Review

Nuclear Organization Changes and the Epigenetic Silencing of FLC during Vernalization

Danling Zhu, Stefanie Rosa and Caroline Dean Department of Cell and Developmental Biology, John Innes Centre, Norwich Research Park, Norwich NR4 7UH, UK

Correspondence to Caroline Dean: [email protected] http://dx.doi.org/10.1016/j.jmb.2014.08.025 Edited by E. Soutoglou

Abstract Changes in nuclear organization are considered an important complement to trans-acting factors, histone modifications and non-coding RNAs in robust and stable epigenetic silencing. However, how these multiple layers interconnect mechanistically to reinforce each other's activity is still unclear. A system providing long timescales facilitating analysis of these interconnections is vernalization. This involves the Polycombmediated epigenetic silencing of flowering locus C (FLC) that occurs as Arabidopsis plants are exposed to prolonged cold. Analysis of changes in nuclear organization during vernalization has revealed that disruption of a gene loop and physical clustering of FLC loci are part of the vernalization mechanism. These events occur at different times and thus contribute to distinct aspects of the silencing mechanism. The physical clustering of FLC loci is tightly correlated with the accumulation of specific Polycomb complexes/H3K27me3 at a localized intragenic site during the cold. Since the quantitative nature of vernalization is a reflection of a bistable cell autonomous switch in an increasing number of cells, this correlation suggests a tight connection between the switching mechanism and changes in nuclear organization. This integrated picture is likely to be informative for many epigenetic mechanisms. © 2014 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/3.0/).

Introduction Plants as sessile organisms are faced with the enormous challenge of how to respond to a constantly changing environment. Many sensing mechanisms have therefore evolved to enable plants to integrate external signals and use them in the regulation of their metabolism, growth and development. Long-term seasonal cues are used to time developmental transitions in order to align specific stages with favorable environmental conditions, thus maximizing reproductive success. For example, the seasonal cues of increasing photoperiod and average daily temperature align flowering with spring. A second aspect of temperature that is recognized as a seasonal cue is the prolonged cold of winter. Flowering is actively repressed until plants have been exposed to sufficient cold (weeks and months depending on the species) so that flowering only occurs after winter. This process, known as vernalization, has been central to the breeding of winter and spring sown varieties, thus

extending the geographical range of many crops [1]. Many native species also require vernalization, and in Arabidopsis thaliana, the process involves the epigenetic silencing of a gene encoding the floral repressor FLOWERING LOCUS C (FLC). This MADS box protein functions as a transcriptional repressor blocking expression of genes required to switch the meristem to a floral fate [2,3]. Vernalization is one of the most well understood examples of environmentally induced epigenetic regulation. Cold exposure results in FLC silencing in a slow and quantitative process, with the degree of silencing reflecting the length of cold exposure [4]. The silencing is then mitotically maintained through subsequent development in warm conditions, enabling floral activator induction by photoperiod and accelerated flowering. The epigenetic stability of FLC silencing is stable through many rounds of cell division, even through cuttings or callus regeneration [5]. The current understanding of vernalization has been established from forward genetic screens combined

0022-2836/© 2014 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/3.0/). J. Mol. Biol. (2015) 427, 659–669

660

Review: Nuclear Organization Changes at FLC

with chromatin biochemistry and more recently with mathematical modeling. In the first part of this review, we introduce the different temporal phases of the vernalization process: establishment of FLC expression before exposure to cold, cold-induced FLC silencing, stable maintenance of the silenced state and then resetting of FLC expression during embryo development (Fig. 1a). We then review more recent work on nuclear organization changes at FLC, and lastly, we speculate on the integration of the

(a)

different mechanistic layers, focusing on how the dynamic chromatin changes are interconnected with changes in nuclear organization. The temporal phases of vernalization Establishment of FLC expression before exposure to cold FLC expression is set early during embryogenesis as the seed develops on the mother plant and this

FLC

COOLAIR

Core-PRC2

VRN5

VIN3

VEL1

FLC

COOLAIR VIN3 Autumn

(b)

(Weeks of cold) Winter

Spring

Summer

High FLC expression state H3K4me2

(A) FRIGIDA

Autonomous pathway

FLD

Low FLC expression state H3K27me3

(A) FCA FY FPA

PRP8

Fig. 1. Changes in expression of the floral repressor FLC regulate flowering and vernalization in Arabidopsis thaliana. (a) Schematic illustration of the FLC locus showing changes in FLC and COOLAIR expression and the dynamic association of different PRC2 complexes over the life cycle of the plant. Temporal changes in FLC and COOLAIR expression are also compared to VIN3 expression. The resetting of FLC expression in developing embryos is shown by the luciferase signal in the seedpods. (b) Schematic illustration of the high and low FLC expression states established in the embryos and mediated by FRIGIDA and the autonomous pathway. The colored balls represent different histone modifications. The gray arrows represent the feedback mechanism linking processing of the COOLAIR transcript, chromatin modification in the gene body and expression level.

661

Review: Nuclear Organization Changes at FLC

level varies depending on life history strategy. There is considerable natural variation for FLC expression, reflected in the different reproductive strategies of different A. thaliana accessions. The winter/spring growth habit is typically determined by allelic variation at FRIGIDA (FRI): a coiled-coil domain protein, which activates FLC expression through a co-transcription regulatory mechanism and chromatin modifications [6,7] (Fig. 1b). FRI directly interacts with the nuclear cap-binding complex, potentially acting as a molecular scaffold in a transcription–activator complex [8,9]. In addition, current data also suggested that FRI is required for the recruitment of AtWDR5a, a conserved component of the H3K4 methyltransferase Compass/ MLL complex, that is targeted to FLC locus and increases H3K4me3 [7]. Rapid cycling accessions such as Columbia (Col) and Landsberg erecta (Ler) carry loss-of-function mutations at FRIGIDA [6,10]. FRIGIDA function requires conserved chromatin modifiers such as the RNA-polymerase-associated factor 1 complex [11–16], Trithorax-like [17,18], radiation sensitivity protein 6/Brefeldin A sensitivity protein 1 [19–21], Facilitates Chromatin Transcription [22], Compass-like [7] and SWR1 complexes [23–30] involved in promoting different aspects of FLC transcription [31]. Antagonizing FRIGIDA is the activity of the autonomous floral pathway involving FCA, FY, FVE, FPA, LUMINIDEPENDENS, FLOWERING LOCUS D and FLOWERING LOCUS K. Molecular analysis has shown that these activities converge to repress FLC rather than function in a “linear” genetic pathway [32]. Their functions link RNA processing steps, both 3′ end processing and splicing, with changes in histone modification at FLC [33,34]. This mechanism involves a set of antisense transcripts produced at the FLC locus, collectively termed COOLAIR [35,36]. These antisense transcripts completely cover the sense transcription unit, initiating downstream of the FLC major poly(A) site and terminating within the FLC promoter. The activities of FCA (an RNA recognition motif protein), FY (a cleavage and polyadenylation specificity factor component) and FPA (an RNA recognition motif protein), together with the conserved cleavage stimulation factors CstF64 and 77, promote proximal polyadenylation of COOLAIR transcripts [34,37,38]. This results in FLD-dependent H3K4me2 demethylation across the body of the gene and low FLC transcription [33]. Loss of FCA and FPA results in increased distal polyadenylation of COOLAIR, increased H3K4 methylation across the gene and high expression [34,37,38]. A role for the conserved spliceosome factor PRP8 in COOLAIR splicing also has recently been described [39]. PRP8 promotes splicing at a weak splice acceptor site in the first short intron in COOLAIR, thus enhancing levels of COOLAIR proximal polyadenylation and thus FLC repression. Dissection of this mechanism revealed a feedback mechanism between gene body histone methylation and COOLAIR processing that reinforces

the low expression state [39]. This can be seen as functioning exactly antagonistically to FRIGIDA activation, which promotes distal polyadenylation, high histone H3K4me2 and high expression. The balance of these activities then achieves a fine molecular balance for quantitative regulation of FLC expression (Fig. 1b). Central to this balance is regulation of COOLAIR transcription, which requires both a positive transcription elongation factor b function [40] and modulation of an extensive R-loop (a three-stranded nucleic acid structure formed by an RNA–DNA hybrid plus a displaced single-stranded DNA strand) that covers the COOLAIR promoter and first exon [41]. Cold induction of FLC silencing As plants are exposed to cold temperature, FLC expression becomes reduced and is epigenetically silenced. Genetic studies identifying trans-factors necessary for vernalization identified many components necessary for this silencing including a suppressor of Zeste 12 homologue, a core component of Polycomb-repressive complex 2 (PRC2). This suggested histone methylation as a major factor in determining the epigenetic silencing underlying vernalization. DNA methylation appears to play no role in vernalization [42,43]. Analysis of the dynamics of association of PRC2 showed that it associated with the FLC locus even in the actively expressed state, thus resembling examples of mammalian Polycomb silencing [44]. Mutagenesis also identified two plant homeodomain (PHD) proteins required for vernalization [45,46]. First, the VERNALIZATION INSENSITIVE 3 (VIN3) is slowly up-regulated by cold—many weeks of cold exposure are required for maximum expression [45]. Second, the VERNALIZATION 5 (VRN5) has a similar structure to VIN3 but is constitutively expressed [46]. VIN3 and VRN5 (also known as VIN3-like 1 [47]) form part of a four-member VEL (vernalization-like) gene family in the A. thaliana genome, which is conserved in monocots [48]. All intact genes encode proteins with a PHD domain, a fibronectin III domain and a conserved C-terminus, which is important for homodimerization/heterodimerization of the proteins with each other [46]. The PHD domains bind preferentially to H3K9me2 in vitro and the VEL family functions to directly repress different subsets of the FLC/MAF (MADS affecting flowering) gene family [49]. Interconnection of the functions of the PHD proteins with the core PRC2 was achieved through analysis of interacting proteins and protein complexes [44,50]. The PHD–PRC2 purified from cold-exposed plants contained VIN3, VRN5 and VEL1, as well as the core PRC2 components [44]. This complex appears analogous to Drosophila Polycomb-like PRC2 and mammalian PHF1–PRC2 [51,52]. PHD–PRC2 accumulates at a specific intragenic site within FLC during the cold and causes cold-dependent accumulation of H3K27me3. The

662 increased H3K27me3 is limited to just one or two nucleosomes in the nucleation region. Additional cold-regulated steps were discovered when FLC expression was still found to decrease during the cold in a vin3 mutant [36]. Use of a customized tiling array identified sense and antisense transcripts that changed at FLC in different genotypes and conditions [36]. Within 2 weeks of cold exposure, the antisense transcripts at FLC accumulate over 10-fold at a time when FLC mRNA has changed rather little. The cold-induced increase in antisense transcript expression led to the name COOLAIR, as a parallel with Hox antisense intergenic RNA (HOTAIR), a long non-coding RNA involved in Polycomb regulation of Hox loci in mammals [53]. A more detailed analysis of COOLAIR induction has shown that it is the proximally induced form that accumulates most strongly, together with the unspliced antisense transcript (T. Csorba, J. Questa, Q. Sun and C. Dean unpublished results). The COOLAIR promoter was shown to drive coldinduced expression of a luciferase reporter and to result in cold-induced reduction in expression of a 35S-GFP reporter suggesting that it is actively involved in the cold-dependent repression of FLC expression [36]. However, analysis of T-DNA insertions in the COOLAIR promoter that reduced COOLAIR expression shows that these transcripts are not absolutely required for vernalization [54]. We are currently dissecting the role of these antisense transcripts and their regulation by multiple pathways, such as RNAi [35], R-loop [41] and cold [36]. Although not identified in the tiling array analysis, a second long non-coding transcript, cold assisted intronic non-coding RNA (COLDAIR), was found by PCR to initiate from a cryptic promoter in the first intron of FLC in a cold-dependent manner [55]. In vivo RNA immunoprecipitation experiments suggested that COLDAIR could interact and recruit the histone methyltransferase subunit (CLF) of PRC2 to the FLC locus [55]. Stable maintenance of the FLC silenced state and then resetting during embryo development A few days after transfer back to warm conditions, the PHD–PRC2 complex (now without the coldexpressed VIN3) spreads across the entire FLC gene [44]. This is associated with very high levels of H3K27me3 blanketing the locus that are required for the epigenetic stability of FLC repression through the rest of development [44,56,57]. Whether increased DNA replication or transcription at the higher temperatures contributes to spreading post-cold is unclear. The mitotic stability of FLC silencing is maintained through many cell divisions but has been shown to be lost when leaves become much older [57]. Proteins identified as playing a role in the maintenance of the silenced state include VERNALIZATION 1 (a non-sequence-specific DNA-binding protein) [58] and LIKE HETEROCHROMATIN PROTEIN 1 (LHP1) (the single Arabidopsis homologue of heterochromatin

Review: Nuclear Organization Changes at FLC

protein 1) [59,60]. LHP1 binds H3K27me3 in vitro through its chromodomain, co-localizes with H3K27me3 in vivo and is required to maintain FLC silencing. This makes LHP1 in some respects the functional equivalent of Polycomb, one of the subunits of Polycomb-repressive complex 1. However, many details to discover how all these activities are integrated to give robust and stable epigenetic silencing of FLC through many cell divisions until embryo development remain to be elucidated. FLC expression is reset during embryo development to ensure a vernalization requirement every generation [61,62]. What regulates this resetting has yet to be determined, but in contrast to the slow silencing during the cold, the reactivation of FLC in the embryo is rapid (Fig. 1a). Whether transcriptional activation is sufficient for resetting will be an important question to address. One feature that distinguishes vernalization from many other epigenetic silencing mechanisms is its quantitative nature. For example, plants given 2 weeks of cold flower later than plants given 4 weeks of cold. The acceleration in flowering is therefore proportional to the length of the cold exposure. This quantitative basis has been explored using a combination of mathematical modeling, based on that used to analyze switching of the mating type locus in Schizosaccharomyces pombe and high-resolution chromatin immunoprecipitation analysis [56,63]. The model assumes a dynamic chromatin environment at FLC with nucleosome turnover occurring at a high rate. Bistable epigenetic states are established by specific protein complexes and histone modifications and these are reinforced through positive feedback mechanisms. At FLC, high levels of H3K36me3 or H3K27me3 appear to reflect the active and silent epigenetic states [64]. Cold results in a cell autonomous switching of the active state to the silent state through the localized nucleation of a PHD–PRC2 complex. The PHD–PRC2 complex spreads across the locus on return to warm and this spreading is required for stability of the epigenetic silencing. Using a GUS reporter gene inserted in FLC, Angel et al. demonstrated this “digital” memory experimentally [56]. After longer cold treatment, an increasing number of cells are switched off, and thus, the overall FLC level in a plant is gradually turned down. The quantitative nature of vernalization after different lengths of cold exposure is thus achieved through a cell population average. This digital memory is translated into a quantitative response in acceleration of flowering at least partially through FLC directly regulating expression of flowering locus T (FT), which encodes a mobile protein that moves through the plant [65]. Nuclear organization changes during the different phases of vernalization Over the past few years, nuclear organization has been established as an important parameter in gene

Review: Nuclear Organization Changes at FLC

regulation. The importance of three-dimensional changes in chromatin conformation for gene regulation has been shown in different systems [66–69] and many studies have correlated transcriptional activity or repression of genes with relative position to nuclear domains, such as the nuclear periphery, the nucleolus or heterochromatin clusters [70–73]. Understanding the role played by this type of organization in regulating gene expression represents a major challenge. Two major changes in nuclear organization during the different temporal phases of FLC silencing involve disruption of a gene loop at FLC and physical association of the FLC loci. Our understanding of these nuclear organizational changes is reviewed below. Disruption of a gene loop in the early phase of vernalization Gene loops were first identified in yeast [74] and were regarded as short-range chromatin interactions where the promoter and 3′ terminator genic regions came into physical contact as transcription proceeded [75]. They are considered to enhance transcription by recycling RNA Pol II from the end of the gene back to the promoter but have also been suggested to play a role in maintaining transcriptional memory [76,77]. In order to address whether gene looping was a component of FLC regulation, chromatin conformation capture was performed in a series of mutants and at different time points of the vernalization process [78]. A robust gene loop at FLC was detected reflecting the physical interaction between a genomic fragment I (including the FLC promoter, first exon and key regulatory elements in intron I) with fragment V, a region ~ 355–970 bp downstream of the FLC poly(A) site that contains promoter elements necessary for COOLAIR antisense transcription (Fig. 2a). Thus, the FLC gene loop has similarities to the promoter– terminator loops found in yeast. The formation of the FLC gene loop was found to be independent of expression level as no major difference in chromosome interaction was found in genotypes with high FLC expression (Col-FRI, fcafpa) or low expression (arp6 and atx1atx2). However, the gene loop was efficiently disrupted within the first 2 weeks of vernalization (2WT0) and did not reform 7 days after transfer of plants back into warm condition (2WT7) (Figs. 2a and 3). A subunit of the SWI/SNF chromatin/ remodeling complex BAF60 has been found to be required for the formation of the FLC gene loop, via modulation of histone density, composition and post-translational modification [79]. Long-range chromatin interactions While gene loop structures represent interactions at short distances, Polycomb targets are also able to

663 engage in long-range chromatin interactions [80,81]. In Drosophila, FISH analysis revealed that cosilenced genes from the two distant Hox complexes can associate in the three-dimensional nuclear space into so-called Polycomb bodies [67]. These long-distance nuclear co-associations strengthen PcG-mediated silencing, indicating that they might play a functional role. This PcG-dependent chromosome contact required components of the RNAi machinery [82]. To investigate whether similar long-range chromatin interactions were involved in the PcG regulation of FLC, we developed a live cell imaging using the Lac I/O system (Fig. 2b) [83]. To visualize FLC in vivo, we inserted an array of 120 copies of the lacO DNA sequence (each copy separated by a random ~ 10-bp sequence to minimize recombination and transgene instability) into an FLC transgene downstream of the poly(A) site. This insertion did not influence FLC function as the transgene fully complemented the mutant phenotype. A control transgene carrying the lacO array but no FLC sequence was included in the experiment. Transgenic lines carrying the FLC-lacO and lacO alone transgenes were crossed with a line expressing a lacI-YFP-NLS fusion. This was under a leaky ethanol inducible promoter system and showed constitutive expression in the region of the root containing endoreduplicated nuclei in epidermal cells. FLC expression is epigenetically silenced in all tissues and not just those making floral tissue. lacO/ lacI foci were only detected in plants carrying both transgenes and could be monitored in root epidermal cells from both meristematic region and endoreduplication zone, where the cells had differentiated. In non-vernalized plants, multiple FLC-lacO foci were detected in this region of the root, whereas in meristematic cells, only two foci were detected (Fig. 2c). After exposure to 2 weeks, cold many nuclei show only one FLC-LacO focus in both endoreduplicated and meristematic cells (Figs. 2c and 3). This only occurred in FLC-lacO transgenic lines, not in lacO alone, and these foci were larger in size than those from multiple foci nuclei in non-vernalized plants, consistent with the occurrence of a physical clustering. The number of nuclei showing clustering of the FLC-LacO alleles increased quantitatively with increasing cold exposure and then remained similar after plants were returned to warm. Notably, this time-dependent increase in clustering mirrors the increase in H3K27me3 at the nucleation site (Fig. 2d). The clustering was reminiscent of the formation of Polycomb bodies in mammalian cells [84–86], but whether RNA Pol II exclusion and/or Polycomb target concentration co-localizes with the FLC clusters remains to be established. Immunofluorescence experiments aimed at localizing RNA Pol II and H3K27me3 were unsuccessful likely due to the harsh pretreatment conditions required for these experiments with plant cells [83]. No obvious

664

(a)

Relative Interaction frequency

Review: Nuclear Organization Changes at FLC

0.35

NV 2WT0 2TW7

0.30 0.25 0.20 0.15 0.10 0.05 0.00 0

1000

2000

3000

4000

5000

6000

7000

8000

Distances to anchor region I (bp) I

II

III

IV

V

VI

(b) FLC-Lac O

Lac O array

X Detector

alcA

YFP-Lac I-NLS

Nos

YFPYFPYFPYFP Lac ILac I Lac I Lac I

FLC-Lac O

2WT0

1.0

100

H3K27me3 % Clusters

80

0.8

60

0.6

40

0.4

20

0.2

H 3 K 2 7 m e 3 in c r e a s e

NV

(d) C lu s te r in g fr e q u e n c y ( % )

(c)

Lac O array

0

0

NV

2WT0

4WT0

6WT0

Fig. 2. Two changes in FLC nuclear organization occur at different phases of the vernalization process. (a) An FLC gene loop that forms between the 5′ promoter region (fragment I) and downstream of the poly(A) site (fragment V) in non-vernalized (NV) seedlings is disrupted by 2 weeks of cold exposure (2WT0). The loop does not reform 7 days (2WT7) after transfer back to warm. The different fragments tested in a chromatin conformation capture experiment are shown with vertical lines [78]. (b) Cold exposure induces physical clustering of the FLC loci. FLC loci were tagged with 120 copies of a lacO DNA sequence downstream of FLC poly(A) site (FLC-LacO). The transgenic line was then crossed to a line carrying a YFP-lacI fusion. lacI/lacO binding enables live imaging of the FLC locus in the nucleus. (c) Representative images of FLC loci either in endoreduplicated root cells (long) or meristematic cells (round) grown without cold (NV) or exposed to 4 °C for 2 weeks (2WT0). (d) Temporal changes in the frequency of cells showing clustering of FLC loci compared to the cold-induced accumulation of H3K27me3 at the nucleation site [83].

concentration in a particular nuclear region or near to the nucleolus was observed in the live imaging experiments performed to date.

The exact mechanism by which clustering occurs remains to be elucidated. Both the Polycomb transfactors (VRN5 and VERNALIZATION 2) [46,87] are

665

Review: Nuclear Organization Changes at FLC

(a)

Disruption of an FLC gene loop

FLC locus chromatin modifications

Interaction of FLC loci

FLC

H3K27me3 COOLAIR

H3K36me3

Nucleation region

(b)

(c)

Fig. 3. Integration of the different regulatory functions at FLC. Comparison of the FLC gene loop, FLC and COOLAIR transcription, the histone modifications H3K36me3 [64] and H3K27me3 [56] and the clustering status. (a) Plants grown in the warm; (b) plants exposed to 2 weeks cold. The nucleation region is indicated with the pink line; (c) after transfer back to warm conditions.

required; however, clustering still occurred in the presence of a mutation in LHP1, which is important for maintenance of FLC repression rather than initiation of the silencing [59]. Integration of the different regulatory functions at FLC The long timescales involved in vernalization provide the opportunity to integrate the different aspects of changes in nuclear organization with other regulatory layers of the silencing mechanism (Fig. 3). The earliest nuclear organizational change is disruption of the FLC gene loop. This coincides with the reduction of FLC sense transcription but whether this is cause or effect remains unknown. Our current hypothesis is that loop disruption facilitates COOLAIR expression by revealing cis-elements that contribute to the cold induction of antisense transcription. An FLC transgene that carries different 3′ regulatory sequences, thus lacking the COOLAIR promoter and up-regulation of antisense transcription in the cold, shows slower FLC transcriptional reduction in the cold. COOLAIR thus has an important functional role but this does not seem to be through a simple transcriptional interference mechanism (T. Csorba, J. Questa, Q. Sun and C. Dean, unpublished results). Similar interconnections among antisense transcription, chromatin contacts and transcriptional activity have been analyzed at the yeast FLO11 locus [88,89]. Different non-coding

transcripts and loops are associated with switching between expression states. Whether similar toggle-like mechanisms are involved at FLC is unknown but the limited analysis undertaken so far has not revealed any additional high frequency internal contacts. The gene loop could also identify a chromosomal domain similar to the topologically associating domains characterized in mammalian genomes. These define chromatin domains that facilitate interaction between cis-regulatory elements [90]. A spatial element to FLC regulation has previously been demonstrated; cold reduces expression of FLC and both flanking genes, but only FLC silencing is epigenetically maintained during subsequent development of the plant in the warm, and the adjacent genes reactivate [91]. In contrast to the gene loop, the physical clustering of FLC loci appears more central to the silencing mechanism (Fig. 3). The tight correlation between clustering with nucleation of the PHD–PRC2 complex and accumulating H3K27me3 suggests that nuclear organization is an important component of the mechanism. The next level of understanding of this inter-relationship will emerge when we know more about nucleation. Nucleation is likely to be limited by expression of the cold-induced PHD protein VIN3 (Fig. 1a), which is induced more slowly than COOLAIR. This temporal separation argues for a requirement for reduced sense and antisense transcription before efficient nucleation can occur. However, a prevailing argument suggests a role for non-coding RNA in

666 recruitment of PRC2 complexes to specific sites in the genome [92]. We think that it is unlikely that COOLAIR is required for recruitment of PRC2 to the nucleation region given that the COOLAIR dynamics do not match nucleation dynamics; we only see low-affinity PRC2 binding to COOLAIR, and plants not expressing COOLAIR show no alteration in H3K27me3 accumulation during the cold (T. Csorba, J. Questa, Q. Sun and C. Dean, unpublished results). The sense non-coding FLC transcript COLDAIR has also been suggested to recruit PRC2 to the nucleation site in FLC [55]. In addition, we need to investigate whether the CpGisland-mediated PcG recruitment, shown to be important in mammals, is involved at FLC [93–95]. It is currently unknown whether the nucleation region is defined by sequence and/or chromatin signature and if it is analogous to Drosophila Polycomb response element (PRE). In Drosophila, PREs are usually located outside the transcribed region of the target gene in the proximal promoter region. Deletion of different regions of intron 1, either a 2.8-kb DNA fragment [96] or a 288-bp DNA fragment [60], suggests that the nucleation region does function as expected of a PRE, namely, it is not required for the initial repression of FLC expression during cold, but it is required to maintain the repressed state. Site-directed mutagenesis experiments and analysis of A. thaliana natural variants where cis variation at FLC confers different lengths of cold requirement should prove informative in defining requirements for efficient nucleation [97,98]. The correlation in the timing of nucleation and physical clustering raises the very interesting possibility that nucleation reflects a digital switch associated with a change in nuclear organization. In the case of FLC, the PHD–PRC2 and accumulating H3K27me3 modification might trigger association of homologous loci. A digital switch could reflect a transfer between nuclear domains, which might correspond to “transcription factories” and “Polycomb bodies” [99–102]. Intuitively, sequence specificity would seem to be important for such a mechanism; thus, roles of the various non-coding RNAs at FLC need to be explored in this respect. Long-distance contacts that enhance PcG-dependent silencing in Drosophila require RNAi components. In Arabidopsis, mutations in the RNAi components DICER-like 1 and DICER-like 3 affect FLC expression but not the epigenetic silencing by vernalization, as assayed by flowering time [103]. One could imagine nucleation and clustering function to reinforce each other ensuring robust silencing [104]. This may be quite general given other associations between induction of silencing and changes in nuclear organization [68,84,105,106]. Perspective The cold-induced epigenetic silencing of the floral repressor FLC underlying vernalization in Arabidopsis

Review: Nuclear Organization Changes at FLC

involves many regulatory layers that are common with other silencing systems. These regulatory layers integrate the noisy temperature signals found in nature and decode them into developmental timing information. The more we learn about these mechanisms, the more parallels—e.g., nucleation and spreading of chromatin complexes, nuclear reorganization and non-coding RNA function—emerge with silencing systems in a whole range of organisms [85,107,108]. A combination of molecular genetics, cell biology and computational modeling is enabling the full dissection of vernalization. We need more systems where multiple approaches are focused on a few key targets, to complement genome-wide approaches, if we are to fully understand the interconnections between these different regulatory layers.

Acknowledgements The authors acknowledge funding from Biotechnology and Biological Sciences Research Council grant BB/K00008X/1 and Biotechnology and Biological Sciences Research Council Industry Strategy Panel grant BB/J004588/1 and thank Judith Irwin, Jo Hepworth and Robert Ietswaart for critically reading the manuscript. Received 21 May 2014; Received in revised form 26 August 2014; Accepted 26 August 2014 Available online 30 August 2014 Keywords: vernalization; FLC; gene loop; epigenetic silencing Abbreviations used: FLC, FLOWERING LOCUS C; PRC2, Polycomb-repressive complex 2; PHD, plant homeodomain; VIN3, VERNALIZATION INSENSITIVE 3; VRN5, VERNALIZATION 5; LHP1, LIKE HETEROCHROMATIN PROTEIN 1; PRE, Polycomb response element.

References [1] Henderson IR, Dean C. Control of Arabidopsis flowering: the chill before the bloom. Development 2004;131:3829–38. [2] Michaels SD, Amasino RM. FLOWERING LOCUS C encodes a novel MADS domain protein that acts as a repressor of flowering. Plant Cell 1999;11:949–56. [3] Sheldon CC, Burn JE, Perez PP, Metzger J, Edwards JA, Peacock WJ, et al. The FLF MADS box gene: a repressor of flowering in Arabidopsis regulated by vernalization and methylation. Plant Cell 1999;11:445–58.

667

Review: Nuclear Organization Changes at FLC

[4] Sheldon CC, Rouse DT, Finnegan EJ, Peacock WJ, Dennis ES. The molecular basis of vernalization: the central role of FLOWERING LOCUS C (FLC). Proc Natl Acad Sci USA 2000;97:3753–8. [5] Burn JE, Bagnall DJ, Metzger JD, Dennis ES, Peacock WJ. DNA methylation, vernalization, and the initiation of flowering. Proc Natl Acad Sci USA 1993;90:287–91. [6] Johanson U, West J, Lister C, Micheal S, Amasino RM, Dean C. Molecular analysis of FRIGIDA, a major determinant of natural variation in Arabidopsis flowering time. Science 2000; 290:344–7. [7] Jiang D, Gu X, He Y. Establishment of the winter-annual growth habit via FRIGIDA-mediated histone methylation at FLOWERING LOCUS C in Arabidopsis. Plant Cell 2009;21: 1733–46. [8] Geraldo N, Baurle I, Kidou S, Hu X, Dean C. FRIGIDA delays flowering in Arabidopsis via a cotranscriptional mechanism involving direct interaction with the nuclear cap-binding complex. Plant Physiol 2009;150:1611–8. [9] Choi K, Kim J, Hwang HJ, Kim S, Park C, Kim SY, et al. The FRIGIDA complex activates transcription of FLC, a strong flowering repressor in Arabidopsis, by recruiting chromatin modification factors. Plant Cell 2011;23:289–303. [10] Toomajian C, Hu TT, Aranzana MJ, Lister C, Tang C, Zheng H, et al. A nonparametric test reveals selection for rapid flowering in the Arabidopsis genome. PLoS Biol 2006;4:e137. [11] Zhang H, van Nocker S. The VERNALIZATION INDEPENDENCE 4 gene encodes a novel regulator of FLOWERING LOCUS C. Plant J 2002;31:663–73. [12] Oh S, Zhang H, Ludwig P, van Nocker S. A mechanism related to the yeast transcriptional regulator Paf1c is required for expression of the Arabidopsis FLC/MAF MADS box gene family. Plant Cell 2004;16:2940–53. [13] Zhang H, Ransom C, Ludwig P, van Nocker S. Genetic analysis of early flowering mutants in Arabidopsis defines a class of pleiotropic developmental regulator required for expression of the flowering-time switch Flowering Locus C. Genetics 2003;164:347–58. [14] He Y, Doyle MR, Amasino RM. PAF1-complex-mediated histone methylation of FLOWERING LOCUS C chromatin is required for the vernalization-responsive, winter-annual habit in Arabidopsis. Genes Dev 2004;18:2774–84. [15] Yu X, Michaels SD. The Arabidopsis Paf1c complex component CDC73 participates in the modification of FLOWERING LOCUS C chromatin. Plant Physiol 2010;153: 1074–84. [16] Park S, Oh S, Ek-Ramos J, van Nocker S. PLANT HOMOLOGOUS TO PARAFIBROMIN is a component of the PAF1 complex and assists in regulating expression of genes within H3K27ME3-enriched chromatin. Plant Physiol 2010;153:821–31. [17] Pien S, Fleury D, Mylne JS, Crevillen P, Inze D, Avramova Z, et al. ARABIDOPSIS TRITHORAX1 dynamically regulates FLOWERING LOCUS C activation via histone 3 lysine 4 trimethylation. Plant Cell 2008;20:580–8. [18] Saleh A, Alvarez-Venegas R, Yilmaz M, Le O, Hou G, Sadder M, et al. The highly similar Arabidopsis homologs of trithorax ATX1 and ATX2 encode proteins with divergent biochemical functions. Plant Cell 2008;20:568–79. [19] Cao Y, Dai Y, Cui S, Ma L. Histone H2B monoubiquitination in the chromatin of FLOWERING LOCUS C regulates flowering time in Arabidopsis. Plant Cell 2008;20:2586–602. [20] Xu L, Menard R, Berr A, Fuchs J, Cognat V, Meyer D, et al. The E2 ubiquitin-conjugating enzymes, AtUBC1 and AtUBC2, play

[21]

[22]

[23]

[24]

[25]

[26]

[27]

[28]

[29]

[30]

[31]

[32] [33]

[34]

[35]

redundant roles and are involved in activation of FLC expression and repression of flowering in Arabidopsis thaliana. Plant J 2009;57:279–88. Gu X, Jiang D, Wang Y, Bachmair A, He Y. Repression of the floral transition via histone H2B monoubiquitination. Plant J 2009;57:522–33. Lolas IB, Himanen K, Gronlund JT, Lynggaard C, Houben A, Melzer M, et al. The transcript elongation factor FACT affects Arabidopsis vegetative and reproductive development and genetically interacts with HUB1/2. Plant J 2010; 61:686–97. Noh YS, Amasino RM. PIE1, an ISWI family gene, is required for FLC activation and floral repression in Arabidopsis. Plant Cell 2003;15:1671–82. Deal RB, Kandasamy MK, McKinney EC, Meagher RB. The nuclear actin-related protein ARP6 is a pleiotropic developmental regulator required for the maintenance of FLOWERING LOCUS C expression and repression of flowering in Arabidopsis. Plant Cell 2005;17:2633–46. Choi K, Kim S, Kim SY, Kim M, Hyun Y, Lee H, et al. SUPPRESSOR OF FRIGIDA3 encodes a nuclear ACTINRELATED PROTEIN6 required for floral repression in Arabidopsis. Plant Cell 2005;17:2647–60. Martin-Trillo M, Lazaro A, Poethig RS, Gomez-Mena C, Pineiro MA, Martinez-Zapater JM, et al. EARLY IN SHORT DAYS 1 (ESD1) encodes ACTIN-RELATED PROTEIN 6 (AtARP6), a putative component of chromatin remodelling complexes that positively regulates FLC accumulation in Arabidopsis. Development 2006;133:1241–52. March-Diaz R, Garcia-Dominguez M, Florencio FJ, Reyes JC. SEF, a new protein required for flowering repression in Arabidopsis, interacts with PIE1 and ARP6. Plant Physiol 2007;143:893–901. Choi K, Park C, Lee J, Oh M, Noh B, Lee I. Arabidopsis homologs of components of the SWR1 complex regulate flowering and plant development. Development 2007;134: 1931–41. Lazaro A, Gomez-Zambrano A, Lopez-Gonzalez L, Pineiro M, Jarillo JA. Mutations in the Arabidopsis SWC6 gene, encoding a component of the SWR1 chromatin remodelling complex, accelerate flowering time and alter leaf and flower development. J Exp Bot 2008;59:653–66. Deal RB, Topp CN, McKinney EC, Meagher RB. Repression of flowering in Arabidopsis requires activation of FLOWERING LOCUS C expression by the histone variant H2A.Z. Plant Cell 2007;19:74–83. Crevillen P, Dean C. Regulation of the floral repressor gene FLC: the complexity of transcription in a chromatin context. Curr Opin Plant Biol 2011;14:38–44. Baurle I, Dean C. The timing of developmental transitions in plants. Cell 2006;125:655–64. Liu F, Quesada V, Crevillen P, Baurle I, Swiezewski S, Dean C. The Arabidopsis RNA-binding protein FCA requires a lysine-specific demethylase 1 homolog to downregulate FLC. Mol Cell 2007;28:398–407. Liu F, Marquardt S, Lister C, Swiezewski S, Dean C. Targeted 3′ processing of antisense transcripts triggers Arabidopsis FLC chromatin silencing. Science 2010;327: 94–7. Swiezewski S, Crevillen P, Liu F, Ecker JR, Jerzmanowski A, Dean C. Small RNA-mediated chromatin silencing directed to the 3′ region of the Arabidopsis gene encoding the developmental regulator, FLC. Proc Natl Acad Sci USA 2007;104:3633–8.

668 [36] Swiezewski S, Liu F, Magusin A, Dean C. Cold-induced silencing by long antisense transcripts of an Arabidopsis Polycomb target. Nature 2009;462:799–802. [37] Hornyik C, Duc C, Rataj K, Terzi LC, Simpson GG. Alternative polyadenylation of antisense RNAs and flowering time control. Biochem Soc Trans 2010;38:1077–81. [38] Hornyik C, Terzi LC, Simpson GG. The spen family protein FPA controls alternative cleavage and polyadenylation of RNA. Dev Cell 2010;18:203–13. [39] Marquardt S, Raitskin O, Wu Z, Liu F, Sun Q, Dean C. Functional consequences of splicing of the antisense transcript COOLAIR on FLC transcription. Mol Cell 2014;54:156–65. [40] Wang ZW, Wu Z, Raitskin O, Sun Q, Dean C. Antisensemediated FLC transcriptional repression requires the P-TEFb transcription elongation factor. Proc Natl Acad Sci USA 2014; 111:7468–73. [41] Sun Q, Csorba T, Skourti-Stathaki K, Proudfoot NJ, Dean C. R-loop stabilization represses antisense transcription at the Arabidopsis FLC locus. Science 2013;340:619–21. [42] Finnegan EJ, Genger RK, Kovac K, Peacock WJ, Dennis ES. DNA methylation and the promotion of flowering by vernalization. Proc Natl Acad Sci USA 1998;95:5824–9. [43] Finnegan EJ, Kovac KA, Jaligot E, Sheldon CC, Peacock WJ, Dennis ES. The downregulation of FLOWERING LOCUS C (FLC) expression in plants with low levels of DNA methylation and by vernalization occurs by distinct mechanisms. Plant J 2005;44:420–32. [44] De Lucia F, Crevillen P, Jones AM, Greb T, Dean C. A PHDpolycomb repressive complex 2 triggers the epigenetic silencing of FLC during vernalization. Proc Natl Acad Sci USA 2008;105:16831–6. [45] Sung S, Amasino RM. Vernalization in Arabidopsis thaliana is mediated by the PHD finger protein VIN3. Nature 2004; 427:159–64. [46] Greb T, Mylne JS, Crevillen P, Geraldo N, An H, Gendall AR, et al. The PHD finger protein VRN5 functions in the epigenetic silencing of Arabidopsis FLC. Curr Biol 2007;17:73–8. [47] Sung S, Schmitz RJ, Amasino RM. A PHD finger protein involved in both the vernalization and photoperiod pathways in Arabidopsis. Genes Dev 2006;20:3244–8. [48] Fu D, Dunbar M, Dubcovsky J. Wheat VIN3-like PHD finger genes are up-regulated by vernalization. Mol Genet Genomics 2007;277:301–13. [49] Kim DH, Sung S. Coordination of the vernalization response through a VIN3 and FLC gene family regulatory network in Arabidopsis. Plant Cell 2013;25:454–69. [50] Wood CC, Robertson M, Tanner G, Peacock WJ, Dennis ES, Helliwell CA. The Arabidopsis thaliana vernalization response requires a polycomb-like protein complex that also includes VERNALIZATION INSENSITIVE 3. Proc Natl Acad Sci USA 2006;103:14631–6. [51] Nekrasov M, Klymenko T, Fraterman S, Papp B, Oktaba K, Kocher T, et al. Pcl-PRC2 is needed to generate high levels of H3-K27 trimethylation at Polycomb target genes. EMBO J 2007;26:4078–88. [52] Sarma K, Margueron R, Ivanov A, Pirrotta V, Reinberg D. Ezh2 requires PHF1 to efficiently catalyze H3 lysine 27 trimethylation in vivo. Mol Cell Biol 2008;28:2718–31. [53] Rinn JL, Kertesz M, Wang JK, Squazzo SL, Xu X, Brugmann SA, et al. Functional demarcation of active and silent chromatin domains in human HOX loci by noncoding RNAs. Cell 2007;129:1311–23. [54] Helliwell CA, Robertson M, Finnegan EJ, Buzas DM, Dennis ES. Vernalization-repression of Arabidopsis FLC requires

Review: Nuclear Organization Changes at FLC

[55] [56]

[57]

[58]

[59]

[60]

[61]

[62]

[63]

[64]

[65]

[66]

[67]

[68]

[69]

[70]

[71]

[72]

promoter sequences but not antisense transcripts. PLoS One 2011;6:e21513. Heo JB, Sung S. Vernalization-mediated epigenetic silencing by a long intronic noncoding RNA. Science 2011;331:76–9. Angel A, Song J, Dean C, Howard M. A Polycomb-based switch underlying quantitative epigenetic memory. Nature 2011;476:105–8. Finnegan EJ, Dennis ES. Vernalization-induced trimethylation of histone H3 lysine 27 at FLC is not maintained in mitotically quiescent cells. Curr Biol 2007;17:1978–83. Levy YY, Mesnage S, Mylne JS, Gendall AR, Dean C. Multiple roles of Arabidopsis VRN1 in vernalization and flowering time control. Science 2002;297:243–6. Mylne JS, Barrett L, Tessadori F, Mesnage S, Johnson L, Bernatavichute YV, et al. LHP1, the Arabidopsis homologue of HETEROCHROMATIN PROTEIN1, is required for epigenetic silencing of FLC. Proc Natl Acad Sci USA 2006;103:5012–7. Sung S, He Y, Eshoo TW, Tamada Y, Johnson L, Nakahigashi K, et al. Epigenetic maintenance of the vernalized state in Arabidopsis thaliana requires LIKE HETEROCHROMATIN PROTEIN 1. Nat Genet 2006;38:706–10. Sheldon CC, Hills MJ, Lister C, Dean C, Dennis ES, Peacock WJ. Resetting of FLOWERING LOCUS C expression after epigenetic repression by vernalization. Proc Natl Acad Sci USA 2008;105:2214–9. Choi J, Hyun Y, Kang MJ, Yun HI, Yun JY, Lister C, et al. Resetting and regulation of Flowering Locus C expression during Arabidopsis reproductive development. Plant J 2009; 57:918–31. Dodd IB, Micheelsen MA, Sneppen K, Thon G. Theoretical analysis of epigenetic cell memory by nucleosome modification. Cell 2007;129:813–22. Yang H, Howard M, Dean C. Antagonistic roles for H3K36me3 and H3K27me3 in the cold-induced epigenetic switch at Arabidopsis FLC. Curr Biol 2014;24:1793–7. Corbesier L, Vincent C, Jang S, Fornara F, Fan Q, Searle I, et al. FT protein movement contributes to long-distance signaling in floral induction of Arabidopsis. Science 2007; 316:1030–3. Lanzuolo C, Roure V, Dekker J, Bantignies F, Orlando V. Polycomb response elements mediate the formation of chromosome higher-order structures in the bithorax complex. Nat Cell Biol 2007;9:1167–74. Bantignies F, Roure V, Comet I, Leblanc B, Schuettengruber B, Bonnet J, et al. Polycomb-dependent regulatory contacts between distant Hox loci in Drosophila. Cell 2011;144:214–26. Noordermeer D, Leleu M, Splinter E, Rougemont J, Laat WD, Duboule D. The dynamic architecture of Hox gene clusters. Science 2011;334:222–5. Nora EP, Lajoie BR, Schulz EG, Giorgetti L, Okamoto I, Servant N, et al. Spatial partitioning of the regulatory landscape of the X-inactivation centre. Nature 2012;485: 381–5. Fraser P, Bickmore W. Nuclear organization of the genome and the potential for gene regulation. Nature 2007;447: 413–7. Gartenberg MR, Neumann FR, Laroche T, Blaszczyk M, Gasser SM. Sir-mediated repression can occur independently of chromosomal and subnuclear contexts. Cell 2004;119: 955–67. Towbin BD, Gonzalez-Aguilera C, Sack R, Gaidatzis D, Kalck V, Meister P, et al. Step-wise methylation of histone H3K9 positions heterochromatin at the nuclear periphery. Cell 2012;150:934–47.

Review: Nuclear Organization Changes at FLC

[73] Brickner JH. Transcriptional memory at the nuclear periphery. Curr Opin Cell Biol 2009;21:127–33. [74] O'Sullivan JM, Tan-Wong SM, Morillon A, Lee B, Coles J, Mellor J, et al. Gene loops juxtapose promoters and terminators in yeast. Nat Genet 2004;36:1014–8. [75] Ansari A, Hampsey M. A role for the CPF 3′-end processing machinery in RNAP II-dependent gene looping. Genes Dev 2005;19:2969–78. [76] Laine JP, Singh BN, Krishnamurthy S, Hampsey M. A physiological role for gene loops in yeast. Genes Dev 2009; 23:2604–9. [77] Tan-Wong SM, Wijayatilake HD, Proudfoot NJ. Gene loops function to maintain transcriptional memory through interaction with the nuclear pore complex. Genes Dev 2009;23: 2610–24. [78] Crevillen P, Sonmez C, Wu Z, Dean C. A gene loop containing the floral repressor FLC is disrupted in the early phase of vernalization. EMBO J 2013;32:140–8. [79] Jegu T, Latrasse D, Delarue M, Hirt H, Domenichini S, Ariel F, et al. The BAF60 subunit of the SWI/SNF chromatinremodeling complex directly controls the formation of a gene loop at FLOWERING LOCUS C in Arabidopsis. Plant Cell 2014;26:538–51. [80] Bantignies F, Grimaud C, Lavrov S, Gabut M, Cavalli G. Inheritance of Polycomb-dependent chromosomal interactions in Drosophila. Genes Dev 2003;17:2406–20. [81] Cheutin T, Cavalli G. Polycomb silencing: from linear chromatin domains to 3D chromosome folding. Curr Opin Genet Dev 2014;25:30–7. [82] Grimaud C, Bantignies F, Pal-Bhadra M, Ghana P, Bhadra U, Cavalli G. RNAi components are required for nuclear clustering of Polycomb group response elements. Cell 2006;124:957–71. [83] Rosa S, De Lucia F, Mylne JS, Zhu D, Ohmido N, Pendle A, et al. Physical clustering of FLC alleles during Polycombmediated epigenetic silencing in vernalization. Genes Dev 2013;27:1845–50. [84] Masui O, Bonnet I, Le Baccon P, Brito I, Pollex T, Murphy N, et al. Live-cell chromosome dynamics and outcome of X chromosome pairing events during ES cell differentiation. Cell 2011;145:447–58. [85] Pollex T, Heard E. Recent advances in X-chromosome inactivation research. Curr Opin Cell Biol 2012;24:825–32. [86] Splinter E, De Wit E, Nora EP, Klous P, Van de Werken HJ, Zhu Y, et al. The inactive X chromosome adopts a unique three-dimensional conformation that is dependent on Xist RNA. Genes Dev 2011;25:1371–83. [87] Gendall AR, Levy YY, Wilson A, Dean C. The VERNALIZATION 2 gene mediates the epigenetic regulation of vernalization in Arabidopsis. Cell 2001;107:525–35. [88] Octavio LM, Gedeon K, Maheshri N. Epigenetic and conventional regulation is distributed among activators of FLO11 allowing tuning of population-level heterogeneity in its expression. PLoS Genet 2009;5:e1000673. [89] Bumgarner SL, Dowell RD, Grisafi P, Gifford DK, Fink GR. Toggle involving cis-interfering noncoding RNAs controls variegated gene expression in yeast. Proc Natl Acad Sci USA 2009;106:18321–6.

669 [90] Giorgetti L, Galupa R, Nora EP, Piolot T, Lam F, Dekker J, et al. Predictive polymer modeling reveals coupled fluctuations in chromosome conformation and transcription. Cell 2014; 157:950–63. [91] Sheldon CC, Finnegan EJ, Peacock WJ, Dennis ES. Mechanisms of gene repression by vernalization in Arabidopsis. Plant J 2009;59:488–98. [92] Tsai MC, Manor O, Wan Y, Mosammaparast N, Wang JK, Lan F, et al. Long noncoding RNA as modular scaffold of histone modification complexes. Science 2010;329: 689–93. [93] Mendenhall EM, Koche RP, Truong T, Zhou VW, Lssac B, Chi AS, et al. GC-rich sequence elements recruit PRC2 in mammalian ES cells. PLoS Genet 2010;6:e1001244. [94] Farcas AM, Blackledge NP, Sudbery I, Long HK, McGouran JF, Rose NR, et al. KDM2B links the Polycomb Repressive Complex 1 (PRC1) to recognition of CpG islands. Elife 2012;1:e00205. [95] Riising EM, Comet I, Leblanc B, Wu X, Johansen JV, Helin K. Gene silencing triggers Polycomb repressive complex 2 recruitment to CpG islands genome wide. Mol Cell 2014;55: 1–14. [96] Sheldon CC, Conn AB, Dennis ES, Peacock WJ. Different regulatory regions are required for the vernalization-induced repression of FLOWERING LOCUS C and for the epigenetic maintenance of repression. Plant Cell 2002;14:2527–37. [97] Coustham V, Li PJ, Strange A, Lister C, Song J, Dean C. Quantitative modulation of polycomb silencing underlies natural variation in vernalization. Science 2013;337:584–7. [98] Shindo C, Lister C, Crevillen P, Nordborg M, Dean C. Variation in the epigenetic silencing of FLC contributes to natural variation in Arabidopsis vernalization response. Genes Dev 2006;20:3079–83. [99] Papantonis A, Cook PR. Fixing the model for transcription: the DNA moves, not the polymerase. Transcription 2011;2:41–4. [100] Sutherland H, Bickmore WA. Transcription factories: gene expression in unions? Nat Rev Genet 2009;10:457–66. [101] Edelman LB, Fraser P. Transcription factories: genetic programming in three dimensions. Curr Opin Genet Dev 2012;22:110–4. [102] Pirrotta V, Li HB. A view of nuclear Polycomb bodies. Curr Opin Genet Dev 2012;22:101–9. [103] Schmitz RJ, Hong L, Fitzpatrick KE, Amasino RM. DICERLIKE 1 and DICER-LIKE 3 redundantly act to promote flowering via repression of FLOWERING LOCUS C in Arabidopsis thaliana. Genetics 2007;176:1359–62. [104] Cavalli G, Misteli T. Functional implications of genome topology. Nat Struct Mol Biol 2013;20:290–9. [105] Xu N, Donohoe ME, Silva SS, Lee JT. Evidence that homologous X-chromosome pairing requires transcription and Ctcf protein. Nat Genet 2007;39:1390–6. [106] Donohoe ME, Silva SS, Pinter SF, Xu N, Lee JT. The pluripotency factor Oct4 interacts with Ctcf and also controls Xchromosome pairing and counting. Nature 2009;460:128–32. [107] Buhler M, Gasser SM. Silent chromatin at the middle and ends: lessons from yeasts. EMBO J 2009;28:2149–61. [108] Moazed D. Mechanisms for the inheritance of chromatin states. Cell 2011;146:510–8.