Inoculation of Ensifer fredii strain LP2/20 immobilized in agar results in growth promotion and alteration of bacterial community structure of Chinese kale planted soil

In our former research, we succeeded in using agar, alginate, and perlite as immobilization materials to maintain long-term survival of the inoculant, Ensifer fredii LP2/20, in a controlled glasshouse. Therefore the information on the establishment and activity of the inoculant to promote plant growth under field conditions, the effects of the inoculant on the soil microbial communities and specific microbial taxa, and the association between the inoculant and soil elements merit further studies. Here, we found that agar was the most suitable material that supported the establishment of the inoculant under field conditions. RNA-based analysis showed that E. fredii LP2/20 immobilized in agar was still metabolically active at day 50 after being introduced into soil. Inoculation of E. fredii LP2/20 immobilized in agar conferred the highest plant dry weight (up to 89.94%) and all plant elements including total N (9.55%), P (17.94%), K (68.42%), Ca (39.77%), Mg (30.76%), Fe (29.85%), and Zn (22.44%). Inoculation of E. fredii LP2/20 immobilized in agar increased soil chemicals including soil organic matter (99.02%), total N (272.48%), P (31.75%), K (52.74%), Fe (51.06%), and Zn (63.10%). High-throughput next-generation sequencing of bacterial 16S rRNA amplicons showed that the Proteobacteria, Acidobacteria, Bacteroidetes, and Firmicutes were dominant phyla in Chinese kale field soil. Inoculation of E. fredii LP2/20 significantly affected the soil bacterial community structure by decreasing total bacterial richness and diversity. The numbers of alpha- and gamma-Proteobacteria were significantly increased while the number of delta-Proteobacteria was significantly decreased due to E. fredii LP2/20 establishment. Soil total P, K, and Ca and soil pH were the important factors that shaped the soil bacterial community composition.


Material and methods
Experimental design and soil sample collection. The experiment was conducted in the field at Pho Kao Ton subdistrict, Mueang district, Lopburi province of Thailand (latitude 14° 47′ 34.2″ N longitude 100° 36′ 13.7″ E) where Chinese kale has been regularly and consecutively cultivated twice a year from November to April for over 5 years. Two kg of each immobilization material (agar, alignate, and perlite) encapsulating E. fredii LP2/20 at a final concentration of 10 9 cells/kg were prepared as described by Nimnoi et al. 9 and then introduced into soil. The immobilization materials were sowed and cell cultures were sprayed directly over the soil by using a knapsack sprayer. The soil was mixed by shoveling over several times. Five treatments (three replicates of each treatment, 15 plots in total) were arranged in a completely randomized design. The treatments included SA (uninoculated control), SB (inoculation with E. fredii LP2/20 in liquid medium), SC (inoculation with E. fredii LP2/20 immobilized in agar), SD (inoculation with E. fredii LP2/20 in alginate), and SE (inoculation with E. fredii LP2/20 immobilized in perlite). Chinese kale seeds were planted in 1.50 m × 7.50 m plots with a 1-m interval between plots to avoid cross-contact among the different treatments. Watering was performed around 1.5 mm/day and a temperature range was between 27 and 38 °C. Fifty days after planting, Chinese kale planted soil from each individual plot were collected from a 0-to 20-cm depth with sterile soil augers according to a staggered-grid method 11 and mixed thoroughly. The mixed soil samples were divided into two portions; one was subjected to soil physicochemical analyses and the other was subjected to soil DNA and RNA extraction. For the former portion, 6 kg of soil from each treatment were collected using sterile trowels, placed in a sterile plastic bag, and immediately stored in an ice box. For the latter portion, 20 g of soil from each treatment were preserved with RNAlater Stabilization Solution (ThermoFisher Scientific, Waltham, MA, USA) and immediately stored in an ice box. After arrival at the laboratory, soil samples of both portions were immediately stored in a − 80 °C refrigerator.
Soil physicochemical analyses. Two kg of soil from each plot were air-dried, sieved through a 2-mm/10 mesh, and then subjected to soil physicochemical characteristic study. The samples were analyzed for particlesize distribution following the method described by Bouyoucos 12 . Organic matter (OM) content was analyzed as described by Walkley and Black 13 . Amounts of P, Fe, Ca, and Zn were analyzed by the Bray II method 14 . Amounts of K and Mg were determined by the flame photometric method 15 . Quantity of available N was estimated as described by Li et al. 11 . pH and electrical conductivity (EC e ) values were measured 15 . Plant growth promotion and plant nutrient analyses. Plant growth-promoting abilities were determined based on plant dry weight and plant nutrients including total N, P, K, Ca, Mg, Fe, and Zn. Plant materials from triplicate plots were dried in an oven for approximately 72 h at 65 °C until constant weight was obtained. Dried plant materials were ground in a mortar and digested with a mixture of H 2 SO 4 , Se, and Na 2 SO 4 modified from the Bergersen method 16 . The total N was determined by the Kjeldahl (wet oxidation) method 17 . Quantity of P was determined spectrophotometrically according to the vanadate-molybdate procedure 18  Change in soil physicochemical characteristics. The physicochemical characteristics of soil from each treatment at 50 days after planting were investigated. The results ( Table 2) show that all of the immobilization treatments significantly increased soil OM and four soil elements including total K, Ca, Fe, and Zn when compared with those of liquid inoculation and uninoculated control. The SC was the only treatment able to significantly increased all of soil elements when compared with those of liquid inoculation and uninoculated control. The SC resulted in the significantly highest amounts of soil OM, total N, and P. The SC and SE resulted in the significantly highest amounts of total K and Zn. The SC and SD resulted in the significantly highest amount of Sequencing analysis, bacterial diversity, and richness indices. In this study, a total of 1,198,141 raw reads were obtained from 15 DNA samples (three replicates/treatment). After tag merge and quality control, a total of 1,159,168 qualified tags (96.74% of raw reads) were obtained. Potential chimera tags were removed from qualified tags by the UCHIME algorithm and 647,864 taxon tags were remained. The tags with ≥ 97% similarity were grouped into the same OTUs. A total of 7317 OTUs were observed from all treatments, with a mean Good's coverage of 98.00 ± 0.00%. The Venn diagram (Fig. 1) illustrates the numbers of common, overlapping, and unique OTUs among treatments. The result exhibits the common 2687 OTUs in all treatments. The SB had the highest unique OTUs (299 OTUs), followed by the SA, SD, SC, and SE, respectively. In addition, Shannon-Weaver and Simpson that indicate diversity, Chao1 and ACE (Abundance-based Coverage Estimator) that represent richness as well as the number of observed OTUs in each treatment were analyzed ( Table 3).
The results show that the SA, SB, and SE shared the highest rank of the observed OTU numbers which were significantly different when compared with that of the SC. The observed OTU numbers of the SC and SD were not significantly different. As higher Shannon-Weaver and Simpson indices indicate greater bacterial diversity, the SC exhibited the lowest bacterial diversity which was significantly different from that of other treatments. The SA possessed the highest bacterial diversity which was not significantly different from that of the SB, SD, and SE. The bacterial richness indices (Chao1 and ACE) of the SA and SC were highest and lowest, respectively, which were significantly different from each other. Both bacterial richness indices of remaining treatments were The ordination of samples from each treatment by PCoA shown in Fig. 3 exhibited that the bacterial communities of the SC formed a separate cluster from those of other treatments. AMOVA indicated that the bacterial community structures of different treatments were significant from each other (P < 0.05). In addition, the variations in the inter-group and inner-group bacterial community compositions were evaluated by ANOSIM. The variations of the inter-group bacterial communities of the SC were considered significant (P < 0.05) from the other treatments and were larger than those of the inner-group (R = 1). On the contrary, the variations of the inner-group bacterial communities of the remaining treatments were not significantly different (P > 0.05) and were larger than those of the inter-group (R = −0). The UPGMA dendrogram of the relative abundance at the phylum level depicted in Fig. 4 was divided into two main clusters. The first main cluster contained only triplicate samples of the SC. The second main cluster contained samples from the remaining treatments, in which the SA and SE were closer to each other as well as the SB and SD were closer to each other. The dendrogram confirmed that inoculation of E. fredii LP2/20 immobilized in agar altered the soil bacterial community structure, whereas inoculations of E. fredii LP2/20 in liquid medium, alginate, and perlite did not affect the soil bacterial community structure. This finding may result from the persistence and cell survival of E. fredii LP2/20 immobilized in agar. Agar was proved to be the most suitable immobilization material that prolonged cell survival and supported the establishment of E. fredii LP2/20 after being introduced into soil rather than other immobilization materials and liquid medium. Thus, we further determined the establishment and activity of E. fredii LP2/20 in each treatment at day 50 after being introduced into soil by using RNA based techniques.

RT-PCR-DGGE analysis.
The establishment and activity of E. fredii LP2/20 at day 50 after inoculation into soil were determined by using RT-PCR-DGGE. In DGGE analysis, the DGGE patterns among three replicates of each treatment were compared first and found to be identical (see Supplementary Fig. S1-S3). The idea of DGGE is employed for the separation of the fragments that are identical in length, but different in sequence. Fragments that are identical both in length and sequence possess the same migration distance in DGGE gel. As shown in Fig. 5, the 16S rRNA gene of E. fredii LP2/20 was used as the reference band (LP) for verification of the species-specific migration position on DGGE gel. DGGE profiles from each treatment were primarily identified by comparing their relative migration positions on gels with that of the reference band. RNA-based DGGE profiles revealed the difference in metabolically active bacteria in soil. Only the SC possessed the 16S rRNA band that migrated to the same position as the reference band, indicating that only E. fredii LP2/20 immobilized in agar was metabolically active on the 50th day of inoculation. To confirm this finding, the band from the SC that migrated the same distance as the reference band in DGGE gel was excised and subjected to cloning and sequencing. Sequence analysis of the cloned DGGE band confirmed that it was derived from E. fredii LP2/20. The sequence of the cloned band has been deposited in the GenBank database (accession number: KU866444).     www.nature.com/scientificreports/ sity by significantly increasing the amounts of alpha-and gamma-Proteobacteria and significantly decreasing the amount of delta-Proteobacteria.

Discussion
PGPB promote plant growth in many ways such as facilitating resource acquisition, modulating plant hormone levels, solubilizing nutrients, and inhibiting various pathogens which hamper the growth and development of plants 2,34 . However, the production of biofertilizer by using free bacterial cells has many disadvantages associated with the viability and stability of PGPB during the production process and storage 35 . Free PGPB cells that are directly inoculated into soil may not effectively colonize plant roots because they are susceptible to variable environments 36    www.nature.com/scientificreports/ using E. fredii HH103 to increase nodulation and plant dry weight of G. max. Temprano-Vera et al. 40 used E. fredii HH103 and NGR234 to promote the growth of G. max and Glycine soja and found that the strain HH103 significantly increased shoot dry weight of G. max and G. soja up to five-and four-fold, respectively when compared with uninoculated control. The strain NGR234 also significantly increased shoot dry weight of G. max and G. soja up to 0.5-and 2.5-fold, respectively when compared with uninoculated control. However, there are limited reports about the use of PGPB to promote plant growth under filed conditions, the traceability and activity of the inoculant as well as the impact of the inoculant on the microbial community under natural conditions that was analyzed by high-throughput sequencing to provide new insights into soil microbial community and diversity. Thus, in this study, we evaluated the potential of using agar, alginate, and perlite as cell immobilization materials to prolong survival of the inoculant and promote plant growth under field conditions. The results show that E. fredii LP2/20 immobilized in agar was the best approach that mostly increased all plant growth parameters including plant dry weight (up to 89.94%) and all plant elements (total N, P, K, Ca, Mg, Fe, and Zn) which were significantly different from those of uninoculated control. When compared to uninoculated control, inoculation of E. fredii LP2/20 in alginate significantly increased plant dry weight, total K, Mg, and Fe and inoculation of E. fredii LP2/20 immobilized in perlite significantly increased only plant dry weight and total Fe, whereas inoculation of E. fredii LP2/20 in liquid medium failed to promote plant growth in all parameters. Perlite and peat have been used as carrier materials to prevent bacterial cells from soil competitive conditions 41,42 . However, there are shortages of natural peat and perlite in some countries or the peat-and perlite-mines are located in forbidden areas 43 . Due to these limitations, alternative immobilization materials such as alginate have been evaluated 44 . Alginate is the most common polymer for the encapsulation of bacteria for various industrial 45,46 and agricultural purposes 47,48 . A few studies on the effect of alginate-encapsulated bacteria on the cotton (Gossypium hirsutum L.) growth and bacterial community under normal and salinity stress conditions have been reported 49,50 . However, alginate-encapsulated inoculants are more expensive than peat-based inoculants and require more complicated technical handling 9 . Therefore the use of agar as an immobilization material has been proposed to offer advantages over the other immobilization materials such as cheaper cost, sufficient availability, and friendliness towards environments. Minaxi 51 and Jain et al. 52,53 used agar as an immobilization material for phosphate-solubilizing bacterial and fungal strains. The experiments showed that cells immobilized in agar had significantly higher phosphate solubilization than did free cells and cells immobilized in other materials. Kiran et al. 45 and Sankaralingam et al. 54 reported that agar was the most effective and suitable matrix to immobilize bacterial cells, with the advantages of a higher enzyme activity, greater resistance to environmental perturbations, and a lower production cost.
Our results agree with Schoebitz et al. 35 who reported that peat, perlite, and clay had high variability in chemical composition which affected the stability and survival of microorganisms and decreased the shelf life of the inoculants. Even though liquid inoculation simplifies inoculant production and application for the farmers, bacterial survival is decreased because there is no cell protective agent against environmental stresses. In this study we also found that all of immobilization treatments significantly increased all soil elements, except total Mg when compared with those of uninoculated control. The SC led to the highest increases of soil OM, total N, P, K, Fe, and Zn. These increases could result from the establishment, proliferation, and metabolic activity of E. fredii LP2/20. Agar was the best immobilization material that prolonged survival and supported its establishment in soil competitive environments. Soil is complex and dynamic in which its biological activity is mostly governed by dominant bacterial species. The dominant bacterial genera such as Bacillus, Pseudomonas, Rhizobium, and Azospirillum influence soil microenvironments and play roles in soil biogeochemical cycling and ecological processes which confer beneficial effects on soil and crop productivity 1,11 .
The 16S rRNA amplicon deep sequencing showed that the bacterial diversity across all treatments was dominated by the phyla Proteobacteria, Acidobacteria, Bacteroidetes, and Firmicutes. Our result corresponds to those in previous studies with agricultural soils 55,56 . The bacterial community of the SC formed a well-separated cluster from that of other treatments. Inoculation of E. fredii LP2/20 immobilized in agar also affected the bacterial communities and diversity by significantly increasing the amounts of alpha-and gamma-Proteobacteria and significantly decreasing the amount of delta-Proteobacteria. To understand the reason why only the SC significantly impacted the soil bacterial communities and diversity, while the other treatments didn't, RT-PCR-DGGE was employed to investigate the establishment and activity of E. fredii LP2/20 in soil. The result from RNA-based analysis revealed that agar was the only one among immobilization materials that was capable of preserving and maintaining E. fredii LP2/20 alive throughout the 50-day cultivation period. On the contrary, alginate and perlite were less suitable immobilization materials under field conditions. This finding still corresponds to our former report that exhibited the superior ability of agar in maintaining cell survival of the inoculant 9 . Agar is a natural polymeric hydrogel extracted from seaweed that protects cells from stresses, desiccation, and environmental factors 35 . This study also showed that immobilization in agar maintained long-term cell survival, leading to promotion of metabolic activity of the inoculant and effects on the soil bacterial community and diversity. Usually PGPB inoculants affect the soil bacterial communities and diversity after their establishment by both direct and indirect mechanisms such as digestion of soil composition to release ammonia, phosphorus, and iron, modulation of the effects of environmental stresses, modulation of phytohormones such as IAA that stimulates plant root development, and production of exudates which subsequently affect soil bacteria 38 . E. fredii LP2/20 has been reported to produce extracellular polysaccharides (EPS) 9 containing many carbon and nitrogen compounds such as ammonium sulfate, potassium nitrate, glucose, and mannose that affected other soil bacteria and subsequently affected soil nutrient recycling 57 . In addition, the relations between soil physicochemical properties and bacterial community were addressed. Soil total P, K, and Ca and soil pH were the dominant factors influencing the soil bacterial community structure. This result agrees with the reports of Li et al. 11 , O'Brien et al. 58 , and Zhang et al. 59 which showed that available nutrient, P, and pH were ones of important factors affecting the soil bacterial community. The pH value and soil N, P, and K contents were correlated with the soil microbial composition 60 .

conclusions
Our results show the new potential use of E. fredii LP2/20 immobilized in agar as a bacterial biofertilizer to promote plant growth and increase plant nutrition and soil fertility in the field. Inoculation of E. fredii LP2/20 immobilized in agar significantly increased the highest plant dry weight up to 1.89-fold over uninoculated control, followed by E. fredii LP2/20 immobilized in alginate (1.49-fold) and perlite (1.42-fold), respectively. In addition, inoculation of E. fredii LP2/20 immobilized in agar also mostly increased all plant elements including total N, P, K, Ca, Mg, Fe, and Zn and altered soil chemicals by increasing soil OM, total N, P, K, Fe, and Zn. High-throughput sequencing of 16S rRNA gene exhibited the effect of E. fredii LP2/20 immobilized in agar on the soil bacterial community structure. The proportions of the alpha-, beta-, gamma-, and delta-Proteobacteria, unidentified Cyanobacteria, Acidobacteria subgroup 6, Bacilli, and SCG were altered in response to the E. fredii LP2/20 establishment. Soil total P, K, and Ca and soil pH were the dominant factors influencing the soil bacterial community structure.

Data availability
All data generated or analyzed during this study has been included in this article. Sequence data has been deposited in the National Center for Biotechnology Information under genomic accession number KU866444 and BioProject accession number SRP154296. www.nature.com/scientificreports/ Publisher's note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creat iveco mmons .org/licen ses/by/4.0/.