Corn Responsiveness to Azospirillum: Accessing the Effect of Root Exudates on the Bacterial Growth and Its Ability to Fix Nitrogen

Corn has shown different degrees of positive response to inoculation with the nitrogen- fixing bacteria of the genera Azospirillum. Part of it has been attributed to the plant genotypic variation, including the root exudates, that are used by the bacteria as energy source. In this study, we grew two corn hybrids that differ for their response to Azospirillum, to investigate the effect of different exudates profiles on the bacteria growth and nitrogenase activity. Employing high performance liquid chromatography, we identified nine amino acids (asparagine, aspartic acid, serine, glutamic acid, valine, phenylalanine, threonine, tryptophan and alanine), six sugars (glucose, sucrose, xylose, arabinose, fructose and galactose) and four organic acids (citrate, malate, succinate and fumarate). The less responsive corn genotype showed reduced plant growth (root volume, shoot dry mass and shoot N content), a lower concentration of Azospirillum cells within the root tissues, a higher content of asparagine and glucose and a reduced amount of metabolites that serve as bacterial energy source (all organic acids + five sugars, excluding glucose). The genotypes did not interfere in the ability of Azospirillum to colonize the substrate, but the metabolites released by the less responsive one reduced the nitrogenase activity.


Introduction
Symbiosis is a biological phenomenon involving changes in the genome and metabolism of organisms from different species, usually with benefits to one or both [1]. In the plant-microbe relationships, two root symbiotic systems have been actively studied, the arbuscular mycorrhizae and root-associated bacteria [1]. While the first mentioned interaction is probably the most widespread in the ecosystem, the plant association with diazotrophic bacteria has been the more exploited symbiotic relationship in plant production [1,2].
Diazotrophic bacteria comprehend groups of free living, associated and nodule forming species capable of enzymatically reducing the atmospheric nitrogen (N) into plant bioavailable N compounds and, in the agriculture, have been shown capable of enhancing crop yield, while reducing the environmental impacts caused by mineral fertilizers [3,4], such as the production of phytohormones,
The genotypes showed comparable values for total content amino acids (Table 1), but the abundance of each class in the exudates differed with the corn lines (Appendix A). The more responsive one displayed lower content of aspartic acid upon full (100% N), lower alanine concentration under partial (75% N) and higher tryptophan abundance in both the partial N supply and inoculation (75% N + Azo) (Appendix A). In the inoculated plants, comparable results of alanine and threonine were found between the genotypes, while a higher release of asparagine and a lower concentration of aspartic acid, serine, glutamic acid, valine, phenylalanine and tryptophan were identified in the less responsive genotypes (Appendix A).
In relation to sugars and organic acids, the genotypes differed for both, the total content (Table 1) and profile (Appendix B), with the less responsive ones showing a reduced concentration of both and, thus, of chemotactic compounds, i.e., metabolites that serve as chemoattracts and an energy source for Azospirillum (sum of all four organic acids + five sugars, excluding glucose). For the less responsive corn line, individual classes of sugar and organic acid were released in lower concentration, apart from glucose, whose concentration was higher upon only mineral N supply (full and partial ) and statistical equivalent to the more responsive one under inoculation (Appendix B). Table 1. Means of total amino acids content (TAA), total sugar (TS), total organic acids (TOA), total of chemotaxis compounds (TCC), most probable number of diazotrophic bacteria (MPN), root volume (RV), shoot dry mass (SDM) and its N content (NCSDN) of two corn genotypes of high (HResp) and low (LResp) responsiveness to A. brasilense under full N mineral supply (100% N), N deprivation (75% N) as well as partial N supply coupled with inoculation (75% N + Azo). According to the Tukey test, means followed by the same capital letter in the column belong to the same group at 5% probability. Means followed by the same lowercase letters in the row do not differ from each other by the F-test at a 5% probability level. * Sum of sugars (excluding glucose) + organic acids.

Treatments
Regarding plant growth, the full supply of N in the form of mineral fertilizer provided plants showing equivalent results for root volume, shoot dry mass and shoot N content (Table 2). However, upon inoculation, the high responsive genotype showed comparable or higher growth parameters than the full N mineral supply (Table 2), while in the less responsive Azospirillum, it was unable to deliver growth parameter, at least, equivalent to mineral fertilization (Table 3). Table 2. Means of root volume (RV), shoot dry mass (SDM), shoot N content (NCSDN) and of the most probable number of Azospirillum (MPN) in the substrate and root tissue of two corn genotypes of high (HResp) and low (LResp) responsiveness to A. brasilense under full N mineral supply (100% N), N deprivation (75% N), as well as partial N supply coupled with inoculation (75% N + Azo). According to the Tukey test, means followed by the same capital letter in the column belong to the same group at 5% probability. Means followed by the same lowercase letters in the row do not differ from each other by the F-test at a 5% probability level.

RV
Regarding the colonization of the rhizosphere by the bacteria, the genotype background did not play a part in the Azospirillum establishment in the substrate, but significantly interfered in the bacteria ability to colonize internal tissues (Table 2), as observed in the reduced most probable number (MPN) of bacteria within the root tissues of the less responsive corn ( Table 2). In artificial media (Table 3), the addition of synthetic compounds that mimic the profile of each genotype had no effect on Azospirillum count, but interfered in the N-fixing mechanism, with the less responsive genotype showing lower nitrogenase activity (Table 3). Table 3. Most probable number of Azospirillum (MPN) and nitrogenase activity of A. brasilense grown in neat NFb medium and NFb media supplemented with synthetic compounds mimicking the more responsive genotype (NFb + HResp) as well as the less responsive one (NFb + LResp). According to the Tukey test, means followed by the same capital letter in the column belong to the same group at 5% probability.

Discussion
Even though sugars are reported as the main class of root-released compounds by plants [13,14], in the present study, this trend has not been confirmed, as our root exudates contained 60% of amino acids, followed by 30% sugars and 10% organic acids. This contrasting result may be related to the high rates of amino acids released by plant roots at early growth stages, as we sampled the plants at seedling stages [15]. However, this higher amino acid content may also be attributed to the Azospirillum, which, as a diazotrophic microorganism, release high rates of N-derivate compounds [16].
Corroborating previous works, all N-depleted plants, as well as those of the less responsive upon inoculation (75% N + Azo), displayed lower amount of sugars, organic acids and chemotactic compounds [17]. All classes of sugars we identified have previously been reported in exudates of corn [18], rice [19] and of the model grass specie Brachypodium [20]. Regarding the organic acids, malate, citrate, succinate and fumarate are considered the most common classes released by corn roots [18], and they were all identified in our study.
In the case of Azospirillum, the plant's root exudates are involved in the bacterial cells migration in the soil [21], in a process governed by energy taxis, i.e., the bacterial moves toward a higher concentration of a nutrient to seek a position at which the energy level is favorable [7,8]. Compared to sugars and organic acids, Azospirillum species have a limited reliance on amino acids to grow [4], which, thus, display null or weak attractiveness on the cells [11]. Concerning A. brasilense, the species used in this work, all classes of organic acids are considered to exert chemotaxis, while, among the sugars, glucose is not considered a chemoattractant [4] and galactose exerts induced taxis, i.e., only when the cells have previously been grown in media containing it [7,8].
However, chemotaxis should not be taken for granted in analyzing corn-Azospirillum symbiosis, as some metabolites that do not serve as energy source, or chemoattracts may interfere in the bacterial growth or metabolism, impacting, thus, the benefits delivered by it to plants [4,22,23]. In this context, asparagine that, such as in our work, has been reported as the most abundant amino acid released by corn [18], wheat [24] and the model grass Brachypodium [20] roots, has already been reported to impair the activity of the nitrogenase of Azospirillum [23]. In cereals, asparagine is involved in N storage and transportation, but its increase within root tissues has been documented as a metabolic response to stress, including nutritional deficiency [25,26]. This is consistent with other plant-growth promoting bacteria, such as Bacillus, that had a suppression of bacterial genes involved in the synthesis of protein after changes in the profile of corn exudates [18]. In this work, however, we only investigated the effects of the whole exudates on bacteria growth and nitrogenase activity, but as the less responsive genotype remained, showing a high amount of asparagine, even under inoculation, we suggest that this amino acid may have contributed to reduce the enzyme activity.
Concerning the other amino acids, while aspartic acid have null effects on Azospirillum species, the tryptophan benefits the bacterial synthesis of plant growth regulators [23], as this latter compound is structurally and functionally related to the synthesis of auxin, the major phytohormone produced by A. brasilense [10,27]. In the literature, an increase in lateral roots due to inoculation has been well-documented and it is, often, ascribed to the bacterial auxin [1,2,5]. In this study, provided the higher concentration of tryptophan in the rhizosphere is of the high responsive genotype, it is plausible that Azospirillum may have had been capable of producing higher concentration of auxin [28], compared to the other genotype. This hypothesis is consistent with the results found for plant growth, as the less responsive genotype displayed reduced root volume data upon inoculation.
Despite the different chemical profiles, the substrate's MPN values suggested that any genotype favored Azospirillum establishment in the substrate, a trend that was later confirmed by our data on bacterial count in supplemented media. This evidences the great degree of versatility of Azospirillum in using different carbon based compounds as energy source, a trait that enables the organism to adapt to a wide array of environments [3], including different agricultural practices and soil types, where cereal crops are grown [4].
Azospirillum colonizes all plant parts, but it exhibits a decreasing pattern concentration from roots to leaves [9]. For this reason, the bacteria are generally regarded as a root-surface microorganism, but the species colonizes the roots in a specie-specific way [10]. A. brasilense, the species used in this work, can further colonize the interior of root tissues [10], where the nitrogenase is more protected against the irreversible inactivation caused by the oxygen. Supporting Fukami et al. [9], our data on internal root colonization, as estimated by the MPN technique, suggest a genotype-dependency regarding tissues colonization, which, among other factors, depends on the bacterial ability to degrade the structural polysaccharide of cell walls pectin, whose content and structure may differ with the plant genetic background [29].
Overall, we showed that just as for their responsiveness, the genotypes differed for the composition of their exudates as well as for the ability of Azospirillum to stablish within root tissues. This difference in profile, however, did not affect the ability of A. brasilense to stablish in the substrate, or to grow in artificial medium, but it interfered in the nitrogenase activity. In this sense, it is likely that the weak growth performance of the less responsive genotype may be related to its exudates profile (lower relative amount of chemotaxis compounds, lower tryptophan concentration and high relative asparagine content), but also to its lower concentration of Azospirillum in the root tissues.
Our results corroborate that the benefits delivered by Azospirillum to corn are affected by the genotype background, and further confirm the involvement of root-released compounds in the degree of response. It is noteworthy that the plant-Azospirillum interaction is settled within a dynamic and complex ecosystem constituted by other compounds not identified in this study, but that may interfere in the behavior of microbial communities [14]. Therefore, further researches on the role of root exudates as mediators of the corn-Azospirillum symbiosis need to be carried out, including regarding the amino acids classes, as despite their null chemotaxis, they seemed to be determinant in the degree of the benefits provided by A. brasilense. Additionally, rather than being the cause of the responsiveness, root exudates are more likely to reflect the specificity of transcriptional signaling involved in the mutual recognitions between Azospirillum and its host plant [12,13].

Materials and Methods
We tested two inbred lines showing different N use efficiency under Azospirillum inoculation, as described by Zeffa et al. [30] and Vidotti [13]. The lines belong to the germplasm bank of the Universidade Estadual de Maringá (Brazil), and were obtained from successive self-pollinations of commercial hybrids with tropical background. Based on the work reported by Zeffa et al. [30], we selected the line L7 as a high responsiveness hybrid to inoculation, while L16 was taken as the low one.
As bacterial source, a liquid formulation of Azospirillum brasilense containing a blend of the AbV5 and AbV6 strains was purchased at the local market (Masterfix Gramíneas, Stoller do Brasil, Campinas, Brazil). These strains derive from a Brazilian nationwide selection program for corn, and allow for a replacement of 25% of the total input of the needed nitrogenous fertilizer [3]. The crop inoculation was manually performed by mixing 100 mL of the commercial product with 1 kg of corn seeds in order to obtain 1.2 × 10 5 of viable cells seed −1 . We tested three N management conditions: (i) full solution, i.e., supplying 2.0 mmol L −1 Ca (NO 3 ) 2 (100% N); (ii) partial N supply, i.e., only 75% of Ca(NO 3 ) 2 used in the full solution (75% N) and (iii) 75% of N + Azospirillum brasilense inoculation (75% N + Azo). Under natural light and temperature conditions, the plants were grown in a greenhouse, where the recorded daily mean temperature ranged between 25 • C and 31 • C, and the daily mean air relative humidity ranged between 62% and 78%. After 25 days, we performed the following evaluations: root volume, shoot dry mass, shoot N content, rhizosphere bacterial count (soil and root tissues) and root exudate profiling.

Plant Growth Parameters
The root volume was determined as the difference between the water volume within a graduated cylinder before and after insertion of the two-times washed fresh roots [30]. The shoot parts were stored in paper bags and dried in a forced ventilation oven at 60 • C for 72 h for subsequent determination of the dry mass and its N content by the Kjeldahl digestion method [32]. Next, after weighting, the dried shoot was ground in a mill for 60 s at 17,000 rpm; then 0.2 g of the flour was placed in test tubes containing 2 g of a catalysis (copper sulfate and selenium powder) and 5 mL of concentrated sulfuric acid. Then, the tubes were gradually heated up to 350 • C on a block digestor for 2.5 h, up to the digestion phase for organic matter. Afterwards, the released ammonia distillation phase was started by a reaction to sodium hydroxide (50%), which was then collected in 4% boric acid solution. Finally, titration was carried out in standard chlorohydric acid solution (1 mol L −1 ) and the difference of the amount of nitrogen was calculated.

Rhizosphere Colonization
For bacterial count we used the most probable number technique (MPN) described by Dobereiner et al. [33]. To evaluate the substrate colonization, 10 g of the root surrounded substrate were collected and suspended in 45 mL saline solution (NaCl 0.85%), while, for assessing the internal the colonization of internal tissues, roots tips were collected and five times washed with sterile distilled water and surface-disinfected by immersion in 70% ethanol (30 s), followed by an 8-min gentle agitation in 2% sodium hypochlorite. Then, 10 g of this root tissue were macerated and suspended in 45 mL saline solution (NaCl 0.85%).

Exudates Collection and Analyses
The exudates were collected by immersing the whole root system in 100 mL of sterile ultra-pure water (Water Ultrapur, Supelco, Darmstadt, Germany), where it remained for 3 h, under stirring in an orbital shaker (60 rpm). Then, following Kawasaki et al. [20], the liquid solution was filtered (PHENEX RC syringe filter, Allcrom, São Paulo, Brazil) and submitted high performance liquid chromatography (HPLC) to quantify the amino acids, the sugars and the organic acids contents released by each genotype. The HPLC was performed as follow: HP 1100, chromatographic column
In accordance with Kim et al. [34], we used the acetylene-reduction assay to assess the nitrogenase activity, which is an indirect method that uses the enzyme's ability to reduce acetylene gas to ethylene. Then, we incubated the NFb media for 48 h at 30 • C in the dark to then reduce their gas phase's partial pressure of oxygen with a mixture of acetylene-air-N (10:10:80 v/v/v) (Bovine Serum Albumin, Sigma-Aldrich). Finally, after 24 h, the rate of ethylene production using gas chromatograph with a flame ionization detector HP-PLOT/AL203 column, and the MPN of Azospirillum were measured.

Experimental Design and Statistical Analyses
In the greenhouse, the experiment was conducted in the completely randomized design in a 2 × 3 factorial scheme, with four replications: two genotypes of different responsiveness × three N managements. From individual HPLC results we further calculated the content of chemotactic compounds, i.e., the concentration of metabolites that serve as bacterial energy source (all organic acids + five six sugars, excluding glucose). When investigating the impact of genotype exudates on Azospirillum growth and N fixation, we tested three treatments in the completely randomized design: one neat NFb medium as a control and two supplemented NFb media, each, containing compounds representing the exudates upon 75% N. After checking the normality and homogeneity of variances, the means were compared using the Tukey test at a 5% probability level. To individually compare the content of each class of metabolites in the exudates, the F-test was conclusive.

Conflicts of Interest:
The authors declare no conflict of interest.
Appendix A Table A1. Means of asparagine, aspartic acid, serine, glutamic acid, valine, phenylalanine, threonine, tryptophan and alanine of two corn genotypes of high (HResp) and low (LResp) responsiveness to A. brasilense under full N mineral supply (100% N), partial N (75% N) as well as partial N supply coupled with inoculation (75% N + Azo). According to the F-test, means followed by the same letter in the row do not differ from each other at a 5% probability level.
Appendix B Table A2. Means of sugars (glucose, sucrose, xylose, arabinose, fructose and galactose) and organic acids (citrate, malate, succinate and fumarate) of two corn genotypes of high (HResp) and low (LResp) responsiveness to A. brasilense under full N mineral supply (100% N), partial N (75% N) as well as partial N supply coupled with inoculation (75% N + Azo).

N Management
Sugars (nmol g −1 ) Organic Acids (nmol g −1 ) HResp According to the F-test, means followed by the same letter in the row do not differ from each other at a 5% probability level.