http://wwwsoc.nii.ac.jp/jsme2/ doi:10.1264/jsme2.ME08558

Microbes Environ. Vol. 24, No. 2, 105–112, 2009

Phylogenetic Diversity and Symbiotic Effectiveness of Root-Nodulating

Bacteria Associated with Cowpea in the South-West Area of Japan

PAPA S. SARR1*, TAKEO YAMAKAWA2, SYUNSEI FUJIMOTO

3, YUICHI SAEKI4, HOANG T.B. THAO1, and AUNG K. MYINT

1

1Laboratory of Plant Nutrition, Division of Bioresource and Bioenvironmental Sciences, Graduate School, Kyushu

University, 6-10-1 Hakozaki, Higashi-ku, Fukuoka 812-8581, Japan; 2Laboratory of Plant Nutrition, Division of Soil

Science and Plant production, Department of Plant Resources, Faculty of Agriculture, Kyushu University, 6-10-1

Hakozaki, Higashi-ku, Fukuoka 812-8581, Japan; 3Laboratory of Plant Nutrition, School of Agriculture, Kyushu

University, 6-10-1 Hakozaki, Higashi-ku, Fukuoka 812-8581, Japan; and 4Department of biochemistry and Applied

Biosciences, Faculty of Agriculture, Miyazaki University, Miyazaki, 889-2192, Japan

(Received November 6, 2008—Accepted February 11, 2009—Published online March 10, 2009)

The phylogenetic diversity of cowpea root-nodulating bacteria in the South-West of Japan was investigated using 60isolates. Seeds of cowpea were aseptically sown in vermiculite and inoculated with a suspension of Cowpea Soil (CS)or Bean Soil (BS) or without a soil suspension as a control. CS and BS were collected from the Kyushu University’sfarm (Japan) at sites where cowpea and bean, respectively, have been cultivated previously. Based on an analysisof the 16S rRNA gene and the Internal Transcribed Spacer (ITS) sequence between the 16S and 23S rRNA genes,56 isolates were assigned to the genus Bradyrhizobium, while one isolate was found to be closely related to the genusRalstonia. The ITS-based phylogeny showed 53 isolates, 2 isolates, and 1 isolate, to be closely related to B.yuanmingense, B. elkanii and B. japonicum, respectively, suggesting that B. yuanmingense strains predominated in thesoils. Among the isolates tested, B. yuanmingense TSC10 and TTC9 exhibited a greater symbiotic activity and couldbe considered efficient inoculants for cowpea.

Key words: phylogenetic diversity, Bradyrhizobium yuanmingense, cowpea root-nodulating bacteria, Ralstonia

Cowpea (Vigna unguiculata L. Walp.), a legume native to

Africa, is an important annual crop in tropical and sub-tropi-

cal regions worldwide, especially in Sub-Saharan Africa,

Asia, and Central and South America (8, 23). The young

leaves, pods, and seeds of this plant are good sources of

dietary protein, vitamins, and minerals for humans and ani-

mals (23). Dakora et al. (7) reported that in Ghana, the bene-

fit of nodulated cowpea to soil nitrogen supply was 60 kg N

ha−1 when residues from the crop were incorporated into the

soil. Therefore, in soils experiencing a decline in soil nitro-

gen status which is a threat to food production, biological

nitrogen fixed via rhizobia-legume symbiosis has been rec-

ommended for the sustenance of traditional agriculture.

However, the common approach to improving the symbiotic

fixation of nitrogen and legume productivity, using superior

or very effective exotic rhizobial strains as inoculants, often

fails to achieve the desired responses (4, 33). The failure has

been attributed to the poor competiveness of the introduced

rhizobia (28), the non-specificity of the bacteria, and the

occurrence and resistance to stress of ineffective indigenous

rhizobial strains in soils (9). To improve the yield and quality

of cowpea, the use of rhizobia, selected from the indigenous

community with the capacity for nodulation and strong nitro-

gen fixing activity, is an important agronomic approach (39).

Thies et al. (29) and Hunt et al. (10) reported that several

indigenous strains of Bradyrhizobium were superior to

commercial varieties for inoculating cowpea and soybean.

According to microbe-host specificity, rhizobia isolated

from cowpea are generally placed in the cowpea-cross-

inoculated group (1) and species of this heterogeneous group

were assigned, based on phylogenetics, to the genus

Bradyrhizobium (13). Created in 1982 (12), the genus

Bradyrhizobium now includes seven species: Bradyrhizobium

japonicum (12), Bradyrhizobium elkanii (17), Bradyrhizobium

liaoningense (36), Bradyrhizobium yuanmingense (37),

Bradyrhizobium betae (21), Bradyrhizobium canariense (32),

and Bradyrhizobium denitrificans (30). Despite this number,

however, studies on cowpea bradyrhizobia are limited. Based

on a polymerase chain reaction-restriction fragment length

polymorphism (PCR-RFLP) analysis of the16S-23S rRNA

intergenic spacer region (IGS), Krasova-Wade et al. (15)

divided isolates from three cowpea cultivars from Senegal,

West Africa into four genetic profiles; IGS type’s I, II, III

and IV. Results of a phylogenetic analysis showed that type I

exhibited low similarity with sequences in databases and

could represent a new species. Insufficient rainfall and the

low moisture content of semi-arid soils limit biological nitro-

gen fixation (26). According to Krasova-Wade et al. (15),

among the three cowpea cultivars, the most water sensitive

(B-21) was exclusively nodulated by the IGS type I strains,

while the Mouride cultivar showing good drought resistance,

harbored greater rhizobial diversity. In Japan, although envi-

ronmental conditions differ from those in traditional agro-

ecological zones of cowpea (tropical areas), this plant is

grown in some areas. However, until recently, there has been

little investigation of the diversity and phylogeny of its

rhizobia. The Kyushu area located in the South-West of Japan

and characterized by high levels of rainfall and different

soil properties compared to less irrigated tropical zones, may* Corresponding author. E-mail: sarr.papa@voila.fr; Tel: +81–92–

642–2847; Fax: +81–92–642–2848.

SARR et al.106

represent a valuable site for investigation’s of the diversity of

cowpea rhizobia in Japan. Therefore, this study was carried

out to (i) determine the phylogenetic diversity of cowpea-

nodulating bacteria in the South-West of Japan, (ii) select

highly effective rhizobia as inoculants for cowpea, and (iii)

investigate whether a relation between cultivar and rhizobia

could be expected in this zone. For this purpose, the 16S

rRNA gene and the Internal Transcribed Spacer (ITS) region

between the 16S and 23S rRNA genes were sequenced in

indigenous bacteria isolated from the roots of cowpea culti-

vars grown in pots with suspensions of soil samples collected

from the Kyushu University’s farm. A cross-inoculation test

was also performed to identify strains with higher nodulation

competitiveness and symbiotic effectiveness.

Materials and Methods

Bacterial strains and DNA isolation

A total of 60 indigenous bacterial strains were isolated fromthe root nodules of two drought-tolerant (Dan IIa and Tvu-11986)cowpea cultivars and one drought-sensitive (Tvu-7778) cultivar.These cultivars were provided by Dr. H. OMAE of JIRCASOkinawa Subtropical Station (Japan). The plants were grown inpots filled with vermiculite and inoculated with two suspensions ofsoil samples collected from the Kyushu University (33°39' W,130°21' E; Fukuoka, Japan) farm or without any soil suspension as acontrol. The two soils, named Cowpea Soil (CS) and Bean Soil(BS), were collected from sites where cowpea and bean have beencultivated previously, respectively. Selected chemical properties ofthe soils before the cultivation of cowpea and bean were are asfollow: pH(H2O) 7.22, total-N 0.048%, total-P 0.045%, total-K0.135%, CEC 19.41 cmolc kg−1. Plant cultivation and bacterial isola-tion were conducted mainly according to procedures described byVincent (31). The possibility of contamination with non-relevantcowpea-nodulating bacteria was excluded by confirmation of theabsence of nodules in negative control cultures without soil (3 pots).Twenty root nodules picked out from each cultivar (Table 1) weresurface sterilized in 70% ethanol, 5% hydrogen peroxide andhomogenized in sterile 0.9% NaCl, crushed and streaked on yeastmanitol agar (YMA) (31), Congo red (CR) or Bromothymol blue(BTB) plates. Following incubation at 30°C for 3 or 7 days, a singlecolony recognized on the medium was removed and conserved in aglycerol-stock culture at −84°C. Total DNA of strains cultured inAIE liquid medium (16) was later extracted, using an ISOPLANTkit (Nippon Gene, Tokyo, Japan) and following the instructions ofthe manufacturer.

Production of acid/alkaline substances on YMA-BTB and growth properties

The isolates were distinguished based on the duration of theirculture (3 or 7 days) and production of acid/alkaline substances onYMA media, and the generation times. On BTB plates, one isolate(TSC1) produced colonies (3–4 mm) after 3 days and the mediumturned yellow (production of acid substances). For the other iso-lates, colonies were visible at 5 to 7 days and the BTB mediumturned blue (production of alkaline substances). The acid-producingisolate TSC1 and representative alkaline-producing isolates, DTB1,DTB4, and DTC9, were selected for growth tests. Sinorhizobiumfredii USDA 194 and Bradyrhizobium japonicum USDA 110 wereused as reference strains for fast- and slow-growing rhizobia,respectively. One colony per isolate was grown on AIE medium at30°C with shaking at 100 rpm and optimum density (OD660nm) wasrecorded every 24 h for DTB1, DTB4, DTC9, and B. japonicumUSDA 110 or every 12 h for S. fredii USDA 194 and TSC1. Growthcurves were drawn based on the results of three independentexperiments. Generation time was calculated from the exponentialphase of the growth curve.

PCR of 16S rRNA and 16S-23S rRNA ITS regions

PCRs were performed as described by Saeki et al. (22). Theprimer sets 16S-F and 16S-R2, and ITS1512F and ITSLS23R, wereused to amplify the 16S and ITS rRNA genes, respectively. PCRproducts were then purified using the Wizard Gel and PCR Clean-up System (Promega, Madison, WI, USA), and the correspondingconcentrations were estimated after agarose gel electrophoresis(1.5% agarose gel in 1×TAE buffer) and staining with SYBR SafeDNA gel stain (Invitrogen, Carlsbad, CA, USA) (Fig. 1. A, B).

Sequencing and phylogenetic analysis

The purified PCR products of the 16S and ITS rRNA genes ofthe 57 isolates were ligated (TaKaRa DNA ligation ver. 2.1 kit) intoa plasmid T-A vector (constructed using the ampicilin resistancepGEM-5Zf(+), Promega) and cloned into competent cells (11)using standard methods. Three clones per isolate were selectedfor plasmid extraction using the Wizard Plus SV Minipreps DNAPurification System (Promega). The plasmid concentrationswere assessed by NIH image 1.62 (National Institutes of Health,Bethesda, MD, USA) after a 1% agarose gel electrophoresis withλHind III marker (Fig. 1C), and 20 μL of plasmid solution (100 ngμL−1) was used for sequencing (Macrogen, Seoul, Korea) with theprimers T7 promoter and SP6. DNA sequences were edited byDNASIS-Mac Ver. 2.0 (Hitachi, San Bruno, CA, USA) to createthe 16S rRNA gene or ITS sequence fragments. Bidirectionalsequences were aligned using GENETYX-MAC ver. 10.1 (Soft-ware Development, Tokyo, Japan) to obtain consensus sequences

Table 1. Numbering and growth characteristics of cowpea-nodulating bacteria

Host cultivar Soil Isolates Color (BTB) Colony size

Dan IIa

CSDTC1, DTC2, DTC3, DTC4, DTC5

Blue

1.5–2 mmDTC6, DTC7, DTC8, DTC9, DTC10

BSDTB1, DTB2, DTB3, DTB4, DTB5

(DTC9: 3–4 mm)DTB6, DTB7, DTB8, DTB9, DTB10

Tvu-7778

CSTTC1, TTC2, TTC3, TTC4, TTC5

Blue 1.5–2 mmTTC6, TTC7, TTC8, TTC9, TTC10

BSTTB1, TTB2, TTB3, TTB4, TTB5

TTB6, TTB7, TTB8, TTB9, TTB10

Tvu-11986

CSTSC1, TSC2, TSC3, TSC4, TSC5

Blue 1.5–2 mmTSC6, TSC7, TSC8, TSC9, TSC10

BSTSB1, TSB2, TSB3, TSB4, TSB5

TSC1 (yellow)3–4 mm

(TSC1, TTC5)TSB6, TSB7, TSB8, TSB9, TSB10

CS and BS are soil samples collected from sites where cowpea (Cowpea Soil: CS) and bean (Bean Soil: BS) were cultivated previ-ously. All isolates showed an entirely pulvinate shape except for DTC9 and TTC5 which had an undulating flat shape. Colony sizeswere obtained after 7 days incubation on YMA plates except for TSC1 (3 days).

Diversity of Cowpea-Nodulating Bacteria 107

for each clone. Sequences were compared with the DDBJ/EMBL/GenBank databases using the BLAST search program (2), and theclosely related sequences found were included in a phylogeneticanalysis using the unweighted pair group method with averages(UPGMA) (24). Multiple alignments of sequences and the calcula-tion of evolutionary distance were performed by the two-parametermethod of Kimura (14).

Cross-inoculation and symbiotic effectiveness

Dan IIa, Tvu-7778 and Tvu-11986 described above and Melakh(6) grown in the tropical zone of Senegal were used as host culti-vars. Nine bacterial strains were selected from the different groupsof isolates identified. Plants were grown in pots filled with vermicu-lite moistened with a nitrogen-free solution (31). The selected bac-terial strains were grown to 1×107 cells mL−1 in AIE liquid mediumand 3 mL was used to inoculate each of the 5 seeds sown per pot.Pots were placed into a 25°C growth chamber under natural lightand watered every 7 days with sterilized de-ionized water. The nod-ulation and nitrogen fixing abilities of the plants were assessed after28 days using the Acetylene Reduction Assay (Shimadzu Gas Chro-matograph GC-14A, Kyoto, Japan), the number nodules and the dryweights of nodules, shoots and roots from 3 plants per pot. All datawere subjected to a statistical analysis using one-way ANOVA, at aprobability level of 5%. Mean separation was performed using theLeast Significant Difference (LSD) whenever a significant result(P<0.05) was obtained.

Nucleotide sequence accession numbers

The nucleotide sequences of the 16S rRNA and the ITS regionbetween the 16S and 23S rRNA genes of the 57 sequenced isolateswere deposited under accession numbers listed in the Table S1.

Results

Isolation and growth properties of cowpea-nodulating

bacteria

Ten isolates were obtained from the root nodules of each

host and soil combination: DTC1 to DTC10 (Dan IIa, CS),

DTB1 to DTB10 (Dan IIa, BS), TTC1 to TTC10 (Tvu-7778,

CS), TTB1 to TTB10 (Tvu-7778, BS), TSC1 to TSC10

(Tvu-11986, CS), and TSB1 to TSB10 (Tvu-11986, BS).

Consequently, there were 60 isolates in total (Table 1).

Except TSC1, all isolates were considered slow-growing

strains based on the culture period (7 days), the production

of alkaline substances (blue color on BTB medium) and

generation times. The generation times (hours±S.E.) of B.

japonicum USDA 110, DTB1, DTB4 and DTC9 were

17.21±0.23, 16.80±0.29, 16.31±0.09 and 16.72±0.45, respec-

tively. The generation times of the fast-growing reference

S. fredii USDA 194 and TSC1 were 7.88±0.10 and 7.74±

0.37, respectively. Two slow-growing isolates (TTC5 and

DTC9) showed colonies with an undulated-flat shape while

the remaining 57 slow-growing isolates had an entirely-

pulvinate shape.

Sequence analysis of the 16S rRNA gene

The isolates DTB5, DTC7 and TTC7, showing signs of

contamination on agar plates, were omitted from further

analysis. A BLAST analysis of the 16S rRNA gene (1.452

kbp) sequences confirmed that all 57 remaining isolates

belonged to the genus Bradyrhizobium, except TSC1 (1.5

kbp) which was close to members of the genus Ralstonia.

Because of the extensive taxonomic difference between

Bradyrhizobium and Ralstonia, an individual phylogenetic

tree was constructed for each group. The tree for the 56

Bradyrhizobium isolates is shown in Fig. 2a. DTC9 and

TTC5 were closely related to B. elkanii and shared 99%

similarity with the reference strains B. elkanii S127 and

B. elkanii SEMIA 6175. The isolate DTB4 showed 99%

similarity with the 16S rRNA gene portion of the complete

B. japonicum USDA 110 gene sequence. The remaining

53 isolates were also closely related to the reference strain B.

japonicum HF7. TTB3 was separated from the other 52

isolates but shared 99% similarity with B. japonicum HF7 as

well as with B. liaoningense LYG10. Of the 53 isolates closely

related to B. japonicum HF7, thirty showed 100% to 99.58%

homology with DTB6, 7 showed 100% homology with DTB3,

6 shared up to 99.72% homology with TSC10, and 4 isolates

had 99.38% similarity with DTC5 (Fig. 2a). The phylogenetic

tree for TSC1 (Fig. 2b) confirmed this isolate to be a member

of the genus Ralstonia. It shared 99% homology with the 16S

rRNA gene sequences of the Ralstonia detusculanense and

Ralstonia pickettii TA reference strains.

Sequences analysis of 16S-23S rRNA ITS genes

For the ITS region, two distinct phylogenetic trees corre-

sponding to Bradyrhizobium (850 bp) and Ralstonia (607 bp)

Fig. 1. 1.5% agarose gel electrophoresis in 1×TAE buffer stained withSYBR safe DNA gel stain (Invitrogen) of the 16S rRNA gene (A) andITS sequences between the 16S and 23S rRNA genes (B) of 13 selectedisolates. C: 1% agarose gel electrophoresis of 8 plasmid clones for 2isolates (4 clones per isolate). M: marker (100 bp ladder for A and B,λHind III for C), p: pGEM-5Zf(+) non-digested plasmid (control).

M I1 I2 I3 I4 II1 II2 II3 II4 Mp

SARR et al.108

were also constructed. Sequences used to construct the tree

(Fig. 3a) for Bradyrhizobium isolates, contained portions of

the end and start of the 16S and 23S rRNA genes, respec-

tively. This topology confirmed that DTB4 was closely

related to B. japonicum USDA 110 with which it shared

100% similarity, and it was named B. japonicum DTB4.

DTC9 and TTC5 were strains of B. elkanii as they shared

99% similarity with the ITS sequences of the reference

strains USDA 121 and USDA 23. B. elkanii DTC9 and B.

elkanii TTC5 were distinguished from the other isolates

based on the shape and size of the colonies on YMA plates.

In contrast, the group of 53 isolates closely related to B.

Fig. 2. Phylogenetic tree constructed by the UPGMA analysis based on the 16S rRNA gene sequences showing the relationships between our(a) Bradyrhizobium-like isolates (1.452 kbp) or (b) the Ralstonia-like isolate (1.5 kbp) and their phylogenetic relatives (in italics) retrieved fromGenBank. Isolates in brackets harbor similar 16S rRNA gene sequences with the isolate(s) they are facing in the dendrogram. Accession numbers ofthe reference strains are shown in parentheses. B: Bradyrhizobium.

Fig. 3. Phylogenetic tree constructed by the UPGMA analysis based on the ITS sequences between the 16S and 23S rRNA genes showing therelationships between our (a) Bradyrhizobium-like isolates (850 bp) or (b) the Ralstonia-like isolate (607 bp) and their phylogenetic relatives (initalics) retrieved from GenBank. Isolates in brackets harbor similar ITS sequences with the isolate(s) they are facing in the dendrogram. Accessionnumbers of the reference strains are shown in parentheses. Our ITS sequences are flanked on each side by the 16S rRNA gene’s last and the 23SrRNA’s first nucleotides. B: Bradyrhizobium.

B. elkanii

Diversity of Cowpea-Nodulating Bacteria 109

japonicum HF7 in the 16S rRNA gene phylogeny, was clus-

tered with B. yuanmingense in the ITS phylogeny. This

group shared at least 99% similarity with the reference strain

B. yuanmingense TAL760. In this group, thirty six isolates

including DTB3, TSC10 and TTB3 shared 100% similarity

with DTB1. DTC5 shared up to 99.7% similarity with related

isolates, and DTB6 was 100% similar to 8 isolates (Fig. 3a).

Moreover, 22 isolates clustered in the DTB6 sub-group of

the 16S rRNA gene phylogeny were transferred into the

DTB1 sub-group of the ITS phylogeny with which they

showed 100% sequence similarity. The ITS-based phylo-

genetic tree for TSC1 indicated that this isolate was a

member of the genus Ralstonia (Fig. 3b). TSC1 shared 95%

similarity with the ITS rRNA sequence of Ralstonia pickettii

RP273DL and was submitted to GenBank under the name

Ralstonia sp. TSC1.

Nodulation effectiveness and nitrogen fixing potential

Eight bradyrhizobial isolates: DTB1, DTB3, DTB4,

DTB6, DTC6, DTC9, TSC10, TTC9 and TSC1, were

selected as inoculants for the cross-inoculation test. Sym-

biotic properties with the four host cultivars of the eight

isolates are shown in Table 2. The Acetylene Reduction

Activity (ARA) of TSC10 was greater than that of the other

strains. TTC9 had a high ARA value, especially when inocu-

lated on Dan IIa and Melakh cultivars. There was no signifi-

cant difference in ARA between the cultivars when they

were inoculated with DTC5. Melakh was poorly co-related

Table 2. Symbiotic characteristics of selected isolates on the four cowpea host cultivars

Isolate HostNodule number

(plant−1)

Dry weight (mg plant−1) ARA

(μmol h−1 g−1 nodule)Nodule Shoot Root

DTB1 Tvu-11986 24.0 ab 9.93 a 206 a 93 a 4.73 b

Tvu-7778 31.0 a 15.96 a 202 a 95 a 2.27 b

Dan IIa 16.3 b 11.60 a 218 a 90 a 12.95 ab

Melakh 16.3 b 12.30 a 210 a 95 a 17.21 a

LSD0.05 12.6 12.04 180 84 11.21

DTB3 Tvu-11986 13.0 b 8.66 b 148 b 74 c 1.08 c

Tvu-7778 32.6 a 16.53 a 204 b 105 b 1.38 c

Dan IIa 37.3 a 23.70 a 307 a 144 a 8.64 b

Melakh 23.6 ab 19.36 a 321 a 143 a 24.77 a

LSD0.05 14.4 7.18 64 19 5.64

DTB4 Tvu-11986 39.6 a 11.00 b 183 b 125 a 3.97 ab

Tvu-7778 32.6 ab 15.26 b 174 b 137 a 0.65 b

Dan IIa 40.0 a 28.40 a 287 a 143 a 5.24 a

Melakh 9.0 b 6.63 b 211ab 116 a 1.96 ab

LSD0.05 25.9 12.32 98 38 3.98

DTB6 Tvu-11986 16.3 b 12.96 b 215 a 85 b 5.24 b

Tvu-7778 33.0 ab 22.90 b 212 a 126 ab 10.18 b

Dan IIa 46.3 a 39.33 a 367 a 153 a 9.37 b

Melakh 39.3 ab 23.30 b 336 a 136 ab 22.06 a

LSD0.05 17.6 14.88 193 64 10.63

DTC5 Tvu-11986 17.6 b 9.60 b 231 a 93 a 2.89 a

Tvu-7778 37.6 ab 21.03 ab 248 a 116 a 1.19 a

Dan IIa 44.3 a 23.76 a 317 a 132 a 11.46 a

Melakh 28.3 ab 16.23 ab 267 a 106 a 15.98 a

LSD0.05 23.5 11.76 157 58 21.45

DTC9 Tvu-11986 4.6 c 8.13 c 130 b 74 b 0.81 c

Tvu-7778 42.6 a 26.86 a 215 a 110 a 6.19 b

Dan IIa 37.0 a 20.76 b 216 a 100 ab 8.47 ab

Melakh 22.3 b 18.10 b 236 a 111 a 11.42 a

LSD0.05 11.7 5.77 70 30 4.47

TTC9 Tvu-11986 14.6 b 10.50 c 145 b 68 b 1.73 b

Tvu-7778 37.3 ab 20.56 b 226 b 114 ab 2.47 b

Dan IIa 38.6 a 29.46 a 384 a 147 a 15.30 ab

Melakh 36.0 a 24.06 ab 402 a 164 a 31.25 a

LSD0.05 22.8 8.85 141 69 4.47

TSC10 Tvu-11986 19.3 b 10.50 c 162 b 87 b 22.77 ab

Tvu-7778 30.0 b 16.36 bc 206 b 72 b 7.43 b

Dan IIa 45.6 a 32.13 a 377 a 147 a 16.68 ab

Melakh 22.3 b 22.36 b 371 a 171 a 35.90 a

LSD0.05 13.9 6.56 54 57 22.89

For each isolate, means with the same letter in a column are not significantly different at the 5% level. One-way ANOVA was performed using thepair-wise Student t test.

SARR et al.110

with DTB4 (1.96 µmol ethylene h−1 g−1 nodule) compared to

TSC10 (35.90 µmol ethylene h−1 g−1 nodule). The lowest

nitrogen fixing potential was generally obtained with DTB4

and DTC9 inoculants. Although shoot and root dry weights

of cultivars inoculated with DTB1, DTC5 and DTB4 were

not significantly different, Melakh and Dan IIa showed in

general, the highest dry weights compared to Tvu-11986 and

Tvu-7778. This result correlated overall with the higher

ARA measured throughout the root system of these two cul-

tivars. On the other hand, although Malakh and Tvu-7778

had similar nodule dry weights except when inoculated with

DTC9, ARA was significantly greater in Melakh than in

Tvu-7778 (Table 2). Cross-inoculation with TSC1 (Table 3)

indicated that this strain cross-reacted with the four host

cultivars. No significant difference was observed between

cultivars in terms of the number or dry weight of nodules.

However, the Tvu-11986 and Melakh cultivars showed the

highest affinity for TSC1 as their ARA values were signifi-

cantly higher than those of Dan IIa and Tvu-7778.

Discussion

In this study, we have examined the phylogenetic diver-

sity of the indigenous population of cowpea-nodulating

bacteria in the South-West of Japan. Among the 57 isolates

sequenced, some minor discordance was observed between

the 16S and ITS rRNA gene sequences of a group of 53

isolates. This group was closely related to B. japonicum

HF7 in the 16S rRNA gene phylogeny and to B.

yuanmingense TAL760 in the ITS phylogeny. However, for

B. yuanmingense TAL760, only ITS sequences were sub-

mitted to the GenBank database. This may explain why the

closest relative of the 53 isolates in the 16S rRNA gene

phylogeny was B. japonicum HF7. Furthermore, since it was

reported that 16S rRNA genes have limited sequence diver-

gence, that evolutionary relationships are better resolved

when reconstructions are made using the sequence diver-

gence of the spacer region between the 16S rRNA and the

23S rRNA (30, 35), this group of 53 isolates were considered

relatives of B. yuanmingense. However, both the 16S rRNA

and ITS phylogenies clustered DTB4 with B. japonicum,

DTC9 and TTC5 with B. elkanii, and TSC1 with the genus

Ralstonia (Ralstonia sp. TSC1). Therefore, this study revealed

the predominance of B. yunamingense in the soils of the

Kyushu University’s farm. Strains of this species seem to

have a naturally broad distribution. B. yunamingense

CCBAU10071 has been recovered from China in association

with Lespedeza cuneata, a native tree (37). Some strains of

B. yuanmingense were isolated from South America;

LMRT28 was associated with Phaseolus lunatus in Peru (20)

and TAL760 was isolated from Indigofera hirsute in Mexico

(25). B. yuanmingense was also found in association with

cowpea in Botswana (Africa). Information on B. yuanmingense

has thus expanded regarding (i) its range of hosts (P. lunatus,

L. cuneata, I. hirsute, and V. unguculata); and (ii) its geograph-

ical range (China, Peru, Mexico and Botswana) to include

another Asian locality in Japan. Since B. yuanmingense was

found nodulating cowpea in Africa, our findings are confir-

mation of this point and further extend the geographical dis-

tribution of this species in Japan. A non-human-mediated

wide geographical distribution has been also reported previ-

ously for B. canariense (32) and Mesorhizobium plurifarium

(34).

Furthermore, host-microbe specificity is observed in many

legumes and one of the important factors affecting the distri-

bution of indigenous rhizobia. For example, the predominant

root-nodule bacteria of Genistioid legumes are mostly B.

japonicum and B. canariense, while Milletioid legumes

appear to be more commonly nodulated by B. yuanmingense

and B. elkanii (27). Krasova-Wade et al. (15) also reported

that the distribution of indigenous rhizobia was affected

by crops. In this study, the drought-tolerant cultivars har-

bored more diverse rhizobial strains (B. yuanmingense, B.

japonicum and B. elkanii) than the drought-sensitive cultivar

(B. yuanmingense, B. elkanii). This result is consistent with

the finding with Krasova-Wade et al. (15) of greater rhizo-

bial diversity for the drought-resistance Mouride cultivar

than drought-sensitive B-21 cultivar.

This is the first report describing the isolation of a beta-

proteobacterial strain (Ralstonia sp. TSC1) from cowpea

nodules. The Nitrogen-fixing bacteria were limited to the

alpha-subclass of Proteobacteria (α-rhizobia) until the dis-

covery of legume-nodulating Burkholderia strains (20) and

Ralstonia taiwanensis (5) which are betaproteobacterial

strains. R. taiwanensis was found to represent 93% of the

Mimosa isolates in Taiwan, indicating that Betaproteobacteria

can be specific symbionts of a legume. Many nitrogen-fixing

endophytes including strains of Burkholderia (18) and

Herbaspirillum (38) have also been isolated from seeds, and

stems as well as roots of several plants. In the present study,

B. elkanii DTC9, B. elkanii TTC5 and Ralstonia sp. TSC1

were isolated from root nodules of cowpea inoculated with

the cowpea soil (CS), indicating that the bacteria inhabiting

the cowpea soil type were more diverse than those in the

bean soil type.

The cross-inoculation test revealed that B. japonicum

DTB4 and B. elkanii DTC9 had less nitrogen-fixing poten-

tials (ARA) than the B. yuanmingense strains used as inocu-

lants (Table 2). B. yuanmingense TSC10 and TTC9 showed

significantly greater nitrogen-fixing potential and could be of

great importance for further inoculations of cowpea in the

field. Moreover, although B. yunamingense TSC10, DTB1

and DTB3 shared 100% similarity in sequence, they differed

in nitrogen-fixing characteristics. This indicated that strains

of B. yunamingense may harbor the same ITS sequences but

differ in other composite genes, making the ITS analysis

Table 3. Symbiotic characteristics of TSC1 isolate on the four hostcowpea cultivars

Host

ARA

(μmol h−1

g−1 nodule)

Number of nodules

(plant−1)

Nodule Dry weight

(mg plant−1)

Tvu-11986 53.54 a 13.0 a 2.70 a

Tvu-7778 33.82 b 17.0 a 4.95 a

Dan IIa 15.70 c 22.5 a 3.90 a

Melakh 59.41 a 16.5 a 3.50 a

LSD0.05 11.79 17 3.37

In a column, means with the same letter are not significantly differentat the 5% level. One-way ANOVA was performed using the pair-wiseStudent t test.

Diversity of Cowpea-Nodulating Bacteria 111

insufficient to accurately separate them. On the other hand,

the Melakh cultivar was more compatible with inoculants

based on the ARA and gave the highest dry weights together

with Dan IIa, while the cultivar (Tvu-11986) had the lowest

dry weight. Similar results regarding host specificity were

obtained for 16 sorted Senegalese cowpea cultivars (19)

among which B-21 was the worst nitrogen-fixing cultivar,

while the Mouride cultivar showed moderate nitrogen-fixing

potential. Although, cultivar-strain interactions in cowpea

are not well known (1, 9), effects of cultivars on nodule

occupancy have been observed in other plant species includ-

ing soybean and bean (3, 28).

In conclusion, it appears that in a single restricted area,

cowpea can be nodulated by at least four different bacterial

species. Strains of B. yuanmingense predominated in the

zone studied and showed the greatest nitrogen-fixation

potential. Further molecular analysis should be carried out on

the novel species isolated from cowpea (Ralstonia sp.

TSC1), to better understand its symbiotic characteristics.

Acknowledgements

We are grateful to the Ministry of Education, Sciences, Sportsand Culture, Japan, for providing a PhD scholarship.

References

1. Allen, O.N., and E.K. Allen. 1981. The Leguminosae: A Source Bookof Characteristics, Uses and Nodulation. University of WisconsinPress, Madison, USA.

2. Altschul, S.F., T.L. Madden, A.A. Schäffer, J. Zhang, Z. Zhang, W.Miller, and D.J. Lipman. 1997. Gapped BLAST and PSI-BLAST: Anew generation of protein database search programs. Nucleic AcidsRes. 25:3389–3402.

3. Balatti, P.A., and S.G. Pueppke. 1990. Cultivar-specific interactionsof soybean with Rhizobium fredii are regulated by genotype of theroot. Plant Physiol. 94:1907–1909.

4. Brockwell, J., P.J. Bottomley, and J.E. Thies. 1995. Manipulation ofrhizobia microflora for improving legume productivity and the soilfertility. A critical assessment. Plant Soil 174:143–180.

5. Chen, W-M., L. Moulin, C. Bontemps, P. Vandamme, G. Bena, andC. Boivin-Masson. 2003. Legume symbiotic nitrogen fixation by β-Proteobacteria is widespread in nature. J. Bacteriol. 185:7266–7272.

6. Cisse, N., M. Ndiaye, S. Thiaw, and A.E. Hall. 1997. Registration of‘Melakh’ cowpea. Crop Sci. 37:1978.

7. Dakora, F.D., R.A. Aboyinga, Y. Mahama, and J. Apaseku. 1987.Assessment of N2 fixation in groundnut (Arachis hypogeae L.) andcowpea (Vigna unguiculata L. Walp) and their relative N contributionto a succeeding maize crop in Northern Ghana. MIRCEN J. 3:389–399.

8. Duke, J.A. 1981. Handbook of Legumes of World Economic Impor-tance. Plenum Press, New York, USA.

9. Figueiredo, M.V.B., J.J. Vilar, H.A. Burity, and F.P. de França. 1999.Alleviation of water stress effects in cowpea by Bradyrhizobium spp.inoculation. Plant Soil 207:67–75.

10. Hunt, P.G., T.A. Matheny, and A.G. Wollum. 1988. Yield and Naccumulation responses of late-season determinate soybean toirrigation and inoculation with various strains of Bradyrhizobiumjaponicum. Soil Sci. Plant Anal. 19:1601–1612.

11. Inoue, H., H. Nojima, and H. Okayama. 1990. High efficiency trans-formation of Escherichia coli with plasmids. Gene 96:23–28.

12. Jordan, D.C. 1982. Transfer of Rhizobium japonicum Buchanan 1982to Bradyrhizobium gen. nov., a genus of slow-growing, root nodulebacteria from leguminous plants. Int. J. Syst. Bacteriol. 32:136–139.

13. Jordan, D.C. 1984. Family III. Rhizobiaceae Conn 1938. p. 234–244.In N.R. Krieg, and J.G. Holt (ed.), Bergey’s Manual of SystematicBacteriology, vol. 1. The Williams & Wilkins Co., Baltimore, USA.

14. Kimura, M. 1980. A simple method for estimating evolutionary rates

of base substitutions through comparative studies of nucleotidesequences. J. Mol. Evol. 16:111–120.

15. Krasova-Wade, T., I. Ndoye, S. Braconnier, B. Sarr, P. de Lajudie,and M. Neyra. 2003. Diversity of indigenous bradyrhizobia associ-ated with three cowpea cultivars (Vigna unguiculata (L.) Walp.)grown under limited and favorable water conditions in Senegal (WestAfrica). African J. Biotech. 2:13–22.

16. Kuykendall, L.D. 1987. Isolation and identification of geneticallymarked strains of nitrogen-fixing microsymbionts of soybeans,p. 205–220. In G.H. Elkan, (ed.), Symbiotic Nitrogen FixationTechnology, Marcel Dekker Inc., New York, USA.

17. Kuykendall, L.D., B. Saxena, T.E. Devine, and S.E. Udell. 1992.Genetic diversity in Bradyrhizobium japonicum Jordan 1982 and aproposal for Bradyrhizobium elkanii sp. nov. Can. J. Microbiol.38:501–505.

18. Mano, H., and H. Morisaki. 2008. Endophytic bacteria in the riceplant. Microbes Environ. 23: 109–117.

19. Ndiaye, M.A.F., M.M. Spencer, and M. Gueye. 2000. Genetic vari-ability in dinitrogen fixation between cowpea (Vigna unguiculata L.Walp.) cultivars determined using nitrogen-15 isotope dilution tech-nique. Biol. Fert. Soils 32:318–320.

20. Ormeño, E., P. Vinuesa, D. Zúñiga-Dávila, and E. Martinez-Romero.2006. Molecular diversity of native bradyrhizobia isolated from limabean (Phaseolus lunatus L.) in Peru. Syst. Appl. Microbiol. 29:253–262.

21. Rivas, R., A. Willems, J.L. Palomo, P. Garcia-Benavides, P.F.Mateos, E. Martinez-Molina, M. Gillis, and E. Velazquez. 2004.Bradyrhizobium betae sp. nov., isolated from roots of Beta vulgarisaffected by tumour-like deformations. Int. J. Syst. Evol. Microbiol.54:1271–1275.

22. Saeki, Y., A. Kaneko, T. Hara, K. Suzuki, T. Yamakawa, M.T.Nguyen, Y. Nagatomo, and S. Akao. 2005. Phylogenetic analysis ofsoybean-nodulating rhizobia isolated from alkaline soils in Vietnam.Soil Sci. Plant Nutr. 51:1043–1052.

23. Singh, B.B., H.A. Ajeigbe, S.A. Tarawali, S. Fernandez-Rivera, andM. Abubakar. 2003. Improving the production and utilization of cow-pea as food and fodder. Field Crops Res. 84:169–177.

24. Sneath, P.H.A., and R.R. Sokal. 1973. Numerical Taxonomy: ThePrinciples and Practice of Numerical Classification. W. H. Freeman& Co., San Francisco, USA.

25. So, R.B., J.K. Ladha, and J.P. Young. 1994. Photosynthetic sym-bionts of Aeschynomene spp. from a cluster with bradyrhizobia on thebasis of fatty acid and rRNA analyses. Int. J. Syst. Bacteriol. 44:392–403.

26. Sprent, J.I. 1976. Nitrogen fixation by legumes subjected to water andlight stresses, p. 405–420. In P.S. Nutman (ed.), Symbiotic NitrogenFixation in Plants. Cambridge Univ. Press, London, UK.

27. Stepkowski, T., C.E. Hughes, I.J. Law, L. Markiewicz, D. Gurda,A. Chlebicka, and L. Moulin. 2007. Diversification of lupineBradhyrhizobium strains: Evidence from nodulation gene trees. Appl.Environ. Microbiol. 73:3254–3264.

28. Streeter, J.G. 1994. Failure of inoculant rhizobia to overcome thedominance of indigenous strains for nodule formation. Can. J. Micro-biol. 40:513–522.

29. Thies, J.E., B.B. Bohlool, and P.W. Singleton. 1991. Subgroups ofcowpea miscellany symbiotic specificity within Bradyrhizobium spp.For Vigna unguiculata, Phaseolus lunatus, Arachis hypogaea, andMacroptilium atrpurpureum. Appl. Environ. Microbiol. 57:1540–1545.

30. van Berkum, P., and J.J. Fuhrmann. 2000. Evolutionary relationshipsamong the soybean bradyrhizobia reconstructed from 16S rRNA geneand internally transcribed spacer region sequence divergence. Int. J.Syst. Evol. Microbiol. 50:2165–2172.

31. Vincent, J.M. 1970. A Manual for the Practical Study of The Root-Nodule Bacteria. Blackwell Scientific Publications, Ltd., Oxford, UK.

32. Vinuesa, P., M. Leon-Barrios, C. Silvam A. Willems, A. Jarabo-Lorenzo, R. Perez-Galdona, D. Werner, and E. Martinez-Romero.2005. Bradyrhizobium canariense sp. nov., and acid-tolerantendosymbiont that nodulates genistoid legumes (Papilionoideae:Genisteae) from the Canary Islands, along Bradyrhizobiumjaponicum bv. genistearum, Bradyrhizobium genospecies α andBradyrhizobium genospecies β. Int. J. Syst. Evol. Microbiol. 55:569–575.

33. Vlassak, K.M., and J. Vanderleyden. 1997. Factors influencing

SARR et al.112

nodule occupancy by inoculant rhizobia. Crit. Rev. Plant Sci. 16:163–229.

34. Wang, E.T., F.L. Kan, Z.Y. Tan, I. Toledo, W.X. Chen, and E.Martinez-Romero. 2003. Diverse Mesorhizobium plurifarium popula-tions native to Mexican soils. Arch. Microbiol. 180:444–454.

35. Willems, A., R. Coopman, and M. Gillis. 2001. Comparison ofsequence analysis of 16S-23S rDNA spacer regions, AFLP analysisand DNA-DNA hybridizations in Bradyrhizobium. Int. J. Syst. Evol.Microbiol. 51:623–632.

36. Xu, L.M., C. Ge, Z. Cui, J. Li, and H. Fan. 1995. Bradyrhizobiumliaoningense sp. nov. isolated from root nodules of soybeans. Int. J.Syst. Bacteriol. 45:706–711.

37. Yao, Z.Y., F.L. Kan, E.T. Wang, G.H. Wei, and W.X. Chen. 2002.Characterization of rhizobia that nodulate legume species of the genusLespedeza and description of Bradyrhizobium yuanmingense sp. nov.Int. J. Syst. Evol. Microbiol. 52:2219–2230.

38. Zakria, M., J. Njoloma, Y. Saeki, and S. Akao. 2007. Colonizationand nitrogen-fixing ability of Herbaspirillum sp. strain B501 gfp1 andassessment of its growth-promoting ability in cultivated rice.Microbes Environ. 22: 197–206.

39. Zilli, J.E., R.R. Valisheski, F.R.F. Filho, M.C.P. Neves, and N.G.Rumjanek. 2004. Assessment of cowpea rhizobium diversity inCerrado areas of Northeasten Brazil. Brazil. J. Microbiol. 35:281–287.