To evaluate the effects of alkaline stress on the growth of P. tenuiflora, shoot length, leaf fresh weight, dry weight, and water content were determined (Figure 1). The shoot length declined gradually with the increase in Na₂CO₃ concentration (Figure 1A). The fresh weight and water content decreased when seedlings were treated with 95 mM Na₂CO₃ (Figure 1B,D).

The photosynthesis indexes of P. tenuiflora under Na₂CO₃ treatment were analyzed. After seven days of 38 mM and 95 mM Na₂CO₃ treatments, the P. tenuiflora seedlings did not show obvious damage to leaf morphology (data not shown), implying the high capacity of P. tenuiflora seedlings to tolerate Na₂CO₃. However, photosynthesis was affected by Na₂CO₃ stress. Stomatal conductance (Gs) (Figure 2A), photosynthetic rate (Pn) (Figure 2B), and transpiration rate (Tr) (Figure 2C) exhibited little changes under 38 mM Na₂CO₃ treatment, but showed marked decreases under 95 mM Na₂CO₃. In addition, chlorophyll fluorescence parameters were monitored to determine the performance of photosystem II (PSII) photochemistry. The maximum quantum efficiency of PSII photochemistry (Fv/Fm) (Figure 2D) and the PSII maximum efficiency (Fv′/Fm′) (Figure 2E) were not significantly altered under 38 mM Na₂CO₃, but were reduced remarkably under 95 mM Na₂CO₃. The non-photochemical quenching coefficient (qNP) (Figure 2F) remained constant under 38 mM and then increased under 95 mM Na₂CO₃ treatment.

Leaf osmotic potential showed a significant decrease under Na₂CO₃ treatments (Figure 3A), indicating the seedlings suffered from osmotic stress. The electrolyte leakage ratio (Figure 3B) and malondialdehyde (MDA) contents (Figure 3C) were increased significantly under Na₂CO₃. This indicates that the plasma membrane integrity was damaged by Na₂CO₃ treatment, probably resulting from ROS generated under high pH and ion stress conditions. The activities of representative antioxidative enzymes were altered with different patterns under Na₂CO₃ stress. The superoxide dismutase (SOD) activity decreased under Na₂CO₃ stress (Figure 3D), but the peroxidase (POD) activity increased under 38 mM Na₂CO₃ (Figure 3E), and the catalase (CAT) activity increased obviously under both Na₂CO₃ concentrations (Figure 3F).

Ion homeostasis is important in plant response to salt stress. Leaf Na content increased with increases in Na₂CO₃ concentration (Figure 4A). Leaf K contents also increased under 95 mM Na₂CO₃ treatment (Figure 4B). This led to declined K/Na ratios under the Na₂CO₃ treatment (Figure 4C). Furthermore, Na and K contents on the leaf surface increased gradually with the increases in Na₂CO₃ concentration, and K/Na ratio was reduced on the leaf surface (Figure 4D,E,F). Moreover, as an important signaling messenger, calcium contents in different subcellular compartments of epidermal cells and mesophyll cells were affected by Na₂CO₃. In epidermal cells, calcium content in the cell wall and cytoplasm increased gradually with the increases in Na₂CO₃ concentration, but reduced in the vacuoles under 38 mM Na₂CO₃ treatment (Figure 4G). Similarly, in mesophyll cells, calcium contents in the cell wall and cytoplasm increased under 95 mM Na₂CO₃ treatments, while the calcium content in vacuoles reduced sharply in the Na₂CO₃ treated samples (Figure 4H).

The alkali-responsive protein profiles were obtained using 2-DE analysis of leaf samples under 0 mM, 38 mM, and 95 mM Na₂CO₃ treatments (Figure 5, Supplemental Figure S1). Approximately 1000 Coomassie Brilliant Blue-stained protein spots were detected on the pI 4–7 gels. A total of 43 protein spots were identified using LC ESI Q-TOF MS/MS and Mascot database searching (Figure 5, Table 1). Based on Gene Ontology, BLAST alignments, and information from literature, the protein identities (IDs) were classified into 10 functional categories including photosynthesis, stress and defense, membrane and transport, carbohydrate and energy metabolism, amino acid metabolism, transcription, protein synthesis, protein folding and transporting, protein degradation, and signaling (Table 1). Among these categories, carbohydrate and energy metabolism (41%), transcription and protein metabolism (27%), photosynthesis (14%), as well as stress and defense (9%), were over-represented.

One important goal of system biology is to understand the interdependence of proteins and their expression profiles in a certain tissue or other biological samples [6,20]. An effective method to determine the regulatory mechanisms for protein interactions is the application of hierarchical clustering algorithms used in DNA microarray experiments. With this method, the proteins appearing on the same branches are assumed to be involved in related molecular functions [6,20]. Thus, to analyze the expression characteristics of proteins involved in each functional category, we performed hierarchical clustering analysis of the 43 IDs, which revealed two main clusters. Cluster I included 32 alkali-induced IDs and cluster II contained 11 alkali-reduced IDs (Figure 6). The number of alkali-induced proteins was obviously larger than that of reduced proteins. The analysis of protein functional categories showed a heterogeneous distribution between the two clusters (Figure 6). For example, most of the IDs involved in carbohydrate and energy metabolism were grouped in cluster I, whereas photosynthesis-related proteins were mainly in cluster II (Figure 6). This suggests that a switch of biological processes occurred in the course of alkaline treatment. Interestingly, the clustering result supports the previous notion that 38 mM is the turning point concentration of Na₂CO₃ for P. tenuiflora[12,21,22], because there were 28 out of 43 protein IDs whose protein abundances changed. Among them, 22 proteins (sub-cluster I-1) reached the maximum levels, and six proteins (subcluster II-1) got to the minimum levels under 38 mM Na₂CO₃ treatment (Figure 6). Such protein patterns correlated well with our aforementioned physiological results, e.g., the photosynthetic capability (e.g., Pn, and qP) and antioxidant-related indexes.

The effects of salinity and alkalinity on plant growth and development vary amongst plants. For glycophytes, salt stress generally reduces plant growth and development, but for most halophytes, moderate salt accumulation promotes the plant growth [23]. However, the less-tolerant dicotyledonous halophytes and monocotyledons halophytes, especially grasses, grow better in non-salinized conditions [23,24]. Our results showed that the growth of P. tenuiflora, a monocot halophyte grass, was not obviously affected by low Na₂CO₃ concentration, but was inhibited by higher Na₂CO₃ concentration (Figure 1). This correlates well with the biomass changes of P. tenuiflora under NaCl stress [6]. In this study, Gs, Pn, and Tr of P. tenuiflora seedlings exhibited little changes at 38 mM Na₂CO₃, but decreased significantly at 95 mM Na₂CO₃ (Figure 2A,B,C), indicating photosynthesis was reduced at the higher Na₂CO₃ concentration. Moreover, Fv/Fm and Fv′/Fm′ were stable at 38 mM Na₂CO₃, but reduced at 95 mM Na₂CO₃ (Figure 2D,E). This implied that the efficiency of PSII photochemistry was not affected by the low Na₂CO₃ concentration, but inhibited by the high Na₂CO₃ concentration. The reduced Fv/Fm of seedlings under 95 mM Na₂CO₃ implied the occurrence of photoinhibition in P. tenuiflora under the higher concentration of Na₂CO₃. The decrease of Fv/Fm was usually accompanied by increase of thermal dissipation, which was evaluated by nonphotochemical quenching of Chl fluorescence (qNP) [25]. Here qNP was maintained almost constant at 38 mM Na₂CO₃, but increased significantly at 95 mM Na₂CO₃ (Figure 2F), which corresponded with the change of Fv/Fm. This result implied that thermal dissipation remained steady at 38 mM Na₂CO₃, but increased at 95 mM Na₂CO₃.

Previous studies have found that salt and alkali stresses affected photosynthetic carbon fixation [1,26]. Our present proteomics data revealed that some of the enzymes in Calvin cycle were reduced by Na₂CO₃ stress, which implied that the decrease in Pn was due to the less efficient carbon fixation under Na₂CO₃ stress. These enzymes included carbonic anhydrase (CA), RuBisCO, phosphoribulokinase (PRK), and ferredoxin-thioredoxin reductase (FTR). CA can help increase the concentration of CO₂ within the chloroplast in order to increase the carboxylation rate of RuBisCO [27]. The change tendency of CA is similar to those of Gs, Pn, and Tr in P. tenuiflora. These results indicate that the decrease of Pn is mainly resulted from the declined carbon fixation. RuBisCO catalyzes the first major step of carbon fixation in C3 plants, and PRK is the key enzyme that functions in phosphorylating RuP into ribulose-1,5-bisphosphate (RuBP) in the Calvin cycle [28]. FTR is an iron-sulfur enzyme, which links light to enzyme regulation in oxygenic photosynthesis, catalyzing the activation of fructose 1,6-bisphosphatase [29]. The decline of these enzymes in the Calvin cycle under Na₂CO₃ treatment could lead to the decrease of carbon fixation.

Salt and alkali stresses enhance the production of ROS, resulting in various ROS-associated perturbations in the seedlings [2]. Chloroplasts are key intracellular ROS generators. In chloroplasts, the production of O₂⁻ is mainly determined by the balance between absorption and utilization of light energy [25]. The energy consumption for CO₂ assimilation suppressed by Na₂CO₃ stress led to ROS imbalance, which would cause oxidative damage to enzymes and thus the photosynthetic apparatus [2,25,30]. In the present study, the significantly increased electrolyte leakage ratio and MDA contents indicate that plasma membrane was damaged by lipid peroxidation under Na₂CO₃ stress [30].

Our results have revealed several mechanisms of light energy balance and antioxidation used to cope with Na₂CO₃ stress in P. tenuiflora seedlings. Salt accumulation on the leaf surface would prevent excessive light absorption in Na₂CO₃ stressed plants. The contents of Na and K on the surface of P. tenuiflora leaves increased gradually with increasing Na₂CO₃ concentrations (Figure 4A,B,D,E), which suggested that P. tenuiflora possesses the ability to secrete salts under Na₂CO₃ stress. Other ions such as Ca, Mg, and Si have also been found to increase in concentration on the P. tenuiflora leaf surface with the increasing Na₂CO₃ concentrations [10]. The accumulation of salts on leaf surfaces developed greater surface reflectance, contributing to reduce light absorption [31]. At the same time, thermal dissipation was enhanced to remove excess energy at high Na₂CO₃ concentration. The increase of qNP demonstrates that the thermal dissipation was enhanced to protect the photochemical apparatus under Na₂CO₃ stress (Figure 2F). In addition, the enhancement of the xanthophyll cycle would increase the thylakoid pH, which is helpful for induced thermal dissipation [25]. The increase of temperature-induced lipocalin (TIL) in our proteomics results supported this notion. Lipocalin was found to be a key enzyme in the xanthophyll cycle responsible for protection against photo-oxidative damage [32]. It was also found to be increased in Solanum lycopersicum under salt stress [33].

The ROS scavenging system was activated in seedlings to cope with Na₂CO₃ stress. The activities of POD (Figure 3E) and CAT (Figure 3F), the two enzymes involved in the removal of H₂O₂, were increased under Na₂CO₃ treatments, especially under 38 mM Na₂CO₃. However, SOD activity was reduced dramatically under Na₂CO₃ stress (Figure 3D). SOD is an enzyme for dismutation of O₂⁻ to produce H₂O₂. Thus, in P. tenuiflora seedlings under Na₂CO₃, H₂O₂ might be mainly produced in peroxisome from oxidation of glycolate during photorespiration rather than from dismutation of O₂⁻[6]. In addition, our proteomics data showed that other antioxidant and detoxification mechanisms were enhanced in seedlings to cope with Na₂CO₃ stress. Glyoxalase (GLO) was found to be induced in seedlings with increasing concentrations of Na₂CO₃. GLO is a member of the glyoxalase system that carries out the detoxification of methylglyoxal and other reactive aldehydes produced in plant metabolism. Previous studies have documented that salinity stress induced high accumulation of methylglyoxal, a potent cytotoxin in various plant species [34]. In transgenic tobacco plants, overexpression of glyoxalase I can tolerate an increase in methylglyoxal and maintain high levels of reduced glutathione under salinity stress [35]. Our results implied that GLO might be an important candidate for conferring high alkali tolerance in P. tenuiflora. Furthermore, germin-like protein (GLP) was also induced under Na₂CO₃ stress. Germin was first detected in germinating wheat seeds, but its homologs have now become ubiquitous in the plant kingdom and have various functions, not only during embryogenesis, but also in biotic or abiotic stress conditions [36]. The increased GLPs were detected in Nicotiana tabacum leaf apoplast [37] and Arabidopsis thaliana roots [38] after exposure to salt stress. Wheat germin has been shown to display oxalate oxidase activity; this activity is shared among most plants [36]. Germin-like oxalate oxidase is involved in degrading the oxalic acid, a highly toxic chemical, through production of H₂O₂[36]. Thus, the Na₂CO₃ increased GLP might provide an explanation for the decrease of oxalic acids in shoots and roots of P. tenuiflora[9]. Additionally, an aluminum-induced protein-like protein (AIPLP) was induced under 95 mM Na₂CO₃ treatment. Previous studies have shown that the AIPLP was not specific to aluminum stress, but also involved in other metal, wounding [39], and drought stresses [40]. This protein might also contribute to the tolerance to Na₂CO₃ stress in P. tenuiflora when at a high concentration.

The intracellular ion homeostasis is fundamental to living cells. Under salinity conditions, high apoplastic levels of Na⁺ would alter the aqueous and ionic thermodynamic equilibrium, resulting in hyperosmotic stress, ionic imbalance, and toxicity [26]. Thus, proper regulation of ion flux is necessary for cells to keep the concentrations of toxic ions low and to accumulate essential ions. Our results implied that P. tenuiflora developed some protective mechanisms to reestablish cellular ion homeostasis through selective salt accumulation or exclusion, in vivo compartmentalization, and Ca²⁺ signaling.

Maintaining a high cytosolic K⁺/Na⁺ ratio is one of the most important mechanisms for plant salt tolerance [8]. In our study, the Na content in leaves increased remarkably in all alkaline treatments. However, the K content did not increase under 38 mM Na₂CO₃, but increased slightly under 95 mM Na₂CO₃ (Figure 4A,B). This led to the decreased intracellular K/Na ratio in P. tenuiflora leaves (Figure 4C), although some Na ions were secreted to the leaf surface (Figure 4D). It is obvious that the ion homeostasis in P. tenuiflora leaves was affected under Na₂CO₃ stress. In previous studies, P. tenuiflora had lower net Na⁺ uptake rates than wheat (less than 50% under 150 mM NaCl) [8], which indicates that P. tenuiflora has a greater capacity than wheat to restrict unidirectional Na⁺ influx to maintain low net Na⁺ uptake[9]. Besides, the increased contents of K and Na on the leaves surface with the increase of external Na₂CO₃ concentrations (Figure 4D,E) support the hypothesis that P. tenuiflora leaves could exude salts through stomata or together with wax secretion [10].

Ion compartmentalization in different tissues can facilitate their metabolic functions [2]. The salt-inducible Na⁺/H⁺ antiporter is in charge of Na removal from the cytoplasm or compartmentalization in the vacuoles [2]. The vacuolar Na⁺/H⁺ antiporters were induced by NaHCO₃ in P. tenuiflora, suggesting its key role in pH regulation under alkaline conditions [41]. Vacuolar-type Na⁺/H⁺ antiporter was mainly driven by the proton gradient across the vacuolar membrane generated by vacuolar type H⁺-ATPases (V-ATPases) [2,42]. The V-ATPase is indispensable for plant growth under normal conditions due to its roles in energizing secondary transport, maintaining solute homeostasis, and facilitating vesicle fusion. Under stress conditions (e.g., salinity, drought, cold, acid, anoxia, and heavy metals), the survival of the cells depends strongly on maintaining or adjusting the activities of the V-ATPases [2,42]. In the present study, a subunit of V-ATPase was induced under Na₂CO₃ stress. The corresponding increase of Na content in vacuoles under Na₂CO₃ stress was much higher than in cytoplasm of epidermal cells and mesophyll cells [43]. These findings suggest that the Na₂CO₃-induced V-ATPase was required to energize the tonoplast for ion uptake into the vacuoles.

Importantly, calcium content was changed with diverse cellular structures (e.g., cell wall, cytoplasm, and vacuole) in epidermal cells and mesophyll cells in leaves of P. tenuiflora when exposed to Na₂CO₃ stress (Figure 4G,H). Calcium is a principal signaling molecule for salinity tolerance [2,3]. High salinity leads to increased cytosolic Ca²⁺, which initiates the stress signal transduction pathways [3]. In this study, the calcium content in cytoplasm of epidermal cells (Figure 4G) and mesophyll cells (Figure 4H) increased significantly under 95 mM Na₂CO₃. In contrast, the calcium content in vacuoles of epidermis cells (Figure 4G) and mesophyll cells (Figure 4H) decreased. This indicates that the increased cytosolic Ca²⁺ might be transported from the apoplast and intracellular compartments [2]. In addition, our proteomics data revealed that a developmentally regulated plasma membrane polypeptide (DREPP PM)-like protein increased under Na₂CO₃ stress. DREPP-like protein contains a possible Glu-rich site at the C terminus responsible for calcium binding. DREPP-like protein has been found to be increased temporarily in rice under cold acclimation [44] and salt stress [45]. This result suggests that DREPP-like protein may be associated with the Ca²⁺ signal transduction pathway in the seedlings of P. tenuiflora under Na₂CO₃ stress.

In this study, nine protein IDs were carbohydrate metabolism-related enzymes, and nine were involved in energy production (Table 1). Among them, seven IDs (representing six unique proteins) were enzymes in glycolysis, including fructokinases (FRK), fructose-bisphosphate aldolase (FBA), triosephosphate isomerase (TIM), glyceraldehyde 3-phosphate dehydrogenase (GAPDH), phosphoglycerate kinase (PGK), and enolase. They were all induced under Na₂CO₃ stress. In fact, glycolysis shared a series of reversible reactions with gluconeogenesis. Interestingly, except for FRK, the remaining five enzymes catalyzing the reversible reactions were shared between glycolysis and gluconeogenesis. FRK, an enzyme that irreversibly catalyzes the transfer of a phosphate group from ATP to fructose in glycolysis, is the most important gateway in the control of sugar influx into glycolysis. Thus, the increase of FRK under Na₂CO₃ along with the other five enzymes would contribute to glucose breakdown for energy generation to cope with Na₂CO₃ stress. In addition, malate dehydrogenase (an enzyme in citric acid cycle) and cytosolic 6-phosphogluconate dehydrogenase (an enzyme involved in pentose phosphate pathway) were induced by Na₂CO₃ stress. Moreover, five homologs of ATP synthases and four proteins containing the AAA ATPase family protein domain were induced with similar patterns to aforementioned carbohydrate metabolism-related enzymes in leaves under Na₂CO₃ stress. This implies that the enhancement of the metabolism provides ATP for plant adaptation to Na₂CO₃ stress. Similar results have also been found in salt-stressed Salicornia europaea[20] and P. tenuiflora[6].

Two amino acid metabolism-related proteins, aspartate aminotransferase (AST) and methionine synthase (MS), increased in levels under Na₂CO₃ treatment (Table 1). AST catalyzes the interconversion of aspartate and α-ketoglutarate to oxaloacetate and glutamate, and plays an important role in nitrogen assimilation and biosynthesis of amino acids [20]. MS is an enzyme that catalyzes the final step in the regeneration of methionine from homocysteine [46]. l-methionine is the substrate for the synthesis of S-adenosyl-l-methionine (AdoMet), which is the major methyl-group donor for numerous transmethylations important in plants secondary metabolism. Lignin biosynthesis has been suggested to represent a major sink for AdoMet consumption in vascular plants. Thus, the increase of MS in Na₂CO₃-stressed P. tenuiflora could reflect an enhanced demand of AdoMet in the synthesis of lignin. In addition, AST and MS have been thought to be relevant to cell wall lignification [20], and a previous study has reported the increased deposition of lignin in the vascular tissues of plants under salinity stress [46]. Therefore, the increase of these two proteins might imply that Na₂CO₃ treatment can increase the vessel development in P. tenuiflora. The enhanced cell wall lignification increased the mechanical rigidity of cell wall, strengthening the vascular tissue and subsequently permitting water to be conducted through the xylem under negative pressure without collapse of the vessels. This structural alteration may enhance the cell-to-cell pathway for water transport, impart greater selectivity and reduced ion uptake, and compensate for diminished bulk flow of water and solutes along the apoplastic pathway [6,20,46].

Additionally, a number of proteins involved in transcription, protein synthesis, protein folding and transport, as well as protein degradation showed Na₂CO₃ response (Table 1). Among them, two Na₂CO₃-induced proteins were involved in protein folding. One was protein disulfide-isomerase (PDI), a member of a thioredoxin superfamily inserting disulfide bonds into folding proteins. PDI is specialized to accommodate the structural features of the membrane and secreted proteins. The other was a putative SecA, the mobile subunit of an integral membrane transporter in Escherichia coli, consuming ATP during the insertion and deinsertion phases of its catalytic cycle while guiding preprotein segments across the membrane [47]. While the exact function of SecA in plants exposed to akalinity stress is not known, its function is likely to affect membrane protein dynamics in P. tenuiflora under alkaline stress.

Seeds of Puccinellia tenuiflora (Turcz.) scribn. et Merr. were sowed on pearlite and cultured in a controlled environment chamber under white fluorescent light (300 μM m⁻² s⁻¹; 13 h light/11 h dark) at 25 °C and about 75% relative humidity for 50 days, as previously described [6,12]. The plants were divided into three groups. In order to correlate with a number of our previous studies that used 0, 0.4%, 1.0% Na₂CO₃ treatments, these plants were treated with 0 mM, 38 mM and 95 mM Na₂CO₃, respectively. Three biological replicates were grown and treated with different concentrations. To maintain stable Na concentrations in Hoagland medium, we changed the culture medium daily and monitored the solution ion content and the osmotic potential using a SevenMulti Neutral meter (Mettler Toledo, London, UK) and a vapor pressure osmometer (5520, Wescor Inc., Logan, UT, USA), respectively. Seven days later, the treated and untreated leaves were harvested and used fresh or immediately frozen in liquid nitrogen and stored at −80 °C for physiological and proteomic analysis. Shoot length and fresh weight were measured right after harvesting. Dry weight was determined after dehydration at 60 °C until a constant weight was maintained. Leaf water content was estimated as the difference between the fresh weight and dry weight divided by the fresh weight [6].

Net photosynthetic rate (Pn), stomatal conductance (Gs), and transpiration rate (Tr) were determined at 10:00 am using a portable photosynthesis system LI-COR 6400 (LI-COR Inc., Lincoln, NE, USA). The induction kinetics of chlorophyll fluorescence was recorded at room temperature using a pulse modulation chlorophyll fluorometer (FMS-2, Hansatech, King’s Lynn, UK). After dark adaptation for 30 min, the initial fluorescence yield (Fo) in weak modulated light (0.12 μmol photons m⁻¹ s⁻¹, 600 Hz), and maximum fluorescence yield (Fm) emitted during a saturating light pulse (5000 μmol photons m⁻² s⁻¹, 0.7 s) were measured. Variable fluorescence (Fv) was calculated by the formula: Fv = Fm−Fo. Then Fm′ was measured using an irradiance of 400 μmol photons m⁻² s⁻¹ as the actinic light and 5000 μmol photons m⁻² s⁻¹ as the saturating flashes. Fo′ was measured using an irradiance of far-red light (1.67 μmol photons m⁻² s⁻¹, 3 s). The variable fluorescence (Fv′) was also calculated by Fv′ = Fm′−Fo′. Maximal photochemical efficiency of PSII (Fv/Fm), photochemical efficiency of PSII in the light (Fv′/Fm′), non-photochemical quenching coefficient (qNP; 1−(Fm′−Fo′)/(Fm−Fo)), could be acquired directly from the instrument [12].

Osmotic potential was determined by a vapor pressure osmometer (5520, Wescor Inc., Logan, UT, USA). The MDA content was determined using the thiobarbituric acid (TBA) reaction as described by Wang et al.[48]. The electrolyte leakage ratio was obtained according to Akram et al.[49]. The three antioxidant enzymes (SOD, POD, and CAT) were extracted and detected according to the method of Yu et al.[6].

For measuring the relative contents of K and Na on the leaf surface, fresh leaves (1 × 2 cm) were cemented to a sample plate, then observed and analyzed with XL-30 surrounding scanning electron microscope (The Netherlands Phillips Company, Eindhoven, Netherlands) and Kevex energy peatrometer (Thermo Company, Schaumburg, IL, USA). The working voltage was 20 kV, and the time difference for the measurement of samples was 60 s. The relative contents of elements were represented as CPS (counts per second), which were computed by subtracting the background value from peak values of varied elements.

The sample preparation and microscopic analysis for Ca determination was done according to Qi et al.[50]. The relative Ca contents in diverse cellular structures (including cell wall, cytoplasm, vacuole) were detected by transmission electron microscope (TEM)-X ray microanalysis of 1 μM sections after freeze-drying and water-free embedding in plastic. The leaves (2 mm × 2 mm) were placed in aluminum mesh boxes, and then frozen quickly in the pre-cooling isopentane and propane mixture (the volume ratio is 1:3). Once frozen, the material was placed in a freeze-dryer for removing all of the water. Dry materials were then loaded into the T-type vacuum permeable tube, and infiltrated with diethyl ether in vacuum at 27 °C for 24 h. After being infiltrated with phenylethylene-butyl methacrylate under normal temperature, the materials were transferred into capsules for polymerization. The embedding materials were cut into 1 μm thickness sections on an ultramicrotome (Reichert Ultracut E, Vienna, Austria). For X-ray microanalysis, dry sections were transferred to folding copper grids, coated with carbon and stored over silica gel. Sections were analyzed in a Hitachi 800 transmission electron microscope (fitted with an energy-dispersive X-ray analyzer EDAX 9100). The accelerating voltage was 150 kV, the angle between sample and probe was 33°, and the counting time was 60 s. The relative Ca contents in diverse cellular structures (including cell wall, cytoplasm, vacuole) were also represented by CPS.

For measuring the concentrations of K⁺ and Na⁺ in the seedling leaves, 0.5 g fresh leaves of each treatment were cut to 2 cm, and put into a small beaker, then immersed in 30 mL deionized water and shaken for 30 min. The leaves were taken out and put into another small beaker with about 30 mL deionized water, then cooked for 15 min in the autoclave with high temperature (121 °C) in order to kill the cells and release all the ions. Finally, after the temperature of the solution was cooled to room temperature, the leaves were filtered from the solution, and the extracted solution was added to the volume of 50 mL. We used the SevenMulti tester to determine the concentrations of Na⁺ and K⁺.

The proteins from leaves under different treatment conditions were extracted according to the method of Wang et al.[48]. Protein samples were prepared independently from three different batches of plants. Protein concentration was determined using a Quant-kit according to manufacture’s instructions (GE Healthcare, Piscataway, NJ, USA). The protein samples were separated and visualized using 2DE approaches according to Yu et al.[6]. Protein samples were separated on 24 cm IPG strips (pH 4–7 linear gradient) through isoelectric focusing (IEF) in the first dimension, followed by 12.5% SDS-PAGE gels in the second dimension. About 1.6 mg protein was loaded per gel. Gels were stained by Coomassie Brilliant Blue (CBB). Three biological replicates for each sample and three technical replicates for each biological replicate were run on 2D gels. Gel image acquisition and analysis were conducted as previously described [48]. For image acquisition, the gels were scanned using an ImageScanner III (GE Healthcare) at a resolution of 300 dpi and 16-bit grayscale pixel depth. The images were analyzed with ImageMaster 2D software (version 5.0) (GE Healthcare, Piscataway, NJ, USA).The average vol% values were calculated from three technical replicates to represent the final vol% values of each biological replicate. Spots with more than 1.5-fold change among the treatments and a p value smaller than 0.05 were considered to be differentially expressed.

The differentially expressed spots were excised from the gels and digested with trypsin [51]. MS/MS spectra were acquired on a LC ESI Q-TOF MS/MS (AB Sciex, Framingham, MA, USA) according to the method of Yu et al.[6]. The MS/MS spectra were searched against the NCBInr protein databases [52] using Mascot software (Matrix Sciences, London, UK). The taxonomic category was green plants. The searching criteria were according to Yu et al.[6].

The identified proteins were searched against the NCBI database [52] and UniProt database [53] to determine if their functions were known. Combined with the result of BLAST alignments, these proteins were classified into different categories based on biochemical functions. Self-organizing tree algorithm hierarchical clustering of the expression profiles was performed on the log transformed fold change values of protein spots [20,54].

All results were presented as means ± standard error (SE) of at least three replicates. Data were analyzed by one-way ANOVA using the statistical software SPSS 17.0 (SPSS Inc., Chicago, IL, USA). The treatment mean values were compared by post hoc least significant difference (LSD) test. A p value less than 0.05 was considered statistically significant.