Looping by RNA: Dynamic control of chromatin loop by long non-coding RNA in plants

In eukaryotes, the three-dimensional (3D) structure of genome plays a critical role in the transcriptional regulation. Even though the 3D structure of genome is mainly determined by DNA and protein components, increasing evidence has demonstrated that RNA also occupies a large portion of chromatin and influences genomic architecture (Ramirez-Colmenero et al., 2020; Wierzbicki et al., 2021). Long non-coding RNA (lncRNA) is a class of non-coding RNAs longer than 200 nucleotides but lack protein-coding potential. LncRNAs are key regulators of gene expression through their capacity to harbor regulatory proteins and to establish molecular interactions. In animals, LncRNAs play roles in dynamic chromatin remodeling through their contribution to the formation of chromatin loop, the maintenance of heterochromatin, and the function of chromatin-modifying complexes and affect various biological processes, including cellular developments, immunity, and cancer (Ramirez-Colmenero et al., 2020). Similarly, lncRNAs in plants have been implicated to function in the gene regulation by mediating chromatin modification and remodeling. Here, we highlight a novel lncRNA that modulates the recruitment of transcription factors and affects chromatin structure to control transcription in response to long-term cold (known as vernalization), in wheat. We discuss this finding in the context of similar studies in plants and animals.

A recent work in hexaploidy wheat (Triticum aestivum) has identified a lncRNA, VAS (TaVRN1 Alternative Splicing), that regulates the transcription of TaVRN1 in part by affecting the chromatin structure (Figure 1a) (Xu et al., 2021). TaVRN1 promotes floral transition in response to vernalization. The VAS is an alternatively spliced transcript of TaVRN1, which contains a part of the 5’ UTR, the entire first exon and a part of the first intron. The VAS transcript is induced by cold exposure, with the maximal expression observed at 21 days of cold in winter wheat accessions. However, it is not detectable in spring wheat accessions even after the vernalization treatment, possibly due to the deletion of critical sequence in the first intron of TaVRN1. Cold-induced VAS expression triggers vernalization-mediated activation of TaVRN1 by recruiting a heterodimer of basic leucine zipper (bZIP) transcription factors (TaRF2a-TaRF2b) to the TaVRN1 promoter. Chromatin conformation capture (3C) assay showed that a chromatin loop between the promoter region near the TaRF2b binding sequence motif (Sp1 motif) and the first intron region, which overlaps with the VAS sequence, is gradually reduced during vernalization. The chromatin isolation by RNA purification associated with mass spectrometry (ChlRP-MS) assay to uncover potential protein factors associated with VAS has identified 11 proteins, including TaRF2b. Therefore, it appears that VAS provides a molecular structure to facilitate the recruitment of protein complexes to their target loci. Candidate VAS-associated proteins also include splicing factor-like protein, sister chromatid cohesion protein and kinase family proteins, implying that VAS may be involved in various chromatin remodeling processes. However, mechanistic details on how VAS and candidate VAS-associated proteins contribute to chromatin remodeling, including chromatin loop formation, at the TaVRN1 locus remain to be understood.

Figure 1. Dynamic chromatin loop formation by lncRNAs in plants.

(A) A chromatin loop formed at the TaVRN1 locus in wheat. Upon vernalization, chromatin loop is lost, and more VAS transcripts are produced. Subsequently, VAS promotes the TaVRN1 expression by recruiting the transcription factors, TaRF2b-TaRF2a, which bind to the Sp1 motif in the TaVRN1 promoter to increase the transcription of TaVRN1. Activation of TaVRN1 eventually promotes flowering in winter wheat.

(B) Chromatin loops mediated by APOLO in Arabidopsis. Auxin induces DNA demethylation at the APOLO locus and opens the loop encompassing the PID promoter region, resulting in increases in both PID and APOLO transcripts. APOLO lncRNA binds to its neighboring PID locus or distant loci, such as auxin responsive genes. APOLO decoys LHP1 and forms local chromatin loops to finetune the expression of target genes.

(C) An intragenic chromatin loop by COLDWRAP and PRC2 in Arabidopsis. COLDAIR and COLDWRAP are gradually induced by vernalization and supports PRC2 occupation to the promoter region to repress FLC expression by H3K27me3 enrichments.

In Arabidopsis, there are at least two examples in which lncRNAs contribute to the chromatin structures by forming chromatin loops, APOLO (Ariel et al., 2014) and COLDWRAP (Kim and Sung, 2017). The APOLO is an intergenic lncRNA transcribed from 5 kb upstream of the gene PINOID (PID). PID encodes a Ser/Thr protein kinase responsible for the polar localization of auxin transporter. Auxin-induced APOLO expression controls PID expression by affecting a chromatin loop formation between the PID promoter and APOLO locus, in cooperation with a component of Polycomb Repressive Complex 1 (PRC1), LHP1 (Figure 1b). APOLO interacts with LHP1 and LHP1 is associated with Histone H3 lysine 27 trimethylation (H3K27me3)-marked chromatin. Therefore, the expression of APOLO triggers dynamic formation of chromatin loop by APOLO-LHP1 (Ariel et al., 2014). Auxin-induced APOLO expression is achieved by decreases in the levels of DNA methylation at APOLO and PID loci by auxin, although how auxin triggers the DNA demethylation remains to be understood. Interestingly, APOLO also contributes to chromatin looping at multiple trans target loci by targeting not only to spatially associated loci but also to distal loci through the sequence complementarity and DNA-RNA duplexes (R-loops) formation (Ariel et al., 2020). The R-loops driven by APOLO decoy LHP1, resulting in subsequent dynamic regulation of chromatin loop at target loci. Another example is COLDWRAP, a lncRNA transcribed from the proximal promoter region of FLOWERING LOCUS C (FLC), which is a major floral repressor in Arabidopsis. COLDWRAP interacts with a component of PRC2, CURLY LEAF (CLF) to form an intragenic chromatin loop between the promoter and 3’ end of the first intron of FLC and confers repression of FLC expression by vernalization (Figure 1c) (Kim and Sung, 2017). In addition, another lncRNAs derived from the first intron of FLC, COLDAIR which is involved in the PRC2-mediated silencing of FLC by vernalization through its capacity in the recruitment of PRC2 to FLC, also contributes to the intragenic repressive chromatin loop accomplished by COLDWRAP and PRC2 (Figure 1c) (Kim and Sung, 2017).

Interestingly, both APOLO and COLDWRAP interact with proteins which belong to two major Polycomb group (PcG) repressive complex (PRCs; PRC1 and PRC2). Since PRCs are widely involved in the transcriptional repression by heterochromatin formation (Piunti and Shilatifard, 2021), those lncRNAs are expected to contribute to chromatin condensation at target loci by interacting with PcG proteins which affect heterochromatin formation. Indeed, in animals, many of lncRNAs have been identified to interact with PRCs including Xist, Kcnq1ot1 and HOTAIR for their repressive chromatin loop formation (Ramirez-Colmenero et al., 2020).

Studies using Hi-C, a genome-wide technique to determine chromatin interactions, have revealed that eukaryotic genome in the nucleus is highly organized. A basic unit of 3D organization of genome is “Topologically associating domains” (TADs) in many eukaryotes. TADs are compartments of chromatin folded into domains with preferential internal interactions among chromatin loops (Bonev and Cavalli, 2016). TADs are the architectural units of chromatin that define regulatory landscapes and the TAD-TAD boundaries contribute to the gene regulation by limiting interactions of cis-regulatory elements to their target genes. Consistent with the roles of lncRNAs in chromatin loop formation, there are examples of lncRNAs that contribute to the TAD formation. The ncRNAs, Eleanor, delineate TAD of the ESR1 locus in the active nuclear compartment of LTED cells and activate the ESR1 gene which encodes estrogen receptor-α (ER) (Abdalla et al., 2019). Another lncRNA, Firre, is located at the borders of TADs within CTCF binding regions and is necessary for super loop formation of inactive X-chromosome (Barutcu et al., 2018). Similarly, several studies defined TADs in some plant species (Ouyang et al., 2020). Arabidopsis genome possesses relatively smaller range of chromatin loops but lacks apparent TADs. Interestingly, the repressive histone mark, H3K27me3 is favored for genes with promoter-promoter interactions over long ranges in Arabidopsis (Ouyang et al., 2020), suggesting the role of chromatin modification in defining 3D chromatin structure in plants. However, our understanding of regulatory and structural elements controlling chromatin loops in plants is still at its infancy, although some chromatin regulators have been implicated in the formation of chromatin loop in Arabidopsis (Li et al., 2018; Zhao et al., 2021).

Increasing numbers of lncRNAs are being identified in plants, and their functional studies will undoubtfully unravel diverse functions of lncRNAs as regulatory elements. The identification of VAS in wheat and its role in the dynamic control of a chromatin loop by vernalization at TaVRN1 locus provide an excellent example of lncRNAs that mediate changes in chromatin architecture although how this new lncRNA is directly involved in the chromatin loop remains to be elucidated. Given that lncRNAs can recruit, decoy, or scaffold protein complexes to mediate chromatin looping and the inter- or intra-chromosomal interactions, it is expected that more lncRNAs that participate in the regulation of gene expression and genomic stability will be uncovered in plants. Therefore, studying potential interactions between chromatin-remodeling factors and lncRNAs will be the promising first step to elucidate mechanisms underlying how lncRNAs contribute the 3D structure of plant genome.