Nucleosome rotational setting is associated with transcriptional regulation in promoters of tissue-specific human genes
© Hebert and Roest Crollius; licensee BioMed Central Ltd. 2010
Received: 12 November 2009
Accepted: 12 May 2010
Published: 12 May 2010
The position of a nucleosome, both translational along the DNA molecule and rotational between the histone core and the DNA, is controlled by many factors, including the regular occurrence of specific dinucleotides with a period of approximately 10 bp, important for the rotational setting of the DNA around the histone octamer.
We show that such a 10 bp periodic signal of purine-purine dinucleotides occurs in phase with the transcription start site (TSS) of human genes and is centered on the position of the first (+1) nucleosome downstream of the TSS. These data support a direct link between transcription and the rotational setting of the nucleosome. The periodic signal is most prevalent in genes that contain CpG islands that are expressed at low levels in a tissue-specific manner and are involved in the control of transcription.
These results, together with several lines of evidence from the recent literature, support a new model whereby the +1 nucleosome could be more efficiently disassembled from gene promoters by H3K56 acetylation marks if the periodic signal specifies an optimal rotational setting.
Nucleosomes, composed of 147 bp of DNA wrapped around a histone octamer, play a fundamental role of compacting DNA molecules inside the nucleus of eukaryotic cells , but also in the regulation of gene expression [2, 3]. Elucidating the molecular mechanisms that specify the position of nucleosomes in a genome is important to understand their role at the crossroads of essential cellular functions.
Factors influencing nucleosome positioning likely include DNA sequence-based information (either to specify a favorable or unfavorable DNA structure or to allow for DNA-histone interactions), contacts between neighboring nucleosomes, and chromatin remodeling proteins. The extent and the modalities of these contributions are still being investigated, and different models have been proposed to explain whole genome nucleosome mapping data in different organisms [4–7]. These results, while primarily focusing on the translational positions of nucleosomes along the DNA molecule, also show that the rotational position of the histone octamer with respect to the DNA molecule is important. High-resolution maps indicate that individual nucleosomes tend to settle at approximately 10-bp intervals around an average position in the genome [4, 6, 8]. Histone cores, when forming a nucleosome with the DNA, thus appear to locally select one of several alternative positions on the DNA, as long as they are separated by distances multiple of a helical turn. Importantly, selecting one position rather than the next will translate the nucleosome by 10 bp, but will not change the rotational angle of the histone core with respect to the DNA molecule and its molecular environment. To wedge histones in their preferred rotational setting, the main theoretic constraint is a periodic occurrence of specific dinucleotides at approximately 10-bp intervals in phase with nucleosome positions [9–11]. This signal is significantly different between species. In yeast, it has been characterized as periodic frequencies of dinucleotides containing only adenine and/or thymidine (WW dinucleotides), with antiphased periodic frequencies of dinucleotides containing cytidines and/or guanines (SS dinucleotides) . In mammalian genomes, the most consistent 10-bp periodic signal is composed of periodic purine dinucleotides (A or G, abbreviated RR), with antiphased pyrimidine dinucleotide frequencies (C or T, abbreviated YY) [13–16], although other combinations of di- and trinucleotides have also been observed [17, 18].
In yeast, high resolution mapping of nucleosomes containing the H2A.Z histone variant, which is typically found in nucleosomes flanking the transcription start site (TSS) of genes [19, 20], led to a model where this rotational setting could be important to present the histone H3 tail in a favorable position at the promoter, or to expose transcription factor binding sites at the nucleosome surface . In the human genome, a high-resolution map of H2A.Z nucleosomes recently led to the conclusion that, in contrast to the yeast genome, a pronounced 10-bp periodicity of specific dinucleotides is absent  near the TSS. Here we examine sequences flanking human TSSs, and we find that a 10-bp periodicity of the same magnitude as that seen in yeast, but of RR rather than WW dinucleotides, does coincide with the first nucleosome after the TSS (+1 nucleosome). Importantly, the signal is specifically in phase with the TSS, suggesting a direct link between transcription and the +1 nucleosome. We analyze the periodic signal with respect to CpG island density, gene expression level and breadth, gene functional annotations, and histone modification marks. We conclude that the periodic signal is likely to play a role in setting the rotational angle of the histone core in the +1 nucleosome, and we propose a model where nucleosome interacting proteins, such as the EP300 histone acetylase, may efficiently trigger histone disassembly prior to RNA polymerase II (RNA pol II) elongation if the rotational setting of the nucleosome is optimal.
A periodic dinucleotide frequency in phase with the TSS coincides with +1 nucleosomes
After the TSS, the frequencies of both G and C remain elevated for approximately 200 bases, thus forming a plateau, before slowly decreasing. Closer examination of the nucleotide composition across the plateau reveals a striking pattern of oscillating frequencies of all four nucleotides, with A and G in phase, and C and T shifted by 5 bp in counter phase (Figure S1A in Additional file 1). The period of the regular pattern is approximately 10 bases and the purine nucleotide peaks are separated from the TSS by a distance multiple of 10 bases, thus residing on the same side of the DNA double helix as the TSS. To better characterize the signal, we analyzed the period of the 16 possible dinucleotide frequencies using discrete Fourier transform (DFT; Figure S1B in Additional file 1; see Materials and methods) and found that mainly purine-purine (RR) and pyrimidine-pyrimidine (YY) dinucleotides contribute to the periodic signal (Figure 1a, inset) in phase and counter-phase, respectively, with the TSS. Randomly shifting the sequences by 1 to 9 bases relative to the TSS completely abolishes the signal (average power spectral density (PSD) magnitude at 10 bp = 0.015; P-value = 2.2 × 10-16, Wilcoxon rank sum test).
If this signal is linked to nucleosome positioning, it should coincide with experimentally defined nucleosome positions from genome-wide mapping efforts. To verify this, we realigned the sequence tags from a recent ChIP-seq experiment aiming at defining the positions of all nucleosomes in human CD4+ cell lines , and we focused on the region immediately downstream of the TSS positions used in our study. Remarkably, the 5' ends of the sequence tags of the forward and reverse strands from the ChIP-seq experiment, which define the boundaries of the nucleosome-bound DNA, show maximal densities that precisely flank the periodic signal (Figure 1b). Thus, DNA sequences of +1 nucleosomes immediately downstream of human TSSs display periodic purine-purine (RR) and pyrimidine-pyrimidine (YY) frequencies.
The periodic signal is correlated with CpG islands
Despite our attempts, the periodic RR and YY signal cannot be detected in individual sequences beyond those periodic dinucleotides one would expect by chance alone, even using standard autocorrelation analysis (data not shown). This lack of significant periodic dinucleotide patterns in individual human H2A.Z sequences has been noted previously using autocorrelation analysis, in contrast to yeast nucleosomal sequences, where periodic patterns appear readily  using these approaches. However, a more sensitive autocorrelation analysis, called autocorrelation spectral estimation, recently showed that 10- and 11-bp periodic AA/TT dinucleotide signals exist in human nucleosomal sequences, while the 11-bp signal is specific to the regions flanking the TSS .
In each group of promoters, we performed a DFT analysis on each of the 16 dinucleotide average frequency profiles between positions +40 and +190 after the TSS (Figure S3 in Additional file 1). A differential comparison between the sets of promoters with and without CpG islands (Supporting information in Additional file 1) should identify those dinucleotides that contribute most to the periodic pattern. Interestingly, in CpG island-containing promoters, GA and AG rank highest among RR dinucleotides, and their complementary CT and TC rank highest among YY dinucleotides (Table S1 in Additional file 1). Notably, dinucleotides AA, TT and TA, which show strong periodic patterns in yeast nucleosome-bound DNA [4, 12], do not contribute to the periodic pattern seen here in human CpG-containing promoters. Within promoters with CpG islands, the strength of the periodic signal is not, however, correlated with the overrepresentation of CpG dinucleotides (Supporting information in Additional file 1).
The periodic signal is most prevalent in tissue-specific genes involved in transcription control
EP300 activity is correlated with increased periodic RR/YY dinucleotides
Genes coding for tissue-specific transcription factors are themselves highly regulated, and given their significant association with a nucleosome rotational positioning signal, we hypothesized that the control of their transcription and information carried by the first nucleosome are somehow connected. Histone modifications are obvious candidates for this potential connection. Histones transiently harbor acetylation and methylation marks deposited by chromatin-modifying enzymes recruited by a diverse array of proteins. One such modifying enzyme is EP300, which directly associates with the pre-initiation complex that includes RNA Pol II , and also binds DNA at a known consensus sequence . EP300 is known to acetylate histones at the following sites: H3K14, H3K18, H4K5, H4K8, H2AK5, H2BK12, H2BK15 . Of these seven marks, six were recently part of a genome-wide mapping of histone modifications in human CD4+ cells . We first tested for the presence of EP300 DNA binding sites in the 13,622 TSSs studied here, and found that they are significantly associated with genes where the first nucleosome carries at least one of the six acetylation marks (P-value = 3 × 10-5, randomization test), in line with expectations. Second, we also searched for the EP300 DNA binding site in all 13,622 TSSs independently of their histone modification status and found that it is significantly associated with the periodic 10-bp RR frequency signal (P-value = 1 × 10-3, randomization test). Third, the intensity of histone acetylations by EP300 on the first nucleosome, as measured by the ChIP-seq sequence tag counts, is also correlated with an increasing magnitude of the periodic signal (P-value = 2 × 10-15, Pearson correlation test; Figure S4 in Additional file 1). Most strikingly, this is also verified for an acetylation mark recently attributed to EP300 on H3K56 , in the globular domain of histone H3. Using recent ChIP-chip results obtained using H3K56ac in the human genome , we show here that the level of H3K56 acetylation is correlated with an increased 10-bp periodicity (Figure 3c, d; low H3K56ac enrichment ratio group versus high H3K56ac enrichment ratio group P-value = 2.2 × 10-16, one-sided Wilcoxon rank sum test). This evidence strongly supports the above hypothesis that a histone-modifying enzyme such as EP300 involved in the first steps of transcription elongation may require a specific rotational setting of the first nucleosome to efficiently carry out its functions (see Discussion).
Conservation of the periodic signal in eukaryotic genomes
The periodic signal observed here appears to be universally present in eukaryotes, albeit involving different dinucleotides. The same periodic RR/YY dinucleotide frequency is seen in human and mouse promoters, but interestingly the medaka fish Oryzias latipes displays a strong periodic signal contributed by AA and TT dinucleotides downstream of the TSS, similar to yeast (Figure S5 in Additional file 1). In yeast, however, the periodic signal appears shorter and is immediately downstream of the TSS , instead of being shifted to the +40 position as in vertebrates.
We describe here a new 10-bp periodic signal present downstream of human TSSs that is concentrated in genes that possess CpG islands, that are expressed at low level in a tissue specific pattern, and that are enriched in functions related to transcription control. Importantly, the signal is centered over the position of experimentally mapped nucleosomes. This result contrasts with a recent study describing the mapping of H2A.Z-containing nucleosomes in the human genome, which concluded that such a periodic signal is essentially absent in human promoters, whereas it had been previously observed in yeast . However, this former study aligned promoters on the predicted +1 nucleosome dyad position, not on experimentally annotated TSSs as here. Tolstorukov et al.  discuss the possibility that a periodic dinucleotide profile may arise in the average frequencies of a set of sequences, even if the periodic signal is not directly related to nucleosome positioning. Such a signal may occur if, for example, a short motif has strong nucleosome positioning properties, but would still allow the histone core to shift by a few base pairs along the sequence to settle in the most favorable configuration in terms of deformation energy cost. Once sequences are obtained by the ChIP-seq technology and aligned at the dyad, their average nucleotide profile may theoretically show such a periodic pattern as a consequence of nucleosome rotational positioning rather than as a cause. Here, however, we align nucleosome sequences independently of the ChIP-seq technology, using the TSS as sole reference. The above scenario may only be applicable to our data if a strong nucleosome positioning motif is itself aligned to the TSS, unrelated to the periodic pattern which, in this case, would be secondary to the motif. Even under this non-parsimonious scenario, however, the conclusion that the rotational setting of the nucleosome is linked to the TSS remains unchanged.
Our work thus underlines a tight coupling between the periodic signal and transcription. We show that the strength of the periodic signal can be correlated with promoters that contain EP300 binding sites, and histones of the +1 nucleosome that are acetylated at residues known to be targets of EP300. Based on these results, we propose a theoretical model that explains how EP300 may efficiently trigger transcription elongation in genes that require rapid and coordinated expression.
A different model was recently proposed to account for H2A.Z-related dinucleotide periodicities near the yeast TSS [3, 4]. In this model, the preferred rotational setting exposes transcription factor binding sequences on the surface of the nucleosome that would otherwise be facing the histone core. Binding of transcription factors would play a role in regulating the translational displacement of the nucleosome, which may be important for gene activation. While our findings are not incompatible with this model developed in yeast, we did not find evidence for specific periodic transcription factor binding site occurrences downstream of human promoters (Supporting information and Figures S9 and S10 in Additional file 1).
Several observations may explain why one or several RR/YY dinucleotides placed at positions separated by multiples of 10 bp along the wrapped DNA can direct the histone core to settle in a specific position and thus specify the rotational setting of the nucleosome. These include: strong stacking interactions between purines facilitating the collapse of the minor groove, and weaker interactions between the complementary pyrimidines facilitating their deformation in the major groove ; the GG = CC and AG = CT steps are, of all steps, the only two that form cross-chain hydrogen bonds in the minor groove, which is probably a determinant of the energetically more favorable smooth versus kinked bending of the DNA ; and an arginine side-chain is located in the minor groove of all histone-DNA binding sites except for one, where the potential discriminator for direct read out is the adenine C2 group versus the guanine N2 group  (Figure S8 in Additional file 1). However, any structural explanation for the RR/YY periodicity in human and mouse should account for the fact that different eukaryotic species appear to rely on different combinations of dinucleotides in the periodic signal.
The RR and YY periodic signals described here suggest a new model where sequence information is directly exploited to create an optimal spatial topology between at least three entities: the RNA Pol II associated with cofactors and EP300, the DNA molecule and the +1 nucleosome (Figure 5). The convergence of many observations leading to this model is striking, yet it is possible that EP300 and nucleosome rotational orientations are not mechanistically linked as suggested, because EP300 activity may be linked to CpG island-containing TSSs due to their role as transcriptional co-activators. Our ability to design experiments that would directly test the model is limited because we currently lack a good understanding of the structural basis for the rotational preference for specific dinucleotides. In particular, we do not know the minimal number of RR (or YY) dinucleotides in phase with the TSS that would be required to specify this spatial topology, but the model nevertheless suggests that if mutations eliminate the crucial RR (or YY) dinucleotides, elongation may not proceed with the required efficiency and may decrease the expression of the gene, thus potentially causing abnormal phenotypes.
Materials and methods
Transcription start site database
All TSSs were extracted from the DBTSS database version 6, 15 September 2007 . In case TSSs were within 200 bp of each other, we considered the most frequent only. TSSs supported by less than two cDNAs mapping to the exact same position were not considered. Each TSS was mapped to the NCBI36 human genome assembly and assigned to the nearest Ensembl gene (version 49). The final dataset contains 13,622 TSSs associated with 12,028 Ensembl genes.
Power spectral analysis
We applied DFT to compute the PSDs or 'periodograms' of the periodic signals using R and Python/Numpy functions. The periodogram magnitude is the squared modulus of the Fourier coefficient divided by the length of the series. Each PSD area is normalized to 1 before extracting the magnitude of the periodicity at 10 bp. To reduce the noise caused by the small size of the genomic region over which the measures are performed (+40 to +190 after the TSS), we applied a 3-bp smoothing window and multiplied the signal with a Hamming window prior to the DFT analysis.
Alignment to the transcription start site
To test the specificity of the phasing of the signal to the TSS, regions from position +40 to +190 where extracted from all 13,622 sequences and a random number (between 1 and 9) of bases was added at their 5' end to introduce a random shift. The average RR frequency was then measured at each position and used to compute the PSD magnitude at 10 bp. The process was repeated 500 times to obtain a distribution, which was compared to the PSD magnitude at 10 bp of the compositional profile of the real sequences (without shift).
ChIP-seq and ChIP-chip data
Nucleosome tags  were downloaded from the NCBI Short Read Archive (SRA) repository under accession number [SRA:SRA000234]. We considered only the human activated CD4+ T cell experiment. Histone methylation and acetylation marks  were downloaded from the SRA repository - [SRA:SRA000206] and [SRA:SRA000287], respectively. Raw sequences were aligned on the human genome assembly (NCBI36) using the Soap 2.01 alignment tool with default options; we only considered exact matches. To evaluate if the strength of the periodicity and the intensity of the acetylation are correlated (Figure S4 in Additional file 1), we computed the distribution of tag counts in the +40 to +200 region after the TSS for the six histone marks linked to EP300 (see above), for the 12,270 sequences that possessed at least one tag. The distribution was divided into quartiles, the RR periodicity at 10 bp was computed for each quartile and a Pearson correlation test was performed between tag count and magnitude of the periodicity at 10 bp. The H3K56 acetylation data  consist of ChIP-chip results on a 244K Agilent Human promoter microarray using immunoprecipitated DNA sequences bound to H3K56 acetylated nucleosomes. Of the 13,622 TSSs used in our study, 6,518 possessed at least one 244K microarray probe positioned between +40 and +200 bp after the TSS in the region overlapping the +1 nucleosome.
Gene expression and Gene Ontology analyses
Human gene expression data from the HG-U133A and GNF1B Affymetrix chips were obtained from the Genomics Institute of the Novartis Research Foundation . After filtering and remapping of probes (Supplementary information in Additional file 1) we obtained 4,372 genes that were also present in the dbTSS dataset and were used for the analysis. The distribution of the median of the normalized expression levels  across the 72 tissues for each gene showed a bimodal distribution that we partitioned using a Gaussian mixture model (Figure S6 in Additional file 1). The two sets of low (L; 1,846 genes) and high (H; 2,526 genes) expression level were analyzed by randomization tests (see below). The quantification of tissue specificity is described in Supporting information in Additional file 1. The distribution of the tissue specificity scores for the 4,372 genes was divided into 3 groups containing 1,199, 2,159 and 1,014 genes (Figure S7 in Additional file 1) with low, medium and high tissue specificity levels, respectively, and the periodicity was measured for each group by bootstrapping as described for the median expression level. The enrichment of Gene Ontology terms in the H versus the L groups was performed using FatiGO+ software , available through the Babelomics site .
The 7-bp consensus sequence binding site of EP300  (matrix from Transfac [M00033] release 7; Figure S11 in Additional file 1) was searched between positions +1 and +40 after the TSS using the position specific weight matrices method . We computed a local probability score using a sliding window of 7 bp (that is, the consensus size) from -500 bp to +500 bp around the TSS. A match for a putative EP300 binding site occurs if the local probability is higher than the average probability computed along the region (probability density estimation ). A total of 198 sequences have a unique match located between 0 and +40 bp after the TSS, and these were used for further analysis. To further account for possible compositional biases in this region, we performed multiple random shuffling of the position within the matrix and computed the distribution of occurrences in the same region as above. None of the iterations generated a significant number of matches in the 0 to +40 region, confirming that the 198 sequences are highly enriched in specific matches to the EP300 position weight matrix.
In several steps of the analysis described here, we wished to test the strength of the PSD magnitude at 10 bp (see 'Power spectral analysis' section above) on a set of TSS sequences that share a specific property (for example, low gene expression level, EP300 binding, and so on). Because calculating a single average PSD value for the whole set of TSSs that share a given property does not provide any means to calculate statistical significance, we performed random sampling with replacement of a subset of TSSs from this population, and calculated PSD values from each sample based on its average RR frequencies as a proxy for both RR and YY frequencies (the 'RR/YY signal'). The distributions obtained in this way are normal, and can be compared to assess if they are statistically different between two populations of sequences. The size of each random sample used here is composed of between 500 and 1,000 sequences (depending on the initial size of the promoter group). The number of samplings required to reach a normal distribution (P-value < 1 × 10-5, Kolmogorov-Smirnoff test) is between 2,000 and 5,000.
ChIP with DNA sequencing
discrete Fourier transform
power spectral density
- RNA pol II:
RNA polymerase II
dinucleotide composed of purine bases (A or G)
Short Read Archive
transcription start site
dinucleotide composed of A or T
dinucleotide composed of pyrimidine bases (C or T).
We wish to thank David Enard for discussions, Stéphane Lecrom and Alexandra Louis for technical help, Fiona Francis for critical reading of the manuscript and Yoichiro Nakatani and Shinichi Morishita for providing medaka TSS mapping data. This work is funded by the ATIP program of the Centre National de la Recherche Scientifique (HRC) and by the French Ministère de l'Enseignement Supérieur et de la Recherche (CH).
- Woodcock CL: Chromatin architecture. Curr Opin Struct Biol. 2006, 16: 213-220. 10.1016/j.sbi.2006.02.005.PubMedView ArticleGoogle Scholar
- Henikoff S: Nucleosome destabilization in the epigenetic regulation of gene expression. Nat Rev Genet. 2008, 9: 15-26. 10.1038/nrg2206.PubMedView ArticleGoogle Scholar
- Jiang C, Pugh BF: Nucleosome positioning and gene regulation: advances through genomics. Nat Rev Genet. 2009, 10: 161-172. 10.1038/nrg2522.PubMedView ArticleGoogle Scholar
- Albert I, Mavrich T, Tomsho L, Qi J, Zanton S, Schuster S, Pugh B: Translational and rotational settings of H2A.Z nucleosomes across the Saccharomyces cerevisiae genome. Nature. 2007, 446: 572-576. 10.1038/nature05632.PubMedView ArticleGoogle Scholar
- Chung HR, Vingron M: Sequence-dependent nucleosome positioning. J Mol Biol. 2008, 386: 1411-1422. 10.1016/j.jmb.2008.11.049.PubMedView ArticleGoogle Scholar
- Valouev A, Ichikawa J, Tonthat T, Stuart J, Ranade S, Peckham H, Zeng K, Malek JA, Costa G, McKernan K, Sidow A, Fire A, Johnson SM: A high-resolution, nucleosome position map of C. elegans reveals a lack of universal sequence-dictated positioning. Genome Res. 2008, 18: 1051-1063. 10.1101/gr.076463.108.PubMedPubMed CentralView ArticleGoogle Scholar
- Mavrich T, Jiang C, Ioshikhes I, Li X, Venters B, Zanton S, Tomsho L, Qi J, Glaser R, Schuster S, Gilmour D, Albert I, Pugh B: Nucleosome organization in the Drosophila genome. Nature. 2008, 453: 358-362. 10.1038/nature06929.PubMedPubMed CentralView ArticleGoogle Scholar
- Tolstorukov MY, Kharchenko PV, Goldman JA, Kingston RE, Park PJ: Comparative analysis of H2A.Z nucleosome organization in human and yeast genome. Genome Res. 2009, 19: 967-977. 10.1101/gr.084830.108.PubMedPubMed CentralView ArticleGoogle Scholar
- Drew HR, Travers AA: DNA bending and its relation to nucleosome positioning. J Mol Biol. 1985, 186: 773-790. 10.1016/0022-2836(85)90396-1.PubMedView ArticleGoogle Scholar
- Richmond T, Davey C: The structure of DNA in the nucleosome core. Nature. 2003, 423: 145-150. 10.1038/nature01595.PubMedView ArticleGoogle Scholar
- Travers A, Drew H: DNA recognition and nucleosome organization. Biopolymers. 1997, 44: 423-433. 10.1002/(SICI)1097-0282(1997)44:4<423::AID-BIP6>3.0.CO;2-M.PubMedView ArticleGoogle Scholar
- Segal E, Fondufe-Mittendorf Y, Chen L, Thåström A, Field Y, Moore I, Wang J, Widom J: A genomic code for nucleosome positioning. Nature. 2006, 442: 772-778. 10.1038/nature04979.PubMedPubMed CentralView ArticleGoogle Scholar
- Kato M, Onishi Y, Wada-Kiyama Y, Abe T, Ikemura T, Kogan S, Bolshoy A, Trifonov EN, Kiyama R: Dinucleosome DNA of human K562 cells: experimental and computational characterizations. J Mol Biol. 2003, 332: 111-125. 10.1016/S0022-2836(03)00838-6.PubMedView ArticleGoogle Scholar
- Kogan S, Trifonov EN: Gene splice sites correlate with nucleosome positions. Gene. 2005, 352: 57-62. 10.1016/j.gene.2005.03.004.PubMedView ArticleGoogle Scholar
- Kogan SB, Kato M, Kiyama R, Trifonov EN: Sequence structure of human nucleosome DNA. J Biomol Struct Dyn. 2006, 24: 43-48.PubMedView ArticleGoogle Scholar
- Fraser RM, Keszenman-Pereyra D, Simmen MW, Allan J: High-resolution mapping of sequence-directed nucleosome positioning on genomic DNA. J Mol Biol. 2009, 390: 292-305. 10.1016/j.jmb.2009.04.079.PubMedView ArticleGoogle Scholar
- Dalal Y, Fleury TJ, Cioffi A, Stein A: Long-range oscillation in a periodic DNA sequence motif may influence nucleosome array formation. Nucleic Acids Res. 2005, 33: 934-945. 10.1093/nar/gki224.PubMedPubMed CentralView ArticleGoogle Scholar
- Pedersen AG, Baldi P, Chauvin Y, Brunak S: DNA Structure in Human RNA Polymerase II Promoters. J Mol Biol. 1998, 281: 663-673. 10.1006/jmbi.1998.1972.PubMedView ArticleGoogle Scholar
- Barski A, Cuddapah S, Cui K, Roh T, Schones D, Wang Z, Wei G, Chepelev I, Zhao K: High-resolution profiling of histone methylations in the human genome. Cell. 2007, 129: 823-837. 10.1016/j.cell.2007.05.009.PubMedView ArticleGoogle Scholar
- Raisner R, Hartley P, Meneghini M, Bao M, Liu C, Schreiber S, Rando O, Madhani H: Histone variant H2A.Z marks the 5' ends of both active and inactive genes in euchromatin. Cell. 2005, 123: 233-248. 10.1016/j.cell.2005.10.002.PubMedPubMed CentralView ArticleGoogle Scholar
- Wakaguri H, Yamashita R, Suzuki Y, Sugano S, Nakai K: DBTSS: database of transcription start sites, progress report 2008. Nucleic Acids Res. 2008, 36: D97-101. 10.1093/nar/gkm901.PubMedPubMed CentralView ArticleGoogle Scholar
- Touchon M, Arneodo A, d'Aubenton-Carafa Y, Thermes C: Transcription-coupled and splicing-coupled strand asymmetries in eukaryotic genomes. Nucleic Acids Res. 2004, 32: 4969-4978. 10.1093/nar/gkh823.PubMedPubMed CentralView ArticleGoogle Scholar
- Schones D, Cui K, Cuddapah S, Roh T, Barski A, Wang Z, Wei G, Zhao K: Dynamic regulation of nucleosome positioning in the human genome. Cell. 2008, 132: 887-898. 10.1016/j.cell.2008.02.022.PubMedView ArticleGoogle Scholar
- Reynolds SM, Bilmes JA, Stafford Noble W: On the relationship between DNA periodicity and local chromatin structure. Proceedings of the Twelfth International Conference on Research in Computational Molecular Biology (RECOMB): May 18-21; Tucson, Arizona. 2009, Berlin, Heidelberg: Springer-Verlag, 5541: 434-450. Lecture Notes in BioinformaticsGoogle Scholar
- Bird AP: DNA methylation and the frequency of CpG in animal DNA. Nucleic Acids Res. 1980, 8: 1499-1504. 10.1093/nar/8.7.1499.PubMedPubMed CentralView ArticleGoogle Scholar
- Saxonov S, Berg P, Brutlag D: A genome-wide analysis of CpG dinucleotides in the human genome distinguishes two distinct classes of promoters. Proc Natl Acad Sci USA. 2006, 103: 1412-1417. 10.1073/pnas.0510310103.PubMedPubMed CentralView ArticleGoogle Scholar
- von Mikecz A, Zhang S, Montminy M, Tan EM, Hemmerich P: CREB-binding protein (CBP)/p300 and RNA polymerase II colocalize in transcriptionally active domains in the nucleus. J Cell Biol. 2000, 150: 265-273. 10.1083/jcb.150.1.265.PubMedPubMed CentralView ArticleGoogle Scholar
- Rikitake Y, Moran E: DNA-binding properties of the E1A-associated 300-kilodalton protein. Mol Cell Biol. 1992, 12: 2826-2836.PubMedPubMed CentralView ArticleGoogle Scholar
- Kouzarides T: Chromatin modifications and their function. Cell. 2007, 128: 693-705. 10.1016/j.cell.2007.02.005.PubMedView ArticleGoogle Scholar
- Wang Z, Zang C, Rosenfeld JA, Schones DE, Barski A, Cuddapah S, Cui K, Roh TY, Peng W, Zhang M, Zhao K: Combinatorial patterns of histone acetylations and methylations in the human genome. Nat Genet. 2008, 40: 897-903. 10.1038/ng.154.PubMedPubMed CentralView ArticleGoogle Scholar
- Das C, Lucia MS, Hansen KC, Tyler JK: CBP/p300-mediated acetylation of histone H3 on lysine 56. Nature. 2009, 459: 113-117. 10.1038/nature07861.PubMedPubMed CentralView ArticleGoogle Scholar
- Xie W, Song C, Young NL, Sperling AS, Xu F, Sridharan R, Conway AE, Garcia BA, Plath K, Clark AT, Grunstein M: Histone h3 lysine 56 acetylation is linked to the core transcriptional network in human embryonic stem cells. Mol Cell. 2009, 33: 417-427. 10.1016/j.molcel.2009.02.004.PubMedPubMed CentralView ArticleGoogle Scholar
- Field Y, Kaplan N, Fondufe-Mittendorf Y, Moore IK, Sharon E, Lubling Y, Widom J, Segal E: Distinct modes of regulation by chromatin encoded through nucleosome positioning signals. PLoS Comput Biol. 2008, 4: e1000216-10.1371/journal.pcbi.1000216.PubMedPubMed CentralView ArticleGoogle Scholar
- Williams SK, Truong D, Tyler JK: Acetylation in the globular core of histone H3 on lysine-56 promotes chromatin disassembly during transcriptional activation. Proc Natl Acad Sci USA. 2008, 105: 9000-9005. 10.1073/pnas.0800057105.PubMedPubMed CentralView ArticleGoogle Scholar
- Masumoto H, Hawke D, Kobayashi R, Verreault A: A role for cell-cycle-regulated histone H3 lysine 56 acetylation in the DNA damage response. Nature. 2005, 436: 294-298. 10.1038/nature03714.PubMedView ArticleGoogle Scholar
- Xu F, Zhang K, Grunstein M: Acetylation in histone H3 globular domain regulates gene expression in yeast. Cell. 2005, 121: 375-385. 10.1016/j.cell.2005.03.011.PubMedView ArticleGoogle Scholar
- Cho H, Orphanides G, Sun X, Yang XJ, Ogryzko V, Lees E, Nakatani Y, Reinberg D: A human RNA polymerase II complex containing factors that modify chromatin structure. Mol Cell Biol. 1998, 18: 5355-5363.PubMedPubMed CentralView ArticleGoogle Scholar
- Gilmour DS, Lis JT: RNA polymerase II interacts with the promoter region of the noninduced hsp70 gene in Drosophila melanogaster cells. Mol Cell Biol. 1986, 6: 3984-3989.PubMedPubMed CentralView ArticleGoogle Scholar
- Rougvie AE, Lis JT: Postinitiation transcriptional control in Drosophila melanogaster. Mol Cell Biol. 1990, 10: 6041-6045.PubMedPubMed CentralView ArticleGoogle Scholar
- Cheng B, Price DH: Properties of RNA polymerase II elongation complexes before and after the P-TEFb-mediated transition into productive elongation. J Biol Chem. 2007, 282: 21901-21912. 10.1074/jbc.M702936200.PubMedView ArticleGoogle Scholar
- Gilmour DS: Promoter proximal pausing on genes in metazoans. Chromosoma. 2009, 118: 1-10. 10.1007/s00412-008-0182-4.PubMedView ArticleGoogle Scholar
- Rasmussen EB, Lis JT: In vivo transcriptional pausing and cap formation on three Drosophila heat shock genes. Proc Natl Acad Sci USA. 1993, 90: 7923-7927. 10.1073/pnas.90.17.7923.PubMedPubMed CentralView ArticleGoogle Scholar
- Rasmussen EB, Lis JT: Short transcripts of the ternary complex provide insight into RNA polymerase II elongational pausing. J Mol Biol. 1995, 252: 522-535. 10.1006/jmbi.1995.0517.PubMedView ArticleGoogle Scholar
- Corey LL, Weirich CS, Benjamin IJ, Kingston RE: Localized recruitment of a chromatin-remodeling activity by an activator in vivo drives transcriptional elongation. Genes Dev. 2003, 17: 1392-1401. 10.1101/gad.1071803.PubMedPubMed CentralView ArticleGoogle Scholar
- Davey C, Sargent DF, Luger K, Maeder AW, Richmond T: Solvent mediated interactions in the structure of the nucleosome core particle at 1.9 a resolution. J Mol Biol. 2002, 319: 1097-1113. 10.1016/S0022-2836(02)00386-8.PubMedView ArticleGoogle Scholar
- Su AI, Wiltshire T, Batalov S, Lapp H, Ching KA, Block D, Zhang J, Soden R, Hayakawa M, Kreiman G, Cooke MP, Walker JR, Hogenesch JB: A gene atlas of the mouse and human protein-encoding transcriptomes. Proc Natl Acad Sci USA. 2004, 101: 6062-6067. 10.1073/pnas.0400782101.PubMedPubMed CentralView ArticleGoogle Scholar
- Wu Z, Irizarry RA: Stochastic models inspired by hybridization theory for short oligonucleotide arrays. J Comput Biol. 2005, 12: 882-893. 10.1089/cmb.2005.12.882.PubMedView ArticleGoogle Scholar
- Al-Shahrour F, Minguez P, Tarraga J, Medina I, Alloza E, Montaner D, Dopazo J: FatiGO +: a functional profiling tool for genomic data. Integration of functional annotation, regulatory motifs and interaction data with microarray experiments. Nucleic Acids Res. 2007, 35: W91-96. 10.1093/nar/gkm260.PubMedPubMed CentralView ArticleGoogle Scholar
- Al-Shahrour F, Minguez P, Tarraga J, Montaner D, Alloza E, Vaquerizas JM, Conde L, Blaschke C, Vera J, Dopazo J: BABELOMICS: a systems biology perspective in the functional annotation of genome-scale experiments. Nucleic Acids Res. 2006, 34: W472-476. 10.1093/nar/gkl172.PubMedPubMed CentralView ArticleGoogle Scholar
- Ben-Gal I, Shani A, Gohr A, Grau J, Arviv S, Shmilovici A, Posch S, Grosse I: Identification of transcription factor binding sites with variable-order Bayesian networks. Bioinformatics. 2005, 21: 2657-2666. 10.1093/bioinformatics/bti410.PubMedView ArticleGoogle Scholar
- Scott DW: Multivariate Density Estimation. Theory, Practice and Visualization. 1992, New-York: WileyView ArticleGoogle Scholar
This article is published under license to BioMed Central Ltd. This is an open access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.