Diversity of Fast-Growth Spore-Forming Microbes and Their Activity as Plant Partners

Biofertilizers are agricultural materials capable of reducing the usage amounts of chemical fertilizers. Spore-forming microorganisms (SFM) could be used for plant growth promotion or to improve plant health. Until now, biofertilizers based on SFM have been applied for rice and other crops. In this study, we isolated and characterized SFM, which were colonized on the Oryza sativa L. roots. SFM were analyzed regarding the short-term effects of biofertilization on the nursery growths. Analysis was performed without nitrogen or any inorganic fertilizer and was divided into two groups, including bacteria and fungi. SF-bacteria were dominated by the Firmicutes group, including species from Viridibacillus, Lysinibacillus, Solibacillus, Paenibacillus, Priestia, and mainly Bacillus (50%). The fungi group was classified as Mucoromycota, Basidiomycota, and mainly Ascomycota (80%), with a predominance of Penicillium and Trichoderma species. In plant performance in comparison with B. pumilus TUAT1, some bacteria and fungus isolates significantly improved the early growth of rice, based on 48 h inoculum with 107 CFU mL−1. Furthermore, several SFM showed positive physiological responses under abiotic stress or with limited nutrients such as phosphorous (P). Moreover, the metabolic fingerprint was obtained. The biofertilizer based on SFM could significantly reduce the application of the inorganic fertilizer and improve the lodging resistances of rice, interactively enhancing better plant health and crop production.


Introduction
The nutrient supplies of plants are significantly influenced by microorganisms such as bacteria and fungi [1][2][3][4]. Those microorganisms that have a positive effect and stimulate plant growth are defined as plant growth-promoting (PGP) microorganisms [5][6][7], including those that form spores. They are distributed in the rhizosphere, on root surfaces (rhizoplane), within root tissues (root cortex), leaves, stems, and as free-living microorganisms in the soil. Currently, some mechanisms related to plant growth promotion have been clarified, such as biological nitrogen fixation, phytohormone production, nutrient solubilization, and siderophore production [5,8]. However, many factors and pathways are still unknown, especially under stress conditions. Many environmental factors affect spore-forming microorganisms (SFM), such as spore formation and plant-interaction effectiveness. Some responses and interactions vary sterilized seeds were incubated for pre-germination for 48 h at 28 °C and incubated in darkness under sterile conditions soaked into sterile distilled water (Millipore-Sigma CA, Germany). Then, the seeds were inoculated with 10 mL of rhizosphere/soil suspension solutions.
Previous to inoculation, the suspension solutions were prepared as 10 g of rhizosphere/soils in 50 mL of sterile MQ (Milli-Q water Direct-Q® 3 UV system, Millipore-Sigma CA, Darmstadt, Germany) per each suspension, and shake during 1 h at 150 rpm ( Figure 1). Then transferred into 100 mL Tray-pot (58 × 45 × 54 mm; Ando-Chemical©, Nara, Japan) with 100 g of sterilized low-nutrient soil Inaho-baido (Inaho-Kako Co., Ltd. Toyama, Japan). The Inaho-baido soil was autoclaved at 121 °C/0.2 MPa for 30 min. Each sample site included 10 pre-germinated seeds. The plants were grown in a controlled environment chamber room at the Tokyo University of Agriculture and Technology (Tokyo, Japan). Plants were grown for 15 days in a photoperiod of 16 h light (5000~7000 LUX) and 8 h dark at 28 °C. A sterilized distilled water was added every day following 100% (v/v) of the maximum water-holding capacity of soil and was maintained at this level throughout the growth period ( Figure 1). After 15 days, the plants were harvested and washed with sterile distilled water. The whole roots were separated and crushed in 5 mL sterile MQ water until obtaining homogenate root solutions and raised up to 10 mL. Subsequently, an aliquot (50 µL) of the root homogenate solution was streaked onto Potato dextrose (PD) Agar (Difco Bacto; Becton, Dickinson and Company, NJ, USA) and incubated for 3 days at 28 °C. The remaining solution was incubated for 1h at 65 °C, then 50 µL of root homogenate solution was transferred in Trypticase soy (TS) Agar (Difco Bacto; Becton, Dickinson and Company, NJ, USA) and incubated 24 h at 28 °C ( Figure 1). The single colonies of strains were re-streaked onto fresh plates of their respective media to obtain pure colonies. Afterward, each strain was incubated in TSB (TS-broth), and PDB (PD-broth) (Difco Bacto; Becton, Dickinson and Company, NJ, USA), and an aliquot (1 mL) of strain-broth was kept with 1 mL of 50% glycerol stock at −80 °C for further characterization [15]. All isolates After 15 days, the plants were harvested and washed with sterile distilled water. The whole roots were separated and crushed in 5 mL sterile MQ water until obtaining homogenate root solutions and raised up to 10 mL. Subsequently, an aliquot (50 µL) of the root homogenate solution was streaked onto Potato dextrose (PD) Agar (Difco Bacto; Becton, Dickinson and Company, NJ, USA) and incubated for 3 days at 28 • C. The remaining solution was incubated for 1h at 65 • C, then 50 µL of root homogenate solution was transferred in Trypticase soy (TS) Agar (Difco Bacto; Becton, Dickinson and Company, Franklin Lakes, Franklin Lakes, NJ, USA) and incubated 24 h at 28 • C ( Figure 1). The single colonies of strains were re-streaked onto fresh plates of their respective media to obtain pure colonies. Afterward, each strain was incubated in TSB (TS-broth), and PDB (PD-broth) (Difco Bacto; Becton, Dickinson and Company, Franklin Lakes, NJ, USA), and an aliquot (1 mL) of strain-broth was kept with 1 mL of 50% glycerol stock at −80 • C for further characterization [15]. All isolates were re-inoculated onto the host for further selection and characterization.

Plant Assay
Three hundred twenty bacterial strains and sixty-six fungal strains were grown in PD and TS broth (respectively) between 1-3 days at 120 rpm, 28 • C to obtain 10 8 cells mL −1 . The cell inoculants were obtained by collecting the cell at 10,000 rpm for 5 min at 4 • C. Before inoculation, the Oryza sativa L. seeds were surface-sterilized, as previously mentioned.
The rice seeds were sown into 200 g of sterilized low-nutrient soil Inaho-baido (Inaho-Kako Co., Ltd., Toyama, Japan) contained in a plant box (7.6 × 7.6 × 10.2 cm). Then, five milliliters of each bacterial or fungal suspension were used to inoculate the seeds, each suspension containing 10 7 CFU mL −1 . The plant box contained 5 seeds, and each treatment had 3 replications.
The plants were cultivated for 3 weeks under a controlled growth chamber room at 16 h light (5000~7000 LUX) and 8 h dark photoperiod at 28 • C. A sterilized RO water was added every day following 100% (v/v) of the maximum water-holding capacity of the soil. Additionally, two control were used in this step: one was non-inoculated treatment as a negative control (named control). The other was positive control, an inoculated treatment based on SF-bacteria (Biofertilizer "Yume-bio"), and it was named TUAT1(+), which contains Bacillus pumilus TUAT1 (AP014928.1) as rice plant growth promoting rhizobacteria (PGPR) [16]. After 3 weeks, shoot lengths and fresh mass of roots and shoots of seedlings were measured and weighed.

Isolation of Genomic DNA
A total of 33 strains between bacteria and fungus were selected based on their plant assay performance compared to the controls. Isolates were grown in TS (for bacteria) and PD (for fungus) broth at 28 • C for 2 days and the collected cells were washed twice with PBS buffer at a ratio of 1:1 (v/v). Afterward, the cells were kept together with 1 bead per sample for 1 h at −80 • C and then crushed using Tissue-lysser II (QIAGEN®, Austin, TX, USA) for 2 min. The genomic DNA was extracted using a Wizard Genomic DNA purification kit (Promega, Madison, WI, USA). The DNA concentrations were confirmed using a NanoDrop 2000 UV-vis spectrophotometer (Thermo Fisher Scientific, Wilmington, DE, USA).

DNA Amplification and Sequencing
The PCR amplification and sequencing analyses were by using different housekeeping genes. For bacteria, the primer sets used for the 16S rRNA gene were carried out as described by Habibi et al. [17]. The RNA polymerase gene subunit β (rpoB) primers were described by Ki et al. [18]. For fungus, the primer sets used for the 18S rRNA gene were described by Wu et al. [19], as ITS-1 was carried out as described by Wang et al. [20]. Amplifications were performed using a thermal cycler GeneAmp PCR system 9700 (Applied Biosystems, Waltham, CA, USA). The PCR conditions were denaturation at 95 • C for 4 min, 30 cycles of denaturation at 94 • C for 1 min, annealing at 60 • C (ITS, rpoB) or 55 • C (16S and 18S rRNA) for 45 s, and extension at 72 • C for 2 min, followed by a final extension at 72 • C for 5 min [21]. The PCR products were checked using a 1.5% (w/v) agarose gel with 0.5 × TBE buffer mixed with 0.5 µg mL −1 ethidium bromide [21]. A 1 kb DNA ladder was used as a marker. Subsequently, the bands were excised for all genes, and DNA was purified using a FastGene ® agarose gel/PCR extraction kit (Nippon Genetics, Tokyo, Japan). According to the manufacturer's protocols, PCR products were sequenced using an ABI Prism 3500 Genetic Analyzer (Applied Biosystems, Waltham, CA, USA). The obtained sequences were aligned using the ClustalW method (Genetix version 11), then compared in the GenBank database by using the online software BLAST algorithm-based sequence alignment (https://www.ncbi.nlm.nih.gov/genbank/, accessed on 15 January 2022). Then, the phylogenetic trees were constructed with the Maximum Composite Likelihood model (no tree-topology was used) by using the software MEGA version 12.0.

Physiological Activity Profile of Fukushima SFM
The selected isolates were first grown in PD and TS broth for 24 h for bacteria and 3 days for fungus at 28 • C, and 5 µL of cell suspensions at 10 7 cells mL −1 were transferred onto agar and broth media and incubated for one week at 28 • C. Growth of isolates was estimated in comparison to the control treatment (non-stress) as follows: -, no growth; +, weak growth (10-20% compared to control); ++, good growth (30-60% compared to control); and +++, excellent growth (similar to/same as control) [21]. After one week, the colony-forming units (CFUs) were calculated by plate counting under stress conditions. These experiments were performed thrice for each isolate.
The isolates were evaluated for growth under different abiotic-stress conditions: alkaline (10 pH, NaOH 1M), acidic pH (4.5, HCl 5M), and high salinity (up to 4% NaCl). The ability of isolates to grow at different stress conditions was compared with the control, which was pH 6.8 and 0.1% NaCl. For this step, Yeast Mannitol Agar and TS Agar were used (Wako Pure Chemical Industries, Co., Ltd., Osaka, Japan).

Phosphorous (P) and Potassium (K) Solubilization Assess
Bacterial isolates were grown in the TS broth medium at 28 • C for 24 h, then 5 µL (10 7 cells mL −1 ), and the fungus isolates were grown in the PD broth medium at 28 • C for 48 h. Each culture was spotted onto Pikovskaya's media [22] to test P solubilization and Aleksandrow's agar [23] to test K solubilization, and the following steps were as described by Artigas et al. [21], for fungi were incubated for 5 days, and bacterial strains were incubated 120 h, both at 28 • C in the darkness. The recording was every 24 h. The formation of a clear halo zone around the bacterial colony indicated the solubilization activity on Pikovskaya's or Aleksandrow's agar, and their solubilization index (SI) was calculated as [Halozone diameter + colony diameter (mm)/colony diameter (mm)] [22,23]. Experiments were performed with triplicates from the incubation of each strain.

Iron (Fe) Sequestration Performance and Growth in N-Free Media
The iron sequestration and the siderophore production were performed by following the CAS assay by using CAS agar plates described by Louden et al. [24]. The isolates were grown in their respective broth medium (PDB: fungus and TSB: bacteria) at 28 • C, then 5 µL (10 7 cells mL −1 ) of each culture was spotted onto CAS agar plates. Growth of species and CAS reaction were evaluated every 24 h up to 120 h, and compared with the control (non-iron media), all strains were non-stressed. The isolate's growth under nitrogen-free media (N-free media) was recorded, and Ashby agar plates were used as described by Allen [25].

Metabolic Fingerprint of SFM from Fukushima
SFM isolates were grown in TS/PD broth medium at 28 • C for 48 h (in the vegetative stage). Subsequently, the strains were centrifuged at 4× g rpm at 10 • C; the pellet was kept and re-suspended in 1 mL sterile MQ water. An aliquot of 100 µL (10 7 cells mL −1 ) of each strain suspension was spotted onto each well of the Biolog ™ GN2 Microplates, according to the manufacturer's protocols (BIOLOG ™ Inc., Hayward, CA, USA). The microplates were read by using SpectracMax Paradigm Microplate reader (Molecular Devices LLC., San Jose, CA, USA), and the data Acquisition and analysis were obtained by SoftMax Pro 6.3 (Molecular Devices© LLC., San Jose, CA, USA).
Statistical analyses included Dunnett's, Tukey's, and other tests, which were undertaken using the EZR 1.52 software (Saitama Medical Center, Jichi Medical University, Saitama, Japan), which is a graphical user interface for R (The R Foundation for Statistical Computing, Vienna, Austria) [26]. A value of p ≤ 0.05 was considered an indication of statistical significance.

Results
The characteristics of the site sampling and type of source used for spore-isolation strains are shown in Table S1. Twenty samples were taken with different vegetation histories, from forestal areas to rice fields (indicated as paddy fields). The samples included soils taken 20 cm or deeper and rhizosphere, including roots and organic matter distinguished as litter in this study.
A total of 386 isolates were collected from the roots of both rice cultivars. From soils, 117 isolates were obtained, and 269 isolates (71% of the total) were collected from the rhizosphere (Table S1, Figure 2a). Within bacteria or fungi, rhizosphere samples have a large number of spore-forming (SF) microorganisms (SFM) with fast growth within 24 h (for bacteria) or 72 h (for fungi). In the collected rhizosphere, 222 bacterial strains and 47 fungi were present. In contrast, there were 98 bacterial strains and 19 fungus isolates from soils ( Figure 2b). Moreover, the number of isolates was predominated by SF-bacteria independently of the sample (Figure 2b). The litter samples from the forest area (Sample 4) had the highest among the sporulating microorganisms. However, fungal strains could not be collected from samples 12, 14, and 17 (Table S1).

Results
The characteristics of the site sampling and type of source used for spore-isol strains are shown in Table S1. Twenty samples were taken with different veget histories, from forestal areas to rice fields (indicated as paddy fields). The sam included soils taken 20 cm or deeper and rhizosphere, including roots and organic m distinguished as litter in this study.
A total of 386 isolates were collected from the roots of both rice cultivars. From 117 isolates were obtained, and 269 isolates (71% of the total) were collected from rhizosphere (Table S1, Figure 2a). Within bacteria or fungi, rhizosphere samples ha large number of spore-forming (SF) microorganisms (SFM) with fast growth within (for bacteria) or 72 h (for fungi). In the collected rhizosphere, 222 bacterial strains an fungi were present. In contrast, there were 98 bacterial strains and 19 fungus isolates soils ( Figure 2b). Moreover, the number of isolates was predominated by SF-bac independently of the sample (Figure 2b). The litter samples from the forest area (Sa 4) had the highest among the sporulating microorganisms. However, fungal strains c not be collected from samples 12, 14, and 17 (Table S1). All SFM were re-streaked onto their correspondent media for pure colo confirming the fast growth selected term in this study; it is important to mention th some cases, the SFM continued to appear after keeping the plates under longer incuba This study tested 320 bacterial strains and 66 fungal isolates for screening. As a trap, these strains were re-inoculated into the 'Koshihikari' rice cultivar. At least strains from each sample type showed the best plant performance confirmation betw them, and in comparison with the positive control (the biofertilizer) and negative co (non-inoculation), the top representative groups with the best plant performance selected. However, only 28% of the SFM were selected, and 18 bacterial strains isolated from both sample types (4 from soil and 14 from rhizosphere). Furthermo representative group of fungal isolates was selected (15 strains, 23% of the total). All SFM were re-streaked onto their correspondent media for pure colonies confirming the fast growth selected term in this study; it is important to mention that in some cases, the SFM continued to appear after keeping the plates under longer incubation.
This study tested 320 bacterial strains and 66 fungal isolates for screening. As a host trap, these strains were re-inoculated into the 'Koshihikari' rice cultivar. At least two strains from each sample type showed the best plant performance confirmation between them, and in comparison with the positive control (the biofertilizer) and negative control (noninoculation), the top representative groups with the best plant performance were selected. However, only 28% of the SFM were selected, and 18 bacterial strains were isolated from both sample types (4 from soil and 14 from rhizosphere). Furthermore, a representative group of fungal isolates was selected (15 strains, 23% of the total).

Phylogenetic Analysis of Fukushima Sporulating Isolates
The 16S rRNA gene and rpoB gene sequences of eighteen bacterial strains were used for phylogenetic analysis. In the case of fungal isolates, ITS1 and 18S rRNA gene sequences of fifteen fungal strains were used for their identification (Table 1). On the one hand, the phylogenetic tree results of "Bacteria isolates" were similar between both genes (Figures S1 and S2), which showed that Fukushima isolates were clustered into the Bacilliales group. Interestingly, in this study, the SF-bacterial isolates were closely related to the Firmicutes phylum, only the Bacilli class, independently of the gene used. Then, the clustering analysis results for both genes divided the isolates into six genera groups (Figures S1 and S2) as follows: Bacillus (50% of total bacterial isolates), Priestia (11%), Peanibacillus (22%), Viridibacillus (6%), Solibacillus (6%), and Lysinbacillus (6%).
Bacillus genus is dominant in this study, and according to the species, 9 isolates were divided into 3 sub-groups, establishing a relationship between Fukushima isolates and B. pumilus (4 strains), Bacillus sp. close to B. thuringiensis (3 strains) and B. simplex (2 strains) formerly known as Peribacillus simplex (Figures S1 and S2). The subsequent group was the Paenibacillus genus which included 4 strains and somehow closely correlated to Paenibacillus amylolyticus; most of the strains were classified as Paenibacillus sp. (Figures S1 and S2) in consensus with all genetic results ( Table 1). The remaining isolates are divided into one strain for each genus (Table 1, Figures S1 and S2).
Regarding the frequency of bacterial isolates, the Bacillus isolates were mainly collected from soil samples with Japanese pampas grass or rice cultivation history. However, two strains, Fuk 10_10 and Fuk 4_10 were obtained from the rhizosphere. In contrast, Priesta isolates were only from rhizosphere samples. In the case of Paenibacillus, these strains were collected equally from the soil and rhizosphere. Furthermore, the Viridibacillus strain was collected from the rhizosphere (Table S1), and up to date, it is the first time to describe these genera as plant growth promoting (Table 1).
On the other hand, "Fungal isolates" were genetically analyzed based on the sequences of ITS1 ( Figure S3) and 18S rRNA ( Figure S4) gene sequences, and their consensus species per each selected isolate is summarized in Table 1.
In this study, the fungus strains were divided into 3 big groups in the fungi taxonomy. The predominant group in this study is closely related to the Ascomycota division (80% of the total fungal strains), followed by the Mucoromycota phylum (13%) and Basidiomycota division (7%), independently of the phylogenetic gene used. The clustering analysis results for both genes divided Ascomycota division into five genera groups (Figures S3 and S4) as follows: Trichoderma (27% of total fungus strains), Penicillium (27%), Fusarium (13%), Sarocladium (7%), and Bisporella (7%). Trichoderma was mainly related to three species: T. polysporum, T. lixii, and T. reesei ( Figure S3). In the case of Penicillium genus, the isolates were closely related to P. decumbens or P. brasilianum and P. oxalicum ( Figure S3). Furthermore, the Ascomycota isolates were related to two species as F. equiseti and F. oxysporum. The remaining Ascomycota strains were closely related to Sarocladium bactrocephalum and Bisporella discendens.
In contrast, Mucoromycota and Basidiomycota isolates were closely related to one specie, respectively ( Figures S3 and S4). Fuk 38 and Fuk 39 were related to Mortierella alpina (Mucoromycota), and Fuk 6 was related to Cutaneotrichosporon moniliiforme (formerly known as Trichosporon moniliiforme). Regarding the occurrence of fungal isolates, the highest number was rhizosphere samples, with 87% of the total (Table 1). Only two isolates were collected from soil; both were found in soil with Japanese pampas grass cultivation history. Furthermore, to date, it is the first time that Cutaneotrichosporon, and Bisporella isolates were described as plant growth promoting (Table 1). Additionally, several fungal strains have been described as pest biocontrol, cellulose degradation, endophytic fungus, or even a pathogen with phytohormone production activity.

Physiological Profile of SFM Based on Plant Growth Requirement
All genetically categorized isolates were phenotypically characterized and assessed for different abiotic stress. First, all isolates were tolerant to high-temperature, especially SF-bacteria. Salinity tolerance was determined by recording growth rates compared to the control (media without any modifications). A total of 89% of bacterial isolates and 80% of fungi could grow at high NaCl concentrations (4%) and showed good or the same growth as the control (Table S2). However, two bacterial isolates, Fuk 11_7 (from the soil) and Fuk 13_15 (from the rhizosphere) could not tolerate NaCl at 4%; these bacterial isolates belonged to Bacillus and Viridibacillus, respectively (Table 1). Similarly, three fungal strains, Fuk 44, Fuk 49, and Fuk 36 could not grow at 4% NaCl but could grow at lower salinity (Table S2); all of them were isolated from rhizosphere samples and belonged to the Trichoderma genus (Table 1).
Fukushima SFM showed tolerance to different pH conditions, from acidic to alkaline. Surprisingly, fungal isolates were not affected by pH changes. However, bacterial isolates were more affected by acidic than alkaline pH. The isolates Fuk 11_7 and Fuk 11_10 (from rhizosphere) could not grow under acidic conditions. Moreover, Fuk 11_7 could not grow at 10 pH. Figure 3(a1,b1) showed the relation between the stress abiotic and the SFM strains, indicating that one strain is out in the SF-bacteria strains, which is identified as Fuk 11_7 (Bacillus sp.); it is a very sensitive bacterial strain to any abiotic stress. NaCl only affects the growth of fungus (Figure 3(b1), Table S2).
The isolates growth was accessed based on nutrient assimilation as plant partners such as P or K solubilization, Fe chelation, and capacity of growth in N-free media (Table S2). The correlation results between nutrient and isolates are shown in the Venn diagram (Figure 3(a2,b2)). In the case of bacterial isolates, most of the isolates could grow under different media with inorganic K and P source, N-free media, but no isolate could grow and assimilate Fe in the media used in this study (Figure 3(a2), Fe is absent in the group). For K and P, solubilization results showed that almost all isolates (89%) could grow under these conditions. The K solubilization ratio is low, and no significant difference was found between strains. In contrast, P solubilization in the isolates showed significate differences, and six isolates (33% of the total) could grow but did not show solubilization. Those isolates with high P solubilizing index were 56%; these ten isolates also shared a positive correlation with another tested nutrient (Figure 3(a2)). In ascending order of isolates' solubilization index for P as follows: Fuk 4_8, Fuk 4_31, Fuk 10_6 and Fuk 10_10, Fuk 5_23 and Fuk 13_14 with P index of 1.0, Fuk 11_10 was 1.2, Fuk 7_4 and Fuk 8_6 were 1.3, and the highest P index was the isolate Fuk 11-2 (Peribacillus; Table 1 and Table S2). In the case of N, most of the isolates (94%) showed N 2 fixation (in Ashby media). Similar to previous stress results, the Fuk 11_7 (Bacillus sp.) could not grow with inorganic K or P and under N-free media.
Microorganisms 2023, 11, 232 9 of 20 the SFM strains, indicating that one strain is out in the SF-bacteria strains, which is identified as Fuk 11_7 (Bacillus sp.); it is a very sensitive bacterial strain to any abiotic stress. NaCl only affects the growth of fungus (Figure 3b.1, Table S2).
. The isolates growth was accessed based on nutrient assimilation as plant partners such as P or K solubilization, Fe chelation, and capacity of growth in N-free media (Table  S2). The correlation results between nutrient and isolates are shown in the Venn diagram (Figure 3(a2,b2)). In the case of bacterial isolates, most of the isolates could grow under different media with inorganic K and P source, N-free media, but no isolate could grow and assimilate Fe in the media used in this study (Figure 3(a2), Fe is absent in the group). For K and P, solubilization results showed that almost all isolates (89%) could grow under these conditions. The K solubilization ratio is low, and no significant difference was found between strains. In contrast, P solubilization in the isolates showed significate differences, and six isolates (33% of the total) could grow but did not show solubilization. Those isolates with high P solubilizing index were 56%; these ten isolates also shared a positive correlation with another tested nutrient (Figure 3(a2)). In ascending order of isolates' solubilization index for P as follows: Fuk 4_8, Fuk 4_31, Fuk 10_6 and Fuk 10_10, Fuk 5_23 and Fuk 13_14 with P index of 1.0, Fuk 11_10 was 1.2, Fuk 7_4 and Fuk 8_6 were 1.3, and the highest P index was the isolate Fuk 11-2 (Peribacillus; Tables 1 and S2). In the case of N, most of the isolates (94%) showed N2 fixation (in Ashby media). Similar to previous stress results, the Fuk 11_7 (Bacillus sp.) could not grow with inorganic K or P and under N-free media.
On the other hand, fungus isolates could not grow in media supplemented with On the other hand, fungus isolates could not grow in media supplemented with inorganic K (Figure 3(b2), K is the out-group). For Fe chelation in the vegetative stage, only 3 isolates could grow (20%) and showed a good Fe index at 2.0 as Fuk 58 (Fusarium), Fuk 10, and Fuk 66 (Penicillium). In the case of P solubilization, 93% could grow in media supplemented with inorganic P, but only 20% (3 isolates) showed good solubilization. These isolates were Fuk 6 (Cutaneotrichosporon with 1.0 index), Fuk 50 (Penicillium with 1.1 value), and Fuk 59 (Bisporella with 1.3 value) with a significative P solubilization index. Only Fuk 49 (Trichoderma) could not grow in media supplemented with inorganic P. N 2-fixation (in media) showed in 80% of the isolates; Fuk 36, Fuk 40 (Trichoderma) and Fuk 63 (Penicillium) could not grow in N-free media.

Metabolic Profile of Fukushima Isolates at the Vegetative Stage
Biolog TM microplates were used in this study as phenotyping and metabolic fingerprint on the Fukushima spore-forming microorganisms. The results showed some correlation between some substrate and phylogenetic clustered groups. Firstly, SFM isolates in this study are correlated for being incapable of metabolizing or assimilating D-psicose and L-arabinose; those that showed some kind of reaction were inconclusive (Figure 4: SF-bacteria, Figure 5: Fungi , Tables S3 and S4). glucose-phosphate ( Figure 5). Sarocladium could metabolize 65 compounds, but there were 23 compounds where there was no response with D-fructose, α-D-glucose, Dmannitol, sucrose, L-asparagine, hydroxy-L proline, and L-pyroglutamic acid. Bisporella isolates seem to be sensitive to any media source, and it is distinguished by metabolized adonitol, D-fructose, L-fucose, α-D-lactose, and the organic acids such as pyruvic, citric, D-galactonic, lactone glucosaminic, urocanic acids. Cutaneotrichosporon isolate could not grow in D-galactose, α-D-glucose, D-mannose, and succinic acid ( Figure 5, Table S4).    In this order, the SF-bacterial isolates as shown in Figure 4. Bromosuccinic acid, ierythritol, and β-methyl-D-glucoside were assimilated in almost all strains isolated in this study, except for Fuk 10_10 and Fuk 5_22, with inconclusive results, which are in the borderline according to the Biolog ™ protocols.
In detail, the Priestia group is capable of metabolizing 77 compounds; however, these isolates could not grow or use the following sources such as α-D-glucose, D-mannose, and the organic acids such as formic, γ-Hydroxybutyric, and glutamic acids. While the isolates related to B. pumilus could metabolize acetic and citric acid, but they could not grow with D-galactose as the only carbon source. In the remaining compounds, the response varied depending on the isolate. The B. thuringiensis group shares certain characteristics with each strain as incapable of metabolizing α-D-glucose and D-gluconic and D-saccharic acids. These strains could metabolize 18 compounds, such as gentiobiose, raffinose, and some acids, such as glucosaminic, α-ketoglutaric, and malonic acids. Peribacillus isolates can synthesize around 26 compounds such as glycerol-phosphates, citric and phydroxyphenlyacetic acids. However, these isolates cannot grow in media with carbon sources such as D-galactose, D-mannose, D-glucuronic and succinic acids.
Furthermore, Viridibacillus, Lysinibacillus, and Solibacillus isolates could not grow with glucosamine-bound compounds, D-mannose, and some organic acids such as malonic and D-saccharic acids ( Figure 4, Table S3). Lastly, Paenibacillus share the characteristics of metabolizing some organic acids, maltose, and β-methyl-D-glucoside. However, Paenibacillus isolates cannot metabolize 18 compounds such as organic acids such as succinic acid, and sugars such as D-galactose, and D-mannose ( Figure 4, Table S3).
On the other hand, the fungal isolates' metabolic profiles were examined, as shown in Figure 5. These fungal isolates were correlated by the positive responses to metabolize or assimilate dextrin, glycogen, D-melibiose, cis-aconitic, α-hydroxybutyric, α-ketobutyric, succinamic acids, and L-ornithine. In the case of malonic acid, lactic, acetic, propionic, formic, and α-ketoglutaric acids were compounds assimilated or metabolized by almost all fungal isolates except Fuk 59.
The fungal isolates share certain characteristics by the phylogenetic group as follows: Trichoderma isolates metabolize 57 compounds ( Figure 5). However, these isolates cannot grow in D-cellobiose, L-fucose, and D-galactose. Penicillium isolates have in common that they could metabolize 34 compounds, but these isolates are not capable of growing in the presence of D-cellobiose, D-galactose, α-D-glucose, and L-rhamnose. The isolates' responses in front of the remaining compounds depend on the strains, including inconclusive ones ( Figure 5, Table S4).
Fusarium could not metabolize α-D-glucose and D-glucuronic acid; however, the isolates could metabolize the remaining compounds. Mortierella could not grow with 23 compounds such as using α-D-glucose and D-mannitol as thesole carbon source, and organic acids such as D-saccharic, succinic, L-glutamic acids, and compounds with glucosephosphate ( Figure 5). Sarocladium could metabolize 65 compounds, but there were 23 compounds where there was no response with D-fructose, α-D-glucose, D-mannitol, sucrose, L-asparagine, hydroxy-L proline, and L-pyroglutamic acid. Bisporella isolates seem to be sensitive to any media source, and it is distinguished by metabolized adonitol, D-fructose, L-fucose, α-D-lactose, and the organic acids such as pyruvic, citric, D-galactonic, lactone glucosaminic, urocanic acids. Cutaneotrichosporon isolate could not grow in D-galactose, α-D-glucose, D-mannose, and succinic acid ( Figure 5, Table S4).

Plant Growth-Promoting Performance of SFM from Fukushima
The selected SFM from Fukushima were used as inoculants during the vegetative phase. Almost all selected isolates promoted plant growth except for two, Fuk 7_4 (Paenibacillus sp.) and Fuk 38 (Mortierella sp.), in comparison with the controls as shown in Figure 6 (SF-bacteria) and Figure 7 (Fungi). The length is positively affected in the case of the bacteria inoculants, which showed rice plant growth of >12 cm, equal to TUAT1 (positive control) or higher than the negative control ("Control: non-inoculation", Figure 6a). In con-trast, the fungus inoculants did not show any improvement in the length of the rice plants; these results are not significantly different from the negative control (Control, Figure 7a). However, we could not find any correlation between the length and biomass promotion.   (a) length (cm), (b) total fresh biomass (mg), (c) total fresh shoot mass (mg), and (d) total root mass (mg). Control means non-inoculation, and TUAT1(+) means positive control from based sporeforming biofertilizer. Dunnet's test was used for length and control as based comparative, and Tukey's test was applied for other parameters *, **, and *** means a value considered statistical significance with p ≤ 0.05, p ≤ 0.01, and p ≤ 0.001, respectively. No annotation means no significance.

Discussion
The diversity of SFM in spore form are dormant species with great importance worldwide and have been examined with various techniques and criteria [14,27,28]. (a) length (cm), (b) total fresh biomass (mg), (c) total fresh shoot mass (mg), and (d) total root mass (mg). Control means non-inoculation, and TUAT1(+) means positive control from based sporeforming biofertilizer. Dunnet's test was used for length and control as based comparative, and Tukey's test was applied for other parameters *, **, and *** means a value considered statistical significance with p ≤ 0.05, p ≤ 0.01, and p ≤ 0.001, respectively. No annotation means no significance.
The total biomass showed significant biomass with more than 85 ± 0.9 mg per nursery plant; in the case of bacteria, one isolate, Fuk 8_6 (B. pumilus) showed the highest promotion with 148.2 ± 0.1 mg per plant (Figure 6b). Other isolates related to B. pumilus (Fuk 11_1, Fuk 11_7), Peribacillus (Fuk 11_2, Fuk 5_23), and Viridibacillus (Fuk 13_15) showed significant promotion in comparison with the TUAT1 (positive control which is a commercial biofertilizer based on spore-forming Bacillus pumilus) with more than 103.1 ± 0.5 mg per plant. On the other hand, fungal inoculants promoted the biomass with more than 100.3 ± 0.1 mg per plant, those isolates with significate biomass promotion were classified as Penicillium (Fuk 66, Fuk 63) and Fuk 36 (Trichoderma) with more than 160 mg per plant; subsequently by Fuk 58 (Fusarium) and Mortierella (Fuk 39) with more than 150 mg per plant (Figure 7b). The remaining isolates showed high biomass compared with the TUAT1 inoculant, except for Fuk 38 (Mortierella) and Fuk 6 (Cutaneotrichosporon).
As mentioned before, biomass has been improved by SFMs as inoculants. However, the best performance in each plant group in some cases differs, such as Fuk 5-23, which showed the highest shoot mass with 76.5 ± 0.2 mg ( Figure 6c); but in the case of fungi, there is no higher promotion than TUAT1 (Figure 7c). In the root case, two bacterial isolates (Fuk 11_7: 66.5 ± 0.3 mg and Fuk 11_1: 61.2 ± 0.2 mg) showed significant difference compared with TUAT1 (45.8 ± 0.1 mg), as shown in Figure 6d. However, the fungal inoculants showed great improvement in root mass and root hairs, being significantly positive and higher with Fuk 6 (77.9 ± 0.2 mg), Fuk 9 (74.5 ± 0.3) mg, and Fuk 16 (73.7 ± 0.1 mg) (Figure 7d).
In summary, according to the analysis of the top rice promoters at the seedling stage, the SF-bacterial isolates were Fuk 8_6 (B. pumilus), Fuk 11_7 (P. megaterium) and Fuk 13_15 (Viridibacillus sp.); the first two strains were isolated from soil and the last one from the rhizosphere. For fungus isolates, they were Fuk 66 (Penicillium sp.), Fuk 63 (Penicillium sp.), and Fuk 36 (Trichoderma); these strains were mainly isolated from the rhizosphere. Interestingly, we correlated that these SF isolates have the ability or the metabolic pathway to work under inorganic P corroborated by media and by assimilation of phosphates compounds.

Discussion
The diversity of SFM in spore form are dormant species with great importance worldwide and have been examined with various techniques and criteria [14,27,28]. However, most of the descriptions are focused on public health and poorly describe them from the perspective of plant-microbe interactions or as an inoculant [4,14]. Our study focused on fast-growing SFM isolation and their characteristics as future plant partners with fast break dormancy strains, and their positive effects on the plant growth or health. This is because a slow-growing group of microbes could not compete with fast-growing bacteria and fungi, especially during the early phases of crop development where microbial diversity establishes itself, even after harvest [3,4]. Additionally, our study will increase the knowledge of SFM in Fukushima prefecture after a long-term halt to crop cultivation in many sampling areas.
The SFM were found independently from rhizosphere and soil samples due to the large range of environmental conditions [27][28][29] in Fukushima prefecture. However, low diversity of SFM or absence was observed in some samples, which indicates that the method used may not be the most suitable for studying SFM diversity. Other reasons, such as the selection method, also need a longer period for isolation (≥48 h), and out of the isolates reported in this study other species may need to take more than 72 h to break spore dormancy. Lastly, some SFM might be uncultivable because the media use is unsuitable or the strains become uncultivable after changing their ecological niche [13,[28][29][30]. Our results indicate that the rhizosphere is a better source than the soil and bacterial isolates are more frequent and faster growers than the fungal type [3,13]. This is attributable to the ability of these SF-bacteria to better adapt than fungi, which gives them the capability to survive abiotic stresses, nutrient deficient, and compound productions such as biofilms [27,31].
In this study, we tried to cover the fast-growing SFM independently from the season due to decomposition deceleration; bacterial and fungal biomass and sporulation rates change according to the type of origin sample or area in one region whether it is a wetland, temperate stream, or arable land, in concordance to previous reports [13,[32][33][34][35][36]. SFM populations with different ecological or agricultural conditions are active colonizers of these ecosystems, thereby showing some specialized physiological functions associated. It was also observed that in the field with different stages of plant succession or rice development, bacteria and fungi seem to be the main colonizers, which was in agreement with our results wherein the bacteria were dominant [3,28,29,[34][35][36][37][38][39][40][41][42].
Similarly, our results suggested that the class Bacilli dominated fast growth and fastbreaking spore dormancy [34]. Garbeva et al. [34] described SF-bacteria as Bacilli-related community structures with high diversity among the order Bacillales, including Bacillus as main genus, their changes have been found under different agricultural regimes such as in permanent grassland, arable land, and under long term presence of grass in soil [34][35][36]. In contrast, our results are somewhat similar but include the predominance of filamentous fungi (Ascomycota) as the main group, which is supported by Sharma et al. [3]; it described that in paddy fields under different growth or decomposition stages, the fungal community and their ratio changes, and not only SF-bacteria [15]. Thus, these results could be associated more with the significant interaction between the time of decomposition and emersion for sporulation of fungal communities, which is related to the enzymatic process of C mineralization, N, P, and S cycles [3,28,29,[33][34][35].
Furthermore, SFM development is commonly attributed to the level of adaptation to a specific niche in terms of the ability to utilize available nutrient resources and tolerate metabolite accumulation or deficiency, such as proline, glycine, trehalose, sucrose and mannitol; on the other hand, these compounds cause oxidative and osmotic stresses in plants [9,15]. Our results indicated that SFM isolates utilize or metabolize these compounds, as demonstrated by Biolog ™ , and indirectly suggested that the presence of SFM in the soil might improve the rice seedling under stress or stress tolerance in plants. Moreover, the Biolog ™ method was used as an Eco-Plates, which can provide an overview and screening of the metabolic fingerprint characteristics of SFM in a short time [31,34,38,39].
The physiological profiling is accompanied by microbial identification for future investigations of microbial community structure and applications in the field [13,30,37,38]. The immediate results can be compared with other studies; for example, until the 2020 outcome, pathogenic spore-formers do not grow below pH 4.5 and the species encountered in spoilaged foods are not the same as in our study [14,27,39]. Therefore, the metabolic profile and other physiological results supported the adequate selection of SFM and the isolates' ability for use as a possible response to climatic changes or different crop cultivation due to its adaptability [13]. Moreover, our results suggested that SFM positively promoted plant growth through various mechanisms, such as plant hormone and ESP production, regulating nutrient exchange, and releasing stress-protectant compounds. At the same time, our SFM metabolic results support, for example, cytokinin production [1,16] by the assimilation of amino acids or sugar-bound and phosphate-bound compounds [15]. The stress-protectant compounds could be the result of SFM biosynthesis of total soluble sugar, proline, and trehalose or by regulating metabolites such as pyruvic acid, succinic acid, and uridine [9,15]. Future experimentation is needed for different SFM strains, which will help elucidate the underlying mechanisms and ecology, such as signaling, colonization, and SFM community interactions.
On the other hand, the spore resistance of SFM is related to the sporulation performance at optimal vegetative growth conditions resulting in the highest sporulation efficiency and spore resistance [14]. Until now, the major influence of sporulation was well reported to be the temperature conditions; based on this, knowing other factors could help increase their resistance to many abiotic stress factors and improve the predictability of spore behavior [1,8,14]. Additionally, this can change according to the succession in the environmental conditions and plant stage, similar to the ripening stages of SF-bacteria in cheese manufacturing [28,37].
Consequently, our results indicate that Bacilli isolates from Fukushima can tolerate low temperatures and acidic conditions and can also be heat-stable spores, as reported pre-viously [1,4]. Moreover, the SFM isolates are mainly aerobically forming with a mesophilic preference (optimal growth temperature between 10 • C and 30 • C) with stable spores forming at 65 • C. Hence, previous studies indicated that endophytic Penicillium, Aspergillus, Trichoderma, Bacillus, and Paenibacillus reduced pathogenic fungi and also induced the biomass yield of cultivated plants [12,28,29,36,37,40] which is similar to our results in the promotion of growth for Oryza sativa L.; however, further experiments are required for assertive use as a future pathogenic antagonist.
Finally, physiological results and their interaction with rice can be correlated; for example, those SFM isolates capable of metabolizing D,L,α-glycerol-phosphate, α-D-glucose-1-phosphate or D-glucose-6-phosphate, also have good P abilities reported in the Pikoskaya medium. At the same time, these results work as guidelines for other unknown strains, such as Viridibacillus, Lysinibacillus, and Bisporella reported in this study. Those strains are reported as PGPR or are known to be P solubilizers such as P. megaterium [4]. Altogether, this suggests that some species could solubilize P, K, and other minerals and fix nitrogen to promote plant growth. Similarly, it could control plant diseases [3,4,13,41], especially in Fukushima area, with the relationship between K/Cs uptake. Thus, their physiological role was one factor behind their plant association or interactions throughout [1,42]. In summary, these results suggest that SF-bacteria development and sporulation may be more resilient in shorter growth periods, and their effect as inoculants may be shorter than fungi; however, the effect of the fungi could remain longer than bacteria [13].

Conclusions
The SFMs from Fukushima prefecture selected were Fuk 8_6 (B. pumilus), Fuk 11_7 (P. megaterium), Fuk 13_15 (Viridibacillus sp.), Fuk 66 (Penicillium sp.), Fuk 63 (Penicillium sp.), and Fuk 36 (Trichoderma); these SF isolates could improve plant biomass under limited nutrients and showed good abilities under different abiotic stress and nutrients restriction. Moreover, further research is necessary to confirm the effectiveness of strains for rice and other crops in field conditions with low nutrient supply as a biofertilizer or against pathogens as biocontrol.
Supplementary Materials: The following supporting information can be downloaded at: https://www.mdpi.com/article/10.3390/microorganisms11020232/s1, Figure S1: Phylogenetic analysis of Fukushima isolates based on 16S rRNA sequences. Figure S2: Phylogenetic analysis based on the rpoB sequences of Fukushima isolates. Figure S3: Phylogenetic tree based on 18S rRNA sequences of Fukushima isolates, constructed with 1000 bp DNA fragment gene sequences. Figure S4: Phylogenetic tree based on ITS sequences of Fukushima isolates, constructed with a 150 bp DNA fragment from sequences. Table S1: Soil and rhizosphere sample characteristics.   Data Availability Statement: GenBank (http://www.ncbi.nlm.nih.gov/genbank, accessed on 15 January 2022). Accession Numbers: the obtained sequences for the different genes found in this study have been deposited in the DNA Databank of Japan (DDBJ) under accession numbers as follows LC606518-LC606535 for 16S rRNA sequences, LC606536-LC606553 for the rpoB gene, LC606568-LC606554 for the 18S rRNA gene, and LC606569-LC606580 for the ITS-1 region.