Introduction

Biological nitrogen fixation (BNF) is a process in which some prokaryotes called diazotrophs transform nitrogen (N2) from the air into NH3 (Stein and Klotz 2016). As a consequence, diazotrophs provide a large nitrogen input to the soil, reducing the need for applying synthetic fertilizers in crop areas. Soybean (Glycine max. L. [Merr.]) obtains between 50 and 60% of nitrogen in average from BNF (Salvagiotti et al. 2008). This is achieved due to the symbiotic relationship between this legume and diazotrophic bacteria known as rhizobia. An infection of the soybean roots by rhizobia is preceded by a species-specific chemical communication and involves the exudation by the roots of a particular type of flavonoids, which generate chemotaxis and the release of lipo-chitooligosaccharids in rhizobia (Abdel-Lateif et al. 2012; Antolín-Llovera et al. 2012). More recently, Compton et al. (2020) challenged the role of flavonoids suggesting that other components released by the seed may be responsible for attraction. Generally, each legume species is only infected by a few species of rhizobia (Wang et al. 2019). In fact, only Bradyrhizobium and Sinorhizobium are publicly recognized as capable of establishing an effective symbiosis with soybean (Wang et al. 2019). Among Bradyrhizobium species, ten are reported as soybean-nodulating: B. elkanii, B. japonicum, B. diazoefficiens, B. daqingense, B. liaoningense, B. huanghuaihaiense, B. ottawaense, B. brasiliense, B. shewense and B. ferriligni (Amsalu et al. 2017; Costa et al. 2020; Wang et al. 2019; Yao et al. 2015).

In Uruguay, soybean accounts for more than half of the crop-cereal sowing area and is one of the most relevant products in export income (DIEA 2020; OPYPA 2020). In virtue of the environmental and economic benefits of inoculating with rhizobia, all soybean crops in our country are inoculated with a mix of B. elkanii U1301 (SEMIA587) and U1302 (SEMIA5019), which are recommended since 1984 by the Uruguayan Ministry of Livestock, Agriculture and Fisheries (Ministerio de Agricultura y Pesca, MGAP). It has been reported that rhizobia suffered a phenotypic and genotypic variation due to frequent multiplication in the laboratory and in the soil (Brunel et al. 1990; Flores et al. 1988). Provided that rhizobia persist in soils and that there is a high rate of genetic modification of the inoculated strains, new evolved strains should arise, which may be more adaptable to the local environment and likely to alter its symbiotic properties. Indeed, many authors have reported naturalized rhizobia with better symbiotic performance, isolated from the soil and nodules (Arsese et al. 2012; Chibeba et al. 2017; de Almeida Ribeiro et al. 2015). In addition, the naturalized rhizobia population could compete with inoculants in soybean nodulation, thus reducing its efficacy (Hungria et al. 1998; Iturralde et al. 2019).

Considering that Uruguay has been using soybean inoculants for many decades, the presence of naturalized soybean-nodulating rhizobia is highly possible. Thus, the objective of this study was to gain knowledge about the rhizobia population indigenous to Uruguay that had not been previously researched and is a key factor in soybean yield. More specifically, we aimed to characterize, both symbiotically and genetically, naturalized soybean-nodulating rhizobia from Uruguayan fields.

Material and methods

Rhizobial strains

Origin of commercial and naturalized rhizobial strains

Four commercial strains were used as a reference in the genetic characterization and symbiotic efficiency trial: B. elkanii strains U1301 (SEMIA587) and U1302 (SEMIA5019), recommended by MGAP for commercial soybean inoculants in Uruguay, B. japonicum strain E109 (SEMIA5085), used in commercial soybean inoculants in Argentina, and B. diazoefficiens SEMIA5080, used in Brazilian inoculants. These strains were obtained from the official collection of the Soil Microbiology Laboratory (Laboratorio de Microbiología de Suelo) of INIA Las Brujas.

Ten naturalized rhizobia strains were selected from a collection obtained by Lagurara (2018). Rhizobia were isolated from soybean nodules collected throughout six field experiments conducted in five Uruguayan departments, using soil either with or without soybean cultivation history and inoculated with reference strains U1301, U1302, SEMIA5080 and E109 (Table 1). Naturalized rhizobia were selected using PCR-BOX fingerprinting (Lagurara 2018). For this study, the isolates whose similarity to the reference strains (U1301, U1302, SEMIA5080 and E109) was less than 80% were considered as naturalized rhizobia. Ten new BOX-PCR profiles were identified and one isolate from each profile was chosen to continue its genetic characterization and symbiotic efficiency study (Table 1).

Table 1 Origin of naturalized rhizobia strains including site (with geographical coordinates), strain inoculated in the treatment plot or no inoculation, and whether it has previous history with soybean
Full size table

Growing conditions and storage of the rhizobia

Rhizobia strains were stored at −20 °C in 25% (v/v) gyicerol:TY at the Microbiology Laboratory (Laboratorio de Microbiología) of Facultad de Agronomía (Universidad de la República, Uruguay). The strains were grown on yeast extract-mannitol agar (YEMA) medium and were incubated at 28 °C between 5 and 10 days until they reached sufficient colony growth levels to observe morphological features and purity. For the DNA extraction and the inoculum preparation, reference and naturalized strains were grown in TY broth (Howieson and Dilworth 2016) for 4 days at 28 °C under stirring at 120 rpm until reaching 107 cells mL−1.

Genetic characterization

DNA extraction

DNA from four reference strains and ten naturalized strains was extracted using GeneJET Genomic DNA Purification Kit #K0721 (Thermo Scientific, USA). DNA purity was confirmed by electrophoresis in agarose 1% (w/v) gels and Tris/EDTA buffer (TE). Good View Nucleic Acid Stain (SBS Genentech, USA) was incorporated in gels to visualize fragments under UV light.

Housekeeping (16S rRNA, atpD, gyrB and rpoB) and symbiotic gene (nifH and nodY/K) sequencing

The amplification conditions of housekeeping (16S rRNA, atpD, gyrB, rpoB, recA and dnaK) and symbiotic genes (nifH and nodY/K) are shown in Supplementary Table S1 (Arsese et al. 2012; Lane 1991; Martens et al. 2008; Rivas et al. 2009; Stepkowski et al 2005). PCR products were purified using GeneJET PCR Purification Kit #K0701 (Thermo Scientific, USA), except for the nodY/K gen, which was purified using Monarch DNA Gel Extraction Kit #T1020S (Biolabs, USA). Gene purity was confirmed by electrophoresis in agarose 1% (w/v) gels and Tris/EDTA buffer (TE). Good view Nucleic Acid Stain (SBS Genentech, USA) was incorporated in gels to visualize fragments under UV light.

Genes were sequenced by Macrogen Inc. (Seoul, South Korea) using Capillary Electrophoresis Sequencing. The sequences obtained were deposited in GenBank under the following accession numbers: MT859674-MT859687 (16S rRNA), MT904311-MT904324 (gyrB) MT904325-MT904338 (rpoB), ON262778 (recA), ON262777 (dnaK) and MT904339-MT904352 (atpD), MT904353-MT904366 (nifH) and MT904367-MT904379 (nodY/K).

Phylogenetic analysis

The phylogenetic analysis of our sequences and the sequences available in GenBank (Supplementary Tables S2 and S3) were performed using MEGA-X v10 software. Sequences from several Bradyrhizobium species reported as Soybean-nodulating or not, were included, such as B. elkanii, B. japonicum and B. yuanmingense. Soybean-nodulating Sinorhizobium fredii LMG 6218T was also used as an outgroup for housekeeping genes. ClustalW was used for alignments, and phylogenetic trees were built with Maximum-Likelihood. Statistical using the Tamura-Nei model, support for tree nodes was evaluated by bootstrap analyses using 1000 replicates.

Symbiotic performance test

Seed preparation, trial management and experimental design

The ten naturalized rhizobial strains were evaluated for symbiotic efficiency in soybean in Leonard jars (Vincent 1970). The experiment was carried out in a greenhouse with a period of 12 h of light at 28 °C for 30 days in a completely randomized design including fifteen treatments with five replicates each. Soybean (Glycine max. L. [Merr.]) seeds of the Nidera 5909 variety were used. Seeds were surface-disinfected in ethanol (70%) for 1 min, then in sodium hypochlorite (2%) for 2 min and finally washed three times with sterile distilled water. Disinfected seeds were pre-germinated in a Petri dish containing 0.6% agar and incubated at 28 °C for 48 h. After germination, two seedlings were transplanted into sterilized Leonard jars, with sterilized sand in the upper part and supplemented with sterilized Broughton and Dilworth´s N-free nutrient solution (Somasegaran and Hoben 1994) in the lower part. Plants were thinned five days after sowing (DAS), leaving one plant per jar.

Five DAS, the inoculated treatments received 1 mL of a bacterial culture containing 107 CFU mL−1. The ten naturalized rhizobia were compared with three positive control treatments, inoculated separately with the reference strains Bradyrhizobium elkanii (U1301 and U1302 1:1), B. japonicum (E109), and B. diazoefficiens (SEMIA 5080), and two negative control treatments without inoculation, with or without N fertilization.

During the experiment, the contents of the jars were periodically completed with a sterilized nutrient solution according to the uptake rate of the plants. A sterilized Broughton and Dilworth´s N-free mineral solution was used to refill the Leonard jars of the inoculated treatments and the control treatment without inoculation, whereas the solution for the N-fertilized treatment included ammonium nitrate (0.35 mM). Soybean plants were collected 30 DAS.

Determination of symbiotic efficiency and statistical analysis

Red and well-developed nodules per plant were counted to determine the number of nodules (NN). Nodules, shoots and roots were dried at 65 °C and then weighed to determine nodules dry weight (NDW), shoots dry weight (SDW) and roots dry weight (RDW). Dried shoots were sent to the Animal Nutrition Laboratory (Laboratorio de Nutrición Animal, Facultad de Agronomía) to analyze total nitrogen in shoots (TNS).

Infostat software (Di Rienzo et al. 2018) and a threshold of p < 0.05 were used for all statistical analyses. Normality and homoscedasticity were verified with Shapiro–Wilk and Levene tests, respectively. An analysis of variance (ANOVA) followed by a Scott-Knoch test was used for identifying mean differences. A principal component analysis was performed to integrate NN, NDW, SDW and TNS.

Results

Genetic characterization

Phylogenetic trees for housekeeping genes 16S rRNA (799 bp), gyrB (568 bp), rpoB (422 bp) and atpD (432 bp) are available in the Supplementary material (Figs. S1–S4). The topology and resolution of 16S rRNA and rpoB phylogenetic trees were similar: four naturalized strains clustered with B. japonicum reference strains and type strains, whereas the other six were in a cluster containing B. elkanii reference strains and type strains from various species (B. elkanii, B. brasiliense, B. pachyrizi and B. ferriligni). On the other hand, the gyrB and atpD analysis distinguishes between those species, having one cluster for B. elkanii including five naturalized strains, and UYS-CA02 clustered with Bradyrhizobium sp. UFLA06-06 (as reported by de Almeida Ribeiro et al. 2015) and a B. ferriligni type strain, the latter only in gyrB phylogeny, since there are no atpD sequences available for B. ferriligni strains.

To achieve a more robust result, a multilocus sequence analysis (MLSA) was performed by concatenating four housekeeping genes sequences (16S rRNA, gyrB, rpoB and atpD) with a total of 2251 bp. The phylogenetic tree obtained is shown in Fig. 1. This phylogenetic tree showed higher bootstrap values than the separate housekeeping genes, although its tree topology was similar to the gyrB and atpD trees. Naturalized rhizobia were placed in three different clusters: five naturalized strains integrated the same cluster as the B. elkanii strains, four naturalized strains were combined with B. japonicum strains, and UYS-CA02 was clustered with Bradyrhizobium sp. UFLA06-06. The nucleotide identity of MLSA for naturalized rhizobia (UYS-CL01, UYS-CL02, UYS-CA01, UYS-CO01 and UYS-SO01) in the first group and B. elkanii U1301 and U1302 vary between 99.73 and 99.91%. In the second group, naturalized rhizobia (UYS-CL03, UYS-CO02, UYS-SJ01 and UYS-SO02) showed a greater similarity to B. japonicum LMG 6138 T (nucleotide identity 99.24–99.28%) than B. japonicum E109 (98.26–98.97%). On the other hand, UYS-CA02 was 97.57% similar to Bradyrhizobium sp. UFLA06-06. Since B. ferriligni CCBAU 51502 T could not be included in this MLSA due to the lack of its atpD sequences, another MLSA was constructed with available gene sequences from 16S rRNA, gyrB, rpoB, dnaK and recA (2383 bp) for B. ferriligni CCBAU 51502 T, Bradyrhizobium sp. UFLA06-06 and other type sequences of the Bradyrhizobium species and the ones obtained for UYS-CA02 (Fig. 2). This analysis showed nucleotide identity of 97.04% between B. ferriligni CCBAU 51502 T and UYS-CA02, 97.43% between B. ferriligni CCBAU 51502 T and Bradyrhizobium sp. UFLA06-06, and 98.8% between Bradyrhizobium sp. UFLA06-06 and UYS-CA02. Considering that UYS-CA02 is close to Bradyrhizobium sp. UFLA06-06, and both are also close to B. ferriligni CCBAU 51502 T in the second MLSA, chances are that UYS-CA02 and Bradyrhizobium sp. UFLA06-06 may belong to the same species, which could be Bradyrhizobium ferriligni.

Fig. 1
figure 1

Maximum likelihood phylogenetic tree based on the concatenated sequence data for housekeeping genes (16S rRNA, gyrB, rpoB and atpD) of soybean-nodulating rhizobia. Data from fourteen strains (indicated with an asterisk) were obtained: four commercial strains (U1301, U1302, SEMIA5080 and E109) and ten naturalized strains. Type strains from the Bradyrhizobium species were included, as well as Sinorhizobium fredii LMG 6218 T as an outgroup. Bootstrap values higher than 70, based on 1000 trials, are indicated at the nodes. Mega-X was used for this analysis

Full size image
Fig.2
figure 2

Maximum likelihood phylogenetic tree based on the concatenated sequence data for housekeeping genes (16S rRNA, gyrB, rpoB, dnaK and recA) of soybean-nodulating rhizobia. Data from naturalized strain UYS-CA02 (indicated with an asterisk) was obtained. Type strains from Bradyrhizobium species were included, as well as Sinorhizobium fredii LMG 6218 T as an outgroup. Bootstrap values higher than 70, based on 1000 trials, are indicated at the nodes. Mega-X was used for this analysis

Full size image

Symbiotic genes were also studied and trees were obtained for nifH (601 bp) and nodY/K (315 bp), as shown in Figs. 3 and 4, respectively. Two main groups can be distinguished in both trees: the first cluster (I) contains B. elkanii U1301, B. elkanii U1302, the five naturalized rhizobia clustered with B. elkanii in MLSA (UYS-CA01, UYS-CL01, UYS-CO01, UYS-CL02 and UYS-SO01), and UYS-CA02; the other cluster (II) includes B. japonicum E109, B. diazoefficiens SEMIA5080 and naturalized rhizobia similar to B. japonicum strains in MLSA (UYS-CO02, UYS-SJ01, UYS-SO02 and UYS-CL03 [nodY/K of the latter could not be amplified]). Cluster I had a nucleotide identity of 100% for both genes among their sequences. On the other hand, nucleotide identity between UYS-CA02 and B. ferriligni CCBAU 51502 T was 97.50% for nifH (there was no nodY/K sequence available for B. ferriligni CCBAU 51502 T). In cluster II, nucleotide identity was 100% between strains for nifH. For the nodY/K gene, B. japonicum E109, B. diazoefficiens SEMIA5080 and UYS-SO02 had 100% of nucleotide identity, whereas UYS-SJ01 and UYS-CO02 had 99.67% of nucleotide identity with B. japonicum E109.

Fig. 3
figure 3

Maximum likelihood phylogenetic tree based on a nifH sequence of soybean-nodulating rhizobia. Data from fourteen strains (indicated with an asterisk) were obtained: four commercial strains (U1301, U1302, SEMIA5080 and E109) and ten naturalized strains. Different strains from Bradyrhizobium species were included. Bootstrap values higher than 70, based on 1000 trials, are indicated at the nodes. Mega-X was used for this analysis

Full size image
Fig. 4
figure 4

Maximum likelihood phylogenetic tree based on the intergenic region of a nodY/K sequence of soybean-nodulating rhizobia. Data from 13 strains (indicated with an asterisk) were obtained: four commercial strains (U1301, U1302, SEMIA5080 and E109) and nine naturalized strains. Different strains of the Bradyrhizobium species were included. Bootstrap values higher than 70, based on 1000 trials, are indicated at the nodes. Mega-X was used for this analysis

Full size image

Symbiotic efficiency

Nodulation, biomass and nitrogen trial results are shown in Table 2. The highest NN and NDW values were obtained from treatments inoculated with a mix of B. elkanii U1301:U1302 (1:1), a commercial strain used in Uruguay, and inoculated separately with UYS-CA02, UYS-SO01, UYS-CO01, UYS-CL01, UYS-CA01 and UYS-CL02. In the case of SDW and TNS, the treatment fertilized with nitrogen had the best performance. However, among inoculated treatments plants inoculated with B. elkanii U1301:U1302 (1:1) produced similar SDW and TNS than those inoculated with four other naturalized strains (UYS-CO01, UYS-CL01, UYS-CA01 and UYS-CL02), which were significantly higher than treatments inoculated with B. japonicum E109, B. diazoefficiens SEMIA5080, UYS-SO02, UYS-CL03 and UYS-CO02.

Table 2 Number of nodules (NN, no. plant−1), nodule dry weight (NDW, g plant−1), shoot dry weight (SDW, g plant−1) and total nitrogen in shoots (mg plant−1) of soybean in a Leonard jar trial
Full size table

To visualize all the variables in the same plot, a principal component analysis (PCA) was performed (Fig. 5). The PCA was able to represent 98.5% of the variability and show two different groups of treatments. The mix of B. elkanii U1301:U1302, UYS-CA01, UYS-CL01, UYS-CO01, UYS-CA02, UYS-SO01 and UYS-CL02 appear on the right side of CP1, producing a better symbiotic performance, and the treatments inoculated with B. japonicum E109, B. diazoefficiens SEMIA5080, UYS-SO02, UYS-CO02, UYS-CL03 and UYS-SJ01 appear on the left side. In summary, the PCA showed a correlation with clusters formed in the trees of symbiotic genes (nifH and nodY/K).

Fig. 5
figure 5

Principal component analysis of symbiotic parameters including the number of nodules (NN), nodule dry weight (NDW), shoot dry weight (SDW) and total nitrogen in shoots (TNS). Points for treatments inoculated with B. elkanii or Bradyrhizobium sp. UYS-CA02 are shown in black, and those inoculated with B. japonicum or B. diazzoefficens in gray

Full size image

Discussion

Genetic characterization

While the 16S rRNA gene is widely used as a molecular marker in prokaryotes taxonomy (Howieson and Dilworth 2016), our results confirm that it fails in distinguishing between Bradyrhizobium species, in agreement with previous reports by Delamuta et al. (2012) and Menna et al. (2009). In contrast, gyrB and atpD genes were more sensitive than 16S rRNS and rpoB, in line with results reported by de Almeida Ribeiro et al. (2015). Indeed, the gyrB and atpD phylogenetic trees succeed in separating B. elkanii from other similar species (B. ferriligni, B. brasiliense and B. pachyrizi). MLSA had the same topology as the atpD gene, but it presented a more robust result, which makes it a useful tool to identify Bradyrhizobium species (Arsese et al. 2012; Chibeba et al. 2017; de Almeida Ribero et al. 2015; Delamuta et al. 2017;).

The MLSA allowed us to identify five naturalized strains as belonging to B. elkanii and four to B. japonicum. These species are the ones most reported worldwide as naturalized strains isolated from soybean nodules (Arsese et al. 2012; Chibeba et al. 2017; de Almeida Ribeiro et al. 2015; Helene et al. 2020; Iturralde et al. 2019; Mason et al. 2017), probably because strains of these species are used as soybean inoculants (Hungria et al. 2005). As for the other naturalized strain (UYS-CA02) and another soybean-nodulating strain isolated by de Almeida Ribero et al. (2015) (Bradyrhizobium sp. UFLA06-06), we suggest that they may belong to B. ferriligni, due to similarities found in the second MLSA analysis. This would be the first report of rhizobia belonging to B. ferriligni isolated from soybean nodules, although B. ferriligni CCBAU 51502 T isolated from nodules of Erythriphleum fordii has been reported as capable of nodulating Glycine max (Yao et al. 2015). Considering that soybean is a promiscuous plant that nodulates with a wide variety of species such as B. elkanii, B. japonicum, B. diazoefficiens, B. daqingense, B. liaoningense, B. huanghuaihaiense and B. ottawaense (Wang et al. 2019), nodulation by new species is expected to occur. For instance, Costa et al. (2020) reported new strains isolated from soybean nodules belonging to B. brasiliense (symbiovar Sojae), although B. brasiliense type strains were isolated from nodules of Vigna unguiculata (Costa et al. 2017).

In addition, the nifH sequence of Bradyrhizobium sp. UYS-CA02 was identical to that of the B. elkanii type and naturalized strains identified by the MLSA, while B. ferriligni CCBAU 51502 T was in the same cluster with only a few differences in its sequence. On the other hand, B. japonicum-like and type strains were grouped in a different cluster, including B. diazoefficiens strains. A similar phylogeny was obtained for the nodY/K tree. Many authors have reported that nodulation and fixation genes present identical taxonomy (Arsese et al. 2012; Delamuta et al. 2017; Mason et al. 2017; Tartaglia et al. 2019). Our results suggest a monophyletic origin of symbiotic genes for B. elkanii- and B. japonicum-like strains, while UYS-CA02 or its ancestor may have received symbiotic genes from B. elkanii strains by horizontal transfer (Arsese et al. 2012; Menna and Hungria 2011). Symbiotic genes are usually located in symbiotic islands that have a great number of insertions and a high horizontal transfer rate (Kaneko et al. 2011).

Symbiotic efficiency

In light of our results, the mix of B. elkanii U1301 (SEMIA587) and U1302 (SEMIA5019) strains that are recommended for soybean inoculants in Uruguay by MGAP had a better symbiotic performance in axenic conditions than the other evaluated commercial strains (B. japonicum E109 and B. diazoefficiens SEMIA5080). Although this result is consistent with de Paiva Barbosa et al. (2017) and Okito et al. (2004), who evaluated soybean plant in axenic conditions inoculated with B. diazoefficiens SEMIA5080 and B. elkanii U1302 (the former also included B. elkanii U1301), it differs from the results of Chibeba et al. (2017), who reported a higher shoot biomass for plants inoculated with B. diazoefficiens SEMIA5080 than those inoculated with B. elkanii U1301 and U1302. Although all these trials were performed in similar and controlled conditions, the differences between results could be attributed to the different soybean varieties (Chibeba et al. 2017) or to genetic modification during trials (Ribeiro Torres and Kaschuk 2012).

In addition to the evaluated biomass and nodulation parameters, PCA appears as an efficient tool for visualizing and integrating all the parameters more simply (Chibeba et al. 2017). Indeed, this shows a better performance of B. elkanii-like naturalized strains, including Bradyrhizobium sp. UYS-CA02, beyond B. japonicum-like naturalized strains and B. diazoefficiens SEMIA5080. This result is consistent with clusters formed in the phylogenetic trees of symbiotic genes. On the other hand, Hungria et al. (1998) evaluated a great number of B. elkanii and B. japonicum commercial and naturalized strains in axenic conditions. Their results show a better symbiotic performance in soybean plants for B. japonicum strains than for those of B. elkanii; as noted above, differences in varieties or genetic modifications may account for the contradictions with our results. Furthermore, the high similarity in genetic characterization between commercial and naturalized strains suggests that the latter may stem from inoculants applied for many years, which is supported by a similar symbiotic performance (Ribeiro Torres and Kaschuck 2012). Finally, we highlight the considerably good symbiotic performance of B. elkanii-like naturalized strains that were as efficient as the commercial inoculants used in Uruguay. Other authors have reported naturalized rhizobia with the same or even better results than commercial strains (Chibeba et al. 2017; Iturralde et al. 2019), which encourages research on naturalized strains as an alternative to the inoculants currently in use. Considering that the capacity of rhizobia to promote plant growth is associated to competitiveness and environmental adaptation, symbiotic efficiency should be tested in soil assays, including competence performance, as well as studying the tolerance to abiotic factors such as temperature, salinity and pH with in-vitro assays (Chibeba et al. 2017; Leite et al. 2018). This information may contribute to the characterization of these native strains, and it represents a possible strategy for increasing soybean yields since naturalized rhizobia can be adapted to local conditions.

Conclusions

Using MLSA, we were able to identify five likely naturalized strains belonging to the B. elkanii species and four to the B. japonicum species. The other naturalized strain is phylogenetically related to B. ferriligni and may belong to that species, while having received symbiotic genes from B. elkanii strains by horizontal transfer. In addition, symbiotic genes show two main clusters that were correlated with nodulation and nitrogen fixation in soybean. It is also worth noting that B. elkanii likely naturalized strains may have derived from commercial inoculant used in Uruguay due to their high identity in housekeeping and symbiotic genes, and also a similar symbiotic performance in axenic conditions. Differences among strains can be associated with the adaptation to edaphic environments, some of them are being addressed by our group in new studies.