Genome-wide identification and analysis of the CNGC gene family in maize

Identification of CNGC genes in the maize genome

A total of 20 Arabidopsis and 16 rice CNGC protein sequences were retrieved from the Arabidopsis Information Resource (TAIR10) database (http://www.arabidopsis.org/) and the Rice Genome Annotation Project (RGAP) database (http://rice.plantbiology.msu.edu/), respectively. This information was then used to identify the CNGC genes in maize. Two methods were utilized to search the maize protein sequences. One used a Hidden Markov Model (HMM) to search against maize protein sequences and the other used the local BLASTP method with a threshold e-value <1e−5. After the searches were conducted, a manual correction was performed to remove any redundancy and proteins without PBC and hinge regions within the CNBD of the CNGC proteins. To further confirm whether the ZmCNGC proteins contained the CNBD domain, the putative ZmCNGC protein sequences were submitted to the SMART (Letunic & Bork, 2018) and NCBI-CDD databases (Marchler-Bauer et al., 2017). The proteins without CNBD domains or with amino acid (aa) numbers below 200 were removed and the ZmCNGCs confirmed. Another 11 gramineae plant CNGCs were identified by applying the same method as that described above.

The PI (theoretical isoelectric point), MW (molecular weight), and GRAVY (grand average of hydropathy) of the ZmCNGCs were predicted by ExPASy (Artimo et al., 2012). CELLO v.2.5 software was used to predict the subcellular location of the ZmCNGCs (Yu, Lin & Hwang, 2010). Information about the chromosome distribution of ZmCNGCs and genetic sequences, such as DNA sequences, CDS, cDNA, and up-stream 1500 base pair (bp) DNA sequences obtained from a BLASTN search of the Ensembl Plant database (Bolser et al., 2016).

Multiple alignments, phylogenetic analysis and gene duplication analysis

Multiple sequence alignments were performed using the T-COFFEE method (Di Tommaso et al., 2011) and visualized by ESPript using the default program setting (Robert & Gouet, 2014). A maximum likelihood (ML) phylogenetic tree was constructed using the MEGA 7 software program with 1,000 bootstrap replications and the Jones-Taylor-Thornton model (Kumar, Stecher & Tamura, 2016). To further validate the accuracy of the ML tree, an un-rooted phylogenetic tree was constructed with 1,000 bootstrap replications using the MEGA 7 software program and was based on full-length protein sequence alignments. The ML analysis of 12 gramineae plants was performed using the IQTree program with the LG  + G4 and state frequencies were determined from the amino acid matrix and other default parameters (Lam-Tung et al., 2015). The tree was visualized by Evolview online (He et al., 2016). Segmental duplication between maize genes and the synteny block between maize and Sorghum, and rice and Brachypodium were obtained from the Plant Genome Duplication database (Lee et al., 2013). The substitution rates (Ka/Ks) for duplication events were calculated by the DnaSP v5 software program (Librado & Rozas, 2009), and the divergence times (Mya) were calculated using the formula: Mya = Ks ∕2λ × 10⁻⁶, where λ = 6.5 × 10⁻⁹ (Lynch & Conery, 2000).

Gene structure and conserved motif analyses

The gene structures (exon-intron) of the ZmCNGC genes were determined by the Gene Structure Display Server (Hu et al., 2015) using the CDS and DNA sequences from the ZmCNGC genes. The conserved motifs of the ZmCNGC proteins were subjected to the MEME Suite web server (Bailey et al., 2015) with the maximum number of motifs set at 10 and an optimum width range for the motifs of 6–200 aa.

Cis-acting regulatory elements and the prediction of protein–protein interaction in ZmCNGCs

The up-stream 1,500 bp DNA sequences in the ZmCNGC genes were used to locate cis-acting regulatory elements by the ‘Signal Scan Search’ programs in the NEW PLACE database (Higo et al., 1999). The interactions between ZmCNGCs and other maize proteins were predicted using the STRING 10 online program (Szklarczyk et al., 2017) and visualized by the Cytoscape v3.4.0 software program (Shannon et al., 2003).

ZmCNGC gene expression profiles and network interaction analysis

To understand the expression of ZmCNGC genes in different tissues, two high-throughput datasets for of maize were obtained from the Expression Atlas datasets (https://www.ebi.ac.uk/gxa/home/) under accession numbers E-MTAB-3826 and E-MTAB-439. These data were used to analyze the expression of ZmCNGC genes in six different tissues (ear, embryo, endosperm, pollen, root, and tassel) and at different developmental stages (embryo, endosperm, and seed). The FPKM values were used to calculate the ZmCNGC genes expressions and these were visualized by OmicShare tools, which is a free online platform for data analysis (http://www.omicshare.com/tools).

Identification of the CNGC genes in maize

A total of 20 Arabidopsis and 16 rice CNGC protein sequences which had been BLAST aligned against maize protein sequences, were used to obtain an overview of the CNGC genes in maize. After BLAST alignment, 18 putative ZmCNGC genes were identified in the maize genome. The 18 putative ZmCNGC genes were confirmed by using SMART and NCBI CDD to determine whether they contained the CNGC-specific domains (CNBD and TM). After removing redundant genes, 12 ZmCNGC genes were finally identified, which was lower than the number of CNGC genes in rice and Arabidopsis (Paterson, Bowers & Chapman, 2004; Yu et al., 2005). The predicted ZmCNGC genes were designated as ZmCNGC1 to ZmCNGC12 based on their family classification (Table 1). To further confirm the existence of the ZmCNGCs, we identified all the expressed sequence tags (ESTs) that had aligned to the ZmCNGC genes by using the BLASTN program provided by the NCBI. The results demonstrated that only ZmCNGC3 had showed no EST hits, whereas the other ZmCNGCs had more than 13 representative matches to ESTs. Five of them were located on chromosome 5, whereas the others were unevenly located on chromosomes 1, 2, 4, 6, 7, and 8. The characteristic features of these 12 ZmCNGC genes are listed in Table 1. The ZmCNGC protein lengths ranged from 326 to 745 aa with an average of 612 aa. The molecular weights of these proteins ranged from 38.63 kDa (ZmCNGC2) to 85.52 kDa (ZmCNGC8) and the calculated pI values ranged from 8.92 (ZmCNGC4) to 9.75 (ZmCNGC12). Subcellular localization prediction analysis showed that all the ZmCNGCs were localized in the plasma membrane except for ZmCNGC3 which was localized in the nuclear fraction. These results are similar to Arabidopsis (Lemtiri-Chlieh & Berkowitz, 2004).

Multiple alignments of maize CNGCs and potassium AKT/KAT channel genes

Many ion transporters other than CNGCs also have a CNBD in the C-terminus and a hexa-TM in the N-terminus. For example, potassium AKT/KAT channels (Shaker type) also contain a CNBD and a TM domain. All AKT/ KAT-type channels consist of six TM regions with one P region (Su et al., 2001). Therefore, 11 maize AKT/KAT proteins were obtained from the NCBI resource and aligned with the 12 ZmCNGC protein sequences (File S1). The results showed that the proteins were highly conserved and that all of them contained six TM domains (S1-S6) and a pore region. The PBC and hinge domain were also highly conserved. A ML phylogenetic tree indicated that maize CNGC and AKT/KAT-type channel genes were clustered into two separate sections (File S1).

The CNBD is a gene structural feature element in plant CNGCs that contains the PBC and hinge region (Diller et al., 2001). Figure 1 illustrates that within the maize PBCs, a conserved phenylalanine (F), a stabilizing glycine (G) and an acidic residue (D or E), and two aliphatic leucines (L) were 100% conserved inside the PBCs. Additionally, aromatic phenylalanine (F) and leucine (L) were 100% conserved within the hinge region. These two conserved regions occurred between the CNBD and CaMBD regions. Based on corresponding alignments with other plants, a stringent motif (L-X(2)-G-[ED]-ELL-[TSG]-W-[ACY]-L-X(10,20)-[SA]-X-T-X(7)-[EQ]-[AG]-F-X-L) which was recognized in the 12 maize CNGCs that included the PBC and hinge domain, was found to be consistent with other plant species (Saand et al., 2015b; Nawaz et al., 2014). The maize, rice, and Arabidopsis CNGCs were also aligned. The results showed that no positions were specific to the maize CNGC consensus, which suggested that the PBCs and hinge domain were highly conserved among plants (File S2).

Phylogenetic and duplication analyses of ZmCNGCs

Large phylogenetic trees with minimal homologous characters increase the likelihood of confounding relationships between different species. Therefore, we constructed a ML phylogenetic tree based on the alignment of CNGC proteins from 12 gramineae plants. The tree included 24 CNGC proteins in Aegilops tauschii, 16 in Brachypodium distachyon, 20 in Hordeum vulgare, 23 in Leersia perrieri, 21 in Nicotiana attenuata, 16 in Oryza sativa, 28 in Setaria italic, 13 in Sorghum bicolor, 79 in Triticum aestivum, 21 in Triticum urartu, and 12 maize CNGCs identified by this study (Table S1). The phylogenetic tree (File S4), was used to cluster these plant CNGC proteins into six groups, which were named Groups I, II, III, IV, IVa, and IVb, all of which had significant bootstrap values. The results showed that the CNGC proteins from B. distachyon, O. sativa, S. bicolor, and maize did not cluster in Group IV, and the N. attenuata CNGCs did not cluster in Group I. Additionally, the number in each group was unevenly distributed. Group IV was the largest with 86 genes, followed by 69 in Group III, 44 in Group II, 33 in Group IVb, 25 in Group I, and 16 in Group IVa. These data demonstrate that gramineae plant CNGC gene expansion appeared to have occurred unequally during evolution.

To better understand the evolutionary relationship among CNGC proteins, a ML phylogenetic tree was created based on the full-length protein alignments of 12 ZmCNGC s, 19 AtCNGCs and 16 OsCNGCs (Nawaz et al., 2014; Maser et al., 2001). The phylogenetic tree shows that the 47 CNGC proteins could be classified into five groups with significant bootstrap values (Fig. 2). These were named Groups I, II, III, IVa, and IVb. This was consistent with what has been previously reported for flowering plant CNGCs (Saand et al., 2015b). Group I contained three maize CNGC genes (ZmCNGC1, ZmCNGC2, and ZmCNGC3), five Arabidopsis, and two rice CNGC genes; Group II contained two maize (ZmCNGC4 and ZmCNGC5), five Arabidopsis, and three rice CNGC genes. Group III contained three maize CNGC genes (ZmCNGC6, ZmCNGC7, and ZmCNGC8); and Group IV contained five rice and four Arabidopsis CNGC genes. The maize genes in Group IV were separated into two groups, Group IVa contained ZmCNGC9 and Group IVb contained three (ZmCNGC10, ZmCNGC11, and ZmCNGC12).

The phylogenetic tree also showed that the maize CNGC genes could also be grouped into five groups (Fig. 3A). Consistent with other plant CNGC genes, Group IVa contained only one or two gene members and formed the smallest group (Saand et al., 2015b). The results suggested that during the evolution of CNGCs, Group IV CNGC genes may have emerged first whereas Groups I and II CNGC genes appeared later. Furthermore, the tree topology produced by the neighbor joining analyses was the same as the ML tree in Fig. 2, and it contained all the groups (File S3).

Additionally, only one segmental duplication gene pair, ZmCNGC10-ZmCNGC11, was formed in the maize genome (Table 2). The evolutionary process between maize CNGCs and other gramineae plants was further explored by investigating the genome synteny among Sorghum, rice, and Brachypodium. The results showed that there were two, two and one ZmCNGC genes that showed syntenic bias towards particular Sorghum, rice, and Brachypodium chromosomes, respectively (Table 2). In addition, Ka/Ks was used to evaluate their specific positions under positive selection pressure after duplication (Mayrose et al., 2007). Ka/Ks values that =1, <1 or >1 indicate neutral, purifying, and positive selection, respectively (Lynch & Conery, 2000). The Ka/Ks value of each gene pair was calculated and the Ka/Ks value for all gene pairs was less than 1, which suggested that these genes had evolved under strong purifying selection. Furthermore, the results indicated that the divergence of maize CNGCs from other gramineae plants did not occur evenly.

ZmCNGC gene structures and the conserved motif analyses

Gene structure analysis could improve our understanding of gene function and evolution. The number of introns ranged from 0 to 7 (Fig. 3B), which was different from rice and Arabidopsis CNGCs. In rice, the OsCNGC s introns numbers ranged from 1 to 11, whereas Arabidopsis CNGCs ranged from 4 to 10 (Nawaz et al., 2014). The Group IVa and IVb ZmCNGCs had distinct gene structures compared to those of the other groups, with more introns at different phases and lengths, which is consistent with most flowering plant species Group IV CNGC genes (Saand et al., 2015b).

The motif-based recognition of proteins can improve understanding of their the evolutionary history (Seoighe & Gehring, 2004). Ten putative motifs were characterized and named as motif1 to motif10 in the ZmCNGCs. The relative positions of the motifs in the five groups were found to have various patterns (Figs. 3C and 3D), for example, all the ZmCNGCs contained motif1, motif2, motif3, and motif4, which meant they presented typical ZmCNGC domains. Motif3 was a combination of the calmodulin binding (CaMB) domain and the IQ domain (QWRTWAA[CV]FIQ[AL]AW[RH]RY), and motif4 was the cyclic nucleotide-binding (CNB) domain, which was located in the C-terminal. Furthermore, motif10, 9, 5, 8 (or 6), 2, and 1 were the motif logos of the transmembrane domains. They represented the S1, S2, S3, S4, S5, and S6 regions of the transmembrane domain in the N-terminal. Motif7 represented the ion transport protein (ITP) domain. In addition, 10 of the ZmCNGCs, but not ZmCNGC2 and ZmCNGC3, possessed motif 5 and 6 that are associated with ion transport (Nawaz et al., 2014). Although other motifs have not been reported in plants or animals, they undertake important functions within the organism.

Prediction of cis-acting regulatory elements and protein-protein interactions that involve ZmCNGC proteins

Cis-acting regulatory elements are important molecular switches that are associated with the transcriptional regulation of genes when environmental stresses are encountered (Nakashima, Ito & Yamaguchi-Shinozaki, 2009). To better understand the possible biological processes undertaken by the ZmCNGCs involved, 1.5 kb sequences upstream of the ZmCNGC gene genomic sequences were used to identify cis-regulatory elements. A total of 137 different putative cis-elements were found to be associated with the identified ZmCNGC genes and only 12, including CACTFTPPCA1, EBOXBNNAPA, DOFCOREZM, MYCCONSENSUSAT, CAATBOX1, GTGANTG10, WRKY71OS, GT1CONSENSUS, ROOTMOTIFTAPOX1, POLLEN1LELAT52, MYBCORE, and OSE2ROOTNODULE, were in the promoter region of the ZmCNGC genes (Table S2), which was highly consistent with rice CNGCs. Additionally, five cis-elements were gene-specific. ACGTCBOX, TATABOX3, CTRMCAMV35S, and HDZIP2ATATHB2 were unique to ZmCNGC6, ZmCNGC7, ZmCNGC8, and ZmCNGC11, respectively. Additionally, some cis-elements were involved in responses to different abiotic/biotic stimuli, including hormones (e.g., abscisic acid, auxin, ethylene), stress (e.g., drought, temperatures, disease) and development (e.g., mesophyll specific, tissue specific) responses, which indicated that these ZmCNGC genes might be involved in regulating diverse stress responses.

The ZmCNGC protein-protein interactions were predicted in order to gain a better understanding of ZmCNGC protein roles in the plant. A total of 120 protein pairs were predicted to interact between the 12 ZmCNGC proteins and the other 11 maize proteins, and ZmCNGC8 was found to interact with ZmCNGC10 by prediction analysis (Fig. 4 and Table S3). We also used genes in Arabidopsis that were homologous to the ZmCNGC genes to predict the protein-protein interactions. The results showed that three were validated by the experimental data (Table S3).

Expression profiles of ZmCNGC genes in different tissues

The physiological function of ZmCNGC genes was investigated using transcriptome sequencing which evaluated the tissue-specific expression levels of ZmCNGC genes in different tissues (see Materials and Methods section). The expression levels among the ZmCNGC genes are shown in Fig. 5A and Table S4 suggested that their expression levels were tissue-specific. For example, ZmCNGC5 was specifically expressed in pollen compared to other tissues, which implied that it played a particular role in pollen physiological development, whereas ZmCNGC2, ZmCNGC4, ZmCNGC6, and ZmCNGC8 had higher expression levels in the roots than in the other tissues, which suggested that they had important roles in root growth and development. All Group IVb ZmCNGC genes, including ZmCNGC10, ZmCNGC11, and ZmCNGC12, were relatively highly expressed in the embryo, which implied that these genes played crucial roles in the growth and development of the maize embryo.

We also evaluated the expression of several ZmCNGC genes in the embryo, endosperm, and seed for several days after pollination. ZmCNGC3, ZmCNGC5 and ZmCNGC7 were not detected or had no expression in any of the tissues. However, ZmCNGC5 was expressed in pollen (Figs. 5A, 5B, 5C, and 5D). Cis-acting regulatory elements analysis showed that only ZmCNGC5 and ZmCNGC7 did not contain CANBNNAPA, which is the element required for embryo and endosperm-specific transcription (Ellerström et al., 1996). This might be the reason why they did not show any expression in these tissues. The embryo specific-expression gene ZmCNGC10 gradually increased over time in the embryo (Fig. 5B). ZmCNGC8 was highly expressed in the embryo, endosperm, and seeds after pollination, and the Group IVb gene showed a similar expression pattern.

Features and evolution of plant CNGC family genes

Plant CNGC family genes are characterized by the presence of a CNBD in the C-terminal and a hexa-TM in the N-terminal (Saand et al., 2015b). A total of 12 ZmCNGC genes were identified after a BLAST search against the maize genome protein sequences. Among them, ZmCNGC3 showed no ESTs and was not expressed in all the tissues, which suggested that it was a non-expressed pseudogene. In Arabidopsis, AtCNGC16 (AT3G48010) had no EST alignments and might also be a pseudogene (Maser et al., 2001). Many other ion transporters also possess these domains. For example, potassium AKT/KAT channels (Shaker type) contain both a CNBD domain and a TM domain. All AKT/ KAT-type channels consist of six transmembrane (TM) regions with one P region (Su et al., 2001). The ML phylogenetic tree showed that the maize CNGC and the AKT/KAT-type channel genes were clustered into two separate sections (File S1). Previous studies have shown that the CNGC-specific motif with a PBC and a hinge domain (L-X(2)-G-[ED]-ELL-[TSG]-W-[ACY]-L-X(10,20)-[SA]-X-T-X(7)-[EQ]-[AG]-F-X-L) only exists in maize plant CNGCs and does not occur in the ion transporters found in other plants, such as rice, Arabidopsis, and tomato (Saand et al., 2015b; Nawaz et al., 2014; Zelman, Dawe & Berkowitz, 2013). In this study, 12 ZmCNGCs were identified and the PBCs and hinge domains were highly conserved after alignment, which further confirmed the previous hypothesis (Saand et al., 2015b). The conserved motif analysis showed that the motif3 (QWRTWAA[CV]FIQ[AL]AW[RH]RY) pattern was the IQ domain among these 12 ZmCNGCs. The IQ domain is conserved among plant CNGCs and enhances the changeable Ca²⁺-dependent channel control mechanisms in plants (Fischer et al., 2013). The results from this study showed that all CNGC proteins contained the IQ motif, which suggested that they bind CaM in a Ca²⁺-dependent manner. Notably, motif4 is the sequence logo of the CNBD domain, which is conserved in most plants and animals (Jackson, Marshall & Accili, 2007).

The results showed that the CNGC family in 12 gramineae plants at various evolutionary nodes is adequate for analyzing the phylogeny and evolution of the CNGC gene family in gramineae plants. In Table S1 shows that 4 out of 12 plants under analysis contained <20 CNGC genes. These were B. distachyon (16), O. sativa (16), S. bicolor (13) and maize (12). The ML tree for the 273 CNGCs clearly indicated that gramineae CNGCs clustered into six groups (I, II, III, IV, IVa, and IVb) with significant bootstrap values; Groups IV, IVa and IVb had particularly high bootstrap support values, but group IV CNGCs did not cluster with B. distachyon, O. sativa, S. bicolor, and maize CNGCs. This suggested that there are missing CNGCs that need to be identified or there was duplication during evolution. Davidson, Guthrie & Lipsick (2013) showed that gene duplication can lead to the generation of significant numbers of novel genes. Therefore, these results imply that duplication events play a principal role in gene evolution. Phylogenetic, gene structure, and conserved motifs analyses were used to classify the 12 ZmCNGC s into five groups with significant bootstrap values. Among these five groups, Group IV ZmCNGC s emerged first and had more introns than the genes in the other groups.

CNGC genes play an important role in plant development

The CNGCs are involved in the regulation of plant growth and development (Chin, Moeder & Yoshioka, 2009b). This study focused on the role that maize CNGCs play in different tissues by investigating its expression in plant embryo, endosperm and seed expression. The results showed that some ZmCNGC genes have tissue-specific expression. For example, ZmCNGC2, ZmCNGC4, ZmCNGC6 and ZmCNGC8 were highly expressed in the roots; ZmCNGC5 was specifically expressed in pollen; and all Group IVb genes were expressed in the embryo. Group IVb genes are the evolutionary ancestors of CNGC genes and were mainly expressed in the embryo and seeds after pollination, which suggested that Group IVb genes play a significant role in embryo development. In addition, all Group IVb genes were linked to gene duplication and they had a similar expression pattern in the different tissues, which indicated that their function was to enhance adaptability during gene evolution.

Most previous studies have shown that CNGC genes are related to pollen development and responses to environmental stimuli. For example, Arabidopsis CNGC16 is critical for pollen fertility under heat and drought stress (Tunc-Ozdemir et al., 2013b), and CNGC18 has been shown to play a role in pollen tube tip growth (Frietsch et al., 2007). In rice, OsCNGC13 promotes the seed-setting rate by facilitating pollen tube growth in stylar tissues (Xu et al., 2017). ZmCNGC1 and ZmCNGC5, two homologous genes of CNGC16 and CNGC18, are mainly expressed in pollen, which indicates that they are predominantly involved in pollen development. Previous studies have shown that AtCNGC3 is mainly expressed in the embryo, leaves, and roots, and ZmCNGC4 expression level is similar to that of AtCNGC3 which was highly expressed during plant development (Kaplan, Sherman & Fromm, 2007). These results imply that these genes play crucial roles during the growth and development of maize.