Distribution and correlation between phylogeny and functional traits of cowpea (Vigna unguiculata L. Walp.)-nodulating microsymbionts from Ghana and South Africa

Cowpea (Vigna unguiculata L. Walp.) is indigenous to Africa, and highly valued for its N2-fixing trait and the nutritional attributes of its grain and leaves. The species’ ability to establish effective symbiosis with diverse rhizobial populations gives it survival and growth advantage in N-limited environments. To explore the functional diversity and phylogenetic positions of rhizobia nodulating cowpea in Africa, nodules were collected from various cowpea varieties grown in soils from the Guinea savanna and Sudano-sahelian agroecologies of Northern Ghana, and from the lowveld and middleveld areas of Mpumalanga Province in South Africa. Box-PCR profiling and multilocus sequence analysis revealed the presence of diverse microsymbionts responsible for cowpea nodulation across the study sites. BOX-PCR amplifications yielded variable band sizes, ranging from 618 bp to 5354 bp, which placed the isolates in six major clusters (Cluster A–F). Phylogenetic analysis based on 16S rRNA, atpD, glnII, gyrB, rpoB, nifH and nodC genes revealed the presence of diverse Bradyrhizobium sp. closely related to Bradyrhizobium daqingense, Bradyrhizobium subterraneum, Bradyrhizobium yuanmingense, Bradyrhizobium embrapense, Bradyrhizobium pachyrhizi, Bradyrhizobium elkanii and novel Bradyrhizobium species in the soils studied, a finding that could be attributed to the unique edapho-climatic conditions of the contrasting environments. The test isolates exhibited distinct symbiotic efficiencies, and also induced variable (p ≤ 0.001) photosynthetic rates, leaf transpiration, total chlorophyll and shoot biomass accumulation on cowpea (their homologous host). Canonical correspondence analysis showed that the distribution of these microsymbionts was influenced by the concentrations of macro- and micronutrients in soils. The pairwise genetic distances derived from phylogenies and nodule functioning showed significant (p < 0.05) correlation, which suggests that local environmental factors played a major role in the cowpea-Bradyrhizobium symbiosis.

planted at Nelspruit, South Africa ( Fig. 1; Table 1). There was a general tendency for isolates within each cluster to group in close proximity based on their locations of origin (Table 1). In this study, however, cowpea genotypes did not influence isolate groupings since most clusters contained isolates from different cowpea genotypes (Table 1).   (Table S1). The 16S rRNA gene-based phylogeny assigned all the selected isolates to the genus Bradyrhizobium (Fig. 2), and placed them in four groups (Group I -IV) within the phylogenetic tree. In Group I, isolates TUTVUGH2, TUTVUGH6, TUTVUGH9, TUTVUGH17, TUTVUGH22, and  TUTVUGH23 obtained from Ghana and TUTVUSA31 and TUTVUSA33  Concatenated sequence and phylogenetic analysis. Due to inconsistencies in the number of gene sequences obtained for the test isolates, two concatenated sequences and phylogenetic analyses were done. The variations in number of sequences obtained were due to failure to amplify the gene in some isolates, as well as to poor quality of sequences obtained in some instances. Two phylogenetic trees were constructed based on concatenated sequences of atpD-glnII-gyrB-rpoB genes for 11 isolates, and atpD-glnII-rpoB for 14 isolates (Figs 3 and 4). The phylograms obtained for concatenated sequences of four ( Fig. 3) or three (Fig. 4) 6). On the other hand, isolates TUTVUGH25 and TUTVUSA44 formed outgroups of isolates in Group II in the nodC phylogeny, but respectively shared 80.7% and 79.6% sequence homology with B. embrapense SEMIA 6208 T , the closest type strain (Fig. 6). followed by isolate TUTVUSA28 (21.5 µmol ms −2 s −1 ), and TUTVUGH18 (20.8 µmol ms −2 s −1 ). All test isolates, including the commercial Bradyrhizobium strain CB756 elicited greater total leaf chlorophyll and photosynthesis when compared to the 5 mM KNO 3 -fed plants. In general, increased photosynthetic rates were mostly associated with higher leaf transpiration, stomatal conductance and greater chlorophyll in leaves ( Table 2). The test isolates induced variable (p ≤ 0.001) nodule number and nodule dry matter on cowpea plants ( Table 2). Isolates TUTVUSA41 and TUTVUSA50 elicited the highest nodule formation on cowpea (the homologous host), while TUTVUSA43 and TUTVUSA45 induced the least nodule number (Table 2) even though the four isolates shared the same cluster in all phylogenetic trees. In fact, high shoot dry weights were also recorded by isolates TUTVUSA41 and TUTVUSA50, which exhibited superior symbiotic efficiency with relative effectiveness values of 166% and 159%, respectively (Table 2).
Soil influence on bradyrhizobial distribution. CCA analysis was used to correlate bradyrhizobial distribution with environmental variables such as the concentrations of micro-and macronutrients (namely, B, Fe, Mn, Ca, Mg, Na, N, P, K, and soil pH) in soils (Fig. 7). A CCA ordination graph was constructed with only those variables which showed significant influence on bradyrhizobial distribution (p < 0.001 at permutation 999). In the CCA plot, the total mean square contingency coefficient (inertia) was 13.16, of which 10% was constrained (explainable) and 90% was unconstrained (unexplainable). Only 6.2% of the variation was explained by both CCA1 and CCA2 axes. As shown in the CCA plot, the concentrations of Mn, Fe and N in soils showed stronger correlations with the first canonical axis (CCA1), although the effect of N was in the opposite direction (Fig. 7). Similarly, soil B and Na showed significant correlations with the second canonical axis (CCA2), although in opposite directions. Nevertheless, both B and Na also contributed to determine the first canonical axis but to a lower extent compared to Mn, N and Fe (Fig. 7). On the other hand, the concentrations of Mg and Ca in soils showed slight correlations with the CCA1 axis. However, soil Ca and Na contents were orthogonal to each other (Fig. 7). The CCA plot revealed that, the concentration of N, Mg, Ca, Mn and Fe exerted a stronger influence on the genomic variations among the bradyrhizobial isolates with the CCA1, followed by B, P and Na with the CCA2 axis. Among the test locations, the isolates from Nyankpala in Ghana were closely related to soil Mn. On the other hand, the isolates obtained from Kliplaadrift in South Africa were highly influenced by increasing soil N, Mg, Ca and P while those from Nelspruit were influenced by the concentration of Na. Furthermore, the concentration of B in soils showed a strong influence on the isolates obtained from Garu and Damongo both in Ghana (Fig. 7). Phylogenetic and functional correlations. The correlations performed between genetic distances obtained from the phylogenies of concatenated housekeeping genes and nifH genes and soil chemical properties or symbiotic parameters were significant (p < 0.05). The results showed that concatenated housekeeping gene phylogeny and soil chemical properties were highly correlated (r = 0.75, p = 0.01) (Fig. 8a). The relationship between genetic distances of both concatenated housekeeping genes and nifH gene phylogenies versus the genetic distance of symbiotic parameters was assessed and found to be significant (r = −0.36, p = 0.04; r = −0.54, p = 0.03, respectively), though negatively (Fig. 8b,c).

Discussion
The isolation and subsequent symbiotic characterization of naturally adapted rhizobial symbionts of legumes constitute the first step to strain identification for inoculant production 20 . This study assessed the genetic diversity and phylogeny of rhizobial symbionts of cowpea from different locations in Ghana and South Africa, using rep-PCR (Box-PCR) fingerprinting and multilocus sequence analysis. The Box-PCR profiles revealed high genetic diversity among the native bacterial symbionts nodulating cowpea in its region of origin in Africa (Fig. 1). Isolates from Kliplaadrift in South Africa had the most diverse population, and occupied five major clusters (Figs 1 and 9). These results support a previous report which showed high genetic variation among bacterial symbionts nodulating cowpea in Botswana, Ghana and South Africa, with South African soils harbouring more diverse rhizobial populations 21 . Except for isolate TUTVUSA45 which stood alone in the dendrogram (Fig. 1), the remaining six major clusters identified in this study contained isolates from different locations in the two countries studied ( Fig. 1; Table 1). There was however a general tendency for isolates within a cluster to group closely to each other based on their country of origin. The observed geographic distribution of the isolates in this study has been reported for cowpea-nodulating microsymbionts in Senegal, Greece and Mozambique 10,13,23 . Although the cowpea genotypes had no marked effect on the clustering of rhizobial isolates in this study (Table 1), they nevertheless influenced the diversity of nodule occupants. And this could be due to differences in the profile of nod gene-inducers present in seed and root exudates which differentially shaped the rhizobial microbiome in the rhizosphere and hence nodule occupancy 31 .
To infer the phylogenetic positions of cowpea isolates, 16S rRNA, protein-coding housekeeping (atpD, glnII, gyrB and rpoB) and symbiotic (nifH and nodC) genes were sequenced and analysed. As found in other studies of microsymbionts nodulating cowpea in Africa 10,13,28 , the isolates in this study clustered with B. elkanii, B. daqingense, B. kavangense, B. subterraneum, B. yuanmingense and B. vignae from sequence analysis of the 16S rRNA gene (Fig. 2). Despite the success of Box-PCR fingerprinting in determining genetic differences among cowpea isolates in this study, the approach was incongruent with the 16S rRNA-based phylogeny, a finding reported by Menna et al. 32 . However, the topologies of the phylograms constructed from the concatenated sequences of atpD-glnII-gyrB-rpoB and atpD-glnII-rpoB genes were consistent, and congruent with the 16S rRNA-based phylogram, thus allowing for a clearer delineation of the phylogenetic positions of the isolates used in this study. For example, isolates TUTVUSA41, TUTVUSA43, TUTVUSA45 and TUTVUSA48 grouped together in the single atpD, glnII, gyrB and rpoB gene phylogenies (Figs S1-S4). Isolates TUTVUGH17, TUTVUGH18, TUTVUGH24 from Ghana as well as TUTVUSA36 and TUTVUSA44 from South Africa similarly formed separate clusters ; Figs S1-S4). Isolates TUTVUSA41, TUTVUSA43, TUTVUSA45 and TUTVUSA48 from South Africa were tightly grouped in the two concatenated trees (Figs 3 and 4), and away from any reference type strains, despite being closer to B. kavangense, B. subterraneum, B. centrolobii and B. yuanmingense with 99.2-100% sequence identity in the 16S rRNA phylogeny. Although this could suggest that they are novel strains, a recent report by Helene et al. 33 showed that high 16S rRNA sequence identity per se does not confirm isolate delineation with reference type strains. But isolates TUTVUSA36 and TUTVUSA44 were closely related to B. pachyrhizi PAC48 T (which was originally isolated from Pachyrhizus erosus in Costa Rica) with 97.9-98.3% sequence identity in the concatenated gene phylogenies 34 . These findings are consistent with previous reports which showed that B. pachyrhizi is associated with cowpea nodulation in the soils of Angola and Mozambique 13,20 . Isolates TUTVUGH17, TUTVUGH18, TUTVUGH24, TUTVUGH22 and TUTVUGH6 (all from Ghana) clustered with the B. elkanii group in the 16S rRNA phylogeny, but not closely with any known species in the other phylograms, suggesting that they could be novel species of Bradyrhizobium. Furthermore, our results also showed that isolate TUTVUSA28 consistently clustered in all phylogenies with B. daqingense, a strain originally isolated from soybean nodules in China and subsequently found to induce effective nodulation on cowpea 35 . The presence of B. daqingense in African soils could be caused by its introduction with soybean seeds from China, the country of origin of the crop 36 . But TUTVUGH25 was closely related to B. vignae which was initially isolated from cowpea in Namibia 37 . Interestingly, isolates TUTVUSA41, TUTVUSA43, TUTVUSA45, TUTVUSA48 in the concatenated atpD-glnII-gyrB-rpoB phylogeny, together with TUTVUSA50 in the atpD-glnII-rpoB phylogeny (see Group III) could be novel cowpea-nodulating symbionts since they shared low sequence homology (94.5-95.7%) with their closest related reference type strains in those phylograms (Figs 3 and 4). The nifH and nodC phylogenies of cowpea isolates in this study were consistent with each other, and congruent with the 16S rRNA-based phylogeny which suggests coevolution of these symbiotic genes. Isolates TUTVUGH17, TUTVUGH18, TUTVUGH22 and TUTVUGH24 which shared 98.9-100% sequence identity in the symbiotic gene phylogenies grouped together and stood alone, but shared 96.5-97.5% sequence homology with B. tropiciagri and B. embrapense, which is an indication that they might represent novel species of the symbiovar tropici. Nevertheless, cowpea is a promiscuous host, and was recently reported to nodulate with different symbiovars in Africa and Europe 24 . Furthermore, the congruency of isolates TUTVUSA41, TUTVUSA43, TUTVUSA48 and TUTVUSA50 from South Africa and isolates TUTVUGH17, TUTVUGH18, TUTVUGH22 and TUTVUGH24 from Ghana in all the phylogenetic trees could suggest the maintenance of symbiotic genes through vertical gene transfer. Also, the close relationship of isolates TUTVUSA33 and TUTVUSA36 with B. pachyrhizi and the presence of many novel Bradyrhizobium species could be due to the unique edapho-climatic conditions of the study sites in Africa. Species of B. pachyrhizi bacteria had previously been isolated from cowpea in Mexico and Greece where the climates are warm and the soils acidic 23,38 , as found at the study sites in Africa. A previous study demonstrated the effect of soil factors on the survival, persistence and diversity of rhizobia in soils 39 , which in turn influenced plant nodulation and symbiotic N 2 fixation 40 . In this study, soil mineral nutrients were found to influence the diversity of bradyrhizobial populations in South African and Ghanaian soils. For example, the South African isolates were highly influenced by the concentrations of N, P and Na in the soil. In the phylogenetic tree, those isolates (i.e. TUTVUSA41, TUTVUSA43, TUTVUSA45, TUTVUSA48 and TUTVUSA50) from South Africa were also found to be closely related to B. subterraneum, B. kavangense, B. centrolobii and B. yuanmingense. However, in China, the distribution of B. yuanmingense was found to be affected by soil K, although other Bradyrhizobium species were also associated with soils with high P concentration 41 . Han et al. 42 also showed that rhizobial diversity was altered by soil phosphorus. In this study, the distribution of isolates from Ghana were strongly influenced by high soil B, Fe and Mn, suggesting the possible roles of these elements in the survival and functioning of microsymbionts in the Ghanaian environment.
Glasshouse evaluation of the cowpea isolates studied revealed marked differences in their symbiotic effectiveness as evidenced by the significantly large variations in plant nodulation, shoot biomass, leaf chlorophyll and photosynthetic rates induced by the bacterial isolates (Table 2). Cowpea nodulated by the test isolates showed markedly higher leaf chlorophyll which probably stimulated greater photosynthesis, stomatal conductance and transpiration rates when compared to the 5 mM KNO 3 -fed plants. The C sink strength of the nodulated cowpea plants induced greater photosynthetic rates which in turn increased biomass accumulation and promoted plant growth relative to nitrate-feeding 16 . The nitrate-fed plants recorded much lower chlorophyll content and leaf photosynthesis despite having similar shoot biomass as some inoculated plants, suggesting that N 2 fixation stimulated greater photosynthesis than nitrate-feeding. A recent report by Kaschuk et al. 43 also found that inoculating Figure 9. African test locations and rhizobial occupation in different Box-PCR clusters. For each location, the number of segments indicate the number of Box-PCR clusters occupied by isolates from that site. Uppercase letters in segments represent the labels of Box-PCR clusters (see Table 1). The area of each segment is proportional to the number of isolates from a given location occupying that cluster. soybean plants with effective Bradyrhizobium strains maintained leaf chlorophyll content, stimulated greater photosynthetic rates and delayed senescence over nitrate-fed plants. In this study, effective N 2 fixation in the nodulated cowpea plants therefore provided the needed N for chlorophyll and Rubisco biosynthesis, which resulted in enhanced photosynthetic rates 43 . Furthermore, a comparison of the phylogenetic and functional genetic distances between test isolates showed a clearly significant (p < 0.05) relationship. The positive correlation found between concatenated gene phylogeny and soil properties could be due to niche conservation and mineral nutrition of the test isolates. The housekeeping genes used in this study are conserved in the isolates, which suggest that the phylogenetically-clustered isolates were similar in their environmental distribution. A significant but negative correlation was found when symbiotic parameters were plotted against nifH and concatenate genes phylogenies (Fig. 8b,c), indicating that functional similarity of isolates differed from their phylogenetic closeness.

Conclusion
Taken together, this study represents an important contribution to the literature about microsymbionts nodulating grain legumes in Africa, especially with the evidence provided here on the presence of high genetic diversity of cowpea microsymbionts in Ghanaian and South African soils. Phylogenetic studies of rhizobial isolates from cowpea planted in Ghana and South Africa revealed the presence of potentially novel groups of rhizobia in those environments that are still waiting to be properly identified. The significant correlation between phylogeny and functional genetic distances suggests possible impact of complex host-symbiont dynamics on the stability of the cowpea-Bradyrhizobium symbiosis.

Materials and Methods
Origin of cowpea root nodules. The nodules used in this study were collected from cowpea plants grown at four locations in the Northern Region (Guinea savanna agroecology) and one location in the Upper East Region (Sudano-sahelian agroecology) of Ghana, and others from two locations, Nelspruit and Klipplaatdrift, respectively, in the lowveld and middleveld areas of the Mpumalanga Province of South Africa (Table 3; Fig. 9). The soil chemical properties as well as geographic coordinates of the sampling sites are shown in Table 3. The nodules used in this study were harvested from either field-grown plants (F) or trapped in the glasshouse (G) using soils collected from the field (Table 3). To trap rhizobia directly in field soils, five cowpea genotypes (Apagbaala, Padi-tuya, Songotra, Omandaw and IT90K-277-2) were planted at Garu, and seven genotypes (Apagbaala, Bawutawuta, Nhyira, Padi-tuya, Songotra, Omandaw and IT90K-277-2) at Kliplaadrift. Rhizobia in soils collected from Nyankpala, Damongo and Nelspruit were trapped in the glasshouse using three cowpea genotypes (Apagbaala, Songotra and IT90K-277-2).
Trapping rhizobia in the field. To trap soil rhizobia under field conditions, cowpea seeds were surface-sterilized in 95% ethanol (3-5 minutes), followed by immersion in NaOCl for 3 minutes, and rinsed 5 times with sterile distilled water. The sterilized seeds of each test cowpea genotype were sown in 3 × 2 m plots with three replications at Garu and Kliplaadrift. After germination, weeds were controlled using hand hoes when necessary. At flowering (at 45 days after planting), 3 plants were dug out per plot and separated into shoots and nodulated roots, placed in pre-labelled brown paper bags and transported to the laboratory, where the roots were gently washed in running tap water to remove soil and adhering debris. Healthy nodules were detached with small root segments and dehydrated on silica gel prior to bacterial isolation.
Trapping rhizobia in the glasshouse. Where field-planting could not be carried out, soils were sampled from those locations (namely, Nyankpala, Damongo and Nelspruit) for trapping rhizobia under glasshouse conditions using three cowpea genotypes (Apagbaala, Songotra and IT90K-277-2) as host legumes. Two sterilized seeds of each genotype were planted in triplicate pots containing sterile (autoclaved) sand. Soil inocula were prepared by suspending 20 g of each soil sample in 1000 mL of sterile distilled water. Each pot was then inoculated with soil suspension from different locations. The plants were supplied with sterile (autoclaved) N-free nutrient Isolation of bacteria from nodules. Bacterial isolation from cowpea root nodules was carried out, as described by Somasegaran and Hoben 45 . Healthy and functional nodules were selected from each source for bacterial isolation. Briefly, the nodules were surface-sterilized by immersion in 75% ethanol for 1 minute, then washed in 3.5% NaOCl for 3 minutes, followed by rinsing five times with sterile distilled water. Each surface-sterilized nodule was then crushed in a loop of sterile distilled water in a sterile petri dish, and the nodule macerate streaked on yeast mannitol agar (YMA) plates, and incubated at 28 °C. Colony appearance was observed from two to twelve days after incubation, and single colonies were selected and streaked onto YMA plates for further characterization.
Bacterial authentication in the glasshouse. In fulfilment of Koch's postulates, single-colony cultures were evaluated for their ability to form root nodules on their homologous host 46 . Cowpea genotype Padi-tuya was used to test the nodulation ability of the bacterial isolates. Before planting, seeds were surface-sterilized 46 and two seeds planted in sterile autoclaved sand contained in sterile pots in a naturally lit glasshouse. The seedlings were thinned to one plant per pot after germination and grown in average temperature of 28 °C in the glasshouse. Four replicate pots were used for each isolate. Six-d-old seedlings were inoculated with 1 mL broth suspension of bacterial culture grown to exponential phase (10 6 -10 7 cells/ml) using sterile micropipettes. Uninoculated plants and 5 mM NO 3 −1 -fed plants were included as controls. The inoculated seedlings were fed with sterile N-free nutrient solution 44 and sterile distilled water in alternation. All the cowpea treated plants were harvested at 50 days after planting (DAP) and assessed for nodulation.

Gas-exchange studies and symbiotic effectiveness.
To assess symbiotic effectiveness of the bacterial isolates, photosynthetic rates (A), stomatal conductance (gs) and leaf transpiration (E) were measured on young and fully expanded trifoliate leaves of each replicate plant at 50 DAP using portable infrared red gas analyzer, version 6.2 (LI 6400XT, Lincoln, Nebraska, USA). The chamber conditions used included photosynthetic flux density of 1000 μmolm −2 s −1 , reference CO 2 concentration of 400 μmolmol −1 and flow rate of 500 μmols −1 . Gas-exchange measurements were carried out between 8:30 am and 12:30 pm. The same leaves used for gas-exchange measurements were plucked for chlorophyll determination. For each leaf sample, chlorophyll was extracted from 6 leaf discs (each with an area of 0.786 cm 2 and weighing ≈ 19.3 mg) using preheated (65 °C) dimethyl sulfoxide 47 . The absorbance of leaf chlorophyll extracts were measured at 645 nm and 663 nm on a Jenway 7300 spectrophotometer and total chlorophyll calculated using the equations described by Richardson et al. 47 . The plants were then uprooted and assessed for nodule number (NN), nodule dry matter (NDM), and shoot dry matter (SDM) after oven-drying at 60 °C for 72 h. The relative effectiveness of isolates (RE) was calculated as the shoot biomass of inoculated plants expressed as a percentage of the shoot biomass of the 5 mM KNO 3 -fed cowpea plants.
Extraction of bacterial genomic DNA and BOX-PCR fingerprinting. Bacterial genomic DNA was extracted using Sigma's Bacterial Genomic DNA Kit following the manufacturer's instructions (GenElute TM ). DNA integrity was checked on 1% agarose gel stained with ethidium bromide.

PCR amplification of the
Sequencing and phylogenetic analysis. For sequencing, amplified PCR products were purified using PCR Cleanup kit (NEB, USA) and following the manufacturer's instruction. The purified amplified DNA was sent to Macrogen (Netherlands) for sequencing. Thereafter, the quality of sequences was verified using the software BioEdit 7.0.9.0 48 . The BLASTn program was used to identify closely related species in the NCBI database. Pairwise and multiple sequence alignments were done with CLUSTALW, and phylogenetic trees constructed by means of the maximum likelihood statistical method using MEGA 7 software 49 . The evolutionary history was inferred by using the Maximum Likelihood method based on the Kimura 2-parameter model 50 . The robustness of branching was estimated using 1000 bootstrap replicates 51 . Phylogenetic and functional correlation. Quantitative data generated from pair-wise genetic distances obtained from gene phylogeny were compared with parameters of nodule functioning elicited by the cowpea isolates. The pair-wise genetic distances of isolates generated from concatenated genes (atpD + glnII + gyrB + rpoB) and nifH gene phylogenies were determined using Kimura 2-parameter matrices in MEGA7 software 49 . The genetic distances of the same isolates based on soil chemical properties and symbiotic parameters (namely, nodule number, nodule dry matter, shoot dry matter, photosynthetic rate, leaf transpiration and total chlorophyll) were also determined using NTSYSpc software, version 2.21 52 . The relationship between genetic distances based on gene sequences, soil chemical properties and symbiotic parameters of isolates were explored using correlation and regression analysis by means of the software STATISTICA version 10.0 53 .

Statistical analysis.
The effect of soil factors on the distribution of cowpea-nodulating bradyrhizobia was examined using canonical correspondence analysis (CCA) with vegan (version 2.4-2) 54 of R software 55 .
Here, we determined which soil or environmental factor was frequently related to the distribution of the test cowpea-nodulating bradyrhizobia. The graph analysis was done for only the soil factors that showed significant contribution. The general permutation test was used to assess the statistical significance of the ordination axes. Quantitative data including nodule number, nodule DM, shoot DM, photosynthetic rate (A), leaf transpiration (E) and total chlorophyll were subjected to normality test by determining the skewness, kurtosis, mean and median values of each dataset (n = 72) using the Data Analysis component in Excel 2016. For each dataset, the mean ≈ median and the skewness and kurtosis values ranged between −1.51 to +0.75 and −0.9 to +0.44, respectively, which are within the range of values (±2) consistent with a normal distribution 56 . The data were then subjected to a 1-way ANOVA using STATISTICA 10.0 program 53 . Where there were significant treatment differences, the Duncan multiple range test was used to separate the means at p ≤ 0.05. Pair-wise genetic distances of concatenated and nifH gene sequences, as well as that generated from symbiotic and plant growth data of cowpea isolates were subjected to correlation analysis using STATISTICA 10.0 program.

Data Availability
Datasets generated and/or analysed during the current study are available from the corresponding author on reasonable request. Nucleotide sequences have been deposited in NCBI GenBank (accession numbers MH339749 -MH339859).