Plant Growth Promotion Abilities of Phylogenetically Diverse Mesorhizobium Strains: Effect in the Root Colonization and Development of Tomato Seedlings

Mesorhizobium contains species widely known as nitrogen-fixing bacteria with legumes, but their ability to promote the growth of non-legumes has been poorly studied. Here, we analyzed the production of indole acetic acid (IAA), siderophores and the solubilization of phosphate and potassium in a collection of 24 strains belonging to different Mesorhizobium species. All these strains produce IAA, 46% solubilized potassium, 33% solubilize phosphate and 17% produce siderophores. The highest production of IAA was found in the strains Mesorhizobium ciceri CCANP14 and Mesorhizobium tamadayense CCANP122, which were also able to solubilize potassium. Moreover, the strain CCANP14 showed the maximum phosphate solubilization index, and the strain CCANP122 was able to produce siderophores. These two strains were able to produce cellulases and cellulose and to originate biofilms in abiotic surfaces and tomato root surface. Tomato seedlings responded positively to the inoculation with these two strains, showing significantly higher plant growth traits than uninoculated seedlings. This is the first report about the potential of different Mesorhizobium species to promote the growth of a vegetable. Considering their use as safe for humans, animals and plants, they are an environmentally friendly alternative to chemical fertilizers for non-legume crops in the framework of sustainable agriculture.


Introduction
The plant growth promoting rhizobacteria (PGPR), also named plant probiotic bacteria, are promising biofertilizers for sustainable and environmentally friendly agriculture since they allow the total or partial substitution of chemical fertilizers added to the crops alone or together with organic amendments [1][2][3]. These bacteria take part in the plant microbiome and can live in the rhizosphere about the effect of strains from the genus Mesorhizobium on tomato seedlings. Therefore, the final aim of this study was to analyze the effect of the inoculation of two selected Mesorhizobium strains showing in vitro PGP-activities on the growth of tomato seedlings. The results obtained showed for the first time that, similar to those of Rhizobium, Mesorhizobium strains are promising biostimulants for tomato plants.

Bacterial Strains
The rhizobia used in this study were isolated from effective root nodules of C. canariense from the Canary Islands in a previous work [50].

Phylogenetic Analysis
In this work, we performed the phylogenetic analysis of the atpD gene, amplified and sequenced using the primers and conditions previously described [51]. These sequences and those of the type strains of Mesorhizobium species described to date were aligned by using ClustalW software [52]. Distances calculated according to Kimura's two-parameter model [53] were used to infer phylogenetic trees with the Neighbor-joining method [54] with MEGA7 software [55]. Confidence values for nodes in the trees were generated by bootstrap analysis using 1000 permutations of the data sets. The atpD sequences were deposited in the GenBank database under the accession numbers MN999428-MN999451.

Analysis of In Vitro Plant Growth Promoting (PGP) Mechanisms
For indole-acetic acid production, bacterial cultures were grown at 28 • C in YMB medium [56] supplemented with 2.5 mM L-tryptophan until reaching a stationary phase. Cells were eliminated by centrifugation and the IAA and IAA-like compounds were measured in the supernatants using a colorimetric method [57]. The phosphate-solubilizing ability was tested by growing the bacteria for 7-9 days in NBRIP medium [58] containing tricalcium phosphate as a source of insoluble phosphate and observing the formation of a transparent halo around the colony. Phosphate-solubilizing effectiveness was calculated as the ratio between the halos around the colony with respect to colony size [59]. Siderophore synthesis was evaluated by growing the bacteria for seven days in Chrome Azurol S medium (CAS)-agar medium [60], in which siderophore-producing strains had a yellowish halo around the colony [61]. The ability to solubilize potassium was tested using the Aleksandrov medium [62], which contains potassium aluminum silicate as K source. The presence/absence (+/−) of the halo was recorded at 14 days.

Tomato Root Colonization and Biofilm Production Assays
Strains CCANP14 and CCANP122 were labeled with the green fluorescent protein (GFP) following a previously described protocol [63]. Tomato seeds were surface disinfected with 70% ethanol for 30 s followed by 5 min in 50% diluted commercial bleach. After six washes with sterile-distilled water, seeds were spread on 0.75% agar plates. Two days after germination, seedlings were transferred to 1.5% agar 10 cm × 10 cm square plates and each seedling was inoculated with 250 µL of a bacterial suspension (0.5 of OD 600 ; 4 × 10 8 CFU mL −1 ) and incubated in a growth chamber (16 h-light/8 h-dark cycle). Mock-inoculated controls were also included. Seedlings were viewed under a confocal microscope (Leica TCS SPE) five days after inoculation. Propidium iodide (7.5 µM) was used to counterstain plant root cells. The ability of strains CCANP14 and CCANP122 to form biofilms in abiotic surfaces was measured using the method of microtiter plate assay with crystal violet post-staining, following the protocol described by Robledo et al. [32]. The strains were grown in the minimal medium [32] and measurements were taken at 24, 48 and 72 hours. Biofilm data were treated with one-way ANOVA and the Tukey's post hoc test at p ≤ 0.05, using RStudio version 1.1.463. Cellulase production was tested onto CMC double-layer plates as described previously [64] and the presence/absence of the halo was recorded at 7 days. Cellulose detection assays were performed as described by Robledo et al. [32].

Microcosm Plant Assays
For these assays, tomato (Solanum lycopersicum var. cherry) seeds were surface disinfected with 70% ethanol for 30 s followed by diluted commercial bleach (2.5% sodium hypochlorite) for 5 min. After six washes with distilled water, seeds were placed on 1% agar plates. Four days after germination, seedlings were planted in (12 cm × 12 cm) plastic pots containing 500 g of sterilized peat (Pro-line green peat, Compo, Münster, Germany). Sixteen plants per treatment were inoculated with the strains CCANP14 and CCANP122, independently, with 1 mL of bacterial cultures (five units in the McFarland standard, 1.5 × 10 9 CFU·mL −1 ). As a control, a set of uninoculated plants were grown in the same conditions. Plants were irrigated with water every two days once a week with a nutrient solution [65] supplemented with 0.4 g·L −1 KNO 3 . Tomato plants were grown in a plant growth chamber with an 8 h-light/16 h-dark cycle. Five weeks after inoculation, the plants were harvested, roots washed with distilled water and fresh and dry weight (70 • C in an oven until constant weight was reached) of tomato shoots and roots were measured. For dry-weight measurement, the samples were dried in an oven at 80 • C. The dry plants were used for ionomic analyses, which were performed by the Ionomic service at IPNA (CSIC) using an ICP-OES AVIO500, Perkin Elmer equipment. Prior to analysis, the obtained data were checked for normality (Shapiro-Wilk test) and for homogeneity of variances (Levene's test), and then, they were subjected to one-way ANOVA, using Fisher's test (p = 0.05) by SPSS (version 21.0) statistical software (IBM, Chicago, IL, USA).

Phylogenetic Analysis of the atpD Gene
In this work, we analyzed the atpD gene of 24 Mesorhizobium strains isolated in the Canary Islands from nodules of C. canariense, because this gene, which was not sequenced in our previous study [61], is commonly used for the differentiation of Mesorhizobium species. Furthermore, it is available for three recently described species, Mesorhizobium denitrificans, Mesorhizobium carbonis and Mesorhizobium zhangyense, for which the glnII gene, commonly included in the identification schemes of Mesorhizobium strains, is not available. The phylogenetic analysis of this gene showed that the analyzed strains belong to six clusters (I to VI) and four lineages (A to D) ( Figure 1). Some of them can be confirmed as belonging to already described species (as they showed around 99% similarity values), namely, CCANP14, CCANP48, CCANP79 and CCANP82 to M. ciceri, CCANP3, CCANP99 and CCANP113 to M. opportunistum, CCANP1 to M. australicum and CCANP122 to M. tamadayense ( Figure 1 and Table 1). The remaining strains were phylogenetically related to several Mesorhizobium species, but the similarity values were equal to or lower than 98% ( Table 1). The strains CCANP11, CCANP29, CCANP33, CCANP68, CCANP78 and CCANP96 were closely related to M. muleiense with similarity values equal to or lower than 97%. The strains CCANP84 and CCANP87 have M. septentrionale as the closest relatives with 97.7% similarity value. The strains CCANP34, CCANP35 and CCANP38 show similarity values near to 98% to M. caraganae. The strains CCANP55 and CCANP61 were closest related to M. jarvisii with 96.6% similarity. Finally, the strains CCANP63 and CCANP130 formed independent phylogenetic lineages having 95.4% and 96.4% similarity, respectively, with respect to their closest related species M. robiniae and M. shonense. Mesorhizobium shonense HAMBI 3295 T (JQ972731)

In Vitro PGP Mechanisms
We have analyzed four of the most commonly PGP properties found in rhizobia, namely, IAA and siderophore production and phosphate and potassium solubilization, in the 24 Mesorhizobium strains from this study. The obtained results showed that they had at least one of these mechanisms (Figure 2). The IAA production was the most widespread trait of all strains synthesized in this auxin, although they varied greatly in the produced amounts from 5 to 69 µg mL −1 ( Table 1). None of the strains possessed the four tested mechanisms; however, nine strains belonging to M. ciceri, M. tamayadense and M. australicum displayed three of them ( Figure 2 and Table 1). Solubilization of tricalcium phosphate was detected in the four strains of M. ciceri, one strain of M. opportunistum and three strains related to M. muleiense (33%). Solubilization of potassium was detected in the four strains of M. ciceri, two strains of M. opportunistum, one strain of M. australicum, the strain of M. tamadayense, one strain related with M. shonense and three strains related to M. muleiense (46%). Both phosphate and potassium solubilization showed the highest indexes for the strains of species M. ciceri ( Figure 2 and Table 1). The less frequent PGP trait in Mesorhizobium is the siderophore production, which was positive only in four strains (17%), two of them belonging to M. australicum and M. tamadayense and the other two strains related to M. jarvisii and M. shonense ( Figure 2 and Table 1).

In Vitro PGP Mechanisms
We have analyzed four of the most commonly PGP properties found in rhizobia, namely, IAA and siderophore production and phosphate and potassium solubilization, in the 24 Mesorhizobium strains from this study. The obtained results showed that they had at least one of these mechanisms (Figure 2). The IAA production was the most widespread trait of all strains synthesized in this auxin, although they varied greatly in the produced amounts from 5 to 69 μg mL −1 ( Table 1). None of the strains possessed the four tested mechanisms; however, nine strains belonging to M. ciceri, M. tamayadense and M. australicum displayed three of them ( Figure 2 and Table 1). Solubilization of tricalcium phosphate was detected in the four strains of M. ciceri, one strain of M. opportunistum and three strains related to M. muleiense (33%). Solubilization of potassium was detected in the four strains of M. ciceri, two strains of M. opportunistum, one strain of M. australicum, the strain of M. tamadayense, one strain related with M. shonense and three strains related to M. muleiense (46%). Both phosphate and potassium solubilization showed the highest indexes for the strains of species M. ciceri ( Figure 2 and Table 1). The less frequent PGP trait in Mesorhizobium is the siderophore production, which was positive only in four strains (17%), two of them belonging to M. australicum and M. tamadayense and the other two strains related to M. jarvisii and M. shonense ( Figure 2 and Table 1).

In vitro Biofilm Formation and Tomato Root Colonization
The biofilm formation, as we previously mentioned, is essential to colonize the plant roots, a process that also involves bacterial cellulose and cellulases. For the analysis of biofilm formation and cellulose and cellulase production, we selected two strains, M. ciceri CCANP14 and M. tianshanense CCANP122, because both strains produced the highest amounts of IAA; 68 μg mL −1 and 69 μg mL −1 , respectively. Furthermore, CCANP14 showed the highest phosphate solubilization index and a high potassium solubilization index; CCANP122 also solubilizes potassium and produces siderophores. The in vitro biofilm formation assay revealed that the two Mesorhizobium strains are able to form biofilms ( Figure 3A). An increase in the biofilm formation along the time was also found for both strains with significant differences only between the biofilm formation after 24 and 48 h for the strain CCANP14 (Figure 3Ab). Despite the lack of statistical significance, CCANP122 appeared to exhibit better biofilm formation than CCANP14, particularly at 24 h (Figure 3Aa). Both strains were able to

In vitro Biofilm Formation and Tomato Root Colonization
The biofilm formation, as we previously mentioned, is essential to colonize the plant roots, a process that also involves bacterial cellulose and cellulases. For the analysis of biofilm formation and cellulose and cellulase production, we selected two strains, M. ciceri CCANP14 and M. tianshanense CCANP122, because both strains produced the highest amounts of IAA; 68 µg mL −1 and 69 µg mL −1 , respectively. Furthermore, CCANP14 showed the highest phosphate solubilization index and a high potassium solubilization index; CCANP122 also solubilizes potassium and produces siderophores. The in vitro biofilm formation assay revealed that the two Mesorhizobium strains are able to form biofilms ( Figure 3A). An increase in the biofilm formation along the time was also found for both strains with significant differences only between the biofilm formation after 24 and 48 h for the strain CCANP14 (Figure 3Ab). Despite the lack of statistical significance, CCANP122 appeared to exhibit better biofilm formation than CCANP14, particularly at 24 h (Figure 3Aa). Both strains were able to produce cellulose in media containing Congo Red (Figure 3 Ba and Bb) and to produce cellulases, although the strain CCANP14 showed the highest activity ( Figure 3 Bc and Bd).
Confocal scanning laser microscopy (CSLC) of GFP-tagged mesorhizobial strains showed that the two strains can colonize cherry tomato roots to different degrees and formed the typical three-dimensional biofilm structure ( Figure 3C). Cherry tomato roots inoculated with strain CCANP14 displayed mature biofilms on the entire roots and root hairs (Figure 3Cb) five days post-inoculation, while CCANP122 (Figure 3Cc) colonized roots and root hairs in a more discrete manner.
produce cellulose in media containing Congo Red (Figure 3 Ba and Bb) and to produce cellulases, although the strain CCANP14 showed the highest activity ( Figure 3 Bc and Bd).
Confocal scanning laser microscopy (CSLC) of GFP-tagged mesorhizobial strains showed that the two strains can colonize cherry tomato roots to different degrees and formed the typical threedimensional biofilm structure ( Figure 3C). Cherry tomato roots inoculated with strain CCANP14 displayed mature biofilms on the entire roots and root hairs (Figure 3Cb) five days post-inoculation, while CCANP122 (Figure 3Cc) colonized roots and root hairs in a more discrete manner.

Microcosm Plant Assays
The obtained results of the tomato inoculation with the strains CCANP14 and CCANP122 are shown in Table 2. Fresh and dry shoot and root lengths and weights of the inoculated plants were significantly higher than those of the uninoculated ones. Additionally, significant differences were found between the two inoculation treatments, except in the shoot length. The inoculation with CCANP14 produced the highest values in the remaining parameters (Table 2). Therefore, the inoculation of tomato seedlings with either of these two bacteria had a positive effect on plant growth, although the strain CCANP14 yielded the best results.

Microcosm Plant Assays
The obtained results of the tomato inoculation with the strains CCANP14 and CCANP122 are shown in Table 2. Fresh and dry shoot and root lengths and weights of the inoculated plants were significantly higher than those of the uninoculated ones. Additionally, significant differences were found between the two inoculation treatments, except in the shoot length. The inoculation with CCANP14 produced the highest values in the remaining parameters (Table 2). Therefore, the inoculation of tomato seedlings with either of these two bacteria had a positive effect on plant growth, although the strain CCANP14 yielded the best results.  The inoculation has an insignificant effect on the N and P content of plants (Table 3). Nevertheless, a significantly higher content of Ca was found in the shoots of plants inoculated with CCANP14 and the content of K and Na were significantly higher in those inoculated with CCANP122. The Zn content was significantly lower in the inoculated plants with respect to the control plants and the Mn content was significantly lower in the plants inoculated with the strain CCANP122 with respect to the control plants. The Fe content was significantly different between the plants from the two inoculation treatments and lower than in the control plants (Table 3). Table 3. Effect of inoculation of Mesorhizobium ciceri strain CCANP14 and Mesorhizobium tamadayense strain CCANP122 on the mineral content of tomato shoots.

Discussion
The strains analyzed in this study were previously distributed into 12 phylogenetic groups after the analysis of the housekeeping genes dnaK, gyrB, truA, glnII, thrA, recA and rpoB. The strains from some of these groups were identified as M. australicum, M. ciceri, M. opportunistum and M. tamadayense, but most of them belonged to an undescribed species of genus Mesorhizobium [66]. From these seven housekeeping genes, glnII and recA have been traditionally used for Mesorhizobium species differentiation and the description of new species. Nevertheless, for some recently described species of this genus, despite their genomes having been sequenced, the glnII gene is not available, whereas the atpD gene sequences are available in GenBank.
Therefore, the atpD gene sequences allowed us to compare our mesorhizobial strains with all described species within the genus Mesorhizobium. The results obtained in the present work showed that the atpD gene phylogeny was overall congruent with that obtained after the analysis of other protein-coding genes [66]. Thus, the atpD gene phylogeny confirmed the previous identification of the strains belonging to M. australicum, M. ciceri, M. opportunistum and M. tamadayense (clusters IV and VI and lineages B and C) (Figure 1) [66] as well as the phylogenetic location of strains within the clusters of M. caraganae (cluster III) and M. septentrionale (cluster II) (Figure 1) [66]. However, the phylogenetic position of some strains has changed because more than 10 Mesorhizobium novel species have been described since the publication of our previous work in the year 2014 [66]. This occurred with the strains CCANP55 and CCANP61 (cluster V), which had M. huakuii as their closest relative [66], whereas we show that currently the closest species is M. jarvisii [34] and confirm that the strains within this cluster V belong to a new Mesorhizobium lineage ( Table 1). The strains CCANP11, CCANP29, CCANP33, CCANP68, CCANP78 and CCANP96 (cluster I) belong to a very divergent cluster that was originally defined as a M. tianshanense-like group [66]. Moreover, according to the atpD gene, their closest related species is M. muleiense (Table 1), which is included within a big clade of species with M. tianshanense (Figure 1). The strain CCANP63 (lineage A) formed an independent lineage related to M. caraganae [66] and has atpD gene sequence more similar to M. robiniae (Table 1), (Figure 1). Finally, the strain CCANP130 (lineage D) formed an independent lineage in our previous work [66] as well as in the atpD gene phylogeny, although the sequence of this gene was more similar to that of M. shonense (Table 1). Therefore, the results of the atpD gene analysis are congruent with those previously obtained after the analysis of other protein-coding genes [66].
Noteworthy is that the phylogenetic distances in the sequences of some housekeeping genes for several recently described Mesorhizobium species are lower than those found in older described species, as occurred for example for Mesorhizobium japonicum and M. jarvisii, whose classification into different species was mainly supported by the DNA-DNA relatedness values found between them [67]. Taking these two species as reference (Figure 1), the strains from clusters I, II, III and V and from lineages A and D belong to putative new species of genus Mesorhizobium, confirming the high taxonomic diversity of the Mesorhizobium strains occupying the Cicer canariense nodules ( Table 1).
The capacity of species from genus Mesorhizobium to fix atmospheric nitrogen in symbiosis with legumes is widely known [68]; nevertheless, the presence of other in vitro plant growth promoting mechanisms has been less studied [12,19]. Among these mechanisms, IAA production is one of the most widespread PGP traits among Mesorhizobium strains [19] and this finding has been confirmed in this study, where the levels of IAA produced by some strains, particularly M. ciceri CCANP14 and M. tamadayense CCANP122 are similar to those found in Rhizobium strains able to promote the tomato growth [14]. The phosphate solubilization of tricalcium phosphate was one of the first plant growth promotion mechanisms reported for genus Mesorhizobium, specifically for a strain nodulating Cicer arietinum [69]. Later, it was confirmed that Mesorhizobium strains nodulating this legume were effective phosphate solubilizers [19,21], which is in agreement with the results from this study, where we found that strains related to M. muleiense and those of M. ciceri, two species originally isolated from C. arietinum nodules, showed the highest phosphate solubilization indexes ( Table 1). The solubilization of potassium has only been reported to date for one strain of genus Mesorhizobium closely related to Mesorhizobium plurifarium isolated from rape rhizospheric soil [24]; now, this is the first report about the capacity to solubilize potassium of strains from different Mesorhizobium species isolated from legume nodules. The siderophore production was reported for some strains nodulating C. arietinum in the last years of the past decade [70] and later, Brígido et al. [19] reported this mechanism for several Portuguese strains nodulating this legume. In our work, siderophore production was detected in only four strains (Table 1). Therefore, it might be concluded that IAA production is a common PGP mechanism in Mesorhizobium strains, whereas phosphate solubilization and siderophore production are variable in agreement with the results of Brígido et al. [19]. These results are in agreement not only with those found for Mesorhizobium strains, but also for other rhizobia [14].
Some of these rhizobia can colonize the roots of non-legumes, as occurs with two Rhizobium strains that can colonize, amongst others, the roots of tomato [14]. In the case of genus Mesorhizobium, the root colonization of Arabidopsis thaliana, Daucus carota (carrots) and Lactuca sativa (lettuce) by two strains isolated from Lotus has been reported [27,71]; nevertheless, this is the first report about the ability of Mesorhizobium strains to colonize tomato roots. Root colonization is an essential step for plant growth promotion, and commonly, the better bacterial growth promoters also are good root colonizers, as occurs with several strains assayed on tomato plants, such as Paenibacillus and Bacillus [28], Rhizobium [14] and Burkholderia [29]. Some plant growth promoting strains are particularly effective in the promotion of tomato seedlings, whose production is carried out in nurseries exclusively dedicated to the commercialization of different seedlings, representing one of the most important economic resources in the agronomic field. For this reason, in many works the effect of different plant growth promoting bacteria on tomato seedlings have been evaluated, such as Burkholderia [40], Azotobacter and Serratia [41], Rhizobium [14], Bacillus [40,41,43,47], Pseudomonas [37,38,40,41] and Arthrobacter [48]. Nevertheless, this is the first study about the effect of Mesorhizobium strains on the growth and development of tomato seedlings. The obtained results showed that both of the assayed strains significantly improve the growth of shoots and roots of tomato seedlings without significantly affecting the percentages of the main macronutrients, N and P. This increase can be due to the biostimulant effect mediated by the IAA produced by these strains, as was reported for other bacteria producing this phytohormone [40]. Moreover, the significant increase of Ca is remarkable in seedlings inoculated with the strain CCANP14 because tomato plants have high requirements for this element involved in nutrition and tomato resistance to bacterial wilt diseases [72]. Additionally, there is a notable increase of K in seedlings inoculated with the strain CCANP122 since this element is involved in plant water regulation [73]. In the case of Mn and Zn, although their content was lower in inoculated plants, the values were higher than those considered enough for a suitable growth of tomato plants [74]. Although the ability to promote barley by a strain of Mesorhizobium mediterraneum [69] and Indian mustard by other strains of Mesorhizobium loti [75] had been previously reported, this is the first report about the ability of strains from different Mesorhizobium strains to promote the growth of the seedlings of a vegetable widely cultivated worldwide.

Conclusions
The results from this study showed that strains of different Mesorhizobium species have several plant growth-promoting mechanisms in addition to symbiotic N fixation, including IAA and siderophores production and phosphate and potassium solubilization. Selected PGPR strains of M. ciceri and M. tamadayense produce cellulases and cellulose that are able to form biofilms in abiotic surfaces and in roots of tomato increasing the growth of the seedlings of this plant. This is the first report about the potential of different Mesorhizobium species to promote the growth of a vegetable and considering their safety for human, animal and plant health after decades of use as bioinoculants. They are environmentally-friendly alternatives to chemical fertilizers for non-legume crops in the framework of sustainable agriculture.  Acknowledgments: The authors thank the Strategic Research Programs for Units of Excellence from Junta de Castilla y León (CLU-2O18-04). The authors dedicate this article to the memory of Prof. Solange Oliveira, who dedicated her last years to the study of the genus Mesorhizobium.

Conflicts of Interest:
The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.