Introduction

Plant growth promoting bacteria represent a fraction of prokaryotes that compose the wide microbial diversity of soils. The growth promotion mechanisms used by these bacteria include the ones that act directly on plant development such as biological nitrogen fixation (BNF), solubilization of insoluble inorganic phosphate, production of plant-growth-regulating substances, and mechanisms that act indirectly on plant development such as antagonism against phytopathogenic agents or antibiosis.

Some studies show that Rhizobium that belong to the nitrogen fixing nodule-forming bacterial genera can promote the solubilization of insoluble inorganic phosphate, which can then increase the production of corn (Zea mays L.) and lettuce (Lactuca sativa L.) (Chabot et al. 1996). Mesorhizobium, another nodule-forming bacterial genus that can solubilize insoluble inorganic phosphate, can promote an increase in the phosphorus content of chickpeas (Cicer arietinum L.) and barley (Hordeum vulgare L.) (Peix et al. 2001a). These bacteria are also actively involved in the synthesis of plant growth hormones such as auxin; which was confirmed in culture medium (Boeiro et al. 2007) and they promoted growth of radish plants (Raphanus sativus L.) (Antoun et al. 1998). A high positive correlation between the in vitro production of indol-3-acetic acid (IAA) derived from l-tryptophan and the yield of mustard seeds (Brassica juncea L.) has also been observed (Asghar et al. 2002).

BNF is relevant to agricultural systems and natural ecosystems because it increases the nitrogen level in these environments. Among the nitrogen-fixing bacteria, the strains that fix nitrogen through symbiosis and nodule formation in leguminous plants are highlighted; they are currently restricted to 15 genera in addition to diverse species that live in association with plants or that live in a free-living state in the soil. Among the 15 nodule-forming genera, only Burkholderia and Azorhizobium were reported as being also able to fix nitrogen in a free-living state (Dreyfus et al. 1988, 1983; Elliott et al. 2007; Moreira et al. 2006; Oliveira-Longatti et al. 2013; Silva et al. 2012).

Several studies have been performed to select bacterial strains that, when in symbiosis with cowpea, exhibit high BNF and increase their productivity. These studies reveal a wide diversity of strains with the potential for use as inoculants (Lacerda et al. 2004; Soares et al. 2006; Soares et al. unpublished results). Indeed, some Bradyrhizobium strains (UFLA 3-84, INPA 3-11B, BR3267) are already approved by the Brazilian Ministry of Agriculture, Livestock and Supply as cowpea inoculants.

There is agronomic interest in identify other important processes in these bacteria that might directly or indirectly promote the growth of leguminous and non-leguminous plants. Because most studies reporting on promoters for plant growth usually focus on the associative and endophytic rhizosphere bacteria such as Pseudomonas, Azospirillum, Pantoea, Paenibacillus and Acinetobacter (Collavino et al. 2010; Ikeda et al. 2013; Lugtenberg and Kamilova 2009; Monteiro et al. 2009; Ogut et al. 2010) there is a need for further research on symbiotic bacteria.

To perform these processes in the soil, the bacteria have to compete for resources with the huge microbial diversity and overcome the presence of a widely varying scope of antibiotics. Thus, the aim of this study was to verify the in vitro ability to perform different biotechnological processes and to display antibiotic resistance of the three strains currently approved as inoculants for cowpea, besides other 26 symbiotic strains and 2 non-symbiotic bacterial strains. All these 31 strains, including the inoculant ones, were isolated from nodules of the trap species cowpea and siratro inoculated with soils from Amazonian or from rehabilitated bauxite-extraction areas in Minas Gerais state. We also identified these strains by 16S rRNA sequencing.

Materials and methods

Strains studied

This study investigated 28 strains and UFLA 03-84, INPA 03-11B and BR3267 strains. These last three strains are approved by MAPA (Ministério da Agricultura, Pecuária e Abastecimento—Ministry of Agriculture, Livestock and Supply) as inoculant for cowpea bean cultures based in published results (Lacerda et al. 2004; Martins et al. 2003; Soares et al. 2006). Table 1 shows the origin and symbiotic and cultural characteristics of these strains. The majority of these strains are able to establish symbiosis with the cowpea bean (Soares et al. unpublished results) (Table 1). In addition to these strains, other strains were used as positive controls for some of the tests: BR5401T and ORS571T served as positive controls for free-living biological nitrogen fixation, and BR11001T and BR11080T served as positive controls for auxin production.

Table 1 Origin, symbiotic efficiency, identification and accession number of the strains UFLA 03-84, INPA 03-11B and BR 3267, which are currently approved as inoculants for the cowpea bean (Vigna unguiculata (L.) Walp) and of the 28 other strains
Full size table

Genetic identification of strains

The DNA of each strain was isolated using the ZR Fungal/Bacterial DNA Kit (Zymo Research Corp.). The 16S rRNA of the strains was amplified by PCR using the primer set 27F (5′-AGAGTTTGATCCTGGCTCAG-3′) and 1492R (5′-GGTTACCTTGTTACGACTT-3′) (Lane 1991). A 2 μL aliquot of extracted DNA was used in polymerase chain reaction (PCR) with a final volume of 50 μL per reaction. The final concentrations of the reagents, per reaction, were 0.2 μM of each primer 27F and 1492R, 2.5 μM of magnesium chloride, PCR buffer at 1×, 0.2 μM of each dNTP, and 0.02 U Taq DNA polymerase (Taq DNA polymerase, Invitrogen). The amplification reaction was carried out in an Eppendorf Mastercycle® thermocycler under the following conditions: initial denaturation at 94 °C for 5 min, 30 cycles of denaturation (94 °C for 40 s), annealing (55 °C for 40 s), extension (72 °C for 1.5 min), and one final extension at 72 °C for 7 min. An aliquot of each PCR reaction (20 μL) was analyzed using a 1 % (w/v) agarose gel with TAE buffer and ethidium bromide staining (5 μg mL−1). PCR products purification and DNA sequencing (with the 27F primer) was performed by Macrogen with a 37 30 × 1 sequencer.

Sequences were compared to GenBank database to obtain most similar accessions and their similarities.

Free-living nitrogen fixation

To evaluate free-living nitrogen fixation, the strains and positive controls BR 5401T and ORS 571T, which are Azorhizobium doebereinerae and Azorhizobium caulinodans strains (Dreyfus et al. 1988; Moreira et al. 2006), respectively, were inoculated in the centre of vials (total volume 10 mL) containing 5 mL of semi-solid, nitrogen-free LO culture medium (Dreyfus et al. 1983), which composition L−1 was: 10 g sodium lactate, 1.67 g K2HPO4, 0.87 g KH2PO4, 0.05 g NaCl, 0.1 g MgSO4·7H2O, 40 mg CaCl2, 4 mg FeCl3, 5 mg MoO4Na·2H2O, 10 mg biotin, 20 mg nicotinic acid, 10 mg pantothenic acid, 2 mL micronutrient solution (0.2 g Na2MoO4·2H2O, 0.235 g MnSO4·H2O, 0.28 g H3BO3, 0.008 g CuSO4·5H2O and 0.024 g ZnSO4·7H2O dissolved in 200 mL of distilled water), 5 mL bromothymol blue (0.5 % in 0.2 N KOH), pH 7.0. Mannitol was also tested as sole carbon source by substituting sodium lactate in the LO medium by it. Each strain was tested in triplicate (3 vials/strain). The flasks were incubated for 3–7 days in the dark at 28 °C until a typical pellicle was formed near the surface of the medium. The inoculated samples were compared to the positive controls; inoculating strains that lead to pellicle formation during this period were considered free-living nitrogen-fixing bacteria, whereas those that did not form a pellicle were not considered free-living nitrogen-fixing bacteria.

Solubilisation of insoluble calcium and aluminium inorganic phosphates

Two experiments were carried out to establish whether the strains could solubilise calcium phosphate (P–Ca) or aluminium phosphate (P–Al). Solubilising activity (solubilisation ability and potential) was evaluated in GES medium, which composition L−1 was: 10 g glucose, 0.1 g KNO3, 100 mL soil extract (The filtered supernatant of 1 kg soil in 1 L of distilled water, autoclaved and allowed to stand 48 h), 2 mL MgSO4 (10 %), 2 mL CaCl2 (1 %), 1 mL NaCl (10 %), 2 mL micronutrient solution (the same used in LO medium), 4 mL Fe-EDTA (1.64 %), and 15 g agar (Sylvester-Bradley et al. 1982). In the first experiment, P–Ca was obtained by adding 50 mL of a 10 % K2HPO4 solution and 100 mL of a 10 % CaCl2 solution in 850 mL of culture medium with nutrient composition equivalent 1 L (all autoclaved separately) to produce an insoluble phosphate precipitate. In the second experiment, 3.04 g L−1 of AlH6O12P3 was added. In the treatment containing P–Ca, the pH was adjusted to 6.8, whereas the pH was adjusted to 4.5 in the treatment with P–Al.

To obtain each inoculum, the strains were grown in liquid culture medium 79 (Fred and Waksman 1928), which composition L−1 was: 0.1 g K2HPO4, 0.4 g KH2PO4, 0.2 g MgSO4·7H2O, 0.1 g NaCl, 10.0 g mannitol and 0.4 g yeast extract, at a pH of 6.8. Saline solution (0.85 %) was added to the cultures to adjust the concentration of cells to an optical density at 600 nm (OD600) of 0.5. Twenty microlitres of cell suspension was spotted at three equidistant points on a plate containing medium with the phosphate precipitate, thus resulting in three colonies per plate, with the experiment being repeated in triplicate for each strain (three plates). The diameter of the solubilisation halo (translucent area around the colony) was measured using a digital calliper daily during 18 days. These measurements were used to obtain the solubilisation index (SI), which was determined by the following equation: S.I = Halo diameter (mm)/Colony diameter (mm) (Berraquero et al. 1976). Based on the S.I, the strains were classified as having a low (S.I < 2.00), medium (2.00 ≤ S.I < 4.00) or high (S.I ≥ 4.00) solubilisation ability. Based to the onset of solubilisation, the strains were also classified as early (when the onset of solubilisation occurred until the third day), late (when the onset of solubilisation occurred after the third day) or non-solubilising (when solubilisation was not visible within 18 days).

Production of the growth hormone auxin

To determine whether the bacteria could produce indole-3-acetic acid (IAA), the experimental strains and the positive controls Azospirillum brasilense (BR 11001T) and A. lipoferum (BR 11080T) (Radwan et al. 2002; Tarrand et al. 1978) were grown in Dygs medium, which contains L−1: 2 g glucose, 2 g malic acid, 1.5 g bacteriological peptone, 2 g yeast extract, 0.5 g K2HPO4, 0.5 g MgSO4·7H2O and 1.5 g glutamic acid. After growth, the cultures were centrifuged, resuspended and adjusted to an OD600 of 0.5 using saline solution (0.85 %), as described before. Aliquots of the bacterial solutions (500 μL) were inoculated in 20 mL of Dygs medium (without l-tryptophan or supplemented with 100 mg L−1 of l-tryptophan) and incubated for 72 h at 30 °C with constant stirring. To quantify the indolic compounds produced after this period, the cultures were centrifuged at 17,792g for 10 min and 3 mL of supernatant was removed and mixed with 2 mL of Salkowski reagent (Sarwar and Kremer 1995). This mixture was placed in the dark for 30 min to develop a pink colour, which is indicative of IAA production. The colour intensity was determined in a spectrophotometer at 535 nm, following the methods described by Asghar et al. (2002). The concentration of IAA was estimated using a standard curve previously prepared with 0, 25, 50, 100, 150, 200 and 300 μg IAA mL−1 (Sigma-Aldrich) in sterilised, uninoculated culture media (Radwan et al. 2002).

Antifungal activity

The antagonistic activity of the strains against Fusarium oxysporum f. sp. phaseoli, a common bean pathogen, was studied following the methods of Peix et al. (2001b) with some modifications. Fungal mycelia grown on 5 mm discs in PDA (Potato Dextrose Agar, Difco) with pH 6.8 were placed in the centre of a Petri dish containing the same culture medium. Isolated colonies from three bacterial strains were streaked onto three areas of the plate around the disc. Each strain was tested in triplicate (3 plates/strain). Mycelia growth in the absence of bacteria served as the control. The plates were incubated for 3–7 days at 28 °C depending on the genus. The results were evaluated by looking for a zone of inhibition in fungal growth in the presence of bacteria.

Bacterial antibiotic resistance

Bacterial resistance to different antibiotics was evaluated by saturated disc diffusion technique in Petri dishes containing 79 solid medium (15 g agar L−1, pH 6.8). The antibiotics studied were azithromycin (15 μg), streptomycin (10 μg), erythromycin (15 μg), ampicillin (10 μg), chloramphenicol (30 μg), rifamycin (30 μg), kanamycin (30 μg), nalidixic acid (30 μg), clarithromycin (15 μg), amoxicillin (10 μg), gentamicin (10 μg) and vancomycin (30 μg) (Cecon™, Brazil).

Bacteria were grown in 79 liquid medium for 3 days with constant stirring. After the incubation, 100 μL of each bacterial culture was spread onto Petri dishes containing the 79 solid medium. Subsequently, using sterile forceps, three discs saturated with different antibiotics were added to each plate. Each strain was tested in triplicate (3 plates/strain). The discs were lightly pressed and kept equidistant from one another to prevent the inhibition zones from overlapping. The plates were inverted and incubated for 3–7 days at 28 °C. After this period, the diameter of the growth inhibition halo (a translucent area around the disc) was measured using a digital calliper.

Statistical analysis

The statistical analyses were performed using the Sisvar program, version 5.3 (Ferreira 2008) and the Scott-Knott test (5 % probability) to compare mean values. All tests were completely randomised.

Results

Genetic identification of strains

The 16S rRNA sequences from the 26 bacterial isolates evaluated were aligned against sequences available from GenBank and revealed the presence of five different genera of bacteria: Pseudomonas, Enterobacter, Acinetobacter, Burkholderia, and Bradyrhizobium. Of these genera, only the last two were previously known as nodule-forming genera. Two strains (UFLA 03-05, UFLA 03-21) were not identified because the amplification of its 16S rRNA gene was not possible. The similarity among the sequences studied through GenBank access was between 98 and 100 % (Table 1).

Free-living nitrogen fixation

The strains UFLA 03-14 (Enterobacter sp.), UFLA 03-17, UFLA 04-0110, UFLA 04-242, UFLA 04-243, UFLA 04-0314, UFLA 04-546, and UFLA 04-559 (Burkholderia fungorum) were able to fix nitrogen while in their free-living state. The nitrogen-fixing strains formed a pellicle close to the LO culture medium surface in the presence of the two carbon sources; lactate and mannitol. The positive controls BR 5401T and ORS 571T did not form a pellicle when sodium lactate was replaced by mannitol.

Solubilization of insoluble calcium and aluminum inorganic phosphates

The ability to solubilize calcium and aluminum phosphates varied among strains in each genus (Tables 2, 3). Eighteen of the 31 strains (58 %), which belonged to all the five genera, could solubilize Calcium phosphate (P–Ca) in solid culture medium, and UFLA 03-15 and UFLA 04-0314 (B. fungorum) exhibited the highest solubilization indexes (SI) (2.75 and 2.25, respectively) (Table 2). All of the P–Ca solubilizing strains showed early solubilization behavior except for UFLA 03-20 and UFLA 04-242 (B. fungorum), which showed late solubilization behavior. The ability to solubilize aluminum phosphate (P–Al) was seen in 13 (42 %) strains belonging to all the five genera; all of the strains had low solubilization ability, with indexes varying between 1.00 and 1.81 mm and showed late solubilization behavior (Table 3) except for UFLA 04-559 (B. fungorum), which showed early solubilization behavior. Strains BR 3267 (B. japonicum), UFLA 03-165 (Acinetobacter genomosp.3), and UFLA 03-154, UFLA 03-20, UFLA 03-17, UFLA 04-559, and UFLA 04-242 (B. fungorum) solubilized both P–Ca and P–Al.

Table 2 Onset of solubilisation (On. Sol.) and solubilisation index (S.I) of calcium phosphate by strains grown in GES medium
Full size table
Table 3 Onset of solubilisation (On. Sol.) and Solubilisation index (S.I) of aluminium phosphate by strains grown in GES medium
Full size table

Production of the growth hormone auxin

Among the 31 strains studied, 29 (94 %) showed the ability to synthesize indole-3-acetic acid (IAA) when l-tryptophan was added to the culture medium (Fig. 1). However, 23 (74 %) of the strains and both of the control strains (BR 11080T and BR 11001T) synthesized IAA in the absence of l-tryptophan. Strains UFLA 03-153 and UFLA 03-170 (Bradyrhizobium spp.) did not produce IAA either in the presence or the absence of l-tryptophan. In the absence of l-tryptophan, UFLA 03-163, UFLA 03-172, and INPA 03-11B (Bradyrhizobium elkanii) and UFLA 04-559 (B. fungorum), UFLA 03-21 and UFLA 03-05 (unidentified) did not produce IAA. Production of IAA among the species varied between 0.00 and 12.59 μg mL−1 in culture medium that was not supplemented with l-tryptophan; and between 0.00 and 42.28 in culture medium that was supplemented with l-tryptophan. The maximum production of IAA was attained by the strain UFLA 03-14 (Enterobacter sp.), while UFLA 03-17 and UFLA 03-15 (B. fungorum) produced slightly less IAA than the positive control BR 11001T (Azospirillum brasilense) in the absence of l-tryptophan. Greater significant differences in the production of IAA were observed when the culture medium was supplemented with l-tryptophan. In media supplemented with l-tryptophan, strain UFLA 04-321 (Bradyrhizobium japonicum) exhibited a significantly larger production of IAA relative to the strain BR 11001T.

Fig. 1
figure 1

Production of indole-3-acetic acid (IAA) by strains grown in Dygs medium in either the presence of 100 mg L−1 of l-tryptophan (grey bars) or in the absence of the amino acid (black bars). Values followed by the same letter under the same treatment were not significantly different according to the Scott-Knott test at 5 % probability

Full size image

Antifungal activity and bacterial antibiotic resistance

None of the strains exhibited the potential to inhibit the growth of Fusarium oxysporum f. sp. phaseoli.

Virtually all of the strains were resistant at least to four of the 12 antibiotics studied except for UFLA 04-0314 (B. fungorum), which was resistant only to three (Table 4). Four strains, UFLA 03-153, UFLA 03-164, UFLA 03-170, and UFLA 03-172 (Bradyrhizobium), were resistant to 11 of the 12 antibiotics studied. UFLA 03-84 (Bradyrhizobium sp.) exhibited resistance to all of the antibiotics studied. In general, the Bradyrhizobium genus was the most resistant to antibiotics. The strains were resistant most often to chloramphenicol and vancomycin, followed by ampicillin. Only UFLA 03-84 and UFLA 03-172 (Bradyrhizobium spp.) were resistant to rifamycin.

Table 4 Resistance of strains to different antibiotics
Full size table

Discussion

The strains UFLA 03-153 and UFLA 03-164, which are highly efficient in BNF when grown in symbiosis with the cowpea bean (Soares et al. unpublished results) (Table 1), and the strains currently approved as inoculants, UFLA 03-84, INPA 03-11B, and BR 3267, exhibited the ability to perform additional plant-growth-promoting processes. Tables 1 and 2 show that the strain BR 3267 can solubilize P–Ca and produce IAA independently from the addition of l-tryptophan. UFLA 03-153 and UFLA 03-84 solubilized P–Al and synthesized IAA, whereas the former only synthesized IAA in the presence of l-tryptophan. UFLA 03-11B solubilized P–Ca and synthesized IAA independently with the addition of l-tryptophan. All of the strains that were efficient in BNF when grown in symbiosis with the cowpea exhibited an initial resistance to more than eight antibiotics and eventually to all 12 antibiotics studied. None of the BNF strains that grow efficiently in symbiosis with the cowpea fixed nitrogen in the free-living state. The strains UFLA 04-0110 and UFLA 04-0321, which exhibited the greatest potential to synthesize IAA in both the absence and presence of l-tryptophan, could both solubilize phosphates (P–Ca and P–Al). UFLA 04-0110 was resistant to four antibiotics, and UFLA 04-0321 was resistant to nine antibiotics.

Several authors have previously demonstrated that some strains of the genus Burkholderia can fix nitrogen when grown either symbiotically (Elliott et al. 2007; Vandamme et al. 2002) or in association with plants (Caballero-Mellado et al. 2004). However, reports on nitrogen fixation by B. fungorum species grown either in symbiosis with the common bean (Ferreira et al.2012) or in the free-living state are quite recent (Oliveira-Longatti et al. 2013; Silva et al. 2012). This strain was isolated from Amazon region by using siratro as trap species. In our work we show that B. fungorum able to establish symbiosis with cowpea is also able to fix N2 in the free living state. Our strain was isolated from bauxite mining areas. The positive controls BR 5401T and ORS 571T, which did not form a pellicle when sodium lactate was replaced by mannitol, confirm the results found by Dreyfus et al. (1998) and Moreira et al. (2006) that the strains of Azorhizobium do not use mannitol as a carbon source. The pellicle is only formed in medium with lactate.

The solubilization of P–Ca by the strains from Burkholderia and Enterobacter has also been previously reported (Collavino et al. 2010; Peix et al. 2001b). When inoculated into common bean plants, the strain SAOCV2 from the Burkholderia cepacia species caused a significant (44 %) increase in the P content of plants that were cultivated in the soil relative to plants that were cultivated in non-inoculated soil (Peix et al. 2001b). Collavino et al. (2010) showed that the strain Enterobacter aerogenes R4M-A significantly increased the phosphorus content of the aerial part of the beans grown in insoluble phosphate conditions compared to beans inoculated with the strain Burkholderia sp. R4M-F or non-inoculated plants.

The three strains currently approved for the inoculation of the cowpea bean were tested by Marra et al. (2011) in GELP culture medium (yeast extract, glucose, peptone, soil extract, and mineral salts) and UFLA 03-84 and BR 3267 showed a solubilization index (SI) of 2.12 and 1.78 for P–Ca and 1.22 and 1.20 for P–Al, respectively. INPA 03-11B did not solubilize P–Al and P–Ca medium. In our study, UFLA 03-84 did not grow in GES culture medium containing P–Ca. However, it showed a SI for P–Al similar to that reported by Marra et al. (2011). The SI for P–Ca and P–Al in the BR 3267 strain were similar between the studies. In our study, the strain INPA 03-11B grew, but did not solubilize the culture medium with P–Al, and had a SI of 1.56 for P–Ca. The main difference between these two culture media is that GELP medium contains peptone and yeast extract, which seems to affect the solublization ability of INPA 3-11B and UFLA 3-84 depending on the type of phosphate.

The presence of P–Ca-solubilizing bacteria with a high solubilization index (SI > 4) has previously been observed in bacteria from Amazonian soils. This high solubilization potential was observed by Hara and Oliveira (2004) in a non-identified nodule-forming bacterium, indicating that bacteria with high potential to solubilize insoluble phosphate might be found in Amazonian soils.

The strains of the genera Burkholderia and Bradyrhizobium that did not produce IAA in the absence of l-tryptophan probably did not have an active indole-3-pyruvate pathway (IpyA) under these culture conditions, because the ability of nodule-forming bacteria to synthesize IAA involves three synthesis pathways: indole-3-acetamide (IAM), indole-3-pyruvate (IpyA) and tryptamine (TAM) (Theunis et al. 2004). However, the IpyA pathway is independent from l-tryptophan, which is a precursor amino acid in the other two pathways. In soil conditions, the exposure of the roots to exogenous bacterial IAA could affect the plant growth in different ways including pathogenesis, growth inhibition, or phytostimulation (Spaepen et al. 2007). The IAA, as well as other hormones, stimulates plant growth only within a narrow concentration range; the lower concentrations of IAA are not effective, and the higher concentrations are toxic (Biswas et al. 2000). Studies have verified that the Rhizobium leguminosarum strain TV-13, which produces 171.17 μg mL−1 of IAA, negatively affected the growth of lettuce seedlings. Conversely, strains of Bradyrhizobium sp. (T6-4, T6-12, C-3, V-10), which produce from 1.2 to 3.3 μg mL−1 of IAA, were shown to increase the vigor of seedlings relative to treatment without inoculation (Schlindwein et al. 2008). In our study, values as high as those mentioned by other authors were not observed for the nodule-forming and endophytic bacteria, indicating that these strains might act as phytostimulants; this hypothesis needs to be verified in studies testing the effect of inoculation in plants.

Our study corroborates the study of Florentino et al. (2012) showing that strain, UFLA 03-84 (Bradyrhizobium sp.), has resistance to all of the studied antibiotics; this strain is an approved inoculant for the cowpea bean (Lacerda et al. 2004; Soares et al. 2006). This demonstrates a selective advantage over the remainder of the microorganisms, making this strain more competitive in the soil and thus indicating an indispensable trait for the establishment of symbiosis; this may explain its success as an inoculant. All of the Bradyrhizobium strains in this study, including the other two strains approved for the cowpea bean (BR 3267 and INPA 03-11B), had considerable resistance to at least eight among the 12 antibiotics tested. These results indicate that the cowpea symbiotic nitrogen-fixing strains are well adapted to overcome amensalistic relationships among the biological populations in the soil that can produce antibiotics, as eight out of the 12 antibiotics tested are produced by microorganisms.

The performance of other biotechnological processes that both contribute to plant growth and have the ability to adapt to several types of stress adds considerable value to the diazotrophic free-living and/or symbiotic bacteria. To evaluate the joint action of these processes on plant growth under the complex, heterogeneous, and dynamic conditions of the field, edaphic systems is being investigated in the next phase of this study.