GsCHX19.3, a member of cation/H+ exchanger superfamily from wild soybean contributes to high salinity and carbonate alkaline tolerance

Cation/H+ exchangers (CHX) are characterized to be involved in plant growth, development and stress responses. Although soybean genome sequencing has been completed, the CHX family hasn’t yet been systematically analyzed, especially in wild soybean. Here, through Hidden Markov Model search against Glycine soja proteome, 34 GsCHXs were identified and phylogenetically clustered into five groups. Members within each group showed high conservation in motif architecture. Interestingly, according to our previous RNA-seq data, only Group IVa members exhibited highly induced expression under carbonate alkaline stress. Among them, GsCHX19.3 displayed the greatest up-regulation in response to carbonate alkaline stress, which was further confirmed by quantitative real-time PCR analysis. We also observed the ubiquitous expression of GsCHX19.3 in different tissues and its localization on plasma membrane. Moreover, we found that GsCHX19.3 expression in AXT4K, a yeast mutant lacking four ion transporters conferred resistance to low K+ at alkali pH, as well as carbonate stress. Consistently, in Arabidopsis, GsCHX19.3 overexpression increased plant tolerance both to high salt and carbonate alkaline stresses. Furthermore, we also confirmed that GsCHX19.3 transgenic lines showed lower Na+ concentration but higher K+/Na+ values under salt-alkaline stress. Taken together, our findings indicated that GsCHX19.3 contributed to high salinity and carbonate alkaline tolerance.


Results
Genome-wide identification and characterization of the CHX family genes in G. soja. To identify the CHX family genes in G. soja, we carried out Hidden Markov Model (HMM) search against G. soja proteome to identify all CPA superfamily members. A total of 55 non-redundant CPA genes were obtained, and were used to construct a phylogenetic tree with Arabidopsis CPAs (AtCHXs 8 , AtKEAs 5 and AtNHXs). Phylogenetic analysis revealed that 34 of the G. soja CPA genes formed a cluster with AtCHXs, belonging to GsCHXs (Fig. 1). Then these GsCHXs were designated according to their homologous and evolutionary relationship to AtCHXs. Basic information of all GsCHXs (including gene name, DNA, CDS, protein length, molecular weight, isoelectric point, transmembrane number, predicted localization and Na + /H + exchanger domain) is provided in Table S1. Notably, all these GsCHXs contained a Na + /H + exchanger domain in their N-terminus. What's more, they all shared 7-13 transmembrane domains and were predicted to localize on plasma membrane or other membrane-containing organelles, indicating their potential function as membrane transporters.
Phylogenetic and structural analysis of soybean CHXs. To explore the evolutionary relationship of GsCHXs, a phylogenetic tree was constructed with CHX protein sequences from G. soja and Arabidopsis. As depicted in Figs 1 and 2, similar to AtCHXs, GsCHXs were also classified into five groups (Group I-V) 8 . Among them, Group II could be further divided into two subgroups Group IIa (AtCHX3-12) and Group IIb (GsCHX3-4.4). Group IIa were previously reported to be specific to Arabidopsis 2 , we speculated that Group IIb might be peculiar to soybean, indicating the diversification of CHXs during soybean speciation.
To further examine the conservation of GsCHX protein sequences and structures, the MEME program was used to predict conserved motifs (Figs 3 and S1). Expectedly, GsCHXs displayed high similarity in protein sequences, and GsCHXs within each individual group shared common motif architecture. Among the checked 15 motifs, motif 4, 6, 7, 8, 9, 12 and 15 in the N-terminus were fragments of the well-known Na + /H + exchanger domain. Motif 2, 5, 10 and 11 composed the AANH-like domain 7 , which is beside the C-terminus of Na + /H + exchanger domain. Besides, another four conserved motifs (Motif 1, 3,13,14) were identified in the C terminus, but their functions were still unknown. Taken together, CHX family was evolutionarily conserved in terms of motif distribution, especially in each individual group.
Expression profiles of GsCHXs under carbonate alkaline stress. The CHX family genes are found to be involved in plant growth and development and it has been documented that CHX family genes functioned in response to adversity stress 4,5,12,16 . However, little is known about the roles of GsCHXs in carbonate alkaline stress. Hence, we analyzed the expression profiles of GsCHX family genes under carbonate alkaline stress (50 mM NaHCO 3 , pH8.5) based on our previous RNA-seq data ( Fig. S2 and Table S2). As depicted in Figs S2 and 4A, only five genes from Group IVa (GsCHX18.1, GsCHX18.2, GsCHX19.2, GsCHX19.3, GsCHX20) exhibited high expression levels and responded to carbonate alkaline stress (Figs S2 and 4A). Among them, GsCHX19.3 exhibited the greatest induction by carbonate alkaline stress, and the value was up to 12 fold (Fig. 4A). Therefore, we focused on GsCHX19.3 for further study.

Examination of GsCHX19.3 expression characteristics in G. soja seedlings.
To get better understanding of GsCHX19.3 expression, we firstly confirmed the induction of GsCHX19.3 under carbonate alkaline stress. In accordance with the RNA-seq data, GsCHX19.3 expression was greatly induced by carbonate alkaline stress (50 mM NaHCO 3 ) (Fig. 4B). After carbonate alkaline treatment, the transcript levels of GsCHX19.3 increased within 6 h to nearly 10 folds, and then decreased at 12 h (Fig. 4B). Considering the fact that high salt stress always occurred simultaneously with carbonate alkaline stress, we further checked the expression of GsCHX19.3 under high salt treatment (200 mM NaCl). As expected, the expression of GsCHX19.3 showed an obvious increase after high salt treatment (Fig. 4C). Overall, results represented here showed that GsCHX19.3 expression was greatly induced by both carbonate alkaline and high salt stresses (Fig. 4B,C), suggesting an important role of GsCHX19.3 in plant responses to salt-alkaline stress.
It has been reported that CHXs exhibited diverse expression patterns in different tissues during plant growth and developmental processes 8,13 . Then we carried out qRT-PCR to detect the expression patterns of GsCHX19.3 in different tissues of G. soja (including old shoot, young shoot, old leaf, young leaf, hypocotyl, root, old pod and flower). The results showed that GsCHX19.3 expressed in all detected tissues, and exhibited higher expression levels in flower and young leaf (Fig. 4D). This result was accordant with that of AtCHX19, which also displayed high expression levels in flower and leaf 8 .

GsCHX19.3 localized to the plasma membrane in onion epidermal cells. As ion transporters,
CHXs usually localize on the endomembrane system to regulate cation and pH homeostasis. Hence, to localize GsCHX19.3 protein, we constructed expression vector with GsCHX19.3 tagged at the N-terminus of yellow fluorescent protein (YFP). The AtPIP2-CFP construct was used as plasma membrane marker 20 . Then GsCHX19.3-YFP and AtPIP2-CFP were transiently co-expressed in onion epidermal cells. As shown in Fig. 5, the signal of GsCHX19.3-YFP protein was exactly overlapped with AtPIP2-CFP on plasma membrane, which was consistent with the result of PSORT prediction (Table S1). This finding was also consistent with previous report that AtCHX19 protein localized on plasma membrane 21 .
Functional complementation assays of GsCHX19.3 in yeast mutant. To test the transport activity of GsCHX19.3, we performed the functional complementation analyses by using the Saccharomyces cerevisiae strain AXT4K. Strain AXT4K was generated by deleting plasma membrane Na + -ATPase (ScENA1-4), plasma membrane Na + , K + /H + antiporter (ScNHA1), vacuolar Na + , K + /H + antiporter (ScNHX1) 22 , and endoplasmic reticulum K + /H + exchanger (ScKHA1) 10 in the W303-1B background. AXT4K was sensitive to high salt and low potassium with alkali pH 5 . Besides, AXT4K was also sensitive to hygromycin B, an aminoglycosidic antibiotic produced by Streptomyces hygroscopicus, which inhibits protein synthesis in both prokaryotic and eukaryotic cells 23 . It has been reported that ScNHX1 and AtNHX1 expression could restore AXT4K tolerance to hygromycin  Table 1. B 5 . In this study, to express GsCHX19.3 in AXT4K, its CDS region was cloned into the pDR196 expression vector and introduced into AXT4K. Growth performance of yeast cells revealed that GsCHX19.3 expression in AXT4K could also recover its hygromycin B resistance (Fig. 6A), indicating GsCHX19.3 could function normally in AXT4K.
We further checked the effects of GsCHX19.3 on ion transporting. To do this, transformed yeasts were grown on Arg phosphate (AP) or YPD medium with different levels of KCl and NaCl. As shown in Fig. S3, GsCHX19.3 expression didn't restore yeast tolerance to high K + and Na + stress under normal pH condition. We also observed that all transformants showed no growth difference on mediums with 0.01 (low K + ), 0.1 (low K + ) and 1mM KCl (normal) at normal pH (pH5.8) (Fig. 6B). These findings suggested GsCHX19.3 didn't function under either low K + , or high K + , or high Na + stress.
Previous studies have shown that AtCHX19 functioned at low K + alkalinity condition 9,11 . Hence, to further explore the function of GsCHX19.3 at low K + and high pH, the transformed yeasts were grown on low-KCl AP medium (0.01 and 0.1 mM KCl) at either pH4.5 (acidic), or pH5.8 (normal), or pH7.5 (alkali). As shown in Fig. 6C, all yeasts exhibited no growth difference on normal AP medium (pH 5.8 with 1 mM K + ). When pH lowered to 4.5, no significant difference was observed for the growth performance of different transformed yeasts. However, when the pH value was up to 7.5, only GsCHX19.3 expressed yeasts grew as well as W303-1B. Especially, only GsCHX19.3 and W303-1B could survive on AP medium with 0.01 mM KCl at pH7.5. Based on these results, we believe that GsCHX19.3 functions as an active K + uptake transporter under alkali pH.
In the view of induced expression of GsCHX19.3 under carbonate stress, we further tested the yeast growth on YPD medium with 0, 7 or 11 mM NaHCO 3 (Fig. 6D). When supplemented with 11 mM NaHCO 3 , W303-1B and GsCHX19.3 transformed yeasts grew much better than other yeast cells, which implied that GsCHX19.3 expression in AXT4K could also facilitate yeast tolerance to carbonate stress. In conclusion, results from yeast complementation assays revealed that GsCHX19.3 functioned as a K + uptake transporter at alkali pH, and also participated in carbonate stress responses in yeast cells.

GsCHX19.3 overexpression in
Arabidopsis increased plant tolerance to salt-alkaline stress. The stress induction of GsCHX19.3 expression in G. soja and its functional complementation in yeast mutant AXT4K preliminarily indicated its potential roles in K + uptake and stress responses. Hence, to further evaluate the  function of GsCHX19.3 in plants, its CDS region was cloned into pCAMBIA330035Su 24 , and transformed into wild type Arabidopsis (Columbia background) through floral dip method (Fig. 7A). Three overexpression (OX) transgenic lines were identified through semi-quantitative RT-PCR analyses (Fig. 7B), and two of them (#2 and #6) with different expression levels were used for further phenotypic assays.
By using the root length assay, we checked whether GsCHX19.3 overexpression in WT could alter high salt (NaCl) and carbonate alkaline (NaHCO 3 ) tolerance. As shown in Fig. 7C,D, all seedlings (WT, and OX lines) showed similar growth under normal condition. Under salt and alkaline stresses, primary roots of OX seedlings were significantly longer than those of WT (Fig. 7E,F) (p < 0.05 by one-way ANVOA).
To further evaluate the salt-alkaline tolerance at the adult stage, the 4-week-old WT and OX plants were watered with 150 mM NaHCO 3 (pH = 9.0). After salt-alkaline treatment for 15 days, WT exhibited chlorosis and wilting, even death, but OX plants appeared relatively healthy (Fig. 7G). Quantification analysis of the survival rate revealed that only 39.8% of WT plants survived, while the survival rates for OX lines were 70.9% (line 2) and 62.5% (line 6) (

GsCHX19.3 mediates K + uptake and Na + excretion under carbonate alkaline stress in plants.
During the long term evolution, plants have developed unique mechanisms to resist salt-alkaline stress. When suffering salt stress, plants would promote K + absorbtion in order to alleviate Na + toxic and maintain K + /Na + 25, 26 . Considering the function of CHXs in cation transporting, we further examined the Na + and K + content in WT and OX plants under both normal and salt-alkaline stress. Under normal growth condition, GsCHX19.3 OX lines exhibited significantly higher K + content than WT, but showed similar Na + accumulation to WT (Fig. 8A,B) (*p < 0.05, **p < 0.01 by Student's t-test). This implied a positive role of GsCHX19.3 in K + uptake in plants.
After carbonate alkaline treatment, all plants showed an obvious decrease in K + content, but a great increase in Na + accumulation (Fig. 8A,B) (*p < 0.05, **p < 0.01 by Student's t-test). Compared with WT, OX lines displayed much lower Na + content under stress treatment (Fig. 8B) (*p < 0.05, **p < 0.01 by Student's t-test). However, the K + content of WT is similar to that of OX lines (Fig. 8A). As a consequence, under normal condition, WT displayed lower K + /Na + ratio, compared with OX lines (Fig. 8C). After carbonate alkaline treatment, all plants showed decreased K + /Na + ratio, but the K + /Na + ratio of OX was remarkably higher than WT (Fig. 8D) (**p < 0.01 by Student's t-test).
In conclusion, above results indicated that GsCHX19.3 overexpression could increase K + absorbtion under normal condition, and reduce Na + absorbtion or increase Na + excretion to alleviate ion poison caused by salt-alkaline stress.

Discussion
Till now, soil salinization-alkalization has become a worldwide problem limiting agricultural production. Salt stress results in increased cytoplasmic Na + concentration and disrupts ion homeostasis. Consequently, excess Na + inhibits enzyme activity 27 . Therefore, once upon salt stress, plants have to start a series of signaling pathways to maintain low Na + but high K + content 26,[28][29][30] . For this reason, research on CHX family is of great importance to understand the responsive mechanism of saline-alkaline stress. The CHX family has been gradually studied in Arabidopsis 2 , rice 8,31 , and Physcomitrella patens 32 . However, no systematical study was performed for CHXs in legume plants. Glycine soja, as a sibling species of Glycine max, exhibits high carbonate tolerance and has become one of the most popular materials to study saline-alkaline stress signaling pathway 21,33,34 . In this study, we identified a total of 34 CHXs in soybean genome and found that soybean possessed about 1.2 times CHXs as many as Arabidopsis. Gene duplications were one of the reasons to force gene family expansion and generate homologous genes with similar structures and functions 35 . Soybean experienced twice gene duplications during evolution, but only once in Arabidopsis 36 , which probably contributed to more CHXs in soybean than Arabidopsis.
Studies have shown that CHX family contains a Na + /H + exchange domain in the N terminus and an AANH_ like domain close to the C terminus 7,8,12 . Besides, in this study, through searching conserved motifs, we also found another conserved domain in the long C tail of soybean CHXs (Fig. 3B). Although the function of this domain was mysterious, we posited that it might be involved in phosphorylation or protein localization, like NHX family 37 . Previous reports also showed that deletion of AtCHX17 C terminus inhibited its localization in endomembrane system 11 . Here, we proved that GsCHX19.3 also localized on plasma membrane in plants (Fig. 5). Therefore, further experiments are needed to investigate the role of the C-terminal domain of GsCHX19.3 in protein localization or activity.
Date to now, studies on CHX family genes were mainly focused on plant growth and development, but it has been reported that CHX family genes responded to adversity stress 6, 11, 12, 32, 34, 38-40 . Combined with our previous RNA-seq data 15 , we found that expression of five GsCHXs from Group IV responded to carbonate alkaline stress (Fig. S2). Interestingly, even though three GsCHX19s showed high sequence identity and close evolutionary relationship, only GsCHX19.3 displayed a greatly increased expression under carbonate alkaline treatment (Fig. 4A). Further qRT-PCR results verified its up-regulated expression under both carbonate alkaline and high salt stress (Fig. 4B,C). Besides, AtCHX16/17/18/19 from Group IV in Arabidopsis were all highly expressed in flower, leaf and root 8 . Here, we also found that GsCHX19.3 was highly expressed in flower and leaf (Fig. 4D), indicating its roles in reproductive tissue development. Above findings suggested the involvement of GsCHX19.3 in plant responses to salt-alkaline stress.
In this study, we further demonstrated the positive role of GsCHX19.3 in salt-alkaline stress responses by mediating ion transport. To identify the ion transport activity and specificity, we used a yeast mutant AXT4K, in which the plasma membrane Na + -ATPase (ScENA1-4), plasma membrane Na + , K + /H + antiporter (ScNHA1), vacuolar Na + , K + /H + antiporter (ScNHX1) 22 , and endoplasmic reticulum K + /H + exchanger (ScKHA1) 10 were knockout. Here, we proved that GsCHX19.3 expression could rescue the growth of AXT4K on the medium with normal K + (1 mM) or low K + (0.01 mM) at pH 7.5 (Fig. 6C). In previous studies, CHXs were functionally associated with K + transport, especially when K + deficiency met with high external pH 8,11,12,32 . We also showed that GsCHX19.3 overexpression in Arabidopsis increased the K + uptake under normal growth condition (Fig. 8A). When K + was limited, active transport through H + -cotransporters is needed to assist K + uptake 41 . Thus, when the medium pH was slightly alkaline, cells could use a K + /H + antiporter to accumulate K + in cytosol and release H + to extracellular to remit the alkaline damage.
Interestingly, GsCHX19.3 expression in AXT4K couldn't confer tolerance to high KCl or NaCl, even it made AXT4K more sensitive to high salt (Fig. S3). Similarly, previous studies have shown that AtCHX20 and AtCHX17 expression in yeast mutants also led to increased sensitivity to high salt stress 5,12 . In contrast, NHX1 from grape was able to rescue the growth of nhx1 yeast mutant under high salt stress 42 , and AtKEAs could mediate K + transport under high K + at normal pH (pH 5.8) 5 . Thus, the function of GsCHX19.3 was distinct from that of NHX and KEA.
However, in plants, high salt stress up-regulated the expression of GsCHX19.3 (Fig. 4C), and its overexpression also enhanced salt tolerance of transgenic Arabidopsis (Fig. 7C,E). Hence, it was worth noting that GsCHX19.3 functioned under salt stress in plants, which was different from yeast. Previous studies have shown that H + pump could generate pH gradient and thereby drive secondary transporters 43,44 . Recent studies showed that the proton pumped out by H + -ATPase could act as substrates of SOS1 (an NHX subfamily member) to promote Na + efflux, and this cooperation thereby is critical for salt stress responses at high pH 45 . Thus, we speculated the phenomenon of decreased Na + content found in transgenic Arabidopsis might be secondary effects of GsCHX19.3 transporter activity. Therefore, further work need be carried out to identify whether GsCHX19.3 could cooperate with other genes to defense salt stress in plants.
We further suggested the biological function of GsCHX19.3 under carbonate alkaline stress. First, AXT4K expressed GsCHX19.3 grew much better than AXT4K under NaHCO 3 stress. At the same time, GsCHX19.3 overexpression in Arabidopsis also enhanced carbonate tolerance (Fig. 7D,F,G,H). As we know, Na + concentration in cytoplasm of plant cells will dramatically increase under stress condition 46 . On one hand, plants will take effective measures to exclude excess cytoplasmic Na + and lower cytoplasmic Na + concentration. On the other hand, plants absorb more K + to maintain the K + /Na + balance. In this study, we found that carbonate alkaline stress increased cytoplasmic Na + content, but GsCHX19.3 OX lines showed lower Na + content than WT (Fig. 8B). This finding suggested that GsCHX19.3 overexpression could help to exclude cytoplasmic Na + , and result in a relative higher K + /Na + values under salt-alkaline stress (Fig. 8D). However, till now, no electrophysiology evidence for K + , Na + transporting by CHX proteins was reported. Further studies are needed to directly characterize the ion transport activity and specificity of GsCHX19.3 by using electrophysiology technologies, such as two-electrode, giant-patch, patch-clamp and double-barreled proton-selective microelectrode.

Materials and Methods
Identification and bioinformatics analysis of G. soja CHX gene family. The whole genome and proteome sequence of G. soja were obtained from NCBI Genome (https://www.ncbi.nlm.nih.gov/genome/13239) 34 . To identify GsCHXs, protein sequences of Arabidopsis CPA family (28 AtCHXs 2 , 6 AtKEAs 5 and 8 AtNHXs) were used to build a Hidden Markov Model (HMM) profile by using HMMER 3.0. Then the HMM profile was used in local searches of G. soja proteome. All obtained CPA protein sequences from G. soja were compared with Arabidopsis CPA protein sequences by alignment with Clustal X and analyzed with MEGA6.0 2, 47 . These proteins were examined for the existence of an intact Na + /H + exchanger domain (Pfam: PF00999) by Pfam (http://pfam.xfam.org/), and were named according to their similarity to Arabidopsis homologous genes and phylogenetic relationship. SOSUI (http://harrier.nagahama-i-bio.ac.jp/sosui/sosui_submit.html) and PSORT (http://psort.hgc.jp/) were used to predict transmembrane domains and protein localization, respectively.
The phylogenetic tree with the neighbor-joining method with 1000 bootstrap values was constructed by MEGA 6.0. Expression values of GsCHX genes under 50 mM NaHCO 3 treatment were acquired from our previous RNA-seq data 15 . The heatmap was built by using excel and adobe illustrator. The conserved motifs of GsCHXs were analyzed by MEME (http://meme.nbcr.net/meme/cgi-bin/meme.cgi).
Quantitative Real-Time PCR. G. soja (07256) was grown and treated as described 14 . 50 mM NaHCO 3 (pH = 8.5) for carbonate alkaline treatment, and 200 mM NaCl was used for high salt treatment.

Subcellular localization of GsCHX19.3-YFP fusion protein in onion epidermal cells. GsCHX19.3
without the stop codon was amplified with the forward primer (GGCTTAAUATGATGGCGACGAGTAACAAC) and the reverse primer (GGTTTAAUCCACTCGTTGGTGTGTCTGG), and then inserted into pCAM-BIA2300YFPu vector to generate GsCHX19.3-YFP with the help of USER enzyme 24 . The plasma membrane marker (AtPIP2-CFP) was obtained from the Nebenführ Lab (http://nebenfuehrlab.utk.edu/markers/default. htm) 20 . And these two vectors were introduced into Agrobacterium tumefaciens strain GV3101 together. Then the transformants were used to infect onion inner epidermal cells as described 49 . Confocal laser-scanning microscope Leica SP8 (Leica, Wetzlar, Germany) was used to detect the fluorescence at 514 nm (YFP) and 433 nm (CFP).
For functional complementation analysis in yeast, GsCHX19.3 was cloned into pDR196 by using gene specific primers containing Sal I and Xho I sites (F: 5′-ACGCGTCGACATGATGGCGAC-3′, R: 5′-CCGCTCGAGTTAACTCGTTGGTGTG-3′). Plasmids were transformed into yeast cells by using lithium acetate method. Transformants were grown on selective AP medium as described 5 . For stress tolerance tests, concentration of yeast culture were normalized to OD 600 =0.4 and then serially diluted. 1 μl aliquots of each 10-fold serial dilution were spotted onto AP plates supplemented with different concentration of KCl, or YPD plates supplemented with different contents of NaCl or hygromycine B, and incubated at 30 °C for 3 days. Drop test media should add 20 mM MES, and pH was adjusted with arginine 9 or phosphoric acid 51 .
Phenotypic analysis under salt-alkaline stress. For the root length assay, the 11-day-old WT and OX seedlings grown on normal 1/2 MS medium were transferred to fresh medium with either 0, or 100, or 115 mM NaCl or with 0, or 8, or 9 mM NaHCO 3 . After vertical growth for another 7 days, the root length was measured. Fifteen seedlings of each line were used for each experiment and the experiments were repeated for three times.
For adult stage, WT and OX were directly grown in soil pots under control conditions for 4 weeks. Then plants were watered with 150 mM NaHCO 3 (pH = 9.0) solution for salt-alkaline treatment for additional 15 days. Photographs were taken to record plant growth performance, and the survival rates were recorded.
For measurement of Na + and K + content, Arabidopsis leaves were harvested and dried at 80 °C in the oven, ground into powder, and digested with 1 mol L −1 HCl overnight. The mixture was centrifuged at 11000 rpm for 3 min, diluted with sterilized milli-Q water and analyzed for Na + and K + content in flame photometer (Aosong 6400A, Shandong, China) as described.
All of the above numerical data were subjected to statistical analyses using EXCEL 2010 and SPSS 17.0 (SPSS, Chicago, USA) statistical software by one-way ANOVA and Student's t-test.