Comprehensive In Silico Characterization and Expression Profiling of Nine Gene Families Associated with Calcium Transport in Soybean

Abstract

Calcium (Ca²⁺) plays a critical role in the regulation of growth and development and environmental stress responses in plants. The membrane-associated Ca²⁺ transport proteins are required to mediate Ca²⁺ signaling and maintain Ca²⁺ homeostasis. Ca²⁺ channels, pumps (ATPases), and antiporters are three major classes of Ca²⁺ transporters. Although the genome-wide analysis of Ca²⁺ transporters in model plants Arabidopsis and rice have been well documented, the identification, classification, phylogenesis, expression profiles, and physiological functions of Ca²⁺ transport proteins in soybean are largely unknown. In this study, a comprehensive in silico analysis of gene families associated with Ca²⁺ transport was conducted, and a total of 207 putative Ca²⁺ transporter genes have been identified in soybean. These genes belong to nine different families, such as Ca²⁺-ATPase, Ca²⁺/cation antiporter, cyclic nucleotide-gated ion channel (CNGC), and hyperosmolality induced cytosolic Ca²⁺ concentration channel (OSCA). Detailed analysis of these identified genes was performed, including their classification, phylogenesis, protein domains, chromosomal distribution, and gene duplication. Expression profiling of these genes was conducted in different tissues and developmental stages, as well as under stresses using publicly available RNA-seq data. Some genes were found to be predominantly expressed in specific tissues like flowers and nodules, and some genes were found to be expressed strongly during seed development. Seventy-four genes were found to be significantly and differentially expressed under abiotic and biotic stresses, such as salt, phosphorus deficiency, and fungal pathogen inoculation. In addition, hormonal signaling- and stress response-related cis-elements and potential microRNA target sites were analyzed. This study suggests the potential roles of soybean Ca²⁺ transporters in stress responses and growth regulation, and provides a basis for further functional characterization of putative Ca²⁺ transporters in soybean.

1. Introduction

Calcium (Ca²⁺) is an essential macronutrient required for various physiological and biochemical processes in plants [1]. As the divalent cation, Ca²⁺ is indispensable for plant growth and development by playing a structural role in the cell wall, acting as a counter-cation for inorganic and organic anions in the vacuole, and functioning as a universal intracellular second messenger in the cytosol [2]. A high ratio of total Ca²⁺ in plant tissues is located in cell walls, which is required for cell wall stabilization [1]. Within plant cells, a high amount of Ca²⁺ is usually sequestered in vacuoles and endoplasmic reticulum (ER), but the concentration of Ca²⁺ in the cytoplasm and nucleus is always very low (~100–200 nM). High levels of cytosolic Ca²⁺ are toxic to cells by forming a complex with macromolecules and forming insoluble compounds with phosphate derivatives [1,2]. Plants have thus evolved Ca²⁺-transporting and Ca²⁺-buffering systems that work in concert to tightly regulate cytosolic Ca²⁺ concentration ([Ca²⁺]_cyt) [3,4].

Ca²⁺ signal plays a critical role in plant responses to environmental stresses, as well as hormonal and developmental cues [5,6,7,8]. The typical Ca²⁺ signaling process can be divided into several steps: A stimulus activates Ca²⁺-permeable channels to generate specific Ca²⁺ signals after it is detected by its receptor; then Ca²⁺ binds to specific proteins (Ca²⁺ sensors), and in turn, modulates the effector proteins and elicits changes in cellular activities; and finally, the resting state of Ca²⁺ is restored by removing Ca²⁺ from the cytoplasm [9,10]. It is well-known that various environmental and developmental cues can trigger specific and distinct spatial-temporal patterns of changes of [Ca²⁺]_cyt [6,11,12]. The [Ca²⁺]_cyt changes triggered by stimuli are stimulus-specific in terms of frequency, amplitude, duration, and location, which is designated as “Ca²⁺ signature” [9,13]. Stimuli-specific Ca²⁺ signature is generated by spatial and temporal Ca²⁺ fluxes [3,5,14]. Diverse Ca²⁺ transporters are involved in the establishment of the Ca²⁺ signature and Ca²⁺ gradient across the plasma membrane and the inner membranes of plant cells, and thus, play important roles in plant growth and development, and plant responses to various stresses [2,9,15,16]. Ca²⁺ channels, Ca²⁺ pumps (also named Ca²⁺-ATPases), and Ca²⁺ antiporters (also named Ca²⁺ exchangers) are three major kinds of Ca²⁺ transport proteins [2,3].

Ion channels can be broadly classified into voltage-gated, ligand-gated, and mechanical-gated ion channels. A voltage-gated ion channel is regulated by voltage difference across a membrane, a ligand-gated ion channel responds to a ligand-receptor interaction, and a mechanosensitive (MS) ion channel responds to mechanical stimuli [14]. The cyclic nucleotide-gated ion channels (CNGCs) belonging to the superfamily of voltage-gated ion channels, are Ca²⁺-permeable cation-conducting channels that transport Ca²⁺, sodium, and potassium cations across membranes. They have been demonstrated to be involved in Ca²⁺-conducting signal transduction pathways, and are essential for many physiological functions, such as tip growth of pollen tubes and root hairs, leaf senescence, pathogen defense, and abiotic stress tolerance [9,17,18]. There are 20 and 16 CNGC members in Arabidopsis (Arabidopsis thaliana) and rice (Oryza sativa), respectively [19,20].

Hyperosmolality-induced [Ca²⁺]_cyt channel 1 (OSCA1) is a recently identified unique Ca²⁺ channel that was indicated to be an osmosensor by mediating the elevation of [Ca²⁺]_cyt via activation of Ca²⁺ influx conductance in Arabidopsis [15]. Its homolog OSCA1.3 was also indicated to be a Ca²⁺-permeable channel controlling stomatal closure during immunity [21]. OSCA1 belongs to a gene family with 15 members in Arabidopsis and 11 members in rice [15,22]. Two-pore cation channel (TPC) is a special class of voltage-gated Ca²⁺ channels [16]. Arabidopsis genome has a single TPC1 gene, which has been described as a tonoplast localized channel responsible for the Ca²⁺ triggered slow vacuolar current [16]. There is also only one TPC1 gene in rice genome; it was demonstrated to be a putative voltage-gated Ca²⁺-permeable channel [23]. Glutamate receptor-like (GLR) proteins are nonselective cation channels that can be activated by amino acids like glutamate and glycine [24]. Arabidopsis GLR3.4 was identified to be Ca²⁺ selective channel, and GLR1.2 and GLR3.3 were found to be closely related to the generation of Ca²⁺ signals [25,26,27]. In Arabidopsis and rice, 15 and 13 GLR family members have been identified, respectively [24,28].

MS channels respond to a wide range of mechanical stimuli, and play an important role in many physiological functions like osmotic pressure, touch and pain sensation, cell volume regulation, and gravity sense [29]. The first family of putative MS ion channels in plants was identified based on the channels’ similarity to the E. coli MS channel of small conductance (MscS) [30]. Plant MscS like proteins (MSLs) have been suggested to be involved in mechanical osmotic stimuli-triggered Ca²⁺ signals [3]. MSL family genes have been characterized in Arabidopsis and rice, wherein ten and six MSL genes have been found, respectively [31,32]. The yeast Mid1 is a stretch-activated MS ion channel required for Ca²⁺ influx [30]. Mid1-complementing activity (MCA) proteins in plants were identified by a functional screen for cDNAs capable of rescuing the mid1 mutant strain of yeast Saccharomyces cerevisiae [33]. In Arabidopsis, there are two MCA members, while there is only one MCA in rice [33,34]. These MCAs were found to be required for Ca²⁺ influx in response to mechanical events and mediate Ca²⁺ homeostasis [30,33,34,35]. Annexins (ANNs) are multifunctional lipid-binding proteins, encoded by a large multigene family spread through the Eukaryote and Prokaryote domains [36]. Plant ANN family proteins have been suggested to function as unconventional Ca²⁺-permeable channels, with roles in linking Ca²⁺, redox, and lipid signaling to coordinate development with environmental stress response [37,38,39,40]. Eight genes encoding putative ANNs were detected in Arabidopsis; 11 and 25 genes were found in rice and wheat, respectively [41,42,43].

Elevation of [Ca²⁺]_cyt is triggered by various stimuli, and the recovery of basal [Ca²⁺]_cyt is mediated by Ca²⁺ extrusion and sequestration mechanisms, which requires energy-consuming Ca²⁺ transporters that work against the electrochemical gradient [3]. Plant Ca²⁺-ATPases are members of the P-type ATPase superfamily implicated in the restoration and maintenance of ion homeostasis by pumping Ca²⁺ out of the cytosol [44]. Plant P-type Ca²⁺-ATPases are divided into two groups, type IIA (ER-type Ca²⁺-ATPase, ECA) and type IIB (autoinhibited Ca²⁺-ATPase, ACA). ACAs contain an N-terminal autoinhibitory domain that binds to calmodulin and activates the Ca²⁺ pump [45,46]. Plant genomes encode multiple isoforms of both ECAs and ACAs: four ECA genes in Arabidopsis and three in rice, and 10 ACA genes in Arabidopsis and 11 in rice have been detected [47,48]. The Ca²⁺/cation antiporter (CaCA) is also involved in Ca²⁺ transport and homeostasis. The CaCA superfamily can be classified in four distinct subfamilies in higher plants: The H⁺/cation exchanger (CAX), the cation/Ca²⁺ exchanger (CCX), the Na⁺/Ca²⁺ exchanger (NCX) related Mg²⁺/H⁺ exchanger (MHX), and the CAX-related Na⁺/Ca²⁺ exchanger-like (NCL) proteins [49]. Thirteen CaCA genes in Arabidopsis and 16 genes in rice have been identified [49,50].

Soybean (Glycine max) is an important economic crop which provides protein- and oil-rich foods for us human. Soybean is also an important crop for producing renewable fuel of biodiesel. However, the yield of soybean all over the world is frequently affected by environmental stresses [51,52,53]. Although the roles of diverse Ca²⁺ transporters in abiotic and biotic stress responses, hormone signaling, and development have been well documented [9,15,54,55], the majority of these studies were carried out in model plants like Arabidopsis and rice. The identification, classification, phylogenesis, expression profile, and physiological functions of Ca²⁺ transport proteins in soybean are largely unknown. In this study, we performed a genome-wide survey, identified and analyzed all the families encoding putative Ca²⁺ transport proteins in the soybean genome.

2. Materials and Methods

2.1. Identification of Ca²⁺ Transport Proteins in the Soybean Genome

Proteins encoding putative transport proteins in the soybean genome were searched from the Phytozome’s soybean database (https://phytozome.jgi.doe.gov/pz/portal.html) using different keywords, such as “calcium transporter”, “calcium channel”, “calcium pump”, “calcium exchanger”, “cyclic nucleotide-gated channel”, “glutamate receptor”, “annexin”, “two-pore cation channel”, and “mechanosensitive-like channel”. In addition, the amino acid sequences of Arabidopsis CNGCs, Ca²⁺-ATPases, Ca²⁺/cation antiporters, OSCAs, TPCs, ANNs, GLRs, MSLs, and MCAs were used as queries to blast against the soybean genome database using the BlastP program. The nucleotide and protein sequences were obtained and analyzed for protein domains using HMMScan (https://www.ebi.ac.uk/Tools/hmmer/search/hmmscan) [56], and InterPro database (http://www.ebi.ac.uk/interpro/) [57]. Proteins without the characteristic domains and motifs of these calcium transporter families were eliminated. The genes encoding putative Ca²⁺ transport proteins in each family were named according to Arabidopsis orthologues and their positions in the phylogenetic trees.

2.2. Phylogenetic Analysis and Chromosomal Mapping

Sequences of Ca²⁺ transport proteins of Arabidopsis, rice, and soybean were obtained from TAIR (http://www.arabidopsis.org/), RGAP (http://rice.plantbiology.msu.edu/index.shtml), and SoyBase (http://www.soybase.org/), respectively. The amino acid sequences of Ca²⁺ transport proteins were aligned using ClustalW, and a phylogenetic tree was constructed by the neighbor-joining method using the software MEGA6 [58]. The genes encoding putative Ca²⁺ transport proteins were mapped on these 20 soybean chromosomes based on their physical positions on chromosomes. If the distance between two neighboring paralogous genes were less than 100 kb and separated by five or fewer genes, they were considered to be tandemly duplicated genes [59].

2.3. Gene Duplication and Calculation of Ka/Ks Values

The duplication of putative Ca²⁺ transporter genes on segmentally duplicated regions was determined by searching using the plant genome duplication database [60], and according to the results of phylogenetic trees. Non-synonymous (Ka) to synonymous (Ks) substitution rates were used to estimate the selection mode for all putative gene pairs of Ca²⁺ transporter families in soybean. TBtools was used to calculate Ks and Ka [61]. The divergence time (T) was calculated by T = Ks/(2 × 6.1 × 10⁻⁹) × 10⁻⁶ MYA, where 6.1 × 10⁻⁹ is the divergence rate in millions of years translated from Ks value [62].

2.4. Identification of Conserved Domains and Cis-Acting Regulatory Elements

Conserved domains of Ca²⁺ transport proteins were predicted using HMMScan (https://www.ebi.ac.uk/Tools/hmmer/search/hmmscan) [56], and InterPro database (http://www.ebi.ac.uk/interpro/) [57]. Promoter sequences of 1.5 kb upstream to the transcription start site of the genes encoding Ca²⁺ transport proteins were retrieved from SoyBase (http://www.soybase.org/), and the location of cis-acting regulatory elements was analyzed using Regulatory Sequence Analysis Tools (http://rsat.eead.csic.es/plants/) [63].

2.5. RNA-Seq Datasets

Genome-wide public RNA-seq datasets (Reads/Kb/Million (RPKM) normalized data) in different tissues of soybean were retrieved from two soybean RNA-seq datasets [64,65]. High-throughput sequencing datasets for soybean in responses to salt, dehydration, flooding, drought, deficiencies of phosphorus and zinc, and inoculations of pathogenic (Fusarium oxysporum) isolate and symbiotic bacteria (Bradyrhizobium japonicum) were also retrieved from published datasets [66,67,68,69,70,71].

3. Results

3.1. Identification of Putative Ca²⁺ Transport Proteins in the Soybean Genome

By keyword searching and using the blast tool as described in the “materials and method” section, a total of 207 proteins were identified to be potential Ca²⁺ transport proteins, including 39 CNGCs, 33 Ca²⁺-ATPases, 30 CaCAs, 21 OSCAs, 35 GLRs, 26 ANNs, 16 MSLs, two TPCs, and five MCAs (Tables S1–S11 (Supplementary Materials)). The total number of Ca²⁺ transport proteins belonging to nine different families in soybean is about 2.0 and 2.3 folds that of Arabidopsis and rice, respectively (Figure 1).

3.2. CNGC Gene Family in Soybean

In this study, 39 CNGC gene members were identified in the soybean genome (Table S2 (Supplementary Materials)). CNGC protein sequences from soybean, Arabidopsis, and rice were aligned using ClustalW, and a phylogenetic tree was constructed using the neighbor-joining method to determine the similarity and homology between the CNGC families of these plants. The topology of the phylogenetic tree revealed that the GmCNGC gene family can be divided into four major groups (Groups I-IV) (Figure 2), which is based on the Arabidopsis groups [20]. Groups I-III are monophyletic, whereas Group IV is subdivided into two distinct clades (Groups IV-a and IV-b). Group IV contains 13 GmCNGC genes, while the other groups contain five to 12 members. Most of these proteins contained a cyclic nucleotide-binding domain (CNBD) and an ion transport domain (Figure S1 (Supplementary Materials)). The sequence alignment of the 39 GmCNGC proteins indicated that the two most conserved regions within the CNBD domain are a phosphate-binding cassette (PBC), and an adjacent hinge region (Figure S2 (Supplementary Materials)), which is the characteristics of CNBD [72]. Additionally, a relatively conserved IQ motif was shown in all the 39 GmCNGC proteins.

3.3. Ca²⁺-ATPase Gene Family in Soybean

The Ca²⁺-ATPases are P-type ATPases that can be classified in two distinct groups: Type IIA (ECA) and type IIB (ACA). Here, we identified eight ECA genes and 25 ACA genes in the soybean genome (Table S3 (Supplementary Materials)). The soybean Ca²⁺-ATPase family exhibited little sequence variation in terms of amino acid length (752 to 1103 aa), but presented high variation in terms of intron number (from 0 to 36) (Table S3 (Supplementary Materials)). All the soybean Ca²⁺-ATPase proteins were predicted to contain five to ten transmembrane domains (Table S3 (Supplementary Materials)). Domain analysis showed that five domains characteristic for Ca²⁺-ATPases were found to be conserved among most of the soybean Ca²⁺-ATPase proteins. These domains include the N-terminal autoinhibitory domain, N-terminus cation transporter/ATPase, E1-E2 ATPase, haloacid dehalogenase-like hydrolase, and C-terminus cation transporter/ATPase (Figure S3 (Supplementary Materials)). Most of the soybean ACAs were predicted to contain an N-terminal autoinhibitory domain, with the exception of six GmACAs (Figure S3 (Supplementary Materials)). To analyze the evolutionary relatedness among the members of Ca²⁺-ATPases from soybean, rice, and Arabidopsis, a phylogenetic tree was constructed using the deduced amino acid sequences. The phylogenetic analysis revealed that Ca²⁺-ATPases formed two distinct groups, namely, type IIA and type IIB (Figure 3). Type IIB is divided into four subgroups, based on the groups of Arabidopsis. The Ca²⁺-ATPases of the same group seemed to have a high degree of evolutionary relatedness. The investigation for the evolution of Ca²⁺-ATPases among these plant species revealed that the gene members in the same clade have a high degree of similarity (Figure 3), suggesting the common ancestry of P-type II Ca²⁺-ATPases in these plant species.

3.4. CaCA Gene Family in Soybean

Here, a total of 30 genes encoding putative CaCAs were identified in the soybean genome, including 18 CAX genes, eight CCX genes, three NCL genes, and one MHX gene (Table S4 (Supplementary Materials)). The intron number of GmCaCA family genes varied from 0 to 11. The protein length of the GmCaCA family ranged from 182 to 718 amino acids (Table S4 (Supplementary Materials)). All the GmCaCA proteins were predicted to contain three to 14 transmembrane domains (Table S4 (Supplementary Materials)). All the GmCaCA proteins contain NCX domains with the exception of GmNCL3, which has a high sequence and phylogenetic similarity to other NCL members (Figure 4, Figures S4 and S5 (Supplementary Materials)). In addition, all the GmNCLs were predicted to contain EF-hand motifs (Figure S4 (Supplementary Materials)), which is similar to the NCL proteins in Arabidopsis and rice [73]. The presence of EF-hand motifs suggests that these proteins may bind Ca²⁺ and regulate cellular Ca²⁺ homeostasis. Phylogenetic analysis indicated that GmCaCA members are closely related to the CaCA proteins in Arabidopsis and rice (Figure 4).

3.5. OSCA Gene Family in Soybean

A total of 21 putative OSCA genes were identified, and they were named GmOSCA1.1 to GmOSCA4.2 in accordance with Arabidopsis orthologues (Table S5 (Supplementary Materials)). A phylogenetic tree was generated to compare the evolutionary relationship of OSCA proteins among Arabidopsis, rice, and soybean. The OSCA family members were separated into four distinct clades, designated I, II, III, and IV (Figure 5). Clade I, II, III, and IV in soybean contained nine, eight, two, and two members, respectively. The amino acid number of GmOSCAs varied from 500 to 803 (Table S5 (Supplementary Materials)). Seven to 11 transmembrane domains were predicted for the different GmOSCAs (Table S5 (Supplementary Materials)). In addition, Ca²⁺ dependent channel domains, late exocytosis domains, and cytosolic domains were contained by most of the GmOSCAs (Figure S6 (Supplementary Materials)).

3.6. GLR Gene Family in Soybean

Thirty-five putative GLR genes were identified in the soybean genome (Table S6 (Supplementary Materials)). The encoded proteins were predicted to contain three to eight transmembrane domains (Table S6 (Supplementary Materials)). All the proteins were predicted to contain a ligand-gated ion channel, and most of them contain a receptor family ligand-binding domain (Figure S7 (Supplementary Materials)). To detect the evolutionary relationships of GmGLRs, a phylogenetic tree was generated from alignments of all the GLRs in soybean, Arabidopsis, and rice (Figure 6). These GLRs could be grouped into four clades, namely, I, II, III, and IV. There were 4, 16, and 15 GmGLR members in clade I, III, and IV. In clade II, there were nine Arabidopsis GLR members and four rice GLR members, but no soybean GLR member was found. In clade IV, only one rice GLR was shown, and no Arabidopsis GLR was found (Figure 6). Therefore, the phylogenetic relationships among these plants are different in clade II and clade IV, suggesting a diverging trend in the evolution of GLR family across different species.

3.7. ANN Gene Family in Soybean

A total of 26 ANN genes were identified in the soybean genome; they were designated as GmANN1 to GmANN26 (Table S7 (Supplementary Materials)). Three to seven introns were found in these genes, and their protein length ranged from 141 to 362 amino acids, with a typical length of 296 amino acids (Table S7 (Supplementary Materials)). All the GmANN proteins were predicted to contain annexin domains, and their sequences are highly conserved (Figures S8 and S9 (Supplementary Materials)). A phylogenetic tree was constructed using full-length amino acid sequences of all the ANNs in soybean, Arabidopsis, and rice. These ANNs could be grouped into six clades (Figure 7). Clade I contains 11 GmANN members, clade V contains five GmANN members, while other clades contain one to four members. In each clade, proteins from soybean, Arabidopsis, and rice are included, with the exception that there is no Arabidopsis ANN in clade VI (Figure 7). These results suggest the similar clustering of ANN proteins in these three plants.

3.8. TPC Proteins in Soybean

Although there is only one TPC member in Arabidopsis and rice, two orthologues were identified in the soybean genome, and designated as GmTPC1 and GmTPC2 (Table S8 (Supplementary Materials)). These two genes both contain 23 introns and encode proteins with an amino acid length of 738–739. Twelve to thirteen transmembrane domains were predicted for these two proteins (Figure 8A). Similar to AtTPC1 and OsTPC1, ion transport protein domain and EF-hand pair were found in GmTPC1 and GmTPC2 (Figure 8A). Phylogenetic analysis indicated that GmTPC1 and GmTPC2 are closely related to the AtTPC1 and OsTPC1 (Figure 8B).

3.9. MSL Gene Family in Soybean

A total of 16 genes were identified to be putative MSL proteins in the soybean genome (Table S9 (Supplementary Materials)). These proteins have a length from 387 to 947 were predicted to contain three to seven transmembrane domains (Table S9 (Supplementary Materials)). All the soybean MSL proteins were predicted to contain a mechanosensitive ion channel domain (Figure S10 (Supplementary Materials)). A phylogenetic tree constructed to analyze the evolutionary relationships among MSL proteins in soybean, Arabidopsis, and rice, showed that the MSLs could be classified into three clades. There were two, six, and eight GmMSL members in clade I, II, and III, respectively (Figure 9).

3.10. MCA Gene Family in Soybean

Five loci were found to be potential MCA genes; these genes designated as GmMCA1 to GmMCA5 contain five to eight introns and encode proteins with a length of 403 to 418 amino acids (Table S10 (Supplementary Materials)). The protein sequences of MCAs are very conserved (Figure S11 (Supplementary Materials)). All the GmMCAs contain a putative transmembrane domain and a Cys-rich domain of the unknown function, called the PLAC8 motif (Figure 10A). A phylogenetic tree indicated that GmMCAs are closely related to the three proteins in Arabidopsis and rice (AtMCA1, AtMCA2, and OsMCA1) (Figure 10B).

3.11. Chromosomal Distribution and Gene Duplication

The identified 207 genes encoding putative Ca²⁺ transport proteins were variously distributed on all the 20 chromosomes of soybean (Figure 11). The number of genes on each chromosome ranged from 3 to 24. A maximum of 24 genes was found on chromosome 13, while only three genes were found on chromosome 20. Meeting the criteria of distance less than 100 kb and separation less than five intervening genes, 20 groups of genes exhibited tandem duplication (Figure 11). Out of the 20 groups, six groups had two genes, each duplicated; the remaining groups had three to five genes duplicated, thus forming clusters on the chromosomes. The GLR, CNGC, and ANN gene family members formed four, two, and two clusters, respectively.

In this study, 79 duplicated gene pairs of putative Ca²⁺ transporters were identified with higher bootstrap values (<90%) (Table S11 (Supplementary Materials)). The duplication of genes resulted in gene family expansion. The synonymous substitution rates (Ks), the non-synonymous substitution rates (Ka), and the Ka/Ks ratio for the 79 duplicated gene pairs revealed high similarities in their coding sequence alignments. The Ks values of these 79 genes ranged from 0.05 for gene pair GmCNGC13/GmCNGC14 to 0.48 for gene pair GmANN3/GmANN4 with an average Ks of 0.139 (Table S11 (Supplementary Materials)). According to the Ka/Ks ratio, the evolutionary history of selection acting on different genes could be measured [74]. The direction and magnitude of natural selection enforcing on the different protein-coding genes could be interpreted by the Ka/Ks ratio. A pair of sequences having Ka/Ks < 1 indicates purifying selection; Ka/Ks = 1 implies both sequences are drifting neutrally; and Ka/Ks > 1 indicates positive or Darwinian selection [62]. The Ka/Ks of 79 duplicated gene pairs was found to be less than 0.64 (Table S11 (Supplementary Materials))—which indicates the influence of purifying selection on the evolution of these gene pairs. Based on the divergence rate of λ = 6.1 × 10⁻⁹ proposed for soybean [62], the duplication time for each gene pairs was calculated. It is observed that all the segmental duplicated pairs showed a time frame between 4.49 and 39.39 Mya, with an average of 11.38 Mya (Table S11 (Supplementary Materials)).

3.12. Expression Profile of Genes Encoding Putative Ca²⁺ Transport Proteins during Development

The expression profiles of genes encoding putative Ca²⁺ transport proteins in different developmental stages of soybean were analyzed by using two publicly available RNA-seq datasets. The first dataset contains 14 tissues, including three vegetative tissues (leaves, root, and nodules) and whole seed at 11 stages of reproductive tissue development (flower, pod, and seeds) [64]. The second dataset contains nine tissues (root hair cells isolated 84 and 120 h after sowing (HAS), root tip, root, mature nodules, leaves, shoot apical meristem (SAM), flower and green pods) [65]. A total of 166 and 194 genes had detectable expression in at least one tissue in the first and second datasets, respectively (Figure 12 and Figure S12 (Supplementary Materials)). There is no corresponding locus name of version 1.0 (Wm82.a1.v1 used by the RNA-seq data) for 13 genes (GmCNGC23, GmACA18, GmACA24, GmCAX7, GmCAX12, GmOSCA2.4, GmOSCA2.8, GmTPC1, GmANN2, GmANN12, GmANN20, GmGLR4.1, GmGLR4.2). In addition, the expression of 28 genes could not be found in the first dataset (Figure 12). These genes could be expressed exclusively under specific developmental stages or environmental conditions. Consistently, these genes showed very low expression levels in tissues of soybean in the second dataset, such as GmCNGC4/10/11/21, GmACA3/4/7/8, and GmOSCA1.7/2.2/2.6 (Figure S12 (Supplementary Materials)). As shown in both datasets, some genes were expressed ubiquitously in diverse tissues, whereas some genes are tissue-specific. For example, GmCNGC12/13/14/25/26, GmACA5/9/11/13/17, GmCAX1/2/8, and GmOSCA3.1/3.2/4.1/4.2 were ubiquitously expressed in various tissues, GmCNGC33, GmACA21/22, GmCAX4, GmCCX4, and GmGLR4.6 were predominantly expressed in root nodule; GmCNGC20, GmACA16, GmOSCA1.6, and GmGLR1.2/4.8/4.9/4.10 were expressed exclusively in the flowers; GmCNGC29 and GmANN5 had strong expression during seed development (Figure 12 and Figure S12 (Supplementary Materials)).

3.13. Responses of Genes Encoding Putative Ca²⁺ Transport Proteins to Stresses and Rhizobia Infection

To investigate the transcriptional responses of putative Ca²⁺ transport genes to abiotic and biotic stresses, we retrieved the published transcriptome data reflecting soybean responses to various stresses, like dehydration, salt, drought, flooding, phosphate deficiency, zinc deficiency, and inoculations of pathogenic isolates of Fusarium oxysporum [66,67,68,69,70]. Among the Ca²⁺ transport genes, 74 of them (35.7%) were found to be differentially expressed under one or multiple stresses (Figure 13). For example, GmOSCA1.1/1.2/2.5/2.7/3.1/3.2, GmCNGC17, and GmNCL2, GmMHX were induced by drought stress; GmCNGC2/3/32/33/36, GmGLR1.1/3.11/4.3/4.5/4.9/4.10/4.15, and GmACA12/13/22/26 were induced by short-term salt stress; GmACA1/2/25 and GmOSCA1.5 were induced by both dehydration and salt stresses; GmCNGC30 and GmANN10 were repressed by both dehydration and salt stresses. In addition, six genes were found to be responsive to rhizobia (Bradyrhizobium japonicum) infection (12–48 h after inoculation, HAI) [71] (Figure 13). GmCNGC15, GmCNGC34, and GmANN24 were induced, whereas GmGLR3.13 and GmACA19 were repressed by rhizobia infection. GmCNGC27 was induced at 12 HAI, but repressed at 24 HAI, suggesting its dynamic response to rhizobia infection.

3.14. Hormone and Stress-Related Cis-Elements in the Promoters

The presence of cis-elements associated with hormonal signal and stress responses were surveyed in the −1500 bp promoter regions upstream to the predicted transcription start sites of these putative Ca²⁺ transport genes. A total of 13 cis-elements have been surveyed, such as dehydration and cold response element DRE/CRT, ABA response element ABRE, ethylene-responsive element GCC-box, SA-responsive element SARE, and WRKY-binding site W-box (Table S12 (Supplementary Materials)). The exact positions of cis-elements located in the promoters were shown in Table S13 (Supplementary Materials). All of the Ca²⁺ transport genes contain at least one kind of cis-element; some genes contain up to nine kinds of cis-elements, like GmCNGC16, GmOSCA2.8, GmCAX18, GmANN4, GLR3.3, and GmMCA2 (Table S13 (Supplementary Materials)). Among the Ca²⁺ transport genes, most of them contain W-box and/or sulfur-responsive element SURE, around half of them contain SARE, MYC2-binding site, and/or I-box, 36 of them contain ABRE, 68 of them contain DRE, while only 8 of them contain GCC-box (Figure S13 (Supplementary Materials)). It is notable that some cis-elements have many copies in the promoter regions (Figure 14 and Figure S14 (Supplementary Materials)). For example, three and four ABRE elements were found in the promoter of GmOSCA2.7 and GmOSCA2.8, respectively; five DRE elements were found in the promoter of GmCAX18; five W-boxes, two CG-boxes, and two I-boxes were found in the promoter of GmCNGC32 (Figure 14 and Figure S14 (Supplementary Materials)). It is interesting that the drought-responsive element DRE/CRT is more enriched in the drought/dehydration-responsive genes encoding putative Ca²⁺ transporters (Figure 13); the percentage of drought/dehydration-responsive genes containing DRE/CRT element is about 1.6 times more than that of the total Ca²⁺ transporter genes and about 1.8 times more than that of drought/dehydration-irresponsive genes (Figure S15 (Supplementary Materials)), suggesting the possible involvement of DRE/CRT-binding transcription factors in the drought/dehydration response of Ca²⁺ transporter genes in soybean.

3.15. Analysis of Potential MicroRNA Target Sites

microRNAs (miRNAs) have been indicated to be involved in growth and development control, stress response, and nutrient homeostasis by mediating post-transcriptional regulation [75,76]. To detect potential miRNA target sites within the putative Ca²⁺ transporter genes, the mRNA sequences of putative Ca²⁺ transporter genes were acquired from SoyBase (http://www.soybase.org/) and analyzed with the psRNATarget server (http://plantgrn.noble.org/psRNATarget/) [77]. To decrease the number of false-positive predictions, miRNA/target site pairs with an expectation score, and cut-off threshold of 3.0 were considered. Consequently, 61 miRNA members belonging to 37 families were predicted to target 47 genes (Table S14 (Supplementary Materials)). Many of these miRNAs have multiple target genes. For example, gma-miR156b has potential target sites on GmACA22 and GmOSCA2.2; gma-miR164 has potential target sites on GmMSL13 and GmMSL14; gma-miR9750 has potential target sites on GmMCA3 and GmMCA4 (Table S14 (Supplementary Materials)). These results suggest that miRNAs are potentially involved in Ca²⁺ signaling and Ca²⁺ homeostasis regulation.

4. Discussion

The physiological and biochemical functions of Ca²⁺ are enabled by its elaborate transportation across cell membranes, which is mediated by diverse Ca²⁺ transporters, including Ca²⁺ channels, pumps, and antiporters [3,5]. Ca²⁺ transporters, which differ in their cellular distribution and mechanism of transport, execute the complex and tight regulation of Ca²⁺ homeostasis [2]. Many proteins belonging to the families of CNGC, GLR, TPC, OSCA, and MCA have been identified to function as Ca²⁺ channels [15,16,17,26,33]. Some proteins belonging to protein families of ANN and MSL were also suggested to be Ca²⁺-permeable channels [3,37]. Although the physiological roles of many Ca²⁺ channels, pumps, and antiporters have been well documented in model plants like Arabidopsis and rice [12,18,21,34,78,79], there is still limited characterization of Ca²⁺ transport gene families in soybean. In the present study, we have performed a genome-wide in silico analysis of soybean to identify diverse Ca²⁺ transport gene families, including Ca²⁺-ATPase, CaCA, CNGC, OSCA, GLR, TPC, MCA, MSL, and ANN families. The chromosomal location, gene and protein structure, domain architecture, expression profiles, promoter cis-elements, and potential miRNA target sites of genes belonging to these families have been analyzed. The total number of putative Ca²⁺ transport protein-coding genes is 104 in Arabidopsis and 89 in rice, while a total of 207 genes encoding putative Ca²⁺ transport proteins were identified in soybean (Figure 1, Tables S1–S11 (Supplementary Materials)). The number of total Ca²⁺ transport protein-coding genes in soybean is about 2.0 and 2.3 times more than that of Arabidopsis and rice, respectively. For each of the nine Ca²⁺ transport protein families, the number of family members in soybeans was around 2.0 times more than that in Arabidopsis or in rice (Figure 1). For example, the number of CNGC genes in Arabidopsis and rice is 20 and 16, respectively, while the number in soybean is 39, which includes 15 segmentally duplicated gene pairs and five tandem duplicated gene groups (Figure 11, Table S11 (Supplementary Materials)). The possible reason behind this significant increase in gene number might be the two duplication events of soybean that has occurred after the monocot/dicot split approximately 59 and 13 million years ago [80]. In this study, phylogenetic trees were constructed to study the phylogenesis and the expansion of Ca²⁺ transport protein gene families in the course of evolution (Figure 2, Figure 3, Figure 4, Figure 5, Figure 6, Figure 7, Figure 8, Figure 9 and Figure 10). The Ca²⁺ transport protein-coding genes from Arabidopsis, rice, and soybean form common sub-clades with high bootstrap values (Figure 2, Figure 3, Figure 4, Figure 5, Figure 6, Figure 7, Figure 8, Figure 9 and Figure 10), suggesting the conserved evolution of these gene families in monocots and eudicots via common ancestors [4,55]. There are three principal patterns of gene duplications, such as tandem duplication, segmental duplication, and transposition. A total of 79 segmentally duplicated gene pairs were identified in all the putative Ca²⁺ transporter families (Table S11 (Supplementary Materials)). In addition, 20 gene groups showed tandem duplication (Figure 11). The expansion of putative Ca²⁺ transport protein-coding genes by gene duplication could be associated with the adaptation to various adverse environmental conditions during the evolution of soybean.

It is clear that various Ca²⁺ transport proteins play important roles in plant growth and development by mediating the generation of Ca²⁺ signaling. For example, Arabidopsis CNGC5, CNGC6, CNGC9, and CNGC14 function as central regulators of Ca²⁺ oscillations in root hair tip growth [9,81]; Arabidopsis CNGC7, CNGC8, CNGC18, GLR1.2, and GLR3.3 are required for pollen tube growth [26,82]; GLR3.2 and GLR3.4 were indicated to form heteromeric channels to regulate lateral root development via Ca²⁺ signaling in the phloem of Arabidopsis [83]; Ca²⁺-ATPase genes, ACA8, ACA10, ACA12, and ACA13, were suggested to be involved in vegetative growth, inflorescence growth, and seed setting [84,85]. In the present study, the tissue expression patterns of putative Ca²⁺ transporter genes showed that these genes were differentially expressed in diverse tissues of soybean during growth and development (Figure 12 and Figure S12 (Supplementary Materials)). Some ubiquitously expressed genes like GmCNGC12/13/14/25/26, GmACA5/9/11/13/17, GmCAX1/2/8, and GmOSCA3.1/3.2/4.1/4.2 may function in diver tissues during the whole life of soybean. Some genes showed strong expression during seed development, such as GmCNGC29 and GmANN5 (Figure 12 and Figure S12 (Supplementary Materials)), suggesting their potential roles in seed setting. Some genes expressed distinctively in the flowers, such as GmCNGC20, GmACA16, GmOSCA1.6, and GmGLR1.2/4.8/4.9/4.10, which may be associated with the reproductive growth of soybean. It is interesting that some genes, such as GmCNGC33, GmACA21/22, GmCAX4, GmCCX4, and GmGLR4.6, were predominantly expressed in root nodule. In addition, GmCNGC15, GmCNGC27, GmCNGC34, and GmANN24 were found to be induced by the inoculation of rhizobia (B. japonicum) (Figure 13). These results suggest their putative roles in symbiosis establishment and nodule development. In Medicago truncatula, the Ca²⁺-ATPase MCA8 and Ca²⁺ channel CNGC15 have been indicated to be required for the generation of Ca²⁺ spiking in the nucleus, where the Ca²⁺ spiking is induced by symbiosis [86,87].

In this study, 35.7% of the putative Ca²⁺ transporter genes in the soybean genome were found to be significantly and differentially regulated under diverse abiotic and biotic stresses, such as short-term salt, short-term dehydration, flooding, drought, phosphorus deficiency, zinc deficiency, and fungal pathogen inoculation (Figure 13). These differentially expressed genes included members from all groups of Ca²⁺ transport proteins. All of the putative Ca²⁺ transporter genes contain diverse stress- and/or hormone-related cis-elements in their promoter regions (Figure 14 and Figure S14 (Supplementary Materials)). The stress-responsive expression patterns of the putative Ca²⁺ transporter genes suggest the involvement of diverse Ca²⁺ transporters in abiotic and biotic stress responses. It is well-known that numerous environmental and developmental cues can trigger fluctuations of [Ca²⁺]_cyt, and various Ca²⁺ transporters are involved in these processes [13]. It is notable that 48 genes are responsive to short-term salt stress treatment and most of them are Ca²⁺-ATPase genes and GLR genes (Figure 13). In soybean, eight Ca²⁺-ATPase genes (GmACA1/2/12/13/22/23/25/26) were found to be quickly induced by salt stress (Figure 13). This is similar to previous findings that some plant Ca²⁺-ATPase genes were up-regulated by salt stress [45,78,88]. AtACA2 and AtACA4 in Arabidopsis, PCA1 in the moss Physcomitrella patens, and OsACA6 in rice have been identified to be involved in salt stress tolerance [44,78,89,90,91]. The enhanced capacity of Ca²⁺-ATPase could help in restoring [Ca²⁺]_cyt, which is transiently elevated by salt treatment, and thus, maintaining Ca²⁺ homeostasis.

In addition to salt stress, various genes of Ca²⁺ channels, pumps, and antiporters were also indicated to be involved in responses and tolerance to numerous stresses [12,15,21,79,92,93]. However, whether the stress-responsive Ca²⁺ transporter genes are possibly involved in stress response and tolerance in soybean deserves further researches.

5. Conclusions

In this study, a total of 207 genes encoding putative Ca²⁺ transporters have been identified in soybean genome, including 33 Ca²⁺-ATPase genes, 30 CaCA genes, 39 CNGC genes, 21 OSCA genes, 35 GLR genes, two TPC genes, 26 ANN genes, 16 MSL genes, and 5 MCA genes. A comprehensive analysis of these genes was preformed, including phylogenetic relationships, chromosomal distribution, protein domains, gene duplication, promoter cis-elements, potential miRNA target sites, and tissue and stress-responsive expression patterns. The comprehensive in silico characterization and expression analyses would provide preliminary evidence of the potential of role of Ca²⁺ transporter genes in growth and development, symbiosis, and responses to multiple environmental stresses in soybean. Considering the important roles of diverse Ca²⁺ transporters in developmental regulation and environmental stress adaptation that have been documented in model plants, further in-depth functional characterization of the orthologous genes in soybean through the combination of biochemical, molecular, physiological and genetic approaches would facilitate the development of high-yielding and stress-resistant crops.