Analysis of the Interaction between Pisum sativum L. and Rhizobium laguerreae Strains Nodulating This Legume in Northwest Spain

Pisum sativum L. (pea) is one of the most cultivated grain legumes in European countries due to the high protein content of its seeds. Nevertheless, the rhizobial microsymbionts of this legume have been scarcely studied in these countries. In this work, we analyzed the rhizobial strains nodulating the pea in a region from Northwestern Spain, where this legume is widely cultivated. The isolated strains were genetically diverse, and the phylogenetic analysis of core and symbiotic genes showed that these strains belong to different clusters related to R. laguerreae sv. viciae. Representative strains of these clusters were able to produce cellulose and cellulases, which are two key molecules in the legume infection process. They formed biofilms and produced acyl-homoserine lactones (AHLs), which are involved in the quorum sensing regulation process. They also exhibited several plant growth promotion mechanisms, including phosphate solubilization, siderophore, and indole acetic acid production and symbiotic atmospheric nitrogen fixation. All strains showed high symbiotic efficiency on pea plants, indicating that strains of R. laguerreae sv. viciae are promising candidates for the biofertilization of this legume worldwide.


Introduction
Pisum sativum L. (pea) is one of the most cultivated legumes worldwide due to the high protein content and quality of its seeds [1], with a cultivated area higher than 8 million ha and a production of over 16 million tonnes per year, about 44% of which corresponds to Europe [2]. Spain is within the top 10 countries producing dry peas, with a production in 2018 of 262,567 tonnes of dry peas and 109,270 tonnes of green peas. Around 30% were produced in the Castilla y León region, the Valladolid province being the highest producer of dry peas within this region with 39,980 tonnes (52% of the total production) according to the Spanish Ministry of Agriculture, Fishing and Food

RAPD Fingerprinting
The RAPD fingerprinting has been widely used for the diversity analysis of fast-growing rhizobia, particularly those isolated from different legumes of the Vicia cross-nodulation group [16][17][18][19][20]. Therefore, we analyzed the RAPD patterns displayed by 24 rhizobial strains isolated from P. sativum nodules [12]. The results of this analysis showed that the strains from this study were genetically diverse and distributed into six groups with similarity values lower than 70% (Figure 1). A representative strain from each one of these groups was subjected to gene sequencing.

Analysis of the rrs Gene
The rrs gene has been sequenced in many strains nodulating Pisum and other legumes from the same cross-inoculation group, such as Vicia, Lens, and Lathyrus, some of them isolated in different Spanish regions [20,21]. These strains belong to the phylogenetic group of R. leguminosarum, which contains several species whose type strains have identical rrs genes ( Figure 2). However, slight differences have been found in the rrs gene sequences of other strains as in the case of strains belonging to the species R. laguerreae in which some strains harbored an insert of about 70 bp at the beginning of their rrs genes [21]. The results of this study showed that the analyzed strains belong to different phylogenetic clusters, each of which also contain other strains isolated from Pisum, Vicia, or Lens in different countries and continents ( Figure 2).

.2. Analysis of the rrs Gene
The rrs gene has been sequenced in many strains nodulating Pisum and other legumes from the ame cross-inoculation group, such as Vicia, Lens, and Lathyrus, some of them isolated in different panish regions [20,21]. These strains belong to the phylogenetic group of R. leguminosarum, which ontains several species whose type strains have identical rrs genes ( Figure 2). However, slight ifferences have been found in the rrs gene sequences of other strains as in the case of strains elonging to the species R. laguerreae in which some strains harbored an insert of about 70 bp at the eginning of their rrs genes [21]. The results of this study showed that the analyzed strains belong to ifferent phylogenetic clusters, each of which also contain other strains isolated from Pisum, Vicia, or ens in different countries and continents ( Figure 2).
All strains from this study presented differences in their rrs genes with respect to those of the ype strains of the species from the R. leguminosarum group. Concretely, the rrs genes of the strains MPS34, AMPS17, AMPS23, and AMPS04, representative of RAPD groups I, IV, V, and VI, espectively, have two different nucleotides with respect to these strains (positions 942 and 955, aking the rrs gene sequence of R. laguerreae FB206 T with accession number JN558651 as a reference). he rrs gene sequences of these strains were identical to those of several strains of R. laguerreae solated from nodules of Vicia and Lens in different countries, highlighting the strain MLS17 isolated All strains from this study presented differences in their rrs genes with respect to those of the type strains of the species from the R. leguminosarum group. Concretely, the rrs genes of the strains AMPS34, AMPS17, AMPS23, and AMPS04, representative of RAPD groups I, IV, V, and VI, respectively, have two different nucleotides with respect to these strains (positions 942 and 955, taking the rrs gene sequence of R. laguerreae FB206 T with accession number JN558651 as a reference). The rrs gene sequences of these strains were identical to those of several strains of R. laguerreae isolated from nodules of Vicia and Lens in different countries, highlighting the strain MLS17 isolated from nodules of L. culinaris in Spain [21] and the strain FB403 included in the original description of R. laguerreae, which was isolated from V. faba nodules in Tunisia [22]. The rrs gene sequence of strain AMPS05, representative of RAPD type II, presented four different nucleotides (positions 67, 69, 942, and 955) and contains an intervening sequence (IVS) of about 70 bp located at the beginning of this gene between nucleotides 70 and 143. The rrs gene sequence and the IVS of strain AMPS05 were identical to those found in strains R. laguerreae MLS05 and MLSC04, which nodulated L. culinaris in Spain [21], in strain CVIII4, included in the original description of R. laguerreae, which nodulated V. sativa in Spain [22], and in other strains isolated from nodules of Pisum, Vicia, and Lens in different countries ( Figure 2). Finally, the rrs gene sequence of strain AMPS22, representative of RAPD group III, has a different nucleotide at the position 1070, and it was identical to those of two strains isolated from nodules of P. sativum in India and a strain isolated from the same host in Russia ( Figure 2). identical to those found in strains R. laguerreae MLS05 and MLSC04, which nodulated L. culinaris in Spain [21], in strain CVIII4, included in the original description of R. laguerreae, which nodulated V. sativa in Spain [22], and in other strains isolated from nodules of Pisum, Vicia, and Lens in different countries ( Figure 2). Finally, the rrs gene sequence of strain AMPS22, representative of RAPD group III, has a different nucleotide at the position 1070, and it was identical to those of two strains isolated from nodules of P. sativum in India and a strain isolated from the same host in Russia ( Figure 2). It is remarkable that, although most strains with rrs gene sequences identical to those of our strains have been assigned to the species R. laguerreae, some of them are named R. leguminosarum or Rhizobium sp. The elucidation of the taxonomic affiliation of these strains cannot be based on the rrs gene sequences since they are identical in several type strains of different species, namely R. leguminosarum, R. laguerreae, R. anhuiense, R. sophorae, R. acidisoli, R. hidalgonense, R. ruizarguesonis, and R. indicum ( Figure 2). It should be based on the analysis of housekeeping genes, such as recA and atpD, which allow us to differentiate among species with identical rrs gene sequences. In this study, we revised the affiliation of some of these strains for which the recA and atpD gene sequences are currently available.

Analysis of recA and atpD Genes
The atpD and recA genes were the first housekeeping genes applied to species differentiation within the family Rhizobiaceae [23] and therefore have been sequenced for the type strains of all Rhizobium species described to date, allowing for the assignment of new isolates to these species. In addition, their sequences are available for many other strains, particularly for those nodulating P. It is remarkable that, although most strains with rrs gene sequences identical to those of our strains have been assigned to the species R. laguerreae, some of them are named R. leguminosarum or Rhizobium sp. The elucidation of the taxonomic affiliation of these strains cannot be based on the rrs gene sequences since they are identical in several type strains of different species, namely R. leguminosarum, R. laguerreae, R. anhuiense, R. sophorae, R. acidisoli, R. hidalgonense, R. ruizarguesonis, and R. indicum ( Figure 2). It should be based on the analysis of housekeeping genes, such as recA and atpD, which allow us to differentiate among species with identical rrs gene sequences. In this study, we revised the affiliation of some of these strains for which the recA and atpD gene sequences are currently available.

Analysis of recA and atpD Genes
The atpD and recA genes were the first housekeeping genes applied to species differentiation within the family Rhizobiaceae [23] and therefore have been sequenced for the type strains of all Rhizobium species described to date, allowing for the assignment of new isolates to these species. In addition, their sequences are available for many other strains, particularly for those nodulating P. sativum and other legumes from the Vicia cross-inoculation group in different geographical locations ( Figure 3).   The results of the analysis of recA and atpD genes showed that the strains analyzed in this work present similarity values higher than 96% and 98%, respectively, with respect to the type strain of R. laguerreae (Table 1). These values showed that all of these strains belong to the species R. laguerreae, even though they matched with score values lower than 2.3 with the type strain of this species after MALDI-TOF MS analysis [12]. The phylogenetic analysis of the concatenated recA and atpD genes showed that the analyzed strains belong to different clusters, some of which also contain strains included in the official description of the species R. laguerreae [22]. The strain AMPS17 belongs to the same phylogenetic cluster of the type strain of this species FB206 T , strains AMPS05 and AMPS34 belong to that of the strain CVIII4, and strains AMPS04 and AMPS23 belong to that of the strain FB310 ( Figure 3).
In the phylogenetic analysis of the recA and atpD genes, we also included strains closely related to those from this study, when the sequences of these two genes are available at GenBank (Figure 3). Several of these strains isolated from nodules of Pisum, Vicia, and Lens, including those strains carrying an IVS in their rrs gene, have been named Rhizobium leguminosarum, but the analysis of their recA and atpD genes placed them within the species R. laguerreae ( Figure 3). This analysis confirmed the phylogenetic complexity of this species, which was already shown at the time of its description, since it encompassed strains phylogenetically divergent [22] and led Flores-Félix et al. [21] to point out the existence of several genospecies within R. laguerreae after the analysis of several strains isolated from nodules of different legumes [21]. Nevertheless, in the present work, where we analyzed more strains from the phylogenetic group of R. laguerreae, we observed a high complexity within this species, indicating that it is difficult to establish reasonable limits for differentiating genospecies (or subspecies) and that it will be necessary to do so based on the analysis of whole genomes.
The strains nodulating P. sativum analyzed in this study fall within four clusters obtained after the recA and atpD gene analysis that also contain other strains isolated in different countries from nodules of Lens, Vicia, and Pisum ( Figure 3). Strain AMPS17 and strains AMPS04 and AMPS23 belong to two clusters, together with closely related strains, most of which are already named R. laguerreae and have been isolated in different countries of Asia, Africa, Europe, and America from nodules of L. culinaris (Turkey, Algeria, and Morocco), V. faba (Tunisia, Algeria, and Peru), and P. sativum (Algeria and India) [9,15,17,22,24,25]. Some of these strains were named R. leguminosarum at the time of their publication because the species R. laguerreae was described later [9,25]. Strains AMPS05 and AMPS34 belong to a cluster containing strains isolated from Vicia sativa in Asia and Europe [21,22,26]. Finally, the strain AMPS22 was closely related to a strain isolated by Rashid et al. [25] from nodules of L. culinaris in Germany ( Figure 3). It is remarkable that, from the strains isolated from P. sativum nodules whose rrs sequences were almost identical to that of the strain AMPS22, only the strain BIHB1132 belongs to R. laguerreae. The strains A1 and JKLM13E, were related to the type strains of R. ruizarguesonis and R. indicum, respectively, and the strain BIHB1160 formed and independent lineage ( Figure 3). In the case of the strains carrying the IVS and whose rrs gene sequences were identical, we also found important divergences for recA and atpD genes, particularly in the case of the strain JHI2449, which showed 96.3% and 98.2% similarity, respectively, with respect to the type strain of R. laguerreae, these values being within the limit for species differentiation. Nevertheless, we also found the opposite case, since strains with some differences in the rrs gene, such as the strain AMPS17, were closely related to the type strain of R. laguerreae in the recA and atpD gene analysis. These results confirmed again that the rrs gene is not a suitable marker for differentiation of closely related species, other core protein-coding genes being more adequate for species delineation.  [27]). ¥ : halo size (mm) around colony. ƒ : ratio between the diameter of the halo around the colony with respect to colony size [28].

Analysis of the nodC Gene
The nodulation genes are responsible for the legume nodule formation, and among them, the nodC gene is the most common phylogenetic marker to define rhizobial symbiovars [29,30]. The symbiovar concept (formerly biovar) is related to both the rhizobial host range and the legume promiscuity degree and hence to the old cross-inoculation group concept [30]. In this way, a promiscuous legume can be nodulated by various symbiovars, and a restrictive legume can only be nodulated by one symbiovar [30]. The legumes from the Vicia cross-inoculation group, which includes Vicia, Pisum, Lens, and Lathyrus, are considered restrictive hosts for nodulation [31], nodulated by strains of the symbiovar viciae [32]. It is remarkable that a single species can contain several symbiovars; the first reported case was the species R. leguminosarum, which contains three symbiovars, two of which nodulate restrictive hosts such as Vicia, Pisum, Lens, Lathyrus (viciae), and Trifolium (trifolii) and one of which nodulates the very promiscuous host Phaseolus (phaseoli) [32]. These three symbiovars are clearly separated through nodC gene analysis [33,34], and for this reason it has been analyzed in many strains isolated from nodules of legumes from the Vicia cross-inoculation group in different geographical locations (Figure 4).
The analysis of the nodC gene showed that, as expected, the closest related strains to those from this study are those of the symbiovar viciae, with similarity values higher than 92%. Nevertheless, the phylogenetic analysis of the nodC gene showed that these strains belong to two different clusters ( Figure 4). Strains AMPS04, AMPS34, AMPS05, and AMPS22 belong to a cluster that encompasses strains isolated from nodules of Pisum, Vicia, and Lens in several countries in Europe and Africa, and strains AMPS17 and AMPS23 belong to a wide cluster encompassing strains nodulating the same legumes in different countries in Europe, America, Asia, and Africa. These results confirmed the existence of an important genetic variability within the symbiovar viciae worldwide, with similarity values lower than 95% in the nodC gene. This value is within the range of internal similarity of most defined symbiovars, but it is lower than the values reported for the strains of the novel symbiovar aegeanense recently described within the genus Ensifer, which showed more than 95% similarity with respect to the symbiovar fredii [35].

Biofilm Formation and Quorum Sensing
Cellulases and cellulose produced by Rhizobium strains are involved in plant colonization and biofilm formation in legume roots [36], and they are produced by several Rhizobium species that are able to form biofilms [36], including some strains from R. laguerreae isolated from P. vulgaris nodules [37]. In agreement with these results, all strains analyzed in this study produced cellulases and cellulose (Table 1, Figure 5A). They were able to form biofilms on abiotic surfaces, and an increase in the biofilm formation over time was found in all selected strains, with significant differences between the biofilm formation after 24 and 48 h, except in the strain AMPS23. Significant differences were also found after 72 h in strains AMPS04, AMPS05, AMPS23, and AMPS34 ( Figure 5(Ba)). The strain AMPS05 exhibits the highest biofilm formation at 72 h and the strain AMPS22 at 24 and 48 h ( Figure 5(Bb)). These results are in agreement with those found for Rhizobium strains nodulating clover [36] or common bean [37].
Rhizobial strains from populations formed on the legume roots communicate among them through signaling systems based on the quorum sensing compounds of the N-acyl-homoserine lactone groups (AHLs), which stimulate the genetic exchange among bacteria in the legume rhizosphere and play a direct role in the legume nodulation process [38,39]. Symbiotic rhizobia have different AHL-based quorum sensing systems, and within the genus Rhizobium, they have been studied in R. etli and R. leguminosarum [39]. In this study, we detected quorum sensing signals in culture supernatants of all strains, although some differences in active molecule production were found (Table 1, Figure 5C). Strains AMPS17 and AMPS22 were able to produce the highest estimated concentration of AHL, followed by strains AMPS04 and AMPS05, strains AMPS23 and AMPS34 being those producing the lowest AHL concentrations ( Figure 5C). Differences in the production of AHL have been found in different strains and symbiovars of R. leguminosarum [40], but this is the first work in which the AHL Plants 2020, 9, 1755 8 of 16 production is compared in different Rhizobium strains isolated from legume nodules using the micro method described in the work of Trovato et al. [27].

In Vitro Plant Growth Promotion Mechanisms
The ability to fix nitrogen in symbiosis with legumes was the first plant growth promotion mechanism studied in members of the genus Rhizobium [41], but other mechanisms have been more recently reported, such as the ability to solubilize phosphate, to synthesize indole acetic acid, or to produce siderophores [42]. In this work, we analyzed these mechanisms in selected strains from each RAPD group showing that all of them produced IAA and siderophores, but only two of them were able to solubilize phosphate (Table 1). Strains AMPS04, AMPS05, and AMPS34 produced higher IAA amounts, which were similar or higher than those reported for other Rhizobium strains [37,43,44]. Concerning siderophore production, the strain AMPS34 produced the halo with the largest radius in the CAS medium, but in general the results were similar to those previously found for other Rhizobium strains [37,43,44]. The ability to solubilize phosphate was found in only two strains, AMPS23 and AMPS34 (Table 1), which agrees with previous works where this mechanism was less frequently found in Rhizobium strains [43]. All these results are in agreement with those obtained for other Rhizobium strains [43,44], including two R. laguerreae strains isolated from common bean and clover nodules [37,45].

Effectiveness on Pea Plants
The effectiveness of rhizobia-legume symbiosis is directly related to the symbiotic nitrogen fixation process carried out by the rhizobia inside the nodules [46]. Therefore, the nodulation ability is also directly related to the effectiveness of rhizobial strains, since nodule formation and growth are regulated in response to the N present in the legume phloem and/or in the soil [47,48]. Moreover, as previously indicated, symbiotic nitrogen fixation was the first studied mechanism of plant growth promotion; therefore, classically, the parameters analyzed to determine the effectiveness of rhizobial strains have been the number of nodules, the N content, and the fresh and dry matter of the legume shoots. The obtained results for these parameters in P. sativum plants inoculated with selected strains from this study are displayed in Table 2. They show that, as expected, all strains were able to nodulate P. sativum, and significant differences in the number of nodules per plant were found between the strain AMPS04 and the type strain of R. pisi. All inoculation treatments yielded P. sativum plants with significantly higher fresh and dry weights than those from the uninoculated control without nitrogen addition. Both parameters were significantly lower in all treatments, including the positive control plants inoculated with the type strain of R. pisi, with respect to the positive noninoculated control plants watered with a nutrient solution containing 0.5g/L of ammonium nitrate as N source. Among the inoculation treatments, those performed with the type strain of R. pisi and with the strain AMPS34 yielded higher values for fresh and dry weights. Both values were not significantly different between these treatments and with respect to those corresponding to the inoculation with the strains AMPS05 and AMPS17. Concerning the percentage of fixed N, all treatments presented significant differences with respect to the uninoculated control without added N, and no significant differences were found among the remaining treatments. The total N content per plant showed the same behavior as that of the shoot dry weight (Table 2). These results imply that, although some strains were more effective than others, in general the strains of R. laguerreae sv. viciae showed high symbiotic efficiency in pea plants, with no significant differences between plants inoculated with the type strain of R. pisi and those inoculated with the strain AMPS34. In this last case, the percentage of fixed N per plant was only 2.5% lower than in the plants watered with a nutrient solution containing N sources. These results imply that R. laguerreae strains established highly effective symbiosis with P. sativum and can be used as inoculants for this legume. Values followed by the same letter in each treatment are not significantly different from each other at p = 0.05 according to Fisher's protected LSD test (least significant differences). S.E. = standard error. * Plants were watered with Rigaud and Pupo solution [49] containing 0.5 g/L of ammonium nitrate.

Diversity Analysis of Strains by RAPD Fingerprinting
The strains were isolated in a previous work from effective nodules (pink colour) of P. sativum plants at the flowering stage in Aldea de San Miguel (41 • 27 37" N 4 • 37 01" W), located in Valladolid province (Spain), where this legume is commonly cultivated [12].
The genetic diversity of isolated strains was assessed by RAPD fingerprinting performed as previously described [50] with the primer M13 (5 -GAGGGTGGCGGTTCT-3 ) and the Dream-Taq™ DNA Green PCR Master Mix (Fisher Scientific, Waltham, MA, USA). PCR conditions were the following: preheating at 95 • C for 9 min, 35 cycles of denaturing at 95 • C for 1 min, annealing at 45 • C for 1 min, and extension at 75 • C for 2 min, with a final extension at 72 • C for 7 min. Aliquots of 17 µL of each PCR product were electrophoresed on 1.5% (w/v) agarose gel in TBE buffer (100 mM Tris, 83 mM boric acid, 1 mM EDTA, pH 8.5) at 6 V/cm. The gels were stained in a solution containing 0.5 mg/L ethidium bromide, and photographed under UV light. Standard VI (Roche, Basilea, Switzerland) was used as a size marker. A dendrogram was constructed based on the matrix generated using the UPGMA method and Pearson's coefficient with Bionumerics version 4.0 (Applied Maths, Austin, TX, USA).

Gene Sequencing and Analysis
The amplification and sequencing of rrs, recA, atpD, and nodC genes were carried out as indicated by Carro et al. [51], Gaunt et al. [23], and Laguerre et al. [33], respectively. These genes were sequenced at the Sequencing DNA Service (NUCLEUS) from Salamanca University (Salamanca, Spain). The obtained sequences were compared with those from GenBank using the BLASTN program [52]. Sequences were aligned using the ClustalX software [53]. The distances were calculated according to Kimura's two-parameter model [54]. Phylogenetic trees were inferred using the neighbor-joining (NJ) analysis [55]. MEGA7 software [56] was used for all these analyses.

Quorum Sensing Signals and Biofilm Production
For the evaluation of the production of quorum sensing signals, strains were grown in 20 mL of YMB medium [57] for 48 h. Then 2 mL were transferred to 18 mL of YMB medium and were incubated for 24 h at 28 • C. The cultures were centrifuged 4 min at 4500× g, and the supernatants were filtered with a 0.22 µm pore filter. The strain Agrobacterium tumefaciens NTL4 (pZRL4) was employed as a reporter of AHL (acylated homoserine lactone) quorum sensing molecules. This strain was inoculated in 20 mL of liquid AB medium [58] supplemented with 30 µg/mL gentamycin and was incubated for 24 h at 28 • C. Afterwards, 10 mL of this culture were mixed with AB medium with agarose (1% w/v) at 43 • C and 150 µL of X-gal (5-bromo-4-chloro-3-indolyl-beta-D-galactopyranoside) stock solution for a final concentration of 60 µg/mL. Aliquots of 200 µL of this suspension were dispensed in each well of a 96-well microtiter plate. Upon medium solidification, 10 µL of a solution containing 10 ng/µL of OHL (N-Octanoyl-DL-homoserine lactone, Sigma Aldrich Co., St. Luis, MO, USA), filtered supernatants, and their serial dilutions were added. OHL was chosen as the AHL positive control. After 24 h of incubation, the digital image was acquired by an HP Scanjet 8200 flatbed scanner.
The ability of strains to form biofilms on abiotic surfaces was measured using the microtiter plate assay crystal violet post-staining as described by Robledo et al. [36]. The strains were grown in the minimal medium until reaching a value of 2.0 at 600 nm OD (ca. 1 × 10 7 cells/mL). A decimal dilution was used for subsequent biofilm formation measurements [36]. Biofilm data were treated with one-way ANOVA and the Fisher's post hoc test at a p value ≤ 0.05, using StatView 5.0 (Abacus Corporation, Berkeley, CA, USA). Data were analyzed by one-way ANOVA, and mean values were compared with Fisher's protected LSD test (least significant differences) (p ≤ 0.05).
Cellulase production was tested on CMC double-layer plates as was previously described [59] and after 7 days the presence/absence of hydrolysis halos was recorded. Cellulose detection assay was carried out as described by Robledo et al. [36].

In Vitro Plant Growth Promoting Mechanisms
The solubilization of insoluble phosphate was analyzed on YED-P plates containing 2% CaHPO 4 or Ca 3 (HPO 4 ) 2 , which were incubated for 15 days at 28 • C [60]. Phosphate-solubilization effectiveness was calculated as the ratio between the halos around the colony with respect to colony size [28]. Siderophore production was evaluated in M9-CAS-AGAR [61] modified with the addition of a cationic solvent, HDTMA, which stabilizes the Fe-CAS complex and gives the medium its characteristic color [62]. Indole acetic acid production was evaluated in JMM medium [63] supplemented with 107 mg/L of tryptophan. After 7 days of incubation, the supernatants were recovered by centrifugation at 5000× g and filtered using 0.22 µm Millipore filters (Millipore Co., Burlington, MA, USA). Afterwards, 1 mL of Salkowsky reagent was added to 2 mL of supernatant, and the red color formed was measured by spectrophotometry at 550 nm using an ATI Unicam 8625 Spectrometer (Mattson, Madison, WI, USA) [64].

Effectiveness Assays in Pisum sativum
For these assays, seeds of Pisum sativum var. "Rondo" were surface-sterilized with 5% sodium hypochlorite for 25 min. The sterilized seeds were germinated in sterile vermiculite and then inoculated with 1 mL per plant of a suspension containing approximately 1 × 10 8 CFU/mL of each strain. Cells were obtained from cultures on YMA plates incubated for 5 days at 28 • C and were suspended in sterile water. The inoculated plants were placed in a plant growth chamber with mixed incandescent and fluorescent lighting (400 microeinsteins m −2 s −1 ; 400-700 nm), programmed for a 16 h photoperiod, day-night cycle, with a constant temperature varying from 25 to 27 • C, and 50-60% relative humidity. Plants were watered with nitrogen-free Rigaud and Puppo [49] nutrient solution. A positive control was watered weekly with the same solution, with 0.5 g/L ammonium nitrate. Plants inoculated with the type strain of Rhizobium pisi DSM 30132 T were included as a reference. Four weeks after inoculation, plants were harvested, the number of nodules was counted, the fresh and dry weight, and the nitrogen content of the aerial parts of plants were measured. This last analysis was performed in the Ionomic Service from CEBAS-CSIC (Murcia, Spain). Data were analyzed by one-way ANOVA, and mean values were compared by Fisher's protected LSD test (least significant differences) (p ≤ 0.05).

Conclusions
Collectively, the results of this study showed the following: (i) although several type strains of different species from the group of R. leguminosarum have identical rrs gene sequences, some differences have been found in this gene with respect to other strains of these species, including the presence of an insert (IVS) of 75 bp at the beginning of the rrs gene. (ii) Strains carrying identical IVS in their rrs genes were isolated from nodules of Pisum, Lens, and Vicia in different countries from Europe and Asia. (iii) All these strains belong to the species R. laguerreae according to the results of the recA and atpD gene analysis, which suggests that several of these strains are currently misnamed in GenBank. (iv) All strains analyzed in this study have been identified as R. laguerreae after recA and atpD gene analysis, and they belong to different clusters within this species, whose complexity was confirmed in this work. (v) Further studies based on whole genome analysis are necessary to establish cut-off values for genospecies or subspecies differentiation as well as to propose the correction of the misnamed strains. (vi) R. laguerreae is widely distributed in nodules of Pisum, Vicia, and Lens, three grain legumes cultivated for human and animal feeding worldwide, but some strains assigned to this species presented similarity values in their recA and atpD genes lower than those currently accepted for species differentiation. (vii) The analysis of the nodC gene showed that all strains analyzed in this study belong to the symbiovar viciae, which is distributed worldwide in the nodules of legumes from Pisum, Lens, and Vicia. (viii) This symbiovar is phylogenetically very complex and currently encompasses strains showing nodC gene similarity values lower than those presented by recently described symbiovars, evidencing that a consensus is necessary to establish cut-off values for symbiovar delineation. (ix) All strains from this study produced molecules involved in plant colonization and quorum sensing, such as cellulose, cellulases, and N-acyl-homoserine lactones. (x) These strains also produced IAA and siderophores, but the phosphate solubilization of tricalcium phosphate was only observed for two strains. (xi) All strains were effective on pea plants, with the strain AMPS34 as the most efficient, without significant differences with respect to the type strain of R. pisi used as a reference in this study. (xii) Taking into account the wide distribution of R. laguerreae sv. viciae strains in different countries and continents in P. sativum nodules and their high effectiveness in symbiosis with this legume, R. laguerreae sv. viciae is a promising candidate for the biofertilization of peas worldwide.