Genetic and Symbiotic Diversity of Nitrogen-Fixing Bacteria Isolated from Agricultural Soils in the Western Amazon by Using Cowpea as the Trap Plant

Abstract

Cowpea is a legume of great agronomic importance that establishes symbiotic relationships with nitrogen-fixing bacteria. However, little is known about the genetic and symbiotic diversity of these bacteria in distinct ecosystems. Our study evaluated the genetic diversity and symbiotic efficiencies of 119 bacterial strains isolated from agriculture soils in the western Amazon using cowpea as a trap plant. These strains were clustered into 11 cultural groups according to growth rate and pH. The 57 nonnodulating strains were predominantly fast growing and acidifying, indicating a high incidence of endophytic strains in the nodules. The other 62 strains, authenticated as nodulating bacteria, exhibited various symbiotic efficiencies, with 68% of strains promoting a significant increase in shoot dry matter of cowpea compared with the control with no inoculation and low levels of mineral nitrogen. Fifty genotypes with 70% similarity and 21 genotypes with 30% similarity were obtained through repetitive DNA sequence (BOX element)-based PCR (BOX-PCR) clustering. The 16S rRNA gene sequencing of strains representative of BOX-PCR clusters showed a predominance of bacteria from the genus Bradyrhizobium but with high species diversity. Rhizobium, Burkholderia, and Achromobacter species were also identified. These results support observations of cowpea promiscuity and demonstrate the high symbiotic and genetic diversity of rhizobia species in areas under cultivation in the western Amazon.

INTRODUCTION

The Brazilian Amazon Forest covers the states of Acre, Amapá, Amazonas, Maranhão, Mato Grosso, Pará, Rondônia, Roraima, and Tocantins, corresponding to 60% of the national territory and an area of approximately 5,000,000 km². Although the diversity of the fauna and flora of this extensive region is well studied, little is known about its soil microbiota. The few existing studies on the subject suggest a high level of diversity among the nitrogen-fixing bacteria that nodulate different species of legumes found in this region (9, 13, 14, 18). Several studies further indicate the potential of this area to harbor currently undescribed cultivable and noncultivable prokaryotes (3, 8, 19).

Several studies that have examined the diversity of the nitrogen-fixing Leguminosae-associated nodulating bacteria have used cowpea [Vigna unguiculata (L.) Walp] as the trap plant species. Cowpea is an important agronomic plant; it is also considered promiscuous, capable of establishing symbiotic relationships with a variety of nodulating bacteria (20) at various degrees of efficiency (14). Because of symbiotic promiscuity, it has long been assumed that cowpea did not respond well to inoculation with field-selected strains. However, experiments using Amazonian strains of Bradyrhizobium have shown significant results in soils from Minas Gerais, Brazil (24). These strains are currently approved for cowpea inoculation by the Ministry of Agriculture, Livestock and Supply (Ministério da Agricultura, Pecuária e Abastecimento [MAPA]) and have been successfully tested in other parts of the country (2). Thus, evaluation of the symbiotic diversity and efficiency of native strains represents an important step toward obtaining novel inoculant strains.

Cultural characteristics have been used successfully for the initial characterization and screening of nodulating bacteria; however, molecular techniques, such as repetitive DNA sequence (BOX element)-based PCR (BOX-PCR) and 16S rRNA gene sequencing, are strongly recommended because their results are more precise in terms of identification and the evaluation of diversity.

The purpose of our study was to evaluate the cultural, genetic, and symbiotic diversity of nitrogen-fixing bacteria isolated from cowpea nodules [Vigna unguiculata (L.) Walp] taken from soils under agricultural use in the region of the upper Solimões River, western Amazon.

MATERIALS AND METHODS

Strain origin.

Strains were obtained from the area between coordinates 4°21′ to 4°26′ S and 69°36′ to 70°1′ W in the municipality of Benjamin Constant, Amazonas State, which encompasses the town of Benjamin Constant and the localities of Guanabara II and Nova Aliança. This area, known as upper Solimões, is located in the triple frontier of Brazil, Colombia, and Peru.

The sampling area includes six windows: windows 1 and 2 in Guanabara II, windows 3, 4, and 5 in Nova Aliança, and window 6 in Benjamin Constant, where several studies of biodiversity and soils have been conducted (http://www.biosbrasil.ufla.br/). These windows were chosen to include the different land use systems in the region: primary forest, secondary forest (late regeneration state), secondary forest (early regeneration state), agroforestry systems, agriculture, and pasture. In each window, sampling points were placed 100 m apart and in some cases 50 m apart, totaling 98 sampling points. Soil samples were collected in March 2004, and each composite sample consisted of 12 simple samples: four sampled in a 3-m radius and eight in a 6-m radius from the sampling point, at a depth of 0 to 20 cm.

A total of 119 strains previously isolated in 2004 from nodules were used in this study. Nodules were surface disinfected by a brief immersion in 95% alcohol, followed by a longer immersion (3 min) in H₂O₂ and washing in several rinses of sterile water (27). These nodules were obtained after the inoculation of soil samples collected from agriculture sampling points, with cowpea cultivar BR14 Mulato serving as the trap plant species. Soil samples were collected at a depth of 0.0 to 0.20 m from the following sampling points under agriculture systems: 18, 19, 21, 26, 27, 28, and 32 at window 2 (Guanabara II), 49 at window 4 (Nova Aliança), and 72 at window 5 (http://www.biosbrasil.ufla.br/). The ranges of chemical characteristics of the soil samples at a depth of 0 to 20 cm collected in this land use system were as follows: pH in water: 4.7 to 6.0; K⁺, 42 to 136 mg dm⁻³; P, 2.3 to 9.3 mg dm⁻³; S, 2.1 to 10.3 mg dm⁻³; Al³⁺, 0 to 5.4 cmol dm⁻³; Ca²⁺, 5.6 to 17.5 mg dm⁻³; and Mg²⁺, 1.6 to 3.7 mg dm⁻³. Micronutrient levels were as follows: Fe²⁺, 10.2 to 162.0 mg dm⁻³; Zn²⁺, 1.9 to 11.5 mg dm⁻³; Mn²⁺, 20.9 to 116.4 mg dm⁻³; B, 0.3 to 0.6 mg dm⁻³; and Cu²⁺, 0.7 to 1.8 mg dm⁻³. The organic matter contents varied from 1.4 to 2.2 dag · kg⁻¹, H plus Al from 2.6 to 21.4; sum of bases (SB) from 8.3 to 21.3 cmol dm⁻³; and base saturation (V) from 32.4 to 85.5%. Further details of the fertility of these soils compared with that of other local land use systems are available in the work of Moreira et al. (17). Amendments, fertilizers, or pesticides have not been applied to any of the land use systems (LUS), and there is no record of using commercial bacterial inoculants for legumes.

Legume-nodulating species of the following genera were found at soil sampling sites 18, 19, 21, 26, 27, 28, and 32 (Guanabara II): Acacia, Entada, Inga, Mimosa, Swartzia, and Tachigali, with those of the tree species Inga edulis being the most abundant. In sampling points 49 and 72 (Nova Aliança), only Piptadenia sp. occurs.

The following cultural characteristics of each strain were evaluated in petri dishes with culture medium (three petri dishes with culture medium by each strain) containing mannitol, yeast extract, mineral salts, and bromothymol blue at pH 6.8 (medium 79) (5), similar to the well known YMA (27): growth rate measured by time to appearance of isolated colonies (fast, 2 to 3 days; intermediate, 4 to 5 days; slow, 6 to 10 days; or very slow, more than 10 days); alteration of culture medium pH (acidification, alkalinization, and neutralization) according to the method of Moreira et al. (18); exopolysaccharide production (minimal, light, moderate, and heavy); and colony color according to the work of Jesus et al. (9). Only pH and growth rate were used to determine groups of phenotypic similarity. The distribution of strains in different cultural groups and relative efficiency classes was analyzed graphically using descriptive statistics.

Strain authentication and symbiotic efficiency.

To examine nodulation capacity (authentication), i.e., the ability to establish symbiosis with its original host, and the symbiotic efficiencies of the 119 nitrogen-fixing bacteria strains isolated from cowpea nodules (trap species), one experiment was performed in a greenhouse at the Laboratory of Soil Microbiology, Department of Soil Science, Federal University of Lavras, Lavras, Brazil. The experiment was conducted over a period of 35 days (3 November to 8 December 2008). During this period, the maximum daily temperature registered varied from 20 to 34°C and the relative air humidity varied from 70 to 80%.

Cowpea (BR17 Gurgueia cultivar) was grown in 500-ml recyclable amber glass bottles wrapped in aluminum foil with a 4-fold dilution of modified Hoagland nutrient solution (6). The inoculated plants and the uninoculated control plants had a low nitrogen concentration (5.25 mg · liter⁻¹) in the nutrient solution, which is considered a starting dose for, and not an inhibitor of, the process of biological nitrogen fixation. The following quantities of stock solutions were added to 4 liters of water: 0.4 ml of 236.16 g · liter⁻¹ CaN₂O₆ · 4H₂O; 0.1 ml of 115.03 g · liter⁻¹ NH₄H₂PO₄; 0.6 ml of 101.11 g · liter⁻¹ KNO₃; 2.0 ml of 246.9 g · liter⁻¹ MgSO₄ · 7H₂O; 3.0 ml of 87.13 g · liter⁻¹ K₂SO₄; 10 ml of 12.6 g · liter⁻¹ CaH₄P₂O₈ · H₂O; 200 ml of 1.72 g · liter⁻¹ CaSO₄ · 2H₂O; 1 ml of 10 g · liter⁻¹ FeCl₃; and 1 ml of micronutrients (2.86 mg · liter⁻¹ H₃BO₃; 2.03 mg · liter⁻¹ MnSO₄ · 4H₂O; 0.22 mg · liter⁻¹ ZnSO₄ · 7H₂O; 0.08 mg · liter⁻¹ CuSO₄ · 5H₂O; and 0.09 mg · liter⁻¹ Na₂MoO₄ · H₂O).

Controls without inoculation and with nitrogen supplementation were also included. In the control with nitrogen supplementation, complete Hoagland solution was used, with 52.5 mg · liter⁻¹ nitrogen.

Two strips of filter paper 2 cm wide and of a length corresponding to the height of the bottle were placed inside each bottle to promote contact between the nutrient solution and the cowpea seeds, in addition to a small amount of cotton in the mouth of the bottle to support the seed. Subsequently, all bottles were autoclaved for 40 min at 1.5 kg/cm² and 127°C.

Cowpea seeds were surface sterilized with 98% alcohol for 30 s and with 2% sodium hypochlorite for 2 min. Seeds were subsequently washed six times with sterile distilled water, immersed in water for 1 h, and then placed in petri dishes with moistened sterile cotton in a growth chamber at 28°C for 24 h or until radicle emission, at which point they were transferred to bottles containing the nutrient solution.

To generate the treatments, liquid medium 79 (5) was inoculated with bacterial cells previously grown on solid medium 79 using a platinum needle and was incubated at 28°C with constant agitation for 3 days for fast-growing strains, 5 days for intermediately growing strains, and 8 days for slow-growing strains. At planting, each seed was inoculated with 1 ml of culture containing about 10⁹ cells.

The study was completely randomized and performed in triplicate. Three positive controls inoculated with the reference strains UFLA 03-84, INPA 03-11B (24), and BR 3267 (15), which had been approved as cowpea inoculants by the Ministry of Agriculture (http://www.in.gov.br/visualiza/index.jsp?data=10/08/2004&jornal=1&pagina=17&totalArquivos=72), and two uninoculated negative controls with low and high nitrogen content (as described previously) were used in each experiment.

To evaluate the symbiotic efficiency of nitrogen-fixing bacteria, plants were harvested 35 days after the commencement of experiments to determine the dry matter of shoots (DMS), number of nodules (NN), and dry matter of nodules (DMN). After the determination of NN, the shoots and nodules were placed in paper bags and dried in a forced-air oven (65 to 70°C) to a constant weight for the determination of dry matter content. The relative efficiency (RE) of each treatment was calculated using the following formula: RE = (inoculated DMS/DMS with N) × 100, where inoculated DMS is the dry matter of shoots after inoculation with the respective strain, and DMS with nitrogen is the dry matter of shoots in the treatment that received a large amount of mineral nitrogen.

All data were tested for normality. The results were analyzed by analysis of variance (ANOVA), with the NN transformed to the square root of (x + 1) as recommended by the software program SAS Learning Edition 2.0. Mean values were grouped by the Scott-Knott test (23) at 5% significance using the software program SISVAR.

Characterization of genetic diversity by BOX-PCR.

The genetic diversity of the 62 authenticated strains was evaluated by BOX-PCR. The following type and reference strains were included: Cupriavidus taiwanensis (LMG19424^T), Burkholderia sabiae (BR3405), Azorhizobium doebereinerae (BR5401^T), Bradyrhizobium sp. (UFLA03-84), Bradyrhizobium elkanii (INPA 03-11B), Mesorhizobium plurifarium (BR3804), and Azorhizobium caulinodans (ORS571^T).

To prepare the samples, isolated colonies from strains grown in medium 79 were placed in 2-ml microtubes containing 1 ml of ultrapure sterile water, heated to 100°C for 10 min, and then cooled on ice. A 25-μl amplification reaction was carried out with the following components: 9.45 μl of ultrapure sterile water; 1.25 μl of 100 mM deoxynucleoside triphosphates (dNTPs), 5.0 μl of Gitschier 5× buffer (21), 0.4 μl of 20 mg · ml⁻¹ bovine serum albumin (BSA), 2.5 μl of 100% dimethyl sulfoxide (DMSO), 1.0 μl of 0.3 μg · μl⁻¹ BOX primer (5′-CTACGGCAAGGCGACGCTGACG-3′) (26), 0.4 μl of 5U · μl⁻¹ Taq DNA polymerase (Fermentas), and 5.0 μl of DNA, and the cycling programs were as previously described (21). The amplified fragments were separated by electrophoresis at 45 V on a 20- by 20-cm 1.5% agarose gel in 0.5× Tris-acetate-EDTA (TAE) buffer for 15 h at room temperature. The 1 kb Plus DNA ladder (Invitrogen) was used as a molecular weight marker. The gel was stained with ethidium bromide and photographed.

The genetic diversity of the strains was analyzed by the presence or absence of polymorphic bands in the gel. The data were grouped by the unweighted pair group mean arithmetic method (UPGMA) algorithm and Jaccard coefficient using the software program BioNumerics 6.5 (Applied Maths, Sint-Martens-Latem, Belgium).

Characterization of genetic diversity by sequencing of the 16S rRNA gene.

A total of 23 authenticated strains, including at least one from each of the 8 cultural groups and representatives of the nine genotypes determined by BOX-PCR at 30% similarity, were randomly selected for sequencing of the 16S rRNA gene. Bacteria were grown in medium 79 at 28°C for the predetermined growth interval of each strain until logarithm phase. Genomic DNA was extracted from cell cultures according to the protocol of the ZR Fungal/Bacterial DNA extraction kit (Zymo Research Corp).

A 5-μl aliquot of extracted DNA was added to a 50-μl PCR mixture containing 0.2 mM dNTP, 2.5 mM MgCl₂, 0.2 μM 27F primer (5′-AGAGTTTGATCCTGGCTCAG-3′) (11), 0.2 μM 1492R primer (5′-GGTTACCTTGTTACGACTT-3′) (11), 1 U Taq DNA polymerase (Fermentas), 1× PCR buffer, and ultrapure sterile water. Amplification was performed in an Eppendorf thermal cycler under the following conditions: one initial denaturation step at 94°C for 5 min; 40 cycles of denaturation at 94°C for 40 s, annealing at 55°C for 40 s, and extension at 72°C for 1.5 min; and a final extension at 72°C for 7 min. The amplified products were separated on a 1% agarose gel, stained with ethidium bromide, and visualized on a transilluminator. Purification of the products was carried out with Microcon (Millipore) filters. Sequencing was performed with primer 27F in a 3730xl sequencer.

The quality of sequences was verified using the software program Phred and submitted to BLAST for comparison with GenBank sequences (National Center for Biotechnology Information, 2010) using the Basic Local Alignment Search Tool (http://www.ncbi.nlm.nih.gov/GenBank/). Only sequences greater than 600 bp in length were used in the phylogenetic analysis. Sequence alignment was performed with the software program ClustalW, and the phylogenetic tree was constructed using the neighbor-joining method in the Kimura 2 model (10) using the parameters in the software program MEGA version 4 (25). A bootstrap confidence analysis was performed with 1,000 repetitions.

Nucleotide sequence accession numbers.

The sequences determined in this work have been deposited in GenBank under accession numbers JX284216 to JX284238.

RESULTS

The 119 strains examined here were phenotypically clustered according to growth rate (time for appearance of visible isolated colonies) and the ability to change the pH of the culture medium. A total of 11 distinct phenotypes were observed: fast-growth, total medium acidification (FA), fast-growth, no alteration of medium pH (FN), fast-growth, medium alkalinization (FAL), fast-growth, acidification localized, i.e., around the colonies (FAN), intermediate growth, total medium acidification (IA), intermediate growth, no alteration of medium pH (IN), intermediate growth, total medium alkalinization (IAL), intermediate growth, acidification localized pH, i.e., around the colonies (IAN), slow growth and medium acidificationm slow growth, no alteration of medium pH (SN); and slow growth, medium alkalinization (SAL) (Fig. 1).

Fig 1.

Distribution of the 119 strains isolated from soil under agricultural use into culture groups according to time (days) to appearance of isolated colonies and pH change of the medium: FA, fast growth, medium acidification; FN, fast growth, no alteration of medium pH; FAL, fast growth, medium alkalinization; FAN, fast growth, localized acidification, i.e., around the colonies; IA, intermediate growth, medium acidification; IN, intermediate growth, no alteration of medium pH; IAL, intermediate growth, medium alkalinization; IAN, intermediate growth, localized acidification, i.e., around the colonies; SA, slow growth and medium acidification; SN, slow growth, no alteration of medium pH; SAL, slow growth, medium alkalinization.

In a greenhouse experiment, nodulation was not observed for the control treatments (without inoculation and 52.5 mg liter⁻¹ or 5.25 mg liter⁻¹ of mineral nitrogen), indicating the absence of contamination. This result allowed the authentication of symbiosis and the evaluation of the symbiotic efficiencies of the selected strains. Our results showed that cowpea established symbiosis with 62 of the 119 strains tested (51%). Strains UFLA 03-214, UFLA 03-142, UFLA 03-200, UFLA 03-183, and UFLA 03-195, along with the reference strains INPA 03-11B and BR 3267, demonstrated the highest means for nodule numbers.

All of the inoculation treatments resulted in a shoot dry mass statistically different from that of the nitrogen control. This included those inoculated with the reference strains UFLA 03-84, INPA 03-11B, and BR 3267. The mean shoot dry mass of the control with no inoculation and a low level of mineral nitrogen (5.25 mg liter⁻¹) was 0.28 g, while that of the control with the optimal dosage of mineral nitrogen for plant development (52.5 mg liter⁻¹) was 0.95 g.

Figure 2 shows the relative efficiency (RE %) of the strains clustered into groups according to the Scott-Knott test with 5% similarity: efficient (group “b”), intermediate efficiency (groups “c” and “d”), low-efficiency (group “e”), and inefficient (groups “f,” “g,” and “h”). The last cluster represents 30% of the isolates studied; these strains did not differ significantly from each other or exhibit lower values than the control without inoculation and a low level of nitrogen. The remaining isolates were clustered with the reference strains UFLA 03-84 (group e), INPA 03-11B (group d), and BR 3267 (group b).

Fig 2.

Distribution of 62 nodulating strains into groups by relative efficiency (RE %) according to the Scott-Knott test at 5% similarity. RE = (inoculated DMS/DMS with N) × 100, where inoculated DMS is the dry matter of shoots in the treatment with inoculation of respective strain and DMS with N is the dry matter of shoots in the control treatment with mineral nitrogen. The RE of each strain was the mean for three replicates, and each replicate had one plant.

The analysis of genetic diversity using BOX-PCR revealed high diversity in 52 of the 62 strains that established symbiosis with cowpea. The DNA from UFLA 03-178, UFLA 03-181, UFLA 03-185, UFLA 03-186, UFLA 03-188, UFLA 03-189, UFLA 03-190, UFLA 03-200, UFLA 03-219, and UFLA 03-222 was not sufficiently amplified by BOX-PCR, and these strains were not included in the clustering (Fig. 3). Fifty genotypes were grouped with 70% similarity (Fig. 3), though the majority consisted of a single strain. Similarity to the reference and the type strains BR 5401^T (Azorhizobium doebereinerae), ORS 571^T (Azorhizobium caulinodans), LMG 19424^T (Cupriavidus taiwanensis), BR 3405 (Burkholderia sabiae), BR 3804 (Mesorhizobium plurifarium), UFLA 03-84 (Bradyrhizobium sp.), and INPA 03-11B (Bradyrhizobium elkanii) was lower than 50%. Only two groups with 100% similarity were formed: UFLA 03-148/UFLA 03-176 and UFLA 03-173/UFLA 03-150.

Fig 3.

Dendrogram showing the genetic similarity (based on BOX-PCR profiles) of bacterial strains that nodulated cowpea and of type and reference strains of known rhizobium species. These bacteria were isolated from soils under agricultural use in the western Amazon. Groups were obtained at 70% similarity. “∗” indicates isolates for which the 16S rRNA gene has been sequenced (see Table 1).

Figure 4 shows the comparison of 16S rRNA gene sequences from 23 strains that nodulated cowpea, representing eight cultural groups (RA, RAL, IA, IN, IAL, SA, SN, and SAL) and nine genotypes. Groups were formed according to BOX-PCR profiles with 30% similarity to the sequences of known species of the alpha- and betaproteobacteria deposited in GenBank; similarities between the studied strains and the GenBank strains ranged from 97 to 100% (Table 1). In addition to various species of Bradyrhizobium, GenBank sequence comparisons revealed strains with high similarity to Rhizobium (UFLA 2-186 and UFLA 2-188) of the alphaproteobacteria and to Burkholderia (UFLA 2-216) and Achromobacter (UFLA 03-205, UFLA 03-183, UFLA 03-206, and UFLA 03-202) of the betaproteobacteria.

Fig 4.

Phylogenetic relationships based on 16S rRNA sequences among strains isolated from cowpea nodules and strains representative of alpha- and betaproteobacteria. Phylogeny was determined by the neighbor-joining method. Bootstrap values were based on 1,000 trials.

Table 1.

Origins (sampling points), cultural characteristics, relative efficiencies, and identification (based on 16S rRNA sequences extant in GenBank) of strains that nodulate and fix nitrogen in symbiosis with cowpea

Strain	Sampling pointa	Cultural characteristicsb	RE %c	Length (bp) of 16S rRNA sequence	Most similar sequence found in GenBank
					Species	Accession no.	% similarity
UFLA 03-205	27	FAL	46b	850	Achromobacter xylosoxidans	HQ676601	100
				850	Achromobacter sp.	HM151970	100
UFLA 03-183	18	FAL	41c	855	Achromobacter xylosoxidans	HQ676601	100
				855	Achromobacter sp.	HM151970	100
UFLA 03-206	27	IA	26f	784	Achromobacter xylosoxidans	HQ676601	99
				784	Achromobacter sp.	HM151970	99
UFLA 03-202	26	IN	45b	829	Achromobacter xylosoxidans	HQ676601	100
				829	Achromobacter sp.	HM151970	100
UFLA 03-216	32	SN	26f	796	Burkholderia sp.	AY914317	97
UFLA 03-173	18	IAL	29f	737	Bradyrhizobium liaoningense	EU145999	100
				737	Bradyrhizobium yuanmingense	AB601663	100
				737	Bradyrhizobium japonicum	GU552901	100
				737	Bradyrhizobium sp.	HQ233244	100
UFLA 03-144	21	SAL	34e	813	Bradyrhizobium elkanii	GU552899	100
				813	Bradyrhizobium sp.	AB531432	100
UFLA 03-139	19A	SN	19g	780	Bradyrhizobium elkanii	GU433457	100
				780	Bradyrhizobium sp.	AB513461	100
				780	Bradyrhizobium pachyrhizi PAC48T	AY624135	100
UFLA 03-174	18	SN	40c	869	Bradyrhizobium elkanii	GU433457	100
				869	Bradyrhizobium sp.	AB513461	100
				869	Bradyrhizobium pachyrhizi PAC48T	AY624135	100
UFLA 03-143	32	SN	41c	814	Bradyrhizobium elkanii	GU552899	100
				814	Bradyrhizobium sp.	HQ233232	100
UFLA 03-149	18	SAL	13h	797	Bradyrhizobium elkanii	GU433465	100
				797	Bradyrhizobium sp.	AB513461	100
				797	Bradyrhizobium pachyrhizi PAC48T	AY624135	100
UFLA 03-214	32	SN	35d	779	Bradyrhizobium elkanii	GU433457	100
				779	Bradyrhizobium sp.	AB513461	100
				779	Bradyrhizobium pachyrhizi PAC48T	AY624135	100
UFLA 03-140	32	SN	40c	765	Bradyrhizobium elkanii	GU433457	100
UFLA 03-142	32	SN	37d	741	Bradyrhizobium elkanii	GU433457	100
UFLA 03-192	19	SAL	41c	725	Bradyrhizobium elkanii	GU433457	100
UFLA 03-182	18	SN	38d	636	Bradyrhizobium elkanii	GU433457	100
				636	Bradyrhizobium sp.	GU433446	100
UFLA 03-150	49	IAL	43b	722	Bradyrhizobium japonicum	HQ231282	100
				722	Bradyrhizobium yuanmingense	AB601663	100
				722	Bradyrhizobium liaoningense	HM446270	100
				722	Bradyrhizobium canariense	AB195986	100
				722	Bradyrhizobium sp.	DQ113663	100
UFLA 03-147	27	SN	33e	759	Bradyrhizobium sp.	EU364699	100
				759	Bradyrhizobium japonicum	FJ025100	100
				759	Bradyrhizobium iriomotense EK05T	AB300992	100
				759	Bradyrhizobium liaoningense	FJ418695	100
UFLA 03-197	21	IAL	34e	804	Bradyrhizobium japonicum	FJ025100	100
				804	Bradyrhizobium liaoningense	FJ418695	100
				804	Bradyrhizobium sp.	FJ390936	100
UFLA 03-148	26	IN	18g	745	Bradyrhizobium japonicum	GU552901	100
				745	Bradyrhizobium liaoningense	GU433468	100
				745	Bradyrhizobium yuanmingense	HM446269	100
				745	Bradyrhizobium canariense	AB195986	100
				745	Bradyrhizobium sp.	EU364719	100
UFLA 03-145	28	SN	44b	803	Bradyrhizobium japonicum	GU552901	100
				803	Bradyrhizobium liaoningense	GU433468	100
				803	Bradyrhizobium yuanmingense	AB601663	100
				803	Bradyrhizobium canariense	AB195986	100
				803	Bradyrhizobium sp.	AF514794	100
UFLA 03-186	19A	FA	32e	827	Rhizobium sp.	HM151908	99
UFLA 03-188	19A	FA	37d	725	Rhizobium sp.	JF740052	99

^aGPS (www.biosbrasil.ufla.br).

^bCultural characteristics in medium 79: FA, fast growth, medium acidification; FAL, fast growth, medium alkalinization; IA, intermediate growth, medium acidification; IN, intermediate growth, no alteration of medium pH; IAL, intermediate growth, medium alkalinization; SN, slow growth, no alteration of medium pH; SAL, slow growth, medium alkalinization.

^cMeans of relative efficiency based on SDM of inoculated treatment compared with SDM of control with mineral N by the following formula: RE = (SDM inoculated/SDM control mineral N) × 100. The same letters in the same column belong to the same group at a 5% significance level (Scott-Knott test). The RE of each strain was the mean for three replicates, and each replicate had one plant.

DISCUSSION

Cowpea is a relevant food crop, and it is extremely useful for diversity studies because of its promiscuity. Our results from the current study support the observed promiscuity of this plant species through the demonstration of high symbiotic and genetic diversity among the bacterial strains studied.

The nonauthenticated strains isolated from nodules (i.e., those strains that did not nodulate) were predominantly fast-growing and acidifying strains, indicating the presence of endophytic bacteria that grew faster than the rhizobia during the isolation process (12). The nodules were not senescent, because they were stiff with no observed decomposition and were harvested from actively growing plants.

Our results for the symbiotic efficiencies of the inoculant strain treatments, UFLA 03-84 (low efficiency), INPA 03-11B (intermediate efficiency), and BR 3267 (efficient), which should be similar to that for the treatment with a large amount of mineral N, may be related to their different tolerances for the high temperatures. During the period in which the experiments were performed, the temperature outside the greenhouse was approximately 35°C (http://www.inmet.gov.br), indicating that the indoor temperature was even higher. Temperatures above 34°C are one factor that may affect the infection process of nodulating bacteria. Plants fertilized with mineral nitrogen show a higher tolerance for abiotic stress than plants that must acquire this nutrient through biological nitrogen fixation (28), which can also explain the highest mean efficiency of the control treatment with a high mineral N supply regarding the treatments that received the efficient inoculant strains mentioned above.

The BOX-PCR results suggest a high genetic diversity among the nodulating strains isolated from soil under agricultural use and corroborate the results from previous studies, in which higher diversity was observed in cultivated lands than in primary forests (9, 13). This finding may be explained by the greater demand for nitrogen that arises in cultivated lands; demand stimulates nodulation and consequently the proliferation of rhizobia (16).

The high genetic diversity (Fig. 3 and 4) of the strains observed in the present study was similar to that reported previously (13) for the diversity of bacteria trapped by the siratro (Macroptilium atropurpureum) trap plant in the same sampling points. However, these authors also reported a higher level of diversity among the nitrogen-fixing bacteria of legumes (Bradyrhizobium, Azorhizobium, Mesorhizobium, Sinorhizobium, Rhizobium, and Burkholderia). In contrast, our study found higher diversity within the Bradyrhizobium and Achromobacter species. The higher prevalence of bacteria exhibiting slow growth and the ability to turn the pH of the culture medium alkaline or neutral is characteristic of Bradyrhizobium species that nodulate cowpea and has been observed previously (4, 30).

Window 2 (Guanabara II) contained most of the sampling points responsible for providing the largest number of strains capable of establishing symbiosis with cowpea. This area also accounts for the largest number of bacterial species identified, which in turn may be related to the high diversity of host legume plant species present in the area.

Two strains each of Bradyrhizobium (UFLA 03-145 and UFLA 03-150) and Achromobacter (UFLA 03-205 and UFLA 03-202) (Table 1) were among the strains considered to be efficient. Bradyrhizobium is also the genus to which the strains UFLA 3-84, INPA 03-11B, and BR 3267, approved as inoculants by MAPA, belong.

Strains identified as belonging to Bradyrhizobium showed 100% similarity to as many as four different species (Table 1). However, 16S rRNA gene sequencing does not offer good species-level resolution among members of the Bradyrhizobium, thus requiring further testing to identify species belonging to this genus (29). For example, during the identification of Bradyrhizobium pachyrhizi and Bradyrhizobium jicamae by 16S rRNA sequence analysis, similarities of 99.1 and 99.4% to Bradyrhizobium elkanii, respectively, were observed. Species differentiation was only possible through the phylogenetic analysis of the 16S-23S intergenic spacer (ITS) and the housekeeping genes glnII and atpD, with subsequent confirmation through homology testing (22).

The Achromobacter strains UFLA 03-202 and UFLA 03-205 were distinctive in terms of their symbiotic efficiency and clustered with the reference strain BR 3267. Benata et al. (1) were the first to report nodulation of Prosopis juliflora by a species of Achromobacter, but species of the genus were reported as human pathogens (7). Here we report the occurrence of Achromobacter as a cowpea symbiont for the first time. Further studies should be conducted to evaluate the efficacy of this symbiosis under field conditions and to verify the reliable identification to the species level.

In conclusion, the strains isolated from agricultural soils in the upper Solimões River region of the western Amazon showed high genetic and symbiotic diversity. Strains were found with an efficiency similar to those of reference strains approved for cowpea inoculation, demonstrating their potential as inoculants. BOX-PCR was found to be useful for discriminating strains and revealed high diversity among them, especially among species of Bradyrhizobium. Achromobacter species are also able to nodulate cowpea and are efficient in biological nitrogen fixation.