The Functional Topography of the Arabidopsis Genome Is Organized in a Reduced Number of Linear Motifs of Chromatin States

Abstract

Chromatin is of major relevance for gene expression, cell division, and differentiation. Here, we determined the landscape of Arabidopsis thaliana chromatin states using 16 features, including DNA sequence, CG methylation, histone variants, and modifications. The combinatorial complexity of chromatin can be reduced to nine states that describe chromatin with high resolution and robustness. Each chromatin state has a strong propensity to associate with a subset of other states defining a discrete number of chromatin motifs. These topographical relationships revealed that an intergenic state, characterized by H3K27me3 and slightly enriched in activation marks, physically separates the canonical Polycomb chromatin and two heterochromatin states from the rest of the euchromatin domains. Genomic elements are distinguished by specific chromatin states: four states span genes from transcriptional start sites (TSS) to termination sites and two contain regulatory regions upstream of TSS. Polycomb regions and the rest of the euchromatin can be connected by two major chromatin paths. Sequential chromatin immunoprecipitation experiments demonstrated the occurrence of H3K27me3 and H3K4me3 in the same chromatin fiber, within a two to three nucleosome size range. Our data provide insight into the Arabidopsis genome topography and the establishment of gene expression patterns, specification of DNA replication origins, and definition of chromatin domains.

Figure 1.

Genome-Wide Annotation of the Arabidopsis Chromatin Defined by Specific Signatures. (A) Optimization of parameters. The smoothing parameter (left panel), number of PCs (middle panel), and window size (right panel) were optimized based on the minimum similarity between clusters. See Methods for further details. (B) Prevalent chromatin states, as a result of a PCA, as described in the text, were defined by a combinatorial pattern of genomic features. They are characterized by a unique combination of values (positive and negative z-score indicate values above or below the average in the genome, respectively) of each of the chromatin features considered. Error bars represent the se of the mean. This number is estimated as the total number of windows divided by the correlation length of the mark considered. (C) Relationship between genomic elements and chromatin states. The overlap (in base pairs) between the indicated genomic elements and each chromatin state was computed and expressed as a percentage. A promoter region of 0.65 kb was considered. Note that TEs are large genomic elements that may have one or more TE genes associated with them. Here, the class TE refers to genomic regions that contain TEs but do not overlap with TE genes. (D) Relationship between gene expression level and chromatin states. RNA sequence reads normalized per kilobase and million reads (RPKM) obtained in whole seedlings (15 d old; see Methods) were computed for each chromatin state. Note the agreement with data presented in (B).

Figure 2.

Representative Genomic loci with Distinct Features Defining Each Chromatin State. Integrated Genome Browser (Nicol et al., 2009) views illustrating genomic regions containing each color-coded chromatin state characterized by their combinatorial profiles of chromatin and DNA sequence features. Note that each panel contains one to four domains of each type (blank regions between them are occupied by other chromatin states that have been omitted in the figure for simplicity).

Figure 3.

ChIP and Sequential ChIP Analyses. (A) Browser views of genomic loci containing unexpected combinations of H3K4me3 and H3K27me3 observed in chromatin state 2 (a to d). The rightmost panel represents a randomly chosen control region with none of the mentioned histone modifications. At the bottom of each panel are depicted the positions of the primers used for the quantitative PCR analysis. (B) Real-time PCR enrichment ratios of the indicated sites for H3K27me3 modification relative to the control region, detected by ChIP. (C) Real-time PCR ratios reflecting the fold enrichment of the analyzed regions after sequential chromatin immunoprecipitations with H3K27me3 antibody and subsequently H3K4me3 antibody. Controls for the ChIP specificity (Control IgG) and for the second ChIP (H3K27me3 → no Ab) are presented. (D) Real-time PCR enrichment ratios of the indicated sites for H3K4me3 modification relative to the control region, detected by ChIP. (E) Real-time PCR ratios reflecting the fold enrichment of the analyzed regions after inverted order of sequential chromatin immunoprecipitations (first H3K4me3 antibody and second H3K27me3 antibody). Error bars in (B) to (E) represent the sd of the duplicates in one representative experiment. Controls for the ChIP specificity (Control IgG) and for the second ChIP (H3K4me3 → no Ab) are also presented.

Figure 4.

Genome-Wide Annotation of the Arabidopsis Chromatin Defined by Specific Signatures. (A) Fraction of the genome (indicated as a percentage in parenthesis) occupied by each of the nine chromatin states. (B) Fraction of chromatin domains (indicated as a percentage in parenthesis) occupied by each of the nine chromatin states. (C) Size distribution of chromatin domains. See text for details. Box and whiskers show the minimum to the maximum of all data in each domain. The bar within the box depicts the median. The sample size for each domain appears in Supplemental Figure 5.

Figure 5.

Localization of Chromatin States and Features Relative to Genomic Elements. (A) The distribution of each chromatin state was determined around the TSSs (left panel) and the TTSs (right panel) of Arabidopsis genes. Colocalization analysis was performed taking into account the orientation of the transcription unit. (B) Estimation of the relative enrichment of histone marks and DNA sequence features around the TSS and TTS.

Figure 6.

Accessibility of Chromatin States. Propensity of colocalization of each chromatin state with DNase I accessible and inaccessible chromatin fractions, as described (Shu et al., 2012), compared with genome average probability/to random. [See online article for color version of this figure.]

Figure 7.

Transition Properties between Chromatin Domains. (A) A karyotype view of Arabidopsis showing the relative location of color-coded chromatin domains identified in the five chromosomes. Asterisks indicate the position of centromeres. (B) Frequency of transition between chromatin states. Bar graphs for each chromatin state show the conditional probability of a given state having another as a neighbor. See text for further details. (C) Network propensity diagram of the frequency of transition between the nine chromatin states. Diamonds and circles represent AT-rich and GC-rich states, respectively. Symbol size represents the deviation in GC content with respect to the average genomic content. Circles are states with GC content larger than average, and diamonds are states with low GC content. The thickness of lines connecting chromatin states is proportional to the propensity of transition between two given states.

Figure 8.

Local Relationships of Chromatin States. (A) The distribution of each chromatin state determined around the center of the domain taking into account the transcription-based orientation of each domain. (B) Genome browser view of the color-coded chromatin domain annotation over an ∼50-kb region of the Arabidopsis Chromosome 1 (coordinates correspond to the TAIR10 version). Representative chromatin states in gene-dense and gene-poor regions, the chromatin state transition propensities, and the chromatin features of specific loci are highlighted in this epigenomic landscape.

Figure 9.

Domain Combinations. (A) Most frequent combinations of chromatin states according to propensities between neighbors (total number in the genome at the right side of the panel). (B) Relationship between chromatin domain combination and gene length (in kilobases; indicated at the right side of the panel). The size of each combination was made proportional to the gene length. Note that combinations containing the pair 7-6 tend to be associated with longer genes.