Elite rhizobia strains and growing conditions
The elite rhizobia strains used were: Bradyrhizobium elkanii – 29 W (= SEMIA 5019) and B. japonicum – CPAC 15 (= SEMIA 5079), efficient in symbiosis with soybean, selected after adaptation to Cerrado soils and released as inoculant strains in 1979 and 1992, respectively (Peres and Vidor 1980; Peres et al. 1993); as well as B. elkanii – INPA3-11B (= SEMIA 6463) and Bradyrhizobium viridifuturi symbiovar tropici – UFLA3-84 (= SEMIA 6461), efficient in symbiosis with cowpea, isolated from nodules of Centrosema sp. in Manaus and from pasture soils in the Amazon, respectively, using cowpea as a trap plant, and released as inoculant strains in 2004 (Lacerda et al. 2004). The strains were cultivated, until log phase (±109 CFU mL-1), in 79-medium (Fred and Waksman 1928), also called YMA (Vincent 1970) (10 g L-1 of mannitol, 0.01 g L-1 of K2HPO4, 0.04 g L-1 of KH2PO4, 0.05 g L-1 of MgSO4.7H2O, 0.01 g L-1 of NaCl, 0.4 g L-1 of powdered yeast extract, and 15 g L-1 of agar, with pH adjusted to 6.8-7.0), at 28°C.
In this study, we used three exopolysaccharides: PEPS, extracted from Paraburkholderia sp. – UFLA 04-269, isolated from Macroptilium atropurpureum in rocky field soils (Araújo et al. 2017); REPS2, extracted from Rhizobium tropici – UFLA 05-16, isolated from Crotalaria spectabilis in gold mining soils (Rangel et al. 2017); and REPS1, extracted from Rhizobium tropici – CIAT 899T, isolated from acid soils of South America (Graham et al. 1994). In previous studies (data not published), these three EPS exhibited physicochemical characteristics that make them excellent vehicles for the formulation of liquid inoculants.
After growth at liquid 79-medium for six days, bacterial cells were removed from the culture medium by centrifugation at 10,000 g for 10 minutes at 4°C, and cold 96% ethanol (4°C) was added to the cell-free supernatant mixture at a ratio of 3:1 (v/v) (Castellane et al., 2014). At this stage, the formation of a supernatant gel and a precipitate was immediately observed. The mixture was cooled to 4°C for 24 h. Ethanol was evaporated in a drying oven at 60°C. The precipitation solvent enabled partial purification of the polymer by eliminating the soluble components of the culture medium (Castellane & Lemos, 2007).
The precipitated product was dried to a constant weight using a Labconco FreeZone 2.5 lyophilizer to verify the number of EPS obtained (g of EPS per L of culture medium).
Antibacterial trial of the three EPS
We tested the possible antibacterial activity of the EPS to the elite inoculant strains 29 W, CPAC 15, INPA3-11B, and UFLA3-84 according to the procedures described by Bauer et al. (1966), with small modifications. In short, each strain was grown in solid 79-medium in a plate and exposed to a sterilized disk impregnated with 200 µL of exopolysaccharide solution (10 mg mL-1) and to control treatments with sterile saline solution (disk imprenated with NaCl 0.85%, w/v) and two antibiotics (chloramphenicol and sulfazotrim). Antibiotics were made available in commercial disks at the concentration of 1 mg mL -1. These antibiotics were used as positive controls of antibacterial activities. Each strain was tested in triplicate. After incubation of the dishes at 28°C over a period of 144 h, the inhibition zone diameter was measured.
Preparation of liquid polymeric formulations of an exopolysaccharide base
We used a modified 79-medium as a solution succinctly described as 40 g L-1 mannitol and 10 g L-1 glycerol as carbon sources, and 10 mg L-1 Fe-EDTA, 0.02 mg L-1 CuSO4.5H2O, and 0.01 mg L-1 H2MoO4.H2O as metal ions. After performing tests with different EPS-concentrations (2%, 1%, 0.5%, and 0.1%), we found the ideal concentration of each one of the EPS in the liquid formulations was 0.1% (w/v). We prepared the formulations from each one of the exopolysaccharides, which constituted three formulations of liquid inoculants: formulation 1 – REPS1; formulation 2 – PEPS; formulation 3 – REPS2; as well as a reference formulation with the original 79-medium (10 g L-1 mannitol) without the metal ions and the EPS. After preparation, the formulations were sterilized in an autoclave at 120°C at 1.0 kg cm2. Standardization of the concentration of the inocula was carried out as described above. Aliquots of 0.1 mL of the culture of the strains 29 W, CPAC 15, INPA3-11B, and UFLA3-84 were inoculated in each one of the formulations. After that, the formulations were transferred to cone-bottom polypropylene centrifuge tubes, with a leak-proof elongated lid and 50 mL capacity, previously sterilized by gamma radiation. The samples were kept in an orbital shaker at 110 rpm and 28°C.
Evaluation of rhizobia growth and survival in the different liquid formulations
The growth and survival of the Bradyrhizobium strains in each liquid formulation described above was evaluated at 5, 10, 20, 30, 60, and 90 days of storage at ambient temperature (28°C). The number of viable cells was determined by counting of colony forming units, using the serial decimal dilution method, in which a 1.0 mL aliquot of each formulation was successively diluted from 10-1 to 10-8 in saline solution (0.85%, w/v). After that, 20 µL aliquots were inoculated on Petri dishes with 79-medium through the microdrop technique and then dried in a bacteriological chamber for 20 minutes before being inverted and then incubated at 28°C for 144 hours. Evaluations were made in quadruplicates. The number of colony forming units per milliliter was calculated using the the following formula:
Evaluation of EPS enriched inoculants in symbiosis with cowpea and soybean in a greenhouse
The studies were conducted in a greenhouse at the Instituto Federal do Maranhão, Campus São Luís – Maracanã, São Luís, MA, Brazil, from May to June 2017 to evaluate the performance of the liquid formulations of inoculants after one month of storage. The soil used, a Latossolo Amarelo Distrófico típico (Oxisol) (Santos 2018), was collected at a depth of 0 to 20 cm. The soil was air dried, sieved in a four-millimeter mesh, homogenized, and placed in 8 kg pots. Soil samples were collected for chemical and textural characterization with the following results: pH of 4.8; organic matter of 1.97 dag kg-1; K, P, and S of 21.36, 2.78, and 4.5 mg dm-3, respectively; Ca, Mg, Al, H+Al, SB, t, and T of 0.73, 0.15, 0.70, 4.37, 0.93, 1.63, and 5.30 cmolc dm-3, respectively; V and m of 17.64 and 42.94 %, respectively; Zn, Fe, Mn, Cu, and B of 1.00, 232.74, 1.61, 0.19, and 0.07 mg dm-3, respectively; and clay, silt, and sand fractions of 10, 3, and 87 dag ka-1, respectively. Soil pH was determined in water at the soil:water proportion of 1:2.5; H+Al was determined by the Ca(OAc)2 method in 0.5 mol L-1; pH was 7.0; exchangeable Ca2+, Mg2+, and Al3+ were extracted with 1 mol L-1 KCl and determined by titration; P and K were extracted by Mehlich-1 and analyzed by colorimetry (P) and flame photometry (K); organic carbon was determined by oxidation with potassium dichromate; and Zn, Mn, and Cu were extracted by Mehlich-1 and determined by atomic absorption spectrophotometry. The values of effective CEC (t), CEC at pH 7.0 (t), sum of bases (SB), and the percentages of base saturation (V%) and of aluminum saturation (m) were indirectly obtained using the values of potential acidity, exchangeable bases, and exchangeable aluminum. Liming was performed to increase base saturation to 70%, with application of 10.5 g per pot of dolomitic limestone (32% of CaCO and 15% of MgCO – total neutralizing power (TNP) = 91%) and subjected to 90 days of incubation. After that, basic fertilization was carried out as recommended by Malavolta (1980) with all the other nutrients, except for N, through commercial compound sources. The application rates of macro- and micronutrients for each pot were 200 mg P, 150 mg K, 50 mg S, 5 mg Zn, 1.5 mg Cu, 3.6 mg Mn, 0.5 mg B, and 0.1 mg Mo per dm3 of soil.
We used a completely randomized experimental design (CRD) with seven treatments, namely, three formulations of EPS, one commercial inoculant, one reference with original 79-medium without metal ions and without EPS, one control without mineral nitrogen (-N), and one control with mineral nitrogen (+N) (300 mg dm-3 de N, using urea as a fertilizer for all the formulations), with four replications. The symbiotic efficiency of the formulations was studied in soybean (29 W and CPAC 15) and cowpea (INPA3-11B and UFLA3-84) and compared with commercial inoculants for the two crops, namely, one liquid commercial formulation for soybean (SEMIA 5079 and SEMIA 5080, with 5.6 x 109 CFU mL-1) and one solid commercial formulation (peat inoculant) for cowpea (SEMIA 6461, with 1.0 x 109 CFU g-1), as well as the reference inoculants for both the strains of soybean and cowpea, and the controls without and with mineral N. The experimental plot consisted of one 8 kg pot containing two plants. The cultivars used were BRS 9180IPRO of soybean and BRS Guariba of cowpea. Seeds were inoculated in lots of 200 seeds with 200 μL of liquid inoculant or 0.2 g of peat inoculant. The seeds inoculated with peat were pre-coated with 100 μL of adhesive sucrose solution (10 g of sucrose in 100 mL of sterile H2O).
The plants were grown in a mean temperature of 31.7°C and 67 ± 8% relative humidity. They were harvested 45 days after planting for determination of the parameters of symbiotic efficiency (number of nodules – NN; nodule dry matter – NDM; shoot dry matter – SDM; relative efficiency – RE; shoot nitrogen concentration – SNC; and shoot nitrogen accumulation – SNA).
We also conducted a study in a greenhouse for the purpose of determining the density of the native populations of rhizobia (most probable number - MPN) in the soil of the previous experiment. Leonard pots were used with sterile sand and Hoagland and Arno (1950) solution and cowpea as a trap plant. At 30 days after germination, presence of nodules was recorded and bacterial quantification estimated in 1.4 x 103 MPN, using the Cochran (1950) table.
The data were analyzed using the statistical analysis software Sisvar® (Ferreira, 2019). Before analysis of variance (ANOVA), the presupposed requirements of normality and homogeneity were checked. Firstly, when data evidenced lack of normality, they were transformed (Y+0.5)0.5. The differences between the treatments were compared by the Scott Knott test (p < 0.01 or p < 0.05).