Skip to main content
NCBI home page
The Plant Cell logoLink to The Plant Cell
. 2020 May 29;32(8):2457–2473. doi: 10.1105/tpc.19.00716

Genome-Wide Characterization of DNase I-Hypersensitive Sites and Cold Response Regulatory Landscapes in Grasses

Jinlei Han a, Pengxi Wang a, Qiongli Wang a, Qingfang Lin a, Zhiyong Chen a, Guangrun Yu a, Chenyong Miao b, Yihang Dao a, Ruoxi Wu a, James C Schnable b, Haibao Tang a,1, Kai Wang a,1
PMCID: PMC7401015  PMID: 32471863

Compiling of genome-wide DNase I-hypersensitive site maps from root, stem, and leaf tissues of multiple grasses reveals regulatory DNA landscapes and the core cold-stress response regulatory network.

Abstract

Deep sequencing of DNase-I treated chromatin (DNase-seq) can be used to identify DNase I-hypersensitive sites (DHSs) and facilitates genome-scale mining of de novo cis-regulatory DNA elements. Here, we adapted DNase-seq to generate genome-wide maps of DHSs using control and cold-treated leaf, stem, and root tissues of three widely studied grass species: Brachypodium distachyon, foxtail millet (Setaria italica), and sorghum (Sorghum bicolor). Functional validation demonstrated that 12 of 15 DHSs drove reporter gene expression in transiently transgenic B. distachyon protoplasts. DHSs under both normal and cold treatment substantially differed among tissues and species. Intriguingly, the putative DHS-derived transcription factors (TFs) are largely colocated among tissues and species and include 17 ubiquitous motifs covering all grass taxa and all tissues examined in this study. This feature allowed us to reconstruct a regulatory network that responds to cold stress. Ethylene-responsive TFs SHINE3, ERF2, and ERF9 occurred frequently in cold feedback loops in the tissues examined, pointing to their possible roles in the regulatory network. Overall, we provide experimental annotation of 322,713 DHSs and 93 derived cold-response TF binding motifs in multiple grasses, which could serve as a valuable resource for elucidating the transcriptional networks that function in the cold-stress response and other physiological processes.

INTRODUCTION

In eukaryotes, transcription is regulated in a spatial, temporal, and cell type-specific manner via the complex interplay between transcription factors (TFs) and cis-regulatory DNA elements. The precise determination of the genomic locations and properties of regulatory DNA elements has led to a better understanding of the complexity of gene expression in diverse biological systems (Amit et al., 2009; Vierstra and Stamatoyannopoulos, 2016). In the past decade, chromatin immunoprecipitation with high-throughput DNA sequencing (ChIP-seq) has become the gold standard method for genome-wide detection of TF binding sites (Wei et al., 2006; Johnson et al., 2007). However, ChIP assays require both prior knowledge of TFs and high-quality antibodies, making it a substantial undertaking to explore comprehensive sets of DNA-protein interactions in genomes.

graphic file with name TPC_201900716R3_fx1.jpg

Open in a new tab

A well-known general feature of active regulatory chromatin regions that are open or accessible for TFs (i.e., open or accessible chromatin) is increased sensitivity to cleavage by DNase I endonuclease. This accessibility creates a DNase I-hypersensitive site (DHS; Garel and Axel, 1976; Wu, 1980), which has become the hallmark of active regulatory DNAs in eukaryotic genomes (Elkon and Agami, 2017). DHS mapping provides an efficient method for identifying regions of the genome involved in DNA-protein interactions and has long been considered the gold standard for comprehensive identification of transcriptional regulatory DNA elements. The advent of DHS identification combined with high-throughput sequencing (DNase-seq) has enabled genome-wide characterization of the repertoire of diverse cis-regulatory DNA elements in a single experiment (Crawford et al., 2006; Hesselberth et al., 2009), an objective that is intractable with the ChIP-seq assay. In the last decade, large-scale DNase-seq and associated chromatin accessibility analysis technologies, such as formaldehyde-assisted isolation of regulatory elements followed by sequencing (FAIRE-seq; Giresi et al., 2007), micrococcal nuclease digestion; followed by high-throughput sequencing (MNase-seq; Schones et al., 2008), sonication of cross-linked chromatin sequencing (Sono-seq; Auerbach et al., 2009), and assay of transposase accessible chromatin sequencing (ATAC-seq; Buenrostro et al., 2013), have been employed successfully to generate genome-wide open-chromatin maps in humans and animals and have provided detailed insight into the biological basis of the regulation of gene expression (Klemm et al., 2019). However, few genome-scale data sets are available that illustrate the distribution and functions of regulatory DNA elements in plants (Zhang et al., 2014; Jiang, 2015; Sullivan et al., 2015) and are mostly limited to a small number of model organisms across a rather limited set of tissue types and biological conditions (Zhang et al., 2012a, 2012b; Sullivan et al., 2014; Qiu et al., 2016; Rodgers-Melnick et al., 2016; Lu et al., 2017, 2019; Maher et al., 2018; Alvarez et al., 2019; Ricci et al., 2019).

Low temperature is a major environmental cue that affects the growth and development of plants and can significantly reduce crop yields under extreme conditions. Many plants from temperate regions, including various species of grasses, have evolved sophisticated regulatory mechanisms for modulating gene expression to adapt to cold stress (Zhu, 2016). In the model plant Arabidopsis (Arabidopsis thaliana), for example, cold stress rapidly induces the expression of many TF genes, including those encoding APETALA2 domain-containing C-repeat binding factors (CBFs). These TFs in turn affect the reprogramming of the transcriptome, involving more than 20,000 genes (Hannah et al., 2005; Chinnusamy et al., 2007). However, it remains unclear whether transcriptional and regulatory responses to cold stress are conserved or divergent across species (Kenchanmane Raju et al., 2018). The gene regulatory networks that respond to cold stress in nonmodel plants are largely unknown, especially in Poaceae (grasses), which constitute the most economically important plant family (Grass Phylogeny Working Group II, 2012). Grasses are ideal plants for investigating the mechanistic basis of variation in tolerance to cold and freezing stress due to their wide range of adaptations to climates and ecologies across the globe (Clayton and Renvoize, 1986; Prasad et al., 2005).

Here, we conducted a comprehensive analysis of the DHS landscapes in diverse tissues from three different grass species, Brachypodium distachyon, Setaria italica (foxtail millet), and Sorghum bicolor (sorghum), under both control and cold-stress conditions. Our results present an alternative view of the fundamental characteristics of plant open chromatin by comparing multiple species and multiple tissue types in vivo. The dynamics of the DHS landscapes and DHS-derived regulatory elements in response to cold presented here offer valuable resources for elucidating cold-responsive regulatory mechanisms, thereby facilitating the engineering of strong cold resistance in cereal crops via genetic modification of the corresponding regulatory elements.

RESULTS

Genome-Wide Mapping of Open Chromatin in Three Grasses

We performed DNase I digestion using untreated and cold-treated tissues from B. distachyon (inbred line Bd21), S. italica (inbred line Yugu1), and S. bicolor (inbred line BTx623) to identify the genome-wide landscape of open chromatin under normal and cold-stress conditions in grasses. The genomes of these inbred lines have been fully sequenced, and the public reference genotypes were used for each species (Paterson et al., 2009; International Brachypodium Initiative, 2010; Bennetzen et al., 2012). A total of 54 DNase-seq libraries were developed using root, stem, and leaf tissues (three tissues) and with three biological replicates for both the cold-treated and control samples (Supplemental Table 1). We obtained a total of 2.0 billion sequence reads from these 54 libraries, and 882 million sequence reads could be uniquely mapped to the corresponding reference genomes (Supplemental Table 1). Analyses of biological replicates revealed Pearson’s correlation coefficients ranging from 0.81 to 0.97, indicating that the results were highly reproducible (Supplemental Figure 1). We then combined the sequence reads of three replicates to improve the data obtained from the same tissue for increased accuracy of peak calling.

In total, we obtained 99,944, 106,944, and 115,825 distinct DHSs (including both untreated and cold-treated samples), covering 7.0, 4.9, and 2.8% of the genome in B. distachyon, S. italica, and S. bicolor, respectively (Figure 1A; Supplemental Table 1). Of these, 12,606 to 16,329 DHSs (11.8 to 14.2%) were specific to a single tissue, 16,556 to 23,388 DHSs (16.6 to 21.7%) were shared by all three tissues in a given species (referred to as ubiquitous DHSs), and 3390 to 4455 DHSs (3.4 to 3.8%) were shared by two of the three tissues (referred to as common DHSs; Figure 1B). However, we found that leaf and stem tissues consistently shared more DHSs than roots shared with either of these tissues (∼1.2 times; Figure 1B). The close relationship between leaf and stem tissues was further confirmed by hierarchical clustering analysis of genomic location (Supplemental Figure 2), reflecting the similarities of the gene regulatory networks.

Figure 1.

Figure 1.

Open in a new tab

Identification of Open Chromatin in B. distachyon, S. italica, and S. bicolor.

(A) DNase-seq data surrounding a syntenic locus shows both ubiquitous and specific DHSs from the 18 tissue samples in the three species. The DNase-seq results from a previous study (Zhang et al., 2016) for this corresponding syntenic region are shown at the top. Dotted lines indicate homologous genes. DHS peaks are indicated by colored vertical boxes.

(B) Venn diagrams showing pairwise comparisons of DHSs identified from three different tissue types in B. distachyon, S. italica, and S. bicolor.

(C) Protoplast transformation using a construct with GFP expression driven by DHSs. 35S-GFP is a positive control transformed with a construct with GFP expression driven by the 35S promoter. 35S Enhancer-GFP is a negative control transformed with a construct with GFP expression driven by 35S Enhancer only. Due to the lack of a promoter, the 35S Enhancer alone cannot drive the transcription of GFP in Bd21 protoplasts. 35S Enhancer-DHS#1-GFP and 35S Enhancer-DHS#2-GFP are protoplast transformations conducted using a construct with GFP expression driven by the two selected DHSs, namely DHS#1 and DHS#2, respectively. GFP expression in the constructs with DHSs indicates that the DHSs promote gene expression. Bars = 100 µm.

To assess the reliability of our DHS data, we conducted comparative analyses between our data and the published DHS data (Zhang et al., 2016) for the leaf tissues of the same B. distachyon line. Overall, we found good concordance between the two data sets (Figure 1A), with 26,947 (89%) previously identified DHSs also identified by our data (Supplemental Figure 3). However, 33,147 (57%) DHSs that we identified were not defined as accessible regions in the prior study (Supplemental Figure 3). In addition to technical or sample variations, this difference may be attributed to the ∼2.6-fold sequencing depth differences between our and prior data (89 versus 34 million sequencing reads). Given the relatively small fragments that these chromatin accessibility experiments generated and the stringent unique mapping criteria that we adopted, we expect that certain sites might have been skipped by the pipeline, particularly in large and highly repetitive genomes.

To further validate our predicted DHSs, we performed a protoplast-based transient transformation assay (Zhao et al., 2018) to validate the promoter and enhancer functions of DHSs. We randomly selected 15 DHSs from B. distachyon leaves (Supplemental Figure 4) and amplified the target DNAs of these DHSs from Bd21 genomic DNA. The target DNAs were then cloned into reporter gene (GFP) constructs. Of the 15 DHSs examined, 12 (80%) constructs resulted in consistent expression of the reporter gene GFP when the DHSs were integrated (Figure 1C; Supplemental Figure 4).

DHSs Are Enriched around the Transcription Start Site and Closely Associated with Adjacent Gene Expression

We first analyzed the data under regular growth conditions to understand the baseline for the comparisons. In all three grasses, DHSs were distributed throughout the genomes and showed a strong positive correlation with gene density (Supplemental Figure 5A; R2 > 0.85). By plotting the positions of the DHSs relative to nearby genes, we found that DHSs showed a similar distribution pattern among the grasses; most of the DHSs were concentrated around the transcription start site (TSS) regions (Supplemental Figure 5B). In B. distachyon, S. italica, and S. bicolor, at least one DHS was present within 1 kb upstream or downstream of the annotated TSS for 86.5, 78.9, and 81.5% of the annotated gene models, respectively. We then divided the DHSs into two exclusive classes: genic and distal intergenic DHSs (>1 kb from any genes). Genic DHSs include those located upstream (within 1 kb upstream of the TSS) of a gene, within a gene body (including exons, introns, and 5′ and 3′ untranslated regions [UTRs]), and downstream (within 1 kb downstream of the transcription termination site) of a gene (Figure 2A).

Figure 2.

Figure 2.

Open in a new tab

Genomic Distributions of DHSs in Grasses.

(A) Genomic locations of DHSs in various genes. The DHS percentage from each tissue is shown in a circle and is indicated by the size of each circle. Dotted boxes highlighted the 5′ UTR and intergenic DHSs from B. distachyon and other species.

(B) Association of gene expression levels with DHS numbers in B. distachyon. The expressed genes were divided into five quintiles (from high expression [bin1] to low expression [bin5]). Inactive genes (FPKM = 0) were classified as bin6 genes. The y axes show the percentage of genes associated with genic DHSs (left; histograms) and the average DHS number per associated gene (right; lines).

Interestingly, B. distachyon has ∼8% fewer DHSs located in intergenic regions than S. italica and S. bicolor (Figure 2A, dotted box; P < 0.01, odds ratios 0.61 to 0.70, Fisher’s exact test). By contrast, B. distachyon has an ∼5% greater proportion of 5′ UTR DHSs (Figure 2A, dotted box) than the other two species (P < 0.01, odds ratios 1.21 to 1.45, Fisher’s exact test), indicating a gene-proximal distribution of DHSs in the grass with the smallest genome of the three. Tissue examination showed similar DHS distribution trends (Figure 2A), except in the exonic regions, where the roots consistently contained fewer DHSs (10.4 to 13.1%) than leaves (18.9 to 20.0%) and stems (17.2 to 18.9%) in all three species (P < 0.01, odds ratios 0.46 to 0.65, Fisher’s exact test).

To explore the relationship between DHSs and the elevated expression of nearby genes, we generated 308, 354, and 484 million uniquely mapped RNA-seq reads from B. distachyon, S. italica, and S. bicolor, respectively (three replicates of each tissue; Pearson correlation coefficient > 0.86). Active genes, defined as those with fragments per kilobase of exon model per million mapped fragments (FPKM) > 0, were grouped into five quintiles according to expression levels (bin1 to bin5), with genes lacking observed expression (FPKM = 0) placed in a sixth bin. In all the tissues and species, the proportion of DHS-containing genes increased with increasing intensity of gene expression (Figure 2B; Supplemental Figure 6). For example, ∼94% of the most highly expressed genes (bin1; Figure 2B; Supplemental Figure 6) contained genic DHSs. The values decreased from ∼91 to ∼59% for genes expressed at relatively low levels (bin2 to bin5), ending at ∼35% for inactive genes (bin6). Moreover, highly expressed genes frequently contained a high number of DHSs (an average of 2.3 times more in bin1 than in bin5) in all three grasses (Figure 2B; Supplemental Figure 6), suggesting that highly expressed grass genes may be the subjects of complex regulatory machinery involving the binding of multiple TFs. Alternatively, as expected, very highly expressed genes tend to be located in open chromatin regions.

DHSs Show Low Sequence Conservation

Comparative analyses have suggested that the conservation of DHSs among plants is rare, even among B. distachyon, S. italica, and S. bicolor leaves (Maher et al., 2018; Burgess et al., 2019). To further evaluate this observation and examine DHS conservation among stem and root cells, we conducted comparative analyses by mapping DHS regions across genomes. The DHSs fell into three categories (Figure 3A): homologous sequences that were present in and annotated as DHSs in the other species (conserved and shared by both species); homologous sequences that were present in but not annotated as DHSs in the other species (conserved but species-specific); and no homologous sequences were found (not conserved). Comparative analyses revealed that 1430 to 9498 DHSs (2.2 to 14.6%) and 1676 to 9979 DHSs (2.5 to 15.0%) were shared by two species or conserved but species-specific, respectively (Figure 3B), pointing to the low conservation of open chromatin landscapes among the grass tissues examined.

Figure 3.

Figure 3.

Open in a new tab

Conservation of DHSs in Grass Species and Correlation of Gene Expression and DHSs in Grasses.

(A) Scenarios of DHS conservation between species. The homologous regions of DHSs from one species (Species 1) are indicated by dashed lines in the other species (Species 2).

(B) Bar plot representing DHS conservation by pairwise comparisons among tissues. Sb, S. bicolor; Si, S. italica; Bd, B. distachyon.

(C) Correlation between gene expression variation and regulatory complexity in B. distachyon. The diagram shows the distributions of the dynamic ranges of expression for the genes in groups with different complexities.

(D) Fold enrichment of tissue-specific genes in different complexity categories. Fold enrichment in each complexity category was calculated as the proportion of tissue-specific genes/proportion of all types of genes. The proportion of tissue-specific genes is the number of tissue-specific genes in the given complexity category/total number of tissue-specific genes in all complexity categories. The proportion of all types of genes is the number of all types of genes in the given complexity category/total number of all types of genes in all complexity categories. Enriched or Depleted indicates the preference for tissue-specific genes relative to the background. The results show that tissue-specific genes are disproportionately present in high-complexity categories.

Notably, conserved DHSs (conserved and shared and conserved but species-specific DHSs) were always observed at greater frequencies between B. distachyon and S. italica than between B. distachyon and S. bicolor (Figure 3B), although these two pairwise comparisons share the same most common recent ancestor and hence the same length of evolutionary divergence. The repetitive sequence content of the S. bicolor genome is greater than that of the S. italica genome (Paterson et al., 2009; International Brachypodium Initiative, 2010; Bennetzen et al., 2012), perhaps due to the larger number of repeat-associated DHSs in S. bicolor than in the other species (Supplemental Figure 7A). These repeat-associated DHSs were masked when conducting searches for conserved sequences between species. We adjusted our sets of DHSs by excluding repeat-associated DHSs and reran the cross-species comparisons. However, conserved DHSs were still more common between B. distachyon and S. italica than between B. distachyon and S. bicolor (Supplemental Figure 7B). This finding is consistent with the DHS conservation trend in another recent study in the same grasses (Burgess et al., 2019). We speculate that this pattern may result from the higher conservation of noncoding regulatory DNA elements between B. distachyon and S. italica than between B. distachyon and S. bicolor, as was previously observed between rice (Oryza sativa) and S. italica and O. sativa and S. bicolor, two pairwise comparisons that share the same most recent common ancestor as the two pairwise comparisons employed here (Turco et al., 2013).

We were also interested in determining whether there were motifs specific to or enriched in conserved DHSs rather than nonconserved DHSs. To this end, we searched the DHSs for known TF binding motifs from the Cis-BP (Weirauch et al., 2014) and DAP-seq (O’Malley et al., 2016) databases. However, we observed no significant differences in the abundance of these motifs between conserved and nonconserved DHSs (Supplemental Figure 8). To investigate the prevalence of each motif, we ranked the individual motifs by their frequencies in conserved and nonconserved DHSs. Frequency comparisons showed that the majority of motifs detected in conserved DHSs also showed comparative prevalence in nonconserved DHSs (Kendall’s Tau of 0.789 to 0.871; Supplemental Figure 8), indicating that there may be limited functional bias when comparing TFs in conserved and nonconserved DHSs.

Characteristics of DHSs for Genes with High Regulatory Complexity

To define gene regulatory complexity, we assembled an atlas of DHSs by combining the peaks from all three tissues in each species. For any given gene, we counted the number of genic DHSs and defined the regulatory locus complexity as the total number of genic DHSs assigned. We grouped the genes into three equally sized classes based on regulatory complexity (Supplemental Figure 9): low-complexity genes, assigned zero to one DHS; medium-complexity genes, assigned two to three DHSs; and high-complexity genes, assigned four or more DHSs. As one might expect, the genes showed a significant increase in the dynamic range of expression from low to medium to high complexity (P < 2.2e-16, Cohen’s d = 0.33 to 0.48, one-sided t tests for high- versus medium-complexity genes and medium- versus low-complexity genes; Figure 3C; Supplemental Figure 10). Moreover, when examining tissue-specific genes (FPKM > twofold change in contrast to other tissues, q < 0.05), we found a significant enrichment of high-complexity genes (Figure 3D; P < 0.01, odds ratios 1.70 to 1.91, Fisher’s exact test), whereas active tissue-common genes (FPKM > 0 in either tissue, FPKM < twofold change) largely belonged to the medium-complexity gene class (Supplemental Figure 11; P < 0.01, odds ratios 1.14 to 1.33, Fisher’s exact test). Gene Ontology (GO) enrichment analysis revealed that genes with a large number of DHSs (i.e., high-complexity genes) were significantly more likely to be annotated with the terms “TF activity” (enriched in B. distachyon and S. bicolor only) and “protein kinase activity” than the other genes (Supplemental Data Set 1). Moreover, we observed significant overlap in the enriched GO terms among species (Supplemental Figure 12), indicating that high-complexity genes tend to be consistently associated with the same set of biological functions.

Tissue-Specific DHSs Largely Serve as Distal Control Elements

Uncovering the mechanisms responsible for the regulation of tissue-specific gene expression is challenging in both plants and animals. To explore spatial correlation, we examined tissue-specific DHSs from the nearest tissue-specific genes (from the genic region). We found that 17.2% of the tissue-specific genes in B. distachyon, 11.3% in S. italica, and 20.4% in S. bicolor have tissue-specific DHSs (Figure 4A; Supplemental Figure 13A). By contrast, non-tissue-specific genes were significantly less likely to be associated with tissue-specific DHSs (P < 0.01, odds ratios 3.32 to 12.91, Fisher’s exact test; Supplemental Figure 13B). Notably, the proportions increased to 64.5, 45.9, and 64.2%, for B. distachyon, S. italica, and S. bicolor, respectively, with the distance increasing to as high as 50 kb (Figure 4A), indicating that a considerable portion of the tissue-specific genes had potential distal regulators, which may be located as far as 50 kb away. To further validate this hypothesis, we conducted chromatin interaction analysis using Hi-C (high-throughput chromosome conformation capture) data from leaf mesophyll cells of S. italica and S. bicolor (Dong et al., 2017). We found that 58.5 and 92.3% of leaf-specific genes in S. italica and S. bicolor, respectively, had distal-interacting chromatins (10 kb to 2 Mb), exhibiting significantly higher long-range interaction levels than those of non-tissue-specific genes (P < 0.01, odds ratios 1.49 to 1.61, Fisher’s exact test; Supplemental Figure 14). Furthermore, we detected significant enrichment for tissue-specific DHSs in leaf-specific gene-interacting long-range chromatins in both S. italica and S. bicolor (P < 0.01, odds ratios 2.46 and 2.58, Fisher’s exact test; Supplemental Figure 14), indicating a tight correlation between remote tissue-specific DHSs and tissue-specific genes.

Figure 4.

Figure 4.

Open in a new tab

Spatial Correlation of Tissue-Specific DHSs and Tissue-Specific Genes.

(A) Cumulative analysis of tissue-specific genes with tissue-specific DHSs. The proportions of tissue-specific genes with tissue-specific DHSs within cumulative distances were calculated. The distance (represented in kilobases) is the distance from a tissue-specific DHS to the closest tissue-specific gene.

(B) Representative examples of tissue-specific genes with tissue-specific DHSs. The CSLD2 family genes Bradi1g50170, Sobic.010G008600, and Seita.4G008800 showed relatively high expression in roots. A root-specific DHS was identified in the upstream region of the respective gene. The NAC recognition motif CATGTG (color coded) was identified from the root-specific DHSs. DHS peaks are indicated by colored vertical boxes.

Identifying the distal residence of tissue-specific DHSs can help predict the regulators of tissue-specific genes. For example, CSLD2 (Bradi1g50170), a member of the cellulose synthase-like gene family with a role in root hair tip development (Park et al., 2011), had significantly higher expression levels in roots than in leaves and stems in B. distachyon (Figure 4B). We observed two DHSs in the upstream sequence of this gene: one was located in the promoter region and shared by leaves and stems, and the other was root-specific and located 2.3 kb upstream of the gene CSLD2. Interestingly, the CSLD2 orthologs in S. bicolor (Sobic.010G008600) and S. italica (Seita.4G008800) also showed higher expression in roots versus leaves and stems (Figure 4B). Root-specific DHSs were also found in the >2-kb upstream region of each gene. Motif scanning revealed that the root-specific DHSs contain the MYC-like sequence CATGTG and are potentially targeted by NAC TFs (Tran et al., 2004; Figure 4B). In plants, several NAC TFs act as master switches that are capable of activating secondary cell wall synthesis (Bhatia et al., 2017) by increasing the expression of cellulose synthase genes (Valdivia et al., 2013). Therefore, it is expected that these root-specific DHS-associated NAC elements play a role in the differential expression of CSLD2 in roots. Further functional assays should confirm the roles of these motifs embedded in these root-specific DHSs.

Dynamics of Open Chromatin under Cold Stress across Tissues and across Taxa

We conducted DNase-seq analyses using cold-treated tissues for each species to investigate the dynamic of the open chromatin landscape in response to cold stress (Supplemental Table 1; Supplemental Figure 1). To avoid introducing biases due to the differences in the total number of reads used in DHS calling, we performed DHS calling using equivalent effective reads from each sample (Supplemental Figure 15). By comparing the DHSs before and after cold treatment (referred to as natural- and cold-DHS, respectively), we detected 400 to 2,543 DHSs (0.7 to 4.7%) whose numbers varied in response to this treatment in five tissues (B. distachyon stems and roots, S. italica stems, and S. bicolor stems and roots) and 6923 to 24,355 DHSs (13.7 to 38.7%) in the remaining tissues (Supplemental Figure 15). By examining gained and lost DHSs, we detected similar changes (i.e., small DHS numbers in the former five tissues and larger DHS numbers in the four remaining tissues; Supplemental Figure 16). These results illustrate the distinct dynamics of open chromatin under cold conditions among tissues and species.

Restructuring of Open Chromatin during the Cold-Stress Response Is Both Tissue- and Species-Specific

To further characterize the cold-DHSs, we separated these DHSs into two groups: DHSs that occurred only in cold-treated tissues (referred to as cold-induced DHSs) and those that were shared between cold-treated and untreated tissues (referred to as cold-shared DHSs). We identified an average of 35,625 cold-shared DHSs from each tissue, which is 18 times more than the number of cold-induced DHSs (average of 1883 DHSs; Figure 5A). For each grass, an average of 14,801 (∼44%) cold-shared DHSs were also shared by all three tissues and an average of 11,426 (∼33%) were shared by two of three tissues (Figure 5A). GO enrichment analysis using the genes adjacent to cold-shared DHSs revealed enriched terms involved in intracellular molecular transport, cellular metabolic process, and membrane production (Supplemental Data Set 2), pointing to possible regulatory roles of the cold-shared DHSs in these types of genes.

Figure 5.

Figure 5.

Open in a new tab

Dynamics of the Chromatin Accessibility Landscape after Cold Treatment.

(A) Venn diagrams of the number of specific and common cold-shared and cold-induced DHSs in leaf, stem, and root tissues in the three grasses.

(B) Bar plot indicating the proportions of cold-induced DHSs within different conservation categories determined by pairwise comparisons. Sb, S. bicolor; Si, S. italica; Bd, B. distachyon. Blue boxes, conserved and shared DHSs by both species; magenta boxes, conserved but species-specific DHSs; green boxes, not conserved DHSs.

(C) Correlation between DEGs and adjacent cold-induced DHSs. The DEGs in different tissues were identified, and the relative proportions of DEGs that were the closest genes related to cold-induced DHSs in a given tissue are indicated (y axis).

Notably, we did not detect any cold-induced DHSs shared by all three tissues (leaf, stem, and root tissue) in the grass species studied; in addition, only 10 to 145 cold-induced DHSs were shared by two tissues (Figure 5A). Next, we questioned whether the same tissues from different grasses had conserved responsive regulatory networks represented by the cold-induced DHSs. Pairwise comparisons of cold-induced DHSs from the same tissues showed that less than 15 cold-induced DHSs exhibited sequence conservation and were shared by both species (Figure 5B). By contrast, 271 to 5067 (72.4 to 95.7%) cold-induced DHSs did not exhibit nucleotide sequence conservation. These data indicate that changes in accessible chromatins in response to cold were not directly conserved in the grass genomes examined.

Differential Gene Expression Is Associated with Nearby Cold-Induced DHSs

To functionally annotate the cold-induced DHSs, we performed RNA-seq analyses under cold treatment and identified differentially expressed genes (DEGs) by comparing gene transcription between treated and untreated tissues (FPKM > twofold changes, q < 0.05). Totals of 10,895, 9197, and 14,157 DEGs were identified, accounting for 36.1, 31.2, and 46.9% of the expressed genes (FPKM > 0 in at least one of the three tissues) under cold conditions in B. distachyon, S. italica, and S. bicolor, respectively. GO enrichment analysis revealed that the DEGs were enriched in biological processes associated with the cold-stress response (Supplemental Data Set 3).

We then asked to what extent cold-induced DHSs affect adjacent DEG expression, as DHSs showed the tendency to be proximal to targets (Jiang, 2015; Sullivan et al., 2015). To this end, we assigned each DHS to the closest TSS and examined the overlap between assigned genes and DEGs. An average of 20.5% of the cold-induced DHS-adjacent genes were DEGs (Figure 5C). Although only 20.5% of adjacent genes were DEGs, Fisher’s exact test indicated that this frequency was significantly higher than the overall prevalence of DEGs (15.1%, the average percentage of DEGs relative to total genes in each sample; P < 0.05, odds ratios 1.22 to 2.18, Fisher’s exact test). Thus, these results highlight the potential importance of expression regulatory events mediated by cold-induced DHSs.

Known and Novel Regulatory Motifs inside the DHSs

To identify transcriptional regulators that play a role in the cold-stress response, we employed MEME-ChIP analysis (Machanick and Bailey, 2011) to identify overrepresented (E < 0.05) putative binding motifs from the cold-induced DHSs. We identified 93 putative motifs (Supplemental Data Set 4). We compared these DHS-derived motifs with publicly available motifs, including those in the Cis-BP (Weirauch et al., 2014) and DAP-seq (O’Malley et al., 2016) databases, to annotate their functions. We found 60 motifs that matched motifs with known functions (Figure 6A; Supplemental Data Set 4). Among these motifs, some are associated with TFs involved in the cold-stress response, such as CBF1, CBF3, CYTOKININ RESPONSE FACTOR4 (CRF4), and ABSCISIC ACID REPRESSOR1 (ABR1), which act as transcriptional activators to enhance freezing tolerance in plants (Pandey et al., 2005; Zhao et al., 2016; Zwack et al., 2016). Some TF motifs with known function, but not limited to the cold-stress response, were also enriched in cold-treated tissues. For example, the TFs NAC WITH TRANSMEMBRANE MOTIF1 (NTM1) and WRKY62, which function as regulators of cell division (Kim et al., 2006) and pathogen defense (Fukushima et al., 2016), respectively, were also enriched in cold-response DHSs (Figure 6A; Supplemental Data Set 4). In addition, 33 motifs did not match any known functional motifs in plants, suggesting that the roles of a large proportion of TFs putatively involved in regulating the cold-stress response in grass are uncharacterized.

Figure 6.

Figure 6.

Open in a new tab

Characterization of the Cistromes Involved in the Cold-Stress Response in Grasses.

(A) De novo motif discovery in cold-induced DHS sequences. Diagrams show three examples of motifs that were predicted in cold-induced DHSs. Motif similarity and match significance (P and q values) are indicated under each motif.

(B) Venn diagrams showing the number of TFs co-occurring between various tissues and species.

(C) The 17 TFs shared by all nine cold-treated tissues. The heat map depicts the enrichment of each TF in cold-induced DHSs. TFs that were reported to function in plant responses to cold stress (Supplemental Table 2) are indicated in red.

We then addressed the conservation of these cold-induced TFs among various tissues and species. We found significant enrichment for the TFs showing a co-occurrence pattern in all three tissues in a given grass (22/245 in B. distachyon, 58/273 in S. italica, and 139/367 in S. bicolor; Figure 6B; P < 0.01, fold enrichment 4.55 to 8.69, SuperExactTest; Wang et al., 2015). TFs that co-occurred in a given type of tissue in all three species also showed significant enrichment (111/375 in leaves, 97/315 in stems, and 21/225 in roots co-occurred in all three grasses; Figure 6B; P < 0.01, fold enrichment 5.13 to 10.22, SuperExactTest). These findings suggest that regulators of the cold-stress response are generally conserved across these grasses in a tissue-specific manner.

We identified 17 TFs that were shared among all three cold-treated tissues in all three species (Figure 6C). Each of the 17 TFs was detected in at least 17 (17 to 852) cold-induced DHSs and collectively accounted for more than 86 (86 to 1,715) of the cold-induced DHSs (Figure 6C). Some of the 17 TFs, including ERF6 (Wang et al., 2013), CRF4 (Zwack et al., 2016), and ERF105 (Bolt et al., 2017), are known activators of low-temperature-responsive genes (Supplemental Table 2). In addition to these ubiquitous cold-induced TFs, some TFs were tissue-specific (64 in leaves, 20 in stems, and 22 in roots) or species-specific (12 in B. distachyon, 36 in S. italica, and 91 in S. bicolor). For example, a number of basic LEUCINE ZIPPER (bZIP) family members (e.g., bZIP3, bZIP48, and bZIP53) involved in responses to environmental stimuli (Satoh et al., 2004; Hossain et al., 2016; Sanagi et al., 2018) were identified exclusively in S. bicolor and B. distachyon leaves. By contrast, the NAC (NAM/ATAF1,2/CUC) family protein NAC083, which is induced under abiotic stress in an abscisic acid-dependent manner (Yang et al., 2011), was identified only in S. italica.

Conservation of Core Gene Networks in Grasses in Response to Cold

We were interested in the 17 conserved cold-responsive TFs and whether there was a common cold response network in various grasses. We first sought to define the target genes for each of the 17 TFs. The putative binding sites of each TF were identified from DHSs. Then, the target genes for each TF were defined by assigning the TF binding site to the nearest TSS. By examining the expression of defined target genes before and after cold treatment, we found that the DEGs were enriched in the vicinity of the binding sites (P < 0.01, odds ratio 1.14 to 1.80, Fisher’s exact test; Supplemental Data Set 5), suggesting that the expression regulatory functions of these TFs drive transcriptomic differences during cold stress.

To gain further insight into the interactions of the 17 conserved TFs in response to cold stress, we conducted TF-to-TF network analysis using a strategy previously published by Sullivan et al. (2014). A putative functional network was created, comprising all edges (TF interactions) connecting the 17 TFs before and after cold stress for each tissue (Figure 7A; Supplemental Figure 17). We identified diverse TF network dynamics in response to cold stress. For example, ERF73, ERF104, and GLABROUS1 ENHANCER BINDING PROTEIN (GeBP) showed zero or one edge change upon exposure to cold stress in S. bicolor leaf tissue (Figure 7A; Supplemental Data Set 6). By contrast, CRF4 gained 55 regulatory edges and ERF1, ERF2, and ERF9 lost 35, 93, and 179 regulatory edges, respectively, under cold stress. However, we observed zero or only two net gain edges in S. bicolor stem tissue for ERF1, ERF2, and ERF9 (Supplemental Figure 17; Supplemental Data Set 6), indicating that the cold-response interactions between the 17 TFs are tissue-specific to a certain degree.

Figure 7.

Figure 7.

Open in a new tab

Conserved TF Network Associated with Cold Stress.

(A) The dynamics of conserved TF interaction networks (input edges only) in S. bicolor leaves under cold stress. TF interactions (edges) are indicated by red, blue, and gray lines, which represent gained, lost, and common edges, respectively. Differential regulation can be observed for some TFs, such as ERF1, ERF2, and ERF9, which lost many regulators in response to cold stress.

(B) TF feedback loops of conserved TF members in B. distachyon leaves. The indicated TFs (gray boxes) are regulated by one or more conserved TFs (yellow boxes) and in turn regulate at least one specific conserved TF.

We then explored the cold-induced feedback loops representing regulation of the 17 TF genes by other TFs or TF-regulated TFs. We observed the formation of several novel feedback loops under cold stress (Figure 7B; Supplemental Figure 18). For example, in B. distachyon leaves, the ethylene-response TF SHINE3 (SHN3) was rewired to regulate the stress-responsive activator HEAT SHOCK TRANSCRIPTION FACTOR A6B (HSFA6B; Huang et al., 2016), which in turn regulates ERF1 (Figure 7B). Meanwhile, ABSCISIC ACID INSENSITIVE4 (ABI4) also indirectly regulates ERF1 by a NAC-LIKE, ACTIVATED BY AP3/PI (NAP)-mediated feedback loop. Moreover, SHN3 was identified in feedback loops in all tissues (Figure 7B; Supplemental Figure 18), suggesting that it might play a central role in the cold-stress response. In addition, ERF2 and ERF9, which are involved in the response to low-temperature stress (Fujimoto et al., 2000; An et al., 2012), were also ubiquitous in all tissues examined except B. distachyon roots. Taken together, these data suggest that these 17 conserved cold-induced TFs, or at least SHN3, ERF2, and ERF9, may be core regulators involved in the cold-response transcriptional cascade from upstream to downstream in the grasses examined.

DISCUSSION

We present a set of genome-wide chromatin hypersensitivity maps comprising more than 320,000 DHSs from three grasses, namely B. distachyon, S. italica, and S. bicolor. For each species, we developed regular and cold treatment-derived DHS maps from roots, stems, and leaves for comparison to elucidate changes in the regulatory landscapes in response to cold stress. DHSs are expected to encompass key regulatory elements, including promoters, enhancers, silencers, and insulators, associated with gene expression in the native chromatin structure under physiological conditions. Thus, the DHS maps presented here offer a global view of the transcriptional regulatory elements in the grasses examined in vivo and can be expected to serve as a valuable resource for elucidating the transcriptional networks that control the cold-stress response and other physiological processes.

Understanding the characteristics of open chromatin, such as genomic distribution and correlation with gene expression, especially genes with tissue- or cell-specific expression, is fundamental for deciphering expressional regulatory networks. However, this information is largely unknown for plants (Jiang, 2015; Sullivan et al., 2015) because few such studies have been conducted in plants (Zhang et al., 2012a, 2012b; Sullivan et al., 2014; Qiu et al., 2016; Rodgers-Melnick et al., 2016; Lu et al., 2017, 2019; Maher et al., 2018; Alvarez et al., 2019; Ricci et al., 2019). In particular, there is a lack of comparative analyses between diverse tissues in multiple species (Maher et al., 2018). In this study, we revealed that genes with tissue-specific expression showed high regulatory complexity among the three grasses. This finding is consistent with the notion that multiple regulatory elements mediate cell-specific gene expression in collaboration and facilitate large developmental shifts in gene expression (González et al., 2015; Arendt et al., 2016), which should be of particular interest to those studying the regulation of genes that are expressed preferentially with tissue-dependent patterns. However, studying the regulatory repertoire for tissue-specific genes is challenging because the representative regulators of tissue-specific DHSs are largely distant from the tissue-specific genes (Figure 4). Given that this approach could be prone to errors, especially for distal intergenic enhancers, studies in both plants (Zhu et al., 2015; Maher et al., 2018) and animals (Dowen et al., 2014; González et al., 2015; Mifsud et al., 2015) have demonstrated that a majority of regulators are proximal to target genes. This finding suggests that identifying the correct regulators by examining the adjacent tissue-specific DHSs from tissue-specific genes is feasible, as exemplified by our study (Figure 4B). Thus, the candidate tissue-specific DHSs presented here, in combination with the corresponding sets of TF binding maps and chromatin topological associations (Bonev and Cavalli, 2016), could help elucidate the underlying transcriptional networks that control tissue-specific gene expression.

Moreover, our analysis of open chromatin and the representative transcriptional regulatory motifs in multiple tissues has uncovered several novel features of plant transcriptional regulation, two of which are rather striking.

First, we observed that open chromatin induced by cold presented in a tissue- and species-specific pattern (Figures 5A and 5B), with no cold-induced DHSs commonly shared across all three tissues in a given species and, consequently, no common cold-induced DHSs detected across three species. This finding highlights the diversity of open chromatin landscapes in response to cold or abiotic stress within diverse plants. A plausible explanation is that the regulatory noncoding sequences could be highly variable due to moving and shuffling (Wallbank et al., 2016), generating distinct tissue- or species-selective regulatory modules (Arendt et al., 2016). In fact, results from both our study (Figures 3B and 5B) and another recent study (Maher et al., 2018) revealed no strong conservation among open chromatin sites at orthologous genes from multiple plants, supporting the high diversification of the DHS-representing module. However, the low-consistency pattern of DHSs across species underscores the feasibility of deriving the core regulatory network by merely comparing DHSs when addressing specific biological questions.

Second, given that genome-wide cold-induced DHSs show low commonalities among tissues and species, putative cold-induced TF binding motifs exhibited a high degree of co-occurrence, with ∼23% common motifs across either three tissues or three species. The features of the common set of motifs in response to cold are consistent with the notion that different cell types share similar core transcriptional regulatory networks in specific biological functions (Neph et al., 2012; Maher et al., 2018), while cis-regulatory elements could rapidly migrate and recombine to generate distinct regulatory modules (Wittkopp and Kalay, 2011; Wallbank et al., 2016), highlighting the conservation of regulatory TFs in the cold-stress response across grasses. Therefore, although the DHSs do not necessarily occur in orthologous locations, the TF binding motifs or short stretches of sequences underlying these DHSs remain highly similar, suggesting that most of these control elements were not simply vertically transmitted as conserved noncoding sequences (Burgess and Freeling, 2014; Liang et al., 2018) but are highly dynamic during grass genome evolution.

METHODS

Plant Materials and Growth Conditions

Seeds of Brachypodium distachyon (inbred line Bd21), Setaria italica (inbred line Yugu1), and Sorghum bicolor (inbred line BTx623) were germinated on wet filter paper at 30°C on Petri plates (in the dark). The seedlings were transferred to potting soil and grown under environment-controlled greenhouse conditions set to 16 h/8 h of light/dark, 22°C light/20°C dark, 60% humidity, and 270 µmol m−2 s−1 light intensity (cool-white fluorescent bulbs). After 15 d, half of the plants were directly transferred to an incubator set at 4°C treatment for 24 h (16 h/8 h of light/dark, 60% humidity, and 270 µmol m−2 s−1). Three biological replicates were generated from tens of individuals from each treatment. Leaf, stem, and root tissues from control and cold-treated plants were collected and frozen immediately in liquid nitrogen for further analysis.

DNase-Seq and RNA-Seq

The “end-capture” strategy in which a 20-bp fragment is extracted from the DNase I-digested chromatin end was used to isolate the DHSs (Boyle et al., 2008). DNase-seq experiments, including nuclei isolation, DNase I digestion, and DNase-seq library construction, were performed exactly as previously described by (Zhang and Jiang 2015). DNase-seq libraries were developed from three biological replicates for each tissue and were sequenced using the Illumina HiSeq platform with a 50-bp single-end model. Total RNA from three biological replicates for each tissue was extracted using TRIzol reagent. RNA-seq libraries were prepared following the manufacturer’s protocol and sequenced with an Illumina HiSeq sequencing system.

Read Mapping, DHS Identification, and Visualization

Genome sequences and annotations for B. distachyon (v.3.1), S. italica (v.2.2), and S. bicolor (v.3.1) were downloaded from Phytozome 12.1 (http://phytozome.jgi.doe.gov/pz/portal.html). The DNase-seq reads for each species were aligned to their respective genomes, and DHSs were called using methods with some modifications previously described by (Zhang et al. 2012a). In brief, reads were quality filtered and trimmed using Sickle (https://github.com/najoshi/sickle). Reads that were 15 to 20 bp long after adapter removal and quality trimming were mapped to the reference genome with a 1-bp mismatch allowed using Bowtie v.1.1.2 (Langmead et al., 2009). Reads that mapped to multiple locations were removed. We used Popera (https://github.com/forrestzhang/Popera) to identify DHSs with the parameter “-b 200 -l 20.” To estimate the false discovery rate (FDR), we randomly generated 10 data sets that contained the same read number as the DNase-seq data set. The FDR was calculated as the ratio of the number of DHSs identified based on random data sets to the number of DHSs from the DNase-seq data. We set the cutoff in Popera to control the FDR < 0.01. Annotation of DHSs relative to genes was performed using the ANNOVAR tool (http://wglab.org/software/9-annovar) with default parameters. The average plots that exhibited the DHS distributions around genic regions were generated using deepTools v.3.2.0 (Ramírez et al., 2016) with a bin size of 1 bp. The SPOT score (the number of subsampled mapped reads falling in DHSs/total number of subsampled mapped reads [5 million]; John et al., 2011; Burgess et al., 2019) was calculated using BEDTools (Quinlan and Hall, 2010) to determine the number of mapped reads possessing an at least 1-bp overlap with a DHS site.

The tissue-specific and cold-induced DHSs were identified as previously described by (Sijacic et al. 2018). For each species, DHSs from all tissues were merged to create a union set of DHSs. The read number for each tissue in the union DHSs was counted using BEDTools (Quinlan and Hall, 2010). Three replicates from each tissue were counted, and the counts were processed using the Limma-Voom framework (Law et al., 2014). Those DHSs that had a fold change > 2 and an adjusted P < 0.05 for a specific tissue were identified as tissue-specific DHSs in that tissue type. For cold-induced DHSs, DHSs from untreated and cold-treated tissues in a given species were merged to create the union set of DHSs. The read count from each replicate in the corresponding union DHS was obtained and processed using Limma-Voom to identify cold-induced DHSs (fold change > 2, adjusted P < 0.05).

For RNA-seq data, sequencing reads were analyzed using TopHat v.2.1.1 (Trapnell et al., 2009) and Cufflinks v.2.2.1 (Trapnell et al., 2010). GO enrichment analysis was performed using an online resource (www.omicshare.com/tools) with the default instructions. The GO terms that had FDRs of 0.05 or less were considered significant. For visualization, the alignments in the BAM files were converted to bigwig format using deepTools v.3.2.0 (Ramírez et al., 2016) with a bin size of 10 bp and reads per kilobase per million mapped reads normalization. Genome browser images were made using Integrative Genomics Viewer v.2.3.92 with bigwig files processed as described above.

Conservation Analysis of DHSs

To investigate the cross-species relationships between B. distachyon, S. italica, and S. bicolor DHSs, we used bnMapper (https://bitbucket.org/james_taylor/bx-python/wiki/bnMapper) with the default setting to map DHSs across genomes. The chain files were constructed according to http://genomewiki.ucsc.edu/index.php/Whole_genome_alignment_howto using tools from the UCSC Genome Browser, including trfBig, faSize, faSplit, lavToPsl, axtChain, and LASTZ (Harris, 2007). Briefly, whole-genome alignment was performed using LASTZ with the parameter “–format=lav,” matching alignments were chained together using axtChain, and the chains were filtered with chainPreNet. To unambiguously map DHSs from one species to others, we also produced reciprocal-best chains using the UCSC Genome Browser tools chainSwap, chainSort, chainNet, netSyntenic, and netChainSubset, representing the most likely set of one-to-one orthologs between two species. The detailed procedures are described at http://genomewiki.ucsc.edu/index.php/HowTo:_Syntenic_Net_or_Reciprocal_Best. The reciprocal-best chains were used to map each DHS definitively to the corresponding homologous location across species using bnMapper (Denas et al., 2015). Once we identified the homologous regions of DHSs in one species, we overlapped the homologous regions with the DHSs in the other species to determine whether the mapped homologous region was also annotated as a DHS. For this purpose, we used the “intersect” command of the BEDTools v.2.27.1 suite (Quinlan and Hall, 2010). DHSs were then classified into three categories: homologous and annotated as DHSs in the other species (conserved and shared by both species); homologous but not annotated as DHSs in the other species (conserved but species-specific); and no homologous sequences were found (not conserved).

Classification of Gene Regulatory Complexity

In each species, the DHSs from root, stem, and leaf tissues were merged using the “merge” command of the BEDTools v.2.27.1 suite (Quinlan and Hall, 2010) to build an atlas comprising peaks from all tissue types. For any given gene, the number of DHSs located in the genic region within the atlas was quantitated. Based on the total number of DHSs assigned, we grouped genes into three classes: low-complexity genes (with zero to one DHS), medium-complexity genes (with two to three DHSs), and high-complexity genes (with four or more DHSs).

Hi-C Data Analysis

Hi-C data sets from the leaf mesophylls of S. italica and S. bicolor were downloaded from the Short Read Archive (accession number PRJNA391551). Chromatin interactions (10 kb to 2 Mb from the TSS regions) were identified as previously described, with some modifications, by (Wang et al. 2017). Briefly, raw FASTQ files from all biological replicates were combined for each species and subjected to quality filtration using Trimmomatic v.0.38 (Bolger et al., 2014). Clean reads were mapped to the corresponding genome assemblies, and 5-kb resolution contact maps were constructed with HiC-Pro v.2.11.1 software (Servant et al., 2015). Significant intrachromosomal interactions (q < 0.05) were identified using Fit-Hi-C v.2.0.5 software (Ay et al., 2014) with parameter settings “-L 10,000 –U 2,000,000 –p 2 –b 200.”

Motif Calling from DHS Sequences

Motifs were identified within DHS sequences using MEME-ChIP (Machanick and Bailey, 2011). Motifs with E < 0.05 were retained. Motif functional annotation was performed by comparison against the Cis-BP (Weirauch et al., 2014) and DAP-seq (O’Malley et al., 2016) databases using TOMTOM v.5.0.4 (Gupta et al., 2007) with parameter settings “-dist pearson –thresh 0.3 –no-sse.”

Network of TF-to-Target and TF-to-TF Interactions

We used FIMO v.5.0.2 (Grant et al., 2011) to identify TF motif occurrences (TF binding sites) within the DHS sequences. Predicted binding sites for a given TF were assigned to putative target genes using the “closest” command of the BEDTools v.2.27.1 suite (Quinlan and Hall, 2010). TF-to-TF interaction analysis was conducted as previously described by (Sullivan et al. 2014). Briefly, an edge (TF-to-TF interaction) will be created when a TF binding site occurred within a target TF gene and 2 kb upstream and downstream of the target TF gene. The TF-to-TF network connections between TF genes were visualized by Cytoscape v.3.4.0 (https://cytoscape.org/).

Protoplast Transformation

Functional validation of DHSs by protoplast transformation was conducted as previously described by (Zhao et al. 2018). Briefly, the target DHSs (Supplemental Figure 4C; Supplemental Data Set 7) were cloned into the 35S Enhancer-GFP vector and the mini35S-GFP vector (Zhao et al., 2018) for validation of their promoter and enhancer functions, respectively. The constructs were then transformed into protoplasts from B. distachyon leaves for functional validation by analyzing GFP signals. Protoplast transformation was conducted as previously described, with some modifications, by (Priyadarshani et al. 2018). Briefly, leaf tissue was immersed in 20 mL of enzyme solution (20 mM MES KOH, 1.5% [w/v] Cellulase R-10, 0.5% [w/v] Macerozyme R-10, 10 mM CaCl2, 20 mM KCl, 0.1% [w/v] BSA, and 0.5 M mannitol, pH 5.7) to release protoplasts. After enzymatic digestion, protoplasts were collected by filtering through a 100-μm nylon mesh, followed by washing three times with WS2 solution (154 mM NaCl, 125 mM CaCl2, 5 mM KCl, and 2 mM MES KOH, pH 5.7). Purified protoplasts were suspended in 2 mL of MMG solution (0.4 M mannitol, 15 mM MgCl2, and 4 mM MES, pH 5.7) and mixed with vectors containing DHSs and 40% (w/v) polyethylene glycol solution. After 15 min of incubation, the protoplasts were washed with WS2 solution and suspended in W1 solution (0.5 M mannitol, 20 mM KCl, and 4 mM MES, pH 5.7). After incubation for 12 to 16 h in the dark, GFP signals were observed and recorded using a confocal microscope (Leica TCS SP8). The functional validation for each DHS construct was confirmed by at least three separate transformations, at least three images were examined per transformation.

Accession Numbers

All data sets have been deposited in the European Nucleotide Archive (ENA) under accession number PRJEB32944.

Supplemental Data

DIVE Curated Terms

The following phenotypic, genotypic, and functional terms are of significance to the work described in this paper:

Acknowledgments

We thank Wenli Zhang from Nanjing Agriculture University for technical assistance with DNase-seq and protoplast transformation, Yufeng Wu from Nanjing Agriculture University for advice on bioinformatic analysis, and Qiulin Liu for assistance with the DNase-seq experiment. We also thank Jiming Jiang from Michigan State University and Tao Zhang from Yangzhou University for their valuable comments on the manuscript. This work was supported by the National Natural Science Foundation of China (grants 31771862 and 31471170) and the International Cooperation Project of Fujian Agriculture and Forestry University (grant KXGH17002).

AUTHOR CONTRIBUTIONS

K.W. and H.T. acquired financial support and provided overall direction of the project; K.W., J.H., H.T., and J.C.S. designed the research; P.W., Q.W., Q.L., G.Y., Z.C., and Y.D. performed the experiments; K.W., J.H., H.T., C.M., R.W., and J.C.S. analyzed the data and wrote the article.

References

  1. Alvarez J.M., Moyano T.C., Zhang T., Gras D.E., Herrera F.J., Araus V., O’Brien J.A., Carrillo L., Medina J., Vicente-Carbajosa J., Jiang J., Gutiérrez R.A.(2019). Local changes in chromatin accessibility and transcriptional networks underlying the nitrate response in Arabidopsis roots. Mol. Plant 12: 1545–1560. [DOI] [PubMed] [Google Scholar]
  2. Amit I., et al. (2009). Unbiased reconstruction of a mammalian transcriptional network mediating pathogen responses. Science 326: 257–263. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. An D., Yang J., Zhang P.(2012). Transcriptome profiling of low temperature-treated cassava apical shoots showed dynamic responses of tropical plant to cold stress. BMC Genomics 13: 64. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Arendt D., Musser J.M., Baker C.V.H., Bergman A., Cepko C., Erwin D.H., Pavlicev M., Schlosser G., Widder S., Laubichler M.D., Wagner G.P.(2016). The origin and evolution of cell types. Nat. Rev. Genet. 17: 744–757. [DOI] [PubMed] [Google Scholar]
  5. Auerbach R.K., Euskirchen G., Rozowsky J., Lamarre-Vincent N., Moqtaderi Z., Lefrançois P., Struhl K., Gerstein M., Snyder M.(2009). Mapping accessible chromatin regions using Sono-Seq. Proc. Natl. Acad. Sci. USA 106: 14926–14931. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Ay F., Bailey T.L., Noble W.S.(2014). Statistical confidence estimation for Hi-C data reveals regulatory chromatin contacts. Genome Res. 24: 999–1011. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Bennetzen J.L., et al. (2012). Reference genome sequence of the model plant Setaria. Nat. Biotechnol. 30: 555–561. [DOI] [PubMed] [Google Scholar]
  8. Bhatia R., Gallagher J.A., Gomez L.D., Bosch M.(2017). Genetic engineering of grass cell wall polysaccharides for biorefining. Plant Biotechnol. J. 15: 1071–1092. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Bolger A.M., Lohse M., Usadel B.(2014). Trimmomatic: A flexible trimmer for Illumina sequence data. Bioinformatics 30: 2114–2120. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Bolt S., Zuther E., Zintl S., Hincha D.K., Schmülling T.(2017). ERF105 is a transcription factor gene of Arabidopsis thaliana required for freezing tolerance and cold acclimation. Plant Cell Environ. 40: 108–120. [DOI] [PubMed] [Google Scholar]
  11. Bonev B., Cavalli G.(2016). Organization and function of the 3D genome. Nat. Rev. Genet. 17: 661–678. [DOI] [PubMed] [Google Scholar]
  12. Boyle A.P., Davis S., Shulha H.P., Meltzer P., Margulies E.H., Weng Z., Furey T.S., Crawford G.E.(2008). High-resolution mapping and characterization of open chromatin across the genome. Cell 132: 311–322. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Buenrostro J.D., Giresi P.G., Zaba L.C., Chang H.Y., Greenleaf W.J.(2013). Transposition of native chromatin for fast and sensitive epigenomic profiling of open chromatin, DNA-binding proteins and nucleosome position. Nat. Methods 10: 1213–1218. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Burgess D., Freeling M.(2014). The most deeply conserved noncoding sequences in plants serve similar functions to those in vertebrates despite large differences in evolutionary rates. Plant Cell 26: 946–961. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Burgess S.J., Reyna-Llorens I., Stevenson S.R., Singh P., Jaeger K., Hibberd J.M.(2019). Genome-wide transcription factor binding in leaves from C3 and C4 grasses. Plant Cell 31: 2297–2314. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Chinnusamy V., Zhu J., Zhu J.K.(2007). Cold stress regulation of gene expression in plants. Trends Plant Sci. 12: 444–451. [DOI] [PubMed] [Google Scholar]
  17. Clayton W.D., Renvoize S.A.(1986). Genera Graminum: Grasses of the World.. (London: Royal Botanic Gardens; ). [Google Scholar]
  18. Crawford G.E., et al. (2006). Genome-wide mapping of DNase hypersensitive sites using massively parallel signature sequencing (MPSS). Genome Res. 16: 123–131. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Denas O., Sandstrom R., Cheng Y., Beal K., Herrero J., Hardison R.C., Taylor J.(2015). Genome-wide comparative analysis reveals human-mouse regulatory landscape and evolution. BMC Genomics 16: 87. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Dong P., Tu X., Chu P.-Y., Lü P., Zhu N., Grierson D., Du B., Li P., Zhong S.(2017). 3D chromatin architecture of large plant genomes determined by local A/B compartments. Mol. Plant 10: 1497–1509. [DOI] [PubMed] [Google Scholar]
  21. Dowen J.M., Fan Z.P., Hnisz D., Ren G., Abraham B.J., Zhang L.N., Weintraub A.S., Schujiers J., Lee T.I., Zhao K., Young R.A.(2014). Control of cell identity genes occurs in insulated neighborhoods in mammalian chromosomes. Cell 159: 374–387. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Elkon R., Agami R.(2017). Characterization of noncoding regulatory DNA in the human genome. Nat. Biotechnol. 35: 732–746. [DOI] [PubMed] [Google Scholar]
  23. Fujimoto S.Y., Ohta M., Usui A., Shinshi H., Ohme-Takagi M.(2000). Arabidopsis ethylene-responsive element binding factors act as transcriptional activators or repressors of GCC box-mediated gene expression. Plant Cell 12: 393–404. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Fukushima S., Mori M., Sugano S., Takatsuji H.(2016). Transcription factor WRKY62 plays a role in pathogen defense and hypoxia-responsive gene expression in rice. Plant Cell Physiol. 57: 2541–2551. [DOI] [PubMed] [Google Scholar]
  25. Garel A., Axel R.(1976). Selective digestion of transcriptionally active ovalbumin genes from oviduct nuclei. Proc. Natl. Acad. Sci. USA 73: 3966–3970. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Giresi P.G., Kim J., McDaniell R.M., Iyer V.R., Lieb J.D.(2007). FAIRE (formaldehyde-assisted isolation of regulatory elements) isolates active regulatory elements from human chromatin. Genome Res. 17: 877–885. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. González A.J., Setty M., Leslie C.S.(2015). Early enhancer establishment and regulatory locus complexity shape transcriptional programs in hematopoietic differentiation. Nat. Genet. 47: 1249–1259. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Grant C.E., Bailey T.L., Noble W.S.(2011). FIMO: Scanning for occurrences of a given motif. Bioinformatics 27: 1017–1018. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Grass Phylogeny Working Group II (2012). New grass phylogeny resolves deep evolutionary relationships and discovers C4 origins. New Phytol. 193: 304–312. [DOI] [PubMed] [Google Scholar]
  30. Gupta S., Stamatoyannopoulos J.A., Bailey T.L., Noble W.S.(2007). Quantifying similarity between motifs. Genome Biol. 8: R24. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Hannah M.A., Heyer A.G., Hincha D.K.(2005). A global survey of gene regulation during cold acclimation in Arabidopsis thaliana. PLoS Genet. 1: e26. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Harris, R.S. (2007). Improved Pairwise Alignment of Genomic DNA. PhD dissertation (College of Engineering Pennsylvania, The Pennsylvania State University).
  33. Hesselberth J.R., Chen X., Zhang Z., Sabo P.J., Sandstrom R., Reynolds A.P., Thurman R.E., Neph S., Kuehn M.S., Noble W.S., Fields S., Stamatoyannopoulos J.A.(2009). Global mapping of protein-DNA interactions in vivo by digital genomic footprinting. Nat. Methods 6: 283–289. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Hossain M.A., Henríquez-Valencia C., Gómez-Páez M., Medina J., Orellana A., Vicente-Carbajosa J., Zouhar J.(2016). Identification of novel components of the unfolded protein response in Arabidopsis. Front. Plant Sci. 7: 650. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Huang Y.C., Niu C.Y., Yang C.R., Jinn T.L.(2016). The heat stress factor HSFA6b connects ABA signaling and ABA-mediated heat responses. Plant Physiol. 172: 1182–1199. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. International Brachypodium Initiative (2010). Genome sequencing and analysis of the model grass Brachypodium distachyon. Nature 463: 763–768. [DOI] [PubMed] [Google Scholar]
  37. Jiang J.(2015). The ‘dark matter’ in the plant genomes: Non-coding and unannotated DNA sequences associated with open chromatin. Curr. Opin. Plant Biol. 24: 17–23. [DOI] [PubMed] [Google Scholar]
  38. John S., Sabo P.J., Thurman R.E., Sung M.H., Biddie S.C., Johnson T.A., Hager G.L., Stamatoyannopoulos J.A.(2011). Chromatin accessibility pre-determines glucocorticoid receptor binding patterns. Nat. Genet. 43: 264–268. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Johnson D.S., Mortazavi A., Myers R.M., Wold B.(2007). Genome-wide mapping of in vivo protein-DNA interactions. Science 316: 1497–1502. [DOI] [PubMed] [Google Scholar]
  40. Kenchanmane Raju S.K., Barnes A.C., Schnable J.C., Roston R.L.(2018). Low-temperature tolerance in land plants: Are transcript and membrane responses conserved? Plant Sci. 276: 73–86. [DOI] [PubMed] [Google Scholar]
  41. Kim Y.S., Kim S.G., Park J.E., Park H.Y., Lim M.H., Chua N.H., Park C.M.(2006). A membrane-bound NAC transcription factor regulates cell division in Arabidopsis. Plant Cell 18: 3132–3144. [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Klemm S.L., Shipony Z., Greenleaf W.J.(2019). Chromatin accessibility and the regulatory epigenome. Nat. Rev. Genet. 20: 207–220. [DOI] [PubMed] [Google Scholar]
  43. Langmead B., Trapnell C., Pop M., Salzberg S.L.(2009). Ultrafast and memory-efficient alignment of short DNA sequences to the human genome. Genome Biol. 10: R25. [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Law C.W., Chen Y., Shi W., Smyth G.K.(2014). voom: Precision weights unlock linear model analysis tools for RNA-seq read counts. Genome Biol. 15: R29. [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Liang P., Saqib H.S.A., Zhang X., Zhang L., Tang H.(2018). Single-base resolution map of evolutionary constraints and annotation of conserved elements across major grass genomes. Genome Biol. Evol. 10: 473–488. [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Lu Z., Hofmeister B.T., Vollmers C., DuBois R.M., Schmitz R.J.(2017). Combining ATAC-seq with nuclei sorting for discovery of cis-regulatory regions in plant genomes. Nucleic Acids Res. 45: e41. [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Lu Z., Marand A.P., Ricci W.A., Ethridge C.L., Zhang X., Schmitz R.J.(2019). The prevalence, evolution and chromatin signatures of plant regulatory elements. Nat. Plants 5: 1250–1259. [DOI] [PubMed] [Google Scholar]
  48. Machanick P., Bailey T.L.(2011). MEME-ChIP: Motif analysis of large DNA datasets. Bioinformatics 27: 1696–1697. [DOI] [PMC free article] [PubMed] [Google Scholar]
  49. Maher K.A., et al. (2018). Profiling of accessible chromatin regions across multiple plant species and cell types reveals common gene regulatory principles and new control modules. Plant Cell 30: 15–36. [DOI] [PMC free article] [PubMed] [Google Scholar]
  50. Mifsud B., et al. (2015). Mapping long-range promoter contacts in human cells with high-resolution capture Hi-C. Nat. Genet. 47: 598–606. [DOI] [PubMed] [Google Scholar]
  51. Neph S., Stergachis A.B., Reynolds A., Sandstrom R., Borenstein E., Stamatoyannopoulos J.A.(2012). Circuitry and dynamics of human transcription factor regulatory networks. Cell 150: 1274–1286. [DOI] [PMC free article] [PubMed] [Google Scholar]
  52. O’Malley R.C., Huang S.C., Song L., Lewsey M.G., Bartlett A., Nery J.R., Galli M., Gallavotti A., Ecker J.R.(2016). Cistrome and epicistrome features shape the regulatory DNA landscape. Cell 165: 1280–1292. [DOI] [PMC free article] [PubMed] [Google Scholar]
  53. Pandey G.K., Grant J.J., Cheong Y.H., Kim B.G., Li L., Luan S.(2005). ABR1, an APETALA2-domain transcription factor that functions as a repressor of ABA response in Arabidopsis. Plant Physiol. 139: 1185–1193. [DOI] [PMC free article] [PubMed] [Google Scholar]
  54. Park S., Szumlanski A.L., Gu F., Guo F., Nielsen E.(2011). A role for CSLD3 during cell-wall synthesis in apical plasma membranes of tip-growing root-hair cells. Nat. Cell Biol. 13: 973–980. [DOI] [PubMed] [Google Scholar]
  55. Paterson A.H., et al. (2009). The Sorghum bicolor genome and the diversification of grasses. Nature 457: 551–556. [DOI] [PubMed] [Google Scholar]
  56. Prasad V., Strömberg C.A., Alimohammadian H., Sahni A.(2005). Dinosaur coprolites and the early evolution of grasses and grazers. Science 310: 1177–1180. [DOI] [PubMed] [Google Scholar]
  57. Priyadarshani S.V.G.N., et al. (2018). Simple protoplast isolation system for gene expression and protein interaction studies in pineapple (Ananas comosus L.). Plant Methods 14: 95. [DOI] [PMC free article] [PubMed] [Google Scholar]
  58. Qiu Z., Li R., Zhang S., Wang K., Xu M., Li J., Du Y., Yu H., Cui X.(2016). Identification of regulatory DNA elements using genome-wide mapping of DNase I hypersensitive sites during tomato fruit development. Mol. Plant 9: 1168–1182. [DOI] [PubMed] [Google Scholar]
  59. Quinlan A.R., Hall I.M.(2010). BEDTools: A flexible suite of utilities for comparing genomic features. Bioinformatics 26: 841–842. [DOI] [PMC free article] [PubMed] [Google Scholar]
  60. Ramírez F., Ryan D.P., Grüning B., Bhardwaj V., Kilpert F., Richter A.S., Heyne S., Dündar F., Manke T.(2016). deepTools2: A next generation web server for deep-sequencing data analysis. Nucleic Acids Res. 44: W160–W165. [DOI] [PMC free article] [PubMed] [Google Scholar]
  61. Ricci W.A., et al. (2019). Widespread long-range cis-regulatory elements in the maize genome. Nat. Plants 5: 1237–1249. [DOI] [PMC free article] [PubMed] [Google Scholar]
  62. Rodgers-Melnick E., Vera D.L., Bass H.W., Buckler E.S.(2016). Open chromatin reveals the functional maize genome. Proc. Natl. Acad. Sci. USA 113: E3177–E3184. [DOI] [PMC free article] [PubMed] [Google Scholar]
  63. Sanagi M., Lu Y., Aoyama S., Morita Y., Mitsuda N., Ikeda M., Ohme-Takagi M., Sato T., Yamaguchi J.(2018). Sugar-responsive transcription factor bZIP3 affects leaf shape in Arabidopsis plants. Plant Biotechnol. (Tokyo) 35: 167–170. [DOI] [PMC free article] [PubMed] [Google Scholar]
  64. Satoh R., Fujita Y., Nakashima K., Shinozaki K., Yamaguchi-Shinozaki K.(2004). A novel subgroup of bZIP proteins functions as transcriptional activators in hypoosmolarity-responsive expression of the ProDH gene in Arabidopsis. Plant Cell Physiol. 45: 309–317. [DOI] [PubMed] [Google Scholar]
  65. Schones D.E., Cui K., Cuddapah S., Roh T.Y., Barski A., Wang Z., Wei G., Zhao K.(2008). Dynamic regulation of nucleosome positioning in the human genome. Cell 132: 887–898. [DOI] [PMC free article] [PubMed] [Google Scholar]
  66. Servant N., Varoquaux N., Lajoie B.R., Viara E., Chen C.J., Vert J.P., Heard E., Dekker J., Barillot E.(2015). HiC-Pro: An optimized and flexible pipeline for Hi-C data processing. Genome Biol. 16: 259. [DOI] [PMC free article] [PubMed] [Google Scholar]
  67. Sijacic P., Bajic M., McKinney E.C., Meagher R.B., Deal R.B.(2018). Changes in chromatin accessibility between Arabidopsis stem cells and mesophyll cells illuminate cell type-specific transcription factor networks. Plant J. 94: 215–231. [DOI] [PMC free article] [PubMed] [Google Scholar]
  68. Sullivan A.M., et al. (2014). Mapping and dynamics of regulatory DNA and transcription factor networks in A. thaliana. Cell Rep. 8: 2015–2030. [DOI] [PubMed] [Google Scholar]
  69. Sullivan A.M., Bubb K.L., Sandstrom R., Stamatoyannopoulos J.A., Queitsch C.(2015). DNase I hypersensitivity mapping, genomic footprinting, and transcription factor networks in plants. Curr. Plant Biol. 3–4: 40–47. [Google Scholar]
  70. Tran L.-S.P., Nakashima K., Sakuma Y., Simpson S.D., Fujita Y., Maruyama K., Fujita M., Seki M., Shinozaki K., Yamaguchi-Shinozaki K.(2004). Isolation and functional analysis of Arabidopsis stress-inducible NAC transcription factors that bind to a drought-responsive cis-element in the early responsive to dehydration stress 1 promoter. Plant Cell 16: 2481–2498. [DOI] [PMC free article] [PubMed] [Google Scholar]
  71. Trapnell C., Pachter L., Salzberg S.L.(2009). TopHat: Discovering splice junctions with RNA-seq. Bioinformatics 25: 1105–1111. [DOI] [PMC free article] [PubMed] [Google Scholar]
  72. Trapnell C., Williams B.A., Pertea G., Mortazavi A., Kwan G., van Baren M.J., Salzberg S.L., Wold B.J., Pachter L.(2010). Transcript assembly and quantification by RNA-seq reveals unannotated transcripts and isoform switching during cell differentiation. Nat. Biotechnol. 28: 511–515. [DOI] [PMC free article] [PubMed] [Google Scholar]
  73. Turco G., Schnable J.C., Pedersen B., Freeling M.(2013). Automated conserved non-coding sequence (CNS) discovery reveals differences in gene content and promoter evolution among grasses. Front. Plant Sci. 4: 170. [DOI] [PMC free article] [PubMed] [Google Scholar]
  74. Valdivia E.R., Herrera M.T., Gianzo C., Fidalgo J., Revilla G., Zarra I., Sampedro J.(2013). Regulation of secondary wall synthesis and cell death by NAC transcription factors in the monocot Brachypodium distachyon. J. Exp. Bot. 64: 1333–1343. [DOI] [PMC free article] [PubMed] [Google Scholar]
  75. Vierstra J., Stamatoyannopoulos J.A.(2016). Genomic footprinting. Nat. Methods 13: 213–221. [DOI] [PubMed] [Google Scholar]
  76. Wallbank R.W., Baxter S.W., Pardo-Diaz C., Hanly J.J., Martin S.H., Mallet J., Dasmahapatra K.K., Salazar C., Joron M., Nadeau N., McMillan W.O., Jiggins C.D.(2016). Evolutionary novelty in a butterfly wing pattern through enhancer shuffling. PLoS Biol. 14: e1002353. [DOI] [PMC free article] [PubMed] [Google Scholar]
  77. Wang M., et al. (2017). Asymmetric subgenome selection and cis-regulatory divergence during cotton domestication. Nat. Genet. 49: 579–587. [DOI] [PubMed] [Google Scholar]
  78. Wang M., Zhao Y., Zhang B.(2015). Efficient test and visualization of multi-set intersections. Sci. Rep. 5: 16923. [DOI] [PMC free article] [PubMed] [Google Scholar]
  79. Wang P., Du Y., Zhao X., Miao Y., Song C.P.(2013). The MPK6-ERF6-ROS-responsive cis-acting element7/GCC box complex modulates oxidative gene transcription and the oxidative response in Arabidopsis. Plant Physiol. 161: 1392–1408. [DOI] [PMC free article] [PubMed] [Google Scholar]
  80. Wei C.L., et al. (2006). A global map of p53 transcription-factor binding sites in the human genome. Cell 124: 207–219. [DOI] [PubMed] [Google Scholar]
  81. Weirauch M.T., et al. (2014). Determination and inference of eukaryotic transcription factor sequence specificity. Cell 158: 1431–1443. [DOI] [PMC free article] [PubMed] [Google Scholar]
  82. Wittkopp P.J., Kalay G.(2011). Cis-regulatory elements: Molecular mechanisms and evolutionary processes underlying divergence. Nat. Rev. Genet. 13: 59–69. [DOI] [PubMed] [Google Scholar]
  83. Wu C.(1980). The 5′ ends of Drosophila heat shock genes in chromatin are hypersensitive to DNase I. Nature 286: 854–860. [DOI] [PubMed] [Google Scholar]
  84. Yang S.D., Seo P.J., Yoon H.K., Park C.M.(2011). The Arabidopsis NAC transcription factor VNI2 integrates abscisic acid signals into leaf senescence via the COR/RD genes. Plant Cell 23: 2155–2168. [DOI] [PMC free article] [PubMed] [Google Scholar]
  85. Zhang T., Marand A.P., Jiang J.(2016). PlantDHS: A database for DNase I hypersensitive sites in plants. Nucleic Acids Res. 44: D1148–D1153. [DOI] [PMC free article] [PubMed] [Google Scholar]
  86. Zhang W., Jiang J.(2015). Genome-wide mapping of DNase I hypersensitive sites in plants. Methods Mol. Biol. 1284: 71–89. [DOI] [PubMed] [Google Scholar]
  87. Zhang W., Wu Y., Schnable J.C., Zeng Z., Freeling M., Crawford G.E., Jiang J.(2012a). High-resolution mapping of open chromatin in the rice genome. Genome Res. 22: 151–162. [DOI] [PMC free article] [PubMed] [Google Scholar]
  88. Zhang W., Zhang T., Wu Y., Jiang J.(2012b). Genome-wide identification of regulatory DNA elements and protein-binding footprints using signatures of open chromatin in Arabidopsis. Plant Cell 24: 2719–2731. [DOI] [PMC free article] [PubMed] [Google Scholar]
  89. Zhang W., Zhang T., Wu Y., Jiang J.(2014). Open chromatin in plant genomes. Cytogenet. Genome Res. 143: 18–27. [DOI] [PubMed] [Google Scholar]
  90. Zhao C., Zhang Z., Xie S., Si T., Li Y., Zhu J.K.(2016). Mutational evidence for the critical role of CBF transcription factors in cold acclimation in Arabidopsis. Plant Physiol. 171: 2744–2759. [DOI] [PMC free article] [PubMed] [Google Scholar]
  91. Zhao H., Zhang W., Chen L., Wang L., Marand A.P., Wu Y., Jiang J.(2018). Proliferation of regulatory DNA elements derived from transposable elements in the maize genome. Plant Physiol. 176: 2789–2803. [DOI] [PMC free article] [PubMed] [Google Scholar]
  92. Zhu B., Zhang W., Zhang T., Liu B., Jiang J.(2015). Genome-wide prediction and validation of intergenic enhancers in Arabidopsis using open chromatin signatures. Plant Cell 27: 2415–2426. [DOI] [PMC free article] [PubMed] [Google Scholar]
  93. Zhu J.K.(2016). Abiotic stress signaling and responses in plants. Cell 167: 313–324. [DOI] [PMC free article] [PubMed] [Google Scholar]
  94. Zwack P.J., Compton M.A., Adams C.I., Rashotte A.M.(2016). Cytokinin response factor 4 (CRF4) is induced by cold and involved in freezing tolerance. Plant Cell Rep. 35: 573–584. [DOI] [PubMed] [Google Scholar]

Articles from The Plant Cell are provided here courtesy of Oxford University Press

RESOURCES