Soil samples were collected from vineyards of Yangling (34°27′N, 108°08′E) and Jingyang (34°26′N, 108°29′E) in Shaanxi, Yongji (34°34′N, 110°15′E) in Shanxi, Wuwei (37°23′N, 101°59′E) in Gansu, and Yinchuan (38°70′N, 106°27′E) in Ningxia during May-June, 2012. The vineyard of Yangling belongs to the College of Enology of Northwest A&F University. The other four sampling sites are commercial vineyards.

Samples were taken from ~30 cm away from the grapevine stem. Approximately 10 g of the soil adhering to the roots of each individual grape plant, considered the rhizospheric soil, was used for the bacterial isolation procedures. Plant residues and stones were removed, and the fresh soil samples were stored in sealed sterile bags at 4°C (Karagöz et al., 2012).

Serially diluted soil samples were plated on the National Botanical Research Institute’s phosphate (NBRIP) medium, which contained 5.0 g tricalcium phosphate (TCP) as the sole P source (Nautiyal, 1999). Each sample was plated in triplicate with suitable soil concentrations. After incubation at 28°C for 72 h, all isolates forming clear halo zones were selected as PSB.

The PSA in liquid medium was measured by following the method of Johri et al. (1999) with modifications. Briefly, 1 mL sample of bacterial culture (10⁶ CFU/mL) was added to a 250-mL flask containing 50 mL of NBRIP. After incubation at 28°C on a rotary shaker (150 rpm) for 7 days, cultures were centrifuged at 10,000 rpm for 20 min. Supernatants were collected for pH and soluble P analyses. Phosphorus in the culture was determined by the Mo-blue method using a spectrophotometer at a wavelength of 700 nm (Watanabe & Olsen, 1965). A separate broth medium inoculated with sterile water served as the control treatment.

Bacterial genomic DNA was extracted using a bacterial genomic DNA extraction kit, and the 16S rRNA genes were amplified by polymerase chain reaction (PCR) using the universal primers fD1 (5’-AGAGTTTGATCCTGGCTCAG-3’) and rD1 (5’-AAGGAGGTGATCCAGCC-3’) (Weisburg et al., 1991). The 50-μL reaction mixture consisted of 5 μL of 10× reaction buffer, 4 μL of 200 mmol/L dNTP mixture, 2.5 μL of 10 μmol/L primer, 2.5 U of Taq DNA polymerase, and 50 ng of genomic DNA. The reaction conditions were as follows: 95°C for 3 min, followed by 35 cycles of 95°C for 1 min, 55°C for 1 min, and 72°C for 2 min, with a final extension step at 72°C for 3 min. The amplified products (nearly 1500 bp in length) were purified using an agarose gel DNA purification kit, and 16S rRNA sequencing was conducted at Beijing Huada Biological Company of China.

The top ten PSB strains, based on their PSA, were identified by 16S rRNA sequencing, and the acquired sequences were submitted to GenBank (http://www.ncbi.nlm.nih.gov/genbank/) with the accession numbers KC776273-KC776282 (Table 1). Based on the results of the database searches, sequences were aligned with representative bacterial sequences from the GenBank database by using ClustalW2 (http://www.ebi.ac.uk/Tools/msa/clustalw2/) to determine their approximate phylogenetic affiliations and 16S rRNA gene sequence similarities. The phylogenetic tree was constructed using the neighbor-joining method with distance matrices and the program MEGA (v6) after bootstrap analysis of 1000 replications (Chen et al., 2006).

Production of indole acetic acid (IAA). The selected strains were grown in a minimal medium (50 mM KH₂PO₄, 50 mM K₂HPO₄, 5 mM MgSO₄, 25 mM (NH₄)₂SO₄, 1% glucose) with 0.05% l-tryptophan at 28°C for 48 h in a shaking incubator at 150 rpm. A separate broth medium inoculated with sterile water served as the control treatment. The culture was centrifuged at 8,000 rpm for 10 min. The supernatant (1 mL) was vigorously mixed with 4 mL of Salkowski’s reagent. The mixture was incubated for 30 min at 25°C in the dark. The absorbance of the resulting solution was measured at 530 nm (Glickmann & Dessaux, 1995).

Production of ACC deaminase. The selected strains were inoculated on the Dworkin–Foster (DF) salts minimal medium containing 1-aminocyclopropane-1-carboxylate (ACC) as the sole N source. The chemical composition of DF salts minimal medium was as follows: 4 g KH₂PO₄, 6 g Na₂HPO₄, 0.2 g MgSO₄∙7H₂O, 1 mg FeSO₄∙7H₂O, 10 μg H₃BO₃, 10 μg MnSO₄, 70 μg ZnSO₄, 50 μg CuSO₄, 10 μg MoO₃, 2 g glucose, 2 g gluconic acid, 2 g citric acid, and 15 g agar in 1 L distilled water. The amount of ACC added to DF salts minimal medium was 0.3033 g/L. The isolates that grew on this medium were selected for further studies. Quantitative measurement of the ACC-deaminase activity of the selected isolates was performed by measuring the amount of α-ketobutyrate produced when the enzyme ACC deaminase cleaved ACC. The concentration of α-ketobutyrate (μmol) was determined by measuring the absorbance of samples at 540 nm (Penrose & Glick, 2003; Shahzad et al., 2013).

Production of siderophores. Qualitative measurement of the siderophore-producing capacity of the selected PSB strains was performed using the universal chrome azurol S (CAS) agar plate assay. On CAS agar plates, siderophore-producing (Sid+) bacteria form colonies with an orange halo because iron is removed from the original blue CAS–Fe (III) complex during siderophore production. Inoculated plates were evaluated for the formation of siderophore halos after incubation at 28°C for 7 days (Schwyn & Neilands, 1987).

Nitrogen-fixing ability. We made some modifications based on the methods available (Lin et al., 2012). The PSB strains were inoculated on modified nitrogen-deficient Ashby’s agar medium (10 g sucrose, 0.2 g NaCl, 0.2 g KH₂PO₄, 0.2 g MgSO₄·7H₂O, 0.1 g CaSO₄·2H₂O, 5 g CaCO₃, and 15 g agar in 1 L distilled water; pH 7.0). The strains that grew on this medium after incubation at 28°C for 7 days were considered to possess the ability to fix nitrogen.

Production of HCN.HCN production was detected according to the method developed by Bakker & Schippers (1987); all isolates were streaked on 10% tryptic soya agar with 4.4 g/L glycine. After 24 h of growth at 28°C, the plates were inverted, and an autoclaved filter paper soaked with picric acid (0.5%) and Na₂CO₃ (2%) solution was placed on each cover. The plates were sealed and incubated at 28°C for 48 h. HCN production is indicated by a change in coloration from orange to red. A strain of Pseudomonas sp. was used as a positive control (Bakker & Schippers, 1987).

Production of chitinase. Chitinase activity was evaluated by measuring the width of the edge of the translucent haloes surrounding the colonies grown on the plates after 48 h of incubation (Rojas-Avelizapa et al., 1999). The medium was prepared with 13-14% (wet weight) colloidal chitin and 2.3% agar dissolved in Castaneda medium (0.625 g diammonium citrate, 0.250 g NaCl, 0.375 g KH₂PO₄, 0.125 g MgSO₄∙7H₂O, 0.375 g Na₂CO₃, 6.5 mL glycerol, in 1 L distilled water; pH 6.5-7.0).

Production of protease. The strains were inoculated on a medium composed of 1% casein and 2.3% agar dissolved in Castaneda medium; the pH was adjusted to 6.5-7.0 to avoid casein precipitation. Casein hydrolysis was detected by the formation of a whitish, opaque halo (coagulated casein) around a translucent area (totally hydrolyzed casein) surrounding the colony (Rojas-Avelizapa et al., 1999).

In a 100-mL flask containing 50 mL of NBRIP, 0.5 mL of the bacterial culture (concentration, 10⁸ CFU/mL) was added. The samples were incubated for 3 days at 150 rpm under various conditions of pH (7.0, 9.0, and 11.0), NaCl concentration (0%, 2.5%, and 5.0%) and temperature (15°C, 28°C, and 40°C). A separate broth medium inoculated with sterile water served as the control treatment. Soluble P content was detected under all these conditions. All treatments were performed in triplicate, and the data are presented using the means of these triplicates (Banerjee et al., 2010).

Experiments were conducted in the greenhouse (average air temperature between 18°C and 30°C; average relative humidity between 50% and 60%) of the College of Enology, Northwest A&F University, Yangling, Shaanxi, China (34°27′N, 108°08′E) during February-July, 2013.

The soil samples used in this experiment were a mixture of dark loessial soil and river sand (1:1, v/v), high-temperature sterilized at 121°C for 20 min. The basic soil properties were 0.84 g/kg total N, 0.63 g/kg total P, 25.46 g/kg total K, 67 mg/kg available N, 14 mg/kg available P, 111 mg/kg available K, 11.4 g/kg organic matter and pH 8.08 (Bao, 2000).

Cuttings of V. vinifera cv. Cabernet Sauvignon (from the vineyard of the College of Enology, Northwest A&F University) were soaked in 0.1% HgCl₂ for 5 min, repeatedly washed with sterilized water, soaked in 25 mg/kg 1-naphthaleneacetic acid solution for 12 h, and then planted in nutritive bowls. After 3-4 leaves had grown on the seedlings, they were transplanted to ethanol-disinfected plastic pots (inner diameter, 23 cm; height, 17 cm).

Two strains of Bacillus aryabhattai, JY17 and JY22, were grown separately in the nutrient broth at 28°C in a shaker (150 rpm) for 24 h. Then, 50 mL of the bacterial suspensions (10⁸ CFU/mL) of two strains were separately inoculated into the middle part of the seedling roots. Sterilized nutrient broth (30 mL) was applied as the control treatment. Experiments were performed in a completely randomized block design. Three treatments (B. aryabhattai JY17, B. aryabhattai JY22 and control) were performed with 5 independent replications, 10 pots per replication. The PGP effects of bacterial treatment were assessed by measuring plant height, root and shoot dry weight (Yu et al., 2012) and stem thickness after the grapevine was grown for 30 days.

A statistical analysis was conducted using an analysis of variance (ANOVA) in the Statistical Package for Social Sciences (SPSS), version 21, followed by a comparison of multiple treatment levels with the control. Duncan’s multiple range test was used for multiple mean comparisons at p <0.05.

Sixty-six PSB strains were obtained from the rhizospheres of grapevine plants and screened based on halo formation in NBRIP medium. Among them, 23 strains were isolated from Jingyang, 14 from Wuwei, 11 from Yangling, 11 from Yinchuan, and 7 from Yongji. These PSB showed different abilities in TCP solubilization and soluble P concentrations varying from 19.41-673.99 mg/L, significantly higher than the control (11.80 mg/L). JY22 showed the highest P solubilization, followed by JY17 with a soluble P concentration of 671.64 mg/L. The ten strains with soluble P concentrations above 500 mg/L were all isolated from Jingyang.

The aforementioned ten PSB strains screened by the authors belonged to 4 genera and 5 species: Bacillus aryabhattai, Bacillus megaterium, Klebsiella variicola, Stenotrophomonas rhizophila, and Enterobacter aerogenes (Table 1). The strains JY3 and JY8 exhibited a sequence identity of up to 99% with K. variicola, whereas the sequences of JY10 and JY15 exhibited a strong similarity to the 16S rRNA sequence of S. rhizophila. The sequences of JY2, JY17, JY22, and JY26 corresponded to those of B. aryabhattai, and the closest identified phylogenetic relatives of JY5 and JY11 were B. megaterium and E. aerogenes, respectively. The phylogenetic tree expressing their relationships is shown in Fig. 1.

The selected bacterial isolates were tested for their PGP traits (Table 2). All isolated strains had the ability to produce IAA and siderophores. Production of IAA by the isolates ranged from 2.23 to 31.03 mg/L. JY22 showed the highest IAA production, followed by JY10, JY2, and JY17 (27.54 mg/L, 23.24 mg/L and 21.21 mg/L, respectively). Except for JY10, the other isolates produced ACC deaminase, with activity varying from 4.84 to 245.89 nmol/mg·h, and JY2 had the highest activity. The isolates JY2, JY3, JY5, JY8, JY17, JY22, and JY26 were able to grow on the N-free Ashby agar medium, signifying that they possessed nitrogen-fixing abilities. The PSB isolates producing HCN included JY3, JY8, JY11, and JY22. The chitinase activity could be detected in seven strains, and all the ten isolates could secrete protease.

Under alkali stress conditions, the soluble P content in NBRIP broth significantly decreased with increasing pH, indicating that the PSA of PSB isolates was inhibited by the high initial pH of the broth medium (Fig. 2A). However, the soluble P concentration at all pH values was >200 mg/L. Under non-stress conditions, there were no significant differences between the soluble P concentrations of JY17 and JY22, but when the initial pH was 9.0, P solubilization by JY22 was significantly lower than that by JY17, showing that JY17 had better tolerance to alkali stress. When the initial pH was 11.0, the soluble P content of JY17 and JY22 was reduced by 37.6% and 40.1%, respectively, which was lower than that during the non-stress condition.

These two strains showed efficient phosphate solubilization at a NaCl concentration of 5% with the amount of solubilized phosphate exceeding 200 mg/L (Fig. 2B). Soluble P levels in JY17 and JY22 at a NaCl concentration of 2.5% were even higher than those under non-stress conditions; this indicates that moderate salt may stimulate the P-solubilizing abilities of certain PSB strains.

The soluble P levels of JY17 and JY22 increased with elevated temperature, and low temperature (15°C) clearly inhibited their PSA (Fig. 2C). However, high temperature (40°C) had no negative impacts on their PSA. The amount of soluble P at 40°C was significantly higher than that at 28°C and 15°C (increased by 33.39% and 38.23%, respectively). The PSA of JY22 at 40°C was the highest, reaching 462.30 mg/L.

PSB treatment increased the biometric parameters of V. vinifera cv. Cabernet Sauvignon (plant height, stem thickness, root and shoot dry weight) over those of the control plants (Table 3). The degree of stimulation varied with respect to the growth parameter. PSB treatment had the maximum stimulatory effect on shoot dry weight, whereas the plant height was only slightly influenced compared with other parameters. The root/shoot ratio of PSB-treated plants was significantly smaller than that of the control plants.