Application of Trichoderma asperellum T34 on maize (Zea mays) seeds protects against drought stress

Coating maize seeds with the microbial plant protection product Trichoderma asperellum strain T34 is an effective form of inoculation that enhances plant performance when faced with drought stress, and it improves nutrient and kernel parameters differently in drought and non-stressed conditions. Drought is currently one of the biggest threats to maize production. Trichoderma spp. is mainly used in agriculture as plant protection product with secondary beneficial effects on plants: improved growth, nutrient uptake and plant immunity. Here, we studied the physiological performance of maize plants under two different water regimes (fully irrigated and drought conditions) and three different seed treatments: application of Trichoderma asperellum strain T34, application of a chemical fungicide (CELEST XL) or the combination of both. Regardless of water regime, T34 treatment improved kernel P and C, kernel number and dry weight. Higher populations of T34 on the rhizosphere (T34 treatment) alleviated water stress better than lower T34 populations (T34+Q treatment). Under drought, T34 treatment improved leaf relative water content, water use efficiency, PSII maximum efficiency and photosynthesis. T34-treated maize seeds maintained sufficient T34 populations to alleviate drought throughout crop development suggesting an optimal dose of 104 and 105 colony forming units g−1 dry weight of rhizosphere under the studied conditions. This work helps to demonstrate the beneficial interaction between T. asperellum strain T34 and maize plants under drought.


Introduction
Seed treatments are a suitable way to protect seeds against environmental adversities and ensure plant stand and seedling health. They mainly consist of the external application of a given compound (or more than one), which stays attached to the seeds (Zeun et al. 2013). Currently, most seed treatments are chemical formulations (Krzyzinska et al. 2005;Zeun et al. 2013; Moya-Elizondo and Jacobsen 2016), the goal of which is to protect the seed against seedand soil-borne diseases, and they are extensively used in all agricultural areas and with almost all crops (Zeun et al. 2013). However, agricultural practices are changing, and more natural and sustainable seed treatments are required. Slowly but steadily, seed treatments have been evolving to include the application of active biological substances, including Azospirillium, Pseudomonas, Azotobacter, Bacillus spp. and Trichoderma spp., together with various adhesive agents (Taylor et al. 1991;Gholami et al. 2009;Hu et al. 2011;Moya-Elizondo and Jacobsen 2016). Microbial Communicated by Anastasios Melis.

Electronic supplementary material
The online version of this article (https ://doi.org/10.1007/s0042 5-020-03404 -3) contains supplementary material, which is available to authorized users. 8 Page 2 of 12 inoculants have several advantages over chemical compounds: greater safety, potentially reduced environmental and human damage, targeted activity, effectiveness in small amounts, and self-multiplication controlled by the plant and the indigenous microbial populations (Singh et al. 2011).
Maize (Zea mays) is grown on over 170 million ha worldwide, of which 130 million ha are in less-developed countries (FAO, 2014). Low yields of these crops in sub-Saharan Africa are largely associated with drought stress and low soil fertility. Water stress decreases plant growth and development leading to a reduction in crop yield (Farooq et al. 2009). Plant tolerance to this abiotic stress is the result of extremely complex mechanisms involving molecular, biochemical and physiological processes (Medina et al. 2016). The degree of the reduction in crop yield due to this stress is dependent not only on the genotype but also on the severity and duration of the drought, the phase of plant growth and development, and environmental interactions with other factors (Medina et al. 2016).
Here we aim to determine the effect of T34 on maize growth, kernel development, and both physiological and nutritional status under two irrigation conditions (wellwatered and water-stressed conditions) via the application of three types of seed treatments: biological, chemical, and their combination. Our work focuses on a rapidly growing hybrid that has some tolerance to drought and which is widely used commercially in the Mediterranean region.

Experimental setup
The experiment was performed in a glasshouse at the facilities of Camps Experiementals, Universitat de Barcelona (41°23′6.259′′N; 2°7′12.434′′E, 60 m above sea level) with maize (Zea mays) seeds treated or not with T34, over four months from June to October. The maize genotype was the hybrid SF9C031 from Semillas Fitó (Barcelona, Spain). T34 was coated on the seeds at a concentration of 10 10 CFU g −1 and applied at 4 kg Tn −1 of seed; the chemical fungicide was applied at 10 L Tn −1 of seed. Temperature and relative humidity were measured throughout the experiment with a datalogger (PCE-HT 71N, PCE Ibérica, Albacete, Spain) located in the center of the trial at canopy level, protected from direct light. During the 113 days after sowing (DAS), the average temperature was 25.3 °C (maximum 38.0 °C and minimum 14.5 °C) and the relative humidity was 63%; (maximum 92.5% and minimum 27.0%). The plants were covered with a shading screen extended in the hottest hours of the day, 12:00-16:00 hours (UTC+2: CEST), and when necessary a cooling system was activated to maintain a suitable temperature.
Three seed pools were used: seeds treated with the commercial biological fungicide (T34), active ingredient T. asperellum strain T34 (Biocontrol Technologies, SL, Barcelona, Spain); seeds treated with a commercial chemical fungicide (Q), CELEST XL active ingredient fludioxonil 25% and metalaxyl-M 1% (Syngenta Crop Protection, Basel, Switzerland); and a combination of both products (T34+Q). Previous to seed treatment, all seeds were surface disinfected before any application with 2% sodium hypochlorite (commercial bleach) for 10 min, and then rinsed three times with sterile distilled water before air drying in a laminar flow hood at room temperature for 2 h. Maize seeds (one per pot) were sowed 10 cm from the top of the substrate, using 10 L pots with coconut fiber:perlite:vermiculite (2:1:1, v/v) as an inert substrate with a pH of 6.5 (25 ºC) and an electrical conductivity of 1 mS cm −1 . The plants were watered with Hoagland's solution diluted 1:2. There were two irrigation conditions: fully irrigated and drought condition, with 20 pots per treatment, giving a total of 120 pots. The pots were placed in two blocks to fit two independent lines of drip irrigation. The pots within each block were placed randomly. The chemical fungicide was applied in such a way as to mimic commercial practices with a well-used broadspectrum fungicide.
The irrigation was the same in all pots for 35 DAS [from VE to V8 maize scale (Hanway and Ritchie 1986)] using two drips per pot, which provide ~160 ± 8.3 mL of water per application three times a day. The fertirrigation was applied automatically in the early morning, at the middle of the day, and in the late afternoon. Then, half watering was applied by removing one of the drips in half of the pots (mild water-stress) for a 28-day period [35-63 DAS, from V8 to R1 (silking) maize stage (Hanway and Ritchie 1986)]. In this period, fertirrigation was applied three times per day, at a dose of 146 ± 5.2 mL pot −1 application −1 , except for the stressed plants, where half of this watering was applied. After that, the drought stress was intensified by removing the remaining drip from the drought treatment pots (intensified water-stress) from 63 to 113 DAS [from R1 to R5 maize scale (Hanway and Ritchie 1986)]. The well-watered plants (WW) continued receiving the fertirrigation three times per day with two drips per pot at a dose of 355 ± 7.6 mL pot −1 application −1 . Nevertheless, both WW and water-stressed (WS) plants received additional water supply (1 application day −1 , during 5 days distributed uniformly in this last 50-day period) to avoid complete plant desiccation in WS. The drought stage lasted 78 days and the study ended at 113 DAS. A soil moisture meter (Jellas Soil pH Meter, Jellas Technology Limited, Hong Kong) was regularly used to monitor the soil moisture in the pots. Normal soil moisture levels (4-7 on a scale of 0-10) were always maintained for WW, while in WS values reached "dry" (< 3) during the measurements and sampling of material.

T34 population
To assess the establishment and maintenance of the fungus, its population was determined by the serial dilution method and plate counting in Trichoderma spp. selective medium (Chung and Hoitink. 1990). Populations were evaluated in seeds at sowing time (3-5 seeds per treatment, T34, Q and T34+Q) as well as at the end of the experiment in the plant rhizosphere for five pots per treatment and condition and repeated twice. Dilutions were initiated with 1 g of seeds in a tube containing 9 mL of autoclaved saline solution (NaCl 9 g L −1 ), or with 10 g of the substrate from the soil rhizosphere (without removing the roots) in 90 mL of autoclaved saline solution and agitated for 30 min in a rotatory shaker.

Scanning electron microscopy images
To check for the effective adhesion of T34 on the seed surface, seed samples were fixed in 2% paraformaldehyde and 2.5% glutaraldehyde (pH 7.4) at 4 °C, then washed in phosphate buffer 5 times at 4 °C and further processed with a solution containing 1% osmium tetroxide and 0.8% potassium ferrocyanide in phosphate buffer for 2 h at 4 °C and later washed with distilled water. Afterwards, the samples were dehydrated with ethanol and dried using the critical point procedure (Polaron critical point drier). The samples were mounted in a standard scanning electron microscope (SEM) support and coated with Au (Fisons Instruments SEM coating system). The secondary electron images were obtained using a SEM JEOL 7001 FEG at the electron microscopy unit.

Gas exchange and chlorophyll fluorescence measurements
Gas exchange parameters were determined to provide key information about plant physiology and photosynthetic machinery. Maize plants were measured with a Li-Cor 6400 system (Li-Cor, USA) using the youngest fully developed leaves. The measurements were recorded from an intermediate leaf position on one side of the central nerve. Four to five plants were used per treatment and water condition at three different stages: 63 DAS, at the end of the mild water stress (R1, silking), and at two points during intensified water stress, at 85 DAS (R3, milk stage) and, 104 DAS (R4, dough stage). The measurements were photosynthesis rate (A n ), transpiration (E), stomatal conductance (g s ), ratio of intercellular/extracellular CO 2 concentration (C i /C a ), maximum quantum efficiency of photosystem II ( F ′ v F ′ m ), and water use efficiency (WUE), calculated as the ratio A n /E. These measurements were taken at saturating 1,500 µmol m −2 s −1 irradiance and 25 ºC using a 2 cm 2 leaf chamber, and performed on clear days between 3 and 9 h after dawn when photosynthesis was most likely to peak.

Water status, yield parameters and mineral nutrients
A simple and classic method for evaluation of plant hydric status is the leaf relative water content (RWC). It was measured on the same leaves as for the gas exchange measurements. RWC was calculated using the formula: RWC (%) = [(FW−DW)/(TW−DW)] × 100; where FW is the leaf fresh weight, TW is the leaf turgid weight after 24 h imbibition in the water at 4 °C, and DW is the leaf dry weight after drying the sample to constant weight in an oven at 60 °C (Barrs and Weatherley 1962).
Since drought and T34 application can affect nutrient absorption, mineral nutrients were determined in leaves at two different growth stages (85 and 104 DAS; R3 and R4, respectively) and in kernels at harvest (113 DAS) in the Centres Científics i Tecnològics de la Universitat de Barcelona (CCITUB, Spain). Moreover, at the end of the experiment (113 DAS) the dry weight/number of kernels per plant, plant height and shoot fresh/dry weight were determined for each treatment and condition combination. Leaf samples were taken from the dried leaves used for the gas exchange, chlorophyll fluorescence and RWC measurements. All the samples were finely powdered using a Mixer Mill MM400 (Retsch GmbH, Germany). Then 100 mg per sample was digested with 2 mL concentrated HNO 3 and 2 mL H 2 O 2 in a Teflon container at 90 °C for 3 days. Analysis of Ca, K, Mg, Fe, P, Na and S was performed by inductively coupled plasma optical emission spectrometry (ICP-OES) using an Optima-3200RL (Perkin Elmer). Analysis of B, Cu, Zn and Mn was performed by ICP mass spectrometry (ICP-MS) using an Elan 6000 (Perkin Elmer).

Carbon and nitrogen content and isotope composition
The carbon and nitrogen concentration and the stable isotope composition (‰) of carbon ( 13 C/ 12 C) and nitrogen ( 15 N/ 14 N) were determined in the leaves and kernels using an elemental analyzer (EA1108, Series 1, Carlo Erba Instruments) coupled to an isotope ratio mass spectrometer (IRMS, Delta C, Finnigan MAT) as described in Medina et al. (2016). The 13 C/ 12 C and 15 N/ 14 N ratios were expressed in δ notation for carbon and nitrogen as follows: δ (‰) = [(sample isotope ratio/standard isotope ratio)-1] × 1000; where the standard refers to the international secondary standards of known stable carbon and nitrogen ratios (Medina et al. 2016).

Statistical analysis
Treatment main effects and interactions were determined by one-or two-way analysis of variance (ANOVA) followed by Fisher's least significance difference (LSD) test (P < 0.05). Results of T34 populations, nutrient elements and isotope parameters are mean values of three replicates per treatment and condition. Nutrient and isotope determination in the kernel was performed on three ears of different plants per treatment and condition, being each of them a pooled sample of all grains in each ear. Results of gas exchange and water parameters are mean values of three to five replicates per treatment and condition. Results of kernel dry weight and number are mean values of four to seven replicates per treatment and condition. Results of plant height and plant shoot weight are mean values of seven and three replicates per treatment and condition, respectively. Normality and homoscedasticity of the data were assumed to perform the ANOVA. Moreover, each plant represented an independent sample. The analysis was performed with GenStat 7th Edition software (VSN International Ltd., UK), while the graphics were produced with Sigma-Plot 11 software (Systat Software Inc., USA).

Seed treatment and T34 population
SEM images (Fig. 1) showed the conidia of fungus attached to the seed in a stable and uniform manner. The adhesive applied surrounded the conidia and adhered it to the seed ( Fig. 1a, b) while developing hyphae and conidia can be seen in the seeds once germinated (Fig. 1c, d).
Fungal population per seed (Table 1) just before sowing showed equal populations in both inoculated treatments (T34 and T34+Q) as expected. The Q treatment, on the other hand, presented negligible populations of T34, probably due to accidental contributions, but insignificantly in comparison to the treated seeds. In the rhizosphere, T34 populations were higher in T34 seed treatment, followed by seeds with T34+Q, and then Q. Interestingly, the application of the chemical reduced the proliferation of T34 by a factor of 10. Apart from that, significant results regarding water regime and treatment unexpectedly differentiated T34 populations, but to a lesser extent than for treatment only.

Physiological plant measurements
Gas exchange and chlorophyll fluorescence measurements were firstly recorded at a very early stage (35 DAS) before the application of the different water regimes (Suppl .  Table S1). At this stage, these parameters were not significantly modified by the treatment factor. At 63 DAS, A n and E increased significantly in the T34 treatments in comparison to Q treatment, especially under water stress conditions in T34 (Fig. 2). C i /C a ratio and leaf RWC were slightly reduced by WS compared to WW and C i /C a ratio and g s increased in T34 compared to T34+Q and Q (Fig. 3, Suppl. Table S1). At a more advanced reproductive stage (85 DAS), A n , E, g s and F ′ v F ′ m decreased under WS compared to WW (Fig. 2, Suppl. Table S1). The highest E value was reached in the treatment T34+Q under WS conditions. C i /C a ratio decreased under WS, with the highest value obtained in T34 under WW, although it was significantly decreased by WS. Leaf RWC was similar between treatments under WW conditions, but interestingly it increased in T34 treatments (T34 and T34+Q) while it decreased in Q under WS (Fig. 3). These changes were similarly achieved for WUE, although they were not statistically significant. In the last growth stage, at 104 DAS, when the cob was ripening and the plant was starting to dry out, we found that A n and g s (Fig. 2) decreased compared to previous growth stages. We were surprised to find that An in T34 and T34+Q under WS was higher than in the Q treatment. However, g s was similarly decreased by WS in the different treatments but not significantly (Suppl . Table S1). Associated with these changes under WS, we observed higher F ′ v F ′ m and WUE in the T34 and T34+Q treatments compared to Q, with these values being even higher when only T34 was applied. Leaf RWC was strongly decreased by WS compared to WW, while it was higher in T34 treatments, particularly in T34+Q (Suppl .  Table S1).

Elemental nutrient composition and isotopic measurements
Leaf mineral nutrient content and C and N isotope compositions at 85 and 104 DAS (Table 2) showed that the water regime significantly affected nutritional components including N, C/N, K, P, S, Mn, B, and Zn, as well as δ 15 N. Particularly, WS decreased N, P, S, Mn, B and Zn in all the treatments at 85 DAS which continued decreased at 104 DAS but now also K was increased. The δ 15 N increased under WS at both times, while the significant decrease in N led to an increase in the C/N ratio in leaves (Suppl . Table S2). Meanwhile, treatment and interaction effects were barely significant: S and Zn were significantly modified by treatment, increasing their content in leaves in T34+Q at 85 DAS, but this effect disappeared at 104 DAS (Suppl.  Table S2). Results of kernel mineral nutrient content and C and N isotope compositions showed significant differences by treatment and water regime (Table 3, Suppl. Tables  S3, S4). C and P contents were higher in T34 and T34+Q treatments than in Q treatment, being particularly higher in T34. WS conditions altered kernel nutrient and isotope composition (Table 3): N, P, Mg, Zn and δ 13 C decreased by WS while Fe, C/N and δ 15 N increased. The treatment × water regime interaction did not modify these parameters in the kernel, except for the δ 13 C (Suppl. Tables S3, S4). It was strongly affected by WS in Q treatment and in a lesser extent in T34+Q, being not affected in T34.

Yield parameters
WS did not affect plant height but decreased shoot fresh/dry weight and dry weight of kernels per plant (Fig. 4, Suppl. Table S5). Shoot fresh weight was higher in T34 and T34+Q Fig. 1 Scanning electron microscopy (SEM) images of corn seed treated with T. asperellum strain T34. a, b Conidia adhered to the surface of a seed corn just after T34 treatment. c, d Germinated hyphae and/or conidia of T34 on germinated corn seed 8 Page 6 of 12 compared to Q, while a similar tendency was observed for shoot dry weight although it did not reach statistical significance. T34 treatment showed the highest number of kernels per plant compared to Q and T34+Q (Fig. 4), while the effects of water regime or the interaction were not significant (Suppl . Table S5). Dry weight of kernels per plant was significantly higher in T34 and T34+Q compared to Q under WW, while a similar non-significant tendency was observed under WS.

Discussion
The seed coating applied in this assay with T. asperellum, strain T34 was effective and the presence of the chemical fungicide did not affect the initial populations of maize plants (Fig. 1, Table 1). In general, many previous studies (Lifshitz et al. 1986;Harman 1991;Prasad et al. 2002;Mastouri et al. 2010) showed that seed coatings excipients used in scientific publications (talc, pelgel, gum arabic, corn starch, methyl cellulose, polyethylene glycol) differ from those used industrially in seed companies (synthetic microplastic-based polymers like the used here) and, therefore, are not comparable with the seed treatment used in our work. Chemical treatment alone presented low T34 populations due to initial cross-contamination, but it was insignificant compared to the treated seeds. Nevertheless, those populations in the rhizosphere remained in a similar range to naturally occurring populations of Trichoderma spp. in soil: 10 2 -10 3 CFU g −1 dry weight of soil (Longa et al. 2009) and were not altered by water regime (Table 1). At the end of the assay, the rhizosphere populations were significantly much higher in the seed treatments with T34 than in T34+Q, demonstrating the capacity of T34 to survive and establish itself in an inert soil from only an impregnated seed. As expected, the presence of the chemical fungicide lowered its populations. Interestingly, the interaction showed that populations increased under WS in T34+Q, but decreased in T34 in comparison to their respective WW treatment. This could mean that the plant may modulate T34 populations as needed, suggesting also that 10 4 CFU g −1 may be inadequate for WS and 10 5 CFU g −1 may be preferable. Other experiments with field trials showed that T34 can be found at 10 4 CFU g −1 rhizosphere dry weight at the beginning of the maize reproductive stage (Fitó seeds, unpublished data) with the same maize hybrid, seed coating (T34+Q) and in optimal conditions. In Pandey et al. (2016), dose-dependent T. harzianum strain Th-56 showed a differentiated water-stress response in different rice genotypes, suggesting the need Table 1 T34 population in the seeds at the beginning of the experiment and in the rhizosphere at 104 days after sowing T34 T. asperellum T34 seed treatment, Q chemical fungicide seed treatment, T34+Q combination of T. asperellum T34 and chemical fungicide, WW well-watered, WS water-stressed (WS), CFU colony forming units of T34 per gram of dry weight of substrate a,b For the treated seed and rhizosphere populations, values were log-transformed mean values ± SE of 3-5 seeds or five pooled pots, respectively, for each treatment and condition, repeated twice. Significant effects were determined by two-way ANOVA (P < 0.05) plus Fisher's LSD. Different letters in each section indicate significant differences (P < 0.05) Beginning of assay (treated seeds) a P value LSD value for a minimal dose of T. harzianum to influence crop physiology under drought, which correlates with the effects observed in our experiment. In our experiment, water and treatment effects were more evident in late stages (grain filling) when the drought was more evident and can strongly decrease kernel yield, in agreement with previous reports (Anami et al. 2009;Bista et al. 2018). However, in the case of photosynthetic capacity, An was significantly modified in early (63 DAS) and late (104 DAS) stages, but not at 85 DAS, which could be associated with the specific genotype used in this experiment. In this sense, the fact that F ′ v F ′ m and g s were mainly affected in late growth stages when water supply was very limited demonstrated that the photosynthetic machinery of our maize hybrid showed a certain drought tolerance. WS only decreased the C i /C a ratio and leaf RWC compared to WW at 63 DAS, changes that did not modify photosynthetic capacity. At 85 DAS, An decreased slightly under WS together with E, mainly due to stomatal closure (lower g s ) and a lower maximum quantum efficiency of photosystem II ( F ′ v F ′ m ; Fig. 1, Table S1). When water deprivation was intensified, coinciding with grain filling stage (104 DAS), we observed the same changes on photosynthetic capacity as before but more severe, also explained by a decrease in leaf water status (WUE, RWC) and CO 2 available for its fixation by Rubisco (C i /C a ratio). It is known that A n decreases when water is scarce, due mainly to stomatal closure or inhibition of CO 2 metabolism (Asada et al. 2000;Javed et al. 2011;Vitale et al. 2011;Aslam et al. 2013), which is in agreement with our results. δ 13 C values are known to decrease with increasing water stress in maize and other C 4 plants (Dercon et al. 2006). In our study, kernel δ 13 C decreased under WS (Table 3, Suppl. Table S3), highlighting that the plants suffered water stress during the experiment regardless of the treatment.
Regarding the treatment effects, at early stages (63 DAS) T34 treatments (T34 and T34+Q) promoted photosynthesis and transpiration compared to Q, particularly under WS (Fig. 1, Suppl. Table S1). Anyhow, T34 treatments showed some features suggesting a better water status, as an increase in E in T34+Q under WW or leaf RWC in T34 and T34+Q compared to Q. During the key growth stage of grain filling (104 DAS), and especially when water was severely limited in WS conditions, mimicking real conditions for field-grown maize in the Mediterranean basin, greater differences were observed. Although T34 seemed to not have an effect on photosynthetic capacity when water supply was optimal, it really influenced it under WS conditions. Indeed, application of T34 (T34 and T34+Q treatments) increased A n and E by an improvement of light harvesting ( F ′ v F ′ m ) and leaf water Plants were subjected to mild stress for 28 days (until 63 DAS), and intensified stress for the rest of the assay (until 85 and 104 DAS). Each value is the mean of 3-5 replicates per treatment and condition. Significant effects were determined by two-way ANOVA (P < 0.05). Different letters denote statistical significance for the interaction. Figures with no letters showed no significant differences in the interaction. Mean values in Suppl. Table S1 Fig. 3 Water use efficiency (WUE) and leaf relative water content (RWC) were recorded in well-watered (WW) and water-stressed (WS) plants at three time points (63, 85 and 104 days after sowing: DAS) with three different seed treatments [T. asperellum strain T34 (T34); chemical fungicide (Q) and their combination (T34+Q)]. Plants were subjected to mild stress for 28 days (until 63 DAS), and intensified stress for the rest of the assay (until 85 and 104 DAS). Each value is the mean of 3-5 replicates (n = 3-5). Significant effects were determined by two-way ANOVA (P < 0.05). Different letters denote statistical significance for the interaction. Figures with no letters showed no significant differences in the interaction. Mean values in Suppl. Table S1 8 Page 8 of 12 status (WUE and RWC). These changes can also suggest the absence or reduction of damage to the photosynthetic machinery caused by drought stress. Similar improvements in quantum efficiency were observed with T. harzianum T22 in Arabidopsis (Shoresh et al. 2010). Moreover, kernel δ 13 C results showed higher values in the T34 and T34+Q treatments than in Q under WS (Suppl . Table S3), which may indicate that T34 improved water status throughout the experiment, since δ 13 C provides information on the photosynthetic carbon assimilation and is a time-integrated indicator of the A n /g s ratio and WUE (Serret et al. 2018). The increase in shoot fresh weight in T34 and T34+Q compared to Q, but not significantly in shoot dry weight, clearly indicated that water content at whole plant level was promoted by T34 application. In carnation plants, enhancement of water uptake related to T34 is known, even in the presence of biotic pressure (Sant et al. 2010).
Leaf mineral nutrient content was modified by WS, indicating that the uptake of several nutrients (i.e., N, P, S, Mn, B and Zn) decreased due to this stress, except in the case of K at 104 DAS (Table 2, Suppl. Table S2). δ 15 N was modified by water conditions in leaves at two growth stages and in kernels at harvest. The nitrogen isotope composition (δ 15 N) is used to trace shifts in nitrogen metabolism, although it could be related to several processes and the underlying biochemical mechanisms are still not very well understood (Medina et al. 2016;Serret et al. 2018). In our experiment, its changes highlighted that it is a good indicator of changes in N metabolism since the N content was affected significantly in leaves and kernels (Tables 2, 3). Drought decreases nutrient uptake in most plants for several reasons, including a reduction of nutrient supply through mineralization and reduced nutrient diffusion in the soil (Bista et al. 2018). Related with this, E and g s decreased in our experiment, which could directly influence nutrient uptake. These changes affected final kernel quality, decreasing the content of nutrients such as N, P and Zn, although Fe increased (Table 3). In many cases, drought stress leads to Fe deficiencies in plant (Ahanger et al. 2016). However, in our case its higher kernel content could reflect a greater Fe uptake and/or allocation during this experiment and can be related with the lower content of P and Zn, which are known to have opposite effects on Fe uptake (Ahanger et al. 2016).
The effects of T34 application on mineral nutrient content in leaves was scare. Although some nutrients (S and Zn) increased in T34+Q leaves compared to the other treatments at 85 DAS, these changes disappeared at 104 DAS. In the case of B, it increased in leaves of T34+Q under WW and T34 under WS. Its uptake is mostly a passive process so it is greatly determined by the rate of water uptake by roots (Ahanger et al. 2016). Its higher content in leaves, at least in T34 under WS, could be related with the improvement of photosynthesis by the stabilization of membrane integrity and reduction of ROS (Ahanger et al. 2016), although T seed treatments (T. asperellum strain T34, chemical fungicide and their combination), C water condition (well-watered and water-stressed), and T × C interaction Each value is the mean of 3 replicates. Significant effects were determined by two-way ANOVA. Different letters denote a statistical significance in Fisher's LSD post hoc (P < 0.05). C, N, K, Ca, Mg, P, S, Na and Fe: mg of nutrient· (g leaf dry weight) −1 . Mn, Cu, B and Zn: µg of nutrient·(g leaf dry weight) −1 . Mean values can be found in Suppl. further studies are needed to corroborate this hypothesis. C and P contents were higher in kernels in T34 and T34+Q compared to Q (Table 3), suggesting higher C assimilation, P uptake and/or translocation to kernels. Similarly, García-López et al. (2015) showed that plants treated with T34 improved P absorption in the presence of non-available P forms in cucumber. In Fernández et al. (2014), improvements of nutrient assimilation and C allocation, as well as other growth parameters were reported in T34-treated tomato plants. In spite of this better C content in kernels and photosynthetic machinery and leaf water status under WS, T34 did not prevent a decrease in kernel weight, suggesting that the drought stress performed in this study was severe. The maize responses to T34 under WS reported here demonstrate that the T34 population levels we measured are adequate to prevent water loss and promote photosynthetic capacity and stress avoidance in maize leaves but not sufficient to improve kernel dry weight and, then, grain yield under WS conditions. Nevertheless, T34 did promote kernel yield when the water regime was more favorable (i.e. WW conditions), highlighting the potential of T34 to improve maize yield under future climate change scenario and/or to reduce environmental contamination. Speculating about the mechanism by which T34 could alleviate drought stress in our experiment, authors such as Alwhibi et al. (2017) observed that T. harzianum increased phenol and flavonoid content in tomato plants affected by drought. Meanwhile, T. asperelloides T203 ameliorated plant growth under biotic stress by lowering deleterious ethylene (that accumulates under stress) thanks to increased antioxidant activity (Shoresh and Harman 2008;Brotman et al. 2013) which is again related to higher phenol synthesis (Ahmad et al. 2017). Increased antioxidant levels would degrade higher amounts of ROS (reactive oxygen species) and, therefore, protect photosynthesis. Harman et al. (2004) suggested that Trichoderma spp. metabolites or root colonization could modulate cellular and molecular changes in plants through modification of plant gene expression, thus causing a shift in plant responses.
In conclusion, in this work we observe the potential of T34 seed treatment under optimal conditions and abiotic stress in a key crop such as maize, which is susceptible to water stress around the world. T34 enhanced hydration and photosynthetic capacity under WS conditions. However, it was not sufficient to avert kernel dry weight decrease, probably due to the severity of the stress. In the context of climate change in Mediterranean regions, where periods of drought are increasingly present, the use of biological plant protection products against soil diseases (such as T34) can have secondary benefits for crops either under optimal conditions or under drought stress. Further studies are necessary to evaluate the positive effects of T34 application on leaf water status and photosynthetic capacity under different Table 3 C and N content and stable isotopic composition (‰), C/N ratio, and content of other macro-and micronutrients in kernels at the end of the assay (113 days) Measurements were taken in well-watered (WW) and water-stressed (WS) plants, with the following seed treatments: T. asperellum strain T34 (T34), chemical fungicide (Q) and their combination (T34+Q). Each value is the mean of 3 replicates. Significant effects were determined by two-way ANOVA. Different letters denote a statistical significance in Fisher's LSD post hoc (P < 0.05). C, N, K, Mg, P and S (mg of nutrient) and water regimes from mild to severe water stress, to understand its effects on plant metabolism and yield improvements, particularly in field-grown conditions and using a wider spectrum of maize genotypes with differentiated drought sensibility. Author contribution statement VE performed the assays, collected kernels, sampled and analyzed tissues, analyzed the data and wrote the manuscript. RV and OV performed the gas exchange measurements and contributed to the writing and critical revision of the manuscript. JJN helped design and perform the seed treatments and critically reviewed the manuscript. MIT provided technical advice concerning T34 application, experimental design and analysis of the results, and reviewed the manuscript.  (T34+Q)]. The plants were subjected to mild stress for 28 days (until 63 DAS), and intensified stress for the rest of the assay (until 85 and 104 DAS). Each value is the mean of 4-7 replicates. Significant effects were determined by twoway ANOVA (P < 0.05). Different letters denote statistical significance for the interaction. Mean values in Suppl. Table S5