The Sinorhizobium meliloti Nutrient-Deprivation-Induced Tyrosine Degradation Gene hmgA Is Controlled by a Novel Member of the arsR Family of Regulatory Genes

Abstract

The regulation of the nutrient-deprivation-induced Sinorhizobium meliloti homogentisate dioxygenase (hmgA) gene, involved in tyrosine degradation, was examined. hmgA expression was found to be independent of the canonical nitrogen regulation (ntr) system. To identify regulators of hmgA, secondary mutagenesis of an S. meliloti strain harboring a hmgA-luxAB reporter gene fusion (N4) was carried out using transposon Tn1721. Two independent Tn1721 insertions were found to be located in a positive regulatory gene (nitR), encoding a protein sharing amino acid sequence similarity with proteins of the ArsR family of regulators. NitR was found to be a regulator of S. meliloti hmgA expression under nitrogen deprivation conditions, suggesting the presence of a ntr-independent nitrogen deprivation regulatory system. nitR insertion mutations were shown not to affect bacterial growth, nodulation of Medicago sativa (alfalfa) plants, or symbiotic nitrogen fixation under the physiological conditions examined. Further analysis of the nitR locus revealed the presence of open reading frames encoding proteins sharing amino acid sequence similarities with an ATP-binding phosphonate transport protein (PhnN), as well as transmembrane efflux proteins.

Rhizobia are soil bacteria capable of engaging in a symbiotic interaction with their host plants, generally legumes. Nitrogen-fixing nodules are formed on the roots of the host plant or combined on the stems in a few instances. Within these novel plant organs, the rhizobia provide the plant with organic nitrogen sources (reviewed in reference 10). Several environmental conditions, including nitrogen deprivation, have been shown to affect the interaction of rhizobia with their host plants (reviewed in reference 50).

Bacteria, such as rhizobia, can use a variety of nitrogenous compounds as sole nitrogen sources; these include dinitrogen, ammonia, nitrate, amino acids, or nucleosides. The preferred source for many bacteria is ammonia, since it supports the highest growth rate in many gram-negative species (37). Nitrogen-limited growth or growth on nitrogen sources other than ammonia induces the synthesis of proteins that transport or degrade a variety of other, less commonly used nitrogenous compounds. This response has been shown to be coordinately regulated by the Ntr (nitrogen regulation) system which, when nitrogen becomes limiting, controls the assimilation of a number of nitrogen sources in many gram-negative bacteria.

Most knowledge of the Ntr system is based on research on enteric bacteria (reviewed in references 25, 28, and 37). The central regulatory proteins of the Ntr response are NtrC (NRI), NtrB (NRII), and NtrA (sigma 54). NtrB and NtrC form a two-component system, and their activity is regulated by the cellular nitrogen status. NtrB phosphorylates NtrC when nitrogen is limiting. The phosphorylated NtrC product activates the transcription of a number of genes, including those encoding other transcriptional regulators (13, 24). In Klebsiella aerogenes and Escherichia coli, a subset of the nitrogen-controlled genes is regulated by the nac gene product, a member of the LysR family of transcriptional regulators. Nac has been shown to be involved in the regulation of the hut, put, ure, gdh, and glt genes (24, 41). In contrast to NtrC activity, Nac activity is not regulated by nitrogen availability; instead, nac gene expression is controlled by NtrC.

The Ntr and Nac control systems have also been found in several other eubacterial genera (28). However, in addition, evidence for the existence of alternative nitrogen regulatory pathways is emerging. For example, a nitrogen regulatory system which is very different from the Ntr system has been described for Bacillus subtilis (15).

Components of the Ntr system have also been identified in diazotrophic bacteria, such as Sinorhizobium, Bradyrhizobium, Azorhizobium, Azospirillum, and Azotobacter (28). The nac gene has been reported only for Azorhizobium caulinodans (30). However, a systematic search for ntr or nac independent nitrogen regulatory systems in diazotrophic organisms has not been carried out thus far. In a previous study, it was shown that Sinorhizobium meliloti responds to nitrogen deprivation by inducing pathways for alternative nitrogen sources, such as tyrosine, alanine, or nitrate (32). Here we show that the regulation of an S. meliloti gene identified during that study and encoding homogentisate dioxygenase (hmgA), which is involved in the utilization of tyrosine as a nitrogen source, is independent of the ntr system. We describe a Tn1721-mediated secondary mutagenesis protocol for S. meliloti and the isolation and characterization of a novel regulatory locus (nitR) involved in nitrogen control of hmgA gene expression.

MATERIALS AND METHODS

Bacterial strains, plasmids, and growth conditions.

The bacterial strains and plasmids used in this study are listed in Table 1. S. meliloti 1021 was grown at 28°C in TY (3) or GTS (20, 32) medium. (NH₄)₂SO₄ was used as a nitrogen source at a final concentration of 0.2%. Nitrogen or carbon deprivation conditions (GTS-N or GTS-C medium, respectively) were created by using GTS medium devoid of sources of nitrogen or carbon (32). GTS and GTS-N media were supplemented with 1% succinate to ensure that only nitrogen was limiting. Escherichia coli strains were grown at 37°C in LB medium (43). Antibiotics were used at the following concentrations: streptomycin, 250 μg ml⁻¹ for S. meliloti; kanamycin (KAN), 200 μg ml⁻¹ and 20 μg ml⁻¹ for S. meliloti and E. coli, respectively; tetracycline (TET), 5 μg ml⁻¹ for S. meliloti and E. coli; and spectinomycin, 50 μg ml⁻¹ for E. coli.

TABLE 1.

Bacterial strains, plasmids, and bacteriophage

Strain, plasmid, or bacteriophage	Characteristic(s)	Reference or source
Strains
S. meliloti
1021	Reference strain; Smr	27
N1, N3, N5, N25, N30,   N110, N112, N113,   N127, N149, N150,   N161	1021 with Tn5luxAB insertions at loci that are induced by N deprivation; Smr Kmr	32
N4, C4, C22, C37, C101	1021 with Tn5luxAB insertions at loci that are induced by N and C deprivation; Smr Kmr	32
N4/46, N4/103, N4/105, N4/55, N4/100, N4/118, N4/101, N4/117, N4/83, N4/43, N4/116	N4 with Tn1721-tagged loci; Smr Kmr Tcr	This work
S10, S11	1021 with Tn5luxAB insertions at loci that are induced by stachydrine; Smr Kmr	35
N4D	N4 with reconstructed nitR::Tn1721; Smr Kmr Tcr	This work
103S, 46S	1021 nitR::Tn1721; Smr Tcr	This work
Rm5001	1021 ntrC::Tn5; Smr Nmr	45
Rm1680	1021 ntrA::Tn5; Smr Nmr	38
E. coli DH5α	Δ(lacZYA-argF) recA1	18
Plasmids
pJOE105	Carries Tn1721; Apr Tcr	40
pRK2013	Mob+ Tra+ IncP Kmr	12
pACYC177	Mob+ Tra− IncW Apr Kmr	4
p4/46	pACYC177 carrying nitR::Tn1721	This work
p4/103	pACYC177 carrying the 5′ end of nitR::Tn1721	This work
pBluescript II SK	Apr	Stratagene, La Jolla, Calif.
pLAFR1	Mob+ Tra− IncP Tcr	16
pRmNitR	pLAFR1 carrying nitR; Tcr	This work
pMB393	Broad-host-range plasmid; Cmr Spr	17
pNitR1	pMB393 carrying nitR; Cmr Spr	This work
Bacteriophage φM12	S. meliloti phage	14

DNA manipulations.

Plasmid DNA was prepared by the alkaline lysis method (39) or using a kit from Qiagen Inc, Santa Clarita, Calif. Total genomic DNA was isolated from S. meliloti strains (21). Restriction digestions, ligations, and Southern blotting were carried out as described previously (39). Plasmids were introduced into E. coli cells by transformation (CaCl₂ method) or electroporation (39) and into S. meliloti cells by triparental conjugation (8).

Tn1721 mutagenesis and screening for putative regulatory mutants.

Plasmid pJOE105, carrying Tn1721, was transferred by triparental mating from E. coli to S. meliloti reference strain 1021 and mutant strains N112 and N4 (see also reference 7). 1021::Tn1721 mutants were selected on TY medium supplemented with TET (resistance conferred by Tn1721). Selection of N112::Tn1721 and N4::Tn1721 double mutants was carried out on TY medium containing KAN (resistance conferred by Tn5luxAB) and TET. The resulting isolates were colony purified, grown in liquid TY medium, and stored in microtiter plates at −80°C.

N4::Tn1721 isolates were grown in microtiter plates containing TY medium supplemented with KAN and TET for 48 h; the resulting cultures were spotted, using an inoculating manifold, in groups of 48 on two sets of filters on GTS plates. After 36 h of incubation, one set of filters was transferred to GTS-N medium and the other was transferred to standard GTS medium (control set). The cell patches on the filters were analyzed for luciferase activity using a photonic camera as described previously (32). The isolates were analyzed for a reduction or a loss of luciferase activity in the absence of nitrogen and for constitutive luciferase activity in the presence of nitrogen. Quantitative measurements of the luminescence of S. meliloti cells were carried out using a photonic camera or a luminometer as described previously (31).

Phage ΦM12 transduction experiments.

Transduction of S. meliloti Tn5luxAB insertion mutations was carried out using bacteriophage ΦM12 as described for E. coli phage P1 (2). ΦM12 lysates of strains N4/46 and N4/103 were used to infect S. meliloti strain 1021 to obtain single nitR::Tn1721 insertion mutants (46S and 103S) and strain N4 to reconstruct the nitR::Tn1721 insertion (N4D). To analyze the effect of the nitR mutation upon previously created Tn5luxAB fusions, the ΦM12 lysate obtained from strain N4/46 was used to transduce S. meliloti strains harboring N- and C-deprivation-induced luxAB reporter gene fusions. The transductants obtained were purified on LB medium with TET and tested for growth on LB medium with TET and KAN.

Cloning of the tagged loci.

Genomic DNA from strain N4/46 was restricted with BamHI, and DNA fragments of approximately 20 kb were purified from a preparative agarose gel and inserted into vector pACYC177. Genomic DNA from strain N4/103 was restricted with HindIII, and DNA fragments of 17 to 20 kb were purified from a preparative agarose gel and inserted into vector pACYC177. Clones p4/46 and p4/103 containing the tagged loci were selected by growth on media containing TET. BamHI and HindIII enzymes were chosen because of the following properties. Restriction with BamHI generates fragments that contain the tagged locus with the entire Tn1721 transposon (no BamHI site within Tn1721), and restriction with HindIII generates a fragment containing the part of Tn1721 that contains the genes for TET resistance. Smaller fragments were cloned into pBluescript SK for sequence analysis and hybridization purposes. Cosmids carrying the genomic regions corresponding to the tagged loci of strains N4/46 and N4/103 were isolated by probing a genomic S. meliloti 1021 cosmid library (9) by hybridization using plasmids p4/46 and p4/103 as probes.

Complementation tests.

A 1-kb XhoI/PstI fragment of pRMNitR, carrying the nitR locus, was inserted into vector pMB393. The construct obtained, pNitR1, was transferred by triparental mating to strains N4/103 and N4/46. These strains were tested for restoration of the luciferase activity of the hmgA-luxAB fusion under nitrogen deprivation conditions.

DNA sequence analysis.

Sequencing of double-stranded plasmid DNA was carried out at the Biotech Institute at Yale University (New Haven, Conn.). Analysis of the data was performed using the Sequencher program (Gene Codes Corporation). Codon preference profiles were generated with the CodonUse 3.1 program (C. Halling, University of Chicago, Chicago, Ill.). DNA and protein similarity searches were carried out using the BLAST server (National Center for Biotechnology Information, Bethesda, Md.) (1), and motifs were analyzed using the websites of InterProScan (http://www.ebi.ac.uk/interpro/interproscan/ipsearch.html) and BLAST ProDom 2000.1 (http://protein.toulouse.inra.fr/prodom). Alignments of deduced amino acid sequences were obtained using the PILEUP program (Genetics Computer Group, Madison, Wis.).

Nodulation and nitrogen fixation assays.

Alfalfa (Medicago sativa) plants were inoculated with S. meliloti strains. After 6 to 8 weeks, nodulation was analyzed and nitrogen fixation was measured as described previously (32).

Nucleotide sequence accession number.

The GenBank/EMBL accession number for nitR is AF323118.

RESULTS

Selected nitrogen-deprivation-induced luxAB reporter gene fusions are expressed independently of the ntr system.

Twenty-one derivatives of S. meliloti 1021 carrying Tn5luxAB reporter gene fusions to genes induced under nitrogen deprivation conditions were previously isolated (32). In S. meliloti, the primary regulatory pathway controlling genes responding to the cellular nitrogen status is the nitrogen regulation (ntr) system, including the ntrA (37) and the ntrBC (45) genes. In order to examine if the ntr system was involved in the regulation of expression of nitrogen-deprivation-induced S. meliloti luxAB reporter gene fusions, DNA fragments carrying the corresponding Tn5luxAB-tagged loci were inserted into broad-host-range plasmid pLAFR1 and introduced into S. meliloti reference strain 1021, strain Rm1680 (ntrA), and strain Rm5001 (ntrC). The luciferase activities of 17 plasmid-borne fusions, described in detail in a previous study (32), were measured under nitrogen deprivation conditions in the three different genetic backgrounds. Three distinct groups of expression patterns were identified (Table 2). Group 1 contained fusions displaying luciferase activity only in the reference strain 1021 and not in the ntr mutant genetic backgrounds. Therefore, the expression of this class of luxAB reporter gene fusions appears to be dependent on the ntr system. Group 2 consisted of fusions displaying luciferase activity in both the reference strain and the ntr mutant backgrounds. Members of this group therefore appear to be independent of ntr-mediated regulation. Group 3 consisted of fusions devoid of luciferase activity in all three different genetic backgrounds, suggesting that the complete promoter region of the Tn5luxAB-tagged loci may not be present on the DNA fragment inserted into pLAFR1. One of the strains belonging to group 2, N4, carries a Tn5luxAB fusion in the homogentisate dioxygenase gene hmgA, involved in tyrosine degradation (31). This strain was selected for further analysis.

TABLE 2.

Ntr regulation of nitrogen-deprivation-induced genes in S. meliloti

Group	Strain(s)	Proposed function and/or amino acid similarity	Luciferase activity in the following strain:			Regulation
			1021	ntrA	ntrC
1	N5, N9, N30	Nitrite assimilation	+	−	−	NtrA and NtrC dependent
	N8	Nitrate assimilation	+	−	−	NtrA and NtrC dependent
2	N112, N149	Exopolysaccharide synthesis	+	+	+	Not NtrA and NtrC dependent
	N1, N113	Putative ATP-binding cassette transporter and protease	+	+	+	Not NtrA and NtrC dependent
	N4, N3	Amino acid degradation	+	+	+	Not NtrA and NtrC dependent
	N150, N161	No similarity	+	+	+	Not NtrA and NtrC dependent
3	N12, N110	Amino acid transporter	−	−	−	Inconclusive
	N25, N127	Amino acid synthesis and degradation	−	−	−	Inconclusive
	N111	No similarity	−	−	−	Inconclusive

Development of a secondary Tn1721 mutagenesis protocol to identify regulatory genes.

In order to identify regulatory genes involved in controlling the expression of the S. meliloti hmgA-luxAB fusion, a secondary mutagenesis protocol with a Tn3 derivative, Tn1721 (40), was developed. To determine the utility of this transposon for the mutagenesis of S. meliloti, the transposition parameters of Tn1721 were first tested with reference strain 1021. The frequency of transposition of Tn1721 in S. meliloti was found to be relatively low (10⁻⁶ per cell generation) but sufficient to generate thousands of Tn1721 insertion mutants in a single experiment. To investigate the specificity of insertion of Tn1721 in S. meliloti, as well as to determine if the presence of a Tn5luxAB transposon in the recipient strain would affect the transposition frequency or insertion specificity of Tn1721, a pilot secondary mutagenesis experiment was carried out using Tn5luxAB-containing strain N112. Total genomic DNA of 12 TET-resistant transconjugants was isolated, digested with the enzymes SalI and HindIII (internal restriction sites of Tn1721) or BamHI (does not cut within Tn1721), and analyzed by Southern blotting using Tn1721 and Tn5luxAB DNA sequences as hybridization probes. Single hybridizing fragments of different sizes were found in all the strains analyzed, suggesting a relatively random Tn1721 insertion pattern (data not shown). Moreover, this analysis revealed that Tn1721 was not inserted into the resident Tn5luxAB transposon. Thus, Tn1721 was found to be a suitable vehicle for secondary transposon mutagenesis of S. meliloti.

Identification of a novel S. meliloti trans-acting regulatory gene (nitR) controlling the luciferase activity of strain N4 (hmgA-luxAB).

Strain N4, carrying a Tn5luxAB insertion in the hmgA gene, was used for secondary mutagenesis analysis. The luxAB fusion in this strain is induced not only under nitrogen deprivation conditions but also under carbon deprivation conditions and in the presence of tyrosine. A collection of 3,600 N4::Tn1721 derivatives was isolated and analyzed for luciferase activity under nitrogen deprivation versus nitrogen excess conditions. Repeated plate screening experiments revealed 11 N4::Tn1721 strains exhibiting a lack or severe reduction of luciferase activity under nitrogen deprivation conditions (Fig. 1A). The degree of luciferase activity reduction was found to vary substantially among the 11 N4::Tn1721 strains. To investigate the effect of carbon deprivation as well as the presence of tyrosine on the expression of the hmgA-luxAB fusion, the 11 N4::Tn1721 strains were examined for luciferase activity under these conditions. All N4::Tn1721 strains, except for one strain (N4/105), expressed the hmgA-luxAB reporter gene fusion at levels similar to those of parental strain N4 under carbon deprivation conditions. These results suggest that the Tn1721-tagged genes in 10 of the 11 strains analyzed are specifically involved in the nitrogen deprivation response. Southern hybridization analysis of the 11 strains showed that in each strain, Tn1721 was inserted in a different position of the genome and the structure of the Tn5luxAB-tagged locus was not affected (data not shown).

FIG. 1.

Luciferase activities. (A) Luciferase activities of N4:: Tn1721 strains under nitrogen (panel A) or carbon (panel B) deprivation conditions or in the presence of tyrosine (panel C). For each strain, the luminescence of a single patch of cells is shown, as described previously (32). Strains N4/89 and N4/90 were randomly chosen from the N4::Tn1721 collection and display the same luminescence pattern as parental strain N4. (B) Luciferase activities of strain N4 and N4:: Tn1721 strains N4/46 and N4/103 under nitrogen deprivation conditions, as determined with a luminometer. The data represent the average and standard error for three independent luminescence measurements. RLU, relative light units; OD, optical density units.

Two of the N4::Tn1721 strains exhibiting a severe reduction of luciferase activity under nitrogen deprivation conditions only (Fig. 1B) were selected for further analysis. The Tn1721-tagged fragments of strains N4/46 and N4/103 were cloned and analyzed by DNA sequencing. In both strains, the expected Tn1721-generated 5-bp target site duplications were found at the site of Tn1721 insertion. Comparison of the DNA sequences of both tagged loci with the corresponding genomic loci showed that Tn1721 was inserted in the target gene without rearrangements (data not shown). Both Tn1721 insertions were found to be located in a single open reading frame (ORF) (ORF2; Fig. 2). A comparison of the deduced amino acid sequence of this ORF with the sequences of proteins in the GenBank database (Fig. 3) revealed significant similarity with several proteins of Streptomyces coelicolor strain A3 (2), including the products of the genes 2SC638.01C and ORFJ12 (2SC638.01c: 44% identity, 53% similarity; ORFJ12: 44% identity, 51% similarity) (33, 36; N. R. Cerdena, J. Parkhill, B. G. Borrell, and M. A. Rajandream, unpublished data), as well as two proteins of B. subtilis, YceK and YczG (YceK: 27% identity, 50% similarity; YczG: 34% identity, 53% similarity) (22, 23). Although the functions of these proteins have not been determined, most of them have been proposed to be members of the ArsR family of transcriptional activators, based on their amino acid similarities (6, 42, 47). This family includes several metalloregulatory proteins, and one of its members is E. coli ArsR, which is a repressor for arsenic resistance (49). An amino acid sequence motif search revealed that the deduced protein encoded by ORF2 contains a helix-turn-helix motif typical of the ArsR family of bacterial regulatory proteins (motifs 1 and 3; InterProScan IPR001845, ProDom PD001992 and PD001992). Therefore, we suggest that the Tn1721-tagged ORF identified in this study constitutes a novel member of the ArsR transcriptional activator family. Since the S. meliloti gene appears to be involved in nitrogen deprivation responses, we propose to designate this gene as nitrogen regulation gene nitR.

FIG. 2.

Structure of the nitR locus and neighboring ORFs, showing deduced amino acid sequence similarities. The position of the Tn1721 insertion is indicated by black triangles. The e values were determined from BLASTX scores and indicate the probability that the match observed would occur merely by chance. Accession numbers are from GenBank. MFS (major facilitator superfamily) is a family of efflux pumps.

FIG. 3.

Alignment of the deduced S. meliloti NitR protein sequence with the deduced amino acid sequences of the S. coelicolor ORFJ12 and B. subtilis yceK and yczG genes. Identical amino acids are indicated by black boxes; similar amino acids are indicated by white boxes. The predicted ArsR helix-turn-helix motif is indicated by lines over the sequences (1: motif 1; 2: motif 3; PR00778: PRINTS accession number).

Downstream of nitR, an ORF (ORF1) encoding a polypeptide with similarity to the PhnN protein of E. coli was found (GenBank accession number P16690) (Fig. 2). The phnN gene is part of a 17-gene cluster involved in phosphonate utilization, although its exact function is not known (5, 26, 29). Recently, other phn-like genes (but not phnN) have been described for S. meliloti (34). Upstream of the nitR gene, an ORF (ORF3) with a divergent transcriptional orientation was detected; this ORF encodes a protein with similarity to putative transmembrane efflux proteins of Pseudomonas aeruginosa (GenBank accession number AAG07743) (44) and S. coelicolor (GenBank accession number CAB66188) (36; Cerdena et al., unpublished data), as well as several chloramphenicol resistance proteins, including a Bacillus halodurans protein (GenBank accession number BAB05835.1) (46; H. Tokami, K. Kakasone, and Y. Tkaki, unpublished data) and a polypeptide of Rhodococcus fascians (GenBank accession number S2183) (11). These proteins share one common feature, namely, the presence of several transmembrane regions. Analysis of the S. meliloti ORF3-encoded protein predicted the presence of 10 possible transmembrane helices (data not shown). This information suggests that ORF3 encodes a transmembrane protein possibly involved in efflux. Analysis of the promoter regions of nitR, ORF1, and ORF3 failed to reveal the presence of classical −35- and −10-type promoter consensus sequences.

Functional analysis of the S. meliloti nitR gene.

In order to show that the effect on luxAB reporter gene expression in strains N4/46 and N4/103 was indeed due to insertional inactivation of the nitR gene, the Tn1721-induced mutation was reconstructed and complementation experiments were performed. The nitR::Tn1721 mutation of strain N4/46 was introduced into parental strain N4 via phage transduction as described in Materials and Methods. The allelic exchange of nitR with nitR::Tn1721 in the transductants was verified by Southern hybridization analysis using the nitR fragment as a probe. When cultivated in TY or GTS medium, the rate of growth of the reconstructed strain (N4D) was found to be the same as that of strain N4/46. In addition, the locus corresponding to the nitR gene was isolated from a genomic S. meliloti cosmid clone library (pRMNitR). A 1-kb XhoI/PstI fragment carrying the entire nitR ORF was inserted into broad-host-range plasmid pMB393 and introduced into strains N4/46 and N4/103. Complementation of the altered regulation of luciferase reporter gene expression in both N4/46 and N4/103 was found (Fig. 4). Therefore, we conclude that the nitR gene is indeed responsible for the altered regulation phenotype.

FIG. 4.

Luciferase activities of strain N4 and nitR strains N4/103 and N4/46 carrying a plasmid-borne nitR gene, as determined with a luminometer. See the legend to Figure 1B for details.

In order to examine the effect of a nitR insertion mutation on other nitrogen-deprivation-induced Tn5luxAB reporter gene fusions in our collection (32, 35), the nitR::Tn1721 insertion of strain N4/46 was introduced into 19 Tn5luxAB-tagged strains. Interestingly, the expression of none of the fusions in these strains was found to be affected in the nitR::Tn1721 genetic background (data not shown), suggesting that nitR-mediated regulation may be specific for the hmgA locus (tagged in strain N4).

In order to analyze the effect of a nitR mutation on the physiology of S. meliloti, single nitR::Tn1721 insertion mutants were constructed via phage transduction, generating strains lacking the Tn5luxAB reporter gene construct (strains 46S and 103S). Strains 46S and 103S were found to grow at the same rate as strains 1021 and N4 in TY and GTS media (data not shown).

In the presence of tyrosine, strain N4 produces a brown pigment due to a Tn5luxAB insertion in the hmgA gene, resulting in the accumulation and oxidation of homogentisate (31). However, the two strains carrying Tn1721 insertions in the nitR gene (46S and 103S) were not found to produce any brown pigment, in spite of the substantial reduction of hmgA-luxAB reporter gene activity observed in strains N4/46 and N4/103 (Fig. 1B). Therefore, we suggest that the residual amount of hmgA expression observed in a nitR mutant background is still large enough to provide the cell with sufficient homogentisate dioxygenase to degrade homogentisate, so that homogentisate does not accumulate. Alternatively, the nitR gene may also be involved in controlling other steps in the tyrosine degradation pathway, such as tyrosine aminotransferase and/or hydroxyphenylpyruvate dioxygenase. These enzymes are likely involved in two consecutive steps of the tyrosine degradation pathway that generate homogentisate (31). Reduced or abolished expression of these enzymes due to mutated NitR could prevent the production and accumulation of homogentisate. However, when we tested double mutants N4/46 and N4/103 for melanin production, we observed that these mutants also made this pigment. Therefore, NitR does not seem to be involved in the regulation of the production of homogentisate but only in its degradation.

The nitR strains (46S and 103S) were found to nodulate alfalfa plants with the same efficiency as S. meliloti reference strain 1021. The resulting nodules displayed wild-type levels of nitrogen fixation, suggesting that the nitR mutations did not affect symbiotic nitrogen fixation under the conditions examined (data not shown).

DISCUSSION

Rhizobia, such as S. meliloti, must be able to persist and compete for scarce nutrients in the bulk soil, compete for colonization of the rhizosphere and plant infection, and adapt their metabolism to the nutritionally more favorable, distinct conditions within the plant cells in the nodule. These three different modes of existence exemplify the need for a high degree of physiological adaptability, specific genetic mechanisms to sense changes in environmental conditions, and the ability to respond rapidly. These characteristics led to a search for S. meliloti genes specifically expressed under nutrient limitation conditions, using Tn5luxAB, and to an examination of the role of the tagged loci in persistence and competition (31, 32). Here we have taken these studies one step further by examining the regulation of one of the previously identified nitrogen- and carbon-deprivation-induced genes, hmgA, which is involved in the degradation of tyrosine as an alternative nitrogen source (31). We found that although an S. meliloti hmgA-luxAB reporter gene fusion is strongly induced under nitrogen limitation conditions, it does not appear to be controlled by the canonical ntr system (ntrA and ntrBC). This system has previously been found to be involved in nitrate assimilation, nitrogen fixation, and glutamine synthetase II (glnII) gene regulation but not in amino acid catabolism (9, 38, 45). Therefore, we developed a method for secondary transposon mutagenesis of the original hmgA-luxAB reporter strain, N4, to identify a hitherto-unidentified novel regulatory pathway(s) involved in nitrogen deprivation responses.

To ensure the absence of transposition immunity, we chose a transposon of the Tn3 family for the second round of mutagenesis. Putative regulatory mutants were selected by screening the double mutants (hmgA::Tn5luxAB and Tn1721) for either reduced luminescence under nitrogen deprivation conditions, indicative of a mutation in a putative activator gene, or the appearance of luminescence under nondeprivation conditions, indicative of a mutation in a putative repressor gene. We were able to isolate 11 strains with reduced luminescence. Two strains with severely reduced luciferase activity under nitrogen deprivation conditions only were analyzed in detail.

The Tn1721 insertions of these two strains were found to reside in the same gene, which we designated nitR (nitrogen regulator). The protein encoded by this gene, NitR, has not yet been described for S. meliloti. It shares significant similarity with several transcriptional regulatory homologs of the ArsR family. Most members of this group are metalloregulatory proteins that bind specific metals and contain conserved cysteine residues involved in metal binding: motif 2 of the ArsR helix-turn-helix motif (consisting of four motifs) (6, 42, 47). It is believed that binding of a metal changes the conformation of the protein, preventing it from binding to its target. These residues (motif 2), however, are not found in NitR or in the other S. coelicolor and B. subtilis ArsR-like regulators (Fig. 3). Most members of the ArsR family described to date function as repressors of heavy metal resistance operons, with the exception of two. HlyU is a transcriptional activator of the hemolysin hlyA gene of Vibrio cholerae (48), and NolR of S. meliloti is a regulator of nod gene expression (19). Thus, it seems that a subset of the ArsR family members, including NitR, HlyU, NolR, YceK, YczG, and several S. coelicolor ArsR homologs, is not involved in the regulation of metal resistance.

So far, we do not know how NitR regulates hmgA gene expression. Because of the similarity with the ArsR family of regulators, NitR might be a DNA-binding protein interacting with the hmgA promoter region. However, it is also possible that NitR regulates the hmgA gene indirectly by controlling the expression of another regulator which, in turn, controls hmgA gene expression.

Thus, the regulation of the S. meliloti hmgA gene appears to be complex, involving at least one activator (NitR) under nitrogen deprivation conditions. The isolation of the activator gene, nitR, involved in the control of nitrogen deprivation in S. meliloti and belonging to the ArsR family of regulators, proves that previously unknown nitrogen response regulatory pathways exist in gram-negative bacteria. NitR seems to be specific for hmgA gene activation, since it is not involved in the regulation of other S. meliloti nitrogen-deprivation-induced genes, such as the exo and spe genes (for exopolysaccharide synthesis and arginine degradation, respectively), which are not under Ntr control. Therefore, more ntr-independent regulators remain to be discovered for S. meliloti.