Introduction

The adverse impacts of climate change, including elevated temperatures and water scarcity, are increasingly affecting human life and ecosystems1. In the agricultural sector, these factors can harm plant growth and physiological processes2,3,4resulting in a substantial decrease in crop yield5. Global temperatures are expected to increase by 2–4 °C by the end of the 21 st century6,7. Despite the heat, severe drought affects many regions worldwide8. The increase in temperature causes rapid water evaporation, leading to catastrophic droughts, especially in dry regions9. A 1 to 3 °C warming scenario reduces wheat yields by 8.5 to 28.5% and moderate to severe Drought by 25.8% and 32.0%, respectively10,11. Heat and drought stress are recognized as major climatic factors with substantial detrimental impacts on wheat growth and development12. These stresses adversely affect plant growth, including organ development, grain filling, photosynthesis, transpiration, and nutrient uptake13. Additionally, the combined effects of heat and drought have significantly contributed to yield losses in many countrie14,15.

There are several bacterial species identified as plant growth-promoting rhizobacteria (PGPR), including Pseudomonas, Azospirillum, Azotobacter, Klebsiella, Enterobacter, Alcaligenes, Arthrobacter, Burkholderia, Serratia, and Bacillus, with the potential to enhance crop growth and yield16,17,18,19,20,21,22,23,24,25. PGPR promote crop growth through direct and indirect mechanisms. The direct effects include the production of plant hormones, nitrogen fixation, nutrient supply, and phosphorus and potassium solubilization26,27,28,29,30. In contrast, indirect effects include the production of antibiotics and lytic enzymes, inducing resistance to abiotic stresses, inhibiting pathogens, hydrogen cyanide production, and the alleviation of drought stress31,32,33,34,35. Also, PGPR exhibit mechanisms that produce osmoprotective compounds, increase enzymes that reduce oxidative stress, and stimulate production of growth regulators, like ethylene, which promote plant tolerance to abiotic stress36.

PGPR comprise a diverse group of microorganisms, some of which can form endospores (bacterial spores)37. Spore-forming bacteria (SFB) are typically gram-positive and are found within the genera Bacillus and Clostridium38,39. Nutrient depletion typically triggers bacterial cells to transform into endospores induced by fluctuations in environmental conditions40,41. The spores have minimal metabolism, respiration, and reduced enzyme production42. Thus, spores can remain stable without water and nutrients in extreme temperatures and pH, UV light exposure, and noxious chemicals43. When environmental conditions become favorable, the dormant spores germinate and transition back to a vegetative state44. Studies showed that inoculation of vegetative cells of Bacillus species enhanced the growth of various crop species; including wheat, corn, soybean, sugarcane, and rice45,46,47,48,49. The ease of endospore production, their resistance to environmental stress conditions, their suitability for long-term storage, and their suitability for field application, make them attractive as inoculants. Previous studies primarily focused on the use of vegetative cells as PGPR inoculants. In contrast, previous work demonstrated the superior performance of Bacillus altitudinis TUAT1 spores to promote rice growth compared to the use of vegetative cells47,50,51,52. The ability of spores to maintain viability under heat and dry conditions, as well as their stability under storage (up to 2 years) enabled development of a successful biofertilizer. A biofertilizer using spores of B. altitudinis TUAT1 is manufactured and marketed as “Yume-bio” (Asahi-agria Co., Ltd, Japan). This biofertilizer was specifically developed for rice plants in paddy fields. However, while previous research demonstrated strong growth promotion effects of the use of spore-based inoculants, studies were not performed to specifically examine the ability to use such inoculants to increase the tolerance of crop plants to heat and drought stress conditions. Abiotic stresses induce reactive oxygen species (ROS) accumulation within plant cells and high concentrations of ROS disrupt cellular homeostasis, damaging DNA, proteins, lipids, and polysaccharides, leading to dysfunction and programmed cell death53,54. Therefore, plants utilize various mechanisms to avoid the effects of ROS through catalase (CAT), which scavenges hydrogen peroxide produced by superoxide dismutase55. Indeed, heat stress upregulates ascorbate peroxidase (APX) activity, which is also a ROS scavenger, in wheat varieties56. Additionally, proline is an essential osmoprotectant, antioxidant, and protein stabilizer that aids plant stress tolerance. It is synthesized through a pathway regulated by the P5CS (Δ1-Pyrroline-5-Carboxylate Synthetase) gene57.

To assess the abiotic stress tolerance and PGP traits of SFB, we isolated novel strains from the rhizospheres of diverse crops that successfully colonize wheat roots. We then evaluated their ability to mitigate stress by applying spores under control, heat, drought, and combined stress conditions. Finally, to elucidate underlying mechanisms, we quantified transcript levels of the stress-responsive genes APX, CAT, and P5CS (Fig. 1).

Fig. 1
figure 1

Schematic representation of the experimental flow, illustrating SFB isolation, characterization, and inoculation to wheat under abiotic stress conditions

Full size image

Results

Isolation and characterization of spore-forming bacteria as PGPR

The isolation process yielded a total of 243 SFB, comprising 161 epiphytic isolates obtained from root surfaces and 82 endophytic isolates from crushed root tissues. Their phenotypic characterization revealed diverse colony morphology on TSA medium, as illustrated in Figure S1. The frequency of tolerance to abiotic stresses was higher in endophytes compared to the epiphytic isolates (Figure S2). Based on selected round morphology and tolerance to different pH levels and salt concentrations, 62 promising isolates were selected for further study. These isolates were classified into three genera based on the 16 S rRNA analysis: Bacillus, Paenibacillus, and Priestia. Of the 62 isolates, 28 were identified as different species of Bacillus, 20 revealed maximum similarity to Paenibacillus, and 14 displayed high similarity to Priestia (Fig. 2 and Table S2). All the isolates and their type strains clustered into the corresponding clades. Of these, 39 isolates (56.5%) exhibited phosphate-solubilizing activity with P-solubilizing index values ranging from 1.1 to 1.7 (Figs. 2 and S3-A). The isolates showed the following distribution: 17.9% belong to Bacillus, 35.8% are associated with Paenibacillus, and 46.1% come from the Priestia group. In addition, 44 isolates (70.9%) showed K-solubilizing activity with index values ranging from 1.1 to 1.9 (Figs. 2 and S3-B). Siderophore production was observed in 12 out of 62 isolates (19.4%), with the diameter of the orange zones formed by these isolates ranging from 1.2 to 2.9 mm, and ten of twelve siderophore-producing isolates (83.3%) were clustered into Bacillus. Notably, isolates FRREn1 (Bacillus amyloliquefaciens) and TSTs-WSREn1 (Bacillus velezensis) demonstrated the highest siderophore production, with orange zone diameters of 2.9 mm and 2.5 mm, respectively (Figs. 2 and S3-C). Regarding IAA production, most isolates (93.5%) produced IAA. The levels of IAA production varied substantially, ranging from a minimum of 0.1 µg/mL to a maximum of 12.6 µg/mL, while Priestia isolates showed relatively higher production compared with others (Figs. 2 and S3-D). Among all isolates, TCREp2 and TSEn-CSREp1 from Paenibacillus group, and TSEnREp1, FTREn1, and TTREn1 from the Priestia group produced the highest levels of IAA. Meanwhile, Bacillus species showed lower production. Among 62 isolates, 11 (17.7%) showed acetylene reduction activity under nitrogen-free conditions regardless of the genus. The levels of ethylene production varied among these 11 nitrogen-fixing isolates, as illustrated in Figs. 2 and S3-E. Based on their capabilities, 13 isolates were selected for further research and characterized using additional housekeeping genes, such as rpoB and gyrA3 (Table S2). Among them, TTREn1 and TC-CSREp1 were chosen for whole-genome sequencing as representatives of Priestia and Paenibacillus, respectively.

Fig. 2
figure 2

The maximum likelihood phylogenetic tree of 62 SFB isolates was constructed based on their 16 S rRNA sequences and type strains. The Bradyrhizobium japonicum was used as an outgroup. Bootstrap support values, determined from a 1000-replicate neighbor-joining analysis of a 1500 bp DNA fragment. This tree shows the biological variation of groups, isolation location, host crop used for isolation, colonization pattern during isolation, P and K solubilization, siderophores production, IAA production, ARA activity, and abiotic stress tolerance. According to Artigas Ramírez et al. (2023), the growth of isolates was evaluated relative to a non-stressed control using a scale ranging from no growth (-) to excellent growth (3, similar to the control), with intermediate levels of weak growth (1) and good growth (2) corresponding to 10–20% and 30–60% growth compared to the control, respectively. The consecutive blank rows in the reference strains mean not tested in our study

Full size image

Whole genome sequence analysis

Whole genome sequences of TTREn1 and TC-CSREp1 were analyzed with the relevant statistics shown in Table S3. The genome sequence predicted 11,842 and 12,128 genes in TTREn1 and TC-CSREp1, respectively. The TTREn1 isolate showed a unique GC content % of 37.5, which is lower compared to TC-CSREp1 and TUAT1. The average nucleotide identity (ANI) with the reference strain for TTREn1 and TC-CSREp1 was 95.4% and 84.5%, respectively. TC-CSREp1’s ANI suggests our isolate is a novel strain of the Paenibacillus group (Table S3). The functional annotation of genomes was performed using the RAST-SEED viewer58which categorized the predicted proteins into various subsystems (Fig. 3A and B). The subsystem coverage of the genomes of TUAT1, TTREn1, and TC-CSREp1 was 29%, 23%, and 18%, respectively (Fig. 3A). Functional subsystem feature counts showed that amino acids and derivatives, followed by carbohydrates, were the most abundant categories across all strains (Fig. 3B). Genes associated with IAA production, auxin synthesis, and stress/drought response are shown in Figs. 3C and D. Details on genes, such as stress response, functional categories, and roles, are presented in Table S4. Additionally, several genes (ppk; polyphosphate kinase, nox; NADH oxidase, ureA; urease, atpE; ATPase subunit c, and poxB; pyruvate oxidase), perhaps playing a role in drought stress tolerance, were found in TTREn1 and TC-CSREp1 but absent in TUAT1.

Fig. 3
figure 3

Using the RAST-Seed viewer, the functional gene annotation connected to subsystems and their distribution in different categories on the whole genome sequences of TUAT1 and the two other isolates was performed. (A) the subsystem coverage and (B) subsystem feature count comparison of the isolates. (C and D) The circular genome maps of TTREn1 and TC-CSREp1 isolates

Full size image

Effect of SFB isolates on plant growth under a single abiotic stress

The spores from 13 selected isolates, along with B. altitudinis TUAT1, were inoculated onto wheat plants. Shoot dry weight significantly increased in six isolates under control conditions, while the remaining isolates also exhibited relatively higher values compared to the uninoculated control (Fig. 4A). A similar trend was observed under heat and drought stress, with five (TTREp1, TC-CSREp1, TSEn-CSREp1, TTREn1, and FP-WSREp1) and eight isolates (TCoREn1, TCo-SSREp1, TSTs-WSREn1, TC-CSREp1, TP-WSREp1, TTREn1, TC-WSREn1, and FP-WSREp1) showing significant increases compared to the uninoculated treatment. Notably, B. altitudinis TUAT1 significantly enhanced shoot dry weight under heat stress but not under drought conditions. Root dry weight increased in four isolates (TTREp1, TSTs-WSREn1, TTREn1, and TC-WSREn1) in control and heat stress, and five isolates (TTREp1, TCo-SSREp1, FTREn1, TTREn1, and TC-WSREn1) under drought stress, while B. altitudinis TUAT1 did not increase root dry weight under Drought conditions. The total dry biomass was increased in three isolates under control, nine isolates under heat stress, and 11 isolates under drought stress. Again, B. altitudinis TUAT1 did not improve these parameters under dry conditions. As for SPAD values, two, 12, and 4 isolates significantly increased under control, heat, and drought stress, respectively, indicating that inoculation of spores is effective in keeping high chlorophyll content, especially under heat stress conditions (Table S5). In contrast, tiller numbers were increased by inoculation of spores from four and six isolates under control and drought stress, respectively, but no isolates increased tiller numbers under heat stress. Leaf numbers were increased by spore inoculation of four, eight, and three isolates under control, heat, and drought stress, respectively. Additionally, the application of spores from some SFB significantly increased shoot length, root length, and root number compared to the uninoculated treatment (Table S5). The highest shoot length was observed in the TCo-SSREp1 isolate under control conditions, FP-WSREp1 under heat stress, and FRREp2 under drought stress conditions.

The colonization of the SFBs on the wheat roots was investigated (Fig. 4B and S4). For all isolates, the number of vegetative cells was higher than that of spore cells in epiphytically associated populations, whereas endophytic colonization did not follow the same pattern. Notably, most of the isolates, except for Bacillus, did not colonize endophytically as spores under drought stress, whereas some such as TSEn-CSREp1 and FR-WSREn1 (Paenibacillus), and FTREn1, and TTREn1 (Priestia) strictly colonized as epiphytes. Bacillus species were stably colonized as spores both endophytically and epiphytically under drought conditions (Fig. 4B). Bacillus isolates exhibited higher biofilm production in vitro, whereas all isolates belonging to Paenibacillus and Priestia genera showed negligible levels of biofilm formation (Figure S5).

Fig. 4
figure 4

(A) The effects of spores’ application of different SFB isolates on the dry weight of wheat shoots, roots, and total biomass under single abiotic stress conditions. Asterisks *, **, and *** denote p-values of < 0.05, < 0.01, and < 0.001, respectively. Red and black basslines separate all the graphs to compare high and lower isolates with B. altitudinis TUAT1 and uninoculated, respectively. (B) The heat map illustrates biofilm formation and the colonization of vegetative cells and spores in both epiphytic and endophytic sites of isolates

Full size image

Wheat response and the expression of stress tolerance genes with spore inoculation under the combined heat and drought stress condition

Six isolates were selected for further studies based on their better performance under the single abiotic stress treatments. The ability of these inoculants to affect wheat growth was tested under conditions of both heat and drought stress. Notably, spore inoculation with all six strains significantly increased total biomass under heat + drought stress treatment, meanwhile five isolates increased shoot dry weight and root dry weight, except TCo-SSREp1 in shoot dry weight and TP-WSREp1 in root dry weight (Fig. 5A). Bacillus altitudinis TUAT1 did not show a clear difference when compared to the uninoculated control. The application of spores significantly increased SPAD values compared to the uninoculated treatment. In contrast, the lowest SPAD value was observed in the uninoculated treatment under heat + drought and drought stress compared with the control and single abiotic stresses (Table S6). Tiller numbers and leaf numbers were not affected by the spore inoculation under the double abiotic stress. Shoot lengths were significantly increased by two isolates, TCo-SSREp1 and TTREn1, under heat + drought stress conditions. Root length significantly increased by inoculation with two isolates, TSEn-CSREp1 and TC-WSREn1, under the heat + drought stress condition. Notably, two isolates increased their root numbers under double abiotic stress, while all strains increased them under single drought condition. Under the control conditions, the expression level of APX was unchanged by inoculations, while the expression of CAT and P5CS was decreased by inoculation (Fig. 5B). Under heat stress, all isolates except for the TC-CSREp1 induced expression of all genes. In contrast, under drought stress, only B. altitudinis TUAT1 conferred remarkable induction of these genes, while inoculation of other isolates did not. The expression of APX was suppressed by TTREn1 inoculation and expression of CAT and P5CS was reduced by TCo-SSREp1. Under combined heat and drought stress condition, only B. altitudinis TUAT1 induced the expression of APX and P5CS, while TCo-SSREp1 decreased the expression of P5CS, TTREn1 suppressed those of APX and P5CS, and TC-CSREp1 downregulated all of them. Hence, these results failed to show a strong and direct connection between the transcriptional response of these stress-responsive genes and the ability of specific PGPR strains (e.g., TUAT1) to promote wheat growth under the various stress conditions tested.

Fig. 5
figure 5

(A) The effects of spores’ application of different SFB isolates on the dry weight of wheat shoots and roots and its overall biomass under double abiotic stress conditions. The stars on the bar plots indicate significant differences using the Dunnet test based on their p-values. The black baselines separate all the graphs to compare the clear difference between isolates and uninoculated treatment. Bars are color-coded to indicate isolate groups: grey for controls, green for Bacillus, purple for Paenibacillus, and orange for Priestia. (B) The gene expression level of wheat leaves under control, heat stress, drought stress, and heat + drought stress conditions. The relative mRNA levels of APX, CAT, and P5CS in leaves were determined and normalized to those of Actin. Means + standard deviations of three biological replicates are shown, with the mean of uninoculated under control condition set as 1

Full size image

Discussion

Endospore formation is limited to certain bacterial species. These SFB have evolved unique survival mechanisms that enable them to adapt to environmental fluctuations. Moreover, some plant-associated SFBs can enhance host plant tolerance to environmental stresses through endophytic colonization and other plant-beneficial traits. In our study, we isolated and characterized various SFB strains belonging to the genera Bacillus, Paenibacillus, and Priestia from the rhizospheres of corn, cotton, peanut, rice, sorghum, soybean, and taro. While the isolates were broadly distributed across host species, certain trends emerged. For example, relatively few isolates were obtained from sorghum, and most Paenibacillus strains were initially isolated as epiphytes. However, since all strains isolated epiphytically from various crops were capable of colonizing wheat endophytically, their origin or initial isolation type may not necessarily constrain their potential applications. This is supported by a previous study demonstrating endophytic colonization by all Paenibacillus species tested59. Our PGP assay among the 62 selected isolates showed that over 50% exhibited P and K solubilization abilities, with Priestia accounting for the largest share of both P-solubilizing (46.1%) and K-solubilizing (40.9%) isolates, suggesting Priestia isolates were particularly effective at mobilizing macronutrients, likely through the secretion of organic acids or phosphatases (Fig. 2)60,61. Siderophore production was observed only in Bacillus strains, suggesting that iron chelation is not a universal trait among our SFB. Furthermore, we were able to determine a clear pattern in the distribution of the biosynthesis of bacillibactin, which is a high-affinity siderophore produced by Bacillus species to scavenge iron under iron-limited conditions, and exporter genes among our isolates (Table S7)62. The IAA synthesis trait underscores the widespread PGPR capacity of SFB to modulate root architectures63. Although we observed very low ARA activities, multiple attempts to amplify nifH genes using established primer sets consistently failed64,65suggesting that these strains may possess different coding sequences. Perhaps, this discrepancy in previous reports detecting nifH in Bacillus might stem from inadvertent detection of contaminated diazotrophs, which we eliminated by heat treatment during the screening, yielding pure SFB isolates.

The gene comparison analysis between Bacillus altitudinis TUAT1 and our new isolates, Paenibacillus sp. TC-CSREp1 and Priestia aryabhattai TTREn1 revealed significant differences. Unlike B. altitudinis TUAT1, the isolates TC-CSREp1 and TTREn1, which demonstrated the strongest PGPR activity in wheat under stress condition, harbor the ppk, ureA, atpE, nox, and poxB genes. Inorganic polyphosphate (poly P) synthesized by poly P kinase (PPK) contributes to survival functions such as motility, quorum sensing, biofilm formation, and the ppk locus enhances tolerance to desiccation and osmotic stress66. ureA, which encodes an enzyme involved in urea hydrolysis, and atpE, encoding a component of ATP synthase, contribute to pH homeostasis via ammonia and ATP production, respectively60,67. nox and poxB, encoding NADH oxidase and pyruvate oxidase, mitigate oxidative stress68,69. The presence of these stress-related genes in both TTREn1 and TC-CSREp1, which exhibited superior performance compared to TUAT1 under drought stress, may confer functional advantages that enhance plant tolerance to abiotic stress. Further analysis of gene expression characterization is necessary to elucidate the relationship between these bacterial genes and stress tolerance mechanisms, as certain Bacillus species have demonstrated plant growth promotion at levels comparable to TTREn1 and TC-CSREp1.

While some SFBs have been reported to enhance wheat growth under stress conditions where they employed vegetative cells as an inoculant, the robust benefits of spore application observed across normal and stress environments in this study further validate the efficacy of spore-based inoculation strategies70,71,72. The application of Bacillus altitudinis TUAT1 spore results in higher growth promotion than vegetative cells, using rice, Arabidopsis, and Setaria plants, with a certain number of spores detected, indicating either that some spores did not germinate or some germinated spores formed spores on the plants47,50,51,73. The current study also significantly increased wheat both shoot and root biomass, tiller and leaf numbers, as well as shoot and root lengths, under both control and stress conditions and showed colonized spores after their application. This reinforces our proposal that utilization of spores as an inoculant outperforms vegetative cells due to their stability and their intrinsic plant growth-promoting activities.

SFB-associated compounds, such as flagellin, peptidoglycan (PGNs), lipopeptides including surfactin, volatile organic compounds (VOCs), could trigger induced systemic resistance (ISR) of the host plants, thereby enhancing their tolerance to abiotic stresses74,75,76,77. Recently, ionome analysis revealed that the spores of TUAT1 contain higher levels of Na, Ca, and Mn compared to vegetative cells78. Endospores are known to accumulate several key components that contribute to their remarkable resistance and stability. Among these, calcium ions (Ca²⁺), typically present as calcium-dipicolinate (Ca-DPA), play a crucial role in stabilizing DNA and dehydrating the spore core. Dipicolinic acid (DPA), a compound unique to endospores, is essential for heat resistance and DNA protection. In addition, minerals such as Mn and Mg are believed to enhance resistance to oxidative stress. The spore cortex contains a specialized form of PGNs, which maintains dormancy and structural integrity. TUAT1 spores and their extracted peptidoglycan were found to promote root growth in Arabidopsis, but this effect was not observed in the lym3 mutant, which lacks a functional peptidoglycan receptor79. These findings suggest that PGNs also contribute to plant growth promotion. Although the precise mechanism by which spores of the newly isolated strains promote plant growth remains unclear, our results indicate that endospore-specific components may also play an important role in enhancing plant growth and abiotic stress tolerance.

Our research also indicated that abiotic stresses affect bacterial colonization of the rhizosphere (Figure S4). We propose some possibilities: the expansin-like EXLX1, which is a gene proposed as regulating endophytic colonization of Bacillus species, is absent in isolates of Paenibacillus and Priestia80alteration of the metabolic composition in root exudates by drought creates unfavorable conditions for bacterial survival within root tissues or disrupts exopolysaccharide (EPS) biosynthesis, which alleviates drought stress of plants and facilitates bacterial colonization81,82,83,84,85,86,87,88. However, although four of five of our Bacillus isolates, which maintained the levels of endophytic colonization under the drought condition, produced more biofilm in vitro. Yet no clear correlation was found between plant growth promotion and colonization levels. Additionally, within the genomes across the three genera, there was no clear pattern with regard to the EPS biosynthesis genes and the EXLX1 gene (Table S7), suggesting that the alleviation of drought stress by these isolates was caused by a different mechanism other than colonization of bacteria.

PGPR inoculation primes the plant’s oxidative stress response, thereby enabling more rapid and robust activation of defense mechanisms upon exposure to stimuli such as abiotic stresses89. In our study, spore application primed the plant response, illustrated by CAT and P5CS expression, under normal conditions, which is supported by the previous studies. Sarkar et al. reported that Bacillus safensis was shown to transiently suppress P5CS expression in wheat at the early phase of heat stress (at 0 h and 4 h during treatment), which effectively primed the plants for a stronger P5CS induction at 8 h and 12 h90. Lee et al. also published that rice plants inoculated with Priestia megaterium strain CACC109 exhibited elevated P5CS expression and consequent proline accumulation during the early stage of drought stress (at 10 days), which in turn led to a reduced P5CS induction at the later stage (35 days), while expression of P5CS, APX, and CAT under normal conditions was suppressed91. CACC109 strain also suppressed APX expression at the late stage of drought stress, while it induced CAT, suggesting a regulatory network system of ROS detoxification to maintain homeostasis. Similarly, Bacillus amyloliquefaciens 5113 maintained reduced APX expression in wheat seedlings throughout several days of drought stress92. TUAT1, TCo-SSREp1, and TTREn1 showed higher expression of P5CS under heat stress upon priming, while only TTREn1 reduced P5CS expression under drought stress. While we did not observe distinct responses APX and CAT expression under drought stress as previously reported, the isolates TTREn1 and TCo-SSREp1, which promoted plant growth, suppressed APX and CAT expression, respectively. In contrast, TUAT1, which did not exhibit a PGPR effect under drought, increased the expression of all genes compared to the mock treatment. In the case of combined stress with heat and drought, our transcriptional results did not display the dramatic changes to inoculation, which may be attributed to the attenuation of plant responses to combined factors compared to those of individual stressors93,94. Tyagi et al. also observed that wheat did not exhibit a synergistic induction in APX activity; instead, the APX response was more subdued than under either heat or drought alone95. Similarly, Raja et al. reported that tomato plants under combined stress attenuated APX activity compared with individual treatments, while CAT activity exceeded that seen under heat stress but showed comparable levels under drought, whereas these conditions drove the greatest H₂O₂ and proline accumulation96. This accumulation may explain the more severe defect phenotypes in the plant than found with a single stress97. Further research is needed in order to discover the mechanisms of abiotic stress alleviation by PGPR, especially when plants are simultaneously challenged with multiple stresses.

Materials and methods

Sampling and isolation of epiphytic and endophytic SFB

The soils and roots of various upland and paddy plant rhizospheres from the fields of Tokyo University of Agriculture and Technology (TUAT) and Fukushima prefecture were used for the isolation of the SFBs (Table S1). The sampling depth varied from 0 to 25 cm depending on the depth of the crop roots. Soil pH and electroconductivity (EC) were determined98. To isolate the epiphytic SFB, the plant roots were carefully washed with sterilized reverse osmosis (RO) water to remove soil from their surface. To isolate the aerobic epiphytic and endophytic SFB, the protocol was followed by Megías et al. and Alves-Júnior et al., respectively99,100. To obtain bacterial spores, an aliquot (1.5 ml) of both the epiphytic solution and homogenized root solution (endophytic bacteria) was heated for one hour at 65°C101. Then, 100 µl of each solution (10–1 dilution) and two serial dilutions (10–2 and 10–3) were plated on Trypticase Soy Agar (TSA) plates containing cycloheximide fungicide (200 ppm) and incubated at 28 °C for 1–2 days to obtain single colonies. To obtain pure isolates, the colonies were streaked onto new TSA plates for purification. Afterward, all isolates were cultured in TS broth. Finally, 500 µl of each culture was combined with 200 µl of 50% (V/V) glycerol and stored at − 80 °C for future use.

In addition, the soil from the rhizosphere of TUAT and Fukushima samples was inoculated to wheat, corn, and soybean to isolate more diverse SFB. The seeds were sterilized in 70% ethanol for 30 s, followed by 3 min in 3% (V/V) sodium hypochlorite, and rinsed five times with sterilized RO water102. Afterward, the seeds were placed into the petri dish surrounded with paper to retain the moisture and incubated for two days at 28 °C. The germinated seeds were then transferred to plastic boxes (10 × 12 cm, width and height, respectively) containing sterilized vermiculite (Hiruishi-tech Co., Osaka, Japan). To prepare the soil inoculum, 10 g of rhizosphere soil collected from crop samples were suspended in 40 ml of sterile RO water and agitated on a rotary shaker at 180 rpm for 1 h to obtain a homogeneous suspension. After transferring the germinated seeds into the plastic boxes, 10 ml of suspension solution was immediately applied to each seed. The plants were cultivated under controlled conditions in a growth chamber and regularly irrigated using a sterilized Murashige and Skoog (MS) medium without vitamins. The isolation process of epiphytic and endophytic SFB was followed as described above, with slight modification. Briefly, the endophytic bacteria were obtained by sterilizing the roots of three-week-old seedlings with 1.5% (V/V) sodium hypochlorite for 30 s. The schematic flow of the experiment, from bacterial isolation to the final plant assay and real-time PCR. The naming method of the isolates is in the Table S2 footnote.

Morphological characterization and tolerance of isolates to abiotic stress conditions

The epiphytic and endophytic isolates were streaked onto the TSA plates and incubated for 24 h at 28 °C to determine their morphological features100. After growing single colonies on plates, the colony color, elevation, shape, and margin were observed under the Stereo microscope (Olympus SZX2–FOF, Tokyo, Japan).

The isolates were elucidated for response to abiotic stress conditions under three pH levels, including pH 9 (alkaline), pH 5 (acidic), and pH 7.2 (control). Salt tolerance was assessed by growing colonies of plates with increasing NaCl concentration; that is, high salinity (4% NaCl), moderate salinity (2% NaCl), and control (0.5% NaCl, from TSA itself). The growth and performance of isolates were compared with their respective controls. Growth was ranked as (i) no growth, (ii) weak growth, (iii) good growth, and (iv) excellent growth, as described by Artigas Ramires et al.101. Based on the morphological features and abiotic stress tolerance for all 243 isolated SFB, a subset of 62 promising isolates, including epiphytic and endophytic isolates, were chosen for further investigation.

Phosphorus (P) and potassium (K) solubilization assay

The determination of P and K solubilization of 62 isolates was tested using NBRIP and Alexandrov medium, respectively103,104. The isolates were grown in TS broth medium for 24 h at 28 °C. Then, 5 µl (10–7 cells/ml) of each culture was added in triplicate to plates containing their respective medium. The plates were subjected to the same incubation conditions as previously mentioned, maintained for 2 days to observe P solubilization and for 1 week to observe K solubilization. This duration allowed for the development of clear halos around the colonies105.

Siderophore and indole-3-acetic acid (IAA) production

To determine siderophore and IAA production, the isolates were first grown in TS broth medium for 24 h at 28 °C. 100 ml of autoclaved Chrome Azurol S (CAS) reagent were added to 900 ml of Luria Bertani (LB) medium after autoclaving. Then, 5 µl (10–7 cells/ml) of fresh culture were dropped in triplicate onto the Luria Bertani (LB) plates and incubated for 2 days at 28 °C for the siderophore production assay. The Burkholderia spp. JSF10 strain was used as a positive control106. After incubation, the orange zone index around the colonies was measured.

To assess IAA production, the 62 SFB isolates were tested following the procedure107. One milliliter of 24-hour bacterial culture was added to 9 ml LB broth containing 0.1% L–tryptophan and incubated for 2 days at 28 °C. Thereafter, 1 ml of the bacterial solution was centrifuged at 10,000 rpm for 10 min. Then, 100 µl of supernatant and 200 µl of Salkowski reagent were added to the 96-well microtiter plate and incubated for 30 min in the dark. After incubation, O.D. 530 nm absorbance was measured using a Thermo Scientific Multiskan SkyHigh Microplate Spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA).

Acetylene reduction assay (ARA)

Nitrogenase activity was measured using the acetylene reduction assay using a gas chromatograph (GC-2014, Shimadzu Corporation, Kyoto, Japan) equipped with a Porapak N column (Agilent Technologies, Santa Clara, USA). One milliliter of fresh TS broth culture was centrifuged at 10,000 rpm for 8 min at 4 °C. The pellet was resuspended in 200 µl of sterilized RO water and added to 10 ml of nitrogen-free (NFb) medium into the 15 ml tubes. The mixture was then incubated for 2 days at 28 °C. After incubation, 1 ml (10% (V/V)) of air from the tubes was replaced with acetylene gas, and the mixture was incubated for an additional 1 h at 28 °C. Thereafter, 1 ml of gas was injected into the GC machine to measure the ethylene production as the standard rate108.

DNA extraction, sequencing of housekeeping genes

DNA was extracted from bacterial cells using the Promega Wizard® HMW DNA Extraction Kit, following the manufacturer’s protocol (Promega, Madison, WI, USA). The purity and concentration of the extracted DNA were measured using a Nanodrop spectrophotometer (Thermo Fisher Scientific, USA), and the DNA was stored at − 30 °C for subsequent molecular analyses.

In addition to the 16 S rRNA gene, the RNA polymerase subunit β (rpoB) and DNA gyrase subunit A (gyrA3) genes were sequenced to assist bacterial identification and taxonomic classification. The PCR reaction was optimized in a total volume of 50 µl, consisting of 25 µl of GoTaq® G2 Green Master Mix (Takara Bio Inc. Japan), 1 µl each of the forward and reverse primers (10 µM), 1 µl of template DNA (215 ng), and 22 µl of sterilized Milli–Q water. The primer sets used were as follows: the universal primers 27 F (5´–AGAGTTTGATCCTGGCTCAG–3´) and 1492R (5´–GGTTACCTTGTTACGACTT–3´) for 16 S rRNA109the rpoB–F (5´–ATCGAAACGCCTGAAGGTCCAAACAT–´3) and rpoB–R (5´– ACACCCTTGTTACCGTGACGACC–3´) for rpoB110and gyrA3–F (5´–GCDGCHGCNATGCGTTAYAC–3´) and gyrA3–R (5´–ACAAGMTCWGCKATTTTT TC–3´) for gyrA3111. The PCR conditions included an initial denaturation at 95 °C for 4 min, followed by 35 cycles at 95 °C (with 30 s, 45 s, 30 s, and 1 min for 16 S rRNA, rpoB, and gyrA3, respectively), annealing temperature was 55 °C (with 30 s, 45 s, 30 s, and 1 min for 16 S rRNA, rpoB, and gyrA3, respectively), and extension at 72 °C for 2 min, and final extension at 72 °C for 7 min. The PCR products were subjected to electrophoresis using a 1.5% (w/v) agarose gel containing 0.5 µg/mL ethidium bromide. The target DNA bands were excised from the gel and purified using a FastGene® agarose gel/PCR extraction kit (Nippon Genetics, Tokyo, Japan) for further analysis. The amplified products were sequenced by Eurofin genomics (Tokyo, Japan), and the results were compared to the respective genes deposited in the GenBank database using the BLAST tool (https://www.ncbi.nlm.nih.gov/balst/, accessed on 10 October 2024). The phylogenetic tree was constructed using the neighbor-joining tree and bootstrap method with 1000 replications through MEGA version 11 software112. The Newick file was used to create the map using tvBOT version 2.6.1113 (https://www.chiplot.online/tvbot.html).

Whole genome sequencing and analysis

The TTREn1 and TC-CSREp1 isolates were cultured overnight in TSA medium at 28 °C from a single colony to extract genomic DNA. The DNA isolation was performed from a 1 ml aliquot using Promega Wizard® HMW DNA Extraction Kit, following the manufacturer’s protocol (Promega, Madison, WI, USA). Genomic DNA quality was assessed with a Qubit 2.0 fluorometer and 1% gel electrophoresis. The quality of the raw paired-end sequence reads was assessed with FastQC (Version 0.11.7; https://www.bioinformatics.babraham.ac.uk/projects/fastqc/). Low-quality (< 20) bases and adapter sequences were trimmed by Trimmomatic software (Version 0.38). The trimmed reads were aligned to the indexed reference genome using bwa-mem (v0.7.17-r1188). SAM files were converted to BAM files and sorted with Picard SortSam (Version 2.18.11, http://broadinstitute.github.io/picard/). The duplicates were calculated using Picard MarkDuplicates. After removing duplicates, single nucleotide variants (SNVs) and short indels were identified using FreeBayes (version 1.3.4). Identified variants were annotated with SnpEff (Version 4.3t) to assess their putative effects on protein translation and identify high-impact mutations. The fastq files were processed and assembly was performed as SPAdes in the Proksee web server (https://proksee.ca/)114. Following this, annotation was carried out using Prokka and RAST (Rapid Annotation using Subsystem Technology). The subsystems were compared with The Seed Viewer (version 2.0). Finally, Proksee was used to create a circular genome map of isolates and Bacillus altitudinis TUAT1 negative control.

Bacterial spore production and plant assay under stress conditions

The 13 high-performing SFB isolates were cultured in Difco Sporulation Medium (DSM) for 1 day, then transferred to fresh DSM and incubated for another 2 days to induce spore production47. The spores were collected by centrifugation at 10,000 rpm for 10 min, washed 2–3 times with sterilized RO water, and resuspended in sterilized RO water. The purity of the spores was confirmed using phase-contrast microscopic observation, with Bacillus altitudinis TUAT1 used as a positive control. The spore concentration of the inoculants was adjusted to 107 cells/ml for use in the plant growth assay. Wheat cv. Yumechikara seeds were sterilized in 70% ethanol for 30 s, followed by 3 min in 3% (V/V) sodium hypochlorite, and rinsed five times with sterilized RO water, pre-germinated for 2 days, and then transferred to plant boxes containing autoclaved Shinano soil (Shinano Baiyoudo Co., Ltd., Nagano, Japan), following the procedure described in the isolation section. The Shinano soil contains approximately 375 mg N kg–1, 750 mg P2O5 kg– 1, and 375 mg K2O kg– 1. The four germinated wheat seeds were planted in each box, with three replications per strain under three different conditions, namely (1) control, (2) heat stress, and (3) drought stress, and (4) heat + drought treatment was added only in the double abiotic stress experiment. After placing the seeds in the plant boxes, 1 ml of bacterial spore suspension (107 spores/ml) was applied to each seed51. The plant boxes were maintained in a controlled growth chamber with specified conditions of 25 °C, a 16 h light (300 µmol−1 m−2 s−1)/8 h dark cycles106 for 2 weeks. After two weeks, the plant boxes were treated under three different conditions. For the control, the plant boxes remained in the same controlled conditions as the initial setup, with no change in temperature or water supply. For heat stress, three boxes for each isolate were transferred to a growth chamber to induce heat stress at 40 °C for 6 h/day for 1 week115. For Drought stress, the plant boxes were kept under the same controlled conditions as the control treatment, but water was withheld for 1 week. Soil moisture in the drought-applied boxes was monitored using a soil moisture meter (pF meter) and maintained between 50 and 60 kPa (2.7–2.8 pF) at the maximum depth of the plant box116. For combined stress of heat and Drought, the plant samples were treated both as described above, at the same time together. After 3-weeks, several growth parameters were measured, including the SPAD value, plant height, number of tillers per plant, number of leaves per plant, shoot and root fresh weight, root length, number of primary roots, and dry weight of shoot and roots.

Bacterial root colonization assay

After harvesting the crops, the roots were gently detached from the soil and washed to remove the surface soil. As described in Sect. "Isolation and characterization of spore-forming bacteria as PGPR", similar procedures were used for isolating epi-and endophytic SFB. Each epiphytic and endophytic solution was divided into two groups for further measurement of vegetative cells and spores. The spores from both the epiphytic and endophytic solutions were heated at 65 °C for one hour to kill the vegetative cells, leaving only the spores for measurement with serial dilution.

Biofilm formation assay

A quantitative biofilm production assay was performed by adding 1 µl of a stationary phase culture of LB broth to 95 µl of LB broth in a 96-well microtiter plate, followed by incubation at 28 °C for 4 days49. The quantification of biofilm production was determined, with a modification involving the use of 0.1% (W/V) crystal violet117. To stain the biofilm, 30 µl of crystal violet was added to the wells and incubated for 30 min at room temperature. The wells were then gently washed twice with 100 µl of 70% (V/V) ethanol, followed by rinsing with RO water. Subsequently, the stained biofilm was solubilized in 100% ethanol, and the absorbance was measured at 550 nm to quantify biofilm production.

Total RNA extraction and real-time PCR analysis

The leaf samples were collected from wheat plants two weeks after sowing. To induce Drought stress, water was withheld from the plants for two days prior to sample collection, resulting in a soil moisture tension of 2.7–2.8 pF at the time of sampling. For heat stress treatment, plants were subjected to a 30-minute heat stress period immediately before sampling. Each treatment included three biological and three technical replications. Subsequently, the collected leaf samples were immediately frozen in liquid nitrogen and ground into a fine powder. The total RNA was extracted from wheat leaves using NucleoSpin® RNA Plant Kit (Takara-bio, Shiga, Japan), following the manufacturer’s instructions. The reverse transcription of RNA into the cDNA was performed using PrimeScript™ RT reagent Kit with gDNA eraser (Perfect RealTime) (Takara Bio). The real-time PCR was performed with LightCycler 96 (Roche Diagnostics, Basel, Switzerland) and KAPA SYBR FAST qPCR Master Mix (2×) (KAPA Biosystems, Wilmington, MA, USA) following the manufacturer’s protocols. The primers used in a real-time PCR reaction for ascorbate peroxidase (APX), catalase (CAT), Δ¹-pyrroline-5-carboxylate synthetase (P5CS) genes are shown in Table 1, with the Actin gene used as the internal standard. Relative expression levels were calculated using the 2-ΔΔCT method118.

Table 1 The wheat-related gene primer sequences, product size, and accession numbers
Full size table

Statistical analysis

Statistical analysis was conducted using R 4.2.2. software (https://cran.r-project.org/bin/windows/base/, accessed: on 25 October 2024). A one-way analysis of variance (ANOVA) was performed on parameters to understand the differences among treatments. The post-hoc Tukey’s test was applied to determine significant differences between treatment means at a significance level of 5%. In addition, the Dunnet test was used to determine the significant difference between the respective treatment and negative control.