Nitrogen fixation island and rhizosphere competence traits in the genome of root-associated Pseudomonas stutzeri A1501

Abstract

The capacity to fix nitrogen is widely distributed in phyla of Bacteria and Archaea but has long been considered to be absent from the Pseudomonas genus. We report here the complete genome sequencing of nitrogen-fixing root-associated Pseudomonas stutzeri A1501. The genome consists of a single circular chromosome with 4,567,418 bp. Comparative genomics revealed that, among 4,146 protein-encoding genes, 1,977 have orthologs in each of the five other Pseudomonas representative species sequenced to date. The genome contains genes involved in broad utilization of carbon sources, nitrogen fixation, denitrification, degradation of aromatic compounds, biosynthesis of polyhydroxybutyrate, multiple pathways of protection against environmental stress, and other functions that presumably give A1501 an advantage in root colonization. Genetic information on synthesis, maturation, and functioning of nitrogenase is clustered in a 49-kb island, suggesting that this property was acquired by lateral gene transfer. New genes required for the nitrogen fixation process have been identified within the nif island. The genome sequence offers the genetic basis for further study of the evolution of the nitrogen fixation property and identification of rhizosphere competence traits required in the interaction with host plants; moreover, it opens up new perspectives for wider application of root-associated diazotrophs in sustainable agriculture.

The Pseudomonas genus belongs to the gamma subgroup of Proteobacteria. Until now, complete genome sequencing has been established for Pseudomonas aeruginosa [PAO1 (1) and PA14], Pseudomonas fluorescens [Pf-5 (2), PfO-1 and SBW25], Pseudomonas putida KT2440 (3), Pseudomonas syringae [DC3000, B728a (4), and 1448A], and Pseudomonas entomophila (5). Although phylogenetically close to P. aeruginosa, Pseudomonas stutzeri belongs to the group of nonfluorescent Pseudomonas. The taxonomic status and biology of this species, isolated from a large diversity of terrestrial and marine environments, have been recently reviewed (6). Some strains have attracted particular attention because of specific metabolic properties, such as denitrification, degradation of aromatic compounds, and synthesis of polyhydroxyalkanoates (6). A1501 (China General Microbiological Culture Collection Center accession no. 0351), formerly known as Alcaligenes faecalis A15, was isolated from rice paddy soils and has been widely used as a crop inoculant in China (7, 8). This strain belongs to genomovar 1 of P. stutzeri (9). Until recently, the ability to fix nitrogen was not recognized within the Pseudomonas genus sensu stricto. This property has now been reported in a few strains (6, 9, 10) and was best documented in the case of strain A1051 (11).

Biological nitrogen fixation, a key step in global nitrogen cycling, is the ATP-dependent reduction of dinitrogen to ammonia by the nitrogenase enzyme complex. The ability to fix nitrogen is widely distributed among Bacteria and Archaea. In addition to symbiotic nitrogen-fixing microorganisms, which induce differentiated structures on the host plant (root nodules of legumes and actinorhizal plants), there exist a variety of free-living nitrogen-fixing bacteria capable of association with the root system of graminaceous plants, such as Klebsiella pneumoniae, Azotobacter vinelandii, and Azospirillum brasilense (12). New species, such as Azoarcus and Gluconacetobacter, can also develop endophytic associations and survive within plant tissues without causing disease symptoms (12, 13).

P. stutzeri A1501 can survive in the soil, colonize the root surface, and invade the superficial layers of the root cortex (7, 8, 14). This is why it was also reported as being an endophyte (14, 15). We report here the analysis of the complete genome of A1501 and compare it with other Pseudomonas species. We also report clustering of the nitrogen fixation region in a putative nitrogen fixation island and the identification of several new genes required for optimal nitrogenase activity. The genome survey also revealed the presence of a number of interesting gene clusters that may be involved in rhizosphere competence and colonization of the host plant.

Results

General Features of the Genome and Comparative Genomics.

The P. stutzeri A1501 genome is composed of a single circular chromosome of 4,567,418 bp, encoding 4,146 probable proteins, 59 tRNA genes, and four rRNA operons (Fig. 1, Table 1). No intact (or nearly intact) prophages were found in the genome of P. stutzeri A1501, but it carries six genes encoding site-specific recombinases of the phage integrase family. We also identified a total of 42 copies of repeat sequences (covering 1.35% of the genome) grouped into 10 types encoding 57 transposases. Four distinct regions with atypical GC content and flanked by tRNA genes (Fig. 1) had general characteristics of typical genomic islands (16), suggesting recent transfer of genetic material into A1501. A fifth region containing the nif genes could also be considered a genomic island, as discussed later.

Fig. 1.

Circular representation of the P. stutzeri A1501 genome and gene expression profiles in nitrogen fixation conditions in comparison with nitrogen excess conditions. Outer scale is marked in 200 kb. (From the outer to the inner concentric circle) Circles 1 and 2, genes encoded by leading and lagging strands, respectively. Coding sequences are color-coded by COG categories. Circle 3, log₂ ratio of the expression level under nitrogen fixation conditions versus ammonia excess conditions for each gene; values are color-coded as shown in the right bottom bar. Circle 4, location of the genomic islands. Green, PST0244 to PST0260; yellow, PST0573 to PST0711; blue, PST1463 to PST1490; pink, PST3342 to PST3471; red, nif island PST1302 to PST1359. Circle 5, transposase, recombinase, and integrase genes. Circles 6 and 7, G + C content and GC skew (G − C/G + C), respectively, with a 10-kb window size. Circles 8 and 9, distribution of tRNA genes and rrn operons, respectively.

Table 1.

General features of Pseudomonas genomes

Species	Size, Mb	CDS, no.	G + C content, %	Coding density, %	tRNA, no.	rRNA operon, no.	Plasmid, no.	Ohologs shared with A1501, no.	Average similarity of orthologs, %
Pst	4.56	4,146	63.8	89.8	59	4	—	—	—
Pae	6.26	5,570	66.6	89.4	63	4	—	2,770	72.2
Pfl	7.07	6,144	63.3	88.8	71	5	—	2,674	70.6
Psy	6.39	5,763	58.4	86.8	63	5	2	2,472	70
Ppu	6.18	5,420	61.6	87.7	74	7	—	2,604	69.9
Pen	5.88	5,169	64.2	89.1	78	7	—	2,531	70.4

Pst, P. stutzeri A1501; Pae, P. aeruginosa PAO1; Pfl, P. fluorescens Pf-5; Psy, P. syringae pv. tomato DC3000; Ppu, P. putida KT2440; Pen, P. entomophila L48.

The size of the genome was much smaller than those reported for other Pseudomonas species sequenced to date. Deduced translation products of A1501 were compared with the predicted proteomes of the five other representative Pseudomonas species to identify interspecies ortholog pairs (Table 1), and orthologous gene synteny was found to be severely disordered [supporting information (SI) Fig. S1]. A1501 exhibited the highest overall similarity to P. aeruginosa compared with other species, because 66.8% (2, 770) of the A1501 genes had counterparts in the P. aeruginosa PAO1 genome. This result is consistent with phylogenetic analysis based on 12 housekeeping proteins in the 11 sequenced Pseudomonas strains (Fig. S2). Based on genome comparison, the number of genes previously recognized as being part of the Pseudomonas core genome may be limited to 1,997. Indeed, most of the genes related to virulence and pathogenicity in P. aeruginosa PAO1 were absent in A1501, such as type III/VI secretion systems, synthesis of both types of quorum sensing molecules, alginate polymer synthesis, siderophores, and antibiotic biosynthesis pathways (1, 17).

Nitrogen Fixation.

All genetic information specific to nitrogen fixation was clustered in the A1501 genome in a 49-kb nitrogen fixation region (PST1302–PST1359) consisting of 59 genes. The G + C content of this region was higher than the average of the entire genome (66.8% vs. 63.8%) (Fig. 1). This favored the hypothesis that the nif region in P. stutzeri constitutes a nitrogen fixation island, as discussed in other nitrogen fixers (16, 18, 19). For example, nif genes are part of an island in Wolinella succinogenes (16, 19), and the existence of several genomic islands in Rhizobium leguminosarum was based on differences in G + C content (18). The nif region in A1501 is flanked by genes coding for cobalamin synthesis (cobalamin 5′-phosphate synthase, PST1301) on one side and glutathione peroxidase (PST1360) on the other (Fig. S3). High gene conservation in the flanking regions was noted in other Pseudomonas species (unable to fix nitrogen), raising the possibility that there exists an insertion hotspot flanked by cobalamin 5′-phosphate synthase and glutathione peroxidase in Pseudomonas species (Table S1) in which the nitrogen fixation island in A1501 was acquired from lateral gene transfer.

The nitrogenase complex comprised the MoFe-protein encoded by nifDK and the Fe protein encoded by nifH. Full assembly of the complex required the products of a dozen other nif genes extremely conserved in free-living and symbiotic diazotrophs, in particular for processing of nitrogenase metalloclusters and catalytic stability (nifMZ, nifW, and nifUS) and for synthesis of a specific molybdenum cofactor (FeMo-co) bound to the MoFe protein (nifB, nifQ, nifENX, nifV, and nifH) (20). Although these genes are common to most systems, their organization, content, and regulation differ between genera.

We compared the organization of A1501 nif genes with those of four other well characterized nitrogen-fixing bacteria relatively phylogenetically close to Pseudomonas, A. vinelandii AvOP (21), K. pneumoniae M5a1 (22), A. brasilense Sp7 (12), and Azoarcus sp. BH72 (13). The general organization of the nif genes in A1501 showed a high degree of similarity to that of A. vinelandii, except that the nif genes were not contiguous in A. vinelandii, but were distributed into two portions of the genome (Fig. 2). A first group of genes (from PST1302 to PST1323 in A1501) contained nifB and nifQ involved in synthesis of FeMo-co and nifLA. NifL and NifA were the negative/positive regulatory proteins for nif operon expression (23, 24). Four additional genes associated with nifB and nifQ (Fig. 2) were also conserved in A. vinelandii (21). The rnfABCDGEF operon, which encodes subunits of a membrane-bound protein complex involved in electron transport to nitrogenase (25), was located next to nifLA in AvOP, BH72, and A1501 (11). A second group (contiguous to group 1 in the case of A1501) contained a set of genes (nifHDKTY, nifNEX, nifUSV, and nifWZMF) in the same order in A. vinelandii, K. pneumoniae, and A1501. These genes were contiguous in K. pneumoniae, whereas additional coding sequences (CDSs) were found interspaced between or within the nif operons both in A. vinelandii and A1501. In A1501, very few nucleotides were present in the noncoding regions between genes, and translational overlapping between start and stop codons of some of the genes was observed (Fig. S4). For example, the 10 genes from PST1349 (hesB-like gene) to PST1358 were likely to be part of the same 7.5-kb transcriptional unit, suggesting a superoperon with translational coupling. Although nitrogenase on-off switching was reported in A1501 (11), draT and draG genes known to control the nitrogenase ADP ribosylation process were not found in A1501, suggesting an inactivation mechanism different from that reported in Azoarcus and A. brasilense (12, 13).

Fig. 2.

Schematic representation of the P. stutzeri A1501 nif gene cluster compared with syntenic nif clusters of A. vinelandii AvOP, K. pneumoniae M5a1, Azoarcus sp. BH72, and A. brasilense Sp7. Connecting lines indicate corresponding orthologous gene clusters with high homology at the protein level. Identically colored segments indicate orthologs. Open boxes represent additional genes of unknown function outside or inside nif gene clusters. Red wedges indicate location of additional genes. fdxN between nifB (PST1306) and PST1305 corresponds to PST1305a.

Transcriptomic analysis of the whole A1501 genome was implemented by using genome-wide microarrays. Significant changes were observed in expression of a number of genes. Among these, 255 genes, including most of the genes within the nitrogen fixation island, showed increased transcription (ratio ≥2) under nitrogen fixation conditions compared with nitrogen excess conditions (Fig. 1 and Table S2). This suggested that some of the CDSs interspaced between nif genes might be involved in nitrogen fixation. We further chose 16 uncharacterized genes within the nif region. Up-regulation under nitrogen-fixing conditions was confirmed by quantitative real-time RT-PCR (using nifH as a control), whereas mutant construction by gene disruption (enabling transcription of downstream genes) led to a significant decrease in nitrogenase activity for all except two of them in PST1308 and PST1331 (Table 2 and Fig. S4). This suggests that the corresponding gene products, although not essential, are required for full nitrogenase activity under the assay conditions used.

Table 2.

Identification of 16 uncharacterized genes within the nitrogen fixation island with increased expression under nitrogen fixation conditions and nitrogenase activity of corresponding mutant strains

Gene ID	Probable function (genes)	Involvement in nitrogen fixation in other systems	Nitrogenase activity, %	Up-regulation factor in microarrays*	Up-regulation factor in real time RT-PCR†
PST1302	Glutaredoxin-related protein	Unknown	76	++++	23.95 ± 4.99
PST1303	Thiosulfate sulfurtransferase	Unknown	27	++++	15.08 ± 2.32
PST1305	Arsenate-reductase-related protein	Unknown	52	++++	28.46 ± 9.56
PST1308	LysR, transcriptional factor	Unknown	91	+	2.52 ± 0.54
PST1312	Thiopurine S-methyltransferase (tmpA)	Unknown	60	+	2.82 ± 0.13
PST1325	Unknown	Unknown	24	+++	27.86 ± 7.92
PST1326	Nitrogenase Fe protein (nifH)‡	Yes	0	++++	132.5 ± 18.3
PST1331	Unknown	Unknown	86	+++	10.09 ± 1.14
PST1332	Unknown	Unknown	60	+	3.69 ± 0.78
PST1336	Unknown	Unknown	39	++	3.71 ± 1.23
PST1337	Unknown	Unknown	72	++++	15.28 ± 1.70
PST1342	Unknown	Unknown	57	+	3.31 ± 0.23
PST1343	Unknown	Unknown	32	+	4.01 ± 0.21
PST1344	Unknown	Unknown	66	++	18.91 ± 2.23
PST1349	Fe-S cluster assembly protein (hesB)	Yes	2	++++	221.51 ± 24.40
PST1353	Serine acetyltransferase (cysE-like)§	Yes	19	++++	4.42 ± 0.12
PST1354	Unknown	Unknown	68	+++	19.31 ± 1.29

All uncharacterized genes have counterparts in A. vinelandii, except for PST1312. Nitrogenase activity of mutants (enabling transcription of downstream genes) is expressed as percentage of wild-type A1501 activity.

*+, 2–4; ++, 4–8, ++; +++, 8-16; ++++, >16.

^†Values are mean ± SD.

^‡nifH was used as control.

^§Also designated nifP or iscA.

Rhizosphere Competence Traits and Root Surface Colonization.

P. stutzeri A1051 is a soil bacterium, but it can colonize rice roots in an aquatic environment (7, 8, 14). It can also colonize wheat (14). Bacteria are found in the root tip mucilage, in root hair, at the emergence of lateral roots and, in some cases, in the first layers of the root cortex (8, 14). Analysis of the genome led to identification of genes coding for different properties that presumably permit A1501 to adapt to the highly competitive environment in the rhizosphere (26).

Some specific metabolic capacity may contribute to fitness in diverse environment and the genome carries a wide spectrum of genes for carbon and nitrogen source utilization, including a large number of transporter genes. Transport systems are key components in communication of bacteria with their environment. A1501 possesses >300 transporter-encoding genes, including at least 138 members of the ATP binding cassette (ABC) transporter superfamily (Fig. 3). This number is in the range of that observed in P. aeruginosa and P. putida. However, 87 genes (including 33 genes for ABC-type transporters) present in A1501 were not found in other Pseudomonas genomes. We found many transport systems for carbohydrates, amino acids and dicarboxylic acids, which are the major components of root exudates (Fig. 3). Acquisition of iron is an important trait for rhizosphere competition (27). A1501 does not carries pathways for the synthesis of siderophores, and it contains fewer TonB-dependent siderophore receptors than other Pseudomononas species. (Table S3). A1501 should get iron from siderophores produced by other soil bacteria in the rhizosphere, a feature also reported in Azoarcus (13).

Fig. 3.

Schematic overview of metabolic pathways and transport systems in P. stutzeri A1501. Gene identifiers are shown by numbers in green. Predicted transporters are grouped by substrate specificity: light blue, inorganic ions; emerald green, sugar; dark blue, organic sources. Reactions for which no candidate enzyme was confidently predicted are indicated by dashed arrows. Pathways that involve multiple reactions are shown by double arrows. Final biosynthetic products are indicated with pink boxes. Arrows indicate direction of transport. Crosses indicate pathways or reactions that are apparently not present in strain A1501.

Pathways for aromatic degradation through the protocatechuate (pca) and catechol (ben) routes found in the A1501 genome are similar to those found in other Pseudomonas, such as P. entomophila (5) and P. putida (28). Using in vivo expression technology, one of the pca gene was reported to be up-regulated in the rhizosopere of the P. stuzeri A15 (14). In general, P. stutzeri strains are not reported to accumulate polyhydroxyalkanoates (PHA) granules, except for one strain, P. stutzeri 1317 (6, 29). The A1501 genome contains phaCABR genes involved in the three-step polyhydroxybutyrate biosynthetic pathway, with phaR coding for a transcriptional regulator, but it contains a single phasin gene (phaP), in contrast to Ralstonia, which carries many (30).

Denitrification is another general property of many Pseudomonas species. The organization of genes involved in nitrate metabolim in A1501 is close to that of the P. stutzeri Zobell strain (31), with a super cluster containing nor, nir, and nos genes (Table S4), except that A1501 does not harbor the regulatory gene nirR. A recent report revealed that a mutant of narG, coding for a component of nitrate reductase, was more competitive than the wild-type strain in colonization of rice and wheat (32).

Analysis of the genome revealed that A1501 has a number of genes reported to play a role in osmotolerance (33). In particular, it carries genes coding periplasmic glucan biosynthesis proteins MgoG and MgoH found in most aerobic bacteria. Except for betA and betB, other genes of the glycine-betaine pathway found in other Pseudomonas species are missing in A1501. The genome carries an ectABC cluster responsible for synthesis and accumulation of ectoine in the moderate halophile Halomonas but (except for ectB) absent in other Pseudomonas. The genome contains a number of genes that can play a role in detoxification of reactive oxygen species commonly found in Pseudomonas genomes. It includes five catalases, two superoxide dismutases, four peroxidases, four hydroperoxide reductases, and nine glutathione S-transferases (34) (Table S5).

Chemotaxis and bacterial surface components (such as flagella, pili, and surface polysaccharides) are of importance in adhesion to a surface and biofilm formation (12, 26). A1501 has a complex chemosensory protein system by which the bacterium can sense and respond to the presence of certain substances, with 31 genes coding for methyl accepting receptors widespread over the genome, a situation close to that of P. putida KT2440 (3). It also carries 13 genes for signal excitation and adaptation and methyl removal proteins (che) and clusters containing flg and fli genes responsible for flagella biogenesis and motility. Indeed, an rpoN mutant that was devoid of flagella displayed less surface colonization than the wild type (12). A set of 28 genes responsible for type IV pili biogenesis are present in A1501, including genes known to be required for natural DNA transformation in other P. stutzeri strains (pilA, comL, pilE, pilC, pilT, and pilU) (35). However, type IV pili are also required for twitching motility, and pilA was found to be essential for root surface colonization and for infection of plant tissue in Azoarcus (13).

Alginate is the major exopolysaccharide found in fluorescent Pseudomonas. However, except for algA (phosphomannose isomerase) and algI (acetylase), all of the other genes involved in alginate biosynthesis are missing in A1501, whereas most of the regulatory genes known to control its biosynthesis (36) are present in A1501 (Table S6). This suggests that the ability to synthesize alginate was lost in strain A1501. We noted a cluster of genes for cellulose biosynthesis in A1501 sharing some similarities with that of P. putida KT2440 that may play a role in attachment to the plant surface. There also exists a gene (PST2494) encoding a hydrolytic enzyme resembling an endoglucanase, which is known to play a role in endophytic colonization in Azoarcus (13); however, no gene coding for pectinolytic activity, which is often found in pathogens, was detected in the A1501 genome. We have not identified genes possibly involved in indole-3-acetic acid (IAA) synthesis. But the genome contains acdS, which codes for a deaminase of the precursor of ethylene, a phytohormone known to inhibit root elongation. This gene is a good candidate for involvement in plant growth promotion by preventing ethylene formation (12, 37).

Discussion

The P. stutzeri A1501 genome is much smaller than that of fluorescent Pseudomonas species sequenced to date (Table 1). This might be due to reductive evolution and could reflect a more stringent association with its plant host, although smaller size is not specific to A1501 and was reported for other P. stutzeri strains that do not necessarily have the same lifestyle (38). In addition, P. putida KT2440, which is also a rhizospheric strain, has a much larger genome (3, 39) (Table 1). We identified a set of 487 genes common to other Pseudomonas and absent in the A1501 genome, suggesting loss of functions in A1501, as noted for the alginate biosynthesis gene cluster and for fluorescent siderophore production.

Lateral gene transfer appears to be a mechanism for continuous adaptation of soil bacteria to environmental changes (16). The A1501 genome contains four typical genomic islands (Fig. 1), suggesting recent transfer. Some genes associated with these genomic islands encode conserved proteins sharing similarity with transport and heavy metal resistance proteins, whose functions may be related to better adaptation to the rhizosphere. We also noted that genes for type I restriction-modification (R-M) systems (Table S7) were present in three of the genomic islands but nowhere else in the genome, suggesting recent acquisition of R-M by lateral gene transfer.

The presence of nif genes in this particular strain of P. stutzeri, when most other Pseudomonas species lack this property, raises the question of their origin. The nitrogen fixation process is believed to have played a crucial role in early cellular evolution, in particular when the geochemical reserves of fixed nitrogen in the biosphere became depleted. Because nitrogen fixation is maintained in Bacteria and Archaea, it is often hypothesized that nif genes may originate from the last common ancestor, even though other scenarios are proposed (40, 41). It can be presumed that nif genes were lost in most lineages of Bacteria and Archaea and that lateral gene transfer played a role in recent acquisition of nif genes in some lineages (16, 42). The rareness of nitrogen fixation properties in true Pseudomonas suggests that nif genes were lost during evolution in most (if not all) Pseudomonas species. Packing of the nif genes in A1501, and the difference in G + C content from the rest of the genome, favor the hypothesis that the 49-kb nif region is a genomic island. The K. pneumoniae 24-kb nif gene cluster comprises the tightest organization of nif genes identified to date (22). In all other systems studied, additional genes associated with core nif genes have been identified, but in most cases, their role in nitrogen fixation remains hypothetical (20). From the data reported here, it is likely that many of the associated genes play a role in some steps of the nitrogen fixation process. We presume that many of these genes, which do not exist in other systems except for Azotobacter, might encode functions for adapting the nitrogen fixation process, such as electron transport chains to nitrogenase and oxygen protection mechanisms. Because the physiology of nitrogen fixation is more restrictive in K. pneumoniae (anaerobiosis) than in P. stutzeri (microaerobiosis) and A. vinelandii (aerobiosis), it is tempting to speculate that there exists lateral transfer of the nif gene from a common ancestor and subsequent acquisition of extra genes so as to adapt to the physiology of the respective hosts.

We have identified, in the genome of A1501, a large number of genes that might be responsible for survival in the soil and colonization of the host plant. Except for a few of them (narG, pcA, and rpoN) identified in refs. 12 and 14, their actual involvement in colonization remains to be established. Recent work in the rhizospheric P. pudtia KT2440 identified 91 genes up-regulated in the presence of maize exudates (39), and 31 of them are conserved in A1501 (including regulators, transport systems, stress response, and chemotaxis) (data not shown). Present knowledge of the genome should help in elucidating regulatory mechanisms that control the association with the host plant, thus enabling better comprehension of mechanisms of nitrogen transfer. An understanding of the mechanisms by which root-associated diazotrophs improve productivity of nonleguminous crops is an important issue. Our research is aimed at improving associative nitrogen fixation for sustainable agriculture and for reducing crop chemical fertilizer requirements. Finally, the availability of the A1501 genome sequence provides insights into the evolution of nitrogen-fixing organisms and opens up new perspectives for molecular interaction between the microbe and host plant.

Methods

Genome Sequencing.

The genome sequence was determined by using whole-genome shotgun strategy as described in ref. 43. Plasmid libraries with one small insert (2–4 kb) and one large insert (4–6 kb) were constructed in an alkaline-phosphatase-treated pUC19 vector after random mechanical shearing (ultrasonic nebulization) of genomic DNA. Finally, 50,408 valid reads and 30.7 Mb of total length were established, providing 7-fold genome coverage. For a detailed description of genome assembly, annotation and comparisons, see SI Methods.

Microarray.

The microarray used in this study featured 3,988 of the 4,146 CDSs identified in P. stutzeri A1501. Full details are in SI Methods.

Quantitative Real-Time RT-PCR.

Quantitative real-time PCR experiments were performed with three independent RNA preparations, using the 7000 Sequence Detection system (Applied Biosystems) and SYBR Green PCR Master Mix (Applied Biosystems) according to the manufacturer's recommendations. First-strand cDNAs were synthesized from 5 μg of total RNA in a 20-μl reaction volume, using the SuperScript first-strand synthesis system for RT-PCR (Invitrogen).

Construction of Mutants.

Uncharacterized genes within the nif island were inactivated by homologous suicide plasmid integration (44), using pK18mob as a vector. DNA fragments of ≈200 bp were amplified by using total DNA of A1501 as template and oligonucleotide primers designed to generate amplicons for the creation of mutations, enabling the transcription of downstream genes. Amplicons were ligated into pK18mob, and the resulting plasmids were introduced into P. stutzeri A1501; recombination at the correct location was checked by PCR. Nitrogenase activity was assayed with bacterial suspensions incubated at an OD₆₀₀ of 0.1 in N-free minimal K medium at 30°C under an argon atmosphere containing 0.5% oxygen and 10% acetylene according to the protocol described in ref. 11. Nitrogenase activity is expressed as nanomoles of ethylene per min per (mg protein).