5B ). fig ft0 fig mode=article f1 fig/graphic|fig/alternatives/graphic mode="anchored" m1 Open in a separate window Figure 5 caption a7 caption a8 Analysis of EP400 Knockdown on Enhancer Occupancy and Gene Expression (A) Heatmaps of H3K4me1, H3K18ac, EP400, H3.3, H2AZ and Pol II at active distal intergenic U2OS enhancers (−/+5 kb from center of H3K4me1 peaks) in KD and/or Mock cells. Active enhancers were scored by the H3K4me1 and H3K18ac marks, excluding +3 kb upstream of TSS or downstream of TTS, and sorted by H3K4me1 enrichment in Mock and EP400 siRNA KD cells. (B) Browser track of putative enhancers. Those showing markers consistent with active enhancers are indicated by arrows below the tracks, i.e., identifiable peaks of Pol II, Mediator, EP400, H3.3, H2AZ, H3K18ac and H3K4me1. (C) Fold change in enrichment of H3.3, H2AZ and Pol II as in Fig. 3D for all EP400-bound enhancers by significant peaks, ranked by H3K4me1 abundance. Wilcoxon Rank-Sum p values for change in H3.3 and H2AZ are both <2E-308 (saturated), and for Pol II is 2E-148. (D) Boxplots of -log 2 Poisson p-values of H3.3, H2AZ and Pol II in M and KD conditions for top 10% enhancers with p values. (E) Nearest neighbor effects by GREAT analysis. Average FPKM was plotted for nearest enhancer-proximal genes for Top, Middle and Bottom 10% of putative enhancers as sorted by H3K4me1 enrichment. The fold-change ( Figure 5C ) and boxplot analyses ( Figure 5D ) of the relationships reveal a strong decrease in H3.3 levels upon EP400 KD. The effect of EP400 KD on H2AZ levels was also evident by the enrichment heatmaps in (A), fold-change heatmaps in (C), and boxplots in (D), consistent with a previous report ( Gevry et al., 2009 ). Pol II occupancy also decreased at enhancers upon EP400 KD although its overall enrichment at enhancers was lower than promoters (note scales used Figure 5A vs. vs.3A 3A ). Genomic Regions Enrichment of Annotation Tool (GREAT) was used to identify the nearest neighboring genes to enhancers. The mRNA FPKM values from the enhancer-neighboring genes decrease upon EP400 KD cells ( Figure 5E ). Collectively, our data establish a link between EP400 and H3.3 occupancy at both promoters and enhancers and suggest that EP400 bound at enhancers might contribute to full levels of expression from the nearest neighboring genes. However, without establishing a direct enhancer-gene relationship, the data are simply correlative.
H2AZ and H3.3 Exchange Activity of EP400 In Vitro
The data from the genome-wide study raised the possibility that EP400 is necessary for H3.3 deposition during transcription. It is difficult to address in vivo whether H3.3 deposition is a direct effect of EP400 or an indirect effect due to other EP400-affected transcription steps that, in turn, influence histone deposition. However, others have shown that EP400 mediates H2AZ exchange in standard in vitro assays ( Gevry et al., 2007 ). In order for an H2AZ-H3.3 double variant nucleosome to be deposited during gene activation, it would have to replace an entire canonical nucleosome. We imagined two ways that a nucleosome could be replaced. In one model, the presence of a canonical nucleosome would stimulate the swapping out with a variant nucleosome. In the other, the canonical nucleosome would be evicted by an alternate process, which would then create space for the assembly of the histone variants on naked DNA. The relevant in vitro predictions of the two models are that histone variants would be deposited onto a template in either a chromatin or DNA stimulated manner. To test these predictions directly, we incubated magnetic bead-immobilized templates, bearing either canonical chromatin or naked DNA, with free variant H2AZ-H2B dimers, H3.3-H4 tetramers or double variant octamers in the presence of recombinant EP400 plus or minus the addition of ATP ( Figure 6 ). As previously noted ( Gevry et al., 2007 ), EP400 co-purifies with ATP and the enzyme Apyrase was necessary to reduce exchange activity in the absence of added ATP. Although a modest amount of background binding by the histone variants was to be expected, EP400 stimulated the reproducible deposition of both H2AZ-containing histone dimers ( Figure 6A ) and H3.3 tetramers ( Figure 6B ) onto canonical nucleosome-bound beads in an ATP-stimulated manner. ATPγS did not support the exchange. Importantly, we detected clear H2A and H3.1 eviction into the supernatant (Sup) in parallel to the deposition of H2AZ and H3.3 ( Figure 6A and B ). In contrast, EP400 deposited variant chromatin less efficiently on a naked DNA template versus a canonical chromatin template ( Figure 6C ). This result is in agreement with the idea that EP400 functions via interaction with chromatin ( Ruhf et al., 2001 ). Although EP400 could perform the reverse exchange of H2A and H3.1 into variant chromatin, its activity was substantially lower in side-by-side reactions than exchange of histone variants into canonical chromatin ( Figure 6D , compare panel I for H2AZ with panel II for H2A; Figure 6E , compare panel I for H3.3 with panel II for H3.1). In contrast, EP400 efficiently exchanged a His-tagged H3.3 onto Flag-tagged H3.3-containing chromatin ( Figure 6E , panel III). We conclude that assembly of variant chromatin is stimulated by the presence of chromatin on the template and proceeds in a direction that favors deposition of histone variants in place of canonical ones. Our data confirm previous reports that EP400 is an H2AZ exchange factor and importantly, extend that functionality to an H3.3 exchange factor and provide a plausible explanation for how H3.3 is deposited during transcription. fig ft0 fig mode=article f1 fig/graphic|fig/alternatives/graphic mode="anchored" m1 Open in a separate window Figure 6 caption a7 caption a8 Histone Exchange by EP400 In Vitro 300 ng immobilized canonical chromatin was incubated as indicated with Apyrase-treated EP400 and 300 ng H3.3-H4 tetramers or H2AZ-H2B dimers +/−1 mM ATP. Chromatin was captured, washed and subjected to immunoblotting for H2AZ (A) or H3 (B). FLAG-H3.3 is distinguished from H3.1 by mobility and detected using a pan H3 antibody; H2B on beads as a control. (C) Alternatively, 150 ng of immobilized naked DNA or 300 ng canonical chromatin was incubated with EP400 and 300 ng of histone variant octamers as above and FLAG-immunoblotted for H3.3. Graphs of (A–C) represent ≥3 independent experiments. Input and conditions for (D) and (E) were as above. (D) I. shows the H2AZ exchange into bead-bound H2A-containing chromatin; II. shows reverse reaction of H2A deposition into H2AZ chromatin. (E). I. shows H3.3 exchange into H3.1 chromatin; II. shows the reverse reaction of H3.1 deposition into H3.3 chromatin; III. shows exchange of His-H3.3 into Flag-H3.3 chromatin.
Discussion
Our study shows that i.) acetylated double variant chromatin is stimulatory for transcription in vitro in an EP400-dependent manner ( Figure 1 ) and ii.) an EP400 KD dependent decrease in deposition of H3.3, and to a lesser extent H2AZ, leads to defects in enhancers and promoters that primarily result in lower levels of transcription in vivo ( Figures 3 , ,4 4 and and5). 5 ). The KD elicited a widespread decrease in genomic mRNA ranging from 1.5- to 20-fold for ~4000 genes (median effect ~ 2.4-fold) ( Figure 4 ). The decrease in expression of the integrated Tet-VP16 responsive reporter gene fell within this range ( Figure 2 ). The results from our in vitro transcription system using both mock and EP400 depleted extracts are particularly relevant as they represent direct effects and as such implicate EP400 and double variant chromatin in stimulation of transcription ( Figure 1 ). We did note many upregulated genes upon KD suggesting these may be repressed but their mRNA expression levels were lower, and they represented half the number of genes positively regulated by EP400. Acetylation of double variant chromatin by Gcn5 and Esa1 is required for EP400 dependent transcription stimulation in HeLa extract. This is relevant because Gcn5, as part of SAGA, and Tip60, the human homologue of Esa1 and the HAT subunit of the Tip60-EP400 complex, are the major HATs stably recruited to PICs in vitro by our model activator GAL4-VP16 ( Chen et al., 2012 ). The SAGA and Tip60-EP400 HAT complexes are recruited by activator in a Mediator-independent manner and are not required for assembly of Pol II and the GTFs at a promoter. SAGA functions downstream of PIC assembly by enabling efficient transcription ( Chen et al., 2012 ). EP400, as a subunit of the Tip60-EP400 complex, functions analogous to SAGA in that its loss does not greatly affect Mediator, GTF or Pol II recruitment at the promoter in vitro but does affect transcription specifically on acetylated double variant chromatin ( Figure 1 ). These data suggest that like SAGA, EP400 functions downstream of PIC assembly by facilitating initiation or elongation by Pol II. Indeed, on both our model reporter template and genome-wide, Mediator recruitment, which is central to PIC assembly, was far less affected by EP400 KD than was H3.3. The yeast homologues of SAGA and Tip60 (NuA4), are necessary for the efficient recruitment of the SWI family member RSC to gene bodies in S. cerevisiae ( Spain et al., 2014 ). In turn, RSC plays a role in both recruitment of Pol II to gene bodies and transcription elongation ( Carey et al., 2006 ; Spain et al., 2014 ). Indeed, RSC binds with high avidity to chromatin acetylated by Gcn5 and Esa1 probably through one or more of its bromodomains. By analogy, the SWI-like EP400 displays a 2.5-fold enrichment on acetylated double variant chromatin and the Tip60-EP400 complex contains a bromodomain factor Brd8, which is enriched similarly ( Doyon et al., 2004 ). It is plausible that Brd8 binds directly to acetylated histones and is responsible for targeting EP400 to acetylated double variant chromatin in vivo. Our in vitro data show that EP400 has the ability to exchange/deposit variant histones onto chromatin more efficiently than naked DNA suggesting that its role is to deposit H3.3 into canonical chromatin during gene transcription ( Figure 6 ). In side-by-side comparisons in vitro, EP400 exchanged variant histones into canonical chromatin more effectively than canonical histones into variant chromatin. However, interestingly, a His-tagged H3.3-H4 bearing tetramer was exchanged very efficiently into double variant chromatin bearing Flag-tagged H3.3. Thus, the reaction seems to favor insertion of soluble H3.3 irrespective of whether the chromatin is canonical or variant. This observation might suggest that EP400 exchanges H3.3 present in the crude HeLa nuclear extracts for the chromatin bound H3.3. Perhaps this continuous exchange process favors a chromatin state that stimulates transcription in vitro. The discovery of EP400 at both promoters and enhancers correlated with the presence of double variant nucleosomes ( Gevry et al., 2009 ; Taubert et al., 2004 ). Because EP400 is a subunit of the Tip60 complex, it is likely recruited to the DNA in part through activators interacting with the TRRAP subunit of Tip60, which is also present in SAGA ( Vassilev et al., 1998 ), and in part through its interaction with chromatin, possibly through Brd8. Thus, the TRRAP subunit likely contributes to recruitment of both the Tip60 and SAGA HATs to enhancers and promoters allowing effective activator-targeted acetylation of all four histones ( Brown et al., 2001 ). Because numerous activators recruit Mediator and SAGA individually through distinct subunit interactions, it makes sense that SAGA and Tip60-EP400 could function independently of Mediator but be necessary for its ability to generate productive mRNA synthesis in a chromatin environment. Indeed, the separate recruitment events may ensure the fidelity of combinatorial control as a mechanism for differential gene regulation on chromatin ( Carey, 1998 ). A similarity shared among promoters and enhancers is the strong effect EP400 KD has on H3.3 deposition. The KD has a smaller effect on deposition of H2AZ. Interestingly, the effect of the KD on H2AZ enrichment was apparent at enhancers bearing high levels of H3K4me1 and was paralleled by the effect on Pol II ( Figure 5 ). We presently do not understand this phenomenon but note that the H2AZ enrichment is lower at enhancers containing high levels of H3K4me1 and Pol II enrichment is lower at enhancers compared with promoters. Although we observed a mild decrease in H2AZ upon EP400 KD, we note that the metazoan homologue of yeast SWR1, SRCAP also exchanges H2AZ for H2A in vitro and in vivo. It is plausible that SRCAP is redundant with EP400 accounting for the smaller effect of KD on H2AZ versus H3.3 ( Wong et al., 2007 ). Chd1 and Chd2 have been reported to control H3.3 incorporation at active chromatin in metazoans ( Konev et al., 2007 ; Siggens et al., 2015 ). Chd1 bears a SWI-like ATPase with adjacent chromodomains that recognize H3K4me3 near the TSS [reviewed in ( Persson and Ekwall, 2010 )]. Mammalian Chd1 binds with higher avidity to H3K4me3 chromatin in vitro. Moreover, Chd1 is recruited to chromatin via Mediator ( Khorosjutina et al., 2010 ; Lin et al., 2011 ) and its transcriptional stimulatory effect is largely on H3K4me3 chromatin in vitro. Thus, Chd1 is an H3K4me3 and Mediator stimulated event, whereas EP400 recruitment is neither, so we did not pursue Chd1’s role in variant chromatin transcription. In conclusion, our study combines biochemical and cell-based reporter assays, with genome-wide analyses. In totality, the data provide a link between H3.3 deposition, EP400 binding and transcription.
Experimental Procedures
Protein Purification Esa1 and Gcn5 were expressed and purified from E. coli as described ( Barrios et al., 2007 ). The EP400 cDNA was overexpressed using baculovirus infection of Sf9 cells for 72h. Sf9 extracts were prepared as described, incubated with HA-beads, and EP400 eluted in 1 mg/ml HA peptide ( Lin et al., 2011 ). In Vitro Transcription and Immobilized Template (IT) Assays The preparation of chromatinized G5E4T, normalization of chromatin amounts, in vitro transcription reactions on chromatin templates, immunodepletion and IT procedures have been described in our previous publications ( Black et al., 2006 ; Lin et al., 2011 ). Briefly, transcription or IT assays were carried out on 50 ng of chromatin templates that were bound with saturating levels of GAL4-VP16. Where indicated, immobilized chromatin was acetylated with 20 ng of Esa1, 20 ng Gcn5 and 1 mM Acetyl-CoA for 45 min. Transcription was measured by 32 P-primer extension and visualized on an Amersham Biosciences Typhoon 9400. Immunoblots were processed on a LI-COR Odyssey and quantitated by Image Studio 2. Antibodies included pan-acetyl Histone H3 (Active Motif), MED23 (BD Pharmingen), Pol II (QED Bioscience), TFIIB (Tantin et al., 1996), EP400, Tip60 (Abcam), ASH2L, CHD1 (Bethyl Laboratories) and all others from Santa Cruz Biotechnology. ChIP Analysis in the Tet-VP16 U2OS Cell Line Stable U2OS Tet-On Luciferase cells were obtained from Clontech Laboratories Inc., Mountain View, CA, and cultured in DMEM with 10% Tet-system approved FBS (Clontech). After addition of 1 μg/ml Doxycycline, transcription and ChIP assays were as previously described ( Black et al., 2006 ). Antibodies used in ChIP included MED26 (Santa Cruz Biotechnology, sc-48776), Pol II (QED Bioscience), H3.3 (Abnova, H00003021-M01), EP400 (Abcam, ab70301), H2AZ (Abcam, ab4174), and H3K4me1 (Abcam, ab8895). Antibody against H3K18ac was in house ( Ferrari et al., 2008 ). Immunoprecipitates were washed and DNA was quantitated by real-time semi-quantitative PCR. Alternately, mRNA was isolated and quantified by Reverse Transcriptase-PCR. siRNA Knockdown of EP400 The web based program DEQOR was used to determine unique regions of EP400 ( Henschel et al., 2004 ) to minimize cross-silencing activities. esiRNAs were produced according to the method described ( Fazzio et al., 2008 ) and validated by PAGE. Mock siRNA was prepared essentially as described for EP400 but was from GFP. Cells were treated for 72 hours with either EP400 or mock siRNA. Chromatin immunoprecipitation, Library Preparation and Analysis ChIP-Seq was performed as described ( Ferrari et al., 2012 ). Libraries were prepared with a KAPA LTP kit using 2 ng DNA, sequenced by the Illumina HiSeq 2000 platform (50 bp reads) and mapped to human genome version hg19 using bowtie 0.12.9 ( Langmead et al., 2009 ) with 100-bp windows for all proteins except EP400 (50 bp). Windows with p values <1.0×10 −3 were deemed significant peaks. Average ChIP-Seq signals, 3kb upstream and downstream of TSS, and metagene plots of average ChIP-Seq signals across gene bodies were calculated by CEAS (Shin et al., 2009). Heatmaps plot p values of enrichment. Significance between Mock and KD was calculated using the Wilcoxon Rank-Sum test. RNA Extraction, mRNA-Seq Library Preparation and Analysis RNA extraction, library preparation and Seq analysis were performed essentially as described in ( Ferrari et al., 2012 ). Libraries were sequenced as above to obtain single end 50 bp-long reads. Reads were aligned as for ChIP but using tophat 2.0.8 ( Trapnell et al., 2009 ). FPKM (fragment per kilobase per million mapped reads) values were determined using Cuffdiff 2.0.2 ( Trapnell et al., 2010 ). In Vitro Histone Exchange Assays Histone exchange assays were performed as described ( Mizuguchi et al., 2004 ). 150 ng of immobilized G5E4T DNA or 300 ng of immobilized canonical chromatin was incubated in 100-μl with 40 ng recombinant EP400 and 300 ng of H3.3-H4 tetramers, or H2AZ-H2B dimers with 1 mM ATP for 1 h at 37°C. Chromatin was captured on beads, washed twice with exchange buffer (0.4 M KCl) and once with buffer containing 70 mM KCl. Bound proteins were eluted with SDS-PAGE loading buffer. The washes and supernatants were combined, TCA precipitated, and fractionated on SDS gels alongside bound fractions, and immunoblotted.
Protein Purification
Esa1 and Gcn5 were expressed and purified from E. coli as described ( Barrios et al., 2007 ). The EP400 cDNA was overexpressed using baculovirus infection of Sf9 cells for 72h. Sf9 extracts were prepared as described, incubated with HA-beads, and EP400 eluted in 1 mg/ml HA peptide ( Lin et al., 2011 ).
In Vitro Transcription and Immobilized Template (IT) Assays
The preparation of chromatinized G5E4T, normalization of chromatin amounts, in vitro transcription reactions on chromatin templates, immunodepletion and IT procedures have been described in our previous publications ( Black et al., 2006 ; Lin et al., 2011 ). Briefly, transcription or IT assays were carried out on 50 ng of chromatin templates that were bound with saturating levels of GAL4-VP16. Where indicated, immobilized chromatin was acetylated with 20 ng of Esa1, 20 ng Gcn5 and 1 mM Acetyl-CoA for 45 min. Transcription was measured by 32 P-primer extension and visualized on an Amersham Biosciences Typhoon 9400. Immunoblots were processed on a LI-COR Odyssey and quantitated by Image Studio 2. Antibodies included pan-acetyl Histone H3 (Active Motif), MED23 (BD Pharmingen), Pol II (QED Bioscience), TFIIB (Tantin et al., 1996), EP400, Tip60 (Abcam), ASH2L, CHD1 (Bethyl Laboratories) and all others from Santa Cruz Biotechnology.
ChIP Analysis in the Tet-VP16 U2OS Cell Line
Stable U2OS Tet-On Luciferase cells were obtained from Clontech Laboratories Inc., Mountain View, CA, and cultured in DMEM with 10% Tet-system approved FBS (Clontech). After addition of 1 μg/ml Doxycycline, transcription and ChIP assays were as previously described ( Black et al., 2006 ). Antibodies used in ChIP included MED26 (Santa Cruz Biotechnology, sc-48776), Pol II (QED Bioscience), H3.3 (Abnova, H00003021-M01), EP400 (Abcam, ab70301), H2AZ (Abcam, ab4174), and H3K4me1 (Abcam, ab8895). Antibody against H3K18ac was in house ( Ferrari et al., 2008 ). Immunoprecipitates were washed and DNA was quantitated by real-time semi-quantitative PCR. Alternately, mRNA was isolated and quantified by Reverse Transcriptase-PCR.
siRNA Knockdown of EP400
The web based program DEQOR was used to determine unique regions of EP400 ( Henschel et al., 2004 ) to minimize cross-silencing activities. esiRNAs were produced according to the method described ( Fazzio et al., 2008 ) and validated by PAGE. Mock siRNA was prepared essentially as described for EP400 but was from GFP. Cells were treated for 72 hours with either EP400 or mock siRNA.
Chromatin immunoprecipitation, Library Preparation and Analysis
ChIP-Seq was performed as described ( Ferrari et al., 2012 ). Libraries were prepared with a KAPA LTP kit using 2 ng DNA, sequenced by the Illumina HiSeq 2000 platform (50 bp reads) and mapped to human genome version hg19 using bowtie 0.12.9 ( Langmead et al., 2009 ) with 100-bp windows for all proteins except EP400 (50 bp). Windows with p values <1.0×10 −3 were deemed significant peaks. Average ChIP-Seq signals, 3kb upstream and downstream of TSS, and metagene plots of average ChIP-Seq signals across gene bodies were calculated by CEAS (Shin et al., 2009). Heatmaps plot p values of enrichment. Significance between Mock and KD was calculated using the Wilcoxon Rank-Sum test.
RNA Extraction, mRNA-Seq Library Preparation and Analysis
RNA extraction, library preparation and Seq analysis were performed essentially as described in ( Ferrari et al., 2012 ). Libraries were sequenced as above to obtain single end 50 bp-long reads. Reads were aligned as for ChIP but using tophat 2.0.8 ( Trapnell et al., 2009 ). FPKM (fragment per kilobase per million mapped reads) values were determined using Cuffdiff 2.0.2 ( Trapnell et al., 2010 ).
In Vitro Histone Exchange Assays
Histone exchange assays were performed as described ( Mizuguchi et al., 2004 ). 150 ng of immobilized G5E4T DNA or 300 ng of immobilized canonical chromatin was incubated in 100-μl with 40 ng recombinant EP400 and 300 ng of H3.3-H4 tetramers, or H2AZ-H2B dimers with 1 mM ATP for 1 h at 37°C. Chromatin was captured on beads, washed twice with exchange buffer (0.4 M KCl) and once with buffer containing 70 mM KCl. Bound proteins were eluted with SDS-PAGE loading buffer. The washes and supernatants were combined, TCA precipitated, and fractionated on SDS gels alongside bound fractions, and immunoblotted.
Acknowledgments
This work was supported by CIHR grant MOP-64289 to J.C. and NIH grants GM074701 to M.F.C. and CA178415 to S.K.K. L.Y was supported by a Ruth L. Kirschstein National Research Service Award GM007185. We thank Kostas Chronis for discussions and technical assistance and Swami Venkatesh for manuscript comments.
Footnotes
GEO Accession The genomics data are accessible through GEO accession number {"type":"entrez-geo","attrs":{"text":"GSE73742","term_id":"73742"}} GSE73742 . Author Contributions S.K.P. and M.F.C. conceived and designed the experiments; S.K.P. performed the experiments. S.K.P. and T.S. performed informatics analyses. C.H. determined MED26 ChIP conditions. L.Y., J.C., and K.J. performed cloning, expression and purification of recombinant proteins. J.C. and K.C. provided critical reagents and technical advice. S.K.P. and M.F.C. wrote the manuscript with input from J.C. and S.K.K. Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Article Images (0)
×![]()
Bioz is the world’s most comprehensive AI search engine for scientific experimentation. The Bioz search engine offers researchers billions of data-driven product, technique, and protocol recommendations. Bioz taps into the latest advances in Artificial Intelligence (AI) – including Natural Language Processing (NLP), Machine Learning (ML) and Generative AI to mine and structure hundreds of millions of pages of complex and unstructured scientific papers. By harnessing the power of AI, Bioz provides researchers with an unprecedented amount of summarized scientific experimentation knowledge right at their fingertips, ultimately helping to speed up drug discovery and the rate of success in finding cures for diseases. Bioz is used by over 15 Million researchers from over 17,000 different universities and companies in 196 countries.
