Beneficial Microorganisms Affect Soil Microbiological Activity and Corn Yield under Deficit Irrigation

Abstract

Water scarcity is one of the main factors that decrease the growth and productivity of corn, since it negatively affects gas exchange and the general metabolism of the crop. The use of beneficial microorganisms (BM) has been considered a potential attenuator of water stress. This study aimed to evaluate the effect of BM and water deficit on growth, gas exchange, grain yield, and soil microbial activity. A field experiment was carried out, in which the treatments were composed of a 2 × 4 factorial scheme, corresponding to two irrigation levels (100% of ETc and 50% of ETc) and to four treatments (T) referring to the soil inoculation with BM (C: control; T1: Bacillus amyloliquefaciens + Azospirillum brasiliense; T2: B. subtilis; and T3: A. brasiliense). The evaluations were carried out in the flowering phase (plant growth, gas exchange, and foliar nitrogen content) and at the end of the plant cycle (grains yield, mineral nitrogen, and microbiological activity). The 50% reduction in irrigation depth severely restricted corn growth and gas exchange and decreased the grain yield by 38%. The water deficit increased the protein content in the grains and the concentration of mineral nitrogen in the soil when the plants were inoculated with BM. Under water stress, inoculation with BM increased corn productivity by 35% and increased soil microbial activity. The inoculation of plants with BM, either in combination (Bacillus amyloliquefaciens + A. brasiliense) or alone (B. subtilis), attenuated the adverse effects of water deficit in maize.

1. Introduction

The water deficit affects not only the arid and semi-arid regions of the world but also areas located in countries with a humid tropical climate, such as Brazil [1]. In these areas, prolonged water scarcity periods can reduce the productivity of important crops, such as corn [2]. Water stress negatively affects plants’ physiological and biochemical processes that restrict growth, development, and productivity [3,4]. Drought causes a decrease in cell turgor, which is essential for proper cellular metabolisms, such as photosynthesis, enzyme activity, and nutrient uptake [4,5].

Corn (Zea mays L.) is one of the most important crops in the world due to its multiplicity of uses, especially as raw material for the food industry. However, in areas where water scarcity is more pronounced, as in many developing countries and the semi-arid region of Brazil, the use of corn in human food as a main diet component is expressive [6]. Thus, in these areas, water deficit is one factor that threatens food security. Therefore, it is necessary to find strategies to mitigate the adverse effects of water deficit in essential crops such as corn.

Previous research has demonstrated that beneficial microorganisms (BM) (also called plant growth-promoting bacteria) have the potential to attenuate environmental stresses in plants, such as water deficit [7,8,9,10]. BM in the soil can induce plants to produce osmoregulant substances such as organic acids, amino acids, and soluble sugars and, thus, act synergistically, contributing to drought tolerance [11,12]. These microorganisms can produce auxins such as indole acetic acid, increasing the length of plant roots and, thus, leading to the greater uptake of water and nutrients from the soil [8]. In this sense, it has been observed that the inoculation of corn with Bacillus amyloliquefaciens increased the nutrient uptake and promoted growth mechanisms in plants, increasing the concentration of amino acids such as tryptophan, isoleucine, alanine, valine, and tyrosine and sugars such as fructose and glucose [11]. Likewise, Lima and collaborators [13] observed that the inoculation of corn with B. subtilis increased the leaf water content and stomatal regulation without impairing the photosynthetic rates. In addition to the effects on plants, BM can affect the biological activity of the soil, favor specific populations of microorganisms that act in key soil processes such as mineralization and the nitrification of soil nitrogen, and thus reduce the losses of this nutrient through leaching and volatilization processes [14,15].

The use of BM in agriculture is advantageous, because it is an environmentally friendly technology, since it can increase crop productivity and soil fertility without exerting any toxic effect on the environment [16,17]. The mechanisms of action of BM on plants such as nitrogen fixation, phosphate solubilization, the synthesis of phytohormones (especially IAA), and osmoregulant substances (such as amino sugars), especially in a controlled environment, have been demonstrated in some studies [9,15,18]. However, it is necessary to expand these studies, especially under field conditions, for a better understanding of the effects of BM on the physiological aspects of plants and on soil microbiological activity, which could represent modes of action of BM as attenuating water deficit in plants. In this sense, we hypothesized that BM improves the gas exchange of corn under a water deficit, promotes an increase in soil biological activity, and thus attenuates the adverse effects of irrigation deficit, increasing growth and productivity.

This study aimed to evaluate the effect of beneficial bacteria on gas exchange, growth, production, and the protein content in corn grains and soil microbiological activity under a water deficit.

2. Materials and Methods

2.1. Experimental Site Description

This research was carried out from July to November 2021 in field conditions at the experimental farm belonging to the Center for Agro-Food Science and Technology—CCTA of UFCG, located in the Sertão Paraibano mesoregion in the municipality of São Domingos—Paraíba. According to the Köppen classification adapted for Brazil, the climate is tropical semi-arid (Bsh), with an average annual temperature above 26.7 °C and an average yearly rainfall of 872 mm [19]. Climatological data for the experimental period (Figure 1) were collected using the AGRITEMPO agrometeorological monitoring system [20]. During the experimental period, no rainfall was recorded in the area.

Before installing the experiment, a composite sample of soil from the area was obtained from 15 collection points, randomly obtained in the layer from 0 to 20 cm. The soil of the experimental site, classified as Planosol [21], was analyzed for its chemical and physical attributes (

Table 1. Physical and chemical attributes of the soil samples used in the experiment.

Chemicals Attributes	Value	Physical Attributes	Value
pH (CaCl2)	6.20	total sand (g kg−1)	444
P (mg kg−1)	291	silt (g kg−1)	353
K+ (cmolc dm−3)	1.19	clay (g kg−1)	203
Na+ (cmolc dm−3)	0.54	BD (g cm−3)	1.36
Ca2+ (cmolc dm−3)	5.80	PD (g cm−3)	2.59
Mg2+ (cmolc dm−3)	3.40	TP (m3 m-3)	0.47
H + Al (cmolc dm−3)	2.30	FC (%)	12.87
SOM (g kg−1)	6.40	PWP (%)	5.29
V (%)	83.0	AW (%)	7.58

SOM: soil organic matter, V: saturation of bases, BD: bulk density, PD: soil particle density, TP: total porosity, permanent wilting point (PWP), and AWC: available water content.

) at the Laboratory of Soils and Plant Nutrition of CCTA/UFCG, according to the methodology described by Emprapa [22].

2.2. Treatments and Experimental Design

A field experiment was carried out, in which the treatments were composed of a 2 × 4 factorial scheme, corresponding to two irrigation levels (100% of ETc and 50% of ETc) and to four treatments (T) referring to the application of BM (C: control; T1: Bacillus amyloliquefaciens BV 03 + Azospirillum brasiliense; T2: B. subtilis BV-09; and T3: A. brasiliense). A randomized block design was used (Appendix A) with five replications, making a total of 40 subplots in the experiment as a whole.

2.3. Soil Tillage and Experiment Installation Details

Soil tillage consisted of carrying out two harrowing, the first being heavy and the second light, before sowing, aiming to break down and level the soil; after which, the planting furrows would be opened. The plants were grown at a spacing of 0.2 m × 1.0 m, corresponding to an estimated stand of 5 plants per linear meter. The spacing between plots was 1.0 m, while the blocks were separated by 2.0 m. The plots consisted of five cultivation lines measuring 4.0 m × 4.0 m. The portion of plot for data collection consisted of the three central lines measuring 2.0 m in length, resulting in an estimated total of 30 plants (Appendix A).

Sowing was carried out on 7 September 2021 using seeds of the hybrid corn cultivar K9555VIP3, which was chosen because it has a short cycle (about 110 days), high productivity, and resistance to the fall armyworm. The planting was carried out manually one day after planting fertilization (7 September 2021), sowing one seed per hole with spacings of 0.2 m.

The fertilization was carried out according to the Fertilization Recommendation Manual for the State of Pernambuco [23] based on the interpretation of the soil analysis of the experimental area. In the planting, 30 kg of N, 60 kg of P₂O₅, 45 kg of K₂O, and 24 kg of S in the form of magnesium sulfate were applied. Supplementary fertilization with nitrogen and potassium was applied at the rate of 50 kg of N per ha in the V4 stage (plants with four expanded leaves) and 50 kg of N, 45 kg of K₂O, 1 kg of B, and 2 kg of Zn in the V6 stage (plants with six expanded leaves). The nutrient sources used were urea (45% of N), potassium chloride (58% of K₂O), simple superphosphate (18% of P₂O₅), magnesium sulfate (9% of Mg and 12% of S), boric acid (17% B), and zinc sulfate (20% Zn). The N and K supplementary fertilizations were carried out via fertigation using the irrigation system.

2.4. Treatments Composition Composition and Application of Treatments

The inoculation with BM was carried out exclusively via fertigation using suspensions of microorganisms in a proportion of 4 L ha⁻¹. The T1 treatment (Bacillus amyloliquefaciens + A. brasiliense) consisted of 1.0 L ha⁻¹ of the commercial product containing 3.0 × 10⁹ CFU/mL of B. amyloliquefaciens and 3.0 L ha⁻¹ of the commercial product containing 1.0 × 10⁹ CFU/mL of A. Brasiliense. Treatment T2 corresponded to the commercial product containing 1.0 × 10⁸ CFU/mL of Bacillus subtilis BV-09, and treatment T3 corresponded to the commercial product containing 1.0 × 10⁹ CFU/mL of A. Brasiliense applied at a dose of 4 L ha⁻¹. Treatments were performed seven days after seedling emergence. The doses of each product (or combination) were diluted in water at a rate of 4 L per 2500 L of water and were applied using a manual sprayer on the root zone of the plants. The same procedure was adopted in the control treatment but used only water without any product.

2.5. Irrigation Management

The plants were drip irrigated, with drippers spaced 0.20 m. After the emergence and standardization of the number of plants per plot, the plants were irrigated following the different water regimes. The total irrigation volume required (TIR) of each irrigation level was obtained by the following Equation (1) [24]:

$$TIR=\frac{\left(FC-PWP\right)\times Z\times BD\times f}{10\times Ea}$$

(1)

where TIR corresponded to the initial total water depth to be applied in mm, FC was the soil moisture corresponding to the field capacity in %, PWP was the soil moisture corresponding to the wilting point in %, Z was the effective corn root system depth (30 cm), BD was the soil bulk density in g cm⁻³, f was the water availability factor for maize (0.5), and Ea was the application efficiency (0.90). During the experiment, meteorological data were obtained from the automatic meteorological station in the municipality of São Gonçalo, Paraíba, as it is the closest to the experiment site, through the website [25].

The control of the volume of water corresponding to each water regime was performed daily at a standardized time according to the ratio of the flow rate of the drippers by the time to reach the proportions of crop evapotranspiration (ETc). As the time interval for each volume of the respective water regime was reached, successive disconnections of the drip strips corresponding to each irrigation level were performed. The irrigation depth corresponding to 100% of ETc was calculated according to Jesen’s equation [26] using the following expression: ETc = Kc × ETo. ETc is the crop evapotranspiration in mm day⁻¹, ETo is the reference evaporation in mm day⁻¹, and Kc is the crop coefficient. The Kc values adopted for corn (initial stage: 0.13, vegetative stage: 0.55, flowering: 1.00, reproductive stage: 1.20, and final stage: 0.90) as a function of its phenological phases were based on [27]. The daily supply of irrigation depths was carried out through the irrigation time considering the characteristics of the cultivation system and the irrigation system according to Equation (2):

$$Ti=\frac{Eto\times Kc\times A}{Ea\times n\times q}$$

(2)

where Ti is the irrigation time in hours, ETo is the reference evaporation in mm day⁻¹, A is the area occupied by a plant in m², n is the number of drippers per plant, q is the dripper flow in L h⁻¹, and Ea is the application efficiency (0.90). Water application uniformity tests were determined according to the Christiansen Uniformity Coefficient (CUC) evaluation methodology proposed by Christiansen [28].

2.6. Phytosanitary Control

Weed control was done mechanically and manually using simple tools such as hand hoes. Regarding insects, there was no need to apply any product for pest control due to the low incidence in the area.

2.7. Assessment of Growth, Gas Exchange, and Leaf Nitrogen Content

In the female inflorescence stage gas exchange, the culm diameter, plant height, and leaf area index (LAI) were determined. On this occasion, photosynthesis (A) (µmol CO₂ m⁻² s⁻¹), stomatal conductance (gs) (mol m⁻² s⁻¹), the transpiration rate (E) (mmol H₂O m⁻² s⁻¹), and intercellular CO₂ concentration (Ci) (mol CO₂ m⁻² s⁻¹), with an infrared gas analyzer (IRGA) (LCpro Analytical Development, Kings Lynn, UK) with a constant light source of 2000 µmol of photons and ambient CO₂ concentration, were evaluated. Readings were taken from 7:00 a.m. to 9:00 a.m. using the diagnostic sheet (leaf opposite to the cob) [29]. The LAI was estimated using a photosynthetically active radiation meter (AccuPAR model LP-80). Readings were taken from 8:00 a.m. to 11:00 a.m. In each plot, five readings were performed below the leaves close to the ground, corresponding to the four cardinal points of the plot. The diagnostic leaves of five plants of each useful plot were collected on the same day to determine the total nitrogen (N) content. The leaves were dried in a forced circulation oven at 60–65 °C and then ground in a Willey-type knife mill. Then, sulfuric digestion was performed, followed by distillation and titration [29].

2.8. Mineral Nitrogen Contents and Soil Microbiological Activity

In each useful plot, four soil subsamples were collected close to the planting line under the influence of the corn rhizosphere in the layer from 0 to 20 cm. After homogenization, 20 g of the composite samples were immediately placed in plastic flasks containing 100 mL of a 1.0 mol L⁻¹ KCl (potassium chloride) solution, then placed in a refrigerator [30]. After defrosting, shaking at 180 rpm, and filtering, 20 mL of the extract were placed in distillation tubes to determine the NH₄⁺ (ammonium) and NO₃⁻ (nitrate) contents by the Kjeldahl nitrogen micro still method [30]. Initially, 0.2 g of calcined MgO was added to each distillation tube. After distillation, the ammonium fractions were obtained by titration with HCl (hydrochloric acid) 0.07143 mol L⁻¹ after being collected in indicators with boric acid. Nitric nitrogen was determined using the same extract (same tube) used for ammonium distillation and then adding 0.2 g of Devarda’s alloy and sending it to a new distillation. Then, it was titrated with the same acid used for ammonium. Mineral nitrogen was calculated by adding NH₄⁺ + NO₃⁻.

The remaining soil composite samples from each plot were frozen to further evaluate the soil respiration rate, microbial biomass carbon, and metabolic quotient. Soil microbial respiration was measured by capturing the C-CO₂ produced in the soil by NaOH (sodium hydroxide) in a hermetically sealed environment [31]. Biomass carbon was evaluated using the irradiation/extraction method, which has, as its basic principle, the elimination of microorganisms by electromagnetic radiation from a microwave oven [32,33,34]. Each soil sample was subdivided into irradiated and nonirradiated samples. After irradiation, the samples were transferred to a 125 mL Erlenmeyer flask and identified according to the procedure. Then, 80 mL of K₂SO₄ (potassium sulfate) extracting solution was added. The samples were shaken for 30 min in a horizontal shaker at 150 rpm and then kept at rest for 30 min. Subsequently, the samples were filtered in recipients with filter paper. The carbon present in the extracts was determined by pipetting 10 mL of the filtered extract into a 125 mL Erlenmeyer flask, where 2 mL of 0.066 mol L⁻¹ K₂Cr₂O₇ (potassium dichromate) solution was added. Then, 10 mL of H₂SO₄ (sulfuric acid) was added. After the samples were lowered to room temperature, 50 mL of distilled water was added to each Erlenmeyer flask, and the titration was performed by adding three drops of ferroin as an indicator and ammoniacal ferrous sulfate 0.03 mol L⁻¹ ammoniacal ferrous sulfate.

2.9. Grain Yield and Protein Content

The ears were harvested manually five to eight days after the grains reached physiological maturity, with a moisture level below 15% (wet basis). The harvest was carried out using the useful plot, collecting 10 ears to assess the characteristics of the ears, the thousand-grain weight, and the grain yield. Based on the grain yield in the 16 m² plots, by extrapolation to 10,000 m², the yields in kg ha⁻¹ were calculated. The yield data were corrected to 13% moisture (wet basis). The moisture content of the grains was evaluated using the oven method at 105 °C for 24 h [35]. The protein contents in the grains were determined by the Kjeldahl micro still method [30] with sulfuric extraction and subsequent distillation. For this, the total nitrogen (N) contents were initially determined in 0.5 g of ground grains and later converted into crude protein by a multiplication factor of 6.25 [36]. The protein yield was obtained by multiplying the protein content by the grain yield and adjusting the units.

2.10. Statistical Analysis

Data referring to the measured variables were submitted to an analysis of variance (ANOVA) and Tukey’s test at the 0.05 probability level using SISVAR^® statistical software version 5.6 [37]. A multivariate analysis of the results was performed using the principal components analysis (PCA), synthesizing the amount of relevant information contained in the original data set in a smaller number of dimensions [38]. From the reduction of the dimensions, the original data of the variables of each component were submitted to a multivariate analysis of variance (MANOVA) using Hotelling [39] at the 0.05 probability level for the irrigation levels (I) and the related treatments’ beneficial microorganisms (BM), as well as for the I × BM interaction. Only variables with a correlation coefficient greater than or equal to 0.6 were kept in each principal component (PC) [40]. For the multivariate statistical analysis, software Statistica v. 7.0 was used [41].

3. Results

3.1. Plant Growth and Leaf Nitrogen Content

According to the analysis of variance (

Table 2. Summary of the analysis of variance for the leaf area index (LAI), plant height (PH), stem diameter (CD), and nitrogen content in leaves (N leaf).

Source of Variance	DF	PH	LAI	CD	Leaf N
	Mean Square
Irrigation levels (I)	1	45,611.032 **	1.917 **	27.400 **	290.005 **
Blocks (replications)	4	490.803	0.112	22.98	20.256
Microorganisms (BM)	3	82.916 ns	0.129 *	0.448 ns	13.0878 **
I × BM	3	605.121 **	0.074 *	1.493 ns	1.898 ns
Error	28	131.743	0.034	1.066	2.488
CV (%)	-	6.89	9.37	3.94	7.20

CV: coefficient of variation. ** p < 0.01 and * p < 0.05 by F test. ^ns Stands for nonsignificant data at the 0.05 probability level. DF: degrees of freedom.

), the irrigation levels influenced the plant height (PH), leaf area index (LAI), stem diameter (CD), and leaf nitrogen content (leaf N), while the treatments related to inoculation with plant growth-promoting bacteria (BM) influenced only the index and the variables LAI and leaf N (Table 2). There were interactions between irrigation levels and BM treatments only for the LAI and PH. The ETc irrigation level of 50% decreased the PH, LAI, CD, and leaf N values by approximately 20%, 32%, 6%, and 22%, respectively (Figure 2). Under water deficit (50% of ETc), treatment C (without inoculation) provided a higher PH value compared to treatment T3 (Azospirillum brasiliense). Still, both did not differ between treatments T1 (Bacillus amyloliquefaciens + A. brasiliense) and T2 (B. subtilis). Under full irrigation (100% of ETc), the treatments with BM did not change the plant height. The T1 treatment provided a higher LAI value than the control treatment. However, these treatments did not differ from the others. Treatment T2 increased the leaf N content when compared to the control and treatment T1 but not when compared to treatment T3.

3.2. Gas Exchange

The irrigation levels influenced the photosynthetic rate (A), stomatal conductance (Gs), transpiration rate (E), and intercellular concentration of CO₂ (Ci) (

Table 3. Summary of the analysis of variance for the CO₂ assimilation rate (A), internal CO₂ concentration (Ci), stomatal conductance (gs), and transpiration rate (E).

Source of Variance	DF	A	Ci	Gs	E
	Mean Square
Irrigation levels (I)	1	1192.326 **	603.049 *	0.509 **	16.593 **
Blocks (replications)	4	405.009	56.204	0.165	25.570
Microorganisms (BM)	3	71.679 ns	567.319 **	0.039 ns	6.732 **
I × BM	3	63.089 ns	191.019 ns	0.029 ns	0.778 **
Error	28	61.239	85.196	0.021	0.814
CV (%)	-	17.57	15.11	32.90	12.38

CV: coefficient of variation. ** Significant (p < 0.01) and * Significant (p < 0.05) by F test. ^ns Stands for nonsignificant data at the 5% probability level. DF: degrees of freedom.

). The treatments related to BM only influenced the values of Ci and E. There was an interaction between the irrigation levels and the treatments with BM only for the transpiration rate, which was also affected by the treatments with BM. On average, a water deficit reduced the photosynthetic rate, stomatal conductance, and transpiration rate by 22%, 41%, and 16%, respectively (Figure 3). Treatment T3 provided the lowest value of Ci without differing from treatment T2. Treatments C and T1 provided similar values of Ci. On average, water restriction increased the Ci by 13% compared to full irrigation. At the 50% ETc irrigation level, the T3 treatment provided a 35% increase in the E value compared to the control treatment without differing from the T1 and T2 treatments. Under full irrigation, treatment T2 stood out, increasing the evapotranspiration rate by 32% compared to the control. However, this treatment was similar to treatments T1 and T3.

3.3. Grain Yield and Protein Yield

According to the analysis of variance, the irrigation levels influenced the cob length and weight, thousand-grain weight (TGW), grain yield (GY), protein concentration in the grain (PC), and grain protein yield (GPY) (

Table 4. Summary of the analysis of variance for the cob length, cob weight, thousand-grain weight (TGW), grain productivity (GY), grain protein concentration (PC), and grain protein production (GPY).

Mean Square
Source of Variance	DF	Cob Length	Cob Weight	TGW
Irrigation levels (I)	1	317.109 **	88,258.659 **	5808.908 **
Blocks (replications)	4	16.567	1506.904	2367.481
Microorganisms (BM)	3	2.565 ns	2141.909 **	234.574 ns
I × B	3	7.064 ns	696.635 ns	862.145 *
Error	28	4.262	376.260	232.861
CV (%)	-	11.86	8.72	5.01
Source of variance	DF	GY	PC	GPY
Irrigation levels (I)	1	121,010,430.8 **	9.844 **	558,005.970 **
Blocks (replications)	4	1,748,237.0	0.069	14,160.126
BM	3	2,916,790.2 **	0.580 ns	21,538.689 **
I × BM	3	1,160,782.8 *	0.593 ns	18,627.769 *
Error	28	318,616.4	0.406	4391.232
CV (%)	-	7.75	7.13	10.31

CV: coefficient of variation. ** p < 0.01 and * p < 0.05 by F test. ^ns Stands for nonsignificant data at the 0.05 probability level. DF: degrees of freedom.

). The treatments related to BM influenced the cob length, GY, and GPY. There was an I × BM interaction for TGW, GY, and GPY. At both irrigation levels, the treatments related to BM did not influence the ear length (Figure 4). On average, a water deficit decreased the ear length by 28%. The ear weight was higher in the T3 treatment, which differed only from the T2 treatment, and water deficit reduced the ear weight by 35%. Under irrigation restriction, TGW was superior in the T2 treatment but differed only from treatment T1, while, under full irrigation, the BM treatments did not change the results of this variable. The full irrigation level provided higher values and TGW in treatments T1 and T2. On average, water restriction decreased the grain yield by 38%. However, at this level of irrigation, the T1 treatment provided an increase of 34% compared to the control, which represented 74% of the average productivity of the treatments without water restriction. At both irrigation levels, the treatments related to BM did not affect the protein concentration in the grains. However, the water restriction increased the value of this variable by 12%. The protein yield followed the same trend as the grain yield; that is, under an irrigation deficit, the highest value was provided by treatment T1, while, under full irrigation, the treatments related to BM did not affect the value of this variable.

3.4. Mineral Nitrogen Contents and Soil Microbiological Activity

According to the analysis of variance, the irrigation levels (I), treatments related to BM, and the interaction I × BM affected the levels of ammonium (NH₄⁺), nitrate (NO₃⁻), and mineral nitrogen (NH₄⁺ + NO₃⁻) in the ground (

Table 5. Summary of the analysis of variance for the soil nitrogen concentrations of ammonium (NH₄⁺), nitrate (NO₃⁻), and mineral nitrogen (NH₄⁺ + NO₃⁻); soil respiration (Sresp); biomass carbon (C-mic); and metabolic quotient (qCO₂).

Mean Square
Source of Variance	DF	NH4+	NO3−	Mineral N
Irrigation levels (I)	1	5790.039 **	11,945.664 **	29,787.444 **
Blocks (replications)	4	268.351563	262.609375	443.141
Microorganisms (BM)	3	522.539 *	2132.265 **	2556.707 **
I × BM	3	685.872 **	1225.638 **	1493.382 **
Error	28	134.914063	153.234	133.017
CV (%)	-	26.16	27.21	13.45
Source of variance	DF	Sresp	C-mic	qCO2
Irrigation levels (I)	1	78.167 *	363,778.833 **	0.077 **
Blocks (replications)	4	4.322	8329.880	0.001
BM	3	30.192 ns	6113.058	0.009 **
I × BM	3	230.833 **	16,440.665 *	0.012 **
Error	28	13.722	3724.769	0.001
CV (%)	-	20.28	24.86	23.38

CV: coefficient of variation. ** p < 0.01 and * p < 0.05 by F test. ^ns Stands for nonsignificant data at the 0.05% probability level. DF: degrees of freedom.

). The irrigation levels and I × BM interaction influenced the soil respiration (Sresp), microbial biomass carbon concentration (C-mic), and respiratory quotient (qCO₂). The irrigation deficit provided the highest levels of NH₄⁺, NO₃⁻, and mineral nitrogen (NH₄⁺ + NO₃⁻) in the soil, with increments of 81%, 106%, and 93%, respectively, compared to full irrigation (Figure 5). In water restriction conditions, the T1 treatment was superior to the other treatments and provided an increase of 47% compared to the control. On the other hand, the NO₃⁻ levels were higher in treatments T2 and T3, which, together with the T1 treatment, were higher than the control.

In the same way, the mineral nitrogen contents were higher in the treatments constituted by the inoculation of BM. Under full irrigation, the treatments related to BM did not change the concentrations of nitrate, ammonium, and mineral nitrogen. Under irrigation deficit, the highest soil respiration rates (Sresp) were obtained in treatments T1 and T2, which were superior to the control and T3 treatment (Figure 5). Under full irrigation conditions, the effect was inverse; the control and T3 treatments were superior to the T1 and T2 treatments. There was no difference between irrigation levels in the T1 treatment. The concentration of C-mic under irrigation deficit was lower in the control treatment but differed only from the T2 treatment. Under full irrigation, BM treatments did not change the C-mic values. On average, there was a 44% reduction in C-mic under full irrigation. The qCO₂ values did not differ between the BM treatments under full irrigation, while, under irrigation deficit, the lowest values were provided by treatments T1 and T2. However, the T1 treatment did not differ from the control treatment. In all BM treatments, the qCO₂ values were higher at the full irrigation level.

3.5. Principal Component Analysis

The multidimensional space of the original variables was reduced to two principal components (PC1 and PC2) with eigenvalues greater than λ > 1.0, according to Kaiser (1960). The eigenvalues and percentages that explained the variation for each component represented 87.21% of the total variation (

Table 6. Eigenvalues, percentages of the total variance explained, in the multivariate analysis of variance (MANOVA), and correlations (r) between the original variables and the principal components.

Parameter	Principal Components
	PC1	PC2
Eigenvalues (λ)	11.42	1.66
Percentage total variance (S2%)	76.12	11.09
Hotelling test (T2) for irrigation levels (L)	0.01	0.01
Hotelling test (T2) for beneficial microorganisms (BM)	0.01	0.01
Hotelling test (T2) for interaction (L × BM)	0.01	0.01
Variables	Correlation coefficient
Grain yield (GY)	0.96	0.08
Internal CO2 concentration (Ci)	−0.84	0.26
Transpiration rate (E)	0.65	−0.75
stomatal conductance (GS)	0.92	−0.29
CO2 assimilation rate (A)	0.95	−0.28
Leaf área index (LAI)	0.93	−0.14
Plant heigth (PH)	0.94	0.27
Culm diameter (CD)	0.96	0.12
Grain protein concentration (GPC)	−0.82	−0.46
Grain protein production (GPY)	0.91	−0.09
Nitrogen content in leaves (Nleaf)	0.90	−0.08
Nitrogen concentration as ammonium (NH4+)	−0.78	−0.06
Nitrogen concentration as nitrate (NO3−)	−0.67	−0.68
Microbial biomass carbon (C-mic)	−0.94	−0.23
Metabolic quotient (qCO2)	0.86	−0.02

). PC1 explained 76.12% of the total variance, formed by most of the variables analyzed. PC2 represented 11.09% of the remaining variance. The two-dimensional projections of the effects of the treatments and variables in the first and second main components (PC1 and PC2) are shown in Figure 6. According to the groupings of the variables presented, in PC 1, it was observed that most of the variables evaluated in the plants were positively correlated with the irrigation level of 100% of ETc, regardless of the BM treatments. In turn, the variables assessed in the soil were associated with the level of irrigation at 50% of ETc, with an emphasis on treatments T2 and T3.

4. Discussion

In the present study, the possibility of BM attenuating the adverse effects of water deficit in corn was evaluated through the plant variables (growth, gas exchange, productivity, and protein content) and soil (mineral nitrogen and microbiological activity). The water deficit imposed by the 50% ETc water depth reduced corn’s growth and leaf area, as well as the nitrogen content in the leaves. Drought stress causes a decrease in cell turgor, which is essential for proper cellular metabolism, such as photosynthesis, enzymatic activity [3,4], and nutrient absorption [5]. In addition, under water deficit, nitrogen contact with the roots, through diffusion and mainly by mass flow, can be reduced [42], decreasing the uptake of this nutrient by the plant [5]. On the other hand, regardless of the irrigation level, the inoculation of plants with Bacillus subtilis promoted an increase of 13% in the leaf area index and 12% in the N content of the leaves compared to the treatment without the inoculation. B. subtilis also increased the leaf area in sweet peppers [43] and leaf N content in sugarcane [44]. Aquino [45] observed that five strains of B. subtilis increased the foliar N content in corn crops compared to uninoculated plants. This indicates that B. subtilis fixed atmospheric N and improved corn nitrogen nutrition [13]. However, the effect of treatments containing BM on maize growth was not well defined, as observed in other studies [7,13].

The gas exchange measurement showed that corn plants closed their stomata (decreased stomatal conductance) under water deficit to reduce water loss. This result was accompanied by a decrease in the photosynthetic rate, an increase in the internal concentration of CO₂, and a decrease in the transpiration rate. In water restriction conditions, corn inoculation with Azospirillum brasiliense increased the transpiration rate by 35% compared to the treatment without inoculation. Physiological changes with the use of BM have also been reported in previous works [7,46]. BM application in the soil can lead to stimulating the production of osmoregulatory substances by the plant and, thus, act synergistically, contributing to drought tolerance [11,12]. These organisms can produce auxins such as indole acetic acid, increasing the length of plant roots, thus leading to a greater absorption of water and nutrients from the soil [8]. In this sense, it has been observed [12] that the inoculation of corn with B. amyloliquefaciens increased the nutrient absorption and promoted growth mechanisms in plants, such as an increased concentration of amino acids such as tryptophan, isoleucine, alanine, valine, and tyrosine and sugars such as fructose and glucose.

The negative effects of water restriction on maize growth and gas exchange were also reflected in a lower grain yield (38% reduction). Water scarcity negatively affects corn development at all phenological stages, promoting an increase in flowering days, maturation days, and anthesis interval and a decrease in leaf area, negatively affecting flowering and grain filling and seriously compromising corn production [6]. At the irrigation level of 50% of ETc, the inoculation of plants with B. amyloliquefaciens and with A. brasiliense provided a productivity increase of 35% in relation to stressed plants without the inoculation. In previous works, it was observed that the inoculation of B. amyloliquefaciens and B. subtilis, associated or not with other BM, was efficient in attenuating water stress in several cultures, promoting an increase in growth and production [9,15,18]. Water restriction increased the protein content (12% increase) in the grains. Other studies have also reported this effect [16,45]. Probably, the increase in protein concentration was due to the decrease in the thousand-grain weight (8% on average in the BM treatments) compared to the level of complete irrigation combined with a decrease in the rate of carbon assimilation and, consequently, the starch synthesis, increasing the proportion of proteins in the grains [46]. Despite the increase in the protein content in the grains, the water deficit reduced the protein yield by 31% due to the severe decrease in grain yield. Although the water deficit caused this antagonistic effect between the protein content and the yield, this is a relevant aspect to be addressed in future research, because there is a possibility that there is a balancing point between the increase in the protein content in the grains caused by water stress and adequate corn grain yield through irrigation depth management during the phenological phases of the crop [47,48].

The concentration of mineral nitrogen (N), as ammonium (NH₄⁺) or NO₃⁻, increased due to the water deficit. In previous works, higher concentrations of mineral N were also observed in soil under water deficits [5,8,49]. Water restriction possibly decreased the rate of N absorption and accumulation by corn, providing higher levels of mineral N in the soil and decreasing the levels of this nutrient in the leaves. The inoculation of plants with B. amyloliquefaciens + A. brasiliense consistently increased the NH₄⁺ content and respiration rate under water restriction but not under full irrigation. The role of BM in increasing mineral N in the soil needs to be better understood. Previous research demonstrated that B. subtilis decreased N volatilization in the form of NH₃ after mineralization [15]. In addition, an inoculation with B. amyloliquefaciens + A. brasiliense could provide an increase in the diversity of microorganisms involved in the mineralization of organic matter [50] or even stimulate the decomposition of organic matter and N mineralization [51] or stimulate root growth and increase the soil organic matter content [52]. The increase in productivity in the T1 treatment is possibly due to the synergistic effect between B. amyloliquefaciens and A. brasiliense. In another work, with a corn crop, the authors observed that the solubilization of phosphorus bound to calcium and iron, and the mineralization of sodium phytate was greater when they were inoculated together in comparison to the separate inoculations, enhancing the release of organic acids. In the present work, the increase in microbial biomass carbon under the water deficit, mainly in the T2 treatment (B. subtilis), was due to the higher microbiological activity. However, at this level of irrigation, the metabolic quotient was not altered. In this sense, Gebauer and collaborators [17] observed that a water deficit favored certain groups of microorganisms that promoted plant growth in soil cultivated with wheat and barley, which was interpreted as an adaptive strategy of plants to water stress.

The multivariate principal component analysis (PCA) demonstrated, more emphatically, a clear separation of the average variables in the plants (except the internal CO₂ and protein content) from the variables measured in the soil. Thus, according to PC1, the variables measured in the plants were closely related to the level of full irrigation (100% ETc). In comparison, most of the variables measured in the soil benefited from the irrigation deficit (50% ETc). The PCA also showed that the inoculation of corn with BM favored soil biological activity and promoted an increase in the grain yield, especially under the water deficit.

5. Conclusions

Our results suggest that combined Bacillus amyloliquefaciens co-inoculated with Azospirillum brasiliense contributed to attenuating the adverse effects of water deficit in maize. Under the water deficit, Bacillus subtilis and A. brasiliense inoculated separately did not prevent water stress in corn but increased the mineral nitrogen content in the soil. In the present study, a 50% reduction in irrigation depth severely restricted corn growth, gas exchange, and decreased the grain yield by 38%. On the other hand, the water deficit increased the protein content in the grains and the concentration of mineral nitrogen in the soil, especially when the plants were inoculated with plant growth-promoting bacteria (BM). In water stress conditions, the inoculation with BM increased corn productivity by 35% and increased soil microbial activity. The findings of this research reinforced the results of previous research and represented a breakthrough in understanding the role of the interactions between plants and microorganisms in adapting to environmental adversities such as water stress.