Phylogeny and symbiotic effectiveness of indigenous rhizobial microsymbionts of common bean (Phaseolus vulgaris L.) in Malkerns, Eswatini

In most legumes, the rhizobial symbionts exhibit diversity across different environments. Although common bean (Phaseolus vulgaris L.) is one of the important legumes in southern Africa, there is no available information on the genetic diversity and N2-fixing effectiveness of its symbionts in Malkerns, Eswatini. In this study, we assessed the phylogenetic positions of rhizobial microsymbionts of common bean from Malkerns in Eswatini. The isolates obtained showed differences in morpho-physiology and N2-fixing efficiency. A dendrogram constructed from the ERIC-PCR banding patterns, grouped a total of 88 tested isolates into 80 ERIC-PCR types if considered at a 70% similarity cut-off point. Multilocus sequence analysis using 16S rRNA, rpoB, dnaK, gyrB, and glnII and symbiotic (nifH and nodC) gene sequences closely aligned the test isolates to the type strains of Rhizobium muluonense, R. paranaense, R. pusense, R. phaseoli and R. etli. Subjecting the isolates in this study to further description can potentially reveal novel species. Most of the isolates tested were efficient in fixing nitrogen and elicited greater stomatal conductance and photosynthetic rates in the common bean. Relative effectiveness (RE) varied from 18 to 433%, with 75 (85%) out of the 88 tested isolates being more effective than the nitrate fed control plants.


Phylogenies based on housekeeping genes (dnaK, glnII, gyrB and rpoB)
Four housekeeping genes (dnaK, glnII, gyrB and rpoB) were selected for a robust multilocus sequence phylogenetic analysis.The PCR amplification of the dnaK, glnII, rpoB, and gyrB genes yielded bands of about 650, 700, 700 and 500 bp, respectively for the selected rhizobial population.The primers used and temperature conditions are listed in supplementary Table S3.The maximum likelihood phylogenetic trees based on the individual gene sequences placed the isolates in different clusters within the genus Rhizobium.Some isolates consistently grouped together in the 16S rRNA phylogram as well as in the phylograms based on individual housekeeping genes, while others showed discrepancies.For example, isolates such as TUTPvES 7 (2) and TUTPvES 43(2) which grouped with other isolates in Cluster IV in the 16S rRNA gene phylogeny showed a high divergence from those isolates in the dnaK, glnII, gyrB and rpoB phylogenies (Fig. 2; Supplementary Fig. S1-S4).Nevertheless, the grouping of isolates in the individual housekeeping genes were largely consistent.For example, isolate TUTPvES 14(1) aligned with R. tropici in the glnII, gyrB and rpoB phylogenies, but stood separately in the dnaK phylogeney due to the absence of the type strain in that phylogram (Supplementary Fig. S1-S4).

Nodulation and plant growth induced by isolates
The results showed significant differences (p ≤ 0.05) in the nodule number, nodule weight and shoot dry matter (SDM) of common beans inoculated with the different rhizobial isolates (  2).Nevertheless, there was a significant positive correlation between nodule number and nodule dry matter induced by the isolates (Fig. 6a).Generally, the rhizobia that induced higher nodule numbers, also elicited greater shoot dry matter in the common bean host plant, with a few exceptions.There was therefore a significant positive correlation when nodule number and nodule dry matter were each plotted against shoot DM (Fig. 6b Treatment  and c).The values of shoot DM ranged from 0.23 g. plant -1 in plants inoculated with isolate TUTPvES 53(4) to 5.77 g. plant -1 in plants inoculated with isolate TUTPvES 20(1).Of all tested isolates, 65 induced significantly higher shoot biomass than the nitrate fed plants, whereas 24 of the isolates induced lower shoot biomass than the nitrate fed plants.
Relative symbiotic effectiveness (RE) ranged from 18% in the plants inoculated with isolate TUTPvES 53(4) to 433% in the plants inoculated with isolate TUTPvES 20(1).The data showed that over 85% of the isolates were highly effective (with RE between 103 and 433%) while five isolates were less effective (Table 2).

Morpho-genetic diversity of rhizobial symbionts of common bean
While the diversity of rhizobia nodulating common bean has been studied worldwide 9,11 , little is known about the crop's rhizobial microsymbionts in Swati (of Eswatini) soils.The differences in morphological features of bacteria can be used as a preliminary assessment of their diversity.In this study, the diversity of rhizobial symbionts of common bean from Eswatini was evidenced by differences in their growth rate and colony appearance.www.nature.com/scientificreports/However, because distantly related rhizobia may share similar morpho-physiological characteristics, ERIC PCR fingerprinting is often used as a robust tool in distinguishing closely related species 12 .The genetic diversity of common bean rhizobia in Swati soils was further assessed using ERIC-PCR fingerprinting.The dendrogram constructed from ERIC-PCR profiles placed a total of 88 common bean isolates in 11 major clusters when considering a 20% similarity level, which represented 80 ERIC PCR types if considered at a 70% similarity level.The diversity of bacteria nodulating various legumes, including common bean, have previously been analysed using ERIC-PCR profiles 13 .For example, the study by Zinga et al. 9 observed high genetic diversity among rhizobial symbionts of common bean in South African soils.Although the number of authenticated rhizobia in this study varied between common bean genotypes, the isolates from genotype DAB 381 showed greater diversity, as they [e.g., TUTPvES 4(1), TUTPvES 4(2), TUTPvES 4(3) and TUTPvES 4(4)] were distributed in four different ERIC clusters (i.e., cluster II, IV, VII and X) (Fig. 1).Similarly, the isolates from genotype DAB 155 appeared in three of the clusters.Within the major clusters, there was a general tendency for isolates from the same field or genotype to group closely together.These observations are consistent with earlier reports that similarly reported the clustering of rhizobial symbionts of cowpea in Mozambique based on their geographic location of origin 14 .As the fields used to trap rhizobia in this study were in proximity at the Malkerns Research Station, the soil chemical properties did not show marked variation; except for the levels of available P which was slightly higher in the soil planted to SARBEN common bean genotypes (Supplementary Table S2).It may therefore be worthwhile to explore the symbionts of the test legume across contrasting agroecologies of Eswatini to provide more insights into the impact of environmental variables on the diversity of the crop's symbionts.

Phylogenetic positions of rhizobia nodulating common bean in Eswatini
Common bean has been widely reported to be promiscuous to a wide range of rhizobia and can be nodulated by different species including R. etli, R. mayense, R pusense, R lusitanum and R. fabae 9,15 .The fields in Eswatini used in this study had no history of inoculation; thus, the rhizobial isolates obtained were considered indigenous to the area, even though common bean seeds may carry viable rhizobial cells from different geographic environments 16 .
The rhizobial isolates in this study aligned with different species belonging to the Rhizobium genus based on the 16S rRNA gene sequence analysis.As shown in Fig. 2, some of the indigenous isolates shared close relations with R. freirei, R. paranaense, R. pusense and R. tropici.However, because the 16S rRNA gene tends to show low resolution at the species/genus level, different housekeeping genes are often used to decipher closely related species 17,18 .In this study, the phylogeny of rhizobial isolates based on the 16S rRNA gene was mostly congruent with those of the housekeeping genes (dnaK, glnII, gyrB and rpoB), with the test isolates aligning closely with Rhizobium type strains in the different phylograms.As a result, the phylogram based on concatenated glnII + dnaK + rpoB was also congruent with those of individual housekeeping genes and showed close similarity with the type strains R. paranaense and R. pusense (Fig. 3).However, although isolate TUTPvES 14(1) consistently aligned with R. tropici in the 16S rRNA phylogeny and in the phylograms based on the individual housekeeping genes, the isolate stood away from the other type strains in the concatenated gene phylogeny due to the absence of R. tropici in that tree.www.nature.com/scientificreports/These findings are similar to those of Zinga et al. 9 who also found close relatives of R. tropici in the root nodules of common bean in South African soils.Interestingly, although R. phaseoli and R. etli are known symbionts of common bean 19 , the indigenous isolates in this study showed low similarity with those type strains.Rhizobium tropici is a well-known microsymbiont for common bean and has been reported as the main symbiont of the crop in Latin America and in East, West and southern Africa 20 .However, as observed in this study, R. paranaense has previously been isolated from root nodules of common bean in sub-Saharan Africa 20 .The other test isolates which did not align closely with the known type strains will require a detailed description via whole genome sequencing.
As to be expected, the phylograms based on the two symbiotic genes (nifH and nodC) were incongruent with the 16S rRNA and the housekeeping genes phylogenies.For example, although the isolates grouped in different clusters in the housekeeping gene phylogenies, they formed a single cluster in the nifH and nodC phylogenies, an observation that could be attributed to the acquisition of those genes via horizontal gene transfer from a common ancestor since they are located on transmissible plasmids in some Rhizobium species 21,22 .A similar observation was made by Zinga et al. 22 in a study on common bean symbionts from South Africa and Mozambique.Aserse et al. 23 also reported incongruency between the phylogenies inferred from two symbiotic genes and that of the 16S rRNA gene in Ethiopia due to possible inter strain gene transfer and gene recombination.
Dlamini et al. 24 had earlier reported the influence of soil pH and nutrient levels of the distribution of rhizobia associated with Bambara groundnut from different locations in Eswatini.Since the present study focused on the symbionts of common beans grown at one location, it may be worthwhile to explore the diversity of the crop's symbionts across contrasting agroecologies of Eswatini in subsequent works.

Symbiotic efficiency of common bean isolates from Eswatini
Aside from variability in morpho-genetic characteristics, the diversity of common bean isolates in this study were also shown by the variable nodulation, plant growth and photosynthetic parameters they elicited in the  www.nature.com/scientificreports/host plants.Whereas some isolates induced both greater nodule number and nodule weight in the host, others elicited greater nodule weight despite inducing fewer nodules in the host; these observations were probably due to differences in nodule size as well as the N 2 -fixing efficiency of the rhizobial symbionts 25 .For example, despite forming few nodules, isolate TUTPvES 45(1) induced higher nodule dry matter, which led to higher photosynthetic rates and shoot dry matter when compared to the 5 mM nitrate fed plants.In some instances, nodule symbionts elicited lower plant growth and photosynthesis in the host despite inducing high nodulation, an indication that some of the nodule symbionts were less effective in fixing nitrogen (N 2 ).Some studies have observed that legumes can sometimes form ineffective nodules that harbour low N 2 -fixing rhizobia 26 .Nevertheless, nodule number and nodule weight were both positively correlated with shoot biomass, indicating that the nodule symbionts contributed to plant growth promotion.As the uninoculated control plants expectedly showed the least plant growth, the effectiveness of isolates was assessed by comparing the biomass of inoculated plants with that of nitrate fed plants 27 .Of the 88 isolates, 74% elicited greater shoot biomass in the host common bean when compared to the 5 mM nitrate-fed plants, with relative effectiveness values ranging from 103 to 433%.Thus, the soils in Eswatini are home to highly effective rhizobia that can potentially be used to formulate commercial inoculants for increased common bean cultivation upon extensive testing in the field to assess their adaptation to prevailing abiotic conditions.

Conclusion
Based on their ERIC PCR banding patterns, this study revealed a high genetic diversity among common bean symbionts in Eswatini, an observation consistent with several reports on the diversity of rhizobia in African soils.Moreover, multilocus sequence analysis based on the sequences of 16S rRNA, rpoB, dnaK, gyrB, and glnII and symbiotic (nifH and nodC) genes aligned the test isolates with the type strains of Rhizobium muluonense, R. paranaense, R. pusense, R. phaseoli and R. etli.Some of the indigenous isolates showed a high divergence from the known reference type strains, and may require further description via whole genome sequencing.Aside from their diversity, a glasshouse experiment showed that most of the isolates were efficient in fixing nitrogen, and elicited greater stomatal conductance and photosynthetic rates in the common bean host.Relative symbiotic effectiveness (RE) of the isolates varied from 18 to 433%, with 75 out of the 80 tested isolates producing greater shoot biomass than the nitrate fed control plants.

Nodule collection and bacterial isolation
Root nodules were collected from different common bean genotypes grown at the Malkerns Research Station in Eswatini during the 2017/2018 cropping season.The soil chemical properties as well as the climatic data of the Malkerns site during the 2017/2018 cropping season are presented in Supplementary Tables S1 and S2.For this, the plants were carefully dug out at the early podding stage and the nodulated roots separated from the shoots.The roots were then transported to the laboratory in prelabelled zip-lock plastics in a cooler box with ice.The roots were rinsed in running tap water to remove debris, and the nodules attached to small root segments www.nature.com/scientificreports/removed and preserved on silica gel in plastic vials prior to bacterial isolation.Bacteria were isolated from the nodules according to the procedure described by Somasegaran and Hoben 28 .For this, surface sterilized nodules were crushed in a loop of sterile water in a petri dish, and the nodule macerate streaked on yeast mannitol agar (YMA) plates and incubated at 28 °C.The plates were monitored daily for colony appearance.The number of days taken for colonies to appear as well as other morphological characteristics (colony size/shape, colour, and texture) were recorded 29,30 .

Nodulation bioassay in the glasshouse
A total of 162 bacterial isolates were obtained from the root nodules of the different common bean genotypes.
To fulfil Koch's postulates, single-colony cultures from these bacterial isolates were assessed for their ability to elicit root nodules on their homologous Kranskop common bean genotype in a naturally lit glasshouse under aseptic conditions.For this, the common bean seeds were surface sterilized as described by 29 and planted in sterile (autoclaved) sand which were contained in sterile plastic pots.After germination, the seedlings were inoculated with 1 mL broth culture of the different bacterial isolates grown to the exponential phase, with three replicate pots per isolate.The plants were watered with autoclaved nitrogen-free nutrient solution and sterile distilled water in alternation.Uninoculated plants and plants treated with 5 mM KNO 3 were used as negative and positive controls, respectively.At 60 days after planting, the plants were harvested and assessed for nodulation; isolates which elicited nodules on three replicate plants were considered as rhizobia.The use of plant materials in different aspects of this study complied with relevant institutional, national and/or institutional guidelines.

Assessment of plant growth and photosynthetic rates induced by isolates
The authenticated rhizobia were further assessed for their symbiotic efficiency using plant nodulation, growth, and photosynthetic rates as reference parameters.For this, photosynthetic rates (A), stomatal conductance (gs) and transpiration rates (E) were measured on the youngest fully expanded trifoliate leaves of common bean plants inoculated with the different isolates using a portable infrared red gas analyser, version 6.2 (LI 6400XT, Lincoln, Nebraska, USA).The chamber conditions were set as follows: photosynthetic flux density of 1000 μmolm -2 s -1 , reference CO 2 concentration of 400 μmolmol -1 and flow rate of 500 μmols -1 .The gas exchange measurements were performed between 9.00 am and 12.00 pm at 60 days after planting 31 .Thereafter, the plants were harvested, separated into shoots and nodulated roots, and the nodules collected from the roots.The shoots and nodules were separately oven-dried in brown paper envelopes, and weighed to determine shoot and nodule dry matter, respectively.The relative effectiveness (RE) of isolates was calculated by expressing the shoot biomass of plants inoculated with the rhizobial isolates as a percentage of the shoot biomass of plants treated with the 5 mM KNO 3 32 .

Molecular characterization of isolates
Genomic DNA extraction and ERIC PCR fingerprinting Bacterial genomic DNA extraction was carried out using the GenElute bacterial DNA isolation kit by following the manufacturer's instructions (Sigma Aldrich, USA).The quality of DNA was assessed on a 1% agarose gel stained with ethidium bromide.The genomic DNA of rhizobial isolates were subjected to ERIC-PCR fingerprinting.The final PCR reaction volume was 25 μL and contained 1 μL of genomic DNA,12.5 μL 2× myTaq PCR master mix (Bioline USA), 1 μL each of the forward and reverse ERIC primers and 9.5 mL double distilled water.The DNA amplification was carried out in a Thermal cycler (T100 Bio-Rad, USA) using standard temperature profiles 24,33 , and the PCR products were subjected to gel electrophoresis on a 1% agarose gel at 85 V for 5 h.A cluster analysis to determine the similarities among isolates using the Jaccard's similarity coefficient was performed with the Bionumerics software (version 8.1).

PCR amplification of 16S rRNA, housekeeping (rpoB, dnaK, gyrB, and glnII) and symbiotic (nifH and nodC) genes
The amplification of the 16S rRNA, housekeeping genes (rpoB, dnaK, gyrB, and glnII), and symbiotic (nifH and nodC) genes were carried out in a 25 μL PCR reaction mixture which contained 1 μL bacterial DNA, 3 μL of 5× myTaq buffer, 1 μL each of the forward and reverse primers of the gene of interest, 0.1 μL Taq polymerase (Bioline, USA), and 18.9 μL sterile ultrapure water using standard temperature profiles (Supplementary Table S3) in a T100 Bio-Rad Thermal Cycler, USA.The amplified gene products were confirmed by gel electrophoresis in a 1% agarose gel and the image captured using the Geldoc Tm XR + system (Bio-Rad, USA).The primers used and their temperature profiles are shown in Supplementary Table S3.

Sequencing and Phylogenetic analysis of amplified genes
The PCR-amplified gene products were purified using a PCR cleanup kit (NEB, USA) following the manufacturer's instructions.Purified DNA was sequenced at Macrogen laboratories (The Netherlands).The software BioEdit 7.0.9.0 was used to confirm the quality of sequences 34 .The sequences of each gene were subjected to BLASTn in the National Centre Biotechnology Information (NCBI) database to identify closely related rhizobial species.The alignment of the reference strain sequences with the test rhizobial isolates were done with CLUSTAL W and phylogenetic trees were inferred using MEGA 7 software 35 .The Kimura 2-paramete model with uniform rates among the sites was used to calculate evolutionary distances and evolutionary history was inferred using www.nature.com/scientificreports/ the maximum likelihood method.The robustness of tree branching was estimated using 1000 bootstrap replicates of the sequence 36 .

Statistical analysis
All quantitative data collected from the glasshouse experiment were subjected to a one-way ANOVA using the STATISTICA program (Version 10).The quantitative datasets such as nodule number, nodule dry matter, shoot dry matter, photosynthetic rates (A), stomatal conductance (gs), leaf transpiration (E) and relative symbiotic effectiveness (RE) were tested for normality by calculating skewness and kurtosis values using the Data Analysis component of Excel.The skewness and kurtosis values ranged from − 0.10 to + 1.51 and −1.67 to + 2.30, respectively, and are consistent with values of a normal distribution 37 .Where there were significant treatment differences, the Duncan's multiple range test was used to separate the means at p ≤ 0.05.Correlation analyses were done to assess the relationship between measured parameters.

Figure 1 .
Figure 1.Dendrogram of ERIC fingerprints revealing the presence of high genetic diversity among the 129 indigenous rhizobial microsymbionts of common bean from Malkerns in Eswatini.Red vertical line indicates 70% similarity cut of point.

Figure 2 .
Figure 2. Maximum-likelihood phylogenetic tree inferred from 16S rRNA gene sequences of common bean symbionts from Eswatini.Phylogenetic trees were inferred using MEGA 7 software 35.The Kimura 2-paramete model with uniform rates among the sites was used to calculate evolutionary distances.

Figure 3 .
Figure 3. Maximum-likelihood phylogeny of microsymbionts of common bean from Eswatini inferred from concatenated glnII + dnaK + rpoB gene sequences.Phylogenetic trees were inferred using MEGA 7 software 35.The Kimura 2-paramete model with uniform rates among the sites was used to calculate evolutionary distances.

Figure 4 .
Figure 4. Maximum-likelihood phylogeny of microsymbionts of common bean from Eswatini inferred from nodC gene sequences.Phylogenetic trees were inferred using MEGA 7 software 35.The Kimura 2-paramete model with uniform rates among the sites was used to calculate evolutionary distances.

Figure 5 .
Figure 5. Maximum-likelihood phylogeny of microsymbionts of common bean from Eswatini inferred from nifH gene sequences.Phylogenetic trees were inferred using MEGA 7 software 35.The Kimura 2-paramete model with uniform rates among the sites was used to calculate evolutionary distances.

Table 1 .
Isolates used in the diversity study, their host genotypes, and morphological characteristics.

Table 2 .
Plant growth, nodulation and photosynthesis parameter of common bean inoculated with indigenous rhizobial strains in the glasshouse.Values (Mean ± SE) with dissimilar letters in a column are significantly different at **p < 0.01.