Characterization of a Major Cluster of nif, fix, and Associated Genes in a Sugarcane Endophyte, Acetobacter diazotrophicus

Abstract

A major 30.5-kb cluster of nif and associated genes of Acetobacter diazotrophicus (syn. Gluconacetobacter diazotrophicus), a nitrogen-fixing endophyte of sugarcane, was sequenced and analyzed. This cluster represents the largest assembly of contiguous nif-fix and associated genes so far characterized in any diazotrophic bacterial species. Northern blots and promoter sequence analysis indicated that the genes are organized into eight transcriptional units. The overall arrangement of genes is most like that of the nif-fix cluster in Azospirillum brasilense, while the individual gene products are more similar to those in species of Rhizobiaceae or in Rhodobacter capsulatus.

Biological nitrogen fixation occurs in species of more than 100 genera distributed among several of the major phylogenetic divisions of prokaryotes (Eubacteria and Archaea) (25). Sequence and mutational analyses of the genes necessary for nitrogen fixation (nif) in many diazotrophs indicate that their products have common structures and functions, while the degree of linkage and arrangement of specific nif and associated genes vary considerably (5, 8, 17). In addition, nif genes and genes involved in plant invasion and nitrogen fixation effectiveness, such as nod and fix in species of Rhizobiaceae, are often linked.

The identification of nitrogen-fixing bacteria with endophytic habitats raises the possibility of a new classification of symbiosis (3). The relationship of a proteobacterial α group member, Acetobacter diazotrophicus (syn. Gluconacetobacter diazotrophicus), with sugarcane represents a promising model system for the study of an association between a monocot and an endophytic nitrogen-fixing bacterium (13, 22). The ability of A. diazotrophicus to enhance sugarcane growth has been documented, and while the benefit to plant growth might be due at least in part to the transfer of bacterially fixed N, another plant growth-stimulating factor(s) is indicated, possibly auxin production by A. diazotrophicus (22). Because of its potential agronomic use and unique status as the only diazotrophic species of Acetobacter so far identified, it was of interest to isolate and characterize genes that are involved in nitrogen fixation and regulation.

Identification of a major cluster of nif and associated genes.

A genomic library of A. diazotrophicus constructed in the wide-host-range cosmid pLAFR3 (Tet^r) (4) was transferred by conjugation from Escherichia coli to several different nif mutants of Azotobacter vinelandii (12). Two mutants that yielded Nif⁺ Tet^r transconjugants were DJ71 (nifV) and DJ35 (nifE), and several transconjugants of both strains carried the same cosmid, pAD71, the 22-kb insert fragment of which carries nifD through mcpA (Fig. 1). One end of the insert fragment contained sequences identical to a portion of the nifHDK region cloned previously (M. Sevilla and C. Kennedy, submitted for publication) and also to a portion of pAD101, another cosmid from the pLAFR3 library, isolated by its ability to complement nifA mutants of A. vinelandii (24). A third cosmid, pAD102, which also complemented nifA mutants, provided the region further upstream of nifA.

FIG. 1.

Major nif gene map of A. diazotrophieus. Arrows indicate the positions and direction for transcription initiation sites. •→, ς⁵⁴-, NifA-dependent promoters. The thick lines underneath represent the DNA fragments used as probes in Northern analysis.

Thirty-two open reading frames (ORFs) were identified in the>30 kbp of sequence obtained (Fig. 1). Gene characteristics, potential regulatory sequences, and gene products are summarized in Table 1. Genes not known to be directly involved in nitrogenase structure, function, or regulation of gene expression include mcpA, encoding a methyl-accepting chemotaxis protein that responds to extracellular signals for chemotactic responses (19), and modABCD, which encode a high-affinity molybdate transport system in E. coli (9), Staphylococcus carnosus (20), and A. vinelandii (18). FixABCX may comprise an electron transfer chain, and, while the fixABCX genes were first identified in Rhizobium meliloti (6) and subsequently in other diazotrophs, some or all are also present in the genomes of E. coli (fixABCX) and Mycobacterium tuberculosis (fixAB), neither of which is a diazotroph. The Fix proteins in E. coli are required for carnitine breakdown, an anaerobic function related to respiration (2), while their function in other organisms is unknown (7). Also of unknown function but similar to gene products in other diazotrophs are the ferredoxins FdxN and FdxB and the products of ORFs 1, 4, and 6, named on the basis of ORF numbers originally assigned. The only ORF without significant similarity to known genes is the one between rpoN and modC, orf11. The nif-fix and associated gene cluster characterized here for A. diazotrophicus is unique in representing the largest single grouping of genes required for nitrogenase structure and function, nif and fix gene regulation (nifA and rpoN), and associated functions (molybdenum uptake and electron transfer) found in any diazotroph so far studied.

TABLE 1.

Comparison of nif-associated gene products

Genea	Function	Regulatory feature	Product size (kDa)	Organisms with most similar gene productsb (% amino acid identity c)
rpoN	ς54		51.3	B. japonicum (34), A. caulinodans (31), C. crescentus (30)
modC	Molybdenum transport; ATP-binding protein		36.8	R. capsulatus(48), A. vinelandii (47), E. coli (46)
modB	Molybdenum transport; permease protein		24.7	R. capsulatus (57), A. vinelandii (40), A. aeolicus (35)
modA	Molybdenum transport; molybdate-binding periplasmic protein		26.9	R. capsulatus (42), A. vinelandii (40), A. aeolicus (32)
orf11	Unknown	NifA-, ςn-binding sites	20.2
nifA	Transcriptional activator	NifA-binding sites	62.2	A. caulinodans (51), B. japonicum (50), R. etli (47)
nifB	Fe-Mo cofactor synthesis	NifA-, ςn-binding sites	51.9	R. capsulatus (67), R. meliloti (67), Rhizobium sp. strain NGR234 (66)
fdxN	Ferredoxin		6.8	R. palustris (77), Rhizobium sp. strain NGR234 (67), R. meliloti (67)
ynfl	Unknown		26.4	R. capsulatus (31)
nifZ	Maturation and activation		6.9	Rhizobium sp. strain NGR234 (63), R. capsulatus (59), E. agglomerans (47)
fixU	Unknown		8.6	R. leguminosarum (56), Rhizobium sp. strain NGR234 (55)
orfl2	Unknown		36.4	A. brasilense (42), A. vinelandii (28)
nifH	Nitrogenase structure; Fe protein	NifA-, ςn-binding site	31.9	R. phaseoli (89), Rhizobium sp. strain NGR234 (85), R. meliloti (85)
nifD	Nitrogenase structure; Fe-Mo protein alpha subunit		55.9	Rhizobium sp. strain NGR234 (80), P. rhizobium (75), H. seropedicae (75)
nifK	Nitrogenase structure; Fe-Mo protein beta subunit		57.2	Rhizobium sp. strain NGR234 (71), B. japonicum (66), T. ferrooxidans (62)
nifE	Fe-Mo cofactor synthesis	ςn-binding site	51.8	B. japonicum (60), Rhizobium sp. strain NGR234 (58), R. capsulatus (57)
nifN	Fe-Mo cofactor synthesis		49.5	R. meliloti (50), Rhizobium sp. strain NGR234 (49), B. japonicum (46)
nifX	Fe-Mo cofactor synthesis		17.4	Rhizobium sp. strain NGR234 (55), R. capsulatus (46), H. seropedicae (37)
orf4	Unknown		16.9	A. caulinodans (53), Frankia spp. (50), F. alni (45)
orf1	Unknown		7.6	R. capsulatus (50), Rhizobium sp. strain NGR234 (48), Anabgena PCC7120 (39)
fdxB	Ferredoxin		9.9	R. capsulatus (41), Rhizobium sp. strain NGR234 (41), P. boryanum (34)
nifQ	Fe-Mo cofactor synthesis	ρ-independent terminator	22.9	R. capsulatus (41), A. vinelandii (40), Rhizobium sp. strain NGR234 (33)
orf6	Unknown	NifA-, ςn-binding sites	10.9	R. capsulatus (48), R. sphaeroides (46), B. japonicum (45)
nifU	Maturation and activation; assembly of iron-sulfur clusters		32.9	A. vinelandii (55), A. brasilense (53), A. chroococcum (52)
nifS	Maturation and activation; homodimeric cysteine desulfurase		43.0	A. vinelandii (59), Rhizobium sp. strain NGR234 (59), R. sphaeroides (59)
nifV	Fe-Mo cofactor synthesis; homocitrate synthase		92.2	Z. mobilis (61), R. sphaeroides (51), F. alni (49)
nifW	Maturation and activation; oxygen protection of the Mo-Fe protein	ρ-independent terminator	12.0	A. caulinodans (48), Rhizobium sp. strain NGR234 (42), A. brasilense (41)
fixA	Electron transfer; electron transfer flavoprotein beta subunit	NifA-, ςn-binding sites	30.7	A. caulinodans (69), Rhizobium sp. strain NGR234 (67), B. japonicum (66)
fixB	Electron transfer; electron transfer flavoprotein alpha subunit		38.9	A. caulinodans (68), Rhizobium sp. strain NGR234 (67), B. japonicum (67)
fixC	Electron transfer; electron transfer flavoprotein-quinone oxidoreductase		48.1	Rhizobium sp. strain NGR234 (63), B. japonicum (62), A. caulinodans (61)
fixX	Electron transfer; ferredoxin-like protein		10.7	R. leguminosarum (73), R. meliloti (71), Rhizobium sp. strain NGR234 (70)
modD	Molybdenum transport		26.7	Z. mobilis (51), R. capsulatus (41), E. coli (39)
mcpA	Chemotaxis		69.3	C. crescentus (63), R. sphaeroides (51), R. leguminosarum (44)

^aOrganization of gene expression is indicated by arrows in Fig. 1. 

^bOrganism in which the gene product most similar to that of A. diazotrophicus was found. Organisms: Aquifex aeolicust, Azorhizobium caulinodans, Azotobacter chroococcum, Azotobacter vinelandii, Bradyrhizobium japonicum, Caulobacter crescentus, Escherichia coli, Frankia alni, Herbaspirillum seropedicae, Parasponia rhizobium, Plectonema boryamum, Rhizobium etli, Rhizobium leguminosarum, Rhizobium meliloti, Rhizobium phaseoli, Rhodobacter capsulatus, Rhodobacter sphaeroides, Rhodopseudomonas palustris, Thiobacillus ferrooxidans, and Zymomonas mobilis. 

^cIdentity of the deduced A. diazotrophicus gene product sequence to the gene product of the three organisms to which it is most related. 

Comparison of organization of nif genes and gene products.

The individual A. diazotrophicus gene products are generally most similar to those found in other α group proteobacteria, with 17 gene products being most like those in members of the Rhizobiaceae and 9 gene products being most closely related to Rhodobacter capsulatus proteins (Table 1). NifU and NifS were most similar to the gene products of Azotobacter species, members of the γ group of proteobacteria. McpA was most similar to the mcpA gene product of the unrelated Caulobacter crescentus (67% identity).

Although none of the individual gene products were most similar to those of Azospirillum species, the overall organization of genes in the A. diazotrophicus nif-fix cluster is most like that of Azospirillum brasilense. These are the only two diazotrophs so far characterized that have an mcpA-like gene associated with the nif-fix cluster (J. Frazzon and I. S. Schrank, personal communication; also this work). The McpA protein is involved in chemotaxis in several organisms. Because both A. diazotrophicus and A. brasilense are found naturally associated with monocot plants, it will be of interest to determine whether McpA is responsible for chemotactic responses to plant exudates for signaling or nutrition. An mcpA mutant strain of A. diazotrophicus lost chemotaxis toward a wide range of attractant stimuli (unpublished result).

Differences between the clusters are that nifA and nifB of A. brasilense are not linked to the other nif genes and nifQ, nifW, and fdxB are absent from the latter (8, 15). nifY is not found in the A. diazotrophicus cluster, but its requirement for nitrogen fixation is uncertain in other diazotrophs; it may be involved in Fe-Mo cofactor insertion into nitrogenase enzyme or in fixed N sensing (11, 23). One portion of the nif-fix cluster of A. diazotrophicus is more like that of R. capsulatus (nifE to nifW) than to A. brasilense, although no ORF1-like gene is found in the latter. The nifU gene in R. capsulatus is truncated compared to most other nif genes, encoding only the C-terminal end of NifU (16).

Transcriptional and translational organization of the cluster.

The transcriptional and translational organization of genes in the nif-fix cluster of A. diazotrophicus shows several interesting features. Northern analysis of mRNA was successful in identifying the cotranscription of nifHDK, as occurs in most other diazotrophs, according to size of transcripts hybridizing to a nifD probe, and also of nifENX orf4 orfl fdxB nifQ, indicated by hybridization of RNA to a nifE probe (Fig. 1 and 2). The former was predicted by sequence analysis, which revealed ςⁿ- and NifA-binding sites upstream of nifH (Sevilla and Kennedy, submitted) required, respectively, for nif promoter recognition and for nif gene transcriptional activation in all other proteobacterial diazotrophs. A ςⁿ recognition sequences but no NifA recognition sequence was found upstream of nifE. Northern blots hybridized with [³²P]dCTP-labeled orf6, nifU, or fixA probes showed only smears or broad bands after autoradiography, suggesting rapid degradation of mRNA or a very low level of transcription. The putative transcriptional organization of the other genes was suggested by sequence analysis (Fig. 1). Locations of possibly significant ςⁿ, NifA recognition sites upstream of genes and of potential transcription terminators downstream of genes are given in Table 1.

FIG. 2.

Northern analysis with probes for nifD (A) and nifE (B). Probes were radiolabeleled with [α-³²P]dCTP by random priming (Stratagene). −, cultures grown with low concentrations of fixed N (0.5 mM NH₄); +, cultures grown with high concentrations of fixed N (10 mM NH₄).

An unusual degree of overlap between the 3′ and 5′ ends of adjacent genes, indicating translational coupling, was found for nifN-nifX (30 bp). Other cases of overlap were found for fixC-fixX (1 bp), rpoN-modC (3 bp), modC-modB (3 bp), modB-modA (3 bp), nifZ-fixU (3 bp), and fixU-orfl (3 bp). Several other cases of translational coupling are found between nif and/or related genes in other diazotrophs, including one large overlap of 34 bp between nifN and nifX in the archeal diazotroph Methanococcus maripaludis and smaller overlaps in M. maripaludis (nifD-nifK, 7 bp) (14), A. vinelandii (orf8-nifW, 3 bp; nifZ-nifM, 10 bp; nifM-orf9, 7 bp) (12), and Klebsiella pneumoniae (nifB-nifQ, 1 bp; nifL-nifA, 4 bp; nifZ-nifM, 4 bp; and nifN-nifX, 14 bp) (1). Of the A. diazotrophicus genes described here, only nifN and modC had the translational initiation codon GTG. This initiation codon also occurs in several genes in various diazotrophs, for example, nifB and nifX of Frankia alni and ORF1 of Plectonema boryanum PCC 73110 (10, 21).

Characterization of genes related to nitrogen fixation and/or plant colonization is important for elucidating these processes in endophytic diazotrophs. This work examines a major cluster of genes that are certainly or potentially important for the ability of A. diazotrophicus to fix nitrogen inside its plant host. This knowledge may be relevant for efforts to use this organism or others like it, such as the endophyte Herbaspirillum seropedicae, to better benefit growth of sugarcane and possibly other monocot plants.

Nucleotide sequence accession number.

Sequences in previous GenBank files have been compiled and combined with new data under accession no. AF030414.