Search

Rnf proteins are proposed to form membrane-protein complexes involved in the reduction of target proteins such as the transcriptional regulator SoxR or the dinitrogenase reductase component of nitrogenase. In this work, we investigate the role of rnf genes in the nitrogen-fixing bacterium Azotobacter vinelandii. We show that A. vinelandii has two clusters of rnf-like genes: rnf1, whose expression is nif-regulated, and rnf2, which is expressed independently of the nitrogen source in the medium. Deletion of each of these gene clusters produces a time delay in nitrogen-fixing capacity and, consequently, in diazotrophic growth. Deltarnf mutations cause two distinguishable effects on the nitrogenase system: (i), slower nifHDK gene expression and (ii), impairment of nitrogenase function. In these mutants, dinitrogenase reductase activity is lowered, whereas dinitrogenase activity remains essentially unaltered. Further analysis indicates that deltarnf mutants accumulate an inactive and iron-deficient form of NifH because they have lower rates of incorporation of [4Fe-4S] into NifH. Deltarnf mutations also cause a noticeable decrease in aconitase activity; however, they do not produce general oxidative stress or modification of Fe metabolism in A. vinelandii. Our results suggest the existence of a redox regulatory mechanism in A. vinelandii that controls the rate of expression and maturation of nitrogenase by the activity of the Rnf protein complexes. rnf1 plays a major and more specific role in this scheme, but the additive effects of mutations in rnf1 and rnf2 indicate the existence of functional complementation between the two homologous systems.

In recent years, it has become evident that [Fe-S] proteins, such as hydrogenase, nitrogenase and aconitase, require a complex machinery to assemble and insert their associated [Fe-S] clusters. So far, three different types of [Fe-S] cluster biosynthetic systems have been identified and these have been designated nif, isc and suf. In the present work, we show that the nif-specific [Fe-S] cluster biosynthetic system from Azotobacter vinelandii, which is required for nitrogenase maturation, cannot functionally replace the isc [Fe-S] cluster system used for the maturation of other [Fe-S] proteins, such as aconitase. The results indicate that, in certain cases, [Fe-S] cluster biosynthetic machineries have evolved to perform only specialized functions.

Azotobacter vinelandii is a soil gamma-proteobacteria that fixes nitrogen and forms desiccation-resistant cysts. The exopolysaccharide alginate is an integral part of the layers surrounding the cysts. Here, we reported the cloning of A. vinelandii algC, encoding the enzyme catalyzing the second step of alginate pathway. We showed that AlgC is involved not only in alginate production, but also in lipopolysaccharide (LPS) synthesis and that it seems to have both phosphomannomutase and phosphoglucomutase activities. The transcriptional analysis of the A. vinelandii algC gene showed that it contained two start sites, one of which was dependent on the alternative sigma factor AlgU/AlgT. This finding explains why alginate biosynthesis is dependent on AlgU activity, since all other alginate biosynthetic genes have been characterized previously and algC is the only alginate structural gene that is directly transcribed by this sigma factor.

Transcription of algD, encoding GDP-mannose dehydrogenase, the key enzyme in the alginate biosynthetic pathway, is highly regulated in Azotobacter vinelandii. We describe here the characterization of a Tn5 insertion mutant (AC28) which shows a higher level of expression of an algD::lacZ fusion. AC28 cells were morphologically abnormal and unable to encyst. The cloning and nucleotide sequencing of the Tn5-disrupted locus in AC28 revealed an operon homologous to the Escherichia coli ampDE operon. Tn5 was located within the ampD gene, encoding a cytosolic N-acetyl-anhydromuramyl-L-alanine amidase that participates in the intracellular recycling of peptidoglycan fragments. The ampE gene encodes a transmembrane protein, but the function of the protein is not known. We constructed strains carrying ampD or ampE mutations and one with an ampDE deletion. The strain with a deletion of the ampDE operon showed a phenotype similar to that of mutant AC28. The present work demonstrates that both alginate production and bacterial encystment are greatly influenced by the bacterial ability to recycle its cell wall.

In diazotrophic organisms, nitrogenase synthesis and activity are tightly regulated. Two genes, nifL and nifA, are implicated as playing a major role in this regulation. NifA is a transcriptional activator, and its activity is inhibited by NifL in response to availability of excess fixed nitrogen and high O(2) tension. It was postulated that NifL binds to NifA to inhibit NifA-mediated transcriptional activation of nif genes. Mutational analysis combined with transcriptional activation studies clearly is in agreement with the proposal that NifL interacts with NifA. However, several attempts to identify NifA-NifL interactions by using methods such as coimmunoprecipitations and chemical cross-linking experiments failed to detect direct interactions between these proteins. Here we have taken a genetic approach, the use of a yeast two-hybrid protein-protein interaction assay system, to investigate NifL interaction with NifA. A DNA fragment corresponding to the kinase-like domain of nifL was PCR amplified and was used to generate translation fusions with the DNA binding domain and the DNA activation domain of the yeast transcriptional activator GAL4 in yeast two-hybrid vectors. Similarly, a DNA fragment corresponding to the catalytic domain of nifA was PCR amplified and used to generate translation fusions with the DNA-binding domain and the DNA-activation domain of GAL4 in yeast two-hybrid vectors. After introducing appropriate plasmid combinations in yeast cells, the existance of direct interaction between NifA and NifL was analyzed with the MATCHMAKER yeast two-hybrid system by testing for the expression of lacZ and his3 genes. These analyses showed that the kinase-like domain of NifL directly interacts with the catalytic domain of NifA.

The iron-molybdenum cofactor (FeMo-co) of nitrogenase contains molybdenum, iron, sulfur, and homocitrate in a ratio of 1:7:9:1. In vitro synthesis of FeMo-co has been established, and the reaction requires an ATP-regenerating system, dithionite, molybdate, homocitrate, and at least NifB-co (the metabolic product of NifB), NifNE, and dinitrogenase reductase (NifH). The typical in vitro FeMo-co synthesis reaction involves mixing extracts from two different mutant strains of Azotobacter vinelandii defective in the biosynthesis of cofactor or an extract of a mutant strain complemented with the purified missing component. Surprisingly, the in vitro synthesis of FeMo-co with only purified components failed to generate significant FeMo-co, suggesting the requirement for one or more other components. Complementation of these assays with extracts of various mutant strains demonstrated that NifX has a role in synthesis of FeMo-co. In vitro synthesis of FeMo-co with purified components is stimulated approximately threefold by purified NifX. Complementation of these assays with extracts of A. vinelandii DJ42. 48 (DeltanifENX DeltavnfE) results in a 12- to 15-fold stimulation of in vitro FeMo-co synthesis activity. These data also demonstrate that apart from the NifX some other component(s) is required for the cofactor synthesis. The in vitro synthesis of FeMo-co with purified components has allowed the detection, purification, and identification of an additional component(s) required for the synthesis of cofactor.

The cloning and expression of a family of five modular-type mannuronan C-5-epimerase genes from Azotobacter vinelandii (algE1 to -5) has previously been reported. The corresponding proteins catalyze the Ca2+-dependent polymer-level epimerization of beta-D-mannuronic acid to alpha-L-guluronic acid (G) in the commercially important polysaccharide alginate. Here we report the identification of three additional structurally similar genes, designated algE6, algE7, and algY. All three genes were sequenced and expressed in Escherichia coli. AlgE6 introduced contiguous stretches of G residues into its substrate (G blocks), while AlgE7 acted as both an epimerase and a lyase. The epimerase activity of AlgE7 leads to formation of alginates with both single G residues and G blocks. AlgY did not display epimerase activity, but a hybrid gene in which the 5'-terminal part was exchanged with the corresponding region in algE4 expressed an active epimerase. Southern blot analysis of genomic A. vinelandii DNA, using the 5' part of algE2 as a probe, indicated that all hybridization signals originated from algE1 to -5 or the three new genes reported here.

We have identified the rfbG from a non-symbiotic and non-pathogenic soil bacterium, Azotobacter vinelandii. The nucleotide sequence analysis of the rfbG revealed an open reading frame that encodes a peptide of 360 amino acids. This deduced peptide shares 57% homology with the RfbG of Synechocystis and 47% homology with the RfbG of Yersinia pseudotuberculosis. The previously identified short-chain dehydrogenases/reductases family signature sequence is conserved in the sequence of the RfbG of A. vinelandii. Southern blotting analysis of A. vinelandii chromosome by probed with 1.1 kb PstI DNA fragment corresponding to rfbG revealed that it is present as single copy on A. vinelandii chromosome. Disrupting the rfbG present on the chromosome of A. vinelandii, by insertion of kanamycin resistance marker via homologous recombination, resulted in drastic changes in the growth characteristics. The rfbG-negative A. vinelandii grown in liquid medium exhibited agglutination that is characteristic of rfb- mutants of other bacteria, suggesting that we have cloned the functional copy of the rfbG of A. vinelandii.

The lep of Azotobacter vinelandii is an 852-base-pair open reading frame (ORF) which encodes a protein of 284 amino acid residues. The translated protein shares 75% homology with leader peptidase I isolated from Pseudomonas fluorescens and 37% homology with leader peptidase I isolated from Escherichia coli. Five highly conserved regions found in the family of leader peptidase I proteins are conserved in A. vinelandii Lep. The putative membrane topology of the protein seems similar to that of E. coli leader peptidase I based on the hydrophobicity analysis of the predicted amino acid sequence. Southern blotting analysis of the A. vinelandii chromosome by probing with lep specific DNA revealed that lep is present as a single copy per the chromosome. A multicopy plasmid carrying A. vinelandii lep could complement a temperature sensitive lep mutant of E. coli strain IT41, suggesting that we have identified the functional copy of lep present on A. vinelandii genome., (Copyright 1997 Academic Press.)

The structural genes for the cyanide-insensitive terminal oxidase (CIO) of Pseudomonas aeruginosa were sequenced. The locus comprised two open reading frames, cioA and cioB, coding for gene products of 488 and 335 amino acid residues with predicted molecular masses of 54241 and 37016 Da respectively. These genes were encoded by a 2.7 kb transcript and probably comprise an operon. Upstream of a major transcriptional start site is a -10 promoter region and, approximately at nucleotides -50 and +13, there are sequences homologous to the binding site of the transcriptional regulator Anr. The deduced amino acid sequences of CioA and CioB are homologous to the cytochrome bd quinol oxidases of Escherichia coli and Azotobacter vinelandii. However, no cytochrome d-like signals were found in wild-type P. aeruginosa strains. An atypical cytochrome d-like signal was seen under low-aeration growth conditions but only in strains in which the cioAB genes were present on a high-copy-number plasmid. The appearance of these cytochrome d-like signals was not paralleled by a concomitant increase in CIO activity. These data support the hypothesis that the CIO of P. aeruginosa does not contain haem d. This raises the possibility that there is a family of bacterial quinol oxidases related to the cytochrome bd of E. coli that can differ in their haem composition from the E. coli paradigm.

Alginate is an unbranched polysaccharide composed of the two sugar residues beta-D-mannuronic acid (M) and alpha-L-guluronic acid (G). The M/G ratio and sequence distribution in alginates vary and are of both biological and commercial significance. We have previously shown that a family of highly related mannuronan C-5-epimerase genes (algE1 to -E5) controls these parameters in Azotobacter vinelandii, by catalyzing the Ca2+-dependent conversion of M to G at the polymer level. In this report, we describe the cloning and expression of a new A. vinelandii epimerase gene (here designated algG), localized 29 nucleotides downstream of the previously described gene algJ. Sequence alignments show that algG does not belong to the same class of genes as algE1 to -E5 but that it shares 66% sequence identity with a previously described mannuronan C-5-epimerase gene (also designated algG) from Pseudomonas aeruginosa. A. vinelandii algG was expressed in Escherichia coli, and the enzyme was found to catalyze epimerization in the absence of Ca2+, although the presence of the cation stimulated the activity moderately. Surprisingly, all activity was blocked by Zn2+. P. aeruginosa AlgG has been reported to contain an N-terminal export signal sequence which is cleaved off during expression in E. coli. This does not happen with A. vinelandii AlgG, which appears to be produced at least partly in an insoluble form when expressed at high levels in E. coli. DNA sequencing analyses of the regions flanking algG suggest that the gene is localized in a cluster of genes putatively involved in alginate biosynthesis, and the organization of this cluster appears to be the same as previously described for P. aeruginosa.

Azotobacter vinelandii presents a differentiation process leading to the formation of desiccation-resistant cysts. Alginate, the exopolysaccharide produced by this bacterium, has been postulated to have a role in cyst formation. Here, we report the cloning and characterization of the A. vinelandii gene coding for the enzyme GDP-mannose dehydrogenase (algD), which is the key enzyme for alginate synthesis in Pseudomonas aeruginosa. This gene has a high degree of similarity with the algD gene from P. aeruginosa, and similar proteins seem to be involved in algD regulation in both bacteria. We show the existence of two mRNA start sites; one of these sites corresponds to a promoter transcribed by RNA polymerase containing a sigma E subunit. An A. vinelandii algD mutant which is completely impaired in alginate production and which is unable to form desiccation-resistant cells was constructed. The effects of NH4, NO3, and NaCl concentrations on algD transcription for three A. vinelandii strains producing different alginate levels were evaluated. We found a strict correlation between alginate production and algD transcription for the three strains studied; however, the effects on algD transcription under the conditions studied were different for each strain. The nitrogen source regulates algD expression in the wild-type strain.

A gene encoding rhodanese (rhdA) was cloned from Azotobacter vinelandii on a 2.3-kb SphI fragment. This fragment was identified by its hybridization to a PCR product obtained by amplification of genomic DNA using degenerate primers encoding the N-terminal sequence of rhodanese purified from A. vinelandii. The sequence of a 1.2-kb region revealed an 813-bp open reading frame that encoded a polypeptide of 271 amino acids, the N-terminal sequence of which was identical to that of A. vinelandii rhodanese. In a search of database entries, eukaryotic rhodaneses and rhodanese-like proteins from bacteria gave the highest scores of identity (27-30%) with the predicted product of the 813-bp open reading frame. A. vinelandii RhdA shows less sequence similarity to vertebrate rhodaneses than it does to prokaryotic rhodanese-like proteins which did not show typical rhodanese activity. Basic residues thought to be catalytically important in bovine rhodanese are not conserved in A. vinelandii rhodanese. The sequence similarity between the two structurally similar domains of rhodanese is more pronounced for the A. vinelandii enzyme than the bovine enzyme, and supports the hypothesis that the complete structure was originally generated by gene duplication. When rhdA was overexpressed in Escherichia coli, rhodanese represented 30% of total cell protein and thiosulfate:cyanide sulfurtransferase activity increased >600 fold in cell-free extracts. A. vinelandii rhdA insertion/deletion mutants had no discernible phenotype distinct from the wild-type strain with respect to growth on various sulfur sources or nitrogenase activity. Mutants retained 20% of wild-type rhodanese thiosulfate:cyanide sulfurtransferase activity suggesting the presence of redundant sulfurtransferase enzymes in A. vinelandii.

In Azotobacter vinelandii, the anfHDGK operon encodes the subunits for the third nitrogenase complex. Two open reading frames (orf1 and orf2) located immediately downstream of anfK were shown to be required for diazotrophic growth under Mo- and V-deficient conditions. We have designated orf1 and orf2 anfO and anfR, respectively. Strains (CA115 and CA116) carrying in-frame deletions in anfO and anfR accumulate the subunits for nitrogenase 3 under Mo-deficient diazotrophic conditions. AnfO and AnfR are required for nitrogenase 3-dependent diazotrophic growth and 15N2 incorporation but not for acetylene reduction. AnfO contains a putative heme-binding domain that exhibits similarity to presumed heme-binding domains of P-450 cytochromes. Amino acid substitutions of Cys-158 show that this residue is required for fully functional AnfO as measured by diazotrophic growth under Mo- and V-deficient conditions. The nucleotide sequence of the region located immediately downstream of anfR has been determined. A putative rho-independent transcription termination site has been identified 250 bp from the 3' end of anfR. A third open reading frame (orf3), located downstream of anfR, does not appear to be required for diazotrophic growth under Mo- and V-deficient conditions.

The L-guluronic acid residues in the Azotobacter vinelandii polysaccharide alginate originate from a post-polymerization reaction catalysed by the enzyme mannuronan C-5-epimerase (ME). We have previously reported the cloning and expression of an A. vinelandii gene encoding this enzyme, and we show here that the organism encodes at least four other ME genes originating from a common ancestor gene by a complex rearrangement process. The biological function of the corresponding enzymes is probably to catalyse the formation of alginates with a variety of physical properties. This model may explain the origin of the structural variability found in alginates isolated both from prokaryotic and eukaryotic organisms. The A. vinelandii enzymes may also potentially be useful for certain medical and biotechnological applications of this commercially important polysaccharide.

An Azotobacter vinelandii mannuronan C-5-epimerase gene was cloned in Escherichia coli. This enzyme catalyzes the Ca(2+)-dependent epimerization of D-mannuronic acid residues in alginate to the corresponding epimer L-guluronic acid. The epimerase gene was identified by screening a bacteriophage EMBL3 gene library of A. vinelandii DNA with a synthetic oligonucleotide probe. The sequence of this probe was deduced after determination of the N-terminal amino acid sequence of a previously reported extracellular mannuronan C-5-epimerase from A. vinelandii. A DNA fragment hybridizing against the probe was subcloned in a plasmid vector in E. coli, and the corresponding recombinant plasmid expressed intracellular mannuronan C-5-epimerase in this host. The nucleotide sequence of the gene encoding the epimerase was determined, and the sequence data showed that the molecular mass of the deduced protein is 103 kDa. A module consisting of about 150 amino acids was repeated tandemly four times in the C-terminal part of the deduced protein. Each of the four repeats contained four to six tandemly oriented nonameric repeats. The sequences in these motifs are similar to the Ca(2+)-binding domains of functionally unrelated secreted proteins reported previously in other bacteria. The reaction product of the recombinant epimerase was analyzed by nuclear magnetic resonance spectroscopy, and the results showed that the guluronic acid residues were distributed in blocks along the polysaccharide chain. Such a nonrandom distribution pattern, which is important for the commercial use of alginate, has previously also been identified in the reaction product of the corresponding enzyme isolated from A. vinelandii.

The nucleotide sequence was obtained for the hypE gene in the cluster of structural and accessory genes required for the assembly and functioning of the membrane-bound, dimeric, (NiFe)hydrogenase in Azotobacter vinelandii. The hypE gene encodes a polypeptide of 341 amino acid residues which is rich in alanine, glycine, valine and proline and appears to be involved in maturation of the enzyme because chromosomal mutations in hypE block O2-dependent H2-oxidation and affect the amount, processing and localization of the (NiFe) hydrogenase alpha-subunit. The complete nucleotide sequence for the hydrogenase gene cluster in A. vinelandii has now been assembled into a contiguous sequence of 13,914 bp containing 16 potential genes which appear to be transcribed undirectionally. They are arranged in the order hoxK, hoxG, hoxZ, hoxM, hoxL, hoxO, hoxQ, hoxR, hoxT, hoxV, hypA, hypB, hypF, hypC, hypD and hypE. This cluster closely resembles those described for comparable (NiFe) hydrogenases in other bacteria.

Several regulatory gene mutants of Azotobacter vinelandii were tested for ability to synthesize functional nitrogenase-1 (Nif phenotype), nitrogenase-2 (Vnf), or nitrogenase-3 (Anf). While nifA mutants were Nif-, Vnf+, and Anf+/-, and ntrC mutants were Nif+, Vnf+, and Anf+, nifA ntrC double mutants were Nif-, Vnf-, and Anf-. A vnfA mutant was Nif+, Vnf+/-, and Anf+/-, and an anfA strain was Nif+, Vnf+, and Anf-. lacZ fusions in the nifH, vnfH, vnfD, anfH, and nifM genes of Azotobacter vinelandii were constructed and introduced into wild-type and regulatory mutants of A. vinelandii. Expression of these operons correlated with the growth phenotype of the regulatory mutants. Apparently either NifA or NtrC can activate expression of nifM. Also, expression of the anf operon required the NifA transcriptional activator, although there are no NifA binding sites at appropriate locations upstream of anfH (or anfA). The results confirm previous reports that VnfA and AnfA are required for expression of vnf and anf genes, respectively, and that VnfA is involved in repression of the nifHDK operon in the absence of molybdenum and of the anfHDGK operon in the presence of vanadium.

The genes coding for two structurally different isocitrate dehydrogenase isozymes (IDH-I and IDH-II) of a psychrophilic bacterium, Vibrio sp. strain ABE-1, were cloned and sequenced. Open reading frames of the genes (icdI and icdII) are 1,248 and 2,229 bp in length, respectively. The amino acid sequences predicted from the open reading frames of icdI and icdII corresponded to the N-terminal amino acid sequences of the purified IDH-I and IDH-II, respectively. No homology was found between the deduced amino acid sequences of the isozymes; however, the IDH-I, a dimeric enzyme, had a high amino acid sequence identity (74.3%) to the Escherichia coli IDH. The deduced amino acid sequence of the IDH-II, a monomeric enzyme, was not related to any known sequence. However, the IDH-II had an amino acid sequence homologous to that of a cyanogen bromide-cleaved peptide containing a putative active-site methionyl residue of the monomeric IDH of Azotobacter vinelandii. The two genes (icdlI and icdII) were found to be tandemly located in the same orientation. Northern (RNA) blot analyses showed that the two genes are transcribed independently. Primer extension experiments located single transcriptional start sites 39 and 96 bp upstream of the start codons of icdI and icdII, respectively. The amount of icdI transcript but not icdII increased when Vibrio sp. strain ABE-1 cells were cultured in acetate minimal medium.

Anabaena variabilis ATCC 29413 is a heterotrophic, nitrogen-fixing cyanobacterium that has been reported to fix nitrogen and reduce acetylene to ethane in the absence of molybdenum. DNA from this strain hybridized well at low stringency to the nitrogenase 2 (vnfDGK) genes of Azotobacter vinelandii. The hybridizing region was cloned from a lambda EMBL3 genomic library of A. variabilis, mapped, and sequenced. The deduced amino acid sequences of the vnfD and vnfK genes of A. variabilis showed only about 56% similarity to the nifDK genes of Anabaena sp. strain PCC 7120 but were 76 to 86% similar to the anfDK or vnfDK genes of A. vinelandii. The organization of the vnf gene cluster in A. variabilis was similar to that of A. vinelandii. However, in A. variabilis, the vnfG gene was fused to vnfD; hence, this gene is designated vnfDG. A vnfH gene was not contiguous with the vnfDG gene and has not yet been identified. A mutant strain, in which a neomycin resistance cassette was inserted into the vnf cluster, grew well in a medium lacking a source of fixed nitrogen in the presence of molybdenum but grew poorly when vanadium replaced molybdenum. In contrast, the parent strain grew equally well in media containing either molybdenum or vanadium. The vnf genes were transcribed in the absence of molybdenum, with or without vanadium. The vnf gene cluster did not hybridize to chromosomal DNA from Anabaena sp. strain PCC 7120 or from the heterotrophic strains, Nostoc sp. strain Mac and Nostoc sp. strain ATCC 29150. A hybridizing ClaI fragment very similar in size to the A. variabilis ClaI fragment was present in DNA isolated from several independent, cultured isolates of Anabaena sp. from the Azolla symbiosis.

Biological nitrogen fixation is catalyzed by nitrogenase, a complex metalloenzyme composed of two separately purifiable component proteins encoded by the structural genes nifH, nifD, and nifK. Deletion of the Azotobacter vinelandii nifS gene lowers the activities of both nitrogenase component proteins. Because both nitrogenase component proteins have metallocluster prosthetic groups that are composed of iron- and sulfur-containing cores, this result indicated that the nifS gene product could be involved in the mobilization of the iron or sulfur required for metallocluster formation. In the present work, it is shown that NIFS is a pyridoxal phosphate-containing homodimer that catalyzes the formation of L-alanine and elemental sulfur by using L-cysteine as substrate. NIFS activity is extremely sensitive to thiol-specific alkylating reagents, which indicates the participation of a cysteinyl thiolate at the active site. Based on these results we propose that an enzyme-bound cysteinyl persulfide that requires the release of the sulfur from the substrate L-cysteine for its formation ultimately provides the inorganic sulfide required for nitrogenase metallocluster formation. The recent discovery of nifS-like genes in non-nitrogen-fixing organisms also raises the possibility that the reaction catalyzed by NIFS represents a universal mechanism that involves pyridoxal phosphate chemistry, in the mobilization of the sulfur required for metallocluster formation.

Five major anfH-hybridizing mRNA species accumulated in Azobacter vinelandii cells derepressed for nitrogenase-3 (an alternative nitrogenase, which appears to lack Mo and V). Using anfH-, anfD-, anfG-, anfK-, and orflorf2-specific probes and mutant strains of A. vinelandii these mRNA species have been identified as encoding anfHDGKorflorf2 (6.0 kb), anfHDGK (4.3 kb), anfHD (2.6 kb), vnfHorfFd (1.3 kb), and vnfH and (or) anfH (1.0 kb). A 0.6-kb mRNA species, which hybridized only to the orflorf2-specific probe, and a 3.5-kb mRNA species, which hybridized to anfD or anfK, also accumulated under these conditions. Northern blot analysis and S1 nuclease mapping indicate that transcription of the anf structural gene cluster initiates at a unique nif consensus promoter situated 127 base pairs upstream from the anfH coding region. Observation of anfH-hybridizing mRNA species that accumulate in strains derepressed for nitrogen fixation demonstrates that transcription of the anfHDGKorflorf2 cluster is normally repressed by Mo, V, and NH4+, whereas transcription of the vnfHorfFd cluster does not require the presence of V and is repressed only by Mo, but not NH4+. Analysis of the accumulation of mRNAs in a tungsten-tolerant strain revealed that Mo and V repression of anf transcription must occur by different mechanisms.

The roles of the product of the hoxZ gene immediately downstream of the hydrogenase gene (hoxKG) in Azotobacter vinelandii were investigated by constructing and characterizing a mutant with the center of the hoxZ gene deleted. The strain lacking the functional hoxZ gene product exhibited a low rate of H2 oxidation with O2 as the electron acceptor relative to that of the wild-type strain. Nevertheless, when the enzyme was exogenously activated and methylene blue was used as the electron acceptor from hydrogenase, rates of H2 oxidation comparable to those in the wild-type strain were observed. These results suggest that the gene product of hoxZ plays a role in activating and maintaining hydrogenase in a reduced active state. The product of hoxZ could also be the linkage necessary for transfer of electrons from H2 to the electron transport chain.

Strains with mutations in 23 of the 30 genes and open reading frames in the major nif gene cluster of A. vinelandii were tested for ability to grow on N-free medium with molybdenum (Nif phenotype), with vanadium (Vnf phenotype), or with neither metal present (Anf phenotype). As reported previously, nifE, nifN, nifU, nifS and nifV mutants were Nif- (failed to grow on molybdenum) while nifM mutants were Nif-, Vnf- and Anf-. nifV, nifS, and nifU mutants were found to be unable to grow on medium with or without vanadium, i.e. were Vnf- Anf-. Therefore neither vnf nor anf analogoues of nifU, nifS, nifV or nifM are expected to be present in A. vinelandii.

The nucleotide sequence (3,600 bp) of a second copy of nifENX-like genes in Azotobacter vinelandii has been determined. These genes are located immediately downstream from vnfA and have been designated vnfENX. The vnfENX genes appear to be organized as a single transcriptional unit that is preceded by a potential RpoN-dependent promoter. While the nifEN genes are thought to be evolutionarily related to nifDK, the vnfEN genes appear to be more closely related to nifEN than to either nifDK, vnfDK, or anfDK. Mutant strains (CA47 and CA48) carrying insertions in vnfE and vnfN, respectively, are able to grow diazotrophically in molybdenum (Mo)-deficient medium containing vanadium (V) (Vnf+) and in medium lacking both Mo and V (Anf+). However, a double mutant (strain DJ42.48) which contains a nifEN deletion and an insertion in vnfE is unable to grow diazotrophically in Mo-sufficient medium or in Mo-deficient medium with or without V. This suggests that NifE and NifN substitute for VnfE and VnfN when the vnfEN genes are mutationally inactivated. AnfA is not required for the expression of a vnfN-lacZ transcriptional fusion, even though this fusion is expressed under Mo- and V-deficient diazotrophic growth conditions.

Azotobacter vinelandii is a free-living nitrogen-fixing bacterium that has one of the highest respiratory rates of all aerobic organisms. Based on various physiological studies, a d-type cytochrome has been postulated to be the terminal oxidase of a vigorously respiring but apparently uncoupled branch of the electron transport system in the membranes of this organism. We cloned and characterized the structural genes of the two subunits of this oxidase. The deduced amino acid sequences of both subunits of the A. vinelandii oxidase have extensive regions of homology with those of the two subunits of the Escherichia coli cytochrome d complex. Most notably, the histidine residues proposed to be the axial ligands for the b hemes of the E. coli oxidase and an 11-amino-acid stretch proposed to be part of the ubiquinone binding site are all conserved in subunit I of the A. vinelandii oxidase. The A. vinelandii cytochrome d was expressed in a spectrally and functionally active form in the membranes of E. coli, under the control of the lac or tac promoter. The spectral features of the A. vinelandii cytochrome d expressed in E. coli are very similar to those of the E. coli cytochrome d. The expressed oxidase was active as a quinol oxidase and could reconstitute an NADH to oxygen electron transport chain.