Rhizosphere soil samples used in the study were collected from robust mulberry trees located in the hydro-fluctuation belt in Longjiao Town, Yunyang County, Chongqing Municipality, China (30°49′26″ N, 108°52′14″ E), at an elevation of approximately 172 m. Soil adhering to the surface of the root (approximately 2–5 mm) was obtained from three individual mulberry trees in September 2016 (following several months of drought stress). The primary root diameter was 2.0–2.5 cm, and the sampling depth was 15 cm below the surface. The water content of the soil was 2.8%, organic carbon was 19.2 g/kg, available phosphorus was 8.3 mg/kg, pH was 6.7, and available potassium was 134 mg/kg. The three replicated samples were transported back to the laboratory and stored at 4°C prior to further processing.

Rhizobacteria were isolated from the serial dilutions of rhizosphere soil. Briefly, 1 g of mixed soil was added to 100 mL of sterile distilled water to create a suspension that was then serially diluted by 10-fold. Subsequently, 100 μL suspensions of 1:1,000 to 1:100,000 dilutions were transferred into water agar (WA), Gause’s agar (GA), Luria-Bertani agar (LB), nutrient agar (NA), and potato dextrose agar (PDA) and incubated at 30°C. The cultures were observed daily and colonies with different morphological characteristics were purified as they appeared on the plates. Each isolate was cultured in potato dextrose broth (PDB) at 180 revolutions per minute (rpm) at 28°C for 16 h. Glycerol was then added to a final concentration of 50%, and the suspensions were stored at −80°C.

Molecular identification of the obtained isolates was performed as previously described by Xu et al. (2019). Genomic DNA was extracted using a PrepMan Ultra Sample Preparation Reagent kit (Applied Biosystems, Palo Alto, CA, United States), and the 16S rRNA gene was amplified using primers 27F/1492R. The amplified products were sequenced at Sangon Biotechnology Co., Ltd., Shanghai, China, and the generated sequences were analyzed against the NCBI database using BLAST to determine the sequence homology with ribosomal genes of closely related organisms. The isolates were classified at the genus level when the BLAST results indicated >97% identity.

The ability of isolates to produce IAA and solubilize phosphate was assessed in LB (Gordon and Weber, 1951) and Pikovskaya (PVK) medium (Ames, 1964), respectively. Each purified isolate was cultured in an LB medium and a PVK medium at 28°C with three replicates. The culture suspension was then centrifuged at 12,000 rpm for 20 min. IAA activity of the supernatant was assayed daily by adding 2 mL of Salkowsky’s reagent (50 mL of 35% perchloric acid and 1 mL of 0.5 M FeCl₃ solution) into 1 mL of supernatant and maintaining the solution in the dark for 20 min. Phosphate-solubilizing ability was assessed by adding 3 mL of ammonium vanadate-molybdate reagent (100 mL of 17.7% ammonium molybdate solution and 100 mL of 0.6% ammonium metavanadate solution with 33.3% nitric acid) into 1 mL of supernatant. The optical density (OD) of the prepared solutions was measured at 530 and 490 nm to determine the content of IAA and phosphate, respectively.

HGS7 strain was selected for further examination based on the plant growth-promoting traits exhibited in vitro. Genomic DNA was extracted using PrepMan Ultra Sample Preparation Reagent (Applied Biosystems, United States) according to the manufacturer’s instructions. The genome was sequenced using a Pacific Biosciences RSII (PacBio, Menlo Park, CA, United States) platform. De novo assembly of the genome from PacBio sequence reads was performed using the hierarchical genome-assembly process (HGAP) software. GeneMark software was used to predict the open reading frames in the genome sequence. tRNAs and rRNAs were predicted using tRNAscan and rRNAmmer software, respectively. Gene functions were annotated using a variety of function-related databases, including the Non-Redundant (Nr) protein database, the Swissprot database, the Pfam database, and the Clusters of Orthologous Groups (COG) protein database. A circular chromosomal genome map with COG functional annotations was plotted using Circos v0.62, and secondary metabolite synthesis gene clusters were predicted using antiSMASH 4.0. The general features and sequences of other representative B. megaterium strains were downloaded from the NCBI database.

The morphological characteristics of HGS7 grown on LB plates at 28°C for 24 h were recorded. Gram staining and spore staining were performed as described by Benson (2002) and observed under an optical microscope. A series of biochemical tests, including Voges-Proskauer and starch hydrolysis tests, were conducted using an HK-MID-66 kit (HUANKAI, China) following the protocol provided by the manufacturer. Phylogenetic trees of HGS7 based on 16S rRNA gene sequences and the single-copy nuclear genes of 15 bacterial strains related to the Bacillus genus were constructed by applying the neighbor-joining method using MEGA version 6.0 with 1,000 replicates of bootstrap values (Tamura et al., 2011).

The antagonistic activity of the HGS7 strain against several plant pathogens was evaluated in vitro. HGS7 cells were cultured in modified PDB (MPDB: 200 g of potato, 20 g of maltose, 10 g of peptone, 5 g of (NH₄)₂SO₄, and 1.5 g of Na₂HPO₄ per liter) at 28°C for 3 days at 180 rpm. Cell-free supernatant was obtained by passing it through a 0.22-μm micropore filter that was then introduced into a sterile PDA medium (10%, v/v). Fresh plugs of pathogens (5 mm diameter) grown in PDA were inoculated onto a PDA containing cell-free supernatant and incubated at 25°C. Pathogens cultured on PDA plates without supernatant served as controls. Each treatment was carried out in triplicate. The diameter of each fungal colony was determined from the average of measurements made in three different orientations, and the level of inhibition was calculated as follows: percent inhibition = (control colony diameter – treated colony diameter)/control colony diameter × 100%.

The abiotic stress tolerance of HGS7 was assessed in vitro. Freshly cultured HGS7 cells were inoculated in LB broth amended with different concentrations of NaCl (0, 2, 4, 6, 8, 10, and 12%; w/v) and then incubated at 28°C at 180 rpm. HGS7 strain was also cultured in LB in which pH was adjusted with 1 M HCl and 1 M NaOH to obtain a pH range of 3.0–10.0 (Rajnish et al., 2015) and incubated at 28°C at 180 rpm. Finally, the HGS7 strain was cultured in LB at 180 rpm at different temperatures (8, 15, 22, 29, 36, 43, 50, and 57°C). Growth was determined in all cases by measuring the optical density of cultures at 600 nm after 24 h.

HGS7 cells were cultured in King’s B medium for 18 h at 28°C at 180 rpm. The fresh culture was centrifuged at 5,000 rpm for 10 min and resuspended in sterilized distilled water to obtain final concentrations of 1.0 × 10⁵, 1.0 × 10⁶, 1.0 × 10⁷, and 1.0 × 10⁸ colony forming units per milliliter (CFU/mL).

Healthy mulberry seeds (“Guisangyou 62”) were disinfected and germinated as described by Xie et al. (2017). The surface-disinfected mulberry seeds were immersed in bacterial suspensions of different concentrations for 24 h. Seeds soaked in sterilized distilled water were used as a control. Each treatment comprised five replicates of 30 seeds that were placed in a 9-cm-diameter petri dish containing sterilized moistened filter paper and maintained at 25°C with a photoperiod of 12 h at 200 μmol⋅m^–2⋅s^–1 and 70% humidity. The germination potential implying the vitality of seeds was determined 4 days after inoculation. The seed germination rate and radicle/plumule length of the germinated seeds were calculated after 15 days.

Surface-sterilized seeds were sown in plastic pots (12 × 12 cm² diameter) containing 300 g of sterilized soil and cultivated at 25°C with a photoperiod of 12 h at 200 μmol⋅m^–2⋅s^–1 and 70% humidity. At the two-leaf stage, 30 mL of bacterial suspensions of different concentrations or sterile water alone were applied to each pot. Five independent replicates were used for each treatment. Fifty days after the application of HGS7 suspension to the soil, the root/shoot length, fresh weight (FW), and dry weight (DW) of three randomly selected individual seedlings were measured.

Polyethylene glycol 6000 (PEG) was used to simulate drought stress (Caruso et al., 2008) and was used at concentrations of 5, 10, 15, and 20% PEG (w/v) to yield −0.035, −0.166, −0.325, and −0.534 MPa water potential, respectively. As mentioned earlier, 90 days after inoculation of suspension of HGS7 (1.0 × 10⁷ CFU/mL) onto mulberry seedlings, 30 mL of PEG of different concentrations or sterilized water alone were applied to each pot. Mulberry trees were cultivated at 25°C with a photoperiod of 12 h at 200 μmol⋅m^–2⋅s^–1 and 70% humidity. Each treatment was replicated 10 times. Then, 7 days after the treatment with PEG, four seedlings were randomly collected and immediately frozen in liquid nitrogen. Proline levels and the activity of antioxidant enzymes, including superoxide dismutase (SOD), catalase (CAT), and peroxidase (POD), were then analyzed using biochemical kits (Suzhou Comin Biotechnology Co., Ltd., China) following the manufacturer’s instructions. The remaining seedlings were irrigated with sterilized water to evaluate the recovery potential. Fifty milliliters of water were initially applied to each pot to flush out the PEG from the soil. Thereafter, 30 mL of sterilized water was added to each pot every 3 days. Shoot and root length and fresh and dry weight of four randomly selected mulberry seedlings from each treatment were calculated 105 days after the initiation of recovery.

Data analyses were carried out using SPSS (version 17.0). Differences between treatments in the seed germination, seedling growth, and the effect of HGS7 on mulberry drought tolerance were analyzed using Tukey’s one-way analysis of variance (ANOVA). ***, **, and * indicated significant differences at P < 0.001, P < 0.01, and P < 0.05, respectively.

A total of 74 isolates were obtained in September after drought stress from rhizosphere soil of robust mulberry growing in the hydro-fluctuation belt of the Three Gorges Reservoir. The bacteria were classified into four phyla and 16 genera (Table 1) based on their near full-length 16S rRNA gene sequences which were deposited in the GenBank under accession numbers MK602381-MK602394, MK602519-MK602577, and MK583533. The predominant phylum was Proteobacteria (48 of the 74 isolates, 64.9%), followed by Firmicutes (19 of the 74 isolates, 25.7%). Pseudomonadales (60.4%), Enterobacteriales (31.3%), and Burkholderiales (8.3%) were the most frequently detected taxonomic orders within the 48 Proteobacterial isolates, while only the order Bacillales was detected within Firmicutes. The dominant genera in the collection of 74 isolates were Pseudomonas (32.4%), Bacillus (25.7%), and Klebsiella (12.2%). Members of Acinetobacter, belonging to the family Moraxellaceae, also accounted for 5.4% of the total culturable bacterial population. The relative abundance of the remaining genera was less than 5%, with single isolates of Rugamonas, Enterobacter, and Leclercia.

A total of 12 isolates showed P solubilization (Figure 1A). After 1 day of growth, Bacillus sp. HGS7, Pseudomonas sp. HNS5, and Acinetobacter sp. HLS7 solubilized more phosphate of 60.4, 51.7, and 54.1 μg/mL, respectively. While slower, Bacillus sp. HPS2 and Lelliottia sp. HNS13 also solubilized substantial phosphorous (over 50 μg/mL) after 2 days of incubation. Overall, Bacillus sp. HGS7, Bacillus sp. HPS2, Lelliottia sp. HNS13, and Acinetobacter sp. HLS7 exhibited the greatest phosphorous-solubilizing capacity, yielding soluble phosphate concentrations of 70, 87.4, 76.1, and 93.4 μg/mL after 3 days of incubation, respectively.

A total of 10 out of the 74 isolates produced IAA (Figure 1B), with Klebsiella sp. HGS6 and Pantoea sp. HNS2 accumulating IAA at concentrations of 487.4 and 435.5 μg/mL by 1 day of incubation, respectively. The production of IAA by Klebsiella sp. HGS6 and Pantoea sp. HNS2, however, did not increase with time. In contrast, IAA accumulation in Bacillus sp. HGS7 significantly increased over time, measuring 909.7 and 887.7 μg/mL by days 2 and 3, respectively.

Bacillus sp. HGS7 was screened for more extensive analyses due to its prominent traits previously associated with plant growth promotion, such as IAA production and phosphate-solubilizing activity. A quantitative analysis of phosphate concentration released in cultured HGS7 cells revealed concentrations of 70 μg/mL by day 3, after which the concentration gradually declined (Supplementary Figure 1A). This strain also produced high amounts of IAA, reaching concentrations of 242.8 μg/mL by day 1 and 1,073.6 μg/mL by day 4 (Supplementary Figure 1B).

The HGS7 strain forms white colonies on LB plates (Figure 2A) and is a Gram-positive (Figure 2B), spore-forming (Figure 2C), and rod-shaped bacterium. Biochemical assays indicated that HGS7 was negative for the Voges-Proskauer reaction and starch hydrolysis but positive for the utilization of sucrose and lactose (Supplementary Table 1). The phylogenetic analysis of the sequence of the full-length 16S rRNA gene of HGS7 indicated that it clustered most closely to Bacillus megaterium and in the same minimal clade as B. megaterium (NR117473) (Figure 2D). A phylogenetic tree based on the single-copy nuclear genes revealed that HGS7 was most closely related to B. megaterium (CP003017) (Figure 2E). Therefore, the HGS7 strain was identified as Bacillus megaterium based on its morphological, biochemical, and molecular characteristics.

The complete circular genome of HGS7 comprises 5.03 Mb, and this strain also has three circular plasmids: p01 (141,167 bp), p02 (75,747 bp), and p03 (16,871 bp) (Table 2 and Figure 3). The G + C contents of the chromosome and three plasmids are 38.27, 33.99, 36.17, and 34.74%, respectively. The chromosome possesses 5,214 predicted coding sequences (CDSs) with an average length of 806 bp. Among the predicted genes, 1,913 (36.7%) could be classified into COG families composed of 21 categories (Supplementary Table 2). Annotations were assigned to 5,169 (99.14%) putative genes based on similarity searches within the Nr database, while 45 (0.86%) CDSs were predicted to encode proteins with unknown or hypothetical functions (Table 2). The general features of B. megaterium HGS7 genome and other representative B. megaterium strains are summarized in Supplementary Table 3. Moreover, the bioinformatic analysis showed that the HGS7 genome contained seven putative gene clusters for the biosynthesis of secondary metabolites using antiSMASH, which included three terpenes, a phosphonate, a type III polyketide synthase (T3pk), a lasso peptide, and a siderophore (Supplementary Table 4).

The analysis of the whole annotated genome sequence revealed that HGS7 contained several genes that likely contribute to plant growth promotion and stress tolerance enhancement. Genes involved in the synthesis of amidohydrolase and aldehyde dehydrogenase were identified, which could convert indole-3-acetamide and indole-3-acetaldehyde to IAA (Table 3 and Supplementary Table 5; Shao et al., 2015; Liu et al., 2016). A potential phosphate solubilization pathway was also detected, wherein gdhII and Hhipl2 encode a glucose dehydrogenase, which could catalyze glucose to gluconic acid. This product could solubilize mineral phosphates (Gupta et al., 2012). Additionally, rhbC, yfmD, and hemH genes were involved in the biosynthesis of ferrochelatase and siderophore ABC transporter permease. Other genes (alsS, alsD, bdhA, and aco) enable the production of volatile organic compounds (VOCs) (Table 3 and Supplementary Table 5). Moreover, HGS7 contained genes that enable the metabolism of antibiotics, such as chalcone (pks11), bacitracin (bcr, bce, nprM), and thiazole (GLX3, thiF). In addition to the plant growth-promoting and antifungal traits, B. megaterium HGS7 also contained several genes that encoded for metabolites contributing to abiotic stress tolerance (Table 3 and Supplementary Table 5). For example, the tre gene cluster encoded trehalose production, a compound that functions as an osmoprotectant (Duan et al., 2013). The genes ytlC, spe, and pot involved in the production of spermidine, which are capable of scavenging free radicals, were also annotated in the genome (Zhang H. W. et al., 2017). Additionally, several genes responsible for the production of cold/heat/alkaline-shock proteins and oxidative stress enzymes, such as superoxide dismutase (sod) (Wu et al., 2016), catalase (kat) (Shin et al., 2008), and peroxidase (PRXQ, per, ykuU) (Su et al., 2018), were also identified in the genome, all of which potentially protect bacterial cells from exogenous oxidative stress.

Given that the HGS7 strain contained multiple genes encoding antimicrobial substances, its antagonistic effect on the growth of various phytopathogens was assessed in vitro. Several plant pathogens were inhibited by the cell-free supernatant of HGS7 (Supplementary Figure 2). It strongly inhibited the hyphal growth of Sclerotinia sclerotiorum and Scleromitrula shiraiana, causal agents of mulberry fruit disease (Table 4). Cell-free supernatants of HGS7 also inhibited Fusarium oxysporum, Fusarium solani, and Phoma exigua, but to a little extent.

The abiotic tolerance of HGS7 was assessed to determine whether it had the ability to grow in stressful environments. This strain exhibited exceptional tolerance to salt in the media containing up to 8% NaCl and even showed limited growth in 10% NaCl (Figure 4A). Moreover, HGS7 could grow well in media with pH ranging from 5.0 to 9.0 (Figure 4B) and at temperatures from 15 to 50°C, although the growth rate was less at temperatures above 43°C (Figure 4C).

Given that many genes are involved in the biosynthesis of potential plant growth-promoting and stress mitigation substances, the effects of HGS7 on the growth and survival of mulberry seedlings were evaluated.

The immersion of seeds in a cell suspension of HGS7 at various concentrations significantly (P < 0.05) increased the rate of seed germination (Figure 5A). Seedling growth was also enhanced after inoculation with cells at different concentrations (Figures 5B,C) compared to control plants. The greatest germination potential was observed when the seeds were immersed in a 1.0 × 10⁷ CFU/mL suspension of HGS7 (Figure 5A). The concentration of 1.0 × 10⁶CFU/mL suspension also significantly (P < 0.01) increased the radicle length by 21.67% (from 26.42 to 33.73 mm) (Figure 5B). An increase in the germination potential/rate and biomass (FW and DW) was also detected when mulberry seeds were immersed in a 1.0 × 10⁵ CFU/mL bacterial suspension of HGS7 (Figures 5A,C). However, the radicle length was lower than that of the control at this concentration of HGS7 (Figure 5B).

The application of HGS7 cell suspension to mulberry seedlings also resulted in an increase in the root and shoot growth relative to water control (Figure 5 and Supplementary Figure 3). The highest growth promotion was observed in 1.0 × 10⁷ CFU/mL HGS7 treatment, where length, fresh weight, and dry weight of roots increased by 56.11, 229.88, and 213.25%, respectively (Figures 5D–F). This treatment also enhanced the fresh and dry weights of shoots by 297.29 and 369.77%, respectively (P < 0.001) (Figures 5E,F). Moreover, the 1.0 × 10⁶ CFU/mL HGS7 treatment facilitated seedling growth, with the dry and fresh weight of shoots increasing by 166.74 and 135.95%, respectively (Figures 5E,F).

The capability of HGS7 to enhance drought stress tolerance in mulberry was assessed by analyzing the biochemical properties of mulberry leaves after inoculating the plants. Although the concentration of proline increased with escalating drought stress even in control plants, seedlings inoculated with HGS7 had significantly higher proline concentrations than control plants in the presence of the highest PEG concentrations (Figure 6A). The largest increase in proline concentration (51%) was seen in plants treated with HGS7 that were not exposed to PEG, while it was 15 and 41% higher after the application of 15 and 20% PEG, respectively (Figure 6A). Moreover, HGS7-inoculated mulberry also exhibited higher reactive oxygen species (ROS) scavenging capacity, as indicated by the elevated levels of antioxidant enzyme activity. Specifically, SOD activity was significantly (P < 0.01) higher in HGS7-inoculated mulberry than in control mulberry treated with 5% PEG, where SOD increased from 29.5 to 89.1 U/g (Figure 6B). Additionally, HGS7 significantly (P < 0.05) enhanced mulberry POD activity, which increased by 59.39, 44.76, 44.90, and 36.29% compared to that of control plants treated with 0, 5, 10, and 15% PEG, respectively (Figure 6C). CAT activity was also significantly (P < 0.01) higher in HGS7-inoculated plants than in controls, increasing by 80.20 and 53.01% in plants treated with 5 and 10% PEG (Figure 6D).

Seven days after drought stress, mulberry trees were irrigated with sterile water to test their ability to recover from drought stress. Mulberry seedlings inoculated with HGS7 recovered from drought stress, conferred by higher concentrations of added PEG, and were much better than control plants. About 105 days after the end of imposition of drought stress, in the groups applied with 5% PEG, mulberry treated with HGS7 had 45.7% longer roots than control plants (Figure 7A). Moreover, the application of HGS7 enhanced the biomass of mulberry roots compared to the controls. The fresh weight of HGS7-treated plants was increased (45.7, 83.0, and 74.7%) (Figure 7B), and their dry weight was also increased (84.6, 87.8, and 85.2%) in plants treated with 0, 5, and 10% PEG, respectively (Figure 7C). In addition, mulberry seedlings treated with HGS7 had longer shoots than controls, that is, 42.5, 45.1, 46.3, 30.9, and 37.01% (P < 0.05) longer in plants treated with PEG, respectively (Figure 7D). The shoot biomass of the HGS7-inoculated plants treated with 0, 5, 10, 15, and 20% PEG was also significantly higher (Figures 7E,F).