RESEARCH ARTICLE Horizontal gene transferand recombination shape mesorhizobial populations in the gene centerof the host plantsAstragalusluteolus and ernestii in Sichuan, China Qiongfang Li1,2, Xiaoping Zhang1, Ling Zou1, Qiang Chen1, David P. Fewer 3 & Kristina Lindstrom¨ 3

1Department of Resources and Environment, Sichuan Agricultural University, Ya’an Sichuan, China; 2College of Life Science and Engineering, Southwest University of Science and Technology, Mianyang, Sichuan, China; and 3Department of Applied Chemistry and Microbiology, University of Helsinki, Helsinki, Finland Downloaded from https://academic.oup.com/femsec/article/70/2/227/532724 by guest on 23 September 2021

Correspondence: Xiaoping Zhang, Abstract Department of Resources and Environment, Sichuan Agricultural University, Ya’an Sichuan Thirty-three rhizobial strains isolated from the root nodules of Astragalus luteolus 625014, China. Tel.: 1 86 835 2882710; and Astragalus ernestii growing on the west plateau at two different altitudes in fax: 186 835 2883166; e-mail: Sichuan province, China, were characterized by amplified rDNA restriction [email protected] analysis (ARDRA), amplified fragment length polymorphism (AFLP), and by sequencing of rrs, glnA, glnII and nifH. The ARDRA analysis revealed considerable Received 10 November 2008; revised 27 August genomic diversity. In AFLP analysis, 20 of 33 Astragalus formed three 2009; accepted 27 August 2009. distinct clades, with others dispersed into different groups with the reference Final version published online strains. Phylogenetic analysis of the rrs gene of six representative strains showed 30 September 2009. that the isolates were members of the genus . Three of the isolates

DOI:10.1111/j.1574-6941.2009.00776.x formed a sister clade to Mesorhizobium loti and Mesorhizobium ciceri, whereas the other three formed a sister clade to a clade harboring the species Mesorhizobium Editor: Philippe Lemanceau huakuii, Mesorhizobium plurifarum, Mesorhizobium septentrionale and Mesorhizo- bium amorphae, indicating the existence of two new species. Phylogenetic analysis Keywords of glnA and glnII confirmed the rrs phylogenies for four strains, but the trees were rhizobia; Astragalus; phylogeny; gene transfer. incongruent. The nifH sequences of the strains formed a monophyletic clade and were typical of those of mesorhizobia forming symbioses with inverted repeat lacking clade legume species. The incongruent phylogenies of the genes studied suggest that horizontal gene transfer and recombination shape mesorhizobial populations in the gene center of the host .

Introduction rapidly being destroyed, characterization and biogeographi- The genus Astragalus consists of up to 3000 species, which cal studies of rhizobial strains isolated from Astragalus are distributed mainly in cool to warm arid and semiarid species are urgently required. Astragalus luteolus and Astra- regions of the northern hemisphere, South America and galus ernestii are wild plants that grow at altitudes above tropical East Africa. It is especially diverse in the south- 3000 m. Astragalus luteolus are indigenous endemic plants of western and the Sino-Himalayan regions of Asia, with the western Sichuan province. Astragalus ernestii grows only approximately 1500–2000 recorded species (Sanderson & in western Sichuan, northwest of Yunnan and in eastern Wojciechowski, 2000; Osaloo et al., 2005). Astragalus species Tibet. Astragalus ernestii is also used in Chinese medicine. can form nodules and fix nitrogen in symbiosis with The classification of rhizobia into taxa is strongly based rhizobia and some representatives of this perennial legume on phylogenies of rrs genes (encoding 16S rRNA gene or family are used in China. Astragalus adsurgens is used for SSU), but phylogenies of several protein encoding house- forage and Astragalus sinicus as green manure. Astragalus keeping genes have recently been used for resolving taxa

MICROBIOLOGY ECOLOGY MICROBIOLOGY membranaceus and Astragalus mongolicus (also called especially at the species level (e.g. Gao et al., 2001; Vinuesa A. membranaceus ssp. mongolicus) are used in Chinese et al., 2005a; Martens et al., 2007, 2008; Nandasena et al., traditional medicine. Astragalus species are also important 2007; Chen et al., 2008). Multilocus sequence analysis is a bee plants. Because environments for Astragalus plants are more reliable classification method than methodology based

FEMS Microbiol Ecol 70 (2009) 227–235 c 2009 Federation of European Microbiological Societies Published by Blackwell Publishing Ltd. All rights reserved 228 Q. Li et al. on solely ribosomal sequences for a number of reasons. Table 1. Strains used in this study and results from ARDRA and rrs Firstly, several unlinked genes dispersed in the core genome sequence analysis better represent the true genealogy of the organism than just rrs sequence one single sequence or sequences from a locus that might Strain code rrs type type (species) show within strain variation (Young & Haukka, 1996). Astragalus luteolus (3650 m) Secondly, especially in rhizobia, ribosomal sequences show SCAU1 A ND mosaicism as a consequence of homologous recombination, SCAU2 A M. huakuii clade which interferes with phylogenetic tree construction (Ter- SCAU3 A ND efework et al., 1998; Van Berkum et al., 2003; Eardly et al., SCAU4 B ND SCAU5 CND 2005). Thirdly, rrs genes of rhizobia often display low SCAU6 D ND polymorphism in comparison with other taxonomic mar- SCAU7 E M. loti/ciceri clade Downloaded from https://academic.oup.com/femsec/article/70/2/227/532724 by guest on 23 September 2021 kers. They are thus often unreliable for species delineation. SCAU12 G ND This last feature is especially true for the genus Mesorhizo- SCAU13 E M. huakuii clade bium, a prevalent symbiont of Astragalus species. SCAU14 H ND Sampling of Chinese rhizobia has mainly been carried SCAU15 B ND out from the provinces Xinjiang, Gansu, Shanxi, Hubei, SCAU17 A ND SCAU18 A ND Hainan, Qinhai and Tibet, and some novel species have been SCAU20 H ND reported to date (Tan et al., 1997; Wei et al., 2003; Kan et al., SCAU21 I ND 2007; Han et al., 2008a, b; Tian et al., 2008). In this study, we SCAU22 C ND wished to obtain information about hitherto unexplored SCAU23 E ND ecosystems at high elevations on the southeast fringe of the SCAU24 D ND Qinghai-Tibet Plateau and obtain more information about SCAU26 F ND rhizobial diversity. We obtained root nodules from SCAU30 E ND SCAU31 E ND A. luteolus and A. ernestii, growing in the Aba district on SCAU32 B ND the northwest plateau of the Sichuan province in south- SCAU33 A ND western China, which is an important catchment area for Astragalus ernestii (4240 m) the Yellow River and Yangtze River. The area has a long and SCAU8 A ND cold winter and a cool and moist summer. The mean annual SCAU9 A M. huakuii clade temperature is 7.8 1C, the rainfall level is low (average SCAU10 A ND 613.9 mm per year) and the soil is barren. The plants SCAU11 F M. loti/ciceri clade SCAU16 B ND sampled grew at elevations of 3650 m (A. luteolus) and SCAU19 H ND 4250 m (A. ernestii), respectively. Thirty-three nodule iso- SCAU25 I ND lates were studied by molecular systematic methods, includ- SCAU27 A M. loti/ciceri clade ing amplified rDNA restriction analysis (ARDRA), SCAU28 A ND amplified fragment length polymorphism (AFLP) and se- SCAU29 D ND quencing of PCR products corresponding to rrs, glnA, glnII Strains used in the phylogenetic studies are in bold. and nifH genes, in order to observe diversity, phylogenetic M., Mesorhizobium; ND, not determined. patterns and speciation in these populations.

N-free plant nutrient solution (Vincent, 1970) and main- Materials and methods tained in a growth chamber at 28 1C with a photoperiod of 15 h. Nodulation was observed after 1 month. Bacterial strains and culture conditions The bacterial strains used in this work are listed in Table 1. DNA isolation Bacterial strains were isolated using the methods described Total genomic DNAs of all isolates were obtained after by Weir et al. (2004). Strains were grown on yeast extract- lysozyme–sodium dodecyl sulfate lysis, followed by phe- mannitol medium (Sy et al., 2001) at 28 1C, and culture nol–chloroform extraction and ethanol precipitation as purity was checked by repeated streaking. Strains were described (Laguerre et al., 1992). stored at À 70 1C in 20% (v/v) glycerol for long-term preservation. In order to check the capacity of the isolates ARDRA to nodulate their original hosts, surface-sterilized seeds of A. luteolus and A. ernestii were inoculated and plants were Complete 16S rRNA genes were amplified with the universal grown in pots filled with vermiculite moistened with a forward primer P1 and the universal reverse primer P6

c 2009 Federation of European Microbiological Societies FEMS Microbiol Ecol 70 (2009) 227–235 Published by Blackwell Publishing Ltd. All rights reserved Horizontal gene transfer and recombination 229

Table 2. Oligonucleotides used as PCR primers and for sequencing Target gene Primer sequenceà Positionw Size of amplified fragment Reference rrs AGAGTTTGATCCTGGCTCAG AACGAACCCT 8–37 c.1400bp Tanet al. (1997) TACGGCTACCTTGTTACGACTTCACCCC 1479–1506 glnA AAGGGCGGCTAYTTCCCGGT 532–551 c. 530 bp Turner & Young (2000) GTCGAGACCGGCCATCAGCA 1143–1124 glnII AACGCAGACAAGGAATTCG 69–88 c.600bp ATGCCCGAGCCGTTCCAGTC 686–667 nifH TACGGNAARGGSGGNATCGGCAA 25–50 c. 840 bp Laguerre et al. (2001) AGCATGTCYTCSAGYTCNTCCA 787–808 Ã

Y = C or T; H = A, C or T; R = A or G; S = C or G; K = G or T; N = A, C, G or T; I = inosine. Downloaded from https://academic.oup.com/femsec/article/70/2/227/532724 by guest on 23 September 2021 wPosition of the primers in the corresponding rrs sequence of Escherichia coli, glnA and glnII sequence of Sinorhizobium meliloti and nifH sequence of Rhizobium leguminosarum bv. trifolii.

(Tan et al., 1997). The PCR products were digested in cloned into vector pGMT (Takara). Ligated plasmids were separate reactions with the restriction endonucleases MspI, transformed into Escherichia coli DH5a. Plasmids contain- HinfI, HaeI and HindIII (Takara) and the fragments were ing glnA, glnII and nifH sequences were extracted and separated in 3.0% agarose gels. purified as described previously (Tiesman & Rizzino, 1991), and then sequenced using an ABI3730 sequencer AFLP analyses (ABI, Foster City, CA) using the primers listed in Table 2. The AFLP procedure was performed as described (Vos et al., 1995) with 100–500 ng of DNA digested with EcoRI and Nucleotide sequence accession numbers MseI restriction enzymes, and in the same step, ligated with The sequences obtained in this study have been deposited in double-stranded adapters specific for each restriction half- the GenBank sequence database under accession numbers site. Two sets of primers were used in separate PCRs: one set EU514524–EU514545. with two selective nucleotides [MseI-gc (GAT GAG TCC TGA GTA AGC) and EcoRI-gc (GAC GTC GTA CCA ATT Data treatment and sequence analysis CGA GC)] and another set with a mixture of two and three selective nucleotides [MseI-gc and EcoRI-gag (GAC GTC The ARDRA gels were scored manually, and bands were GTA CCA ATTCGA GAG)]. The PCR conditions were as marked as absent or present. Distances were calculated (sj) described by Terefework et al. (2001). The PCR products and dendrograms were constructed using the UPGMA were separated on 5% polyacrylamide gels. Electrophoresis method in NT-SYS. The AFLP gels were analyzed using the was performed in Tris–borate–EDTA buffer at 120 W and GELCOMPAR 4.0 program (Applied Maths, Kortrijk, Belgium) 50 1C for 2 h. Silver-stained polyacrylamide gels were and UPGMA dendrograms obtained based on distances scanned and analyzed. calculated with the Dice index. The gene sequences obtained were aligned in CLUSTAL_X (version 1.83) (Thompson et al., PCR amplification and nucleotide sequencing 1997) and edited using DNAMAN (version 6.0). Seven representative strains were chosen for partial DNA Phylogenetic analysis sequencing of three housekeeping core genes, rrs, glnA Ã encoding glutamine synthetase I (GSI) and glnII, encoding Phylogenetic analyses were conducted using PAUP (Swof- glutamine synthetase II (GSII). The nifH gene encoding ford, 2001). Primer sequences and ambiguous regions of the dinitrogenase reductase was chosen to represent the sym- alignments were excluded. Phylogenetic trees were inferred biotic genes. rrs gene fragments were amplified as described using the neighbor-joining method. One thousand boot- using primers P1 and P6 and an annealing temperature of strap replicates were analyzed to test the stability of mono- 56 1C(Tanet al., 1997). glnA and glnII gene fragments were phyletic groups in neighbor-joining analysis. The amplified with primer pairs GSI-1, GSI-2 and GSII-1, GSII- incongruence of the glnA and glnII genes was evaluated 2, respectively, using an annealing temperature of 55 1C using the conservative length difference test (Farris et al., Ã (Turner & Young, 2000). The nifH gene was amplified using 1994) implemented with PAUP (Swofford, 2001). We used the primers nifHI and nifHF and an annealing temperature random taxon addition (10 replicates), tree bisection–re- of 57 1C (Laguerre et al., 2001). Purified rrs PCR products connection branch swapping and heuristic searches with were directly sequenced and the other gene fragments were 10 000 repartitions of the data.

FEMS Microbiol Ecol 70 (2009) 227–235 c 2009 Federation of European Microbiological Societies Published by Blackwell Publishing Ltd. All rights reserved 230 Q. Li et al.

Results 5055 60 65 70 75 80 85 90 95 100 SCAU9 S. meliloti 102F28 Nodulation S. arbuis HAMBI1552 SCAU8 Pure cultures of all 33 isolates were observed to nodulate the M. huakuii CCBAU2609 original host plants. SCAU10 SCAU14 SCAU20 SCAU31 Genetic diversity R. leguminosarum 127K17 SCAU5 RFLP patterns were produced for rrs from the 33 Astragalus R. mongolense USDA1844 SCAU24 rhizobia (Table 1). Three of the four restriction enzymes, R. etli CFN42 S. saheli USDA4102

Msp1, HaeIII and HindIII, produced two patterns and HinfI Downloaded from https://academic.oup.com/femsec/article/70/2/227/532724 by guest on 23 September 2021 SCAU33 produced three. The tested strains were divided into eight R. leguminosarum USDA2370 rrs genotypes A-I (Table 1), revealing considerable genetic SCAU2 M. ciceri USDA3383 diversity. M. loti NZP2234 R. sullae IS123 AFLP banding patterns were obtained from all strains. An S. xinjiangensis CCBAU110 UPGMA dendrogram was constructed from the AFLP S. meliloti USDA1002 R. gallicum USDA2918 fingerprints of the 33 strains isolated from Astragalus plants Agrobacterium.sp IAM13129 and 18 reference strains (Fig. 1). Three distinct clades were SCAU19 SCAU26 formed by the novel isolates and the rest was dispersed SCAU22 among the reference strains. Neither the ARDRA nor the SCAU29 SCAU4 AFLP fingerprinting could differentiate the strains based on SCAU28 SCAU18 host plant or site of isolation. SCAU16 SCAU17 SCAU11 rrs phylogeny SCAU12 SCAU13 A 1400-bp fragment of rrs was obtained from six selected SCAU15 SCAU21 representative strains: SCAU2, SCAU7, SCAU9, SCAU11, SCAU27 SCAU13 and SCAU27 (Table 2). In the phylogenetic tree SCAU25 SCAU23 constructed, genus Mesorhizobium was included in a mono- SCAU32 SCAU6 phyletic group containing all Astragalus isolates and Amino- SCAU7 bacter species (Fig. 2). The Astragalus isolates formed two S. fredii USDA205 SCAU30 separate clades. The relationships among the defined refer- S. medicae USDA1037 ence species were similar to those in previously reported R. hainanense I66 SCAU1 trees (Wei et al., 2003). SCAU2, SCAU9 and SCAU13 SCAU3 grouped together and formed a sister clade to Mesorhizo- Fig. 1. Dendrogram generated with the GELCOMPAR 4.0 computer module bium huakuii, Mesorhizobium plurifarum, Mesorhizobium using Pearson’s coefficient clustering by the unweighted pair group septentrionale and Mesorhizobium amorphae. SCAU7, method using arithmetic averages. Eighteen reference strains and 33 SCAU11 and SCAU27 formed a sister clade to Mesorhizo- isolates are included. The scale indicates percent similarity. bium loti and Mesorhizobium ciceri. The novel isolates represent two distinct phylogenetic branches in the rrs tree. phylogenetic tree based on glnII, the genera Rhizobium and Sinorhizobium separated out as two monophyletic clades. Phylogenetic analysis of glnA and Genus Mesorhizobium was divided into two pleophyletic glnII sequences clades (Fig. 3b), with the Astragalus strains SCAU2, SCAU9 In the phylogenetic tree based on glnA gene sequences, and SCAU13 and strain MAFF303099 positioned within one Genus Mesorhizobium formed a well-separated clade of them and strains SCAU7 and SCAU27 within the other, as (Fig. 3a). Strains SCAU2 and SCAU9 clustered with Mesor- a sister clade to M. ciceri. Unfortunately, no glnII sequences hizobium sp. MAFF303099, forming a sister clade to M. were obtained for SCAU5 and SCAU11. huakuii. The sequences of strains SCAU5 and SCAU11 were The length incongruence test indicated that the two identical to that of SCAU2. Strains SCAU7, SCAU13 and datasets were incongruent with one another (P = 0.0001), SCAU27 formed a sister clade to M. ciceri and M. loti. The supporting the topological incongruence we observed. For topology of the glnA tree was different from that of the rrs the Astragalus strains, the glnII tree was congruent with the tree, with the positions of strains SCAU11 and SCAU13 rrs tree, albeit more deeply branched, the exception being swapped in relation to the other Astragalus isolates. In the the position of the strains in relation to M. loti.

c 2009 Federation of European Microbiological Societies FEMS Microbiol Ecol 70 (2009) 227–235 Published by Blackwell Publishing Ltd. All rights reserved Horizontal gene transfer and recombination 231

73 Mesorhizobium amorphae ACCC 19665T (AF041442) Mesorhizobium septentrionale SDW014T (AF508207) 71 Mesorhizobium plurifarium LMG 11892T (Y14158) Mesorhizobium huakuii IFO 15243T (D13431) Mesorhizobium sp. SCAU13 (EU514528) Mesorhizobium sp. SCAU2 (EU514529) 94 Mesorhizobium sp. SCAU9 (EU514526) T 97 Mesorhizobium mediterraneum UPM-Ca36 (L38825) Mesorhizobium temperatum SDW018T (AF508208) 68 Mesorhizobium tianshanense A-1BST (AF041447) Mesorhizobium albiziae CCBAU 61158T (DQ100066) 100 Aminobacter aminovorans DSM7048T (AJ011759) Downloaded from https://academic.oup.com/femsec/article/70/2/227/532724 by guest on 23 September 2021 Aminobacter aganoensis DSM7051T (AJ011760) 92 Aminobacter niigataensis DSM7050T (AJ011761) Mesorhizobium chacoense PR5 (AJ278249) Fig. 2. Neighbor-joining phylogenetic tree based 68 Mesorhizobium sp. SCAU27 (EU514524) on 1338 bp aligned bases of rrs sequences 86 showing the relationship of the novel isolates Mesorhizobium sp. SCAU7 (EU514525) 90 Mesorhizobium sp. SCAU11 (EU514527) with reference strains. Bootstrap values above 60 are indicated. A scale bar indicates 0.01 79 Mesorhizobium ciceri UPM-Ca7 (U07934) T substitutions per site. Strains isolated in this study 91 Mesorhizobium loti LMG 6125 (X67229) T are in bold. A number of Rhizobium and Mesorhizobium thiogangeticum SJT (AJ864462) Sinorhizobium species were used as the 100 Phyllobacterium sp. DSM 17640 (DQ431466) outgroup. However, only the Mesorhizobium/ Phyllobacterium sp. CCBAU 83356 (EF035066) Phyllobacterium clade is shown in this paper. 0.01 substitutions per site

Fig. 3. Neighbor-joining phylogenetic tree based on 506 bp aligned bases representing glutamine synthetases: (a) glnA and (b) glnII. Bootstrap values above 60 are indicated and scale bars represent 0.01 substitutions per site. The glnA and glnII trees are rooted at the midpoint. Strains isolated in this study are in bold.

Phylogenetic analysis of nifH gene sequences obtained. Sequences of the Astragalus strains were all closely related and formed a well-defined, monophyletic clade The nifH fragments amplified were approximately 750 bp in among sequences from described species of Mesorhizobium size. The nifH sequences of SCAU13 and SCAU27 were not (Fig. 4).

FEMS Microbiol Ecol 70 (2009) 227–235 c 2009 Federation of European Microbiological Societies Published by Blackwell Publishing Ltd. All rights reserved 232 Q. Li et al. Downloaded from https://academic.oup.com/femsec/article/70/2/227/532724 by guest on 23 September 2021

Fig. 4. Neighbor-joining phylogenetic tree based on 528 bp aligned bases of nifH. Only bootstrap probability values 4 50% are indicated at the branching points. Scale bars represent 0.02 substitution per site. Strains isolated in this study are in bold.

Accession numbers isolated clades might represent two new species. Mesorhizo- bium sp. MAFF303099, whose genome has been fully The GenBank accession numbers for the rrs, glnA, glnII and sequenced and that is still erroneously called M. loti (Turner nifH gene sequences reported in this paper are et al., 2002), is part of the clade of Astragalus isolates. This EU514530–EU514546. confirmed that the Astragalus isolates belonged to the Mesorhizobium genus. Discussion The bacterial species has been defined in numerous ways. We conducted a molecular systematic study of 33 rhizobia Considering the biology of the and species concepts isolated from two different Astragalus species, A. luteolus in general, an evolutionary species concept seems desirable. and A. ernestii, growing at two different elevations in the This primary concept should be followed by secondary highlands situated on the southeast fringe of the Qinghai- concepts that guide the taxonomist in the classification work Tibet Plateau in Sichuan, China. (Lindstrom¨ & Gyllenberg, 2006). So far, DNA : DNA hybridi- AFLP fingerprinting revealed that the isolates were zations have been considered the gold standard in diverse, but belonged to the same population, because the and rrs phylogeny desirable for species delineation. Consider- genomic profiles for isolates from the two plants or sites ing the non conserved nature of especially mesorhizobial rrs were intermingled. The AFLP dendrogram does not indicate (de Lajudie et al., 1998), the species affiliation of the isolates the actual relationship between the recognized strains should not be resolved based on the rrs phylogeny alone. Also because it fails to resolve Rhizobium from Sinorhizobium taking the disadvantages of DNA : DNA hybridizations into and Mesorhizobium. Terefework et al. (2001) proposed that account (Lindstrom¨ & Martınez-Romero,´ 2007), it becomes AFLP is an excellent technique to differentiate strains or very important to determine the phylogenies of several house- closely related species, but it is by no means a good keeping genes for rhizobial species assignments. Glutamine phylogenetic tool. Our results support this view. synthetase exists in at least two related forms, GSI and GSII, ARDRA analysis showed that the Astragalus isolates were whose sequences were used in evolutionary molecular clock diverse, but belonged to the same rrs clade. Sequencing of rrs studies to analyze the phylogenies of 16 species of rhizobia by of representative strains confirmed that the strains belong to Turner & Young (2000). Their phylogenetic studies indicated the Mesorhizobium genus and indicated that the novel that glnA was congruent with rrs. glnII followed the same

c 2009 Federation of European Microbiological Societies FEMS Microbiol Ecol 70 (2009) 227–235 Published by Blackwell Publishing Ltd. All rights reserved Horizontal gene transfer and recombination 233 pattern, but M. huakuii showed an aberrant position in the close relatives in the IRLC clade, is Eurasia and specifically protein phylogeny. Our study confirmed the findings of the steppes and mountains of southwestern to south-central Turner & Young (2000) with respect to the general shapes of Asia and the Himalayan plateau. Sampling of rhizobia in the the trees obtained with named species. The better resolution center of diversity of Rhizobium galegae biovar orientalis in of these sequences in comparison with rrs suggests that the the Caucasus yielded a collection with greater diversity than glutamine synthetase genes are good phylogenetic markers for a collection of R. galegae biovar officinalis from the same rhizobial systematics. area (Andronov et al., 2003). Bailly et al. (2007) stressed that The positions of the new isolates from this study, how- empirical studies focusing on the interplay between ecologi- ever, showed some aberrant features in the glnA tree. Two of cal adaptation and sex in bacteria are needed to gain better the strains (SCAU2, SCAU9) that grouped as a putative new knowledge of the processes leading to bacterial diversifica-

species in the rrs tree maintained their position in both the tion. Vinuesa et al. (2005a) stressed the importance of Downloaded from https://academic.oup.com/femsec/article/70/2/227/532724 by guest on 23 September 2021 glnA and the glnII trees, whereas the third member of this rrs sufficiently dense sampling of strains when achieving reli- clade, SCAU13, had a different glnA gene, more closely able gene phylogenies for taxonomic purposes. They related to that of strains SCAU7 and SCAU27, which form showed, using concatenated atpD-recA sequences, that three the second putative new species. Thus, glnA seems to have named Rhizobium species, Rhizobium gallicum, Rhizobium been involved in recombinational events and is not a mongolense and Rhizobium Yanglingense, should in fact be suitable phylogenetic marker for genus Mesorhizobium. considered as belonging to the same species. On the other The work by Gao et al. (2001, 2004) with isolates from hand, Vinuesa et al. (2005b) demonstrated that strains A. adsurgens and that of Zhang et al. (2000) with A. sinicus belonging to the new species Bradyrhizobium canariense strains demonstrated that populations from Astragalus could be separated from Bradyrhizobium japonicum using a species are genetically diverse, but mainly belong to the population genetics approach. Mesorhizobium genus. The phylogenies of glnA sequences Several new mesorhizobial species were described re- obtained for A. adsurgens isolates were congruent with those cently, but unfortunately, they were not always based on of rrs, and two new species, Mesorhizobium temperatum and extensive sampling of bacteria and analysis of sufficient M. septentrionale, were described. The results indicated that housekeeping genes. Thus, the true nature of the population the populations representing the species were not under- genetics of mesorhizobia is still not obvious. Our results going genetic mixing for these housekeeping genes. In this indicate that horizontal gene transfer and recombination study, we show that if traditional, minimum criteria were to shape mesorhizobial populations in the gene center of the be used, two clades of our Astragalus rhizobia in the rrs tree host plants A. luteolus and A. ernestii. We propose that wider would be potentially new species. However, both the AFLP sampling of rhizobial populations in the gene centers is fingerprinting results and the incongruence of the rrs and GS likely to reveal interesting evolutionary patterns leading to gene trees indicate that we are dealing with samples from a speciation of symbionts nodulating the large Astragalus population that is undergoing genetic exchange and recom- plant family. bination at a core genomic level. Thus, species delineation is not advisable at this stage. Symbiotic rhizobial genes are commonly found on po- Acknowledgements tentially transmissible elements (MacLean et al., 2007; This research was supported by the National Natural Science Nandasena et al., 2007). Gao et al. (2004) sequenced Foundation of China, the Specialized Research Found for symbiotic nodA from A. adsurgens rhizobia belonging to the Doctoral Program of Higher Education (20060626006) the species M. septentrionale and M. temperatum and found and the Academy of Finland. that most of them were closely related, indicating horizontal gene transfer during evolution. We sequenced the symbiotic nifH encoding one nitrogenase subunit and showed that References isolates from the two Astragalus species under study were Andronov EE, Terefework Z, Roumiantseva ML, Dzyubenko NI, phylogenetically close. They were also part of a large clade Onichtchouk OP, Kurchak ON, Dresler-Nurmi A, Young JP, harboring nifH sequences of bacteria nodulating plant Simarov BV & Lindstrom¨ K (2003) Symbiotic and genetic species that are members of the inverted repeat lacking clade diversity of Rhizobium galegae isolates collected from the (IRLC) of the Hologalegina in the Papillionideae (Wojcie- orientalis gene center in the Caucasus. Appl Environ chowski et al., 2004). The bootstrap values of the branches of Microb 69: 1067–1074. this larger part of the mesorhizobial nifH tree are, however, Bailly X, Olivieri I, Brunel B, Cleyet-Marel JC & Bena´ G (2007) low. Horizontal gene transfer and homologous recombination According to Polhill (1981a, b), the geographic center of drive the evolution of the nitrogen-fixing symbionts of diversity and presumed origin of Astragalus, like most of its Medicago species. J Bacteriol 189: 5223–5236.

FEMS Microbiol Ecol 70 (2009) 227–235 c 2009 Federation of European Microbiological Societies Published by Blackwell Publishing Ltd. All rights reserved 234 Q. Li et al.

Chen WF, Guan SH, Zhao CT, Yan XR, Man CX, Wang ET & Martens M, Dawyndt P, Coopman R, Gillis M, De Vos P & Chen WX (2008) Different Mesorhizobium species associated Willems A (2008) Advantages of multilocus sequence analysis with Caragana carry similar symbiotic genes and have for taxonomic studies: a case study using 10 housekeeping common host ranges. FEMS Microbiol Lett 283: 203–209. genes in the genus Ensifer (including former Sinorhizobium). de Lajudie P, Willems A, Nick G et al. (1998) Characterization of Int J Syst Evol Micr 58: 200–214. tropical tree rhizobia and description of Mesorhizobium Nandasena KG, O’Hara GW, Tiwari RP, Willlems A & Howieson plurifarium sp. nov. Int J Syst Bacteriol 48: 369–382. JG (2007) Mesorhizobium ciceri biovar biserrulae, a novel Eardly BD, Nour SM, Van Berkum P & Selander RK (2005) biovar nodulating the pasture legume pelecinus L. Rhizobial 16S rRNA and dnaK genes: mosaicism and the Int J Syst Evol Micr 57: 1041–1045. uncertain phylogenetic placement of Rhizobium galegae. Appl Osaloo SK, Maassoumi AA & Murakami N (2005) Molecular Environ Microb 71: 1328–1335. systematics of the Old World Astragalus () as inferred Farris JS, Kallersj¨ o¨ M, Kluge AG & Bult C (1994) Testing from nrDNA ITS sequence data. Brittonia 57: 367–381. Downloaded from https://academic.oup.com/femsec/article/70/2/227/532724 by guest on 23 September 2021 significance of congruence. Cladistics 10: 315–319. Polhill RM (1981a) Papilionoideae. Advances in Legume Gao JL, Terefework Z, Chen WX et al. (2001) Genetic diversity of Systematics. Part 1 (Polhill R.M. & Raven P.H., eds), rhizobia isolated from Astragalus adsurgens growing in pp. 191–208. Royal Botanic Gardens, Kew. different geographical regions of China. J Biotechnol 91: Polhill RM (1981b) Galegeae. Advances in Legume Systematics. 155–168. Part 1 (Polhill RM & Raven PH, eds), pp. 357–363. Royal Gao JL, Turner SL, Kan FL et al. (2004) Mesorhizobium Botanic Gardens, Kew. septentrionale sp. nov. and Mesorhizobium temperatum sp. Sanderson MJ & Wojciechowski MF (2000) Improved bootstrap nov., isolated from Astragalus adsurgens growing in the confidence limits in large-scale phylogenies, with an example northern regions of China. Int J Syst Evol Micr 54: 2003–2012. from Neo-Astragalus (Leguminosae). Syst Biol 49: 671–685. Han TX, Wang ET, Wu LJ, Chen WF, Gu JG, Gu CT, Tian CF & Swofford DL (2001) PAUPÃ: Phylogenetic Analysis Using Chen WX (2008a) Rhizobium multihospitium sp. nov., isolated Parsimony (Ã and Other Methods), Version 4.0b8. Sinauer, from multiple legume species native of Xinjiang, China. Int J Sunderland, MA. Syst Evol Micr 58: 1693–1699. Sy A, Giraud E, Jourand P et al. (2001) Methylotrophic Han TX, Han LL, Wu LJ, Chen WF, Sui XH, Gu JG, Wang ET & Methylobacterium bacteria nodulate and fix nitrogen in Chen WX (2008b) Mesorhizobium gobiense sp. nov. and symbiosis with legumes. J Bacteriol 183: 214–220. Mesorhizobium tarimense sp. nov., isolated from wild legumes Tan ZY, Xu XD, Wang ET, Gao JL, Martinez-Romero E & Chen growing in desert soils of Xinjiang, China. Int J Syst Evol Micr WX (1997) Phylogenetic and genetic relationships of 58: 2610–2618. Mesorhizobium tianshanense and related rhizobia. Int J Syst Kan FL, Chen ZY, Wang ET, Tian CF, Sui XH & Chen WX (2007) Bacteriol 47: 874–879. Characterization of symbiotic and endophytic bacteria Terefework Z, Nick G, Suomalainen S, Paulin L & Lindstrom¨ K isolated from root nodules of herbaceous legumes grown in (1998) Phylogeny of Rhizobium galegae with respect to other Qinghai-Tibet plateau and in other zones of China. Arch rhizobia and agrobacteria. Int J Syst Bacteriol 48: 349–356. Microbiol 188: 103–115. Terefework Z, Kaijalainen S & Lindstrom¨ K (2001) AFLP Laguerre G, Masurier SI & Amarger N (1992) Plasmid profiles fingerprinting as a tool to study the genetic diversity of and restriction fragment length polymorphism of Rhizobium Rhizobium galegae isolated from Galega orientalis and Galega leguminosarum bv. viceae in field populations. FEMS Microbiol officinalis. J Biotechnol 91: 169–180. Ecol 101: 17–26. Thompson JD, Gibson TJ, Plewniak F, Jeanmougin F & Higgins Laguerre G, Nour SM, Macheret V, Sanjuan J, Drouin P & DG (1997) The Clustal X windows interface: flexible strategies Amarger N (2001) Classification of rhizobia based on nodC for multiple sequence alignment aided by quality analysis and nifH gene analysis reveals a close phylogenetic relationship tools. Nucleic Acids Res 25: 4876–4882. among Phaseolus vulgaris symbionts. Microbiol 147: 981–993. Tian CF, Wang ET, Wu LJ, Han TX, Chen WF, Gu CT, Gu JG & Lindstrom¨ K & Gyllenberg HG (2006) A cross-disciplinary Chen WX (2008) Rhizobium fabae sp. nov., a bacterium that species concept. WFCC Newsletter, July 2006, 14–28. nodulates Vicia faba. Int J Syst Evol Micr 58: 2871–2875. Lindstrom¨ K & Martınez-Romero´ E (2007) International Tiesman J & Rizzino A (1991) A rapid and reliable method for the committee on systematics of prokaryotes, subcommittee on purification of high-quality plasmid DNA for double-stranded the taxonomy of Agrobacterium and Rhizobium. Int J Syst Evol sequencing. Biotechniques 10: 326–328. Micr 57: 1365–1366. Turner SL & Young JP (2000) The glutamine synthetases of MacLean AM, Finan TM & Sadowsky MJ (2007) Genomes of the rhizobia: phylogenetics and evolutionary implications. Mol symbiotic nitrogen-fixing bacteria of legumes. Plant Physiol Biol 17: 309–319. 144: 615–622. Turner SL, Zhang XX, Li FD & Young JP (2002) What does a Martens M, Delaere M, Coopman R, De Vos P, Gillis M & bacterial genome sequence represent? Misassignment of MAFF Willems A (2007) Multilocus sequence analysis of Ensifer and 303099 to the genospecies Mesorhizobium loti. Microbiology related taxa. Int J Syst Evol Micr 57: 489–503. 148: 3330–3331.

c 2009 Federation of European Microbiological Societies FEMS Microbiol Ecol 70 (2009) 227–235 Published by Blackwell Publishing Ltd. All rights reserved Horizontal gene transfer and recombination 235

Van Berkum P, Terefework Z, Paulin L, Suomalainen S, Wei GH, Tan ZY, Zhu ME, Wang ET, Han SZ & Chen WX (2003) Lindstrom¨ K & Eardly BD (2003) Discordant phylogenies Characterization of rhizobia isolated from legume species within the rrn loci of Rhizobia. J Bacteriol 185: 2988–2998. within the genera Astragalus and Lespedeza grown in the Loess Vincent JM (1970) A Manual for the Practical Study of Root Plateau of China and description of Rhizobium loessense sp. Nodule Bacteria. Blackwell Scientific, Oxford. nov. Int J Syst Evol Micr 53: 1575–1583. Vinuesa P, Silva C, Lorite MJ, Izaguirre-Mayoral ML, Bedmar EJ Weir BS, Turner SJ, Silvester WB, Park DC & Young JM (2004) & Martınez-´ Romero E (2005a) Molecular systematics of Unexpectedly diverse Mesorhizobium strains and Rhizobium rhizobia based on maximum likelihood and Bayesian leguminosarum nodulate native legume genera of New phylogenies inferred from rrs, atpD, recA and nifH sequences, Zealand, while introduced legume weeds are nodulated and their use in the classification of Sesbania microsymbionts by Bradyrhizobium species. Appl Environ Microb 70: from Venezuelan wetlands. Syst Appl Microbiol 28: 702–716. 5980–5987. Vinuesa P, Leon-Barrios´ M, Silva C, Willems A, Jarabo-Lorenzo Wojciechowski MF, Lavin M & Sanderson MJ (2004) A phylogeny Downloaded from https://academic.oup.com/femsec/article/70/2/227/532724 by guest on 23 September 2021 A, Perez-Galdona´ R, Werner D & Martınez-Romero´ E (2005b) of legumes (Leguminosae) based on analysis of the plastid Bradyrhizobium canariense sp.nov., an acid-tolerant matK gene resolves many well-supported subclades within the endosymbiont that nodulates endemic genistoid legumes family. Am J Bot 91: 1846–1862. (Papilionoideae: Genisteae) growing in the Canary Islands, Young JP & Haukka KE (1996) Diversity and phylogeny of along with B. japonicum bv. genistearum, Bradyrhizobium rhizobia. New Phytol 133: 87–94. genospecies a and Bradyrhizobium genospecies b. Int J Syst Zhang XX, Turner SL, Guo XW, Yang HJ, Debelle´ F, Yang GP, Evol Micr 55: 569–575. Denari´ e´ J, Young JP & Li FD (2000) The common nodulation Vos PR, Hogers M, Bleeker M et al. (1995) AFLP: a new technique genes of Astragalus sinicus Rhizobia are conserved despite for DNA fingerprinting. Nucleic Acids Res 23: 4407–4414. chromosomal diversity. Appl Environ Microb 66: 2988–2995.

FEMS Microbiol Ecol 70 (2009) 227–235 c 2009 Federation of European Microbiological Societies Published by Blackwell Publishing Ltd. All rights reserved