Phenotypic and genotypic characteristics of cowpea rhizobia from soils of Ethiopia

Phenotypic and genotypic characteristics of cowpea rhizobia indigenous to soils of Ethiopia are unknown. Forty indigenous cowpea rhizobial isolates were collected from cowpea-growing areas of the country and were characterized for their growth and genetic properties. Based on their cultural characteristics, the isolates were categorized into fast (FG), slow (SG), and extraslow-growings (ESG). The FG, SG, and ESG isolates had mean generation time (h)/colony diameter (mm)/date of turbidity formation (d) in the range of 2.5-7.5/2-4/2-3, 7.5-30/0.5-3.5/3-5, and 30-50/0.5-1/5-7, respectively. Thirty two and sixteen percentages of the isolates were FG and ESG, respectively. Most of the isolates (87%) could grow on culture medium of pH 4.5, but were intolerant of pH 8. The intrinsic antibiotics resistance (IAR) pattern was FG>SG>ESG for ampicillin, ciprofloxacin, chloramphenicol, erythromycin, neomycin sulfate and penicillin, whereas the pattern was FG<SG<ESG for gentamycin sulfate salt and nalidixic acid. The C-source utilization pattern was FG>SG>ESG for dextrin, dextrose, glucose, starch and sucrose whereas it was FG<SG<ESG for arabinose and galactose. Only FG isolates grew on culture medium containing methionine as sole N source. Cluster analysis of the isolates based on their phenotypic and genotypic characteristics matched with their growth categories in which 65% of the isolates that had mean generation time of >7.5 h were grouped together at 70% of similarity. Partial sequence analysis of 16S rRNA gene showed the existence of isolates most similar to rhizobial species of Bradyrhizobium species, Bradyrhizobium japonicum, Bradyrhizobium elkanii, Rhizobium rubi, and Mesorhizobium species. In general, cowpea rhizobial isolates from soils of Ethiopia in this study were mainly SG and sensitive to stress in vitro conditions, but versatile in utilization of varieties of C and N substrates. Such studies are important in Ethiopia to identify rhizobial isolates that could be amendable for use as inoculants to improve cowpea production.


INTRODUCTION
Cowpea is a food and forage tropical legume that develops effective symbiosis with atmospheric nitrogen fixing root nodule bacteria (Dakora and Keya, 2000).Previously, rhizobia nodulating tropical legumes were characterized as slow-growing and alkaline producing under in vitro conditions and were generally grouped under Bradyrhizobium species.Later, the species was classified into soybean rhizobia (Bradyrhizobium japonicum) and "cowpea miscellany" rhizobia (tropical bradyrhizobia) (Jordan, 1982).Cowpea miscellany rhizobia cross-nodulate Vigna unguiculata, Arachis hypogaea, Macroptilium atropureum and Phaseolus lunatus (Allen and Allen, 1981), although their degree of symbiotic effectiveness varies among the legumes (Thies et al., 1995).
The general categorization of cowpea rhizobia as the slow-growing tropical bradyrhizobia has led to misconception about their diversity because reports are showing the existence of both slow-growing and fastgrowing rhizobial isolates in the root nodule of V. unguiculata (Mpepereki et al., 1996).However, the proportion of fast-growing rhizobia in the root nodule of cowpea is lower than the slow-growing species (Martins et al., 1997;Zhang et al., 2007).The fast-growing cowpea rhizobia from different countries were identified under the genera of Rhizobium, Sinorhizobium, and Mesorhizobium (Zhang et al., 2008;Leite et al., 2009;Balasubramani et al., 2014).
Cowpea rhizobia show distinct variations in their cultural and physiological characteristics, which have been used for identification and diversity studies.Fastgrowing and slow-growing isolates vary in their growth and colony characteristics (Jordan, 1982), tolerance to temperature, antibiotics, pH and salts (Zablotowicz and Focht, 1981;Florentino et al., 2010;Abdelnaby et al., 2015).Fast-growing cowpea rhizobial isolates are more tolerant to in vitro stresses than the slow-growing isolates (Leite et al., 2009;Abdelnaby et al., 2015).Reports about versatility in utilization of C and N substrates by cowpea rhizobia are inconsistent (Hadad and Loynachan, 1985;Abdelnaby et al., 2015).
Recent genetic studies based on 16S rRNA gene have revealed the true phylogenetic association among rhizobial isolates.Particularly, analysis of more than one gene (housekeeping and accessory genes) using different molecular tools is revealing high diversity among cowpea rhizobial isolates (Zhang et al., 2007(Zhang et al., , 2008;;Guimarães et al., 2012;Silva et al., 2012).In general, cowpea rhizobial diversity study is more plausible when phenotypic and genotypic characteristics are combined in polyphasic pairing (Zhang et al., 2008).In Ethiopia, cowpea rhizobial diversity is yet to be studied.Therefore, this study is to broadly investigate the phenotypic and genotypic characteristics of cowpea rhizobia indigenous to soils of Ethiopia.

Rhizobial isolation
A total of 87 soil samples were collected from cowpea growing areas of Ethiopia (Figure 1) to obtain rhizobial isolates.Cowpea seeds of variety-"Bole" (from Melkasa Agricultural Research Center, Ethiopia) were planted in the soil samples in a pot culture under greenhouse conditions.After 45 days of planting, the root nodules were collected separately, surface sterilized and root nodule bacteria were isolated on yeast extract mannitol agar medium (YEMA) containing Congo red dye using standard methods (Somasegaran and Hoben, 1994).Individual colonies were periodically sub-cultured in YEMA medium to confirm purity and uniformity of colonies and the pure-cultures were preserved at 4°C.

Experimental conditions for laboratory tests
All experiments were set under the following conditions unless stated otherwise.Culture medium was adjusted to pH 6.8 before sterilization at 121°C for 15 min.The experiments were conducted on YEMA medium that contained the following ingredients in liters of distilled water: 10 g of Mannitol; 0.5 g of Yeast extract; 0.5 g of K2HPO4; 0.2 g of MgSO4.7H2O;0.025 g of Congo red; 0.1 g of NaCl, 15 g of agar; and pH of 7 (Somasegaran and Hoben, 1994).A loopful of active broth culture was inoculated and incubated at 28±2°C for 4 to 11 days.The tests were done in triplicates and presence or absence of growth of isolates on the culture medium was recorded as "+" or "-", respectively.The experiments included controls of their respective growth conditions.

Presumptive tests of isolates
The isolates were checked for growth on peptone-glucose agar medium that contained the following ingredients in liters of distilled water: 10 g of peptone; 5 g of glucose; 15 g of agar; 0.1 g of bromocresol purple (Somasegaran and Hoben, 1994).The isolates were also seeded on yeast extract lactose agar medium that contained the following ingredients in liters of distilled water: 10 g of lactose; 1 g of yeast extract; 20 g of agar.After appearance of colony, cultures of yeast extract lactose agar were flooded with Benedict's reagent to distinguish Agrobacterium (a yellow ring forming colonies) from Rhizobium (Schaad, 1980).

Authentication of isolates
The isolates were checked for renodulation of the host legume on *Corresponding author.E-mail: girmayek@gmail.com.
Author(s) agree that this article remain permanently open access under the terms of the Creative Commons Attribution License 4.0 International License sterile potted sand culture according to standard methods (Vincent, 1970).Surface sterilized cowpea seeds ("Bole") were germinated on water agar and planted in the potted sand.The sufficiently grown rhizobial isolates in broth culture medium were inoculated into each seedling (1 ml per seedling) after the appearance of secondary leaf.Pots containing uninoculated seeds were also included as controls.All pots were fertilized with sterile quarter strength of Broughton and Dilworth (1971) N-free medium twice per week and the plants were checked for nodulation after 45 days of planting.The bacterial isolates that renodulated the host plants were designated as Ethiopian Cowpea Rhizobia (ECR) with different numbers.

Cultural characteristics of rhizobial isolates
Colony characteristics of the rhizobial isolates such as color, size in diameter, shape, and texture were determined according to Sinclair and Eaglesham (1984).The isolates were checked for acid/alkaline production by growing them on YEMB medium containing 25 μgml -1 of bromothymol blue (BTB) (Jordan, 1982).Growth rate of the isolates was determined spectrophotometrically according to White (1995).

Stress tolerance tests of the rhizobial isolates
Fresh cultures of the isolates were incubated at a temperature of 32, 35, 37, 40, and 45°C for 8 days to determine temperature tolerance of the isolates.Performance of the isolates at different pH was tested on Keyser liquid medium adjusted to pH of 4, 4.5, 5, 8, and 9.Moreover, tolerance of the isolates to metal toxicity was tested on sterile Al (50 µM) and Mn (200 µM) separately and in combination on Keyser liquid medium adjusted to pH 5 (Keyser and Munns, 1979).Stock solution of the metals were filter sterilized using 0.22 μm pore size disk-syringe filter and mixed with sterilized agar medium (cooled to 47°C) on rotary shaking water bath as indicated in Ayanaba et al. (1983).

DNA extraction
Genomic DNA of the bacterial isolates was extracted by DNAzol kit (Genomic DNA isolation reagent, Molecular Research center, Inc., State, USA) according to the manufacturer"s protocol.One hundred microliters of the bacterial broth culture from its late exponential phase was mixed with 1 ml of DNAzol reagent.After repeated pipetting for mixing, the homogenates were centrifuged at 10,000g for 10 min at room temperature.The resulting viscous supernatant was transferred into a fresh tube and the DNA was precipitated by addition of 0.5 ml of 100% ethanol.After centrifuging at 10,000 rpm for 10 min, the supernatant was removed and the precipitated DNA was washed with 1 ml of 75% ethanol at least two times by gentle pipetting and centrifuging.The remaining ethanol was removed from the bottom of the tube with pipette and the DNA pellet was solubilized with 0.25 ml of 8 mM NaOH.Finally, quality and quantity of the DNA extracts were determined by Nanodrop (Thermo Scientific, Nanodrop 2000, Wilmington, USA).

BOX polymerase chain reaction (BOX-PCR) profiles of isolates
The BOX-PCR profiles of the isolates were analyzed using the BOX-A1R primer (5"-CTA CGG CAA GGC GAC GCT GAC G-3") to evaluate the genetic similarity among them.The PCR mix contained 2 µl template, 12.5 µl GoTaq Hot start colorless master mix, 2 µl of 5 mM primer, with final volume was 25 µl.All of the PCR were conducted with a BIO-RAD MyCycler thermocycler (BIO-RAD Laboratories, Hercules, CA, USA) with the following conditions: initial denaturing at 94°C for 4 min; 29 cycles of 94°C for 3 s, 92°C for 45 s, 50°C for 1 min and 65°C for 6 min (Rademarker et al., 2004).The electrophoresis of the BOX-PCR product was performed on 2% high resolution agarose gel at 4°C with 70 V for 13 h.The 1 kb plus DNA ladder (Promega, City, USA) was used as a molecular marker in each gel.The gel was stained with ethidium bromide for 30 min and destained with distilled water thereafter.Finally, the gel was viewed under UV light and photographed with CCD camera (Gel Logic 100 imaging system, Kodak, Rochester, NY, USA).

16S rRNA gene amplification and sequencing
The 16S rRNA gene of the isolates was amplified using a universal bacterial primer pair 968F (5′-AAC GCG AAG AAC CTT AC-3′) and 1401R (5′-CGG TGT GTA CAA GAC CC-3′) (Felske et al., 1996).Final volume of the PCR mix was 25 µl (2 µl template, 12.5 µl GoTaq Hot start master mix, 2 µl of each primer at 5 µM and 6.5 µl PCR quality water).The PCR condition was initial denaturation of 2 min at 94°C followed by 30 cycles of 1 min of 94°C, 1 min of 58.5°C, and 2 min of 72°C.The final extension was for 10 min at 72°C.All PCR were conducted with a BIO-RAD MyCycler thermocycler (BIO-RAD Laboratories, Hercules, CA, USA).The PCR amplicons (433 bp) were purified with a PCR purification kit (Wizard PCR Preps DNA Purification System, Promega, Madison, WI, USA) and were sequenced with the forward primer 968F at the Georgia Genomics Facility (http://dna.uga.edu/).The sequence data were compared against the existing sequences in the GenBank database of the National Center for Biotechnology Information (NCBI) (http://blast.ncbi.nlm.nih.gov/) to identify taxonomic groups that were most similar to the isolates.The sequences were deposited in the GenBank and set of accession numbers in the range of KY368577-KY368594 except accession number KY378912 for the isolate ECR-78.

Cluster analysis of the isolates
Physiological characteristics such as tolerance to temperature, pH, salts, metal toxicity, IAR, pesticides, and substrate utilization were used to cluster the rhizobial isolates.Cluster analysis of the BOX-PCR profiles (banding patterns) of the isolates was also done to compare their genetic similarities.The results were changed into binary matrices where the digits 1/0 represent the presence/absence of a phenotypic characters or DNA bands.The similarity matrix was generated by Euclidean distances, which were used to build a tree with the unweighted pair group mean averages (UPGMA) algorithm.Analysis of data was performed using the software PAleontological Statistics (PAST) version 2.17c (Hammer et al., 2001).
The 16S rRNA gene partial sequence of cowpea rhizobia representative from the phenotypic and genotypic grouping were used for the construction of phylogenic tree.Multiple alignment of the nucleotide sequences were conducted by ClustalW and were trimmed for gaps.Tree was constructed from subsection of the alignment (337 bp existing at the center) by neighbor-joining method in Kimura-2 (Kimura, 1980) using software program MEGA Version-6 (Tamura et al., 2007).Stability of the groupings was estimated by bootstrap confidence analysis with 1,000 repetitions and the cut-off value for condensed tree was 50%.Some root nodule bacterial species were included in the tree construction from the National Center for Biotechnology Information (https://www.ncbi.nlm.nih.gov/).

Rhizobial isolation
Eighty seven soil samples were collected from cowpea growing areas of the country.Eighty soil samples (92%) had isolates that induced nodules on the cowpea, indicating the wide distribution of cowpea rhizobia in soils of Ethiopia.The fact that most of the soil samples contained cowpea nodulating rhizobia suggests that the sampling areas had history of cowpea cultivation.According to Woomer and Asano (1990), cropping history of the land has impact on abundance of root nodule bacteria that are related to specific host legume.
In some nodules more than one kind of bacterial isolates were identified.This is in agreement with Abdelnaby et al. (2015) in which multiple occupants of different rhizobial isolates were recovered from root nodule of cowpea in Egypt.Totally, 77 isolates were collected from pinkish/red color root nodules.During the presumptive tests, 61 isolates (80%) were gram negative, rod shaped, grew poorly on peptone glucose medium, and negative for 3-ketolactose test, which are the distinctive characteristics of root nodule rhizobial isolates.These characteristics separate them from other endophytes like Agrobacterium species (Murugesan et  al., 2010; Kaur et al., 2012).Only 51 isolates (83% of the total) were able to reinfect the host plant as indicated by confirmatory test for nitrogen fixing rhizobia.This shows that only 66% of the isolates that were initially obtained from the nodules were true cowpea rhizobial isolates.This is higher as compared to the 46% of nodule inducing cowpea rhizobia reported from soils of Brazil (Marra et al., 2012).
It is interesting to note that about 34% of isolates failed to induce nodule on the host legume and considered as cowpea root nodule endophytes.Several previous reports also showed the occurrence of non-rhizobial bacteria in the root nodule of cowpea (Sarr et al., 2010;Abdelnaby et al., 2015).Finally, a total of 40 indigenous cowpea rhizobial isolates that are representative of the different geographical locations of the country were selected for further study.Majority of them were from the northern (43%) and central parts (33%) of the country as these regions are the major cowpea producing areas (FAO, 1996;Etana et al., 2013).

Growth properties of the rhizobial isolates
Cowpea rhizobial isolates showed variation in their cultural characteristics.They were categorized into three major groups based on color and texture of the colonies: white-dry (WD), white-wet (WW), and yellow-dry (YD).About 90% of the isolates displayed white colony, of which 60% was dry in colony texture (Table 1).Chagas et al. (2013) also reported about 93% of cowpea rhizobial isolates with similar colony texture.On the other hand, the proportion of white to yellow colonies of cowpea rhizobia was equal to those from Brazil (Leite et al., 2009;Silva et al., 2012).Although most of the rhizobial isolates in this experiment were dry in colony texture, one-third of the isolates were wet.This is similar to the cultural characteristics of cowpea rhizobia from sub-Saharan Africa where few of the isolates produced large gum colonies (Abdelnaby et al., 2015) but in contrast to Brazilian cowpea rhizobia in which more than 90% of the isolates displayed mucoid colony texture (Silva et al., 2012;Chagas et al., 2013).Cowpea rhizobia from soils of Ethiopia showed wide range of cultural characteristics compared to the previous reports from different countries.This could be associated with the differences in edaphoclimatic conditions among the sampling areas.
The isolates also showed variation in colony shape-border which was recorded as either circular-smooth (CS), circular-rough (CR), or irregular-smooth (IS).Sixty five percent of the isolates were CS (Table 1).Similarly, more than 55% of cowpea rhizobial isolates from Brazil was CS (Silva et al., 2012).The colony characteristics of most of the present isolates (85%) were similar to the genus Bradyrhizobium that are characterized as circular, opaque, rarely translucent, white, convex, and granular in colony properties (Jordan, 1982;Hadad and Loynach, 1985).Moreover, more than 60% of the isolates formed visible colony (VC) after five days of incubation.This is a characteristic of the genus Bradyrhizobium as indicated in Jordan (1982).Twenty percent of the isolates formed VC within three days of incubation.Isolates that formed colonies within five days of incubation also formed TB in liquid medium within two to four days of incubation (Table 1).
In this study, only ECR-95 changed color of BTB-YEMA medium into yellow, which is the characteristic of fast-growing rhizobial isolates.Unlike of the present study, previous reports have shown a clear distinction between fast-growth/acid production and slow-growth/alkaline production in cowpea rhizobia.For instance, all cowpea rhizobial isolates from Brazil that formed colonies after the 4th day of incubation increased pH of culture medium (Zilli et al., 2004).This is a typical characteristic of slow-growing isolates (Norris, 1959).In the same study, 62% cowpea rhizobial isolates that showed visible colonies within three days of incubation acidified the culture medium (Leite et al., 2009), showing higher abundance in fast-growing than the slow-growing isolates.Although more than half of the present rhizobial isolates were slow-growing (MGT>7.5 h) (Table 1), they did not increase pH of the culture medium.The correlation of fast-growth to acid production and slowgrowth to alkaline production was previously questioned by Hernandez and Focht (1984) that partly invalidated the concept.
MGT and CD of Ethiopian cowpea rhizobial isolates were in the ranges of 3 to 50 h and 0.5 to 4 mm, respectively.The MGT range was out of the range that has been used for the traditional taxonomy of rhizobia as fast-growing and slow-growing isolates (Jordan, 1982).MGT of cowpea rhizobia collected from different countries were in the ranges of 1.4 to 44.1 h (Hernandez and Focht, 1984).Although there is no distinct boundary in MGT for slow-growing and fast-growing cowpea rhizobial isolates (Hernandez' and Focht, 1984), some authors recommend the doubling time in the range of 3.4 to 8.3 and 14.3 to 33.3 h for fast-growing and slowgrowing rhizobial isolates, respectively (Martinez-de and Arias, 1977; Kennedy and Greenwood, 1982).
Based on their MGT, CD, VC and TB, cowpea rhizobial isolates from soils of Ethiopia were grouped into three major growth categories: fast (FG), slow (SG), and extraslow-growings (ESG).The MGT/CD/VC/TB of FG, SG, ESG isolates were in the range 2.5 to 7.5 h/2 to 4 mm/2 to 5 days/2 to 3 days, 7.5 to 30 h/0.5 to 3.5 mm/3 to 6 days/3 to 5 days, and 30 to 50 h/0.5 to 1 mm/5 to 8 days/5 to 7 days, respectively (Table 1).Similarly, Leite et al. (2009) classified cowpea rhizobia from soils of Brazil into three groups based on pH change of the culture medium, colony size, and colony color.Although, the categories overlapped, Hernandez and Focht (1984) also grouped cowpea rhizobia into FG and SG based on MGT boundary of 7.8 h.According to the present growth categories, 32 and 16% of the Ethiopian cowpea rhizobial isolates were FG and ESG, respectively (Table 1).

Stress tolerance of the rhizobial isolates
Greater than 80% of the isolates were resistant to the test dose of Nal, Neo and Pen.Up to 75% of the isolates was less sensitive to Amp, Chl, and Gent, and 50% of the isolates were sensitive to Cipr, Ery and Str.IAR of the isolates showed trend in relation to their growth categories except for streptomycin in contrast to Xavier et al. (1998) that reported the absence of correlation between cowpea rhizobial growth properties and their IAR.The IAR pattern was FG>SG>ESG for Amp, Cipr, Chl, Ery, Neo, and Pen whereas it was FG<SG<ESG for Gent and Nal (Figure 2).The IAR pattern has been used for identification and diversity studies in cowpea rhizobia (Zilli et al., 2004).In this study, greater than 50% of the isolates were resistant to six of the nine antibiotics tested.
All the isolates were very sensitive to the test dose of the fungicides, except ECR-14, ECR-68, and ECR-101 that were resistant to curzate.Similarly, the isolates were sensitive to the test dose of glyphosate, except for ECR-10 and ECR-69.Although, pesticide resistance of the isolates did not correlate with their growth properties, Sawicka and Selwet (2008) reported the sensitivity of SG rhizobia compared to FG rhizobia.Therefore, sensitivity of the present isolates to the pesticides could be associated to their growth properties as majority of them (68%) were SG (Table 1).
Majority of the isolates (75%) were tolerant to pH 4.5, but sensitive to pH of 8 except for ECR-95 (Table 2).Previous studies have also shown sensitivity of cowpea rhizobia to alkaline medium compared to acidic medium (Keyser et al., 1979;Zablotowicz and Focht, 1981).
Particularly, cowpea rhizobia were tolerant to strong acidity of pH 4.5 (Zablotowicz and Focht, 1981) but intolerant to pH 10 (Hadad and Loynachan, 1985).This is in agreement with the present findings in that isolates were intolerant to grow in medium of pH 4 and pH greater than 8.The tolerance of ECR-95 to alkaline medium (pH 8) could be associated with its acid production during its growth that can bring down pH of the medium to neutral.FG isolates were resistant to acidic environment compared to the other growth categories (Table 2).However, the relationship between MGT and their pH tolerance was irregular (data not shown).In general, the present isolates had narrow range of pH tolerance (pH 4.5 to 8) compared to previous studies on cowpea rhizobia that were able to grow in pH range of 4 to 10 (Zablotowicz and Focht, 1981;Florentino et al., 2010;Abdelnaby et al., 2015).This might be associated with soil pH of the sampling areas as isolates from stressed soil conditions are likely to be tolerant to similar in vitro conditions.
Resistance of the isolates to toxicity of Al and Mn under acidic in vitro condition varied in that 200 µM of Mn did not affect their growth whereas only 55 and 15% of the isolates were able to grow with 50 µM Al and on the combination of the two metals under acidic conditions, respectively (Table 2).Previous studies also showed toxicity of metals on cowpea rhizobia under acidic growth condition (Paudyal et al., 2007;Keyser and Munns, 1979).
Incubation temperature of greater than 28°C affected growth of the isolates under in vitro conditions.About 75% of the isolates were tolerant to temperatures of 35°C, but sensitive to temperature greater than 40°C except for isolates ECR-71, ECR-86, and ECR-99.The general sensitivity of cowpea rhizobial isolates to temperature higher than the optimum growth temperature (28±2°C) was previously reported (Florentino et al., 2010;Abdelnaby et al., 2015).In this study, tolerance of isolates to high temperature did not show pattern in relation to their growth categories (Table 2).
The isolates were sensitive to salt in the form of NaCl at concentration greater than 0.25% (Table 2).This is similar to the findings of Mensah et al. (2006).However, studies from Brazil (Florentino et al., 2010) and Libya (Abdelnaby et al., 2015) showed tolerance of cowpea rhizobia up to 3% (w/v) of NaCl in growth medium.At a concentration of 0.5% NaCl, the tolerance pattern of the isolates was FG>SG>ESG (Table 2).

Substrate utilization tests of the rhizobial isolates
The isolates utilized fructose, mannose, sorbitoal, maltose, gluconic acid, xylose and glycerol as carbon sources but only few isolates (<20%) grew on inulin, dextrose, inositol, and glucose (Table 3).Cowpea rhizobia from sub-Saharan Africa are poor in utilizing monosaccharide and sugar alcohols (D-glucose, galactose, sorbitol and inositol) but are efficiently in assimilating disaccharides (sucrose and maltose) (Abdelnaby et al., 2015).Nonetheless, SG isolates were reported to be poor in utilizing monosaccharides and disaccharides (Stowers and Elkan, 1984;Eaglesham et al., 1987).Similarly, only less than 30% of the present isolates from the categories of SG and ESG were able to grow on medium containing the disaccharides lactose and sucrose as sole carbon sources (Table 3).
Utilization of some of the carbon sources showed pattern in relation to their growth categories.The pattern in utilization of dextrin, dextrose, glucose, starch and sucrose can be put as FG>SG>ESG whereas for arabinoose and galactose the utilization pattern was ESG>SG>FG.The carbon sources such as CMC, dextrin, dextrose, glucose, inositol and starch were not utilized by the ESG isolates (Table 3).Polysaccharides have been considered as poor sources of carbon for growth of rhizobia especially for the genus Bradyrhizobium (Jordan, 1982).Cellulose is particularly unavailable to SG cowpea rhizobial isolates (Abdelnaby et al., 2015).This is similar to the present study in which isolates from ESG were unable to grow on the polysaccharides (CMC and Starch).
In this study, there were no single carbon sources, which can distinctively be used to identify FG, SG or ESG isolates.However, the FG isolates utilized more than 69% of the tested carbon sources.In addition, some isolates were able to use a wide range of carbon sources as compared to the others.For instance, ECR-0 utilized more than 70% of the total carbon sources tested in Only substrates that showed growth difference among the isolates were indicated in Table 1.
this study (data not shown).This could be indicative of the potential ability of the isolate to be competitive in the rhizosphere because metabolic diversity is presumed to be a survival strategy of microorganisms.
With regard to nitrogen source utilization, more than 80% of the isolates utilized the amino acids Leucine, L-arginine, threonine and tyrosine as sole nitrogen sources in contrast to only 7.5% of the isolates (all FG) that were able to grow on DLmethionine (Table 3).The FG isolates were efficient in utilization of methionine, lysine, and glutamic acid whereas the SG isolates grew on medium containing phenylalanine as sole nitrogen sources.It is interesting to note that the ESG isolates were poor in utilization of N-sources because they failed to utilize more than 50% of the amino acids tested (Table 3).The versatility in utilization of different N-sources by FG cowpea miscellany rhizobial isolates (particularly on methionine) compared to the SG isolates was previously reported (El-akhal et al., 2009).

Cluster analysis of the rhizobial isolates
In this study, phenotypic characteristics were used to construct dendrogram by unweighted pair group method with arithmetic mean (UPMGA).At about 50% of similarity, the rhizobial isolates were categorized into two major phenotypic groups.The larger group (P1) contained 65% of the isolates, which have MGT greater than 7.5 h, and the other group (P2) consisted of isolates with MGT less than 7.5 h.As percentage of similarity increases, diversity among the isolates also increased and six sub-clusters were formed at 70% of similarity, of which five sub-clusters were under the group P1.Previous reports also showed cluster of cowpea rhizobia into six phenotypic groups at 70% of similarity using similar phenotypic characteristics (Abdelnaby et al., 2015).Isolates from P2 did not show differences at 70% similarity.This shows higher relatedness of FG isolates as compared to the other categories (Figure 3).
The phenotypic characteristics showed differences among isolates except for ECR-3/ECR-4, ECR-55/ECR-64, ECR-68/ECR-89, and ECR-6/ECR-9/ECR-92 that showed 100% similarity to each other (Figure 3).These isolates were obtained from different geographical areas with different edaphoclimatic conditions.Therefore, the cluster did not reflect the geographical origin of the isolates.This is different from Steenkamp et al. (2008) that showed the cluster of cowpea rhizobial isolates from different African countries based on their origin of isolation.Obviously, phenotypic grouping of the present isolates was merely dependent on their growth properties in culture medium, particularly based on their growth rate expressed as MGT.Zhang et al. (2007) also reported the cluster of cowpea rhizobia according to both their MGT and the type of host variety used.
The number and size of bands of the BOX-PCR profiles of the isolates ranged from 1 to 11 and 600 to 10,000 bp, respectively (Figure 4).The BOX-PCR profiles of the present isolates are different from those that were reported by Menna et al. (2009) in which rhizobia isolates from different legumes including cowpea had a mean of 20 bands, with sizes that ranged between 200 and 5000 bp.The species that were most similar to Bradyrhizobium species and B. elkanii showed high rates of band polymorphism whereas isolates that were most similar to the species of B. japonicum and R. rubi were monomorphic but varied in band sizes (Figure 4 and Table 4).
Cluster analysis of the isolates based on their BOX-PCR profiles is as shown in  Table 4. Sequence similarities of the rhizobial isolates with known bacterial strains based on their partial 16S rDNA gene sequence.

Isolate
Bacterial species Identity (%) E-value ECR-0, 68, 89, 101 Bradyrhizobium species 99 0.0 4,14,27,62,65,94,97,  The low genetic diversity among isolates in this study could be due to the similar land use systems of the sampling areas (agricultural lands).Previous studies had shown the effects of land use system on "cowpea miscellany" rhizobial diversity (Mwenda et al., 2011).For instance, cultivation of legumes on a certain land has a selective pressure on some of the rhizobial species.Zilli et al. (2004) reported the highest cowpea diversity from lands where neither soybean nor cowpea had ever been planted.Therefore, cropping history of the sampling site can create a selective advantage for some cowpea rhizobial isolates.Besides, the current isolates were obtained using only one cowpea variety ("Bole") although it is advisable to obtain isolates using multiple hosts for diversity study (Zhang et al., 2007).In general, the cluster analysis of cowpea rhizobial isolates based on their phenotypic and genetic (BOX-PCR profiles) characteristics showed similar trends, which is that isolates were clustered mainly based on their growth rate that reflected their physiological traits.This is similar to what was reported by Zhang et al. (2007) in which cowpea rhizobia clustered according to their growth rate.Partial sequence analysis of the 16S rRNA gene of the isolates representative of the phenotypic and BOX-PCR clusters showed the existence of different species of rhizobia in the root nodule of cowpea.The largest genotypic sub-cluster (GSC-IV) that included 25% of the total isolates (ECR-3, 62, 65, 94, 97, and 100) by BOX-PCR was most similar to the species of B. japonicum (Table 4).ECR-68 that represented GSC-III was most similar to Bradyrhizobium spp.Although the GSC-I and II were grouped separately based on BOX-PCR analysis (Figure 5), ECR-31 from I and ECR-50 from II were most similar to B. elkanii (Table 4).The difference between the two subgroups could possibly be discerned by targeting a larger size amplicon other housekeeping and nitrogen fixing genes that are used for identification of closely related cowpea rhizobial species (Krasova-Wade et al., 2003;Zhang et al., 2008).GSC 6-8 were grouped together under the PSC-VI and representative isolate(ECR-78) was most similar to Mesorhizobium species (Table 4).
Partial nucleotide sequence of 16S rRNA gene showed the existence of three cowpea nodule inducing genera: Bradyrhizobium, Rhizobium and Mesorhizobium.The minimum similarity percentage was 98 except for ECR-78 (92%).About 80% of the isolates were most similar to the genus Bradyrhizobium, with ≥99% of similarity, of which 60% of them were most similar to B. japonicum (Table 4).The prevalence of genus Bradyrhizobium in the root nodule of cowpea was previously reported in studies from different countries, including 90% from Chinese soil (Zhang et al., 2008) and 81% from Brazilian soil (Guimarães et al., 2012) and 100% from Senegal soil (Krasova-Wade et al., 2003).
The existence of FG isolates of the genera Rhizobium and Mesorhizobium among the present isolates supports previous reports of cowpea nodulation by FG rhizobia of the genera Rhizobium, Sinorhizobium, and Mesorhizobium (Zhang et al., 2007(Zhang et al., , 2008;;Leite et al., 2009;Balasubramani et al., 2014).This shows the nonspecific nodulation of cowpea by rhizobial species.
The phylogenetic tree of the 16S rRNA gene sequence showed that the isolates were grouped into two alongside the reference species of the genera Bradyrhizobium, Mesorhizobium, Rhizobium, and Sinorhizobium (Figure 6).The isolates were further sub-clustered into four groups together with their most similar species at bootstrap value of greater than >70% (Figure 6).In general, cluster analysis of cowpea rhizobial isolates based on their physiological and genetic characteristics resulted in groupings that reflect their cultural growth properties.

Conclusions
The study showed that cowpea rhizobia are abundant in soils of Ethiopia.They are diverse in phenotypic and genotypic properties.Large proportion of the isolates was SG, but FG isolates were also present.Therefore, cowpea rhizobial isolates indigenous to soils of Ethiopia include fast, slow and extraslow-growing varieties.However, the isolates were neither acid nor alkaline producers under in vitro conditions.This is contrary to the common practice of using acid/alkalkine production as a way of distinguishing rhizobial isolates based on their growth categories.
Colonies of the isolates were mostly white in color, dry in texture, and circular in shape.The stress tolerance and substrates utilization properties of the isolates under in vitro conditions correlated with their growth categories.FG isolates had wider substrate utilization potential than slow growing isolates.The combined effects of acidity, Mn, and Al strongly inhibited growth of the non-fast growing isolates under in vitro conditions.
Based on phenotypic and genotypic characteristics, the isolates were clustered according to their cultural characteristics at a boundary of MGT 7.5 h.Representative isolates from the clusters were identified as Bradyrhizobium species, Bradyrhizobium japonicum, Bradyrhizobium elkanii, Rhizobium rubi, and Mesorhizobium species using partial nucleotide sequences of 16S rRNA gene.Such studies are important in Ethiopia to identify rhizobial isolates that could be amendable for use as inoculants to improve cowpea production.

Figure 1 .
Figure 1.Map of Ethiopia showing soil sampling sites, which are highlighted.

Figure 2 .
Figure 2. The IAR of cowpea rhizobial isolates in relation to their growth categories.

Figure 3 .
Figure 3. UPGMA dendrogram showing the phenotypic diversity of Vigna unguiculata rhizobia isolates from cowpea growing areas of Ethiopia.The dendrogram was constructed based on binary matrices of 54 physiologic characteristics of the isolates.

Figure 4 .
Figure 4. BOX-PCR profiles of cowpea rhizobial isolates from soils of Ethiopia.The numbers on the lanes are the isolate designations without the prefix ECR.The L lanes are DNA ladder that has bands of known sizes in base pairs (bp).

Figure 5 .
Figure 5. BOX-PCR profiles of the rhizobial isolates that nodulated cowpea from soils of Ethiopia.Similarity matrix was generated by Euclidean distances and a tree was built with the unweighted pair group mean averages (UPGMA) algorithm.

Figure 6 .
Figure 6.Neighbor-joining tree based on 16S rRNA gene partial sequence showing the relationships of cowpea rhizobia isolates from the lowland agroecology of Ethiopia.Bootstrap values are based on 1000 replications.Cut-off value for condensed tree was 50%.
CS: Circular colony shape with smooth border; IS: irregular colony shape with smooth border; CR: circular colony shape with rough border; YD: yellow colony color with dry texture; WD: white colony with dry texture; WW: white colony with wet texture; CD: colony diameter; VC: days to form visible colony; TB: days to form turbidity; MGT: mean generation time.

Table 2 .
The percentage rhizobial isolates that grew under different stresses under in vitro conditions.

Table 3 .
Percentage rhizobial isolates that grew on different carbon and nitrogen sources under in vitro conditions.