An Orphan LuxR Homolog of Sinorhizobium meliloti Affects Stress Adaptation and Competition for Nodulation

Abstract

The Sin/ExpR quorum-sensing system of Sinorhizobium meliloti plays an important role in the symbiotic association with its host plant, Medicago sativa. The LuxR-type response regulators of the Sin system include the synthase (SinI)-associated SinR and the orphan regulator ExpR. Interestingly, the S. meliloti Rm1021 genome codes for four additional putative orphan LuxR homologs whose regulatory roles remain to be identified. These response regulators contain the characteristic domains of the LuxR family of proteins, which include an N-terminal autoinducer/response regulatory domain and a C-terminal helix-turn-helix domain. This study elucidates the regulatory role of one of the orphan LuxR-type response regulators, NesR. Through expression and phenotypic analyses, nesR was determined to affect the active methyl cycle of S. meliloti. Moreover, nesR was shown to influence nutritional and stress response activities in S. meliloti. Finally, the nesR mutant was deficient in competing with the wild-type strain for plant nodulation. Taken together, these results suggest that NesR potentially contributes to the adaptability of S. meliloti when it encounters challenges such as high osmolarity, nutrient starvation, and/or competition for nodulation, thus increasing its chances for survival in the stressful rhizosphere.

Bacteria are constantly subjected to environmental changes that require them to coordinate their behavior in order to adapt and survive. This is particularly crucial for bacteria that interact with eukaryotic hosts during the establishment of a pathogenic or symbiotic association. One way to achieve this coordination is through the regulatory system known as quorum sensing. This form of regulation mediates gene expression through the interaction of signaling molecules and regulatory proteins in a population density-dependent manner. Quorum sensing is often found in host-associated bacteria, where it plays a role in human pathogenesis (Pseudomonas aeruginosa), symbiosis (Sinorhizobium meliloti), plasmid conjugation (Agrobacterium tumefaciens), and antibiotic production (Erwinia carotovora) (17, 48, 72, 73). Quorum sensing was first discovered and described for the gram-negative marine bacterium Vibrio fischeri, whose regulation of bioluminescence by LuxR/LuxI serves as a paradigm for most quorum-sensing systems of gram-negative bacteria (19, 20). According to this model, production of the autoinducer signal, N-acyl homoserine lactone (AHL), is controlled by proteins from the LuxI family of synthases, which bind to and activate specific regulators belonging to the LuxR family; those regulators then go on to control specific target genes (20). Typically, members of the LuxR family of proteins share two conserved domains: an amino-terminal autoinducer-binding domain and a carboxy-terminal DNA-binding domain (23, 25, 73). Binding of the LuxR-type protein to its cognate autoinducer causes a conformational change that results in the multimerization and stabilization of the protein (80). The LuxR-AHL complex that is formed recruits RNA polymerase and regulates the transcription of target genes by recognizing and binding to specific DNA sequences at quorum-sensing-regulated promoters (25, 73).

In the typical LuxR/LuxI-type quorum-sensing circuit, the gene encoding the response regulator (luxR) is in close proximity to the gene encoding the AHL synthase (luxI). In several instances, no gene encoding a LuxI homolog can be found associated with the gene encoding the regulator, which therefore is commonly referred to as an orphan LuxR regulator (22). Numerous bacteria have been reported to carry orphan LuxR homologs that regulate a variety of genes and/or functions (2, 14, 21, 56, 79). One example of an orphan LuxR regulator is QscR, which responds to the AHLs produced by the LasR/LasI quorum-sensing system of P. aeruginosa and regulates the synthesis of virulence factors such as phenazine and hydrogen peroxide (14, 42). QscR thus integrates into the complex quorum-sensing regulatory network of P. aeruginosa (40, 42). Another example of an orphan LuxR-type regulator is TrlR of A. tumefaciens (56). Transfer of the virulence genes encoded by the Ti plasmid in A. tumefaciens is controlled by the TraR/TraI quorum-sensing system (24). The Ti plasmid also codes for trlR, and in the absence of conditions conducive to conjugation (e.g., rich carbon sources), TrlR binds to TraR and prevents this resource-intensive process (13, 56). Interestingly, in some bacteria, orphan LuxR regulators are responsive for exogenous signal molecules. For example, although Escherichia coli and Salmonella enterica do not normally produce AHLs, they possess LuxR-type regulators that respond to quorum-sensing signals produced by other microbial species (2). Similarly, the orphan LuxR homologs XccR of Xanthomonas campestris pv. campestris and OryR of Xanthomonas oryzae pv. oryzae respond to host plant exudates in order to mediate virulence (21, 79).

Sinorhizobium meliloti is a gram-negative soil bacterium that can exist in a free-living state or can form a symbiotic association with the plant Medicago sativa (alfalfa). During this association, bacteria fix atmospheric nitrogen for the plant, often a limiting factor in plant growth, and in return receive nutrition from the host (26). Quorum sensing has been extensively characterized for S. meliloti, where SinI and SinR are the homologs of LuxI and LuxR, respectively (26, 47, 48, 70). SinI is involved in the production of a range of AHLs, and SinR serves to regulate AHL production in response to population density (47, 48, 70). In addition to SinR/SinI, S. meliloti has a gene called expR that encodes an orphan LuxR-type response regulator (57). The ExpR regulator, in conjunction with SinR/SinI, controls a myriad of genes involved in motility, chemotaxis, and production of the symbiotically relevant exopolysaccharides EPS II and succinoglycan by S. meliloti (27, 29, 32). In addition to the SinR and ExpR response regulators, the S. meliloti Rm1021 genome contains four additional orphan LuxR homologs (encoded by the SMc04032, SMc00878, SMc00877, and SMc00658 loci) that are not associated with a known AHL synthase of S. meliloti (26).

Here we explore the regulatory role of SMc04032, one of the orphan LuxR homologs of S. meliloti, via global expression analysis and phenotypic studies. Disruption of the SMc04032 locus, which we have named nesR, renders the bacteria unable to cope with specific nutritional, environmental, and stress conditions. We show that nesR is involved in the active methyl cycle within the methionine biosynthesis pathway and impacts general stress and metabolic responses. In addition, we show that the nesR mutant has a competitive disadvantage in nodulation compared to the wild-type strain.

MATERIALS AND METHODS

Bacterial strains and media.

The bacterial strains and plasmids used in this study are listed in Table 1. The wild-type strain used in this study is Rm8530 (SU47 str-21 expR⁺). The nesR mutant (2011mTn5STM.1.11.H02) was obtained from Anke Becker, who generated it by random Tn5 transposon mutagenesis in Rm2011 (60). The nesR mutation was then transferred into different strain backgrounds by generalized transduction of 2011mTn5STM.1.11.H02 using phage φM12 as described previously (28, 60). S. meliloti strains were grown at 30°C in Luria-Bertani (LB) medium supplemented with 2.5 mM MgSO₄ and 2.5 mM CaCl₂ (LB-MC) for routine cultures. Starter cultures for RNA purification were grown in tryptone yeast extract (TY) medium and subcultured in minimal mannitol glutamate low-phosphate medium (MLP) (50 mM morpholinepropanesulfonic acid [MOPS], 19 mM sodium glutamate, 55 mM mannitol, 0.1 mM K₂HPO₄/KH₂PO₄, 1 mM MgSO₄, 0.25 mM CaCl₂, 0.004 mM biotin [pH 7]) as described previously (31). For growth with glycine betaine as the sole source of carbon, S. meliloti was grown aerobically at 30°C in MLP supplemented with 1 mM glycine betaine (Sigma) instead of mannitol. Antibiotics were used at the following concentrations: streptomycin, 500 μg/ml; gentamicin, 5 μg/ml (for E. coli) or 50 μg/ml (for S. meliloti); neomycin, 200 μg/ml in LB medium and 25 μg/ml in MLP.

TABLE 1.

Bacterial strains and plasmids used in this work

Strain or plasmid	Relevant characteristic(s)a	Reference or source
Strains
Sinorhizobium meliloti
Rm8530	SU47 str-21 expR+ Smr	57
2011mTn5STM. 1.11.H02	Rm2011 SMc04032::Tn5	60
Rm8530 nesR	SMc04032::Nm	This work
Rm8530 nesR + pJNesR	SMc04032::Nm + pJNesR	This work
Rm8530 nesR + pJN105	SMc04032::Nm + pJN105	This work
Escherichia coli
DH5α	—b	Life Technologies
MT616	MT607(pRK600)	28
Plasmids
pJN105	araC-PBAD cassette cloned into pBBR1MCS-5	54
pJNesR	pJN105 containing the SMc04032 gene; Gmr	This work

^aSm, streptomycin; Nm, neomycin; Gm, gentamicin.

^bSee the source catalog.

Construction of pJNesR.

To complement the nesR mutation, the SMc04032 gene was cloned into the arabinose-inducible plasmid pJN105 (54). An 800-bp fragment was amplified from Rm8530 chromosomal DNA with primers 5′-CGCGAATTCCCTGTGACCGGACCGGAGAC-3′ and 5′-GCGTCTAGAGGGGTTTCGAGCCGTGGCAG-3′ (underlined nucleotides are EcoRI and XbaI restriction enzyme sites, respectively). The plasmid was then transformed into E. coli, and candidates were selected with gentamicin. The recombinant plasmid was introduced into Rm8530 nesR by triparental mating, and transconjugants were selected with neomycin and gentamicin (28). For the vector-only control, the pJN105 plasmid was introduced into Rm8530 nesR by triparental mating and selected with gentamicin (28). All strain constructs were confirmed by PCR.

Sensitivity to detergent stress.

The ability of S. meliloti to grow in the presence of the hydrophobic detergent sodium deoxycholate (DOC) was tested by plating dilutions of the bacterial strains onto LB agar containing 0.1%, 0.2%, 0.4%, or 0.6% DOC. Viable counts were expressed as CFU and were plotted against the concentrations of DOC. The assays were conducted as three independent experiments.

Growth analysis.

A single colony of each S. meliloti strain was inoculated into 2 ml of TY medium in the presence of the appropriate antibiotics and was grown at 30°C. Saturated cultures were then washed and subcultured (1:100) in either MLP alone, MLP plus 0.5 M NaCl, or MLP (without mannitol) plus 1 mM glycine betaine. Cell density was measured by monitoring the optical density at 600 nm (OD₆₀₀), and all growth curves were performed in triplicate.

AHL extraction and TLC analysis of AHL production.

AHLs were extracted from 25-ml cultures of S. meliloti grown to an OD₆₀₀ of 2 with the appropriate antibiotics as described previously (48). Five-milliliter aliquots of the cultures were acidified to a pH of <4 and were extracted twice with equal amounts of dichloromethane as described previously (66). The AHL-containing dichloromethane extracts were dried in a SpeedVac (Labconco), resuspended in 20 μl of 70% methanol, and spotted onto a Whatman LKC₁₈ analytical thin-layer chromatography (TLC) plate. The TLC plates were resolved in a 70% methanol chamber, and after drying, the plate was overlaid with the A. tumefaciens NTL₄(pZLR4) indicator organism as described previously (47).

RNA purification for microarray and real-time PCR analyses.

Bacteria were grown to saturation in TY medium. A 1:100 dilution was used to inoculate 25 ml of MLP with the appropriate antibiotics. Cultures were then grown aerobically at 30°C until they reached an OD₆₀₀ of 1.2. These culture conditions and optical density have been established previously as optimum conditions for differential expression of quorum-sensing-regulated genes (29, 32). Cells were harvested by centrifugation (10,000 × g for 2 min at 4°C), and the cell pellets were immediately frozen in liquid nitrogen. Cells were then resuspended in 10 mM Tris HCl (pH 8) and in the RLT buffer (supplemented with β-mercaptoethanol) provided with the RNeasy minikit (Qiagen). The cells were transferred to Fast Protein tubes (Qbiogene) and disrupted using the Ribolyser (Hybaid) (45 s; level 6.5) prior to spin column purification according to the RNeasy minikit RNA purification protocol. The RNA samples were then treated with Qiagen on-column RNase-free DNase. Samples were DNase treated a second time with Turbo RNase-free DNase from Ambion according to the manufacturer's protocol, and an additional RNA cleanup step was performed (31). RNA integrity was determined by using an Agilent 2100 Bioanalyzer.

Affymetrix GeneChip hybridization and expression analysis.

cDNA synthesis and hybridization to the GeneChip Medicago genome array (Affymetrix, Santa Clara, CA) were performed, as described previously, at the Core Microarray facility at the University of Texas Southwestern Medical Center (Dallas) (29, 32). Ten micrograms of RNA was used for each experiment, and the GeneChip Scanner 3000 was used to measure the signal intensity of each array. The .CEL files generated were processed by the Affymetrix GeneChip operating software (GCOS, version 1.4). Comparative analysis of control and experimental expression yielded an M value (signal log ratio), which indicated an increase, a decrease, or no change in the expression of a gene in the mutant with respect to that in the wild-type strain. An M value of ≥1 (a twofold change) and a P value of ≤0.05 were considered significant.

Quantitative real-time PCR analysis.

cDNA from each strain was prepared with the RETROscript kit from Ambion by using 1 μg of total RNA per reaction, and 1 μl of the cDNA reaction product was used as a template for the real-time PCR. The oligonucleotide sequences for real-time PCR are listed in Table 2. For real-time PCR analysis using SYBR green dye, each reaction mixture contained 0.3 μM sense oligonucleotide, 0.3 μM antisense oligonucleotide, 0.5× SYBR green 1 (Sigma), and half of an Omni Mix HS PCR bead (each PCR bead contains 1.5 U of Taq DNA polymerase, 10 mM Tris-HCl [pH 9.0], 50 mM KCl, 1.5 mM MgCl₂, 200 μM deoxynucleotide triphosphate, and stabilizers, including bovine serum albumin) in a 25-μl reaction volume. The experiment was performed in a Cepheid SmartCycler, version 2.0c, as previously described (31). The expression of the SMc00128 gene was used as an internal control for normalization as described previously (29). This gene was used as a reference because it was expressed at similar levels under different environmental conditions and/or in different mutant backgrounds, as measured by real-time PCR (38).

TABLE 2.

Sequences of the oligonucleotides used for quantitative real-time PCR

Gene	Sequence (5′ to 3′)
	Forward primer	Reverse primer
SMc03112 (metH)	CCCAATCTATCCGAAAGG	GAGTATCTGAAGGTTGCC
SMc04325 (bmt)	GCCCTTTCCGATCTCCTC	TCCACGAAGTCCTGATGC
SMc01109 (metK)	GGGCATCATGTTCGGCTATG	CGGTCACCTGGCTCTTGG
SMc02755 (ahcY)	CGGACCCTGCTGCTTGAC	TCGGCATCCTTCTCCAACC
SMc01843 (metF)	GATTATTGATCGGCAGTATCC	CTTACGCAGTTCTTCTTCG

Plant symbiosis assay.

To test the nodulation and nitrogen fixation ability of the nesR mutant, plant assays with the symbiotic host M. sativa were carried out in duplicate with at least 15 plants per strain per experiment. S. meliloti cultures were grown in LB-MC with the appropriate antibiotics at 30°C until saturation. A 1:100 dilution was used to inoculate 3-day-old plant seedlings on Jensen's agar as described previously (41). Mock treatment with water or inoculation with the Rm1021 exoY mutant strain was used as the uninoculated or Fix⁻ control, respectively (41). The plates were incubated at 25°C with 60% relative humidity and a 16-h light cycle. Plants were examined weekly, and after 4 weeks, the plant height was measured and pink (nitrogen-fixing) and white nodules were enumerated.

Competition assay for plant symbiosis.

Five ratios (100:1, 10:1, 1:1, 1:10, and 1:100) of the wild-type strain to a mutant strain were tested for competitive nodulation of the host. Saturated cultures of the two strains were washed three times with sterile water, diluted (1:100), and mixed in the appropriate amounts to achieve the desired inoculum ratios. The relative numbers of viable bacteria were enumerated by dilution plating of a portion of the inocula on LB-MC-streptomycin and LB-MC-neomycin plates. One milliliter of each inoculum was applied to 3-day-old plant seedlings on Jensen's agar as described previously (41). The plates were incubated at 25°C with 60% relative humidity and a 16-h light cycle. Each combination of the strains was applied to 10 plants, and 6 nodules per plant were harvested at 4 weeks postinoculation. Nodules from the crown region were picked and sterilized in a 50% bleach solution for 5 min. In a microtiter plate, the nodules were subsequently washed three times with water and then crushed in LB-MC plus 0.3 M glucose. Bacteria released from the crushed nodules were diluted and plated on LB-MC plates. The bacterial strains were further identified by growth in LB-MC-neomycin and/or LB-MC-streptomycin. In addition, PCR analyses were carried out to confirm the genetic backgrounds of the strains. The competition assay for nodulation was carried out twice, and competitiveness was assessed by comparing the proportional representation of each strain in the inoculum to the proportional representation of each strain after recovery from the nodules (8).

RESULTS

Sequence analysis of the SMc04032 locus.

In S. meliloti, the SinR and ExpR regulators have been extensively characterized as part of the quorum-sensing regulon (32, 46, 48). The sequenced S. meliloti Rm1021 genome revealed the presence of a large number of putative luxR-type genes. Of these homologs, four in particular (the SMc04032 [nesR], SMc00878, SMc00877, and SMc00658 genes) shared the highest level of homology to the genes encoding the classical V. fischeri LuxR protein and the S. meliloti SinR and ExpR proteins (26). These genes do not appear to be associated with a cognate luxI-type gene and could be considered to code for orphan LuxR regulators (22, 26).

Genetic and biochemical evidence has identified two conserved domains in LuxR-type proteins: the autoinducer-binding N-terminal domain and the DNA-binding C-terminal domain (73). Our analysis of the NesR protein with SMART (simple modular architecture research tool) revealed that it contains an autoinducer-binding domain (amino acids 17 to 178) and a helix-turn-helix DNA-binding domain (amino acids 189 to 246), characteristic of the LuxR family of proteins (43). The NesR protein is predicted to encode a protein of 260 amino acids with a molecular mass of 29.7 kDa (43).

Crystal structure studies of the TraR quorum-sensing regulator of A. tumefaciens identified the amino acid residues that are involved in AHL binding, TraR dimerization, and DNA binding (80). Through protein sequence alignments of TraR with other LuxR-type proteins, it was observed that nine residues are identical in at least 95% of LuxR-type proteins (73, 80). An amino acid sequence alignment of NesR with SinR and ExpR of S. meliloti as well as with other previously characterized LuxR-type regulators is shown in Fig. 1. The carboxy-terminal DNA-binding domains of all LuxR-type proteins compared showed complete identity at E178, L182, and G188 (with respect to TraR). Among the six conserved residues in the amino-terminal domain, D70, P71, W85, and G113 also showed complete identity. At Y61, a substitution with a highly similar amino acid (Y61W) occurs in NesR. Interestingly, both XccR and OryR, which are LuxR-type regulators from Xanthomonas species, also have a similar Y61W substitution. The W57 residue was not conserved in NesR, SinR, XccR, or OryR. Overall, it seems that the predicted DNA-binding domain of NesR is generally conserved, whereas its predicted signal-binding domain shows some degree of variability. In TraR, the Y61 and W57 residues were shown to be required for binding to AHLs (80). The observation that W57 and Y61 were not conserved in NesR correlates with our findings that NesR does not depend on binding to AHLs for its regulatory role (see below).

FIG. 1.

Alignment of the deduced protein sequence of NesR with those of selected LuxR-type proteins. The sequences of Agrobacterium tumefaciens TraR (Entrez protein database accession no. AAA64793; 14% identity), Vibrio fischeri LuxR (AAA27542; 20% identity), Pseudomonas aeruginosa RhlR (AAG06865; 19% identity), Sinorhizobium meliloti ExpR (NP_387385/NP_387388; 17% identity) and SinR (NP_385944; 22% identity), Xanthomonas oryzae pv. oryzae OryR (Q5H3E9; 42% identity), and Xanthomonas campestris pv. campestris XccR (AAY48364; 44% identity) are aligned with that of S. meliloti SMc04032 (NesR). The alignment was performed using Vector NTI Advance 10 (Invitrogen) software. Shaded black letters represent highly similar residues; shaded white letters represent identical residues. Asterisks indicate conserved residues of LuxR-type proteins (80).

NesR affects the active methyl cycle of S. meliloti.

In order to determine the regulatory role of NesR in S. meliloti, we conducted a genome-wide expression analysis utilizing the GeneChip Medicago genome array. The gene expression profile of the nesR mutant was compared to that of the wild-type strain (Rm8530) (see additional information in the supplemental material). The data generated from the microarray analysis were mined to determine the regulation of genes with similar functions or to identify genes belonging to a single metabolic pathway. We found that genes differentially expressed in the absence of nesR belonged mainly to the active methyl cycle of S. meliloti (Fig. 2). These genes include metH, bmt, metK, metF, and ahcY, and their expression was downregulated in the absence of nesR. This observation was further validated by performing real-time PCR expression analyses (Fig. 3). We complemented the nesR mutant with a functional copy of the gene on a plasmid (pJNesR) or with a vector-alone control (pJN105) and compared the expression of the NesR-regulated genes in these constructs to that in the wild-type strain (Fig. 3).

FIG. 2.

Active methyl cycle and catabolism of glycine betaine in S. meliloti. Cyclic synthesis of methionine to produce the potent methyl donor SAM is known as the active methyl cycle. In S. meliloti, methionine generated from either methyl tetrahydrofolate or glycine betaine is converted to SAM, which, after sequential demethylations, regenerates homocysteine. The production of methionine from glycine betaine also yields dimethylglycine and is the first step of the catabolic degradation pathway of glycine betaine to pyruvate. The pathway and its associated genes are adapted from information provided by the KEGG database, the S. meliloti Rm1021 genome sequence, Barra et al., and Smith et al. (7, 26, 67). M-THF, methyl tetrahydrofolate; THF, tetrahydrofolate; ^5,10-M-THF, N⁵,N¹⁰-methylene tetrahydrofolate.

FIG. 3.

nesR regulates the expression of genes from the active methyl cycle of S. meliloti. Quantitative real-time PCR assays were performed to measure the expression of genes from the active methyl cycle. Changes in expression were calculated as 2^−ΔΔC^T from the C_T values obtained by real-time PCR analyses. Negative change values indicate downregulation in the nesR mutant (Rm8530 nesR) compared to expression in the wild-type strain (Rm8530). The expression of the downregulated genes was restored when the mutant was complemented with nesR on a plasmid (Rm8530 nesR + pJNesR). Rm8530 nesR + pJN105 served as a vector-only control. Results are averages from at least three independent biological experiments, where within the replicates the coefficient of variance of the C_T values was <4%. The experiments included the SMc00128 gene as an internal control (38).

The active methyl cycle involves the conversion of homocysteine to methionine, which in turn generates the potent methyl donor S-adenosyl-l-methionine (SAM) (Fig. 2). Donation of a methyl group from SAM regenerates homocysteine via the intermediate S-adenosyl-l-homocysteine (SAH) (Fig. 2) (49). In S. meliloti, the synthesis of methionine occurs via the cobalamin-dependent methionine synthase, encoded by metH (SMc03112), which transfers the methyl group from methyl tetrahydrofolate to homocysteine (26, 49). Recently, Barra et al. determined that in S. meliloti 102F43, the betaine methyltransferase enzyme encoded by bmt (SMc04325) can also serve as a methionine synthase by methylating homocysteine to methionine using glycine betaine as the methyl donor (7). This enzyme is unique among bacteria, and its closest homolog is encoded by the human bmt gene (7). Our analysis indicated that the expression of both the methionine synthase genes of S. meliloti, bmt and metH, was decreased (three- to fivefold, respectively) in the nesR mutant and that complementation of the nesR mutant increased the expression of bmt fourfold relative to that of the wild-type strain and restored the expression of metH to wild-type levels (Fig. 3).

Further down the pathway, the amino acid methionine is converted to SAM by the activity of S-adenosyl-l-methionine synthase, encoded by metK (SMc01109) (Fig. 2) (26). SAM is often referred to as the “active methionine,” because it serves as a methyl donor in key metabolic pathways, such as those for protein synthesis, nucleic acid methylation, and the generation of metabolites such as polyamines (44). Upon donation of a methyl group by various SAM-specific methyltransferases, SAM is demethylated into SAH. Homocysteine is reintroduced into the cycle by the cleavage of SAH into homocysteine and adenosine by S-adenosyl-l-homocysteine hydrolase, encoded by ahcY (SMc02755) (Fig. 2) (26, 49). Recycled homocysteine can then accept a methyl group from N⁵,N¹⁰-methylene tetrahydrofolate by the activity of methylene tetrahydrofolate reductase, encoded by metF (SMc01843), to generate methionine (26). We observed that the expression of metK and metF was reduced six- and fivefold, respectively, in a nesR mutant and that complementation of the nesR mutant restored expression to levels close to those of the wild-type strain (Fig. 3). Expression of ahcY was downregulated 15-fold in a nesR mutant, and complementation of the nesR mutant led to a threefold upregulation of ahcY expression relative to that in the wild-type strain (Fig. 3).

The nesR mutant does not exhibit a growth defect.

Since the nesR mutant affected the active methyl cycle, which is involved in the synthesis of methionine, we wanted to determine if the mutant was auxotrophic for this amino acid. We measured the growth of Rm8530 (wild type), Rm8530 nesR, Rm8530 nesR + pJNesR, and Rm8530 nesR + pJN105 in minimal low-phosphate medium. We observed that the growth of the mutant was similar to that of the wild-type strain, indicating that the mutant was not auxotrophic and that any differences observed between the mutant and wild-type strains were not due to a growth defect (data not shown).

Sensitivity of the nesR mutant to salt and detergent stresses.

Adaptation to environmental changes as diverse as high osmolarity, variations in pH, and nutrient starvation invokes a general stress response in bacteria (64). Alterations in bacterial membranes have also been observed as a general stress response to environmental insults such as pH changes and osmotic or heat stress (45, 77). An association of the active methyl cycle with the general stress response has been observed in several instances (see Discussion) (4, 34, 53, 62). Therefore, we sought to investigate if nesR played any role in the survival of general salt and detergent stresses.

Within the rhizobia, S. meliloti is relatively salt tolerant, even though within the species there is high strain-dependent variation in the response to hyperosmolarity. In general, different strains of S. meliloti can tolerate NaCl concentrations ranging from 0.3 to 0.7 M (52, 76). We compared the growth of the nesR mutant and the wild-type strain in minimal low-phosphate medium supplemented with 0.5 M NaCl. We observed that in the presence of salt, the nesR mutant grew at a much lower rate than the wild-type strain and that complementation of the nesR mutant restored its growth rate to wild-type levels (Fig. 4A), suggesting that the mutant is less adaptable to osmotic shock.

FIG. 4.

Analysis of sensitivity to stress. The nesR mutant showed reduced efficiency in adapting to an osmotic upshock and detergent stress. (A) Growth curves of the wild-type strain (Rm8530), the nesR mutant, and complemented strains in minimal low-phosphate medium supplemented with 0.5 M NaCl. Results are means for three independent experiments, and calculated standard errors are indicated. (B) To test for detergent stress, the nesR mutant and the wild-type strain were subjected to increasing concentrations of DOC by plating onto LB-DOC agar, and the resulting CFU was determined. Results are means for three independent experiments; standard deviations from the means are shown. The differences between the wild-type and mutant CFU were significant (P < 0.02).

Gram-negative bacteria are fairly resistant to the membrane-dissolving properties of detergents (78). A change in sensitivity to detergents such as DOC typically occurs if the cell is exposed to a stress that causes outer membrane erosion, and such a change indicates an alteration in the bacterial outer membrane (78). To determine the effect of DOC, the nesR mutant and wild-type strains were plated onto LB agar plates containing a range of DOC concentrations. The mutant exhibited increased sensitivity to DOC, and after complementation the effect of DOC was abolished (Fig. 4B). For example, at 0.1% DOC, the survival of the nesR mutant was reduced 26% from that of the wild-type strain. Similarly at 0.2%, 0.4%, and 0.6% DOC, the survival of the nesR mutant was 23%, 35%, and 47% lower, respectively. Overall, it was observed that the sensitivity of the nesR mutant to DOC increased as the concentration of the detergent was increased.

The ability of the nesR mutant to use glycine betaine as a sole source of carbon is impaired.

Usually bacteria utilize glycine betaine exclusively as an osmoprotectant (see Discussion), but several studies of S. meliloti have shown its unique ability to catabolize glycine betaine as a sole source of carbon (52, 67). In S. meliloti, glycine betaine can donate a methyl group to homocysteine to yield methionine and dimethylglycine as end products (Fig. 2) (7, 67). This reaction is also the first step in the catabolism of glycine betaine (67). S. meliloti can catabolize glycine betaine by successive demethylation reactions into pyruvate, which then feeds into the Krebs cycle to generate ATP (67). Therefore, the possible roles of glycine betaine in S. meliloti include osmoregulation and nutrition (7, 52, 67). Barra et al. reported that in Rm1021 (an expR derivative of Rm8530), glycine betaine does not serve as an osmoprotectant, and hence, its main role may depend on catabolic degradation (Fig. 2) (7). In order to assess the abilities of the mutant and wild-type strains to catabolize glycine betaine as a sole source of carbon, we evaluated their growth in minimal low-phosphate medium where mannitol was replaced with 1 mM glycine betaine. This concentration of glycine betaine has previously been shown to stimulate growth in S. meliloti (9). Compared to mannitol, glycine betaine proved to be a poor source of carbon for the growth of the wild-type strain (data not shown). Compared to that of the wild-type strain, the ability of the nesR mutant to grow via the catabolism of glycine betaine was severely impaired (Fig. 5). Interestingly, when the nesR mutant was complemented, the ability to grow in the presence of glycine betaine as a sole source of carbon was enhanced. This finding also correlates with the results of our real-time PCR expression analysis, which show that the expression of bmt (involved in glycine betaine catabolism) increased above wild-type levels in the presence of the nesR-complementing plasmid (Fig. 3).

FIG. 5.

Growth with glycine betaine as a sole carbon source. S. meliloti has the unique ability to catabolize glycine betaine as a sole source of carbon and energy. Growth analyses of the wild-type strain, the nesR mutant, and complemented strains inoculated in minimal low-phosphate medium where mannitol was replaced with 1 mM glycine betaine are shown. Results are means for three independent experiments, and calculated standard errors are indicated.

Individual and competitive establishment of symbiosis by the nesR mutant with the host plant.

Successful candidates for symbiosis in the wild must be effective at nitrogen fixation as well as competitive against locally prevalent soil strains for nodulation (8). Mutants of S. meliloti susceptible to detergents or changes in osmolarity have been reported to display variable abilities to nodulate (7, 12, 39, 65). To evaluate the symbiotic and competitive characteristics of a nesR gene mutant, plant nodulation assays were carried out. Plants inoculated with the wild-type strain and the nesR mutant individually were proficient at establishing symbiosis, as evidenced by healthy green plants with no significant differences in the numbers of pink or white nodules on the roots at the end of 4 weeks (data not shown). However, in coinoculation experiments, where the wild-type strain and the nesR mutant were combined and then applied to the plants at various ratios (100:1, 10:1, 1:1, 1:10, and 1:100), the wild-type strain was consistently more competitive than the mutant in nodule occupancy. When these strains were coinoculated at a 1:1 ratio, the percentages of bacteria recovered from the nodules were 65% and 35% for the wild-type and mutant strains, respectively (Fig. 6). Mathematical models to estimate the competitive abilities of strains have been described previously (3, 8). Competitive analysis using these models indicated that a significant linear relationship exists between the inoculum ratios and the nodule occupancy ratios (data not shown). In addition, the competitive index (C_{wild-type::mutant}) had a positive value, 0.181 ± 0.008, indicating that the wild-type strain was more competitive for nodulation than the nesR mutant (8).

FIG. 6.

Symbiosis and competition for plant nodulation. The nesR mutant was capable of establishing symbiosis but was less proficient at nodule occupancy when competing with the wild-type strain. M. sativa roots were coinoculated with different ratios of the wild-type strain (Rm8530) and the nesR mutant on Jensen's medium (41). Percentages of wild-type and mutant strains recovered from plant nodules at different inoculum ratios were compared. The percentages of bacteria applied to the plants are shown along the x axis, and the percentages recovered from crushed nodules are shown along the y axis. The differences observed were significant (P < 0.05).

NesR and the quorum-sensing system of S. meliloti.

Based on the observation that NesR was implicated in regulating SAM expression, we examined its role in the production of AHLs in S. meliloti. The AHL production profiles of the nesR mutant and the wild-type strain were examined in various quorum-sensing mutant backgrounds. We observed no difference in AHL production for strains lacking the nesR gene (data not shown). In addition, we examined the role of the Sin AHLs in the regulation of genes from the active methyl cycle of S. meliloti. The expression of various genes from the active methyl cycle remained unchanged in the absence or presence of the sinI gene (data not shown), indicating that the regulation of genes by NesR is not dependent on the Sin AHLs.

DISCUSSION

NesR is an orphan LuxR-type protein.

Based on homology, NesR belongs to the LuxR family of proteins, which share two regions of sequence conservation, an autoinducer/signal-binding domain and a DNA-binding domain. Generally, the DNA-binding motif in this family of response regulators is highly conserved, whereas the autoinducer-binding locus is similar but allows for related substitutions, perhaps reflecting the different types of signals to which the LuxR-type regulators can be responsive (16, 21, 79). For example, V. fischeri LuxR is activated by AHLs, whereas FixJ, another LuxR-type protein of S. meliloti, is activated in response to oxygen tension (18, 61). Additionally, these substitutions could reflect an alternative approach to achieving stability, since the binding of the signal to the regulator usually promotes multimerization and stabilization, both of which seem to be required for efficient DNA binding (16, 80). Since NesR is predicted to be a LuxR homolog, we investigated whether it played a role in the production and regulation of the known autoinducers (Sin AHLs) of S. meliloti. We observed no difference in the AHL production pattern of strains lacking the functional nesR gene (data not shown). Usually, the response regulator flanking the AHL synthase on the genome is involved in the production of AHLs, as is the case for S. meliloti SinR (48). Hence, it is not surprising that NesR is not associated with the production of the Sin AHLs. Our genome-wide expression studies and real-time PCR expression analysis indicated that NesR played a role in regulating genes from the active methyl cycle of S. meliloti (Fig. 2 and 3). We observed that the expression profile of the NesR-regulated genes did not change in a sinI mutant, indicating that the AHLs provided by this synthase do not act as the activating signals for the regulatory role of NesR.

Because quorum sensing is a population density-dependent regulatory system, we analyzed the expression patterns of the genes from the active methyl cycle in the wild-type strain (Rm8530) at OD₆₀₀ values of 0.2 and 1.2, which correspond to the early-logarithmic phase and the early-stationary phase of growth, respectively (N. Gurich and J. González, unpublished data). The expression of the active methyl cycle genes did not change as population density increased (data not shown), suggesting that quorum sensing per se does not play a role in their transcription. While on the basis of its characteristic domains and homology NesR belongs to the LuxR family of proteins, it does not seem to rely on the hallmarks of quorum sensing (activation by AHLs and population density) for its regulatory activity.

Activating signal of NesR.

Although AHLs are the most common signals for LuxR-type proteins, additional molecules, such as cyclic dipeptides or plant-derived compounds, have been shown to act as signals as well (21, 33, 79). In the S. meliloti Rm1021 genome, the nesR gene is flanked by two proline iminopeptidase (PIP) genes (pip2 and pip3) (26). PIP catalyzes the removal of the N-terminal proline from peptides, an activity found mainly in bacteria and some plants (51, 63). Association of PIP with LuxR-type proteins has been observed in several plant-related bacteria (21, 79). For instance, in Rhizobium etli CFN 42 and Rhizobium leguminosarum 3841, two pip genes border a putative LuxR-type gene, whereas in several Pseudomonas or Xanthomonas species, a pip gene flanks the putative LuxR-type gene (30, 75, 79). The role of one such association was characterized by Zhang et al. for the plant pathogen X. campestris pv. campestris, where plant-derived exudates activate XccR (an orphan LuxR homolog), which in turn upregulates the transcription of the pip gene (79). PIP, as well as XccR, is required for mediating bacterial virulence (79). Additionally, in the plant pathogen X. oryzae pv. oryzae, the orphan LuxR homolog OryR has been shown to be responsive to host plant exudates (21). Both XccR and OryR are highly homologous to NesR (Fig. 1), though the overexpression or mutation of nesR in S. meliloti does not change the transcription of the pip genes, in contrast to X. campestris pv. campestris XccR (data not shown). The role, if any, of plant-produced or other exogenous signals in activating NesR remains to be explored.

Correlation between the active methyl cycle and the stress response.

In higher eukaryotes and prokaryotic microorganisms, similar biochemical and molecular mechanisms are induced in response to abiotic stresses such as high osmolarity, fluctuating temperatures, or water imbalance (11). Numerous studies with plants have revealed that a prominent physiological response to stress includes the accumulation of methylated metabolites, especially those derived from the active methyl cycle (10). The active methyl cycle provides methyl groups to several key metabolic reactions, including DNA methylation and the production of glycine betaine, inositol derivatives, and polyamines (10, 44). Some of these products play a role in stress adaptation by acting as osmoprotectants (glycine betaine) and as radical scavengers (inositol derivatives) (10, 11). In several studies, changes in the transcription of genes from the active methyl cycle have been correlated with a response to stress. For example, in tomato plants, production of SAM is increased in response to stress, leading in turn to the increased production of protective lignifying tissue (62). Expression of HvMS (a metH equivalent), a methionine synthase gene from barley leaves, has been shown to be upregulated in response to salt, drought, and cold stresses (53).

Association of the active methyl cycle with stress is not limited to plants; it is observed in bacteria as well. In Pseudomonas syringae, methionine biosynthesis is required for epiphytic fitness on leaves, and this requirement was increased under environmentally stressed conditions (4). Within rhizobia, a methionine synthase (metH) mutant of Sinorhizobium fredii RT19 was highly sensitive to salt stress, while in S. meliloti 102F34, growth inhibition due to a metH mutation was relieved by the osmotic stress protector glycine betaine (7, 34). Overall, these studies indicate that the regulation of genes from the active methyl cycle plays an important role in the cell's ability to cope with stress (4, 7, 34, 53, 62). In this study we provide yet another example of the connection between stress and the active methyl cycle. Through microarray and real-time PCR gene expression analyses, we determined that mutation of the nesR gene leads to decreased expression of the active methyl cycle genes (Fig. 3). The nesR mutant is also less able to cope with hyperosmotic and detergent stresses (Fig. 4A and B), suggesting that the active methyl cycle may be critical for stress tolerance in S. meliloti. This study thus identifies the role of a previously unknown protein, NesR, as a modulator of metabolic fitness by its capacity to regulate the bacterial active methyl cycle.

nesR and the nutritional range of S. meliloti.

The complex soil environment niche of S. meliloti selects for genes that provide the microorganism with a broad ability to metabolize a variety of carbon sources (50, 52). We observed a metabolic defect in the nesR mutant in the form of a reduced capacity to catabolize glycine betaine as the sole source of carbon. Outside of nutrition, glycine betaine has recently been shown to be the methyl donor in the synthesis of methionine in S. meliloti (Fig. 2) (7). In addition, glycine betaine is a universal osmoprotectant that is used by both plants and a wide range of bacteria to counteract hyperosmotic shock (6, 35, 52). Typically, osmoprotectants are biologically inert, i.e., they accumulate under high osmotic conditions, but when normal conditions are resumed, the osmoprotectants are extruded from the cells (35, 52). However, S. meliloti is atypical in its response to osmotic stress. Unlike the situation in most bacteria, the osmoprotectants are actively catabolized by S. meliloti even under hyperosmotic conditions, indicating that these compounds are not inert (9, 52). Moreover, glycine betaine uptake in S. meliloti is constitutive, in contrast to other bacteria, where uptake is stimulated by high osmolarity (68). For rhizobia, it has been suggested that the role of glycine betaine is more nutrition focused rather than osmoprotective (71). The capacity to break down glycine betaine holds several potential benefits for S. meliloti. In its free-living form, it can expand its nutritional range, providing it with a competitive edge in the rhizosphere. In a study of several Rhizobium species, Wielbo et al. observed that strains that used a wider range of carbon sources were more competitive for nodulation (74). In its symbiotic association, glycine betaine produced by the plant could also be made available to the bacteroids, where its catabolism could be used for maintaining the symbiotic association (52). Interestingly, the ability to degrade betaines such as trigonelline, glycine betaine, and stachydrine has been shown to play a critical role in plant symbiosis; genes involved in the catabolism of trigonelline and choline to glycine betaine are expressed at all stages of symbiosis, while stachydrine catabolism genes are required for nodulation (58, 68). Various potential sources of glycine betaine for S. meliloti exist in the soil, including choline, the precursor of glycine betaine, as well as preformed glycine betaine; these are released by the surrounding plant and bacteria of the rhizosphere (35, 52). Given these facts, regulation of the active methyl cycle by NesR could serve to widen the nutritional range of S. meliloti in the free-living state and support its symbiotic association with the host.

Competition for plant nodulation.

During plant nodulation assays, we observed that the nesR mutant was as proficient as the wild-type strain at establishing symbiosis or fixing nitrogen (Fix⁺). However, the nesR mutant was less efficient at nodulation when competing with the wild-type strain. The ability to compete for plant nodulation is typically measured by coinoculating two strains at a 1:1 ratio under sterile conditions and comparing the relationships between representation in the nodule and representation in the inocula (8). In this study we not only looked at competition between the mutant and the wild type when inoculated in a 1:1 ratio but also tested four other ratios (Fig. 6). At all the ratios tested, the nesR mutant was consistently the weaker competitor. Rhizobial competitiveness is influenced by numerous factors, such as metabolic fitness, survival under stresses in the rhizosphere, microbial interactions with other prevalent species, and the ability to utilize nutrients in the soil (5, 8, 74). Therefore, factors such as increased sensitivity to stress (osmotic or detergent) and/or reduced efficiency of the active methyl cycle could contribute to the competitive defect of the nesR mutant.

Alterations in the bacterial membrane have previously been reported to occur as a response to stress conditions, including ionic, osmotic, or heat stresses (45, 77). Several studies of S. meliloti have reported a correlation between sensitivity to membrane-solubilizing detergents such as DOC and a decreased ability to compete for nodulation (12, 39, 65). Increased sensitivity to DOC is due either to an altered lipopolysaccharide (LPS) structure or to a loss of membrane integrity, probably resulting from a reorganization of the membrane components (55). For example, in previous nodulation studies of DOC-sensitive S. meliloti mutants, all of the strains were proficient at symbiosis and were Fix⁺ when assayed individually. While some of the mutations were found to affect bacterial LPS structure directly, others had a non-LPS-related effect on the membrane. Nonetheless, all mutants displayed DOC sensitivity, suggesting compromised membrane integrity (12, 39, 65). Analysis of the LPS structure of the nesR mutant showed no alterations from that of the wild-type strain (data not shown), indicating that the nesR mutation has a non-LPS-related effect on the membrane, causing an increased sensitivity to detergents that could possibly contribute to its competitive defect.

Studies of S. meliloti also report nodulation defects in strains where the synthesis of methionine is altered. For S. meliloti 102F34, a metH mutant is Fix⁺ but defective in competing with the wild-type strain for nodulation (7). Similarly, a metA (involved in methionine biosynthesis) mutant of Rm2011 is Fix⁺ but exhibits delayed nodulation (7, 59). Auxotrophs of methionine have also been shown to affect nodulation not only in S. meliloti but in other rhizobia as well (1, 36, 37, 69). Therefore, through these studies, it is well established that methionine is critical for nodule invasion and that the plant provides very small amounts of methionine to the bacteria (1, 7, 36, 37, 59, 69). The active methyl cycle of S. meliloti is the source of methionine, as well as of stress-detoxifying metabolites (see above). Though the nesR mutant is not auxotrophic for methionine, the combined effect of the reduced availability of methionine from the plant, the downregulation of methionine-generating genes (metH and bmt), and the sensitivity to DOC stress could contribute to its impaired nodulation when it competes with the wild-type strain.

Role of the functional nesR gene within the soil environment.

Chemoattractants secreted by plants attract microorganisms from bulk soil toward the rhizosphere. The level of microbial activity around plant roots is estimated to be several orders of magnitude higher than that in bulk soil (17, 52). In this environment, the metabolic and nutritional fitness of the bacteria equips them to compete for nutrition as well as for plant nodulation. Additionally, during growth, plants cause an active uptake of water from the rhizosphere and an expulsion of toxic solutes (52). Due to the combined metabolic activities of the plant and the microorganisms, the rhizosphere tends to be high in osmolarity and thus is a stressful environment for the bacteria (52). The nodulation process also contributes to general stress for rhizobia in terms of high turgor pressure within the infection thread and elevated oxidative and osmotic stresses in the nodules (15, 52). In the rhizosphere, S. meliloti nesR could play a role in the overall ability of the bacterium to tolerate stress and could allow it to utilize variable sources of carbon. Moreover, within the plant environment, nesR could provide the bacterium with the capacity to compete efficiently for symbiotic nodulation. Therefore, NesR could help S. meliloti to survive the rhizosphere environment and to establish an efficient interaction with the host plant.

In conclusion, we have identified the LuxR-type protein NesR as a regulator of the active methyl cycle of S. meliloti. The physiological processes modulated by NesR impact the ability of S. meliloti to survive stresses, catabolize specific carbon sources, and ultimately compete for plant nodulation. Because NesR is a transcriptional regulator, identification of the mechanisms by which it perceives signals and executes gene regulation could be the focus of future studies and may provide insight into alternative activating mechanisms of LuxR-type proteins.