Bradyrhizobia associated with Laburnum anagyroides, an exotic legume grown in Poland

We isolated 18 rhizobial strains from root nodules of a leguminous shrub Laburnum anagyroides (common laburnum) grown in Southeast Poland as an exotic plant. With the use of BOX-PCR fingerprinting, the isolates were clustered into 2 main groups and one separate lineage, which was congruent with the ITS-RFLP results. The phylogenetic trees constructed based on 16S rRNA and combined atpD, dnaK, glnA, and recA gene sequence data separated the representative strains into three evolutionary lineages within the Bradyrhizobium jicamae supergroup, with Bradyrhizobium algeriense and Bradyrhizobium valentinum as the closest relatives. The nodA and nifH gene phylogenies proved that the L. anagyroides symbionts carry a symbiotic gene variant known as Clade IV, representing the symbiovar retamae. Phenotypic characteristics of the isolates and reference strains are also reported. Our study of the rhizobia nodulating L. anagyroides growing in Poland complements earlier few findings on the symbiotic associations of this Genisteae species.


Introduction
Bacteria of the genus Bradyrhizobium (Jordan 1982) and at least 11 other rhizobial genera participate in the process of N 2fixing symbiosis with legume plants classified in the family Fabaceae (Lindström et al. 2015;Shamseldin et al. 2017). They can establish symbiosis with all major legume lineages of Fabaceae spp. and even non-legume Parasponia andersonii, which suggests that bradyrhizobia may have been the ancestor of all rhizobia Parker 2015). Bradyrhizobium communities are also abundant in many soils where legumes are absent, often acting as endophytes of different plants (Vanlnsberghe et al. 2015). The studies have revealed that the genus Bradyrhizobium is separated into seven lineages, including symbiotic as well non-symbiotic bacteria (Avontuur et al. 2019;Ormeño-Orrillo and Martínez-Romero 2019). A large number of species in Bradyrhizobium has been recognized, however the list of species in this genus will probably be still growing due to the rapid development of molecular tools and the huge diversity of this genus (Stępkowski et al. 2018;de Lajudie et al. 2019).
Given the widespread deforestation and the need for increasing soil fertility and quality of poorly productive and anthropogenically affected soils, more importance is now being given to woody and shrubby legumes due to their beneficial properties (Adams et al. 2010;Castro et al. 2017). Laburnum anagyroides is a large shrub from the tribe Genisteae originating from Southern and Central Europe. The partnership of L. anagyroides with N 2 -fixing rhizobia gives this plant an ability to grow in nutrient deficient soils. The common laburnum was first introduced in Poland in 1805 and is now found in many parks, open spaces, and gardens as a decorative plant. Although this plant is widespread in Poland, there have been no studies characterizing its symbiotic bacteria. A few earlier studies identified the Spanish L. anagyroides microsymbionts as Bradyrhizobium canariense (Ruiz-Díez et al. 2009) and these from Belgium and Croatia as Bradyrhizobium spp. (De Meyer et al. 2011;Parker 2015); however, further investigations are necessary to provide better understanding of the L. anagyroides Electronic supplementary material The online version of this article (https://doi.org/10.1007/s13199-020-00668-x) contains supplementary material, which is available to authorized users. microsymbiont diversity and to confirm their positive effects on plant growth.

Isolation of rhizobia from root nodules
The bacterial isolates were obtained from effective (pink) root nodules of L. anagyroides Medic. plants growing in urban green areas in South-Eastern Poland (Table 1). The root systems of 50 plants were examined and about 850 nodules were obtained for further analysis. The bacteria were isolated from surface-sterilized nodules and after sterilization prepared and plated on solid Yeast Mannitol Agar (YMA) as previous described (Somasegaran and Hoben 1994). The bacteria isolated from the nodules were purified by streaking to obtain single colonies and pure cultures were used in further experiments.

Phenotypic analysis
The culture characteristics were assessed in YM Broth (YMB) or on YMA plates. Mean generation times were determined with the spectrophotometric method (480 and 600 nm) after incubation of the bacteria in YM broth at 28°C and pH 7.0. The growth temperature was determined by incubation of the cultures in YMA pH 7.0 from 4°C to 45°C. The growth pH range was determined by incubation of the cultures in YMA at 28°C and pH from 4.0 to 10.0. Basal medium pH was adjusted to required pH with 1 M HCl or 1 M NaOH. The salt tolerance of the isolates was assayed by adding 0.5, 1, 2, and 3% NaCl to the medium. Utilization of 11 different amino acids or assimilation of 13 various carbohydrates at 1% (w/ v) as a sole nitrogen or carbohydrate sources was investigated on a modified liquid YM broth, in which yeast extract or mannitol (respectively) was replaced by the different amino acids or the respective carbohydrates. The intrinsic antibiotic resistance was tested on YMA plates containing 14 different antibiotics at the appropriate dilution.

Plant nodulation tests
The plant test was carried out as previous described (Hoagland and Arnon 1950;Vincent 1970;Somasegaran and Hoben 1994). Briefly, the seeds were sterilized, prepared, germinated and planted on a solid N-free Hoagland's medium (one per jar). After 7-14 days, healthy seedlings were inoculated with 0.5 U McFarland standard (1.5 × 10 8 CFU/ml) cells of an individual strain grown in liquid YM broth, pelleted, and suspended in sterile water. The plants were grown for a minimum of 3 and a maximum of 6 weeks in a greenhouse under natural light supplemented with artificial light (14 h day/10 h night, at 24/19°C). Then, the plants were examined for the presence of nodules, their number, size, and color, as well as appearance of the aerial parts. The tests were performed in triplicate and compared with the negative control (noninoculated plants) and positive control (reinoculated L. anagyroides).

Rep-PCR fingerprinting, ITS-RFLP, and gene sequencing
The genomic DNA was isolated from the bacteria using a Genomic Mini AX Bacteria Spin Kit (A&A Biotechnology). For rep-PCR fingerprinting, DNA was amplified by PCR with the primer BOXA 1R (Versalovic et al. 1994) as described by Kaschuk et al. (2006). To confirm the reproducibility of the BOX-PCR fingerprinting protocol, the isolates were analyzed independently three times. A computer cluster analysis of polymorphic bands was carried out using a simple matching

III
The representative strains used for further phylogenetic analysis are marked in bold coefficient and the UPGMA method in the NTSYSpc program (Sneath and Sokal 1973). PCR amplifications of the 16S-23S rRNA ITS were done using primers FGPS 1490 and FPGS 132 previously described by Laguerre et al. (1996). The PCR products were digested separately with MboI, MspI, Hin6I, HinfI, TaqI, AluI, and RsaI endonucleases and restriction fragments were separated by electrophoresis in 3% agarose gels in TBE buffer. The isolates that had identical RFLP patterns were appointed to the same rDNA genotype. The housekeeping genes coding for recombinase A (recA), 70 kDa chaperone protein (dnaK), ATP synthase subunit beta (atpD), and glutamine synthetase II (glnII) were amplified as described by Stępkowski et al. (2005). 16S rDNA and symbiotic genes nifH and nodA coding for dinitrogenase reductase and N-acyltransferase nodulation protein A, respectively, were amplified as described earlier (Kalita and Małek 2017). The PCR products were ligated into plasmid pJET1.2 (Thermo Fisher Scientific) according to the manufacturer's instructions and transformed in E. coli XL1Blue using a standard method. The inserts of recombinant clones were sequenced in Genomed (Poland).

Phylogenetic analysis
Gene sequences obtained in this study were compared to GenBank nucleotide sequences using the BLAST program. Sequence alignments were created using ClustalW at default configuration and corrected manually. Sequence identity values based on obtained multiple alignments were calculated using BioEdit software. Phylogenetic trees were inferred using maximum likelihood (ML) method in MEGA 6.06 (Tamura et al. 2013). The best-fit nucleotide substitution model for each tree was determined in MEGA 6.06. To determine the statistical support for the branches of the phylogenetic trees, 1000 bootstrap replicates of the data were analyzed.

Isolation and nodulation tests
In this study, 18 rhizobial strains were isolated from root nodules sampled from L. anagyroides grown as a decorative plant in two cities of Southeast Poland (Table 1). All the strains were able to nodulate L. anagyroides plants efficiently. As a result, the big dark green plants developed an average of 8-25 pink root nodules per plant 3-6 weeks after the inoculation with the studied isolates (Fig. 1). These plants were clearly different from the small non-inoculated control plants. Three representative isolates Ba-Z45, Ba-Z95, and Ba-Z102 also induced effective nodules on Colutea arborescens and Lupinus polyphyllus (red nodules). No symbiosis was established with Glycine max.

BOX-PCR and RFLP analysis of the 16S-23S rDNA ITS
After DNA amplification with the BOX A1R primer, genomic DNA profiles consisting of 20 polymorphic bands sized from 200 to 2000 bp were obtained for all 18 strains. For some isolates the BOX A1R-PCR generated identical DNA profiles. Figure 2 shows a dendrogram based on the analysis of BOX A1R-PCR products, which were clustered at the similarity level of 41%. At 70% similarity, the studied strains grouped into two rep-PCR clusters consisting of 14 and 3 strains and 1 independent lineage (Fig. 2). Based on the 16S-23S rDNA ITS restriction fragment length polymorphism (RFLP) analysis, 3 ITS genotypes were distinguished among the 18 isolates (Table 1). Three isolates that represent different ITS RFLP and BOX-PCR types were used in the subsequent characterization.

16S rRNA gene phylogeny
The sequences of 16S rDNA (1253 bp

Housekeeping gene phylogeny
The sequences of atpD (491 bp), dnaK (236 bp), glnII (480 bp), and recA (412 bp) of the representative isolates exhibited 94-96%, 98-99%, 96-98%, and 94-95% similarity to each other, respectively. Compared to the type strains of recognized Bradyrhizobium species, these sequences were   S1-S4). Concatenated sequences of the recA, dnaK, atpD, and glnA genes (1619 bp) for the tested isolates of L. anagyroides shared similarities with each other in a range of 95-96%, likewise with B. algeriense RST89 T . In the phylogenetic tree based on the concatenated sequences of the housekeeping genes, the isolates were clustered in one of a well-supported branch comprising B. algeriense RST89 T and Bradyrhizobium valentinum LmjM3 T (Fig. 4).

Discussion
The bacterial strains analyzed here are rhizobia obtained from nodules of L. anagyroides growing in Poland. To get a tentative insight into the genetic diversity of the isolates we performed rep-PCR fingerprinting with the BOX A1R primer, which had been shown to be a reliable typing method used for differentiation and preliminary characterization of rhizobial strain collections (Koeuth et al. 1995;Kaschuk et al. 2006;Menna et al. 2009;Cardoso et al. 2017). The BOX-PCR analysis provided the basis for separating the isolates into three distinct groups (Fig. 2, Table 1). In order to confirm the diversity pattern of the tested strains obtained by rep-PCR, PCRbased 16S-23S rDNA RFLP analysis was carried out. Previous studies have shown relevant results for searching nodule bacterial biodiversity, and selection of strains for further experiments can be provided by restriction analysis of the ITS region (e.g. Menna et al. 2009;Sachs et al. 2009;Yao et al. 2014). We identified three different RFLP-PCR profiles of 16S-23S rDNA in the analysis of all tested L. anagyroides nodule isolates, corroborating the rep-PCR results (Table 1). Phenotypic characterization of L. anagyroides nodule isolates was performed by assessment of properties that were previously described as useful for Bradyrhizobium species differentiation (Chang et al. 2011;Guerrouj et al. 2013). The three L. anagyroides strains Ba-Z45, Ba-Z95, and Ba-Z102, i.e. representatives of each rep-PCR and ITS-RFLP cluster, did not behave phenotypically as a homogenous group (Table 2). In contrast to strains of B. algeriense RST89 T (Ahnia et al. 2018) and B. icense LMTR13 T (Durán et al. 2014), all isolates studied had a shorter generation time, grew well at 37°C and 42°C but poorly at 45°C, and utilized sucrose as the sole carbon source (Table 2).
In prokaryotes, the 16S rRNA gene sequence is a key phylogenetic marker, in most cases allowing fast identification of new isolates to at least the genus level (Stackebrandt et al. 2002;Tindall et al. 2010). It is generally accepted that nearly complete 16S rRNA gene sequence identity of two bacterial strains <95% and < 98.7% is strong evidence for distinct genera and distinct species, respectively (Yarza et al. 2014). To classify our isolates to the genus level, we sequenced the 16S rRNA gene from three representative strains. The Generation time (h) in YM broth 8-9 8-9 9-10 >20 11-12 >6 Growth at:   phylogenetic analysis of the 16S rDNA sequences (Fig. 3) showed that the tested L. anagyroides nodule isolates were members of the Bradyrhizobium genus with affinity to strains referred to as the B. jicamae supergroup (Avontuur et al. 2019). The previous studies have shown that the B. jicamae supergroup strains are relatively rare and their geographic range covers mainly hot and arid areas all over the world (e.g. Durán et al. 2014;Ahnia et al. 2018;Boulila et al. 2009). In contrast, the sequences of L. anagyroides endosymbionts from Belgium (De Meyer et al. 2011) andSpain (Ruiz-Díez et al. 2009) are grouped with 16S rDNA sequences of Bradyrhizobium japonicum supergroup, pointing to their phylogenetic separateness from the bradyrhizobia isolated in Poland (Fig. 3). Likewise, the prevalence of the B. japonicum supergroup strains among Genisteae bradyrhizobia derived from Central and Northern Europe was indicated in earlier studies (Stępkowski et al. 2018). Therefore, identification L. anagyroides symbionts inhabiting areas of the temperate climate, which markedly differs from the Mediterranean climate, as members of the B. jicamae supergroup, suggest that the legume host, but not the climate type, is a crucial element in the process of the dissemination of the studied bacterial strains. It has been established that human-mediated transport of legume plants, especially seeds, plays an increasingly important role in the process of spreading particular groups of rhizobia, although it was earlier concluded that the wide geographic range detected for many bradyrhizobia might also result from natural processes of dispersal (Parker 2015).
Since the diversity in the 16S rRNA gene sequences of bradyrhizobia is too low to detect differences among species and distinguish them, the recA, dnaK, atpD, and glnA genes were sequenced subsequently to provide more discriminatory analysis of the tested L. anagyroides endosymbionts. It is generally accepted that these protein-coding genes with faster evolution rates than those of the 16S rRNA gene are alternative taxonomic markers for easy and reliable identification of bacterial isolates as well as species delineation replacing DNA:DNA hybridization (Stackebrandt et al. 2002;Glaeser and Kämpfer 2015). Multilocus sequence analysis (MLSA) encompassing combined analysis of several housekeeping genes has been widely used in studies with bradyrhizobia (e.g. Stępkowski et al. 2005;Menna et al. 2009;Lindström et al. 2015;Kalita and Małek 2017). In our studies, the Bootstrap values based on 1000 replicates >70% are indicated at the branching points. The scale bar presents the number of nucleotide substitutions per site taxonomic affiliations of the tested strains were analyzed by MLSA according to the scheme introduced by Menna and Hungria (2011) to establish the phylogenetic relations of diverse collections of Bradyrhizobium strains. Phylogenies were inferred for each atpD, dnaK, glnII, and recA loci separately (Fig. S1-S4) and for combined gene sequences (Fig. 4). In the case of housekeeping genes, their similarity values for the tested strains were much lower compared to those of the 16S rRNA gene sequences, which were identical or almost identical. However, it was earlier found that 16S rDNA could be identical among closely related species from the genus Bradyrhizobium (Degefu et al. 2018). Based on the highlevel similarity of housekeeping gene sequences and the independent phylogenetic position of the tested strains along with the B. algeriense RST89 T , recovered by MLSA, it is reasonable to regard B. algeriense as the closest phylogenetic neighbor of the tested isolates. However, MLSA revealed that the strains Ba-Z45, Ba-Z95, and Ba-Z102, representatives of the three groups delineated by BOX-PCR and ITS-RFLP, represent distinct evolutionary lineages, well differentiated from the currently named species. In fact, the values of the nucleotide identity of the concatenated sequences of recA, dnaK, atpD, and glnII between each other and B. algeriense RST89 T (95-96%) were not higher than those between the type strain of the other recognized species, e.g. B. icense LMTR 13 T /B. retamae Ro19 T (95%), B. viridifuturi SEMIA 690 T /B. tropiciagri SEMIA 6148 T (96%), and B. embrapense SEMIA 6208 T /B. tropiciagri SEMIA 6148 T (96%). Therefore, the tested L. anagyroides isolates might form a novel Bradyrhizobium species, but this should be confirmed by genomic data. In fact, due to the growing availability of genomes for rhizobia, MLSA has now been superseded by ANI comparisons between genomes, which has been proven to be adequate for taxonomic purposes (de Lajudie et al. 2019). Nevertheless, MLSA supports many phylogenetic relationships recovered in the genome-based phylogeny for Bradyrhizobium (Avontuur et al. 2019). Unfortunately, there is no defined MLSA threshold to identify bacteria at the species or genus level in rhizobia, in contrast to some other groups of bacteria, e.g. Xenorhabdus (Tailliez et al. 2010) or Bacillus (Liu et al. 2017). However, progress is expected in the next few years related to development of an online database of MLSA allowing comparison of genomic sequences for fast phylogenetic and taxonomic studies of bradyrhizobia (Azevedo et al. 2015).
The studied strains were confirmed to be closely associated with B. algeriense and B. valentinum, i.e. two species comprising nodule endosymbionts of Genisteae legumes Retama sphaerocarpa and Lupinus mariae-josephae, respectively, which grow solely in the Mediterranean (Durán et al. 2014;Ahnia et al. 2018). Additionally, a tight relationship was also found between the tested isolates and Bradyrhizobium strains nodulating wild genistoid legume shrubs Calicotome spinosa and Spartium junceum growing in the same region (Cardinale et al. 2008;Salmi et al. 2018). In our plant nodulation tests, the isolates were able to establish effective symbiosis with the native host L. anagyroides as well as C. arborescens and L. polyphyllus, but not with G. max, which was in line with the report by Ahnia et al. (2018). The data described indicate that these bacteria are not strictly specialized towards nodulation of specific plant hosts. In fact, the promiscuous type of symbiosis, widespread among bradyrhizobia, is regarded to be more ancestral than the host-limited symbiosis restricted to specific niches, which may be a result of coevolutionary relationships between microsymbionts and the host plant (Perret et al. 2000).
Unfortunately, not all L. anagyroides symbionts isolated earlier were included in the phylogenetic analysis due to the lack of complete sequence data. However, as illustrated in the dnaK and recA phylogenetic trees (Fig. S2, S3), the tested isolates, especially Ba-Z95, showed phylogenetic affinity to Bradyrhizobium spp. strains isolated from nodules of L. anagyroides in Croatia, in contrast to endosymbionts isolated from L. anagyroides nodules growing in Belgium and Spain, classified as B. canariense and Bradyrhizobium sp., respectively (Fig. 3).
Symbiotic genes are usually studied in polyphasic rhizobial research, since they allow verification of the nodulation and ni trogen fixation ability. Bacteria in the genus Bradyrhizobium have symbiotic genes displaying a high level of diversity and a different evolutionary history from their housekeeping genes, which suggests their acquisition by lateral gene transfer (LGT) (Menna and Hungria 2011;Lindström et al. 2015;Hungria et al. 2015). It has been reported that sequence analysis of nodA and nifH genes can be useful to examine the symbiotic diversity of Bradyrhizobium strains and elucidate the role of LGT in inheritance of the symbiotic loci (Degefu et al. 2018). The phylogenetic studies of nodA and nifH symbiotic genes, which revealed 16 major clades in the genus Bradyrhizobium, have facilitated the development of the phylogeographic framework for this group of bacteria (Beukes et al. 2016;Stępkowski et al. 2018). In our study the clustering results of the nodA sequences of the tested isolates were essentially similar to those obtained from the nifH gene sequence analysis (Figs. 5 and 6), but slightly different from the results of the analysis of housekeeping genes (Fig. 4). Following the classification scheme described for symbiotic genes of Bradyrhizobium symbionts, the strains nodulating L. anagyroides in Poland, like all other B. jicamae supergroup strains, appear to belong to a group referred to as Clade IV (Beukes et al. 2016).
Moreover, the phylogenetic analysis of nodA and nifH gene sequences allowed a conclusion that the studied L. anagyroides nodule isolates are part of the symbiovar retamae. This symbiovar is represented by B. retamae, B. algeriense, and B. valentinum nodulating Retama spp., R. sphaerocarpa, and L. mariae-josephae, respectively (Guerrouj et al. 2013;Durán et al. 2014;Ahnia et al. 2018). The differences in the symbiotic and core gene phylogenies as well as the average nucleotide identity (ANI) values suggest that the symbiotic genes of rhizobia from the symbiovar retamae have been horizontally transferred among different species (Ahnia et al. 2018). Multiple symbiovars, distinguishing symbiotic capabilities of rhizobia, presented within the same host suggest the promiscuity of legume species. Despite the lack of full sequence data of the symbiotic genes from other L. anagyroides symbiotic strains, it can be concluded that in its root nodules, L. anagyroides can harbor endosymbionts belonging to different symbiovar types, at least glycinearum (Bradyrhizobium sp. R-46310), genistearum (B. canariense), and retamae (the strains isolated in Poland). It was also reported by De Meyer et al. (2011) that L. anagyroides, like most of the other exotic legumes growing in Belgium, is a highly promiscuous host plant engaged in symbioses with indigenous rhizobia nodulating native legumes. This promiscuity could increase its ability to spread into new habitats; hence, L. anagyroides seems to be a good candidate for revegetation purposes. Together, our findings provide useful information on identification of the rhizobial partner of L. anagyroides, a legume originating from Southern Europe that has expanded its range into Poland. This is a necessary step for further research addressing ecological and evolutionary processes that shape these particular mutualistic relationships.
Acknowledgements The authors wish to thank dr Waldemar Kazimierczak for technical assistance.
Funding information This work was supported by the Polish Ministry of Science and Higher Education grant for statutory activity (8028) and European Union from the European Regional Development Fund under the Operational Programme Development of Eastern Poland 2007-2013 (POPW.01.03.00-06-003/09-00).
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.