Epigenetic Mutation of RAV6 Affects Leaf Angle and Seedmuch higher expression in Epi-rav6 (+/–) and Epi-rav6 (–/–) leaves (Fig. 2, B and C). Moreover, the degree of up-regulation

Epigenetic Mutation of RAV6 Affects Leaf Angle and SeedSize in Rice1[OPEN]

Xiangqian Zhang2, Jing Sun2, Xiaofeng Cao, and Xianwei Song*

Guangdong Engineering Research Center of Grassland Science, College of Forestry and LandscapeArchitecture, South China Agricultural University, Guangzhou 510642, China (X.Z.); State Key Laboratory ofPlant Genomics and National Center for Plant Gene Research, Institute of Genetics and DevelopmentalBiology, Chinese Academy of Sciences, Beijing 100101, China (J.S., X.C., X.S.); and Collaborative InnovationCenter of Genetics and Development, Shanghai 200433, China (X.C.)

Heritable epigenetic variants of genes, termed epialleles, can broaden genetic and phenotypic diversity in eukaryotes. Epiallelesmay also provide a new source of beneficial traits for crop breeding, but very few epialleles related to agricultural traits havebeen identified in crops. Here, we identified Epi-rav6, a gain-of-function epiallele of rice (Oryza sativa) RELATED TO ABSCISICACID INSENSITIVE3 (ABI3)/VIVIPAROUS1 (VP1) 6 (RAV6), which encodes a B3 DNA-binding domain-containing protein. TheEpi-rav6 plants show larger lamina inclination and smaller grain size; these agronomically important phenotypes are inherited in asemidominant manner. We did not find nucleotide sequence variation of RAV6. Instead, we found hypomethylation in thepromoter region of RAV6, which caused ectopic expression of RAV6 in Epi-rav6 plants. Bisulfite analysis revealed that cytosinemethylation of four CG and two CNG loci within a continuous 96-bp region plays essential roles in regulating RAV6 expression;this region contains a conserved miniature inverted repeat transposable element transposon insertion in cultivated rice genomes.Overexpression of RAV6 in the wild type phenocopied the Epi-rav6 phenotype. The brassinosteroid (BR) receptor BRINSENSITIVE1 and BR biosynthetic genes EBISU DWARF, DWARF11, and BR-DEFICIENT DWARF1were ectopically expressedin Epi-rav6 plants. Also, treatment with a BR biosynthesis inhibitor restored the leaf angle defects of Epi-rav6 plants. Thisindicates that RAV6 affects rice leaf angle by modulating BR homeostasis and demonstrates an essential regulatory role ofepigenetic modification on a key gene controlling important agricultural traits. Thus, our work identifies a unique rice epiallele,which may represent a common phenomenon in complex crop genomes.

Epigenetic gene variants (epialleles) carry heritablechanges in gene expression that do not result from al-terations in the underlying DNA sequence (Kakutani,2002). In eukaryotes, cytosine DNA methylation, aconserved epigenetic mark, plays essential roles in thesilencing of transposable elements (TEs) and genes(Law and Jacobsen, 2010). In higher plants, the fewknown epialleles involve alterations in DNA methyla-tion, indicating that this epigenetic markermakes a large

contribution to epigenetic diversity. The Arabidopsis(Arabidopsis thaliana) clark kent epiallele, which hashypermethylated cytosines at the SUPERMAN locus,causes increased numbers of stamens and carpels(Jacobsen and Meyerowitz, 1997). Also, DNA hypo-methylation at two direct repeats in the promoter re-gion of FLOWERING WAGENINGEN (FWA; Soppeet al., 2000) causes the late-flowering phenotype inArabidopsis plants carrying the fwa epiallele. Plantscarrying the natural epiallele hypermethylated at Linariacycloidea-like show altered floral symmetry, from bilat-eral to radial, in Linaria vulgaris (Cubas et al., 1999). Intomato (Solanum lycopersicum), the Colorless nonripeningphenotype results from hypermethylation at the pro-moter of SQUAMOSA promoter-binding protein like(Manning et al., 2006). In melon (Cucumis melo), thetransition from male to female flowers results fromDNA hypermethylation in the promoter of CmWIP1, asmediated by a transposon insertion in gynoecious va-rieties (Martin et al., 2009).

Work in rice (Oryza sativa) has found only two epi-alleles, Epi-d1 and Epi-df, both of which show a dwarfphenotype (Miura et al., 2009; Zhang et al., 2012). Epi-d1is a spontaneous epiallele that shows ametastable dwarfphenotype caused by DNA hypermethylation in thepromoter region ofDWARF1 (Miura et al., 2009).Epi-df isa gain-of-function epiallele caused by hypomethylation

1 This work was supported by the National Natural Science Foun-dation of China (grant nos. 31300987 to X.S. and 30900884 to X.Z.), theGenetically Modified Breeding Major Projects (grant no.2014ZX08010–002 to X.C.), the Natural Science Foundation of Guang-dong Province, China (grant no. 2014A030313457), a postdoctoralfellowship (to J.S.), and the State Key Laboratory of Plant Genomics.

2 These authors contributed equally to the article.* Address correspondence to [email protected] author responsible for distribution of materials integral to the

findings presented in this article in accordance with the policy de-scribed in the Instructions for Authors (www.plantphysiol.org) is:Xianwei Song ([email protected]).

X.Z. designed and performed experiments; J.S. analyzed data andmodified the paper; X.C. designed experiments and modified thepaper; and X.S. designed and performed experiments and wrotethe article.

[OPEN] Articles can be viewed without a subscription.www.plantphysiol.org/cgi/doi/10.1104/pp.15.00836

2118 Plant Physiology�, November 2015, Vol. 169, pp. 2118–2128, www.plantphysiol.org � 2015 American Society of Plant Biologists. All Rights Reserved.

https://plantphysiol.orgDownloaded on November 5, 2020. - Published by Copyright (c) 2020 American Society of Plant Biologists. All rights reserved.

mailto:[email protected]

http://www.plantphysiol.org

mailto:[email protected]

http://www.plantphysiol.org/cgi/doi/10.1104/pp.15.00836

https://plantphysiol.org

in the 59 region of FERTILIZATION-INDEPENDENTENDOSPERM1 (FIE1). The ectopic expression of FIE1in Epi-df results in dwarf and various floral defects thatare inherited in a dominant manner (Zhang et al., 2012).In Arabidopsis, DNA methylation occurs in three

sequence contexts: CG, CHG, and CHH (where H = A,C, or T) catalyzed by the de novo DNA methyltrans-ferase DOMAINS REARRANGED METHYLASE2(DRM2; Cao and Jacobsen, 2002). In the symmetricalcontexts, DNAMETHYLTRANSFERASE1 (MET1) andCHROMOMETHYLASE3 maintain methylation in theCG and CHG contexts, respectively (Lindroth et al.,2001; Law and Jacobsen, 2010). Small interfering RNAstrigger de novo methylation in all sequence contextsand also trigger the maintenance of CHH methylationvia RNA-directed DNA methylation predominatelymediated by DRM2 (Cao and Jacobsen, 2002; Law andJacobsen, 2010).DNA methylation may be more prevalent and im-

portant in rice than in Arabidopsis, owing to the largenumbers of TEs in the rice genome. In fact, Arabidopsismutants of variousDNAmethyltransferases orDECREASEINDNAMETHYLATION1 show few or no developmentaldefects (Vongs et al., 1993; Lindroth et al., 2001; Cao andJacobsen, 2002; Saze et al., 2003). By contrast, the ricemet1a null mutant shows either viviparous germinationor early embryonic lethality (Hu et al., 2014; Yamauchiet al., 2014). Moreover, impairment of the RNA-directedDNA methylation pathway proteins Dicer-like 3a andDRM2 causes drastic and pleiotropic developmentalphenotypes (Moritoh et al., 2012; Wei et al., 2014).Leaf angle is an important agronomic trait that di-

rectly affects crop architecture and grain yields (Sinclairand Sheehy, 1999). Crops with erect leaves capturemore light for photosynthesis and are suitable for denseplanting, all of which increase yields (Sakamoto et al.,2006). In rice, the brassinosteroid (BR) phytohormonesparticipate in the determination of leaf angle (Tongand Chu, 2012; Zhang et al., 2014). BR-deficient orBR-insensitive mutants display erect leaves, whileoverexpression of BR biosynthesis genes or signalingcomponents results in less erect leaves with large leafinclination (Yamamuro et al., 2000; Hong et al., 2005;Bai et al., 2007). For example, loss-of-function mutantsof rice BR INSENSITIVE1 (OsBRI1) and EBISU DWARF(D2), which encode the rice BR receptor kinase and aBR synthesis enzyme, respectively, show erect leaves(Yamamuro et al., 2000; Hong et al., 2003). In rice,RELATEDTOABSCISICACID INSENSITIVE3 (ABI3)/VIVIPAROUS1 (VP1; RAV)-LIKE1 (RAVL1), a B3DNA-binding domain-containing protein, maintainsBR homeostasis via the coordinated activation of BRI1and BR biosynthetic genes D2, DWARF11 (D11), andBR-DEFICIENT DWARF1 (BRD1; Je et al., 2010). Theravl1mutant and RAVL1 overexpression lines showederect leaves and large leaf angles, respectively (Je et al.,2010).Here, we identified a natural epiallele of rice RAV6

and found that plants carrying this epiallele show increasedleaf angle and small grains, as well as hypomethylation

in the RAV6 promoter region and ectopic expression ofRAV6. Our work revealed that epigenetic modificationof RAV6 plays an essential role in the regulation ofimportant agricultural traits in rice and found thatRAV6 acts via BR homeostasis.

RESULTS

A Semidominant Mutant Shows Large Lamina Inclinationand Small Seed Size

A spontaneously occurring rice mutant with largeleaf angle and small seedswas isolated from the japonicarice ‘Zhonghua11’ and was named Epi-rav6 based onour subsequent characterization (see below). Comparedwith the wild type (cv Zhonghua11), almost all leaf an-gles throughout all developmental stages were larger inEpi-rav6 plants (Fig. 1A). The mutant had almost thesame amount of rachis as the wild type but had obvi-ously smaller grains (Fig. 1, B–D; Supplemental Fig. S1).Compared with the wild type, the Epi-rav6 seeds weremarkedly smaller, by 12.8% in length (average 7.8mm inthewild type to 6.8mm in Epi-rav6; Fig. 1, C andD), 25%in width (average 3.6 mm in the wild type to 2.7 mm inEpi-rav6; Fig. 1, B and D), and 20% in thickness (average2.4 mm in the wild type to 1.9 mm in Epi-rav6; Fig. 1E).Thus, the 1,000-grain weight of Epi-rav6 seeds was only57% of that of the wild type (Fig. 1E). Together, thesefindings indicate that Epi-rav6 influences leaf angle andgrain size.

To determine the inheritance of Epi-rav6, we exam-ined the phenotypes of 260 progeny of self-pollinatedEpi-rav6 plants. Three phenotypes segregated in theseprogeny: 62were like thewild type, 143were like Epi-rav6,and 55 had severe defects, including fewer tillers, verylarge leaf angles, and sterility (Fig. 1A). The ratio of thewildtype to Epi-rav6 like to severe phenotypes was 1:2.3:0.89(x2 = 2.98, P. 0.05). In addition, 27 F1 plants derived fromthe cross between Epi-rav6 and cv Zhonghua11 showednearly a 1:1 ratio of wild-type progeny and Epi-rav6-likeprogeny (Supplemental Fig. S2). This genetic evidencedemonstrated that the mutant is controlled by a single,semidominant locus and the derived allele is hetero-zygous. Thus, we regarded Epi-rav6-like progeny asheterozygotes and refer to them as Epi-rav6 (+/–); wealso regarded the progeny with severe defects as ho-mozygotes and refer to them as Epi-rav6 (–/–).

Cloning and Characterization of RAV6

To explore the molecular mechanism responsible forthe Epi-rav6 phenotype, we used map-based cloning toisolate the causal gene. Using an F2 population derivedfrom a cross between the Epi-rav6mutant and indica varHuajingxian74, we mapped Epi-rav6 to an 18.5-kb re-gion on chromosome 2. Based on the MSU7.0 annota-tion (http://rice.plantbiology.msu.edu/), this regioncontains only the promoter region of a putative gene,Os02g45850 (Fig. 2A). However, we found no nucleo-tide sequence difference between the mutant and the

Plant Physiol. Vol. 169, 2015 2119

Identification of an Epiallele of RAV6 in Rice


http://www.plantphysiol.org/cgi/content/full/pp.15.00836/DC1



wild type in this region. We next investigated the ex-pression level of Os02g45850, the only annotated genein the mapped region. Compared with its low expres-sion level in wild-type leaves, Os02g45850 showedmuch higher expression in Epi-rav6 (+/–) and Epi-rav6(–/–) leaves (Fig. 2, B and C). Moreover, the degree ofup-regulation was much higher in homozygous thanheterozygous lines (Fig. 2, B and C).

To further confirm that the ectopic expression ofOs02g45850 is responsible for the developmental de-fects in Epi-rav6, we overexpressed Os02g45850 in thewild type under the control of the maize (Zea mays)Ubiquitin promoter. All the positive transformants withhigh expression of Os02g45850 displayed large leaf in-clination and small grains (Fig. 3). Thus, we concludedthat the abnormal phenotype in this mutant resultsfrom high expression of Os02g45850. Os02g45850 en-codes a B3 DNA-binding domain-containing protein,which belongs to the RAV family andwas named RAV6(Romanel et al., 2009). Therefore, we named the alleleEpi-rav6.

RAV6 Regulates Leaf Angle via MediatingBR Homeostasis

The plant-specific RAV family belongs to the B3transcription factor superfamily and consists of 12

members in rice, including RAV6 (Romanel et al., 2009).Phylogenetic analysis showed that the RAV6 (encodedby Os02g45850) is more closely related to RAVL1(encoded by Os04g49230) than to other members of theRAV family (Fig. 4A). Comparison of deduced aminoacid sequences suggested that RAV6 exhibits a highdegree of sequence identity with RAVL1, especially inthe B3 DNA-binding domain (Fig. 4B). These observa-tions indicate that RAV6 and RAVL1may share similarfunctions. RAVL1 maintains BR homeostasis via thecoordinated activation of BR synthetic and receptorgenes in rice (Je et al., 2010). In addition, plants over-expressing RAVL1 showed increased lamina inclination,like the Epi-rav6mutants (Je et al., 2010), suggesting thatRAV6 may also be involved in the BR pathway via ac-tivation of BR biosynthetic and signaling genes. Toconfirm this, we measured expression of representativegenes encoding BR biosynthesis enzymes and the BRreceptor, known RAVL1 targets, in Epi-rav6mutants. Asexpected, expression of BR synthetic genes, includingD2, D11, and BRD1, was dramatically up-regulated inthe Epi-rav6 mutant (Fig. 5A). Similarly, we also ob-served higher levels of induction for the BR receptorgene BRI1 in Epi-rav6 mutants. These data suggest thatRAV6 affects the expression of genes involved in BRsignaling and BR biosynthesis.

To further confirm that the large lamina inclination ofrav6 mutant results from BR overdose, we postulated

Figure 1. Characterization of a semidomi-nant rice mutant with larger leaf angle andsmaller seeds. A, The wild-type (WT), het-erozygous Epi-rav6 (+/–), and homozygousEpi-rav6 (2/2) plants at tilling stage. B andC, Comparison of grain and seed width (B)and length (C) between the wild type andEpi-rav6 (+/–). D, Characterization of seedweight, seed length, seed width, and seedthickness in the wild type and the Epi-rav6mutant. Values are means 6 SD. In eachgraph, statistically significant differencesare indicated by double asterisks (**P ,0.01). Bars = 20 cm (A) and 1 cm (B and C).

2120 Plant Physiol. Vol. 169, 2015

Zhang et al.



that treatment with a BR inhibitor would restore, atleast partially, the mutant phenotype. As expected, inthe presence of propiconazole (Pcz), a BR biosynthesisinhibitor (Hartwig et al., 2012), the lamina inclinationof Epi-rav6 mutant was restored (Fig. 5B). Taken to-gether, these results suggest that RAV6 is involved inBR-mediated developmental processes.

Hypomethylation of the Promoter Region of RAV6 inEpi-rav6 Plants

We found no nucleotide sequence difference betweenthe wild type and Epi-rav6, but we did observe altera-tion of RAV6 expression levels in the Epi-rav6 plants,suggesting that the mutation may result from an epi-genetic modification. We therefore investigated theDNA methylation status of the RAV6 locus. The pro-moter region of RAV6 contains many TEs, including anadh5-like miniature inverted repeat transposable ele-ment (MITE), a short interspersed nuclear element(SINE), two unclassified retrotransposons, and a Snabo-like MITE (Fig. 6A; Supplemental Fig. S3A). TEs, espe-cially those proximal to genes, can act as epigeneticmediators to influence nearby gene expression (Hollister

and Gaut, 2009; Hollister et al., 2011;Wei et al., 2014). So,we performed bisulfite sequencing to analyze methyla-tion in a 2,358-bp genomic region consisting of 573 bpof gene body, 1,321 bp of upstream region, and 464 bp of59 distal retrotransposon sequence (Supplemental Fig.S3A). In the wild type and homozygous Epi- rav6 mu-tants, we found no DNA methylation in all three se-quence contexts in the 573-bp gene body of RAV6.However, we observed higher CG and CHG but notCHHDNAmethylation in proximal and distal upstreamregions of RAV6 in the wild type compared with thehomozygous Epi-rav6 mutant. All the changed methyl-ation sites occurred in a contiguous 96-bp region at –600to –504 bp relative to the transcriptional start site, in-cluding the closestMITE (adh5-likeMITE). This region ishypermethylated in the wild type but hypomethylatedin Epi-rav6, containing four CG sites and two CHG sites(Fig. 6B; Supplemental Fig. S3B). To further confirm thatthe ectopic expression of RAV6 in Epi-rav6 results inhypomethylation of the promoter region, we treated thewild-type seeds with 5-aza-29 deoxycytidine (5-aza-dC),an inhibitor of DNA methylation (Chang and Pikaard,2005). The expression levels of RAV6 were measured in7-d-old seedlings with or without 5-aza-dC treatment.Treatmentwith 5-aza-dC up-regulatedRAV6 expressionto varying degrees in three cultivated rice strains, in-cluding two japonica (cv Zhonghua11 and Nipponbare)and one indica (cv Kasalath) accessions (Fig. 6C). Thesethree cultivated accessions also contain the MITE (seebelow); therefore, these results indicate that the DNAmethylation mediated by the TE in the 59 upstream se-quence of RAV6 plays an essential role in the regulationof RAV6 expression.

Figure 2. Molecular cloning and expression analysis of RAV6. A, Map-based cloning of RAV6. The RAV6 locus was mapped to the long arm ofrice chromosome 2 (Chr. 2) betweenmarkers RM3512 and RM318. Thegene was further delimited to an 18-kb genomic region between themarkers ID8-26 and ID5-3 within the bacterial artificial chromosome(BAC) clones AP004178 and AP004255. The number of recombinants ismarked corresponding to the molecular markers. This 18-kb intervalcontains the candidate gene Os02g45850. Gray and black boxes indi-cate the untranslated regions and the coding sequence, respectively. Band C, Expression analysis of RAV6 by reverse transcription-PCR (B) andreal-time reverse transcription-PCR (C) in the wild type (WT), Epi-rav6(+/–), and Epi-rav6 (–/–). Rice elongation factor 1a (OsEF1a) was used asa control. Values are means 6 SD of three biological replicates.

Figure 3. Overexpression of RAV6 phenocopies the Epi-rav6 mutantphenotype. A, The morphologies of wild-type (WT), Epi-rav6 (+/–), andtransgenic plants overexpressing RAV6 at seedling stages. B, Relativeexpression analysis of RAV6 in wild-type, Epi-rav6 (+/–), and the RAV6-overexpressing plants at seedling stage. C and D, The grain morphologiesof the transformed wild-type plants. OV-1 and OV-2 indicate transgenicplants overexpressing RAV6. Bars = 1 cm (A) and 2 cm (C and D).









The Effect of MITE-Mediated DNA Methylation on RAV6Expression Is Conserved in Cultivated Rice

MITEs are short (less than 600 bp), nonautonomousDNA transposons. As the TEs with the highest copynumbers in the rice genome, MITEs mainly occur in thechromosomal arms, especially in the vicinity of genes(Jiang et al., 2004). According to the Plant Repeat Da-tabase, the japonica rice ‘Nipponbare’ genome contains

2,984 copies of the adh5-like MITE (Ouyang and Buell,2004). The adh5-like MITE in the RAV6 promoter re-gion is 100 bp long and located 198 bp upstream of theRAV6 gene body.

To further explore the evolutionary significance ofthe epigenetic regulation of RAV6, we investigated thenatural variation of the MITE in the promoter region oftheRAV6 locus in 40 accessions of cultivated rice, whichrepresent all of the major groups of Asian cultivated rice

Figure 4. Evolutionary relationship between RAV6 and RAVL1 in rice. A, Phylogenetic relationship of the RAV subfamily of B3DNA-binding domain-containing proteins in rice. The bar indicates substitutions per site. B, Amino acid sequence alignment ofrice RAV6 and RAVL1. Identical and similar amino acids are boxed in black and gray, respectively. The signature B3-bindingdomain is underlined.


Zhang et al.



(Xu et al., 2012). The results showed that the TE in-sertion in the promoter region of RAV6 is conservedin all cultivated rice accessions (data not shown). Tofurther confirm this, we aligned the sequences ofthe MITEs, as well as the 59 region of RAV6, in fourknown assembled rice genomes, including threecultivated rice strains, cv Nipponbare, 93-11, andKasalath, and the wild rice Oryza brachyantha (Yu et al.,2002; International Rice Genome Sequencing Project,2005; Chen et al., 2013; Sakai et al., 2014). The MITEinsertion was conserved in the cultivated rice ge-nomes but was absent in O. brachyantha (Fig. 7A).Consistent with this, the sequence of the 59 promoterregion of RAV6 is also more conserved in cultivatedrice, especially the six cytosine methylation sites,which are essential for regulation of RAV6 expression(Fig. 7A).To further determine the conserved effect of the

MITE-mediated DNAmethylation on RAV6 expressionin cultivated rice accessions, we measured RAV6 tran-script levels and DNA methylation patterns in fourcultivated rice strains, including three japonica (cvNipponbare, Kongyu131, and Koshihikari) and oneindica (cv Kasalath) accessions. Compared with thehigh expression of RAV6 observed in Epi-rav6, weobserved little or no RAV6 expression in leaves at thetillering stage in all cultivated rice strains (Fig. 7B).Like the pattern of wild-type cv Zhonghua11, the other

four cultivated rice accessions also showed higherDNA methylation in the contiguous 96-bp region ofRAV6 promoter, in contrast with very low DNAmethylation in Epi-rav6 (Fig. 7C; Supplemental Fig.S4). Thus, it is reasonable to predict that the TE insertionoccurred prior to divergence of indica and japonicarice and that the TE-mediated epigenetic regulationof RAV6 expression is evolutionarily conserved incultivated rice.

DISCUSSION

In this work, we identified an epiallele of rice RAV6;this epiallele shows hypomethylation of the RAV6promoter region, causing ectopic RAV6 expression,larger lamina inclination, and smaller grains. We foundthat RAV6 expression is associated with extensivemethylation of a MITE along with the nearby sequencein the 59 region of this gene. RAV6 encodes a B3 DNA-binding domain-containing protein that has the mostsequence similarity to RAVL1, which can maintain BRhomeostasis to control rice development (Je et al.,2010). Functional analysis revealed that RAV6 coor-dinately activates the BR receptor BRI1 and BR bio-synthetic genes to control rice leaf angle. This suggeststhat RAV6 performs a similar function to its homolog,RAVL1.

Figure 5. RAV6 regulates leaf angle bycontrolling genes affecting BR homeosta-sis. A, Real-time reverse transcription-PCRanalyses of BRD1, D2, D11, and BRI1 inwild-type (WT), Epi-rav6 (+/–), and Epi-rav6 (–/–) plants. 25S ribosomal RNA wasused as a control. Values are means6 SD ofthree biological replicates. B, Lamina in-clination of 45-d-old seedlings in wild-type, Epi-rav6 (+/–), and Epi-rav6 (–/–)plants subjected to Pcz treatment for 15 d.Bar = 20 cm.







The Rice Mutant Epi-rav6 Shows Hypomethylation ofDNA at RAV6

Ectopic expression of RAV6 in Epi-rav6 mutants in-dicates that it is a gain-of-function allele of RAV6. Al-though we found no DNA sequence changes in RAV6,we did find that Epi-rav6 shows a loss of cytosine DNAmethylation in a MITE along with the nearby sequencelocated in its 59 promoter region (Fig. 6A; SupplementalFig. S3A). Thus, Epi-rav6 has similar features to theArabidopsis fwa mutant, the first DNA-hypomethylatedmutant identified in plants, with hypomethylation at twodirect repeats in the 59 region of FWA and ectopic ex-pression of FWA (Soppe et al., 2000). In rice, only twoepialleles, Epi-d1 and Epi-df, were identified previously,with hyper- and hypomethylation at the causal genes,respectively (Miura et al., 2009; Zhang et al., 2012). Al-though both Epi-df and Epi-rav6 are hypomethylated, themethylation sites of the two epialleles differ. Epi-df showshypomethylation mainly in the coding region and notassociated with repeated sequences (Zhang et al., 2012).However, the gain ofmethylation at theD1 locus inEpi-d1is associated with the upstream repeat sequences (Miuraet al., 2009), similar to fwa and Epi-rav6. Therefore, theserepeat elements around the causal genes in Epi-d1, fwa,and Epi-rav6 might function as epigenetic mediators toinfluence nearby gene expression, further confirmingprevious findings (Hollister et al., 2011; Wei et al., 2014).

The phenotypes of epialleles generally show meta-stable inheritance with a low percentage of revertants,such as in Epi-d1 and Epi-df. Epi-rav6 is a semidominantmutant, and the homozygous lines are sterile. There-fore, it is impossible to identify authentic revertantsbetween generations (Fig. 1).

RAV6 Controls Rice Lamina Inclination via MediatingBR Homeostasis

The B3 DNA-binding domain-containing proteinsare plant-specific transcription factors that play im-portant roles in growth, development, flowering time,seed development, and seed maturation (Swaminathanet al., 2008; Romanel et al., 2009). The B3 domain canbind DNA in a sequence-specific or nonspecific mannerand activate or repress the transcription of specifictarget genes (Je et al., 2010). Both RAV6 and RAVL1belong to the RAV family and have higher sequencesimilarity to each other than to other members of theRAV family (Fig. 4). RAVL1 functions as a transcriptionfactor, directly binding BR biosynthetic and receptorgenes and thus maintaining BR homeostasis (Je et al.,2010). In Epi-rav6, the direct targets of RAVL1, includ-ing BRI1, D2, D11, and BRD1, showed increased ex-pression (Fig. 5A), indicating that RAV6 might alsotarget these genes. The lines overexpressing RAVL1

Figure 6. DNAmethylation analysis of theRAV6 locus. A, Schematic representation ofRAV6with putativeMITE,MITE-adh5-likeDNA transposon. Boxes indicate exon (black), untranslated regions (white), andMITE (green). The red line shows the region usedfor analysis of DNAmethylation by bisulfite sequencing (B). B, DNAmethylation status of bisulfite-sequenced region (as indicatedin A) in wild-type (WT) and Epi-rav6 (–/–) plants. Histograms represent the percentage of CG (red), CNG (blue), and CHH (green).C, Real-time reverse transcription-PCR analyses of RAV6 expression in 7-d-old seedlings treatedwith (+) or without (–) 5-aza-dC,an inhibitor of DNA methylation. Rice elongation factor 1a (OsEF1a) was used as a control. Values are means 6 SD of threebiological replicates.


Zhang et al.





showed larger lamina inclination and hypersensitivityto BR (Je et al., 2010). Consistent with this, increased leafangles were observed in Epi-rav6 mutants and RAV6overexpression lines (Fig. 3A). Moreover, treatmentwith BR inhibitor fully rescued the larger lamina

inclination in Epi-rav6 (Fig. 5B). All these results indi-cated that both RAV6 and RAVL1 control leaf anglesvia maintaining BR homeostasis.

In addition to leaf angle, BRs also influence rice grainsize (Zhang et al., 2014). Generally, BR-deficient and

Figure 7. Genome sequence, gene expression, and DNA methylation analysis of the RAV6 locus in different rice accessions. A,Comparative sequence analysis of the RAV6 locus in four assembled rice genomes. The synteny of the RAV6 locus in cv Nip-ponbare, 93-11, and Kasalath andO. brachyantha is shown in the top section. The bottom section shows an alignment of theDNAsequence 258 bp upstream and the 59 gene body of RAV6 in four assembled rice genomes. TheMITE region and 59 gene body areunderlined by green and black lines, respectively. The asterisk indicates the cytosine region, whose degree of methylation wasreduced in Epi-rav6 (–/–), as shown in Figure 6B. B, Expression analysis of RAV6 by reverse transcription-PCR (top) and real-timereverse transcription-PCR (bottom) in cultivated rice strains and Epi-rav6 (–/–). Zhonghua11, Nipponbare, Kongyu131, andKoshihikari are japonica cultivars, and Kasalath is an indica cultivar. Rice elongation factor 1a (OsEF1a) was used as a control.Values are means6 SD of three biological replicates. C, DNAmethylation status of the bisulfite-sequenced region (as indicated inFig. 6A) in thewild type (WT; cv Zhonghua11), Epi-rav6 (–/–), and four cultivated rice strains. Histograms represent the percentageof methylation in the CG (red), CNG (blue), and CHH (green) contexts.





BR-insensitive mutants, such as d2, d11, and brd1,showed small grains, while overexpression lines suchas Increase Leaf Inclination1 and BRI1-SUPPRESSOR1showed large lamina joints and increased grain size(Hong et al., 2002, 2003; Tanabe et al., 2005; Tanakaet al., 2009). The Epi-rav6 plants showed activation ofBR signaling and large leaf angles, but small grains,which differs from typical mutants affecting BR. Weproposed that decreased grain size in Epi-rav6 may re-sult from an integrated effect of changes in expressionof multiple target genes. Because RAV6 encodes a B3DNA-binding domain-containing protein, lots of targetgenes involved in multiple developmental processesmight be influenced by the gain-of-function changeof RAV6 in Epi-rav6. This also can be confirmed, as theEpi-rav6 homozygotes showed pleiotropic develop-mental defects (Fig. 1A).

MATERIALS AND METHODS

Plant Materials and Growth Conditions

A spontaneously occurring rice (Oryza sativa) mutant Epi-rav6 was isolatedfrom a japonica rice ‘Zhonghua11’ population in Guangzhou, China, in 2006.Transgenic plants overexpressing RAV6 were produced by introducing therespective genes driven by the maize (Zea mays) Ubiquitin promoter into wild-type plants. Briefly, the 1,239-bp coding sequence of RAV6 was amplified withthe primers OsRAV-1BXF and OsRAV-1239RNS (Supplemental Table S1), andthe PCR product was inserted into the BamHI/SpeI sites of pCUbi1390,resulting in a plant expression vector driven by the Ubiquitin promoter, whichwas then transformed into rice. Plants were grown in the field under naturalconditions.

Map-Based Cloning

For map-based cloning ofOsRAV6, 725 recessive individual plants showingnormal leaf inclinations were selected from an F2 population derived from across between the Epi-rav6 mutant and indica var Huajingxian74. Simple se-quence repeat markers and insertion/deletion markers on chromosome 2 wereused for finemapping. TheRAV6 genewas selected from an approximately 18-kbregion as the candidate gene. To find out the mutation site, we amplified thecorresponding fragments from the Epi-rav6 mutant and wild-type plants, respec-tively. Primers used in the map-based cloning are listed in Supplemental Table S1.

Reverse Transcription-PCR and Real-Time PCR

Real-time PCR analysis was performed using the CFX96 Real-Time PCRSystem (Bio-Rad) and SYBR Green I (S-7567; Invitrogen). PCR was performedusing Hot-Start Taq DNA polymerase (DR007B; Takara Bio). For each sample,quantifications were made in triplicate. Melting curves were read at the end ofeach amplification by steps of 0.3°C from 65°C to 95°C to ensure that thequantifications were derived from real PCR products and not primer dimers.The primers used for Figure 5A are listed in Supplemental Table S1.

Bisulfite Sequencing

Genomic DNAs were isolated from 3-week-old rice seedlings or the leaf attillering stage. Bisulfite treatment of genomic DNA was conducted using theMethylation Gold kit according to the manufacturer’s instructions (ZYMOResearch Corporation). Treated DNAs were then used for PCR amplificationusing Hot-Start Taq DNA polymerase (DR007B; TaKaRa). PCR conditions wereas follows: 95°C for 5 min; 35 cycles of 95°C for 30 s, 50°C for 30 s, and 72°C for40 s; and 72°C for 10 min. The PCR products were cloned into the pEASY-T1cloning vector (TransGen Biotech), and individual clones were sequenced. Theanalysis of sequencing data was performed using KISMETH (Gruntman et al.,2008). The primers used for bisulfite sequencing are listed in Supplemental Table S1.

5-aza-dC Treatment

Rice seeds were soaked in water at 37°C for 16 h. The seeds were then im-mersed in 20 mM Tris-HCl, pH 7.5, with or without 0.3 mM 5-aza-dC (A3656;Sigma) at room temperature for 48 h in the dark. After washing, seeds wereplanted in soil. Seven-day-old seedlings were sampled for further analysis.

Phylogenetic Analysis and Alignment

Full-length protein sequences were used for phylogenetic analysis. The se-quence of OsRAV6 was downloaded from TIGR (release 7; http://rice.plantbiology.msu.edu/) and used to search the proteome database of riceavailable at TIGR using BLAST. Alignments of protein sequences were per-formed using MUSCLE (Edgar, 2004) with default parameters (gap opening, –2.9;gap extension, 0; hydrophobicity multiplier, 1.2; and max iteration, 8). The maxi-mum likelihood tree was constructed using MEGA6 (Tamura et al., 2013). Forty-eight different amino acid substitution models were tested, and the JTT+G+Fmodel (Jones et al., 1992) was considered the best model with the lowest BayesianInformationCriterion scores.All sites of the alignmentwere used formodel testing.The –ln likelihood was 10,996.08. Nonparametric bootstrap analyses (Sandersonand Wojciechowski, 2000) consisted of 1,000 replicates. All four known genomesequences of Oryza spp. were collected, including japonica rice ‘Nipponbare’ (In-ternational Rice Genome Sequencing Project, 2005), indica rice ‘93-11’ (Yu et al.,2002), aus rice ‘Kasalath’ (Sakai et al., 2014), and the wild rice Oryza brachyantha(Chen et al., 2013). The genomic DNA sequence of the OsRAV6 gene body and230-bp upstream region was used as a query sequence, and homologous regionswere searched in the other three genomes of Oryza spp. using FASTA (version36.3.5d; Pearson and Lipman, 1988). Multiple sequence alignment was performedusing ClustalW (version 2.1; Thompson et al., 1994; Larkin et al., 2007).

Pcz Treatment

In this study, Pcz, a BR biosynthesis inhibitor (Hartwig et al., 2012), was usedto measure the mutants’ growth and morphological responses to BR by ob-serving inclination of lamina joints after exposure to Pcz. For Pcz treatment, theplants were grown in the fields after incubating seeds for 3 d at 30°C to ensuresynchronized germination. The 30-d-old seedlings were sprayed with 10 mM

Pcz (BannerMaxx; Syngenta) every 3 d. After 15 d, plants were transplanted intopots and photographed. Control plants were grown and not treated with Pcz.

Supplemental Data

The following supplemental materials are available.

Supplemental Figure S1. The panicle and seeds of wild-type and Epi-rav6(+/–) plants.

Supplemental Figure S2. The wild-type, Epi-rav6 (+/–), and hybrid F1plants at the heading stage.

Supplemental Figure S3. Analyses of DNA methylation levels at RAV6 inthe wild type and mutants by bisulfite sequencing.

Supplemental Figure S4. Analyses of DNA methylation levels at RAV6 inthe wild type, Epi-rav6 (–/–), and four cultivated rice strains by bisulfitesequencing.

Supplemental Table S1. List of primers used in this paper.

Received June 2, 2015; accepted September 6, 2015; published September 8,2015.

LITERATURE CITED

Bai MY, Zhang LY, Gampala SS, Zhu SW, Song WY, Chong K, Wang ZY(2007) Functions of OsBZR1 and 14-3-3 proteins in brassinosteroid sig-naling in rice. Proc Natl Acad Sci USA 104: 13839–13844

Cao X, Jacobsen SE (2002) Role of the Arabidopsis DRM methyltransfer-ases in de novo DNA methylation and gene silencing. Curr Biol 12:1138–1144

Chang S, Pikaard CS (2005) Transcript profiling in Arabidopsis revealscomplex responses to global inhibition of DNA methylation and histonedeacetylation. J Biol Chem 280: 796–804


Zhang et al.






http://rice.plantbiology.msu.edu/

http://rice.plantbiology.msu.edu/







Chen J, Huang Q, Gao D, Wang J, Lang Y, Liu T, Li B, Bai Z, LuisGoicoechea J, Liang C, et al (2013) Whole-genome sequencing of Oryzabrachyantha reveals mechanisms underlying Oryza genome evolution.Nat Commun 4: 1595

Cubas P, Vincent C, Coen E (1999) An epigenetic mutation responsible fornatural variation in floral symmetry. Nature 401: 157–161

Edgar RC (2004) MUSCLE: a multiple sequence alignment method withreduced time and space complexity. BMC Bioinformatics 5: 113

Gruntman E, Qi Y, Slotkin RK, Roeder T, Martienssen RA, Sachidanandam R(2008) Kismeth: analyzer of plant methylation states through bisulfite se-quencing. BMC Bioinformatics 9: 371

Hartwig T, Corvalan C, Best NB, Budka JS, Zhu JY, Choe S, Schulz B(2012) Propiconazole is a specific and accessible brassinosteroid(BR) biosynthesis inhibitor for Arabidopsis and maize. PLoS One 7:e36625

Hollister JD, Gaut BS (2009) Epigenetic silencing of transposable elements:a trade-off between reduced transposition and deleterious effects onneighboring gene expression. Genome Res 19: 1419–1428

Hollister JD, Smith LM, Guo YL, Ott F, Weigel D, Gaut BS (2011)Transposable elements and small RNAs contribute to gene expressiondivergence between Arabidopsis thaliana and Arabidopsis lyrata. Proc NatlAcad Sci USA 108: 2322–2327

Hong Z, Ueguchi-Tanaka M, Fujioka S, Takatsuto S, Yoshida S, HasegawaY, Ashikari M, Kitano H, Matsuoka M (2005) The rice brassinosteroid-deficient dwarf2 mutant, defective in the rice homolog of ArabidopsisDIMINUTO/DWARF1, is rescued by the endogenously accumulatedalternative bioactive brassinosteroid, dolichosterone. Plant Cell 17: 2243–2254

Hong Z, Ueguchi-Tanaka M, Shimizu-Sato S, Inukai Y, Fujioka S,Shimada Y, Takatsuto S, Agetsuma M, Yoshida S, Watanabe Y, et al(2002) Loss-of-function of a rice brassinosteroid biosynthetic enzyme,C-6 oxidase, prevents the organized arrangement and polar elongationof cells in the leaves and stem. Plant J 32: 495–508

Hong Z, Ueguchi-Tanaka M, Umemura K, Uozu S, Fujioka S,Takatsuto S, Yoshida S, Ashikari M, Kitano H, Matsuoka M (2003) Arice brassinosteroid-deficient mutant, ebisu dwarf (d2), is caused by aloss of function of a new member of cytochrome P450. Plant Cell 15:2900–2910

Hu L, Li N, Xu C, Zhong S, Lin X, Yang J, Zhou T, Yuliang A, Wu Y,Chen YR, et al (2014) Mutation of a major CG methylase in ricecauses genome-wide hypomethylation, dysregulated genome ex-pression, and seedling lethality. Proc Natl Acad Sci USA 111: 10642–10647

International Rice Genome Sequencing Project (2005) The map-basedsequence of the rice genome. Nature 436: 793–800

Jacobsen SE, Meyerowitz EM (1997) Hypermethylated SUPERMAN epi-genetic alleles in Arabidopsis. Science 277: 1100–1103

Je BI, Piao HL, Park SJ, Park SH, Kim CM, Xuan YH, Park SH, Huang J,Do Choi Y, An G, et al (2010) RAV-Like1 maintains brassinosteroidhomeostasis via the coordinated activation of BRI1 and biosyntheticgenes in rice. Plant Cell 22: 1777–1791

Jiang N, Feschotte C, Zhang X, Wessler SR (2004) Using rice to understandthe origin and amplification of miniature inverted repeat transposableelements (MITEs). Curr Opin Plant Biol 7: 115–119

Jones DT, Taylor WR, Thornton JM (1992) The rapid generation of mu-tation data matrices from protein sequences. Comput Appl Biosci 8: 275–282

Kakutani T (2002) Epi-alleles in plants: inheritance of epigenetic informa-tion over generations. Plant Cell Physiol 43: 1106–1111

Larkin MA, Blackshields G, Brown NP, Chenna R, McGettigan PA,McWilliam H, Valentin F, Wallace IM, Wilm A, Lopez R, et al(2007) Clustal W and Clustal X version 2.0. Bioinformatics 23: 2947–2948

Law JA, Jacobsen SE (2010) Establishing, maintaining and modifyingDNA methylation patterns in plants and animals. Nat Rev Genet 11:204–220

Lindroth AM, Cao X, Jackson JP, Zilberman D, McCallum CM, HenikoffS, Jacobsen SE (2001) Requirement of CHROMOMETHYLASE3 formaintenance of CpXpG methylation. Science 292: 2077–2080

Manning K, Tör M, Poole M, Hong Y, Thompson AJ, King GJ,Giovannoni JJ, Seymour GB (2006) A naturally occurring epigeneticmutation in a gene encoding an SBP-box transcription factor inhibitstomato fruit ripening. Nat Genet 38: 948–952

Martin A, Troadec C, Boualem A, Rajab M, Fernandez R, Morin H, PitratM, Dogimont C, Bendahmane A (2009) A transposon-induced epige-netic change leads to sex determination in melon. Nature 461: 1135–1138

Miura K, Agetsuma M, Kitano H, Yoshimura A, Matsuoka M, JacobsenSE, Ashikari M (2009) A metastable DWARF1 epigenetic mutantaffecting plant stature in rice. Proc Natl Acad Sci USA 106: 11218–11223

Moritoh S, Eun CH, Ono A, Asao H, Okano Y, Yamaguchi K, ShimataniZ, Koizumi A, Terada R (2012) Targeted disruption of an orthologue ofDOMAINS REARRANGED METHYLASE 2, OsDRM2, impairsthe growth of rice plants by abnormal DNA methylation. Plant J 71:85–98

Ouyang S, Buell CR (2004) The TIGR Plant Repeat Databases: a collectiveresource for the identification of repetitive sequences in plants. NucleicAcids Res 32: D360–D363

Pearson WR, Lipman DJ (1988) Improved tools for biological sequencecomparison. Proc Natl Acad Sci USA 85: 2444–2448

Romanel EA, Schrago CG, Couñago RM, Russo CA, Alves-Ferreira M(2009) Evolution of the B3 DNA binding superfamily: new insights intoREM family gene diversification. PLoS One 4: e5791

Sakai H, Kanamori H, Arai-Kichise Y, Shibata-Hatta M, Ebana K,Oono Y, Kurita K, Fujisawa H, Katagiri S, Mukai Y, et al (2014)Construction of pseudomolecule sequences of the aus rice cultivarKasalath for comparative genomics of Asian cultivated rice. DNARes 21: 397–405

Sakamoto T, Morinaka Y, Ohnishi T, Sunohara H, Fujioka S, Ueguchi-Tanaka M, Mizutani M, Sakata K, Takatsuto S, Yoshida S, et al (2006)Erect leaves caused by brassinosteroid deficiency increase biomassproduction and grain yield in rice. Nat Biotechnol 24: 105–109

Sanderson MJ, Wojciechowski MF (2000) Improved bootstrap confidencelimits in large-scale phylogenies, with an example from Neo-Astragalus(Leguminosae). Syst Biol 49: 671–685

Saze H, Mittelsten Scheid O, Paszkowski J (2003) Maintenance of CpGmethylation is essential for epigenetic inheritance during plant game-togenesis. Nat Genet 34: 65–69

Sinclair TR, Sheehy JE (1999) Erect leaves and photosynthesis in rice.Science 283: 1456–1457

Soppe WJ, Jacobsen SE, Alonso-Blanco C, Jackson JP, Kakutani T,Koornneef M, Peeters AJ (2000) The late flowering phenotype of fwamutants is caused by gain-of-function epigenetic alleles of a homeodo-main gene. Mol Cell 6: 791–802

Swaminathan K, Peterson K, Jack T (2008) The plant B3 superfamily.Trends Plant Sci 13: 647–655

Tamura K, Stecher G, Peterson D, Filipski A, Kumar S (2013) MEGA6:Molecular Evolutionary Genetics Analysis version 6.0. Mol Biol Evol 30:2725–2729

Tanabe S, Ashikari M, Fujioka S, Takatsuto S, Yoshida S, Yano M,Yoshimura A, Kitano H, Matsuoka M, Fujisawa Y, et al (2005) A novelcytochrome P450 is implicated in brassinosteroid biosynthesis via thecharacterization of a rice dwarf mutant, dwarf11, with reduced seedlength. Plant Cell 17: 776–790

Tanaka A, Nakagawa H, Tomita C, Shimatani Z, Ohtake M, NomuraT, Jiang CJ, Dubouzet JG, Kikuchi S, Sekimoto H, et al (2009)BRASSINOSTEROID UPREGULATED1, encoding a helix-loop-helixprotein, is a novel gene involved in brassinosteroid signaling andcontrols bending of the lamina joint in rice. Plant Physiol 151: 669–680

Thompson JD, Higgins DG, Gibson TJ (1994) CLUSTAL W: improving thesensitivity of progressive multiple sequence alignment through se-quence weighting, position-specific gap penalties and weight matrixchoice. Nucleic Acids Res 22: 4673–4680

Tong H, Chu C (2012) Brassinosteroid signaling and application in rice.J Genet Genomics 39: 3–9

Vongs A, Kakutani T, Martienssen RA, Richards EJ (1993) Arabidopsisthaliana DNA methylation mutants. Science 260: 1926–1928

Wei L, Gu L, Song X, Cui X, Lu Z, Zhou M, Wang L, Hu F, Zhai J, MeyersBC, et al (2014) Dicer-like 3 produces transposable element-associated24-nt siRNAs that control agricultural traits in rice. Proc Natl Acad SciUSA 111: 3877–3882

Xu X, Liu X, Ge S, Jensen JD, Hu F, Li X, Dong Y, Gutenkunst RN, Fang L,Huang L, et al (2012) Resequencing 50 accessions of cultivated and wildrice yields markers for identifying agronomically important genes. NatBiotechnol 30: 105–111





Yamamuro C, Ihara Y, Wu X, Noguchi T, Fujioka S, Takatsuto S,Ashikari M, Kitano H, Matsuoka M (2000) Loss of function ofa rice brassinosteroid insensitive1 homolog prevents internodeelongation and bending of the lamina joint. Plant Cell 12: 1591–1606

Yamauchi T, Johzuka-Hisatomi Y, Terada R, Nakamura I, Iida S (2014)The MET1b gene encoding a maintenance DNA methyltransferase isindispensable for normal development in rice. Plant Mol Biol 85: 219–232

Yu J, Hu S, Wang J, Wong GK, Li S, Liu B, Deng Y, Dai L, Zhou Y, ZhangX, et al (2002) A draft sequence of the rice genome (Oryza sativa L. ssp.indica). Science 296: 79–92

Zhang C, Bai MY, Chong K (2014) Brassinosteroid-mediated regulation ofagronomic traits in rice. Plant Cell Rep 33: 683–696

Zhang L, Cheng Z, Qin R, Qiu Y, Wang JL, Cui X, Gu L, Zhang X, Guo X,Wang D, et al (2012) Identification and characterization of an epi-alleleof FIE1 reveals a regulatory linkage between two epigenetic marks inrice. Plant Cell 24: 4407–4421


Zhang et al.



Documents

Epigenetic Mutation of RAV6 Affects Leaf Angle and Seedmuch higher expression in Epi-rav6 (+/–) and Epi-rav6 (–/–) leaves (Fig. 2, B and C). Moreover, the degree of up-regulation