Parameter evaluation for developing phosphate-solubilizing Bacillus inoculants

Abstract

Bacterial inoculants have been used in agriculture to improve plant performance. However, laboratory and field requirements must be completed before a candidate can be employed as an inoculant. Therefore, this study aimed to evaluate the parameters for inoculant formulation and the potential of Bacillus subtilis (B70) and B. pumilus (B32) to improve phosphorus availability in maize (Zea mays L.) crops. In vitro experiments assessed the bacterial ability to solubilize and mineralize phosphate, their adherence to roots, and shelf life in cassava starch (CS), carboxymethyl cellulose (CMC), peat, and activated charcoal (AC) stored at 4 °C and room temperature for 6 months. A field experiment evaluated the effectiveness of strains to increase the P availability to plants growing with rock phosphate (RP) and a mixture of RP and triple superphosphate (TS) and their contribution to improving maize yield and P accumulation in grains. The B70 was outstanding in solubilizing RP and phytate mineralization and more stable in carriers and storage conditions than B32. However, root adherence was more noticeable in B32. Among carriers, AC was the most effective for preserving viable cell counts, closely similar to those of the initial inoculum of both strains. Maize productivity using the mixture RPTS was similar for B70 and B32. The best combination was B70 with RP, which improved the maize yield (6532 kg ha⁻¹) and P accumulation in grains (15.95 kg ha⁻¹). Our results indicated that the inoculant formulation with AC carrier and B70 is a feasible strategy for improving phosphorus mobilization in the soil and maize productivity.

Introduction

The phosphorus (P) dynamic and availability are a severe concern in the Brazilian Cerrado Biome, marked by high weathering and P fixation in clays and oxides (Al₂O₃, CaO, Fe₂O₃) [1]. In this scenario, P absorption by plants is reduced, leading farmers to apply high doses of P chemical fertilizers to support crop productivity [2, 3]. However, many factors, such as the high cost of phosphate fertilizers, a decrease of P deposits worldwide, the indiscriminate and long-term use of chemical fertilizers, pollution of soil and water, and a constant world population increase, indicate a need for new approaches focusing on sustainability while maintaining crop productivity [4].

The search for microbial fertilizers has been stimulated for several decades [3, 5]. Biofertilizers are a safe, low-cost technology with great potential to replace or complement chemical fertilizers [6–8]. Phosphate-solubilizing microorganisms (PSM) are a beneficial group capable of mobilizing organic and inorganic P from insoluble compounds in the soil, increasing plant P availability and P uptake [3, 6–11]. Thus, PSMs represent a feasible strategy for improving agroecosystems’ productivity and sustainability. Among the PSM, spore-forming Bacillus species are among the best candidates for cycling inorganic and organic P in the plant rhizosphere [11–13]. In addition, B. megaterium, B. subtilis, B. velezensis, B. pumilus, and Paenibacillus polymyxa have multiple traits that improve plant nutrient acquisition. In addition, many Bacillus produce phytohormones, enzymes, and siderophores that promote plant growth, besides protecting plants against pathogenic and abiotic stressors [14, 15].

Inoculants are commonly formulated by combining a microbial strain and a carrier to provide better shelf life for the formulation [16, 17]. Carriers must preserve microorganism viability (inactive or active state) during transport, storage, and application in the field [18]. Unfavorable temperature and moisture conditions may often decrease viable cells’ recovery rate, affecting the inoculum efficiency [16]. Thus, the carriers must have physicochemical specifications, such as high moisture retention, ease of sterilization and processing, non-toxicity to bacteria, biodegradability and pollutants-free, availability in large quantities, low cost, and reasonable seed adhesion [3, 16]. These conditions are paramount to increasing the shelf life of an inoculant.

Different microbial formulations have been developed using liquid or solid materials as carriers. Solid formulations are made with inorganic or organic compounds, such as peat, agro-industrial wastes, vermiculite, perlite, rock phosphate, calcium sulfate, and polysaccharides, prepared as granules or powder [16, 17]. However, cheaper and more abundant compounds have been successfully tested, such as clay, coal dust, vegetable by-products, and biodegradable polymers [16, 19]. In addition, charcoal and biochar are promising alternatives for inoculant formulation due to their physicochemical properties, such as the excellent capacity for cell coverage, large internal surface area, and pore size, good water retention capacity to increase microbial resistance to desiccation, inertness, and ease manipulation [20]. Thus, the efficacy and long shelf life of inoculants are determined by the precision of the methods employed in their production. Furthermore, these procedures will guarantee the necessary number of viable cells until their application in the field [18].

Finally, but equally important, the choice of strains aiming to develop successful inoculants must consider the microbial adaptability to the environment in which they will be used. For example, in Brazil, the Cerrado biome responds to approximately 70% of agricultural production, and this value continues growing [21]. Among the crops, maize, soybean, sugar cane, and cotton are the most cultivated in the area. The Brazilian Cerrado is the most diverse tropical savanna in the world, characterized by old acidic soils, nutrient-poor, high levels of Al and Fe, and strongly influenced by the water regime. Although regarded as a global biodiversity hotspot [22], the Cerrado biodiversity still needs to be explored [23, 24]. Despite that, the Brazilian Ministry of Agriculture recently authorized the first inoculant for phosphate solubilization formulated with B. megaterium (CNPMS B119) and B. subtilis (CNPMS B2084) isolated from Cerrado Biome [7, 25, 26].

Considering the mentioned aspects, the present study aimed to evaluate the potential of the B70 strain of B. subtilis and B32 of B. pumilus to solubilize and mineralize phosphate and their shelf life in different carriers. We also tested the B70 and B32 inoculants to increase maize productivity and P accumulation in grains in the field fertilized with rock phosphate.

Material and methods

Bacillus strains and culture media

The strains B32 and B70 of B. pumilus and B. subtilis, respectively, were isolated from soil samples collected in the experimental area of the Embrapa Maize and Sorghum Research Center, situated in the Brazilian Cerrado Biome and deposited in the Collection of Multifunctional Microorganisms of Embrapa Maize and Sorghum. The strain B32 was isolated from non-rhizosphere soil and B70 from the maize rhizosphere [7, 27].

This study used Potato Dextrose Agar (PDA) and Nutrient Broth (NB) culture media in two distinct situations. As the strains B32 and B70 form colonies with distinctive border shape, color, and roughness patterns on the PDA medium but not in Nutrient Agar, PDA was used to simplify the identification of the two Bacillus strains and eventual sample contamination by other Bacillus species during carrier survival counts and bacterial recovery from the roots. The NB medium was used for mass production from pure colonies of B32 and B70.

In vitro experiments

Phosphate solubilization and mineralization analysis

The phosphate solubilization and phosphate mineralization analysis were performed by inoculating the two bacterial strains (10⁹ CFU mL⁻¹; OD 540 nm = 1.0) in a culture medium with an inorganic form of P (tricalcium phosphate, Ca-P) and organic P (phytate-P), respectively, according to the methodology described in Nautiyal [28]. Both cultures were incubated at 27 °C and 150 rpm for 12 days. Soluble P and pH of the culture medium were measured every 3 days, according to the methodology described in Murphy and Riley [29]. All treatments were performed in quadruplicate.

Root adhering assessment by Bacillus strains

The root adhering process of B32 and B70 was carried out, as reported by Hozore and Alexander [30]. Root fragments of maize seedlings were aseptically immersed in the bacterial suspension containing approximately 10⁷ cells mL⁻¹ for 20 min. The number of viable cells recovered from the roots was evaluated by their weak or strong adherence to the root. The weak bacterial adhesion was determined by immersion of individual root fragments in 10 mL of saline solution (0.85% NaCl) for 15 s. After, the strong adhesion was assessed by immersing each fragment again in 10 mL of saline solution for 15 min and agitation at 150 rpm. Serial dilutions were prepared (10⁻¹–10⁻⁶) and plated in a PDA medium. Count was performed after 48 h of growth, and the results were registered in percentage. The standard was the number of viable bacteria quantified in the initial inoculum.

Survival test of Bacillus strains in different carriers and temperatures

Pure colonies of B70 and B32, selected in PDA medium, were transferred to the nutrient broth for 72 h, at 29 °C, under shaking at 350 rpm. After 1 week of growth, the bacterial strains were centrifuged for 10 min at 6.000 g. The bacterial suspensions were resuspended in saline solution (0.85% (m/v) NaCl) and adjusted to absorbance to 1 (550 nm), corresponding to 10⁹ CFU mL⁻¹. Later, bacterial suspensions were added to the carriers.

The carriers used were activated carbon powder (AC) (particle size of 0.5 mm), cassava starch (CS) (particle size of 0.1 mm), carboxymethyl cellulose (CMC, Sigma-Aldrich®), and commercial peat (peat). CMC and CS were prepared at 1.5% concentration in water. Afterward, these mixtures were gelled (70 °C) under constant agitation, cooled to 25 °C, adjusted to pH 7, and autoclaved. The inoculation of the Bacillus strains was carried out in sequence, forming a colloidal suspension. AC and peat (85 g) were autoclaved in polyethylene bags. Then, 8.5 mL of each culture was diluted ten times (10⁷ CFU mL⁻¹) and transferred to each bag, creating a solid-like inoculum. All inoculants showed 10⁷ CFU mL⁻¹ or 10⁷ CFU g⁻¹ at the end of this process.

The formulations were stored at room temperature and refrigerated at 4 °C for 6 months, and the viable bacterial count was performed monthly. Bacterial cells were extracted from the carriers by saline solution 0.85% (w/v) and plated in PDA, and the colony count was performed after 2 days of incubation at room temperature. Serial dilutions of 10⁻⁴ to 10⁻⁶ assessed the concentration of viable cells. In our study, the viability of the inoculants after 6 months ranges from 6.7 to 8.8 Log UFC g⁻¹ (Table 1). All treatments were performed in triplicate.

Table 1.

Cell viability and range variation of B. pumilus B32 and B. subtilis B70 in different carriers, storage temperatures, and time

	Cell viability after six months (Log UFC g−1)a				Variation range of viable cells (Log UFC g−1)b
Carriers	Room temperature		Refrigerated (4 °C)		Room temperature		Refrigerated (4 °C)
	B32	B70	B32	B70	B32	B70	B32	B70
Peat	6.7 bB*	8.5 aA	6.8 aA	7.4 bA	0.8	1.4	0.7	2.5
CMC	6.9 bB	8.2 aA	6.7 aB	8.4 aA	1.1	1.2	0.5	0.6
CS	7.9 aA	8.2 aA	6.9 aB	8.8 aA	1.7	2.0	1.8	0.9
AC	7.3 abA	7.2 bA	6.6 aA	7.1 bA	1.0	0.9	1.0	1.2

^aAverage cell recovery at 6 months of storage. CMC carboxymethyl cellulose, CS cassava starch, AC activated charcoal

^bThe interval between the maximum and minimum value evaluated in the storage periods

^*Means followed by the same lowercase letters in the column do not differ significantly by the Tukey test (p ≤ 0.05), and means followed by the same uppercase letters in line do not differ significantly by the Tukey test (p ≤ 0.05) for each storage temperature

Field experiment

Production of inoculants

The tested microorganisms were grown in nutrient broth for one week at 28 °C under agitation. After the incubation period, cultures were centrifuged for 10 min at 6000 rpm. Bacterial suspensions were adjusted to absorbance equal to or greater than one at a wavelength of 550 nm in order to obtain approximately 10⁹ cells mL⁻¹. Subsequently, 8.5 mL of suspensions at 10⁸ cells per gram of charcoal was added to the 85 g of charcoal at a final concentration of approximately 10⁷ cells per gram of seed. Then, the inoculant (bacterium + charcoal) was added to the seeds after involving them with a 5% cassava starch solution. AC was the choice for field tests due to better results in Bacillus stability in storage tests. Both strains showed lower variation in cell counts over the storage period in activated charcoal (AC) (Table 1, Fig. 4). Besides, AC presented some characteristics, such as low cost and easy access in several countries.

Fig. 4.

Cell recovery from spores of B. pumilus B32 and B. subtilis B70 strains in different carriers and storage temperatures over 6 months. CS cassava starch, CMC carboxymethyl cellulose, AC activated charcoal

Maize planting area and experimental design

The field experiment was conducted at the experimental station of Embrapa in Sete Lagoas-MG, Brazil (19° 28′ 4″ S, 44° 15′ 08″ W), during the crop season (November–February). The local climate is classified as Cwa under the Köppen system (high altitude tropical, with mild winters and hot, rainy, and humid summers). The driest months are April until September (means of 21 °C and 4 mm rainfall), and the warmest, rainiest months are October until March (means of 25 °C and 237 mm rainfall). The soil area was classified as dystrophic Red Oxisols (Soil Taxonomy). The chemical and physical characteristics of the soil were measured, as reported by Embrapa [31], with the following values: pH H₂O = 6.3; Al = 0.01; Ca = 3.9; Mg = 0.7; CEC (cation exchange capacity) = 6.8 (cmolc dm⁻³); P = 9.7; K = 80.4 (mg dm⁻³); V% (base saturation) = 71.2%; clay content = 740 g kg⁻¹; OM (organic matter) = 34.7 (g dm⁻³). The experiment was irrigated by sprinkle method in function to soil water balance, using crop evapotranspiration, measured in a class A tank as water consumption [32]. The max/min temperature was 37.4 °C and 15.6 °C (Fig. 1).

Fig. 1.

Climatogram of maximum and minimum monthly temperature (°C) and precipitation (mm) for Sete Lagoas municipality, collected by INMET (National Institute of Meteorology, Brazil). https://portal.inmet.gov.br/

The treatments were arranged in a completely randomized design with three replicates and eight treatments as follows: zero P control, triple superphosphate control (TS control), rock phosphate control (RP control), a mix of rock phosphate and TS (RPTS control), B32 + RP, B70 + RP, B32 + RPTS, and B70 + RPTS. The maize hybrid BRS 1055 was grown for 120 days, corresponding to the end of the cycle.

The experimental plots consisted of four rows with 5 m length, with 0.70 m spacing between rows. Fertilization was performed with 300 kg ha⁻¹ of 20–00-20 (60 kg ha⁻¹ of N and 60 kg ha⁻¹ of K₂O). Urea was side-dressed, with two applications of 150 kg ha⁻¹ each (30 and 45 days after planting). As the source of P, RP and the mixture RPTS were applied according to each treatment at a dose of 100 kg P₂O₅ ha⁻¹, except in the control treatment with zero P. In the TS control, 100% of TS was applied. The mixture RPTS comprised 50% RP and 50% TS. The natural rock phosphate used was Itafós, extracted from Brazilian rock phosphate mines and characterized as the igneous type. Itafós presents a very low P solubility (approximately 20% of P₂O₅ content) and predominance of two main phases: quartz and apatite [33].

The plant dry biomass, grain yield, and total P content in grains were performed according to the methodology described by Silva, 2009 [31]. Plants at the R10 stage (physiological maturity, 120 days after planting) were cut at the base of the stem and dried in a forced air circulation oven at a temperature of 65 °C until constant weight to determine the dry matter. The ears of maize were sampled in the central lines of the plots, threshed in bulk, forming the plot sample. Then, the grain moisture was measured in each sample/plot and weighed, to obtain the total weight of grains per plot. Productivity (kg ha⁻¹) was calculated by correcting the moisture to 13%, considering the weight of grains per plot and the area of the plot. Subsequently, grain samples were dried in an oven with forced air circulation at 65 °C until they reached constant mass. The dried samples were weighed, and a subsample was ground in a Willey mill (model MA020 cyclone type, Marconi, Piracicaba, SP), and the analysis of P concentration (g kg⁻¹) was performed according methodology described in [31]. The total P accumulated in the grains was determined by multiplying the nutrient concentration (g kg⁻¹) by the dry weight of the grains and recorded in kg ha⁻¹. The relative yield (Y1, Y2, Y3, Y4) for each treatment was determined by the average yield of the treatment divided by the control treatments: TS (Y1), zero P (Y2), RP (Y3), and RPTS (Y4). The results were converted to percentages (%) by multiplying the decimal by 100 and ranged from 0 to 100%, with values above 100% indicating yields higher than controls.

Data analyses

All data were submitted to variance analysis (ANOVA) and mean comparison test using the software R v.4.2.2 [34]. For the in vitro experiments, Bonferroni´s (p ≤ 0.05, p ≤ 0.01, and p ≤ 0.001) or Tukey’ s tests (p ≤ 0.05) compared means. For field experiments, means were compared by the Scott-Knott test (p ≤ 0.05 or p ≤ 0.1).

Results

Phosphate solubilization and mineralization by Bacillus strains in vitro

Both Bacillus strains could solubilize Ca-P in vitro, with P release more significant than 110 mg L⁻¹ (Fig. 2A). Furthermore, pH and P release were inversely proportional, with a pH reduction (7 to 5) observed after 2 days of incubation (Fig. 2A). Comparatively, the B70 was more efficient in P solubilization than the B32, with P release values around 167.8 mg L⁻¹ after 12 days of incubation (Fig. 2A). Similar results were obtained for phytate-P mineralization, in which B70 showed more outstanding P release than B32. After 12 days of incubation, the B70 released about 105.3 mg⁻¹ of P, while B32 showed a maximum value of about 34 mg L⁻¹ of P (Fig. 2B).

Fig. 2.

Effects of Bacillus strains (B70 and B32) on soluble P releasing from calcium phosphate (A) and phytate (B). Thick lines indicate the release of P soluble (mg L⁻¹), and dashed lines indicate the pH values

Bacillus strains adhering to maize roots

The difference in the percentage of cells of the initial inoculum adhering to maize roots between B32 (0.15%) and B70 (0.025%) was highly significant. Only about 0.15% (9.9 × 103 CFU cm⁻¹) of B32 adhered to the roots, with 0.05% and 0.1% classified as tightly and weakly adhering, respectively (Fig. 3). Considering B70, only 0.014% of the initial inoculum adhered to the roots (6.2 × 102 CFU cm⁻¹), with 0.01% considered tightly and 0.025% weakly adhered (Fig. 3).

Fig. 3.

Adherence of Bacillus strains to maize roots. Average represents the percentage of adhered cells. The viable cell counts at the initial inoculum were considered 100%. Means followed by the symbol * and ** differ significantly by Bonferroni’s test p ≤ 0.05 and p ≤ 0.01, respectively

Carriers, temperatures, and a survival time test

Cassava starch (CS) was the carrier that showed more viable cell counts for both Bacillus strains at the end of the storage period (Table 1). However, there was significant variation over the period with fluctuations in CFU values (Fig. 4). Both Bacillus strains showed lower variation in cell counts over the storage time in activated charcoal (AC). Considering the storage in AC at room temperature, the variation in cell recovery was close to the initial inoculum (Log 7 CFU g⁻), with about 0.9 and 1.0 Log CFU g⁻¹ of variation in viable cell counts for B70 and B32, respectively (Fig. 4; Table 1). Differently, using CMC showed discrepant results for the two strains and formulations with B32 and CMC presented a final CFU count lower than that of the initial inoculum at both storage temperatures (Table 1). The behavior of formulations in peat was similar to CMC, but oscillations in cell counts occurred only for B70 (Fig. 4). Thus, the AC carrier was chosen for field tests due to storage stability (Fig. 4, Table 1) and other peculiar characteristics, such as low cost and ease of access in many developing countries.

Agronomic efficiency of maize inoculated with Bacillus strains

In general, our results showed that Bacillus inoculants positively affected the yield and P accumulation in the grains compared to the non-inoculated treatments (zero P control and TS control) (Table 2). The treatment with B70 + RP had a higher plant biomass value (7680 kg ha⁻¹) than all treatments with B32 (Table 2). The treatments B70 + RP, RPTS (controls), and B32 + RP increased the maize productivity and differed significantly from the other treatments. Here, an increase in the relative yield of about 17%, 37%, 18%, and 28% was observed in B70 + RPTS compared to the controls TS (Y1), zero P (Y2), RP (Y3) and RPTS (Y4), respectively (Table 2). There was also a higher P accumulation in the grains by combining inoculation and RP and RPTS. In this case, the inoculations of B32 and B70 resulted in an increment of approximately 13% and 23%, respectively, compared with the control (RPTS). The B70 with RP treatment increased the P accumulation in the grains by roughly 19% more than the control. The total P accumulated in the grains had significant values only in the combination of B32 + RPTS (Table 2). The TS treatment showed a higher yield value but did not promote more P accumulation than B70 (RP and RPTS) and B32 (RPTS).

Table 2.

Field experiment adjusted averages of maize productivity and P concentration (g kg⁻¹) and total P accumulated in the grains (kg ha⁻¹) inoculated with two strains of B. pumilus (B32) and B. subtilis (B70)

Treatment*	Maize plant	Maize grain
	Dry biomass (kg ha−1)	P (g kg−1)	P accumulated (kg ha−1)	Productivity (kg ha−1)	Y1a (%)	Y2 (%)	Y3 (%)	Y4 (%)
zero P control	6027c	2.42a*	12.53b*	5165.7b**	85	100	86	94
TS control	10322a	2.23a	13.41b	6053.8b	100	117	101	110
RP control	7116c	2.23a	13.39b	5996.4b	99	116	100	109
B32 + RP	6164c	2.19a	11.45b	5212.9b	86	100	87	94
B70 + RP	7680b	2.47a	15.95a	6532.0a	108	126	109	118
RPTS control	5756c	2.17a	11.98b	5517.9b	91	106	92	100
B32 + RPTS	6285c	2.36a	15.19a	6468.9a	107	125	108	117
B70 + RPTS	6581c	2.33a	16.50a	7085.9a	117	137	118	128

^aY1 = relative grain yield to TS control; Y2 = relative grain yield to zero P control; Y3 = relative grain yield to RP control; Y4 = relative grain yield to RPTS control. RP rock phosphate, TS triple superphosphate; RPTS RP + TS

^*Means in each column followed by the same letter are not significantly different by Scott-Knott’s test (p ≤ 0.05), except productivity (**p ≤ 0.1)

Discussion

Carriers and solubilization experiments

The search for indigenous bacterial strains in the Brazilian Cerrado is increasing to obtain the best potential use to apply as an inoculant in plants of regional interest [7, 12]. Our work determined that B. subtilis B70 and B. pumilus B32 isolated from Brazilian Cerrado soils have at least two characteristics vital in P inoculants, i.e., calcium phosphate solubilization and P-organic mineralization. Furthermore, these characteristics are robust indicators of the multifunctionality of these Bacillus strains with optimistic projections for field applications.

In vitro tests showed that the pH decrease in the culture medium was inversely proportional to the increase of P availability, an expected fact for Bacillus species with the potential to solubilize phosphate [33, 34]. Some Bacillus species are recognized as phosphate solubilizers (B. circulans, B. megaterium, and B. subtilis), with a variety of values of around 29 to 957, 3 mg L⁻¹ of P released [15, 35–38]. Our results showed the capacity of these Bacillus strains to solubilize calcium phosphate for 12 days. In this case, the B70 released 160 mg L⁻¹ of P (pH 5) after incubation time, being more efficient than the B32. This value is highest found for Zhang et al. [39], who considered effective PSMs capable of releasing 57.7 mg L⁻¹ of P. Comparatively, the strain of B. subtilis B2084 studied by Abreu et al. [40] released 120 mg L⁻¹ of P bound to calcium (pH 4.8), which both strains (B70 and B32) showed values above up that B. subtilis B2084. In addition, these results may be assigned to the decrease in pH, which is the primary phosphate solubilization mechanism. A decrease in pH values may often be related to producing and releasing organic acids, protons (H +), and inorganic acids [11, 26, 36, 40]. Thus, the phosphate solubilization capacity may depend on the microorganism species, strain type, soil property, plant species, and others [41]. In this case, each strain may have different metabolism staff that interfered in a greater or lesser release of P, with variations in the amounts and type of enzymes and organic acids released. Generally, organic acids decrease the soil pH and chelate cations and heavy metals. The degree of chelation depends on the type of organic acid, number and proximity of carboxyl groups, type of metal, and pH of the solution [42, 43]. For organic phosphate, differences in the types of phosphatases may determine changes in enzymatic kinetics and phosphate mineralization [40]. Phospho-monoesterases and phytases are the main enzymes in mineralizing P-phytate, the soil’s primary form of organic P [15, 44]. The B70 strain had a remarkable ability to access different P sources (inorganic and organic) and probably other mechanisms that may improve plant growth [5, 26, 45, 46].

Complex traits dictate bacterial attachment to the root surface, and each bacterial species may present different root colonization mechanisms [12]. Colonization traits, such as auto-aggregation and flagella movements, chemotaxis, secretion of enzymes and polysaccharides, growth rate, and competition, determine the bacterial ability to colonize roots. These attributes contribute to the bacterial attachment to the root surface and determine their ability and rhizospheric competence [12, 29, 47]. Our study found that the bacterial strain B70, which did not show the best result for root adherence in vitro, exerted the most significant influence on the increase in maize productivity and the accumulation of phosphorus in grains. Thus, bacterial adherence to the roots is not the primary factor defining a better performance of maize under field conditions. Other factors like competition with the better-adapted soil native microflora and interactions with the rhizospheric microbiome are crucial determinants for the adaptability and performance of a bacterium introduced in the soil.

An inoculant formulation must fulfill essential characteristics for its practical applications in the field. Among them, a formulation needs to provide a suitable environment, physical–chemical protection for a prolonged time of storage, and a critical number of cells of approximately 10⁶–10⁷ cells/plant [17, 48]. These characteristics guarantee a reliable source of living cells to interact with plants and soil microbiomes [16, 17]. In this sense, carriers play a central role in formulations, ensuring the stability and maintaining the viable cell counts in stored formulations. In addition, after application in the soil, carriers must protect the bacteria from fertilizer acidity [49]. Thus, the stability of the formulation is essential during production, storage, transport, and application [50].

In our work, AC preserved the viable cell counts for B. subtilis B70 and B. pumilus B32 in 6 months of storage. Paczkowski and Berryhill [51] found similar results with the non-sporulating Rhizobium, where the inoculum viability in AC was maintained at approximately 10⁶ and 10⁷ cells g⁻¹ after 7 and 12 months of storage, respectively. Finely ground charcoal may be a feasible alternative carrier to peat besides maintaining the number of viable cells [52, 53]. Furthermore, AC may be an easily accessible and low-cost option for Bacillus-based inoculant formulation, compared with peat and CMC. Although many studies have shown that peat is one of the leading carriers for inoculants, with outstanding results for several bacteria, especially rhizobia [16, 54, 55], the use of peat for inoculant production is limited due to its high cost and variations in the physicochemical properties. In addition, peat is a non-renewable resource (fossil origin) extracted from fragile environments [56]. Also, peat is rarely available in some countries and not readily obtainable in most developing countries, particularly in the tropics (17, 49]. Furthermore, considering the vast number of publications about carriers for bacterial formulations, they are almost contradictory, and a universal carrier for an ideal formulation does not exist. Also, there has yet to be a consensus about the best carrier for inoculant formulation for Bacillus species [57]. Thus, what is best for one species of bacteria may not be for others.

In non-spore-forming bacteria, survival is often challenged by nutrient-depleted conditions because, under starvation, they enter the stationary phase, i.e., although the bacterial growth ceases, the cells remain metabolically active [58]. In Escherichia coli, a rejuvenation from the stationary phase is heterogeneous and age-dependent since respiration during the stationary phase induces intracellular damage, leading to delayed regrowth [58]. On the contrary, spore-forming bacteria (Bacillus and Clostridium species) withstand extreme starvation conditions since spores are dormant bodies that do not have an active metabolism [59]. Thus, the bacterium’s adaptability to different stressors determines the appropriate substrate for developing shelf-life formulations. In a 500-year microbial experiment for determining long-term spore survival, data from the first 2 years of Bacillus subtilis storage show no significant decrease in spore viability [60]. However, some factors, such as vacuum and high NaCl concentration, negatively affected spore viability in short-term storage experiments [60]. Unlike B70, the storage at 4 °C reduced the viability of B32 spores. Many factors, such as temperature variation, oxygen, moisture, microbial contamination, and light, may negatively affect the shelf life of the inoculants [61]. Thus, there is no universal carrier or storage temperature since choosing the proper formulation is a specific process for each strain. Thus, the formulation of inoculants with a reliable and consistent effect under field conditions is still a bottleneck for their broader use [19]. In this study, we tested several carriers for Bacillus to determine which was best suited to our soil conditions and tropical climate, aiming to establish the parameters for inoculant production on an industrial scale.

Another critical aspect of defining the commercial success of an inoculant is related to the low cost and easy access to the products that will be used as carriers. In non-sporulating bacterial inoculants, access to glucose is critical for the initial bacterial growth (Lag phase) since it may ensure survival and an increased cell number over time [48, 62]. In these species, the increase of the initial number of cells may not be economically viable due to a rapid deterioration of the product [53].

Field experiment

Many agricultural practices and management are used for maize crops, such as high mechanization and high amounts of fertilizer application or even low input practices, adopted by small and medium-sized farmers in many developing countries [26]. In this sense, the positive results of bacterial strains B32 and B70 in the two P fertilization conditions (RP or RPTS) are essential because these bacteria may colonize plants under different P sources.

However, it is essential to note that using phosphate-solubilizing inoculants does not eliminate the need for phosphate fertilizers. First, the amount of organic P in the soil could be insufficient to supply the growing plants’ needs, which is particularly important in the Brazilian Cerrado soils. Second, seedlings and flowering are critical stages of plant growth that require quick and high amounts of P. In addition, the microbial inoculants have yet to be established at the seedling stage. Thus, crop seedlings must be supplied with essential nutrients [63].

In the field tests, B70 promoted higher grain yield and P content by improving plants’ use of RP and RPTS. The increase in grain yield due to B70 and B32 inoculants was 17% and 7% higher than the control. These results corroborate Sousa et al. [26], which observed an increase of 16% in grain P of maize inoculated with the B70 strain compared with non-P and non-inoculation treatment. Despite the lowest root adherence in in vitro, the B70 strain increased plant biomass and maize productivity under both P sources (RP and RPTS). A possible explanation for this finding is that microbes in the field face multiple challenges for space and resources [64]. In this case, the B70 strain may be better adapted for colonizing the maize rhizosphere and competing with the endogenous microbial community than B32 [64], possessing other colonization characteristics and adherence to the roots that could be explored in future works to unravel these mechanisms. B. subtilis is one of the most studied plant growth-promoting rhizobacteria (PGPR) with broad applications in food industries, medicine, and agriculture. Different metabolites produced by B. subtilis promote plant growth in the agricultural field. This species also confers biotic and abiotic stress tolerance by inducing systemic resistance in plants (ISR) and lipopeptide production with antifungal activity [13, 14, 34, 65]. Thus, multifunctional mechanisms in both Bacillus strains, especially B70, may contribute to plant growth in weathered soils with low P availability. In Brazilian Cerrado soils, most applied P fertilizer forms complex with ions Fe, Al, and Ca [66]. In this case, rock phosphate may be an alternative P source for crops because it dissolves slowly, and nutrients in the rock are released gradually. However, in some soils, the dissolution rate might be faster to support healthy plant growth [10, 66]. In contrast, when soluble phosphate fertilizers are applied to highly weathered soil, the P recovery rates by plants may be low due to phosphate-specific adsorption mechanisms [66]. The combined use of RP, TS, and Bacillus may reduce soil P adsorption through a moderate P release and mobilization. Our work demonstrated the positive effects of these combinations to increase the maize yield and P accumulation in grains, due to the microbial improvement in soil P cycling and P uptake by plants. Thus, multifunctional mechanisms in both Bacillus strains, especially B70, may contribute to plant growth in weathered soils with low P availability. Thus, it is a viable alternative for phosphate correction in agriculture systems, mainly in the long term. Plant growth-promoting rhizobacteria (PGPR) may provide P for plants through solubilization and mineralization.

However, more than the improvement in P availability is needed to explain our results. This statement is because the members of the genus Bacillus have several beneficial traits, such as a symbiotic lifestyle and metabolic versatility that promote plant growth and tolerance to biotic and abiotic stresses [16]. These attributes make Bacillus species a powerful alternative for agrobiotechnological applications [10]. The positive effects of B70 and B32 on maize yield and P accumulation in grains can be attributed to a myriad of beneficial molecules contributing to healthy plant growth, such as nitrogen fixation, production of phytohormones, hydrolytic enzymes, antimicrobial compounds, and molecules inducing the plant defense mechanisms. The genome sequencing of these strains and genome mining will allow the development of further strategies for their better use as plant growth-promoting and crop production improvements.

In conclusion, this study demonstrated that phosphate solubilization and mineralization using the strain B70 of Bacillus subtilis might represent an important alternative to improve maize productivity and P accumulation in grains, and the use of activated charcoal as the carrier in formulations for bacterial inoculants may ensure better durability for the final product and maize crop productivity.

Author contribution

All authors contributed to the study’s conception and design. D. B., B. B. M., J. E. F. F., and C. A. d. O-P. wrote the manuscript; C. A. d. O-P., F. C. d. S., and I. E. M. designed the experiments. C. A. d. O-P. and J. E. F. F. revised the manuscript. B. B. M. conducted the laboratory assays and field tests. D. B. and C. A. d. S. analyzed all the data. All authors read and approved the final version of the manuscript.

Funding

This work was supported by the Empresa Brasileira de Pesquisa Agropecuária (Embrapa), FAPEMIG, FINEP, and Brazilian Council for Scientific and Technological Development (CNPq)/ INCT-Plant-Growth Promoting Microorganisms for Agricultural Sustainability and Environmental Responsibility (Grant n. 465133/20142, Fundação Araucária-STI, Capes). FINEP/CT funded this study—AGRO/FNDCT (Cooperation Agreement n. 01.22.0080.00, Ref. 1219/21).

Declarations

Conflict of interest

The authors declare no competing interests.