Nitrogen fixation genetics and regulation in a Pseudomonas stutzeri strain associated with rice

Nicole Desnoues1, Min Lin2, Xianwu Guo1,3,{dagger}, Luyan Ma1,{ddagger}, Ricardo Carreño-Lopez1,§ and Claudine Elmerich1,4

1 Microbiologie et Environnement, CNRS URA D2172, Institut Pasteur, Paris, France
2 Biotechnology Research Institute, Chinese Academy of Agricultural Sciences, Beijing, PR China
3 Centro de Investigación sobre Fijación de Nitrógeno, Universidad Nacional Autónoma de México, Cuernavaca, Morelos, Mexico
4 Institut des Sciences du Végétal, CNRS UPR 2355, Bâtiment 23, Avenue de la Terrasse, 91198 Gif sur Yvette, France

Correspondence
Claudine Elmerich
Claudine.Elmerich{at}isv.cnrs-gif.fr


   ABSTRACT
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS AND DISCUSSION
REFERENCES
 
The Pseudomonas stutzeri strain A1501 (formerly known as Alcaligenes faecalis) fixes nitrogen under microaerobic conditions in the free-living state and colonizes rice endophytically. The authors characterized a region in strain A1501, corresponding to most of the nif genes and the rnf genes, involved in electron transport to nitrogenase in Rhodobacter capsulatus. The region contained three groups of genes arranged in the same order as in Azotobacter vinelandii: (1) nifB fdx ORF3 nifQ ORF5 ORF6; (2) nifLA-rnfABCDGEF-nifY2/nafY; (3) ORF13 ORF12-nifHDK-nifTY ORF1 ORF2-nifEN. Unlike in A. vinelandii, where these genes are not contiguous on the chromosome, but broken into two regions of the genome, the genes characterized here in P. stutzeri are contiguous and present on a 30 kb region in the genome of this organism. Insertion mutagenesis confirmed that most of the nif and the rnf genes in A1501 were essential for nitrogen fixation. Using lacZ fusions it was found that nif and rnf gene expression was under the control of ntrBC, nifLA and rpoN and that the rnf gene products were involved in the regulation of the nitrogen fixation process.


The GenBank accession numbers for the sequences reported in this paper are AJ297529.2; AJ313205 and AJ320536.

{dagger}Permanent address: Institute of Life Science and Technology, Huazhong Agricultural University, PR China.

{ddagger}Permanent address: National Laboratories for Agrobiotechnology, Beijing Agricultural University, Beijing 100094, PR China.

§Present addresss: Instituto de Ciencias, Universidad Autónoma de Puebla, CP 72000, Puebla, Pue, Mexico.


   INTRODUCTION
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS AND DISCUSSION
REFERENCES
 
Nitrogen fixation is widespread among eubacteria and archaea. All these bacteria possess an enzyme complex, the nitrogenase, which is responsible for the reduction of nitrogen gas to ammonia. Nitrogen fixation (nif) genes have been characterized, and their regulation investigated, in many bacterial species (Elmerich, 1991; Dean & Jacobson, 1992; Merrick, 1992; Fischer, 1994; Lee et al., 2000; and references therein). Extensive analysis has been carried out in Klebsiella pneumoniae (oxytoca) and Azotobacter vinelandii. Little information is yet available about the nif genes of other members of the {gamma}-Proteobacteria, except for a Pantoea agglomerans strain (Siddavattam et al., 1995). It was long believed that there were no nitrogen-fixers among strains of the genus Pseudomomas sensu stricto (Young, 1992). Indeed, consistent with this assumption, most of the strains described as putative nitrogen-fixing Pseudomonas were later reassigned to genera in the {alpha}- and {beta}-Proteobacteria (Chan et al., 1994). It now seems that several strains unambiguously classified as true Pseudomonas spp. can be added to the list of nitrogen-fixers, on the basis of physiological properties, nitrogenase assays, phylogenetic studies and detection of nifH DNA by hybridization or PCR amplification (Chan et al., 1994; Vermeiren et al. 1999). Two nitrogen-fixing isolates belong to the species Pseudomonas stutzeri. P. stutzeri CMT.9.A was isolated from the roots of Sorghum (Krotzky & Werner 1987), whereas strain A15 originated from rice paddies in China (You & Zhou 1989; You et al., 1991). Strain A15 was initially identified as an Alcaligenes faecalis, and was later reassigned to P. stutzeri (Vermeiren et al., 1999). P. stutzeri strains form a heterogeneous group comprising seven genomic subgroups, called genomovars, identified by DNA–DNA hybridization (Sikorski et al., 1999, and references therein). Strain A15 belongs to genomovar 1 (Vermeiren et al. 1999). This strain may be considered as an endophyte of rice and, as it promotes plant growth, it is widely used to inoculate rice in China (You et al., 1991; Hallmann et al., 1997; Vermeiren et al., 1998).

P. stutzeri A1501 (Chinese Culture Collection: CGMCC 0351), used in this work, is similar, if not identical, to the A15 strain used by Vermeiren et al. (1999). The strain was re-isolated from rice roots inoculated with strain A15 (M. Lin, unpublished). We report herein the nitrogen-fixation properties of strain A1501, and describe the organization of nif genes in this strain and the isolation of mutant strains including key regulatory mutants in the nifLA and ntrBC genes. A nif cluster spanning 30 kb was identified. The nifLA genes lie within this cluster. Homologues of rnf/nqr genes, involved in electron transport to nitrogenase in Rhodobacter (Schmel et al., 1993), were identified in stain A1501 and found to be essential for nitrogen fixation. We also report data concerning the regulation of nif, rnf and ntr gene expression.


   METHODS
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS AND DISCUSSION
REFERENCES
 
Bacterial strains, plasmids, growth conditions.
The bacterial strains and key plasmids used in this work are listed in Table 1. A map of the genomic region characterized, indicating the position of the kanamycin cassettes used for mutant construction, is presented in Fig. 3. We used Luria broth (LB) as the rich medium for P. stutzeri and Escherichia coli cultures. The minimal lactate-containing medium (medium K) was as previously described (Galimand et al., 1989). Plasmids were transferred into P. stutzeri recipients by conjugation, using E. coli S17-1 as the donor, as previously described (Galimand et al., 1989). Tetracycline (Tc) was added to the medium at a concentration of 5 µg ml-1, and kanamycin (Km) at a concentration of 100 µg ml-1, as required.


View this table:
[in this window]
[in a new window]
 
Table 1. Strains and plasmids used in this study

 


View larger version (30K):
[in this window]
[in a new window]
 
Fig. 3. Physical map of the nif and rnf region of P. stutzeri strain A1501. Location of genes, putative ORFs and km cassette insertions. Restriction sites (only relevant sites are indicated): Bg, BglII; Cl, ClaI; E, EcoRI; Kpn, KpnI; P, PstI; Sa, SacI; Sm, SmaI; Sp, SphI; X, XhoI.

 
Nitrogenase activity and {beta}-galactosidase assays.
For rapid determination, nitrogenase activity was assayed after culture overnight at 30 °C in 7 ml N-free minimal medium K, with or without 0·5 g agar l-1, in a 10 ml bottle. Otherwise, nitrogenase activity was determined using a derepression protocol, as follows. Cells from an overnight culture in LB medium were centrifuged and resuspended in a 50 ml flask containing 10 ml N-free minimal medium K, at an OD600 of 0·1. The suspension was incubated for 3 h, at 30 °C, with vigorous shaking, under an argon atmosphere containing 1 % oxygen and 10 % acetylene. We determined the ethylene production at regular intervals by gas chromatography, using the acetylene reduction test, as previously described (Galimand et al., 1989). Each experiment was repeated at least three times. In some experiments, the incubation medium was supplemented with a nitrogen source (ammonia, glutamine) or the gas phase was adjusted to give different oxygen concentrations (see Results). Protein concentrations were determined by the Bio-Rad protein assay. For {beta}-galactosidase assays, the bacteria were incubated as for nitrogenase assays, except that no acetylene was added to the atmosphere. {beta}-Galactosidase activity was determined after 4 h, as described by Miller (1972).

Molecular techniques and DNA sequencing.
Plasmid isolation, genomic DNA extraction, gel electrophoresis, restriction mapping, transformation and molecular cloning, Southern blotting and hybridization, amplification by PCR (Amersham kit) and cloning of PCR products (TA cloning kit, Invitrogen) were performed by standard methods (Sambrook et al., 1989) or as recommended by the manufacturers of the products used. Restriction enzymes were purchased from Boehringer and oligonucleotides from Oligo Express Company. We tested for the presence of plasmids in strain A1501 as described by Mavingui et al. (2002). Nucleotide sequencing was performed by Genome Express Company, on double-stranded DNA, using appropriate oligonucleotides as primers. Data were compiled with the Lasergene program. We searched for sequence similarity with the BLAST program of the National Centre for Biotechnology Information Server.

Cloning of the ntrBC region and of the nitrogen-fixation gene cluster
The ntrBC genes.
We constructed a gene library of total DNA from A1501 partially digested with Sau3A in the broad-host-range plasmid pLA2917, linearized with BglII. The mean size of the inserts was 20 kb. This library was used in colony hybridization experiments using, as a probe, a 2 kb BglII–PstI ntrC fragment from the A. brasilense plasmid pAB7. Screening was also performed by genetic complementation of the A. brasilense ntrC mutant 7148 (see text). Clone p58 of the library carried ntrBC on a 5 kb XhoI fragment. This fragment was isolated and inserted into pSUP202 and pVK100 to yield pNTRBC-58-1 and pNTRBC-58-2, respectively.

The nifH region.
The nifH forward primer described by Zehr & McReynolds (1989) and the nifD reverse primer (1337R) described by Ueda et al. (1995) were used with total DNA from strain A1501. They amplified a 2·1 kb fragment, which was inserted first into pCR2.1, and then, in the form of an EcoRI fragment into, pSUP202, to yield pNIFHD. A BamHI fragment carrying the kanamycin (km) cassette from pUC4K was inserted into the BglII site within the nifH coding sequence. The km cassette was then inserted into the host genome by homologous recombination to yield the nifH mutant strain 1502. Total DNA of strain 1502 was then partially digested with Sau3A and DNA fragments 9–11 kb in size, carrying the Km resistance gene and adjacent DNA, were then inserted into pBS digested with BamHI. We obtained overlapping fragments covering the nifH upstream region extending to rnfC (pBS7) and the downstream region extending to nifN (pBS3) (Fig. 3).

The nifB region.
A 650 bp DNA fragment from A. vinelandii strain 104399 was first amplified with the oligonucleotides CTATTCGGAAGAGGCGCACCAC (forward) and CGGGCTCGMRATCAGCGGCATG (reverse) designed from A. vinelandii and Rhodobacter capsulatus nifB nucleotide sequences (accession numbers J03411 and X07567). This PCR product was used to clone a 6·1 kb XhoI fragment, carrying nifBQ and cob, which was inserted into pBS and pVK100 vectors, to yield pNIFB-4 and pNIFB-41, respectively. This PCR fragment was also used as a probe in colony hybridization of the gene library (see above): it identified p64, a plasmid that carried nifB. From p64, we excised the 6·9 kb BglII fragment containing the upstream region of nifB extending to tmp, and inserted it into the BamHI site of pUC19.

The nifLA region.
We amplified nifA by PCR with the oligonucleotides primers nifA-p1 and nifA-p2, designed by Zhang et al. (2000), from total DNA of strain 1508, an ntrBC deletion derivative of A1501. The 450 bp PCR product was then used as a hybridization probe. It enabled us to identify a 5·3 kb ClaI fragment containing nifA, in total DNA of strain 1508, which was then inserted into pBS to yield pNIFA3. This ClaI fragment also carried the C-terminal part of nifL and the tmpA/reg2 region. The nifL gene, and the adjacent DNA extending to rnfG, was obtained from strain 1806, a nifA-km mutant, by partial Sau3A digestion and cloning of DNA fragments carrying the Km resistance gene. The 7·7 kb SacI fragment covering the entire nifLA region was inserted into the pTZ19R vector to yield pNIFLA-7 and into pRK415 to yield pNIFLA-77.

Recombination of the km cassette and transcriptional lacZ gene fusion into the host genome.
For construction of mutant strains, restriction fragments containing a km cassette within the desired coding sequence were inserted into a suicide vector. The resulting plasmid was transferred into the host by conjugation and the mutation was generated by insertion of the cassette into the genome by homologous recombination. Each mutant construction was checked for correct recombination in the host genome, either by Southern hybridization of total DNA with appropriate probes or by PCR amplification with appropriate oligonucleotide primers.

nifH.
See above for strain 1502; for the construction of strain 1502-lacZ we used a lacZ-km cassette from pKOK5 instead of the km cassette.

nifE and nifY.
A 3·5 kb EcoRI fragment from pBS3 was inserted into pSUP202, to yield pNIFE. km cassettes from pUC4KISS and pUC4K were then inserted into the KpnI site within nifE and the BglII site within nifY, to generate the nifE 1504 and nifY 1503 mutant strains, respectively.

nifB.
A 1·3 kb EcoRI fragment from pNIFB-4 carrying nifB was inserted into pBS and a pUC4K km cassette was inserted into the BglII site localized within the nifB coding sequence. The nifB-km fragment was isolated and inserted into pPHU281 vector before recombination into the host genome, to yield strain 1505.

nifA and nifL.
A 4·7 kb KpnI–ClaI fragment of the nifA region of pNIFA3 was recovered as a KpnI fragment, inserted into pPHU281 and the 618 bp BglII intragenic fragment was replaced by a BamHI fragment carrying the pUC4K km cassette to generate 1506 and by a pKOK5 lacZ-km cassette, in both orientations, to generate 1506-lacZF and 1506-lacZR. For nifL, a 1506 bp ClaI fragment covering most of the nifL sequence was replaced by the pUC4KIXX km cassette. We used the Klenow fragment of DNA polymerase I to fill in ends of pNIFLA-7 digested with ClaI and inserted the cassette, as a SmaI fragment, in the opposite orientation of the gene. A SacI fragment containing the nifL-km insertion was then inserted into pPHU281 before conjugation into the host strain, to yield strain 1507.

rnf mutants.
We recovered from pBS7 a 5·5 kb SacI fragment that carries part of rnfC, extending to ORF13, and inserted this fragment into pBS, to yield pRNF-5. A km cassette from pUC4KISS was used to replace a 1179 bp SphI fragment of pRNF-5 containing part of rnfCD, and the resulting fragment was inserted into pPHU281. The pHU281 derivative was used to generate strain 1509. The 5·5 kb fragment was also inserted into pPHU281 as a XbaI–SacI fragment. We then inserted a km cassette from pUC4KIXX into the XhoI site within the rnfG coding sequence in both orientations. The resulting plasmids were used to generate the 1510-1 and 1510-2 mutant strains.

ntrBC mutant.
A 2018 bp SmaI fragment from pNTRBC-58-2, covering part of the ntrB and ntrC coding sequences, was replaced by a km cassette from pUC4KIXX and the resulting plasmid was used to generate the 1508 mutant strain.

Construction of plasmids carrying translational lacZ fusions.
Translational lacZ fusions to the nifH, nifL, ntrB and rnfA promoter regions were obtained in the broad-host-range vector pGD926. The oligonucleotides used to amplify DNA fragments carrying the putative promoter regions and the start of the coding sequences were designed with HindIII (forward primer) and BamHI (reverse primers) restriction sites. After PCR amplification and restriction with HindIII plus BamHI the corresponding fragments were inserted into pGD926 digested with the same enzymes. We obtained pPS-nifH-LacZ with primers (F) GTCAAGCTTCTGCCAAGGATCGTTCACGG and (R) CCTGGATCCTTCCCGTAAATAGCGCATTG (restriction sites italicized). This plasmid carries a 468 bp fragment containing 434 bp of the non-coding sequence upstream from the putative ATG. We obtained pPS-nifL-LacZ with primers (F) GTGAAGCTTCTGCACAGGGTATTTACGGCTCC and (R) GTCGGATCCACGGCTTGCTGGAACACTTCGGG. This plasmid carries a 283 bp fragment containing 134 bp of the non-coding sequence upstream from the putative ATG. We obtained pPS-ntrB-LacZ with primers (F) GTGAAGCTTCCAGCTGAGCAATGTGTCGATCG and (R) GTCGGATCCGGGTTCATGTATTCCAGACGC. This plasmid carries a 504 bp fragment containing 409 bp of the non-coding sequence upstream from the putative ATG. We obtained pPS-rnfA-LacZ with primers (F) GTGAAGCTTGCAAAACCCACAGGGAGGCGC and (R) GTGGGATCCATGAACGGACACAGGCCGAG. This plasmid carries a 329 bp fragment containing 244 bp of the non-coding sequence upstream from the putative ATG. The constructions were checked by nucleotide sequencing to ensure that no errors had been introduced during PCR and that the resulting plasmids carried the correct in-frame fusions.


   RESULTS AND DISCUSSION
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS AND DISCUSSION
REFERENCES
 
Strain A1501
The isolate A1501, used in this work, is a Gram-negative, highly motile, oxidase-positive, aerobic strain. It shows properties reported by Vermeiren et al. (1999), including denitrification in anaerobic conditions, a general feature of P. stutzeri strains. This strain uses a range of organic acids and sugars including maltose and glucose. The use of maltose is typical of P. stutzeri strains (see Sikorski et al., 1999). The results of classical taxonomic biochemical tests performed, as well as the nucleotide sequencing of a small fragment of 16S rDNA (data not shown), were consistent with identification of the A1501 isolate as P. stutzeri. The full sequence of the 16S rDNA of strain A1501 was determined by Lin et al. (2000; GenBank accession number AF143245).

Southern hybridization of total A1501 DNA with nifH and nifDK probes from Azospirillum brasilense resulted in the detection of HindIII, PstI, XhoI and EcoRI bands with similar apparent sizes, suggesting that nifHDK homologues were present in strain A1501 (Fig. 1a). It also showed that the three genes were clustered and that probably A1501 contained only one copy of each. We detected no amplification (data not shown) with oligonucleotides for the PCR amplification of alternative nitrogenase genes (Loveless & Bishop, 1999), suggesting that no alternative nitrogenase was present in strain A1501. This was also the conclusion for the P. stutzeri strain CMT.9.A (Fallik et al., 1991). Hybridization with A. brasilense nifA (Fig. 1a) or ntrC (not shown) intragenic fragments revealed a complex pattern with multiple hybridizing bands, suggesting the presence of a large number of response regulators of the ntrC family in P. stutzeri as described for Pseudomonas aeruginosa (Rodrigue et al., 2000).



View larger version (27K):
[in this window]
[in a new window]
 
Fig. 1. nif genes and plasmids in P. stutzeri A1501. (a) Southern blot of total DNA of strain A1501 with A. brasilense nif probes. Probes: lanes A–D, 0·9 kb XhoI–SalI nifH fragment from pAB3; lanes E–H, 2·6 kb PstI nifDK fragment from pAB3; lanes I–K, 0·4 kb PstI nifA fragment from pAB51. Digestions: lanes A, E, I, restriction with HindIII; lanes B, F, J, restriction with PstI; lanes C, G, K, restriction with XhoI; lanes D, H, restriction with EcoRI. (b) Profile of plasmid extract: A, NGR234; B, A1501; C, 1502 (nifH-km).

 
A plasmid, in the range of 50 kb, that may carry some genetic information for nitrogen fixation has been identified in strain CMT.9.A (Krotzky & Werner, 1987). We searched for plasmids in strain A1501 by a variety of conventional techniques. No plasmid was detected in P. stutzeri A1501 or its nifH-km derivative 1502, whereas, as expected, megaplasmids of 536 kb and over 2000 kb present in the Rhizobium sp. NGR234 (see Mavingui et al., 2002) were detected (Fig. 1b). This is consistent with nif genes having a chromosomal location in strain A1501.

Effects of oxygen and ammonia on nitrogenase derepression and activity
Nitrogenase activity of the wild-type A1501 was detected in semi-solid medium devoid of a nitrogen source after overnight culture in the presence of various carbon sources including lactate, succinate, glucose, gluconate, maltose and ethanol. The bacteria formed a pellicle under the surface, suggesting that nitrogen fixation occurred at low oxygen tension.

If cell density was adjusted to an OD600 0·1, A1501 nitrogenase activity was detected at a range of initial oxygen concentration from 0·5 % up to 4 % (Fig. 2a). At concentrations of 2 % and 4 % oxygen, the derepression of nitrogenase was delayed for several hours (Fig. 2a), unless a higher cell density was used (not shown). No activity was detected with air (not shown), whereas at oxygen tension of 0·5 % or lower, the derepression curve was not linear (Fig. 2a), suggesting that oxygen was limiting. Thus, a balance between the oxygen concentration and cell density was required and, in most experiments, nitrogenase was assayed with the derepression protocol described in Methods, with an initial oxygen concentration of 1 % and OD600 0·1. Specific nitrogenase activity was in the range of 7·5±3·1 nmol ethylene min-1 (mg protein)-1. This activity is lower than that we observed, using a similar derepression protocol, with other species fixing nitrogen in microaerobiosis, such as Azospirillum brasilense or Azorhizobium caulinodans [in the range of 60 nmol ethylene min-1 (mg protein)-1; Bozouklian et al., 1986; Mandon et al., 1998].



View larger version (17K):
[in this window]
[in a new window]
 
Fig. 2. Time-course of nitrogenase activity. (a) Effect of oxygen concentration in the gas phase on A1501 wild-type nitrogenase activity. (b) Switch-off of nitrogenase activity in A1501. The vertical arrow indicates the time at which (0·01, 0·1, 1 or 10 mM final concentration) or glutamine (1·3 mM final concentration) was added. (c) Activity of wild-type and mutant strains: A1501, wild-type; 1508, ntrB-km; 1509, rnfC-km; 1508/pNTRBC, strain 1508 containing the pNTRBC-58-2 plasmid. Results are from one representative of three (a) or two (b and c) independent experiments in which similar kinetics were observed.

 
The ‘switching off and on’ of the nitrogenase activity was also observed, a feature long described in many diazotrophs (Bergersen, 1991), and recently found in Azoarcus (Egener et al., 2001). At concentrations of 10 and 1 mM and a glutamine concentration of 1·3 mM, the nitrogenase activity was rapidly and totally switched off. At 0·1 mM , the rapid and total loss of activity was reversible and switch-on was observed (Fig. 2b). Little effect was found with 0·01 mM (Fig. 2b). Further experiments are required to find if the switch-off–on is due to covalent modification of nitrogenase as reported for photosynthetic bacteria or whether it is related to the ammonia shock response found in Azotobacter (Bergersen, 1991; Ludden, 1994).

Identification of nitrogen-fixation genes
A map of the nitrogen-fixation cluster identified is shown in Fig. 3. This region was cloned as described in Methods, by first cloning fragments containing the nifH, nifB, nifA and nifL genes. By establishing a physical map of the plasmid clones and completely sequencing these clones, we were able to identify overlaps, making it possible to construct the map of this cluster. We used hybridization of total DNA with appropriate probes to check that the physical map derived from the cloned fragments was consistent with that deduced from the chromosome.

As expected, a high degree of conservation was found with the products of the corresponding nitrogen-fixation genes in other diazotrophs. The highest level of identity was observed with A. vinelandii gene products (Table 2; Jacobson et al. 1989). Indeed, A. vinelandii is the most closely related species phylogenetically. A high degree of identity was found with K. pneumoniae (Arnold et al., 1988), another member of the {gamma} subgroup of Proteobacteria, and with Azoarcus, from the {beta} subgroup (Egener et al., 2001, 2002) and Rhodobacter capsulatus, from the {alpha} subgroup (Schmel et al. 1993). The nifK translation product was 90 % identical to that of Azotobacter, 78 % identical to that of Azoarcus, 70 % identical to that of Magnetococcus sp. MC1 (unclassified proteobacteria, genome project NZ_AAAN0 0000 000) and 66 % identical to that of K. pneumoniae.


View this table:
[in this window]
[in a new window]
 
Table 2. Similarity of the products of nif, rnf and other related nitrogen-fixation gene products of P. stutzeri A1501 to the equivalent proteins of A. vinelandii, and possible functions

 
A typical feature of members of the {gamma}-Proteobacteria is the association of nifA, which encodes an activator of the transcription of nif genes, with a nifL gene (Dixon, 1998; Reyes-Ramirez et al., 2001; Rudnick et al., 2002). NifL modulates NifA activity in response to oxygen and fixed nitrogen levels. The deduced amino acid sequence of the translation product of the nifA gene of strain A1501 does not contain the conserved Cys residues between the central region and the C-terminal domain. This NifA is therefore probably oxygen-insensitive, like the NifA molecules of Azotobacter and K. pneumoniae (Merrick, 1992; Fisher 1994; and references therein). The deduced amino acid sequence of the translation product of nifL displayed the conserved domain structure, with the PAS domain reported in Azotobacter (Zhulin et al., 1997; Dixon, 1998). A higher degree of identity was found with NifL (43 %) and NifA (59 %) of Azoarcus, a member of the {beta}-Proteobacteria, than with K. pneumoniae NifLA (32 and 55 %, respectively). Indeed, it has been suggested the nif genes of Azoarcus originate from lateral transfer (Egener et al., 2002). Interestingly, a gene (PY07698) encoding a 937 residue protein resembling NifL in its N-terminal part (41 % identity on 430 residues) and NifA in its C-terminal part (53 % identity on 477 residues) has been found in the Plasmodium yoelii yoelii genome (accession number EAA20209; Carlton et al., 2002). The function of this hybrid NifL–NifA homologue in a eukaryotic organism is unknown.

For the nifHDK gene region, the order of the ORFs from ORF13 to nifN is identical to that of A. vinelandii (Jacobson et al., 1989). A high degree of identity was observed for the translation products of this region, with scores higher for the translation products known to play a role in nitrogen fixation than for other conserved ORFs (Table 2). We have not yet characterized the DNA region downstream from nifN.

Another group of genes displaying a high degree of similarity to those of A. vinelandii contained nifB and nifQ, together with the fdx gene and the ORFs 3, 5 (ORF111) and 6 (ORF119) (Joerger & Bishop, 1988; NZ_AAAD010 00087). In strain A1501 this group of genes is located between two sets of genes that are probably not related to the nitrogen-fixation process, whereas in Azotobacter nifB is linked to nifLA (Joerger & Bishop, 1988). The operon designated as cob (Fig. 3) probably encodes a protein involved in cobalamin synthesis whereas the seven ORFs upstream from nifB encode conserved proteins of unknown function. Two of these proteins, Reg1 and Reg2 (Fig. 3), resemble putative transcriptional regulatory proteins of the LysR and AraC families, respectively, and the deduced product of tmpA displays a high level of identity to thiopurine methyltranferases. The location of the cob region, encoding an essential growth factor (Roth et al., 1996), in the vicinity of the nifBQ region supports the notion of a chromosomal location for the nif genes.

One interesting finding was the detection in P. stutzeri of a group of genes, rnfABCDGEH, highly similar to genes of R. capsulatus (Schmel et al., 1993, Jouanneau et al., 1998). These genes, called rnf, for Rhodobacter nitrogen fixation, are located upstream from the fdx and nifENX genes in Rhodobacter. They are essential for nitrogen fixation and play a role in electron transfer to the nitrogenase (Schmel et al., 1993). The rnf region has been shown to display a high level of identity to the nqr locus of Vibrio alginolyticus and Vibrio cholerae (Kumagai et al., 1997; Häse & Mekalanos 1999). Similar genes have been found in many other non-nitrogen-fixers (Jouanneau et al., 1998); including P. aeruginosa (accession numbers PA3489 to 3495). RnfA (NqrE), RnfD (NqrB) and RnfE (NqrD) are transmembrane proteins similar to components of an NADH ubiquinone oxidoreductase (NQR). RnfG also displays some similarity to an NQR component. RnfB and RnfE contain iron–sulphur binding domains and the function of RnfH is unknown. An rnf cluster was recently detected in Azotobacter (Rubio et al., 2002), and rnfA is probably present in Azoarcus (Egener et al., 2002). Downstream from the rnfH gene, we identified a homologue of the nifY/X family, which we designated nifY2 (nifY-like). This gene encodes a product with high identity to that of nafY, which is found at the same location in Azotobacter (Rubio et al., 2002). The nafY gene has been shown to encode the {gamma} protein, which is involved in stabilization of the nitrogenase (Rubio et al., 2002). Interestingly, rnfA is linked to nifLA but transcribed in the opposite orientation. This also appears to be the case in Azotobacter (Rubio et al. 2002; accession number AF450501) and the organization of the nifLA–rnfABCDGEF–nifY2/nafY region is similar in the two organisms.

Functionality of nif and rnf genes
We constructed mutant strains by inserting km or km-lacZ cassettes into the wild-type genome by homologous recombination. Insertion in nifB (1505), nifH (1502 and 1502-lacZ) and nifE (1504), and insertion/deletion in nifA (1506 and 1506-lacZF) and nifL (1507), led to a Nif- phenotype with a specific activity below 0·1 nmol ethylenemin-1 (mg protein)-1. We checked that plasmid pNIFB-41 restored wild-type nitrogenase activity to the nifB mutant 1505 (not shown). Plasmid pNIFLA-77 restored only 10 % nitrogenase activity to the 1506 (nifA) mutant strain and less than 30 % to the 1507 (nifL) mutant strain.

The insertion in nifY (1503) did not significantly impair nitrogenase activity. This result may be due to the presence of a second copy of nifY, nifY2/nafY, in the A1501 genome, or it may indicate that nifY is not essential, as also reported in the case of Azotobacter (Jacobson et al., 1989; Rubio et al., 2002).

Mutations in rnf genes also strongly decreased nitrogenase activity. An rnfCD deletion (1509) resulted in detectable, but extremely low levels of nitrogenase activity (Fig. 2c), in the range of 0·1 nmol ethylene min-1 (mg protein)-1. There is only 41 bp between rnfH and nifY2/nafY coding sequences and we did not find a putative promoter motif upstream of nifY2. To rule out the possibility that the Nif- phenotype is due to a polarity effect on genes downstream from rnfH, we constructed a non-polar insertion into rnfG by inserting a KIXX cassette (1510-1). The resulting strain, like strain 1510-2, which carries the cassette in the opposite orientation, was also Nif-. Rubio et al. (2002) reported that rnfH and nafY were transcribed independently in Azotobacter, and this is likely to be the case in strain A1501. We also checked that nitrogenase activity was not restored if the assay was performed at lower oxygen concentrations (not shown), indicating that no other electron-transport chain was derepressed and compensated for the loss of the rnf gene products in the conditions used.

Cloning of ntrBC genes and properties of an ntrBC deletion mutant strain
The two-component regulatory proteins NtrB and NtrC play an essential role in the regulation of nitrogen fixation in K. pneumoniae, whereas NtrBC are not required in A. vinelandii (Alvarez-Morales et al., 1984; Rudnick et al., 2002; and references therein). Several plasmid clones carrying two-component sensor-response regulator genes were obtained by colony hybridization with the A1501 gene library, using A. brasilense ntrC fragment as a probe. Putative ntrC-containing plasmids were subsequently identified by genetic complementation of an ntrC mutant strain of A. brasilense, 7148, for nitrate utilization. The plasmid clone p58 was retained. Nucleotide sequencing of the 5 kb XhoI fragment of p58, carrying the putative ntrC gene, showed that this fragment also carries ntrB. The deduced amino acid sequences of the translation products of two genes displayed a high level of identity (78 and 88 %, respectively), to those of the ntrB and ntrC products of P. aeruginosa. The trmH gene (Fig. 3), which encodes a putative RNA methylase and is located downstream from ntrC, is also conserved in P. aeruginosa at the same location. The ntrBC genes in A1501 are not linked to glnA (glutamine synthetase structural gene), whereas such linkage is observed in Klebsiella and Azotobacter (Alvarez-Morales et al., 1984; Toukdarian & Kennedy, 1986). Total DNA digested with various enzymes was hybridized with the 5 kb XhoI fragment as a probe and the results were consistent with there being a single copy of ntrBC in the P. stutzeri genome (not shown). A strain with a deletion covering the 3' end of ntrB and ntrC was constructed. This strain, namely 1508, was impaired in nitrogen source utilization. It did not grow with nitrate, histidine or ornithine and showed a strong defect in nitrogen fixation. The ntrBC mutant strain 1508 was delayed in derepression (Fig. 2c), with the specific activity measured being similar to that of the rnf mutants, in the range of 0·1 nmol ethylene min-1 (mg protein)-1. The final levels of nitrogenase activity, after overnight incubation, were below 5 % of that in the wild-type strain. Complementation with pNTRC5, which carries the XhoI fragment containing the wild-type ntrBC region, restored wild-type nitrogenase activity (Fig. 2c) and growth properties.

Regulation of nif, rnf and ntr gene expression
We assayed the {beta}-galactosidase activity of nifH– and nifLA–lacZ chromosomal fusions in strain 1502-lacZ and 1506-lacZF and found that, in agreement with the data obtained for the nitrogenase in the wild-type, the activity of both fusions was decreased in the presence of air and . Activities with the nifH–lacZ fusion were, respectively, in Miller units: 6000 units under argon/oxygen (99 : 1, v/v) atmosphere; 500 units under air; and 20 units in the presence of 10 mM . The expression of the nifLA–lacZ fusion was 770 units under argon/oxygen (99 : 1, v/v) atmosphere, 240 units under air; and 70 units with . Standard deviation was in the range of 15 %. No activity was observed with strain 1506-lacZR, which carried the fusion in the opposite orientation (data not shown).

The expression of plasmid-borne lacZ fusions to the nifH, rnfA, nifL and ntrB promoter regions was monitored in the wild-type and in various mutant strains, including an rpoN mutant, strain 1550. The nucleotide sequence of the 8 kb fragment containing the rpoN region was established in our laboratory (GenBank accession number AJ496594) and a more detailed characterization of this region and of strain 1550, which carries a km cassette within the rpoN coding sequence, will be reported elsewhere. Strain 1550 was Nif-.

The expression of nif genes is known to be controlled by a common mechanism in Proteobacteria (Dixon, 1998). In this mechanism, the transcription activator NifA and a modified form of RNA polymerase containing the {sigma}54 factor, encoded by rpoN, activate promoters with a conserved -12(GC) -24(GG) structure. The precise way in which NifA synthesis and activity is regulated differs from one system to another (Fischer, 1994; Dixon, 1998; Rudnick et al., 2002).

The level of expression of the nifH–lacZ fusion was only slightly lower in the nifH mutant than in the wild-type, but was much lower in the rpoN (1550) and nifA (1506) mutant strains (Table 3). Indeed putative GG-GC motifs with correct spacing were found upstream from the various putative nif operons and upstream from rnfA. For example, the sequence CTGGCACAAAGGCTGCT, located 73 nucleotides upstream from the presumed translation start site of nifH, may constitute a {sigma}54-dependent promoter, as reported by Barrios et al. (1999). There are also two putative TGT-N10-ACA motifs (at positions -188 and -250 with respect to the ATG), which may correspond to NifA-dependent upstream activator sequences (UAS) (see Merrick, 1992; Dixon, 1998). The expression of nifH is under the control of NtrBC (Table 3). This suggests that NtrC is involved in nifA expression.


View this table:
[in this window]
[in a new window]
 
Table 3. Expression of lacZ fusions to nifH, rnfA, nifL and ntrB promoter regions in the P. stutzeri wild-type and mutant strains

Data are in Miller units and are the means of at least three or six (*) independent determinations ± standard deviation.

 
The fusion to the nifLA promoter region in plasmid pPS-nifL-LacZ was less strongly expressed (59 Miller units) than the chromosome-borne fusion (770 Miller units; see above). Indeed, the two fusions are different: the chromosomal fusion (1506-lacZF) is a transcriptional fusion in nifA coding sequence and the plasmid-borne fusion (pPS-nifL-LacZ) is a translational fusion in the nifL coding sequence. Also, we observed poor complementation of the nifA mutant strain (1506) by plasmid pNIFLA-77. This could account for the poor expression of the fusion. Furthermore, it cannot be ruled out that some regulatory elements were not cloned during the construction of the plasmid-borne fusion. Although additional constructions are required to draw more definite conclusions, it is possible to give a preliminary interpretation of the data obtained in Table 3. Expression levels of the nifLlacZ fusion were lowered by 3·5-fold in the rpoN-mutated background (1550; Table 3). A similar decrease was observed in the wild-type A1501 containing pPS-nifL-LacZ in the presence of ammonia (not shown). This confirmed that the nifLA promoter is N-regulated and is {sigma}54-dependent, like those of Klebsiella and Azoarcus (Egener et al., 2002) but unlike that of Azotobacter, in which nifLA expression is constitutive (Blanco et al., 1993). The pPS-nifL-LacZ activity was similar in the nifA mutant (1506) and in the wild-type (Table 3), suggesting that NifA is not necessary for its own synthesis. The expression of pPS-nifL-LacZ activity was significantly decreased as compared to values observed in the wild-type (P value 0·027) when introduced into strain 1508. This indicates that NtrC is required for nifA expression.

The requirement of rpoN for ntrBC expression (pPS-ntrB-LacZ) indicated that these two genes were expressed under the control of a {sigma}54-dependent promoter (Table 3). Indeed, the expression of the fusion was very low in the ntrBC mutant strain (1508), suggesting that NtrC was required for its own synthesis. In contrast, the fusion was overexpressed (by a factor of 2) in a nifA mutant (1506), suggesting that NifA repressed ntrBC expression, probably in some indirect manner, since NifA has not been shown to directly repress genes in other organisms. This also suggests that NtrC cannot substitute for NifA in nif gene expression.

Expression of the rnfA–lacZ fusion (pPS-rnfA-LacZ) in the wild-type was 75 % lower in the presence of ammonia (data not shown), suggesting a specific role for rnf genes in N-limiting conditions, as reported for Rhodobacter (Schmel et al., 1993; Saez et al., 2001). Indeed, RpoN (1550) and NtrC (1508), and to a lesser extent NifA (1506), controlled the expression of the rnfA–lacZ fusion (Table 3). Inactivation of the rnf genes leads to a decrease in the amount of nitrogenase polypeptides in Rhodobacter (Schmel et al., 1993). In addition, Jouanneau et al. (1998) reported a decrease in the amount of RnfB and RnfC polypeptides in rnfD and rnfG mutant strains, suggesting that the ubiquinone reductase was unstable in the absence of one of its subunits. We therefore evaluated the expression of the various lacZ fusions introduced into strain 1509, which carries an rnfCD (rnfC-km) deletion. Compared to wild-type levels, the level of expression from the nifH promoter (pPS-nifH-LacZ) was low in strain 1509, as was that from the ntrB (pPS-ntrB-LacZ) and rnfA (pPS-rnfA-lacZ) promoters, whereas the level of expression from the nifL (pPS-nifL-LacZ) promoter was increased (Table 3). This suggests a regulatory role for electron-transport-chain components in the expression of nitrogen-fixation genes. As suggested by Dixon (1998), NifL, which is a redox-sensing flavoprotein, may be reduced in vivo by several electron-transport chains. Thus, Rnf proteins may exert their putative regulatory role via NifL oxidation state, in addition to effects on the transfer of electrons to nitrogenase.

Conclusion
The characterization of nif genes in P. stutzeri A1501 is of particular interest, as no studies of nif genetics in this genus have to our knowledge been published. Nitrogen fixation in Pseudomonas sp. associated with plants is probably more common than initially thought. In a recent work, analysis of the bacterial population in the rhizosphere of cordgrass (Spartina alterniflora) based on PCR-based amplification of nifH sequences and separation of the amplicons by denaturing gradient gel electrophoresis, revealed nifH sequences highly similar to that of strains A1501 and CMT.9.A, suggesting that P. stutzeri-related strains are present in the Spartina rhizosphere (Lovell et al., 2000).

We have described here the characterization of nif genes and the general physiology of the regulation of nitrogenase synthesis and activity in our P. stutzeri strain. The nitrogen-fixation ability is probably encoded by the chromosome as no plasmids were detected. The organization of the nif and rnf genes characterized resembled that of equivalent genes in Azotobacter. Our data have revealed several interesting features and major differences with the situation in Azotobacter and Klebsiella. In particular, the ntrBC mutant is impaired in nitrogen fixation in P. stutzeri, whereas this is not the case in Azotobacter. Also, NifA controls ntrBC expression, which may reflect regulatory circuitry different from other organisms. The switching-off of nitrogenase activity and the role of the rnf genes also reflect interesting features of this particular strain.


   ACKNOWLEDGEMENTS
 
This work is dedicated to the memory of Professor Chongbiao You. The authors would like to thank Dr Y. Dessaux for taking an interest in this work and for providing laboratory space to the corresponding author, Dr J. Sappa for improving the language, Ms M. Angeles Moreno for technical assistance in plasmid profile experiments and the Pasteur Institute strain collection for providing the Azotobacter strain. Xianwu Guo received a fellowship from the Chinese Government; Luyan Ma received support from AFIRST, and the Institut Pasteur supported R. Carreño-Lopez. We are grateful for the support from Programme Franco-Chinois de Recherches Avancées (PRA-2B) and the programmes 863 and 973 of the Ministry of Sciences and Technology of China.


   REFERENCES
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS AND DISCUSSION
REFERENCES
 
Allen, L. N. & Hanson, R. S. (1985). Construction of broad-host-range cosmid cloning vectors: identification of genes necessary for growth of Methylobacterium organophilum on methanol. J Bacteriol 161, 955–962.[Medline]

Alvarez-Morales, A., Dixon, R. & Merrick, M. (1984). Positive and negative control of the glnA ntrBC regulon in Klebsiella pneumoniae. EMBO J 3, 501–507.[Abstract]

Arnold, W., Rump, A., Klipp, W., Priefer, U. B. & Pühler, A. (1988). Nucleotide sequence of a 24,206-base-pair DNA fragment carrying the entire nitrogen fixation gene cluster of Klebsiella pneumoniae. J Mol Biol 203, 715–738.[Medline]

Barrios, H., Valderrama, B. & Morett, E. (1999). Compilation and analysis of {sigma}54-dependent promoter sequences. Nucleic Acids Res 27, 4305–4313.[Abstract/Free Full Text]

Bergersen, F. J. (1991). Physiological control of nitrogenase and uptake hydrogenase. In Biology and Biochemistry of Nitrogen Fixation, pp. 76–102. Edited by M. J. Dilworth & A. R. Glenn. Amsterdam: Elsevier.

Blanco, G., Drummond, M., Woodley, P. & Kennedy, C. (1993). Sequence and molecular analysis of the nifL gene of Azotobacter vinelandii. Mol Microbiol 9, 869–879.[Medline]

Bozouklian, H., Fogher, C. & Elmerich, C. (1986). Cloning and characterization of the glnA gene of Azospirillum brasilense Sp7. Ann Inst Pasteur 137B, 3–18.

Carlton, J. M., Anguioli, S. V., Suh, B. & 41 other authors (2002). Genome sequence and comparative analysis of the model rodent malaria parasite Plasmodium yoelii yoelii. Nature 419, 512–519.[CrossRef][Medline]

Chan, Y. K., Barraquio, W. L. & Knowles, R. (1994). N2-fixing pseudomonads and related soil bacteria. FEMS Microbiol Rev 13, 95–118.

Dean, D. R. & Jacobson, M. R. (1992). Biochemical genetics of nitrogenase. In Biological Nitrogen Fixation, pp. 763–834. Edited by G. Stacey, R. H. Burris & H. J. Evans. New York: Chapman & Hall.

de Zamaroczy, M., Delorme, F. & Elmerich, C. (1989). Regulation of transcription and promoter mapping of the structural genes for nitrogenase (nifHDK) of Azospirillum brasilense Sp7. Mol Gen Genet 220, 88–94.[Medline]

Ditta, G., Schmidhauser, T., Yakobson, E., Lu, P., Liang, X. W., Finlay, D. R., Guiney, D. & Helinski, D. R. (1985). Plasmids related to the broad host range vector, pRK290, useful for gene cloning and for monitoring gene expression. Plasmid 13, 149–153.[Medline]

Dixon, R. (1998). The oxygen responsive NifL-NifA complex: a novel two-component regulatory system controlling nitrogenase synthesis in {gamma}-Proteobacteria. Arch Microbiol 169, 371–380.[CrossRef][Medline]

Egener, T., Martin, D. E., Sarkar, A. & Reinhold-Hurek, B. (2001). Role of a ferredoxin gene cotranscribed with the nifHDK operon in N2 fixation and nitrogenase "switch-off" of Azoarcus sp. strain BH72. J Bacteriol 183, 3752–3760.[Abstract/Free Full Text]

Egener, T., Sarkar, A., Martin, D. E. & Reinhold-Hurek, B. (2002). Identification of a NifL-like protein in a diazotroph of the {beta}-subgroup of the Proteobacteria, Azoarcus sp. strain BH72. Microbiology 148, 3203–3212.[Abstract/Free Full Text]

Elmerich, C. (1991). Genetics and regulation of Mo-nitrogenase. In Biology and Biochemistry of Nitrogen Fixation, pp. 103–141. Edited by M. J. Dilworth & A. R. Glenn. Amsterdam: Elsevier.

Fallik, E., Chan, Y. K. & Robson, R. L. (1991). Detection of alternative nitrogenases in aerobic Gram-negative nitrogen-fixing bacteria. J Bacteriol 173, 365–371.[Medline]

Fischer, H. M. (1994). Genetic regulation of nitrogen fixation in rhizobia. Microbiol Rev 58, 352–386.[Medline]

Galimand, M., Perroud, B., Delorme, F., Paquelin, A., Vieille, C., Bozouklian, H. & Elmerich, C. (1989). Identification of DNA regions homologous to nitrogen fixation genes nifE, nifUS and fixABC in Azospirillum brasilense Sp7. J Gen Microbiol 135, 1047–1059.[Medline]

Hallmann, J., Quadt-Hallmann, A., Mahaffee, W. F. & Kloepper, J. W. (1997). Bacterial endophytes in agricultural crops. Can J Microbiol 43, 895–914.

Häse, C. & Mekalanos, J. J. (1999). Effects of changes in membrane sodium flux on virulence gene expression in Vibrio cholerae. Proc Natl Acad Sci U S A 96, 3183–3187.[Abstract/Free Full Text]

Hübner, P., Masepohl, B., Klipp, W. & Bickle, T. A. (1993). nif gene expression studies in Rhodobacter capsulatus: ntrC-independent repression by high ammonium concentrations. Mol Microbiol 10, 123–132.[Medline]

Jacobson, M. R., Brigle, K. V., Bennett, L. T., Setterquist, R. A., Wilson, M. S., Cash, V. L., Beynon, J., Newton, W. E. & Dean, D. R. (1989). Physical and genetic map of the major nif gene cluster from Azotobacter vinelandii. J Bacteriol 171, 1017–1027.[Medline]

Joerger, R. D. & Bishop, P. E. (1988). Nucleotide sequence and genetic analysis of the nifB-nifQ region from Azotobacter vinelandii. J Bacteriol 170, 1475–1487.[Medline]

Jouanneau, Y., Jeong, H. S., Hugo, N., Meyer, C. & Willison, J. C. (1998). Overexpression in Escherichia coli of the rnf genes from Rhodobacter capsulatus: characterization of two membrane-bound iron-sulphur proteins. Eur J Biochem 251, 54–64.[Abstract]

Knauf, V. C. & Nester, E. W. (1982). Wide host range cloning vectors: a cosmid clone bank of an Agrobacterium Ti plasmid. Plasmid 8, 45–54.[Medline]

Kokotek, W. & Lotz, W. (1989). Construction of a lacZ-kanamycin-resistance cassette, useful for site-directed mutagenesis and as a promoter probe. Gene 84, 467–471.[CrossRef][Medline]

Krotkzy, A. & Werner, D. (1987). Nitrogen fixation in Pseudomonas stutzeri. Arch Microbiol 147, 48–57.

Kumagai, H., Fujiwara, T., Matsubara, H. & Saeki, K. (1997). Membrane localization, topology, and mutual stabilization of the rnfABC gene products in Rhodobacter capsulatus and implications for a new family of energy-coupling NADH oxidoreductases. Biochemistry 36, 5509–5521.[CrossRef][Medline]

Lee, S., Reth, A., Meletzus, D., Sevilla, M. & Kennedy, C. (2000). Characterization of a major cluster of nif, fix and associated genes in a sugarcane endophyte, Acetobacter diazotrophicus. J Bacteriol 182, 7088–7091.[Abstract/Free Full Text]

Liang, Y. Y., Kaminski, P. A. & Elmerich, C. (1991). Identification of a nifA-like regulatory gene of Azospirillum brasilense Sp7 expressed under conditions of nitrogen fixation and in the presence of air and ammonia. Mol Microbiol 5, 2735–2744.[Medline]

Liang, Y. Y., Arsène, F. & Elmerich, C. (1993). Characterization of the ntrBC genes of Azospirillum brasilense Sp7: their involvement in the regulation of nitrogenase synthesis and activity. Mol Gen Genet 240, 188–196.[Medline]

Lin, M., Smalla, K., Heuer, H. & van Elsas, J. D. (2000). Effect of an Alcaligenes faecalis inoculant strain on bacterial communities in flooded microcosms planted with rice seedlings. Appl Soil Ecol 15, 211–225.[CrossRef]

Loveless, T. M. & Bishop, P. E. (1999). Identification of genes unique to Mo-independent nitrogenase systems in diverse diazotrophs. Can J Microbiol 45, 312–317.[CrossRef][Medline]

Lovell, C. R., Piceno, Y. M., Quattro, J. M. & Bagwell, C. E. (2000). Molecular analysis of diazotroph diversity in the rhizosphere of the smooth cordgrass, Spartina alterniflora. Appl Environ Microbiol 66, 3814–3822.[Abstract/Free Full Text]

Ludden, P. W. (1994). Reversible ADP-ribosylation as a mechanism of enzyme regulation in procaryotes. Mol Cell Biochem 138, 123–129.[Medline]

Mandon, K., Michel-Reydellet, N., Encarnacion, S., Kaminski, A., Leija, A., Cevallos, M., Elmerich, C. & Mora, J. (1998). Poly-{beta}-hydroxybutyrate turnover in Azorhizobium caulinodans is required for growth and controls nifA expression. J Bacteriol 180, 5070–5076.[Abstract/Free Full Text]

Mavingui, P., Flores, M., Guo, X., Davilla, G., Perret, X., Broughton, W. J. & Palacios, R. (2002). Dynamics of genome architecture in Rhizobium sp. NGR234. J Bacteriol 184, 171–176.[Abstract/Free Full Text]

Merrick, M. J. (1992). Regulation of nitrogen fixation genes in free-living and symbiotic bacteria. In Biological Nitrogen Fixation, pp. 835–876. Edited by G. Stacey, R. H. Burris & H. J. Evans. New York: Chapman and Hall.

Miller, J. (1972). Assay for {beta}-galactosidase. In Experiments in Molecular Genetics, pp. 352–355. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory.

Reyes-Ramirez, F., Little, R. & Dixon, R. (2001). Role of Escherichia coli nitrogen regulatory genes in the nitrogen response of the Azotobacter vinelandii NifL-NifA complex. J Bacteriol 183, 3076–3082.[Abstract/Free Full Text]

Rodrigue, A., Quentin, Y., Lazdunski, A., Méjean, V. & Foglino, M. (2000). Two-component systems in Pseudomonas aeruginosa: why so many? Trends Microbiol 8, 498–504.[CrossRef][Medline]

Roth, J. R., Lawrence, J. G. & Bobik, T. A. (1996). Cobalamin (coenzyme B12) : synthesis and biological significance. Annu Rev Microbiol 50, 137–181.[CrossRef][Medline]

Rubio, L. M., Rangaraj, P., Homers, M. J., Roberts, G. P. & Ludden, P. W. (2002). Cloning and mutational analysis of the {gamma} gene from Azotobacter vinelandii defines a new family of proteins capable of metallocluster binding and protein stabilization. J Biol Chem 277, 14299–14305.[Abstract/Free Full Text]

Rudnick, P., Kunz, C., Gunatilaka, M. K., Hines, E. R. & Kennedy, C. (2002). Role of GlnK in NifL-mediated regulation of NifA activity in Azotobacter vinelandii. J Bacteriol 184, 812–820.[Abstract/Free Full Text]

Saez, L. P., Garcia, P., Martinez-Luque, M., Klipp, W., Blasco, R. & Castillo, F. (2001). Role of draTG and rnf genes in reduction of 2,4-dinitrophenol by Rhodobacter capsulatus. J Bacteriol 183, 1780–1783.[Abstract/Free Full Text]

Sambrook, J., Fritsch, E. F. & Maniatis, T. (1989). Molecular Cloning: a Laboratory Manual, 2nd edn. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory.

Schmel, M., Jahn, A., Meyer zu Vilsendorf, A., Hennecke, S., Masepohl, B., Schuppler, M., Marxer, M., Oelze, J. & Klipp, W. (1993). Identification of a new class of nitrogen fixation genes in Rhodobacter capsulatus: a putative membrane complex involved in electron transport to nitrogenase. Mol Gen Genet 241, 602–615.[Medline]

Schmidhauser, T. J., Ditta, G. & Helinski, D. R. (1988). Broad-host-range plasmid cloning vectors for gram-negative bacteria. Biotechnology 10, 287–332.[Medline]

Siddavattam, D., Steibl, H. D., Kreutzer, R. & Klingmuller, W. (1995). Regulation of nif gene expression in Enterobacter agglomerans: nucleotide sequence of the nifLA operon and influence of temperature and ammonium on its transcription. Mol Gen Genet 249, 629–636.[Medline]

Sikorski, J., Rosello-Mora, R. & Lorenz, M. G. (1999). Analysis of genotypic diversity and relationships among Pseudomonas stutzeri strains by PCR-based genomic fingerprinting and multilocus enzyme electrophoresis. Syst Appl Microbiol 22, 393–402.[Medline]

Simon, R., Priefer, U. & Pühler, A. (1983). A broad host range mobilization system for in vivo genetic engineering: transposon mutagenesis in Gram-negative bacteria. Biotechnology 1, 784–791.

Toukdarian, A. & Kennedy, C. (1986). Regulation of nitrogen metabolism in Azotobacter vinelandii: isolation of ntr and glnA genes and construction of ntr mutants. EMBO J 5, 399–407.[Abstract]

Ueda, T., Suga, Y., Yahiro, N. & Matsuguchi, T. (1995). Genetic diversity of N2-fixing bacteria associated with rice roots. Can J Microbiol 41, 235–240.[Medline]

Vermeiren, H., Hai, W. L. & Vanderleyden, J. (1998). Colonization and nifH expression on rice roots by Alcaligenes faecalis A15. In Nitrogen Fixation with Non-Legumes, pp. 167–177. Edited by K. A. Malik, M. S. Mirza & J. K. Ladha. Dordrecht: Kluwer.

Vermeiren, H., Willems, A., Schoofs, G., de Mot, R., Keijers, V., Hai, W. & Vanderleyden, J. (1999). The rice inoculant strain A15 is a nitrogen-fixing Pseudomonas stutzeri strain. Syst Appl Microbiol 22, 215–224.[Medline]

You, C. & Zhou, F (1989). Non-nodular endorhizospheric nitrogen fixation in wetland rice. Can J Microbiol 35, 403–408.

You, C. B., Song, H. X., Wang, J. P., Lin, M. & Hai, W. L. (1991). Association of Alcaligenes faecalis with wetland rice. Plant Soil 137, 81–85.

Young, J. P. W. (1992). Phylogenetic classification of nitrogen-fixing organisms. In Biological Nitrogen Fixation, pp. 43–86. Edited by G. Stacey, R. H. Burris & H. J. Evans. New York: Chapman & Hall.

Zehr, J. & McReynolds, L. A. (1989). Use of degenerate oligonucleotides for amplification of the nifH gene from the marine cyanobacterium Trichodesmium thiebautii. Appl Environ Microbiol 55, 2522–2526.[Medline]

Zhang, Y., Pohlmann, E. L., Ludden, P. W. & Roberts, G. P. (2000). Mutagenesis and functional characterization of the glnB, glnA, and nifA genes from the photosynthetic bacterium Rhodospirillum rubrum. J Bacteriol 182, 983–992.[Abstract/Free Full Text]

Zhulin, I. B., Taylor, B. L. & Dixon, R. (1997). PAS domain S-boxes in Archaea, Bacteria and sensors for oxygen and redox. Trends Biochem Sci 22, 331–333.[CrossRef][Medline]

Received 29 January 2003; revised 6 May 2003; accepted 15 May 2003.