Accepted Manuscript Title: Revised phylogeny of Rhizobiaceae: Proposal of the delineation of Pararhizobium gen. nov., and thirteen new species combinations Author: Seyed Abdollah Mousavi Anne Willems Xavier Nesme Philippe de Lajudie Kristina Lindstr¨om PII: DOI: Reference:
S0723-2020(14)00191-X http://dx.doi.org/doi:10.1016/j.syapm.2014.12.003 SYAPM 25672
To appear in: Received date: Revised date: Accepted date:
23-9-2014 7-12-2014 12-12-2014
Please cite this article as: S.A. Mousavi, A. Willems, X. Nesme, P. de Lajudie, K. Lindstr¨om, Revised phylogeny of Rhizobiaceae: proposal of the delineation of Pararhizobium gen. nov., and thirteen new species combinations, Systematic and Applied Microbiology (2014), http://dx.doi.org/10.1016/j.syapm.2014.12.003 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
1
Revised phylogeny of Rhizobiaceae: proposal of the delineation of Pararhizobium gen. nov.,
2
and thirteen new species combinations
3
Seyed Abdollah Mousavia,e*, Anne Willemsb, Xavier Nesmec, Philippe de Lajudied, Kristina Lindströme
4 5
a
6
Biocentre 1, Viikinkaari 9, P.O. Box 56, Helsinki, Finland, FIN-00014
7 8
b
9
c
ip t
University of Helsinki, Division of Microbiology and Biotechnology, Department of Food and Environmental Sciences,
cr
Laboratory of Microbiology, Department of Biochemistry and Microbiology, Faculty of Science, Ghent University, K. L. Ledeganckstraat 35, B-9000 Gent, Belgium
University of Lyon; Université Lyon1; Ecologie Microbienne, UMR CNRS 5557 / USC INRA 1364, 16 rue R. Dubois,
F-69622 Villeurbanne cedex, France
11
d
12
Montpellier Cédex 5, France
13
e
14
00014
15
*Corresponding author. email address: [email protected] or [email protected] (Seyed Abdollah
16
MOUSAVI).
17
Department of Environmental Sciences, Viikinkaari 2a, P.O. Box 65, University of Helsinki, Helsinki, Finland, FIN-
18
00014.
19
Running title: Phylogeny of Rhizobiaceae
us
10
an
IRD, Laboratoire des Symbioses Tropicales et Méditerranéennes, Campus de Baillarguet TA A-82/J, 34398
te
d
M
University of Helsinki, Department of Environmental Sciences, Viikinkaari 2a, P.O. Box 65, Helsinki, Finland, FIN-
Ac ce p
20 21
Abstract
22
The family Rhizobiaceae accommodates the seven genera Rhizobium, Neorhizobium, Allorhizobium,
23
Agrobacterium, Ensifer (syn. Sinorhizobium), Shinella and Ciceribacter. However, several so-
24
called Rhizobium species do not exhibit robust phylogenetic positions. Rhizobium is extremely
25
heterogeneous and is in need of major revision. Therefore, a phylogenetic examination of the family
26
Rhizobiaceae by multilocus sequence analysis (MLSA) of four housekeeping genes among 100
27
strains of the family was undertaken. Based on the results we propose the delineation of the new
28
genus Pararhizobium in the Rhizobiaceae family, and thirteen new species combinations:
29
Agrobacterium nepotum comb. nov., Agrobacterium pusense comb. nov., Agrobacterium
30
skierniewicense comb. nov., Allorhizobium vitis comb. nov., Allorhizobium taibaishanense comb.
31
nov., Allorhizobium paknamense comb. nov., Allorhizobium oryzae comb. nov., Allorhizobium
32
pseudoryzae comb. nov., Allorhizobium borbori comb. nov., Pararhizobium giardinii comb. nov.,
1 Page 1 of 15
33
Pararhizobium capsulatum comb. nov., Pararhizobium herbae comb. nov., and Pararhizobium
34
sphaerophysae comb. nov.
35 36
Key words: MLSA, Housekeeping genes, Rhizobiaceae.
37
Introduction
39
The family Rhizobiaceae Conn 1938, belongs to the order Rhizobiales in the class
40
Alphaproteobacteria. The Rhizobiaceae comprises six genera harboring plant-associated species,
41
Rhizobium, Neorhizobium, Allorhizobium, Agrobacterium, Ensifer (syn. Sinorhizobium), Shinella,
42
and the genus Ciceribacter described for strains isolated from chickpea rhizosphere soil and
43
representing one species. This study mainly focuses on the plant-associated (either pathogenic or
44
nitrogen-fixing that were isolated from plant tissues) bacterial members of the Rhizobiaceae family.
45
The current classification in this family is mostly based on phenotypic features, DNA-DNA
46
hybridizations, and 16S rRNA gene sequences. Since the genomes of rhizobia may lose or gain
47
plasmids or genomic islands bearing genes governing catabolic capacities, performing phenotypic
48
tests, mostly for utilization of carbon and nitrogen sources, may not be that informative for rhizobial
49
taxonomy [8, 24]. DNA:DNA hybridization (DDH) is a recommended method to delineate novel
50
species; nevertheless, in practice the obtained results of the DDH are different because of variation
51
between laboratories and methods [10, 18]. The 16S rRNA gene is considered as a suitable
52
molecular marker for phylogenetic analyses. However several studies pointed out that the 16S
53
rRNA gene phylogeny may not provide a perfect snapshot of evolution of prokaryotes, since the
54
gene may be subject to horizontal gene transfer and genetic recombination at high frequency [40].
55
Moreover, a phylogeny based on 16S rRNA genes does generally not allow resolution of closely
56
related species due to the high degree of conservation of the gene. Because of the drawbacks of the
57
above systematics methods, MLSA of housekeeping genes has become more favored and
58
acceptable method to establish taxonomy of prokaryotes. It is now widely assumed that phylogeny
59
based on several well-chosen housekeeping genes, exhibiting sufficient conservation and
60
distribution in bacterial genomes, accurately reflects the phylogeny of the bacteria [20, 21].
61
Some long standing taxonomic uncertainties in the family Rhizobiaceae have been clarified by
62
MLSA results. For instance, the uncertain phylogenetic position of the former R. galegae and three
63
related species was resolved by transferring the four species, R. galegae, R. vignae, R. alkalisoli,
64
and R. huautlense to Neorhizobium gen. nov. [22]. One of the most controversal genera in the
65
Rhizobiaceae is Agrobacterium, a group of occasionally pathogenic bacteria phylogenetically close
66
to Rhizobium. On the basis of 16S rRNA gene phylogeny, Young et al. [47] transferred the
Ac ce p
te
d
M
an
us
cr
ip t
38
2 Page 2 of 15
Agrobacterium species to Rhizobium; which introduced a controversy about the validity of the
68
genus Agrobacterium [9, 47]. However, several studies concluded that Agrobacterium must be
69
retained as a genus name for the species A. radiobacter, A. rubi, “A. fabrum”, and agrobacterial
70
genomovars (G1-G13) that have not yet received a Latin binomial [4, 19]. More species currently
71
classified in Rhizobium, i.e. R. skierniewicense and R. nepotum, are candidates to join
72
Agrobacterium [30], while the former Agrobacterium rhizogenes is now definitively designated
73
Rhizobium rhizogenes [19]. Rhizobium is a heterogeneous genus accommodating two major sub-
74
clusters, around R. tropici and R. leguminosarum respectively. The genus name Rhizobium
75
encompasses 56 species; however, it is not a proper name for all, since some of them are
76
phylogenetically interspersed among members of other genera in Rhizobiaceae. For instance, R.
77
giardinii and R. oryzae did not group with Rhizobium members in several MLSA studies [20, 22].
78
The genus name Allorhizobium was designated for one species, Al. undicola [5]; further studies
79
showed that Agrobacterium vitis and R. taibaishanense clustered with A. undicola [22]. Another
80
controversial group in the Rhizobiaceae family is the genus Sinorhizobium (syn. Ensifer). The
81
genus Ensifer was created for the species E. adhaerens by Casida [2]; However, Willems et al. [42]
82
proposed that the species E. adhaerens is phylogenetically related to the larger group of
83
Sinorhizobium spp. The genus Sinorhizobium was created by Chen et al. [3] and contained 11
84
species by 2007. The phylogenetic relationships of 16S rRNA gene sequences of Ensifer adhaerens
85
and Sinorhizobium spp. were so close that it caused nomenclatural controversies [46, 42]. However,
86
according to the rules of Bacteriological Code giving priority to older name, the genus name Ensifer
87
was adopted for both Sinorhizobium and Ensifer taxa (Opinion 84 of the Judicial Commission). The
88
genus Ensifer now contains 15 species. Although the number of new species of the family
89
Rhizobiaceae have been increasing dramatically, the phylogeny of all species of the family have not
90
been simultaneously studied recently. Including more taxa of the family could be considered as a
91
solution to estimate the phylogeny of the family more robustly.
92
In order to solve the taxonomic uncertainties concerning the plant-associated members of the
93
Rhizobiaceae family, we performed MLSA of 100 strains of the family Rhizobiaceae and 16
94
rhizobial strains from other rhizobial families, using four house-keeping genes namely 16S rRNA,
95
atpD (ATP synthase F1, beta subunit), recA (recombinase A), and rpoB (RNA polymerase, beta
96
subunit).
Ac ce p
te
d
M
an
us
cr
ip t
67
97 98
Material and methods
99
3 Page 3 of 15
Bacterial samples, DNA preparation
101
We studied a total of 100 strains representing 81 species of the family Rhizobiaceae that were
102
described and validated by January 2014 in the present work (Table S1). Strains for which the
103
sequences were not available in GenBank were obtained from LMG and HAMBI culture collections.
104
The strains received from culture collections were grown in 5 ml Tryptone-Yeast Extract (TY)
105
broth at 28oC for 48 hours, and were then cultured on yeast mannitol agar (YMA) medium at 28oC
106
for 2-3 days [34]. Single colonies of the bacteria were grown in 5 ml TY broth for preservation in
107
20% glycerol–YEM broth at -80°C and DNA isolation. The UltraClean Microbial DNA Isolation
108
Kit (MO BIO Laboratories, Inc.) was used for DNA extraction from the samples. The DNA samples
109
were kept at -20oC.
110
PCR amplification and gene sequencing
111
For the MLSA approach, four housekeeping genes, 16S rRNA, atpD, recA, and rpoB, were used in
112
this
113
(http://www.ncbi.nlm.nih.gov/genbank). PCR amplification and sequencing were performed
114
according to Mousavi et al. [22]. The sequences were checked and edited using GAP4 [35], and
115
blasted
116
accession numbers of the sequences of the studied housekeeping genes are listed in Table S1.
117
Sequence analyses
118
Newly obtained nucleotide sequences, together with sequences obtained from GenBank
119
(http://www.ncbi.nlm.nih.gov/genbank), were aligned applying MUSCLE [7] software at EML-EBI
120
[11] for 16S rRNA, and ClustalW [17] in BioEdit version 7.0.5.3 [12] for atpD, recA, and rpoB
121
genes. The alignments were edited manually. The best-fit nucleotide substitution models were
122
selected according to Akaike information criterion (AIC) for each locus using MEGA5 [36]. The
123
Maximum likelihood phylogenetic trees based on the 16S rRNA sequences were constructed with
124
1,000 bootstrap (BS) replicates; and the mean distance between the groups was computed by
125
MEGA5. Bayesian analyses of the concatenated four housekeeping genes were implemented by
126
running the algorithm Metropolis coupled Markov chain Monte Carlo (MCMCMC) for twenty
127
million generations twice with MrBayes 3.2 [32]. The program Tracer v1.5 was used to analyze the
128
trace files generated by Bayesian MCMCMC runs, and the Bayesian trees were visualized by
129
FigTree v1.4.0 (http://tree.bio.ed.ac.uk/software/figtree/).
the
sequences
us
of
were
obtained
an
Most
from
GenBank
M
study.
cr
ip t
100
The
Ac ce p
te
d
(http://blast.ncbi.nlm.nih.gov/Blast.cgi?CMD=Web&PAGE_TYPE=BlastHome).
130 131
Results
132
Analyses of 16S rRNA gene sequences and individual protein-coding housekeeping genes
4 Page 4 of 15
The General Time Reversible plus gamma distribution plus invariable sites (GTR+G+I) model was
134
selected as the best-fit model of phylogenetic analysis of the sequences of the 16S rRNA gene
135
dataset. The 16S rRNA gene sequences (1279 bp) of 160 rhizobial strains (listed in Table S1),
136
including 81 species of the family Rhizobiaceae, were analyzed by Maximum likelihood phylogeny
137
(Fig S1). The taxa belonging to the genera Agrobacterium, Allorhizobium, Ciceribacter, Ensifer,
138
Neorhizobium, Ochrobactrum, Rhizobium, and Shinella stand in a moderately supported clade
139
(BS=71%). The generic clade accommodating A. radiobacter, “A. fabrum”, A. rubi, A. larrymoorei,
140
and the genomovars that have not received a Latin binomial, is strongly supported (BS=84%). The
141
Rhizobium-affiliated species do not form a monophyletic group, with several of them representing
142
separate lineages. Thirty-one plant-isolated Rhizobium species group together and form the
143
Rhizobium clade in the 16S rRNA tree, with two groups, the so-called R. tropici group, consisting
144
of eleven species and the so-called R. leguminosarum sub-cluster, with low confidence branching
145
(
S0723-2020(14)00191-X http://dx.doi.org/doi:10.1016/j.syapm.2014.12.003 SYAPM 25672
To appear in: Received date: Revised date: Accepted date:
23-9-2014 7-12-2014 12-12-2014
Please cite this article as: S.A. Mousavi, A. Willems, X. Nesme, P. de Lajudie, K. Lindstr¨om, Revised phylogeny of Rhizobiaceae: proposal of the delineation of Pararhizobium gen. nov., and thirteen new species combinations, Systematic and Applied Microbiology (2014), http://dx.doi.org/10.1016/j.syapm.2014.12.003 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
1
Revised phylogeny of Rhizobiaceae: proposal of the delineation of Pararhizobium gen. nov.,
2
and thirteen new species combinations
3
Seyed Abdollah Mousavia,e*, Anne Willemsb, Xavier Nesmec, Philippe de Lajudied, Kristina Lindströme
4 5
a
6
Biocentre 1, Viikinkaari 9, P.O. Box 56, Helsinki, Finland, FIN-00014
7 8
b
9
c
ip t
University of Helsinki, Division of Microbiology and Biotechnology, Department of Food and Environmental Sciences,
cr
Laboratory of Microbiology, Department of Biochemistry and Microbiology, Faculty of Science, Ghent University, K. L. Ledeganckstraat 35, B-9000 Gent, Belgium
University of Lyon; Université Lyon1; Ecologie Microbienne, UMR CNRS 5557 / USC INRA 1364, 16 rue R. Dubois,
F-69622 Villeurbanne cedex, France
11
d
12
Montpellier Cédex 5, France
13
e
14
00014
15
*Corresponding author. email address: [email protected] or [email protected] (Seyed Abdollah
16
MOUSAVI).
17
Department of Environmental Sciences, Viikinkaari 2a, P.O. Box 65, University of Helsinki, Helsinki, Finland, FIN-
18
00014.
19
Running title: Phylogeny of Rhizobiaceae
us
10
an
IRD, Laboratoire des Symbioses Tropicales et Méditerranéennes, Campus de Baillarguet TA A-82/J, 34398
te
d
M
University of Helsinki, Department of Environmental Sciences, Viikinkaari 2a, P.O. Box 65, Helsinki, Finland, FIN-
Ac ce p
20 21
Abstract
22
The family Rhizobiaceae accommodates the seven genera Rhizobium, Neorhizobium, Allorhizobium,
23
Agrobacterium, Ensifer (syn. Sinorhizobium), Shinella and Ciceribacter. However, several so-
24
called Rhizobium species do not exhibit robust phylogenetic positions. Rhizobium is extremely
25
heterogeneous and is in need of major revision. Therefore, a phylogenetic examination of the family
26
Rhizobiaceae by multilocus sequence analysis (MLSA) of four housekeeping genes among 100
27
strains of the family was undertaken. Based on the results we propose the delineation of the new
28
genus Pararhizobium in the Rhizobiaceae family, and thirteen new species combinations:
29
Agrobacterium nepotum comb. nov., Agrobacterium pusense comb. nov., Agrobacterium
30
skierniewicense comb. nov., Allorhizobium vitis comb. nov., Allorhizobium taibaishanense comb.
31
nov., Allorhizobium paknamense comb. nov., Allorhizobium oryzae comb. nov., Allorhizobium
32
pseudoryzae comb. nov., Allorhizobium borbori comb. nov., Pararhizobium giardinii comb. nov.,
1 Page 1 of 15
33
Pararhizobium capsulatum comb. nov., Pararhizobium herbae comb. nov., and Pararhizobium
34
sphaerophysae comb. nov.
35 36
Key words: MLSA, Housekeeping genes, Rhizobiaceae.
37
Introduction
39
The family Rhizobiaceae Conn 1938, belongs to the order Rhizobiales in the class
40
Alphaproteobacteria. The Rhizobiaceae comprises six genera harboring plant-associated species,
41
Rhizobium, Neorhizobium, Allorhizobium, Agrobacterium, Ensifer (syn. Sinorhizobium), Shinella,
42
and the genus Ciceribacter described for strains isolated from chickpea rhizosphere soil and
43
representing one species. This study mainly focuses on the plant-associated (either pathogenic or
44
nitrogen-fixing that were isolated from plant tissues) bacterial members of the Rhizobiaceae family.
45
The current classification in this family is mostly based on phenotypic features, DNA-DNA
46
hybridizations, and 16S rRNA gene sequences. Since the genomes of rhizobia may lose or gain
47
plasmids or genomic islands bearing genes governing catabolic capacities, performing phenotypic
48
tests, mostly for utilization of carbon and nitrogen sources, may not be that informative for rhizobial
49
taxonomy [8, 24]. DNA:DNA hybridization (DDH) is a recommended method to delineate novel
50
species; nevertheless, in practice the obtained results of the DDH are different because of variation
51
between laboratories and methods [10, 18]. The 16S rRNA gene is considered as a suitable
52
molecular marker for phylogenetic analyses. However several studies pointed out that the 16S
53
rRNA gene phylogeny may not provide a perfect snapshot of evolution of prokaryotes, since the
54
gene may be subject to horizontal gene transfer and genetic recombination at high frequency [40].
55
Moreover, a phylogeny based on 16S rRNA genes does generally not allow resolution of closely
56
related species due to the high degree of conservation of the gene. Because of the drawbacks of the
57
above systematics methods, MLSA of housekeeping genes has become more favored and
58
acceptable method to establish taxonomy of prokaryotes. It is now widely assumed that phylogeny
59
based on several well-chosen housekeeping genes, exhibiting sufficient conservation and
60
distribution in bacterial genomes, accurately reflects the phylogeny of the bacteria [20, 21].
61
Some long standing taxonomic uncertainties in the family Rhizobiaceae have been clarified by
62
MLSA results. For instance, the uncertain phylogenetic position of the former R. galegae and three
63
related species was resolved by transferring the four species, R. galegae, R. vignae, R. alkalisoli,
64
and R. huautlense to Neorhizobium gen. nov. [22]. One of the most controversal genera in the
65
Rhizobiaceae is Agrobacterium, a group of occasionally pathogenic bacteria phylogenetically close
66
to Rhizobium. On the basis of 16S rRNA gene phylogeny, Young et al. [47] transferred the
Ac ce p
te
d
M
an
us
cr
ip t
38
2 Page 2 of 15
Agrobacterium species to Rhizobium; which introduced a controversy about the validity of the
68
genus Agrobacterium [9, 47]. However, several studies concluded that Agrobacterium must be
69
retained as a genus name for the species A. radiobacter, A. rubi, “A. fabrum”, and agrobacterial
70
genomovars (G1-G13) that have not yet received a Latin binomial [4, 19]. More species currently
71
classified in Rhizobium, i.e. R. skierniewicense and R. nepotum, are candidates to join
72
Agrobacterium [30], while the former Agrobacterium rhizogenes is now definitively designated
73
Rhizobium rhizogenes [19]. Rhizobium is a heterogeneous genus accommodating two major sub-
74
clusters, around R. tropici and R. leguminosarum respectively. The genus name Rhizobium
75
encompasses 56 species; however, it is not a proper name for all, since some of them are
76
phylogenetically interspersed among members of other genera in Rhizobiaceae. For instance, R.
77
giardinii and R. oryzae did not group with Rhizobium members in several MLSA studies [20, 22].
78
The genus name Allorhizobium was designated for one species, Al. undicola [5]; further studies
79
showed that Agrobacterium vitis and R. taibaishanense clustered with A. undicola [22]. Another
80
controversial group in the Rhizobiaceae family is the genus Sinorhizobium (syn. Ensifer). The
81
genus Ensifer was created for the species E. adhaerens by Casida [2]; However, Willems et al. [42]
82
proposed that the species E. adhaerens is phylogenetically related to the larger group of
83
Sinorhizobium spp. The genus Sinorhizobium was created by Chen et al. [3] and contained 11
84
species by 2007. The phylogenetic relationships of 16S rRNA gene sequences of Ensifer adhaerens
85
and Sinorhizobium spp. were so close that it caused nomenclatural controversies [46, 42]. However,
86
according to the rules of Bacteriological Code giving priority to older name, the genus name Ensifer
87
was adopted for both Sinorhizobium and Ensifer taxa (Opinion 84 of the Judicial Commission). The
88
genus Ensifer now contains 15 species. Although the number of new species of the family
89
Rhizobiaceae have been increasing dramatically, the phylogeny of all species of the family have not
90
been simultaneously studied recently. Including more taxa of the family could be considered as a
91
solution to estimate the phylogeny of the family more robustly.
92
In order to solve the taxonomic uncertainties concerning the plant-associated members of the
93
Rhizobiaceae family, we performed MLSA of 100 strains of the family Rhizobiaceae and 16
94
rhizobial strains from other rhizobial families, using four house-keeping genes namely 16S rRNA,
95
atpD (ATP synthase F1, beta subunit), recA (recombinase A), and rpoB (RNA polymerase, beta
96
subunit).
Ac ce p
te
d
M
an
us
cr
ip t
67
97 98
Material and methods
99
3 Page 3 of 15
Bacterial samples, DNA preparation
101
We studied a total of 100 strains representing 81 species of the family Rhizobiaceae that were
102
described and validated by January 2014 in the present work (Table S1). Strains for which the
103
sequences were not available in GenBank were obtained from LMG and HAMBI culture collections.
104
The strains received from culture collections were grown in 5 ml Tryptone-Yeast Extract (TY)
105
broth at 28oC for 48 hours, and were then cultured on yeast mannitol agar (YMA) medium at 28oC
106
for 2-3 days [34]. Single colonies of the bacteria were grown in 5 ml TY broth for preservation in
107
20% glycerol–YEM broth at -80°C and DNA isolation. The UltraClean Microbial DNA Isolation
108
Kit (MO BIO Laboratories, Inc.) was used for DNA extraction from the samples. The DNA samples
109
were kept at -20oC.
110
PCR amplification and gene sequencing
111
For the MLSA approach, four housekeeping genes, 16S rRNA, atpD, recA, and rpoB, were used in
112
this
113
(http://www.ncbi.nlm.nih.gov/genbank). PCR amplification and sequencing were performed
114
according to Mousavi et al. [22]. The sequences were checked and edited using GAP4 [35], and
115
blasted
116
accession numbers of the sequences of the studied housekeeping genes are listed in Table S1.
117
Sequence analyses
118
Newly obtained nucleotide sequences, together with sequences obtained from GenBank
119
(http://www.ncbi.nlm.nih.gov/genbank), were aligned applying MUSCLE [7] software at EML-EBI
120
[11] for 16S rRNA, and ClustalW [17] in BioEdit version 7.0.5.3 [12] for atpD, recA, and rpoB
121
genes. The alignments were edited manually. The best-fit nucleotide substitution models were
122
selected according to Akaike information criterion (AIC) for each locus using MEGA5 [36]. The
123
Maximum likelihood phylogenetic trees based on the 16S rRNA sequences were constructed with
124
1,000 bootstrap (BS) replicates; and the mean distance between the groups was computed by
125
MEGA5. Bayesian analyses of the concatenated four housekeeping genes were implemented by
126
running the algorithm Metropolis coupled Markov chain Monte Carlo (MCMCMC) for twenty
127
million generations twice with MrBayes 3.2 [32]. The program Tracer v1.5 was used to analyze the
128
trace files generated by Bayesian MCMCMC runs, and the Bayesian trees were visualized by
129
FigTree v1.4.0 (http://tree.bio.ed.ac.uk/software/figtree/).
the
sequences
us
of
were
obtained
an
Most
from
GenBank
M
study.
cr
ip t
100
The
Ac ce p
te
d
(http://blast.ncbi.nlm.nih.gov/Blast.cgi?CMD=Web&PAGE_TYPE=BlastHome).
130 131
Results
132
Analyses of 16S rRNA gene sequences and individual protein-coding housekeeping genes
4 Page 4 of 15
The General Time Reversible plus gamma distribution plus invariable sites (GTR+G+I) model was
134
selected as the best-fit model of phylogenetic analysis of the sequences of the 16S rRNA gene
135
dataset. The 16S rRNA gene sequences (1279 bp) of 160 rhizobial strains (listed in Table S1),
136
including 81 species of the family Rhizobiaceae, were analyzed by Maximum likelihood phylogeny
137
(Fig S1). The taxa belonging to the genera Agrobacterium, Allorhizobium, Ciceribacter, Ensifer,
138
Neorhizobium, Ochrobactrum, Rhizobium, and Shinella stand in a moderately supported clade
139
(BS=71%). The generic clade accommodating A. radiobacter, “A. fabrum”, A. rubi, A. larrymoorei,
140
and the genomovars that have not received a Latin binomial, is strongly supported (BS=84%). The
141
Rhizobium-affiliated species do not form a monophyletic group, with several of them representing
142
separate lineages. Thirty-one plant-isolated Rhizobium species group together and form the
143
Rhizobium clade in the 16S rRNA tree, with two groups, the so-called R. tropici group, consisting
144
of eleven species and the so-called R. leguminosarum sub-cluster, with low confidence branching
145
(
