Gram-negative soil bacteria Bradyrhizobium spp. infects and establishes a nitrogen-fixing symbiosis with soybean, which involves the mutual exchange of diffusible signal molecules, resulting in root nodule formation (Ramongolalaina et al., 2018). The initial steps of this symbiotic interaction involve reciprocal communication by means of chemical signals produced by both members of the symbiosis. This process is triggered by the interaction between flavonoid compounds from root exudates and the protein product of the regulatory nodD gene of the bacterium (Naveed et al., 2015).

Several nod gene inducers from leguminous root exudates have been identified as isoflavone and flavanone molecules and their results is Nod factor production (Gautam et al., 2016). The isoflavonoids genistein and daidzein have been shown to be the primary inducers of nod gene expression in Bradyrhizobium japonicum (Kosslak et al., 1987). Nod factors are lipochitinoligosaccharides (LCO), which results in a wide transcriptional reprogramming that leads to coordinated multistep processes of infection and nodule organogenesis, involving many plant and bacterial molecular components (Mbengue et al., 2020).

The use of rizobia inoculants in legume seeds is maybe the oldest agrobiotechnological application. The inoculation of soybean with Bradyrhizobium spp. can fully supply the crop’s demand on N (Hungria & Mendes, 2015). Several works have demonstrated positive effects on soybean growth, nodulation and yield of inoculants based on Bradyrhizobium (Jarecki, 2020; Leggett et al., 2017; Secchi et al., 2019).

Secondary metabolites produce by rhizobia may provide an evolutionary advantage for survival of microbes in soil (Demain, 1998), they also help in the establishment of symbiotic partnerships (Brencic & Winans, 2005), and added quality to the inoculants in their effect on plants (Nápoles et al., 2016). The use of secondary metabolites of rhizobia indicates biotechnological potential to improve growth and grain yields of crops as soybean and maize (Marks et al., 2013).

Great efforts have been applied by several industries to develop new products able to attend the new requirements of the market. The tendency is to apply considerable investment in innovation, searching for new inoculant formulations to hit the ever growing market (Santos et al., 2019). The manipulation of fermentation conditions is among the suggested tools for enhancing the chances of commercial scale production and thus the field applications of these bioproducts.

The induction of LCOs or Nod factors secreted by rhizobia is between the priorities of research on improved delivery systems. Potential commercial LCO products for seed and foliar applications in legumes and nonlegumes, such as corn, are currently available in the marketplace. More exploration in this direction is required to get more benefits from plant growth promotion bacteria (Bashan et al., 2014).

Preliminary studies, have demonstrated the differentiated nod genes induction in time, pointing to the log phase. The expression of the nodD gene per cell appears to increase substantially in the early log phase to Rhizobium trifolii- induced with flavones from clover (Djordjevic et al., 1987). In Rhizobium leguminosarum associated to Pisum sativum and Lens culinaris, one study refer the best inducing time with naringenin and hesperitin at 7 h after culturing cells and the maximal inductions were obtained at 24 h after induction (Begum et al., 2001).

The objective of this study was to determine if the moment to induce the nodulation genes in Bradyrhizobium inoculants affects the synthesis of compounds related to the nodulation process and their interaction with soybean plants.

The research was developed in the company Rizobacter S.A., Argentina and in the laboratories of Fyteko S.A., Belgium, in May-June of 2017.

Bacterial strains and growth conditions

Bradyrhizobium japonicum SEMIA 5079 was obtained from the FEPAGRO (Fundação Estadual de Pesquisa Agropecuária) culture collection. The strain was selected because it is widely used in commercial B. japonicum inoculants in Brazil and Argentina (Hungria et al., 2006). Other Bradyrhizobium sp. strain (537), isolated from Argentinian soils, was used to compare. The cultures were grown at 28 °C in 100 mL of yeast mannitol medium (Vincent, 1970).

Induction experiment

To induce Nod factor (nodulation factors) and other molecules related with nodulation synthesis, the medium was supplemented by the addition of 10 μM daidzein (Sigma) at three moments: since the beginnings (time zero), at 32 hours (when bacteria were in exponential phase) and finally when inoculants were formulated. A not induced culture was used as control (NI). Flasks induced or not were shaken at 150 rpm for 60 hours on an incubator orbital shaker. Bacterial cultures (1.4 - 1.8 x 109 CFU ml^-1) were used to inoculate soybean seeds at rate of 25 mL 100 g^-1 seed.

Extraction of the lipidic fraction

Taking into account that the majority of the signals recognized for their biological activity in the rhizobia-leguminous interaction are lipid, a selective extraction of these molecules from the inoculum treated with daidzein (Induced) or not (Control) was performed, using 180 mL of n-butanol in each sample. They were put in orbital shaker at 150 rpm for 15 min and kept resting overnight at dark and room temperature of 25 ± 2 °C. The organic phase was then extracted in each sample, centrifuged at 12 000 g, 10 °C, for 10 min. All samples were concentrated by rotoevaporation at 50-80 °C until 2 mL of each was obtained, which were used for the detection of signals produced by the bacteria.

Metabolite detection in the lipidic fraction by gas chromatography–mass spectrometry (GC-MS)

Volatile derivatives of extracted metabolites were prepared by silylation, using BSTFA (N, O-Bis (trimethylsilyl) trifluoroacetamide) as reagent in combination with trimethyl chlorosilane (TMCS) (Kit BSTFA + TMCS, Supelco).

For analysis by GC-MS, a gas chromatograph coupled to a Shimadzu GC-MS QP-2010 mass spectrometer was used; system equipped with AOC-20i auto-injector, AOC-20s autosampler and a direct insertion system controlled by the GC-MS solution software. An Optima 5 MS column (30 m × 0.25 mm ID, 0.25 µm film thickness) was used.

The conditions of the chromatographic analyzes were: injector temperature 310 °C, oven temperature 100 °C during 6 min. It was subsequently increased to 320 °C at a rate of 20 °C min-1 and maintained for 5 min. The injection volume was 1 µL and the flow in the column of 0.75 mL/min, using helium as entrainment gas. All the detected signals were normalized and relative values expressed in function of the retention time (min).

Plant growth conditions

Soybean seeds cv. DM 4600 RR were inoculated with the different treatments. Non-inoculated seeds were used as control. A soil in which soybean had not been grown was used for the assay. Plants were cultivated in a growth chamber with 5000 lux of illumination, temperature of 25-30 °C and 60-70 % of relative humidity for 31 days. Plants were harvested at stage V3 (35 days after sowing); corresponding to the vegetative stage with at least three fully developed trifoliate leaves. The number and dry matter of nodules on main and total root were determined. Also shoot and root dry mass, were evaluated.

Experimental design and statistical analysis

The experiments were conducted twice. Results from the second assay are shown. Treatments were a completely randomized design with factorial arrangements composed by 2 strains x 3 moments + 2 references witness, with a total of eight treatments. Also, the control without inoculation and both strains non induced were compared. Ten replicates were used in every treatment. The Info-Stat program was used for two-ways analysis of variance. Where a significant treatment effect (p<0.05) was observed by ANOVA, Multiple ranges of Tukey HSD test was conducted to determine differences among means at p<0.05.

A representative normalized chromatogram to analyse the diversity of metabolites produced by Bradyrhizobium sp. 537 and B. japonicum 5079 without induction (NI) or induced with daidzein at time zero (0h), in exponential phase (32 h) and in the formulated (Form), is shown in Figure 1 and 2, and indicates the different metabolites profiles among the two strains.

Figure 1

Figura 1. Perfil de metabolitos en cromatografía gaseosa-espectrometría de masas (CG-EM) extraídos de la cepa Bradyrhizobium sp. 537 sin inducir (NI) o inducida con daidzeína a tiempo cero (0 h), en la fase exponencial (32 h) y en el formulado (Form). Fyteko S.A., Bruselas, Bélgica, 2017.

In spite of both studied strains belong to the same genus, their produced metabolites profile was different. Similar results were found with diverse exo-metabolomic profile in distinct strains belonging to the same family (Rhizobiaceae) and all forming nitrogen-fixing nodules on Phaseolus vulgaris roots (Ramírez-Puebla et al., 2019).

In general, the effect of strain 5079 induced at 0 h stands out in all variables, although it did not differ from its induction in the formulated. There was no correspondence between the profile of metabolites excreted by these strains and their behavior in nodulation and growth in plants. Apparently, the moment to induce the nodulation genes in these Bradyrhizobium strains affected the synthesis of compounds, but not their interaction with soybean plants.

Preliminary studies, however, have demonstrated the differentiated nod genes induction in time, pointing to the log phase. The expression of the nodD gene per cell appears to increase substantially in the early log phase to Rhizobium trifolii- induced with flavones from Clover (Djordjevic et al., 1987). In Rhizobium leguminosarum associated to Pisum sativum and Lens culinaris, one study refer the best inducing time with naringenin and hesperitin at 7 h after culturing cells and the maximal inductions were obtained at 24 h after induction (Begum et al., 2001).

According to these results, in the production of Bradyrhizobium inoculants there were no great differences if the culture is induced from the beginning, at 32 h, when it was in the exponential phase of growth, or in the final stage, during product formulation. This may be because in each of these stages the bacterium continues to multiply and induce because they were in contact with fresh nutrients from the culture medium and also with the inducer. Therefore, the cells continue dividing and accumulating LCOs and other metabolites as a result of induction. This finding could be interesting to produce soybean induced inoculants, taking into account that the moment to add the inducer is not restricted.

The excreted metabolites profile by Bradyrhizobium depended on the strain and the moment of induction.

The induction of Bradyrhizobium modulated the synthesis of excreted compounds and positively affected the nodulation and growth of soybeans.

The moment to induce the nodulation genes in Bradyrhizobium inoculants does not affect the nodulation neither soybean plants growth.