Isolation of bacteria
In May 2016, 12 strains of bacteria were isolated from root surfaces in the wheat rhizoplane in saline-alkali soil of the Yellow River delta in Shandong Province, China (118°49ʹ15″E, 37°24ʹ31″N). The soil parameters were as follows: pH, 8.535; salt content, 0.2143%; electrical conductivity (EC), 622 μs/cm; organic matter, 25.847 g/kg; total N, 1.67 g/kg; and Olsen-P, 13.42 g/kg. Briefly, to isolate the bacterium, roots (1 g fresh weight) were thoroughly washed, homogenized in 0.5 × phosphate-buffered saline (PBS; 9 mL), serially diluted to 10-7 in sterile nutrient agar (NA) medium with a 4% NaCl concentration, and placed in a 30℃ incubator for 48–72 h . The bacterium was subcultured twice. Finally, the isolates were streaked onto NA medium. A glycerol stock solution (30% v/v) of the isolate was prepared and stored at –80°C for later use. Based on colony morphology differences, 12 isolates were identified. Next, we measured the salt tolerance and other PGP activities of these strains to determine the strains with the most potential to promote plant growth in a pot experiment.
Screening for salt tolerance
The isolates were inoculated into Luria-Bertani (LB) medium (1% NaCl) and cultured for 24 h (30℃, 200 rpm/min, 8 × 108 CFU/mL) as seed solutions. The seed solutions were inoculated, in triplicate, into LB medium with different NaCl concentrations (5%, 8%, 10%, 15%, 20%, 25%, 35%) using a 2% inoculum. After 7 d, the absorbances of the cultures were measured with a TU-1810 spectrophotometer (Beijing Puxi General Instruments Co., Ltd., China) at 600 nm. At 30℃, 1 mL of each seed solution was absorbed and diluted to 104 . Uninoculated medium was used as a blank control. The isolates were tested for their ability to survive and tolerate salt in water-soluble fertilizer provided by Shandong Agricultural University Fertilizer Co., Ltd. The main components of potassium nitrate-containing humic acid water-soluble fertilizer) products were the following: macroelements, ≥ 400 g/L; total N, ≥ 360 g/L; potassium, ≥ 45g/L; humic acid, ≥ 30g/L; nitrate N, ≥ 90 g/L; ammonium N, ≥ 90 g/L; and amide N, ≥ 180g/L. The main components of solid water-soluble fertilizer (sulfuric ammonium yellow silver-humic acid soluble fertilizer) were as follows: N-P2O5-K2O = 17-5-23, macroelements, ≥ 45%; nitrate N, ≥ 9%; and fulvic acid, ≥ 3%. The fermentation broth of isolates was inoculated into fertilizer at 5% (v/v) and 5% (w/v) and placed at room temperature for 2 years. Based on salt tolerance of the isolates in LB medium at different NaCl concentrations, we identified HG-15 as the strain with the highest salt tolerance in NA medium. A glycerol stock (15% w/v) of the isolate was prepared and stored at –80℃ until further use.
Amplification and sequencing of 16S rRNA genes
To identify the bacterium at the molecular level, its 16S rRNA was amplified by PCR using a standard method . Universal primers 27F (5′-AGAGTTTGATCMTGGCTCAG-3′) and 1492R (5′-TACGGYTACCTTGTTACGACT-3′) were used to amplify the 16S rRNA . Amplified sequences were then gel purified using a TIANquick Midi Purification kit (Tiangen Biotech, China) and sequenced. The obtained 16S rRNA sequence and sequences in NCBI were compared. Next, the pairwise evolutionary distance between the 16S rRNA sequence of the test strain and related bacterial strains was calculated, and a phylogenetic tree was constructed using the neighbor-joining method in MEGA software (version 5.0) . Bootstrapping of 1,000 replicates was used to assess clustering of associated taxa.
Bioassays for the promotion of growth and enhancement of salinity tolerance traits
The bacterium was first cultured in enriched medium and then transferred to basic medium with ACC as the whole N source. ACCD was determined by quantifying the alpha-ketobutyric acid produced by ACC decomposition. ACCD activity was determined as previously described by Penrose and Glick . The test organism was grown in 15 mL tryptic soy broth to the late logarithmic growth stage and incubated at 30℃ for 24 h at 200 rpm. The cells were then harvested via centrifugation, washed with 0.1 M Tris-HCl (pH 7.6), added to 7.5 mL of DF-basic medium containing 3 mM ACC as the sole N source, and cultured at 30°C overnight. ACCD was induced by exposing bacterial cells to an oscillating water bath at a rate of 200 rpm for 24 h at 30℃. Then, the cells were harvested by centrifugation, washed with 0.1 M Tris-HCl (pH 7.6), and resuspended in 600 μL of 0.1 M Tris-HCl (pH 8.5). Next, 30 μL of toluene was added to the cell suspension and shaken for 30 s. Then, 100 μL of toluene cells was used to measure protein and stored at 4°C. The remaining toluene cell suspension was immediately used for the ACCD assay. ACCD activity was determined by measuring the amount of α-ketobutyric acid produced by hydrolytic cleavage of ACC. The amount of α-ketobutyrate (KB) was determined at 540 nm by comparing the absorbance of the test sample with a standard curve of pure KB (Sigma-Aldrich, USA). ACCD activity was expressed as the amount of alpha-KB micron produced per hour in the reaction system.
To test the ability of the isolate to fix N, a preliminary test of N fixation was conducted by growing the isolate in Ashby’s medium devoid of fixed N sources as qualitative evidence of atmospheric N fixation. The ability to fix atmospheric N was evaluated by growth on N-free JNFb medium using a standard protocol . For N-fixing bacteria assays, we measured sugars to determine the N fixation efficacy of the isolate based on sugar measurement-anthrone photoelectric colorimetry (620 nm) and N measurement-semi-micro photoelectric colorimetry (420 nm). Here, 50 mL of N-free media (1% sugar) was added to a 250 mL flask and sterilized at 121°C for 30 min. After this, two colonies were scraped into individual bottles and shaken at 200 rpm for 3 d at 30°C, and then the bacterial solution was diluted, and the sugar quantity was measured. The sugar standard curve was generated using glucose, whereas the N standard curve was generated using ammonium sulfate. Bacterial N fixation efficiency was considered the mass of N (mg) fixed from the air per 1 g of carbohydrate consumed by the isolate and expressed as mg N/g sugar .
The alkaline solution of mercury iodide and potassium iodide reacts with ammonia to form a reddish-brown colloidal compound. This color is absorbed over a wide wavelength range (410–425 nm). The LB fermentation broth of the isolate was inoculated into 10 mL of peptone solution (peptone, 10 g; NaCl, 10 g; water, 1 L) and then cultured at 37°C for 48 h. Next, 1.0 mL of sodium potassium tartrate solution and 1.5 mL of Nessler reagent were added and mixed. After the mixture was left standing for 10 min, the control was treated with the same amount of sterile LB. Absorbance was measured at 420 nm in a cuvette with a light path of 20 mm, and the amount of ammonia produced was determined according to an ammonium standard curve .
Mineral silicate dissolution activity
The strain was cultured in silicate solid medium [sucrose, 10 g; yeast extract, 0.5 g; (NH4)2SO4, 1 g; Na2HPO4, 2 g; MgSO4∙7H2O, 0.5 g; CaCO3, 1 g; potassium feldspar, 1 g; agar, 15 g; deionized water 1 L]. A clear area was formed around the colony, which indicated that the strain solubilized potassium feldspar .
Proline in sulfosalicylic acid solution was heated by adding acid ninhydrin, and the solution turned red. After adding toluene, all of the pigments were transferred to toluene, and the depth of the dye indicated the level of proline. This method was modified from one described by Bates et al.  to determine proline content. In addition to proline, glutamate acid also plays a role in osmoregulation. It can be extracted with a special reagent, and the results can be determined at 570 nm after adding a chromogenic agent. Reduced glutathione (GSH) is one of the most important anti-oxidation sulfhydryl substances in cells. It plays an important role in anti-oxidation, sulfhydryl protection, and amino acid transport. DTNB reacts with GSH to form a complex with a characteristic absorption peak at 412 nm. Its absorbance is directly proportional to GSH content. GSH concentration can be determined based on the volume of bacterial solution. The content of glutamic acid and GSH was determined using a kit (Solarbio, China).
Phytohormone identification and quantification
IAA reacted with Salkowski’s colorifier to change the color of the solution to pink, with a maximum absorption peak at 530 nm. The isolate was inoculated into LB liquid medium for 48 h (30℃, 170 rpm/min). Then, the culture solution was centrifuged at 4°C and 11,292 × g for 5 min, and 0.5 mL of supernatant was taken into a glass tube for the addition of 2.0 mL of Salkowski solution. After 20 min of culturing at room temperature, the absorbance of each tube was measured at 530 nm using a spectrophotometer. The concentration of IAA (μg/mL) in the medium was determined by absorbance based on a standard curve .
Bacterial cultures (NFb) in the exponential growth phase were separated into several 20 mL fractions for identification of abscisic acid (ABA), zeatin (ZA), salicylic acid (SA), jasmonic acid (JA), and gibberellins 3 (GA3). These fractions were centrifuged at 7,200 × g for 20 min at 4°C. The extraction method was as follows: a sample was removed, thawed, shaken, and centrifuged. Next, 1 mL of the supernatant and 0.5 mL of petroleum ether were mixed for triple extraction and decolorization, after which the upper ether phase was discarded, and the lower aqueous phase was adjusted to pH 2–3 using an aqueous 1 mM trichloroacetic acid solution. The triple extraction was performed using an equal volume of ethyl acetate, after which the extract was collected and combined with the ethyl acetate layer, mixed with hydrogen, and diluted with 0.5 mL in the mobile phase. Subsequently, an appropriate amount of solution was removed with a needle filter for testing. The peak area of each standard solution was determined successively according to the chromatographic conditions; the peak area was considered the vertical coordinate and the concentration was considered the horizontal coordinate in calculating the standard curve and correlation coefficient of hormones.
ABA HPLC liquid phase conditions were as follows: mobile phase – mobile phase A was methanol, whereas mobile phase B was 1% aqueous acetic acid in an isocratic elution (50% A + 50% B). ZA HPLC liquid phase conditions were as follows: mobile phase – mobile phase A was methanol, whereas mobile phase B was water in an isocratic elution (30% A + 70% B). SA HPLC liquid phase conditions were as follows: mobile phase–mobile phase A was methanol, whereas mobile phase B was 1% aqueous acetic acid in an isocratic elution (60% A + 40% B). JA HPLC liquid phase conditions were as follows: mobile phase – mobile phase A was acetonitrile, whereas mobile phase B was 0.1% aqueous phosphate in an isocratic elution (60% A + 40% B). GA3 HPLC liquid phase conditions were as follows: mobile phase – mobile phase A was methanol, whereas mobile phase B was 1% aqueous acetic acid in an isocratic elution (35% A + 65% B). The computer, detector, and pump were turned on, the column was installed, and the software was opened. Then, in the method group, the injection volume was set to 10 μL with a flow rate of 0.8 mL/min, column temperature of 35°C, sample time of 40 min, and UV wavelength of 254 nm. ABA, ZA, SA, JA, and GA3 were determined using the corresponding HPLC method at UV wavelengths of 254, 254, 294, 210, and 254 nm.
Screening for other PGP traits
Using yeast dextran as a substrate, the glucanase activity of a bacterial strain can be measured . The amount of reducing sugar in a reaction mixture can be determined using a dinitrosalicylic acid solution . The activity of the iron-producing isolate carrier was determined using chrome azurol S, and growth of the isolate on CAS medium produced an orange halo that was likely due to the activity of an iron carrier .
Volatile organic compounds (VOCs) produced by bacteria were determined by GCMS-TQ8050 (X). The fiber extraction head used 65 μm PDMS-DVB fibers, and a Rtx-5MS capillary column (60 m × 0.25 μm ID × 0.25 μm thickness film) with a slope of 5000 was used. The mass spectra of unknown compounds were compared with those in NIST17 and NIST17s (National institute of Standards and Technology) standard mass spectrometry libraries to determine the structure of the substance corresponding to the peaks.
The pathogenic fungi tested were Fusarium oxysporum, Fusarium pseudograminearum, Rhizoctonia solani, Fusarium graminearum, Botryosphaeria ribis, and Botryosphaeria dothidea. These are all soil-borne pathogenic fungi that cause wheat and other crop diseases [71, 72]. We inoculated pure pathogen cultures, applying each fungus (5 mm disk) to the center of a PDA plate (9 cm in diameter). After culture at 30℃ for 24 h, the bacterial isolate was introduced in plates around the fungal cultures, with a distance between the isolate and the center of the dish of 2.5 cm. All plates were cultured in a constant temperature incubator at 30℃ for 72 h, and the experiment was repeated three times. The percent growth inhibition was calculated as follows: inhibition (%) = [1 – (fungal growth/control growth)] × 100% . Each experiment was performed in triplicate, and the results were expressed as the mean with standard deviation.
Physiological and biochemical characterization
We used standard protocols for physiological and biochemical tests of our isolate, including Gram stain, starch agar, IMViC (indole, methyl red, Voges-Proskauer, citrate utilization test), and CAT testing . In addition, we used the BIOLOG identification system (BIOLOG Microstation, Biolog Inc, Hayward, CA) for biochemical testing with different carbon sources. The strain was assessed with 71 carbon sources and 23 chemical susceptibility assays according to BIOLOG instructions.
Test of colonization
Colonization of the strain was determined independently using three methods. (1) An HG-15 strain showing rifampicin and spectinomycin resistance was obtained, and the number of colonies in the wheat rhizosphere soil was evaluated on the 7th, 14th, 21st, and 28th day following inoculation. Colony forming units (CFU) per gram of soil were determined using a method previously described by Islam et al. . Each treatment and experiment were replicated three times. (2) The surfaces and cross sections of wheat roots, stems, and leaves were observed via scanning electron microscopy. HG-15 colonization was compared with that on the uninoculated control. (3) According to Singh and Jha , rep-PCR gene fingerprints of HG-15 strain and strains recovered from wheat rhizosphere soil, roots, stems, and leaves indicates colonization. We used a rep-PCR reaction mixture containing 25 μL 10 × Taq PCR master mix, 1 μL template DNA, 2 μL of BOX-AIR primer (CTACGGCAAGGCGACGCTGACG) (Sangon, China), and 22 μL of ddH2O. PCR amplification was performed under the following conditions: initial denaturation at 95°C for 7 min; 35 cycles of 94°C for 1 min, 53°C for 1 min, and 65°C for 8 min; and a final extension at 65°C for 16 min. (4) The surface of wheat was sonicated, treated with 75% alcohol for 2 min and 2% sodium hypochlorite for 10 min for disinfection, and finally washed five times with sterile water. Treated roots, stems, and leaves were cut separately into small pieces of approximately 0.5 cm, and tissue sections were brought into contact with PDA medium containing rifampicin and spectinomycin. The water in the final wash step was cultured to test the thoroughness of sterilization. The culture was carried out at 30°C for 3–6 d; the noninoculated wheats were used as the control. In this manner the HG-15 colonization of wheat roots, stems, and leaves was detected.
The soil used for potted plants was obtained at 0–20 cm of depth in wheat fields in the Yellow River delta (118°41′07′′E, 37°17′17′′N) (Dongying City, Shandong, China) in October 2018. The soil was brought to the greenhouse and passed through a 0.5 cm sieve. The number of bacteria cultured from the sample was 1.36 × 104 CFU·g-1 of dry weight soil. The upper and lower inner diameters of the flowerpot were 16.5 cm and 12 cm, respectively. The soil in each pot weighed 2.2 kg. The soil was also analyzed for its nutrient content. Various soil parameters were as follows: pH, 8.329; salt content, 0.1492%; EC, 456 μs/cm; organic matter, 23.51 g/kg; total N, 1.072 g/kg; Olsen-P, 0.0104 g/kg; K+, 0.6782 g/kg; Na+, 1.0162 g/kg; Ca2+, 0.23863 g/kg and Mg2+, 0.50805 g/kg.
The HG-15 strain was inoculated in LB liquid medium at 200 rpm and 30℃ for 12 h (logarithmic growth period). The bacterial suspension was then centrifuged at 1,073 × g for 10 min to harvest cells and resuspended immediately in sterile water three times. The HG-15 suspension (1 × 108 CFU/mL) was adjusted to its final concentration with sterile water. The experimental wheat variety used was Jimai 21 (The crop research institute of Shandong Academy of Agricultural Sciences considered (865168 / nongda 84-1109) F1 as the mother parent and ji 84-5418 as the father for sexual hybridization and systematic breeding). Wheat seed surfaces were sterilized with 70% ethanol (v/v) for 2 min and washed with disinfectant water three times. Seeds were then placed in 1% (w/v) NaClO solution for 3 min and rinsed with sterile water three times to remove residual sodium hypochlorite . A total of 72 pots were planted with 10 wheat seedlings in each pot. After 7 d, the above ground height of the wheat was approximately 5 cm; wheat with greater or less height was removed. In the treatment group, 20 mL of cell suspension was applied; in the control group, 20 mL of water was used instead. Plants were irrigated with NaCl solution for salt stress (0.15% 456 μs/cm, 0.25% 722 μs/cm, 0.35% 972 μs/cm). Inoculated and uninoculated bacteria were introduced into 12 pots for each concentration. The pots were placed in a completely random design, with each treatment repeated three times. The position of the pot was changed randomly during the trial to eliminate environmental errors. Greenhouse conditions were 20–28℃, 45-50% humidity, and natural lighting. The planting cycle was 28 d. The plants were watered thoroughly at sowing, and then watered once a week with an equal amount (300 mL) of water in each pot. The plants were not watered at 22–28 d. No nutrient solution or other fertilizer was added during the growth cycle. The combined weight of root and shoot was taken as the fresh weight of each plant. The dry weight of the plant was measured after placing a plant in the oven at 70℃ for 2 d. For accuracy and precision, each sample was tested in triplicate. The growth of plants was measured in terms of root length, plant height, fresh weight, and dry weight.
Effect of HG-15 on soil chemistry properties under NaCl stress conditions
Soil pH and EC values were analyzed with 1:2.5 and 1:5 soil-water ratios using digital pH (FE20) and EC (FE930) meters (Mettler Toledo, Switzerland), respectively. Organic carbon content was determined using 1 N potassium dichromate for titration and 0.5 N ferrous ammonium sulfate for back titration . Olsen P was extracted with 0.5 M NaHCO3 and determined according to the procedure described by Olsen et al. . Total N was measured using the Bremner  method.
Bulk soil was shaken off by uprooting the wheat. Soil still adhering to the root surface (concentrated within 2 mm of the root surface) was considered rhizosphere soil. We used a sterilized brush to collect the soil in a sterile bag on an ultra-clean worktable. For ion analysis, 0.2 g of wheat rhizosphere soil was treated with 1 mL deionized water and 5 mL concentrated sulfuric acid overnight, and then the cooked liquid was fixed to 50 mL. Measurements were carried out on 1 mL of the solution, which was extracted and diluted 10 times. Na+, K+, Ca2+ and Mg2+ content was measured via inductively coupled plasma optical emission spectroscopy (ICP, Thermo Scientific™ iCAP™ 7000 Plus, USA) .
Effects of HG-15 on plant growth under NaCl stress conditions
Photosynthetic characteristics and soluble sugar and proline content
Fresh leaves were treated overnight with absolute ethanol, and then chlorophyll (Chl) and carotenoids were extracted and measured spectrophotometrically using the method described by Arnon . Leaf soluble sugar was extracted from boiling water and quantified via the method by Thomas . Proline content was determined according to the method described by Bates , wherein valine was extracted with 3% sulfosalicylic acid and filtered. Next, an aliquot of the filtrate was supplemented with 1 mL ninhydrin and glacial acetic acid reagent. The mixture was then boiled for 1 h, placed on ice to stop the reaction, and the absorbance of the sample was measured at 520 nm using a UV spectrophotometer.
Net photosynthetic rate (A), stomatal conductance (gsw), intercellular CO2 concentration (Ci), and transpiration rate (E) of fully expanded leaves were analyzed using a LI-6800XT (Li-Cor, USA) portable photosynthetic apparatus after 28 d of treatment. The measurement time was 9:00–11:00, and the photon flux was set to 1200 μmol·m-2·s-1. Five leaves were examined for each treatment. Following an assessment of photosynthetic characteristics, chlorophyll fluorescence parameters were determined using an IMAGING-PAM (Walz, Effeltrich, Germany) fluorometer, and PSII primary light energy conversion efficiency was defined as Fv/Fm. Before measurement, the leaves were dark-adapted for 20 min, and three leaves were selected for each treatment. All operations were carried under background light intensity conditions less than 1 μmol·m-2·s-1. The intensity of saturated pulsed light of the instrument was 2400 μmol·m-2·s-1, and that of the measured light was less than 0.5 μmol·m-2·s-1 .
Biochemical analysis for osmolytes in plants after NaCl and bacterial inoculation
Approximately 0.2 g of fresh leaves was placed in a precooled mortar. First, 1 mL of 50 mmol/L buffer (containing 2% polyvinylpyrrolidone, pH 7.8, 4℃) was added, and the leaves were ground into a homogenate in an ice bath. Then, we washed the mortar with 0.5 mL of the above buffer solution to reach a final volume of 1.5 mL. After centrifugation at 12,000 × g and 4℃ for 20 min, the supernatant was considered the crude enzyme solution.
In the presence of H2O2, POD can oxidize guaiacol to form a tan substance, which can be measured by a spectrophotometer. We removed 200 mL of PBS (0.2 mol/L, pH 6.0), and 0.076 mL guaiacol (2-methoxyphenol) was added to the liquid, heated, and stirred to allow it to dissolve. After cooling, 0.112 mL 30% H2O2 was added, and the mixture was used as the reaction solution. Three milliliters of reaction solution were added to 10 μL crude enzyme solution, and 10 μL PBS was added to 3 mL reaction solution in the control group. POD activity was represented by △OD470/(mg·min).
Using SOD in the presence of oxidized substances, riboflavin can be reduced by light. The reduced riboflavin is easily reoxidized under aerobic conditions to produce O2, which can reduce nitro blue tetrazolium to a blue methyl trace, which has a maximum absorption at 560 nm. SOD can remove O2 and inhibit the formation of methyl hydrazone. The enzyme activity can be calculated by photoreduction. We prepared a phosphoric acid buffer (0.05 mol/L, pH 7.8), with 3.1 mL used in the treatment group and 3.2 mL in the blank group. Then, we added 1 mg/mL EDTA-Na2 (0.2 mL), 20 mg/mL L-methionine (0.2 mL), 0.1 mg/mL riboflavin solution (0.2 mL), and 1 mg/mL nitrogen blue tetrazole solution (0.2 mL) to each group. In the treatment group, 0.1 mL crude enzyme solution was added, whereas, in the blank group, no enzyme solution was added. The total volume of the reaction was 4 mL. The first group was treated with light for 30 min (4000 Lx), and the reaction was terminated with darkness. The second group was treated with darkness for 30 min. The OD560 was measured in the second group. The activity of SOD was based on 50% inhibition of photoreduction of nitroblue tetrazolium as an enzyme activity unit (U), and the activity of SOD was expressed as U·mg-1 .
CAT activity was determined based on ultraviolet absorption. H2O2 has strong absorption at 240 nm wavelength, CAT can decompose hydrogen peroxide, and the absorbance (A240) of reaction solution decreases with reaction time. The activity of CAT can be measured according to the rate of change in absorbance. We obtained 200 mL of PBS (0.15 mol/L, pH 7.0), added 0.3092 mL 30% H2O2, and shook the solution well. We then obtained 3 mL of reaction solution and added 50 μL of crude enzyme solution, (50 μL of PBS was added to the control group with 3 mL of reaction solution), with the absorbance at 240 nm read once every 1 min for 2 min in total. An extinction coefficient of 39.4 mM-1·cm-1 for H2O2 at 240 nm was used to calculate activity. The enzyme content of A240 decreased by 0.1 in 1 min was regarded as the enzyme activity unit (U·mg-1) .
Malondialdehyde (MDA) is one of the most important products of membrane lipid peroxidation. Under high temperature, MDA can react with thiobarbituric acid to form trimethyl Sichuan (3,5,5-trimethyloxazole 2,4-dione), which has an absorption peak at 532 nm and a smaller light absorption at 600 nm. According to the extinction value at 532 nm, the content of MDA in a solution can be calculated. We obtained 0.1 g of plant sample, added 3 mL of 10% TCA for grinding, and centrifuged the sample at 3000 × g for 10 min. We then then obtained 2 mL of supernatant, added 2 mL of 0.65% TBA solution, mixed well, boiled it in water for 15 min and cooled it quickly, measured the supernatant at OD440, OD532, and OD600, and reported the result as μmol·g-1 Pro .
Data analysis was performed using IBM SPSS 19.0. The Q-Q plot method was used to show that soil and plant parameters were normally distributed. For the same salt stress test with different treatments, t-testing (p < 0.05) was used. One-way ANOVA and Dunnett test were used to analyze the data under different salt stress conditions. A redundancy analysis (RDA) of soil and plant parameters was performed using Canoco 4.5.1 (Microcomputer Power, Ithaca, USA) software, and factors with significant explanatory functions were tested with conditional term effect analysis. Pearson test (two-tailed) was used to analyze the correlation between soil and plant indexes, and the results were made into a heat map using Origin 9.0 software.