Calmodulin-Like (CML) Gene Family in Medicago truncatula: Genome-Wide Identification, Characterization and Expression Analysis

Calcium is an important second messenger in mediating adaptation responses of plants to abiotic and biotic stresses. Calmodulin-like (CML) protein is an important calcium-signaling protein that can sense and decode Ca2+ signal in plants. Medicago truncatula is a model legume plant; however, investigations of MtCML proteins are limited. Using genome analysis and BLAST database searches, fifty MtCML proteins that possess EF-hand motifs were identified. Phylogenetic analysis showed that CML homologs between M. truncatula, Arabidopsis thaliana and Oryza sativa shared close relationships. Gene structure analysis revealed that these MtCML genes contained one to four conserved EF-hand motifs. All MtCMLs are localized to eight chromosomes and underwent gene duplication. In addition, MtCML genes were differentially expressed in different tissues of M. truncatula. Cis-acting elements in promoter region and expression analysis revealed the potential response of MtCML protein to abiotic stress and hormones. The results provide a basis of further functional research on the MtCML gene family and facilitate their potential use for applications in the genetic improvement on M. truncatula in drought, cold and salt stress environments.


Introduction
Calcium is a universal second messenger in mediating adaptation responses of plants to abiotic and biotic stresses [1]. Diverse stimuli such as plant hormones, low temperature, drought, salt, and pathogens induce rapid and transient changes in cellular Ca 2+ concentration [2], which are sensed and decoded by calcium-binding proteins (CBP) to activate downstream reactions [3]. The CBPs include calmodulins (CaMs), CaM-like proteins (CMLs), Ca 2+ -dependent protein kinases (CPKs/CDPKs) and calcineurin B-like proteins (CBLs) in plants [4][5][6]. A typical CaM contains four EF-hand motifs [7], while an EF-hand motif is composed of two helices, the E helix and the F helix, flanking a Ca 2+binding loop in a structure that resembles a hand [8]. CMLs are a unique class of EF-hand proteins and specifically existed in plants. The CML gene family has been characterized in some plant species, such as Arabidopsis [9], rice [10], grapevine (Vitis amurensis) [11], Chinese cabbage (Brassica rapa) [12] and tomato (Solanum lycopersicum) [13].

Phylogenetic Relationship of CMLs among M. truncatula, Arabidopsis and Rice
To evaluate the evolutionary relationship, all CML members from M. truncatula (50), Arabidopsis (50) and rice (32) were aligned using the maximum-likelihood (ML) method to generate an unrooted phylogenetic tree ( Figure 1). The sequences of MtCML proteins are shown in Table S1. CML proteins in these species could be clustered into eight groups (Groups 1-8). The results suggest existence of a common ancestor before the divergence of monocots and dicots. More members were grouped in Group 3, 6, 7 and 8 than in others. Group 3 contained Arabidopsis and M. truncatula CMLs members, but not rice CMLs (Figure 1), indicating that there are differences in the evolution of CML proteins between monocots and dicots. In addition, most of the MtCML showed homologous to those in Arabidopsis and rice ( Figure 1)

Gene Structure and Domain Architectures
MtCMLs structure was analyzed based on the arrangement of exon/intron. No intron was found in 40 members of MtCMLs, while one to six introns were found in the others. One intron was found in three MtCMLs (MtCML28, MtCML39 and MtCML46); three introns in four MtCMLs (MtCML7, MtCML10, MtCML43 and MtCML47), four in MtCML11 and six in MtCML34 and MtCML38 ( Figure  2A).

Gene Structure and Domain Architectures
MtCMLs structure was analyzed based on the arrangement of exon/intron. No intron was found in 40 members of MtCMLs, while one to six introns were found in the others. One intron was found in three MtCMLs

Chromosomal Location and Synteny Analysis of MtCML Genes
MtCMLs were mapped onto the chromosomes against the M. truncatula genome database to examine their chromosomal distribution. MtCMLs were located on chromosome 1 to 8. Ten MtCMLs  Figure 3A). The results indicated that MtCMLs were randomly scattered on different chromosomes.
Collinearity diagrams among MtCMLs were further analyzed. The results showed that all MtCML genes underwent gene duplication in M. truncatula genome ( Figure 3A, Table S7). Most MtCML genes were found with 1-3 paralogous genes in M. truncatula. Five MtCML genes (MtCML5, MtCML6, MtCML8, MtCML12 and MtCML20) had four homologous genes in M. truncatula. In addition, a synteny analysis of CML genes among M. truncatula and A. thaliana was performed. Twenty pairs of orthologous CMLs were identified between M. truncatula and A. thaliana ( Figure 3B, Table S7). Three genes (MtCML7, MtCML27 and MtCML50) had two homologous genes in A. thaliana, while two genes (MtCML12 and MtCML18) had three orthologous genes in A. thaliana. The percentage of identity between pairs of paralogous MtCML proteins ranged from 21.69% to 100% in EF-hand motifs are the most prominent structural feature in CML proteins. Most of MtCML members (33) contained four conserved EF-hand motifs, and seven members contained three EF-hand motifs. Two EF-hand motifs were present in nine MtCML members (MtCML1, MtCML2, MtCML3, MtCML4, MtCML11, MtCML19, MtCML24, MtCML32 and MtCML27), while one EF-hand motif in MtCML28 ( Figure 2B).

Chromosomal Location and Synteny Analysis of MtCML Genes
MtCMLs were mapped onto the chromosomes against the M. truncatula genome database to examine their chromosomal distribution. MtCMLs were located on chromosome 1 to 8. Ten MtCMLs (MtCML1, MtCML2, MtCML3, MtCML4, MtCML5, MtCML6, MtCML7, MtCML8, MtCML9 and MtCML10) were located on chromosome 1, while nine (MtCML25, MtCML26, MtCML27, MtCML28, MtCML29, MtCML30, MtCML31, MtCML32 and MtCML33) on chromosome five. There were seven MtCMLs locating on chromosomes four, seven and eight, respectively. In addition, five MtCMLs (MtCML13, MtCML14, MtCML15, MtCML16 and MtCML17) were located on chromosome three, three (MtCML34, MtCML35 and MtCML36) on chromosomes six and two (MtCML11 and MtCML12) on chromosome two ( Figure 3A). The results indicated that MtCMLs were randomly scattered on different chromosomes. (Table S7). Among M. truncatula and A. thaliana, the identity between pairs of orthologous ranged 28.06% to 87.16%. Some high percentages of identity between MtCMLs and AtCMLs suggest that MtCML protein sequences and functions could be highly conserved. These results indicated that MtCML genes may be generated by gene duplication, which may be associated with conserved amino acid sequence in MtCML.

Spatial and Temporal Expression Profiles of MtCMLs
The spatial and temporal expression profiles of MtCMLs were examined based on the microarray data from M. truncatula Gene Expression Atlas (MtGEA, https://mtgea.noble.org/v3/). Twenty-nine MtCMLs have corresponding probe sets in the dataset (Table S2). MtCML13, MtCML27, MtCML36, MtCML38 and MtCML42 exhibited relatively high expression in all tissues (flowers, leaves, petioles, pods, stems, roots and seeds) ( Figure 4A), indicating that their functions may be extensive. MtCML7, MtCML14, MtCML17 and MtCML22 were preferentially expressed in roots, and MtCML47 was only expressed in mature seeds, while MtCML20 and MtCML30 were expressed in leaves. Some MtCMLs (MtCML6, MtCML13, MtCML25, MtCML28 and MtCML30) were highly expressed in floral organs ( Figure 4A). On the other hand, relative expression of MtCML7, MtCML14, MtCML17, MtCML20, MtCML22 and MtCML47 was examined using qRT-PCR ( Figure 4B-G). MtCML7 was mostly expressed in roots, stems and leaves, but not in seeds ( Figure 4B); MtCML14 was highly expressed in leaves and its transcript could be detected in roots, stems, flowers and seeds ( Figure 4C). MtCML17 was highly expressed in roots, and its transcript could be detected in stems, flowers and pods ( Figure  D). MtCML20 was major expressed in leaves and flowers as well as in stems, but not in roots and pods ( Figure 4E), MtCML22 was only highly expressed in roots ( Figure 4E). MtCML47 was expressed in all detected organs with highest transcript in leaves ( Figure 4F). The expression patterns of MtCML7, MtCML17, MtCML20, MtCML22 and MtCML47 were consistent with that in microarray data. The different spatial and temporal expression patterns of the MtCML genes suggest their functional diversity in Medicago development. Collinearity diagrams among MtCMLs were further analyzed. The results showed that all MtCML genes underwent gene duplication in M. truncatula genome ( Figure 3A, Table S7). Most MtCML genes were found with 1-3 paralogous genes in M. truncatula. Five MtCML genes (MtCML5, MtCML6, MtCML8, MtCML12 and MtCML20) had four homologous genes in M. truncatula. In addition, a synteny analysis of CML genes among M. truncatula and A. thaliana was performed. Twenty pairs of orthologous CMLs were identified between M. truncatula and A. thaliana ( Figure 3B, Table S7). Three genes (MtCML7, MtCML27 and MtCML50) had two homologous genes in A. thaliana, while two genes (MtCML12 and MtCML18) had three orthologous genes in A. thaliana. The percentage of identity between pairs of paralogous MtCML proteins ranged from 21.69% to 100% in M. truncatula (Table S7). Among M. truncatula and A. thaliana, the identity between pairs of orthologous ranged 28.06% to 87.16%. Some high percentages of identity between MtCMLs and AtCMLs suggest that MtCML protein sequences and functions could be highly conserved. These results indicated that MtCML genes may be generated by gene duplication, which may be associated with conserved amino acid sequence in MtCML.

Spatial and Temporal Expression Profiles of MtCMLs
The spatial and temporal expression profiles of MtCMLs were examined based on the microarray data from M. truncatula Gene Expression Atlas (MtGEA, https://mtgea.noble.org/v3/). Twenty-nine MtCMLs have corresponding probe sets in the dataset (Table S2). MtCML13, MtCML27, MtCML36, MtCML38 and MtCML42 exhibited relatively high expression in all tissues (flowers, leaves, petioles, pods, stems, roots and seeds) ( Figure 4A), indicating that their functions may be extensive. MtCML7, MtCML14, MtCML17 and MtCML22 were preferentially expressed in roots, and MtCML47 was only expressed in mature seeds, while MtCML20 and MtCML30 were expressed in leaves. Some MtCMLs (MtCML6, MtCML13, MtCML25, MtCML28 and MtCML30) were highly expressed in floral organs ( Figure 4A). On the other hand, relative expression of MtCML7, MtCML14, MtCML17, MtCML20, MtCML22 and MtCML47 was examined using qRT-PCR ( Figure 4B-G). MtCML7 was mostly expressed in roots, stems and leaves, but not in seeds ( Figure 4B); MtCML14 was highly expressed in leaves and its transcript could be detected in roots, stems, flowers and seeds ( Figure 4C). MtCML17 was highly expressed in roots, and its transcript could be detected in stems, flowers and pods ( Figure 4D). MtCML20 was major expressed in leaves and flowers as well as in stems, but not in roots and pods ( Figure 4E), MtCML22 was only highly expressed in roots ( Figure 4E). MtCML47 was expressed in all detected organs with highest transcript in leaves ( Figure 4F). The expression patterns of MtCML7, MtCML17, MtCML20, MtCML22 and MtCML47 were consistent with that in microarray data. The different spatial and temporal expression patterns of the MtCML genes suggest their functional diversity in Medicago development.

Analysis of cis-Acting Element in the Promoter Region of MtCML Genes
To analyze the potential function of MtCML genes, a 1000-bp sequence upstream start codon in the promoter region of MtCMLs was analyzed using the PlantCARE. The major cis-acting elements related to stress and phytohormone responses are shown in Table S5. ABA (ABRE), MeJA (CGTGAmotif) and auxin (AuxRR-core) response elements were enriched in the promoter of MtCMLs, while salicylic acid (TCA-motif), gibberellin (GARE-motif), circadian, drought (MBS) and cold (LTR) response elements were also found in the promoters ( Figure 5). Seven MtCMLs (MtCML8, MtCML20, MtCML28, MtCML33, MtCML35, MtCML37 and MtCML40) have drought response element, and six MtCMLs (MtCML2, MtCML6, MtCML16, MtCML19, MtCML33 and MtCML47) have cold response element in the promoter ( Figure 5). The results indicated that MtCMLs may respond to plant hormones and abiotic stresses. MtActin was used as an internal control to ascertain if the different ciselements found in the promoters of the MtCML genes were or not substantially enriched [41]. Three cis-acting elements including two gibberellin responsive elements (GARE-motif and P-box), three salicylic acid responsiveness elements (TCA-element) and one light responsiveness (Box 4) were observed in the promoter region of MtActin (Table S5), while LTR, ABRE, CGTGA-motif, MBS and AuxRR-core elements were enriched in the MtCML promoters regions (Table S5).

Analysis of cis-Acting Element in the Promoter Region of MtCML Genes
To analyze the potential function of MtCML genes, a 1000-bp sequence upstream start codon in the promoter region of MtCMLs was analyzed using the PlantCARE. The major cis-acting elements related to stress and phytohormone responses are shown in Table S5. ABA (ABRE), MeJA (CGTGA-motif) and auxin (AuxRR-core) response elements were enriched in the promoter of MtCMLs, while salicylic acid (TCA-motif), gibberellin (GARE-motif), circadian, drought (MBS) and cold (LTR) response elements were also found in the promoters ( Figure 5). Seven MtCMLs (MtCML8, MtCML20, MtCML28, MtCML33, MtCML35, MtCML37 and MtCML40) have drought response element, and six MtCMLs (MtCML2, MtCML6, MtCML16, MtCML19, MtCML33 and MtCML47) have cold response element in the promoter ( Figure 5). The results indicated that MtCMLs may respond to plant hormones and abiotic stresses. MtActin was used as an internal control to ascertain if the different cis-elements found in the promoters of the MtCML genes were or not substantially enriched [41]. Three cis-acting elements including two gibberellin responsive elements (GARE-motif and P-box), three salicylic acid responsiveness elements (TCA-element) and one light responsiveness (Box 4) were observed in the promoter region of MtActin (Table S5), while LTR, ABRE, CGTGA-motif, MBS and AuxRR-core elements were enriched in the MtCML promoters regions (Table S5).

Expression Profiles of MtCMLs in Response to Salt, Drought and Cold
Gene expression profiles of MtCMLs in different tissues under salt, drought and cold stresses were obtained from M. truncatula Gene Expression Atlas (MtGEA, https://mtgea.noble.org/v3/). There are 29 MtCML genes that have corresponding probe sets in the dataset [42]. MtCMLs expression patterns were changed in response to salt and drought stress (Table S3). MtCML24 and MtCML50 were significantly upregulated after six hours of salt stress, whereas MtCML17 was downregulated ( Figure 6A). In the hydroponic treatment experiment, MtCML10, MtCML15, MtCML16 and MtCML50 transcripts were upregulated after one hour of salt stress and downregulated at 10 h ( Figure 6B). Most of MtCML transcripts were unaltered in shoot during drought treatment except for MtCML33 and MtCML47, whose expression was increased significantly ( Figure 6C). On the other hand, MtCML15, MtCML16, MtCML23, MtCML47 and MtCML50 transcripts in roots were induced by drought treatment, followed by decrease after rewatering, while MtCML7, MtCML14, MtCML22, MtCML37, MtCML39, MtCML45 and MtCML49 transcripts were significantly reduced after drought treatment followed by increase after rewatering ( Figure 6D). The results indicated that MtCML may participate in drought and salt stress responses.
Expression of six MtCML genes that have LTR cis-acting element in the promoter regions in response to cold was analyzed (Table S4)

Expression Profiles of MtCMLs in Response to Salt, Drought and Cold
Gene expression profiles of MtCMLs in different tissues under salt, drought and cold stresses were obtained from M. truncatula Gene Expression Atlas (MtGEA, https://mtgea.noble.org/v3/). There are 29 MtCML genes that have corresponding probe sets in the dataset [42]. MtCMLs expression patterns were changed in response to salt and drought stress (Table S3). MtCML24 and MtCML50 were significantly upregulated after six hours of salt stress, whereas MtCML17 was downregulated ( Figure 6A). In the hydroponic treatment experiment, MtCML10, MtCML15, MtCML16 and MtCML50 transcripts were upregulated after one hour of salt stress and downregulated at 10 h ( Figure 6B). Most of MtCML transcripts were unaltered in shoot during drought treatment except for MtCML33 and MtCML47, whose expression was increased significantly ( Figure 6C). On the other hand, MtCML15, MtCML16, MtCML23, MtCML47 and MtCML50 transcripts in roots were induced by drought treatment, followed by decrease after rewatering, while MtCML7, MtCML14, MtCML22, MtCML37, MtCML39, MtCML45 and MtCML49 transcripts were significantly reduced after drought treatment followed by increase after rewatering ( Figure 6D). The results indicated that MtCML may participate in drought and salt stress responses.
Expression of six MtCML genes that have LTR cis-acting element in the promoter regions in response to cold was analyzed (Table S4)

Discussion
Camodulin-like proteins play a key role in signal transduction during plant growth and development [3]. A total of 50 CML genes were identified in M. truncatula in this study, which was same to that in Arabidopsis (50 CML members) [9], but different from that in rice (32 CML genes) [10], grapevine (68 CML genes), Chinese cabbage (79 CML genes) [12] and tomato (52 CML genes) [13]. MtCMLs show extensive variations in gene length, predicted protein size and pI. The varying protein size and pI and hydrophilic property of MtCMLs are consistent with that in CMLs of Chinese cabbage [12,43]. The function of CML members are dependent upon their amino acids sequences, number of EF-hand motifs that are implicated in the Ca 2+ -binding properties and their responses to plant hormones and environmental stresses [8,44]. The number of EF-hand motifs varies among plant species. Arabidopsis CML proteins typically possess two to six EF-hand motifs [9], but one to four conserved EF-hand motifs were found in MtCMLs. The diversity of MtCMLs in structure may result in functional diversity. The intron-exon structural analysis showed that most of MtCML genes have no intron. This case was consistent with AtCMLs, OsCMLs and BrCMLs [9,10,12].
MtCMLs were classified into eight groups, which is likely similar to those in other plant species [9,12,13]. Phylogenetic analysis of CML proteins revealed that many MtCMLs were homologous to those in Arabidopsis and rice. In consistence, most of BrCMLs from Brassica rapa were also closely related to their corresponding homologs in Arabidopsis and rice [12]. These results suggest that CMLs are conserved among plant species. Gene duplication events are important in the rapid expansion and evolution of gene families [45]. In our study, chromosomal distribution analysis revealed that all MtCML genes were randomly scattered on different chromosomes. Homologous gene pairs were

Discussion
Camodulin-like proteins play a key role in signal transduction during plant growth and development [3]. A total of 50 CML genes were identified in M. truncatula in this study, which was same to that in Arabidopsis (50 CML members) [9], but different from that in rice (32 CML genes) [10], grapevine (68 CML genes), Chinese cabbage (79 CML genes) [12] and tomato (52 CML genes) [13]. MtCMLs show extensive variations in gene length, predicted protein size and pI. The varying protein size and pI and hydrophilic property of MtCMLs are consistent with that in CMLs of Chinese cabbage [12,43]. The function of CML members are dependent upon their amino acids sequences, number of EF-hand motifs that are implicated in the Ca 2+ -binding properties and their responses to plant hormones and environmental stresses [8,44]. The number of EF-hand motifs varies among plant species. Arabidopsis CML proteins typically possess two to six EF-hand motifs [9], but one to four conserved EF-hand motifs were found in MtCMLs. The diversity of MtCMLs in structure may result in functional diversity. The intron-exon structural analysis showed that most of MtCML genes have no intron. This case was consistent with AtCMLs, OsCMLs and BrCMLs [9,10,12].
MtCMLs were classified into eight groups, which is likely similar to those in other plant species [9,12,13]. Phylogenetic analysis of CML proteins revealed that many MtCMLs were homologous to those in Arabidopsis and rice. In consistence, most of BrCMLs from Brassica rapa were also closely related to their corresponding homologs in Arabidopsis and rice [12]. These results suggest that CMLs are conserved among plant species. Gene duplication events are important in the rapid expansion and evolution of gene families [45]. In our study, chromosomal distribution analysis revealed that all MtCML genes were randomly scattered on different chromosomes. Homologous gene pairs were found in almost all MtCMLs. There are many homologous CML genes between the chromosomes in M. truncatula and Arabidopsis. Whole-genome duplication (WDG) in legumes played a major role in shaping the genome and contributed the raw material for the evolution of nodulation [39,46]; M. truncatula has undergone high rates of local gene duplication and share an ancient round of gene duplications from other legume species [39,47]. Therefore, WGD and tandem duplication probably participated in driving MtCML genes evolution.
The tissue-specific expression patterns of MtCML genes are presumed to be associated with their potential biologic roles. Four MtCMLs (MtCML6, MtCML13, MtCML25 and MtCML28) were highly expressed in floral organs, indicating that they are associated with flowering regulation. This can be supported by that AtCML25, the homolog of MtCML25, is involved in pollen germination and pollen tube elongation via regulating Ca 2+ and K + transmembrane trafficking in pollen grains and pollen tubes in Arabidopsis [25]. MtCML17 and MtCML22 were highly expressed in roots, indicating that they are associated with roots in M. truncatula. However, qRT-PCR showed that MtCML14 and MtCML47 were mostly expressed in leaves and seeds rather than roots ( Figure 4B-G). The expression difference between qRT-PCR and microarray data may result by the difference of plant growth stages.
The temporal, spatial, and cell type-specific expression of genes are associated with the regulatory elements in the promoter region [48]. Some promoter elements (LTR, ABRE, CGTGA-motif, MBS and AuxRR-core) were enriched multiple times in the promoter regions of MtCML genes, but were not found in MtActin. It suggested that MtCML genes can be involved in regulating multiple hormone responses and abiotic stresses. Previous studies have shown that some CML genes participated in the hormonal or abiotic stresses responses when specific cis-elements were found in their promoter regions, like AtCML9 [49], AtCML20 [36], AtCML37 [34] and MtCML40 [40]. This is consistent with our hypothesis. The expression of CML genes varies specifically in response to Ca 2+ signaling and a variety of stress responses, especially abiotic stresses. In the present study, four MtCML genes (MtCML14, MtCML17, MtCML30 and MtCML50) and nearly half of the MtCML genes showed extensive responses to salt and drought stresses, respectively. Similar to these results, studies have shown that AtCML9 was induced under dehydration treatment [2]. Transcripts of AtCML9, AtCML37, AtCML38, AtCML39 and OsMSR2 (a rice CML) were up-regulated substantially following salt stress [2,35,50]. Although MtCML40 is not found in the database, its expression was induced by salt treatment and negatively regulates salt tolerance in M. truncatula [40]. Six MtCML genes showed response to cold stress. AtCML24 transcript is induced after cold treatment in Arabidopsis [23]. The tomato (Solanum habrochaites) CML gene, ShCML44, was found to improve the tomato tolerance in cold, drought and salinity stresses [51]. These previous studies have shown that CML can play a role in response to salt, drought and low temperature stresses, which is consistent with our qRT-PCR results. These results will help further explore the function of MtCML genes in M. truncatula.

Identification of CML Genes in M. truncatula
The M. truncatula Mt4.0v1 protein sequences were downloaded from Phytozome 12 (https: //phytozome.jgi.doe.gov/). The reported CML gene sequences in A. thaliana were downloaded from the TAIR database (https://www.arabidopsis.org/) and used as a query to perform BLASTP searching. The NCBI database was used to search the potential CML genes in M. truncatula. In addition, the protein sequences of OsCMLs was downloaded from TIGR (http://rice.plantbiology.msu.edu/) and used as queries to search against the M. truncatula proteome. To differentiate CaM and CML genes which were homologs in M. truncatula, we follow the principle of the major character of EF-hand-containing proteins [13,43] and screen the candidate CML genes. All resulting non-redundant protein sequences were checked for the presence of the entire EF-hand domain by SMART (http: //smart.embl-heidelberg.de/) and InterProScan (http://www.ebi.ac.uk/interpro/).

Analysis of Conserved Domain, Gene Structure and Characterization of MtCML Genes
The MtCML protein sequences were analyzed for physical and chemical characteristics, including the molecular weight (MW), theoretical point (pI) and grand average of hydropathicity (GRAVY), using the ProtParam tool of ExPASy (http://web.expasy.org/protparam/). The exon-intron structure analysis of MtCML genes was conducted using the TBtools (Toolbox for Biologists) program with default parameters. The conserved motifs were analyzed using the MEME tool (http://meme-suite.org/ tools/meme), with the minimum width of motifs as 10, the maximum width of motifs as 40 and the other parameters as default values.

Phylogenetic Relationships of CML Proteins in M. truncatula, Arabidopsis and Rice
The multiple alignments of MtCML proteins were performed using ClustalX with default parameters. Phylogenetic analysis of CMLs in M. truncatula, Arabidopsis and rice was performed using MEGA X with the maximum-likelihood (ML) method 1000 bootstrap replicates [52].

Chromosomal Locations of MtCML Genes
The sequences of MtCML were used to retrieve their chromosomal locations in M. truncatula genome databases Phytozome 12 (https://phytozome.jgi.doe.gov/). TBtools software was used for analyze the chromosomal locations and homologous relationship of MtCML genes [53]. The Multiple Collinearity Scan toolkit (MCScanX) was used to analyze the gene duplication events [53].

Analysis of cis-Acting Elements of MtCML Genes
A 1000 bp of genomic DNA sequence upstream of the transcriptional start site in each MtCML was obtained from the Phytozome database (https://phytozome.jgi.doe.gov), and the PlantCARE database (http://bioinformatics.psb.ugent.be/webtools/plantcare/html/) [54] was used to identify the cis-acting elements in the promoter region of MtCMLs.

Analysis of Microarray Expression Profile
The genome-wide microarray data of M. truncatula in different tissues at various developmental stages and the response to drought and salt were retrieved from M. truncatula Gene Expression Atlas (MtGEA, https://mtgea.noble.org/v3/). The transcript data of MtCMLs were analyzed using TBtools software. The normalized expression data were used to generate heatmap using the TBtools software [53].

Analysis of Relative Expression of MtCMLs in Different Tissues and Response to Cold Using qRT-PCR
M. truncatula (Jemalong) A17 plants were were grown in 15 cm-diameter plastic pots containing a mixture of peat and perlite (3:1, v/v) in a greenhouse with temperature ranging from 20 to 28 • C under natural light. Four-week-old plants were exposed to low temperature in a growth chamber at 4 • C for 24 h under 12 h light as cold treatment. Total RNA was extracted using RNAprep pure Plant Kit (TIANGEN, Beijing, China) according to the manufacturer's instructions. The cDNA was synthesized from 1 µg of total RNA, using the PrimeScript RT reagent Kit with gDNA Eraser (Takara, Japan). MtCML genes transcript were analyzed using qRT-PCR in Thermal Cycler Dice Real Time System II (Takara, Japan) following the manufacturer's instructions. The PCR reaction mix was consisted of diluted cDNA as template, 200 nM forward and reverse primers and 5 µL SYBR Premix Ex Taq (Takara, Dalian, China). Parallel reactions to amplify actin were used to normalize the amount of template. Relative expression was calculated by 2 −∆∆Ct . The expression levels in the control plants without treatment (0 h) were normalized to 1. The MtActin gene was used as the internal control. The qRT-PCR primers listed in Table S6 were acquired by the PrimerQuest tool (http://www.idtdna.com/Primerquest/Home/Index).

Conclusions
A total of 50 MtCML genes were identified in M. truncatula, which were divided into eight groups. Analysis of intron-exon structure revealed that CML gene family is evolutionarily conserved. The functional divergence between subgroups was mainly attributed to changes in the EF-hand domains, which may confers functional differentiation of MtCMLs. Synteny analysis showed that WGD and tandem duplication probably participated in driving MtCML genes evolution. Expression profile analysis showed that most of MtCML genes are expressed in leaves, flowers and roots and had a tissue and organ expression preference. Many cis-acting element responsive to plant hormones, such as ABA, MeJA and auxin were enriched in the promoter of MtCMLs, while salicylic acid (TCA-motif), gibberellin (GARE-motif), circadian, drought (MBS), and cold (LTR) response elements were found in the promoter of some MtCMLs. Many MtCMLs were regulated by drought and salt stresses. qRT-PCR analysis showed that MtCML16 and MtCML33 were upregulated, MtCML2, MtCML6, MtCML19 and MtCML47 were downregulated in response to cold treatment. Our results provide new information for further understanding of the structure, evolution and function of MtCML gene family. Author Contributions: Z.G. and Q.S. designed this study; Q.S. and S.Y. performed the bioinformatics analysis; Z.G. monitored the experimental work; Q.S. and S.Y. carried out the expression analysis; Q.S. and Z.G. wrote the manuscript. All authors have read and agreed to the published version of the manuscript.

Conflicts of Interest:
The authors declare no conflict of interest.