M

Rm

JKa

b

a

ARRA

KFHFH

I

ns1ttq2tsalgcR

0h

Journal of Plant Physiology 170 (2013) 444– 451

Contents lists available at SciVerse ScienceDirect

Journal of Plant Physiology

jou rna l h om epa ge: www.elsev ier .com/ locate / jp lph

olecular Biology

equirement of histone acetyltransferases HAM1 and HAM2 for epigeneticodification of FLC in regulating flowering in Arabidopsis

un Xiaoa,b,1, Hong Zhanga,b,1, Lijing Xinga, Shujuan Xua,b, Huanhuan Liua,b,ang Chonga, Yunyuan Xua,∗

Key Laboratory of Plant Molecular Physiology, Institute of Botany, Chinese Academy of Sciences, Beijing 100093, ChinaGraduate University of the Chinese Academy of Sciences, Beijing 100049, China

r t i c l e i n f o

rticle history:eceived 11 September 2012eceived in revised form 29 October 2012ccepted 5 November 2012

eywords:loweringistone modificationLCAM1/2

a b s t r a c t

Histone acetylation is an important posttranslational modification associated with gene activation. InArabidopsis, two MYST histone acetyltransferases HAM1 and HAM2 work redundantly to acetylate histoneH4 lysine 5 (H4K5ace) in vitro. The double mutant ham1/ham2 is lethal, which suggests the critical role ofHAM1 and HAM2 in development. Here, we used an artificial microRNA (amiRNA) strategy in Arabidopsisto uncover a novel function of HAM1 and HAM2. The amiRNA-HAM1/2 transgenic plants showed earlyflowering and reduced fertility. In addition, they responded normally to photoperiod, gibberellic acidtreatment, and vernalization. The expression of flowering-repressor FLOWERING LOCUS C (FLC) and itshomologues, MADS-box Affecting Flowering genes 3/4 (MAF3/4), were decreased in amiRNA-HAM1/2lines. HAM1 overexpression caused late flowering and elevated expression of FLC and MAF3/4. Mutationof FLC almost rescued the late flowering with HAM1 overexpression, which suggests that HAM1 regulation

of flowering time depended on FLC. Global H4 acetylation was decreased in amiRNA-HAM1/2 lines, butincreased in HAM1-OE lines, which further confirmed the acetyltransferase activity of HAM1 in vivo.Chromatin immunoprecipitation revealed that H4 hyperacetylation and H4K5ace at FLC and MAF3/4 wereless abundant in amiRNA-HAM1/2 lines than the wild type, but were enriched in HAM1-OE lines. Thus,HAM1 and HAM2 may affect flowering time by epigenetic modification of FLC and MAF3/4 chromatins atH4K5 acetylation.

ntroduction

In eukaryotes, DNA wraps onto a histone octamer to formucleosomes, which are folded through a series of higher-ordertructures to eventually form a chromosome (Kornberg and Lorch,999). The core histones are predominantly globular, except forheir N-terminal “tails”. Distinct types of modifications found onhese tails include acetylation, methylation, phosphorylation, ubi-uitination and ADP-ribosylation (Strahl and Allis, 2000; Cao et al.,008). Combinations of these modifications are thought to consti-ute the so-called “histone code” (Jenuwein and Allis, 2001). Oneuch modification, histone acetylation, has been found to be associ-ted with gene expression. In general, hyperacetylation of histoneysine is considered to be a characteristic of actively transcribed

enes, whereas low-transcription-activity genes are typically asso-iated with hypo-acetylated histones (Sterner and Berger, 2000;ichards and Elgin, 2002).

∗ Corresponding author. Tel.: +86 10 6283 6213; fax: +86 10 8259 4821.E-mail address: xuyy@ibcas.ac.cn (Y. Xu).

1 These authors contribute equally to this article.

176-1617/$ – see front matter © 2012 Elsevier GmbH. All rights reserved.ttp://dx.doi.org/10.1016/j.jplph.2012.11.007

© 2012 Elsevier GmbH. All rights reserved.

The dynamic change of histone acetylation is regulated byantagonistic action of 2 classes of enzymes, histone acetyltrans-ferase (HAT) and histone deacetylase (Shahbazian and Grunstein,2007). Arabidopsis has 12 HATs classified into 4 families on thebasis of sequence similarity and characteristics: Gcn5-related N-acetyltransferase (GNAT), MOZ, Ybt2, Sas2, Tip60- like (MYST),p300/CBP, and TAFII250 families (Sterner and Berger, 2000; Pandeyet al., 2002).

HATs are involved in various biological processes through trans-criptional regulation of numerous genes. For instance, one memberof the GNAT family, HAG1 (also known as AtGCN5), plays a role inregulation of cold tolerance, floral development, embryonic cell-fate patterning, and light responsiveness (Stockinger et al., 2001;Bertrand et al., 2003; Vlachonasios et al., 2003; Benhamed et al.,2006; Long et al., 2006). HAC1, HAC5, and HAC12 of the p300/CBPfamily are involved in regulating flowering time (Han et al., 2006;Deng et al., 2007). Two MYST family members, HAM1 and HAM2,work redundantly to regulate gametophyte development (Latrasse

et al., 2008). However, little is known about their function in flow-ering transition.

Flowering is a major switch from the vegetative to the repro-ductive phase of plant development. In Arabidopsis, this transition

t Phys

ibcoptadSf2tprm2bFmmFfFdDlsa

aglsg

M

P

gMfflwsg2o

Ga

o(HbwIif2wa

J. Xiao et al. / Journal of Plan

s regulated by 5 major pathways: vernalization, photoperiod, gib-erellin, autonomous, and aging (Fornara et al., 2010). Among thisomplex network, FLOWERING LOCUS C (FLC) is a key repressorf flowering and is regulated, both positively and negatively, byosttranslational histone modifications (He, 2009). H3K4, di- andri-methylation, and H3K36, di- and tri-methylation, are associ-ted with actively transcribed FLC. In contrast, H3K9 and H3K27,i- and tri-methylation, are tightly coupled with FLC repression.pecific histone methyltransferases have been proven to respondor the multiple modification sites in FLC chromatin (Berr et al.,011). In addition, histone acetylation plays an essential role inhe regulation of FLC transcription in flowering control. Two com-onents of the autonomous pathway, FLD and FVE, negativelyegulate FLC by participating in the histone deacetylation of chro-atin at FLC loci (He et al., 2003; Ausin et al., 2004; Kim et al.,

004). The vernalization pathway also regulates FLC expressiony changing its chromatin acetylation status (Bastow et al., 2004;innegan et al., 2004; Amasino, 2005). However, unlike the histoneethyltransferase-mediated specific-site methylation in FLC chro-atin, none of the above factors involved in histone acetylation at

LC chromatin is a HAT. So far, only p300/CBP family histone trans-erases HAC1 and 5 were reported to influence flowering throughLC transcription regulation in Arabidopsis. Yet, they are indepen-ent of histone acetylation status at FLC chromatin (Han et al., 2006;eng et al., 2007). Thus, although it is clear that histone acety-

ation/deacetylation regulates FLC expression, the identity of thepecific HAT directly responsible for the histone acetylation statust FLC chromatin is still unknown.

Here, we report that the MYST family members HATs, HAM1nd HAM2, also affect flowering time, besides their function inametogenesis. Our results indicated that HAM1 and HAM2 regu-ate flowering time through epigenetic modification of the H4K5acetatus within chromatins of FLC and MADS-box Affecting Floweringenes 3/4 (MAF3/4).

aterials and methods

lant material and treatments

Arabidopsis thaliana ecotype Col and C24 were used as theenetic backgrounds of mutants and transgenic lines, respectively.utants of ham1 (SALK 147727) and ham2 (SALK 106406) were

rom the SALK T-DNA collection (Alonso et al., 2003). The mutantc-20 was described previously (Helliwell et al., 2002). Plantsere grown at 22 ◦C under long-day (LD, 16-h light/8-h dark) or

hort-day (SD, 8-h light/16-h dark) conditions. Vernalization andibberellic acid (GA) treatments were as described (Bastow et al.,004; Lim et al., 2004). Flowering time was measured as numberf rosette leaves at bolting.

eneration of artificial microRNA of HAM1/2 (amiRNA-HAM1/2)nd HAM1-overexpression (HAM1-OE) transgenic plants

As described (Schwab et al., 2006), we obtained 4ligo nucleotides (I–IV) from the amiRNA designer WMD3http://wmd3.weigelworld.org/cgi-bin/webapp.cgi) to engineerAM1 and HAM2 amiRNA into the endogenous miR319a precursory site-directed mutagenesis. The amiRNA-containing precursoras generated by overlapping PCR with the specific primers

to IV and the common primers A and B. The sequences aren Supplemental Table S1. Then the precursor-containing DNA

ragment was cloned into the binary vector pSN1301 (Li et al.,006) with the enzyme sites of EcoRI and BamHI. The constructas transformed into Agrobacterium tumefaciens strain GV3101,

nd the resulting bacteria were used to transform Arabidopsis

iology 170 (2013) 444– 451 445

ecotype Col by the floral dip method (Clough and Bent, 1998). Thetransformants were selected on 1/2 MS with 20 �g/ml hygromycin.The full-length HAM1 coding sequence was cloned into the binaryvector pSN1301 with the enzyme sites of XbaI and KpnI. Theconstruct was transformed into A. tumefaciens strain GV3101 forlater transformation into the wild-type Col and the flc-20 mutantto generate HAM1-OE and HAM1-OE/flc-20 transgenic plants.Transformant selection was on 1/2 MS with 20 �g/ml hygromycin.

Identification of T-DNA insertion mutants

DNA was isolated from the leaves of ham1, ham2 and their dou-ble mutant. T-DNA border primer LB: 5′-ATT TTG CCG ATT TCGGAAC-3′ was used to amplify DNA from mutants obtained fromthe SALK collection. The gene-specific primers SALK-147727-LPand SALK-147727-RP were used for identifying ham1 and SALK-106046-LP and SALK-106046-RP for ham2. Primer sequences are inSupplemental Table S1.

Real-time RT-PCR

RNA samples were isolated from 2-week-old plants by use ofTRIzol reagent (Invitrogen). cDNA was synthesized from total RNAby use of M-MLV Reverse Transcriptase (Promega) and oligo (dT).Quantitative real-time PCR was performed on Mx3000p (Agilent)by using SYBR Green reagent (Toyobo). Primer sequences used foramplification are listed in Supplemental Table S1. Data were ana-lyzed by the threshold cycle (Ct) method. Specific gene expressionwas normalized to the level of the internal control ubiquitin (UBQ;Mockler et al., 2004) by the formula 2−�Ct, where �Ct is the Ct ofthe target gene subtracted from the Ct of the UBQ gene.

Histone protein extraction and protein gel blot analysis

Histone-rich proteins were extracted from the aerial part ofseedlings at 14 days after germination as described (Bowler et al.,2004). Proteinase inhibitors were added, and histone-rich proteinswere isolated by 15% SDS-PAGE. Then, proteins were transferred topolyvinylidene fluoride (BioTraceTM, USA) membrane, which wasblocked with 5% non-fat milk powder in TBST and then probedwith anti-H4ace or H4K5ace (1:10,000; Millipore, anti-H4ace: 06-598; anti-H4K5ace: 07-327). The blots were developed by use ofhorseradish peroxidase-conjugated goat anti-rabbit IgG secondaryantibody (Zhongshan Golden Bridge Biotechnology, China) and anECL protein gel blot detection system (Amersham, Sweden). Anti-H3 (1:10,000; Millipore, anti-H3: 07-690) was a loading control.

Chromatin immunoprecipitation (ChIP) assay

ChIP experiments were performed as described (Bowler et al.,2004) with 14-days-old seedlings. Antibodies with specificity to H4ace (K5K8K12K16) or H4K5ace were used to immunoprecipitatethe chromatin. The amount of immunoprecipitated FLC chromatinwas determined by real-time PCR on different regions of FLC locusas reported (Liu et al., 2007). For the amount of immunoprecip-itated MAFs chromatin, gene-specific primers were designed atthe promoter-exon1 and -intron1 region. Primer sequences are inSupplemental Table S1.

Results

Knockdown of HAM1 and HAM2 resulted in early flowering

Arabidopsis has 2 members of MYST family HATs: HAM1(At5G64610) and HAM2 (At5G09740) (Earley et al., 2007). HAM1and HAM2 are highly similar in DNA and deduced amino acid

446 J. Xiao et al. / Journal of Plant Physiology 170 (2013) 444– 451

Fig. 1. Knockdown of MYST histone acetyltransferases, HAM1 and HAM2, leads to earlier flowering time. (A) Schematic diagram of the position of artificial microRNA (amiRNA)generated against HAM1 and HAM2 genes. (B) Quantitative PCR analysis of mRNA level of HAM1 and HAM2 in wild-type Col and amiRNA-HAM1/2 transgenic lines R1 andR2. Data are mean ± SD of 3 replicates. (C) amiRNA-HAM1/2 transgenic lines flowered earlier than the wild-type Col under different conditions. LD: long day; SD: short day;S ation

(

stisBtHTrkdo5fmepCR(Ca

a

D + GA: short day with gibberellic acid (GA); SD + V30: SD with 30 days of vernalizmean ± SD of 15 plants for each line). *p < 0.05, **p < 0.01 by Student’s t test.

equences (Fig. S1). Both share a similar expression pattern:hey are constitutively expressed in all tissues, predominantlyn shoot apical meristem (Fig. S2) (Winter et al., 2007), whichuggests their potential function in flowering time regulation.ecause ham1/ham2 is lethal (Latrasse et al., 2008), we generatedransgenic plants with knockdown expression of both HAM1 andAM2 by an artificial microRNA strategy (Schwab et al., 2006).he 21mer amiRNA was designed at the specific and conservedegion (Fig. 1A), 282–303 nt of the HAM1 coding sequence, tonock down both HAM1 and HAM2 endogenous expression asescribed (http://wmd3.weigelworld.org/cgi-bin/webapp.cgi). Webtained several amiRNA-HAM1/2 transgenic lines with more than0% reduction of both HAM1 and HAM2 transcripts and chose 2 forurther analysis (Fig. 1B). These amiRNA-HAM1/2 lines grew nor-

ally as the wild-type Col in the vegetative stage, but they floweredarlier than Col under both LD and SD conditions (Fig. 1C). For exam-le, under LD conditions, rosette leaf numbers were 12.2 ± 0.78 forol, but 9.8 ± 0.85 and 9.7 ± 1.02 for amiRNA-HAM1/2 lines R1 and2, respectively (Fig. 1D). Regardless of vernalization or gibberellinGA3) treatment, amiRNA-HAM1/2 lines still flowered earlier than

ol, although flowering was accelerated for both transgenic linesnd Col (Fig. 1C and D).

Furthermore, we isolated homozygous ham1 (SALK 147727)nd ham2 (SALK 106406) from the SALK collection (Fig. 2A) and

treatment. (D) Rosette leaf number when bolting in Col and amiRNA-HAM1/2 lines

performed a genetic cross. The double heterozygous HAM1/ham1HAM2/ham2 was isolated from F2 progeny plants by PCR genotyp-ing (Fig. 2B) and the expression of HAM1 and HAM2 was analyzedby real-time PCR to ensure knockdown or knockout of HAM1 andHAM2 in mutants (Fig. 2C). We re-checked the flowering timeunder LD condition and found that the double mutant HAM1/ham1HAM2/ham2 flowered earlier than Col and single ham1 or ham2. Therosette leaf number of HAM1/ham1 HAM2/ham2 was 10.2 ± 0.72,about 2 leaves less than Col and ham1 or ham2 (Fig. 2D). Some plantsof amiRNA-HAM1/2 lines showed reduced fertility (Fig. S3), whichwas similar with HAM1/ham1 ham2/ham2 mutant (Latrasse et al.,2008). Therefore, knockdown of HAM1 and HAM2 by amiRNA orT-DNA insertion resulted in earlier flowering.

Overexpression of HAM1 caused delayed flowering

Knockdown of both HAM1 and HAM2 promoted flowering,which suggests their functional redundancy in regulating flower-ing time. Therefore, we wondered whether overexpression of eitherone could result in the opposite phenotype. HAM1-overexpressing

transgenic plants (HAM1-OE) were generated by constitutionalexpression of HAM1, driven by the 35S promoter. We chose 2 trans-genic lines with greater than four-fold enhanced expression ofHAM1 for analysis of flowering time (Fig. 3A and B). The rosette

J. Xiao et al. / Journal of Plant Physiology 170 (2013) 444– 451 447

Fig. 2. Heterozygous ham1/ham2 double mutant shows earlier flowering time. (A) Schematic diagram of the T-DNA insertion site of ham1 (SALK 147727) and ham2( m2 doo ) RoseS

ltrdwfldr

Aa

bt

F(b

SALK 106406). (B) Genome PCR to identify ham1, ham2 and heterozygous ham1/haf mRNA level of Col, ham1, ham2 and heterozygous ham1/ham2 double mutant. (Dtudent’s t test.

eaf number was 12.0 ± 0.75 for Col when bolting under LD condi-ions, but 15.1 ± 0.82 and 14.2 ± 0.91 for HAM1-OE1 and HAM1-OE2,espectively (Fig. 3C). The delayed time was dose dependent. Theelayed-flowering phenotype remained under SD conditions orith GA3 treatment. However, vernalization rescued the delayedowering (Fig. 3B and C). Thus, overexpression of HAM1 led toelayed flowering time, which suggests that HAM1 is a negativeegulator of flowering.

lteration of expression of flowering repressor FLC and MAFs inmiRNA-HAM1/2 and HAM1-OE lines

Genes involved in chromatin modification influence floweringy regulating the expression of several key transcription fac-ors, such as suppressor of overexpression of cytochrome oxidase

ig. 3. Overexpression of HAM1 caused late flowering. (A) Quantitative PCR analysis of mRB) Flowering of HAM1-OE transgenic lines and Col. (C) Rosette leaf number when bolting

y Student’s t test.

uble mutant. The primers’ location is indicated in (A). (C) Quantitative PCR analysistte leaf number of when bolting (mean ± SD of 15 plants for each line). *p < 0.05 by

1 (SOC1), FLC, and its homolog MAFs (Farrona et al., 2008).To investigate the molecular basis of flowering regulation byHAM1 and HAM2, we examined the expression of these factors.The expression of FLC and its homologues MAF3/4 was lower inamiRNA-HAM1/2 lines than in Col (Fig. 4A), with greater reduc-tion for FLC than MAF3/4 (>60% vs. 30–40%). In contrast, the levelsof MAF1, MAF2, and MAF5 did not differ from that of Col. Asexpected, the transcriptional level of SOC1, the direct target geneof FLC, increased more than two-fold in amiRNA-HAM1/2 lines thanCol.

Coincidently, the expression pattern of FLC, SOC1, and MAFs pre-

sented an opposite change manner in the HAM1-OE lines (Fig. 4B).The expression of FLC was three-fold higher in HAM1-OE lines thanCol. The level of SOC1 transcripts was reduced more than 50% ineach HAM1-OE line. The transcriptional levels of FLC homologues

NA level of HAM1 in HAM1-OE lines OE1 and OE2. Data are mean ± SD of 3 replicates.in Col and HAM1-OE lines (mean ± SD of 15 plants for each line). *p < 0.05, **p < 0.01

448 J. Xiao et al. / Journal of Plant Physiology 170 (2013) 444– 451

Fig. 4. HAM1 and HAM2 influenced flowering mostly via FLOWERING LOCUS CONTROL (FLC) transcription. (A and B) Quantitative PCR analysis of mRNA level of floweringm MAD( (C) Mb E/flc-2

Mtf

aohsecOnHT

Hl

tlclisl

is2tCiHalba

arker genes FLC, suppressor of overexpression of cytochrome oxidase 1 (SOC1) andA) and HAM1-OE lines (B) compared with Col. Data are mean ± SD of 3 replicates.olting. (D) Rosette leaf number when bolting in Col, HAM1-OE, flc-20 and HAM1-O

AF3/4 were increased by more than two-fold as well. In contrast,he transcriptional levels of MAF1, MAF2, and MAF5 did not differrom that of Col.

FLC was a key repressor in the flowering regulation network,nd because the change in its mRNA level was greater than thatf other MAFs in amiRNA-HAM1/2 and HAM1-OE lines, HAM1 mayave influenced flowering mainly through FLC. The flc-20 mutanthows extremely early flowering in the C24 ecotype (Helliwellt al., 2002). We generated HAM1-OE/flc-20 transgenic plants byonstitutively expressing HAM1 in the background of flc-20. HAM1-E/flc-20 plants flowered similarly to that of flc-20 (rosette leafumber, 5.6 ± 0.72 vs. 4.2 ± 0.51), but significantly earlier thanAM1-OE lines or Col (14.2 ± 0.91 and 11.5 ± 0.75) (Fig. 4C and D).hus, HAM1 was functionally dependent on FLC for flowering.

istone acetylation changes in amiRNA-HAM1/2 and HAM1-OEines

HAM1 and HAM2 may influence flowering by changing the his-one modification status. Western blot analysis in amiRNA-HAM1/2ines showed a significant decrease in H4K5 acetylation (Fig. 5A). Inontrast, H4K5 acetylation was significantly increased in HAM1-OEines. Furthermore, both changes were dose dependent. Intrigu-ngly, hyperacetylation of H4 (K5, K8, K12, K16 acetylation) wasimilar to that of H4K5ace in amiRNA-HAM1/2 lines and HAM1-OEines (Fig. 5A).

Numerous evidence showed that histone methyltransferasesnfluence flowering by directly regulating the histone methylationtatus at the chromatin of flowering repressor FLC (Wang et al.,007; Pien et al., 2008; Xu et al., 2008; He, 2009). Did histone acetyl-ransferase HAM1 and HAM2 regulate H4 acetylation status on FLC?hIP assay revealed that H4K5ace level and H4 hyperacetylation

n regions A, B, and C of FLC were substantially lower in amiRNA-AM1/2 lines than Col, with little change in regions D and E (Fig. 5B

nd C). While in the HAM1-OE lines, H4K5ace and H4 hyperacety-ation levels were significantly increased in regions A, B, and C,ut still no obvious change in regions D and E (Fig. 5D). Histonecetylation is believed to be enriched around the transcription start

S-box Affecting Flowering genes 1–5 (MAF1-5) in amiRNA-HAM1/2 transgenic linesorphological phenotype of Col, HAM1-OE, flc-20 and HAM1-OE/flc-20 plants when0 plants when bolting. Data are mean ± SD of 15 plants for each line.

site to regulate gene expression, so the histone acetylation changesin regions A and B of FLC might contribute to the transcriptionalchange of FLC in amiRNA-HAM1/2 and HAM1-OE lines.

The mRNA levels of MAF3/4 and FLC were similar in amiRNA-HAM1/2 and HAM1-OE lines, with no difference from Col levelsfor MAF1, MAF2, and MAF5 (Fig. 4A and B). We further analyzedthe histone acetylation status of MAFs chromatin. We used ChIPassay followed by real-time PCR to test the enrichment of H4hyperacetylation and H4K5ace in the promoter-exon1 and intron1regions in MAFs (Fig. 6A). H4 hyperacetylation and H4K5ace inthe promoter-exon1 and intron1 regions of MAF3/4 were lower inamiRNA-HAM1/2 lines, but higher in HAM1-OE lines than in Col,with no significant difference in promoter regions in MAF1, MAF2,and MAF5 in amiRNA-HAM1/2 or HAM1-OE lines (Fig. 6B and C).HAM1 and HAM2 may have influenced the H4 acetylation statusof MAF3/4, as well as FLC, especially in the transcription initiationregions. Therefore, HAM1 and HAM2 might affect flowering timeby influencing the transcription of flowering regulators FLC andMAF3/4 through H4 acetylation of their chromatin structure.

In short, knockdown of acetyltransferases HAM1 and HAM2,either by artificial microRNA strategy (amiRNA-HAM1/2) or T-DNAinsertion (heterozygous mutant HAM1/ham1-HAM2/ham2), mayresult in reduction of H4K5 acetylation and H4 hyperacetylation inFLC chromatin. It caused a decrease of FLC transcription, and thusshowed earlier flowering phenotype. Meanwhile, HAM1 overex-pression lead to controversial effects.

Discussion

HAM1 and HAM2 work redundantly to regulate floweringtransition in Arabidopsis

The MYST family of HAT is highly conserved in eukaryotes. Themembers play critical roles in gene-specific transcription regula-

tion, DNA repair, and replication (Thomas and Voss, 2007; Pillus,2008). Participation in such basic nuclear functions suggests thatalteration of these MYST HATs might lead to multiple-growthdefects.

J. Xiao et al. / Journal of Plant Physiology 170 (2013) 444– 451 449

Fig. 5. H4 acetylation changed in amiRNA-HAM1/2 and HAM1-OE lines at the global level and around the FLC loci. (A) Western blot analysis of H4 hyperacetylation (H4ace)and global level change of acetylated histone H4 lysine 5 (H4K5ace) in amiRNA-HAM1/2 and HAM1-OE lines compared with Col. H3 was used as loading control. (B) Schematicd Relativi me PCa

ttt2opotCge(tbnfl

iagram of the gene structure of FLC and regions checked by ChIP assay. (C and D)

n amiRNA-HAM1/2 (C) and HAM1-OE lines (D) compared with Col. ChIP then real-tind H4K5ace antibodies. Data are mean ± SD of 3 replicates.

High sequence similarity between HAM1 and HAM2 suggestedheir functional redundancy (Fig. S1). Consistently, mutants of nei-her HAM1 nor HAM2 showed obvious developmental defects inhe whole life cycle in our study and other studies (Latrasse et al.,008). However, homozygous ham1/2 were not viable becausef severe defects in the formation of male and female gameto-hytes (Latrasse et al., 2008), which indicates the critical rolef HAM1 and HAM2 in development. However, it also maskedhe function of HAM1 and HAM2 in other developmental stages.onstitutive expression patterns suggested their function besidesametogenesis. High expression at the shoot apical meristem,specially dynamic changes during the floral transition processFig. S2), strongly suggests that they may influence flowering

ime. To address this possibility, we generated transgenic plantsy use of amiRNA. These amiRNA-HAM1/2 transgenic lines grewormally or at least appeared normal at seedling stage, but theirowering transition was accelerated. The flowering phenotype of

e abundance of H4K5ace and H4 hyperacetylation at different regions of FLC geneR of the relative abundance of immunoprecipitated DNA with H4 hyperacetylation

amiRNA-HAM1/2 transgenic lines responded normally to thephotoperiod, GA, and vernalization treatment (Fig. 1). We re-observed the phenotype of ham1/ham2 heterozygous doublemutants and found it also showed early flowering under LD con-ditions. This data further supported that altered HAM1 and HAM2expression influenced flowering transition. Meanwhile, some indi-vidual amiRNA-HAM1/2 plants showed reduced fertility, similar toham1/ham2 heterozygous double mutants (Latrasse et al., 2008).Thus, the amiRNA strategy consequently affected HAM1 and HAM2function. Furthermore, we generated HAM1-OE transgenic plants.Coincident with the amiRNA-HAM1/2 lines, HAM1-OE lines resultedin late flowering phenotype.

These results suggested that knockdown of both HAM1 and

HAM2 promoted flowering besides altering gametogenesis devel-opment, and HAM1 and HAM2 worked redundantly. Here, weuncovered the roles of MYST family HAT in multiple plant devel-opmental stages. The regulation of flowering by HATs is general for

450 J. Xiao et al. / Journal of Plant Physiology 170 (2013) 444– 451

Fig. 6. H4 acetylation changed in amiRNA-HAM1/2 and HAM1-OE lines at MAF1-5 loci. (A) Schematic diagram of the gene structure of representative MAF genes and ther on ana e PCRa

ta

Td

gaeivtmtHi2eligsliswotHo

egions checked by ChIP assay. (B and C) Relative abundance of H4 hyperacetylatimiRNA-HAM1/2 (B) and HAM1-OE lines (C) compared with Col. ChIP then real-timnd H4K5ace antibodies. Data are mean ± SD of 3 replicates.

he MYST family and p300/CBP family HATs such as HAC1, HAC5,nd HAC12 in Arabidopsis (Han et al., 2006; Deng et al., 2007).

he influence of HAM1 and HAM2 on flowering time is mainlyependent on FLC

Flowering is precisely regulated by a complex network inte-rating the multiple inputs from outside inputs such as day lengthnd temperature, as well as inside developmental signals (Fornarat al., 2010). A class of MADS box transcription factors, FLC, andts homologues MAFs are repressors in inhibiting flowering in theegetative development stage. Their expression is programmedo decrease during flowering induction and is related to histone

odification changes in chromatin (Farrona et al., 2008). Methyla-ion, acetylation, and ubiquitination all influence their expression.istone methyltransferase and acetyltransferase influence flower-

ng through epigenetic regulation of FLC transcription (Han et al.,006; Deng et al., 2007; Wang et al., 2007; Pien et al., 2008; Xut al., 2008). Our study of the MADS box genes indicated decreasedevels of FLC and MAF3/4 in amiRNA-HAM1/2 lines, but increasedn HAM1-OE lines. Furthermore, the change in expression wasreater for FLC than MAF3/4. SOC1, the direct target gene of FLC,howed a controversial pattern in amiRNA-HAM1/2 and HAM1-OEines as compared with FLC. Moreover, we overexpressed HAM1n the flc-20 mutant background and found the flowering timelightly delayed in HAM1-OE/flc-20 transgenic lines as comparedith flc-20 lines. These data, together with the flowering phenotype

f amiRNA-HAM1/2 and HAM1-OE plants under LD and SD condi-ions and the response to vernalization, suggested that HAM1 andAM2 influenced flowering mostly by affecting the transcriptionf FLC.

d H4K5ace at different regions of MADS-box Affecting Flowering genes (MAFs) in of the relative abundance of immunoprecipitated DNA with H4 hyperacetylation

HAM1- and HAM2-mediated H4 acetylation influence FLCtranscription and flowering time

Epigenetic modifications, especially histone acetylation, playan essential role in regulating FLC expression (He, 2009). How-ever, although the effect of p300/CBP family members HAC1 andHAC5 on flowering depends on FLC, the histone acetylation sta-tus at FLC chromatin did not change (Han et al., 2006; Deng et al.,2007). HAM1 and HAM2 have HAT activity: they acetylated H4at lysine 5 (H4K5ace) in vitro (Earley et al., 2007). In our results,global level histone acetylation H4K5ace and H4 hyperacetyla-tion (K5K8K12K16ace) level deceased in amiRNA-HAM1/2 lines, butincreased in HAM1-OE lines. Thus, HAM1 and HAM2 may haveacetylated H4 in vivo.

Further ChIP analysis identified that H4 hyperacetylation andH4K5ace around the transcription initiation regions of FLC andMAF3/4 differed in amiRNA-HAM1/2 and HAM1-OE transgenicplants, decreasing in amiRNA-HAM1/2, but increasing in HAM1-OEplants. Histone acetylation is thought to release the compact chro-matin to facilitate access of the RNA polymerase complex with thestart of transcription (Kouzarides, 2007). The differential expres-sion of FLC, MAF3/4 in amiRNA-HAM1/2 and HAM1-OE lines mightbe caused by the changed histone acetylation status around theirtranscription initiation regions. Unlike the previous report of flow-ering regulation by p300/CBP family HATs HAC1 and HAC5 (Hanet al., 2006; Deng et al., 2007), our data suggest that MYST familyHAT members HAM1 and HAM2 directly influenced the chromatin

H4K5 acetylation status of flowering repressor FLC and MAF3/4 tochange their transcription level and induce later flowering. Further,H4K5 acetylation may be a possible histone modification marker forthe transcription activation of FLC in flowering control.

t Phys

A

IS(tgtA

A

fj

R

A

A

A

B

B

B

B

B

C

C

D

E

F

F

F

H

H

J. Xiao et al. / Journal of Plan

cknowledgments

We thank Prof. Elizabeth Dennis (Commonwealth Scientific andndustrial Research Organization, Australia) for flc-20 seeds and theALK Center for ham1 and ham2 seeds. We thank Prof. Detlef WeigelMax Planck Institute for Developmental Biology, Germany) forhe pRS300 vector to construct the amiRNA-HAM1/2 precursor. Weratefully acknowledge funding from the National Science Founda-ion of China (30971624) and the Innovation Grant of the Chinesecademy of Sciences (20100322).

ppendix A. Supplementary data

Supplementary data associated with this article can beound, in the online version, at http://dx.doi.org/10.1016/.jplph.2012.11.007.

eferences

lonso JM, Stepanova AN, Leisse TJ, Kim CJ, Chen H, Shinn P, et al. Genome-wideinsertional mutagenesis of Arabidopsis thaliana. Science 2003;301:653–7.

masino RM. Vernalization and flowering time. Curr Opin Biotechnol2005;16:154–8.

usin I, Alonso-Blanco C, Jarillo JA, Ruiz-Garcia L, Martinez-Zapater JM. Regula-tion of flowering time by FVE, a retinoblastoma-associated protein. Nat Genet2004;36:162–6.

astow R, Mylne JS, Lister C, Lippman Z, Martienssen RA, Dean C. Vernaliza-tion requires epigenetic silencing of FLC by histone methylation. Nature2004;427:164–7.

enhamed M, Bertrand C, Servet C, Zhou DX. Arabidopsis GCN5, HD1, and TAF1/HAF2interact to regulate histone acetylation required for light-responsive geneexpression. Plant Cell 2006;18:2893–903.

err A, Shafiq S, Shen WH. Histone modifications in transcriptional activation duringplant development. Biochim Biophys Acta 2011;1809:567–76.

ertrand C, Bergounioux C, Domenichini S, Delarue M, Zhou DX. Arabidopsis his-tone acetyltransferase AtGCN5 regulates the floral meristem activity throughthe WUSCHEL/AGAMOUS pathway. J Biol Chem 2003;278:28246–51.

owler C, Benvenuto G, Laflamme P, Molino D, Probst AV, Tariq M, et al. Chromatintechniques for plant cells. Plant J 2004;39:776–89.

ao Y, Dai Y, Cui S, Ma L. Histone H2B monoubiquitination in the chromatinof FLOWERING LOCUS C regulates flowering time in Arabidopsis. Plant Cell2008;20:2566–602.

lough SJ, Bent AF. Floral dip: a simplified method for Agrobacterium-mediatedtransformation of Arabidopsis thaliana. Plant J 1998;16:735–43.

eng W, Liu C, Pei Y, Deng X, Niu L, Cao X. Involvement of the histone acetyltrans-ferase AtHAC1 in the regulation of flowering time via repression of FLOWERINGLOCUS C in Arabidopsis. Plant Physiol 2007;143:1660–8.

arley KW, Shook MS, Brower-Toland B, Hicks L, Pikaard CS. In vitro specificitiesof Arabidopsis coactivator histone acetyltransferases: implications for histonehyperacetylation in gene activation. Plant J 2007;52:615–26.

arrona S, Coupland G, Turck F. The impact of chromatin regulation on the floraltransition. Semin Cell Dev Biol 2008;19:560–73.

innegan EJ, Sheldon CC, Jardinaud F, Peacock WJ, Dennis ES. A cluster of Arabidop-sis genes with a coordinate response to an environmental stimulus. Curr Biol2004;14:911–6.

ornara F, de Montaigu A, Coupland G. SnapShot: control of flowering in Arabidopsis.

Cell 2010;141. p. 550 e1–2.

an SK, Song JD, Noh YS, Noh B. Role of plant CBP/p300-like genes in the regulationof flowering time. Plant J 2006;49:103–14.

e Y, Michaels SD, Amasino RM. Regulation of flowering time by histone acetylationin Arabidopsis. Science 2003;302:1751–4.

iology 170 (2013) 444– 451 451

He Y. Control of the transition to flowering by chromatin modifications. Mol Plant2009;2:554–64.

Helliwell CA, Wesley SV, Wielopolska AJ, Waterhouse PM. High-throughput vectorsfor efficient gene silencing in plants. Funct Plant Biol 2002;29:1217–25.

Jenuwein T, Allis C. Translating the histone code. Science 2001;293:1074–80.Kim HJ, Hyun Y, Park JY, Park MJ, Park MK, Kim MD, et al. A genetic link between

cold responses and flowering time through FVE in Arabidopsis thaliana. Nat Genet2004;36:167–71.

Kornberg RD, Lorch Y. Twenty-five years of the nucleosome, fundamental particleof the eukaryote chromosome. Cell 1999;98:285–94.

Kouzarides T. Chromatin modifications and their function. Cell 2007;128:693–705.

Latrasse D, Benhamed M, Henry Y, Domenichini S, Kim W, Zhou DX, et al. TheMYST histone acetyltransferases are essential for gametophyte developmentin Arabidopsis. BMC Plant Biol 2008;8:121.

Li C, Liang Y, Chen C, Li J, Xu Y, Xu Z, et al. Cloning and expression analysis of TSK1,a wheat SKP1 homologue, and functional comparison with Arabidopsis ASK1 inmale meiosis and auxin signalling. Funct Plant Biol 2006;33:381–90.

Lim MH, Kim J, Kim YS, Chung KS, Seo YH, Lee I, et al. A new Arabidopsis gene,FLK, encodes an RNA binding protein with K homology motifs and regulatesflowering time via FLOWERING LOCUS C. Plant Cell 2004;16:731–40.

Liu F, Quesada V, Crevillen P, Baurle I, Swiezewski S, Dean C. The ArabidopsisRNA-binding protein FCA requires a lysine-specific demethylase 1 homolog todownregulate FLC. Mol Cell 2007;28:398–407.

Long JA, Ohno C, Smith ZR, Meyerowitz EM. TOPLESS regulates apical embryonic fatein Arabidopsis. Science 2006;312:1520–3.

Mockler TC, Yu X, Shalitin D, Parikh D, Michael TP, Liou J, et al. Regulation of floweringtime in Arabidopsis by K homology domain proteins. Proc Natl Acad Sci USA2004;101:12759–64.

Pandey R, Muller A, Napoli CA, Selinger DA, Pikaard CS, Richards EJ, et al. Analy-sis of histone acetyltransferase and histone deacetylase families of Arabidopsisthaliana suggests functional diversification of chromatin modification amongmulticellular eukaryotes. Nucl Acids Res 2002;30:5036–55.

Pien S, Fleury D, Mylne JS, Crevillen P, Inze D, Avramova Z, et al. ARABIDOPSISTRITHORAX 1 dynamically regulates FLOWERING LOCUS C activation via histoneH3 lysine-4 trimethylation. Plant Cell 2008;20:580–8.

Pillus L. MYSTs mark chromatin for chromosomal functions. Curr Opin Cell Biol2008;20:326–33.

Richards EJ, Elgin SC. Epigenetic codes for heterochromatin formation and silencing:rounding up the usual suspects. Cell 2002;108:489–500.

Schwab R, Ossowski S, Riester M, Warthmann N, Weigel D. Highly specific genesilencing by artificial microRNAs in Arabidopsis. Plant Cell 2006;18:1121–33.

Shahbazian MD, Grunstein M. Functions of site-specific histone acetylation anddeacetylation. Annu Rev Biochem 2007;76:75–100.

Sterner DE, Berger SL. Acetylation of histones and transcription-related factors.Microbiol Mol Biol Rev 2000;64:435–59.

Stockinger EJ, Mao Y, Regier MK, Triezenberg SJ, Thomashow MF. Transcriptionaladaptor and histone acetyltransferase proteins in Arabidopsis and their inter-actions with CBF1, a transcriptional activator involved in cold-regulated geneexpression. Nucl Acids Res 2001;29:1524–33.

Strahl BD, Allis CD. The language of covalent histone modifications. Nature2000;403:41–5.

Thomas T, Voss AK. The diverse biological roles of MYST histone acetyltransferasefamily proteins. Cell Cycle 2007;6:696–704.

Vlachonasios KE, Thomashow MF, Triezenberg SJ. Disruption mutations of ADA2band GCN5 transcriptional adaptor genes dramatically affect Arabidopsis growth,development, and gene expression. Plant Cell 2003;15:626–38.

Wang X, Zhang Y, Ma Q, Zhang Z, Xue Y, Bao S, et al. SKB1-mediated symmetricdimethylation of histone H4R3 controls flowering time in Arabidopsis. EMBO J2007;26:1934–41.

Winter D, Vinegar B, Nahal H, Ammar R, Wilson GV, Provart NJ. An “Electronic Fluo-rescent Pictograph” browser for exploring and analyzing large-scale biological

data sets. PLoS ONE 2007;2:e718.

Xu L, Zhao Z, Dong A, Soubigou-Taconnat L, Renou JP, Steinmetz A, et al. Di- and tri-but not mono-methylation on histone H3 lysine 36 marks active transcription ofgenes involved in flowering time regulation and other processes in Arabidopsisthaliana. Mol Cell Biol 2008;28:1348–60.