Background

Calcium is a universal secondary messenger that participates in multiple eukaryotic signalling pathways [1]. In plants, Ca2+ signal transduction via calcium-conducting channels is an important mechanism for transducing the signals derived from diverse environmental and developmental stimuli [2, 3]. Additionally, signal transductions contribute to growth, plant biotic interactions, and responses to hormones, light, and salt stress [4]. Cyclic nucleotide-gated ion channels (CNGCs) are components of Ca2+-conducting signal transduction pathways [5]. They are Ca2+-permeable cation-conducting channels that transport sodium, calcium, and potassium cations across membranes. Localized in the plasma membrane [6, 7], vacuole membrane [8], or nuclear envelope [9], CNGCs are controlled from inside the cell by secondary messengers such as Ca2+/calmodulin (CaM) and cyclic nucleotide monophosphates (cNMPs; 3′,5′-cAMP and 3′,5′-cGMP) [3, 6, 10, 11]. The CNGCs are hypothesized to be involved in the uptake of both essential and toxic cations, Ca2+ signalling, development, pollen fertility and tip growth, gravitropism, leaf senescence, innate immunity, pathogen defence, and abiotic stress tolerance [6, 12,13,14,15].

The application of bioinformatics tools (for genes/proteins prediction and phylogenetic analysis), and experimental approaches (gene expression, mutant analysis and overexpression in yeast/Escherichia coli) have led to the identification, characterization, and functional analysis (in exceptional cases) of CNGC family genes in important plant species, including Arabidopsis thaliana [5], rice [16], tomato [17], pear [18], and Physcomitrella patens [19]. Researchers have only recently started to investigate the evolution, function (and underlying regulatory mechanism) of plant CNGCs, as well as their phylogenetic relationships with other channels. Briefly, plant CNGCs are characterised by conserved structural components, including a short cytosolic N-terminus, six transmembrane helices (S1–S6) with a pore-forming region between S5 and S6, and a cytosolic C-terminus containing a cNMP-binding domain (CNBD). The CNBD is the most conserved region of CNGCs carrying a plant CNGC-specific motif spanning the phosphate-binding cassette (PBC) and hinge region, which mediates channel gating by cAMP and/or cGMP [3, 20]. A latest study of the A. thaliana CNGC12 gene suggested that plant CNGCs have multiple CaM-binding domains (CaMBDs) at cytosolic N- and C-termini [3]. Moreover, channel functionality depends on CaM binding to the conserved isoleucine–glutamine (IQ) motif in the C-terminus of the channel, indicating CaM positively and negatively regulates CNGCs [3]. Studies on individual isoforms and the A. thaliana CNGC family revealed that plant CNGC genes may be functionally distinguished in a group-dependent manner. For example, AtCNGC19 and AtCNGC20, which belong to Group IV-a, are involved in salt stress responses [8]. Additionally, AtCNGC2 and AtCNGC4, which are Group IV-b members, affect disease resistance against various pathogens and thermotolerance [21, 22]. Mumtaz et al. [4, 17] recently concluded that Group IV-b SlCNGC genes regulate different types of resistance against diverse pathogens in tomato. It is unclear whether this also applies to other plant species.

Brassica oleracea (2n = 18) is a member of the family Brassicaceae (approximately 338 genera and 3709 species), which consists of many important vegetable and oilseed crops, including brussels sprout, kohlrabi, and kale [23]. Among the cultivated species, B. oleracea exhibits the largest genetic and morphological diversity, making it highly adaptable to different environments. Sexually compatible B. oleracea crops, such as cabbage, cauliflower, and broccoli, are valued for their economic, nutritional, and potent anticancer properties [24]. The whole-genome sequence of this plant species was recently published [24], which enabled us to study the B. oleracea CNGC family. We used in silico and experimental approaches to identify, characterise, and functionally verify CNGC gene family members. We applied multiple tools and programs to complete in-depth analyses of each CNGC gene family member, including an analysis of the physiological and biochemical properties of the encoded proteins. Our objective was to elucidate the diversification, expansion, and evolution of the CNGC gene family. Furthermore, we investigated CNGC expression patterns to clarify the mechanisms underlying their responses to biotic and abiotic stresses, and to identify novel genes potentially useful for breeding.

Results

Genome-wide identification of CNGC genes in Brassica oleracea

For a complete overview of the B. oleracea CNGC gene family, we used the 20 A. thaliana CNGC genes as queries in BLAST searches of the Ensembl Plants database. Out of the 34 non-redundant putative gene sequences retrieved, eight gene accessions with truncated sequences or lacking CNGC-specific domains (CNBD and transmembrane) were eliminated from analyses (Additional file 1). Finally, 26 CNGC genes containing both essential domains (PF00520/PF07885 and PF00027) and a CNGC-specific motif were identified in the B. oleracea genome (i.e., BoCNGC1–26). Of the 26 BoCNGC genes identified in the latest genome assembly version in Ensembl Plants, 16 and 24 were detected in earlier versions from Bolbase (v.1.3) and GenBank (v.2.1) respectively (Table 1).

Table 1 Properties of 29 BoCNGC genes identified in Brassica oleracea
Full size table

The physiological and biochemical properties of the 26 BoCNGC proteins were determined by computing different parameters, and are tabulated in Table 1. These proteins varied in length from 558 to 789 amino acids, with an average of 717 amino acids. The ProtParam tool revealed that there was a considerable range in BoCNGC residue weight (112.795–116.128 g/mol) and molecular weight (63.938–89.775 kDa) depending on the number of atoms present. The computed average pI of majority of BoCNGC proteins was relatively high (range 8.23 to 10.18), signifying that these proteins are localized to membranes, and will supposedly participate in basic buffers. The BoCNGC19, which had pI than 7.4, indicate that this protein likely participate in the acidic buffers. Approximately one third of BoCNGC proteins had a low net charge (<17), while other are composed of more charged amino acids. Nearly all BoCNGC were hydrophilic, with BoCNGC17 and BoCNGC22 being slightly hydrophobic, which endorses its multifaceted role in cellular membrane transport. According to the instability index (II), only two proteins were stable in test tubes, namely BoCNGC2 and BoCNGC3. Aliphatic index showed that most BoCNGC proteins were thermostable at a wide temperature ranges, similar to other globular proteins.

Phylogenetic analysis of BoCNGC genes

Multiple sequence alignments and a maximum likelihood phylogenetic tree constructed between BoCNGCs and AtCNGCs were used to determine the similarity and homology between the B. oleracea and A. thaliana CNGC families. To strengthen the phylogenetic analysis, we identified and included 29 CNGC homolog genes from sister specie Brassica rapa (BrCNGCs) in current analysis. The sequence alignment revealed high similarity between the amino acid sequences of the three species, especially at the conserved domain regions (Additional file 2). The topology of the inferred maximum likelihood scoring tree revealed that the BoCNGC gene family can be divided into four major groups (i.e., Groups I–IV), which are based on the A. thaliana groups (Fig. 1) [5]. Groups I–III are monophyletic, while Group IV is sub-divided into two distinct clades (i.e., Groups IV-a and IV-b). Group IV contains 12 BoCNGC genes, while the other groups contain three to six members. Moreover, individual phylogenetic trees that were constructed based on the aligned B. oleracea and A. thaliana CNGC proteins produced similar clustering patterns (Additional files 3 and 4).

Fig. 1
figure 1

Phylogenetic tree of Brassica oleracea, Arabidopsis thaliana, and Brassica rapa CNGC proteins. A maximum likelihood phylogenetic tree was created with MEGA 6.0, using the Jones–Taylor–Thornton model. The bootstrap values from 1000 replications are provided at each node. The BoCNGC proteins identified in this study are indicated with blue circles, while the AtCNGCs and BrCNGCs are indicated with maroon diamonds and green rectangles, respectively

Full size image

Chromosomal distribution and diversification of BoCNGC genes

The 26 BoCNGC genes were mapped onto B. oleracea chromosomes, and the position of each locus was determined. These genes were randomly distributed across the genome, and were detected on eight of nine chromosomes (i.e., C1–5 and C7–C9). The BoCNGC genes were unevenly distributed, with some chromosomes (i.e., C1 and C5) carrying five genes, while the rest had fewer genes (e.g. C7). Chromosome 6 did not carry any of the BoCNGC genes (Fig. 2a).

Fig. 2
figure 2

Chromosomal localization, synteny, and expansion of the B. oleracea CNGC gene family. a Physical locations and distances of the BoCNGC genes across the eight Brassica oleracea chromosomes. Red and blue lines correspond to forward and reverse orientations of each locus, respectively. b Circos plot presenting gene duplication (tandem and segmental) events and synteny of the BoCNGC genes. The BoCNGC genes are presented as numbers on the B. oleracea chromosomes (red). Tandem and segmental duplications are indicated by white numbers and red lines, respectively. Syntenic relationships with 10 Brassica rapa (A01 to A10) and five Arabidopsis thaliana (Chr1 to Chr5) chromosomes are represented as green and blue lines, respectively. The maps are based on orthologous pair positions, and reveal highly conserved syntenic relationships

Full size image

Gene duplication events

Gene family expansion occurs via the following three mechanisms: tandem duplication, segmental duplication, and whole-genome duplication [25]. We investigated gene duplication events to clarify the genome expansion mechanism of the B. oleracea BoCNGC superfamily. An evaluation of the physical distance between BoCNGC gene loci revealed that eight genes (i.e., BoCNGC18/BoCNGC19, BoCNGC21/BoCNGC22/BoCNGC24, and BoCNGC20/BoCNGC25/BoCNGC26) were tandemly duplicated. These genes were detected on C3, C1, and C5, respectively. The data obtained from the Plant Genome Duplication Database revealed that 13 BoCNGC genes distributed across the B. oleracea genome were associated with segmental duplications (Fig. 2b). The BoCNGC gene clusters likely formed via tandem and segmental duplication events may have expanded and enhanced the functional diversity of the gene family.

Comparative syntenic and evolutionary analyses of orthologous CNGC gene pairs

The B. oleracea and B. rapa genomes are currently divided into three sub-genomes, namely LF (least fractionated), MF-I (most fractionated), and MF-II [26]. We observed that the B. oleracea LF sub-genome contains the most BoCNGC genes (14), followed by sub-genomes MF-I (10) and MF-II (2) (Additional file 5). Because of a Brassica-lineage specific whole-genome triplication (WGT) [27], each A. thaliana CNGC gene was expected to generate three Brassica copies. However, there were 20 A. thaliana CNGC genes, 26 B. oleracea CNGC genes, and 29 B. rapa CNGC genes. To detect the retention or loss of CNGC genes after a WGT event, the syntenic map of BoCNGC genes with the model A. thaliana and B. rapa CNGC genes provided markers for defining the regions of conserved synteny among the three genomes (Fig. 2b). Compared with the ancestral Brassicaceae blocks (A to X) in A. thaliana, the synteny of 15 AtCNGC genes was preserved in Brassica species, based on the number of corresponding genes. Ten of the 20 AtCNGC genes were retained as a single copy in the equivalent blocks of both Brassica species. Three AtCNGC genes (i.e., AT2G23980, AT2G24610, and AT5G54250) located on the I and W syntenic blocks, were preserved as two copies in Brassica genomes, which were asymmetrically fractionated into three sub-genomes. Two AtCNGC genes (i.e., AT3G17690 and AT3G17700) in the F syntenic block were retained as three copies in each species. Two extra gene copies (i.e., BoCNGC20 and BoCNGC22) were located on potential overlap/tandem repeat regions of the B. oleracea genome, thus producing phylogenetic cluster IV-b. Approximately 25 B. oleracea CNGC genes and 24 B. rapa CNGC genes exhibited clear syntenic relationships among the three species. Two gene pairs (i.e., BoCNGC3 and BoCNGC23; Bra034281 and Bra029958) were not part of an A. thaliana syntenic block (Additional file 6), suggesting that these genes originated after the divergence from A. thaliana. The remaining four B. rapa genes were likely generated after the speciation event. In addition, 11 BoCNGC genes exhibited strong syntenic relationships with the genes from other plant species, implying this gene family is important for plant growth, development, and stress resistance (Additional file 6).

The orthologous CNGC gene pairs between the B. oleracea and A. thaliana genomes were used to estimate the Ka, Ks, and Ka/Ks values (Table 2). The mean Ka/Ks value of all orthologous gene pairs in the B. oleracea CNGC gene family was 1.98. Most of the BoCNGC genes had Ka/Ks ratios greater than 1. Additionally, the minimum and maximum Ka/Ks ratios were 1.05 (BoCNGC26) and 7.7 (BoCNGC6), respectively. These findings indicate that the BoCNGC gene family is under positive selection pressure, and might preferentially conserve functions and structures under this selective pressure.

Table 2 Comparative analysis of Ka, Ks and Ka/Ks values for CNGC gene pairs between B. oleracea compared to A. thaliana. Ka/Ks ratio greater than 1 indicates positive selection, a ratio less than 1 indicates functional constraint, and a Ka/Ks ratio equal to 1 indicates neutral selection
Full size table

Domain architecture and alignment of BoCNGC proteins

Domain composition analyses revealed that BoCNGC proteins contain two primary domains, namely CNBD and TM (Additional file 7). The sequence alignment of 26 BoCNGCs indicated that the two most conserved regions within the CNBD domain are a PBC, and an adjacent hinge region (Fig. 3; Additional file 8). The following highly conserved consensus motif was identified: [LI]-X(2)-[GSE]-X-[VFIY]-X-G-X(0,1)-[DE]-L-L-X-W-X-[LQ]-X(10,20)-S-X-[SAR]-X(7)-[VTI]-E-[AG]-F-X-L. This sequence can be used to classify newly annotated or un-annotated candidate sequences as Brassica CNGCs. Additionally, there was a relatively conserved IQ domain and a less conserved CaMBD adjacent to a CNBD present in 24 of the 26 BoCNGC proteins. Two proteins (i.e., BoCNGC18 and BoCNGC19) were observed to lack the CaMBD and IQ domains because their sequences are truncated at the C-terminal end of the CNBD. A high sequence divergence was noted among different groups, particularly between members of Sub-groups IV-a and IV-b. For example, the CaMBD [FLY[−X(10,12)-[AFI]-R-[FY](0,1), was not particularly conserved between Group IV-b and the other groups. However, the IQ motif [IV]-Q-X-X-W-R-X-X-X-[RKQ] was relatively conserved among the BoCNGC proteins (Fig. 3). Alignments between BoCNGCs, AtCNGCs, and BrCNGCs revealed a high sequence divergence at the C-terminal of the CNBD, in which several Group IV-b members lack the CaMBD and IQ motif (Additional files 9 and 10). Overall, our in silico analyses suggest that ion transport and CNBDs along with the PBC and hinge region are conserved in all three species, and are characteristic of plant CNGCs.

Fig. 3
figure 3

Multiple sequence alignment of BoCNGC-specific domains, and a three-dimensional model of BoCNGC1. Cartoon model with characteristic CNGC domains provided on top. The BoCNGC-specific consensus motif keys are listed below the cartoon. Amino acids allowed in a specific position are presented in square brackets. X represents any amino acid, while numbers in round brackets indicate the number of amino acids. The multiple sequence alignment of BoCNGC proteins is presented with the CNBD, CaMBD, and IQ domain indicated with different colours. The CNBD domain includes a conserved PBC and hinge region, followed by the CaMBD. Residues shaded in black and grey indicate 100% and >50% similarity among the 26 BoCNGCs

Full size image

Gene structure and motif composition analysis

To characterise the structural diversity of the BoCNGC family members, we analysed the exon–intron organization of individual BoCNGC genes. The majority of the BoCNGC genes from phylogenetic Groups I–III contained six or seven exons, while the Group IV members had 8–11 exons (Fig. 4). Closely clustered BoCNGC genes in the same clades were similar regarding the number of exons and intron lengths. Most of the introns in BoCNGC genes were phase 0 introns, which occur in between complete codons. Fifty-four phase 2 introns (i.e., located between the second and third nucleotides of a codon) were observed in the BoCNGC family, in which the genes carried two phase 2 introns. The exceptions were BoCNGC1 and BoCNGC2, which contained three phase 2 introns. Only the members of phylogenetic Group IV-b had single phase 1 introns at the terminal end of their sequences. A comparison between the exon–intron organizations of BoCNGC genes and the AtCNGC genes clustered in the same phylogenetic groups revealed several differences (Additional file 11). Most of the phase 1 introns were present in AtCNGC genes, implying that intron loss during evolution resulted in a decrease in the number of introns in BoCNGC genes, particularly those in Groups I–III and IV-a.

Fig. 4
figure 4

Schematic diagram presenting the BoCNGC gene structures and the conserved motifs in the encoded proteins. The neighbour-joining phylogenetic tree is provided on the left side of the figure, followed by the exons–introns, which are indicated as red boxes and black lines, respectively. Motifs are represented by numbers in coloured boxes on the right. Numbers [0, 1 and 2] provided on the gene structures represent the respective intron phases. The length of each exon and intron can be determined using the provided scale. The order of the motifs corresponds to the motif positions in the protein sequence. However, the length of the box does not correspond to the length of the motif

Full size image

The BoCNGC protein sequences were used for domain or motif structure analyses with the Multiple Expectation Maximization for Motif Elicitation suite [28]. Ten conserved motifs were identified. According to Pfam codes [29] and WebLogo, only seven motifs (i.e., 1–5, 7, and 10) encode domains with known functions (Fig. 4; Additional files 12 and 13). Motif 2 was the biggest motif encoding a conserved domain, which is probably associated with peptidase_C50, putative aminopeptidase, or DNA polymerase III subunit tau_4. Motifs 1 and 5, which encode a CNBD and an ion transport domain, respectively, were conserved among all BoCNGC family members. The ion transport domain had the most motifs, including motifs 4, 5, 7, and 10. The IQ CaM-binding motif (PF00612) was conserved among BoCNGC family members, with the exception of BoCNGC18, 19, and 22. Group IV proteins contained the fewest functionally annotated motifs, suggesting that the closely related proteins in each group have similar motifs and are also probably functionally similar. The functions of the remaining motifs (i.e., 6, 8, and 9) remain to be determined.

Post-translational modification and phosphorylation of BoCNGC proteins

When BoCNGC protein sequences were analysed using ScanProsite [30], multiple putative phosphorylation sites were revealed. These sites may act as substrates for several kinases, including casein kinase II, protein kinase C, tyrosine kinase, and cAMP/cGMP kinases. Protein kinase C, a family of ten isoenzymes that play a vital role in cellular signal transduction [31], were the most abundant, with 16 sites in BoCNGC4, BoCNGC5, BoCNGC8, and BoCNGC12. Casein kinase II sites, which were the most abundant in Group IV members, are reported to influence different developmental and stress responsive pathways in Arabidopsis [32]. All BoCNGC proteins had multiple N-myristoylation/N-glycosylation motif sites, which are highly conserved compared with the other PTMs. The lipid modification by N-myristoylation might controls the redox disproportions originating from different stresses in plants [33], while glycosylation is crucial for correct growth [34]. The BoCNGC5 and BoCNGC18 proteins contained the most N-myristoylation (11) and N-glycosylation (10) sites, respectively. Other PTM sites, such as those for amidations, tyrosine kinase, serine- and glutamic acid- rich regions, cell attachment sequences, and leucine zipper patterns, were less conserved and randomly distributed (Table 3). Such phosphorylations deliver effective means to regulate most physiological activities, including metabolism, transcription, DNA replication and repair, cell proliferation [35].

Table 3 The number of predicted post-translational modification sites in BoCNGC protein sequences
Full size table

Prediction of functional association network of BoCNGC proteins

To explore the relationships among different BoCNGC proteins, a hypothetical protein–protein interaction network was in silico predicted with the STRING program (accessed in April 2016) [36] and AtPID (Arabidopsis thaliana Protein Interactome Database), using using orthologous AtCNGCs as query. The STRING interaction network for the first shell of interactors of AtCNGC proteins, supported by confidence score, is presented in Fig. 5a. Fourteen AtCNGCs, having 24 orthologs in B. oleracea, interact with flagellin-sensitive 2 (i.e., FLS2 or MPL12.8), represented by association in curated databases (confidence score: 0.8). This association was traced to manually curated plant–pathogen interaction pathway imported from the Kyoto Encyclopedia of Genes and Genomes database (Additional file 14). Supported by principal component analysis, a positive interaction (confidence score: 0.154) was observed between BoCNGC10 and BoCNGC13, which are the orthologues of AtCNGC17 and AtCNGC18, respectively. In another interaction network, BoCNGC1 interacts with BoCNGC2 and BoCNGC18–26, which are orthologues of AtCNGC13, 2, 19 and 20 respectively. This interaction is based on protein homology, association in curated human pathways (http://www.reactome.org/), or genes encoding these proteins have correlated expression levels. We also observed that the Group IV proteins are associated with constitutive photomorphogenic 1 and CaM proteins (i.e., CaM4, CaM6, and CaM7) (Fig. 5a).

Fig. 5
figure 5

Hypothetical functional association network of CNGC proteins according to the orthologues relationship between Arabidopsis thaliana and Brassica oleracea. a STRING based protein-protein interaction network of CNGC proteins. Blast match results of string-network database showing similarity of CNGC-encoded proteins between of B. oleracea and A. thaliana are given in brackets (i.e., percentage and similarity index score). b AtPID (Arabidopsis thaliana Protein Interactome Database) based protein-protein interaction network of CNGC proteins. The details of each node, interaction type and reference are provided in Additional file 15

Full size image

Using orthologous Arabidopsis CNGCs as query in the AtPID uncover more potential interactions between CNGCs, and to other proteins, which are validated by experimental data from different assays (Fig. 5b; Additional file 15). The results exhibited strong interactions of co-expression and gene fusion between CNGC functional partners belonging to similar clades. For example, AtCNGC10 interacted with AtCNGC1, 3 and 13, while AtCNGC17 interacted with AtCNGC18 as mentioned earlier. AtCNGC10 interacted with more CNGCs than other proteins. In addition, some CNGCs (AtCNGC1, 5, 6, 9, 10, 13, 17, 18 and 19) interacted with many important signaling and stress related regulatory proteins, including calmodulins. These interactions are supported by data from yeast two-hybrid, and Affinity Capture-MS assays. Five CNGC genes (AtCNGC 1–4, and 11) were found to have available phenotypes of mutant data from seedlings, leaves and embryos, showing that these genes play important roles in hyper-sensitivity, pathogen and abiotic stress resistance (Additional file 15).

Additional evidence from experimental/biochemical data detected by protein kinase (MI:0424) and anti tag coimmunoprecipitation (MI:0007) assays in human putative homologs (i.e., Potassium voltage-gated channel 2 and Leucine rich repeat containing 47/Per-Arnt-Sim domain kinase) suggest a functional link between CNGCs and FLS2 [37, 38]. The experimental details and LC-MS/MS, yeast two-hybrid and phosphorylation of peptide arrays of human interacting KCNH2 and LRRC47/PASK proteins can be found in supplementary material of Behrends et al. [38]. Using Mating-Based Split Ubiquitin Assays in A. thaliana, Chen et al. [39] reported strong, positive (in both 500 μM methionine and at least one 150 μM methionine conditions), and statistically significant interaction between these protein pairs, which are required for polarized tip growth of pollen tube [40]. In another interaction network, BoCNGC1 interacts with BoCNGC2 and BoCNGC18–26, which are orthologues of AtCNGC13, 2, 19 and 20 respectively. Additionally, we observed a weak interaction (confidence score: 0.151) between AtCNGC13 (i.e., orthologues of BoCNGC1) and BRI-associated receptor kinase 1 (BAK1), which was previously observed between AtCNGC17 and BAK1 [41]. Though, it is reported that evidence transfer from one model organism to the other seems feasible approach to study interaction conservation, and it has been implemented in several frameworks already [42]. However, these experimental proofs are essential to support this analysis in B. oleracea.

Identification of microRNA target sites

Identifying the targets of the predicted microRNAs (miRNAs) may provide insights into the biological functions of miRNAs influencing plant development, signal transduction, and stress adaptations [43]. We searched for potential miRNA targets in a set of identified BoCNGC transcripts using the plant small-RNA target analysis server (psRNATarget) [44]. Using a cut-off threshold of 5 for the search parameters, we identified 14 miRNAs with target sites in 17 BoCNGC transcripts, with expectation scores of 1.5–5 (Additional file 16). To decrease the number of false positive predictions, small-RNA/target site pairs with an expectation score and cut-off threshold of 3 were considered. Consequently, five miRNAs with target sites in nine BoCNGC genes were identified (Table 4). These miRNAs were localized to the 3′ arm of the stem-loop hairpin structure. Unlike bol-miR838d, which has five target genes, the remaining miRNAs have only one target gene. Moreover, only bol-miR838d has multiple target sites (i.e., complementary regions) on BoCNGC15 and BoCNGC16 transcripts. The accessibility of the target site varied from 2.883 (bol-miR838d) to 16.4 (bol-miR5021), where lower values correspond to a greater possibility of contact between the miRNA and target site. Four miRNAs were determined to be involved in cleaving the target transcript, while two miRNAs were predicted to inhibit the translation of target genes.

Table 4 Putative microRNA targets predicted in 26 BoCNGC transcripts
Full size table

Gene ontology enrichment analysis

Using Blast2GO (v.3.3.5), we assigned 31 gene ontology (GO) classes to 26 BoCNGC genes with BLAST matches to known proteins in the InterPro database. The majority of the genes were assigned to biological process (22), followed by molecular function (7) and cellular components (3). All genes encoded integral membrane components associated with ion channel activity for transmembrane transport. Notably, BoCNGC1 was associated with salicylic acid biosynthesis, negative regulation of defence responses, regulation of plant-type hypersensitive responses, and responses to chitin. Additionally, BoCNGC6 was associated with DNA-mediated transformation (Additional file 17).

The level 2 GO enrichment analysis revealed that all 26 BoCNGC proteins are integral cell membrane components, with four proteins (i.e., BoCNGC1, BoCNGC4, BoCNGC5, and BoCNGC17) forming cell parts, and two proteins (i.e., BoCNGC4 and BoCNGC5) forming macromolecular complexes (Additional files 18-a and 19). These proteins are involved in cellular processes associated with transport, binding, and transduction (Additional files 18-b and 19). The biological process category at GO level 2 indicated that BoCNGC1 and BoCNGC17 are associated with cell death and immune responses to stimuli, while another eight CNGCs, including BoCNGC19, are associated with localization (Additional files 18-c and 19). Moreover, we mapped the 26 annotated sequences to reference pathways in the Kyoto Encyclopedia of Genes and Genomes database [45]. Twenty-four of these genes were defined as “cyclic nucleotide gated channels”, and assigned to the “plant-pathogen interaction” pathway (Additional files 14 and 20).

Expression patterns in different plant parts

We investigated the steady-state B. oleracea BoCNGC expression patterns in seven tissues (i.e., leaf, stem, callus, root, silique, flower, and bud) using Illumina RNA-sequencing data from the Gene Expression Omnibus database. Of the 26 BoCNGCs, 19 were expressed at relatively high levels (fragments per kilobase of exon model per million mapped reads value >1) in at least one tissue, including 15 in the roots and siliques, 16 in leaves, and 17 in stems, buds, and flowers. The 19 genes were also expressed in calli (Fig. 6a). Some of the syntenic duplicates have diverged in expression patterns indicating sunfunctionalization. For example, BoCNGC26 and BoCNGC19 have very similar expression patterns. But their duplicates BoCNGC21 and BoCNGC20 now have different expression patterns. An additional investigation revealed that BoCNGC17 and BoCNGC16 were the most highly expressed genes, especially in flowers, implying they may be important for Brassica species development. Among the other genes, BoCNGC3 was highly expressed in roots, while BoCNGC2 was highly expressed in siliques and calli, suggesting that the expression of this genes is induced by wounding. Most of the Group III and IV genes were expressed at low levels in the leaves, stems, calli, roots, and siliques, while BoCNGC26 was not expressed in any tissue.

Fig. 6
figure 6

BoCNGC expression profiles in different plant parts and in response to stress. a Normalized BoCNGC expression levels (fragments per kilobase of exon model per million mapped reads) in different Brassica oleracea plant parts. b Effect of biotic stress on the expression levels of BoCNGC genes in the leaves of 25 days old seedlings inoculated with Xanthomonas campestris pv. campestris at 24 hpi. c Effect of abiotic stress on the expression levels of BoCNGC genes in the leaves of 25 days old seedlings subjected to cold stress at 4 °C for 24 h. The Y-axis indicates the relative expression levels of treated versus untreated control (CK). The error bars were calculated based on three biological replicates using standard deviation. The asterisks on bars represent the statistical significance of each gene at p ≤ 0.01 based on LSD test

Full size image

A review of the reported expression profiles of orthologus Arabidopsis CNGCs in the tissues of wild and mutant plants suggest that a) the mRNAs of this gene family are expressed in all plant tissues, b) expression in leaves is greater than in roots, stem and flower, c) group-I, II and IV CNGCs are highly expressed in flowers and apex of Arabidopsis mutants (Additional file 21) [46]. Some of these observations have been confirmed during earlier investigation of CNGC mutants in Arabidopsis plants, for example AtCNGC1 [47]. Moreover, the expression patterns of BoCNGC1 and BoCNGC7 were consistent with their orthologs (ATCNGC10 and ATCNGC5), which are highly expressed in roots than leaves [7]. Our results are also corroborated by the findings of Borsics et al. [6], showing that AtCNGC10 mutant plants exhibited reduced mRNA levels in flower than its closest related member AtCNGC13 and WT plants.

Expression patterns in response to abiotic and biotic stresses

Based on the BoCNGC expression patterns in different tissues, we attempted to determine whether these genes were associated with plant defence responses, especially against race- and species-specific Brassica pathogens. Therefore, we analysed the BoCNGC expression profiles in the shoots of 25-day-old Brassica plants infiltrated with Xanthomonas campestris pv. campestris (Xcc). The BoCNGC expression levels at 24 h post-inoculation are presented in Fig. 6b. The pathogen induced considerable changes to BoCNGC expression levels, including the up-regulation of the expression of 10 BoCNGC genes in infiltrated seedlings, with the highest levels observed for BoCNGC21. This was followed by BoCNGC2 and BoCNGC1 from Group I, BoCNGC5 and BoCNGC7 from Group II, and BoCNGC26 and BoCNGC20 from Group IV-b. Interestingly, none of the Group III and IV-a genes were affected.

We also examined the BoCNGC expression levels under cold conditions. The expression of 13 of the 26 BoCNGC genes was up-regulated in cold-stressed plants, although the expression levels were lower than the levels induced by Xcc (i.e., biotic stress) (Fig. 6c). The expression levels of genes from Groups I, II, and IV were significantly induced by cold stress, with the highest levels observed for BoCNGC17 and BoCNGC23. In contrast, the Group III BoCNGCs were expressed at low levels or not at all under cold conditions. Moreover, most of the duplicated gene pairs and genes encoding interacting proteins produced similar expression patterns (especially in response to Xcc). The exception was BoCNGC24 whose expression was not significantly up-regulated like its duplicates (i.e., BoCNGC21 and BoCNGC22).

The expression patterns of many BoCNGCs under pathogen stress were consistent with the expression patterns of their Arabidopsis orthologs obtained from the AtGenExpress project (Additional file 22) [46]. The involvement of group-IV CNGCs in disease resistance and hyper-sensitivity has been documented earlier [21, 22]. However, the cumulative profiles of group-I and IV CNGCs in Arabidopsis seedlings showed apposite trend of down-regulation by cold stress at 4 °C for 24 h, showing specie-specific divergence of expression pattern.

Discussion

The CNGC gene family has been reported for many agriculturally important plants [17, 18, 20]. However, a genome-wide identification and annotation of CNGC genes has not been reported for B. oleracea. In this study, we identified 26 B. oleracea CNGC genes, and determined that the BoCNGC gene family is larger than the CNGC families of most of the reported crops [4]. The isoelectric point (pI) and charge of a protein is important for solubility, subcellular localization, and interaction, depending on both insertion and deletions between orthologs, and the ecology of the organism [48]. It is reported that proteins in cytoplasm possess an acidic pI (pI < 7.4), nuclear proteins have more neutral pI (7.4 < pI < 8.1), while those in membrane have more basic pI [48], where basic residues located on either side of membrane spanning region play a role in the stabilization of the protein in membrane [49]. The net charge of a protein is a fundamental physical property, and its value directly influences the solubility, aggregation, and crystallization of the protein [50]. The 26 BoCNGCs were localized to membranes, greatly varied in physicochemical properties, and will theoretically participate in basic buffers. These variations reflects the changes in protein composition, and their effects on association of receptors with charged ligands, folding and stability, solubilization and precipitation, and selective transport of ions in protein channels [50].

Homologous genes within the same taxonomic group are assumed to exhibit similar structural, functional, and evolutionary properties, which may help clarify the role(s) of B. oleracea CNGC genes. Because of the close relationship between B. oleracea and A. thaliana, the BoCNGC genes were highly similar (>90%) to the corresponding AtCNGC genes regarding plant CNGC-specific domains, amino acid compositions, gene structures, and phylogenetic classifications. Interestingly, we revealed the absence of the CaMBD and IQ domain in BoCNGC18 and BoCNGC19, which raised the possibility that these were abnormal CNGC proteins. However, we found that many of their homologs in A. thaliana, pear and B. rapa reportedly lack the CaMBD [18]. Similar to other CNGCs, these proteins have regular 3D structural and membrane topologies, with conserved binding sites for cGMP/cAMP. Furthermore, the presence of conserved nickel- and zinc-binding sites suggests that BoCNGC18 and BoCNGC19 may have lost their secondary domains during evolution, but gained functional diversity. Additional research is required to clarify this point.

Proteins undergo post-translational modifications (PTMs), which increase the range of their functions through different mechanisms [51]. The associated PTMs likely affected protein function, localization, and stability, as well as the dynamic interactions with other molecules [52]. Following gene annotations and phylogenetic analyses, we predicted the presence of multiple PTM sites in BoCNGCs. Apart from evolutionarily conserved PTMs, other types of modification sites were detected in BoCNGCs, which diversified the functions and underlying mechanisms of CNGC-specific PTMs. Protein–protein interaction networks provide a base for systematic understanding of cellular processes that can be used to filter and assess the functional genomics data and provide an instinctive platform to annotate the structures, functions and evolutionary properties of proteins [53]. Using two different approaches, and orthologous Arabidopsis CNGCs as a reference, we found that most CNGCs were associated with various protein–protein interaction networks involving CNGCs and other proteins related to light signalling [54], regulation of enzyme activities [55] and cellular processes [56], brassinosteroid signal transduction [57], and resistance against pathogens [58]. These aanalyses can offer new information for future experimental research and provide cross-species predictions for efficient interaction mapping [53]. Additionally, of the 26 BoCNGC genes, nine included target sites for diverse groups of novel and conserved miRNAs. These miRNA families are highly conserved in Brassicaceae species, where they are expressed in leaves, siliques, and flowers. These miRNAs are reported to function in regulation of genes related to growth (miR157/171/824) [59], Brassica-specific stomatal organization (miR824), pollen development (miR824) [60], abiotic stress tolerance, and plant–pathogen interactions (miR398) [61].

Gene duplications during evolution increase the genomic content and expand gene functions to optimise the adaptability of plants [25]. Brassica oleracea is an ancient polyploid, whose genome underwent a WGT event approximately 16 million years ago, after diverging from A. thaliana, followed by large-scale chromosomal re-arrangements (i.e., re-diploidisation). As a member of the classical triangle of U [62], the assembled genome of B. oleracea (540 Mb) is larger than that of its sister species, B. rapa (312 Mb) [63] that diverged from a common ancestor nearly 4 million years ago [64]. The less number of CNGC genes in Brassica genomes suggest that most of the duplicated gene copies were lost post-polyploidization. Reversion of the few duplicated CNGC genes to single copy might be due to neutral loss of unnecessary duplicates over time. Another possible explanation could be that CNGC proteins participate in dosage sensitive interactions that is affected by the copy number of each protein subunit (gene balance hypothesis) [24]. Synteny analysis revealed that more than 80% of the BoCNGC genes are located in conserved syntenic blocks, which lost and gained some genes. These results are consistent with the findings of Liang et al. [65]. We presume that functionally redundant gene copies are reportedly lost after genome duplication events, while some copies of functionally important genes are kept [51]. Our findings suggest that the WGT and segmental duplication events were important for the expansion of the B. oleracea CNGC family, where tandem duplications only affected the expansion of Group IV-b. Altogether, the conservation of CNGC genes after substantial genome reshuffling suggests that these genes are crucial for plant development [66]. Finally, the detailed analyses of gene expression in different tissues and under stress conditions further supported the importance of various CNGC genes for B. oleracea growth, development, and survival. To the best of our knowledge, this manuscript is the first to describe a comprehensive and systematic analysis of the B. oleracea CNGC gene family. The generated data may be useful for constructing protein–protein interaction networks and experimentally validating the miRNA targets, which regulate the development of B. oleracea. Besides, our results might help in understanding the functions of BoCNGCs related to the regulation of signal transduction pathways, and elucidate the expression profiles of the corresponding genes during plant development and stress responses. The results of the bioinformatics and comparative genomic analyses are also valuable for studying CNGC protein functions, with potential implications for the economic, agronomic, and ecological enhancement of B. oleracea and other Brassica species.

Conclusions

In conclusion, this work is the first comprehensive and systematic analyses of CNGC gene family in B. oleracea. There are 26 CNGC genes in B. oleracea, which are classified into 4 groups (I-IV) and fractionated into three sub-genomes; this gene family appears to have expanded through WGT, segmental and tandem duplication events; the BoCNGC gene family is under positive selection pressure. All the BoCNGC protein sequences contain a CNGC specific domain CNBD that comprises a PBC and a “hinge” region, featured by a stringent motif: LI]-X(2)-[GSE]-X-[VFIY]-X-G-X(0,1)-[DE]-L-L-X-W-X-[LQ]-X(10,20)-S-X-[SAR]-X(7)-[VTI]-E-[AG]-F-X-L. This study provided comprehensive information about domain structure, exon-intron structure, and the phylogenetic tree and expression analysis of CNGC genes in Chinese cabbage. These data are useful to construct protein-protein interaction network and experimentally validate the miRNA targets, which regulates and induces multiple responses in B. oleracea. The bioinformatics analysis and comparative genomic analysis also provides valuable information in the study of CNGC protein functions for the improvement of the economic, agronomic, and ecological benefits of Chinese cabbage. Furthermore, this study assists to elucidate the functions of differentially expressed candidate genes in the regulation of signal transduction pathway, plant development and stress resistance in B. oleracea.

Methods

Identification of Brassica oleracea CNGC genes

To identify the B. oleracea CNGC genes, 20 Arabidopsis CNGC protein sequences obtained from TAIR10 (https://www.arabidopsis.org/) [67] were used as queries to perform a homology-based search of the Ensembl Plants database (genome version v.2.1) [68]. This search was conducted with the default parameters of the BLASTP program. All non-redundant protein sequences were retrieved, and their domains were analysed with online servers: Simple Modular Architecture Research Tool (SMART) (http://smart.embl-heidelberg.de/) [69] and the Conserved Domains Database (CDD) (http://www.ncbi.nlm.nih.gov/Structure/cdd/wrpsb.cgi) [70]. The analyses were completed with the default cut-off parameters. Sequences containing the cNMP/CNBD (IPR000595) and transmembrane/ion transport protein (PF00520) domains as well as a plant CNGC-specific motif in the PBC and hinge region within the CNBD were recognized as CNGC proteins. The identified BoCNGC genes were named according to their positions in the phylogenetic tree.

Protein characterisation and amino acid properties

Details regarding gene and protein lengths as well as chromosomal locations were obtained from the Ensembl Plants database. Amino acid properties, including charge, molecular weight (kDa), aliphatic and instability indices, isoelectric points (pI), and grand average of hydropathy (GRAVY), were determined using the online available ProtParam tool (http://web.expasy.org/protparam/) [71]. The PTM sites were predicted with the ScanProsite web server (http://prosite.expasy.org/scanprosite/) [30].

Multiple sequence alignment and phylogenetic analysis

The identified CNGC proteins were aligned using the default settings of the ClustalX 2.0 program [72]. The conserved CNGC-specific domains were manually checked and shaded with the DNAMAN program (version 6.0.3.40; Lynnon Corporation, Quebec, Canada). The BoCNGC protein sequences were also aligned with CNGC sequences from A. thaliana and B. rapa (downloaded from the Brassica database; http://brassicadb.org/brad/) [73] using the default settings of the ClustalX 2.0 program. The alignments were viewed with the GeneDoc program [74]. A phylogenetic tree was constructed using the maximum likelihood method of MEGA 6.0 (1000 bootstrap replications) [75].

Chromosomal locations and gene duplication events

Details regarding the chromosomal locations of the BoCNGC genes were obtained from the Ensembl Plants database. The Plant Genome Duplication Database [76] was searched to identify segmentally duplicated genes. BoCNGC genes were defined as tandemly duplicated if the distance between the homologous loci was <50 kb [65]. The syntenic relationships among BoCNGCs, AtCNGCs, and BrCNGCs were evaluated using the Search Syntenic Genes tool in Bolbase [77].

Gene structure, motif composition, and prediction of three-dimensional models

Gene exon/intron structures were predicted with the Gene Structure Display Server (version 2.0) [78], with genomic and coding sequences as the input data. The conserved motifs in the CNGC sequences were identified using the Multiple Expectation Maximization for Motif Elicitation suite and the Motif Alignment and Search Tool [28] with the following parameters: optimal motif width: 6–200; maximum number of different motifs: 10. The detected motifs were annotated with Pfam [29]. Gene ontology enrichment analysis was performed using Blast2GO (v.3.3.5) [79].

Analysis of microRNA target sites and protein–protein interactions

The B. oleracea miRNA sequences obtained from the miRBase database at http://mirbase.org/ [80]. To detect potential miRNA target sites within the BoCNGC genes, the obtained miRNAs were analysed with the psRNATarget server (http://plantgrn.noble.org/psRNATarget/) [44] The information about protein-protein interaction, and available mutant information for Arabidopsis CNGC-encoded proteins was obtained from STRING (v10) [36] and AtPID (http://www.megabionet.org/atpid/webfile/query.php).

Analysis of BoCNGC transcriptome data

To investigate the BoCNGC expression profiles, we used the Illumina RNA-sequencing data available in the Gene Expression Omnibus database (accession number GSE42891) [24]. Transcript abundance was calculated as fragments per kilobase of exon model per million mapped reads, and the resulting values were log2 transformed. A hierarchical cluster was created and a heat map was generated with R language program [81].

Experimental conditions and quantitative real-time polymerase chain reaction assay

We used a quantitative real-time polymerase chain reaction (qRT-PCR) to quantify the BoCNGC expression levels in response to biotic (bacterial pathogen) and abiotic (cold) stresses. Cabbage (B. oleracea var. capitata L.) seedlings were grown for 25 days in a greenhouse at 23 ± 2 °C under natural light. For the cold stress treatment, seedlings were incubated at 4 °C for 24 h. For the bacterial infection, Xcc was first cultured in medium B [82] at 28 °C. Cells were collected by centrifugation, re-suspended in sterilized distilled water, and adjusted to an optical density at 600 nm of 0.1. The midvein of the first fully opened leaf (just above the petiole) was inoculated with the Xcc suspension using a 1-ml syringe. Sterilized ddH2O was used as the control solution. The treated plants were returned to the greenhouse and sampled 24 h later. The extraction of RNA and synthesis of cDNA were completed as previously described [20]. Gene-specific primers were designed with Primer 5.0 (Additional file 23). The qRT-PCR was conducted using a StepOne Real-Time PCR System (Applied Biosystems, USA) and SYBR Premix Ex Taq reagents (TAKARA, Japan) as described by Kabouw et al. [83]. Finally, the 2−ΔΔCt method [84] was used to calculate the relative gene expression values, which were subsequently transformed to log2- expression ratios and plotted in figures. Each experiment was performed with three technical replicates. The Actin gene (AF044573) was used as an endogenous control.

Statistical analysis

The RT-qPCR expression data was subjected to analysis of variance (ANOVA) using computer statistical package (SAS software SAS Institute, Cary, NC). Least significant difference (LSD) test at p ≤ 0.01 was used to check the significant differences between the expression levels of different BoCNGC genes compared to control.