A fresh S. borealis (Hack.) Nakai plant was collected from Kyungpook National University, Daegu, South Korea. To isolate rhizospheric bacteria, a rhizospheric soil sample was processed based on the methods of Fischer et al. (2007). Endophytic bacteria present in the sample were isolated following a previously described method (Albdaiwi et al., 2019). Briefly, healthy roots were washed thoroughly under running tap water to remove adhering soil particles and then rinsed in double distilled water several times. Root samples were cut into small pieces and then surface sterilized in 70% ethanol for 1 min and 3% sodium hypochlorite solution for 10 min. Subsequently, they were rinsed in sterile distilled water three times. To validate root surface sterilization efficiency, an aliquot (20 μl) from the third wash solution was dispersed on Luria–Bertani (LB) medium and incubated at 28°C ± 2°C for 1 week. After that, sterilized roots (1 g) were aseptically crushed with a mortar and pestle using 10 ml of 1% NaCl and serial dilutions were prepared. Here, 1 ml of the root extract and the diluents were spread on LB agar medium. The plates were incubated at 28°C and surveilled up to 1 week for bacterial colony emergence. Morphologically dissimilar colonies of bacteria were designated and refined with streaking on LB agar medium. In total, 11 bacterial pure cultures were selected and preserved in corresponding growth medium with 25% glycerol at −20°C until further use. The selected colonies were named B1–B11, respectively.

All experiments were repeated three times, and each was conducted following a completely randomized design with three replications for each strain. The uninoculated medium served as a negative control.

A citrate test of bacterial strains was conducted by adopting the method of Simmons (1926). Citrate utilization ability was confirmed by the change in the color from green to blue. Siderophore production was examined using chrome azurol S (CAS) agar media following a previously described method (Alexander and Zuberer, 1991). The formation of a clear halo zone around the colonies was considered the result of siderophore production. Nitrogen-free bromothymol blue malate (NFb) medium was used for visual detection of nitrogen-fixing activities of bacterial strains (Baldani et al., 2014). A color change from green to blue indicated that the strain probably had nitrogen-fixing activity in the solid culture conditions. IAA production was determined following the methods of Sarwar and Kremer (1995). The appearance of a pink-red color confirmed IAA production by bacterial strains. The ACC deaminase activity of bacterial strains was assessed on sterile minimal Dworkin and Foster (DF) salts media containing 3-mM ACC (Sigma-Aldrich, United States) as a sole nitrogen source (Penrose and Glick, 2003). Bacterial strains multiplying on the ACC-supplemented plates were considered ACC deaminase producers (Ali et al., 2014).

The ability of bacterial strains to solubilize inorganic phosphate was detected using Pikovskaya agar medium supplemented with tricalcium phosphate (Gaur, 1990). For determining potassium solubilization, the bacterial cultures were spot-inoculated onto Aleksandrov medium containing mica powder (Hu et al., 2006). Zinc solubilizing activity was determined in mineral salts agar medium supplemented with 0.1% zinc oxide (Venkatakrishnan et al., 2003). Si solubilization was assessed in the basal media supplemented with potassium alumina silicate (Sharma et al., 2011). The development of a clear zone around the bacterial colonies was considered to indicate solubilizing activity.

The experiment was executed in triplicate with control and multiple test concentrations. The unmodified medium served as a control.

The intrinsic salt tolerance of bacterial strains was examined by observing their growth on the LB agar medium, which was supplemented with various NaCl concentrations (0.5%, 2.5%, 5%, 7.5%, and 10%). The plates were incubated for 72 h at 28°C ± 2°C (Barra et al., 2016).

The drought endurance of strains was assessed in polyethylene glycol (PEG 6000)-supplemented LB agar medium. The strain growth was surveyed at 28°C ± 2°C for 24 h in LB agar media with different osmotic potentials (−0.05, −0.15, −0.3, −0.49, and −0.73 MPa) (Ali et al., 2014).

The heavy metal tolerance of bacterial strains was examined on LB agar medium supplemented with varied cadmium (Cd), chromium (Cr), and nickel (Ni) concentrations ranging from 0.4 to 1.0 g L^–1. The significant growth of bacterial strains in the presence of heavy metals during 72 h at 28°C ± 2°C was considered heavy metal tolerance (Cervantes-Vega et al., 1986).

The selected bacterial strain (B11) was identified using 16S rDNA and the partial sequence of the gene encoding the subunit B protein of DNA gyrase (gyrB) sequencing and phylogenetic analysis. For this purpose, total genomic DNA was isolated as described by Chang et al. (2011). Polymerase chain reaction (PCR) assays were performed for 16S rDNA and gyrB regions using primers 27F/1492R and UP1F/UP2R (Supplementary Table 1). The preparation of the PCR reaction mixture and PCR amplification were completed following Ki et al. (2009) and Ahaotu et al. (2013). PCR products were purified and sequenced at SolGent Co., Ltd. (Daejeon, South Korea). The phylogenetic analysis of bacterial strain sequences was performed using RAxML GUI v.1.3 (Silvestro and Michalak, 2012).

Based on the results of in vitro assays, the B11 bacterial strain was chosen and used for further assays. Specifically, the growth-promoting and phytoremediation effects of B11 on the growth of pepper under different abiotic stress conditions were assessed.

A log phase culture of the B11 bacterial strain at the optical density (600 nm) was used to prepare the bacterial inoculum suspension. The selected bacteria were prepared in LB medium and maintained at 28°C in a shaker at 150 rpm for 24 h until late exponential growth phase. Afterward, the bacterial culture was centrifuged (8,000 × g, 10 min), and the supernatant was decanted. The obtained pellet was washed three times with sterile distilled water. Afterward, the obtained cell pellets were deferred in sterile distilled water (cell density of 10⁷–10⁸/ml), vortexed, and then used for plant treatment (50 ml/pot).

Seeds of pepper (C. annuum cv. Geumsugangsan) were provided by Takii Korea Ltd. (Seoul, South Korea). They were surface sterilized in 70% ethanol and 1.5% sodium hypochlorite and tested for the efficiency of the sterilization process (Albdaiwi et al., 2019) and viability (Ke et al., 2018). Germinated seeds were planted in plastic pot trays (28 cm × 54 cm) filled with horticultural soil (Shinsung Mineral Co., Ltd., South Korea), with one seed planted per pot. Seeds were then grown in a climate chamber under the following conditions: 250 μmol photons m^–2 s^–1 PAR (photosynthetically active radiation), 65% relative humidity, 25°C ± 2°C, 16:8 h light:dark cycle, and daily watering of seedlings. Three-week-old seedlings with uniform sizes were transferred to pots (10 cm × 10 cm) and subjected to different treatments. All pepper seedlings were randomly divided into two groups: (i) normal control, grown with distilled water (50 ml/pot), and (ii) salinity treatment, irrigated with NaCl (0.5%, 1%, 1.5%, and 2.5%) (50 ml/pot). Each treatment contained three replicates of 15 plants each. Plants treated with each concentration were measured for various plant growth characteristics. Plant growth parameters such as plant height, root length, stem diameter, leaf length/width, number of leaves, chlorophyll content, plant fresh/dry weight, and root fresh/dry weight were measured after day 8 (8DAT) of treatment. Ultimately, 1% NaCl was found to be the optimal treatment concentration for use in further studies.

The optimal PEG treatment concentration was obtained via an experimental method described by Kaya and Doganlar (2019). In addition, we chose to expose pepper plants to 0.1% Cd stress.

Experimental work was performed in a growth chamber at the Department of Applied Bioscience, Kyungpook National University. Plants were grown under the aforementioned conditions. In addition, sterilized pepper seeds (one seed/pot) were planted in plastic pot trays containing autoclave-sterilized horticultural soil (Shinsung Mineral Co., Ltd., South Korea). Seedlings of similar size (one seedling/pot) were selected after 3 weeks and then transferred to pots (10 cm × 10 cm). During the experiment, the temperature was maintained constantly at 25°C ± 2°C (day/night), with a relative humidity of 65% and a light intensity of approximately 200 μmol m^–2 s^–1. Soil moisture (70%), pH (∼7), and electrical conductivity (EC) (≤1.2) were recorded before the pot experiment; at the end of the experiment, soil samples from individual pots (300 g/pot) in each treatment were examined to again determine moisture, pH, and EC using a humidity tester (Model DM-5; Takemura Electric Works, Ltd., Tokyo, Japan) and conductivity meter (YSI Model 32, United States) (Supplementary Table 1).

A 2 days after transplanting, 3-week-old seedlings were subjected to various treatments. They were divided into two groups: the normal control group, irrigated with sterile distilled water (50 ml/pot), and the bacterized group, irrigated with bacterial inoculum suspension (50 ml/pot). Each group was treated for 7 days, after which the groups of unbacterized and bacterized seedlings were subdivided into two groups containing an equal number of seedlings; this produced eight experimental groups, which are described in Table 1. Each treatment was performed twice in triplicate. The soil moisture level of each pot was monitored daily using a humidity tester (Model DM-5, as above). The pepper seedlings were exposed to selected abiotic stresses for 8 days, and sampling was performed at the end of each treatment. The harvested samples were either immediately used or rapidly deactivated in liquid nitrogen and stored at −80°C.

Amino acid content was quantified using the method developed by Waqas et al. (2015). About 50 mg of freeze-dried leaves was hydrolyzed in the presence of hydrochloric acid (1 ml of 6-N HCl) for 24 h at 110°C. The extraction was then concentrated and vacuum dried at 80°C for 24 h. Afterward, the residue was diluted with deionized water (2 ml) and evaporated twice. The concentrated residue was then dissolved with hydrochloric acid (1 ml of 0.02-N HCl), and the mixture was passed through a 0.45-μm filter membrane. Finally, the solution was analyzed using a Hitachi L-8900 Amino Acid analyzer (Hitachi High-Technologies Corporation, Tokyo, Japan).

Soluble protein content in the various treatments was quantified based on the methods of Ashraf and Iram (2005). The freeze-dried leaf samples (0.1 g) were ground and then mixed with a 1-ml phosphate buffer (50 mM, pH 7.0). The mixture was centrifuged at 10,000 rpm for 10 min at 4°C. Subsequently, the supernatant was collected and treated with an appropriate reagent, and the absorbance of each sample was assessed at 595 nm. Protein content was estimated in all enzymatic preparations using the Bradford method (Bradford, 1976), with bovine serum albumin as the standard.

Sugar content was estimated according to the defined method of Khan et al. (2018). Briefly, freeze-dried leaves were crushed, extracted using 80% ethanol, and then vacuum dried. The dried residue was redissolved in 1 ml of deionized water and passed through 0.45-μm nylon-66 syringe filters. Furthermore, the filtered samples were injected into an HPLC system (Millipore Co., Waters Chromatography, Milford, MA, United States).

The activity of antioxidant enzymes such as peroxidase (POD) and polyphenol oxidase (PPO) was analyzed using the method reported by Putter (1974). Superoxide dismutase (SOD) activity was investigated via the method of Sirhindi et al. (2016). To determine flavonoids, 1,1-diphenyl-2-picrylhydrazyl (DPPH) radical scavenging activities, and total polyphenols, samples were processed and analyzed following previously described procedures (Zheng and Wang, 2001; Barka et al., 2006; Wang et al., 2009; Li et al., 2016). The mixture activity and absorbance were measured at selected wavelengths using a Multiskan GO UV/Vis microplate spectrophotometer (Thermo Fisher Scientific, Waltham, MA, United States).

Following various treatments, hydrogen peroxide (H₂O₂) levels were determined as previously described (Jana and Choudhuri, 1982; Tsai et al., 2005). Frozen samples were freeze-dried and then ground with a pestle and mortar. The ground sample (0.3 g) was homogenized with 3 ml of ice-cold phosphate buffer [50 mM, 1 mM ethylenediaminetetraacetic acid (EDTA), 1% polyvinylpyrrolidone (PVP), pH 7.0] and centrifuged at 13,000 rpm for 20 min. The supernatant (2 ml) was mixed with 1 ml of 20% (v/v) H₂SO₄ containing 0.1% titanium chloride, and the mixture was centrifuged at 13,000 rpm for 20 min. Supernatant intensity was measured at 410 nm with a T60 UV-Vis Spectrophotometer (PG Instruments Ltd., Wibtoft, United Kingdom).

The level of lipid peroxidation was estimated through the thiobarbituric acid reaction, as described by López-Serrano et al. (2019). Malondialdehyde (MDA) content was calculated using its extinction coefficient. The amount of lipid peroxidation was shown as the level of MDA created per gram of tissue weight.

The potential bacterial strain was grown in LB broth containing 1% NaCl, 10% PEG, and 0.1% Cd. The bacterial pellet was harvested and prepared for analysis as previously explained by Damodaran et al. (2013). For nutrient analysis of pepper plants, samples were first freeze-dried, and processed into a powder. Subsequently, the prepared samples were used to quantify Cd, sodium (Na), and nutrient (potassium, K; phosphorus, P; and calcium, Ca) uptake in pepper plants using inductively coupled plasma mass spectrometry (Optima 7900DV; Perkin-Elmer, United States). Treatments without bacterial inoculation were used to determine the initial concentrations of salt, heavy metals, and macronutrients. The removal efficiency of salt and Cd by the bacterial strain (B11) was calculated as follows: salt/Cd removal efficiency % = C0-CsC0×100, where C0 and Cs represent the initial and final concentrations of salt and Cd, respectively.

Total RNA was extracted from the pepper leaves (8DAT) and used for cDNA synthesis and quantitative PCR (qPCR) following the procedure of Imran et al. (2018). Specifically, 1 μg of RNA was used to synthesize cDNA with a BioFACT RT-Kit (BIOFACT, Korea) following the manufacturer’s standard protocol. The synthesized cDNA was employed as a pattern in a two-step qRT-PCR reaction, which was performed to determine transcript quantity with an Illumina Eco system (Illumina, United States) (Supplementary Table 2).

The experiments were conducted in triplicate consecutively three times. The results were subjected to statistical analysis by ANOVA using R (version 4.0.3). A least significant difference (LSD) test (p < 0.05) was used to determine significant differences among treatments. The mean and standard deviation were estimated with Microsoft Excel 2017. Graphs were prepared with GraphPad Prism (version 6.01; San Diego, CA, United States).

A total of 11 rhizobacteria were isolated from the rhizosphere and endosphere of S. borealis plants. The 16S rDNA and gyrB gene sequencing and phylogenetic analysis disclosed that the selected PGPR strain (B11) had a 100% similarity with the known sequences in GenBank and belonged to Bacillus amyloliquefaciens. In addition, phylogenetic analysis of the B11 strain using RAxML GUI v.1.3 software revealed its relevance to other strains of respective species (Figure 1). The 16S rDNA and gyrB gene sequences obtained in the present study have been submitted to the NCBI GenBank database and assigned the accession nos. MW599955 and MW602944, respectively.

The results of the in vitro growth-promoting assays in all strains are presented in Table 2 and Supplementary Figure 1. Most bacterial strains exhibited IAA and siderophore production, nitrogen fixation, and ACC deaminase activity (except B1). The strains B3–B8 did not show any citrate utilization activity. Our findings revealed variable solubilization activities: the strains B7, B10, and B11 were able to solubilize potassium, zinc, phosphate, and silicon.

The bacterial strains were screened in vitro for their potential to tolerate salt, drought, and heavy metal stress conditions (Table 3). Most bacterial strains were able to tolerate salinity levels up to 10%, but some bacterial strains (B1, B2, B7, and B10) could not tolerate such levels. The strains B1 and B2 exhibited growth on LB media supplemented with only 0.5% and 2.5% NaCl. Similarly, the strains B7 and B10 could tolerate up to 5% salinity. The growth of strains B3 and B8 was inhibited on LB media supplemented with various concentrations of NaCl. In terms of their response to PEG-amended media, all bacterial strains were able to survive in drought conditions of “-0.05, −0.15, −0.3” MPa, as well as with “−0.49” and “−0.73” MPa osmotic stress (Table 3). Most bacterial strains showed visible growth on LB media supplemented with the tested Cr and Cd concentrations (Table 3). The specific results revealed different growth responses of the bacterial strains exposed to various Ni concentrations. Exceptionally, the strains B3 and B8 did not exhibit heavy metal resistance. The bacterial strain B11 indicated tolerance in all amended media; therefore, this strain was selected for further study (Table 3 and Supplementary Figures 2, 3).

The effect of salinity on pepper seedlings was investigated at NaCl concentrations of 0.5, 1, 1.5, and 2.5%. Treatment with the lowest NaCl concentration (0.5%) produced no remarkable change compared with the control plants (Supplementary Figure 4 and Supplementary Table 3). However, pepper plants treated with NaCl at 1% showed decreased plant height (33.12%), root length (19.35%), stem diameter (50%), leaf length (45.83%), leaf width (40.90%), plant fresh weight (66.66%), plant dry weight (50%), root fresh weight (55.55%), root dry weight (66.66%), and relative leaf number (27%) (p < 0.05) relative to these measures in control plants (Supplementary Figure 4 and Supplementary Table 3). Furthermore, irrigation with the highest level of NaCl (2.5%) caused a drastic reduction in plant growth parameters. Thus, 1% NaCl was the optimal concentration chosen for use in further studies.

1,1-diphenyl-2-picrylhydrazyl content in pepper seedlings increased under stress conditions, whereas DPPH content was reduced in PGPR-treated seedlings exposed to all stresses. Specifically, PGPR inoculation contributed to 9.73%, 59.06%, and 37.99% reductions in DPPH content under salinity, drought, and heavy metal stresses relative to the DPPH detected in uninoculated stressed plants (Figure 4A). Furthermore, total polyphenol content was slightly enhanced under stress conditions, but it decreased under these conditions upon PGPR inoculation (Figure 4B).

Enhanced SOD activity was observed in pepper plants under stress conditions; however, SOD activity was lower in PGPR-treated plants subjected to the salinity (15.25%), drought (9.62%), and heavy metal (31.83%) stresses when compared with that of untreated stressed plants (Figure 4C).

Similar trends were observed with POD, PPO, and flavonoid activities, which increased distinctly under stress conditions. However, PGPR treatment reduced POD, PPO, and flavonoid content under stress conditions; for instance, their activities decreased (POD, 13.69%; PPO, 12.31%; flavonoid, 6.20%) in PGPR-inoculated seedlings subjected to salt stress (p < 0.05) (Figures 4D–F).

To determine the effect of the PGPR inoculant on the macronutrient state of pepper plants and its detoxifying role, five elements, i.e., Ca, K, P, Na, and Cd, were examined (Table 6). Under normal conditions, an increase was detected in the concentrations of K and P in plants inoculated with PGPR relative to the respective concentrations in control plants. Furthermore, Ca content was higher in inoculated plants than in control plants under normal conditions. Compared with the stressed plants and the PGPR-inoculated plants, plant macronutrients were regulated in inoculated stressed plants, which showed increases in K and P concentrations and a decrease in the concentration of Ca under stress conditions. Moreover, PGPR significantly reduced the accumulation of Na and Cd in plants under salinity and heavy metal stresses. Na and Cd uptake by bacterial cells is presented in Table 6. These results validate the detoxifying role of PGPR. The removal efficiency of Na and Cd by bacterial strain B11 was 67.08% and 45.24%, respectively.

Pepper seedlings were used to study the expression of abiotic stress-responsive genes. In total, eight genes (Supplementary Table 2) were analyzed for their change in expression in pepper plant seedlings under abiotic stresses and PGPR inoculation (B11).

This study characterized three binding immunoglobulin protein (BiP) genes (CaBiP1, CaBiP2, and CaBiP3) in pepper plants. These genes showed differential responses in pepper seedlings under abiotic stress and PGPR inoculation. Salt-, drought-, and heavy metal-stressed plants showed an increase in CaBiP1 expression levels relative to expression in unstressed control plants; however, among salt-, drought-, and heavy metal-stressed plants, the PGPR inoculation reduced expression by 76.30, 86.19, and 93.37%, respectively, compared with that in uninoculated plants (Figure 5A). Additionally, abiotic stresses enhanced CaBiP2 gene expression in uninoculated stressed plants in comparison with inoculated control plants. PGPR-inoculated salt-, drought-, and heavy metal-stressed plants showed lower CaBiP2 gene expression relative to the expression in uninoculated stressed plants (Figure 5B). CaBiP3 expression levels were increased significantly in heavy metal-stressed plants in comparison with control plants; however, under heavy metal stress conditions, CaBiP3 expression reduced remarkably (96.53%) in PGPR-inoculated stressed plants (Figure 5C). Furthermore, PGPR-inoculated salt-stressed plants showed a 60.22% decrease in CaBiP3 expression compared with expression levels in uninoculated salt-stressed plants.

The quantitative expressions of CaXTH1 and CaXTH2 genes in pepper seedlings under abiotic stresses and PGPR inoculation are shown in Figures 5D,E. Application of abiotic stresses reduced CaXTH gene expression in pepper plants, whereas the expression of these genes increased in PGPR-treated plants. For instance, PGPR exposure enhanced CaXTH2 gene expression by approximately 75.69%, 66.64%, and 93.26% under salinity, drought, and heavy metal stresses compared with the respective expressions in stressed plants alone.

The effects of salinity, drought, and heavy metal stresses as well as PGPR inoculation on BAX inhibitor 1 (BI-1; anti-PCD) was studied in pepper seedlings through the change in the expression of the BI-1 gene (CaBI-1) (Figure 5F). Expression levels of this gene in control and PGPR-inoculated plants displayed minor differences under unstressed conditions; however, higher CaBI-1 expression was noticed in stressed plants. PGPR-inoculated plants showed a decrease in CaBI-1 expression compared to uninoculated stressed plants (38.79% under salinity, 97.40% under drought, and 95.79% under heavy metal stress conditions).

The expression of the gene related to the transcription factor WRKY2 (CaWRKY2) was also examined. Salinity-, drought-, and heavy metal-stressed plants exhibited a 66.35%, 65.03%, and 75.03% increase, respectively, in CaWRKY2 expression levels compared with the expression observed in unstressed plants. However, PGPR-inoculated stressed plants showed a decrease in gene expression levels relative to the expression recorded in uninoculated stressed plants. CaWRKY2 levels in stressed pepper plants were reduced in PGPR-treated plants in comparison to uninoculated stressed plants (64.06% under salinity, 93.74% under drought, and 90.61% under heavy metal stresses; Figure 5G).

The expression level of CaPTI1 gene was evaluated in pepper seedlings subjected to abiotic stresses and PGPR inoculation. As shown in Figure 5H, salinity-, drought-, and heavy metal-stressed plants demonstrated a 94.82%, 93.84%, and 97.19% rise, respectively, in CaPTI1 expression levels in comparison with the expression detected in unstressed plants while reduced expression level was observed in PGPR-inoculated stressed plants. PGPR treatment lowered the CaPTI1 expression level by 83.44% under salinity, 94.75% under drought, and 82.07% under heavy metal stresses compared to the uninoculated stressed plants (Figure 5H).