Inoculation with Azospirillum sp. and Herbaspirillum sp. Bacteria Increases the Tolerance of Maize to Drought Stress

Stress drought is an important abiotic factor that leads to immense losses in crop yields around the world. Strategies are urgently needed to help plants adapt to drought in order to mitigate crop losses. Here we investigated the bioprotective effects of inoculating corn grown under drought conditions with two types of plant growth-promoting rhizobacteria (PGPR), A. brasilense, strain SP-7, and H. seropedicae, strain Z-152. Plants inoculated with the bacteria were grown in a greenhouse with perlite as a substrate. Two hydric conditions were tested: normal well-watered conditions and drought conditions. Compared to control non-inoculated plants, those that were inoculated with PGPR bacteria showed a higher tolerance to the negative effects of water stress in drought conditions, with higher biomass production; higher carbon, nitrogen, and chlorophyll levels; and lower levels of abscisic acid and ethylene, which are plant hormones that affect the stress response. The oxidative stress levels of these plants were similar to those of non-inoculated plants grown in well-watered conditions, showing fewer injuries to the cell membrane. We also noted higher relative water content in the vegetal tissue and better osmoregulation in drought conditions in inoculated plants, as reflected by significantly lower proline content. Finally, we observed lower gene expression of ZmVP14 in the inoculated plants; notably, ZmVP14 is involved in the biosynthesis of abscisic acid. Taken together, these results demonstrate that these bacteria could be used to help plants cope with the negative effects of drought stress conditions.


Introduction
Corn (Zea mays L.) farming is one of the most important and extensive farming systems in the world because of the myriad products derived from this plant. In 2013, a total of 1,018,111,958 tons of corn were produced globally [1]. Drought is a major abiotic stress factor that affects crop yield. Daryanto et al. [2] collected data from peer-reviewed publications between 1980 and 2015 that examined maize and wheat yield responses to drought using field experiments and concluded that the maize yield reduction was 39.3%. The loss in crop yield during drought depends on the phenological stage of the crops and on the severity of the hydric deficit [3]. In coming years, global warming is predicted to increase the severity and the frequency of drought. According to Food and Agriculture Organization of the United Nations (FAO), the Food and Agriculture Organization of the United Nations, agriculture must adapt to the effects of global warming and improve crop resilience in order for food production farming conditions. One such technology is the use of plant growth-promoting rhizobacteria (PGPR) that reduce the harmful effects of abiotic stress.
Azospirillum sp. are among the most studied PGPR [41][42][43][44]. These bacteria have been used successfully as inoculants in different crop and agro-ecological conditions, and they help increase crop production efficiency [45]. Many groups have reported that Herbaspirillum sp. bacteria can help the growth and productivity of some economically important crops, such as rice, corn, and sugar cane [46][47][48][49]. Some studies have shown that using PGPR can lead to a higher tolerance to abiotic stress conditions, including drought and salinity [50][51][52][53][54][55][56][57][58]. However, little is known about the molecular processes involved in the interaction of plants and bacteria in drought conditions.

Experimental Design and Statistical Analysis
We designed a factorial experiment with two factors. The first factor, bacteria, had three levels: (1) control seeds treated with sterile NFb medium [59]; (2) seeds inoculated with Azospirillum brasilense, strain SP-7 (ATCC 29729™); and (3) seeds inoculated with Herbaspirillum seropedicae, strain Z-152 (ATCC 35894™). The second factor, watering, had two levels: (1) watering once every 24 h, called the well-watered (WW) condition; (2) watering once every 96 h, called the drought (D) condition. The experiment involved six treatments, each composed of three replicates. Each experimental unit or replicate had 36 samples. The data were analyzed using analysis of variance (ANOVA), including the interaction between the bacteria and watering factors as a source of variation. We also performed the Tukey test of multiple comparisons using the statistical software InfoStat [60].

Bacterial Growth and Inoculation
We used NFb medium [59] to grow strain A. brasilense, and JNFb medium [61] to grow strain H. seropedicae. To inoculate the seeds, we used a concentration of 1.6 × 10 9 CFU mL −1 (colony-forming units per milliliter) of A. brasilense and 1.8 × 10 9 CFU mL −1 of H. seropedicae in a volume of 50 mL of the appropriate medium. Before inoculation, the seeds were rinsed for 15 min in a solution of 30% v/v of commercial bleach, 70% sterile distilled water, and 100 µL/L of Triton X-100. The seeds were then rinsed three times for 10 min/rinse in sterile water. For inoculation, the seeds were placed in bacterial medium for 24 h.

Plant Growth Conditions
This experiment was conducted in a greenhouse with temperatures between 25 • C and 30 • C, a 16 h/8 h light/dark cycle, and 50-70% humidity. We used a 36-well tray (Speedling Incorporated, Nipomo, CA, USA). Each well had a volume of 250 mL, and sterile perlite was used as the substrate. Each tray was treated the same way in order to establish the experimental unit. After the seeds were inoculated, we placed one seed in each well. We watered each well with 40 mL of Murashige & Skoog Salt Mixture medium [62] at a concentration of 0.20 in distilled water. During the first 8 days of the test, all plants were watered once every 48 h for every treatment condition. Starting on day 8, the plants were watered once every 48 h in the WW condition and once every 96 h in the D condition. The experiment was complete 20 days after sowing (DAS).

Symbiotic Development
Once the experiment was complete, we took samples of the roots of plants that were inoculated with the bacterial strains. We then cleaned the root surface with 25 mL of the solution 30% v/v of commercial bleach, 70% sterile distilled water, and 100 µL/L of Triton X-100. Using a mortar, we extracted a homogenate from which we prepared serial dilutions. The CFU of each strain was determined after 5 days of incubation at 30 • C in plates with solid Congo red medium [63], using colony shape and color as a way to confirm the identities of each strain.

Total Biomass (TB) Production
In order to determine the dry weight of the TB, we collected six random samples (entire plants) from each treatment condition 20 DAS. The samples were placed in an 80 • C oven for 48 h. When the sample was dry, we determined its weight.

Total Carbon (TC) and Total Nitrogen (TN) Content
In order to analyze the TC and the TN, we collected three random samples (entire plants) from each treatment condition 20 DAS. We used a modified Walkley-Black method [64] for TC determination, and we used the Kjeldahl method [65] for TN determination.

Total Chlorophyll (TChl) Content
The TChl content was determined using the method described by Inskeep and Bloom [66]. We collected three random samples (leaves from three plants) from each treatment condition 20 DAS. Disc-shaped samples (1-cm in diameter) were taken from the leaves, avoiding the central nervation. The weight of four discs per sample was taken. Then the four discs were placed in Eppendorf tubes containing 2 mL of dimethylformamide (DMF). The tubes were covered with aluminum foil and stored in a refrigerator at 4 • C for 4 days. The absorbance of each sample was then read at 647, 652, and 664 nm using a spectrophotometer. Chlorophyll content in mg total chlorophyll per g fresh weight was expressed.

ABA and Ethylene Content
In order to determine the ABA content, we collected three random samples (entire plants) from each treatment condition 20 DAS. Then we performed an extraction using the methods of Kelen et al. [67] and Iriti et al. [68]. For analysis, we used an Agilent 1100 Series HPLC with a Zorbax Eclipse XDB C18 column (150 × 4.6 mm; 5 µm particle size). The mobile phase was MeOH:H 2 O 70:30, pH 4.0; flux was 0.5 mL/min; and UV detection was at 265 nm. The injection volume was 20 µL, and the ABA retention time was 3.9 min.
In order to determine the ethylene content, we randomly collected three leaves per treatment condition 20 DAS. These leaves were placed in 10-mL jars with 1 mL of the regulator solution (50 mM Na 2 HPO 4 /NaH 2 PO 4 , pH 6.8). The jars were sealed with a rubber septum and incubated for 24 h at 25 • C in the dark. Ethylene production was measured using gas chromatography [34].

Relative Water Content (RWC)
In order to determine the RWC, we collected six random samples (the last leaf completely enlarged) from each treatment condition 20 DAS, to assess the fresh weight. We determined the weight of the completely turgid sample and the dry weight according to the methods of Naveed et al. [55]. To evaluate the RWC we used the following equation [69]: Fresh weight − Dry weight Completely turgid weight − Dry weight × 100 (1)

Malondialdehyde (MDA) Content
To determine the MDA content, we collected four random samples (entire plants) from each treatment condition 20 DAS. These fresh samples were homogenized in a 20% p/v trichloroacetic acid (TCA) solution and centrifuged at 3500× g for 20 min. A 1-mL aliquot of the supernatant was added to 1 mL of 20% TCA solution plus 0.5% (p/v) thiobarbituric acid and 100 µL of butylated hydroxytoluene (from a 4% solution in ethanol). This mixture was heated at 95 • C for 30 min, cooled on ice, and centrifuged at 10,000× g for 15 min. We determined the absorbance of an aliquot of the supernatant at 532 nm and subtracted the value for non-specific absorption at 600 nm. The concentration of thiobarbituric acid reactive substances (TBARS) was calculated using an extinction coefficient of 155 mM −1 cm −1 as described by Heath and Packer [70].

Proline Content
To determine the proline content, we collected four random samples (entire plants) from each treatment condition 20 DAS. With a cold (4 • C) mortar, we ground 0.5 g of fresh tissue and homogenized it in 5 mL of 3% (p/v) sulfosalicylic acid to precipitate the proteins. The homogenized sample was filtered through Whatman grade 2 filter paper, and 2 mL of the filtered sample or the proline standard was placed in a test tube and was mixed with 2 mL of glacial acetic acid and 2 mL of acid ninhydrin. The solution was shaken and incubated at 100 • C for 1 h, creating a colored complex. The reaction was stopped by placing the sample in an ice bath, then 4 mL of toluene was added to each test tube and each sample was mixed using a vortex mixer for 15-20 s. Finally, the two phases were allowed to separate. The organic (toluene) phase was recovered, and measurements were performed at 520 nm using a spectrophotometer [71].

Differential Expression of the ZmVP14 Gene
The ZmVP14 gene was amplified and quantified using real time PCR and SYBR ® Green to detect the product. Random samples were collected 20 DAS. The primers used to amplify the ZmVP14 gene were as follows: forward 5 -TCCACGACTTCGCCATCACC-3 and reverse 5 -CGTCTTCTCCTTGTCCAGCACC-3 . The products were quantified relative to the control treatment (WW) using the 2 −∆∆C T method [72]. Expression of the endogenous actin gene served as a control using the following primers: forward 5 -TCCTGACACTGAAGTACCCGATTG-3 and reverse 5 -CGTTGTAGAAGGTGTGATGCCAGTT-3 . We used the following amplification conditions: 50 • C for 2 min; 95 • C for 10 min; (95 • C for 15 s and 54 • C for 1 min) × 45 cycles. We used a dissociation curve (melting) to make sure there was no nonspecific amplification.

Symbiotic Development
In the WW condition, the A. brasilense concentration was 9.2 × 10 5 CFU mL −1 and the H. seropedicae concentration was 6.3 × 10 5 CFU mL −1 . In the D condition, the A. brasilense concentration was 2.5 × 10 4 CFU mL −1 and the H. seropedicae concentration was 4.2 × 10 4 CFU mL −1 . Figure 1 shows the TB production results. In the D condition, plants inoculated with H. seropedicae showed 29.5% greater TB production than control plants, and plants inoculated with A. brasilense showed 26% greater TB production than control plants. In the WW condition, plants inoculated with A. brasilense showed 15% greater TB production than control plants.  Figure 2 shows the TC content results. In the D condition, plants inoculated with either PGPR had significantly (41%) more TC than control plants (p ≤ 0.05). In the WW condition, plants inoculated with either PGPR produced approximately 45% more TC than the control plants. There were significant differences in the TN content according to the inoculated bacterial strain, as shown in Figure 3. In both the D and the WW conditions, plants inoculated with H. seropedicae produced 26% more TN than the controls (p ≤ 0.05).   Figure 2 shows the TC content results. In the D condition, plants inoculated with either PGPR had significantly (41%) more TC than control plants (p ≤ 0.05). In the WW condition, plants inoculated with either PGPR produced approximately 45% more TC than the control plants. There were significant differences in the TN content according to the inoculated bacterial strain, as shown in Figure 3. In both the D and the WW conditions, plants inoculated with H. seropedicae produced 26% more TN than the controls (p ≤ 0.05).  Figure 2 shows the TC content results. In the D condition, plants inoculated with either PGPR had significantly (41%) more TC than control plants (p ≤ 0.05). In the WW condition, plants inoculated with either PGPR produced approximately 45% more TC than the control plants. There were significant differences in the TN content according to the inoculated bacterial strain, as shown in Figure 3. In both the D and the WW conditions, plants inoculated with H. seropedicae produced 26% more TN than the controls (p ≤ 0.05).    Figure 4 shows the TChl content results. In the D condition, plants inoculated with either PGPR had significantly more TChl than control plants (p ≤ 0.05). Specifically, plants inoculated with H. seropedicae produced 41.4% more TChl than controls, and plants inoculated with A. brasilense produced 33% more. In the WW condition, plants inoculated with H. seropedicae produced 41% more TChl than control plants (p ≤ 0.05).   Figure 4 shows the TChl content results. In the D condition, plants inoculated with either PGPR had significantly more TChl than control plants (p ≤ 0.05). Specifically, plants inoculated with H. seropedicae produced 41.4% more TChl than controls, and plants inoculated with A. brasilense produced 33% more. In the WW condition, plants inoculated with H. seropedicae produced 41% more TChl than control plants (p ≤ 0.05).   Figure 4 shows the TChl content results. In the D condition, plants inoculated with either PGPR had significantly more TChl than control plants (p ≤ 0.05). Specifically, plants inoculated with H. seropedicae produced 41.4% more TChl than controls, and plants inoculated with A. brasilense produced 33% more. In the WW condition, plants inoculated with H. seropedicae produced 41% more TChl than control plants (p ≤ 0.05).   Figure 5 shows the ABA content results. In both the WW and D conditions, control plants had significantly higher ABA content than the PGPR-inoculated plants (p ≤ 0.05). In the inoculated plants, the ABA content did not differ significantly in D versus WW conditions. In control plants, the ABA content was 30% higher in plants in D conditions versus WW conditions, but this difference was not significant. Figure 6 shows the ethylene content results. In both the WW and D conditions, control plants showed higher ethylene content than inoculated plants, and the ethylene content was not significantly different in the inoculated plants in D versus WW conditions (inoculation * drought interaction p = 0.0297).

ABA and Ethylene Content
Microorganisms 2017, 5, 41 8 of 17 Figure 5 shows the ABA content results. In both the WW and D conditions, control plants had significantly higher ABA content than the PGPR-inoculated plants (p ≤ 0.05). In the inoculated plants, the ABA content did not differ significantly in D versus WW conditions. In control plants, the ABA content was 30% higher in plants in D conditions versus WW conditions, but this difference was not significant. Figure 6 shows the ethylene content results. In both the WW and D conditions, control plants showed higher ethylene content than inoculated plants, and the ethylene content was not significantly different in the inoculated plants in D versus WW conditions (inoculation * drought interaction p = 0.0297).

ABA and Ethylene Content
As can be seen in Figure 6, the ethylene production is increased in non-inoculated plants under drought conditions.    Figure 5 shows the ABA content results. In both the WW and D conditions, control plants had significantly higher ABA content than the PGPR-inoculated plants (p ≤ 0.05). In the inoculated plants, the ABA content did not differ significantly in D versus WW conditions. In control plants, the ABA content was 30% higher in plants in D conditions versus WW conditions, but this difference was not significant. Figure 6 shows the ethylene content results. In both the WW and D conditions, control plants showed higher ethylene content than inoculated plants, and the ethylene content was not significantly different in the inoculated plants in D versus WW conditions (inoculation * drought interaction p = 0.0297).

ABA and Ethylene Content
As can be seen in Figure 6, the ethylene production is increased in non-inoculated plants under drought conditions.   As can be seen in Figure 6, the ethylene production is increased in non-inoculated plants under drought conditions. Table 1 shows the RWC (%) of the plants according to inoculation status and hydric conditions. The RWC was higher in plants inoculated with either PGPR in both WW and D conditions. In the WW condition, plants inoculated with A. brasilense had the highest RWC (10% higher) versus control plants.

RWC
In the D condition, plants inoculated with H. seropedicae had the highest RWC (5.5% higher) versus control plants. Table 1. Effect of inoculation on the relative water content (RWC) of plants in well-watered (WW) and drought (D) conditions. Means with common letters are not significantly different (p > 0.05). Data are means of six replicates ± standard deviation (SD). The inoculation x drought interaction p < 0.0001.

Proline Content
Proline is an indicator of osmoregulation, and was measured during the experiment; the results are indicated in Figure 7. In the WW condition, control plants had a higher proline concentration than plants inoculated with either PGPR. In the D condition, the proline levels increased significantly over time regardless of inoculation status, but control plants had the highest concentration of proline relative to the WW condition (eight-fold more proline than control plants in the WW condition). In the same condition hydric (D), the plants inoculated with H. seropedicae showed the greatest increase (four-fold) in proline over time, while those inoculated with A. brasilense showed a two-fold increase in the proline level (p ≤ 0.05).

Proline Content
Proline is an indicator of osmoregulation, and was measured during the experiment; the results are indicated in Figure 7. In the WW condition, control plants had a higher proline concentration than plants inoculated with either PGPR. In the D condition, the proline levels increased significantly over time regardless of inoculation status, but control plants had the highest concentration of proline relative to the WW condition (eight-fold more proline than control plants in the WW condition). In the same condition hydric (D), the plants inoculated with H. seropedicae showed the greatest increase (four-fold) in proline over time, while those inoculated with A. brasilense showed a two-fold increase in the proline level (p ≤ 0.05).

MDA Content
MDA was quantified in order to assess the integrity of the cellular membranes, since MDA acts as a lipid peroxidation indicator. Figure 8

MDA Content
MDA was quantified in order to assess the integrity of the cellular membranes, since MDA acts as a lipid peroxidation indicator. Figure 8 shows the MDA content results 20 DAS. In the D condition, plants inoculated with A. brasilense showed better membrane stability than plants inoculated with H. seropedicae and control plants. In the D condition, plants inoculated with H. seropedicae showed a two-fold increase in MDA content over time, while control plants showed a >300% increase in MDA content. In the D condition, plants inoculated with A. brasilense showed a~30% increase in MDA content, which was not significantly different than control plants in the WW condition (p ≤ 0.05).
>300% increase in MDA content. In the D condition, plants inoculated with A. brasilense showed a ~30% increase in MDA content, which was not significantly different than control plants in the WW condition (p ≤ 0.05).  Figure 9 shows the results of the ZmVP14 gene expression analysis. The gene expression level was considered to be 1 in control plants in the WW condition. In the WW condition, ZmVP14 gene expression was higher in control plants than in PGPR-inoculated plants. In control plants, ZmVP14 gene expression was almost five-fold higher in the D condition than in the WW condition. ZmVP14 gene expression was almost undetectable in plants inoculated with A. brasilense in both water conditions.   Figure 9 shows the results of the ZmVP14 gene expression analysis. The gene expression level was considered to be 1 in control plants in the WW condition. In the WW condition, ZmVP14 gene expression was higher in control plants than in PGPR-inoculated plants. In control plants, ZmVP14 gene expression was almost five-fold higher in the D condition than in the WW condition. ZmVP14 gene expression was almost undetectable in plants inoculated with A. brasilense in both water conditions. Microorganisms 2017, 5, 41 10 of 17 >300% increase in MDA content. In the D condition, plants inoculated with A. brasilense showed a ~30% increase in MDA content, which was not significantly different than control plants in the WW condition (p ≤ 0.05).  Figure 9 shows the results of the ZmVP14 gene expression analysis. The gene expression level was considered to be 1 in control plants in the WW condition. In the WW condition, ZmVP14 gene expression was higher in control plants than in PGPR-inoculated plants. In control plants, ZmVP14 gene expression was almost five-fold higher in the D condition than in the WW condition. ZmVP14 gene expression was almost undetectable in plants inoculated with A. brasilense in both water conditions.

Discussion
Determination of the CFU mL −1 in the root samples showed that A. brasilense and H. seropedicae effectively colonized the corn seedlings in plants grown in both the WW and D conditions. Thus, bacterial growth was not significantly affected by the D condition, and our results were similar to those obtained by Ruíz-Sánchez et al. [53] with arbuscular micorrhiza and Azospirillum sp. in rice, by Naveed et al. [55] with other PGPR (Burkholderia sp. and Enterobacter sp.) in corn, and by Cohen et al. [56] with Azospirillum sp. in Arabidopsis thaliana.
The TB results (Figure 1) showed that the bacteria promoted vegetal growth in both hydric conditions, but the differences in growth in inoculated versus control plants were greater in D conditions (p = 0.0213). These results may be related to a lower level of stress in the inoculated plants, which showed higher TB. Our results were similar to those obtained by Ruíz-Sánchez et al. [53], Naveed et al. [55], Cohen et al. [56], and Tiwari et al. [67], using Pseudomonas sp. in Cicer arietinum L. There were other interesting results related to TB production that are shown in Figures 2-4, in terms of the TC, TN, and TChl contents. Specifically, when TB was greater, TC (p = 0.0473) and TN (p = 0.0482) were also higher regardless of the inoculation status or hydric condition. These results may be related to a lower level of stress in inoculated plants, which showed lower ABA content. This allowed the stomata to stay open, even in drought conditions, and also allowed better fixation of atmospheric CO 2 in the carbon compounds used in biomass production. These results are in accordance with those of Naveed et al. [55], who showed higher CO 2 assimilation levels, stomatal conductance, and transpiration rate in inoculated corn plants in a drought condition. On the other hand, the higher nitrogen content of inoculated plants may be related to the nitrogen that is provided by the PGPR, which is the result of the biological fixation of atmospheric N 2 [47]. Regarding the TChl content (p = 0.0273), the available nitrogen could have been used for chlorophyll synthesis, which could explain the differences between the TChl in inoculated plants versus control plants. Our results were consistent with those of Ruíz-Sánchez et al. [53], Naveed et al. [55], and Cohen et al. [56].
In this experiment, we found the highest concentrations of ABA in control plants in the D condition. This shows that stressful conditions trigger a signaling pathway that leads to ABA biosynthesis. The control seedlings in the D condition were pale green, smaller in size (lower aerial and radical biomass), had lower turgidity (the plants had fallen or were low), and showed other symptoms of stress, such as leaf senescence. This is an expected result, according to Salinas-Moreno and González-Hernández [73], who showed that a decrease in the hydric potential of corn leaves increased ABA production. In drought conditions, this process prevents the plant from losing water through transpiration. In turn, this decreases the photosynthesis rate, which leads to the observed loss of color (chlorophyll) and slowed or halted growth. The ABA concentration results ( Figure 5) showed that in the WW condition, control plants had a mean of 0.25 µg/gFW ABA, whereas inoculated plants had a mean of 0.05 µg/gFW ABA. In the D condition, control plants showed an increase in the ABA concentration, whereas inoculated plants did not show a similar response. Our results showed that inoculation with PGPR had a negative effect on ABA synthesis in plants. The inoculated plants had a lower stress level. On the other hand, the inoculated plants did not show significant changes in the level of ZmVP14 expression in both WW and D conditions, except those inoculated with H. seropedicae, which showed significant differences (inoculation * drought interaction p < 0.0001). It has been reported that Bacillus subtilis, strain B26, confers resistance against drought stress in Brachypodium and this is linked to the upregulation of expression of several drought-responsive genes and the modulation of the DNA methylation process [74].
The ethylene content was also lower in inoculated plants (p = 0.0297) ( Figure 6). This could be due to the effects of the ACC deaminase produced by the bacteria, which may mitigate the deleterious effect of ethylene, thereby lowering the stress level in the plant and promoting plant growth [18]. The RWC (Table 1) was higher in inoculated plants in both the WW and D conditions (p < 0.0001). It is possible that the integrity of the plasma membrane was better due to the beneficial effects of inoculation, which may have mitigated the damage by the ROS produced in stressful conditions. This was shown by Naveed et al. [55], Cohen et al. [56], and Tiwari et al. [57]. The MDA results (Figure 8) showed that the PGPR-inoculated seedlings had a lower level of damage during drought stress, because in the D condition, control plants showed a significant increase in the MDA level, which is an indicator of the damage caused by lipid peroxidation. The consequences of lipid peroxidation include deterioration in the cellular membrane, deterioration in the selective permeability of the membrane and, finally, cellular disintegration and death. We observed the same effects in the MDA concentrations in the WW condition regardless of inoculation status, demonstrating that inoculation does not affect the MDA in WW conditions. In contrast, in the D condition, there was a clear difference between inoculated plants and control plants. These results may be related to the RWC (Table 1), because less cellular membrane damage would allow a higher water content level inside the cells. Importantly, MDA analysis is used in tolerance tests of corn crops under hydric stress [75], and these results show that in D conditions, the bacteria help plants reduce membrane damage.
Osmolytes synthesis induced by ABA can also help protect plants against drought stress. Notably, osmolytes can decrease the hydric potential of the cell and thereby help it to avoid losing water. Proline is one type of osmolyte, and in this experiment, changes in the ABA concentration ( Figure 5) were similar to the changes in the proline concentration (Figure 7) in the D condition. It appeared that non-inoculated plants could not use this physiological tool to maintain the RWC because the control plants were not turgid and, in many cases, had fallen. These results are in agreement with those obtained by Paleg et al. [76] and Ashraf and Foolad [77], who showed that proline synthesis can be seen as a symptom of stress in plants, taking into account their pathway of induction. It has also been reported that plants that were inoculated with PGPR (Pseudomonas sp.) bacteria showed a decrease in proline synthesis when they were under water stress [57]. This would be in agreement with our results, that plants inoculated with A. brasilense and H. seropedicae may have a lower level of stress in the D condition than control plants. Finally, Bogges et al. [78] determined that water stress prevents the oxidation of proline, resulting in higher proline levels. In addition, other researchers working with Bacillus megaterium BOFC15 and Arabidopsis found that spermidine improves drought tolerance in plants, which was associated with altered levels of ABA [79].
The ZmVP14 gene expression analysis ( Figure 9) showed that in both the WW and D conditions, control plants showed higher ZmVP14 expression than inoculated plants. Notably, ZmVP14 expression is induced when the seedlings are under hydric stress. Our data showed that there were differences in inoculated versus non-inoculated plants in terms of genetic regulation of ABA production that resulted in differences in ABA concentration. Our ZmVP14 expression results agreed with those of Tan et al. [16] and with those of Chernys and Zeevaart [80], who showed that the inhibition of ZmVP14 affects ABA biosynthesis. Furthermore, the inoculated plants grew well even under conditions of water deficit, in addition the inoculated plants had lower levels of ABA than the control plants. Perhaps because there is some kind of mechanism induced for the presence of bacteria that allows sustaining a good growth and good water ratio in the plant.
Our results are probably in agreement and are explained by the highest RWC, such as that found in References [81,82]. Changes in the distribution of specific fatty acids in the root [82] and polyamines-compounds that regulate the growth of plants, including cadaverine-have been correlated with the growth of the radical system or the mitigation of osmotic stress in some plant species [83].

Conclusions
Our results show that the inoculation of maize plants with A. brasilense or H. seropedicae bacteria had important effects on the tolerance of the maize to drought stress. The bacteria had direct effects on physiological, biochemical, and molecular processes that resulted in reduced stress to the plants. The results very clearly show a large positive effect on growth under drought and a reduced stress response (proline, ethylene), and reduced evidence of stress (MDA). Inoculation with these bacteria also greatly affected % total carbon, ABA content, and ethylene content. These results suggest that these bacteria could be used to reduce the effects of drought stress and thereby improve the productivity of maize crops in drought conditions.