Introduction

As the second most important plant families for agriculture, Legumes (Fabaceae) provide excellent materials for human food, animal feed, and industry use (Graham and Vance 2003). They account for one-third of crop production in the whole world, and play essential roles in symbiotic nitrogen fixation every year (Benedito et al. 2008). So far, many researchers have devoted to the study of breeding and genetic improvements of Legumes, such as Glycine max, Lotus japonicas, Medicago truncatula. In the past decades, with rapid development of microarray and genome sequencing technologies, the genome-wide research has become feasible. Currently, the genome sequencing of L. japonicas and M. truncatula were finished (Sato et al. 2008; Young et al. 2011). The gene expression atlases of both species were also established, respectively (Benedito et al. 2008; Verdier et al. 2013).

To better understand the regulation on growth and development of Legumes, we mainly aimed at the study on transcription factors (TFs), which temporarily and spatially control the expression of their target genes through binding upstream cis-elements (Jin et al. 2017). MYB is one of most abundant plant transcription factor (TF) families, and has been implicated in diverse plant-specific processes (Cedroni et al. 2003). The first plant MYB gene, isolated in maize, encoded a c-MYB-like TF, which was involved in the biosynthesis of anthocyanin (Du et al. 2012). The most common type of plant MYB TFs is R2R3-MYB with two repeats (Jin et al. 2017). To date, substantial data about MYB transcription factors have been shown in both monocotyledonous and dicotyledonous plants (Feller et al. 2011).

MYB transcription factors were defined by the typical helix-turn-helix motifs (HTH) of their DNA-binding domains. Most MYB proteins in animals and plants have three (R1, R2, and R3) and two (R2 and R3) imperfect repeats, despite the fact that MYB genes containing both one and three repeats have been also found in plants (Braun and Grotewold 1999; Kranz et al. 2000). Thus, the MYB genes in Arabidopsis thaliana were classified into three types: (R1)R2R3_Myb, Myb_related and atypical_MYB (Yanhui et al. 2006). Notably, some groups also classified into three or four subfamilies according to the repeat numbers in the MYB domain (Stracke et al. 2001; Du et al. 2013; Li et al. 2016; Salih et al. 2016). Currently, genome-wide studies on R2R3-type MYB have been conducted in various plant species, such as A. thaliana (Stracke et al. 2001), diploid and polyploid cotton (Cedroni et al. 2003), Zea mays (Du et al. 2012), Beta vulgaris (Stracke et al. 2014), Pyrus bretschneideri (Li et al. 2016), Jatropha curcas (Peng et al. 2016), the tomato family Solanaceae (Gates et al. 2016). In spite of the extensive studies on R2R3-type MYB, the evolutionary history of MYB-related proteins in plants remains largely unknown (Du et al. 2013).

MYB TFs play important roles in plant growth and development, abiotic stress tolerance, hormone signal transduction and disease resistance (Jin and Martin 1999; Roy 2016). For example, expression profiles analysis in peanut (Arachis hypogaea L.) identified 30 MYB genes responsive to abiotic stress treatment (Chen et al. 2014). Overexpression of some MYB genes could lead to alternation of abiotic and biotic stress in tobacco (Li et al. 2014). R2R3-type MYB TFs were involved in secondary metabolism, such as phenylpropanoid metabolism (Jin and Martin 1999). MYB transcription factors, particularly R2R3-type MYB, play important roles in regulation of plant developmental processes, such as defense, cell shape, pigmentation, and root formation (Gates et al. 2016). Thus, MYB transcription factors may be related to organ development processes.

Here, we identified and characterized 104 and 166 MYB genes in L. japonicas and M. truncatula, respectively, most of which were R2R3-type MYB. Phylogenetic analysis indicated that MYB genes in M. truncatula underwent species-specific expansion. The expression analysis showed diverged expression profiles of most MYB genes in various organs, suggesting that they might be involved in plant organ growth and development.

Materials and methods

Identification of MYB genes

An extensive search was performed to identify MYB genes based on all protein sequences in the genomes of both L. japonicas (version: Lotus_r3.0) and M. truncatula (version: Mt4.0v1). All the coding sequences and protein sequences of both species were downloaded from the websites: ftp://ftp.kazusa.or.jp/pub/lotus/lotus_r3.0/ and http://www.jcvi.org/medicago/index.php. Based on the classification of MYB genes in A. thaliana (Yanhui et al. 2006), we collected 124 transcripts of R2R3-MYB genes. All the protein sequences were used for HMMER prediction of transcriptional factors under an E value with 1e-30, among which the classification of MYB genes were conducted according to that in A. thaliana. All the MYB genes were classified into three types: (R1)R2R3_Myb, Myb_related and atypical_MYB (Yanhui et al. 2006). Subsequently, all the coding sequences and protein sequences of the predicted MYB genes were extracted based on their gene/protein ID numbers.

The genomic locations of these MYB genes in L. japonicas and M. truncatula were extracted in the GFF files. Then these MYB genes were mapped to each chromosome of both species. We further predicted the MYB genes in both L. japonicas and M. truncatula subject to tandem duplication by MCScanX (Wang et al. 2012a).

Phylogenetic analysis

To characterize the evolutionary features of MYB proteins, we extracted all the predicted MYB protein sequences in both L. japonicas and M. truncatula. The protein sequences of A. thaliana MYB genes were obtained from the website TAIR (http://www.arabidopsis.org/). Using MEGA7.0 (Kumar et al. 2016), we constructed the phylogenetic trees of all MYB protein sequences in these three species. Both Neighbor-Joining (NJ) method and Maximum Likelihood (ML) method were employed for tree generation. In each method, bootstrap analysis with 1000 replicates was used.

Analysis of gene structures and conserved motifs

The GFF files of L. japonicas and M. truncatula were downloaded from the websites: ftp://ftp.kazusa.or.jp/pub/lotus/lotus_r3.0/ and http://www.jcvi.org/medicago/index.php. We then extracted the information of all the predicted MYB genes in both species, and used the Online Gene Structure Display Server (GSDS 2.0: http://gsds.cbi.pku.edu.cn/) (Hu et al. 2015) to obtain the gene structures of four expanded gene clusters.

To analyze the sequence features of MYB repeats in R2R3-MYB proteins, we extracted the amino acid sequences of R2 and R3 repeats of all R2R3-MYB proteins in both L. japonicas and M. truncatula, and aligned them in each species by ClustalOmega (http://www.ebi.ac.uk/Tools/msa/clustalo/) using the default parameters. Using default settings, we employed the WebLogo (http://weblogo.berkeley.edu/logo.cgi) (Crooks et al. 2004) to create the sequence logos for R2 and R3 repeats from the multiple alignment files generated by ClustalOmega.

In order to identify conserved protein motifs in MYB TFs, we used the MEME software (version 4.11.2) (Bailey and Elkan 1994), with the parameter default settings.

Gene expression analysis

The microarray data of L. japonicas and M. truncatula were downloaded from the gene expression atlas website: http://ljgea.noble.org/v2/index.php and http://mtgea.noble.org/v3/index.php. The organ samples were collected in previous studies (Benedito et al. 2008; Verdier et al. 2013). In L. japonicas, the examined organs included Leaf, Nod21, Pod10d, Pod14d, Pod20, Pt (petiole), Root, Root0h, Seed10d, Seed12d, Seed14d, Seed16d, Seed20d, and Stem, while the M. truncatula organs included Flower, Leaf, Petiole, Pod, Root, Seed_10dap, Seed_12dap, Seed_16dap, Seed_20dap, Seed_24dap, Seed_36dap, Stem, and VegBud (vegetative bud). All these organs used for microarray experiments were harvested from plants under standard growth conditions. The microarray data were normalized as described by Benedito et al. (2008). The Z score was calculated according to previous studies (Benedito et al. 2008; Verdier et al. 2013). We determined the mean expression levels from three biological replicates of each organ. The expression data of MYB genes were then extracted in each species. The clustering analysis and the heat map generated were performed by R package heatmaps (https://cran.r-project.org/web/packages/pheatmap/index.html).

Results and discussion

Identification and classification of MYB transcription factors

To identify MYB genes in both the genomes of L. japonicas and M. truncatula, the HMMER prediction was performed using 198 MYB proteins in A. thaliana as queries, including 126 R2R3-MYB, 5 R1R2R3-MYB, 64 MYB-related, and 3 atypical MYB genes (Yanhui et al. 2006). As a result, 104 and 166 MYB genes were predicted in L. japonicas and M. truncatula, respectively. According to previous classification (Yanhui et al. 2006), there are 2 R1R2R3-MYB genes, 101 R2R3-MYB genes and 1 atypical MYB gene in L. japonicas. 5 R1R2R3-MYB genes, 160 R2R3-MYB genes and 1 atypical MYB gene were found in M. truncatula. In this study, the 104 and 166 MYB genes represented the MYB classification in L. japonicas and M. truncatula. Consistent with previous studies in plants (Li et al. 2016), the R2R3-MYB family is the most abundant MYB TFs in L. japonicas and M. truncatula, thus we further analyzed the R2R3-MYB family.

To map the MYB genes to chromosomes, the genomic locations of them were extracted from GFF files of all genes in L. japonicas and M. truncatula. All 104 MYB genes were mapped to 7 chromosomes (chromosome 0–7) (Fig. 1). 155 of 166 MYB genes (93.4%) were mapped to 8 chromosomes (chromosome 1–8) (Fig. 2), while 11 MYB genes were mapped to 9 unanchored scaffolds, including each MYB gene in 8 unanchored scaffolds and 3 MYB genes in one unanchored scaffold (Table 1). Each of 7 and 8 chromosomes contained R2R3-MYB genes, while the distributions of them were not even. Both the chromosome 0 in L. japonicas and chromosome 5 in M. truncatula contained 29 R2R3-MYB genes (Table 1), accounting for the most R2R3-MYB genes in one chromosome. Each of the 5th and 6th chromosomes contained 6 R2R3-MYB genes, and the 6th chromosome contained 11 R2R3-MYB genes (Table 1), representing the fewest numbers of R2R3-MYB genes in one chromosome. The uneven chromosome distribution of MYB genes may be due to uneven rates of gene duplication events.

Fig. 1
figure 1

The chromosome location of MYB genes in L. japonicas

Full size image
Fig. 2
figure 2

The chromosome location of MYB genes in M. truncatula

Full size image
Table 1 Chromosome distributions of three MYB subfamilies in L. japonicas and M. truncatula
Full size table

Segmental duplication and tandem duplication were two of major mechanisms in origin and evolution of large gene families. From the genomic locations of MYB genes in L. japonicas and M. truncatula, we could conclude that both segmental and tandem duplication played important roles in shaping the evolution of MYB genes. Using MCScanX (Wang et al. 2012a), we obtained 9 and 13 MYB genes subject to tandem duplication (Supplementary Table 1). Interestingly, several gene clusters were subject to species-specific tandem duplication, because these tandem duplicated gene clusters were restricted to L. japonicas or M. truncatula.

MYB TFs have been studied in various plants (Jin et al. 2017), showing that the MYB gene numbers varied from 3 in Helicosporidium to 489 in Brassica napus. Even in Legumes, the MYB TFs vary greatly, from 117 in Vigna radiate to 430 in Glycine max (Jin et al. 2017). This indicated that expansion of MYB TFs occurred in the evolution of Legumes, particularly in G. max. In this study, we identified 104 and 166 MYB TFs in L. japonicas and M. truncatula, suggesting MYB expansion in M. truncatula.

Expansion of MYB transcription factors in M. truncatula

To evaluate the evolutionary significance of the MYB genes, we performed phylogenetic analysis of L. japonicas, M. truncatula and A. thaliana MYB proteins using Neighbor-Joining (NJ) method and Maximum Likelihood (ML) method (Fig. 3 and Supplementary Fig. 1). Generally, the topology of them was similar. All the MYB genes of three species in the NJ tree could be divided to 14 subgroups (Fig. 3). Consistent with previous studies (Wang et al. 2015; Salih et al. 2016), the bootstrap values for some subgroups of the NJ tree were low due to relatively large number of gene sequences.

Fig. 3
figure 3

The Neighbor-Joining (NJ) phylogenetic tress of all MYB genes in A. thaliana, L. japonicas and M. truncatula. filled circle denotes MYB genes in L. japonicas and M. truncatula, in which red denote MYB genes in M. truncatula, while green denote MYB genes in L. japonicas; filled triangle denotes MYB genes in A. thaliana

Full size image

In order to validate the NJ tree, the phylogenetic tree of MYB genes was also reconstructed with ML method. The results here showed that the trees constructed by both methods mentioned above were mainly consistent with each other (Fig. 3 and Supplementary Fig. 1). Thus, the phylogenetic trees of MYB genes were reliable.

Subgroup 14 was the oldest MYB subgroup, while the subgroup 1 was the youngest one. Particularly, the subgroup 1, 5, 6, 8, 9 and 13 contained more MYB genes in M. truncatula than that in L. japonicas (Fig. 3), supporting the species-specific expansion of R2R3-MYB genes in M. truncatula. As shown in Supplementary Table 2, four gene clusters with 36 MYB genes were expanded in both NJ and MP trees, supporting the species-specific expansion of R2R3-MYB genes in M. truncatula. No similar expansion events with at least seven genes were observed in L. japonicas (Fig. 3 and Supplementary Fig. 1).

Additionally, the MYB genes in subgroup 7 were specific to L. japonicas and M. truncatula, but absent in A. thaliana (Fig. 3 and Supplementary Fig. 1). Further sequence analysis showed that the part of MYB proteins in the subgroup 7 were homologous to other proteins in G. max, A. thaliana, Z. mays and O. sativa. Therefore, these MYB TFs may not be specific to Legumes. However, other parts of them have no homology in G. max, A. thaliana, Z. mays or O. sativa, suggesting that they might only emerge in these two species through partial gene duplication. Therefore, M. truncatula may evolve novel R2R3-MYB genes to regulate gene expression.

Features of gene structures and conserved domains in MYB genes

To investigate the gene structure features, we collected the exon numbers of all 270 MYB genes in both L. japonicas and M. truncatula. Most of the R2R3-MYB genes in L. japonicas (95.0%) and M. truncatula (98.1%) contained between 1 and 12 introns (Fig. 4). Most R2R3-MYB genes in both L. japonicas (67.3%) and M. truncatula (75.6%) contained two introns, followed by that containing one intron (18.8% in L. japonicas and 13.8% in M. truncatula) (Fig. 4 and Supplementary Fig. 2). Only a small portion of R2R3-MYB genes were intronless. This result was similar to that described in Arabidopsis, Vitis vinifera, Eucalyptus grandis and Gossypium hirsutum (Matus et al. 2008; Soler et al. 2015; Salih et al. 2016). Notably, all R1R2R3-MYB genes and atypical MYB genes in both species contained no less than seven exons (Supplementary Table 3).

Fig. 4
figure 4

The exon numbers of all MYB genes in L. japonicas and M. truncatula

Full size image

We further studied the variation within the conserved motifs of R2R3-MYB genes in both L. japonicas and M. truncatula, using the WebLogo program (Crooks et al. 2004). The R2 motifs showed similar amino acid compositions between L. japonicas and M. truncatula (Fig. 4), suggesting similar structures and functions between both species. A similar result of R3 motifs was observed (Fig. 4). However, the amino acid compositions between R2 and R3 in the same species were different (Fig. 4), suggesting functional divergence between them. Although the functional divergence of these R2R3-MYB genes remained to be identified, they were thought to play important roles in regulation of gene expression and functional diversification in plants.

Additionally, MEME results showed that most of MYB TFs contained at least three motifs, but the motif sequences were not the same between subgroups, also supporting sequence and even functional divergence of these MYB TFs (Supplementary Fig. 2 and Supplementary Table 4). Several MYB genes lost some motifs in both L. japonicas and M. truncatula. Lj2g3v1534080 (subgroup 2) and Lj4g3v1630970 (subgroup 4) lost the 3rd motif, Lj6g3v2095710 (subgroup 5), Medtr5g007300 (subgroup 5), Lj0g3v0195029 (subgroup 6), and Lj2g3v0320640 (subgroup 6) lost the 2nd motif. In subgroup 8, Lj0g3v0334719 and Lj3g3v3054560 lack the 1st motif. These observations suggested structural diversifications of these MYB TFs in L. japonicas and M. truncatula.

Divergent expression of MYB genes

In L. japonicas, 84 of 104 MYB genes (~80.8%), represented by 120 probe sets, were expressed in at least one of all the investigated organs in this study, while 88 of 166 MYB genes (~53.0%), represented by 99 probe sets, were could be detected in the microarrays of M. truncatula organs, suggesting more than 50% of the MYB genes could be detected in the microarray system of previous studies (Benedito et al. 2008; Verdier et al. 2013). 80.3 and 86.1% of all genes in L. japonicas and M. truncatula were expressed in one or more organs (Benedito et al. 2008; Verdier et al. 2013). Then the average expression values of MYB genes and all genes in both species were calculated. As shown in Table 2, the expressed MYB genes in both species exhibited lower average expression levels than all genes (P < 0.05, t test).

Table 2 The expression of all MYB genes and all genes expressed in L. japonicas and M. truncatula
Full size table

Analysis of organ-specific genes may provide insight into specialized organ processes, including biochemical, physiological, developmental and other processes (Verdier et al. 2013). According to this previous study, we calculated Z scores for each probe set to identify organ-specific genes. Similarly, a minimum Z-score of 2.85 and a minimum normalized expression value >100 were used as threshold values. Interestingly, all of MYB genes in L. japonicas and M. truncatula have Z-scores less than 2.85 (Fig. 5); therefore, they were expressed in at least two organs.

Fig. 5
figure 5

The conserved motifs of the R2R3 MYB genes in L. japonicas and M. truncatula

Full size image

We further analyzed the expression pattern in these organs. Some MYB genes within the same subgroup exhibit similar expression patterns in these organs of L. japonicas and M. truncatula. As shown in Supplementary Tables 5, 9 and 12 gene clusters within the same subgroups were observed to exhibit similar expression profiles in L. japonicas and M. truncatula. For example, Lj2g3v0320640 and Lj6g3v1201340 in subgroup 6 were higher expressed in pod than in other organs (Fig. 6). Lj0g3v0214919 and Lj0g3v0115219 in subgroup 11 exhibited similar expression profile (Fig. 6). The expression levels of Medtr3g083540 and Medtr8g020490 in subgroup 6 were higher in seeds than in other organs (Fig. 7). Medtr0140s0030 and Medtr0489s0020 in subgroup 4 were highly expressed in leaf, Vegetative-Bud and pod (Fig. 7). In spite of these observations, most of MYB genes within the same subgroups showed different expression patterns (Figs. 6, 7), indicating expression divergence and even functional divergence. As mentioned above, these MYB genes might arise through segmental and tandem duplication. Previously, many studies have showed expression divergence after gene duplication was a general pattern in the course of plant evolution (Casneuf et al. 2006; Li et al. 2009; Wang et al. 2012b). Our data here also support this expression pattern.

Fig. 6
figure 6

The heatmap of all MYB genes in L. japonicas

Full size image
Fig. 7
figure 7

The heatmap of all MYB genes in M. truncatula

Full size image

Particularly, only a small portion (25.0%) of the 4 expanded 36 MYB genes in 4 subgroups (Supplementary Table 2) has expression data in M. truncatula, suggesting that most of them may be lowly expressed. These species-specific expanded MYB genes might arise through recent duplication events. Newly duplication genes were usually functional redundant, thus the MYB genes here may reduce this redundancy through down-regulating expression levels.

Roles of MYB TFs in plant development

To characterize the functions of MYB TFs in L. japonicas and M. truncatula, we collected functional assignments of 20 MYB TFs in A. thaliana within 14 subgroups, according to previous study (Stracke et al. 2014). As shown in Supplementary Table 6, MYB TFs within these 11 subgroups were involved in cell cycle, defense, development, differentiation and metabolism. We also collected putative functions of 22 MYB TFs in A. thaliana within 14 subgroups as summarized in a previous report about Chinese White Pear (P. bretschneideri) (Supplementary Table 7) (Li et al. 2016). The MYB TFs within the same subgroups may share similar functions, thus we could obtain functional clues of MYB TFs in L. japonicas and M. truncatula. For example, the MYB TFs within Subgroup 1 was related to development and metabolism, such as shoot morphogenesis and leaf patterning, root development, and lignin biosynthesis (Supplementary Tables 6 and 7). Taken the summary together, MYB TFs within Subgroup 1, 3, 4, 5, 6, 8, 9, 11, 12 and 13 were related to development (Supplementary Tables 6 and 7), covering most of the 14 subgroups. Thus, most of the MYB TFs were involved in plant development.

Conclusions

In summary, the present study identified 270 MYB genes in L. japonicas and M. truncatula according to the A. thaliana genome. The phylogenetic relationships between subfamilies, conserved motifs, expression patterns in different organs, were surveyed in detail. Our results provide better understanding about the molecular basis of MYB genes on the organ growth and development of L. japonicas and M. truncatula, which allow us to obtain a better platform for interesting MYB gene research in the future.