Synergy between Rhizobial Co-Microsymbionts Leads to an Increase in the Efficiency of Plant–Microbe Interactions

Combined inoculation of legumes with rhizobia and plant growth-promoting rhizobacteria or endophytes is a known technique for increasing the efficiency of nitrogen-fixing symbiosis and plant productivity. The aim of this work was to expand knowledge about the synergistic effects between commercial rhizobia of pasture legumes and root nodule bacteria of relict legume species. Pot experiments were performed on common vetch (Vicia sativa L.) and red clover (Trifolium pratense L.) co-inoculated with the participation of the corresponding commercial rhizobial strains (R. leguminosarum bv. viciae RCAM0626 and R. leguminosarum bv. trifolii RCAM1365) and seven strains isolated from nodules of relict legumes inhabiting the Baikal Lake region and the Altai Republic: Oxytropis popoviana, Astragalus chorinensis, O. tragacanthoides and Vicia costata. The inoculation of plants with combinations of strains (commercial strain plus the isolate from relict legume) had a different effect on symbiosis depending on the plant species: the increase in the number of nodules was mainly observed on vetch, whereas increased acetylene reduction activity was evident on clover. It was shown that the relict isolates differ significantly in the set of genes related to different genetic systems that affect plant–microbe interactions. At the same time, they had additional genes that are involved in the formation of symbiosis and determine its effectiveness, but are absent in the used commercial strains: symbiotic genes fix, nif, nod, noe and nol, as well as genes associated with the hormonal status of the plant and the processes of symbiogenesis (acdRS, genes for gibberellins and auxins biosynthesis, genes of T3SS, T4SS and T6SS secretion systems). It can be expected that the accumulation of knowledge about microbial synergy on the example of the joint use of commercial and relict rhizobia will allow in the future the development of methods for the targeted selection of co-microsymbionts to increase the efficiency of agricultural legume–rhizobia systems.


Introduction
The term rhizobia can be considered as a historically valid general term for a large group of predominantly soil-and plant-associative α-proteobacteria that form the order Rhizobiales Kuykendall 2006 (synonym Hyphomicrobiales Douglas 1957 emend. Hördt et al. 2020) [1]. Representatives of the first described rhizobia species were isolated from root nodules of leguminous plants (family Fabaceae) and belonged to the genera Allorhizobium, Azorhizobium, Bradyrhizobium, Sinorhizobium, Rhizobium and Mesorhizobium [2]. Subsequently, numerous other genera and species within this order were isolated from nodules, although many of the described taxons did not nodulate the host plant upon inoculation

Bacterial Material
Commercial strains of rhizobia: Rhizobium leguminosarum bv. viciae RCAM0626 nodulating common vetch (V. sativa L.) and R. leguminosarum bv. trifolii RCAM1365 nodulating red clover (T. pratense L.) were used. The rhizobial strains Mesorhizobium japonicum Opo-235, M. japonicum Opo-242, Bradyrhizobium sp. Opo-243 and M. kowhaii Ach-343 were isolated previously from nodules of the relict legumes originating from Buryatia (Baikal Lake region, Russia): Oxytropis popoviana Peschkova (N51 • Table 1. All isolates belong to the order Rhizobiales (Hyphomicrobiales). Strains of nodule bacteria were isolated from the obtained nodules using the standard method described earlier [22] and were cultivated using modified yeast extract mannitol agar (YMA) [23] supplemented with 0.5% succinate (YMSA) [24]. All strains were deposited in the Russian Collection of Agricultural Microorganisms (RCAM) and stored at −80 • C in an automated Tube Store (Liconic Instruments, Lichtenstein) [25]. Table 1. Commercial strains and rhizobial isolates from the relict legumes of the Baikal Lake region and the Altai Republic used in the pot experiment.

Pot Experiments
Seeds of common vetch (V. sativa) and red clover (T. pratense) plants were surface sterilized with H 2 SO 4 for 10 min, washed with sterile tap water and germinated on filter paper in Petri dishes at 25 • C in the dark for 4 days. Germinated seedlings were transferred to polypropylene pots containing 0.5 kg of sterile sand (2 seeds per pot, 7 pots per treatment). Each pot was supplemented with 100 mL of the nutrient solution (g/L): K 2 HPO 4  The nitrogen fixation of nodules was measured using the acetylene-reduction method [26]. Briefly, the roots of each pot (two plants) were collected, washed in tap water, weighted and placed in 50 mL flasks. The flasks were hermetically sealed, supplemented with 5 mL of acetylene and incubated for 1 h in the dark at 20 • C. The amount of ethylene formed in the flasks was determined using a gas chromatograph GC-2014 (Shimadzu, Kyoto, Japan). Roots of control plants without nodules were used as a negative control. Then, nodule number in each flask was calculated. Strains were re-isolated from the obtained nodules and identified using 16S rDNA sequencing as described earlier [27]. The data were processed via the standard method of variance analysis using the software STATISTICA version 10 (StatSoft Inc., Tulsa, OK, USA). Fisher's LSD test was used to evaluate differences between means.
Search for homologs of the symbiotic fix, nif, nod, noe and nol genes as well as genes that promote plant growth (acdRS, gibberellin-and auxin-biosynthesis related) and genes of the T3SS, T4SS and T6SS secretion systems in annotated genomes was performed using CLC Genomics Workbench 7.5.1 (CLC bio, Aarhus, Denmark) software using local BLASTn and tBLASTx (e-value threshold of 1e-30) (https://blast.ncbi.nlm.nih.gov accessed on 22 March 2023).

Pot Experiments
The pot experiments were performed on common vetch and red clover plants. Commercial strains of rhizobia (R. leguminosarum bv. viciae RCAM0626 and R. leguminosarum bv. trifolii RCAM1365), as well as seven strains isolated from the relict legumes of the Baikal Lake region and the Altai Republic (Oxytropis popoviana, Astragalus chorinensis, O. tragacanthoides and Vicia costata), were used for mono-and co-inoculation (Table 1). It should be noted that the climate in the places of growth of relict legumes is sharply continental, with cold winters and hot summers. Soils are poor and low in carbon, nitrogen and phosphorus.
The choice of isolates for pot experiments with different plant species was justified using preliminary data on the positive effect of joint inoculations in plant nodulation assays with Baikal isolates [19,21] and the newly isolated strains of Altai origin. The results are shown in Tables 2-5.
In variants of co-inoculation with the commercial strain R. leguminosarum bv. viciae RCAM0626, three isolates (except the Altai isolate Devosia sp. A8/3-2) resulted in a significant increase in the number of nodules ( Table 2). By themselves, all isolates did not form nodules or tumors on the vetch roots. An enhanced total nitrogen-fixing activity per plant was observed only when the strain R. leguminosarum bv. viciae RCAM0626 was combined with isolates M. kowhaii Ach-343 and Devosia sp. A8/3-2 (by 150 and 157%, respectively, Table 2). Co-inoculation with the last isolate also increased the level of activity calculated per one nodule compared with the commercial strain (by 143%). After co-inoculation with isolates M. japonicum Opo-235 and Devosia sp. A8/3-2, there was a significant increase in the biomass of vetch roots (by 122 and 130%, respectively, Table 3), while the isolate M. kowhaii Ach-343 made a smaller contribution to this parameter (116%). An enhanced total plant biomass (roots and shoots) was also observed in these three co-inoculations, but the change was not statistically different (112-114%). In general, an increase in the number of nodules did not affect the V. sativa biomass. The most effective was the combination of the strain R. leguminosarum bv. viciae RCAM0626 with the Altai isolate Devosia sp. A8/3-2, which, with almost the same number of nodules, led to an increase in nitrogen-fixing activity and plant weight. Table 2. Number of nodules and acetylene reduction activity after inoculation of V. sativa plants with the commercial strain R. leguminosarum bv. viciae RCAM0626 and the isolates M. japonicum Opo-235, Bradyrhizobium sp. Opo-243, M. kowhaii Ach-343 and Devosia sp. A8/3-2 in the pot experiment. Data are means ± standard errors of one representative experiment (n = 7). Asterisks show significant difference against the treatment with the strain RCAM0626 (Fisher's LSD test, * p < 0.05, ** p < 0.01, *** p < 0.001). AV-absolute value, RV-relative value to the strain RCAM0626.

Treatment
Number of Nodules (pot −1 ) Acetylene Reduction Activity Opo-235 The data of the pot experiment on mono-and co-inoculation of red clover with the commercial strain R. leguminosarum bv. trifolii RCAM1365; the isolates are shown in Table 4. Unlike common vetch plants, the isolate M. japonicum Opo-235 itself formed inactive nodules on clover roots. However, none of the isolates contributed to the formation of more nodules when co-inoculated with the commercial strain. In the experiment on T. pratense, the increase in the level of nitrogen-fixing activity per one nodule by 130-218% was detected when the commercial strain was combined with each of four isolates used (Table 4). Table 4. Number of nodules and acetylene reduction activity after inoculation of T. pratense plants with the commercial strain R. leguminosarum bv. trifolii RCAM1365 and the isolates M. japonicum Opo-235 and Opo-242, Phyllobacterium sp. A18/5-2 and P. zundukense A18/3m in the pot experiment. Data are means ± standard errors of one representative experiment (n = 7). Asterisks show significant difference against the treatment with the strain RCAM1365 (Fisher's LSD test, * p < 0.05, *** p < 0.001). AV-absolute value, RV-relative value to the strain RCAM1365.

Treatment
Number of Nodules (pot −1 ) Acetylene Reduction Activity  This parameter was maximal in the variant R. leguminosarum bv. trifolii RCAM1365 + P. zundukense A18/3m. In some cases, higher activity was accompanied by an increase in the plant biomass (Table 5). Root and total plant biomass enhanced by 161 and 135%, respectively, when the strain R. leguminosarum bv. trifolii RCAM1365 was used together with the isolate M. japonicum Opo-242, while the shoot biomass increased by 128% when co-inoculated with the isolate P. zundukense A18/3m. Table 5. Biomass of T. pratense plants after inoculation with the commercial strain R. leguminosarum bv. trifolii RCAM1365 and the isolates M. japonicum Opo-235 and Opo-242, Phyllobacterium sp. A18/5-2 and P. zundukense A18/3m in the pot experiment. Data are means ± standard errors of one representative experiment (n = 14). Asterisks show significant difference against the treatment with the strain RCAM1365 (Fisher's LSD test, ** p < 0.01, *** p < 0.001). Plant, root and shoot biomass in mg fresh weight plant −1 , AV-absolute value, RV-relative value to the strain RCAM1365. Thus, the effect of joint inoculations on symbiosis with two plant species was different. On vetch, an increase in the number of nodules was mainly observed, on clover, an increase in their activity was mainly observed. This was also true for co-inoculations with the isolate M. japonicum Opo-235, which was used on both plants. Higher nitrogenfixing activity did not always have a positive effect on plant biomass, which is probably due to the insufficiently long vegetation period. Nodules obtained in pot experiments were typical elongated in shape. All isolates used for joint inoculation of clover were re-isolated from nodules together with the strain R. leguminosarum bv. trifolii RCAM1365 and identified. Nodules obtained on vetch roots in the variant R. leguminosarum bv. viciae RCAM0626 + Devosia sp. A8/3-2 also contained both strains, while only the commercial strain was found inside the nodules in the remaining co-inoculations of V. sativa.

Whole Genome Sequences of the Studied Strains
The symbiotic and plant growth-promoting genes, as well as the genes of protein secretion systems (T3SS, T4SS and T6SS) involved in the symbiosis formation, were the target genes in all the strains used in pot experiments with common vetch and red clover. The results of the search for the symbiotic fix, nif, nod, noe and nol genes showed that the commercial strains and Baikal isolates M. japonicum Opo-235, M. japonicum Opo-242 and M. kowhaii Ach-343 [21] have common nodABC genes, which are usually necessary for the formation of nodules in legumes [31]. The nif genes (nifHDK and nifENB) as well as fixABCX genes, encoding components of the nitrogenase and playing a central role in electron transfer to nitrogenase [32,33], were also found in these strains. Previously, we showed that Bradyrhizobium sp. Opo-243 had only one cluster of fixKJLNOQPGHIS genes [21]. In contrast, new Altai isolates Devosia sp. A8/3-2, Phyllobacterium sp. A18/5-2 and P. zundukense A18/3m did not have any symbiotic clusters, but only individual genes ( Table 6). Each strain had a unique set of symbiotic genes and possessed some additional nod, nif, fix, nol and noe genes as compared with the commercial strains. Table 6. Presence of the symbiotic genes fix, nif, nod, noe and nol in the commercial rhizobial strains R. leguminosarum bv. viciae RCAM0626 and R. leguminosarum bv. trifolii RCAM1365 used in the pot experiments, as well as three strains isolated from the Altai relict legumes O. tragacanthoides (Devosia sp. A8/3-2) and V. costata (Phyllobacterium sp. A18/5-2 and P. zundukense A18/3m).

Genes
Isolates from the Altai Relict Legumes Commercial Strains Presence of the acdSR genes, as well as genes associated with the biosynthesis of gibberellins and auxins involved in plant growth stimulation, is shown in Table 7. The commercial strains R. leguminosarum RCAM0626 and RCAM1365 contained no acdS gene encoding 1-aminocyclopropane-1-carboxylate (ACC) deaminase, which promotes nodule formation due to decreasing biosynthesis of phytohormone ethylene [34,35]. At the same time, the studied isolates had the acdS genes and/or acdR genes involved in the regulation of ACC deaminase. It can be stated that the commercial strains had significantly fewer genes associated with the biosynthesis of gibberellins and auxins compared with the isolates (Table 7). In general, Altai isolates Devosia sp. A8/3-2, Phyllobacterium sp. A18/5-2 and P. zundukense A18/3m were found to have the largest number of genes involved in auxin biosynthesis. Table 7. Presence of the ACC deaminase genes, as well as genes associated with the biosynthesis of auxins and gibberellins in the commercial rhizobial strains R. leguminosarum bv. viciae RCAM0626, R. leguminosarum bv. trifolii RCAM1365 and seven strains isolated from the relict legumes. Genes only present in isolates are marked in pink.

Features/Genes
Isolates from the Relict Legumes Strain RCAM
Genes of the T6SS secretion system were found in only three Baikal isolates (M. japonicum Opo-235, M. japonicum Opo-242 and M. kowhaii Ach-343), while the last isolate had the most of them ( Table 8).
The secretion of effector proteins, such as cytotoxins, lysozymes, lipoproteins, adherence factors and other compounds, due to the function of T3SS, T4SS and T6SS secretion systems, plays an important role in processes involved in bacterial virulence [38][39][40][41]. It is known that these secretion systems of rhizobia modulate nodule initiation and formation as well as the host specificity of microsymbionts [42][43][44][45].

Comparison of the Genomic Data and Results of Pot Experiments
To explain the role of relict isolates in increasing the efficiency of symbiosis with V. sativa and T. pratense plants, a comparative genomic analysis of the genes of strains used in pot experiments and their effects on plants was carried out. As a summarized result, the genes of rhizobial isolates related to symbiosis along with the effects of the combined inoculations with isolates and commercial strains on common vetch and red clover are presented in Table 9. We propose that at least some of these genes presented in the rhizobial isolates could be expressed and involved in the effects on symbiosis. Further discussion of the results obtained is based on this assumption. However, to confirm the proposed hypotheses, it is necessary to conduct further study of the expression of individual genes during the interaction of bacteria with plants.
Among accessory genes of the isolates that were absent in the commercial strain R. leguminosarum bv. viciae RCAM0626, the nodPZ and nolK genes participating in the modification of Nod-factors (NFs) could affect the formation of nodules and their total number on the common vetch roots. It is known that the nodP gene is involved in the 6-O-sulfation of the reducing end of the NFs [31], and NodP functioned in conjunction with NodQ, which synthesized the donor of the sulfate group [46]. It should be noted that the commercial strain R. leguminosarum bv. viciae RCAM0626 has the nodQ gene, although the importance of NFs' sulfation for the common vetch has not been shown. The gene nodZ carries out fucosilation of the non-reducing end of the NFs, while the nolK gene is involved in the synthesis of the fucosyl group precursor from mannose and is associated with the expression of the nodZ gene [47][48][49]. The nodZ and nolK genes were found in various R. leguminosarum strains, and they are responsible for the modification of NFs, particularly the attachment of a fucosyl group that is of importance for the tribe Fabeae [50]. Isolate Devosia sp. A8/3-2 was the only one that increased the level of nitrogen-fixing activity per vetch nodule (i.e., not by increasing their number). It can be assumed that the symbiotic genes nifMV and fixK of the isolate Devosia sp. A8/3-2 could be considered as the main accessory genes contributing to the increased nitrogen fixation by the co-inoculated plants. It was previously described that the nifM and nifV genes are required, respectively, for the accumulation of active Fe protein and for the maturation of nitrogenase via homocitrate synthesis [51,52]. The coordinated functioning of fixKL and fixJT genes is necessary for optimal expression of genes regulating bacterial growth under microaerophilic conditions [53,54]. Table 9. Symbiotic and non-symbiotic genes present in the isolates but not in the commercial strains of common vetch and red clover and the effects of co-inoculations. SG stands for symbiotic genes, NSGnon-symbiotic genes. Effect: NN, PB, RB and SB-increased number of nodules, plant biomass, root biomass and shoot biomass per plant, respectively; ARAp and ARAn-increased acetylene reduction activity per plant and per nodule, respectively. Joint inoculations of red clover increased specific nitrogen-fixing activity per nodule, which was not observed in the experiment with common vetch. The additional nif and fix genes of isolates could contribute to this parameter in the presence of R. leguminosarum bv. frifolii RCAM1365 (Table 9). For example, genes nifZQXW are involved in the synthesis of the Fe-S-cofactor and fixation of molybdenum, protecting nitrogenase from the negative effects of oxygen [51,[55][56][57][58]. The fixHSQ genes play important role in the biosynthesis of the oxidase complex, which regulates the transport of oxygen through membranes, allowing normal cell breath at low oxygen concentrations [53,56]. The role in the symbiosis of the fixKL genes, as well as the gene nifV, is described above. In all variants of joint inoculation with an increased level of nitrogen-fixing activity, both strains were present in nodules. In our previous work, the isolates M. japonicum Opo-235 and Opo-242 also increased the nitrogen-fixing activity per one nodule when co-inoculated with the strain R. leguminosarum bv. trifolii RCAM1365 in the plant nodulation assay [21]. Another study showed an increase in the nodule activity on Glycyrrhiza uralensis plants co-inoculated with the M. japonicum Opo-235 and M. kowhaii Ach-343 strains carrying the complementary nifQV and fixJKL genes. The important observation was that these isolates, having genetic complementation to each other, were localized in the same plant cells [19].

Isolate
It should be noted that in addition to symbiotic genes, the genes related to plant growth promotion (acdSR, gibberellin-and auxin-biosynthesis related genes), as well as the genes of the T3SS, T4SS and T6SS secretion systems, can also positively influence the process of nodulation and nitrogen-fixing activity (Table 9). It was shown that these genes involved in the biosynthesis of phytohormones and key components of bacterial pathogenicity (e.g., penetrating effectors and cytotoxins) can participate in the integration between symbiotic partners [39][40][41][59][60][61]. Further study of the phenomenon of synergistic rhizobial interaction based on the complementarity of genes in taxonomically different comicrosymbionts will help reveal the mechanisms for the formation of effective integration of partners and the evolution of legume-rhizobium symbiosis.
Thus, this study and the results obtained earlier show that strains isolated from relict legumes have markedly more genes that affect plant-microbial interactions than commercial strains isolated from currently cultivated crops. At the same time, relict isolates can occupy nodules formed by commercial strains, even if they cannot form them on their own. Since the level of cooperation between different co-microsymbionts most likely depends on their spatial arrangement (rhizosphere, rhizoplane or nodule), the localization of rhizobia in the same nodule (or even in the same plant cell) should lead to an increase in the level of their integration. To implement such integration, a necessary condition is the expression of the corresponding genes in relict isolates and commercial strains during co-inoculation. A detailed study of this issue will make it possible to understand the mechanisms of the observed synergy. At this stage, we can only assume that the observed effects may be associated with the modification of nod-factors under the influence of relict isolates and horizontal gene transfer between microsymbionts. This cooperation between different rhizobia, including relict isolates and commercial strains, can be expressed with a positive change in symbiotic parameters (number of nodules, nitrogen-fixing activity and plant biomass) during joint inoculation.
It can be assumed that this study is not only of applied importance for agricultural biotechnology, but is also of great fundamental interest. In particular, based on a comparison of the whole genome analysis of strains and their effect on plants, preliminary conclusions can be drawn about the role of different genes in the formation of effective symbiosis.