• We are sorry, but NCBI web applications do not support your browser and may not function properly. More information
Logo of pnasPNASInfo for AuthorsSubscriptionsAboutThis Article
Proc Natl Acad Sci U S A. Mar 28, 2006; 103(13): 4970–4975.
Published online Mar 20, 2006. doi:  10.1073/pnas.0600681103
PMCID: PMC1458779
From the Cover
Developmental Biology

Control of petal shape and floral zygomorphy in Lotus japonicus


Zygomorphic flowers, with bilateral (dorsoventral) symmetry, are considered to have evolved several times independently in flowering plants. In Antirrhinum majus, floral dorsoventral symmetry depends on the activity of two TCP-box genes, CYCLOIDEA (CYC) and DICHOTOMA (DICH). To examine whether the same molecular mechanism of floral asymmetry operates in the distantly related Rosid clade of eudicots, in which asymmetric flowers are thought to have evolved independently, we investigated the function of a CYC homologue LjCYC2 in a papilionoid legume, Lotus japonicus. We showed a role for LjCYC2 in establishing dorsal identity by altering its expression in transgenic plants and analyzing its mutant allele squared standard 1 (squ1). Furthermore, we identified a lateralizing factor, Keeled wings in Lotus 1 (Kew1), which plays a key role in the control of lateral petal identity, and found LjCYC2 interacted with Kew1, resulting in a double mutant that bore all petals with ventralized identity to some extents. Thus, we demonstrate that CYC homologues have been independently recruited as determinants of petal identities along the dorsoventral axis in two distant lineages of flowering plants, suggesting a common molecular origin for the mechanisms controlling floral zygomorphy.

Keywords: dorsoventral axis, floral development, keeled wings in Lotus, LjCYC2, squared standard

Floral zygomorphy (dorsoventral asymmetry) is an evolutionary adaptation that facilitates outcrossing by attracting pollinators (17). The phenomenal diversity in Leguminosae (Rosid clade of eudicots), the third largest family of flowering plants with ≈20,000 species (8), is often explained by successful coevolution with pollinators. In the subfamily Papilionoideae (12,000 spp.) to which Lotus japonicus (9, 10) belongs, most species have specialized zygomorphic “pea” flowers with three types of petals, which are arranged along a dorsoventral axis: a single bilaterally symmetrical petal (standard) in the dorsal position, two asymmetric lateral petals (“wings”) and two asymmetric ventral petals which form a “keel” (Fig. 1A and B). In contrast, Antirrhinum majus, a well studied member in Lamiales (Asterid clade) (11, 12), possesses two asymmetrical dorsal petals, two asymmetrical lateral petals, and a single bilaterally symmetrical ventral petal (Fig. 1 C and D). It is believed that the zygomorphy of Leguminosae has evolved separately from the Lamiales (6, 13). Although the different internal symmetries of the counterpart petals in L. japonicus and A. majus are consistent with the independent evolution of zygomorphy in these two lineages, both are a response to a dorsoventral axis, suggesting that there could be a divergence in the make up of the mechanism to determine the axis in the control of zygomorphic developments in the two distantly related species.

Fig. 1.
Development and comparison of zygomorphic flowers. Zygomorphic flowers and petals of L. japonicus (A and B) and Antirrhinum (C and D) are shown. In the case of Antirrhinum, only the lobes of petals are shown. Lines with arrow show the direction of floral ...

In A. majus, the development of zygomorphic flower requires the activity of two TCP-box genes (14), CYC and DICH (11, 12), whose function is mediated through an interaction with some specific MYB genes (1517). When both CYC and DICH are mutated, more petals and stamens are developed in the dorsal region, and all petals resemble the shape of ventral petal. Thus, CYC and DICH could have a dual role in the control of zygomorphic development: an early one affecting primordium initiation and controlling floral asymmetry, and a later one affecting organ asymmetry and other morphological characters (11, 12). It has been further proposed that the petal asymmetry arises through a series of steps in which the expression patterns of CYC and DICH are progressively elaborated and maintained (12), and thus development of floral and petal asymmetries are closely related. Because TCP-box genes have been found widely in higher plant genomes (7, 14), it is possible to test whether CYC like genes are involved in the development of zygomorphy in different species. In some close relatives of Antirrhinum, several CYC orthologues have been found to play a similar role in the control of floral symmetry (18, 19), but other studies suggest that the bilaterally symmetrical flowers in some families of the same Asterid clade might not require orthologues or functional analogues of CYC or DICH (6, 20). So far, because of lack of an amenable experimental system, most studies have relied on identification of CYC homologues and investigation of their expression patterns, and little robust functional analysis in the distantly related species has been reported.

In legumes, a number of CYC-like genes have been isolated and found to have undergone repeated duplication events, suggesting that they might have divergent functions (21, 22). To investigate the molecular mechanisms underlying the development of different zygomorphic flowers among angiosperms, we explored a model legume, Lotus japonicus (9, 10), and examined whether the CYC homologues also play a role in the control of floral asymmetry in this papilionoid legume. We demonstrated that a CYC-like gene in L. japonicus, LjCYC2, has a dorsalizing activity during petal development, similar to CYC and DICH, even though the zygomorphy of Leguminosae is believed to have evolved separately from the Lamiales (1, 6, 7, 13). Furthermore, we identified a lateralizing factor, keeled wings in Lotus 1 (Kew1), which plays a key role in the control of lateral petal identity, and found that LjCYC2 interacted with Kew1, resulting in a double mutant bearing all petals with ventralized identity to a different extent. Our data show that CYC homologues have been independently recruited to the control of floral zygomorphy in distantly related lineages, suggesting that their dorsal activity should be a primary and synapomorphic function during the evolution of dorsoventral specification.


Inflorescence and Floral Development in L. japonicus.

L. japonicus (Gifu B-129) (9) is a perennial herb. When starting reproductive growth, each shoot becomes a primary inflorescence (I1), and produces a secondary inflorescence meristem (I2) at each node. Floral meristems are generated from I2. We divided I2 development into seven stages, from I2-1 to I2-7 (Fig. 1 EK). One I2 meristem initiates from the axil of the compound leaf at each node (I2-1 and I2-2, Fig. 1 E and F). Bract primordia initiate at I2-3 (Fig. 1G) and become visible at I2-4 when the meristem of I2 ceases activity and becomes compressed (Fig. 1H). Then, normally three floral meristems are partitioned sequentially from I2 (Fig. 1 I and J) and finally I2 differentiates, becoming covered by trichomes (Fig. 1K). The development of floral dorsoventral asymmetry can be observed at the very beginning of floral development when the floral organs initiate in a unidirectional order (23), and the dorsoventral axis becomes evident (Fig. 1K): organ primordia in the dorsal region develop more slowly than those in the ventral region. During late floral development, three types of petals develop along the dorsoventral axis, and a key morphological landmark is the differentiation of the vascular veins in petals. The various shapes of different petals are distinctive (Fig. 1L) when the primary veins initiate in all petals. Then, the characteristic shapes of dorsal, lateral, and ventral petals become visible (Fig. 1M) when the initiation of minor veins and freely ending veinlets begins and the differentiation of epidermal cell types commences. Further elaboration of petal structures continues, giving rise to mature petals with blades and claws in dorsal petals and blades and stipes in lateral and ventral petals (Fig. 1N), whereas various representative cell types are generated in the epidermal layers of different petals (see Fig. 4 HJ).

Fig. 4.
Flowers of the wild-type (L. japonicus, Gifu B129), transgenic, and mutant plants. (AG) Front and side views of flowers and different petals are shown. (Scale bar, 0.5 cm.) (A) Wild-type flower. Arrow indicates that two ventral petals form a ...

CYC Homologues in L. japonicus.

Four TCP-box genes, LjCYC1, LjCYC2, LjCYC3, and LjCYC5 (GenBank accession nos. DQ202475, DQ202476, DQ202477, and DQ202478), were isolated from L. japonicus. Phylogenetic analysis placed three of them, LjCYC1, LjCYC2, and LjCYC3, in a close clade including CYC and DICH (Fig. 2D), which correspond to the LEGCYC I (LjCYC1 and LjCYC2), and LEGCYC II (LjCYC3) genes, respectively (21, 22). RNA in situ hybridizations were conducted to analyze their expression patterns. It was found that LjCYC1 was expressed in I1 and dorsal regions of floral primordia. Expression of LjCYC3 could not be detected by RNA in situ hybridization, and transcripts of LjCYC5 were located in young leaves and at the base of I2 (data not shown). Therefore, we focus on LjCYC2, which was the most similar to CYC in expression pattern (Figs. 2A and and33CF). However, in contrast to the expression pattern of CYC, which was only observed in the dorsal region of floral meristem and dorsal floral organs, the earliest expression of LjCYC2 was found at the margin between I1 and I2 (Fig. 3A), and also became detectable in the center of the I2 meristem at I2-4, when it began to differentiate (Figs. 1H and and33B). Later, LjCYC2 was expressed in the dorsal region of floral meristems (Fig. 3 C and D), and then persisted in the dorsal organs of developing flowers, including the dorsal sepal, petal and stamen primordia (Fig. 3 E and F). The asymmetrical expression pattern of LjCYC2 in the floral meristem indicates that it could play a similar role to CYC in the control of floral dorsoventral development in L. japonicus, whereas its expression in the developing inflorescence meristem suggests a multifunction role during floral development.

Fig. 2.
Molecular and phylogenetic analysis of LjCYC2 and identification of a mutant allele. (A) Alignment of putative protein sequences of LjCYC2 and CYC. TCP domain is underlined with straight lines and R domain is underlined with waved lines (7). Triangle ...
Fig. 3.
Expression pattern of LjCYC2. (AF) Longitudinal and transverse sections through the main apex of wild type were hybridized with LjCYC2 antisense RNA probes labeled with digoxigenin-UTP. The transcript-specific hybridization signal is visualized ...

Effects of Altered Expression of LjCYC2 in Transgenic L. japonicus.

To test its function, the effects of reduced or constitutive expression of LjCYC2 were investigated in transgenic plants (Fig. 5 and Supporting Text, which is published as supporting information on the PNAS web site). All transgenic plants expressing LjCYC2 constitutively showed a specific effect in the lateral and ventral petals. They displayed abnormal petal shapes and cell types in the adaxial epidermal layer to varying extents: the lateral petals became more symmetrical and more like dorsal petals, possessing dorsal-like conical cells mixed with jigsaw puzzle-shaped; the ventral petals were also more symmetrical, displaying patches of abnormal conical cells, whereas the dorsal petal maintained the wild-ype shape and cell type (Fig. 4B and HJ). This finding indicates that the ectopic expression of LjCYC2 can confer a dorsalizing effect. This effect was greatly enhanced in the GFP tagged LjCYC2 line (SH0578), with all of the petals fully dorsalized and the normal dorsoventral asymmetry abolished (Fig. 4 G and WY). On the other hand, the antisense transgenic plants often displayed a phenotype with abnormal lateral petals which resembled the ventral in both petal shape (Fig. 4C) and epidermal cell type (data not shown), suggesting that reduced LjCYC2 activity leads ventralization. Therefore, the dorsalizing effect of LjCYC2 on petal identity was revealed in both sense and antisense transgenic plants. However, altered LjCYC2 activity also gave rise to other morphological alterations, such as altered inflorescence and floral structures (Fig. 5B and C). In sense transgenic lines, I1 terminated with shortened internodes was observed frequently (Fig. 5B), and in primary antisense transgenic plants, two flowers were frequently fused to various extents (Fig. 5C), although the variable phenotypes were weakened and vanished in succeeding generations. These data indicate that the precise expression pattern of LjCYC2 is important for floral development, suggesting other roles for LjCYC2, which could be distinct from its dorsalizing function.

Fig. 5.
Comparison of primary inflorescence and secondary inflorescence (I2) between wild type and transgenic plants of L. japonicus. (A) In wild type, I2 develops from the axil of the compound leaf at each node (arrows) and usually produces two flowers with ...

Identification of an LjCYC2 Mutant Allele, squ1.

A large scale mutagenesis experiment was conducted in L. japonicus (a total of ≈50,000 M2 families) to screen for mutations with abnormal floral symmetry. Several single recessive mutants with abnormal petal shape were obtained. One mutant was named squared standard 1 for the abnormal shape of the dorsal petal (Fig. 4D). squ1 seems to specifically affect the development of dorsal petal without other notable phenotypes, and its effect on petal development begins during initiation of the primary vein in the petal (data not shown). Genetic analysis showed that squ1 co-segregated with a molecular marker for the LjCYC2 locus in chromosome 2. Sequencing data confirmed that squ1 carried a point mutation at the splicing site of the intron in LjCYC2 (Fig. 2B), which gave rise to an abnormal transcript with lower expression level (Fig. 2C). squ1 has the capacity to encode a protein, in which the seventeen C-terminal amino acids are replaced by five others (Fig. 2C). This finding suggests that squ1 is a mutant allele of LjCYC2. Apart from its phenotype on the shape of dorsal petals, squ1 also has a specific effect on the epidermal cell shape in the dorsal petal. The cells do not all have the typical conical morphology, but include jigsaw puzzle-shaped cells characteristic of the lateral petals in the wild type (Fig. 4 NP), indicating a correlation between the abnormal shape of squ1 dorsal petal and the reduction of dorsal identity. Therefore, the mutant phenotype of ljcyc2 is in agreement with transgenic data, confirming that LjCYC2 has a specific function in the control of dorsal identity. However, squ1 did not display any other detectable phenotype apart from the abnormal dorsal petal, in contrast to the range of phenotypes in transgenic plants with either sense or antisense transgene of LjCYC2.

Isolation of the kew1 Mutant.

Another mutation only affects the appearance of the lateral petal and is named keeled wing in Lotus 1 (kew1) for its similar phenotype to the mutant keeled wing (k) in pea (24, 25), whose flowers bear the abnormal lateral petals with the shape mimicking keel petals. In kew1, the lateral petals resemble the ventral in both petal shape and epidermal cell types (Fig. 4 E and QS), but the unidirectional initiation of floral organ primordial (23), as well as the early development of lateral petals up to the stage when the primary veins are initiated (Fig. 1L), are the same as the wild type (data not shown). This finding suggests that Kew1 is a specific factor in the control of lateral petal identity and functions at a late stage during petal development. Our mapping experiment positioned kew1 in the short arm of chromosome 5, where it shares a macro synteny with the region containing k in pea (data not shown), suggesting that kew1 and k could be the same genetic locus controlling lateral petal identity in papilionoid legumes.

Interaction Between LjCYC2 and Kew1.

We tested for interaction between LjCYC2 and Kew1 by crossing squ1 with kew1, because mutants of both genes displayed effects on the shape and identity of dorsal and lateral petals, respectively. The squ1kew1 double mutant, identified in the F2 population, had enhanced ventral phenotypes on dorsal and lateral petals, leading to all petals becoming ventralized to a different extent: the dorsal petal was much smaller and did not expand outward (Fig. 4F), with some adaxial epidermal cells possessing the characters of normal ventral petals (Fig. 4T); the lateral petals were ventralized but larger than in kew1 (Fig. 4F), and the cell size and morphology were also ventralized (Fig. 4 R and U). Thus, LjCYC2 and Kew1 have roles which are not limited to only the dorsal or lateral petals, respectively, indicating that both genes are used together in the dorsoventral mechanism to determine petal identities. On the other hand, there is no defect in floral organ initiation or inflorescence development in the squ1kew1 double mutant (data not shown), suggesting that both LjCYC2 and Kew1 could have specific role in the control of petal identity along dorsoventral axis.


In this study, we exploited a model legume, L. japonicus, to conduct a robust functional analysis of a CYC-like gene, LjCYC2, during zygomorphic floral development. We have demonstrated that LjCYC2, similar to CYC in A. majus, has an asymmetric expression pattern in floral meristems and a dorsalizing activity on petal identity, even though the zygomorphy of Leguminosae is believed to have evolved separately from the Lamiales (1, 6, 7, 13). A recent study has shown that the expression pattern of an LjCYC2 orthologue is altered in the legume Cadia, which is typical of a series of unrelated papilionoid genera with unusual, more or less radially symmetrical floral morphologies, supporting a role for CYC homologues in the evolution of floral novelty in legumes (H. L. Citerne, R. T. Pennington, and Q. C. B. Cronk, personal communication). Thus, these data reveal that CYC homologues have been recruited in Leguminosae and play an important role in the development of dorsoventral asymmetry.

It may not be coincidence that both LjCYC2 and CYC confer dorsal activities in the two distantly related lineages. Although the precise functions of CYC and LjCYC2 on the development of zygomorphy need to be clarified further, mutants of both genes display similar effects on petals at the cellular level, such as determining the epidermal cell types, size, and shape. This finding is in agreement with the general role of the TCP-box genes, which is linked to the regulation of cell division and proliferation (2628). It suggests that a conserved function shared by TCP-box genes is important for the dorsal activity of both LjCYC2 and CYC. In the Rosid Arabidopsis, which has radially symmetric flowers, the closest homologue to CYC and LjCYC2, TCP1, displays an asymmetric expression pattern transiently in the dorsal domain of floral meristems, as well as in the adaxial region of axillary shoot meristems (29). It has been speculated that an ancestral asymmetry had been created in a common ancestor of A. majus and Arabidopsis by a CYC/TCP1-related gene, which presumably had the radially symmetrical flower, and further evolution of this gene, such as changes in timing or levels of expression and interactions with the target genes, may underlie repeated and independent recruitment of CYC-like genes to control of zygomorphic floral development (29, 30). Consistently, our finding about the dorsalizing function of LjCYC2 indicates that the similar evolutionary course of CYC-like genes have occurred independently in Leguminosae. Thus, the CYC-like genes could help to define the dorsoventral axis throughout angiosperms and there could be a common molecular origin for the mechanisms controlling floral zygomorphy. However, although both CYC and LjCYC2 possess dorsalizing activity, they bring about the distinct zygomorphic flowers in A. majus and L. japonicus, respectively, which produce different internal symmetry in their counterpart dorsal and ventral petals (Fig. 1 B and D), suggesting that LjCYC2 and CYC could have been integrated into different genetic networks with divergent roles controlling the zygomorphic development. It is likely that the evolution of dorsoventral axis to control of zygomorphic flowers does not only depend on the extent to which the ancestral dorsalizing activity has been elaborated, but also rely on the way how it was responded in each zygomorphic clade. Thus, the dorsalizing activity should be a primary and synapomorphic function being recruited independently into the mechanism to control the dorsoventral axis in different lineages.

There are some hints to how the dorsoventral axis might have evolved in the distantly related species independently. Leguminosae consists of three subfamilies, Caesalpinioideae, Mimosoideae, and Papilionoideae. Caesalpinioideae is a paraphyletic assemblage from which the monophyletic Mimosoideae and Papilionoideae arose. Caesalpinioideae flowers are variable in morphology, although often subtly zygomorphic, whereas Mimosoideae has mostly species with actinomorphic flowers, and most species in Papilionoideae display prominent zygomorphic “pea” flowers. It has been shown that the CYC-like genes have undergone two duplication events: an early one giving rise to LEGCYC I and LEGCYC II genes could have occurred before the evolution of Papilionoideae, and a later one resulting in LEGCYC IA and LEGCYC IB have occurred during the early diversification of Papilionoideae (21, 22). So far, only LEGCYC IB genes, such as LjCYC2 in L. japonicus and its homologue in pea (unpublished data), have been found to possess dorsalizing activity and control the development of dorsal petals. In contrast to the expression pattern of CYC, LjCYC2 in L. japonicus is expressed transiently in the inflorescences (Fig. 3 AF), and altered LjCYC2 expression in transgenic plants had effects on both the development of inflorescences and flowers (Fig. 5C), suggesting multiple developmental roles of LjCYC2 that are not shared with CYC in A. majus. Although a mutant of LjCYC2 (squ1) affected only the identity of dorsal petals (Fig. 4D), it is possible that squ1 is a weak mutation giving rise to a partially functional protein, because its transcript is detectable and capable of encoding a mutant protein with the TCP and R domains intact. Alternatively, the function of LjCYC2 in the inflorescences may be redundant, because the LEGCYC IA gene LjCYC1 is also expressed transiently in inflorescence meristems before it is asymmetrically expressed in the dorsal region of floral meristem. It has been shown that petal asymmetry in A. majus could arise through a series of steps in which the expression patterns of CYC and its homologue DICH are progressively elaborated and maintained from early floral meristem initiation through late stages of development, which should potentially reflect a sequence of evolutionary events (12). Similarly, the dynamic expression pattern of LjCYC2 and LjCYC1 in the inflorescences might be necessary for elaborating the dorsoventral axis in floral meristems, and register the evolutionary course of their ancestor in Leguminosae. It is possible that the duplication of CYC-like genes and their functional divergence are generally important for the inherent dorsalizing activity of their ancestor to be elaborated and responded, so that the dorsoventral axis has been evolved to produce the prominent zygomorphic flowers in different lineages. Further analysis of the expression pattern and function of different CYC-like genes should provide more information about how gene duplication could contribute to the elaboration of dorsoventral axis and the development of zygomorphy among angiosperms.

A genetic factor, Kew1 in L. japonicus, was identified in this study, the mutant of which affects the development of lateral petal only, resulting in the lateral petal resembling the ventral in both petal shape and epidermal cell types (Fig. 4 E and QS). Because only the identity of lateral petals is altered and no other defect has been found in the mutant kew1, Kew1 should have a specific role in the control of lateral petal identity. Although their relationship needs further investigation, the similar phenotypes of the mutant kew1 and k suggest that Kew1 and K might be the same genetic component in the same pathway, and represent a lateralizing factor. We found that LjCYC2 could interact with Kew1, and the double mutant bore ventralized petals. The enhanced phenotype of squ1kew1 double mutant in the dorsal petal indicates that the activity of Kew1 could also be needed for the development of dorsal petals, suggesting that there is a link between the dorsalizing and lateralizing activities which should be important for the elaboration of dorsoventral axis. This possibility should be properly addressed with the cloning of Kew1 and K from L. japonicus and pea in the near future.

Both single mutant, kew1 and squ1, and the double mutant display defects on the development of dorsal and/or lateral petals, and their malfunction appears in the late stages of floral development when primary veins are initiated in petals (Fig. 1M). It has been reported that morphologic differentiation of floral organs along the dorsoventral axis in papilionoid legumes occurs in the late stages of floral ontogeny (31). The specific activities and interaction of LjCYC2 and Kew1 to control the identities of dorsal and lateral petals in the late stage of floral development could provide the molecular basis for the characteristic floral ontogeny in legumes. On the other hand, the development of dorsoventral axis is evident from the beginning of floral organ initiation, suggesting that there could be a heterochronic role for the dorsalizing and lateralizing activities in the control the development of dorsoventral axis and the petal identity respectively. In A. majus, the cyc mutant displays partial, and the cycdich double mutant complete, loss of dorsalized petals, with an increased number of dorsal organs, suggesting that CYC/DICH have an early function to determine the floral organ numbers (11, 12). In contrast, although all petals are ventralized to a different extent, there is no increase of floral organ numbers in the squ1kew1 double mutant, indicating that the LjCYC2 does not play the same role as the one of CYC in the control of floral organ initiation. These data further suggest that the independent recruitment of CYC-like genes within each zygomorphic clade might result in the acquisition of different functions, which could in turn have implication for the divergence of the mechanisms and account for the distinct zygomorphy between different lineages.

Materials and Methods

Plant Material and Growth Conditions.

L. japonicus ecotypes Gifu B129 and Miyakojima (MG-20) were used in this study. All plants were grown at 20–22°C with a 16-h light/8-h dark photoperiod at 150 mE·m−2·s−1.

Two mutants, kew1 and squ1 were isolated from a γ ray mutagenized M2 population in the Gifu B129 background. After several backcrosses, in which 3:1 ratios of the wild type and mutant were observed in the segregating populations, line SH0389 for kew1 and line SH0120 for squ1 were established.

Gene Cloning and RT-PCR.

Fragments of CYC-like genes were amplified from L. japonicus by using degenerate oligonucleotide primers Pcyc1 and rPcyc2 (sequences of all primers are given in Supporting Text) in 5′-RACE using the 5′-RACE system (GIBCO/BRL). RACE products were used to screen a genomic library (32). A pair of primers around the intron of LjCYC2 (SL1716 and SL1717) was used to detect the splicing in both wild type and mutant.

Phylogenetic Analysis.

Protein distance analysis was carried out by using mega version 3.1 (33) by the full length of protein sequences from GenBank according to refs. 11, 12, 14, 18, 29, and 34. Distance matrices used the PAM-Dayhoff model of amino acid substitution; the phylogenetic tree was constructed with the neighbor-joining method, and bootstrap analyses used 1,000 resampling replicates. To simplify the analysis, only the full-length TCP protein sequences of some representative and particular interest TCP-box genes were used.


Tissue for in situ hybridization was fixed overnight in 4% (wt/vol) paraformaldehyde in phosphate buffer, pH 7.0, and embedded in Paraplast (Sigma). Nonradioactive in situ hybridization was performed essentially as described (35). Scanning electron microscopy was performed on plastic replicas as described earlier (36). Contrast and color balance were adjusted by using photoshop 8.0 (Adobe, Mountain View, CA).

Transgenic Plants.

Transformation of L. japonicus was carried out as described (9). Details of transgenic constructions and transgenic lines are described in Supporting Text. The alteration of LjCYC2 expression level in some lines were confirmed by real-time PCR (Supporting Text).


kew1 was mapped by using an F2 mapping population, resulting from a cross between kew1 (from the Gifu B129 background) and Miyakojima (MG-20). In total, 6,000 F2 individuals were analyzed with markers flanking kew1. squ1 was mapped with a F2 mapping population whose size was 160 plants. Two primers (LJSSR0527 and LJSSR0528) were used as the SSR markers for the genomic region containing LjCYC2.

Supplementary Material

Supporting Text:


We thank A Hudson, C. Kidner, R. T. Pennington, Q. C. Cronk, D. Bradley, M. Ambrose, P. Cubas, and E. Coen for their help and critical comments on the manuscripts; two anonymous reviewers for providing help and advice; and Z. Xu, J. Li, Y. Xue, G. Wang, Y. Tian, H. Lin, Y. Wang, and S. Ge for their genuine encouragement and support for this experiment. This work was supported by National Natural Science Foundation of China Grant 30430330 and Ministry of Science and Technology of the People’s Republic of China Grant 2003AA222030.

Conflict of interest statement: No conflicts declared.

See Commentary on page 4801


1. Coen E., Nugent J. M., Luo D., Bradley D., Cubas P., Chadwick M., Copsy L., Carpenter R. Philos. Trans. R. Soc. London B. 1995;350:135–138.
2. Endress P. K. Curr. Opin. Plant Biol. 2001;4:86–91. [PubMed]
3. O’Neill S. D. Annu. Rev. Plant Physiol. Plant Mol. Biol. 1997;48:547–574. [PubMed]
4. Neal P. R., Dafni A., Giurfa M. Annu. Rev. Ecol. Syst. 1998;29:345–373.
5. Giurfa M., Dafni A., Neal P. R. Int. J. Plant Sci. 1999;160:S41–S50. [PubMed]
6. Donoghue M. J., Ree R. H., Baum D. A. Trends Plant Sci. 1998;3:311–317.
7. Cubas P. BioEssays. 2004;26:1175–1184. [PubMed]
8. Lewis G. P., Schrire B. D., MacKinder B., Lock J. M. Legumes of the World. Kew, U.K.: Royal Botanic Gardens; 2003.
9. Handberg K., Stougaard J. Plant J. 1992;2:487–496.
10. Udvardi M. K., Tabata S., Parniske M., Stougaard J. Trends Plant Sci. 2005;10:222–228. [PubMed]
11. Luo D., Carpenter R., Vincent C., Copsey L., Coen E. Nature. 1996;383:794–799. [PubMed]
12. Luo D., Carpenter R., Copsey L., Vincent C., Clark J., Coen E. Cell. 1999;99:367–376. [PubMed]
13. Stebbins G. L. Flowering Plants: Evolution Above the Species Level. Cambridge, MA: Harvard Univ. Press; 1974.
14. Cubas P., Lauter N., Doebley J., Coen E. Plant J. 1999;18:215–222. [PubMed]
15. Almeida J., Rocheta M., Galego L. Development (Cambridge, U.K.) 1997;124:1387–1392. [PubMed]
16. Galego L., Almeida J. Genes Dev. 2002;16:880–891. [PMC free article] [PubMed]
17. Corley S. B., Carpenter R., Copsey L., Coen E. Proc. Natl. Acad. Sci. USA. 2005;102:5068–5073. [PMC free article] [PubMed]
18. Cubas P., Vincent C., Coen E. Nature. 1999;401:157–161. [PubMed]
19. Hileman L. C., Kramer E. M., Baum D. A. Proc. Natl. Acad. Sci. USA. 2003;100:12814–12819. [PMC free article] [PubMed]
20. Reeves P., Olmstead R. Mol. Biol. Evol. 2003;20:1997–2009. [PubMed]
21. Fukuda T., Yokoyama J., Maki M. J. Mol. Evol. 2003;57:588–597. [PubMed]
22. Citerne H., Luo D., Pennington R. T., Coen E., Cronk Q. C. Plant Physiol. 2003;131:1042–1053. [PMC free article] [PubMed]
23. Dong Z. C., Zhao Z., Liu C. W., Luo J. H., Yang J., Huang W. H., Hu X. H., Wang T. L., Luo D. Plant Physiol. 2005;137:1272–1282. [PMC free article] [PubMed]
24. Pellew C., Sverdrup A. J. Genet. 1923;13:125–131.
25. Ellis T. H. N., Turner L., Hellens R. P., Lee D., Harker C. L., Enard C., Domoney C., Davies D. R. Genetics. 1992;130:649–663. [PMC free article] [PubMed]
26. Kosugi S., Ohashi Y. Plant Cell. 1997;9:1607–1619. [PMC free article] [PubMed]
27. Kosugi S., Ohashi Y. Plant J. 2002;30:337–348. [PubMed]
28. Tremousaygue D., Garnier L., Bardet C., Dabos P., Herve C., Lescure B. Plant J. 2003;33:957–986. [PubMed]
29. Cubas P., Coen E., Zapater J. M. Curr. Biol. 2001;11:1050–1052. [PubMed]
30. Costa M. M., Fox S., Hanna A. I., Baxter C., Coen E. Development (Cambridge, U.K.) 2005;132:5093–5101. [PubMed]
31. Tucker S. C. Plant Physiol. 2003;131:911–926. [PMC free article] [PubMed]
32. Guo X., Liu Y., Luo D. J. Plant Physiol. Mol. Biol. 2004;30:234–238. [PubMed]
33. Kumar S., Tamura K., Nei M. Bioinformatics. 2004;5:150–163. [PubMed]
34. Cubas P. In: Developmental Genetics and Plant Evolution. Cronk Q. B. C., Bateman R. M., Hawkins J. A., editors. London: Taylor and Francis; 2002. pp. 247–266.
35. Coen E., Romero J. M., Doyle S., Elliott R., Murphy G., Carpenter R. Cell. 1990;63:1311–1322. [PubMed]
36. Green P. B., Linstead P. Protoplasma. 1990;158:33–38.

Articles from Proceedings of the National Academy of Sciences of the United States of America are provided here courtesy of National Academy of Sciences
PubReader format: click here to try


Related citations in PubMed

See reviews...See all...

Cited by other articles in PMC

See all...


Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...