The Common Nodulation Genes of Astragalus sinicus Rhizobia Are Conserved despite Chromosomal Diversity

Abstract

The nodulation genes of Mesorhizobium sp. (Astragalus sinicus) strain 7653R were cloned by functional complementation of Sinorhizobium meliloti nod mutants. The common nod genes, nodD, nodA, and nodBC, were identified by heterologous hybridization and sequence analysis. The nodA gene was found to be separated from nodBC by approximately 22 kb and was divergently transcribed. The 2.0-kb nodDBC region was amplified by PCR from 24 rhizobial strains nodulating A. sinicus, which represented different chromosomal genotypes and geographic origins. No polymorphism was found in the size of PCR products, suggesting that the separation of nodA from nodBC is a common feature of A. sinicus rhizobia. Sequence analysis of the PCR-amplified nodA gene indicated that seven strains representing different 16S and 23S ribosomal DNA genotypes had identical nodA sequences. These data indicate that, whereas microsymbionts of A. sinicus exhibit chromosomal diversity, their nodulation genes are conserved, supporting the hypothesis of horizontal transfer of nod genes among diverse recipient bacteria.

Rhizobia are soil bacteria that can form nodules, in which they fix nitrogen, on leguminous plants in a host-specific manner. Nodulation (nod) genes have been identified that control the specific infection and nodulation of the plant hosts. The initial infection event is regulated by a NodD protein or proteins which activate the transcription of other nod genes in the presence of host-produced flavonoids (12, 25, 37). The nodABC genes are called common nod genes because they are present in all rhizobia. Other nod genes, such as nodFE, nodH, nodSU, and nodZ (12, 25, 37), are present in various combinations in rhizobial species and are called host-specific nod genes.

Expression of common and host-specific nod genes results in the production of lipochitooligosaccharides (Nod factors) that act as morphogenic signal molecules on specific legume hosts (12, 37). All Nod factors have a β-1,4-linked N-acetyl glucosamine oligosaccharide backbone ranging in length from 3 to 5 residues and substituted for by an N-acyl chain at the nonreducing end and other chemical groups on the glucosamine residues. The common nodABC gene products are involved in the synthesis of the N-acylated oligosaccharide core, while the host-specific nod gene products are involved in the decoration of this backbone with substitutions that confer plant specificity. The nodABC genes encode an acyltransferase, a chitin oligosaccharide deacetylase, and a chitin oligosaccharide synthase, respectively (3, 33). The common nod genes are also involved in determining host range specificity to some extent. For example, different NodA proteins recognize and transfer different fatty acid chains to the chitooligosaccharide chain, the length of which is determined by NodC (11, 27, 32). The common nodABC genes are essential for nodule formation. Mutation in any of them abolishes the ability to produce Nod factors and results in a nonnodulating (Nod⁻) phenotype (12).

Astragalus sinicus L. (Chinese milk vetch) is an important winter-growing green manure, traditionally grown in the rice fields of southern China, Japan, and Korea. A. sinicus is a very specific host and usually forms nodules only with rhizobia isolated from itself (4), the only reported case of cross-inoculation being with a rhizobial strain isolated from Astragalus ciceri (24). Chen et al. (5) undertook a taxonomic study of nine A. sinicus isolates and proposed a new species, Rhizobium huakuii, for the rhizobia isolated from this host. On the basis of 16S ribosomal DNA (rDNA) sequence data and other taxonomic criteria, Jarvis et al. (21) proposed a new genus, Mesorhizobium, to which R. huakuii was transferred. Our previous work has shown that A. sinicus rhizobia are diverse (17, 42): of 204 strains analyzed, all are Mesorhizobium, but they belong to four different 16S rDNA genotypes. These four genotypes can be subdivided into seven genotypes when 16S and 23S rDNA data are combined. Mesorhizobium sp. (A. sinicus) strain 7653R belongs to the dominant 16S rDNA genotype, genotype 3, which is different from that of the M. huakuii type strain, CCBAU2609. Strain 7653R contains two large plasmids, and the nod genes are carried by the larger plasmid (43). The structures of the Nod factors produced by three Mesorhizobium sp. (A. sinicus) strains, including 7653R, have been determined: the three strains produce identical pentameric lipochitooligosaccharides that are O sulfated and partially N glycolylated at the reducing end and N acylated at the nonreducing end by a C_18:4 fatty acid (40). The aims of this study were to isolate and characterize the nod genes of Mesorhizobium sp. (A. sinicus) strain 7653R, to compare the common nod genes with those from other species of rhizobia, and to look at nod gene variation among A. sinicus nodule isolates representing the different chromosome types (17, 42).

MATERIALS AND METHODS

Bacterial strains, plasmids, and growth conditions.

Rhizobia were grown at 28°C on either tryptone-yeast (TY) medium (tryptone, 5 g/liter; yeast extract, 3 g/liter; CaCl₂ · 2H₂O, 0.87 g/liter) or yeast extract-mannitol (YEM) medium (39). Escherichia coli strains were grown at 37°C on Luria-Bertani (LB) medium (28). Where appropriate, antibiotics were added to the following concentrations (micrograms per milliliter): streptomycin, 200; ampicillin and kanamycin, 100; phleomycin, 20; and tetracycline, 10. The rhizobial strains from A. sinicus are described in Table 1. The other rhizobial strains and plasmids used in this study are shown in Table 2.

TABLE 1.

Strains and genotypes of A. sinicus rhizobia used in this study

Strain	Geographic origin (China)	16S rDNA typea	16S–23S rDNA IGS type (23S rDNA genotype)b
nodA sequencing
ZJ15B7	Deqin, Zhejing	a3	9 (1)
HB5A4	Wuhan, Hubei	c7	? (10)
HN14A16	Nan, Hunan	c8	? (10)
HN8A1	Henshan, Hunan	b4	5 (1)
JX2B5	Nanchang, Jiangxiu	a3	8 (5)
JS5A15	Wujin, Jiangsu	a3	1 (3)
HN15B23	Nan, Hunan	c7	? (11)
nodDC intergenic region PCR
7653R	Nanjing, Jiangsu	a3	3
ZJ9A16	Nanxi, Zhejiang	a	2
ZJ10B11	Nanxi, Zhejiang	a	3
ZJ13A7	Shaoxin, Zhejiang	b	7
ZJ13A4	Shaoxin, Zhejiang	b	7
ZJ13B16	Shaoxin, Zhejiang	c	7
GX8A7	Guilin, Guangxi	a	3
GX8B23	Guilin, Guangxi	a	3
HB7A6	Wuhan, Hubei	d	3
HB7B4	Wuhan, Hubei	a	6
JS6A16	Wujin, Jiangsu	a	10
JS6A11	Wujin, Jiangsu	a	10
HN2A1	Changsha, Hunan	a	9
HN6A4	Henshan, Hunan	b	5
HN6B13	Henshan, Hunan	b	5
HN8A3	Henshan, Hunan	a	3
HN15A6	Nan, Hunan	c	?
HN15B23	Nan, Hunan	c	?
HN15B4	Nan, Hunan	a	3
JX5B2	Yujiang, Jiangxi	c	?
JX6A2	Xinyu, Jiangxi	d	?
JX6A4	Xinyu, Jiangxi	a	7
JX6A18	Xinyu, Jiangxi	b	5
JX6A19	Xinyu, Jiangxi	b	7
JX6B8	Xinyu, Jiangxi	d	?

^aLetters refer to the 16S rDNA genotype obtained by RFLP analysis with four enzymes, and numbers refer to the 16S rDNA genotype obtained with nine enzymes (42). 

^bQuestion marks indicate that more than one band was produced in the 16S–23S rDNA IGS PCR amplification (17). The number in parentheses is the 23S rDNA genotype (42). 

TABLE 2.

Bacterial strains and plasmids used in this study

Strain or plasmid	Relevant characteristics	Sourcea or reference
Strains
Sinorhizobium meliloti
Rm2011 = SU47	Wild type, Nod+ on Medicago sativa	GMI
Rm1021	Strr derivative of Rm2011	GMI
GMI5382	Rm2011 nodA::Tn5 Strr Phr Nmr	10
GMI5383	Rm2011 nodB::Tn5 Strr Phr Nmr	10
GMI5387	Rm2011 nodC::Tn5 Strr Phr Nmr	10
GMI6227	Rm1021 nodD1::Tn5 nodD2::tm nodD3::sp/g-1	20
GMI5619	Rm2011 Δ(nodH)8 Nmr Phr	9
GMI6348	Rm1021 nodQ1::Tn5-233 nodQ2::Tn5	30
GMI6628	Rm2011 Δ(nodF)13 nodL::Tn5	2
Mesorhizobium (type strains)
M. loti	LMG6125, Nod+ on Lotus tenuis	LMG
M. ciceri	USDA3383, Nod+ on Cicer arietinum	USDA
M. tianshanense	USDA3592, Nod+ on Glycyrrhiza pallidiflora	USDA
M. mediterraneum	USDA3392, Nod+ on Cicer arietinum	USDA
M. huakuii	CCBAU2609, Nod+ on Astragalus sinicus	5
M. plurifarium	HAMBI193, Nod+ on Acacia senegal	HAMBI
Plasmids
pLAFR3	IncP, cosmid derived from pRK290, Tcr	33
pRK2013	Helper plasmid, Kmr	13
pBluescript KS+	Sequencing vector, Apr	28
pHN50	pLAFR3 containing a 30-kb nod gene region Mesorhizobium sp. strain 7653R	This work
pHN71	pLAFR3 containing a 25-kb nod gene region of Mesorhizobium sp. strain 7653R	This work
pHN72	pLAFR3 containing a 21-kb nod gene region of Mesorhizobium sp. strain 7653R	This work
pHN51	4.2-kb EcoRI fragment of pHN50 cloned in pBluescript KS+	This work
pHN61b	2.8-kb EcoRI fragment of pHN72 cloned in pBluescript KS+	This work
pHN64	10-kb HindIII fragment of pHN50 cloned in pBluescript KS+	This work
pRmSL42	pBR325 carrying the common nod genes of S. meliloti	15
pPN25	pLAFR1 carrying 7.1-kb nod genes of M. loti	31
pGMI174	pBR322 derivative carrying S. meliloti 2011 nodFEGH	8

^aUSDA, U.S. Department of Agriculture, Beltsville, Md.; HAMBI, University of Helsinki, Helsinki, Finland; GMI, LBMRPM INRA-CNRS, Castanet-Tolosan, France; LMG, Collection of Bacteria of Laboratorium voor Microbiologie, University of Gent, Ghent, Belgium; CCBAU, Culture Collection of Beijing Agricultural University, Beijing, China. 

Construction of a cosmid gene library.

Total DNA of Mesorhizobium sp. (A. sinicus) strain 7653R was prepared according to the method of Ma et al. (23) and partially digested with the restriction enzyme EcoRI. Fragments of 20 to 30 kb were isolated by centrifugation (25,000 rpm on a Beckman SW28 rotor; 20°C, 24 h) through a sucrose density gradient (10 to 50% [wt/vol]), dialyzed to remove the sucrose, and then concentrated to a suitable volume by precipitation with 2 volumes of ethanol by resuspension in distilled water. DNA was then ligated with a sevenfold excess of EcoRI-digested and phosphatase-treated cosmid vector pLAFR3 (33). The ligated DNA was packaged into lambda bacteriophage particles and introduced into E. coli LE392 by transfection (28). Three thousand single colonies were picked and stored separately in 20% glycerol at −70°C by using 96-well plates. The colonies were mixed before use.

Conjugation and plant experiment.

The cosmid gene library was introduced into Sinorhizobium meliloti mutant strains through triparental mating with the help of plasmid pRK2013 (13); crosses were carried out on TY medium by the method described by Christensen and Schubert (6). Medicago sativa seeds and nodules were surface sterilized by soaking in 95% ethanol for 5 min and in 0.2% acidified HgCl₂ for 3 min and then being rinsed 10 times in sterile water. The surface-sterilized seeds were germinated at 22°C in darkness overnight, and the seedlings were aseptically grown in test tubes on Jensen nitrogen-free agar slants (9). Two seedlings were put into each tube.

DNA manipulations and sequence analyses.

Southern blots were carried out on Hybond-N⁺ membranes (Amersham) according to the manufacturer's instructions. The nod gene probes were labeled with the Promega Multiprimer Labeling kit. The nodD and nodC gene probes were retrieved from pRmSL42, which contains the nod genes of S. meliloti (15). The nodA gene of Mesorhizobium loti cloned in plasmid pPN25 (31) was used for nodA gene hybridization. A 0.6-kb SalI fragment from pGMI174, internal to the S. meliloti 2011 nodE, was used as a nodE probe (8). Sequencing of the common nod genes of Mesorhizobium sp. strain 7653R was achieved by further subcloning the 4.2-kb EcoRI and 10-kb HindIII fragments as several smaller fragments (about 1 to 2 kb) into the sequencing vector pBluescript KS+ (28). Sequencing reactions were performed by Sanger's dideoxy chain termination method, and the extension products were separated and detected on an Automated Laser Fluorescent DNA Sequencer (Pharmacia). Primers were synthesized for sequencing in order to get the whole sequences of both strands.

The DNA sequences were analyzed by using the GCG software version 7.1 (Genetics Computer Group, University of Wisconsin, Madison). The sequences used for phylogenetic studies were first aligned by using PILEUP, and phylogenetic trees were constructed by using ClustalW (36), which uses the neighbor-joining algorithm with the K2P distance correction. The resultant trees were displayed with TreeView (26). The nodABC gene sequences of the following species or strains used in phylogenetic analyses were obtained from the EMBL database (accession numbers in parentheses): Azorhizobium caulinodans ORS571 (L18897), Rhizobium leguminosarum bv. viciae (Y00548), Rhizobium galegae (X87578), S. meliloti Rm1021 (M11268), S. meliloti Rm41 (X01649), Rhizobium etli CE-3 (M58625 and M58626), Sinorhizobium sp. strain NGR234 (X73362), Sinorhizobium fredii USDA257 (M73699), M. loti NZP2037 (X52958), Mesorhizobium sp. strain N33 (U53327), and Rhizobium tropici CFN299 (X98514).

PCR amplifications.

The nodD-to-nodC region of A. sinicus rhizobia was amplified by using primers DC1 (5′-GTA CAG GAG GGC ATC GCG AA-3′) and DC2 (5′-CTG CAG CTG CAG CGA ATC TG-3′). Primer DC1 corresponds to positions 1244 to 1225 in the Mesorhizobium sp. strain 7653R nodDBC sequence (in nodD), and primer DC2 corresponds to positions 3189 to 3170 (in the nodC gene). The PCR was performed in a 60-μl volume containing 1× reaction buffer (10 mM Tris-HCl, 50 mM KCl, 0.1% Triton X-100), 0.5 μM each primer, 1.25 mM MgCl₂, 0.2 mM deoxynucleoside triphosphates (dNTP), 30 ng of template DNA, and 2 U of Taq polymerase (Sangon). The following temperature profile was used: initial denaturation at 94°C for 3 min; 30 cycles of 94°C for 1 min, 56°C for 1 min, and 72°C for 2 min; and final extension at 72°C for 10 min.

Amplification of nodA sequences with nodA-1 and nodA-2 used the protocol of Haukka et al. (18). Primers nodA-3 (5′-TCA TAG CTC YGR ACC GTT CCG-3′) and nodA-4 (5′-ATC ATC KYN CCG GNN GGC CA-3′), corresponding to positions 980 to 960 and 955 to 936 of the Mesorhizobium sp. strain 7653R nodA sequence, were designed with this sequence and those present in the databases. Where necessary, these primers were used in a nested PCR, involving two rounds of amplification, with the product of the first amplification, with primers nodA-1 and nodA-3, being used as a template for the second amplification, with primers nodA-1 and nodA-4. The same amplification conditions were used for each primer pair in 50-μl volumes containing 1× reaction buffer, 50 pmol of each primer, 1.63 mM MgCl₂, 0.2 mM dNTP, and 1 U of Taq polymerase (Promega). The following temperature profile was used: initial denaturation at 97°C for 2 min; 25 cycles of 92°C for 40 s, 55°C for 1 min, and 72°C for 1.5 min; and final extension at 72°C for 5 min. Amplified nodA sequences were cleaned by using the PCR Purification kit (Qiagen) and eluted in 30 μl of H₂O before sequencing with the Dye Terminator Ready Reaction kit (Perkin-Elmer) and an ABI 377 automated sequencer.

Nucleotide sequence accession number.

The DNA sequences of nodDBC and nodA are available in the EMBL database under accession no. AJ249393 and AJ249353, respectively. The nodA sequences are available in the EMBL database under accession no. AJ25040 to AJ25042.

RESULTS

Cloning of common and specific nod genes from Mesorhizobium sp. (A. sinicus) strain 7653R by functional complementation.

Nod factors produced by M. huakuii strains have been shown to be similar to those from S. meliloti (40): both are O sulfated at C-6 of the reducing end and N acylated by α,β-polyunsaturated acyl chains at the nonreducing end. We hypothesized that both species possess a similar set of nodulation genes and that various S. meliloti Nod⁻ mutants could be used to identify clones containing their M. huakuii counterparts by heterologous functional complementation by a genomic cosmid library of strain 7653R. The seven S. meliloti 2011 nod mutant strains listed in Table 2, which exhibit a clear Nod⁻ phenotype on alfalfa, were used. The transconjugants of each cross were used to inoculate 200 seedlings of Medicago sativa, the macrosymbiont of S. meliloti. Efficient nodulation occurred with a high proportion (80 to 97%) of replicates for all crosses, except that of the nodFL mutant (GMI6628), for which nodulation was found in only 18% of tubes.

Cosmids carrying genes complementing the S. meliloti nod mutants were isolated from the surface-sterilized nodules and characterized by restriction analysis. A physical map of the nodulation region was constructed (Fig. 1). Cosmids identical to pHN50 were isolated from Nod⁺ transconjugants of the S. meliloti nodA, nodB, nodC, and nodD₁D₂D₃ mutants, indicating that pHN50 contains all of these genes. Southern hybridizations using the nodD, nodA, and nodC probes located the nodC and nodD genes to a 4.2-kb EcoRI fragment and the nodA gene to a 20-kb EcoRI fragment and a 10-kb HindIII fragment. These results suggest that nodA is separated from the nodBC genes, in contrast to the arrangement seen in most rhizobia. Cosmid pHN71, isolated from a Nod⁺ transconjugant of the S. meliloti nodF nodL mutant (GMI6628), and pHN50 share 5.0- and 0.8-kb EcoRI fragments adjacent to the 20-kb EcoRI fragment carrying nodA. Southern hybridization using the nodE probe indicated that the 5-kb fragment contains a nodE homolog. The pHN72 cosmid, isolated from a Nod⁺ transconjugant of the nodH mutant (GMI5619), overlaps pHN50 by the 4.2-kb EcoRI fragment carrying the nodDBC homologs. Partial sequencing of the adjacent 2.8-kb EcoRI fragment of pHN72 indicated that it contains a nodH homolog and a partial nodI homolog.

FIG. 1.

Physical and genetic map of the nodulation (nod) gene region of Mesorhizobium sp. strain 7653R. Black boxes indicate possible nod box sequences, shaded arrows indicate sequenced genes, open arrows represent partially sequenced genes, and the white box represents the truncated nodB-like sequence downstream of nodA. The positions of primers DC1 and DC2 are indicated by triangles. Only the regions of cosmids pHN71 and pHN72 relevant to this study are shown (→). E, EcoRI; S, SalI; B, BglII; P, PstI; H, HindIII.

Sequence analysis of the common nod genes of strain 7653R.

The 4.2-kb EcoRI fragment that hybridized with the nodD and nodC probes was subcloned into pBluescript KS+ (pHN51) and sequenced (accession no. AJ249393). This revealed four open reading frames, which correspond to the full-length nodD, nodB, and nodC genes and a partial nodI gene (Fig. 1). There is a 47-bp nod box (positions 1484 to 1530) in the nodDB intergenic region (Fig. 2), but a full-length nodA gene is not present in this region. Alignment of nodB gene sequences from 7653R and other rhizobia indicated that the 7653R nodB gene has an extra 54 bp at its 5′ end, which shows 68.3% identity with the 5′ end of the 7653R nodA gene (Fig. 2).

FIG. 2.

Organization of the 7653R nodDB intergenic region. The 5′ end of nodB is similar to the 5′ end of 7653R nodA (sequence written in italic letters). The nucleotide sequence of the nodDB intergenic region is numbered as in accession no. AJ249393. Arrows indicate start codons of the nodD and nodB genes. Boldface letters indicate the consensus nod box (29).

The 10-kb HindIII fragment carrying nodA was subcloned into pBluescript KS+ (pHN64). Southern hybridization with the nodA probe and subcloning localized the gene to a 1.15-kb PstI-EcoRI DNA fragment which was sequenced (accession no. AJ249353). This region contains a full-length nodA gene which has a recognizable nod box upstream (positions 242 to 288) (Fig. 3). These results confirm that nodA is separated from nodBC by a distance of approximately 22 kb. A 107-bp truncated nodB-like sequence (ΔnodB) was identified downstream of nodA, which shares 50.5% sequence identity with the corresponding part of the full-length nodB gene (Fig. 3). The ΔnodB start codon overlaps the nodA stop codon (ATGA). This kind of junction between nodA and nodB has been found in many rhizobia. The presence of ΔnodB immediately downstream of nodA and the presence of a nodA-like extension at the 5′ end of nodB are molecular evidence that genome rearrangements occurred at the nodA-nodB junction in an ancestor of the nodulation region now carried by strain 7653R.

FIG. 3.

Organization of the 7653R DNA sequence upstream and downstream of the nodA gene. DNA downstream nodA shows similarity to a fragment of the nodB gene (sequence written in italic letters). The nucleotide sequence is numbered as in accession no. AJ249353. The start and stop codons of the nodA gene are underlined. The consensus nod box (29) is indicated by boldface letters.

The organization of common nod genes varies among Mesorhizobium species but is conserved in Mesorhizobium isolated from A. sinicus.

To determine whether the separation of nodA from nodBC was unique to strain 7653R or was present in other A. sinicus isolates, primers DC1 and DC2 were designed to amplify the nodDC region, which corresponds to the 2.0-kb PstI fragment of pHN51 (Fig. 1). Twenty-four representative strains were selected from a collection of field isolates isolated from six southern provinces of China. These strains had previously been characterized by restriction fragment length polymorphism (RFLP) analysis of PCR-amplified 16S rRNA genes and 16S-23S rDNA intergenic spacers (IGS) (Table 1) (17, 42). PCR amplification of each of the 24 strains, using primers DC1 and DC2, produced a single band of 2.0 kb. This is the same size as obtained from 7653R and indicates that the separation of nodA from nodBC is a general feature of A. sinicus rhizobia.

Additional evidence that nodA is separated from nodBC was provided by using primers nodA-1 and nodA-2. These primers had been used previously to amplify nodA sequences from tropical tree rhizobia and correspond to residues 14 to 37 of nodA and 65 to 43 of nodB of the S. meliloti 1021 sequence (M112684) (18). Although a nodB-like sequence follows the nodA gene of Mesorhizobium sp. strain 7653R, the nodA-2 primer site in this sequence is different, and amplification would be unlikely. PCR using the nodA-1 and -2 primer pair did not produce visible products with any of the seven A. sinicus isolates tested (Table 1). This finding is further evidence that nodA is separated from nodB in the Mesorhizobium strains of A. sinicus.

In contrast, the M. ciceri, M. mediterraneum, M. plurifarium, and M. tianshanense type strains produced an amplification product with the nodA-1–nodA-2 primer combination, indicating that, in these species, the nodB gene is located immediately downstream of nodA. The nodA-1–nodA-2 primer pair did not amplify nodA in the related species M. loti. This is in agreement with the previous report showing that, in M. loti, the nodB gene is separated from the nodAC genes (31). The fact that the rearrangements in M. huakuii and M. loti are different suggests that these rearrangements occurred independently.

Identity of nodA sequences in A. sinicus rhizobia of various chromosomal types.

Having shown that the general organization of the common nodulation genes was conserved among A. sinicus rhizobia, we wanted to know whether the common nod genes were also conserved at the DNA sequence level in these rhizobia. For this work, we focused on the nodA gene and used only seven strains that represent each of the seven genotypes identified in an earlier study (42) (Table 1). Since the nodA-2 primer could not be used to amplify nodA genes from A. sinicus rhizobia, new reverse nodA PCR primers, nodA-3 and nodA-4, were designed to match the 3′ region of aligned nodA sequences. PCR using the nodA-1–nodA-3 primer pair produced weak amplification products with most of the A. sinicus isolates. However, a nested PCR using the nodA-1–nodA-4 primer pair and the nodA-1–nodA-3 PCR product as a template gave clean amplification products for all seven A. sinicus isolates, and each PCR product was sequenced on both strands. All seven nodA sequences were identical to that of strain 7653R, showing total conservation in spite of the diversity of geographical origins and chromosome backgrounds.

Phylogenetic studies.

Phylogenetic relationships between the nodA genes of A. sinicus rhizobia and the nodA sequences available in the GenBank/EMBL database were studied at both DNA and deduced protein levels. The results indicated that the closest relatives of 7653R nodA genes are those of two other Mesorhizobium strains, Mesorhizobium sp. (Oxytropis arctobia) strain N33 and M. loti. In order to assess whether the nodA genes of all described Mesorhizobium species clustered together, we sequenced the nodA genes from type strains of M. tianshanense, M. ciceri, M. mediterraneum, and M. plurifarium, after PCR amplification of nodA by using the nodA-1 and nodA-2 primers. The dendrograms based on nodA nucleotide sequence are shown in Fig. 4. All Mesorhizobium nodA sequences, except the M. plurifarium sequence, lie in the same clade with 97% bootstrap support. The M. plurifarium type strain was isolated from Acacia senegal, a tropical leguminous tree, which is distantly related to the legume hosts of the other type strains surveyed in this work (14). The nodA sequences of M. ciceri and M. mediterraneum, two species isolated from the same host plant (chickpea), differ at a single position.

FIG. 4.

Phylogenetic position of the Mesorhizobium sp. (A. sinicus) strain 7653R nodA and nodC sequences, calculated by using the neighbor-joining algorithm and the Kimura two-parameter model. Percentage bootstrap values obtained after 1,000 trials are shown, and the scale bars refer to the number of substitutions per site. A., Azorhizobium; S., Sinorhizobium; M., Mesorhizobium.

Comparison of the 7653R nodDBC gene sequences with the corresponding sequences available in databases also indicated that the closest relatives of 7653R genes are those of Mesorhizobium sp. (Oxytropis arctobia) strain N33 and M. loti. The dendrogram based on nodC nucleotide sequences is shown in Fig. 4.

DISCUSSION

The nod gene organization in A. sinicus rhizobia is unusual.

The common nodulation genes in most rhizobial species are organized in a single operon, nodABC, with a copy of nodD upstream and divergently transcribed (15, 37). The exceptions described previously are R. etli (38), M. loti (31), and Mesorhizobium sp. strain N33 (7). M. loti strains NZP2037 and NZP2213 have nodB apart from nodAC (Fig. 5). The nodA gene is separated from nodBC in R. etli and Mesorhizobium sp. strain N33. However, in all cases studied so far, the nodA, -B, and -C genes are in the same orientation. Physical mapping and sequence analysis of the common nodulation gene region of Mesorhizobium sp. (A. sinicus) strain 7653R revealed that nodA is separated from nodBC by about 22 kb, and the two operons are divergently transcribed. Like R. etli, M. loti, and Mesorhizobium sp. strain N33, strain 7653R has a truncated nodB-like sequence immediately downstream of the nodA gene (Fig. 5). The complete 7653R nodB gene has extra sequence at the 5′ end that shows homology with the full-length nodA gene. These two observations support the hypothesis that nodABC was the ancestral operon organization and that genome rearrangements have occurred.

FIG. 5.

Schematic map of selected nodulation genes in rhizobial species where rearrangements of the nodABC genes have been detected (7, 31, 37). Black boxes indicate functional nod genes. Shaded boxes indicate DNA fragments similar to nod genes, while white boxes indicate nod boxes. Arrows indicate the direction of transcription of the nod genes. M., Mesorhizobium.

Horizontal transfer of nodulation genes.

Genetically dissimilar microsymbionts have frequently been isolated from a single legume species (25). This is well documented in the case of the symbionts of soybeans, which belong to two different genera, Bradyrhizobium and Sinorhizobium (25), and of beans which belong to five different species of Rhizobium (R. leguminosarum, R. etli, R. tropici, R. gallicum, and R. giardinii) (1). Our previous work indicated that rhizobia belonging to four different 16S rDNA types nodulate A. sinicus (42). The high genetic diversity of A. sinicus rhizobia has been confirmed by the analysis of other marker genes, such as the 23S rRNA gene (42), recA, glnA, and glnII (our work [unpublished data]). However, the results presented here show that representatives of the different rhizobial rDNA types have the same common nod gene organization and identical nodA gene sequences. The chemical structures of Nod factors produced by Mesorhizobium sp. (A. sinicus) strain 7653R and two M. huakuii strains, Ra5 and Ra98, have been shown to be the same (40). These findings suggest that the nod genes of A. sinicus rhizobia are conserved in contrast to their chromosomal diversity.

The phylogenetic positions of different 16S rDNA genotypes within Mesorhizobium have been shown to be independent of host plant origin (22). Dendrograms based on both 16S and 23S rDNA indicate that Lotus rhizobia, which are known to be diverse, are the closest relatives of the different rDNA types of A. sinicus rhizobia studied in this paper (42). Sullivan et al. (34) observed transfer of the ability to form effective nodules with Lotus corniculatus between different species of mesorhizobia within soil communities. Although the symbiotic genes of Lotus rhizobia are on the chromosome, they can be transferred to nonsymbiotic mesorhizobia in the field environment because they are carried on a symbiotic island (35). Some Lotus rhizobia and A. sinicus rhizobia have the same chromosomal background but different host ranges. For example, Mesorhizobium sp. (A. sinicus) strain 7653R and M. loti NZP2037 belong to the same 16S rDNA genotype, which is different from those of both the M. loti type strain, NZP2213, and the M. huakuii type strain, CCBAU2609. This situation is equivalent to that known for the three biovars of R. leguminosarum, which have the same chromosomal type, but have different symbiotic genes and host ranges. Like R. leguminosarum, the nodulation genes of A. sinicus rhizobia are carried on a large plasmid, which could spread within field populations by horizontal plasmid transfer (17). Rapid horizontal transfer would explain why identical symbiotic genotypes are associated with different chromosomal backgrounds. This work has also revealed that the nodA sequences of the M. mediterraneum and M. ciceri type strains are almost identical. This finding parallels the results for the A. sinicus isolates: both the M. ciceri and M. mediterraneum type strains were isolated from the same host plant, Cicer arietinum (chickpea), again suggesting symbiotic gene transfer between chromosomal types that are sufficiently divergent to be considered separate species.

These findings suggest that symbiotic transfers can occur between different species of Mesorhizobium, in agreement with the findings of Sullivan et al. (34) for Lotus rhizobia. Interspecies transfer has also been documented for Rhizobium species: the same symbiotic genotypes have been isolated from R. leguminosarum bv. phaseoli, R. etli, R. gallicum, and R. giardini (1). If different species of fast-growing rhizobia are exchanging symbiotic genes at detectable frequencies, other genes might also be exchanged. If this were the case, it would bring into question the reliability and usefulness of species designations. However, we have recently sequenced internal fragments of the recA and atpD genes of the type strains of most rhizobium species, and these data support the 16S phylogeny (unpublished data). The three genera Rhizobium, Sinorhizobium, and Mesorhizobium are separated in all three phylogenies, and there is no evidence for intergenus gene exchange. Analysis of the glutamine synthetase genes (glnA and glnII) of rhizobial type strains may indicate an ancient exchange of genes between a Rhizobium species and M. huakuii but does not suggest that intergenus transfers of chromosomal genes are common (41).

We have shown that the functional nodA sequences of A. sinicus rhizobia are totally conserved between strains with different chromosomal backgrounds. While the amino acid sequence of nodA is undoubtedly constrained by functional considerations, the redundancy of the genetic code means that many base substitutions in the DNA sequence will not alter the protein. Such “silent” substitutions usually have little selective effect and accumulate relatively rapidly over evolutionary time. Complete identity at the DNA level is therefore strong evidence for a recent common ancestry of the nodA genes. The nonfunctional ΔnodB sequence is presumably not under selective pressure and will accumulate mutations more rapidly than a protein-coding gene. The ΔnodB sequence could serve as a molecular marker to further investigate symbiotic gene transfers among A. sinicus rhizobia for population dynamic and evolutionary studies. If symbiotic gene transfers are rare, the phylogeny of ΔnodB would be more similar to the 16S rDNA (chromosomal) phylogeny, since both symbiotic and chromosomal genes will have had a shared history. If, on the other hand, there has been a recent rapid spread of the A. sinicus symbiotic genes, sequence variation within ΔnodB will not reflect the chromosomal background. The ΔnodB sequence might therefore give an insight into how common horizontal gene transfer is in the genus Mesorhizobium.