Genetic diversity of

R E S EA RCH AR T I C L E

Genetic diversity of Mimosa pudica rhizobial symbionts in soils

of French Guiana: investigating the origin and diversity of

Burkholderia phymatum and other beta-rhizobia

Ravi P.N. Mishra1,5, Pierre Tisseyre1, Remy Melkonian1, Clemence Chaintreuil1, Lucie Miche1,Agnieszka Klonowska1, Sophie Gonzalez2, Gilles Bena1,4, Gisele Laguerre3 & Lionel Moulin1

1IRD, UMR LSTM, Montpellier, France; 2IRD, Herbier de Guyane, Cayenne, French Guiana; 3INRA, UMR LSTM, Montpellier, France; and4Laboratoire de Microbiologie et de Biologie Moleculaire, University Mohammed V Agdal, Agdal, Morocco

Correspondence: Lionel Moulin, Laboratoire

des Symbioses Tropicales et

Mediterraneennes, TA-A82/J Campus

de Baillarguet 34398 Montpellier Cedex 5,

France. Tel.: +33 4 67 59 37 63;

fax: +33 4 67 59 38 02; e-mail:

[email protected]

Present address: Ravi P.N. Mishra, Novartis

Vaccines Research Centre, Via Fiorentina 1,

Siena, 53100, Italy.

Received 25 July 2011; revised 14 October

2011; accepted 19 October 2011.

DOI: 10.1111/j.1574-6941.2011.01235.x

Editor: Philippe Lemanceau

Keywords

rhizobia; symbiosis; Mimosa; Burkholderia;

Cupriavidus; biodiversity.

Abstract

The genetic diversity of 221 Mimosa pudica bacterial symbionts trapped from

eight soils from diverse environments in French Guiana was assessed by 16S

rRNA PCR-RFLP, REP-PCR fingerprints, as well as by phylogenies of their 16S

rRNA and recA housekeeping genes, and by their nifH, nodA and nodC symbi-

otic genes. Interestingly, we found a large diversity of beta-rhizobia, with Burk-

holderia phymatum and Burkholderia tuberum being the most frequent and

diverse symbiotic species. Other species were also found, such as Burkholderia

mimosarum, an unnamed Burkholderia species and, for the first time in South

America, Cupriavidus taiwanensis. The sampling site had a strong influence on

the diversity of the symbionts sampled, and the specific distributions of symbi-

otic populations between the soils were related to soil composition in some

cases. Some alpha-rhizobial strains taxonomically close to Rhizobium endophyti-

cum were also trapped in one soil, and these carried two copies of the nodA

gene, a feature not previously reported. Phylogenies of nodA, nodC and nifH

genes showed a monophyly of symbiotic genes for beta-rhizobia isolated from

Mimosa spp., indicative of a long history of interaction between beta-rhizobia

and Mimosa species. Based on their symbiotic gene phylogenies and legume

hosts, B. tuberum was shown to contain two large biovars: one specific to the

mimosoid genus Mimosa and one to South African papilionoid legumes.

Introduction

Rhizobia are a functional class of soil bacteria able to

develop a nitrogen-fixing symbiosis with legumes, and are

also termed legume nodulating bacteria (or LNB). The

legume nodulation ability is spread among the Alpha-

and Beta- subclasses of Proteobacteria, and the names

alpha- and beta-rhizobia have been proposed for conve-

nience (Gyaneshwar et al., 2011).

Beta-rhizobia were originally described in 2001 in two

parallel studies: one describing two Burkholderia strains

(STM678 and STM815) isolated from Aspalathus carnosa

(Papilionoideae) in South Africa and Machaerium luna-

tum (Papilionoideae) in French Guiana, respectively

(Moulin et al., 2001), which were subsequently named

Burkholderia tuberum and Burkholderia phymatum

(Vandamme et al., 2002); and one describing Ralstonia

taiwanensis from two Mimosa (Mimosoideae) species in

Taiwan (Chen et al., 2001), which was later renamed Cu-

priavidus taiwanensis (Vandamme & Coenye, 2004). Since

then, several diversity studies have demonstrated the

widespread occurrence of beta rhizobia as symbionts of

Mimosa species in Costa Rica and Texas (Barrett &

Parker, 2006; Andam et al., 2007), Panama (Barrett &

Parker, 2005; Parker, 2008), Brazil and Venezuela (Chen

et al., 2005b; Bontemps et al., 2010), Taiwan (Chen et al.,

2005a), India (Verma et al., 2004), Papua New Guinea

(Elliott et al., 2009), Australia (Parker et al., 2007), and

China (Liu et al., 2011).

To date, several nodulating beta-rhizobia species have

been described, most of them belonging to the Burkholde-

ria genus. The B. tuberum type strain (STM678T) is able

FEMS Microbiol Ecol && (2011) 1–17 ª 2011 Federation of European Microbiological Societies

Published by Blackwell Publishing Ltd. All rights reserved

MIC

RO

BIO

LO

GY

EC

OLO

GY

to nodulate a range of Cyclopia species (Elliott et al.,

2007b). Strains affiliated to B. tuberum were also isolated

from Mimosa pigra in Panama (Barrett & Parker, 2005),

Mimosa spp. in Costa Rica (Barrett & Parker, 2006) and

from several diverse Mimosa species in Central Brazil

(Bontemps et al., 2010). Burkholderia mimosarum has

been isolated from M. pigra nodules in Taiwan, Venezu-

ela, Brazil, China and Australia (Chen et al., 2006; Parker

et al., 2007; Liu et al., 2011), while Burkholderia nodosa

was isolated from nodules on Mimosa scabrella and

Mimosa bimucronata in Brazil (Chen et al., 2007), and

Burkholderia sabiae from Mimosa caesalpiniifolia nodules,

also in Brazil (Chen et al., 2008). So far, only one nodu-

lating species has been described in the Cupriavidus

genus, namely C. taiwanensis (Chen et al., 2001), though

other nodulating strains have been isolated but not yet

described to species level. Cupriavidus strains were iso-

lated from nodules of Mimosa pudica, Mimosa diplotricha

and M. pigra in Taiwan (Chen et al., 2003, 2005a), Costa

Rica (Barrett & Parker, 2006), Texas (Andam et al., 2007)

and India (Verma et al., 2004).

The species B. phymatum was originally described

based on a single strain, STM815T, which was isolated

from M. lunatum nodules collected in French Guiana

(Moulin et al., 2001; Vandamme et al., 2002). This spe-

cies has since been reported as being isolated from M. pu-

dica nodules in Papua New Guinea (two strains, Elliott

et al., 2007a) and China (four strains, Liu et al., 2011),

but has not yet been found as a symbiont of Mimosa in

Australia, or in Central and South America (Chen et al.,

2003, 2005b; Barrett & Parker, 2005, 2006; Bontemps

et al., 2010). It is relevant to note that the Bontemps

et al. (2010) diversity study was performed on 47 native

(mainly endemic) Mimosa species in Central Brazil, and

not a single strain among the 148 isolates belonged to

B. phymatum. On the other hand, this is not so surpris-

ing, as the preferred symbiont of B. phymatum, M. pudic-

a, was only infrequently encountered in Central Brazil by

Bontemps et al. (2010), as its niche was occupied by the

other 200+ Mimosa spp. that are native/endemic to these

centres of Mimosa diversity (Simon & Proenca, 2000;

Simon et al., 2011). In addition, Central Brazilian endem-

ics appear not to nodulate effectively with B. phymatum

(dos Reis Junior et al., 2010). Nevertheless, although the

Cerrado/Caatinga biomes appear not to contain it, until

further sampling of M. pudica is performed in these

regions, the occurrence of B. phymatum in Central Brazil

cannot be excluded. The type strain of B. phymatum,

STM815, was originally isolated from a M. lunatum nod-

ule in a plant nursery in Paracou (French Guiana). The

demonstration of nodulation ability of M. lunatum by

STM815 has proven inconclusive because of the difficul-

ties in germinating its seeds, and growing this tree under

greenhouse conditions (L. Moulin, unpublished data),

although nodulation tests on a different species of

Machaerium (Machaerium brasilense) were also unsuccess-

ful (Elliott et al., 2007b). On the other hand, Elliott et al.

(2007b, 2009) and dos Reis Junior et al. (2010) showed

that STM815 has a very large host range on Mimosa spe-

cies, with higher symbiotic capacities (competition for

nodulation, nitrogen fixation within nodules) than C. tai-

wanensis LMG19424T. The question thus remains open

about the occurrence and diversity of B. phymatum in

South America.

Previous diversity studies have thus shown an affinity

between beta-rhizobia and Mimosa species. Some Mimosa

species also appear to have a broad host range for beta-

rhizobial strains. For example, in Bontemps et al. (2010),

among 118 strains recovered from 47 Mimosa species and

belonging to at least seven Burkholderia species, 98 were

able to nodulate M. pudica. Mimosa pudica thus appears

to be highly promiscuous, capable of nodulating with a

very diverse range of beta-rhizobial species. Based on this

fact, the aim of this study was to uncover the diversity of

beta-rhizobia using a soil-trapping strategy with M. pu-

dica as a trap host in French Guiana, a country located in

South America, which contains the principal centre of

diversification of the Mimosa genus, that is the Cerrado

region of Brazil (Simon & Proenca, 2000; Simon et al.,

2011). The specific questions we adressed were the follow-

ing: (i) What is the intra and interspecific diversity of

beta-rhizobia trapped by M. pudica in diverse soils of

French Guiana? (ii) Is B. phymatum predominant in

French Guiana? (iii) Is there a link between soil parame-

ters and beta-rhizobial diversity? (iv) What are the origins

of the symbiotic genes in the beta-rhizobial populations?

To isolate M. pudica compatible symbionts, we used a

trapping method on dilutions of soils recovered from

eight sites mainly around Cayenne with M. pudica as a

trap host. Strains were isolated from the nodules, tested

for their nodulation ability, and further characterized at

the genetical and genomic level. To determine the speci-

ficity of the symbiosis between beta-rhizobia and M. pu-

dica, we conducted the same trapping strategy on four

soils using Siratro (Macroptilium atropurpureum), a

broad-host-range legume that is able to trap both alpha

and beta-rhizobia (Lima et al., 2009). The diversity of

symbionts trapped by Siratro was then compared with

that of the symbionts trapped by M. pudica. Rhizobial

genetic diversity was assessed using 16S rRNA PCR-RFLP,

REP-PCR genomic fingerprints, as well as by phylogenies

based upon the 16S rRNA and recA housekeeping genes.

Phylogenies of the symbiosis-related genes nodA, nodC

and nifH were also performed in order to determine the

origin of symbiosis in the beta-rhizobial strains that we

isolated in the coastal lowlands of French Guiana.

ª 2011 Federation of European Microbiological Societies FEMS Microbiol Ecol && (2011) 1–17


2 R.P.N. Mishra et al.

Materials and methods

Soil sampling and characteristics

Two hundred grams of soil was sampled under M. pudica

plants on eight sites in French Guiana (see Table 1),

mainly on Cayenne island (on the coastal side east of

Cayenne, near the towns of Remire and Montjoly) and

on the roadside between Cayenne and Kourou cities. GPS

coordinates and Flora composition of sampled soils are

given in Table 1 and Supporting Information Fig. S1.

Altitude of sampled sites ranged from 7 to 21 m.

Soil parameters were analysed by ‘Laboratoire d’analyse

des sols d’Arras’, INRA, France: pH (water and KCL

method), granulometry (five fractions), humidity (at

105 °C), C/N, CEC (Metson method), total nitrogen and

organic carbon, P2O5 (Joret–Hebert method), total

CaCO3, Na and H+ (BaCl2 extraction), Cu, Fe, Zn and

Mn contents (by ICP-AES).

Plant trapping of rhizobia

Mimosa pudica (B&T world seeds; ecotype M. pudica var.

hispida) was used as a legume host for trapping and for

nodulation tests of bacterial strains. Siratro (M. atropur-

pureum; UCAD, Dakar, Senegal) was also used as a trap

host in parallel experiments. Mimosa pudica seeds were

scarified and surface sterilized with 96% H2SO4 and 3%

Calcium hypochlorite (15 min each treatment, followed

by five or six washes with sterile dH2O); while Siratro

seeds were immersed for 3 min in 96% H2SO4 followed

by thorough washing. The seeds were then soaked over-

night in sterile dH2O, transferred to water agar plates

(0.8% Agar) and incubated overnight at 37 °C for germi-

nation. Seedlings were transferred to Gibson tubes or to

deep-well microplates, and incubated in a tropical plant

growth chamber (30 °C, 80% humidity, 16 h day/8 h

night). Five grams of soil was suspended in 50 mL of

sterile water and vortexed thoroughly. Serial dilutions up

to 10�2 were performed, and 0.7 mL of diluted soil sus-

pension was inoculated in to either Gibson tubes (80 mL)

or to deep-well microplates (8 mL per well) filled with

Jensen’s nutrient medium (Vincent, 1970). Eight repli-

cates were used for each dilution level.

Nodules were harvested at 28 days post-inoculation.

Nodules of varying sizes were selected for isolation (60

nodules for soils S1–S4, and 20 nodules for soils S5–S8).

Nodules were thoroughly washed in running tap water,

sterilized by immersing in 3% calcium hypochlorite for

5 min and washed five or six times with sterile dH2O.

Surface-sterilized nodules were individually crushed in

20 lL dH2O. Nodule contents were streaked on yeast

mannitol agar plates (Vincent, 1970) and incubated at

28 °C for 48 h. All colony types were picked and purified

twice by repeated streaking on YMA plates. All isolates

were retested for nodulation with M. pudica (or M. atro-

purpureum) using Gibson tubes under the growth condi-

tions described earlier, except that 1 mL of exponential

bacterial culture was inoculated into the tubes instead of

soil suspension.

All strains isolated from M. pudica and M. atropurpureum

nodules are listed in Table S1. For long-term

Table 1. Characteristics of soils used for rhizobial trapping

Soil Site characteristics and flora at proximity

GPS

coordinates

Altitude

(m)

Soil parameters

Gr pH Corg Ntot OM Cal Pho

S1 East Cayenne, Coastal gardens, Mimosa

pudica and Machaerium lunatum

4°56′46″N

52°18′03″W

8 Loamy sand 7.04 26.2 1.57 45.3 < 1 0.021

S2 East Cayenne, Montjoly town, M. pudica 4°55′19″N

52°16′38″W

8 Sand 5.95 10.6 0.806 18.3 < 1 < 0.01

S3 East Cayenne, Coastal path, M. pudica 4°56′25″N

52°17′41″W

7 Sand 6.14 10.5 0.816 18.2 < 1 0.011

S4 North Cayenne, Elysee savannah,

M. pudica

4°55′32″N;

52°24′50″W

21 Clay loam 5.39 18.2 0.977 31.5 < 1 < 0.01

S5 Cayenne South, Roadside, M. pudica 4°52′51″N;

52°18′46″W

9 Sandy loam 8.58 36.6 2.55 63.3 65.3 0.18

S6 Cayenne North, Roadside, M. pudica 5°02′50″N;

52°36′20″W

17 Clay loam 6.23 25.9 1.88 44.8 < 1 < 0.01

S7 West of Remire town, M. pudica,

Aeschynomene sp.

4°53′03″N;

52°17′34″W

8 Clay 5.44 59 4.65 102 < 1 0.013

S8 Rochambeau savannah, M. pudica 4°49′26″N;

52°21′20″W

6 Sandy loam 6.20 29.7 2.09 51.4 < 1 < 0.01

Gr, granulometry-soil texture according to particle size; pH, soil pH (in water); Corg, organic carbon content (g kg�1); Ntot, total nitrogen content

(g kg�1); OM, organic matter (g kg�1); Cal, CaCO3 content (g kg�1); Pho, phosphate content (g kg�1).



Investigating the origin and diversity of beta-rhizobia 3

maintenance, strains were grown at 28 °C in YM broth

(Vincent, 1970) and preserved in 20% glycerol at �80 °C.

Molecular methods

Isolation of bacterial DNA was performed using the SDS-

proteinase K lysis procedure described by Moulin et al.

(2004). All PCR templates were performed with GO-Taq

Polymerase (Promega) following manufacturer instruc-

tions.

A nearly full-length16S rRNA gene was amplified by

PCR using the primer pair FGPS6 and FGPS1509 (Nor-

mand et al., 1992), as described in Moulin et al. (2001).

Ten microlitres of the 16S rRNA PCR product were

digested separately with the restriction endonucleases

MspI, HinfI and CfoI (Promega) in 20 lL reaction vol-

ume following manufacturer instructions. The digests

were analysed separately by agarose gel electrophoresis

(3% Metaphor; FMC Bioproducts).

The 800-bp recA fragments were amplified by PCR (ini-

tiation step 94 °C, 5 min, followed by 35 cycles of 94 °C,

30 s, 55 °C, 30 s, 72 °C, 30 s; and final step of 7 min at

72 °C) and sequenced using the primers recABur1F and

recABurk1R (Table 2) for beta-rhizobia, while TSrecAF

and TSrecAr were used for alpha-rhizobia under condi-

tions described in N’Zoue et al. (2009). PCR products were

purified using the Qiaquick Gel Extraction kit (Qiagen),

following manufacturers instructions, and were sequenced

using the same primers that were used for amplification.

A 440-bp fragment of nifH was amplified using primers

nifHfor and nifHrev (Table 2), as described in Chen et al.

(2003). A 500-bp fragment of nodA was amplified and

sequenced for all representative strains using nodA1F and

nodArbrad (Table 2), except strains of B. tuberum for

which specific primers pairs were designed (nodAF-190t,

nodAR1). For strains STM3625 and STM3629, two frag-

ments of the nodA gene were detected. For these two

strains, cloning of each nodA copy was performed using

the pGEMT-Easy kit (Promega), transformed into XL2

electrocompetent cells. Plasmids were extracted with the

Wizard Plus SV Miniprep kit (Promega), and inserts were

sequenced using the universal primers M13F and M13R.

An 800-bp fragment of nodC was also amplified and

sequenced for representative strains as described in Bon-

temps et al. (2010), except for C. taiwanensis strains for

which a specific primer set was designed (nodC-Ctai468F/

nodC-Ctai1231R) (Table 2).

DNA sequencing was subcontracted to Macrogen Inc.

and Genoscreen Inc., using ABI chemistry and ABI3730

sequencers, and the same primers as used for PCR.

REP-PCR fingerprints of each strain were obtained by

PCR amplification using primer pairs REP1R and REP2

(Versalovic et al., 1991). PCR mix and conditions were as

described by Kaschuk et al. (2006), except that the primer

annealing temperature was reduced to 40 °C. PCR prod-

ucts were visualized on 1% agarose gel in TAE 19 buffer

after elecrophoresis. Rep fingerprint groups were defined,

one rep-group including identical or nearly identical (up

to one band difference) REP-PCR profiles.

Phylogenetic analysis

Sequences were corrected using CHROMAS PRO v1.33 (Techn-

elysium Pty Ltd), aligned with CLUSTALX and alignments

Table 2. Primers used for amplification and sequencing of 16S rRNA, recA, nodA, nodC and nifH genes

Primer name Primer sequence (5′–3′) Target gene References

FGPS6 GGAGAG TTAGATCTTGGCTCA 16S rRNA Normand et al. (1992)

FGPS1509 AAGGAGGGGATCCAGCCGCA 16S rRNA Normand et al. (1992)

nifHfor AARGGNGGNATYGGHAARTC nifH Chen et al. (2003)

nifHrev GCRTA8ABNGCCATCATYTC nifH Chen et al. (2003)

NodA1f TGCRGTGGAARNTRBVYTGGGAAA nodA Sy et al. (2001)

NodAr.Brad TCACARCTCKGGCCCGTTCCG nodA Sy et al. (2001)

NodAF-190t AGTTGGGCCGGMGCNAGGCCTGA nodA This study

NodAR1 CAACGAACTGTTAATTGGCA nodA This study

nodCBurk2F ACTSATACTYAACGTMGAYTC nodC Bontemps et al. (2010)

nodCBurk2R GMRAAYCCRAGAAATCGAAG nodC Bontemps et al. (2010)

NodCCtai468F TACTCAAGATGCGGGACAGA nodC This study

NodCCtai1231R CCGGAGATTCCGATGATGGA nodC This study

nodCrhizo1F AYGTHGTYGAYGACGGTTC nodC Bontemps et al. (2010)

nodCrhizo1r CGYGACAGCCANTCKCTATTG nodC Bontemps et al. (2010)

TSrecAf CAACTGCMYTGCGTATCGTCGAAGG recA Stepkowski et al. (2005)

TSrecAr CGGATCTGGTTGATGAAGATCACCATG recA Stepkowski et al. (2005)

recABurk1F GATCGARAAGCAGTTCGGCAA recA This study

recABurk1R TTGTCCTTGCCCTGRCCGAT recA This study




were manually curated with GENEDOC software (Nicholas

& Nicholas, 1997). Phylogenies were built by distance on

MEGA4 (Tamura et al., 2007); or by Maximum likelihood

using PHYML (Guindon et al., 2005) or PAUP4. Bootstrapping

or aLRT (approximative Likelihood Ratio Tests, SH-like)

tests were conducted using the www.phylogeny.fr website

(Dereeper et al., 2008).

Statistical analyses

Shannon diversity indexes were calculated using PAST soft-

ware (Hammer et al., 2001). Genetic differentiation tests

were calculated with GENEPOP (Rousset, 2008). A principal

component analysis (PCA) of soils was performed from

the data matrix including soil physico-chemical parame-

ters and relative frequency of each rhizobial species using

XLSTAT (Addinsoft™).

Nucleotide sequence accession numbers

Sequences were deposited in GenBank under the acces-

sion numbers listed in Table S1.

Results

Trapping of M. pudica symbionts from French

Guiana soils

We trapped 221 isolates from eight soils in French

Guiana using M. pudica as a trap host. All isolates

nodulated M. pudica when tested as pure cultures under

axenic conditions, except for four Burkholderia cepacia

isolates that did not form any nodules by 4 weeks post-

inoculation.

The 221 purified isolates were distributed into 10 ribo-

types by 16S rRNA PCR-RFLP using three restriction

enzymes. Assignment of each profile to a rhizobial species

was then performed by the sequencing of the full 16S

rRNA sequence from a representative strain (see Table 3

and Fig. 1a). The isolates were distributed into three gen-

era and all of them could be assigned to a described spe-

cies based on per cent of 16S rRNA nucleotide identity

(1100 bp alignment): B. phymatum (99–100% 16S rRNA

gene identity with STM815T, two ribotypes, 63 strains),

B. tuberum (98–99% 16S rRNA gene identity with

STM678T, three ribotypes, 132 strains), B. mimosarum

(99–100% 16S rRNA gene identity with PAS44T, one rib-

otype, nine strains), Burkholderia sp. (97% with B. phy-

matum as the closest species, one ribotype, one strain),

C. taiwanensis (98–99% 16S rRNA gene identity with

LMG19424T, one ribotype, six strains), Rhizobium sp.

(98% with closest described species Rhizobium endophyti-

cum CCGE1052, one ribotype, four strains), and B. cepa-

cia (one ribotype, six strains).

A 16S rRNA gene phylogeny (Fig. 1a) was built to infer

the evolutionary relationships among representative

strains of our collection, including selected reference

strains (mostly beta-rhizobial species). Mimosa symbionts

previously described in the literature, as well as M. pudica

strains from Central Brazil (Bontemps et al., 2010), were

also added to the analysis. The Burkholderia sequences

either fell into defined species of symbiotic Burkholderia

(B. mimosarum, B. phymatum, B. tuberum) or into clades

of undescribed species, for example STM4206 grouped in

a clade with strains isolated from diverse Mimosa species

(Mimosa acutistipula, M. pudica, Mimosa tenuiflora,

M. pigra) from Central and South America (Brazil, Pan-

ama and Costa Rica). Strain STM3632, however, fell into

Table 3. Mimosa pudica trapping: strain species assignment and distribution in soils

Species*

Representative

strain

16S-RFLP

ribotype Rep groups S1 S2 S3 S4 S5 S6 S7 S8 All

Burkholderia phymatum STM3619 AAA 13 16 – – 36 – 3 – 2 57

B. phymatum STM3675 DEG 1 – – – 2 – 4 – – 6

Burkholderia sp. STM4206 AAE 1 1 – – – – – – – 1

Burkholderia tuberum STM3649 ABF 12 – 45 55 18 1 2 2 5 128

B. tuberum STM3638 ABD 2 – 2 – – – – – – 2

B. tuberum STM6020 FBF 2 – – – – – – – 2 2

Burkholderia mimosarum STM3621 ABB 3 9 – – – – – – – 9

Cupriavidus taiwanensis STM6018 EGI 3 – – – – 6 – – – 6

Burkholderia cepacia STM3632 CDD ND 5 – – 1 – – – – 6

Rhizobium sp. STM3625 BCC 2 4 – – – – – – – 4

Total 35 47 55 57 7 9 2 9 221

All strains samples are listed in Table S1. S1–S8 refer to the soils listed in Table 1, and numbers indicates distribution of 16S-RFLP types across

soils. Burkholderia phymatum type strain STM815 belonged to RFLP profile AAA.

ND, not determined.

*Species assignment was based on 16S rRNA gene phylogeny.




B. tuberum STM 678T (Acar, SA)DUS 833 (Acal, SA)

Mcas7. 1 (Mca, PA)

STM 6020 (Mp, FG)STM 3671 (Mp, FG)

STM 6035 (Mp, FG)

STM 3638 (Mp, FG)hpud12.1 (Mp, CR)

STM 3621 (Mp, FG)B. mimosarum PAS44T (Mpi, TW) MAP3.5 (Mpi, VE)

Br3429 (Mac, BR)JPY389 (Mp, BR) Br3466 (Mt, BR)

mpud5.2 (Mp,PA)hpud10.4 (Mp,CR) tpig4.4 (Mpi, CR)STM 4206 (Mp, FG)SEMIA 6398 (Pip, BR)Br3432 (Mac, BR)

mpa3.10 (Mpi, AU)

B. nodosa Br3437T (Msc,BR)WSM3937 (Rhy,SA)WSM3930 (Rhy, SA)B. terricola LMG20594T

mpa4.1 (Mpi, AU)Br3469 (Mcam, BR) mpa7.4 (Mpi, AU)

B. caledonica LMG19076T

B. megapolitana LMG23650T .

JPY157 (Mp, BR) JPY268 (Mp, BR) JPY264 (Mp, BR) JPY266 (Mp, BR) B. sabiae Br3407T (Mcae, BR)SEMIA6382 (Mcae, BR)SEMIA6167 (Mcae, BR)

B. phymatum STM815T (Mlun, FG)NGR195A (Mp, PNG)STM 3619 (Mp, FG)NGR114 (Mp, PNG) STM 3675 (Mp, FG)B. terrae KMY01T

TJ182 (Mdi, TW)

STM 3632 (not symbiotic, FG)B. cepacia ATCC25416T

B. vietnamensis MGK3T

B. glumae LMG2196T

B. plantarii LMG9035T

B. pseudomallei 1026bB. thailandensis E264

MS1 (Mp, IN)

LMG19425 (Mdi, TW)tpig.6a (Mpi, CR)Ct amp18 (Masp, TX) cmp2 (Masp, TX) MApud10.1 (Mp,CR) C. necator CIP 104763

C. taiwanensis LMG19424T (Mp, TW)PAS15 (Mpi, TW)

STM 6018 (Mp, FG)BHU1 (Mp, IN)

tpud27.6 (Mp, IN)

C. paucula LM3413T

C. metallidurans CIP 104848 C. campinensis WS2T

C. gilardii LM5886T C. basilensis CIP 106792

Ra. pickettii CIP7323T

Ra. solanacearumGMI1000

NGR181 (Mdi, PNG)

R. rhizogenes LMG9509

TJ171 (Mdi,TW)

UPRM8021 (Mcer, PR)

R. tropici CIAT899T

TJ172 (Mdi, TW) JPY491 (Mxan, BR) R. gallicum R602spT

JPY479 (Mxan, BR)

TJ173 (Mdi,TW) R. etli bv. mimosae Mim1 (Maf, MX) R. etli TJ167 (Mdi,TW)

R. leguminosarum IAM 12609T

STM 3625 (Mp, FG)

S. meliloti LMG6133T

Br. japonicum USDA110

Afipia massiliensis 3463

Afipia felis ATCC53690

Br. elkanii USDA76T

BT

BN

BP

B. caribensis MWAP64T

NGR193A (Mp, PNG)

R. endophyticum CCGE2052T (not symbiotic)

STM4804 (Ma, FG)STM4798 (Ma, FG)

STM4800 (Ma, FG)

STM4811 (Ma, FG)STM4814 (Ma, FG)

98

9397

88

64

100

100

65

64

98

5381

775850

96

100

100

76

100

100

100

100

100

69

68

100

99

99

93

62

59

98

96

76

69

95

93

84

86

80

58

99

99

100

91

99

5698

94

72

68

100

82100

100

5064100

50

57

53

0.02

STM 3649 (Mp, FG)

STM 4801 (Ma, FG)

BM

BSP1

BS

BSP2

CT

BSP3

53

Burkholderia

Cupriavidus-

Ralstonia

16S rDNA

tpud22.2 (Mp, CR)tpud40.a (Mp, CR)

99

98

73

CCGE1002 (Mo, MX)

RT

RE

RSP

BSP4

(a)

STM3649

CCGE1002

B. tuberum STM678T

B. phymatum STM815T

STM3619

JPY268

JPY266

JPY264

JPY157

B. sabiae Br3405

B. sabiae Br3407T

JPY389

STM4206

B. caribensis MWAP64

B. graminis LMG18294

B. terricola LMG20594

B. xenovorans LB400

B. fungorum LMG16225

B. caledonica LMG19076

B. phenazinium LMG2247

B. caryophylli LMG2155

B. glathei LMG14190

B. mimosarum PAS44T

STM3621

B. mimosarum NGR190

B. sacchari LMG19450

B. nodosa Br3437

B. kururiensis LMG19447

B. plantarii LMG9035

B. glumae 2196

B. vietnamensis LMG10929

B. cepacia PC184

B. cenocepacia AU1054

B. cepacia ATCC25416

B. mallei ATCC23344

B. thailandensis LMG20219

B. gladioli LMG2216

cmp2

cmp60

MApud3.4

Tpig6.a gi

MApud10.1

Tpud27.6

MApud8.1

STM6018

C. taiwanensis LMG19424T

cmp52

CCBAU65751

R. etli CFN42T

R. etli bv. mimosae IE4771

R. leg. bv. phaseoli USDA2671

R. leg. bv. viciae 3841

R. fabae CCBAU33202T

R. tropici CIAT899T

R. tropici CFN299

R. gallicum R602spT

R. endophyticum CCGE2052

STM3625

R. giardinii H152T

R. galegae LMG6214T

97

90

94

87

95

95

94

88

97

78

87

93

99

92

98

79

64

77

98

77

93

79

73

82

91

73

99

95

74

69

82

94

98

91

90

10092

72

76

98

70

78

71

87

99

0.05

BP

BT

BSP1

recA

BS

BSP3

BM

CT

(b)




the same clade as ATCC25416, the type strain of the non-

symbiotic B. cepacia species. The 16S rRNA sequence

from Cupriavidus STM6018 clustered in the large C. tai-

wanensis branch of the tree, together with strains isolated

from Mimosa species Worldwide, represents the first

C. taiwanensis isolate from South America (the closest

isolate being one from Costa Rica). The alpha-rhizobial

strain STM3625 fell into the Rhizobium genus, and was

very close to two Rhizobium strains isolated from M. pu-

dica in Costa Rica (tpud22.2 and tpud40.a; sharing 99%

identity). These three strains together formed a clade

close to R. endophyticum (98% identity), which is a

nonsymbiotic endophytic strain isolated from Phaseolus

vulgaris in Mexico (Lopez-Lopez et al., 2010).

The recA gene, a house-keeping gene commonly used

in rhizobial phylogeny (N’Zoue et al., 2009; Miche et al.,

2010), was also sequenced for representative strains to

confirm their species assignment based on their 16S

rRNA gene phylogeny. The derived recA phylogeny

showed identical groupings of strains compared with their

16S rRNA gene phylogenies (Fig. 1b), except for the

BSP1 clade, whose phylogenetic postion was better

resolved using recA (and placed it closer to the B. sabiae

and B. phymatum clades).

Distribution of species and ribotypes across

soil samples

The distribution of 16S rRNA PCR-RFLP types in soils is

presented in Table 3. Rhizobial species exhibited a spe-

cific distribution for each soil, and a variable number of

species was sampled depending on the soil of origin. It is

notable that trapping in soils S5–S8 gave significantly less

nodules (c. 10 nodules per soil sample) and hence, fewer

isolates, compared with soils S1–S4.

Soil S1 hosted four species of Mimosa rhizobia (B. phy-

matum, B. mimosarum, Burkholderia sp., Rhizobium sp.),

while soils S2, S3 and S7 hosted only one (B. tuberum).

Soils S4, S6 and S8 hosted the two genotypes of B. phy-

matum and B. tuberum, and the S5 genotypes of C. tai-

wanensis and B. tuberum (Table 3). Amongst all

genotypes isolated, the most abundant were B. tuberum

(132 strains), followed B. phymatum (63), B. mimosarum

(8), C. taiwanensis (6), B. cepacia (5) and Rhizobium sp.

(4). The most frequent ribotypes in the soils were B.

tuberum (ABF) and B. phymatum (AAA). Other geno-

types were found only in one soil, such as Rhizobium sp.

(BCC), Burkholderia sp. (AAE) and B. mimosarum in soil

S1, and C. taiwanensis in soil S5. A B. cepacia genotype

was also found for several isolates (CDD) in two different

soils (S1 and S4).

Shannon diversity indexes were calculated for each site

and compared (excluding sites S5–S8 for which too few

strains could be sampled). At the species level, two sites

(S2 and S3) were monospecific with regard to isolated

M. pudica symbionts, with 100% B. tuberum (H = 0),

while S1 was the most diverse (H = 1.33), and S4 was

intermediate (H = 0.69). The three index values were sta-

tistically different from each other (v2 exact test,

P < 0.001).

As population structures of Mimosa rhizobia differed

significantly across soil samples, we compared their dis-

tribution with soil parameters. The most striking fea-

tures were that soils S2 and S3 exhibited nearly

identical soil parameters, and they hosted only a single

species, B. tuberum. Cupriavidus strains were only found

in soil S5, which, as the only basic soil amongst all the

soils sampled, was characterized by a high pH value

(8.9). Soil S1 exhibited a neutral pH value (7.04), and

hosted the highest diversity at the species level, including

Fig. 1. 16S rRNA (a) and recA (b) gene phylogenetic trees of Mimosa symbionts and related species (among Alpha- and Betaproteobacteria). The

16S rRNA gene tree was built by Neighbour Joining from a matrix corrected by the Kimura-2 parameter method. Numbers at nodes are

bootstrap% from 1000 replicates (shown only when > 50%). The recA tree was built by Maximum Likelihood following a GTR model; numbers

at nodes are aLRT tests values (Approximate Likelihood-ratio Test, SH-like; shown only when > 50%) for branch support. Scale bars indicate

numbers of substitutions per site. Broken tree lines indicate that branch length is not informative (upper branch trees were reduced to fit to

page). Both trees were rooted with Pseudomonas fluorescens Pf5 (Gammaproteobacteria). Strains isolated during this study (French Guiana) are

in bold (from Mimosa pudica and Macroptilium atropurpureum trapping). Defined bacterial species of Mimosa symbionts are indicated on each

side of the trees: BP, Burkholderia phymatum; BT, Burkholderia tuberum; BSP1–BSP3, Burkholderia spp. clade 1–clade 3; BM, Burkholderia

mimosarum; BS, Burkholderia sabiae; CT, Cupriavidus taiwanensis; RT, Rhizobium tropici; RE, Rhizobium etli; R., Rhizobium; B., Burkholderia; Br.,

Bradyrhizobium. Informations about plant host and geographical origin are only given for beta rhizobia. Abbreviations used: for Mimosa species:

Mp: Mimosa pudica, Mo: Mimosa occidentalis, Mca: Mimosa casta, Mt: Mimosa tenuiflora, Mac: Mimosa acutistipula, Mpi: Mimosa pigra, Msc:

Mimosa scabrella, Mcam: M. camporum, Mdi: Mimosa diplotricha, Mcae: Mimosa caesalpiniifolia, Masp: M. asperata, Mxan: M. xanthocentra,

Mcer: Mimosa ceratonia, Maf: Mimosa affinis; others plants: Ma: Macroptilium atropurpureum, Acar: Aspalathus carnosa, Acal: A. callosa, Pip:

Piptadenia, Rhy: Rhynchosia, Mlun: Machaerium lunatum; strain origin: FG: French Guiana, CR: Costa Rica, SA: South Africa, BR: Brazil, PA:

Panama, AU: Australia, VE: Venezuela, TW: Taiwan, PNG: Papouasie New Guinea, IN: India, TX: Texas (USA), PR: Puerto Rico. Scale bar indicates

number of substitutions per site. Accession numbers are available in Table S3.




both alpha and beta-rhizobial genotypes. All other soils

were acidic and only yielded Burkholderia strains.

Results of PCA with regard to soil physico-chemical

characteristics and frequency of each rhizobial species

are shown in Fig. S2a and b. The values for the three

first principal components of soil factors F1, F2 and F3

were 55%, 21% and 18%, respectively. In both PCA

plots, soils S2 and S3 (in which B. tuberum predomi-

nated) were not distinguished, while soil S7 was sepa-

rated from all the other soils. The alkaline pH and the

high phosphate and CaCO3 content discriminated soil

S5 (which contained C. taiwanensis) along F2. More

generally, the main soil physico-chemical characteristics

related to the distribution of rhizobial species were pH,

phosphate and CaCO3 content, and soil texture (espe-

cially broad sand).

Intraspecific variability of isolates by REP-PCR

fingerprinting

To assess the intraspecific variability of each species sam-

pled, REP-PCR fingerprints were produced for all isolates

(except the B. cepacia strains). The distribution across

soils of each REP-PCR group is given in Table S2. For

B. phymatum and B. tuberum species, 14 and 16 REP-PCR

groups could be distinguished (BP1–BP14; BT1–BT16),

respectively, with two genotypes dominating the isolates

(BP1 and BP11 for B. phymatum, BT1, BT2 for B. tube-

rum) with an even distribution across soils. Some

REP-PCR groups were found to be soil specific, such as

BT7–BT11 in soil S2. The type strain of B. phymatum,

STM815, was included in this study, and its REP-PCR

profile was close but not identical to one of the dominant

B. phymatum groups (BP1), which was found in soils S1

and S4. For B. mimosarum, Rhizobium sp. and C. taiwan-

ensis species, only two to four REP-PCR groups were found

(probably due to a low number of isolates of these species).

At the intraspecific level, and including all genotypes

(based on REP-PCR fingerprints), Shannon indexes were

H = 2.175 (S1), 1.921 (S2), 0.9689 (S3), 1.964 (S4). Soil

S3 was statistically different from the other three sites.

Despite only one species being detected in soil S2 (B. tube-

rum), the number of haplotypes was as high as in S1

(hosting four species), and surprisingly had a much

higher number of haplotypes than S3, which contained

the same rhizobial species, B. tuberum (S2: 10 haplotypes;

S3: 4 haplotypes). Burkholderia phymatum was the most

diverse species (H = 2.29). The three other rhizobial spe-

cies (B. mimosarum, B. tuberum and C. taiwanensis) exi-

hibited similar diversity patterns (H = 1.14; 1.94, 1.01,

respectively). Rhizobial populations at each site (S1–S4)

were compared on the basis of their genetic composition

using Chi-square exact tests. All four populations were

highly differentiated (P > 0.001 for each pair), meaning

that the site of sampling had a high influence on the

diversity of rhizobia sampled.

Phylogenies of nodA, nodC and nifH symbiotic

genes

Phylogenies of nodA and nifH gene fragments of repre-

sentative strains of each 16S rRNA PCR-RFLP type were

built to infer the origin and evolutionary relationships of

symbiotic genes among the bacterial collection.

The nodA data set included representative strains from

each genotype previously described plus nodA sequences

from known beta-rhizobia (essentially strains isolated

from Mimosa and Rhynchosia), as well as sequences from

alpha-rhizobia reference strains and those isolated from

Mimosa. Figure 2a shows a nodA phylogenetic tree

inferred by Maximum likelihood and rooted with Azo-

rhizobium caulinodans ORS571. All beta-rhizobial nodA

sequences grouped in a large clade including isolates from

Mimosa species. This clade was monophyletic and subdi-

vided into four subclades corresponding to: (i) B. phyma-

tum, B. sabiae and Burkholderia sp. strains (BP/BS/BSP1-2

in Fig. 2a), (ii) C. taiwanensis strains (CT), (iii) B. mimo-

sarum strains (BM), (iv) B. tuberum Mimosa strains

(BT1). The BP/BS/BSP1-2 subclade included the B. phy-

matum type strain STM815, NGR195A (isolated from

M. pudica in Papua New Guinea) and DUS751 (isolated

from M. pigra in Vietnam), as well as B. sabiae Br3407

Fig. 2. nodA (a) and nifH (B) phylogenetic trees of Mimosa symbionts and related species (among Alpha- and Betaproteobacteria). Both trees

were built by ML following a GTR model (all ML parameters estimated). Scale bar indicates numbers of substitutions per site. Numbers at nodes

are boostrap% from 100 replicates (shown only when > 50%) performed under ML criterion. Arrows indicate the monophyly of nodA and nifH

in Mimosa beta-rhizobia. Broken tree lines indicate that branch length is not informative (upper branch trees were reduced to fit to page). nodA

and nifH trees were rooted with Azorhizobium caulinodans and Azotobacter vinelandii, respectively. Strains isolated during this study (French

Guiana) are in bold (both Mimosa pudica and Siratro trapping). Species affiliations of each strain are given before each strain name following the

16S rRNA gene phylogeny (see Fig. 1 legend for abbreviations). Information about plant host and geographical origin are given for Mimosa

symbionts. Scale bar indicates number of substitutions per site. The arrow indicates the monophyletic origin of nodA and nifH genes for beta-

rhizobia isolated from Mimosa spp. Abbreviations are the same as those given in Fig. 1 legend, except: Mbi: Mimosa bimucronata, Min: M.

invisa, Acang: Acaciella angustissima, B. sp.: undefined Burkholderia species, leg. leguminosarum. Scale bar indicates number of substitutions per

site. Accession numbers are available in Table S3.




Bm NGR190 (Min, PNG) Bm STM3621 (Mp, FG)B. sp. PTK47 (Mpi, TW) Bm PAS44T (Mpi, TW)

B. sp. DUS751 (Mpi) Bm MAP3-5 (Mpi, VE)

Bm Br3454 (Msc, BR) Bn BR3461 (Mbi, BR)

Bt STM6020 (Mp, FG)Bt STM3638 (Mp, FG)Bt STM3649 (Mp, FG)Bt STM4801 (Mp, FG)

B. vietnamensis G4 B. xenovorans CAC124

B. xenovorans LB400

Bt STM678 (Acar, SA)

Br. elkani STM4798 (Ma, FG)Bradyrhizobium sp. ANU289

Br. japonicum STM4811 (Ma., FG) Br. japonicum USDA110

Br. elkanii STM4804 (Ma. atro., FG)

photosynthetic BradyrhizobiumORS278 Azorhizobium caulinodans ORS571

Xanthobacter autotrophicus Py2

Re CFN42T

Re TJ173 (Mdi, TW) Re TJ167 (Mdi, TW)

Re Mim1 (Mafi, MX)

Rsp STM3625 (Mp, FG)R. gallicum R602sp

Rt TJ171 (Mdi, TW)Rt TJ172 (Mdi, TW)

R. leg. bv. viciae 3841R. leg. bv. trifolii K0049

S. meliloti 1021

M. loti MAFF303099

S. fredii NGR234Sinorhizobium sp. ANU240

Br. sp. STM4800 (Ma, FG)

Rhodopseudomonas palustris CGA009

BT1 bv mimosae

BSP4/BT2 bv papilionoidae

Ct MApud8.1 (Mp, CR)



Ct Tpud27.2 (Mp, CR) Ct tpig.6a (Mpi, CR)

Ct STM6018 (Mp, FG)Ct MS1 (Mp, IN)

Ct LMG19424T (Mp, TW) Ct LMG19425 (Mdi, TW)

Ct NGR193A (Mp, PNG)

Ct cmp2 (Masp, TX)

Ct BHU1 (Mp, IN)

Ct PAS15 (Mpi, TW) Ct amp18 (Masp, TX)

CT

Bp STM3619 (Mp, FG)Bp STM3675 (Mp, FG)

Bsp1 STM4206 (Mp, FG)

Bcar TJ182 (Mdi, TW)

Bp NGR195A (Mp, PNG) Bsp2 BR3469 (Mcam, BR) Bp STM815T (Mlun& FG)

Bs Br3407 (Mcae, BR)

BP/BS/BSP1-2

BM

BN

nifH

Bsp1 JPY389 (Mp, BR)

Bsp3 JPY266 (Mp, BR)Bsp3 JPY264 (Mp, BR)Bsp3 JPY268 (Mp, BR)Bsp3 JPY157 (Mp, BR)

Bt CCGE1002 (Mo, MX)

70

62

100

100

47

94

85

59

78

100

94

89

100

83

99

86

100

100

100

94

75

100

76 86

80

97

100

95

100

100

90

54

98

100

94

100

0.1

BSP3

Bsp4 WSM3930 (Rhy, SA) Bsp4 WSM3937 (Rhy, SA)

100

RE

RT

Azorhizobium caulinodans ORS571

Bs Br3407 T(Mcae, BR)Bsp1 STM4206 (Mp, FG)Bp STM3619 (Mp, FG)Bp STM815T (Mlun, FG)Bp NGR195 Bp STM3675 (Mp, FG)Bp DUS751 (Mpi, VT)

Bsp2 Br3469 (Mcam, BR) Ct Tpig6a (Mpi, CR) Ct MApud10.1 (Mp, CR) Ct NGR193A (Mp, PNG)Ct LMG19425 (Mdi, TW)Ct LMG19424T (Mp, TW) Ct Tpud27.6 (Mp, CR)Ct STM6018 (Mp, FG)Ct cmp2 (Masp, TX)Bcar TJ182 (Mdi, TW) Ct PAS15 (Mpi, TW)

Ct amp18 (Masp, USA)Bm PAS44T (Mpi, TW) Bm MAP3-5 (Mpi, VE) )Bm PTK47 (Mpi, TW) B. sp. mpa6.4 (Mpi, AU) B.sp. mpa2-1a (Mpi, AU) Bm STM3621 (Mp, FG)Bm NGR190 (Mdi, PNG)Bm Br3454 (Msc, BR)

Bn Br3461 (Mbi, BR)Bt STM6020 (Mp, FG)Bt STM3638 (Mp, FG)Bt STM3649 (Mp, FG)Bt CCGE1002 (Mo, MX)

Rt TJ171 (Mdi, TW)Rt TJ172 (Mdi, TW)Rt NGR181 (Mdi, PNG)Rt CFN299T

Rt ESH10 (Acang, MX)Rt CFNESH9 (Acang, MX)Rt ESH23 (Acang, MX)Rsp. STM3625 (I) (Mp, FG)Rsp. STM3629 (I) (Mp, FG)

Rsp STM3625 (II) (Mp, FG)Rsp STM3629 (II) (Mp, FG)

R. sp CFNESH34 (Acang, MX) Re Mim1 (Maf, MX)Re TJ167 (Mdi, TW) (AJ505305)

S. mexicanus ITTGS3 S. chiapanecum ITTGS1 S. chiapanecum ITTGR11 S. chiapanecum ITTHS70 Sinorhizobium sp. ITTGS8

M. plurifarium CFNESH18 M. plurifarium CFNESH19

M. plurifarium CFNESH26 S. kostiense HAMBII1489T

B. tuberum STM678T (Acar, SA) Bsp4 WSM3930 (Rhy, SA)Bsp4 WSM3937 (Rhy, SA)

Methylobacterium nodulans ORS2060T Methylobacterium sp. 4-46

S. fredii NGR234 R. etli CFN42 Br. japonicum USDA110

Br. elkanii LMG6134T

M. loti MAFF303099 R. leg. bv. trifoli WSM1325 R. leg. bv. trifoli ANU843S. meliloti 1021

R. leg. bv. viciae 248

100

56

100

100

100

82

66

97

51

85

90

56

97

99

97

83

100

78

90

100

100

96

99

76

63100

98

65

74

72

100

94

100

70

99

100

100

97

78

0.1

BP/BS/BSP1-2

CT

BM

BT1 bv mimosae

BN

nodA

RE

RT

BSP4/BT2 bv papilionoidae

(a) (b)




and Burkholderia sp. Br3469, two strains isolated from

Mimosa species in Brazil. The CT subclade contained

strains from diverse Mimosa species isolated from North

(Texas) to Central and South America, Asia and Indone-

sia. The BT subclade encompassed all B. tuberum strains

from Mimosa including CCGE1002 (isolated from

Mimosa occidentalis in Mexico). Interestingly, B. tuberum

Mimosa isolates did not group with the type strain of

B. tuberum, STM678, which was isolated from a papilio-

noid legume (A. carnosa) in South Africa. Instead,

STM678 forms a clade within a large branch that includes

all the alpha-rhizobia, and it clusters with Rhynchosia

strains (belonging to the BSP4 clade) that were also iso-

lated from South Africa (Garau et al., 2009).

Two nodA fragments from Rhizobium sp. STM3625

could be amplified by PCR. Cloning and sequencing led

to two separate nodA copies sharing 73% nucleotide iden-

tity. These two nodA copies could also be detected in the

second REP-PCR group of Rhizobium sp. (strain

STM3629). Phylogenetic analyses (Fig. 2a) revealed that

the first copy (named nodA-I) grouped together with Rhi-

zobium tropici strains isolated in Mexico from Acaciella

angustissima (Rincon-Rosales et al., 2009), while the sec-

ond copy (nodA-II) formed a new branch in the nodA

tree without close homologues.

The nifH gene tree topology of Mimosa isolates

(Fig. 2b) was very close to that of the nodA one. The only

difference was that the phylogenetic placement found for

B. tuberum isolates from Mimosa (BT1 clade) differed

from those from South African legumes (BT2 clade), that

is the B. tuberum isolates from Mimosa species were

included in the same ‘Mimosa’ panclade in terms of their

nifH and nodA phylogenies, while the B. tuberum isolates

from South Africa grouped in a separate betaproteobacte-

rial subclade with Burkholderia xenovorans and Burkhol-

deria fungorum.

A large survey of Mimosa symbionts from Central Brazil

was published during the course of this study, with the

publication of a large data set of nodC sequences (Bon-

temps et al., 2010). In order to compare the symbiotic

genes of French Guiana isolates with Brazilian ones, we

sequenced a nodC fragment from each representative

strain in our collection. As for nodA, the nodC phylogeny

(Fig. 3) grouped together all the beta-rhizobial Mimosa

strains, with the main difference being in the sub-branch-

ing of the CT subclade compared with the Burkholderia

subclades. Indeed in the nodA and nifH phylogenies, CT

formed a sister clade to (BP/BS/BSP1-2), while in the

nodC tree, the CT subclade was more deeply branched at

the base of the Mimosa beta-rhizobial panclade. Such

phylogenetic incongruence between trees was further stud-

ied using different ML models to determine whether long-

branch attraction or third base codon saturation were a

cause of bias in the nodA, nifH or nodC trees. Trees built

using a ML codon model (as defined in Moulin et al.,

2004) or by excluding the third base codon position (not

shown) produced the same topologies as the trees shown

in Fig. 2 (for nodA and nifH) and Fig. 3 (for nodC).

Trapping of rhizobia with the broad-host-range

legume M. atropurpureum

In this study, we also wondered whether the trapping

results obtained with M. pudica could be explained by

beta rhizobia being the main rhizobial populations in

French Guiana soil, or if they were because of any speci-

ficity between these symbiotic partners.

In order to assess this, we used the same soil dilutions

that were used for Mimosa trapping (only on first four

soils S1–S4), but with a broad-host-range trapping

legume species, Siratro. A summary of the diversity sam-

pled is given in Table 4, and the complete set of trapped

strains is given in Table S1. Most of the rhizobia trapped

with Siratro were Bradyrhizobium strains, and only one

isolate of Burkholderia could be trapped from soil S2

using this plant host (STM4801, a B. tuberum strain).

Discussion

Mimosa pudica has an affinity for diverse beta-

rhizobia

Specificity between beta-rhizobia and Mimosa species is a

feature reported in several studies based on their isolation

from wild nodules (Chen et al., 2003, 2005a, b, Bontemps

et al., 2010). In this study, we trapped the M. pudica

compatible symbiont diversity from soils and observed a

strong affinity of this legume species for diverse beta-

rhizobial species. To test whether this trapping strategy

(and the observed affinity) was not biased by the natural

populations of rhizobia present in the sampled soils, we

performed a parallel trapping experiment with the broad-

host-range legume Siratro, a legume species that has been

previously shown to trap both alpha and beta-rhizobia

Fig. 3. Maximum likelihood nodC phylogenetic tree of Mimosa spp. symbionts and rhizobial reference strains. ML parameters used: GTR,

estimation of all parameters. Scale bar indicates number of subsitutions per site. Numbers at nodes are aLRT values (approximate Likelihood Ratio

test). Names in bold are Mimosa pudica symbionts. Abbreviations are the same as those in Fig. 1. Plant host, geographical origin and accession

numbers used are indicated after each strain name. Accession numbers are available in Table S3. *STM815 is a broad-host-range Mimosa

symbiont, but was isolated from a Machaerium lunatum nodule.




Azorhizobium caulinodans ORS571

C. taiwanensis LMG19424T (M. pudica, TW)Ct STM6018 (M. pudica, FG)

Sinorhizobium sp. NGR234 M. huakui R7A M. loti MAFF303099

Br. japonicum USDA110R. etli bv. mimosae Mim74 (M.affinis, MX)R. etli TJ173 (M. diplotricha, TW)

R. etli CFN42S. meliloti 1021

R. leguminosarum 3841 Methylobacterium nodulans ORS2060Rhizobium sp. JPY491 (M. xanthocentra, BR)Rhizobium sp. JPY479 (M. xanthocentra,BR)Rhizobium sp. TJ172 (M. diplotricha, TW)

Rhizobium sp. NGR181 (M. invisa, PNG)R. tropici UPRM8021 (Mimosa sp.)

Rhizobium sp. STM3625 (M. pudica, FG)

100

10038

99

95

92

79

65

93

94

94

46

100

32

81

0.1

JPY395 (M. adenocarpa, BR) JPY256 (M. setosa, BR) JPY476 (M. somnians, BRBt STM3649 (M. pudica, FG)JPY604 (M. xanthocentra, BR)JPY636 (M. xanthocentra, BR)JPY634 (M. debilis, BR) JPY637 (M. debilis, BR)JPY326 (M. xanthocentra, BR)JPY325 (M. xanthocentra, BR)JPY595 (M. gracilis, BR) JPY489 (M. thermarum, BR)JPY453 (M. melanocarpa, BR)JPY403 (M. lanuginosa, BR)JPY622 (M. gracilis, BR)JPY410 (M. humivagans, BR)JPY418 (M. gracilis, BR) JPY419 (M. discobola, BR)JPY490 (M. callithrix, BR)JPY468 (M. somnians, BR, BR)JPY605 (M. somnians, BR)

JPY407 (M. somnians, BR)JPY623 (M. radula, BR)JPY431 (M. adenocarpa, BR)JPY385 (M. crumenarioides, BR)

JPY382 (M. velloziana,BR) JPY585 (M. quadrivilis, BR)JPY346 (M. cordistipula, BR) JPY381 (M. gemmulata, BR)

JPY388 (M. lewisii, BR)JPY350 (M. gemmulata, BR)

JPY294 (M. setosissima, BR)JPY404 (M. claussenii, BR)JPY414 (M. venatorum, BR)JPY629 (M. cyclophylla, BR)JPY452 (M. gardneri, BR)JPY300 (M. setosa, BR)JPY283 (M. claussenii, BR)JPY160 (M. claussenii, BR)JPY578 (M. blanchetti, BR)JPY587 (M. velloziana, BR)JPY439 (M. ulei, BR) JPY456 (M. laniceps, BR)Bm STM3621 (M. pudica, FG)B. mimosarum PAS44T(M. pigra, TW)JPY321 (M. pigra, BR)

Bp STM3619 (M. pudica, FG)B. phymatum STM815T (Machaerium lunatum*, FG)Bsp1 STM4206 (M. pudica, FG)

Bsp3 JPY268 (M. pudica, BR)Bsp3 JPY380 (M. gemmulata, BR)

84

99

96

98

73

85

99

90

100

99

99

92

98

100

82

100

94

92

BT

BN

BM

Bsp1 JPY389 (M. pudica, BR)JPY461 (M. quadrivilis, BR)JPY359 (M. tenuiflora, BR)

BP/BSP1/

BSP3

CT




from Brazilian soils (Lima et al., 2009). In our study,

Siratro trapped mainly Bradyrhizobium (96% of isolates)

and only one Burkholderia isolate could be detected, while

M. pudica trapped mostly beta-rhizobia (98% of isolates).

Such results clearly show a specificity of M. pudica

towards beta-rhizobia and that this legume species can be

used as a trap host to recover a large number of Burk-

holderia species (and C. taiwanensis) from soil.

Populations of compatible M. pudica symbionts

differ depending on French Guiana soil under

study

Our sampling strategy revealed different assemblages of

species depending on the soil under study. Some species,

such as B. tuberum, were dominant in several soils, while

others had only minor populations (Rhizobium, B. mimo-

sarum). Compared with other published studies, we

found a large number of beta-rhizobial species in French

Guiana and more interestingly, a high differentiation of

populations between sites (at both inter and intraspecific

levels). It is clear that soil parameters as well as the geo-

graphical distribution and competitive ability to nodulate

of the various rhizobial species affected the trapping

results, and these could explain the specific distribution

of the isolates. The PCA analyses of soil parameters and

bacterial species revealed the significance of soil pH,

phosphate, CaCO3 content and granulometry on the dis-

tribution of rhizobial populations, especially on the pres-

ence of C. taiwanensis in alkaline soils. It should be

noted, however, that C. taiwanensis was found only in

one soil, so it is premature to make any definitive conclu-

sions about any affinity of C. taiwanensis strains to alka-

line soil until more soil sampling has been undertaken.

On the other hand, in the case of acidic pH, several stud-

ies have already underlined the preference of Burkholderia

species for acidic soils (Garau et al., 2009; dos Reis Junior

et al., 2010), which is a feature that might explain their

near absence from the basic soil from which Cupriavidus

strains were isolated. Rhizobium sp. strains were only

found in a soil with neutral pH, but as with C. taiwanensis,

more sampling is required to determine whether the

occurrence of Rhizobium in Mimosa nodules has any spe-

cific connection with soil pH.

Burkholderia phymatum and B. tuberum are

common M. pudica symbionts in French Guiana

This study is the first to our knowledge to use a trapping

strategy, with M. pudica as a trap host, to uncover the

genetic diversity of beta-rhizobia in soils of South America.

The diversity of M. pudica compatible symbionts sam-

pled by our trapping strategy revealed four species of

Burkholderia (B. phymatum, B. mimosarum, B. tuberum

and a yet unnamed Burkholderia species), C. taiwanensis

and a Rhizobium sp., with the diversity in species assem-

blages (or dominance of a single one) depending on the

soil under study. We also report strains of B. cepacia

inhabiting some nodules that were unable to induce any

nodulation when inoculated as pure cultures on M. pudic-

a. The most sampled species was B. tuberum which com-

prised 60% of all isolates (131 strains), with a dominant

16S rRNA RFLP type found in all soils (except one), and

highly diversified (in terms of REP-PCR profiles) in soil

S2 (10 haplotypes) but strangely not in S3 (4 haplotypes).

We chose to name all these strains B. tuberum as they

were phylogenetically grouped in the same clade as B.

tuberum STM678T (in both the 16S rRNA gene and recA

trees), although several subgroups can be observed, and so

we cannot exclude the possibility that several species are

actually present. Interestingly, this ‘large’ B. tuberum-like

species was also detected in the diversity study of Bon-

temps et al. (2010) on rhizobial isolates from 47 Mimosa

species in Brazil (constituting 54% of the 143 rhizobial

isolates), with one strain being reported as isolated from a

M. pudica nodule (JPY389). This species has also been

reported in M. pudica nodules in Costa Rica and Panama

(Barrett & Parker, 2005, 2006). Burkholderia tuberum thus

Table 4. Macroptilium atropurpureum trapping: strain species assignment and distribution in soils

Species*

Representative

strain

16S-RFLP

ribotype S1 S2 S3 S4 All

Bradyrhizobium elkanii STM4798 HMH 6 – – – 6

B. elkanii STM4804 IMI 3 6 2 5 16

Bradyrhizobium japonicum STM4811 LOK – – 1 4 5

B. japonicum STM4814 LNI – – – 1 1

Bradyrhizobium sp. STM4800 INI – 1 – – 1

Burkholderia tuberum STM4801 ABF – 1 – – 1

Total 9 8 3 10 30

S1–S4 refer to the soils listed in Table 1, and numbers indicates distribution of 16S rRNA gene-RFLP types across soils.

ND, not determined.

*Species assignment was based on 16S rRNA phylogeny.




seems to be a well-adapted symbiont of Mimosa species in

South America, widespread in the area of origin and

diversification of the genus Mimosa (the Brazilian Cerra-

do), as well as on the French Guianan coast.

Burkholderia phymatum represented 28% of all sampled

strains with 66 isolates, and was found in four soils (out

of eight sampled), including 15 REP-PCR profiles. Burk-

holderia phymatum was named on the basis of only one

strain (STM815, Vandamme et al., 2002) and was recently

enlarged to three recently by the addition of two Mimosa

isolates from Papua New Guinea (Elliott et al., 2007a).

Our study thus enhances the number of worldwide iso-

lates of B. phymatum to a total of 66 strains, with at least

15 different REP-PCR profiles. The STM815 REP-PCR

profile was close to the most abundant one (BP1) recov-

ered from two soils in the present study. STM815 origi-

nated from a M. lunatum nodule, and so we also isolated

the nodule symbionts of this legume species where it was

present in site S1 (which hosted several B. phymatum

strains from M. pudica). The rhizobial symbionts isolated

from seven nodules from M. lunatum trees were all

restricted to the genus Bradyrhizobium (L. Moulin,

unpublished data). Taken together with the fact that

STM815 does not form nodules on another species of

Machaerium (M. brasilense, Elliott et al., 2007a, b),

M. lunatum does not appear to be a likely host for

B. phymatum, and M. pudica can be accepted as the ori-

ginal and specific host for this species. On the other

hand, Bontemps et al. (2010) did not find any B. phyma-

tum among the 147 strains in their collection from

diverse Mimosa species in the Cerrado, a result that can

be linked to the almost absence of M. pudica in the non-

urbanized areas of this region in which the (mainly ende-

mic) native Mimosa spp. grow. Burkholderia phymatum

might thus be a specific symbiont of M. pudica and be

restricted to soils where this species is present. This idea

is strengthened by the fact that dos Reis Junior et al.

(2010) clearly showed that Cerrado/Caatinga Mimosa

natives and endemics do not nodulate effectively (or at

all) with B. phymatum. However, it should be noted that

BP is capable of effectively nodulating a wide range of

widespread and weedy Mimosa species (Elliott et al.,

2007a; dos Reis Junior et al., 2010), which often share the

same environments as M. pudica, and so we cannot

exclude the possibility that these species might also have

B. phymatum as a natural symbiont. Other Mimosa sym-

biont diversity studies in Central and South America

(Costa Rica, Panama, Brazil, Venezuela), which were per-

formed on a more restricted set of Mimosa species

(including M. pigra, M. pudica, M. scabrella, M. caesalpi-

niifolia and M. diplotricha) reported nodulation only

by B. tuberum and various new species of Burkholderia

(B. mimosarum, B. nodosa, B. sabiae), but not by B. phy-

matum (Barrett & Parker, 2005, 2006; Chen et al., 2005a,

b). Taken together, these data suggest that B. phyma-

tum is less frequent as a Mimosa species symbiont

than B. tuberum, which has spread across the whole con-

tinent up to at least Central America. The apparently

weak geographical distribution of B. phymatum is uncor-

related with the broad host range and high competitive

ability of this species (Elliott et al., 2007a, 2009). On the

other hand, a recent study of Liu et al. (2011)

has described B. phymatum (as well as B. mimosarum and

C. taiwanensis) isolates from M. pudica nodules in China,

and thus, B. phymatum might have a broader geographi-

cal distribution than was previously thought.

Several strains of B. mimosarum were also found in our

sampling. This species was reported in several symbiont

diversity studies of Mimosa species in Taiwan (M. pigra;

Chen et al., 2005a, 2006), China (M. pudica, M. diplotri-

cha; Liu et al., 2011), Brazil (M. pigra, M. pudica, M. sca-

brella; Chen et al., 2005b; Bontemps et al., 2010) and

Australia (M. pigra; Parker et al., 2007), which taken

together suggest that B. mimosarum is relatively specific

to M. pigra. This hypothesis is also strengthened by com-

petitive nodulation assays (Elliott et al., 2009) showing

that B. mimosarum PAS44 is more competitive for nodu-

lation of M. pigra than B. phymatum STM815 (highly

competitive) and C. taiwanensis LMG19424, at least in a

solid medium.

The occurrence of C. taiwanensis as a M. pudica symbi-

ont in French Guiana is more enigmatic, as it had not pre-

viously been reported in South America. The detection of

this species in our study was observed in only one soil,

which was characterized by a high pH, which might

explain its predominance at this site over Burkholderia

which appears to be more adapted to acidic soils (Table 1,

Table S1, Garau et al., 2009; dos Reis Junior et al., 2010).

The presence of a low number of alpha-rhizobial iso-

lates in our collection was not surprising, as several other

studies have already found alpha rhizobia in Mimosa

nodules (Chen et al., 2003; Barrett & Parker, 2006; Elliott

et al., 2009; Bontemps et al., 2010), although they always

represented a low% of sampled isolates compared with

beta-rhizobia. The French Guiana Rhizobium isolates were

very close to the Costa Rican strains tpud22.2 and

tpud44A (99% identity in 16S rRNA), which were iso-

lated from nodules collected from wild M. pudica plants

(Barrett & Parker, 2006). All three strains might represent

a new species close to R. endophyticum (both in

16S rRNA and recA phylogenies), which is a nonsymbiot-

ic P. vulgaris endophyte from Mexico (Lopez-Lopez et al.,

2010). Moreover, the same rhizobial genotype was found

in a parallel study on M. pudica symbionts in New

Caledonia, indicating that this Rhizobium species is spe-

cific and distributed worldwide spread as a M. pudica-




interacting bacterium (A. Klonowska, unpublished data).

It is interesting to note that this group of strains is unre-

lated to other Mimosa-compatible Rhizobium species

described so far, such as R. tropici TJ171 and TJ172 iso-

lated from M. diplotricha in Taiwan and NGR181 from

Mimosa ceratonia in Papua New Guinea, or Rhizobium

etli bv. mimosae TJ167 from M. diplotricha in Taiwan and

Mim1 from Mimosa affinis in Mexico (Wang et al., 1999;

Chen et al., 2003; Elliott et al., 2009). The French Guiana

Rhizobium isolates exhibited atypical symbiotic gene

properties, hosting two separate copies of nodA gene

(Fig. 2a). Such a feature has not been reported to date to

our knowledge, and the full genome sequencing of

STM3625 confirmed the presence of two nodA copies in

two separate operons under the control of nodboxes (L.

Moulin, unpublished data). Such features are currently

under study to unravel the putative functional role of

both copies in the symbiotic interaction.

Symbiotic genes of M. pudica beta-rhizobia are

monophyletic

The nodA, nodC and nifH gene phylogenies of M. pudica

beta rhizobia from French Guiana were monophyletic and

mirrored the 16S rRNA and recA gene phylogenies. Such a

phylogenetic pattern indicates that the ancestor of con-

temporary Burkholderia species had already interacted

with legumes (probably with the ancestor of Mimosa) and

that nod genes were vertically inherited during Burkholde-

ria speciation events. Such a pattern had already been

observed in the phylogeny of nodC genes from Burkholde-

ria symbionts of 47 Mimosa species (Bontemps et al.,

2010), and the present study can add B. phymatum to this

pattern, as this species has not been recovered from any

study in Brazil. As with almost all alpha-rhizobia, the nod-

ulation genes of beta-rhizobia are assembled in operons

carried by symbiotic plasmids (Moulin et al., 2001; Chen

et al., 2003; Amadou et al., 2008) and they are supposed

to have co-evolved jointly. Therefore, finding common

evolutionary patterns for both genes (nodA and nodC) is

not surprising and adds support to their vertical inheri-

tance from a common symbiotic ancestor. The nodA and

nifH phylogenies placed the French Guiana Cupriavidus

strain in a sister clade to B. phymatum/B. sabiae, together

with other C. taiwanensis strains sampled in South Eastern

Asia. Such a phylogenetic position suggests that Cupriavi-

dus inherited nod genes from an ancestor of the B. phyma-

tum and B. sabiae species (whch are closely related to each

other according to their 16S rRNA and recA gene phyloge-

nies). However, incongruence was found in the branching

of the CT subclade in the phylogenies of the symbiotic

genes: CT grouped as a sister clade to the BP/BS clade in

the nodA and nifH trees, but in the nodC phylogeny CT

was positioned as an out-group to the Burkholderia

Mimosa panclade. Symbiotic genes in C. taiwanensis thus

appear to have a more complex evolutionary history than

those in Burkholderia, with several nod gene transfers.

As the Mimosa genus diversified into more than 496

species in the Americas (Simon et al., 2011), it is possible

that Burkholderia co-diversified with Mimosa into several

symbiotic species, each exhibiting some specificity in their

host range. Several studies have already shown the affinity

of Burkholderia spp. to Mimosa species, such as B. sabiae

with M. caesalpiniifolia (Chen et al., 2008) and B. mimo-

sarum with M. pigra (Chen et al., 2005a, b, Chen et al.,

2006), which might explain their relatively weak occur-

rence in our M. pudica trapping experiments.

The B. tuberum species contains at least two

biovars

While a pattern of codiversification between Burkholderia

and Mimosa species is suggested, a case of adaptation of a

beta-rhizobial species to different legume subfamilies was

found with B. tuberum. Indeed, B. tuberum Mimosa

isolates grouped in a subclade of the large Mimosa-

beta-rhizobia panclade in the nodA and nodC trees, while

B. tuberum from papilionoid species (STM678 isolated

from A. carnosa and able to nodulate Cyclopia species,

Moulin et al., 2001; Elliott et al., 2007b) grouped in a

branch together with Burkholderia sp. (BSP4) strains from

Rhynchosia ferulifolia (Garau et al., 2009) and was placed

outside the nodA ‘Mimosa’ panclade (with alpha rhizobia,

such as Methylobacterium and Bradyrhizobium). In this

latter case, the nodA phylogeny of beta-rhizobia correlated

with the host range rather than with the bacterial species

phylogeny, thus indicating different origins of nod genes

in B. tuberum. The nifH phylogeny also separated the

papilionoid isolates (which were close to B. xenovorans)

from the mimosoid ones (which grouped with other

Mimosa Burkholderia isolates). It is also relevant to note

that B. tuberum is the only known species of beta-rhizo-

bia with strains able to nodulate two subfamilies of

legumes (Mimosoids and Papilionoids). Burkholderia

tuberum thus appears as a more diversified legume symbi-

ont than B. phymatum in terms of geographical distribu-

tion and legume host range, and two biovars can be

defined according to their host range and nod gene phy-

logeny: biovar mimosae (Mimosa symbionts) and biovar

papilionoideae (Cyclopia symbionts), originating from

two separate horizontal transfers of nodulation genes.

Conclusion

This study has shown that M. pudica is a good trapping

host to study the diversity of soil beta rhizobia and that




it exhibits a high affinity for several Burkholderia species.

Our study also revealed a site-dependant distribution of

beta-rhizobial species. A vertical inheritance of symbiotic

genes among Burkholderia species was also highlighted.

The specific distribution of rhizobia in diverse soils could

be attributed occasionally to environmental factors, and it

showed variable phylogeographical patterns. The large

number of isolates sampled in this study will be further

characterized for their nodulation competitive abilities

and symbiotic efficiencies to understand how and to

which extent diverse beta-rhizobial species have adapted

to Mimosa species in comparison with alpha-rhizobia.

Moreover, the full genome sequencing of Rhizobium sp.

STM3625 and B. tuberum STM3649 at Genoscope as well

as B. mimosarum STM3621 and C. taiwanensis STM6018

as part of the GEBA-RNB project (Genome Encyclopedia

of Bacteria & Archae, Root Nodulating Bacteria) at the

Joint Genome Institute, should give insights into the

adaptation of these various species towards establishing

symbiosis with legumes.

Acknowledgements

We thank Marcelo Simon for identification of the M. pu-

dica ecotype, Euan James for helpful discussions, and the

French National Agency of Research (ANR) for funding

(Project ‘BETASYM’ ANR-09-JCJC-0046). R.P.N.M. was

funded by a postdoctoral grant from IFCPAR (Indo-

French Centre for the Promotion of Advanced Research).

References

Amadou C, Pascal G, Mangenot S et al. (2008) Genome

sequence of the beta-rhizobium Cupriavidus taiwanensis and

comparative genomics of rhizobia. Genome Res 18: 1472–

1483.

Andam CP, Mondo SJ & Parker MA (2007) Monophyly of

nodA and nifH genes across Texan and Costa Rican

populations of Cupriavidus nodule symbionts. Appl Environ

Microbiol 73: 4686–4690.

Barrett CF & Parker MA (2005) Prevalence of Burkholderia sp.

nodule symbionts on four mimosoid legumes from Barro

Colorado Island, Panama. Syst Appl Microbiol 28: 57–65.

Barrett CF & Parker MA (2006) Coexistence of Burkholderia,

Cupriavidus, and Rhizobium sp. nodule bacteria on two

Mimosa spp. in Costa Rica. Appl Environ Microbiol 72:

1198–1206.

Bontemps C, Elliott GN, Simon MF et al. (2010) Burkholderia

species are ancient symbionts of legumes. Mol Ecol 19: 44–52.

Chen WM, Laevens S, Lee TM, Coenye T, De Vos P, Mergeay

M & Vandamme P (2001) Ralstonia taiwanensis sp. nov.,

isolated from root nodules of Mimosa species and sputum

of a cystic fibrosis patient. Int J Syst Evol Microbiol 51:

1729–1735.

Chen WM, Moulin L, Bontemps C, Vandamme P, Bena G &

Boivin-Masson C (2003) Legume symbiotic nitrogen

fixation by beta-proteobacteria is widespread in nature.

J Bacteriol 185: 7266–7272.

Chen WM, James EK, Chou JH, Sheu SY, Yang SZ & Sprent JI

(2005a) Beta-rhizobia from Mimosa pigra, a newly discovered

invasive plant in Taiwan. New Phytol 168: 661–675.

Chen WM, de Faria SM, Straliotto R et al. (2005b) Proof that

Burkholderia strains form effective symbioses with legumes:

a study of novel Mimosa-nodulating strains from South

America. Appl Environ Microbiol 71: 7461–7471.

Chen WM, James EK, Coenye T, Chou JH, Barrios E, de Faria

SM, Elliott GN, Sheu SY, Sprent JI & Vandamme P (2006)

Burkholderia mimosarum sp. nov., isolated from root

nodules of Mimosa spp. from Taiwan and South America.

Int J Syst Evol Microbiol 56: 1847–1851.

Chen WM, de Faria SM, James EK, Elliott GN, Lin KY, Chou

JH, Sheu SY, Cnockaert M, Sprent JI & Vandamme P

(2007) Burkholderia nodosa sp. nov., isolated from root

nodules of the woody Brazilian legumes Mimosa

bimucronata and Mimosa scabrella. Int J Syst Evol Microbiol

57: 1055–1059.

Chen WM, de Faria SM, Chou JH, James EK, Elliott GN,

Sprent JI, Bontemps C, Young JP & Vandamme P (2008)

Burkholderia sabiae sp. nov., isolated from root nodules of

Mimosa caesalpiniifolia. Int J Syst Evol Microbiol 58: 2174–

2179.

Dereeper A, Guignon V, Blanc G et al. (2008) Phylogeny.fr:

robust phylogenetic analysis for the non-specialist. Nucleic

Acids Res 36: 465–469.

dos Reis Junior FB, Simon MF, Gross E et al. (2010)

Nodulation and nitrogen fixation by Mimosa spp. in the

Cerrado and Caatinga biomes of Brazil. New Phytol 186:

934–946.

Elliott GN, Chen WM, Chou JH et al. (2007a) Burkholderia

phymatum is a highly effective nitrogen-fixing symbiont of

Mimosa spp. and fixes nitrogen ex planta. New Phytol 173:

168–180.

Elliott GN, Chen WM, Bontemps C, Chou JH, Young JP,

Sprent JI & James EK (2007b) Nodulation of Cyclopia spp.

(Leguminosae, Papilionoideae) by Burkholderia tuberum.

Ann Bot 100: 1403–1411.

Elliott GN, Chou JH, Chen WM, Bloemberg GV, Bontemps C,

Martinez-Romero E, Velazquez E, Young JP, Sprent JI &

James EK (2009) Burkholderia spp. are the most competitive

symbionts of Mimosa, particularly under N-limited

conditions. Environ Microbiol 11: 762–778.

Garau G, Yates RJ, Deiana P & Howieson JG (2009) Novel

strains of nodulating Burkholderia have a role in

nitrogen fixation with papilionoid herbaceous legumes

adapted to acid, infertile soils. Soil Biol Biochem 41:

125–134.

Guindon S, Lethiec F, Duroux P & Gascuel O (2005)

PHYML Online – a web server for fast maximum

likelihood-based phylogenetic inference. Nucleic Acids Res

33: 557–559.




Gyaneshwar P, Hirsch AM, Moulin L et al. (2011) Legume-

nodulating betaproteobacteria: diversity, host range and

future prospects. Mol Plant Microbe Interact 24: 1276–1288.

Hammer Ø, Harper DAT & Ryan PD (2001) PAST:

Paleontological Statistics Software Package for education

and data analysis. Palaeontologia Electronica 4: 4–9.

Kaschuk G, Hungria M, Andrade DS & Campos RJ (2006)

Genetic diversity of rhizobia associated with common bean

(Phaseolus vulgaris L.) grown under no-tillage and

conventional systems in Southern Brazil. Appl Soil Ecol 32:

210–220.

Lima AS, Nobrega RSA, Barberi A, da Silva K, Ferreira DF &

Moreira FMS (2009) Nitrogen-fixing bacteria

communities occurring in soils under different uses in the

Western Amazon Region as indicated by nodulation of

siratro (Macroptilium atropurpureum). Plant Soil 319:

127–145.

Liu XY, Wu W, Wang ET, Zhang B, Macdermott J & Chen

WX (2011) Phylogenetic relationships and diversity of beta-

rhizobia associated with Mimosa spp. grown in

Sishuangbanna, China. Int J Syst Evol Microbiol 61: 334–342.

Lopez-Lopez A, Rogel MA, Ormeno-Orrillo E, Martinez-

Romero J & Martinez-Romero E (2010) Phaseolus vulgaris

seed-borne endophytic community with novel bacterial

species such as Rhizobium endophyticum sp. nov. Syst Appl

Microbiol 33: 322–327.

Miche L, Moulin L, Chaintreuil C, Contreras-Jimenez JL,

Munive-Hernandez JA, Del Carmen Villegas-Hernandez M,

Crozier F & Bena G (2010) Diversity analyses of

Aeschynomene symbionts in Tropical Africa and Central

America reveal that nod-independent stem nodulation is

not restricted to photosynthetic bradyrhizobia. Environ

Microbiol 12: 2152–2164.

Moulin L, Munive A, Dreyfus B & Boivin-Masson C (2001)

Nodulation of legumes by members of the beta-subclass of

Proteobacteria. Nature 411: 948–950.

Moulin L, Bena G, Boivin-Masson C & Stepkowski T (2004)

Phylogenetic analyses of symbiotic nodulation genes support

vertical and lateral gene co-transfer within the

Bradyrhizobium genus. Mol Phylogenet Evol 30: 720–732.

Nicholas KB & Nicholas BJ (1997) Genedoc: a tool for editing

and annotating multiple sequence alignments. Distributed

by the author.

Normand P, Cournoyer B, Simonet P & Nazaret S (1992)

Analysis of a ribosomal RNA operon in the actinomycete

Frankia. Gene 111: 119–124.

N’Zoue A, Miche L, Klonowska A, Laguerre G, de Lajudie P &

Moulin L (2009) Multilocus sequence analysis of

bradyrhizobia isolated from Aeschynomene species in

Senegal. Syst Appl Microbiol 32: 400–412.

Parker MA (2008) Symbiotic relationships of legumes and

nodule bacteria on Barro Colorado Island, Panama: a

review. Microb Ecol 55: 662–672.

Parker MA, Wurtz A & Paynter Q (2007) Nodule symbiosis of

invasive Mimosa pigra in Australia and in ancestral habitats:

a comparative analysis. Biol Invasions 9: 127–138.

Rincon-Rosales R, Lloret L, Ponce E & Martinez-Romero E

(2009) Rhizobia with different symbiotic efficiencies

nodulate Acaciella angustissima in Mexico, including

Sinorhizobium chiapanecum sp. nov. which has common

symbiotic genes with Sinorhizobium mexicanum. FEMS

Microbiol Ecol 67: 103–117.

Rousset F (2008) genepop’007: a complete re-implementation

of the genepop software for Windows and Linux. Mol Ecol

Resour 8: 103–106.

Simon MF & Proenca C (2000) Phytogeographic patterns of

Mimosa (Mimosoideae, Leguminosae) in the Cerrado biome

of Brazil: an indicator genus of high altitude centers of

endemism? Biol Conserv 96: 279–296.

Simon MF, Grether R, Queiroz LP, Sarkinen TE, Dutra VF &

Hughes CE (2011) The evolutionary history of Mimosa

(Leguminosae): towards a phylogeny of the sensitive plants.

Am J Bot 98: 1201–1221.

Stepkowski T, Moulin L, Krzyzanska A, McInnes A, Law IJ &

Howieson J (2005) European origin of Bradyrhizobium

populations infecting lupins and serradella in soils of

Western Australia and South Africa. Appl Environ Microbiol

71: 7041–7052.

Sy A, Giraud E, Jourand P et al. (2001) Methylotrophic

Methylobacterium bacteria nodulate and fix nitrogen in

symbiosis with legumes. J Bacteriol 183: 214–220.

Tamura K, Dudley J, Nei M & Kumar S (2007) MEGA4:

Molecular Evolutionary Genetics Analysis (MEGA) software

version 4.0. Mol Biol Evol 24: 1596–1599.

Vandamme P & Coenye T (2004) Taxonomy of the genus

Cupriavidus: a tale of lost and found. Int J Syst Evol

Microbiol 54: 2285–2289.

Vandamme P, Goris J, Chen WM, de Vos P & Willems A

(2002) Burkholderia tuberum sp. nov. and Burkholderia

phymatum sp. nov., nodulate the roots of tropical legumes.

Syst Appl Microbiol 25: 507–512.

Verma SC, Chowdhury SP & Tripathi AK (2004) Phylogeny

based on 16S rDNA and nifH sequences of Ralstonia

taiwanensis strains isolated from nitrogen-fixing nodules of

Mimosa pudica, in India. Can J Microbiol 50: 313–322.

Versalovic J, Koeuth T & Lupski JR (1991) Distribution of

repetitive DNA sequences in eubacteria and application to

fingerprinting of bacterial genomes. Nucleic Acids Res 19:

6823–6831.

Vincent JM (1970) A manual for the practical study of root-

nodule bacteria. I.B.P. Handbook, vol. 15. Blackwell Scientific

Publications, Ltd, Oxford.

Wang ET, Rogel MA, Garcia-de los Santos A, Martinez-

Romero J, Cevallos MA & Martinez-Romero E (1999)

Rhizobium etli bv. mimosae, a novel biovar isolated from

Mimosa affinis. Int J Syst Bacteriol 49: 1479–1491.

Supporting Information

Additional Supporting Information may be found in the

online version of this article:




Fig. S1. Localization of sampling sites.

Fig. S2. PCA analysis representation of soils with regard

to soil physico-chemical parameters and M. pudica rhizo-

bial species distribution.

Table S1. Listing of strains trapped on French Guiana

sois by Mimosa pudica and Macroptilium atropurpureum.

Table S2. REP-PCR grouping and geographical distribution.

Table S3. Accession numbers used for reference strains in

16S rRNA, recA, nifH, nodA and nodC gene phylogenies,

by order of appearance from top to bottom in Fig. 1a

(16S rRNA), 1B (recA), 2A (nodA), 2B (nifH) and 3

(nodC).

Please note: Wiley-Blackwell is not responsible for the

content or functionality of any supporting materials sup-

plied by the authors. Any queries (other than missing

material) should be directed to the corresponding author

for the article.




Genetic diversity of

Documents