The pot experiments under different levels of salt stress were conducted in a controlled environment room at the Institute of Subtropical Crops of Zhejiang Province, Wenzhou, China (120°64′E, 28°00′N) in 2015. In March, 1-year-old healthy and homogenous D. superbus young shoots were transferred to plastic pots (15 cm height and 12.5 cm inner-diameter with holes at the bottom, one seedling per pot) that were filled with 2.5 kg of peat (pH 6.11). The peat was imported and had an organic matter content of 3.22%, total N content of 0.282%, total P content of 0.198%, and total K content of 1.663%. Three months later, a completely randomized design with 5 replications per treatment and 5 plants per replication was adopted. Eight treatments consisting of four salt concentrations (0, 0.3, 0.6, and 0.9% NaCl) and two SA concentrations (0 and 0.5 mM) were implemented as follows: T1 (distilled water); T2 (distilled water with 0.5 mM SA); T3 (0.3% NaCl); T4 (0.3% NaCl with 0.5 mM SA); T5 (0.6% NaCl); T6 (0.6% NaCl with 0.5 mM SA); T7 (0.9% NaCl); and T8 (0.9% NaCl with 0.5 mM SA). To avoid osmotic shock, the salt solution was gradually added to the soil through eight steps to achieve the final concentrations of 0.3, 0.6, and 0.9%. The SA was dissolved in ethanol, “Tween-20” (0.1% dilute solution) was added, and double-distilled water was used to obtain 0.5 mM SA. The SA was sprayed on the leaves, both the adaxial and abaxial surfaces, twice daily at 7:00 and 18:00 (4 days before the salt treatment). Analyses were carried out after 45 days, when obvious external differences were observed between the plants subjected to different treatments.

At the end of experiment (45 days), one intact plant (above-ground shoot and below-ground root) from each replicate per treatment was harvested for biomass determination. The leaf area (LA) (cm²) of the third and fourth fully expanded leaves from the top of the shoots was measured at the same time. The fresh weight and biomass, including roots, stems, and leaves, was obtained by electronic scales. And the leaf mass ratio (LMR) was calculated for each individual by dividing the leaf dry mass by total biomass. The specific leaf weight (SLW) was calculated by dividing the fresh leaf weight by the corresponding leaf area (LA) (Tang et al., 2015).

The youngest, fully developed, healthy leaves were randomly selected from the top of the shoots for photosynthetic measurements, using an LI-6400 XT portable photosynthesis system (Li-Cor, Inc., Lincoln, NE, U.S.) with a standard leaf chamber equipped with a 6400-02B LED light source. Measurements were conducted on sunny day from 8:30 to 11:30; the leaf temperature was maintained at 28 ± 2°C, the CO₂ concentration ([CO₂]) in the leaf chamber was 400 ppm, relative humidity (RH) was 50%, and the photosynthetic photon flux density was 1,200 μmol m⁻² s⁻¹. LA was analyzed at the same time and used when calculating the photosynthetic parameters (Ma et al., 2015; Tang et al., 2015).

Approximately 0.1 g of finely cut and well-mixed leaf tissues were repeatedly extracted using 8 mL of 95% alcohol. Chlorophyll was extracted for 24 h in the dark, and the samples were shaken three times until blanched. The absorbance of the supernatant was measured at 649, 665, and 470 nm using a spectrophotometer (Shimadzu UV-2550, Kyoto, Japan) after centrifugation of the mixture. Chlorophyll concentrations were calculated using a standard method (Lichtenthaler, 1987) and expressed as mg g⁻¹ fresh weight (FW).

The proline content was expressed as relative electrolyte leakage rate and determined by the method of Shen et al. (2014). The proline concentration was calculated using proline standards (0–50 mg ml⁻¹) in an identical manner (Shen et al., 2014).

The lipid peroxidation level was assessed based on the malondialdehyde (MDA) content using a method described by Guidi et al. (2000). The MDA concentration was expressed as mol g⁻¹ using the following formula: 6.45(A532–A600)–0.56A450.

Membrane permeability was assessed by measuring the relative electrolyte conductivity of the leaf following the method described by Deshmukh et al. (1991). Leaves (0.2 g) were rinsed using distilled water and immersed in a test tube with 30 mL of distilled water for 12 h. The electrical conductivity (EC1) of the solution was measured using a conductivity meter (Model DJS-1C; Shanghai Analytical Instrument Co. Shanghai, China). Then, the samples were heated at 100°C for 30 min and the conductivity (EC2) in the bathing solution was determined. Membrane permeability was calculated as the ratio of EC1/EC2.

The O2.- production rate was measured by monitoring the nitrite formation from hydroxylamine in the presence of O2.- as described by Wang and Luo (1990) with a slight modification. The O2.- production rate was expressed as μM min⁻¹ g⁻¹ protein.

For analysis of the H₂O₂ content, 0.2 g of leaf tissue was ground finely and homogenized with 25 mL of acetone at 0°C using the method described by Patterson et al. (1984). The H₂O₂ content was calculated using H₂O₂ as a standard and expressed as mM g⁻¹ protein.

To obtain a crude enzyme extract, 0.3 g of frozen leaves were ground in a mortar with 8 mL of 50 mM phosphate buffer solution (pH 7.8) containing 1% polyethylene pyrrole (PVP) at 4°C. The homogenate was centrifuged at 10,000 rpm for 15 min at 4°C. The supernatant was collected to measure the enzyme activities.

SOD activity was analyzed by monitoring its ability to inhibit the photochemical reduction of nitroblue tetrazolium (Giannopotitis and Ries, 1977). One unit of SOD activity was defined as the amount of enzyme that resulted in 50% inhibition of reduction of nitroblue tetrazolium. The specific SOD activity was expressed as U g⁻¹ FW protein. CAT activity was measured by monitoring the disappearance of H₂O₂ (Díaz-Vivancos et al., 2008). One CAT unit was defined as the amount of enzyme needed to decompose 1 mmol H₂O₂ min⁻¹ under these assay conditions. The specific CAT activity was expressed as U g⁻¹ FW min⁻¹. The POD activity in the leaves was estimated using guaiacol as the substrate, following Thomas et al. (1982). Specific POD activity is expressed as U g⁻¹ FW min⁻¹.

Fully expanded leaves from the same leaf position were harvested from all treatment seedlings under daylight conditions. Square samples were cut from the middle of the leaf and fixed with 2.5% glutaraldehyde in phosphate buffer (0.1 M pH7.0) for more than 4 h, then washed three times in the phosphate buffer (0.1 M pH7.0) for 15 min at each step, next postfixed with 1% OsO₄ in phosphate buffer for 1–2 h and washed three times in the phosphate buffer (0.1 M pH7.0) for 15 min at each step. Then the sample was first dehydrated by a graded series of ethanol (30, 50, 70, 80, 90, 95, and 100%) for about 15–20 min at each step, transferred to the mixture of alcohol and iso-amyl acetate (v:v = 1:1) for about 30 min, then transferred to pure iso-amyl acetate for about 1 h or overnight. In the end, the sample was dehydrated in Hitachi Model HCP-2 critical point dryer with liquid CO₂. The dehydrated sample was coated with gold-palladium in Hitachi Model E-1010 ion sputter for 4–5 min and observed in Hitachi Model TM-1000 SEM (Hitachi, Japan). For morphometric estimation of stomatal aperture and density, the methods described by Snider et al. (2009) were used.

To examine the chloroplast ultrastructure of mesophyll cells, leaf samples were immediately fixed in 2.5% (v/v) glutaraldehyde (0.1 M phosphate buffer, pH 7.2) for at least 4 h after being cut from the plants. Then, the samples were immersed in 1% (v/v) osmium acid for post-fixation 1–2 h and washed three times in the phosphate buffer (0.1 M, pH7.0) for 15 min at each step. The specimen was first dehydrated by a graded series of ethanol (30, 50, 70, 80, 90, 95, and 100%) for about 15–20 min at each step, then transferred to absolute acetone for 20 min. Then samples were placed in 1:1 mixture of absolute acetone and the final Spurr resin mixture for 1 h at room temperature, then transferred to 1:3 mixture of absolute acetone and the final resin mixture for 3 h and to final Spurr resin mixture for overnight. And then specimen was placed in eppendorf contained Spurr resin and heated at 70°C for more than 9 h. At the end, the specimen was sectioned in LEICA EM UC7 ultratome and sections were stained by uranyl acetate and alkaline lead citrate for 5–10 min respectively and observed in Hitachi Model H-7650 TEM (H7650, Hitachi, Tokyo, Japan; Deng et al., 2012). The methods of microscopic measurement was described by Ma et al. (2015).

Data were tested using analysis of variance (ANOVA) with SPSS software version 16.0 (SPSS, Chicago, IL, U.S.), Duncan's multiple range test was used to detect differences between means. The P-value was set at 0.05 and 0.01 for the ANOVA and Duncan's multiple range tests, respectively.

Salinity treatments significantly suppressed plant growth and development, resulting in a pronounced reduction in the fresh weight and dry biomass of D. superbus. Compared with the non-salt-treated D. superbus (T1), the 0.3, 0.6, and 0.9% NaCl treatments markedly reduced the fresh weight of the seedlings by 10.7, 28.13, and 38.5%, respectively (Table S1, Figure 1). The SA treatment reduced the decrease in the growth of the salt-stressed D. superbus, and there was no significant difference between non-salt treatment (T1) and the non-salt treatment with SA (T2). Compared with the salt-stressed plants without SA treatment under 0.3%, 0.6% NaCl conditions, the fresh weight and biomass increased by 17.23 and 8.27% under 0.3% NaCl with SA, and by 34.01 and 12.86%, respectively, under 0.6% NaCl with SA. However, there was no significant difference in the fresh weight between the treatments with and without SA under 0.9% NaCl stress. Whereas, SA increased the biomass allocated to the leaves of D. superbus under 0.6 and 0.9% salt-treated (Table S1). Interestingly, the specific leaf weight (SLW) did not change significantly with 0.3% NaCl (P > 0.05), compared with the non-salt-treated D. superbus (T1) but was markedly increased with 0.6 and 0.9% NaCl (P < 0.05).

Salt stress inhibited photosynthesis, with decreases in the net assimilation rate (Pn), transpiration rate (Tr), and stomatal conductance (Gs) of the leaves of 10.2, 12.2, and 22.6% under the 0.3% NaCl condition, 23.1, 20.9, and 32.2% under the 0.6% NaCl condition, and 30.1, 25.8, and 32.2%, respectively, under the 0.9% NaCl condition (Table S2). Notably, the intercellular CO₂ (Ci) under 0.9% NaCl was significantly higher than the other treatments (P < 0.05). The SA treatment increased the Pn, Tr, Ci, and Gs of leaves under 0.3 and 0.6% salt stress. The Pn and Tr of SA-treated plants increased by approximately 9.02 and 10.2% under the 0.3% NaCl condition and 8.7 and 4.8%, respectively, under the 0.6% NaCl condition compared with the non-SA-treated plants. The Ci in the SA-treated plants under the 0.9% NaCl condition was similar to that in the non SA-treated plants (P > 0.05). However, the Ci in the SA-treated plants under 0.3 and 0.6% NaCl stress was markedly higher (8.7 and 3.5%, respectively) than that of the plants without SA treatment (P < 0.05). Furthermore, there were no obvious differences in the Pn-, Gs-, Tr-, and Ci-values between the SA treatment (T2) and the water-treated (without SA) plants (T1) and between the SA (T8) and 0.9% NaCl treatment without SA (T7) (P > 0.05).

Salt stress resulted in a decrease in the chla, chlb, total chlorophyll and carotenoid contents compared with those of the non-salt-treated plants. The chla content decreased by 29.5% under the 0.3% salt condition, 72.3% under the 0.6% salt condition, and 77.1% under the 0.9% saline condition (Figure 2A). The SA treatment under salt stress conditions slowed the decrease in the chla content by approximately 16.2% under the 0.3% saline condition and 41.3% under the 0.6% saline condition compared with the non-SA-treated plants (Figure 2A). The chlb content decreased by 20.4% under the 0.3% salt condition, 71.4% under the 0.6% saline condition, and 77.5% under the 0.9% saline condition (Figure 2B). Under salt stress conditions, the SA treatment slowed the decrease in the chlb content by approximately 14.7% under the 0.3% saline condition and by 7.1% under the 0.6% saline condition compared with the non-SA-treated plants (Figure 2B). Similar responses were observed in the carotenoid content (Figure 2C). However, there were no significant differences in the chla, chlb, and carotenoid contents between the treatments with and without SA under 0.9% NaCl stress.

Salt stress caused membrane lipid peroxidation injury in the leaves of the D. superbus (Figure 3). The MDA content increased by 98.2, 162.1, and 264.2% under the 0.3, 0.6, and 0.9% NaCl conditions, respectively, and the relative electrical conductivity (REC) increased by 4.38, 55.1, and 139.1% under the 0.3, 0.6, and 0.9% NaCl conditions, respectively. In addition, the proline content increased by 28.6, 139.8, and 714.6% under the 0.3, 0.6, and 0.9% NaCl conditions, respectively. Compared with the non-SA-treated plants, the reductions in MDA in the SA-treated plants were approximately 33.0% (P < 0.05) under the 0.3% NaCl condition and 18.9% (P < 0.05) under the 0.6% NaCl condition. However, compared with the non-SA-treated plants, the REC and proline content increased by 76.8 and 53.7% under 0.3% salt stress, and by 52.8 and 54.1%, respectively, under 0.6% salt stress. Interestingly, no significant difference was found between MDA, REC and the proline content between the SA (T2) and water-treated (without SA) plants (T1) (P > 0.05) and between the SA-treated plants with 0.9% NaCl (T8) and those without SA that were subjected to 0.9% NaCl (T7) (P > 0.05).

Salt stress accelerated the production of ROS in the leaf tissue of the D. Superbus (Figure 4). The O2.- production rate was significantly increased by 56.5% (P < 0.05), 111.4% (P < 0.05), and 203.1% (P < 0.05) under the 0.3, 0.6, and 0.9% NaCl conditions, respectively, and the H₂O₂ content increased by 60.0% (P < 0.05), 97.5% (P < 0.05) and 130.0% (P < 0.05) under the 0.3, 0.6, and 0.9% NaCl conditions, respectively. Compared with the non-SA-treated plants, the reductions in the O2.- production rate and the H₂O₂ content in the SA-treated seedlings were approximately 21.5% (P < 0.05) and 31.2% (P < 0.05) under the 0.3% NaCl condition and 18.9% (P < 0.05), 14.1% (P < 0.05) (P < 0.05) under the 0.6% NaCl condition, respectively. In addition, no significant difference was found in the O2.- production rate and the H₂O₂ content between the SA (T2) and water-treated (without SA) plants (T1) (P > 0.05), and between the SA plants (T8) and the 0.9% NaCl without SA plants (T7) (P > 0.05).

The superoxide dismutase (SOD), peroxidase (POD) and catalase (CAT) activities in the D. superbus were significantly affected by salt tress and SA treatment (Figure 5). SOD increased by 46.3% (P < 0.05), 60.3% (P < 0.05), and 58.3% (P < 0.05) under the 0.3, 0.6, and 0.9% NaCl conditions, respectively. In addition, POD increased by 8.6, 18.1% (P < 0.05) and 52.1% (P < 0.05) under the 0.3, 0.6, and 0.9% NaCl conditions, respectively. However, CAT increased by 44.2% (P < 0.05) and 63.4% (P < 0.05) under the 0.6 and 0.9% NaCl conditions, respectively, but no significant difference was found between the non-salt treatment (T1) and the 0.3% NaCl condition (T3). The SA treatment enhanced the superoxide dismutase (SOD), peroxidase (POD) and catalase (CAT) activities compared with the non-SA-treated plants under salt stress. SOD increased by 27.7% (P < 0.05) and 67.18% (P < 0.05) under the 0.3% NaCl conditions and 0.6% NaCl conditions, respectively, compared with the non-SA-treated plants. The peroxidase (POD) and catalase (CAT) activities increased by 22.4%(P < 0.05), 42.6% (P < 0.05), and 42.1% (P < 0.05), 48.8% (P < 0.05) under the respective 0.3 and 0.6% NaCl conditions, compared with the non-SA-treated plants. However, there was no significant difference in the superoxide dismutase (SOD), peroxidase (POD), and catalase (CAT) activities between the SA (T2) and water-treated (without SA) plants (T1) (P > 0.05).

Salt stress resulted in a variation in the stomatal density and aperture compared with the non-salt-treated plants. Different salt stress treatments were associated with differences in the stomatal traits (Figure 6, Table S3). When salt stress increased from 0 to 0.9%, a gradual decrease in the stomatal aperture was observed. However, the SA treatment under salt stress conditions slowed the decrease in stomatal aperture under different salt stress conditions. And the stomata density decreased from the 0% to the 0.9% salinity treatment. While the SA treatment under 0.3 and 0.6% salt stress increased stomata density.

Salt stress caused obvious changes in the shape, size of chloroplasts (Figure 7, Table S4). The numbers of grana decreased with increasing salt treatment, and the shape of the chloroplasts varied from ellipsoid to swollen oblate or spheroidal. The chloroplasts in the leaves of seedlings grown under the non-salt treatment (T1) and non-salt treatment with SA (T2) showed a normal ultrastructure, with typical grana, well-organized stroma thylakoids and a maximum number of grana lamella. However, the stroma thylakoids became more irregular and disordered, with teas and faults, and the grana lamella decreased with salt stress from 0 to 0.9%. The plants grown under SA treatment mostly had grana containing more and well-organized thylakoids than those plants grown under salt stress. In particular, the number of osmiophilic granules increased with salt stress from 0 to 0.9%, whereas the number of osmiophilic granules decreased under SA treatment compared with those plants grown under salt stress. Overall, the chloroplasts of the non-salt-treated seedlings (T1) and the non-salt-treated seedlings with SA (T2) developed the best.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.