The BAF60 Subunit of the SWI/SNF Chromatin-RemodelingComplex Directly Controls the Formation of a Gene Loop atFLOWERING LOCUS C in ArabidopsisW

Teddy Jégu,a David Latrasse,a Marianne Delarue,a Heribert Hirt,b Séverine Domenichini,a Federico Ariel,c

Martin Crespi,c Catherine Bergounioux,a Cécile Raynaud,a and Moussa Benhameda,1

a Institut de Biologie des Plantes, Unité Mixte de Recherche 8618 Université Paris-Sud XI, 91405 Orsay, Franceb Institut des Sciences du Végétal, UPR CNRS, F-91190 Gif-sur-Yvette, FrancecUnité de Recherche en Génomique Végétale Plant Genomics, INRA/CNRS/University of Evry, F-91057 Evry, France

SWI/SNF complexes mediate ATP-dependent chromatin remodeling to regulate gene expression. Many components of thesecomplexes are evolutionarily conserved, and several subunits of Arabidopsis thaliana SWI/SNF complexes are involved in thecontrol of flowering, a process that depends on the floral repressor FLOWERING LOCUS C (FLC). BAF60 is a SWI/SNFsubunit, and in this work, we show that BAF60, via a direct targeting of the floral repressor FLC, induces a change at the high-order chromatin level and represses the photoperiod flowering pathway in Arabidopsis. BAF60 accumulates in the nucleusand controls the formation of the FLC gene loop by modulation of histone density, composition, and posttranslationalmodification. Physiological analysis of BAF60 RNA interference mutant lines allowed us to propose that this chromatin-remodeling protein creates a repressive chromatin configuration at the FLC locus.

INTRODUCTION

DNA/histone complexes generate a barrier, thereby reducing theaccessibility of transcription factors and the general transcrip-tional machinery to DNA. Among the mechanisms that haveevolved to overcome this barrier is chromatin remodeling.Chromatin remodelers, which have been referred to as chromatin-remodeling machines (CRMs), are multiple-subunit complexesthat use the energy of ATP hydrolysis to modify DNA–histoneinteractions and alter the location or conformation of nucleo-somes (Clapier and Cairns, 2009). These multiple-protein com-plexes, with unique subunit compositions, regulate access tochromatin DNA by influencing the structure, the type of histonevariants, and nucleosome positioning. Hence, these complexesmediate a remarkably broad spectrum of biological processesinvolving chromatin. The distinctive feature of all CRMs is thepresence of a central ATPase, belonging to the SWI2/SNF2family (Bork and Koonin, 1993; Eisen et al., 1995). The Arabidopsisthaliana genome encodes 41 Snf2-related proteins distributed intosix superfamilies, among which the Snf2-like family encom-passes the motor subunits for SWI/SNF complexes (Knizewskiet al., 2008).

The SWI/SNF complex is recruited to the promoter region ofa target gene by specific transcription factors in order to facili-tate or repress nearby transcription (Weissman and Knudsen,2009). The structure of this complex is highly conserved, and

SWI/SNF subunits can be subclassified into three categories: (1)the above-mentioned enzymatic ATPase, (2) core subunits, and(3) accessory subunits. In mammals, the chromatin-remodelingactivity of the SWI/SNF complex is essential for developmentalprocesses, hormone responses, and tumor suppression, and theaccessory subunits are hypothesized to dictate the specificity ofSWI/SNF complex function (Wu, 2012).In plants, experimental analyses and database surveys in-

dicate the probable existence of SWI/SNF-type complexes, butno complete plant CRM has been isolated and characterizedso far (Jerzmanowski, 2007). Nevertheless, several componentsof these complexes are evolutionarily conserved in plants, suchas the SWI3 subunits (Sarnowski et al., 2005), the SNF5/BSHsubunit (Brzeski et al., 1999), the nuclear actin-related proteinARP4 (Kandasamy et al., 2005), BRAHMA (BRM), or SPLAYED(SYD). Both of these latter proteins are ATPases of ArabidopsisSWI/SNF complexes and have been shown to participate in thecontrol of flower development and flowering time (Wagner andMeyerowitz, 2002; Farrona et al., 2004; Hurtado et al., 2006;Bezhani et al., 2007; Fornara et al., 2010; Wu, 2012). Likewise,the SWI3B protein interacts with the flowering regulator FCA(Sarnowski et al., 2005). Among plant developmental processes,flowering is one of the most studied. Research over the last 15years, mainly in Arabidopsis, has provided a good understandingof how different environmental and endogenous cues are in-tegrated to cause a developmental switch from the shoot apicalmeristem to a flower (He, 2009). Floral induction is regulated byexogenous (daylength and temperature) and endogenous signals;for example, in the case of Arabidopsis, flowering is induced bylong days and cold treatment (vernalization), but plants main-tained in short-day (SD) conditions will eventually flower, owingto the autonomous pathway of flowering time regulation. Twocentral flowering regulators, CONSTANS (CO) and FLOWERING

1Address correspondence to moussa.benhamed@u-psud.fr.The author responsible for distribution of materials integral to the findingspresented in this article in accordance with the policy described in theInstructions for Authors (www.plantcell.org) is: Moussa Benhamed(moussa.benhamed@u-psud.fr).W Online version contains Web-only data.www.plantcell.org/cgi/doi/10.1105/tpc.113.114454

This article is a Plant Cell Advance Online Publication. The date of its first appearance online is the official date of publication. The article has been

edited and the authors have corrected proofs, but minor changes could be made before the final version is published. Posting this version online

reduces the time to publication by several weeks.

The Plant Cell Preview, www.aspb.org ã 2014 American Society of Plant Biologists. All rights reserved. 1 of 14

LOCUS C (FLC), antagonistically regulate flowering (Putterillet al., 1995; Samach et al., 2000): the CO gene, encoding a zincfinger protein, acts as a floral activator and mediates the pho-toperiod pathway in long-day (LD) conditions, whereas the FLCgene, encoding a MADS box protein, acts as a floral repressorand mediates the autonomous pathway as well as the vernali-zation pathway. In turn, CO and FLC regulate the expression ofdownstream genes, the so-called flowering pathway integratorsFLOWERING LOCUS T (FT), SUPPRESSOROFOVEREXPRESSIONOF CONSTANS1 (SOC1), and LEAFY. These three genes integratesignals from multiple flowering pathways, and their expressionlevels determine flowering time (Simpson and Dean, 2002; Parcy,2005). In SD conditions, the CO pathway does not seem to beinvolved in the control of flowering, since the CO protein is notbeing stably produced (Valverde et al., 2004). Under this condi-tion, induction of SOC1 is mediated by gibberellic acid signaling,as evidenced by the gibberellic acid biosynthetic mutant ga1-3,which fails to flower under SD conditions (Wilson et al., 1992).

Analyses of mutations that activate or silence the expressionof FLC have revealed a conserved regulatory pathway thatmodulates transcription via chromatin regulation by histoneposttranslational modifications. For instance, early-flowering mu-tants with reduced levels of FLC expression display elevatedlevels of trimethylation on lysine-27 of histone H3 (H3K27me3)at this locus (Doyle and Amasino, 2009), whereas FLC expressioncorrelates positively with acetylation on lysine-9 of histone H3(H3K9Ac; Wang et al., 2006).

A series of early-flowering mutants identified a major role forthe SWR1 complex in FLC expression. The chromatin remod-eling SWR1/SRCAP complex belongs to a different class ofchromatin remodelers than SWI/SNF complexes and relies onthe activity of ATPases from the SWR1-like family, comprisingINO80 and SWR1 (Knizewski et al., 2008). This complex is in-volved in the substitution of histone H2A by the histone variantH2A.Z (March-Díaz and Reyes, 2009). Indeed, mutations in PIE1,ARP6 (also known as SUF3 and ESD1 in Arabidopsis), and SEF/AtSW6C show early flowering and reduced FLC expression (Nohand Amasino, 2003; Choi et al., 2005, 2007; Deal et al., 2005;Martin-Trillo et al., 2006; March-Díaz et al., 2007; Lázaro et al.,2008), and ARP6 and PIE1 were found to be required for H2A.Zdeposition at FLC (Deal et al., 2007). Finally, it has been shownthat disruption of BRM results in high FLC expression and thatthe brm mutation restores high accumulation of FLC mRNAin pie and sef mutants, suggesting that SWR1/SCRAP andSWI/SNF complexes play antagonistic roles in FLC regulation(Farrona et al., 2011). Whether BRM acts by modulating H2A.Zincorporation remains to be established, but expression of FLCin the brm mutant does not require the incorporation of H2A.Z,suggesting that this may be the case. H2A.Z is associated withboth gene activation and gene repression and is frequently en-riched at the 59 and 39 ends of genes (Marques et al., 2010). Inthe case of FLC, incorporation of H2A.Z at this locus would leadto relative nucleosome instability, thereby facilitating nucleo-some displacement by the transcription machinery.

More recently, Crevillén et al. (2013) have used quantitativechromosome conformation capture (3C) in a range of Arabi-dopsis genotypes and showed that higher order chromatinstructure is probably involved in FLC regulation. They detected

a gene loop at the FLC locus, reflecting interaction betweensequences in the FLC 59 flanking region with sequences in the 39flanking region, and clearly demonstrated that this loop structurecould be efficiently disrupted during vernalization. This loop didnot reform after transfer of plants back to warm conditions.However, the mechanism involved in the formation and/or sta-bilization of this structure remains an open question.As stated above, several components of the Arabidopsis SWI/

SNF complexes are involved in the control of flowering. In ad-dition, SWI/SNF complexes have been shown to produce in vitroDNA loops in a nucleosome-dependent manner (Zhang et al.,2006). This prompted us to investigate the function of anotherputative subunit of SWI/SNF complexes in the control of flow-ering. The SWP73 component of the yeast (Saccharomycescerevisiae) SWI/SNF complex has a functional role in tran-scriptional activation (Cairns et al., 1996). Orthologs of SWP73,called BAP60 in Drosophila melanogaster and BAF60 in human,have been extensively characterized biochemically and geneti-cally (Tang et al., 2010). The Arabidopsis genome encodes twohighly similar homologs of SWP73: At5g14170 (also namedAt BAF60, At SWP73B, or CHC1 according to the ChromatinConsortium Database [http://chromdb.org]) and At3g01890 (alsonamed At SWP73A). Although BAF60 has been shown to be in-volved in Agrobacterium tumefaciens–mediated root transformation(Crane and Gelvin, 2007), the repair of UV-B light–damaged DNA(Campi et al., 2012), and probably cell cycle regulation (Jéguet al., 2013), its molecular role in chromatin remodeling and plantdevelopment remains unclear.In this work, we investigated the function of Arabidopsis

BAF60 at the chromatin level and in the regulation of floweringtime. We show that BAF60 is required for the repression of FLCexpression. Downregulation of BAF60 results in a late-floweringphenotype in LD conditions, mainly through the increase of FLCexpression in correlation with the decrease of its downstreamtargets. Furthermore, BAF60 regulates histone mark modifications,H2A.Z deposition, chromatin condensation, and loop formation viaa direct binding at the FLC locus. BAF60 thus regulates theexpression of FLC by creating a repressive chromatin configu-ration at this locus and plays a key role in the control of chro-matin structure.

RESULTS

H2A.Z Deposition and Chromatin Condensation AreImpaired in BAF60 RNA Interference Lines

To perform a functional analysis of BAF60 in Arabidopsis, we firstsearched for loss-of-function mutants in publicly available mutantdatabases. Unfortunately, all T-DNA insertion lines showed insertionsin the promoter or 59 untranslated region, and mRNA accumu-lation remained unchanged in homozygous lines (SupplementalFigure 1). Therefore, we used two independent RNA interference(RNAi) lines described previously (Crane and Gelvin, 2007) forwhich the extent of silencing was documented by RT-PCR at theconsortium database (http://chromdb.org). Quantitative real-timeRT-PCR (qRT-PCR) confirmed the down-regulation of BAF60 inthese lines (Supplemental Figure 2).

2 of 14 The Plant Cell

ATP-dependent chromatin-remodeling complexes may act inconcert with various histone-modifying enzymes to modulatechromatin structure (Hargreaves and Crabtree, 2011). The globalpattern of different histone modifications was thus analyzed byimmunoblot assays in BAF60 RNAi lines, but no obvious dif-ference was found. We next asked whether global chromatincondensation could be modified in BAF60 RNAi lines. Micro-coccal nuclease (MNase) digestion provides a useful tool foranalyzing chromatin structure because it cleaves the linker DNAbetween adjacent nucleosomes. Depending on the extent ofdigestion, chromatin fragments with different sizes correspondingto nucleosome number are generated and form a specificDNA ladder reflecting the level of chromatin compaction. Totest the consequences of BAF60 down-regulation on chromatinstructure, we performed a limited MNase digestion assay. Thepatterns of permeability to nuclease digestion (Figures 1A

and 1B; Supplemental Figure 3) indicated a nuclease hyper-sensitivity of the BAF60 RNAi line chromatin compared with thewild type.H2A.Z histone variants are known to facilitate intramolecular

folding of nucleosomal arrays while simultaneously inhibiting theformation of highly condensed structures resulting from in-termolecular associations (Fan et al., 2002). Previous studiessuggest that BRM, the ATPase subunit hypothesized to functionin the same complex as BAF60, might be involved in H2A.Zdeposition (Farrona et al., 2011). Therefore, we tested H2A.Zdeposition in the BAF60 knockdown lines. Immunoblot analysisrevealed the higher accumulation of H2A.Z in both RNAi lines(Figure 1C). The specificity of the H2A.Z antibody used wasdemonstrated using the double mutant hta9 hta11, in which twoof the three genes encoding H2A.Z variants are inactivated. Inour conditions, H2A.Z accumulation was undetectable in this

Figure 1. BAF60 Is Involved in H2A.Z Deposition and Chromatin Condensation.

(A) Nuclei from wild-type or RNAi BAF60 cells were isolated and analyzed in an MNase digestion assay. The reaction was stopped after 1, 4, 7, and10 min, and equal amounts of digested DNA were loaded on agarose gels and stained with ethidium bromide. The red arrows indicate nucleosomes.(B) Curves representing the intensity of the bands corresponding to mononucleosomes as a function of time.(C) Nuclear protein fractions extracted from wild-type or BAF60 RNAi plants were analyzed by immunoblotting using an anti-H2A.Z antibody. Therelative quantifications (right panel) were normalized using an antibody raised against histone H3. Error bars refer to SD of three biological replicates.

BAF60 Controls Gene Loop Formation 3 of 14

mutant (Supplemental Figure 4). To confirm these results, weperformed an immunostaining experiment on the wild type andboth RNAi lines. Quantification of the H2A.Z fluorescent signalrevealed a higher accumulation of H2A.Z in both RNAi lines,confirming our protein gel blot analysis (Supplemental Figure 5).In order to determine whether this global increase was due to anenhanced expression of H2A.Z variants, quantitative analysis ofthe expression of genes encoding H2A.Z was performed. TheArabidopsis genome encodes 13 H2A histones, 3 of which areH2A.Z variants (Deal et al., 2007). Quantitative PCR showed thatin Arabidopsis, H2A.Z mRNA was mainly transcribed from theHTA9 gene (Supplemental Figure 6) and that no increase in theexpression of this gene was detected in BAF60 RNAi lines,suggesting that H2A.Z accumulation is regulated at the proteinlevel. As non-chromatin-bound histones are rapidly degradedby phosphorylation and ubiquitylation-dependent proteolysis(Singh et al., 2009), our results support the hypothesis thatBAF60 is involved in negatively regulating H2A.Z deposition intochromatin, which leads to H2A.Z stabilization when BAF60 isdownregulated.

BAF60 Influences Flowering Time and CO, FT, SOC1, andFLC Expression in Long and Short Days

Among genes known to be regulated via H2A.Z incorporationinto chromatin, several are involved in the control of floweringtime (Deal et al., 2007; Farrona et al., 2011). Hence, we com-pared the flowering time of BAF60 RNAi lines and Wassilewskijacontrol plants in two independent experiments. Both RNAi lines,grown under normal growth conditions (LD photoperiod), dis-played a late-flowering phenotype measured by the number ofdays to bolting and the rosette leaf number at flowering (Figures2A and 2B). Wild-type plants bolted nearly 17 d after sowing,whereas 23 d were required for the RNAi lines, which also hadincreased number of rosette leaves in comparison with the wildtype (Figure 2B). This phenotype was associated with a de-crease in leaf size, as shown in Figure 2A.

This late-flowering phenotype observed in LD conditionssuggests that BAF60 could be involved in the photoperiodpathway of flowering. This pathway is mediated by CO, a tran-scription factor required to activate the expression of FT underLD conditions. FT moves from leaves to the apical meristem totrigger a cascade of events that lead to plant flowering, includingthe activation of SOC1 expression (Fornara et al., 2010). Toaddress the molecular mechanisms underlying the late-floweringphenotype of BAF60 RNAi lines, we quantified the expression ofthese different regulators in these plants. Because CO, FT, andSOC1 mRNA levels show a circadian rhythm, their expressionwas examined during a 24-h period. In both RNAi lines, theaccumulation of CO mRNA was only slightly upregulated com-pared with wild-type plants (Figure 3; Supplemental Figure 7).Consistently, a slight overaccumulation of FT mRNA, in contrastto a 3-fold reduction of SOC1 mRNA, was detected in BAF60RNAi lines compared with the control (Figure 3). To correlate thelate-flowering phenotype observed in the RNAi lines and SOC1downregulation, we performed a time-course analysis of SOC1expression in wild-type and RNAi plants growing under LDconditions. In the wild type, we observed that SOC1 expression

reached a peak 19 d after sowing, whereas this peak was ob-served more than 23 d after sowing in the RNAi lines (SupplementalFigure 8). Hence, SOC1 misregulation was consistent with the late-flowering phenotype.Knowing that FT and SOC1 are targets of FLC, we then as-

sayed the expression of FLC, which was 7-fold upregulated inthe RNAi lines. Taken together, these results provide evidencethat genes involved in the control of flowering time are de-regulated in BAF60 RNAi lines. The resulting balance betweenthe action of activators and repressors of flowering favors the

Figure 2. BAF60 Regulates Flowering Time.

(A) Twenty-day-old wild-type plant and two independent BAF60 RNAilines grown under LD conditions.(B) Flowering time of plants grown under LD photoperiod. Values shownare averages 6 SD of days (top) or number of rosette leaves (bottom) atflowering of at least 30 plants.

4 of 14 The Plant Cell

inhibition of flowering via upregulation of FLC and down-regulation of SOC1.

To further explore this hypothesis, the flowering time of theBAF60 RNAi lines introgressed into the flc background wascompared with that of the flc mutant and the wild-type controlplants. In two independent experiments, the flc BAF60 RNAilines, grown under normal growth conditions (LD photoperiod),displayed an early-flowering phenotype, measured by the daysto bolting and the rosette leaf number at flowering (Figure 2B).These results demonstrate that the late-flowering phenotypeobserved in the RNAi lines was due to the overexpressionof FLC. Since BAF60 affects FLC expression, we also askedwhether BAF60 could be involved in the autonomous pathwayregulating flowering time. Under SD conditions, considering thenumber of days to bolting, the flowering time was the same forall genotypes. However, the number of leaves at bolting waslower in both BAF60 RNAi lines compared with the control(Supplemental Figure 9). CO and FT were initially identified asmutants that flower late in LD conditions but not in SD con-ditions (Koornneef et al., 1991). However, ectopic expression ofCO leads to early flowering, irrespective of daylength (Onouchiet al., 2000), showing that transcriptional regulation of CO ex-pression is an important determinant in the control of flowering.Therefore, we examined the expression of CO, FT, and SOC1 byqRT-PCR in SD conditions. The CO mRNA level was 4-foldupregulated in correlation with a 5-fold increase in accumulationof FT mRNA and increased SOC1 mRNA levels in both RNAi

lines compared with the wild type (Supplemental Figure 10).Although we also observed an increase in the FLC mRNA level,this seemed insufficient to repress FT and SOC1 (SupplementalFigure 10). Taken together, these results suggest that the up-regulation of CO could induce the early flowering observed in SDconditions (Supplemental Figure 11).

BAF60 Influences the Expression of FLC by RegulatingHistone Modifications and H2A.Z Deposition

FLC repressors and activators may act on chromatin structure atthe FLC locus either via modification of specific histones or viaH2A.Z deposition (Crevillén and Dean, 2011). Since BAF60 mayalter the chromatin configuration of its target genes, we ana-lyzed how the absence of BAF60 affects histone modificationsat the FLC locus. Both a transcription-repressing mark (H3K27me3)and a transcription-activating mark (H3K9Ac) were assessedby chromatin immunoprecipitation (ChIP) assays using primersspecific for three regions of the FLC locus, covering the FLCpromoter (distal and proximal regions), the first exon, and thefirst intron of the gene (Figure 4A).In the BAF60 RNAi line, H3K27me3 was decreased while

H3K9Ac was increased in the 59 part of the FLC locus (Figure4B). These results are consistent with the observed FLC upre-gulation caused by the downregulation of BAF60. We next ex-amined the possibility that BAF60 could be involved in H2A.Zdeposition at the FLC locus. ChIP assays were performed on

Figure 3. BAF60 Controls the Expression of Flowering Genes.

qRT-PCR data show the relative expression of the indicated genes in wild-type and BAF60 RNAi plants in LD conditions. Total RNA samples werecollected every 4 h from 14-d-old plants during a 24-h period. mRNA abundance was quantified by qRT-PCR and expressed relative to the abundanceof UBQ10 transcripts. Gray areas behind the traces represent night periods. Average relative quantity 6 SD is shown for each sample.

BAF60 Controls Gene Loop Formation 5 of 14

wild-type and BAF60 RNAi lines using an H2A.Z antibody. In bothgenetic backgrounds, chromatin H2A.Z incorporation was re-stricted to the FLC exons 1, 6, and 7, and this incorporation washigher in the RNAi lines (Figure 4B; Supplemental Figure 12).

H2A.Z-containing nucleosomes facilitate RNA polymerase IIpassage by affecting Pol II elongation complexes and favoringefficient nucleosome remodeling over the gene (Santistebanet al., 2011). To assess the effect of H2A.Z enrichment on RNApolymerase II occupancy, ChIP assays were performed using anantibody raised against RNA polymerase II. We observed thatRNA polymerase II was present on the proximal promoter of FLC in

both genotypes, with a 12-fold increase in RNA polymerase II oc-cupancy in the BAF60 RNAi line (Figure 4B). Altogether, our resultssuggest that BAF60 inhibits H3K9Ac, H2A.Z, and RNA polymeraseII deposition at the FLC locus to promote its repression.

BAF60 Regulates the FLC Loop

Taken together, our results show that BAF60 downregulationresults in a general increase in H2A.Z incorporation into chro-matin, notably at the FLC locus. To determine whether FLC is adirect target of BAF60, we first generated plants overexpressinga cyan fluorescent protein (CFP)–tagged version of BAF60. Thisallowed us to check that the protein accumulates in the nucleus(Supplemental Figure 13). We then examined the ability of BAF60to bind to the FLC locus by ChIP using an anti-GFP antibody. Thesame assay using antibodies against IgG was used as a negativecontrol. As shown in Figure 5, we observed a specific associa-tion of BAF60 with the promoter region of the FLC gene, the firstintron, and the 39 part of the locus.Considering the binding of BAF60 on FLC and the higher in-

corporation of the H2A.Z variant at this locus in the BAF60 RNAilines, we wondered whether BAF60 could influence chromatincondensation in this particular genomic region. In order tomeasure DNA accessibility and to obtain a quantitative as-sessment of chromatin condensation at the FLC locus, weperformed a limited MNase digestion followed by real-time PCR.As shown in Figure 6, in the wild type, we observed a nuclease-hypersensitive site in the proximal part of the FLC promoter,whereas no obvious difference was noticed in the two other partsof the genomic region tested. This result is in accordance witha previous analysis showing an apparent nucleosome-free region

Figure 4. BAF60 Controls Histone Modification State and RNA Poly-merase II Occupancy at the FLC Locus.

(A) Schematic representation of the regions of the FLC locus analyzed.Black boxes correspond to exons, the arrow indicates the site oftranslation initiation, and numbers indicate the positions of primer pairsused.(B) Quantification data of the ChIP results. Nuclei were extracted from14-d-old seedlings grown under LD conditions, and immunoprecipitationwas performed with antibodies specific for H3K27me3, H3K9Ac, H2A.Z,and RNA polymerase II. Average relative quantity 6 SD is shown for eachsample (two biological replicates).

Figure 5. BAF60 Binds to the FLC Locus.

Top, schematic representation of the FLC locus. The position of eachprimer pair used for ChIP-quantitative PCR is indicated. Arrows indicatethe positions of the transcription start sites. Exons are represented asblack boxes and introns as black lines. Bottom, quantification data of ChIPresults. Nuclei were extracted after cross-linking from 14-d-old seedlingsexpressing the BAF60-CFP transgene and sonicated, and chromatin/protein complexes were immunoprecipitated with antibodies specific forGFP or IgG. Error bars refer to SD of three biological replicates.

6 of 14 The Plant Cell

immediately upstream of the transcription start site (Shivaswamyet al., 2008). However, in BAF60 RNAi plants, we observednuclease hypersensitivity for the three genomic regions tested,reflecting a reduction in chromatin condensation and thus probablyin nucleosome occupancy at the FLC locus in BAF60 knockdownlines.

Recently, Crevillén et al. (2013) demonstrated that the FLClocus could make a loop that plays a role in its transcriptionalregulation. Since histone depletion facilitates chromatin loops(Diesinger et al., 2010), we performed 3C experiments to explorewhether BAF60 could regulate the maintenance of this high-order chromatin structure. First, we confirmed the results showinga chromatin loop at the FLC locus in the wild type and comparedit with BAF60 RNAi lines (Figure 7; Supplemental Figure 14). Asa control, we verified that no PCR product was observed whenour plant material was not cross-linked or ligase was not added inthe 3C reaction. Interestingly, we observed that twice as manyFLC loops were present and that this structure was more stable inthe BAF60 RNAi context. These results suggest that BAF60 playsa negative role in loop formation and/or stabilization, at least atthe FLC locus, and allowed us to elucidate the molecular mech-anisms governing chromatin conformation involved in the controlof flowering.

BAF60 Is Necessary for Efficient Disruption of the FLC Loopafter Vernalization

Since a major conclusion from Crevillén et al. (2013) was that theFLC gene loop could be disrupted by vernalization, we analyzedthe flowering phenotype of both RNAi lines after vernalization.The vernalization treatment efficiently reduced flowering time inthe wild type: as expected, plants displayed a significant re-duction of the rosette leaf number at flowering (Student’s t test,P = 0.05) and of the number of days to bolting (Student’s t test,P = 0.05; Figures 8A and 8B). However, we observed that thevernalization did not have any effect on the two RNAi lines

regarding the number of days to bolting and the rosette leafnumber at flowering (Figures 8A and 8B).Having established that BAF60 was necessary for the ver-

nalization response, we investigated the relationship betweenthe phenotype observed, the expression of FLC, and the loopfrequency. qRT-PCR analysis revealed that in the wild type, thedecrease in FLC expression started 5 d after vernalizationtreatment, whereas in both RNAi lines, a clear decrease wasobserved only 15 d after vernalization treatment (Figure 8C).These results correlate with the phenotype displayed by the twoRNAi lines and suggested that BAF60 could be a regulator of thevernalization pathway through its action on FLC expression.Then, knowing that BAF60 was required for the loop de-

stabilization and/or inhibition of its formation in control con-ditions, we hypothesized that during vernalization, the phenotypeobserved in both RNAi lines could be a consequence of thepresence of a more stable FLC loop. As reported previously,loop formation was reduced by vernalization in the wild type(Crevillén et al., 2013). Interestingly, we observed that the FLCloop was more stable after vernalization in both RNAi lines(Figure 8D), and these observations were in agreement with theexpression data and the phenotype observed. Altogether, theseresults suggested that BAF60 is necessary for disrupting the loopafter vernalization treatment to reduce FLC expression, therebyinducing flowering.

DISCUSSION

Despite an intensive study of ATP-dependent chromatin-remodelingcomplexes in yeast, a single-celled organism, the function ofthese complexes in developmental processes in higher eukar-yotes remains largely unknown. Even though a SWI/SNF com-plex has not yet been isolated from plant tissues, homologs ofmost of its components are present in Arabidopsis. In orderto gain further insight regarding the role of these complexes,

Figure 6. BAF60 Is Required for Chromatin Condensation at the FLC Locus.

Quantification of MNase sensitivity was measured using qRT-PCR at the FLC locus after a limited MNase digestion with four time point kinetics. Datawere normalized to the control without MNase digestion, and graphical representation of the data is presented below.

BAF60 Controls Gene Loop Formation 7 of 14

we analyzed the molecular function of one accessory subunit,At BAF60, a homolog of the yeast SWP73 protein. We showthat BAF60 is a positive regulator of chromatin condensation,probably by its negative effect on H2A.Z incorporation intochromatin, and, more specifically, that it controls the formationof a chromatin regulatory loop involving a crucial flowering gene,FLC, thereby modulating flowering time and the vernalizationresponse.

We showed that BAF60 RNAi lines displayed a late-floweringphenotype only in LD conditions, together with upregulation of theFLC gene and downregulation of SOC1. Analysis of flc BAF60RNAi lines confirmed that the late-flowering phenotype was at-tributable to FLCmisregulation. Under SD conditions, the floweringtime of the RNAi lines was not significantly different from the controlin terms of number of days, although the number of leaves wasdecreased compared with the wild type. In this case, this pheno-type was correlated with an upregulation of CO, FT, and SOC1,which could not be prevented by FLC activation. Hence, loss ofBAF60 function has opposing effects on regulators of floweringtime, resulting in the simultaneous upregulation of repressors andactivators both in LD and SD conditions. The resulting balancebetween these regulators leads to delayed flowering in LD con-ditions and early flowering in SD conditions.

Our data indicate that BAF60 is a regulator of the photoperiodvia direct targeting of FLC. Therefore, the regulation of FT andSOC1 by BAF60 might take place in at least two different ways:through CO expression and/or independently of CO. The late-flowering phenotype of the knockdown BAF60 RNAi lines in LDconditions can be explained by the higher levels of expression ofFLC, which lead to a downregulation of SOC1 (Figure 9A).Although in SD conditions the CO pathway does not seem to

be involved in the control of flowering (Valverde et al., 2004),overexpression of either CO or FT causes photoperiod-independentearly flowering (Kardailsky et al., 1999; Kobayashi et al., 1999;Onouchi et al., 2000). In addition, high levels of FLC do not inhibitthe early flowering of CO-overexpressing plants (Hepworth et al.,2002). Hence, the upregulation of CO observed under SD con-ditions in BAF60 RNAi plants is likely to account for the observedearly-flowering phenotype and probably cannot be overcomeby the simultaneous upregulation of FLC. Altogether, our resultssuggest that BAF60 is a repressor of COmainly in SD conditions(Supplemental Figure 11). How this photoperiod specificity occursneeds to be determined in the future.Although BRM, like BAF60, is a repressor of FLC and CO,

the SWI/SNF ATPase mutants brm and syd display precociousactivation of FT both in LD and SD conditions (Wagner and

Figure 7. BAF60 Regulates Gene Loop Formation.

Quantitative 3C of the FLC locus is shown using FI as the anchor region in 14-d-old wild-type and RNAi BAF60-2 seedlings. Relative interactionfrequencies were calculated as described in Methods. Data are averages of three biological replicates each with three technical replicates. In the graph,BamHI and BglII restriction sites are indicated with vertical dotted lines, the analyzed FLC regions are numbered with Roman numerals, and the anchorregion is highlighted with gray shading. A schematic representation of the FLC locus is shown above, with the positions of primers used for the 3Canalysis represented by gray arrowheads. TSS, transcription start site.

8 of 14 The Plant Cell

Meyerowitz, 2002; Farrona et al., 2004, 2011; Hurtado et al.,2006). SYD and BRM may have common targets or participatein common processes (Bezhani et al., 2007; Wu, 2012). How-ever, their interaction with BAF60 has not been reported yet, andother partners to which BAF60 could be associated remainunknown. The observation that both BRM and BAF60 repressthe expression of FLC and CO strongly suggests that these twoproteins function in the same complex. The opposing outcomesof the brm mutation and BAF60 downregulation with respect toflowering time in LD conditions is puzzling. These contrastingeffects of mutations predicted to affect the same complexsuggest distinctive activities of the various subunits even thoughthey regulate the same general process: chromatin remodeling.Indeed, downregulation of BAF60 and BRM or SYD results ina different balance between activators and repressors of thefloral transition, leading to opposite flowering phenotypes. Thisobservation highlights the complexity of fine-tuning mecha-nisms that govern the floral transition and suggests that severalsubcomplexes differing by their composition and target genes

may be at work to modulate the balance of flowering regulators(Srikanth and Schmid, 2011).From a molecular point of view, we showed that BAF60 could

directly repress FLC transcription by modulation of its chromatinstructure. Genome topology has emerged as a key player ingenome functions (Cavalli and Misteli, 2013). A contribution oflocal genome chromatin loops in transcription has long beenappreciated (reviewed in Hou and Corces, 2012). Chromatinloops can regulate transcription via a variety of mechanisms, forexample, by bringing a distant enhancer into contact with apromoter (Tolhuis et al., 2002). On the contrary, they can as-sociate several polycomb response elements to target sites withthe polycomb machinery and repress gene expression (Tiwariet al., 2008). Other types of chromatin loops involve insulatorbinding proteins (Yang and Corces, 2012). In the case of the FLClocus, the association between the 59 and 39 ends of the genecould allow efficient recycling of RNA polymerase II and othergeneral transcription factors inside the loop, which are releasedafter the termination of transcription (Ansari and Hampsey,

Figure 8. BAF60 Regulates Gene Loop Dynamics after Vernalization Treatment.

(A) Flowering time measured as days to bolting after vernalization of wild-type plants and BAF60 RNAi lines. Vernalization significantly reducedflowering time in the wild type but not in RNAi lines.(B) Flowering time measured as number of leaves at the time of bolting after vernalization of wild-type plants and BAF60 RNAi lines. The obtainedresults were the same as in (A).(C) qRT-PCR analysis of FLC expression in the wild type (black bars) and BAF60 RNAi lines (red bars). FLC expression was always higher in BAF60RNAi lines than in the wild type, and it only started to decrease 10 d after vernalization instead of 5 d after vernalization in the wild type.(D) 3C assay of loop formation on the FLC locus after vernalization. Interaction between regions I and V decreased in the wild type after vernalization(black bars). By contrast, this association was more stable in the two RNAi lines (red bars) and started to decrease only 10 d after vernalization.

BAF60 Controls Gene Loop Formation 9 of 14

2005). Surprisingly, the FLC loop was not disturbed by muta-tions that affect several chromatin regulatory pathways and leadto different expression levels (Crevillén et al., 2013), indicatingthat expression of FLC can be modulated independently of thespatial organization at this locus. However, the formation of thisloop may mediate the vernalization pathway for FLC regulation,since its formation was disrupted during vernalization, con-comitant with the downregulation of FLC. We showed here thatBAF60 interacts with the 59 and 39 parts of the FLC locus, and itsdownregulation induces the formation and/or stabilization of theloop structure, expression of FLC, and deposition of active epi-genetic marks in this locus. Altogether, our results suggest that

BAF60 could regulate FLC by negatively affecting the loop dy-namics through the modulation of histone density, composition,and posttranslational modification (Figure 6B). Consistently, BAF60RNAi lines are insensitive to vernalization, providing evidence forthe physiological relevance of the involvement of BAF60 in theregulation of FLC loop stability or formation.Another mechanism accounting for the role of BAF60 as

a negative regulator of FLC expression is the incorporation ofH2A.Z into chromatin. Indeed, we showed that BAF60 is in-volved in chromatin condensation at the FLC locus in part viathe modulation of H2A.Z incorporation into chromatin. This maybe consistent with the ability of S. cerevisiae SWI/SNF RSC toremodel nucleosomes, thereby creating “inactive” nucleosomalstates at the locus. Among the players involved in the control ofthe nucleosomal state, the H2A.Z variant plays important rolesboth in activating transcription and antagonizing the spreadof heterochromatin into euchromatic regions (Marques et al.,2010). Previously, data obtained from Drosophila and humansshowed that H2A.Z levels positively correlate with gene ex-pression (Barski et al., 2007; Mavrich et al., 2008). However,H2A.Z variants may not participate directly in the activation oftranscription but rather commit to a chromatin state competentfor activation by other factors or prevent its silencing by DNAmethylation (Zilberman et al., 2008). In Arabidopsis, recent re-sults have shown that H2A.Z is required for gene activation inresponse to warmer temperature (Kumar and Wigge, 2010). Ourdata show that deposition of the histone variant H2A.Z over theFLC locus is correlated with an increase in its transcription, asevidenced by increased mRNA accumulation and higher poly-merase II binding on its proximal promoter and the stabilization/formation of a loop. This is in agreement with recent resultsshowing that the loss of H2A.Z from chromatin in arp6 and pie1mutants results in reduced FLC expression and prematureflowering, indicating that this histone variant is correlated withhigh expression levels of FLC. BRM represses FLC expression,pointing to an antagonistic role of the SWI/SNF and SWR1/SRCAP complexes, on H2A.Z deposition at the FLC locus.Under such a hypothesis, FLC expression levels would resultfrom the balance between these opposing activities on itschromatin. In agreement with this model, analysis of geneticinteractions between BRM and the histone H2A.Z depositionmachinery demonstrates that brm mutations overcome the re-quirement of H2A.Z for FLC activation. In the absence of BRM,a constitutively open chromatin conformation may render H2A.Zdispensable (Farrona et al. 2011). Regarding specifically the loopstructure, Crevillén et al. (2013) demonstrated that the loop waspresent in the arp6 mutant, suggesting that incorporation ofH2A.Z alone could not explain its formation. MNase sensitivityassays revealed that the FLC locus was more accessible todigestion in BAF60 RNAi lines, suggesting that nucleosomes aremore distant in the absence of BAF60. Whether this role ofBAF60 is mediated by H2A.Z incorporation or also results froma direct effect on nucleosome positioning remains to be es-tablished. More generally, our results show a role for BAF60 inthe modulation of H2A.Z incorporation into chromatin, which isnot confined to the FLC locus but observed at the whole-genomescale, as evidenced by the reduced condensation of the genomein BAF60 RNAi lines.

Figure 9. Model for the Role of BAF60 in the Control of Flowering Time.

(A) Genetic model. Under LD conditions and after a vernalization treat-ment in wild-type plants, FLC is repressed by BAF60 to promote flow-ering. In BAF60 RNAi lines, a high level of FLC expression inducesa repression of SOC1 and leads to a late-flowering phenotype.(B) Molecular model. Under LD conditions in wild-type plants, an equi-librium exists between the formation and the release of the FLC loop inwhich BAF60 is involved. This equilibrium could be destabilized by theknockdown of BAF60.

10 of 14 The Plant Cell

The discovery of SWI/SNF-related complexes possessinghistone deacetylase activity, a function shown to repress tran-scription, suggests that nucleosome-remodeling activities canbe coupled with other chromatin-modifying activities to inducetranscriptional repression (Knoepfler and Eisenman, 1999; Laiand Wade, 2011). Our results show that depletion of BAF60also modifies the deposition of two important histone marks,H3K27me3 and H3K9Ac, involved in the repression and acti-vation of the transcriptional state, respectively. These resultssupport a model whereby a combination of specific histonemodifications associated with particular histone variants couldmodulate the dynamics of gene loop formation, although wecannot rule out the possibility that some of the changes in his-tone modifications we observed could be a consequence ofaltered gene expression rather than its cause. In conclusion,BAF60 functions in the control of flowering time by regulatingthe expression of genes through the incorporation of H2A.Z intochromatin, induction of histone modification, and control ofgene loop formation. Our work reveals an important role forBAF60 in gene loop formation in plant chromatin control andraises the possibility that SWI/SNF complexes, notably thosecontaining BAF60, are key factors governing the equilibriumbetween the formation and dissociation of the gene loop ina developmental context such as the control of flowering.

METHODS

Plant Material and Growth Conditions

Arabidopsis thaliana seeds of both RNAi lines, BAF60-1 (CS30982) andBAF60-2 (CS23961), in the Wassilewskija (wild-type) background, wereobtained from the NottinghamArabidopsis StockCentre. The flc-3mutantline has been described previously (Guyomarc’h et al. 2006). Plants weregrown in chambers at 20°C on soil or on sterile half-strength Murashigeand Skoog medium and 0.8% agar in SD (8 h of light at 20°C, 16 h ofdarkness at 18°C) or LD (16 h of light at 20°C, 8 h of darkness at 18°C)conditions. Seeds were surface-sterilized by treatment with bayrochlorefor 20 min, washed, and imbibed in sterile water for 2 to 4 d at 4°C toobtain homogeneous germination. To vernalize, seeds were grown for7 d at standard warm-growing conditions (16 h of light at 20°C, 8 h ofdarkness at 18°C) before being transferred to cold (16 h of light and 8 h ofdarkness at 5°C) conditions for 2 weeks and then returned to standardwarm-growing conditions.

Production of Plants Overexpressing a CFP-Tagged Versionof BAF60

Total RNA was extracted from seedlings or leaves of Arabidopsis ecotypeColumbia with the Nucleospin RNA II kit (Macherey-Nagel) according tothe manufacturer’s instructions. First-strand cDNA was synthesized from2 µg of total RNA using Improm-II reverse transcriptase (A3802; Promega)according to the manufacturer’s instructions. A 1604-bp BAF60 cDNAfrom the start codon, deleted of the last base of the stop codon, wasamplified by PCR from the cDNA pool using the following primers: BAF60-f(59-GGGGTACCATGTCTGGTAACAACAACAATC-39) and BAF60r (59-CCGCTCGAGCACCAACTCCCTGGCCCATCGTTC-39). The PCR productwas cloned into the pGEM-T vector and sequenced. The cDNA was thenexcised via a KpnI/XhoI restriction and subcloned into the pEntr1C vector.Then, to obtain CFP translational fusions (35S:BAF60-CFP), the BAF60cDNA in pEntr1C vector was recombined according to the Gateway pro-cedure (Invitrogen) downstream of the CaMV35S promoter in the binary

vector pB7CWG2 (Plant Systems Biology, Vlaams Interuniversitair Instituutvoor Biotechnologie, Ghent University). The expression vectors wereintroduced into Agrobacterium tumefaciens strain GV3101 by electro-poration, and transgenic lines were generated by the floral dip method.Ten independent lines from the T1 to T6 generations were successivelyselected on sand supplemented with Basta (10 mg/L phosphinothricin).

RNA Extraction and qRT-PCR Analysis

Total RNA was extracted from seedlings with the RNeasy MiniPrep kit(Qiagen) according to the manufacturer’s instructions. First-strand cDNAwas synthesized from 2 mg of total RNA using Improm-II reverse tran-scriptase (A3802; Promega) according to the manufacturer’s instructions.Four percent of the synthesized cDNA was mixed with 100 nM of eachprimer and LightCycler 480 Sybr Green I master mix (Roche AppliedScience) for quantitative PCR analysis. Products were amplified andfluorescent signals acquired with the LightCycler 480 detection system.The specificity of amplification products was determined by meltingcurves. UBQ10 was used as an internal control for signal normalization.Exor4 relative quantification software (Roche Applied Science) auto-matically calculates relative expression levels of the selected genes withalgorithms based on the DDCt method. Data were from duplicates of atleast three biological replicates. The sequences of all primers used aregiven in Supplemental Table 1.

MNase Sensitivity Assay

Five grams of 14-d-old seedlings was ground, and nuclei were isolatedwith 1% Triton X-100, filtered through a 63-µm filter, and incubated in MNbuffer (20 mM Tris-HCl, pH 7.5, 70 mM NaCl, 20 mM KCl, 5 mM MgCl2,and 3 mM CaCl2). One hundred units of MNase (Invitrogen) was added toinitiate the kinetics. Adding EGTA and EDTA to a final concentration of2 mM stopped the reaction. DNA was extracted with chloroform, pre-cipitated with isopropanol, resuspended in ultrapure water, and quantifiedwith Qubit. Sensitivity toMNase was then analyzed by visualizing the DNAdigestion pattern by agarose gel electrophoresis using 1 µg of DNA or byqRT-PCR using 6 ng of digested DNA.

Immunoblot and ChIP Analyses

Nuclear proteins were extracted from 14-d-old in vitro–grown seedlings.After quantification with the Bradford method, identical amounts ofproteins were resolved by SDS-PAGE and then transferred to a poly-vinylidene difluoride membrane (Bio-Rad) using a Mini-Protean 3 Cell(Bio-Rad). Immunoblot analysis was performed using 1 µg/mL primarypolyclonal antibodies raised against histone H2A.Z (Abcam) and histoneH3 (Abcam) and then with secondary antibodies conjugated to alkalinephosphatase. Antibody complexes were detected by chemiluminescenceusing the Immun-Start AP Substrate kit (Bio-Rad).

ChIP assays were performed on 14-d-old in vitro–grown seedlingsusing anti-GFP (Clontech), anti-IgG (Millipore), anti-H3K27me3 (Millipore),anti-H3K9Ac (Millipore), H2A.Z (Abcam), and RNA polymerase II (SantaCruz) antibodies, using a method modified from Gendrel et al. (2005).Briefly, after plant material fixation in 1% (v/v) formaldehyde, the tissueswere homogenized and the nuclei were isolated and lysed. Cross-linkedchromatin was sonicated using a water bath Bioruptor UCD-200 (Dia-genode; 15-s on/15-s off pulses, 15 times). The complexes were im-munoprecipitated with antibodies overnight at 4°C with gentle shakingand incubated for 1 h at 4°C with 50 mL of Dynabeads Protein A (In-vitrogen). Immunoprecipitated DNA was then recovered using the IPurekit (Diagenode) and analyzed by qRT-PCR. An aliquot of untreatedsonicated chromatin was processed in parallel and used as the total inputDNA control.

BAF60 Controls Gene Loop Formation 11 of 14

3C

Two grams of 14-d-old seedlings was cross-linked in 1% (v/v) formal-dehyde at room temperature for 20 min. The cross-linked plantlets wereground, and nuclei were isolated and treated with 0.3% SDS at 65°C for40 min. SDS was sequestered with 2% Triton X-100. Digestions wereperformed overnight at 37°C with 400 units of BamHI (Fermentas) and 400units of BglII (Fermentas). The restriction enzymes were inactivated by theaddition of 1.6% SDS and incubation at 65°C for 20 min. SDS was se-questered with 1% Triton X-100. DNA was ligated by incubation at 22°Cfor 5 h in a 5-mL volume using 100 units of T4 DNA ligase (Fermentas).Reverse cross-linking was performed by overnight treatment at 65°C.DNA was recovered after proteinase K treatment by phenol/chloroformextraction and ethanol precipitation. Relative interaction frequencies werecalculated by qRT-PCR using 15 ng of DNA. A region uncut byBamHI andBglII in the proximal promoter was used to normalize the amount of DNA.

Confocal Imaging

Plantlets were mounted in 5% glycerol and directly imaged on a TCS-SP2upright microscope (Leica Microsystems) with 488/543-nm excitation, 488/543-nm beamsplitter filter, and 515/615-nm (green channel) and 610/625-nm (red channel) detection windows. Transmitted light was also collected.

Immunofluorescence

Seedlings were fixed in 4% paraformaldehyde in PHEM (60 mM PIPES,25 mM HEPES, 10 mM EGTA, and 2 mM MgCl2, pH 6.9) for 1 h at roomtemperature (vacuum was applied for 20 min to facilitate uptake of thefixation solution). Seedlings were washed 5 min in PHEM and 5 min inPBS, pH 6.9, and chopped in a Petri dish in PBS supplemented with 0.1%(w/v) Triton X-100. The mixture was filtered (50 µm) and centrifuged for10 min at 2000g. The supernatant was carefully removed, and the pelletwas washed once with PBS and gently resuspended in 20 mL of PBS, anda drop was placed on a poly-Lys slide and air-dried. The slides wererehydrated with PBS and permeabilized two times by 10min of incubationin PBST (PBS plus 0.1% [v/v] Tween 20). The slides were placed in amoistchamber and incubated overnight at 4°C with primary anti-histone H2A.Zantibody (Abcam) in PBST supplemented with 3% (w/v) BSA. The slideswere washed 53 10min in PBST (at room temperature) and incubated 1 hat room temperature in the dark with the secondary antibody (A11037;Invitrogen; Alexa Fluor goat anti-rabbit) diluted 1:400 (v/v) in PBST and3% BSA. The slides were washed 5 3 10 min in PBST. The slides werethenmounted with a drop of Vectashield with 49,6-diamidino-2-phenylindoleandobservedwith the suitable cube fluorescence filters (BP340-380,DS400,and BP450-490 for 49,6-diamidino-2-phenylindole, BP570-590, DS595, andBP605-655 for A594).

Accession Numbers

Sequence data from this article can be found in the GenBank/EMBLlibraries under the following accession numbers: BAF60 (AT5G14170),FLC (AT5G10140), SOC1 (AT2G45660), FT (AT1G65480), and CO(AT5G15840).

Supplemental Data

The following materials are available in the online version of this article.

Supplemental Figure 1. qRT-PCR Data Showing Relative Expressionof BAF60 (AT5G14170) in Two Available T-DNA Insertion Lines.

Supplemental Figure 2. qRT-PCR Data Showing Relative Expressionof BAF60 in Wild-Type and BAF60 RNAi Plants Cultivated under LD orSD Conditions.

Supplemental Figure 3. BAF60 Is Involved in Chromatin Condensa-tion.

Supplemental Figure 4. Specificity of H2A.Z Antibody.

Supplemental Figure 5. Immunofluorescence Detection and Quanti-fication of H2A.Z in Isolated Nuclei of the Wild Type, RNAi BAF60-1,and RNAi BAF60-2.

Supplemental Figure 6. qRT-PCR Data Showing Relative Expressionof HTA8, HTA9, and HTA11 in Wild-Type and BAF60 RNAi Plants.

Supplemental Figure 7. qRT-PCR Data Showing Relative Expressionof the Indicated Genes in Wild-Type and BAF60 RNAi Plants in SDConditions.

Supplemental Figure 8. qRT-PCR Data Showing Relative Expressionof SOC1 in Wild-Type and BAF60 RNAi Plants in LD Conditions.

Supplemental Figure 9. Flowering Time of Plants Grown under an SDPhotoperiod.

Supplemental Figure 10. qRT-PCR Data Showing Relative Expres-sion of the Indicated Genes in Wild-Type and BAF60 RNAi Plants inSD Conditions.

Supplemental Figure 11. Genetic Model for the Role of BAF60 inFlowering Time Regulation under SD Conditions.

Supplemental Figure 12. BAF60 Controls Histone H2A.Z Incorpora-tion at the FLC Locus.

Supplemental Figure 13. BAF60 Is Localized in Nuclei.

Supplemental Figure 14. Quantitative 3C of the FLC Locus Using FIas the Anchor Region in 14-d-Old Wild-Type and RNAi BAF60-2Seedlings.

Supplemental Table 1. Sequences of Primers Used in This Study(59 to 39).

ACKNOWLEDGMENTS

We thank Philip Wigge for generously providing the double mutanthta9-hta11.

AUTHOR CONTRIBUTIONS

T.J. designed and performed research. D.L., M.D., S.D., and F.A. performedresearch. C.B. generated new genetic tools. C.R., H.H., and M.C. wrote thearticle. M.B. designed the research and wrote the article.

Received June 3, 2013; revised December 1, 2013; accepted January 20,2014; published February 7, 2014.

REFERENCES

Ansari, A., and Hampsey, M. (2005). A role for the CPF 39-endprocessing machinery in RNAP II-dependent gene looping. GenesDev. 19: 2969–2978.

Barski, A., Cuddapah, S., Cui, K., Roh, T.Y., Schones, D.E., Wang,Z., Wei, G., Chepelev, I., and Zhao, K. (2007). High-resolutionprofiling of histone methylations in the human genome. Cell 129:823–837.

Bezhani, S., Winter, C., Hershman, S., Wagner, J.D., Kennedy, J.F.,Kwon, C.S., Pfluger, J., Su, Y., and Wagner, D. (2007). Unique,

12 of 14 The Plant Cell

shared, and redundant roles for the Arabidopsis SWI/SNF chromatinremodeling ATPases BRAHMA and SPLAYED. Plant Cell 19: 403–416.

Bork, P., and Koonin, E.V. (1993). An expanding family of helicaseswithin the ‘DEAD/H’ superfamily. Nucleic Acids Res. 21: 751–752.

Brzeski J., Podstolski, W., Olczak, K., and Jerzmanowski, A. (1999).Identification and analysis of the Arabidopsis thaliana BSH gene,a member of the SNF5 gene family. Nucleic Acids Res. 11: 2393–2399.

Cairns, B.R., Levinson, R.S., Yamamoto, K.R., and Kornberg, R.D.(1996). Essential role of Swp73p in the function of yeast Swi/Snfcomplex. Genes Dev. 10: 2131–2144.

Campi, M., D’Andrea, L., Emiliani, J., and Casati, P. (2012).Participation of chromatin-remodeling proteins in the repair ofultraviolet-B-damaged DNA. Plant Physiol. 158: 981–995.

Cavalli, G., and Misteli, T. (2013). Functional implications of genometopology. Nat. Struct. Mol. Biol. 20: 290–299.

Choi, K., Kim, S., Kim, S.Y., Kim, M., Hyun, Y., Lee, H., Choe, S.,Kim, S.G., Michaels, S., and Lee, I. (2005). SUPPRESSOR OFFRIGIDA3 encodes a nuclear ACTIN-RELATED PROTEIN6 requiredfor floral repression in Arabidopsis. Plant Cell 17: 2647–2660.

Choi, K., Park, C., Lee, J., Oh, M., Noh, B., and Lee, I. (2007).Arabidopsis homologs of components of the SWR1 complex regulateflowering and plant development. Development 134: 1931–1941.

Clapier, C.R., and Cairns, B.R. (2009). The biology of chromatinremodeling complexes. Annu. Rev. Biochem. 78: 273–304.

Crane, Y.M., and Gelvin, S.B. (2007). RNAi-mediated gene silencingreveals involvement of Arabidopsis chromatin-related genes inAgrobacterium-mediated root transformation. Proc. Natl. Acad. Sci.USA 104: 15156–15161.

Crevillén, P., and Dean, C. (2011). Regulation of the floral repressorgene FLC: The complexity of transcription in a chromatin context.Curr. Opin. Plant Biol. 14: 38–44.

Crevillén, P., Sonmez, C., Wu, Z., and Dean, C. (2013). A gene loopcontaining the floral repressor FLC is disrupted in the early phase ofvernalization. EMBO J. 32: 140–148.

Deal, R.B., Kandasamy, M.K., McKinney, E.C., and Meagher, R.B.(2005). The nuclear actin-related protein ARP6 is a pleiotropicdevelopmental regulator required for the maintenance of FLOWERINGLOCUS C expression and repression of flowering in Arabidopsis. PlantCell 17: 2633–2646.

Deal, R.B., Topp, C.N., McKinney, E.C., and Meagher, R.B. (2007).Repression of flowering in Arabidopsis requires activation of FLOWERINGLOCUS C expression by the histone variant H2A.Z. Plant Cell 19: 74–83.

Diesinger, P.M., Kunkel, S., Langowski, J., and Heermann, D.W.(2010). Histone depletion facilitates chromatin loops on the kilobasepairscale. Biophys. J. 99: 2995–3001.

Doyle, M.R., and Amasino, R.M. (2009). A single amino acid changein the Enhancer of Zeste ortholog CURLY LEAF results invernalization-independent, rapid flowering in Arabidopsis. PlantPhysiol. 151: 1688–1697.

Eisen, J.A., Sweder, K.S., and Hanawalt, P.C. (1995). Evolution ofthe SNF2 family of proteins: Subfamilies with distinct sequencesand functions. Nucleic Acids Res. 23: 2715–2723.

Fan, J.Y., Gordon, F., Luger, K., Hansen, J.C., and Tremethick, D.J.(2002). The essential histone variant H2A.Z regulates the equilibriumbetween different chromatin conformational states. Nat. Struct.Biol. 3: 172–176.

Farrona, S., Hurtado, L., Bowman, J.L., and Reyes, J.C. (2004). TheArabidopsis thaliana SNF2 homolog AtBRM controls shoot developmentand flowering. Development 131: 4965–4975.

Farrona, S., Hurtado, L., March-Díaz, R., Schmitz, R.J., Florencio,F.J., Turck, F., Amasino, R.M., and Reyes, J.C. (2011). Brahma isrequired for proper expression of the floral repressor FLC inArabidopsis. PLoS ONE 6: e17997.

Fornara, F., de Montaigu, A., and Coupland, G. (2010). SnapShot:Control of flowering in Arabidopsis. Cell 141: 550.

Gendrel, A.-V., Lippman, Z., Martienssen, R., and Colot, V. (2005).Profiling histone modification patterns in plants using genomic tilingmicroarrays. Nat. Methods 2: 213–218.

Guyomarc’h, S., Benhamed, M., Lemonnier, G., Renou, J.P., Zhou,D.X., and Delarue, M. (2006). MGOUN3: Evidence for chromatin-mediated regulation of FLC expression. J. Exp. Bot. 57: 2111–2119.

Hargreaves, D.C., and Crabtree, G.R. (2011). ATP-dependentchromatin remodeling: Genetics, genomics and mechanisms. CellRes. 21: 396–420.

He, Y. (2009). Control of the transition to flowering by chromatinmodifications. Mol. Plant 2: 554–564.

Hepworth, S.R., Valverde, F., Ravenscroft, D., Mouradov, A., andCoupland, G. (2002). Antagonistic regulation of flowering-time geneSOC1 by CONSTANS and FLC via separate promoter motifs. EMBOJ. 21: 4327–4337.

Hou, C., and Corces, V.G. (2012). Throwing transcription for a loop:Expression of the genome in the 3D nucleus. Chromosoma 121: 107–116.

Hurtado, L., Farrona, S., and Reyes, J.C. (2006). The putative SWI/SNF complex subunit BRAHMA activates flower homeotic genes inArabidopsis thaliana. Plant Mol. Biol. 62: 291–304.

Jégu, T., et al. (2013). Multiple functions of Kip-related protein5connect endoreduplication and cell elongation. Plant Physiol. 161:1694–1705.

Jerzmanowski, A. (2007). SWI/SNF chromatin remodeling and linkerhistones in plants. Biochim. Biophys. Acta 1769: 330–345.

Kandasamy, M.K., Deal, R.B., McKinney, E.C., and Meagher, R.B.(2005). Silencing the nuclear actin-related protein AtARP4 inArabidopsis has multiple effects on plant development, includingearly flowering and delayed floral senescence. Plant J. 41: 845–858.

Kardailsky, I., Shukla, V.K., Ahn, J.H., Dagenais, N., Christensen, S.K.,Nguyen, J.T., Chory, J., Harrison, M.J., and Weigel, D. (1999).Activation tagging of the floral inducer FT. Science 286: 1962–1965.

Knizewski, L., Ginalski, K., and Jerzmanowski, A. (2008). Snf2proteins in plants: Gene silencing and beyond. Trends Plant Sci. 10:557–565.

Knoepfler, P.S., and Eisenman, R.N. (1999). Sin meets NuRD andother tails of repression. Cell 99: 447–450.

Kobayashi, Y., Kaya, H., Goto, K., Iwabuchi, M., and Araki, T.(1999). A pair of related genes with antagonistic roles in mediatingflowering signals. Science 286: 1960–1962.

Koornneef, M., Hanhart, C.J., and van der Veen, J.H. (1991). Agenetic and physiological analysis of late flowering mutants inArabidopsis thaliana. Mol. Gen. Genet. 229: 57–66.

Kumar, S.V., and Wigge, P.A. (2010). H2A.Z-containing nucleosomesmediate the thermosensory response in Arabidopsis. Cell 140: 136–147.

Lai, A.Y., and Wade, P.A. (2011). Cancer biology and NuRD: A multifacetedchromatin remodelling complex. Nat. Rev. Cancer 11: 588–596.

Lázaro, A., Gómez-Zambrano, A., López-González, L., Piñeiro, M.,and Jarillo, J.A. (2008). Mutations in the Arabidopsis SWC6 gene,encoding a component of the SWR1 chromatin remodellingcomplex, accelerate flowering time and alter leaf and flowerdevelopment. J. Exp. Bot. 59: 653–666.

March-Díaz, R., García-Domínguez, M., Florencio, F.J., and Reyes,J.C. (2007). SEF, a new protein required for flowering repression inArabidopsis, interacts with PIE1 and ARP6. Plant Physiol. 143: 893–901.

March-Díaz, R., and Reyes, J.C. (2009). The beauty of beinga variant: H2A.Z and the SWR1 complex in plants. Mol. Plant 4:565–677.

Marques, M., Laflamme, L., Gervais, A.L., and Gaudreau, L. (2010).Reconciling the positive and negative roles of histone H2A.Z in genetranscription. Epigenetics 5: 267–272.

BAF60 Controls Gene Loop Formation 13 of 14

Martin-Trillo, M., Lázaro, A., Poethig, R.S., Gómez-Mena, C., Piñeiro, M.A.,Martinez-Zapater, J.M., and Jarillo, J.A. (2006). EARLY IN SHORTDAYS 1 (ESD1) encodes ACTIN-RELATED PROTEIN 6 (AtARP6), a putativecomponent of chromatin remodelling complexes that positively regulatesFLC accumulation in Arabidopsis. Development 133: 1241–1252.

Mavrich, T.N., et al. (2008). Nucleosome organization in theDrosophila genome. Nature 453: 358–362.

Noh, Y.-S., and Amasino, R.M. (2003). PIE1, an ISWI family gene, isrequired for FLC activation and floral repression in Arabidopsis.Plant Cell 15: 1671–1682.

Onouchi, H., Igeño, M.I., Périlleux, C., Graves, K., and Coupland, G.(2000). Mutagenesis of plants overexpressing CONSTANS demonstratesnovel interactions among Arabidopsis flowering-time genes. Plant Cell12: 885–900.

Parcy, F. (2005). Flowering: A time for integration. Int. J. Dev. Biol. 49:585–593.

Putterill, J., Robson, F., Lee, K., Simon, R., and Coupland, G. (1995). TheCONSTANS gene of Arabidopsis promotes flowering and encodes a proteinshowing similarities to zinc finger transcription factors. Cell 80: 847–857.

Samach, A., Onouchi, H., Gold, S.E., Ditta, G.S., Schwarz-Sommer,Z., Yanofsky, M.F., and Coupland, G. (2000). Distinct roles ofCONSTANS target genes in reproductive development of Arabidopsis.Science 288: 1613–1616.

Santisteban, M.S., Hang, M., and Smith, M.M. (2011). Histonevariant H2A.Z and RNA polymerase II transcription elongation. Mol.Cell. Biol. 31: 1848–1860.

Sarnowski, T.J., Ríos, G., Jásik, J., Swiezewski, S., Kaczanowski, S.,Li, Y., Kwiatkowska, A., Pawlikowska, K., Koźbiał, M., Koźbiał, P.,Koncz, C., and Jerzmanowski, A. (2005). SWI3 subunits of putativeSWI/SNF chromatin-remodeling complexes play distinct roles duringArabidopsis development. Plant Cell 17: 2454–2472.

Shivaswamy, S., Bhinge, A., Zhao, Y., Jones, S., Hirst, M., and Iyer, V.R.(2008). Dynamic remodeling of individual nucleosomes across a eukaryoticgenome in response to transcriptional perturbation. PLoS Biol. 6: e65.

Simpson, G.G., and Dean, C. (2002). Arabidopsis, the Rosetta stoneof flowering time? Science 296: 285–289.

Singh, R.K., Kabbaj, M.H., Paik, J., and Gunjan, A. (2009). Histonelevels are regulated by phosphorylation and ubiquitylation-dependent proteolysis. Nat. Cell Biol. 11: 925–933.

Srikanth, A., and Schmid, M. (2011). Regulation of flowering time: Allroads lead to Rome. Cell. Mol. Life Sci. 68: 2013–2037.

Tang, L., Nogales, E., and Ciferri, C. (2010). Structure and functionof SWI/SNF chromatin remodeling complexes and mechanisticimplications for transcription. Prog. Biophys. Mol. Biol. 102: 122–128.

Tiwari, V.K., McGarvey, K.M., Licchesi, J.D., Ohm, J.E., Herman,J.G., Schübeler, D., and Baylin, S.B. (2008). PcG proteins, DNAmethylation, and gene repression by chromatin looping. PLoS Biol.6: 2911–2927.

Tolhuis, B., Palstra, R.J., Splinter, E., Grosveld, F., and de Laat, W.(2002). Looping and interaction between hypersensitive sites in theactive beta-globin locus. Mol. Cell 10: 1453–1465.

Valverde, F., Mouradov, A., Soppe, W., Ravenscroft, D., Samach,A., and Coupland, G. (2004). Photoreceptor regulation of CONSTANSprotein in photoperiodic flowering. Science 303: 1003–1006.

Wagner, D., and Meyerowitz, E.M. (2002). SPLAYED, a novel SWI/SNF ATPase homolog, controls reproductive development inArabidopsis. Curr. Biol. 12: 85–94.

Wang, J., Tian, L., Lee, H.S., and Chen, Z.J. (2006). Nonadditiveregulation of FRI and FLC loci mediates flowering-time variation inArabidopsis allopolyploids. Genetics 173: 965–974.

Weissman, B., and Knudsen, K.E. (2009). Hijacking the chromatinremodeling machinery: Impact of SWI/SNF perturbations in cancer.Cancer Res. 69: 8223–8230.

Wilson, R.N., Heckman, J.W., and Somerville, C.R. (1992).Gibberellin is required for flowering in Arabidopsis thaliana undershort days. Plant Physiol. 100: 403–408.

Wu, J.I. (2012). Diverse functions of ATP-dependent chromatinremodeling complexes in development and cancer. Acta Biochim.Biophys. Sin. (Shanghai) 44: 54–69.

Yang, J., and Corces, V.G. (2012). Insulators, long-range interactions,and genome function. Curr. Opin. Genet. Dev. 22: 86–92.

Zhang, Y., Smith, C.L., Saha, A., Grill, S.W., Mihardja, S., Smith,S.B., Cairns, B.R., Peterson, C.L., and Bustamante, C. (2006). DNAtranslocation and loop formation mechanism of chromatin remodelingby SWI/SNF and RSC. Mol. Cell 24: 559–568.

Zilberman, D., Coleman-Derr, D., Ballinger, T., and Henikoff, S.(2008). Histone H2A.Z and DNA methylation are mutually antagonisticchromatin marks. Nature 456: 125–129.

14 of 14 The Plant Cell

DOI 10.1105/tpc.113.114454; originally published online February 7, 2014;Plant Cell

Martin Crespi, Catherine Bergounioux, Cécile Raynaud and Moussa BenhamedTeddy Jégu, David Latrasse, Marianne Delarue, Heribert Hirt, Séverine Domenichini, Federico Ariel,

Arabidopsis in FLOWERING LOCUS CFormation of a Gene Loop at The BAF60 Subunit of the SWI/SNF Chromatin-Remodeling Complex Directly Controls the

 This information is current as of October 29, 2020

 

Supplemental Data /content/suppl/2014/01/27/tpc.113.114454.DC1.html

Permissions https://www.copyright.com/ccc/openurl.do?sid=pd_hw1532298X&issn=1532298X&WT.mc_id=pd_hw1532298X

eTOCs http://www.plantcell.org/cgi/alerts/ctmain

Sign up for eTOCs at:

CiteTrack Alerts http://www.plantcell.org/cgi/alerts/ctmain

Sign up for CiteTrack Alerts at:

Subscription Information http://www.aspb.org/publications/subscriptions.cfm

is available at:Plant Physiology and The Plant CellSubscription Information for

ADVANCING THE SCIENCE OF PLANT BIOLOGY © American Society of Plant Biologists