Introduction

Nitrogen-fixing soil bacteria that establish mutualistic associations with legumes are referred to as rhizobia. Rhizobia produce polysaccharides, such as exopolysaccharides (EPSs), capsular polysaccharides (KPSs), lipopolysaccharides (LPSs), and cyclic β-(1,2)-glucans which are necessary for establishing symbiotic associations (Kannenberg and Brewin 1994; Hoang et al. 2004). In addition to protecting bacteria against environmental stresses, polysaccharides of the rhizobia play vital but poorly understood roles in infection of legume roots (Skorupska et al. 2006). EPS may aid in the attachment of bacteria to roots, and play a structural role in infection thread formation (Skorupska et al. 2006). In addition, numerous pieces of evidence indicate that rhizobial extracellular polysaccharides are crucial to protect bacteria against host defenses (Niehaus and Becker 1998; D’Antuono et al. 2005). The symbiotic phenotype of mutants defective in elements of EPS production depends on the type of nodule ontogeny (Gray et al. 1991; Fraysse et al. 2003). For example, specifically, acidic EPS seems to be essential for the establishment of rhizobium–legume symbiosis on legumes developing an indeterminate type of nodule (Fraysse et al. 2003).

Exopolysaccharide biosynthesis represents a multi-step process, in which many gene products are involved. The synthesis of succinoglycan (EPSI) of Sinorhizobium meliloti has been extensively studied. Many exo genes responsible for EPSI biosynthesis in S. meliloti have been identified, including all the glycosyl transferase genes involved in the synthesis of whole repeating carbohydrate units, as well as some genes involved in secretion (exsA) and polymerization (exoQ, exoT, and exoP) (Heidstra et al. 1994; Gonzalez et al. 1996). Genes homologous to those involved in biosynthesis of EPS II of S. meliloti have also been reported in the synthesis of EPSs of other rhizobial species, such as Rhizobium leguminosarum and Bradyrhizobium japonicum (Heidstra et al. 1994). The process of rhizobial EPS biosynthesis is complicated and regulated at both transcriptional and post-translational levels, with multiple regulatory systems identified in S. meliloti and R. leguminosarum (Heidstra et al. 1994; Skorupska et al. 2006). It is known that environmental factors such as the media osmolarity, and phosphate and nitrogen concentrations can influence the synthesis of the EPS (Bardin and Finan 1998; Mendrygal and Gonzalez 2000), but it is not clear how bacteria recognize and respond to these environmental signals to affect EPS synthesis.

To date there have been few detailed studies about EPS in the Mesorhizobium genus (Chen et al. 1995) of moderately growing rhizobium. In M. loti, exopolysaccharide mutants are fully effective on Lotus uliginosus (determinate nodulating host) but are ineffective on Leucaena leucocephala (indeterminate nodulating host) (Hotter and Scott 1991). In this study, we used transposon mutagenesis to identify seven EPS production defective mutants in Mesorhizobium tianshanense, a bacterium originally isolated from arid, saline, desert soil in northwestern China in 1995 (Chen et al. 1995), and later widely found in dry soils, acting as a nitrogen-fixing symbiont for at least eight different plant species, including species of Glycyrrhiza (licorice) (Tan et al. 1997), whose roots are one of the most important crude medicines in Asia and Europe. We found that exopolysaccharide production in M. tianshanense is important for the rhizobial biofilm-formation and nodulation.

Materials and methods

Bacterial strains, plasmids, and culture conditions

Bacterial strains and plasmids used in this work are shown in Table 1. M. tianshanense strains were grown at 28°C in TY medium (Vincent 1970) and Escherichia coli were grown at 37°C in LB medium (Sambrook et al. 1989). Mesorhizobium exopolysaccharide production defective mutants were screened at 28°C in YMA medium (Vincent 1970). Antibiotics were added to the following final concentrations: streptomycin, 100 mg/ml; kanamycin, 50 μg/ml; chloramphenicol, 30 μg/ml; and ampicillin, 100 μg/ml. The mtpC insertional deletion was constructed by cloning the mtpC internal fragment into pVIK112 (Kalogeraki and Winans 1997). The resulting plasmid, pZJ2, was then introduced into the M. tianshanense chromosome by single crossover homologous recombination. The insertional mutants were confirmed by PCR using a mtpC upstream primer and a lacZ primer. This also simultaneously generated a mtpC–lacZ transcriptional fusion. The mtpR and mtpS deletion mutants were constructed by cloning the internal fragments of mtpR and mtpS into the plasmid pEX18Gm (Hoang et al. 1998), resulting pWP2, and pWP6, respectively. These two plasmids were then transferred into M. tianshanense strains by conjugation and single-crossover events were selected. The insertional mutants were confirmed by PCR.

Table 1 Bacterial strains and plasmids used in this study
Full size table

Screening for exopolysaccharide-deficient mutants

Random transposon mutagenesis of M. tianshanense SmR was carried out by introducing a mariner transposon containing a promoterless lacZ gene (pKGL3) (Xu and Mekalanos, unpublished) from E. coli SM10λpir using filter mating (Sambrook et al. 1989). The transconjugants were selected on YMA medium with kanamycin and streptomycin. Nonmucoid colonies were selected from ten libraries of transposon insertion mutants for further characterization.

Arbitrary PCR and nucleotide sequence analysis

Arbitrary PCR was performed as described previously (Gao et al. 2006) to obtain short fragments of chromosomal DNA flanking transposon ends. The PCR products of the second round were sequenced with the transposon primer used in the second round, and the sequences were compared with the Genbank DNA sequence database using the BLASTX program. The full sequences were obtained by subcloning the transposon insertion flanking regions into pBluescript (Clonetech) and sequencing using the primer walking method.

β-Galactosidase activity assays

Bacteria containing mtpC–lacZ transcriptional reporters were grown in TY medium with appropriated antibiotics. Samples were withdrawn at the time points indicated and β-galactosidase activity was measured as described previously (Miller 1972).

Screening for exopolysaccharide regulatory genes

The M. tianshanense strain containing a mtpC–lacZ transcriptional fusion was mutagenized by the mariner transposon pSC137 (Chiang and Mekalanos 1999) as described previously (Sambrook et al. 1989). Transconjugants were selected on TY plates with appropriate antibiotics and 5-bromo-4-chloro-3-indolyl β-d-galactoside (X-gal) (100 μg/ml) and white colonies, which lost the ability to activate mtpC–lacZ, were isolated.

Biofilm formation assay

Borosilicate glass tubes were filled with 1 ml of MM medium containing 10 μl of saturated overnight cultures as inoculum and appropriate antibiotics (Wang et al. 2004). After 3 days incubation at 28°C with circular agitation, the content of each tube was removed gently. The tubes were rinsed with water and stained with 1.2 ml of 1% crystal violet for 5 min, and tubes were then washed and dried and photographs were taken.

Nodulation assays

Glycyrrhiza uralensis (Asian licorice) seeds were sterilized and germinated as previously described (Zheng et al. 2006). After 6 days, plants were inoculated with 300 μl cultures of wild-type or mutant M. tianshanense grown in TY medium (OD600 = 2.0). The plants were pulled out to count the number of nodules at the time indicated. There were at least six replicates for each inoculation.

Nucleotide sequence accession number

The M. tianshanense polysaccharide biosynthesis gene DNA sequence reported here has been deposited in the GenBank database under accession number EU034647 and EU034648.

Results and discussion

Screening of exopolysaccharide defective mutants in M. tianshanense

Exopolysaccharides play important roles in establishing rhizobium–plant interactions. However, little is known of the regulation of EPS biosynthesis in rhizobia such as M. tianshanense. To study the regulation of EPS production, a transposon insertion mutant library of M. tianshanense was screened for mutants defective in exopolysaccharide production. By screening approximately 10,000 random transposon insertion mutants on YMA agar, we isolated seven nonmucoid mutants (Es2, Es3, Es4, Es5, Es6, Es8, Es10). The colony morphology of these mutants is distinct from that of wild type (Fig. 1a, left panel).

Fig. 1
figure 1

Colony morphology and chromosomal locations of transposon insertions of the exopolysaccharide-deficient mutants. a M. tianshanense and its EPS mutants were streaked on MM agar plates at 28°C for 2 days. b Genetic maps of genes involved in M. tianshanense EPS production and transposon insertion locations (circles). The mtp stands for M. tianshanense polysaccharide gene

Full size image

We then used arbitrary PCR followed by subcloning and DNA sequencing to identify transposon insertion sites in these seven EPS mutants. Sequence analysis revealed that those transposons inserted in two clusters (Fig. 1b). The first gene cluster is similar to pssNOPT of R. leguminosarum bv. viciae (Young et al. 2006). All these genes are highly conserved in many Rhizobium species and involved in the translocation of polysaccharides and polymerization of the repeating subunits of EPS (Skorupska et al. 2006). We thus propose to name these genes mtpABCD (Mesorhizobium tianshanense polysaccharide genes ABCD). Another gene isolated from the transposon screen was located upstream of the mtpABCD operon and divergently transcribed (Fig. 1b). This gene, which we named mtpE, was similar to the exo5 gene in R. leguminosarum bv. trifolii. exo5 encodes a UDP-glucose dehydrogenase responsible for the oxidation of UDP-glucose to UDP-glucuronic acid (Krol and Skorupska 1997); mutations in exo5 mutant causes a pleiotropic phenotype and affects the production of glucuronic acid- and galacturonic acid-containing polysaccharides (Laus et al. 2004). Of note, our screen for EPS mutants may not be saturating, as all seven mutants have different insertion sites and we failed to obtain the regulators isolated from our second screen (see below). It is likely that more genes are involved in EPS biosynthesis in M. tianshanense.

To confirm our screen results, we constructed a strain containing a lacZ insertion (see “Materials and methods”) in one of the EPS genes, mtpC, and measured the exopolysaccharide production in the mutant. As expected, the production of exopolysaccharides was completely abolished in the mtpC::lacZ mutant (Fig. 1a, right panel), indicating that the mtpC gene product is essential in M. tianshanense EPS production.

M. tianshanense eps genes are regulated by a two-component regulatory system

Various genes have been found to regulate the synthesis of exopolysaccharides, but the mechanisms behind their regulatory roles are not very clear. In order to investigate the regulatory genes of exopolysaccharides in M. tianshanense, we constructed a chromosomal mtpC–lacZ transcriptional reporter fusion and examined β-galactosidase activity at the different time points of bacterial growth. Expression of mtpC reached maximal levels at the late-log phase and slightly declined by stationary phase (data not shown).

To investigate how EPS genes are regulated in M. tianshanense, we mutagenized the M. tianshanense strain containing the mtpC–lacZ reporter using a transposon and screened for LacZ− transconjugants (white colonies on X-gal plates). From approximately 10,000 transconjugants screened, five mutants displayed low β-galactosidase activity. Sequence analysis of arbitrary PCR products from those strains indicated that the mariner transposon disrupted the lacZ gene in four mutants, while the transposon insertion in one mutant (ES11) disrupted the sensor histidine kinase gene of a two-component regulatory system. Further sequencing indicated that a response regulator gene is located immediately downstream of the senor kinase. We therefore named them mtpS and mtpR, respectively (Fig. 1b). The predicted gene products are similar to the sensor kinase/regulator ActS–ActR of R. leguminosarum bv.viciae (86% identity at amino acid levels), which has previously been studied for its function in low pH conditions (Boesten et al. 2000). In S. medicae, ActSR has been shown to regulate many genes, including cbbS and narB, and is required for the low pH and microaerobic induction of the nitrogen fixation regulators fixK and nifA (Fenner et al. 2004). To confirm the role of MtpRS in regulation of EPS biosynthesis in M. tianshanense, we constructed mtpR and mtpS insertional deletions in wild type and strains containing mtpC–lacZ, respectively. Both mtpR and mtpS mutants did not produce exopolysaccharides, similar to that of mtpC mutants (Fig. 1a, right panel). Mutations in either mtpR or mtpS gene also abolished mtpC–lacZ activity (Fig. 2), indicating that this two-component regulatory system is crucial for mtpC activation. Interestingly, both MtpR and MtpS did not affect mtpE transcription, as mtpE–lacZ β-galactosidase activity in mtpRS mutants was similar to that in wild type bacteria (data not shown). The actual external signals required to modulate MtpRS activity is currently under investigation.

Fig. 2
figure 2

The effect of the MtpR/MtpS two-component regulatory system on EPS production and mtpC transcription. Both MtpR and MtpS are required for mtpC expression. Strains containing mtpC–lacZ reporters were grown to late-log at 28°C and β-galactosidase activity was measured and reported as Miller Units. The results are average of three experiments ± SD

Full size image

Influence of EPS biosynthesis on biofilm formation and nodulation

Rhizobia can form dense, structurally complex biofilms on abiotic surfaces as well as on root surfaces, and biofilm formation is important for rhizobium–legume symbiosis (Ramey et al. 2004). S. meliloti exopolysaccharide-deficient mutants exhibit reduced biofilm phenotypes that correspond to their reduced nodulation abilities (Fujishige et al. 2006). To examine whether EPS biosynthesis is involved in biofilm formation in M. tianshanense, we compared the biofilm formation of wild type and EPS mutants on glass surfaces. Figure 3 shows that while wild type strains produced significant amount of biofilm, while mtpC, mtpR, and mtpE mutants all formed little biofilms. These data indicate that EPS is indispensable in forming biofilms in M. tianshanense and are accordance with exopolysaccharide mutants in many other bacterial species, such as E. coli (Danese et al. 2000), Pseudomonas aeruginosa (Friedman and Kolter 2004), and Vibrio cholerae (Yildiz and Schoolnik 1999).

Fig. 3
figure 3

The effect of EPS production on M. tianshanense biofilm formation. Biofilm formation of wild type and EPS mutants. Strains were grown in TY medium at 28°C for 3 days. Biofilms were visualized by staining with 1% crystal violet

Full size image

To test whether EPS production in M. tianshanense plays a role in symbiosis, a nodulation assay on Asian licorice (Glycyrrhiza uralensis) was performed using wild type and EPS mutant mtpC, mtpR strains. On average 11.3 nodules were formed per plant about 1 month after inoculation of wild type bacteria. In contrast, no nodules formed on the plant roots inoculated with either mtpC or mtpR mutant cultures, indicating that M. tianshanense EPS mutants are profoundly defective in nodulation. These results are not unexpected as in many other rhizobia mutants that are defective in EPS production usually induce abnormal nodules or fail to nodulate at all (Becker and Pühler 1998; Wielbo et al. 2004).

Mesorhizobium tianshanense, capable of nodulating and fixing nitrogen on various plants grown in arid environments, was only recently discovered, and the genes and gene regulation involved in bacterium–plant symbiosis are largely unknown. We have previously demonstrated that LuxR/LuxI-type quorum sensing regulation is crucial for root hair adherence and nodulation by this bacterium (Zheng et al. 2006). In this study, we discovered genes involved in M. tianshanense exopolysaccharides production and regulation. Given the significant role played by this bacterium in the growth of important plant species, a better understanding of exoploysaccaride function in M. tianshanense symbiotic associations is important from both economic and environmental points of view.