Shilei Xu,Yonghai Luo,Zhigang Cai,Xiangling Cao,Xiaohe Hu,3,Jun Yang and Da Luo,3

1National Key Laboratory of Plant Molecular Genetics,Institute of Plant Physiology and Ecology,Shanghai Institutes for Biological Sciences,the Chinese Academy of Sciences,Shanghai 200032,China

2Graduate University of the Chinese Academy of Sciences,Beijing 100049,China

3School of Life Sciences,Sun Yat-sen University,Guangzhou 510275,China

4Laboratory of Cancer Cell Biology,Tianjin Medical University Cancer Hospital and Institute,Tianjin 300060,China

Introduction

TCP genes have been found in all the plant genomes investigated so far,and encode transcription factors regulating different aspects of plant growth and development.Based on sequence similarity and functional diversity,TCP genes among different plant genomes can be divided into two categories(Cubas et al.1999;Kosugi and Ohashi 2002;Martin-Trillo and Cubas 2010).Class I TCP genes are shown to promote plant growth and proliferation(Kosugi and Ohashi 1997,2002;Li et al.2005).Class II TCP genes can be subdivided to two clades:the CIN clade whose members are involved in lateral organ development(Nath et al.2003;Crawford et al.2004),and the CYC/TB1 clade whose members are involved in the development of flowers or lateral shoots(Doebley et al.1995,1997;Luo et al.1996,1999).To date,it has been shown that TCP genes play multiple roles in different aspects of plant growth and development(Martin-Trillo and Cubas 2010).

TCP genes encode transcription factors which can act as either transcriptional activators or repressors.For example,AtTCP20(a class I TCP transcription factor in Arabidopsis)and TEOSINTE BRANCHED1(TB1,a class II TCP transcription factor in maize(Zea mays))have been found to have strong activation activity(Kosugi and Ohashi 2002;Li et al.2005).On the other hand,AtTCP21 and AtTCP24 belong to class I and II TCP proteins in Arabidopsis,respectively,and act as repressors(Masuda et al.2008;Pruneda-Paz et al.2009).However,some of the TCP proteins,such as the PCF2,PCF5,and PCF6 in rice,show very weak or little transcription activity in in vitro assays(Kosugi and Ohashi 2002).These data suggest that the transcriptional activity of TCP transcription factors depends on both the protein properties and on their interaction with other proteins.

During zygomorphic(bilaterally symmetrical)floral development,it has been shown that the CYC-like TCP genes are key regulators recruited independently in different plants to control dorsoventral(DV)identities(Cubas 2004;Feng et al.2006;Preston and Hileman 2009;Jabbour et al.2009).In zygomorphic flowers,typically three types of petals(dorsal,lateral,and ventral)develop along a floral DV axis(Luo et al.1996;Cubas et al.1999;Endress 1999,2001).The dorsal,lateral,and ventral petals possess different characteristics of DV identities,such as shape and size.Two TCP genes,CYC and its homolog DICH,have been cloned in Antirrhinum majus,and were found to be involved in the control of zygomorphic flower development(Luo et al.1996,1999).The transcription of both CYC and DICH can only be detected in the dorsal region of floral meristems during early floral development.While the lossof-function mutant of CYC compromises the identities of both dorsal and lateral petals,the complete ventralized phenotype was observed in the cyc dich double mutant(Luo et al.1996,1999),suggesting that a redundant function of the CYC-like TCP genes is the control of dorsal and lateral identities during zygomorphic floral development in A.majus.

In Papilionoideae legumes,a gene cluster which is comprised of three highly homologous genes,LjCYC1,LjCYC2,and LjCYC3,and encodes CYC-like TCP transcription factors,has been identified in the genome of Lotus japonicus(Feng et al.2006).Among these genes,LjCYC1 and LjCYC2 are only expressed in the dorsal region of the floral meristem,but LjCYC3 is expressed in both the dorsal and the lateral regions(Feng et al.2006;Wang et al.2010b;Weng et al.2011).The mutant of LjCYC2,squared standard 1(squ1),has been identified,and the dorsal identity in the mutant flowers of squ1 was shown to be compromised.In a transgenic experiment,it was shown that ectopic expression of LjCYC2 is sufficient to confer dorsal identity to petals in the lateral and ventral regions(Feng et al.2006).In contrast,reduction of LjCYC1 expression by an RNA interference(RNAi)construct in the transgenic experiment did not give rise to a detectable phenotype(Wang et al.2010b).These data demonstrate that LjCYC2 is a key factor in determining dorsal identity.It has been shown that another factor,Keeled wings in Lotus 1(KEW1),is responsible for the establishment of lateral identity during zygomorphic development(Feng et al.2006).Furthermore,the squ1 kew1 double mutant in L.japonicus bears all the petals with ventral identity to different extents,suggesting genetic interaction between LjCYC2 and Kew1(Feng et al.2006;Jabbour et al.2009).

Keeled wings(K)is a single locus in pea identified in a classical genetic study.The flowers of the kew1 mutant in L.japonicus,as well as those of k in Pisum sativum L.plants,display abnormal petal appearance in the lateral region,which mimics ventral petals(Figure 1A,Feng et al.2006;Wang et al.2008).When K was cloned,it was found that K may encode a CYC-like TCP protein,PsCYC3,the homolog of LjCYC3(Wang et al.2008).When an RNAi construct was applied to knockdown LjCYC3 in a transgenic experiment,the phenotype of the transgenic flowers was similar to that of kew1(Wang et al.2010b),confirming that LjCYC3 is necessary to control the identity of lateral petals.In a mapping experiment,kew1 was positioned in a 200 kb genomic region containing LjCYC3(Wang et al.2008).However,the fine-mapping of kew1 failed because no recombination was found in the 200 kb region.Furthermore,it was difficult to determine the essence of the kew1 mutation,due to failure to identify the sequence variation between kew1 and the wild type in the genomic sequence of the LjCYC3 locus(Wang et al.2008).These data suggest that kew1 could be a mutant allele of either LjCYC3 or alternatively a regulator of LjCYC3.Recently,another kew mutant,kew3,was identified as an allele of kew1,which provides a good opportunity for future investigation of kew mutants(Wang et al.2010a).

In this study,we analyzed kew1 and its allelic mutant kew3,and demonstrated that kew1 and kew3 are allelic to LjCYC3.The sequence variation among different LjCYC genes was further analyzed,and the genes’diverse functions were investigated in a transgenic experiment.It was found that LjCYC1,similar to LjCYC2,possesses a dorsalizing function,and that LjCYC3 alone is sufficient to confer lateral identity to kew1 plants,supporting the notion that dorsal and lateral identities are controlled by the diverse functions of different LjCYC genes.However,LjCYC3 failed to alter the dorsal identity in transgenic plants,indicating an epistatic effect between the dorsal and lateral functions during zygomorphic floral development.In a transcriptional activation assay,both LjCYC1 and LjCYC2 displayed weak activity,but LjCYC3 exhibited strong activity.In addition,the activity of the LjCYC3 protein can be dramatically reduced by co-expression with LjCYC2,suggesting a modulating effect of the interaction between the LjCYC2 and LjCYC3 proteins.Taken together,our data demonstrate that sequence variation and the subsequent alteration of protein property may be essential for the diverse functions of different LjCYC genes in their control of the zygomorphic floral development in Papilionoideae legumes.

Figure 1.Analysis of kew mutations.(A)Flowers of wild type(Gifu ecotype),kew1,and kew3 mutants.Red arrows indicate lateral petals(the wild type)and the ventralized lateral petals(kew mutants).FF,front view of flowers;FS,side view of flowers;DP,dorsal petal;LP,lateral petal;VP,ventral petal.Bar=0.5 cm.(B)Expression of LjCYC3 in dorsal and lateral petals of wild type,kew1,and kew3.Expression of LjCYC3 is analyzed by quantitative polymerase chain(PCR)reaction.In the wild type,LjCYC3 is expressed in the dorsal and lateral petals,but not in the ventral petals.The expression level is dramatically decreased in kew1.The expression of the abnormal transcript in kew3 is low in the dorsal petals but very high in the lateral petals.(C)Restriction fragment length polymorphism(RFLP)analysis of kew1.Southern hybridization was performed using LjCYC3 open reading frame(ORF)as the probe,and an RFLP of 5.3 kb MspI fragment(the blue arrow)was detected in kew1.(D)Inversion of a large DNA fragment containing the LjCYC3 locus in kew1.A schematic diagram of the regions containing LjCYC3 in the wild type and kew1,respectively:the large arrows indicate the orientation of the DNA sequence in the wild type,the long red arrow represents the ORF of LjCYC3,and the thick blue line indicates the 5.3 kb MspI fragment detected in the RFLP analysis.The breakpoint of the inversion is indicated.P1,P2,P3,and P4:the primers used in the inverse PCR to amplify the self-ligation product from the 5.3 kb MspI fragments(blue line)from kew1.α,β,γ,and ε:the primers used for“breakpoint-specific”P(pán)CR to confirm the breakpoints of the inversion in kew1.(E)One base deletion in the ORF of LjCYC3 in kew3.(F)Confirmation of the DNA rearrangement in kew1 by “breakpoint-specific”P(pán)CR.

Results

Analysis of kew mutants

In the previous mapping experiment,no sequence variation was detected between the wild type and the kew1 mutant in the LjCYC3 locus(Wang et al.2008),and the particularity of the kew1 mutant was unclear.After the first analysis of the kew1 mutant,another allelic mutant,kew3,was identified(Figure 1A,Wang et al.2010a),and these two allelic mutants were investigated in this study.In a quantitative reverse transcription polymerase chain reaction(RT–PCR)experiment,expression of LjCYC3 was found to be abnormal in the dorsal and lateral petals of both kew1 and kew3.The expression level of LjCYC3 was dramatically decreased in kew1(Figure 1B).Although LjCYC3 was expressed at a much lower level in the dorsal petals of kew3,its expression level was much higher in the ventralized lateral petals of kew3 in comparison to its wild-type counterparts(Figure 1B).When the transcript of LjCYC3 from the kew3 mutant was sequenced,a single base deletion was detected at the region encoding the TCP domain(corresponding to position 485 of the LjCYC3 open reading frame(ORF),Figure 1E),which should cause a frame-shift and an abnormal transcript,giving rise to an abnormal product with a truncated TCP domain.Thus,we conclude that kew3 is a mutant allele of LjCYC3.

To characterize the mutation in kew1,Southern blot hybridization was conducted using LjCYC3 ORF as the probe,and restricted fragment length polymorphisms(RFLP)analysis was performed with different restriction endonucleases.An RFLP was detected in kew1 when the restriction enzyme MspI was used(Figure 1C),suggesting a possible rearrangement of the DNA sequence around the LjCYC3 locus.To confirm this rearrangement,inverse PCR(IPCR)was conducted to amplify the self-ligation product of MspI-digested fragments prepared from kew1 genomic DNA,using two pairs of nested primers complementary to the LjCYC3 sequence.An IPCR product with an approximate size of 4 kb was identified in kew1.Sequence analysis of the IPCR product revealed that in kew1,the downstream region of the LjCYC3 locus is linked to a region which should be approximately 144 kb upstream of the promoter region of the LjCYC3 locus in the wild type,confirming that a large DNA fragment inversion occurred in the kew1 mutant(Figure 1D).In order to clarify the breakpoints of the inversion in kew1,“breakpoint-specific”P(pán)CR primers were designed,and the expected PCR products were obtained when the genomic DNA from either kew1 or the wild type was used as the templates(Figure 1D,F).Sequence analysis revealed that the inversion event resulted in deletion of only a few base pairs at the breakpoints of inversion(data not shown).Quantitative RT-PCR was performed to detect the expression level of a dozen genes flanking the LjCYC3 locus or near the breakpoints in the dorsal and lateral petals of the wild type and kew1.No difference in the expression of these genes was observed(data not shown),suggesting that the inversion could specifically affect the expression of LjCYC3 in kew1.Thus,we conclude that kew1 is an allelic mutant of LjCYC3,caused by a large inversion of the fragment containing the locus.

Sequence variation of LjCYC genes at the 3′end of ORFs

The three LjCYC genes,LjCYC1,LjCYC2,and LjCYC3,possess a similar gene structure,all containing two exons and one intron.The conserved TCP domain and the R domain(with rich arginine(R)amino acid residues)can be found near their N-terminus and C-terminus,respectively(Figure 2A).In both LjCYC1 and LjCYC2,the coding sequence(CDS)is extended into the second exon,while in LjCYC3,the stop codon is at the end of the first exon.The similarity in gene structure was further investigated by sequence comparison at the DNA level.It was found that the non-coding sequence of the second exon of LjCYC3 matched well with the coding sequence of the second exon of LjCYC1 and LjCYC2(Figure 2B),indicating that LjCYC3,as a descendent of the LjCYC genes,should lose its coding sequence in the second exon.It is interesting to note that squ1,a mutant of LjCYC2,is caused by a point mutation,which causes the loss of its coding sequence at the second exon and its failure to confer dorsal identity during zygomorphic floral development(Figure 2B,Feng et al.2006).These data indicate that the sequence variation at the 3′end of the CDS plays an important role in the diverse functions and activities of LjCYC genes.

Diverse functions of LjCYC genes in the control of DV identities

A transgenic experiment on L.japonicus was conducted to compare the function of different LjCYC genes.A transgenic construct,35S::LjCYC1-ox,was transformed into the wild-type plant,in which the ORF of LjCYC1 was tagged with GFP(Green Fluorescent Protein)and was under the control of 35S constitutive promoter.The transgenic flowers of 35S::LjCYC1-ox with the dominant phenotype looked similar to those of 35S::LjCYC2-ox which were obtained in a previous study(Feng et al.2006).In both 35S::LjCYC1-ox and 35S::LjCYC2-ox transgenic flowers,all petals possessed the appearance of dorsal petals in the wild type;in other words,the petals in the lateral and ventral positions became dorsalized and resembled the dorsal petals of the wild type in shape and size(Figure 3A).These data demonstrate that both LjCYC1 and LjCYC2 have similar functions in determining dorsal identity during zygomorphic floral development.

Figure 2.Gene structure of LjCYC genes and sequence variation in the 3′end of their open reading frames(ORFs).(A)Schematic diagram of the gene structure of LjCYC genes:exons(the rectangles)and introns(the wave lines)of the three LjCYC genes are indicated.The coding sequences in exons are indicated in black and the non-coding ones in white.The TCP and R domains are highlighted in red and blue,respectively.(B)Comparison of the 3′ends of the ORFs of LjCYC genes.The consensus nucleotides among LjCYC genes are highlighted.The red triangles indicate the position of the splicing sites of introns.Red boxes indicate the stop codon of the LjCYC genes.The C-terminal of LjCYC2 is lost in the squ mutant(the blue star).

Figure 3.Transgenic flowers of LjCYC genes.(A)Flowers of wild type,LjCYC1-ox,and LjCYC2-ox transgenic plants.DP,dorsal petal;LP,lateral petal;VP,ventral petal.All petals in the flowers of LjCYC1-ox and LjCYC2-ox transgenic plants become dorsalized.Bar=1 cm.(B)Flowers of Gifu,kew1,and LjCYC3-frag and LjCYC3-ox plants.DP,dorsal petal;FF,front view of flowers;FS,side view of flowers;LP,lateral petal;VP,ventral petal.LjCYC3-frag and LjCYC3-ox confer lateral identity to the lateral and ventral petals in kew1 transgenic plants.The shape of the lateralized petals in LjCYC3-ox transgenic plants is different from that of those in LjCYC3-frag.Bar=0.5 cm.(C)Expression levels of LjCYC3 in different petals of Gifu,kew1,and LjCYC3 complement and overexpression plants.

Two constructs,LjCYC3-frag13.3 and 35S::LjCYC3-ox,were transformed to kew1 plants.LjCYC3-frag13.3 is a complementary construct and contains a 13.3 kb genomic fragment with a full LjCYC3 locus(beginning from 7.5 kb upstream of the translation start site and ending 4.7 kb downstream of the stop codon of LjCYC3).In 35S::LjCYC3-ox,the ORF of LjCYC3 was tagged with GFP and was under the control of the 35S constitutive promoter.The transgenic plants of both LjCYC3-frag13.3 and 35S::LjCYC3-ox displayed similar phenotypes:the dorsal petals were the same as those of the wild type in shape and size,but the petals in the lateral and ventral positions resembled the lateral ones in the wild type in shape and size to a certain extent(Figure 3B).However,there was a difference between these two types of transgenic flowers:the lateralized petals of 35S::LjCYC3-ox exhibited a deformed shape,while those of LjCYC3-frag13.3 mimicked the lateral petals of the wild type perfectly(Figure 3B).Thus,LjCYC3 alone should be sufficient to determine lateral identity,confirming the diverse functions of the LjCYC genes in controlling dorsal and lateral identities.Furthermore,dorsal identity should be epistatic to lateral identity,because the ectopic expression of LjCYC3 is not able to compromise the development of dorsal petals.

The expression level of LjCYC3 in the petals of transgenic flowers was measured and compared with those in the wild type and in kew1.It was found that 35S::LjCYC3-ox had a much higher expression level of LjCYC3 than the wild type and kew1 in all petals(Figure 3C).However,LjCYC3-flag13.3 had a similar expression level of LjCYC3 in both lateral and ventral petals(Figure 3C),which is consistent with the lateralized petals appearing in both lateral and ventral positions.Therefore,the cis-elements in the regulatory region of the LjCYC3 locus within the 13.3 kb fragment are not sufficient to determine an asymmetrical expression pattern.

In vitro assay of the transcriptional activity of LjCYC proteins

An in vitro assay was conducted to analyze the transcriptional activities of different LjCYC proteins.The GAL4 binding domain(BD)was fused with different LjCYC proteins,and the expression level of the lacZ reporter gene was measured.A transcriptional activation assay indicated that the LjCYC3 fusion protein causes high expression of the reporter gene(Figure 4A),indicating that LjCYC3 has strong transcriptional activity.Conversely,the LjCYC2 and LjCYC1 fusion proteins with BD exhibited very weak activity in the assay(Figure 4A).

The regions containing transcriptional activation domains(AD)in LjCYC3 were analyzed.It was found that both the TCP and R domains did not have activation activity(Figure 4B),indicating that the TCP and R domains are not essential for activation activity in the assay.However,the N-terminus from amino acids 1 to 113,as well as a region in the C-terminus between amino acids 250 and 370,possessed strong activation activity(Figure 4B).Therefore,LjCYC3 should have two separate transcriptional AD separated by TCP and R domains.

It has been previously shown that the TCP transcription factors tend to form a dimer(Kosugi and Ohashi 2002).To test the possible interaction between LjCYC proteins,LjCYC2 was used as bait in a yeast two-hybrid assay.The interactions between LjCYC2 and LjCYC2,LjCYC2 and LjCYC1,and LjCYC2 and Lj-CYC3 were detected,respectively(Figure 4C).To test whether the activation activity of LjCYC3 could be affected when Lj-CYC3 formed a heterodimmer with LjCYC2,the LjCYC fusion proteins were constructed in the co-expression vector pBridge:the LjCYC was fused with either the BD or NLS(the nuclear localization signal).It was found that the transcriptional activity of BD-LjCYC3 was reduced dramatically when co-expressed with NLS-LjCYC2,while co-expression of BD-LjCYC2 and NLS-LjCYC3 did not exhibit any activation activity of LjCYC3(Figure 4D).These data suggest that LjCYC2 has a negative effect on the activation activity of LjCYC3.

There are two possible explanations for the negative effect of LjCYC2 on the activity of LjCYC3:LjCYC2 may exhibit activity as a transcription repressor,and/or direct physical interaction could block LjCYC2 from the LjCYC3 AD.A yeast transcription repression assay was conducted.When introduced to the test promoter in yeast strain 122 which exhibits strong LacZ expression,the LjCYC2 fusion protein did not obviously reduce reporter gene expression when compared with an empty vector alone(Figure 4E),indicating that LjCYC2 itself is not a direct transcription repressor.

Discussion

In legumes,it has been predicted that a series of gene duplication events had resulted in a number of CYC-like TCP genes which may have different functions(Citerne et al.2003).In L.japonicus,it has been found that LjCYC2 exhibits dorsalizing activity to determine the dorsal identity of petals during zygomorphic floral development(Feng et al.2006).In this study,we demonstrate that different members of the LjCYC genes have acquired diverse functions in the control of dorsal and lateral identities.When ectopically expressed,LjCYC1 was shown to play a role in determining dorsal identity,similar to LjCYC2.In the previous study,knockdown of LjCYC1 expression through introduction of an RNAi construct(Wang et al.2010b)resulted in no detectable phenotype in the transgenic plant L.japonicus.Therefore,it is likely that LjCYC1 has a redundant function to determine dorsal identity.This is similar to the situation in Antirrhinum where DICH’s activity is redundant to CYC’s,and their interaction is essential for the control of the dorsal and lateral identities(Luo et al.1996,1999).However,in L.japonicus,the activity of LjCYC2 as well as the redundant activity of LjCYC1,may be restricted in the control of dorsal identity.

Figure 4.Transcriptional activity of different LjCYC proteins.(A)LjCYC proteins possess different transactivation activity in the assay.(B)Detection of the domains in LjCYC3 responsible for transactivation activity.Top panel:schematic of LjCYC3 protein variants.The Gal4 DNA binding domain is illustrated in yellow;the TCP and R domains are highlighted in red and blue,respectively,and the other coding sequences are highlighted in black.Bottom panel:the N-and C-termini of LjCYC3 possess activity similar to the full-length one.BD,Gal4 DNA binding domain;CT,the C-terminus;FL,the full length of ORF;MD,the middle region of the ORF;NT,the N-terminus;ORF,open reading frame.(C)LjCYC2 can interact with LjCYC1,LjCYC2,or LjCYC3 in a yeast two-hybrid assay.AD,Gal4 activation domain;SD-2,SD/-Trp/-Leu;SD-4,SD/-Trp/-Leu/-His/-Ade.(D)The transactivation activity of LjCYC3 is dramatically reduced when co-expressed with LjCYC2.C2,ORF of LjCYC2;C3,ORF of LjCYC3,NLS,nuclear localization signal sequence.(E)Repression activity assay of LjCYC proteins in yeast.No repression activity of LjCYC proteins was detected.

In the previous study,it was found that LjCYC3 may be a lateralizing factor.In this study,we showed that LjCYC3 alone should be sufficient to confer lateral identity in the otherwise ventral and ventralized petals,but it failed to alter the identity of dorsal petals.These data clearly demonstrate the diverse functions of different LjCYC genes:the dorsal and lateral identities are controlled separately by different LjCYC genes.

Phylogenic analysis has revealed that three LjCYC genes originated from two events of gene duplication.The first gene duplication event separated LjCYC3 from its ancestors LjCYC1 and LjCYC2;the second event separated LjCYC1 from LjCYC2(Citerne et al.2003;Feng et al.2006).Sequence analysis revealed that LjCYC genes share a conserved gene structure.It was found that after descending from their common ancestor,a stop codon was introduced at the end of first exon of LjCYC3,and it consequently lost the coding sequence at the small second exon.Coincidently,squ,a mutant of LjCYC2,was caused by a point mutation at the beginning of the second exon,giving rise to a truncated protein and resulting in the loss of dorsalizing activity(Feng et al.2006).At the time of writing,the function of the truncated protein of LjCYC2 with the deletion of amino acids in the second exon is being analyzed in transgenic plants,and the dorsalizing activity of the protein is indeed compromised(Z.Xu et al.unpubl.data,2012).These data suggest that the sequence variation at the C-terminus may be important for the functional diversity of the LjCYC proteins.In addition,the non-coding exon maintained in LjCYC3 suggests that the 3′untranslated region of the LjCYC genes could be important in regulating their functions.

kew1 and kew3 were analyzed in this study.It was found that kew3 is caused by a 1 bp deletion in the coding region of LjCYC3,and that kew1 is caused by an inversion of a large DNA fragment containing the LjCYC3 locus.There should not be a functional LjCYC3 protein in kew3,because the one base deletion should give rise to a frame-shift and cause a putative truncated protein with the N-terminus and most parts of the TCP domain of LjCYC3(～162 amino acids).However,the truncated protein could maintain transcriptional activation activity and lose normal lateral activity.Thus,the high expression level of the abnormal transcript in the ventralized lateral petals in kew3 could be explained by the action of the abnormal kew3 protein.In the kew1 mutant,expression of LjCYC3 is dramatically reduced,and an inversion of a large DNA fragment containing the LjCYC3 locus has been identified.It is possible that the inversion in kew1 blocks the function of some trans elements which are essential for maintaining the correct expression level of LjCYC3.Indeed,there are hints as to the existence of trans elements.In a transgenic experiment,the complementary construct LjCYC3-frag13.3 containing the CDS of LjCYC3,as well as its 7.5 kb upstream and 4.5 kb downstream sequences,were transformed in kew mutant plants.The transgenic flowers bore perfect lateral petals and lateralized ventral petals(Figure 3B).These results demonstrate that LjCYC3-frag13.3 contains all the necessary information to control lateral identity,but does not contain sufficient regulatory information to regulate asymmetrical expression,indicating there should be some trans regulatory elements far away from the LCYC3 locus.

In a transgenic experiment,it was found that dorsal identity conferred by LjCYC1 or LjCYC2 should be epistatic to lateral identity conferred by LjCYC3.There is some evidence for the molecular basis of epistasis.It has been shown that the TCP domain recognizes some consensus DNA elements and the consensus motifs could be different for class I and II TCP proteins(Costa et al.2005;Schommer et al.2008).Furthermore,TCP proteins have been shown to form homoand heterodimers,which may be essential for their binding to DNA(Kosugi and Ohashi 2002;Aggarwal et al.2010).In this study,we found that LjCYC proteins possess different activities for transcriptional activation in an in vitro assay.Furthermore,when co-expressed with LjCYC2,LjCYC3 activity for transcriptional activation is dramatically decreased.These data suggest that the activity of LjCYC proteins could be altered by the interaction between these proteins.It is possible that an epistatic effect of dorsal activity could occur from the direct interaction of LjCYC proteins through the formation of heterodimers.In the lateral petals of the wild-type plants,only LjCYC3 is expressed,and the lateral identity is then determined by LjCYC3;in the dorsal petals of the wild-type plant and in the transgenic petals with ectopic expression of LjCYC2,the activity of LjCYC3 is inhibited by LjCYC2 and the lateral identity cannot be established.Furthermore,it has recently been shown that auto-and cross-regulation occur among CYC-like genes in Primulina heterotricha(Yang et al.2012).The sequences of LjCYC loci have been analyzed,and the binding motif(GGNCCC,Costa et al.2005)for CYC protein can be found in the 5′regulatory regions of LjCYC2 and LjCYC3,but not in that of LjCYC1(S.Xu et al.unpubl.data,2012),suggesting that there could be a self-regulatory mechanism for the expression of LjCYC2 and LjCYC3.However,more experiments–especially in vivo or in planta–should be conducted to confirm the interaction among LjCYC genes in both expression and protein levels.

Taken together,our study demonstrates that the LjCYC genes possess divergent functions in the control of dorsal or lateral activities.The sequence variation and the subsequent alteration of protein property may play important roles in the genes’functional diversity.Our data also suggests that the interaction of LjCYC genes may be important in the control of the zygomorphic flower development in L.japonicus.

CYC-like TCP genes have been independently recruited in diverse angiosperm lineages which have zygomorphic flowers(Preston and Hileman 2009;Jabbour et al.2009).In some of zygomorphic flowers plants,different CYC-like copies have unique expression pattern similar to LjCYC genes in Lotus japonicus,such as BcCYC2A and BcCYC2B in Byrsonima crassifolia(Zhang et al.2010),DipsCYC2A and DipsCYC2B in Lonicera(Howarth et al.2011),PhCYC1D and PhCYC1C in Primulina heterotricha(Gao et al.2008,Yang et al.2012).Similar to a recent report on subfunctionalization of different CYC-like genes in sunflowers(Chapman et al.2012),our study provides insight into detecting the functional diversity of CYC-like TCP genes and the interaction among them in more zygomorphic flower plants,which may help us better understand the evolution of CYC-like TCP genes in the context of floral zygomorphy.

Materials and Methods

IPCR and cloning

The breakpoint was subcloned using the following technique.Genomic DNA from kew1 was digested with MspI.The restricted DNA fragments were purified and then self-ligated with T4 DNA ligase.The IPCR was conducted using two primer pairs,P1/P2 and P3/P4(P1,5′-TGCTCTCTCCCTTGA CTCCTTC-3′;P2,5′-AAGAAACACCAAGACCACAATATATAC-3′;P3,5′-GGTGAGGATAGGGTTGAAGAAGC-3′;and P4,5′-TAATAGATAGAAACTTGGGACTAACCC-3′).After the sequence of one breakpoint junction was obtained,two pairs of breakpoint primers α/β and γ/ε were designed to confirm and obtain sequences of the breakpoints of the inversion(α,5′-GCTGTGGCATAAAGACCAGATTC-3′; β,5′-TCCCGTTCCG CTCTGACGC-3′;and γ,5′-GAACAACACCAGTTAAAACTAA CCG-3′;ε,5′-GTATTCCTGGGGTGTTTGAGATC-3′).

Generation of transgenic lines and plant growth conditions

The transformation vectors pCAMBIA 1302 and pCAMBIA 1300 were used to generate the transgenic constructs.Plant transformation was performed as described by Feng et al.(2006).L.japonicus ecotype Gifu B129 as the wild type and kew1 mutants were used for plant transformation.All plants were grown at 20–22°C with a 16 h light/8 h dark photoperiod at 150 mE·m-2·s-1.

RNA extraction and RT-PCR

Petals were dissected from 5 mm floral buds.Total RNA was extracted from petals using Plant RNA Reagent(Invitrogen).cDNA was synthesized using a First Strand cDNA Synthesis Kit(Ferments).For real-time RT-PCR analysis,the SYBR Premix Ex Taq(Perfect Real Time;TaKaRa)and Rotor-Gene Multifilter Real-time Cycler(Corbett Research)were used according to the manufacturers’instructions.A constitutively-expressed gene in Lotus japonicus,polyubiquitin(LjUBI),was used as a control to normalize the expression data(Wandrey et al.2004).The gene-specific primers for real-time RT-PCR were:LjCYC3,5′-GCTTCTTCAACCCTATCCTCACC-3′,5′-CAATCACATTA AACCCATCTCCAC-3′;LjUBI,5′-TTCACCTTGTGCTCCGT CTTC-3′,5′-AACAACAGCACACACAGACAATCC-3′.

Yeast assays

A transcriptional transactivation assay,a yeast two-hybrid assay,and a transcriptional repression assay were conducted mainly using the Matchmaker System(Clontech).The yeast strains AH109(Clontech)and 122(Saha et al.1993)were used for transformation.The yeast vectors used in the experiments were pGBKT7,pGADT7,pBridge(Clontech),and pMA424(Ma and Ptashne 1987).β-Galactosidase activity was measured according to the manufacturer’s instructions.

Full-length and partial coding sequences of LjCYC genes were amplified by PCR and then fused with the GAL4 BD in pGBKT7 or pMA424 to perform transcription activation and repression assays,respectively.To detect the protein–protein interaction between LjCYC2 and other LjCYC proteins,different LjCYC proteins were fused with the GAL4 AD in pGADT7.

Full-length LjCYC ORFs were cloned using the following primers:for ORF of LjCYC1,5′-GAATTCATGTTCCCTAATTC CAGTTAC-3′,5′-AGATCTTAGTGTAGATGAGGACTGGTG-3′;for ORF of LjCYC2,5′-GAATTCATGTTCCCTTTCAGCTCA AG-3′,5′-GGATCCTAGTGTGGACGTGGATTTGTG-3′;and for ORF of LjCYC3,5′-GAATTCATGTTCTCTTCCACCT CATATC-3′,5′-AGATCTTACCACTGGTTCATGAAACATG-3′.

To obtain variants of the LjCYC3 proteins,different fragments of LjCYC3 ORFs were cloned using the following primers:the N-terminus region,5′-GGAATTCATGTTCTCTTC CACCTCATATC-3′,5′-TTATGTGGTAGCCATGGCTGCAG-3′;the TCP domain,5′-GGAATTCAAGAAAGACAGGCACAG CAAG-3′,5′-TTAAAGCTCTTGGATTGCACTCTCAG-3′;the middle region between the TCP and R domains(MD),5′-GGAATTCGCAAGGAGTAAGAACTGTTC-3′,5′-TTACTTATT CTGAAGAACACAAACATC-3′;the MD plus R domain,5′-GGAATTCGCAAGGAGTAAGAACTGTTC-3′,5′-TTACTTGTA ACAGGTTCTTTCTCTAGC-3′;and the C-terminus region,5′-GGAATTCATGTTGGAAGATCAAAGATGC-3′,5′-TTACCA CTGGTTCATGAAACATG-3′.

To detect the effect of protein–protein interaction between LjCYC2 and LjCYC3,LjCYC2 and LjCYC3 were co-expressed in the pBridge vector.The ORFs of LjCYC2 and LjCYC3 were separately cloned into one of the two multiple cloning sites in pBridge,and fused with either BD or the second NLS.Two combinations of fusion proteins were made:LjCYC2-BD/LjCYC3-NLS and LjCYC3-BD/LjCYC2-NLS.

Acknowledgements

This work was supported by the National Natural Science Foundation of China(30930009),and the Ministry of Agriculture of China for Transgenic Research(2011ZX08009-003 and 2009ZX08009-112).

Aggarwal P,Das Gupta M,Joseph AP,Chatterjee N,Srinivasan N,Nath U(2010)Identification of specific DNA binding residues in the TCP family of transcription factors in Arabidopsis.Plant Cell 22,1174–1189.

Chapman MA,Tang S,Draeger D,Nambeesan S,Shaffer H,Barb JG,Knapp SJ,Burke JM(2012)Genetic analysis of floral symmetry in Van Gogh’s sunflowers reveals independent recruitment of CYCLOIDEA genes in the Asteraceae.PloS Genet.8,1–10.

Citerne H,Luo D,Pennington RT,Coen E,Cronk QC(2003)A phylogenomic investigation of CYCLOIDEA-like TCP genes in the Leguminosae.Plant Physiol.131,1042–1053.

Costa MM,Fox S,Hanna AI,Baxter C,Coen E(2005)Evolution of regulatory interactions controlling floral asymmetry.Development 132,5093–5110.

Crawford BC,Nath U,Carpenter R,Coen ES(2004)CINCINNATA controls both cell differentiation and growth in petal lobes and leaves of Antirrhinum.Plant Physiol Add.135,244–253.

Cubas P(2004)Floral zygomorphy,the recurring evolution of a successful trait.BioEssays 26,1175–1184.

Cubas P,Lauter N,Doebley J,Coen E(1999)The TCP domain:A motif found in proteins regulating plant growth and development.Plant J.18,215–222.

Doebley J,Stec A,Gustus C(1995)teosinte branched1 and the origin of maize:Evidence for epistasis and the evolution of dominance.Genetics 141,333–346.

Doebley J,Stec A,Hubbard L(1997)The evolution of apical dominance in maize.Nature 386,485–488.

Endress PK(1999)Symmetry in Flowers:Diversity and evolution.Int.J.Plant Sci.160,3–23.

Endress PK(2001)Evolution of floral symmetry.Curr.Opin.Plant Biol.4,86–91.

Feng X,Zhao Z,Tian Z,Xu S,Luo Y,Cai Z,Wang Y,Yang J,Wang Z,Weng L,Chen J,Zheng L,Guo X,Luo J,Sato S,Tabata S,Ma W,Cao X,Hu X,Sun C,Luo D(2006)Control of petal shape and floral zygomorphy in Lotus japonicus.Proc.Nat.Acad.Sci.USA 103,4970–4975.

Gao Q,Tao J,Yan D,Wang Y,Li Z(2008)Expression differentiation of CYC-like floral symmetry genes correlated with their protein sequence divergence in Chirita heterotricha(Gesneriaceae).Dev.Genes Evol.218,341–351.

Howarth DG,Martins T,Chimney E,Donoghue MJ(2011)Diversification of CYCLOIDEA expression in the evolution of bilateral flower symmetry in Caprifoliaceae and Lonicera(Dipsacales).Ann.Bot.107,1521–1532.

Jabbour F,Nadot S,Damerval C(2009)Evolution of floral symmetry:A state of the art.C.R.Biol.332,219–231.

Kosugi S,Ohashi Y(1997)PCF1 and PCF2 specifically bind to cis elements in the rice proliferating cell nuclear antigen gene.Plant Cell 9,1607–1619.

Kosugi S,Ohashi Y(2002)DNA binding and dimerization specificity and potential targets for the TCP protein family.Plant J.30,337–348.

Li C,Potuschak T,Colon-Carmona A,Gutierrez RA,Doerner P(2005)Arabidopsis TCP20 links regulation of growth and cell division control pathways.Proc.Natl.Acad.Sci.USA 102,12978–12983.

Luo D,Carpenter R,Vincent C,Copsey L,Coen E(1996)Origin of floral asymmetry in Antirrhinum.Nature 383,794–799.

Luo D,Carpenter R,Copsey L,Vincent C,Clark J,Coen E(1999)Control of organ asymmetry in flowers of Antirrhinum.Cell 99,367–376.

Ma J,Ptashne M(1987)A new class of yeast transcriptional activators.Cell 51,113–119.

Martín-Trillo M,Cubas P(2010)TCP genes:A family snapshot ten years later.Plant Cell 15,31–39.

Masuda HP,Cabral LM,de Veylder L,Tanurdzic M,de Almeida EJ,Geelen D,Inzé D,Martienssen RA,Ferreira PC,Hemerly AS(2008)ABAP1 is a novel plant Armadillo BTB protein involved in DNA replication and transcription.EMBO J.27,2746–756.

Nath U,Crawford BC,Carpenter R,Coen E(2003)Genetic control of surface curvature.Science 299,1404–1407.

Preston JC,Hileman LC(2009)Developmental genetics of floral symmetry evolution.Trends Plant Sci.14,147–154.

Pruneda-Paz JL,Breton G,Para A,Kay SA(2009)A functional genomics approach reveals CHE as a component of the Arabidopsis circadian clock.Science 323,1481–1485.

Saha S,Brickman JM,Lehming N,Ptashne M(1993)New eukaryotic transcriptional repressors.Nature 363,648–652.

Schommer C,Palatnik JF,Aggarwal P,Chételat A,Cubas P,Farmer EE,Nath U,Weigel D(2008)Control of jasmonate biosynthesis and senescence by miR319 targets.PLoS Biol.6,1991–2001.

Wandrey M,Trevaskis B,Brewin N,Udvardi MK(2004)Molecular and cell biology of a family of voltage-dependent anion channel porins in Lotus japonicus.Plant Physiol.134,182–193.

Wang J,Tian Z,Zhao Z,Luo D,Hu X(2010a)Mutations affecting lateral petal development in Lotus japonicus.Mol.Plant Breed.8,6–10.

Wang J,Wang Y,Luo D(2010b)LjCYC genes constitute floral dorsoventral asymmetry in Lotus japonicus.J.Integr.Plant Biol.52,959–970.

Wang Z,Luo Y,Li X,Wang L,Xu S,Yang J,Weng L,Sato S,Tabata S,Ambrose M,Rameau C,Feng X,Hu X,Luo D(2008)Genetic control of floral zygomorphy in pea(Pisum sativum L.).Proc.Natl.Acad.Sci.USA 105,10414–10419.

Weng L,Tian Z,Feng X,Li X,Xu S,Hu X,Luo D,Yang J(2011)Petal development in Lotus japonicus.J.Integr.Plant Biol.53,770–782.

Yang X,Pang H,Liu B,Qiu Z,Gao Q,Wei L,Dong Y,Wang Y(2012)Evolution of double positive autoregulatory feedback loops in CYCLOIDEA2 clade genes is associated with the origin of floral zygomorphy.Plant Cell 24,1834–1847.

Zhang W,Kramer EM,Davis CC(2010)Floral symmetry genes and the origin and maintenance of zygomorphy in a plantpollinator mutualism.Proc.Natl.Acad.Sci.USA 107,6388–6393.