Proteogenomics analysis reveals specific genomic orientations of distal regulatory regions composed by non-canonical histone variants

Letters

11 pages
2 views

Please download to get full document.

View again

of 11
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Share
Description
Histone variants play further important roles in DNA packaging and controlling gene expression. However, our understanding about their composition and their functions is limited. Integrating proteomic and genomic approaches, we performed a
Transcript
  METHODOLOGY Open Access Proteogenomics analysis reveals specific genomicorientations of distal regulatory regionscomposed by non-canonical histone variants Kyoung-Jae Won 1,2 † , Inchan Choi 1,2,5 † , Gary LeRoy 3 † , Barry M Zee 3,4 , Simone Sidoli 4 , Michelle Gonzales-Cope 3,4 and Benjamin A Garcia 4* Abstract Background:  Histone variants play further important roles in DNA packaging and controlling gene expression.However, our understanding about their composition and their functions is limited. Results:  Integrating proteomic and genomic approaches, we performed a comprehensive analysis of the epigeneticlandscapes containing the four histone variants H3.1, H3.3, H2A.Z, and macroH2A. These histones were FLAG-tagged inHeLa cells and purified using chromatin immunoprecipitation (ChIP). By adopting ChIP followed by mass spectrometry(ChIP-MS), we quantified histone post-translational modifications (PTMs) and histone variant nucleosomal ratios in highlypurified mononucleosomes. Subsequent ChIP followed by next-generation sequencing (ChIP-seq) was used to map thegenome-wide localization of the analyzed histone variants and define their chromatin domains. Finally, we included inour study large datasets contained in the ENCODE database. We newly identified a group of regulatory regions enrichedin H3.1 and the histone variant associated with repressive marks macroH2A. Systematic analysis identified both symmetricand asymmetric patterns of histone variant occupancies at intergenic regulatory regions. Strikingly, these directionalpatterns were associated with RNA polymerase II (PolII). These asymmetric patterns correlated with the enhancer activitiesmeasured using global run-on sequencing (GRO-seq) data. Conclusions:  Our studies show that H2A.Z and H3.3 delineate the orientation of transcription at enhancers as observedat promoters. We also showed that enhancers with skewed histone variant patterns well facilitate enhancer activity.Collectively, our study indicates that histone variants are deposited at regulatory regions to assist gene regulation. Background The eukaryotic genome is packaged in the nucleus aschromatin, a dynamic arrangement which serves tocompact the DNA. Chromatin structure is highly com-plex as, while packaged, is accessible for selective geneexpression and DNA repair. Moreover, chromatin ishighly dynamic during chromosome condensation pro-cesses such as mitosis and meiosis [1]. The fundamentalunit of chromatin is the nucleosome. Nucleosomes arecomposed of an octamer of histone proteins comprised of two copies each of H2A, H2B, H3, and H4 [2]. Histone N-terminal tails are exposed outside the nucleosomes, andthey are heavily modified by dynamic post-translationalmodifications (PTMs). The deposition of such PTMsmodulates chromatin structure, which directly affects theabovementioned DNA-related events [3,4]. Histone PTMsare also among the major drivers of epigenetic memory, asthey can be inherited after cell division [5]. Aberrations inPTM relative abundance have been found in several dis-eases [6,7], which highlights the direct link between his-tone marks and cell phenotype.In addition to the canonical histones, there are alsoprotein variants encoded by separate genes [8]. These variants play further important roles in DNA packagingand controlling gene expression [9]. For instance, his-tone H2A.Z replaces canonical H2A at some 5 ′  end of both active and inactive genes [10-15]. Recent studies alsoidentified that H2A.Z is enriched at active enhancers, de-stabilizing the local nucleosome structure and facilitating * Correspondence: bgarci@mail.med.upenn.edu † Equal contributors 4 Epigenetics Program, Department of Biochemistry and Biophysics, PerelmanSchool of Medicine, University of Pennsylvania, Philadelphia, PA 19104, USAFull list of author information is available at the end of the article © 2015 Won et al.; licensee BioMed Central. This is an Open Access article distributed under the terms of the CreativeCommons Attribution License (http://creativecommons.org/licenses/by/4.0), which permits unrestricted use, distribution, andreproduction in any medium, provided the srcinal work is properly credited. The Creative Commons Public DomainDedication waiver (http://creativecommons.org/publicdomain/zero/1.0/ ) applies to the data made available in this article,unless otherwise stated. Won  et al. Epigenetics & Chromatin  (2015) 8:13 DOI 10.1186/s13072-015-0005-9  nucleosome removal [16,17]. Histone H3.3 is specially enriched at transcriptionally active genes as well as regula-tory elements [18-22]. Also, unstable H2A.Z/H3.3 double- variant-containing nucleosomes were reported at activepromoters, enhancers, and insulator regions [23]. AnotherH2A variant, macroH2A, is enriched for the inactive Xchromosome, and therefore, it has been mainly associatedwith heterochromatic regions [24,25]. More recently,macroH2A has also been shown to activate genes, al-though it was still closely associated with the silencingmark H3K27me3 [26]. Such complex panorama of histonePTMs and variants calls for further studies to more accur-ately define the combinatorial preferences of histone vari-ants and their function for gene regulation.In order to understand the strategic deposition of histone proteoforms and their functional roles, we quanti-tatively investigated using chromatin immunoprecipitationcoupled to mass spectrometry (ChIP-MS) the compositionat single-nucleosome resolution of histone-variant-containing mononucleosomes. Moreover, such mononu-cleosomes were genome-wide mapped using chromatinimmunoprecipitation followed by next-generation sequen-cing (ChIP-seq). In these experiments, we employed HeLacells expressing either FLAG-tagged canonical histoneH3, H3.3, canonical histone H2A, and H2A.Z. More-over, FLAG-affinity-purified mononucleosomes wereanalyzed by ChIP-MS (Figure 1A, B) to quantitatively determine histone PTM composition [27]. Hence, ourproteogenomic approach allowed us to define chroma-tin domains containing combinations of histone vari-ants. In particular, we mapped and determined thecomposition of domains enriched in histone H3.1 andmacroH2A, which last is known as repressive signature.More importantly, we observed directional profiles of histone variants, where histone H2A.Z occurred aheadof H3.3 in gene enhancers. This directional pattern co-occurred with the enrichment of RNA polymerase II(PolII), suggesting that PolII has orientation at en-hancers and histone variants reflect its transcriptionaldirection. Results and discussion Determination of the relative abundance of histonevariants Nucleosomes contain two copies of each core histonetype (that is, H4, H3, H2A, and H2B). We used quantita-tive MS to analyze the ratio of the different H3 and H2A variants in the FLAG-purified mononucleosomes. Asexpected, canonical H2A was found to be the mostabundant variant in nucleosomes purified with eitherFLAG-H2A or FLAG-H3.1 (Figure 1C and Additionalfile 1: Table S1). H2A.Z was not observed to be enrichedin mononucleosomes purified with histone H3.3, eventhough previous studies indicate that it should beenriched at enhancers of expressed genes [16,17]. Wefound a portion of H2A.Z co-existing with H3.3, but this Figure 1  Analysis of immunoprecipitated mononucleosomes.  Coomassie staining gel representing protein composition of   (A)  FLAG-H2A-and  (B)  FLAG-H3-immunoprecipitated histone samples. NEG represents negative control.  (C)  Relative abundance of canonical histone H2A (gray),H2A.Z (red), macroH2A (green), canonical histone H3 (H3.1 + H3.2, brown), and H3.3 (violet) calculated from the ChIP-MS analysis of H2A FLAG-tagged histones and  (D)  H3 FLAG-tagged histones. Won  et al. Epigenetics & Chromatin  (2015) 8:13 Page 2 of 11  portion is very small and likely quite specialized. Canon-ical H2A makes up about 50% of the H2A population inmononucleosomes purified with FLAG-H2A.Z orFLAG-macroH2A, suggesting that nucleosomes that con-tain these variants are asymmetric, containing one copy of canonical H2A (Figure 1C and Additional file 1: Table S1).Surprisingly, a small fraction of mononucleosomes weredetected to possess both H2A.Z and macroH2A. On theother hand, analysis of H3.3-containing nucleosomesshows that canonical histone H2A is the most abundantform, followed by H2A.Z, and then finally macroH2A(Figure 1C and Additional file 1: Table S1).We also quantified the relative ratios of H3.1 and H3.2 versus H3.3 in the FLAG-purified nucleosomes by usingMS (Figure 1D). To do so, we utilized a peptide that wascommon in the H3.1 and H3.2 variants (a.a. 27 to 40) yet differed from the H3.3 variant by one amino acid(see the  ‘ Methods ’  section). Therefore, we were not ableto distinguish H3.1 from H3.2. From our calculations,H3.1 FLAG-purified nucleosomes contained histone H3 variants with a ratio of roughly 1:3 (H3.3:canonical H3)(Figure 1D and Additional file 1: Table S1). This suggeststhat most of the nucleosomes containing the variantH3.3 are asymmetric, containing one copy of H3.1 orH3.2 partnered with one copy of H3.3. This conclusionwas further supported by the fact that approximately half of the FLAG H3.3-purified nucleosomes containedeither H3.1 or H3.2 (Figure 1D and Additional file 1:Table S1). The most significant observation regardingthe H3 variants was found in the FLAG macroH2A-purified nucleosomes; these nucleosomes contained only approximately 3% H3.3, suggesting that H3.3 is rarely found in nucleosomes containing the repressivemacroH2A variant (Figure 1D and Additional file 1:Table S1). This result was further supported by the factthat Flag H3.3-purified nucleosomes contained very low levels of macroH2A (approximately 1%). Determination of histone PTM relative abundance By using ChIP-MS results (Figure 2 and Additional file 2:Table S2), we investigated the relative abundance of his-tone PTMs in H2A.Z- and H3.3-containing mononucleo-somes. Briefly, we observed an enrichment of active marksin such nucleosomes as compared to the global chromatinlevels and, in particular, to nucleosomes containing thehistone variant macroH2A. For instance, the activatingmark H3K4me2 mark was highly enriched in nucleosomeChIPed with H3.3 or H2A.Z (11.5-fold in H3.3 and 19.8-fold in H2A.Z as compared with the genomic chromatinlevels). This data was interesting considering that H2A.Zwas not enriched in mononucleosomes purified with theH3 variant H3.3, indicating that H3.3 and H2A.Z may oc-cupy distinct chromatin regions marked by H3K4me2.H3K4me3 was found to be enriched almost 30-fold inH2A.Z-purified nucleosomes as compared to global input(Figure 2A and Additional file 2: Table S2), which wasconsistent with previous observations that investigated ac-tive promoters [11]. A similar trend was observed forH3K36me3 in H3.3- and H2A.Z-purified nucleosomes,which was about 7% and 15% of the total histone H3, re-spectively. In genomic chromatin and macroH2A-containing nucleosomes, H3K36me3 was only 4% and 1%,respectively (Figure 2 and Additional file 2: Table S2). Thiswas not surprising, as H3K36me3 is enriched downstreamto the transcriptional start sites (TSSs) of active genes,which are the same genomic regions where H2A.Z isenriched [11]. Interestingly, H3.3-purified nucleosomeswere less enriched in H3K36me3 than nucleosomes puri-fied with H2A.Z, even though both H3.3 and H3K36me3generally mark actively transcribed regions [11,18-21].As compared to the total chromatin, both H3.3- andH2A.Z-purified nucleosomes were enriched for the acti- vating marks H3K9ac and H3K27ac (4.3- and 2.4-foldchanges for H3K9ac and 1.8- and 2.4-fold changes forH3K27ac, respectively) (Figure 2 and Additional file 2:Table S2). We also observed a dramatic enrichment of H4K16ac in nucleosomes purified with H3.3 (approxi-mately 50%) and H2A.Z (approximately 40%) as com-pared to this modification in genomic chromatin(approximately 20%). This confirmed once again whatwe expected, as H3.3, H2A.Z, and H4K16ac are allenriched in gene bodies [28]. Moreover, H4K16ac ishighly enriched in nucleosomes bound by the BET fam-ily bromodomain containing proteins (Brd2, Brd3, andBrd4), which are bound to and assist gene transcriptionby PolII [29]. The enrichment of H4K16ac was presentin tandem with combinations of H4K5ac, H4K8ac, andH4K12ac in H3.3- and H2A.Z-purified nucleosomes(Figure 2B and Additional file 2: Table S2). Finally, therepressive mark H3K27me3 was enriched approximately threefold changes in macroH2A-purified nucleosomes ascompared to genomic chromatin (approximately 18% inmacroH2A-purified  vs.  approximately 6% in genomic orapproximately 5% in H2A.1-purified nucleosomes).Moreover, we observed an inverse relationship betweenH3K4me2/3 and H3K27me3 on nucleosomes purifiedwith either H2A.Z or macroH2A. Conversely, H3K27aclevels in macroH2A ChIPs were depleted (0.26%) ascompared to the H3K27ac levels found in canonicalH2A.Z ChIPs. Taken together, our data demonstrate thatthe trends of the major PTMs we observed in histonesH3 and H4 were similar for H2A.Z and H3.3. Histone variant genome-wide profiles Our proteomic analyses further questioned how the co-occupying histone variants were represented in the gen-ome. For this, we used deep sequencing approaches tomap the FLAG-tagged histone variants in the genome. Won  et al. Epigenetics & Chromatin  (2015) 8:13 Page 3 of 11  First, we asked how the histone variants are positionedaround genes. After sorting the annotated Refseq genesbased on their expression levels, we investigated the his-tone variant levels around the genes. H2A.Z was highly enriched around the annotated TSSs of active genes butabsent in inactive genes (Figure 3A and Additional file 3:Figure S1). This was consistent with the previousgenome-wide surveys [11,23]. The sharp enrichment of H2A.Z marked the two nucleosomes flanking thenucleosome-free regions (NFRs) at active promoters[11,30]. In the gene body, H2A.Z was absent regardlessof their expression levels (Additional file 3: Figure S1).Both H3.1 and H3.3 were depleted around the centerof TSSs of active genes (Figure 3A and Additional file 3:Figure S1). These results differ from the previous study in mouse embryonic stem cells (mESCs) where H3.3showed strong bimodal enrichment at the TSSs of activegenes [16], but they are in agreement to the previousgenome-wide study in HeLa cells [23]. At transcriptiontermination sites (TTSs), we also found a depletion of H3.1 and H3.3 [31]. In general, H3.1 and H3.3 levelscorrelated with gene expression levels in the gene body.H3.3 levels were increased towards the 3 ′  end of thegenes, consistent with previous studies [21,23]. WhileH2A.Z was sharply enriched around TSSs, the enrich-ments of H3.1 and H3.3 were modest. Also, we con-firmed a negative correlation of the macroH2A withgene expression levels [32]. Figure 2  Relative ratio of histone post-translational modifications in FLAG-IP experiments as compared to the global HeLa extract. (A) Calculated relative abundance of single histone PTMs. Log 2  ratio was calculated between each FLAG-IP sample (listed on top of the heat map)and the HeLa input. Single PTMs were sorted by common regulation into a hierarchical tree.  (B)  Log 2  relative ratio of combinatorial histone PTMs,calculated using the same approach. Won  et al. Epigenetics & Chromatin  (2015) 8:13 Page 4 of 11  We then investigated if histone variants are enrichedat promoter distal (>2 kbp from known TSSs) regulatory regions. For this, we retrieved the map of DNaseI hyper-sensitive sites (DHSs) in HeLa cells from ENCODE [33].DHSs potentially demarcate regulatory elements, includ-ing promoters, enhancers, silencers, insulators, and locuscontrol regions [34]. We identified a total of 94,600DHSs using Homer [35]. Among them, 37,073 DHSswere located distal (>2 kbp) to the known TSSs, TTSs,and outside the body of the annotated mRNA and thelong non-coding RNAs (lncRNAs). Monitoring the fourhistone variants (H2A.Z, macroH2A, H3.1, and H3.3)at the distal regulatory regions, we defined 16 groups(Figure 3B). We then further characterized such clus-ters by examining the histone PTMs H3K4me1 andH3K27ac, markers for enhancers [36,37]. Based on theproteomic study (Figure 2), we expected H2A.Z andH3.3 are with activating histone modification marks.Our results showed various combinations of histone variants at these distal regulatory regions. As expected,majority of the distal DHSs were enriched for H3.3and/or H2A.Z as well as H3K27ac indicating that thesehistone variants are important for enhancer function.Also, we found clusters marked by H3.1 (clusters 2 and7) or even with repressive macroH2A mark (cluster 1).Contrary to our proteomic results (Figure 2B), clusters2 and 7 are enriched for H3.1 and activating H3K27acmark at a certain level. The clustering results indicatediverse epigenetic codes composed of histone variantsat distal regulatory regions. Histone variants have symmetric and asymmetric patternsat distal regulatory regions Histone H3.3 and other histone variants were observedto have asymmetric profiles (Figure 3B, C and Additionalfile 3: Figure S2). The clusters 5 to 7 showed that H3.3and H2A.Z were skewed to one side. We also checkedthe average profiles of p300, H3K27ac, and DHSs at eachcluster. The H3K27ac profiles were enriched on the sidewhere the H3.3 peak was located (Additional file 3:Figure S3) even though the skewness was less dramaticas compared to histone variant profiles. The DNaseI andhistone acetyltransferase p300 profiles, marker for en-hancers [37,38], were centered at the DHSs regardlessof the pattern of the histone variants (Additional file 3:Figure S3), confirming that transcriptional co-factorsare not biased to a single direction. Other transcriptionfactors analyzed did not show asymmetric patterns ei-ther (data not shown), further indicating that the skew-ness in the histone variant profiles was independentfrom transcription factors and their co-factors.To further investigate the association of histone vari-ants with gene regulation, we examined PolII occupancy,which was found to be skewed towards the peak of H3.3 Figure 3  Genomic profiles of histone variants. (A)  Distribution of ChIP-seq reads at annotated TSSs (±2 K) and TTSs (±500) and  (B)  at distalregulatory regions. We clustered DHSs located in the intergenic region. We identified 16 groups and rearranged them to 10 clusters based ontheir profiles. Various compositions of histone variants were found. Clusters 5 to 10 are composed of 2 mirroring groups. After clustering basedon histone variants, we aligned histone modification. Histone variants are off-centered for the mirroring clusters (clusters 5 to 10), suggestingorientation at regulatory regions.  (C)  Symmetric and asymmetric profiles of histone variants. Clusters 1 and 4 show symmetric profiles with variouscompositions of histone variants. Cluster 5 shows mirroring asymmetric profile. All profiles for all clusters are shown in Additional file 3: Figure S2. Won  et al. Epigenetics & Chromatin  (2015) 8:13 Page 5 of 11
Advertisement
Related Documents
View more...
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks