Salt-Sensitive Signaling Networks in the Mediation of K+/Na+ Homeostasis Gene Expression in Glycyrrhiza uralensis Roots

We investigated the effects of salt-sensitive signaling molecules on ionic fluxes and gene expression related to K+/Na+ homeostasis in a perennial herb, Glycyrrhiza uralensis, during short-term NaCl stress (100 mM, 24 h). Salt treatment caused more pronounced Na+ accumulation in root cells than in leaf cells. Na+ ions were mostly compartmentalized in vacuoles. Roots exposed to NaCl showed increased levels of extracellular ATP (eATP), cytosolic Ca2+, H2O2, and NO. Steady-state flux recordings revealed that these salt-sensitive signaling molecules enhanced NaCl-responsive Na+ efflux, due to the activated Na+/H+ antiport system in the plasma membrane (PM). Moreover, salt-elicited K+ efflux, which was mediated by depolarization-activated cation channels, was reduced with the addition of Ca2+, H2O2, NO, and eATP. The salt-adaptive effects of these molecules (Na+ extrusion and K+ maintenance) were reduced by pharmacological agents, including LaCl3 (a PM Ca2+ channel inhibitor), DMTU (a reactive oxygen species scavenger), cPTIO (an NO scavenger), or PPADS (an antagonist of animal PM purine P2 receptors). RT-qPCR data showed that the activation of the PM Na+/H+ antiport system in salinized roots most likely resulted from the upregulation of two genes, GuSOS1 and GuAHA, which encoded the PM Na+/H+ antiporter, salt overly sensitive 1 (SOS1), and H+-ATPase, respectively. Clear interactions occurred between these salt-sensitive agonists to accelerate transcription of salt-responsive signaling pathway genes in G. uralensis roots. For example, Ca2+, H2O2, NO, and eATP promoted transcription of GuSOS3 (salt overly sensitive 3) and/or GuCIPK (CBL-interacting protein kinase) to activate the predominant Ca2+-SOS signaling pathway in salinized liquorice roots. eATP, a novel player in the salt response of G. uralensis, increased the transcription of GuSOS3, GuCIPK, GuRbohD (respiratory burst oxidase homolog protein D), GuNIR (nitrate reductase), GuMAPK3, and GuMAPK6 (the mitogen-activated protein kinases 3 and 6). Moreover, GuMAPK3 and GuMAPK6 expression levels were enhanced by H2O2 in NaCl-stressed G. uralensis roots. Our results indicated that eATP triggered downstream components and interacted with Ca2+, H2O2, and NO signaling to maintain K+/Na+ homeostasis. We propose that a multiple signaling network regulated K+/Na+ homeostasis in NaCl-stressed G. uralensis roots.

We investigated the effects of salt-sensitive signaling molecules on ionic fluxes and gene expression related to K + /Na + homeostasis in a perennial herb, Glycyrrhiza uralensis, during short-term NaCl stress (100 mM, 24 h). Salt treatment caused more pronounced Na + accumulation in root cells than in leaf cells. Na + ions were mostly compartmentalized in vacuoles. Roots exposed to NaCl showed increased levels of extracellular ATP (eATP), cytosolic Ca 2+ , H 2 O 2 , and NO. Steady-state flux recordings revealed that these salt-sensitive signaling molecules enhanced NaCl-responsive Na + efflux, due to the activated Na + /H + antiport system in the plasma membrane (PM). Moreover, salt-elicited K + efflux, which was mediated by depolarization-activated cation channels, was reduced with the addition of Ca 2+ , H 2 O 2 , NO, and eATP. The saltadaptive effects of these molecules (Na + extrusion and K + maintenance) were reduced by pharmacological agents, including LaCl 3 (a PM Ca 2+ channel inhibitor), DMTU (a reactive oxygen species scavenger), cPTIO (an NO scavenger), or PPADS (an antagonist of animal PM purine P2 receptors). RT-qPCR data showed that the activation of the PM Na + /H + antiport system in salinized roots most likely resulted from the upregulation of two genes, GuSOS1 and GuAHA, which encoded the PM Na + /H + antiporter, salt overly sensitive 1 (SOS1), and H + -ATPase, respectively. Clear interactions occurred between these salt-sensitive agonists to accelerate transcription of salt-responsive signaling pathway genes in G. uralensis roots. For example, Ca 2+ , H 2 O 2 , NO, and eATP promoted transcription of GuSOS3 (salt overly sensitive 3) and/or GuCIPK (CBLinteracting protein kinase) to activate the predominant Ca 2+ -SOS signaling pathway in salinized liquorice roots. eATP, a novel player in the salt response of G. uralensis, increased the transcription of GuSOS3, GuCIPK, GuRbohD (respiratory burst oxidase homolog protein D), GuNIR (nitrate reductase), GuMAPK3, and GuMAPK6 (the mitogenactivated protein kinases 3 and 6). Moreover, GuMAPK3 and GuMAPK6 expression INTRODUCTION Excess salts in the soil disrupts ion homeostasis in herbaceous and woody species (Munns and Tester, 2008;Polle and Chen, 2015). Maintaining cellular and whole-plant K + /Na + homeostasis is required for plant adaptation to salt stress (Shabala et al., 2005;Sun et al., 2009aSun et al., ,b, 2010aChen and Polle, 2010;Chen et al., 2014). The plasma membrane (PM)located H + -ATPase and Na + /H + antiporter play crucial roles in maintaining K + /Na + homeostasis in higher plants. The PM Na + /H + antiporter, salt overly sensitive 1 (SOS1), prevents excessive Na + accumulation in the cytoplasm (Zhu, 2001(Zhu, , 2016. The PM H + -ATPase sustains an H + gradient to drive Na + and H + transport across the PM (Blumwald et al., 2000). Moreover, H + -pumps preserve a less-depolarized membrane potential, thus restricting K + efflux through depolarizationactivated outward rectifying K + channels (DA-KORCs) and non-selective cation channels (DA-NSCCs, Sun et al., 2009bSun et al., , 2012aZhang et al., 2015). A large body of evidence suggests that salt-sensitive signaling molecules, such as extracellular ATP (eATP), hydrogen peroxide (H 2 O 2 ), calcium (Ca 2+ ), nitric oxide (NO), and their crosstalk contribute to the regulation of the Na + /H + antiport system (the H + -ATPase and Na + /H + antiporter). This system contributes to K + /Na + homeostasis in a variety of plant species (Zhang et al., 2007;Sun et al., 2010aSun et al., ,b, 2012aLu et al., 2013;Lang et al., 2014).
Glycyrrhiza uralensis Fisch. (Licorice), a perennial herb of the genus Leguminosae, is naturally distributed in the arid and semi-arid areas of eastern Asia (Li et al., 2016). Licorice is frequently used as a crude therapeutic medicine to protect against multiple diseases in Asian populations (Mochida et al., 2017). Apart from its pharmaceutical functions, G. uralensis is ecologically important, both for conserving soil and water and for improving soil structure in semiarid ecosystems (Zhang and Ye, 2009). The deep-rooted nature of G. uralensis plants enables them to survive desert and semi-desert habitats in northwestern China. However, how G. uralensis sustains ionic homeostasis under saline conditions and whether salt-sensitive signals contribute to the demonstrated salt tolerance have not been investigated in this liquorice species.
In the present study, we aimed to characterize the importance of Ca 2+ , H 2 O 2 , NO, and eATP in mediating Na + /H + transport in the salinized roots of G. uralensis. Flux measurements with non-invasive micro-test technology (NMT) revealed that these salt-induced signals were essential for restricting K + efflux and enhancing Na + exclusion in liquorice roots. We also screened for alterations in the transcription of genes involved in various salt-signaling pathways. We aimed to explore the network of multiple interactions among Ca 2+ , H 2 O 2 , NO, and eATP in the regulation of signaling and gene expression related to K + /Na + homeostasis in G. uralensis roots.

Plant Materials and Culture Conditions
Seeds of G. uralensis were obtained from the Mongolian Autonomous County of Hoboksar, Tarbagatay Prefecture, Xinjiang Uygur Autonomous Region (latitude 46 • 82 N, longitude 85 • 75 E). The seeds were planted in plastic pots (5 cm in diameter, 8 cm in height), containing a 2:1 mixture of sand and nursery soil, and placed in a growth chamber at Beijing Forestry University, Beijing, China. The potted G. uralensis were well irrigated, according to evaporation demand, and fertilized with one-quarter-strength Hoagland solution weekly. The temperature and relative humidity were maintained at 25-28 • C and 60-70%, respectively. A photoperiod of 14 h (9:00-23:00) was applied, and photosynthetically active radiation varied from 280 to 350 µmol m −2 s −1 . After 2 weeks of culture, rooted liquorice seedlings were transferred to 300-ml pots containing one-quarter-strength Hoagland's nutrient solution for hydroponic equilibration.

Salt Treatments
Hydroponic-equilibrated seedlings of G. uralensis were subjected to 0 or 100 mM NaCl for 24 h. Na + concentrations in root and leaf cells were examined after 6, 12, and 24 h of treatment. Na + , K + , and H + fluxes were measured along the root axes with the NMT technique. The effects of PM transporter/channel inhibitors were examined in NaCl-treated G. uralensis. A blocker of the Na + /H + antiporter, amiloride (50 µM), and a specific inhibitor of the H + -ATPase, sodium orthovanadate (500 µM), were used to inhibit the Na + /H + antiport system in the PM (Sun et al., 2009a). A typical K + channel inhibitor, tetraethylammonium chloride (TEA, 50 µM), was used to reduce NaCl-elicited K + efflux (Lu et al., 2013;Lang et al., 2014). In our study, control and NaCl-treated roots were treated with these inhibitors for 30 min before the flux recordings. In addition, two series of experiments (described below) were carried out to determine the involvement of Ca 2+ , H 2 O 2 , NO, and eATP in regulating Na + and K + fluxes and gene expression in NaCl-treated G. uralensis roots.

Series 1: Agonist Treatments
We added exogenous agonists, CaCl 2 (10 mM), H 2 O 2 (10 mM), the NO donor, sodium nitroprusside (SNP, 100 µM), and ATP-Na 2 (300 µM), and measured the effects on NaCl-induced Na + and K + fluxes in young roots of G. uralensis. The chemicals were added to one-quarter-strength nutrient solution in the presence and absence of NaCl (100 mM). Control plants treated with or without salt were cultured in nutrient solution without the application of the chemicals mentioned above. The steady-state fluxes of K + and Na + were recorded along the root axis after 24-h NaCl treatments.
We also examined the expression levels of genes involved in salt transport and signaling after salt and agonist (Ca 2+ , H 2 O 2 , SNP, and ATP) treatments. Specifically, we examined expression of the PM H + -ATPase gene, GuAHA; the PM Na + /H + antiporter gene, GuSOS1; the salt overly sensitive 3 gene, GuSOS3; the calcineurin B-like protein (CBL)-interacting protein kinase gene, GuCIPK; the respiratory burst oxidase homolog protein D gene, GuRbohD; the nitrate reductase gene, GuNIR; and the mitogen-activated protein kinases 3 and 6 genes, GuMAPK3 and GuMAPK6.

Series 2: Antagonist Treatments
Control and NaCl (100 mM, 24 h)-stressed G. uralensis seedlings were treated with or without pharmacological agents for 30 min. These agents were: LaCl 3 , an inhibitor of the PM Ca 2+ channel (5 mM); DMTU, a ROS scavenger (5 mM); cPTIO, a scavenger of NO (300 µM); and PPADS, an antagonist of animal PM P2 receptors (300 µM) (Sun et al., 2010a,b;Chen et al., 2013;Zhao et al., 2016). Next, young roots with apices of 2.0-3.0 cm were sampled and equilibrated in measuring solution for 30 min. Then, steady-state fluxes of K + and Na + along the root axes were recorded in plants after treating with NaCl and antagonist (LaCl 3 , DMTU, cPTIO, and PPADS). We also examined the abundances of GuAHA and GuSOS1 transcripts in these roots.

Protocols for NMT Recording
We used the NMT technique (NMT-YG-100, Younger United States LLC, Amherst, MA, United States) to measure the net Na + , K + , and H + fluxes in G. uralensis roots. The microelectrodes were prepared and calibrated as previously described (Sun et al., 2009a,b;Lang et al., 2014).
After roots were exposed to NaCl treatment, with either an agonist (Ca 2+ , H 2 O 2 , SNP, and ATP) or an antagonist (amiloride, sodium orthovanadate, TEA, LaCl 3 , DMTU, cPTIO, and PPADS), root segments with 2.0-3.0 cm apices were selected and washed two or three times with redistilled water. When placed in a buffer with a lower Na + concentration, the preloaded Na + would diffuse from the surface of salt-stressed roots. To decrease the effect of this excess salt release on flux recordings, roots were equilibrated prior to flux recordings in a measuring solution (0.1 mM NaCl, 0.1 mM MgCl 2 , 0.1 mM CaCl 2 , and 0.5 mM KCl) for 30 min. The concentrations of Ca 2+ and K + in the measuring solution were set to 0.1 and 0.5 mM, respectively (Li et al., 2012), to reduce interference from Ca 2+ and K + on the Na + electrodes (Cuin et al., 2011). The pH of the measuring solution was adjusted to 5.7 with HCl and KOH.
After equilibration, roots were immobilized on the bottom of a measuring chamber with 10 ml of fresh measuring solution. Flux measurements were started at 200 µm from the root apex and conducted along the root axis, up to 2700 µm from the root apex, at intervals of 200 or 300 µm (vigorous ion fluxes were typically observed at the apical regions; Lu et al., 2013;Lang et al., 2014). A 6-8 min continuous recording was performed at each measuring point in the apical zones. Five or six individual seedlings were measured from each treatment group.

Na + Visualization within Root and Leaf Cells
To evaluate the NaCl-induced Na + distribution in G. uralensis roots and leaves, we used a specific fluorescent probe, CoroNa-Green AM (Sun et al., 2012a). Two-week-old seedlings were exposed to 0 or 100 mM NaCl for 6, 12, or 24 h. Then, the roots and leaves were exposed to CoroNa-Green AM (20 µM) for 2 h in a 5 mM Mes-KCl loading buffer (pH 5.7). Cellular Na + was visualized with a Leica SP5 confocal microscope (Leica Microsystems GmbH, Wetzlar, Germany). The confocal settings were as follows: excitation 488 nm, emission 510-530 nm, frame = 512 × 512.

Cytosolic Ca 2+ , H 2 O 2 , and NO Levels in Roots
In G. uralensis roots, we used specific fluorescent probes to detect cellular signal contents. We used Rhod-2 AM (Biotium) to detect cytosolic Ca 2+ (Sun et al., 2012a;Zhang et al., 2015); H 2 DCF-DA (Eugene) to detect H 2 O 2 (Sun et al., 2010a,b); and DAF-FM DA (Eugene) to detect NO (Sun et al., 2012a). Briefly, young roots were exposed to 0 or 100 mM NaCl for 30 min. Then, the roots were transferred to a 5 mM Mes-KCl loading buffer (pH 5.7) containing 2 µM Rhod-2 AM, 50 µM H 2 DCF-DA, or 10 µM DAF-FM DA. The staining was performed in the dark for 1 h at room temperature. Next, the roots were washed 4-5 times with Murashige and Skoog (MS) liquid medium prior to confocal microscope measurements. The confocal settings were as follows: excitation 488 nm, emission 510-530 nm for H 2 DCF-DA and DAF-FM DA; and excitation 543 nm, emission 570-590 nm for Rhod-2 AM (frame = 512 × 512).

Extracellular ATP in Roots
Extracellular ATP levels were monitored with the Enlighten ATP assay system bioluminescence kit (Promega, Madison, WI, United States; Sun et al., 2012a;Deng et al., 2015). Briefly, G. uralensis roots were exposed to 0 or 100 mM NaCl at room temperature. The liquid culture medium of control and NaCltreated roots was sampled at 0, 5, 10, 20, 40, 60, 120, and 240 min, then immediately frozen in liquid nitrogen. eATP was measured in an assay with luciferin-luciferase Turner Designs Modulus TM Microplate Multimode Reader (Promega Corp., Madison, WI, United States). The eATP levels were calculated, based on a standard curve created by measuring a linear range (0.01-100 nM) of standard eATP concentrations (Sun et al., 2012a;Deng et al., 2015).

Quantitative Real-time PCR Analysis
The transcription levels of genes related to the PM Na + /H + transport system and salt signaling were evaluated in salt-stressed plants. Quantitative real-time PCR assays were conducted according to Deng et al. (2015) with some modifications. Briefly, total RNA was isolated from G. uralensis roots with TRIzol reagent (Invitrogen). DNA was eliminated by treating for 0.5 h with DNase I (Promega). An aliquot of purified RNA (1 µg) was used as template for first strand cDNA synthesis with M-MLV reverse transcriptase (Promega) and oligo (dT) primers. Specific primers for GuAHA, GuSOS1, GuSOS3, GuCIPK, GuRbohD, GuNIR, GuMAPK3, and GuMAPK6 were designed, based on homologous sequences found in Populus trichocarpa or Arabidopsis. Forward and reverse primers are listed in Supplementary Table 1. Amplification was performed as described by Ding et al. (2010): 95 • C for 5 min, followed by 32 cycles of 94 • C for 30 s, 55 • C for 30 s, and finally, 72 • C for 30 s, with a final step of 72 • C for 10 min. The transcripts of target genes were normalized to the expression level of the G. uralensis β-actin 2 gene (GuACT2), and relative expression was calculated with the 2 − C T method (Livak and Schmittgen, 2001). Each experiment was replicated at least three times, and mean values are shown.

Data Analysis
Ion fluxes were evaluated with JCal V3.0, which was created by Yue Xu 1 . In the present study, positive values denote cation efflux and negative values denote cation influx. All experimental data were processed with SPSS 17.0 for statistical tests. Data were subjected to an Analysis of Variance (ANOVA), and comparisons between means were performed with Duncan's multiple range test. P-values less than 0.05 were considered statistically significant.

Na + Levels in Root and Leaf Cells
The Na + concentrations in roots and leaves of G. uralensis were detected with a Na + -sensitive fluorescent dye, CoroNa-Green AM. The Na + fluorescence in roots and leaves increased with the duration of salt exposure (6, 12, and 24 h) (Figure 1). Intracellular Na + was detected as a bright green fluorescence, which was typically observed in vacuoles (Figure 1). However, Na + levels in roots were 1.69-to 2.40-fold higher than that in leaves over the observation period (Figure 1). This result indicated that G. uralensis roots could take up and accumulate high Na + within a short period of salt treatment. Therefore, the roots were used to evaluate the effects of salt signaling molecules on ion fluxes and gene transcription.

Effect of Signaling Molecules on NaCl-Induced Ion Fluxes
Na + Flux Under no-salt control conditions, G. uralensis roots exhibited stable, constant Na + efflux along the root apex, with a mean value of 37.89 pmol cm −2 s −1 (Figure 4A). After exposure to NaCl (100 mM) for 24 h, Na + efflux along the root tip significantly increased to 315.24 pmol cm −2 s −1 (Figure 4A). Of note, the FIGURE 1 | Na + concentrations in root and leaf cells of NaCl-stressed Glycyrrhiza uralensis. Two-week-old G. uralensis seedlings were exposed to 0 or 100 mM NaCl for 6, 12, and 24 h in one-quarter-strength Hoagland solution, then stained with CoroNa-Green AM to detect Na + concentrations. Representative confocal images of (A,B) roots (scale bar: 250 µm) and (C,D) leaves (scale bar: 50 µm) show the Na + content (bright green fluorescence). The orange-red color is chlorophyll autofluorescence. The mean value (±SD) of 4-5 independent experiments is shown in the left bottom corner of each panel, and different letters (a and b) denote significant differences (P < 0.05) between control (-NaCl) and salt treatment (+NaCl). maturation region (1700-2000 µm from the apex) displayed 10-20% higher Na + efflux than the meristematic zone (200 µm from the apex).
Under NaCl exposure, the addition of 10 mM Ca 2+ markedly increased the Na + efflux by 82% in the measured root regions ( Figure 4A). However, the addition of LaCl 3 (5 mM), an inhibitor of Ca 2+ -channels in the PM, markedly reduced the salt-elicited Na + efflux ( Figure 4A). Compared to NaCl treatment, in no-salt control conditions, exogenously applied CaCl 2 or LaCl 3 had no significant effect on root Na + flux with the exception of a few measuring points (Figure 4A).
Pharmacological experiments revealed that salt-elicited Na + efflux was significantly suppressed by amiloride (an inhibitor of the Na + /H + antiporter) or sodium orthovanadate (a specific inhibitor of the PM H + -ATPase) ( Figure 5A). Moreover, steadystate recordings showed that these inhibitors markedly decreased the H + influx induced by salt treatment (Figure 5B). These results indicated that salt-stimulated Na + efflux was due to active Na + extrusion, i.e., Na + /H + antiport across the PM, in this medicinal plant.
Under short-term NaCl stress, exogenously applied H 2 O 2 (10 mM), SNP (a NO donor, 100 µM), or ATP (300 µM) FIGURE 2 | Cytosolic Ca 2+ , H 2 O 2 , and NO levels within root cells in NaCl-stressed G. uralensis. Two-week-old G. uralensis seedlings were exposed to 0 or 100 mM NaCl for 30 min in one-quarter-strength Hoagland solution, then stained with specific fluorescent probes for detecting Ca 2+ (Rhod-2, orange-red), H 2 O 2 (H 2 DCF, green), and NO (DAF-FM, green). Representative confocal images (scale bar: 250 µm) show (A) control and (B) NaCl-stressed roots. The bright green fluorescence corresponded to the detection of H 2 O 2 and NO, while the orange-red color is the Ca 2+ fluorescence. The mean value (±SD) of 4-5 independent experiments is shown in the left corner of each panel, and different letters (a and b) denote significant differences (P < 0.05) between control (-NaCl) and salt treatment (+NaCl).
FIGURE 3 | NaCl-induced alterations in extracellular ATP (eATP) in G. uralensis roots. Two-week-old G. uralensis seedlings were exposed to 0 or 100 mM NaCl for 4 h in one-quarter-strength Hoagland solution. Extracellular ATP was detected with the Enlighten ATP assay system bioluminescence kit. Each value (±SD) is the mean of 4-5 independent experiments. * P < 0.05, compared to no-salt control.

K + Flux
Non-salinized G. uralensis roots displayed a stable, constant K + efflux with a mean of 50.27 ± 8.49 pmol cm −2 s −1 (Figure 4B). Salt treatment markedly increased the K + efflux, up to 151.35 pmol cm −2 s −1 in the measured regions (200-2700 µm from the apex) ( Figure 4B). Inhibitor experiments showed that the salt-induced K + loss was inhibited by a K + channel blocker, TEA ( Figure 5C). In contrast to TEA, sodium orthovanadate, the specific inhibitor of the PM H + -ATPase, markedly enhanced the salt-elicited K + loss from liquorice roots ( Figure 5C). This indicated that the K + loss in salt-stressed roots was due to activation of DA-KORCs or NSCCs in the PM (Sun et al., 2009b;Zhang et al., 2015).
Of note, Ca 2+ , H 2 O 2 , SNP, or eATP reduced K + efflux by 12-59% in salinized roots, although the effect H 2 O 2 was more pronounced than that of the other agonists (Figures 4B, 6B-8B). In contrast, salt-induced K + efflux was significantly enhanced by all the tested antagonists, LaCl 3 , DMTU, cPTIO, and PPADS (Figures 4B, 6B-8B). In general, none of the signaling molecules (Ca 2+ , H 2 O 2 , SNP, or eATP) or the inhibitors (LaCl 3 , DMTU, FIGURE 4 | Effects of CaCl 2 and LaCl 3 on Na + and K + fluxes in G. uralensis roots under salt stress. Roots were untreated (control, blue) or exposed to CaCl 2 (10 mM, red) for 24 h in the absence (-NaCl) and presence of NaCl (100 mM). For inhibitor treatment, control and NaCl-stressed roots were subjected to LaCl 3 (5 mM, green) for 30 min. Steady-state flux profiles of (A) Na + and (B) K + were measured along the root axis at the apical zones (200-2700 µm from the root tip) in no-salt (left panels) and salt-stressed (center panels) conditions. Each point represents the mean of five to six individual plants. * P < 0.05, * * P < 0.01, * * * P < 0.001, compared to controls. (Right panels) Means of Na + and K + fluxes at all measurement points, in no-salt (-NaCl) and salt-stressed (+NaCl) plants. Bars (±SD) represent the means of five to six individual plants; different letters (a, b, c, and d) indicate significant differences (P < 0.05) between treatments.

Effect of Signaling Molecules on NaCl-Induced Transcription of K + /Na + Homeostasis Genes
GuAHA and GuSOS1 NaCl treatment (100 mM, 24 h) induced significant increases in the expression of Na + /H + antiport system genes, GuAHA (PM H + -ATPase gene) and GuSOS1 (PM Na + /H + antiporter gene) (Figure 9). Interestingly, exogenously applied Ca 2+ , H 2 O 2 , SNP, or eATP increased the expression of GuAHA and/or GuSOS1 under NaCl stress (Figure 9). These data suggested that Ca 2+ , H 2 O 2 , SNP, and eATP were involved in regulating the transcription of the PM Na + /H + antiport system. Accordingly, pharmacological data showed that the salt-elicited upregulation of GuAHA and GuSOS1 could be suppressed by DMTU, cPTIO, or PPADS (Figure 9). However, the Ca 2+ -channel inhibitor, LaCl 3 , did not block the salt-induced upregulation of GuAHA and GuSOS1 transcription (Figure 9). Moreover, we found that these salt signaling molecules and pharmacological agents had no obvious effects on gene expression in the absence of NaCl stress, with the exception of H 2 O 2 , which induced GuAHA expression in control conditions (Figure 9).

Salt-Responsive Genes Related to Signaling Pathways
As shown in Figure 10, NaCl increased the transcription of a series of salt-responsive genes. GuSOS3 is important in Ca 2+ signaling pathways (Zhu, 2001(Zhu, , 2003(Zhu, , 2016Qiu et al., 2002;Yang et al., 2009;Ji et al., 2013); GuCIPK is important in Ca 2+ signaling pathways (Xiang et al., 2007;Hu et al., 2015); GuRbohD is important in H 2 O 2 signaling (Rejeb et al., 2015); GuNIR is important in NO signaling ; and GuMAPK3 and GuMAPK6 are important in eATP signaling (Choi et al., 2014). We found that several signaling molecules changed the expression pattern of the selected salt-responsive genes under salt stress. For example, exposing NaCl-stressed plants to Ca 2+ , H 2 O 2 , or SNP enhanced transcription of GuSOS3 or GuCIPK (Figure 10). Of note, eATP produced a pronounced induction of Ca 2+ signaling pathway genes; the expression levels of both GuSOS3 and GuCIPK were stimulated by eATP in NaCl-stressed roots (Figure 10). Also, GuRbohD transcription was enhanced by these signaling molecules, but Ca 2+ and eATP produced more pronounced effects than H 2 O 2 and SNP (Figure 10). GuNIR expression remained constant in NaClstressed roots, regardless of Ca 2+ , H 2 O 2 , or SNP treatment (Figure 10). However, GuNIR transcription was enhanced with eATP in salinized G. uralensis roots (Figure 10). The abundances of GuMAPK3 and/or GuMAPK6 transcripts increased in the presence of all signaling molecules, but H 2 O 2 and eATP produced more pronounced effects on GuMAPK6 (Figure 10). We also noticed that, in general, Ca 2+ , H 2 O 2 , and eATP increased the expression of the tested salt-responsive genes under nosalt control conditions; in contrast, SNP had less of an effect (Figure 10).

NaCl Increased Salt-Sensitive Signaling Molecules in G. uralensis Roots
A short period of NaCl exposure caused cellular Na + accumulation, which was more pronounced in roots than FIGURE 5 | Effects of amiloride, sodium orthovanadate, and tetraethylammonium (TEA) on Na + , K + , and H + fluxes in G. uralensis roots under salt stress. Roots were exposed to 0 (control) or 100 mM NaCl (NaCl) for 24 h, then exposed to transporter/channel inhibitors for 30 min. Steady-state fluxes were measured along the root axis at the apical zones (200-2700 µm from the root tip). Mean fluxes of (A) Na + and (B) H + were measured in the absence and presence of inhibitors, (left) amiloride (50 µM) and (right) sodium orthovanadate (500 µM). (C) The mean K + flux was measured in the absence and presence of inhibitors, (left) TEA (50 µM) and (right) sodium orthovanadate (500 µM). Bars (±SD) represent the means of five to six individual plants; letters (a, b, c, and d) indicate significant differences between treatments (P < 0.05).
in leaves (Figure 1). The buildup of Na + in root cells resulted in remarkable increases in cytosolic Ca 2+ , H 2 O 2 , NO, and eATP (Figures 2, 3). The rapid increase of these signaling molecules indicated that G. uralensis roots could sense NaCl stress, and they set into motion a wide range of cellular processes required for salt adaptation (Chen and Polle, 2010;Chen et al., 2014;Polle and Chen, 2015). Accordingly, our NMT and RT-qPCR data revealed that eATP, Ca 2+ , H 2 O 2 , NO, and their interactions played crucial roles in regulating ion fluxes and gene transcription (Figures 4-10). These findings were similar to findings from our previous study in a salt-resistant poplar, P. euphratica (Sun et al., 2010a(Sun et al., ,b, 2012a.

Signaling Molecules Upregulated
Expression of the PM H + -ATPase and the Na + /H + Antiporter Na + Homeostasis To avoid toxicity, due to excessive Na + in the cytosol, it is crucial for glycophyte plants to adapt to saline conditions (Shabala et al., 2005;Sun et al., 2009aSun et al., ,b, 2010aChen and Polle, 2010;Chen et al., 2014). The perennial species, G. uralensis, exhibited significant Na + extrusion and a corresponding H + uptake after exposure to 24-h NaCl treatments (Figures 4, 5). However, the salt-induced Na + efflux and H + influx were markedly FIGURE 6 | Effects of H 2 O 2 and DMTU on Na + and K + fluxes in G. uralensis roots under salt stress. Roots were untreated (control, blue) or exposed to H 2 O 2 (10 mM, red) for 24 h in the absence (-NaCl, no-salt) and presence of NaCl (100 mM). For inhibitor treatment, no-salt and NaCl-stressed roots were subjected to DMTU (5 mM, green) for 30 min. Steady-state flux profiles of (A) Na + and (B) K + were measured along the root axis at the apical zones (200-2700 µm from the root tip) in no-salt (left panels) and salt-stressed (center panels) conditions. Each point represents the mean of five to six individual plants. * P < 0.05, * * P < 0.01, and * * * P < 0.001, compared to controls. (Right panels) Bars (±SD) represent the mean of five to six individual plants; letters (a, b, c, and d) indicate significant differences (P < 0.05) between treatments.
FIGURE 7 | Effects of sodium nitroprusside (SNP) and cPTIO on Na + and K + fluxes in G. uralensis roots under salt stress. Roots were untreated (control, blue) or exposed to SNP (100 µM, red) for 24 h in the absence (-NaCl, no-salt) and presence of NaCl (100 mM). For inhibitor treatment, no-salt and NaCl-stressed roots were subjected to cPTIO (300 µM, green) for 30 min. Steady-state flux profiles of (A) Na + and (B) K + were measured along the root axis at the apical zones (200-2700 µm from the root tip) in no-salt (left panels) and salt-stressed (center panels) conditions. Each point represents the mean of five to six individual plants. * P < 0.05, * * P < 0.01, * * * P < 0.001, compared to controls. (Right panels) Bars (±SD) represent the mean of five to six individual plants; letters (a, b, c, and d) indicate significant differences (P < 0.05) between treatments.
blocked by amiloride (an inhibitor of Na + /H + antiporters) or sodium orthovanadate (a specific inhibitor of the PM H + -ATPase) (Figure 5). These results suggested that salinized roots of G. uralensis extruded Na + and took up H + via the activated Na + /H + antiport system in the PM (i.e., the H + -ATPase and Na + /H + antiporter; Shabala et al., 2003Shabala et al., , 2005Sun et al., 2009a;Lu et al., 2013;Lang et al., 2014;Zhao et al., 2016). Notably, we found that the Na + efflux was enhanced by Ca 2+ , H 2 O 2 , NO, and eATP (Figures 4, 6-8). Moreover, salt-induced Na + extrusion could be reduced by pharmacological agents that FIGURE 8 | Effects of eATP and PPADS on Na + and K + fluxes in G. uralensis roots under salt stress. Roots were untreated (control, blue) or exposed to ATP-Na 2 (300 µM, red) for 24 h in the absence (-NaCl, no-salt) and presence of NaCl (100 mM). For inhibitor treatment, no-salt and NaCl-stressed roots were subjected to PPADS (300 µM, green) for 30 min. Steady-state flux profiles of (A) Na + and (B) K + were measured along the root axis at the apical zones (200-2700 µm from the root tip) in no-salt (left panels) and salt-stressed (center panels) conditions. Each point represents the mean of five to six individual plants. * P < 0.05, * * P < 0.01, and * * * P < 0.001, compared to controls. (Right panels) Bars (±SD) represent the mean of five to six individual plants; letters (a, b, c, and d) indicate significant differences (P < 0.05) between treatments.
blocked the pathways regulated by those molecules, i.e., LaCl 3 , DMTU, cPTIO, and PPADS, respectively (Figures 4, 6-8). These results indicated that the signaling molecules were required to activate the PM Na + /H + antiport system in the presence of NaCl salinity.
Our RT-qPCR assays showed that the activated Na + /H + antiport system in salinized roots presumably resulted from the upregulation of GuSOS1 and GuAHA genes (Figure 9). In a previous study, eATP was found to mediate the induction of PeSOS1 and PeAHA in the poplar, P. euphratica, during NaCl stress (Sun et al., 2012a;Zhang et al., 2015). Moreover, NO was found to enhance Na + exclusion by increasing the expression of the PM H + -ATPase and Na + /H + antiporter in a secretor mangrove, Avicennia marina, under high salinity . Our previous study revealed that NO most likely interacted with Ca 2+ and H 2 O 2 in Aegiceras corniculatum to up-regulate FIGURE 10 | Effects of Ca 2+ , H 2 O 2 , SNP, and eATP on relative expression of salt-responsive genes in G. uralensis roots under salt stress. Roots were exposed for 24 h to 0 or 100 mM NaCl, supplemented with or without CaCl 2 (10 mM), H 2 O 2 (10 mM), SNP (a NO donor, 100 µM), or ATP-Na 2 (300 µM). Quantitative RT-PCR results show the relative transcript abundance of homolog genes in G. uralensis, such as GuSOS3 (salt overly sensitive 3), GuCIPK (CBL-interacting protein kinase), GuRbohD (respiratory burst oxidase homolog protein D), GuNIR (nitrate reductase), GuMAPK3 (mitogen-activated protein kinase 3), and GuMAPK6 (mitogen-activated protein kinase 6). GuActin2 served as an internal control for expression normalization. Forward and reverse primers for all tested genes are listed in Supplementary Table 1. Bars (±SD) represent the means of three to five individual plants; letters (a, b, c, and d) indicate significant differences between treatments (P < 0.05).
the PM Na + /H + antiport system . Chung et al. (2008) found that reactive oxygen species mediated SOS1 mRNA stability in Na + -treated Arabidopsis.
In addition to our agonist findings, the pharmacological data also showed that the salt-induced transcription of GuSOS1 or GuAHA could be inhibited by DMTU, cPTIO, or PPADS in saltstressed G. uralensis roots (Figure 9). These findings suggested that the endogenous salt-sensitive messengers, H 2 O 2 , NO, and eATP, contributed to the induction of G. uralensis Na + /H + antiport genes during NaCl stress. However, the Ca 2+ -channel inhibitor, LaCl 3 , did not block the salt-responsive induction of GuAHA and GuSOS1 (Figure 9). This result implied that vacuolar Ca 2+ release might facilitate cytosolic Ca 2+ signaling in the salt response of G. uralensis . Indeed, in a previous study, we showed that a vacuole-generated Ca 2+ signaling pathway participated in the regulation of ionic homeostasis in NaCl-stressed P. euphratica cells .

K + Homeostasis
In G. uralensis roots, NaCl-induced K + efflux was blocked by TEA (a specific inhibitor of K + permeable channels), but enhanced by vanadate (Figure 5). These findings suggested that NaCl-induced K + loss was mediated by depolarization-activated channels, e.g., KORCs and NSCCs (Shabala et al., 2005(Shabala et al., , 2006Sun et al., 2009b;Lu et al., 2013;Lang et al., 2014;Zhao et al., 2016). The addition of Ca 2+ , H 2 O 2 , NO, and eATP reduced the salt-induced K + efflux (Figures 4, 6-8). Presumably, this result was due to the inhibition of K + -channels by the activated PM H + -ATPase, because these signaling molecules upregulated GuAHA transcription in salinized roots (Figure 9). Previous studies have shown that NaCl-induced increases in PM H + -ATPase activity depended on H 2 O 2 production, in P. euphratica (Zhang et al., 2007;Sun et al., 2010a,b) and in secretor and non-secretor mangroves (Lu et al., 2013;Lang et al., 2014). In A. marina leaves, NO remarkably enhanced PM H + -ATPase activity and AHA1 transcription, and conversely, these activities were reduced by NO synthesis inhibitors and NO scavengers . The maintenance of K + homeostasis in P. euphratica cells was attributed to the eATP induction of AHA (Sun et al., 2012a). Moreover, in poplar cells, NaCl-induced K + loss increased, when AHA transcription was inhibited by the glucose-hexokinase trap system or P2 receptor antagonists (suramin and PPADS) (Sun et al., 2012a;Zhao et al., 2016). In the present study, we also found that NaCl-induced K + loss increased (Figures 4B, 6B-8B) and GuAHA expression was inhibited by the four tested antagonists, but LaCl 3 produced less pronounced effects compared to DMTU, cPTIO, and PPADS (Figure 9). We concluded that salt-induced signaling molecules were required for upregulation of the PM H + -ATPase gene in G. uralensis roots. As a result, enhanced H + pumping activity, on one hand, reduced K + loss via depolarization-activated channels, and on the other hand, promoted Na + extrusion via PM Na + /H + antiporters (Chen and Polle, 2010;Chen et al., 2014;Polle and Chen, 2015).
Multiple Signaling Networks Involved in the NaCl-Induced Expression of Salt-Responsive Genes Related to K + /Na + Homeostasis Clear interactions occurred between these stress signals to accelerate the transcription of salt-adaptive signaling pathway genes in G. uralensis roots. Ca 2+ increased the GuSOS3 expression (Figure 10), thus leading to enhanced Na + extrusion via the SOS-signaling pathway (Zhu, 2001). In NaCl-treated roots of G. uralensis, H 2 O 2 , NO, or eATP, promoted the transcription of GuSOS3/GuCIPK (Figure 10), which indicated that these stress signals predominantly activated the Ca 2+ -SOS signaling pathway. A previous study in P. euphratica cells showed that exogenously applied H 2 O 2 increased Ca 2+ influx, which led to elevated cytosolic Ca 2+ (Sun et al., 2010b). Based on our present results in G. uralensis roots, we suggest that H 2 O 2 increased cytosolic Ca 2+ , which then mediated PM Na + /H + antiport upregulation via the SOS-signaling pathway (Zhu, 2001(Zhu, , 2016. Furthermore, we found that NO enhanced the transcription of GuSOS3/GuCIPK in NaCl-stressed liquorice roots (Figure 10). Thus, NO-simulated Ca 2+ -SOS signaling would promote Na + efflux and alleviate cellular Na + toxicity in G. uralensis. Similarly, in the secretor mangrove, A. corniculatum, NO enhanced Na + efflux elicited by Ca 2+ .
Extracellular ATP signaling is a novel player in salt-stress acclimation. We found that eATP increased the expression of GuSOS3, GuCIPK, GuRbohD, GuNIR, GuMAPK3, and GuMAPK6 (Figure 10). Moreover, GuMAPK3 and GuMAPK6 expression levels were enhanced by H 2 O 2 in salinized G. uralensis roots (Figure 10). This indicated that eATP interacted with H 2 O 2 and Ca 2+ signaling to maintain K + /Na + homeostasis. Previously, eATP was shown to interact with H 2 O 2 and Ca 2+ to increase Na + extrusion in two mangrove species, Kandelia obovata and A. corniculatum . In P. euphratica cells, eATP signaling was mediated by H 2 O 2 and cytosolic Ca 2+ in the salt response (Sun et al., 2012a). Accordingly, eATP is thought to bind P2-like receptors in the PM (Choi et al., 2014), which leads to an increase in H 2 O 2 and a transient elevation in cytosolic Ca 2+ (Jeter et al., 2004;Demidchik et al., 2009;Sun et al., 2012a). Thus, eATP could initiate the H 2 O 2 and Ca 2+ signaling cascades and cause an increase in Na + /H + exchange across the PM of G. uralensis roots under NaCl stress.
Extracellular ATP also increased the expression of GuNIR (Figure 10). This finding indicated that NO was a downstream component of eATP signaling. Similarly, in P. euphratica cells, NO was triggered by eATP, although NO played a negligible role in eATP-stimulated cell death (Sun et al., 2012b). There are species-specific interactions between eATP and NO in the mediation of K + /Na + homeostasis . In this study, eATP signaling appeared to be mediated by NO in G. uralensis roots (Figure 10). However, in the non-secretor, K. obovata, NO was redundant in the presence of eATP, because eATP alone exerted a pronounced effect on Na + /H + antiporters .

CONCLUSION
Our findings suggested that salt exposure increased Ca 2+ , H 2 O 2 , NO, and eATP, which served as signaling molecules in mediating K + /Na + balance by elevating Na + efflux and restraining K + loss in G. uralensis. Based on these results, we proposed a multiple signaling network for regulating ionic homeostasis in salinized G. uralensis (Figure 11). The NaClinduced signaling molecules, Ca 2+ , H 2 O 2 , NO, and eATP, upregulated GuSOS1 and GuAHA expression, which increased the numbers of Na + /H + antiporters and H + pumps in the PM. The enhanced Na + /H + antiport system promoted the SOSsignaling pathway. In addition, H + -pump activity preserved the membrane potential, which restricted K + efflux through DA-KORCs and DA-NSCCs. Interestingly, we also found interactions between these stress signaling molecules and the expression of salt-responsive genes in G. uralensis roots. Ca 2+ , H 2 O 2 , NO, and eATP enhanced GuSOS3/GuCIPK genes, which are related to the Ca 2+ -SOS signaling pathway. Moreover, eATP exhibited novel interactions with Ca 2+ , H 2 O 2 , and NO signaling, which contributed to the upregulation of GuSOS3, GuCIPK, GuRbohD, and GuNIR. This crosstalk was thought to contribute to the upregulation of GuSOS1 and GuAHA expression in G. uralensis roots. Further investigations are needed to confirm these interactions.

AUTHOR CONTRIBUTIONS
TL, JX, and SC conceived of the original screening and research plans; SC supervised the experiments; TL, SD, NZ, CD, YnZ, YlZ, HZ, GS, and JY performed most of the experiments; CW, YW, QD, and SL provided technical assistance to TL, SD, and NZ; TL designed the experiments and analyzed the data; TL conceived of the project and wrote the article, with contributions from all the authors; SC supervised and complemented the writing. All authors have read and approved the manuscript.