Genome-wide identification, characterization and expression analysis of the monovalent cation-proton antiporter superfamily, and their function analysis in maize salt tolerance

Background Sodium toxicity and potassium insufficient are important factors affecting the growth and development of maize in saline soil. The monovalent cation proton antiporter (CPA) superfamily comprises Na+/H+ exchanger (NHX), K+ efflux antiporter (KEA), and cation/H+ exchanger (CHX) subfamily proteins, which play vital functions in maize salt tolerance. Results A total of 35 ZmCPA genes were identified in maize, and they were phylogenetically classified into 13 ZmNHXs, 16 ZmCHXs and 6 ZmKEAs. ZmCPA genes have a conserved gene structure, with the determined introns range from 11 to 25, 0 to 5 and 16 to 19 in ZmNHXs, ZmCHXs, ZmKEAs, respectively. All proteins have transmembrane domains, with an average transmembrane number of 8, 10, and 10 in ZmNHX, ZmCHX and ZmKEA proteins, respectively. Transient expression in maize protoplasts showed that ZmCHX16 and ZmNHX8 are located in the cell membrane. All ZmCHX subfamily genes showed lower expression compared to ZmNHX and ZmKEA subfamilies. Diverse expression in the 60 tissues and modulated expression in response to salt stress suggested ZmCPAs’ role in maize development and salt stress. Yeast complementary experiment revealed the function of ZmNHX8, ZmCHX8, -12, -14, -16 and ZmKEA6 in salt tolerance. Maize mutants zmnhx8 and zmkea6 further validated the important function of ZmNHX8 and ZmKEA6 in salt tolerance. Phosphorylation sites and cis-acting regulation elements analyses indicated that phosphorylation and transcriptional regulation may be involved in salt tolerance of ZmCPA genes. Conclusions Our study provides comprehensive information about ZmCPA gene superfamily, which would be useful in their future functional characterization.

ZmCPA proteins are evenly distributed throughout the protein ( Figure S7). On 1 5 average, 20.5, 23 and 18.6 phosphorylated sites were found in each member of the 1 6 three subfamilies, respectively. The cis-acting regulatory elements interacted with specific transcriptional factors 2 0 (TFs) are essential for gene expression regulation [41]. The cis-acting regulatory 2 1 elements in promoter sequences of the ZmCPA genes were predicted in PlantCARE 2 2 database (http://bioinformatics.psb.ugent.be/webtools/plantcare/html/) ( Figure S8). 2 3 On basis of functions annotations, the identified cis-acting elements were divided into 2 4 light, stress, and hormone responsive responsive categories. In light responsive 2 5 category, the G-box, Sp1, ARE, GT1-motif, the MRE, ATCT-motif and the 2 6 GATA-motif were the common elements in all ZmCPA genes. The LTR, ARE, 2 7 GC-motif, MBS and TC-rich repeats were common elements in the stress response 2 8 category. The hormone responsive category contained CGTCA-motif, ABRE, AuxRR-core, TGA-element, P-box, GARE-motif, TATC-box, TCA-element and 1

Expression analysis in different tissues and developmental stages 4
The NimbleGen maize microarray data [42] (ZM37) including 60 tissues representing 5 11 major organ systems and various developmental stages of the B73 maize inbred 6 line was employed to analyze the expression pattern of the ZmCPA genes. The gene 7 expression data of 31 ZmCPA genes including 10 ZmNHXs, 16 ZmCHXs and 5 8 ZmKEAs was used for cluster analysis. As revealed by the heatmap, all ZmNHX genes 9 except ZmNHX6 and ZmNHX13 were highly expressed in all 60 tissues ( Figure 5). 1 0 ZmNHX6 had a much higher expression level in anthers and leaf than other 58 tissues, 1 1 and ZmNHX13 was low in all 60 tissues. In case of ZmKEAs, all genes were highly 1 2 expressed in all 60 tissues. All ZmCHX genes showed low expression compared to 1 3 ZmNHX and ZmKEA subfamilies, and ZmCHX genes only exhibited high expression 1 4 in anthers ( Figure 5). 1 5 To confirm the organ-specific expression of ZmCPA genes revealed by the 1 6 microarray data, semi-quantitative reverse transcription polymerase chain reaction 1 7 (semiq-RT-PCR) of 9 ZmCPA genes was performed with total RNA isolated from the 1 8 roots, leaves, ears, immature tassel, pollens, anthers, silk and whole seed (20 days 1 9 after pollinated) of the B73 inbred line, and the primers for semiq-RT-PCR were listed 2 0 in Table S5. The semiq-RT-PCR analysis showed that the expression of 8 ZmCPA 2 1 genes was consistent with that of microarray ( Figure 6). The ZmNHX4, -5 and -8 2 2 showed specific expression in roots and leaves, while ZmCHX6, -14, -16 were 2 3 specifically expressed in pollens and anthers. ZmKEA1 and ZmKEA4 were 2 4 predominately expressed in all tested tissues except seed (20 days after pollinated). The ZmKEA6 was not included in the microarray data, and it had the same expression 2 6 pattern as other ZmKEA subfamily genes. In order to understand the expression response of ZmCPA genes to salt stress, two 1 gene subfamilies including 4 ZmNHXs and 4 ZmKEAs were chosen for expression 2 profile analysis by real-time quantitative reverse transcription polymerase chain 3 reaction (qRT-PCR) with the primers listed in Table S6 (Figure 7). This study 4 analyzed the gene expression response at 1h, 2h, 4h, and 24h after salt stress. Under 5 100 mM KCl stress condition and in root, ZmKEA4, downregulated, while other genes were upregulated at first then downregulated. In 7 leaf, all 8 ZmCPAs were upregulated. When treated with 100mM NaCl, expression of 8 , were downregulated, and ZmNHX8 were 9 upregulated after 24h, while that of the other genes did not change much in root. In 1 0 leaf, ZmKEA4, -6 and ZmNHX4, -8, -11 were upregulated, while other genes were 1 1 downregulated. In conclusion, these results implied that ZmCPAs might play a role in 1 2 salinity stress tolerance through expression regulation. To test the function of ZmCPAs in salt tolerance, the coding sequences of ZmNHX8, 1 6

ZmKEA1
ZmCHX8, -12, -14, -16 and ZmKEA6 were cloned into the yeast expression vector 1 7 pDR196 with the promoter PMA1 and then vectors were introduced into a 1 8 Saccharomyces cerevisiae mutant strain AXT3K. The strain AXT3K lacks the 1 9 function of plasma membrane Na + -ATPases (ScENA1-4), plasma membrane Na + , 2 0 K + /H + antiporter ScNHA1, and vacuolar Na + , K + /H + antiporter ScNHX1 [43]. 2 1 Therefore, it is sensitive to high Na + . The transformed yeast was grown on Arg 2 2 phosphate (AP) medium with different levels of NaCl ( Figure 8). AXT3K mutants 2 3 failed to grow in medium containing 20 mM NaCl. Expression of ZmNHX8, ZmCHX8, 2 4 -12, -14, -16 and ZmKEA6 enhanced AXT3K salt tolerance ( Figure 8). These results 2 5 indicate that ZmCPAs have the function of salt tolerance. (EMS4-02c2af) , which were produced by EMS mutagenesis of B73 inbred line, from 1 Maize EMS induced Mutant Database (MEMD) [44]. The zmnhx8 and zmkea6 2 mutants had pre-termination mutation in ZmNHX8 (Zm00001d022504) and ZmKEA6 3 gene (Zm00001d026645), causing production of truncated proteins (Figure 9). 4 Phenotype of inbred lines B73 and two maize mutants were analyzed after four days 5 of growth under salt stress. Under normal conditions, the growth status of B73 and 6 mutants was not significantly different. However, the seedling length and dry weight 7 of zmnhx8 mutant under 100 mM KCl condition were significantly lower than B73 8 without salt treatment (P<0.05). Similarly, the seedling length and dry weight of 9 zmkea6 mutant under 100 mM NaCl treatment were significantly lower than B73 1 0 without salt treatment (P<0.05) ( Figure 9). These results further verified that 1 1 ZmNHX8 and ZmKEA6 are important salt tolerance-related genes. In this study, 35 ZmCPAs were identified to analyze the function of this gene family 2 4 in maize. Earlier six NHX genes of maize have been reported in various studies [48, 2 5 49], which were probably named on the basis of their sequence similarity to known 2 6 plant CPA genes. To avoid the ambiguity, we performed nomenclature of each 2 7 ZmCPA gene following their order on the chromosomes. Analysis of chromosomal 2 8 distribution revealed that ZmCPAs were evenly distributed on the 10 chromosomes of 1 2 maize. Similarly, the CPA genes were derived from all chromosomes in wheat [41], 1 and 15 out of 17 chromosomes in pear [38], respectively. Phylogenetic tree was 2 generated using full-length CPA protein sequences of maize, rice and Arabidopsis. 3 The homologous proteins were found tightly clustered due to high homology among 4 them. Classification of the CPA superfamily genes into NHX, KEA and CHX 5 subfamilies and their further categorization into various groups such as N1-N3, 6 K1-K2 and C1-C4 has also been previously performed in Arabidopsis, pear [13,38]. 7 The sub-cellular localization predicted of different species was consistent up to 8 some extent. AtNHXs exhibited vacuole, endosome and plasma membrane 9 localization [12]. Majority of ZmNHX proteins were also predicted for similar 1 0 localization. ZmKEA2 was predicted chloroplast localization, which similar with 1 1 AtKEA1, AtKEA2 and AtKEA3. Most of the ZmCHX proteins were predicted to be 1 2 localized in plasma membrane ,which same as reported for AtCHX13 and AtCHX14 The expression pattern of the ZmCPA genes from the NimbleGen maize microarray 1 8 data showed ZmNHX6 was highly expressed in leaf and anthers, ZmNHX9 was found 1 9 to be grain specific, other ZmNHX and ZmKEA genes exhibited significant expression 2 0 during multiple developmental stages. However, ZmCHXs were specific expression in 2 1 anthers. Similar expression trend has been reported for the CPA genes in other plant 2 2 species. AtNHX1 and AtNHX2 are required for growth and floral development in 2 3 Arabidopsis [19], AtNHX5-6 are essential for normal growth and development in 2 4 Arabidopsis [20]. TaNHX2, TaNHX5 and TaNHX8 genes exhibited significant 2 5 expression during multiple developmental stages, which suggested their crucial role in 2 6 growth and development. TaKEA6 and TaKEA3 group genes were prominently 2 7 expressed in certain developmental stages of root, leaf, stem and spike, which 2 8 suggested their function in tissue development [41]. TaCHX family genes showed 1 3 distinct expression pattern where most of the genes were relatively highly expressed 1 in anthers [41], which similar to maize suggested their role in reproductive organ 2 development. At the same time, specific expression of genes was verified by semi-3 qRT-PCR. We studied expressions of the ZmCPA genes in the control and salinity 4 treatments using qRT-PCR. Four ZmKEAs in high concentration of KCl were 5 upregulated with salinity treatment in leaf, but in root, ZmKEA4, -5, -6 were 6 downregulated in high K + . In Arabidopsis, AtKEA1, AtKEA3 and AtKEA4 expression 7 was enhanced significantly under low K + stress (1mM KCl), but AtKEA2, -5, and -6 8 were not [27]. The differential expression in response to K + stress suggests that 9 ZmKEA1 involved in K + acquisition under K + conditions in maize, whereas ZmKEA4, 1 0 5 and 6 may have different functions. ZmNHX8 was upregulated in NaCl treatment, 1 1 but ZmNHX4, -5, and -11 were downregulated in root. NHX7/SOS1 is critical for 1 2 excluding Na + from plant roots [51] and ZmNHX2 is associated with a major Six ZmCPAs were cloned in the yeast expression vector pDR196 and introduced 1 7 into a yeast mutant strain AXT3K. They restored AXT3K mutant resistance to Na + . 1 8 ZmNHX8 had been verified again for its role in salt stress. AtNHX5 and AtNHX6 1 9 recovered tolerance to salt using a yeast expression system [20]. These results suggest 2 0 that ZmNHXs share a common mode of action and are involved Na + transport in 2 1 maize. Nevertheless, neither AtCHXs nor AtKEAs have been found to improve yeast 2 2 growth in salt stress [12,27]. In this study, ZmCHX8, -12, -14, -16 and ZmKEA6 2 3 recovered tolerance to high Na + . This found was different from Arabidopsis, 2 4 suggesting that the CHX and KEA subfamilies are also resistant to salt stress in maize. In the present study, we performed identification and characterization of ZmCPA 1 superfamily comprising ZmNHX, ZmKEA and ZmCHX subfamily proteins in the 1 genome of maize. Gene and proteins structure analyses suggested conserved nature of 2 evolutionary related molecules in each subfamily, however they significantly differed 3 from the members of other groups. The occurrence of high composition of helices and 4 coils in tertiary structure, and numerous TM regions supported hydrophobic 5 membrane bound nature of these proteins. Diverse occurrence of differential 6 expression in various tissues and under abiotic stress conditions indicated the 7 importance of these genes in growth and development and stress management. The 8 prediction of phosphorylation sites and cis-acting regulatory elements indicates that 9 phosphorylation and transcriptional regulation may be related to the salt tolerance of Further, these genes will also be useful in future crop improvement programs for 1 6 stress tolerance. 1 7 1 8

Materials and methods 1 9
Identification and bioinformatic analysis of ZmCPA gene superfamily. 2 0 The known CPA genes of Arabidopsis were used to query the maize AGPv4 gene set 2 1 (https://download.maizegdb.org/Zm-B73-REFERENCE-GRAMENE-4.0/) using a 2 2 local BLASTP program with an E-value <1e-10. The putatively identified sequences 2 3 were further confirmed by HMMER search for the presence of signature Na + /H + 2 4 exchanger (PF00999) domain.

5
The phylogenetic tree was constructed using full length CPA protein sequences of 2 6 maize, Arabidopsis, and rice. Alignment of the sequences was done using MUSCLE 2 7 v3.8.31 program [54], and a phylogenetic tree was built employing neighbor-joining 2 8 (NJ) method using MEGA 6.0 [55] with the following sets, bootstrap value of 1000, WoLF PSORT (https://www.genscript.com/wolf-psort.html), three in silico programs, 4 were used to predict the putative organellular localization of ZmCPA proteins. All the 5 ZmCPA proteins were modeled using SWISS-MODEL 6 (https://swissmodel.expasy.org/) [40] to simulate their 3D structures. Putative 7 conserved motifs in maize CPA proteins were identified using the MEME Suite 5.1.1 8 (http://meme-suite.org/tools/meme) with the following sets, motif length of 10-50 aa, 9 maximum number of motifs to find is 15. The DNA and transcript sequences of ZmCPA genes obtained from the maize 1 3 sequence annotation database MaizeGDB were used to design gene-specific PCR 1 4 primers with Primer3 (http://primer3.ut.ee/). DNA and cDNA sequences were 1 5 validated using PCR and RT-PCR with B73 genomic DNA and total RNA as 1 6 templates and gene-specific primers shown in Table S2. Alignment of validated DNA 1 7 and cDNA sequences of each maize CPA gene was performed to analyze the gene 1 8 structure of ZmCPA genes. Gene structure display server (GSDS 2.0) was used to 1 9 display the exon-intron structure, and intron phases [58].

Plant materials and treatments 2 2
The maize B73 inbred lines was used in this study. For qRT-PCR, the sterilized seeds 2 3 were plant in a hydroponic equipment described previously [59] with sterile water in a 2 4 greenhouse at 27/23°C with day/night of 12/12h. Four days later, the plants were For maize mutants, the sterilized seeds were cultured hydroponically in a 3 greenhouse at 27/23°C with day/night of 12/12h as above. The 1× Hoagland solution 4 was exchanged every two days. Ten days later, 100 mM KCl and 100 mM NaCl were 5 added. Four days after salt stress, phenotypic analysis was performed. Total RNA was isolated from different tissues of the B73 inbred lines, including 1 5 seedling roots, leaves, 5-cm ears, immature tassels, anthers, pollens, silks, and seeds 1 6 of 20 days after pollination, using the Trizol reagent (Invitrogen, USA) according to 1 7 the manufacturer's protocol. All RNA was purified using the DNase I (Thermo 1 8 Scientific, China). First-strand cDNA was synthesized from 1μg of total RNA (20 μ L 1 9 reaction volume) using PrimeScript™ 1st Strand cDNA Synthesis Kit (Takara, Japan) 2 0 according to the manufacturer's protocol. plasmids harboring the ZmCHX16-GFP and ZmNHX8-GPF fusion constructs each 6 was co-transfected with the OsSCAMP1-RFP construct into the protoplast cells. 7 OsSCAMP1 is a known rice secretory carrier membrane protein and used here as a 8 membrane protein control [39]. The transformed protoplast cells were cultured at 9 room temperature overnight and were observed using an Leica SP8 confocal 1 0 microscope (Leica, USA). The coding sequences of ZmNHX8, ZmCHX8,ZmCHX12,ZmCHX14,ZmCHX16,1 4 ZmKEA6 were cloned into the PDR196 vector, and then transformed into the yeast 1 5 strain AXT3K (ena1-4::HIS3,nha1::LEU2, nhx1::KanMX). The transformed yeast 1 6 cells were cultured overnight at 29°C in YPDA medium containing 1 mM KCl. Cells 1 7 were normalized in water to A 600 of 0.8. For cation tolerance testing, 5μL aliquots 1 8 from yeast cultures or 10-fold serial dilutions were spotted onto AP [65] plates 1 9 supplemented with 1 mM KCl with or without NaCl.