Microbial inoculants improve nutrients uptake and yield of durum wheat in calcareous soils under drought stress in the Mediterranean region

ABSTRACT Despite some evidence in greenhouse experiments, the benefits of Bacillus subtilis and Trichoderma asperellum on nutrients uptake by plants under field conditions are unknown. This work aims to study the effect of these microbial inoculants on durum wheat nutrition and yield under Mediterranean conditions where soils may lead to micronutrients and phosphorus deficiencies and where drought may promote low use efficiencies of applied fertilizers. To this end, three experiments were performed in two growing seasons. B. subtilis increased the uptake of nitrogen, phosphorus, and manganese relative to other treatments in the experiment carried out in the first growing season involving two experimental sites. The increased phosphorus uptake was attributed to the increased activity of alkaline phosphatase. In this first season, T. asperellum increased acid phosphatase activity without effect on P uptake and decreased the P to Zn molar ratio in grains, thus improving grain quality for human consumption. In the second season, the effects of microbial inoculants were less evident possibly due to favourable environmental conditions for crops (higher water availability) or phosphate fertilization. Results evidenced that B. subtilis can be effective improving the development and nutrient uptake of durum wheat under more stressful conditions, such as drought stress.


Introduction
Agriculture depends on expensive and non-renewable resources such as fertilizers. The total worldwide demand for NPK fertilizers was greater than 169 million tons in 2020 with an annual increase between 2015 and 2020 of around 2% (FAO 2017). Increasing demand and the dependence on energy and non-renewable mine resources have led to marked increasing trend in prices, much higher than expected some years ago (e.g. Keyzer 2010;Schnitkey et al. 2022). Thus, it is imperative to increase the nutrient use efficiency in agriculture, i.e. the proportion of applied fertilizer or nutrient present in soil that is finally taken by crops (López-Bellido et al. 2005;Congreves et al. 2021), for ensuring agricultural sustainability. In addition, a higher nutrient use efficiency will contribute to a decrease in environmental problems related to nutrients loss to the atmosphere or to water from applied fertilizers (Chien et al. 2009;Recena et al. 2017). This increased efficiency of use can be achieved through the management of soil, crop, and fertilizer (Dobermann 2007). In this regard, the use of rhizosphere microorganisms can enhance the acquisition of nutrients by plants (Marschner et al. 2006;Hinsinger et al. 2011;Zhang et al. 2016). There is evidence on the positive effects of microbial inoculants on the uptake by crops of nitrogen (Kazi et al. 2016;Masood et al. 2020), phosphorus (Khan et al. 2007;Owen et al. 2015;García-López et al. 2021), and micronutrients (de Santiago et al. 2009(de Santiago et al. , 2011(de Santiago et al. , 2013Moreno-Lora et al. 2019), frequently demonstrated in greenhouse experiments and not under real field conditions. Increased nutrient uptake by crops can improve their quality through biofortification, which includes the enrichment of the edible portion of the crop with minerals (Ramesh et al. 2014;Khande et al. 2017;Shaikh and Saraf 2017). This is particularly relevant in staple crops, such as wheat, which represents the base food of around a third of world population (Cakmak and Kutman 2018). In this sense, increasing zinc (Zn) concentration in grains is a relevant issue due to the impact of Zn malnutrition worldwide (Ramesh et al. 2014;Chen et al. 2017;Khande et al. 2017;Shaikh and Saraf 2017;Singh et al. 2017).
Inoculation with microorganisms capable of improving the uptake of nutrients by plants may be even more relevant under environmental conditions prone to nutrient deficiency and low fertilizer efficiency. Zn and iron (Fe) deficiency are relevant agronomic problems in calcareous soils (Ryan et al. 2013); these soils are frequently low in available phosphorus (P) and there is a low use efficiency of applied P fertilizers (Recena et al. 2017). These deficiencies often occur at the same time (Duffner et al. 2012;Hussain et al. 2017), but antagonistic effects can occur, such as Zn and Fe deficiency induced by excessive fertilization with P (Zhang et al. 2016;Moreno-Lora et al. 2019;Xie et al. 2019). In addition to soil, climatic conditions can also explain nutrient deficiencies or low efficiency of nutrient use by crops. Nitrogen (N) use efficiency decreases with restricted water availability in soil (López-Bellido et al. 2005;Ullah et al. 2019), and higher availability of P in soils is required under drought conditions (Matar et al. 1992;Fink et al. 2016). Both calcareous soils and drought are frequent constraints for agriculture in Mediterranean regions (Jacobsen et al. 2012;Ryan et al. 2013).
Not only crops, but also the success in the use of microbial inoculants can be constrained by soil and climatic factors. Soil biological and physicochemical properties affect the growth and proliferation rates of inoculated microorganisms (Owen et al. 2015;Sruthilaxmi and Babu 2017;Gouda et al. 2018). Additionally, the climatic conditions and chemical properties of soils can also modulate their effects (Khan et al. 2007;García-López et al. 2018). All this contributes to the high variability in their performance, generating contradictory results, and consequently, uncertainty about their effectiveness under field conditions. Economic sustainability in agricultural production requires cost-effective measures that involve different benefits for crops. Therefore, microbial inoculants capable of providing different services will have a growing interest in agriculture. In this regard, Bacillus subtilis and Trichoderma asperellum seem to have potential for their use in agriculture. They are known to promote plant development, control plant diseases, and some abiotic stresses such as salinity or drought (Pishchik et al. 2015;Lastochkina et al. 2020;Miljaković et al. 2020;Rivera-Méndez et al. 2020). In addition, they have been shown to improve wheat nutrition and biofortification (García-López et al. 2018;Moreno-Lora et al. 2019). However, these effects have been mostly studied under environmental-controlled conditions (e.g. growing chamber), and their practical application, providing guarantees in transfer from laboratory to farm, needs to be tested. To this end, it is necessary to consider not only the need for field experiments, but also that studies on the effects of microbial inoculants have rarely considered crop nutrition in an integrated way. This means assessing the effects on different nutrients and the equilibrium between them (e.g. antagonism risks) to address the issues of soil fertility decline and food quality for humans (Prasanna et al. 2021). Consequently, this work aimed to assess the effects of these two microorganisms, B. subtilis QST713 and T. asperellum T34, on wheat yield, nutrition, and grain quality in Mediterranean field conditions where soils and climate can restrict the uptake of nutrients by crops, with an integrated perspective in which not only the effect on the uptake of different nutrients, but also on their equilibrium will be evaluated.

Experimental site and soil properties
The experiments were carried out from December to June during two growing seasons (2016-2017 and 2017-2018). The first experiment (Experiment 1) was carried out during the first crop season (2016)(2017) in two different locations in south Spain; one in the experimental farm Tomejil of the Institute for Agricultural and Fisheries Research and Training of Andalusia (IFAPA) near Carmona in the province of Seville (37°24ʹN, 5°35ʹO) and the other in Espejo in the province of Cordoba (37°48ʹN, 4°27ʹO). During the second crop season (2017-2018), the same experiment was replicated at Tomejil (Experiment 2). Experiments 1 and 2 did not involve the application of P. A third experiment (Experiment 3) was carried out in Espejo (37°47ʹN, 4°35ʹO). This later experiment differed in the fertilization strategy to evaluate the effect of inoculated microorganisms on micronutrients when P was applied as fertilizer.
The mean temperature was 14.2°C and 12.8°C in Tomejil and 14.1°C and 12.3°C in Espejo, in the first and second seasons, respectively. Total annual rainfall was 457 and 486 mm in Tomejil, while in Espejo it was 549 and 555 mm in the first and second seasons, respectively ( Figure 1). Soil samples were taken at a depth of 0-30 cm to determine soil properties (Table 1). Both soils were calcareous with clay texture and classified as Chromic Haploxerert according to the Soil Taxonomy (Soil Survey Staff 2014). The phosphate availability indices were around the threshold values for the response to fertilizers, and the extractable Zn by DTPA (diethylenetriaminepentaacetic acid) was below the threshold value for deficiency in three of four cases ).

Experimental design
The experiments were carried out according to a randomized block design with four replications and involving three inoculation treatments: control without inoculation, B. subtilis, and T. asperellum. Experiment 1 involved 2 factors: inoculation treatments and site (Tomejil and Espejo), and Experiments 2 and 3 involved only one factor: inoculation treatments. The plots were different depending on the site and growing season. In Tomejil, the plot surface was 14.4 m 2 (1.2 × 12 m) and 41.8 m 2 (3.8 × 11 m) in the first and second season, respectively. In Espejo, the surface of the plot surface was 34.2 m 2 (3.8 × 9 m) and 33.4 m 2 (2.4 × 14 m 2 ) in the first and the second seasons, respectively. Durum wheat (Triticum durum, cv Almicar) was seeded in the first week of December, at a seed rate of 220 kg ha −1 and harvested in the last week of May. Before sowing, seeds were treated by applying 3 ml of a common sugar dissolution (15% w/v) and 0.6 g of sifted peat (<150 µm, dried at 45°C) to 100 g of seeds, to provide a source C for inoculated microorganisms and a high specific surface, respectively. The inoculum was prepared from commercial products and applied to the seeds at a rate of 1.5 10 13 ha -1 colony forming units (CFU) for B. subtilis QST713 (Serenade Max, Bayer Cropscience) and 3.75 10 12 ha -1 conidia for T. asperellum strain T34 (Biocontrol Technologies SL). The treated seeds were mixed and homogenized before sowing. N fertilizer (urea) was applied before planting and topdressing, as detailed in Table S1 (supplemental material). Additionally, 35 kg ha −1 of P were applied in at pre-plant in Experiment 3 in the form of superphosphate. The chemical weeding was done according to the usual farmer's practice. For dicot weeds, amine salt of clopyralida 35 g l -1 + amine salt of MCPA at 350 g l -1 was used applying 2 l ha -1 of this mix. For monocot weeds, 300 g ha -1 of an herbicide composed of 45 g kg -1 of de mesosulfuron-methyl, 67.5 of de propoxicarbazona-sodium, and 90 of mefenpir-diethyl, was applied.

Sample collection and analysis
During the growing season, plant sampling were done thrice at: tillering (Z2.5), boot-head visible (Z4.5-Z5.0), and harvest (Z9.2) (Zadoks et al. 1974). In the second and third samplings, rhizospheric soil was collected from each plot. The plant material was collected by harvesting wheat plants in two 0.5 m 2 (0.7 x 0.7 m) area in each plot and the rhizospheric soil was collected in c.a. 10-cm depth. The harvested shoots were dried in a forced air oven at 65°C for 48 h to determine dry biomass (DM). In the last sampling (harvest), the shoot and grain biomass were determined separately and the harvest index was calculated. After weighting, dry material was milled to <1 mm, and an aliquot of 0.25 g was mineralized at 550°C for 8 h in a muffle furnace. The ashes were dissolved in 10 ml of 1 M HCl and heated at 100°C for 15 min. In the resulting digest, the concentration of P was determined colorimetrically (Murphy and Riley 1962), and the concentrations of Fe, Mn, Zn and Cu were determined by atomic absorption spectrophotometry on a Solar M device (Thermo, Madrid, Spain). N concentration was determined by the Dumas method using an elemental analyzer (Leco Instruments, Madrid, Spain). The total amount of nutrients in each plant organ was calculated as the product of its concentration in the organ and the dry biomass. The nutrient harvest indices were calculated as the quotient of their total amounts in grain and in the entire aerial part. The nitrogen utilization efficiency (NUtE) was calculated as the ratio of grain yield to total N uptake. The P to Zn molar ratio was calculated as the quotient between their concentrations in the grain. This ratio was deemed an index of grain quality since a high concentration of P in grains, which is mostly in the form of phytate, decreases the absorption of Zn in the intestine (Miller et al. 2007;Bohn et al. 2008;Khoshgoftarmanesh et al. 2017;Gómez-Coronado et al. 2019). The grain protein concentration was calculated as 5.65 times the concentration of N in the grains.
Rhizospheric soil was sampled according to Wang et al. (2009) as the soil retained in roots after mechanical shaking. Finally, a representative sample of the composite rhizospheric soil was taken of each sampled area in each plot. Rhizospheric soil samples were immediately sieved (< 2 mm) and stored at 8°C (no longer than 72 h) (Nazih et al. 2001) to determine biological and biochemical parameters, and a portion of them was air dried and stored for chemical analyses. The population density of B. subtilis and T. asperellum was estimated by dilution plating as described by Tuitert et al. (1998), to confirm the presence of inoculated microorganisms. To disrupt soil aggregates, 5 g of soil and 90 ml of 0.1% Na-pyrophosphate (Na 4 P 2 O 7 · 10 H 2 O) were shaken at 2.5 s -1 for 30 min. From the resulting soil suspension, dilution series were prepared by mixing 1 ml of this suspension and 9 ml of 0.1% water agar. Then, 0.1 ml of dilution was pipetted into 3 replicate plates containing a semiselective culture medium for T. asperellum (Chung and Hoitink 1990), prepared according to Borrero et al. (2012). The plate count of T. asperellum was performed 4 days after plating and was expressed as CFU per gram of soil. The isolation of B. subtilis was performed in selective culture medium, as described by Turner and Backman (1991). Previously, soil suspensions were heated at 80°C (10 min) in a water bath (Tuitert et al. 1998). The CFUs were counted in two steps; first, 48 h after plating, and 5 days later, a second count was performed to determine which of them corresponded to B. subtilis, based on the morphological criteria described by Aguilar et al. (2007). The dehydrogenase activity was determined based on the procedure of Casida et al. (1964), with some modifications: 1 g of soil, 0.01 g of CaCO 3 , 0.25 ml of 3% 2,3,5-triphenyl-tetrazolium chloride (TTC) and 0.875 ml of water were mixed in a falcon tube and incubated at 37°C in darkness for 24 h. Subsequently, the triphenyl formazan (TPF) produced was sequentially extracted by adding ethanol and centrifuging at 1260 g for 10 min to separate it from the soil. The total volume added was 15 ml and its TPF concentration was determined colorimetrically by measuring the absorbance at 485 nm, using a Lambda 35 spectrophotometer (Perkin Elmer, USA). The ß-glucosidase was determined according to Eivazi and Tabatabai (1988), measuring p-nitrophenol (PNP) produced from soil incubation at 37°C for 1 h, using 0.05 M 4-nitrophenyl-ß-D-glucopyranoside as an enzyme substrate. Acid and alkaline phosphatases were determined by soil incubation using 0.05 M 4-nitrophenyl phosphate as a enzyme substrate (Tabatabai and Bremner 1969;Eivazi and Tabatabai 1977).

Statistical analysis
An analysis of variance (ANOVA) with two factors, microbial treatment and location, was performed for Experiment 1 and treatment as the only factor in Experiments 2 and 3. Location was considered a random factor, and treatment a fixed factor. Before ANOVA, the normal distribution of the residues and the homogeneity of the variance were checked with the Shapiro-Wilks and Levene tests, respectively. If either one or both of these criteria were not met, the data were transformed according to a power function to meet these criteria. When the effect of the main factors was significant, the means were compared using the LSD test. If the interaction between factors was significant, an ANOVA was performed for the combination of the two factors and a comparison of means using the test mentioned above. All statistical analyses were performed with Statgraphics Centurion XVI software (Statgraphics, 2013). Data were analysed separately for each experiment since fertilization management and soil properties were different in each year in Espejo.

Experiment 1
In the first two samples, no significant effects of microbial inoculants on nutrients or plant biomass were observed (not shown). A significant concentration of CFUs was found at harvest, confirming that inoculated microorganisms were present in the soil at the end of the growing season. No CFUs were found in the control plots. At harvest, B. subtilis led to the highest dry biomass straw yield (P = 0.0198; Table 2). The total biomass and grain yield also tended to be higher with B. subtilis than with T. asperellum and the control without inoculation (P = 0.0521 and P = 0.0652, respectively). No significant differences were observed between the treatments in harvest index, grain weight, or Zn concentration in the grains (Table 2). B. subtilis increased total extractions of N, P and Mn by crop compared to T. asperellum and the non-inoculated control (Table 3), without significant differences in the concentration of these nutrients in plant tissues (Table 2; Table S2). The concentration of P in the grain and the molar ratio of P to Zn were decreased by T. asperellum, without an effect on the concentration of Zn in the grains ( Table 2). The Fe concentration in the grain tended to be higher with B. subtilis than with the non-inoculated control (P = 0.0943, Table S2). 14.7 ± 0.8 6.6 ± 0.4 6.2 ± 0.4 0.5 ± 0.0 40 ± 0.7 11.1 ± 0.5 3.1 ± 0.1 21.1 ± 0.5 146 ± 2 Control 13.1 ± 0.9 5.9 ± 0.4 5.4 ± 0.4 0.5 ± 0.0 37.8 ± 1.9 11.3 ± 0.5 3.1 ± 0.1 20.6 ± 0.8 148 ± 3 T. asperellum 13.1 ± 0.8 5.8 ± 0.3 5.5 ± 0.4 0.5 ± 0.0 39.7 ± 0.7 11.1 ± 0.5 3 ± 0. The site had a significant effect on the yield and nutritional variables. The biomass yield was higher in Tomejil than in Espejo due to increased straw DM production ( Table 2). The protein content in the grain was 24% higher in Tomejil than in Espejo. Overall, crop nutrient extractions were higher in Tomejil than in Espejo, particularly in the case of Fe (Table 3). With the exception of Zn, the concentrations of nutrients in plant tissues were higher in Tomejil than in Espejo (Table S2). NUtE and N and P harvest indices were higher in Espejo than in Tomejil (Table 3).
In general, there were no differences in enzyme activities in the rhizosphere between locations in the second sampling (Z4.5-5.0), but a significant interaction between the site and the treatments was observed in the case of β-glucosidase (Table 4). This interaction is explained because B. subtilis and T. asperellum increased β-glucosidase activity relative to the control in Tomejil, without significant differences between them (not shown). In Espejo, both inoculants also increased this enzyme activity compared to the control, but in this case, the activity promoted by T. asperellum was significantly higher than that promoted by B. subtilis (not shown). At this stage of growth (Z4.5-5.0), the activity of acid phosphatases was not affected by inoculated microorganisms, while alkaline phosphatase was increased by B. subtilis compared to the other treatments (Table 4). No significant effect on enzyme activity was observed in the rhizospheric soil at harvest (not shown).

Experiment 2
In general, biomass, straw, and grain yield in Experiment 2 in Tomejil during the second season was much higher than in the first season in this location ( Table 5). The presence of both inoculants was confirmed in the soil at harvest. B. subtilis tended to increase straw production relative to T. asperellum (P = 0.0657), but no differences were observed in grain production or grain-related parameters ( Table 5). The total extraction of Zn by crops was also higher with B. subtilis compared to T. asperellum (Table 6). Both microorganisms decreased the extraction of Fe from crops by more than 25% compared to the non-inoculated control (Table 6), tending to decrease its concentration in straw (P = 0.0564) ( Table S3). The concentration of Mn in the straw was significantly lower with inoculated microorganisms (Table S3). Trichoderma asperellum tended to increase the harvest index of P relative to B. subtilis (Table 6).
Acid phosphatase was highest with T. asperellum in the second sampling (Z4.5-5.0; Table 7) or at harvest (not shown), but there were no differences in alkaline phosphatase or dehydrogenase activities between treatments in both samples. β-glucosidase activity was the lowest with B. subtilis in the second sampling (Table 7).

Experiment 3
Inoculation with both microorganisms did not affect biomass or grain production when a soluble source of P was applied (Table 5) and no differences were observed in the extraction of nutrients by crops (Table 6). Both inoculants were present in the soil at harvest. Only a detriment to the Cu concentration in the straw was observed (Table S3). T. asperellum increased acid phosphatase activity compared to B. subtilis and the non-inoculated control at harvest (not shown). There were no significant differences between treatments in alkaline phosphatase, β-glucosidase nor dehydrogenase activities in any of the samplings (Table 7 shows data for the second sampling; at harvest are not shown).

Discussion
Inoculation with B. subtilis improved crop development compared to T. asperellum and the control without inoculation in Experiment 1, as revealed by the significantly higher straw biomass and the trend to increase total aerial biomass and grain yield with B. subtilis compared to the other treatments. This effect may be ascribed to the effect of B. subtilis on phytohormone production and distribution in plants (Arkhipova et al. 2005;Turner and Backman 1991). It may also be related to improved crop nutrition since B. subtilis increased N, P, and Mn uptake relative to other treatments. Although the theoretical N supply (soil + fertilizer N) was considered appropriate for the expected yields, the grain protein content at the Espejo site was not high in Experiment 1. This was ascribed to an inadequate uptake of N by the crop, which in fact was 22 kg of N per ton of grain. Thus, the crop may benefit from an enhanced N uptake promoted by B. subtilis. In agreement with this, N recovery by crop was low taking into account residual N and fertilizer rates applied, and perhaps B. subtilis may contribute to an improved use of soil and fertilizer N. Although B. subtilis increased N uptake, this was not reflected in an increase in NHI and, consequently, in an increase in protein concentration in grains or in an increase in NUtE.
B. subtilis improved P uptake in Experiment 1, which may contribute to better crop development. This agrees with previous evidence obtained under environmental controlled conditions (García-López et al. 2018). The increased P uptake may be ascribed to the increased alkaline phosphatase activity observed with B. subtilis. This increase in phosphatase activity seems to be related to increased microbial activity, as revealed by increased β-glucosidase activity with B. subtilis in both soils. Increasing hydrolytic enzyme activity is a common mechanism of some microorganisms to mobilize P through the hydrolysis of organic P compounds, leading to increased P uptake and plant growth (García-López et al. 2018). However, other mechanisms may affect the mobilization of P. B. subtilis is known to release low molecular weight organic compounds that promote P solubilization from insoluble sources . This latter mechanism involved in P mobilization can also affect the availability of micronutrients to plants (Moreno-Lora et al. 2019; Moreno-Lora and Delgado 2020), which explains the increased Mn uptake with B. subtilis in Experiment 1. The lack of an effect on P nutrition with T. asperellum contradicts previous evidence (Zhao et al. 2017;García-López et al. 2018). In fact, this microorganism increased acid phosphatase activity at both sites. However, this activity may have a low impact on organic P hydrolysis in soils with a basic pH.
Increased P uptake with B. subtilis did not lead to an antagonistic effect on Zn or other nutrients, according to previous evidence (Moreno-Lora et al. 2019). Furthermore, B. subtilis appeared to have positive effects on Fe concentration in grains. The increased uptake of nutrients was not the effect of increased concentrations of these nutrients in plant tissues, in agreement with Nguyen et al. (2019), who reported the same effect for strains of B. velezensis. The benefits of Fe concentration in grains Table 4. Enzyme activities in the second sampling   β-glucosidase activity measured as the amount of p-nitrophenol released from M 4-nitrophenyl-ß-D-glucopyranoside, and Phosphatase activity, alkaline and acid, measured as the amount of p-nitrophenol released from p-nitrophenol phosphate; Dehydrogenase activity measured as the amount of triphenyl-formazan produced; LSD, least significant difference The decreased P content in grains with T. asperellum relative to the control in Experiment 1 was not a negative effect, since this led to a decrease in the molar ratio of P to Zn, which is indicative of improved Zn digestibility (Miller et al. 2007;Khoshgoftarmanesh et al. 2017;Gómez-Coronado et al. 2019). Therefore, this is a positive aspect of grain quality attributed to an effect of the inoculant on P homeostasis in plants rather than to decreased P uptake by the crop. All this reveals that inoculants may affect not only the uptake of nutrients but also its distribution within plants, which may significantly affect the quality of edible parts.
The microbial inoculants increased the overall soil microbial activity as estimated with βglucosidase activity in Experiment 1. This may be attributed to the development of the inoculants themselves or to the effects on soil microbial communities. Although both sites had similar soil properties, the significant interaction in β-glucosidase activity between the site and inoculants may reflect different microorganisms of the native soil and the structure of the community, which may be differently affected by microbial inoculants (Owen et al. 2015;Sruthilaxmi and Babu 2017).
In general, in the experiments performed in the second growing season (Experiments 2 and 3), the effects on plant development and yield were much more limited than in the previous year (Experiment 1). The climatic conditions in the second season were very different from those in the first. The cool and rainy spring (Figure 1) led to higher biomass and grain production in the second season compared to the first, especially at the Tomejil site (Experiment 2), where the total biomass production increased by 60%. Furthermore, no effects on P and N uptake were observed. This reveals that the benefits of inoculants on crop performance and nutrition may be less evident when environmental conditions are favorable for crop development. However, the inoculants also changed biochemical properties in the experiments performed during the second year, revealing changes in microbial activity. However, in Experiment 2, the effect of inoculants on β-glucosidase activity was the opposite to that of Experiment 1, likely revealing that the effects on microbial activity in the soil can be different depending on environmental conditions. In this second year, B. subtilis did not increase alkaline phosphatase activity, which may contribute to explaining the non-significant effect on P uptake by the crop. Thus, favorable environmental conditions for crops or fertilization with P can restrict the P mobilizing capacity of microbial inoculants. In this regard, increased rainfall in Mediterranean environments, such as that observed in spring in the second year (Experiments 2 and 3), leads to an increased soil water content that facilitates P movement to roots through diffusion, allowing an improved use of soil P by crops (Fink et al. 2016). Table 7. Enzyme activities in the second sampling   β-glucosidase activity measured as the amount of p-nitrophenol released from M 4-nitrophenyl-ß-D-glucopyranoside, and Phosphatase activity, alkaline and acid, measured as the amount of p-nitrophenol released from p-nitrophenol phosphate; Dehydrogenase activity measured as the amount of triphenyl-formazan produced; LSD, least significant difference In Experiment 2 performed in the second year, both inoculants reduced Fe uptake and the concentration of Fe and Mn in the whole plant compared to the control. This effect has been previously reported and ascribed to the potential competition between microorganisms and plants for nutrients, which is more evident under conditions of limited micronutrient availability such as this calcareous soil (de Santiago et al. 2011;Marschner et al. 2011). Both T. asperellum and B. subtilis have efficient Fe acquisition mechanisms that allow them to compete positively under conditions of restricted availability (de Santiago et al. 2011(de Santiago et al. , 2013García-López et al. 2018). In Experiment 2, Fe uptake was 39% higher than at the same location in Experiment 1, revealing higher crop needs due to increased development. This may explain why competition for Fe between microorganisms and plants was only observed when crop needs are increased under favorable climatic conditions in soils with restricted Fe availability.

Conclusions
B. subtilis increased straw biomass yield and uptake of N, P, and Mn under field conditions during the first growing season. In this season, this microorganism tended to increase the yield of grain and total biomass. The increased P uptake was attributed to the increased alkaline phosphatase activity. In the first season, T. asperellum decreased the P-to-Zn molar ratio in grains, thus improving grain quality for human consumption. However, in the second season, favorable environmental conditions for crops or fertilization with P restricted the benefits of microbial inoculants on crop nutrition and development. Therefore, this work demonstrated the benefits of B. subtilis inoculation in wheat nutrition and development under drought stress conditions.