Identifications of a novel group of intra-genic interactive putative enhancers reveals a distinctive enhancer architectures in plants

ABSTRACT: Enhancers are important regulators of gene expression in eukaryotes; however, only a few enhancers have been identified in plants, and the genome-wide identifications are lacking. To investigate enhancers in Arabidopsis thaliana, we analyzed the chromatin landscape, RNA Pol II occupancy, and the polyadenylated and non-polyadenylated Arabidopsis transcriptomes, using exosome-deficient lines to capture stable and unstable transcripts. Here we identified the set of unique >1900 regions, termed Putative Enhancers (PEs), carrying canonical chromatin signatures of metazoan transcriptional enhancers. Almost of PEs are intra-genic and transcriptionally active. Unexpectedly, most of the PE carrying these signatures are associated with the 3′ ends of protein-coding genes, enriched in transcription factor binding, with binding motifs resembling polyadenylation signals, thus linking links two hubs of very different, functionally and mechanistically distinct processes – cleavage/polyadenylation and transcription termination of protein-coding genes with transcriptional activation/regulation of different genes. And may present a novel DNA elements with regulatory potential. Additionally, according to high resolution Arabidopsis Hi-C data 24% of identified PEs interact with other protein-coding genes intrachromosomally, through binding by the same set of transcription factors, supporting the notion of co-regulation of expression between PEs and their targets. We also confirmed that PEs are highly conserved in the recently sequenced 1035 different Arabidopsis accessions. Identifications and characterization of a novel putative intragenic enhancers (PEs) in Arabidopsis. To determine if the identified PEs are actively transcribed we profiled the RNA expression in WT and the exosome deficient Arabidopsis seedlings. Since eRNAs (transcripts arising at enhancer locus) varied by their polyadenylation status, we collected both the polyadenylated RNAs (polyA(+)) and the pool of ribo-minus RNA depleted of rRNAs to increase our detection for non-polyadenylated RNA, and subjected to high throughout sequencing. The bioinformatics protocols used for processing RNA-seq datasets, transcriptome assembly, and profiling transcriptional activity are described in further details in the associated publication.