Effects of Plant Growth Promoting Rhizobacteria Microbial Inoculants on the Growth, Rhizosphere Soil Properties, and Bacterial Community of Pinus sylvestris var. mongolica Annual Seedlings

[Objective] Determine the ability of three plant growth promoting rhizobacteria (PGPR) strains (Pseudomonas Mandelli A36, Serratia plymuthica A13 and Pseudomonas koreensis A20) to promote plant growth, evaluate the effect of inoculation with PGPR strains on seedling biomass, root structure, nutrient index, and enzyme activity, and assess the effect of PGPR inoculation on soil nutrient index, enzyme activity, and the soil microecological environment. [Method] The ability of the three PGPR strains to secrete indole-3-acetic acid (IAA), dissolve inorganic phosphorus, and produce siderophore and hydrolase was determined by the medium color change method, pot experiment to determine the effects of three PGPR strains on plant biomass, physicochemical properties, soil physicochemical properties and microbial diversity. [Result] The three PGPR strains had the ability to secrete IAA, solubilize inorganic phosphorus, and produce siderophore, the results of the pot experiment showed that inoculation with PGPR strain had a signicant effect on plant biomass, root index, nutrient index and enzyme activity, as well as soil nutrient index, enzyme activity and bacterial diversity. [Conclusion] This study provides basic data references for PGPR strains to improve the soil microecological environment and promote the growth and development of Pinus sylvestris var. Mongolica seedlings. P. koreensis JDM-2 ACC deaminase activity, that it signicant antibacterial effect on Bacillus subtilis the genome sequence of P. koreensis CRS05-R5 strain One study found that S. BU09 obvious control effect on potato scab discovered that S. plymuthica A21-4 could cucumber quality and regulate the micro-ecological environment of cucumber rhizosphere In this study, three treatment methods, including inoculation, compound inoculation, and non-inoculation, were used. The bacterial was perforated and injected into the rhizosphere soil. The results showed that the seedling height, ground diameter, fresh weight, and dry weight of the P. sylvestris var. mongolica seedlings increased signicantly after PGPR inoculation, which was consistent with previous results whereby the PGPRs Serratia proteamaculans 1-102 and Serratia liquefaciens texture, and prevents and controls plant diseases. In this study, after inoculation with PGPR, some growth-promoting bacterial genera in the rhizosphere soil increased signicantly, such as Arthrobacter, Ramlibacter, Gemmatimonas, Bacillus, and Serratia. Of these, Arthrobacter is a benecial functional bacteria that can improve the IAA content and salt tolerance of plants (Velázquez-Becerra et al. 2011). One of the important characteristics of the genus Bacillus is that it can produce spores with special resistance under adverse conditions, thus playing an active role in preventing plant diseases, improving plant resistance, and promoting plant growth (Oliveira et al. 2010; Probanza et al. 2002). Serratia is a benecial bacteria that can tolerate heavy metals and repair plants, playing an important role in protecting the environment (76). Actinobacteria is a Gram-positive bacteria that can degrade cellulose and chitin as the main resource for soil nutrient supply. A recent study found that after applying nitrogen fertilizer, the abundance of Actinobacteria in the Gurbantünggüt Desert soil increased signicantly (Huang et al. 2018), which is similar to the results of this study. After inoculation with PGPRs, the relative abundance of Actinobacteria in the rhizosphere soil was signicantly positively correlated with soil AN content (p < 0.05). Chloroexi is a Gram-negative bacteria that can potentially autotrophically metabolize through photosynthesis. This study found that after inoculation with PGPR, the relative abundance of Chloroexi in the rhizosphere soil was signicantly negatively correlated with soil AN content (p < 0.05), which is similar to the results of Ren et al. (Ren et al. 2020) who found that after inoculation with biochar + PGPR, the soil nitrogen content increased, while the abundance of Chloroexi decreased. The cluster analysis results showed that the bacterial communities treated by strains A20 and A36 differed signicantly from those treated with strains A13 and CK at the genus level, indicating that inoculation with strains A20 and A36 altered the soil bacterial community. The RDA results showed that soil AP, TP, AK, and OM indicators signicantly affected the bacterial community composition, suggesting that there are the main environmental factors that contribute to the differences in microbial community composition.


Introduction
Soil microorganisms promote material cycling and energy ow in the ecosystem. They play important roles as both producers and decomposers in the ecosystem (Wu et al. 2013; Zhang and Yu, 1990). In addition, soil microorganisms perform processes such as oxidation, nitrogen xation, nitri cation, and ammoniation in the soil to promote the decomposition of soil organic matter and nutrient conversion. Soil microorganisms are widely distributed in the plant rhizosphere and are most speciose in a dynamically changing environment. In 1904, the German scientist Lorenz Hiltner rst proposed the concept of the "rhizosphere" which refers to the soil around the root system. In the rhizosphere, plant root activity alters the physical and chemical properties of the soil, providing a special ecological environment for interaction between plants and soil microorganisms (Compant et al. 2010; Kloepper et al. 1980;Liu 2005). Since the concept of the "rhizosphere" was proposed, there have been increasing numbers of studies on the plant rhizosphere, mainly involving the physiological structure of the root system, rhizosphere soil nutrients, rhizosphere soil enzyme activities, and rhizosphere soil microorganisms, as well as the connection between them (Li 2002). There are a large number of active microorganisms, such as fungi, bacteria, and actinomycetes, in the rhizosphere soil, approximately 28-48 times higher than in non-rhizosphere soil. There are about 104-106 fungi genera in every 1 g of soil, and the main groups include lamentous fungi such as Penicillium, Fusarium, Aspergillus, and Trichoderma (Barreto et al. 2008;I. G et al. 2002;Sreevidya et al. 2016). There are about 104-108 actinomycetes genera in every 1 g of soil, primarily composed of Streptomyces, Micromonospora, and Nocardia (Cocking 2003). Bacteria are the most abundant rhizosphere soil microorganisms, with each 1 g of soil containing approximately 106-1010 bacteria genera. The main bacterial groups include Bacillus, Pseudomonas, Flavobacterium, Serratia, Rhizobium, and Azotobacter. Rhizosphere bacteria can promote plant growth, increase plant biomass, promote the absorption and utilization of soil nutrients by plants, improve the microecological environment of the rhizosphere soil, and inhibit or antagonize pathogenic bacteria (Vessey 2003). In addition, PGPR can degrade pollutants in the soil, improve soil fertility, control pests and diseases, and reduce the environmental pollution and soil compaction caused by the use of pesticides and fertilizers (Zhang et al. 2013). One study found that Brevibacillus brevis DZQ3 could signi cantly promote the growth of tobacco (Zhu et al. 2012), while another discovered that Azospirillum could promote the growth of corn (Zhu et al. 2012). A further study found that Pseudomonas CBT1, CBT6, and Cbt7 has a certain control effect on cucumber fusarium wilt (Yue and Zhang, 2009). Although PGPR has a signi cant impact on plant growth, the in uence of PGPR on perennial tree species and their associated rhizosphere communities are still poorly investigated, and the number of related PGPR reported is also very limited. At present, Bacillus and Pseudomonas are the two most studied and most important plant growth-promoting bacteria (Myresiotis et al. 2012). In addition to these two genera, Serratia has also been reported on, though not extensively so. Serratia sp. sy5 was found to increase the biomass of corn (Koo and Cho 2009). Serratia sp. CDP-13 could enhance the induced resistance and salt tolerance of wheat (Singh and Jha 2016), while Serratia Sp. A21-4 could promote the growth and development of capsicum and also demonstrated strong rhizosphere colonization ability (Yayou et al. 2016).
Pinus sylvestris var. mongolica is a geographical variety of Pinus sylvestris. It has a developed root system and can fully absorb and utilize water. Due to its fast growth, cold resistance, and drought resistance, as well as its aesthetic qualities, it has been widely introduced into the three northern areas of China where it is the main tree species used in shelter and sand control engineering (Zhang and Li, 2003). However, due to the large-scale use of pesticides and chemical fertilizers in recent years, P. sylvestris var. mongolica populations have declined in many areas, causing great economic loss. The introduction of rhizosphere microorganisms to replace pesticides and chemical fertilizers could address this decline as well as promote the growth and development of P. sylvestris var. mongolica, improve its root structure, and enhance its resistance to stress. The application of PGPR is pollution-free, residue-free, and more conducive to human and animal safety (Fiorentino et al. 2018;). After the introduction of rhizosphere microorganisms, the population and distribution of soil microorganisms will change, the basic physical and chemical properties of the soil will be altered, and the growth and development of plants will also be affected. As different rhizosphere microorganisms have different effects on the soil, it is essential that the impacts of rhizosphere microorganisms on the community structure, species composition, spatial distribution, and diversity of the soil microecological environment are explored. Therefore, the objectives of this study were to 1. Determine the ability of three PGPR strains (P. mandelii A36, P. koreensis A20, and S. plymuthica A13) to promote plant growth; 2. Evaluate the effect of inoculation with PGPR strains on seedling biomass, root structure, nutrient index, and enzyme activity; and 3. Assess the effect of PGPR inoculation on soil nutrient index, enzyme activity, and the soil micro-ecological environment.
Our overall aim was to explore the use of rhizobacteria in changing the community structure of soil microorganisms, promoting the growth and development of P. sylvestris var. mongolica, and improving the stress resistance of P. sylvestris var. mongolica.

Identi cation of PGPR strains
The three PGPR strains A13, A36, and A20 were isolated from the rhizosphere soil collected from a P. sylvestris var. mongolica forest at the Zhanggutai Experimental Forest Farm in Liaoning Province (42°43′-42°51′ N, 121°53′-122°22′ E), China. These three strains are highly e cient strains with multiple growth-promoting characteristics. Analysis of 16S rRNA sequences was used to identify the three strains. The three strains were inoculated in nutrient broth (NB) (peptone 1%; beef extract; 1%; sodium chloride 0.5%, Haibo Biotechnology, China) liquid medium and incubated at 37 °C with shaking (180 rpm) for 24 h, following which the genomic DNA was extracted from each strain using the bacterial genomic DNA isolation Kit DP302 (Beijing Tiangen Biochemical Technology Co., Ltd., China). The 16S rRNA gene primers 27F (5'agagttgatcctggctcag-3') and 1541R (5'-aaggaggtgatcccacgcca-3') were used for ampli cation (Galkiewicz and Kellogg, 2008). The polymerase chain reaction (PCR) was performed in 25-µL reactions containing 12.5 µL 2 × Taq PCR mix solution,1 µL forward primer, 1 µL reverse primer, 1 µL template DNA, and 9.5 µL double-distilled H 2 O. The reaction conditions were: 94 °C for 1 min, 30 cycles of 94 °C for 20 s, 56 °C for 20 s, and 72 °C for 2 min, followed by a nal extension at 72 °C for 5 min. The reaction products were quali ed by 0.8% agarose gel electrophoresis and then sent to Ruibo Biotech Co., Ltd (Harbin, China) for sequencing. The NCBI Blast server (http://www.ncbi.nlm.nih.gov) was used to compare the sequences with the registered sequences in the GenBank database. A phylogenetic tree was constructed using the maximum likelihood method in the software package MEGA (version 6.0), and the topology of the phylogenetic tree was evaluated using 1,000 bootstrap replicates.

Determination of the plant growth promotion characteristics of the PGPR strains
The molybdenum blue colorimetric method was used to determine the ability of the strains to dissolve inorganic phosphorus (Shekhar 1999). Speci cally, the strain suspension was spot-inoculated on sterile National Botanical Research Institute phosphate (NBRIP) solid medium and cultured at 28 °C for 3 d, during which the formation of a transparent ring was observed. The strains producing transparent circles were inoculated in sterile NBRIP liquid medium and cultured at 28 °C with shaking (180 rpm) for 7 d. A non-inoculation control was set, and then the soluble phosphorus content of the strain culture solution was measured.
The ability of the strains to produce indole-3-acetic acid (IAA) was tested using the methods of Bric (1991) and (Bric et al. 1991). The strain was inoculated in R 2 A liquid medium (0.5 g yeast, 0.5 g peptone, 0.5 g casein, 0.5 g glucose, 0.5 g soluble starch, 0.3 g K 2 HPO 3 , 0.05 g MgSO 4 , 0.3 g sodium pyruvate, and 1000 mL H 2 O) supplemented with 200 mg/L L-tryptophan and incubated at 28 °C with shaking (180 rpm) for 4 d, following which the strain culture was centrifuged at 12,850 × g for 10 min. One milliliter of the culture solution was used to determine the OD 600 value. The same volume of Salkawaski's reagent (50 mL 35% HClO 4 + 1 mL 0.5 M FeCl 3 ) was added, and the solution was placed in the dark for 30 min. A color change to red indicated that the strain had produced IAA. The IAA content of the sample was measured at Od 530 , and the yield was calculated as (mg · L − 1 · OD 600 − 1 ) = OD 600 / OD 530 . All tests were conducted in triplicate.
The ability of the strain to act as an iron-producing carrier was assessed using the improved Chrome Azurol S (CAS) test plate (Shin et al. 2001). First, Mannitol Salt Agar (MSA) medium (0.5 g KCl, 0.5 gMgSO 4 , 4 g glucose, 5 g casein peptone, 15 g agar, and 1 L distilled water), chromeazurol S dye solution (1 mM CAS, 0.1 mM FeCl 3 , and 4 mM cetyltrimethylammonium bromide), and phosphate solution (0.5905 g NaH 2 PO 4 · 2H 2 O, 2.427 g Na 2 HPO 4 · 12H 2 O, 0.25 g NH 4 Cl, 0.075 g KH 2 PO 4 , 0.125 g NaCl, and 100 mL distilled water) were prepared, and then the three solutions were sterilized at 115 °C for 25 min. Before pouring the medium, 50 mL of CAS dye solution and 50 mL phosphate solution were added into 1000 mL MSA medium and mixed well to produce the MSA-CAS detection medium (Chen 2006;Pérez-Miranda et al. 2007;Xiang 2006). The strain suspension was inoculated on the MSA-CAS detection plate and cultured at 28 °C for 7 d, and a non-inoculation control was set. If an orange-yellow circle formed on the blue CAS plate, it indicated that the strain had produced siderophore.
The ability of the strain to produce HCN was measured using the improved method of Kloepper et al. (Kloepper et al. 1991). The strain was inoculated on King's B medium supplemented with 4.4 g/L glycine (20 g peptone, 1.5 g K 2 HPO 4 , 1.5 g MgSO 4 · 7H 2 O, 10 g glycerol, 15 g agar, and 1000 mL distilled water). A lter paper strip soaked in picric acid solution (2.5 g picric acid, 12.5 g Na 2 CO 3 , and 1000 mL distilled water) was placed on top of the plate cover and cultured at 28 °C for 3 d. The change in the color of the lter paper strip from yellow to brown to red indicated the generation of HCN. The intensity of the color was visually recorded.
The ability of the strain to produce hydrolase was measured using the method of Cappuccino and Sherman (Cappuccino 2010). The strain suspension was inoculated on skim milk agar medium (100 g skim milk, 5 g peptone, 15 g agar, and 1000 mL distilled water) and starch agar medium (10 g soluble starch, 5 g peptone, 3 g beef extract, 15 g agar, and 1000 mL distilled water) and cultured at 28 °C for 3 d to observe any transparent areas around the spots. The medium was prepared by adding 1% cellulose, 0.5% chitin, and 1% pectin to the basic medium (1 g glucose, 0.5 g yeast extract, 0.5 g MgSO 4 , 1 g KCl, 1 g NaNO 3 , 1 g K 2 HPO 4 , 15 g agar, and 1000 mL distilled water), and the strain suspension was inoculated into cellulose medium, chitin medium, and pectin medium at 28 °C for 5 d. The cellulose medium was soaked with 0.01% Congo red solution for 15 min, following which the solution was poured out and decolorized with 1% NaCl solution for 5 min to observe whether any transparent areas appeared on the red background. Gram iodine liquid was then poured into the chitin medium and pectin medium, and the appearance of clear areas on the dark blue background was observed.

Pot experiment
The seeds of P. sylvestris var. mongolica (purchased from the Zhanggutai Experimental Forest Farm in Zhangwu County, Liaoning Province, China) were surface-sterilized with potassium permanganate (0.5%, v/v) for 30 min and then washed ve times with sterile distilled water. They were then germinated on sterile moistened gauze at 25 °C for 5 d. After germination, the seedlings were transferred to plastic pots ( Before plant inoculation, the three strains were separately inoculated into 250 mL Erlenmeyer asks containing 100 mL of NB liquid medium and maintained in 28 °C with shaking (180 rpm) for 48 h. The bacterial cultures were centrifuged at 8000 g for 10 min at 4 °C, and the collected bacteria were repeatedly centrifuged, washed with sterilized water three times, and then diluted with sterilized water to achieve an OD 600 of 0.6, which was the nal volume ratio used as the inoculum.
For all treatments, including the control, 10 pots (15 seedlings per pot) were prepared, giving a total of 150 seedlings per treatment. There were four treatments: (1) inoculation with sterile water (CK); (2) inoculation with A13; (3) inoculation with A20; and (4) inoculation with A36. The inoculations were performed by transferring 100 mL of the bacterial inoculum into the planting hole (Yang et al. 2019), where it was introduced at the root system level. The control plants were inoculated with 100 mL of sterile water. All treatments were arranged at random under the greenhouse conditions given above.

Sampling and biomass analysis of the seedlings
Samples were taken at three months after seedling inoculation. Fifty seedlings from each treatment group were randomly selected, of which 30 were used to measure biomass index, including seedling height (SH), ground diameter (GD), fresh weight of seeding (SFW), dry weight of seeding (SDW), fresh weight of root (RFW), and dry weight of root (RDW). During sampling, damage to the root system of the seedlings was minimized. The root system was washed to remove the soil, following which 10 randomly selected seedlings were used for scanning and grading the root system using an Epson v 700 root scanner. The indexes of root length, surface area, average diameter, number of root tips, bifurcation number, and root volume were analyzed.

Nutrient and physiological parameter analysis of the seedlings
Ten seedlings of P. sylvestris var. mongolica were randomly selected for drying. After drying, the roots, stems, and leaves were ground separately and stored in test tubes at room temperature for future use. Total nitrogen (TN) was determined using the Kjeldahl method, and available nitrogen (AN) was determined using the alkaline hydrolysis diffusion method (LY/T 1228-2015). Total phosphorus (TP) was determined by Mo-sb anti-colorimetry, and available phosphorus (AP) was determined using the sodium bicarbonate extraction method (LY/T 1232-2015). Total potassium (TK) was determined by ame photometry, and available potassium (AK) was determined using the NH4OAc leaching ame photometer method (LY/T 1234-2015). Organic matter (OM) was determined using the potassium dichromate oxidationexternal heating method (LY/T 1237-1999). Superoxide dismutase (SOD), catalase (CAT), peroxidase (POD), malondialdehyde (MDA), plantsoluble sugar (PSS) and proline (PRO) were determined using a kit from the Nanjing Jiancheng Bioengineering Company (Patterson et al. 1984).

Soil enzyme activities and physiochemical properties analysis
Upon collection of the plants, the topsoil was removed, and the rhizosphere soil was collected and sieved with a 20-mesh screen. Three replicates were tested in each treatment group. Ten grams of rhizosphere soil was obtained from each treatment group and stored at 4 °C for the determination of rhizosphere soil enzyme activity. Fifteen grams of soil from each treatment group was divided into three test tubes on average and stored at -80 °C for measuring soil microbial diversity. The rest of the rhizosphere soil was air-dried and stored at room temperature for measuring rhizosphere soil nutrients. The determination methods of soil TN, AN, TP, AP, TK, AK, and OM were the same as those described in 2.5. The soil acid phosphatase (APA), catalase (CA), sucrase (SA), and urease (UA) activities were measured using the Nanjing Jiancheng Biological Engineering Company kit.

Bacterial diversity analysis
Three replicate rhizosphere soil samples of the P. sylvestris var. mongolica seedlings from each treatment group were sieved with a 10-mesh screen and stored in a refrigerator at -80 °C for high-throughput sequencing. High-throughput sequencing was performed by Majorbio Bio-Pharm Technology Co., Ltd (Shanghai, China) on the Illumina Miseq Sequencer platform. FLASH (version 1.2.11) software was used to splice the paired-end sequences (https://ccb.jhu.edu/software /FLASH/index. shtml) (Magoc and Salzberg, 2011

Accession number(s)
The 16S rRNA gene sequences of the three strains were submitted and stored in the GenBank database with the accession numbers MT280201 (Pseudomonas koreensis A20), MT280202 (Serratia plymuthica A13), and MT280203 (Pseudomonas mandelii A36).

Identi cation of PGPR strains and plant growth-promoting characteristics
The 16S rRNA gene sequences of the three strains were analyzed, and phylogenetic trees were constructed (Fig. 1). The similarity between strain A13 and Serratia plymuthica DSM 4540 (AJ233433) was 99.00%, and thus the strain was preliminarily identi ed as Serratia plymuthica. The similarity between strain A36 and Pseudomonas mandelii (AF058286) was 99.72%, and thus the strain was preliminarily identi ed as Pseudomonas mandelii. The similarity between strain A20 and Pseudomonas koreensis Ps 9-14 (AF468452) was 98.80%, and thus the strain was preliminarily identi ed as Pseudomonas koreensis.
The three strains had various plant growth-promoting properties (Table 1). Strains A36, A13, and A20 were able to dissolve inorganic phosphorus, and the solubility of the inorganic phosphorus was 189.50 µg/mL, 211.00 µg/mL, and 429.50 µg/mL, respectively, and thus strain A13 exhibited the highest inorganic phosphorus solubility of 429.50 µg/mL. Strains A36, A13, and A20 all had the ability to secrete IAA. The IAA content of the three strains was 18.88 µg/mL, 4.59 µg/mL, and 22.66 µg/mL, respectively. All three strains could produce siderophore. Strains A36 and A20 produced light yellow circles on the CAS detection plate, while strain A13 produced orange-yellow circles. The lter paper strips of strains A13 and A20 changed from yellow to brown after 24 h of inoculation, and from brown to red after 60 h of inoculation. The lter paper strip of strain A36 remained unchanged, indicating that strains A13 and A20 had the ability to produce HCN, while A36 did not.
As indicated in Table 1, the three strains produced transparent areas on skim milk agar medium and starch agar medium, indicating that all three strains could produce protease and amylase. Strains A36 and A20 produced transparent areas on the cellulose medium, indicating that these two strains had the ability to produce cellulose. With the exception of strain A36, which was negative on the chitin medium, the remaining strains were positive and could produce chitinases. All three strains produced clear areas against the dark blue background of the pectin medium, indicating that all three strains had the ability to produce pectinase.

Effects of PGPR inoculation on seedlings and soil nutrients.
OM and N, P, and K are essential nutrients for plant growth. Inoculation with PGPR had a certain effect on the nutrient content of roots, stems and leaves of P. sylvestris var. mongolica seedlings (Fig. 3a). The content of OM and TN in the roots of inoculation with A36 increased the most, 52.79% and 21.36%, respectively (p < 0.05), the TP content of roots inoculation with A13 increased the most by 28.94% (p < 0.05); The content of OM and TN in the stems of inoculation with A36 increased the most, 21.16% and 89.87%, respectively (p < 0.05), the TK content of stems inoculation with A13 increased the most by 37.75%; The TN and TK content of leaves inoculation with A20 increased the most, 61.07% and 14.04%, respectively (p < 0.05), and inoculation with A20 increased the TP content of leaves by up to 26.85% (p < 0.05).
Inoculation with PGPR increased soil nutrient content (Fig. 3b). Inoculation with A13 increased the soil OM, TN and AN content the most, 39.58%, 49.18% and 47.62%, respectively (p < 0.05); Inoculation with A20 increased soil TP content by 97.44%; Inoculation with A36 increased the soil AP and TK content the most, 83.32% and 20.16%, respectively (p < 0.05); Inoculation of three PGPRs increased the soil AK content signi cantly. Among them, the inoculation of A36 had the best effect, increasing by 98.10% (p < 0.05).

Effects of PGPR inoculation on seedlings and soil enzyme activities
Inoculation with PGPR has a certain effect on plant enzyme activities (Fig. 4a, b, c Rhizosphere soil enzymes promote soil metabolism, change the form of soil nutrients, improve soil characteristics, and help increase the productivity of plant rhizosphere soil. PGPR inoculation had a certain effect on soil enzyme activity (Fig. 4g, h,  3.5 High-throughput sequencing analysis of microbial diversity in the rhizosphere soil 3.5.1 Effect of PGPR on the composition of the soil bacterial community The Venn diagram re ected the similarity and overlap of the samples. Following OTU clustering, separation, and elimination, a total of 3191 bacterial OTUs were obtained with a similarity of 97% (Fig. 5a). The unique bacterial OTUs of the A36, A13, A20, and CK groups were 174, 252, 192, and 200, respectively. There were more unique bacteria in the soil treated with A13, indicating that inoculation with A13 provided a more favorable environment for microbial diversity. CK and A36 shared 1,550 OTUs, CK and A13 shared 1,731 OTUs, and CK and A20 shared 1,550 OTUs, indicating that the bacterial communities of the CK and A13 groups were more similar than those of the other groups.
Analysis of the sequencing results indicated that soil samples from the four treatment groups comprised a total of 27 phyla, 68 classes, 186 orders, 343 families, 641 genera, and 1310 species of soil bacteria. The bar chart shows the species composition of different groups at the phylum level (Fig. 5b). The most abundant bacterial phyla were Acidobacteria, Patescibacteria, Bacteroidetes, Chloro exi, Gemmatimonadetes, Firmicutes, Actinobacteria, and Proteobacteria, which contributed almost 97.2% of the bacterial sequences. In the A36, A13, and A20 treatment groups, the relative abundance of Acidobacteria and Firmicutes was signi cantly lower than that of the CK group (p < 0.01), while the opposite was observed regarding Patescibacteria (p < 0.01), the relative abundance of which after PGPR treatment was signi cantly higher than that of the CK group. There was no signi cant difference in the relative abundance of Proteobacteria. The relative abundance of Bacteroidetes in the A20 group was not signi cantly different from that of the CK group, but was signi cantly higher than that of the A36 and A13 groups (p < 0.01).
The relative abundance of Gemmatimonadetes in the A36 group was not signi cantly different from that of the CK group, but it was signi cantly higher than that of A20 and A13 groups (p < 0.01). The relative abundance of Actinobacteria in the A36 and A20 groups was signi cantly higher than that of CK and A13 groups (p < 0.01).
The cluster analysis by heat map showed the community composition among the samples (Fig. 5c). The sample clustering results showed that the A13 group and the CK group were relatively similar, and the A36 group and the A20 group were relatively similar. The A36, A20, and CK groups were clearly separated, indicating that the soil bacterial community changed after the addition of PGPR. The relative abundance of Arthrobacter in the CK group was 3.48%, while the relative abundance of Arthrobacter was the highest in the A36 and A20 groups at 9.61% and 12.15%, respectively, showing a signi cant increase (p < 0.05). In the A13 group, the relative abundance of Serratia was the highest at 5.42%, while in the CK group, the relative abundance of Serratia was 0, indicating that the relative abundance of Serratia increased signi cantly in the A13 group (p < 0.05).

Relationship between soil bacterial community composition and soil environmental factors
The relative abundance of bacterial dominant genera in 12 soil samples (Table 2), a total of 28 bacterial genera accounted for more than 1% of the total community of all treatment groups, and the bacterial dominant genera of each treatment group showed differences (p < 0.01). Compared with CK, inoculation with PGPR signi cantly reduced the relative abundance of Burkholderia-Caballeronia-Paraburkhol, Nocardioides and Blastococcus (p < 0.01), and signi cantly increased the relative abundance of Mesorhizobium, Tumebacillus, Conexibacter and Caenimonas (p < 0.01). There were also differences in the dominant genera of bacteria.

Diversity of bacterial community
The coverage of the four soil libraries ranged from 0.987 to 0.990 (Fig. 6a), which was close to 1, indicating that the sequencing depth covered all the species of the soil sample, suggesting that the sequencing results represent the true situation of the soil bacteria. The order of the Chao index was A13 > CK > A20 > A36 (Fig. 6b), and except for the A13 group, there was no signi cant difference between the other groups. This indicated that the total number of soil bacterial species in the A13 group was higher than the other groups, and the community richness was also higher than the other groups. The Shannon index performance was: A13 > CK > A36 > A20 (Fig. 6c). The larger the Shannon value, the higher the community diversity. The Simpson index was as follows: A20 > A36 > CK > A13 (Fig. 6d). The larger the Simpson value, the lower the community diversity. The A13 group index was signi cantly higher than other groups, indicating that the A13 group community diversity was higher.
Pearson analysis was used to analyze the correlation between the α diversity index and environmental factors (Fig. 6e). Chao and Shannon indexes were signi cantly positively correlated with TN and AN content (p < 0.05), and were extremely signi cantly positively correlated with OM content (p < 0.01). The Coverage index was signi cantly positively correlated with TP and AK content (p < 0.01). The Simpson index was positively correlated with TP content, but was negatively correlated with OM content. These ndings indicate that PGPR has a crucial effect on soil nutrient cycling.

Redundancy analysis of soil bacteria (RDA)
The relationship between soil physicochemical properties and the relative abundances of dominant bacteria was assessed using RDA at the genus level (Fig. 7). The A20 sample was positively correlated with urease activity, but negatively correlated with TN and OM. A36 was positively correlated with TK and AP, and negatively correlated with AN, acid phosphatase activity, and CAT activity. A13 was positively correlated with OM and TN, and negatively correlated with TP. The RDA showed that the relative abundance of Arthrobacter, Sphingomonas was positively correlated with AP; the relative abundance of Arthrobacter and Sphingomonas was positively correlated with TP; the relative abundance of Arthrobacter was positively correlated with AK; and the relative abundance of Serratia and Tumebacillus was positively correlated with OM. The Monte-Carlo permutation test indicated that the AP, TP, AK, and OM indicators of the soil were signi cantly related to the bacterial community composition at the genus level (AP: pseudo-F = 14.35, p = 0.004; TP: pseudo-F = 7.96, p = 0.02; AK: pseudo-F = 6.13, p = 0.03; OM: pseudo-F = 10.30, p = 0.009), indicating that soil AP, TP, AK, and OM were the main environmental factors in uencing the composition of the microbial communities (at the genus level).

Discussion
Current agricultural production practices need to prioritize environmental sustainability, and thus the use of soil microorganisms has been suggested as a promising alternative to harmful pesticides and fertilizers, as well as for increasing crop yield (Wertz et al. 2007). PGPRs not only promote plant growth, control plant disease, and increase crop yield, but also exhibit strong rhizosphere colonization ability and speci c microecological functions, and are thus potentially important biocontrol microorganisms (Schippers et al. 2003). Currently, the most representative PGPR genera that have been discovered are Pseudomonas, Enterobacter, Clostridium, Arthrobacter, Achromobacter, Micrococcus, Flavobacterium, Azospirillum, Azotobacter, and Bacillus (Cristiana et al. 2008;Swain and Ray, 2009). The three types of PGPR used in this study were P. mandelii, S. plymuthica, and P. koreensis. Pseudomonas rhizosphere bacteria have been extensively studied, though the two species used in the present experiment have been studied comparatively less. For instance, Ródenas et al. found that P. mandelii strain 29 could signi cantly increase mycorrhizal colonization (Navarro-Ródenas et al. 2016), while another study found that P. mandelii was associated with rice plant nitrogen xation. Pseudomonas koreensis has not been reported on much, though P. koreensis Ps 9-14T has been isolated from the soil (Kwon et al. 2003), and another study discovered that P. koreensis could enhance the drought resistance of Helianthus annuus (Macleod et al. 2015). A P. koreensis JDM-2 strain was isolated from Eucommia ulmoides roots and exhibited ACC deaminase activity, and found that it had signi cant antibacterial effect on Bacillus subtilis (Gong 2011). Additionally, the genome sequence of P. koreensis CRS05-R5 strain has been analyzed (Lin et al. 2016). One study found that S. plymuthica BU09 had an obvious control effect on potato scab (Zhang and Liu, 2017), while another discovered that S. plymuthica A21-4 could improve cucumber quality and regulate the micro-ecological environment of cucumber rhizosphere soil (Ding et al. 2018). In this study, three treatment methods, including inoculation, compound inoculation, and non-inoculation, were used. The bacterial suspension was perforated and injected into the rhizosphere soil. The results showed that the seedling height, ground diameter, fresh weight, and dry weight of the P. sylvestris var. mongolica seedlings increased signi cantly after PGPR inoculation, which was consistent with previous results whereby the PGPRs Serratia proteamaculans 1-102 and Serratia liquefaciens 2-68 could promote the emergence and growth of corn (Pan et al. 1999). Thiery et al. found that the PGPR Bacillus amyloliquefaciens FZB 42 could promote cotton growth (Alavo et al. 2015), indicating that PGPRs can effectively promote plant growth.
In this study, the root length, surface area, number of root tips, and root volume of the seedlings increased signi cantly after PGPR inoculation, which may be related to the secretion of IAA by the three PGPR strains. A previous study showed that the effect of PGPR on root development is related to the secretion of IAA, as IAA can promote root growth and development, increase root surface area, and promote root metabolism (Glick 2014). Additionally, the inoculation of IAA-secreting PGPR on canola signi cantly increased the number of stems and branches (Asghar et al. 2004). The PGPR strains CA1001 and CA2004, which can secrete IAA, were inoculated into crops to enhance root and stem biomass (Chandra et al. 2018). Therefore, the increase in the root structure index of P. sylvestris var. mongolica seedlings may be mainly due to the IAA secreted by the PGPR. There is further evidence to illustrate this point; that is, among the three PGPR strains, A20 secreted the highest amount of IAA (22.66 µg/mL), and inoculation with strain A20 increased the root length, root surface area, root average diameter, and the number of root tips of the seedlings most signi cantly.
Plant enzymes such as SOD, POD, and CAT can effectively remove the reactive oxygen generated during metabolism, thus balancing the production and removal of active oxygen in the plant and preventing reactive oxygen from causing membrane peroxidation and other damage (Gururani and  . This was con rmed by another study (Jha and Subramanian, 2013) that showed that after PGPR inoculation, the antioxidant enzyme activity of rice increased, thereby increasing plant resistance to salt stress.
PGPRs not only promote plant growth but also improve the rhizosphere environment of the plants. Soil enzyme activity can re ect the transformation ability of soil nutrients and the health of soil to a certain extent ). Inoculation with PGPRs was previously found to signi cantly increase soil enzyme activity, total nitrogen, available nitrogen, and soil organic matter content . In this study, the same trend was observed. After inoculation with PGPR, the nutrient content of TP, AP, AK, and other nutrients in the rhizosphere soil of P. sylvestris var. mongolica increased signi cantly, and enzyme activities such as invertase and acid phosphatase were also increased. Some studies have demonstrated a close relationship between soil enzyme activity and soil available nutrients (Lian et al. 2011). For example, Li et al.
found that soil CAT had a signi cant negative correlation with available phosphorus, and that urease had a signi cant positive correlation with available potassium (Ning et al. 2014). Another study found that soil CAT and phosphatase were signi cantly positively correlated with available nitrogen (Chen et al. 2014). In this study, after inoculation with the PGPRs, the enzyme activity of the rhizosphere soil increased, following which the content of available nutrients increased. The content of available nitrogen, phosphorus, and potassium increased the most, namely by 109.52%, 143.50%, and 98.10%, respectively. To some extent, the content of nitrogen, phosphorus, potassium, and other nutrients in the soil represents the potential fertility of the soil (Wang et al. 2015). Inoculation with PGPRs effectively improved the nutrient content of the rhizosphere soil of P. sylvestris var. mongolica and improved the ability of the plant to tolerate its external environment. In this study, it was found that after PGPR inoculation, in addition to the signi cant increase in biomass, the nutrient content and enzyme activity of the seedlings also increased. Many reports have suggested that PGPRs improve the biomass and the nutrient and enzyme activity of plants, thereby promoting plant growth and metabolism, such as in wheat and spinach (Akmak et al. 2007), pea (Akhtar 2014), and organically-grown raspberry (Orhan et al. 2006 (76). Actinobacteria is a Gram-positive bacteria that can degrade cellulose and chitin as the main resource for soil nutrient supply. A recent study found that after applying nitrogen fertilizer, the abundance of Actinobacteria in the Gurbantünggüt Desert soil increased signi cantly (Huang et al. 2018), which is similar to the results of this study. After inoculation with PGPRs, the relative abundance of Actinobacteria in the rhizosphere soil was signi cantly positively correlated with soil AN content (p < 0.05). Chloro exi is a Gram-negative bacteria that can potentially autotrophically metabolize through photosynthesis. This study found that after inoculation with PGPR, the relative abundance of Chloro exi in the rhizosphere soil was signi cantly negatively correlated with soil AN content (p < 0.05), which is similar to the results of Ren et al. (Ren et al. 2020) who found that after inoculation with biochar + PGPR, the soil nitrogen content increased, while the abundance of Chloro exi decreased. The cluster analysis results showed that the bacterial communities treated by strains A20 and A36 differed signi cantly from those treated with strains A13 and CK at the genus level, indicating that inoculation with strains A20 and A36 altered the soil bacterial community. The RDA results showed that soil AP, TP, AK, and OM indicators signi cantly affected the bacterial community composition, suggesting that there are the main environmental factors that contribute to the differences in microbial community composition.