Pseudomonas Oxalaticus Research Articles

Taxonomic Position of “ Pseudomonas oxalaticus” Strain Ox1 T (DSM 1105 T) (Khambata and Bhat, 1953) and its Description in the Genus Ralstonia as Ralstonia oxalatica comb. nov.

"Pseudomonas oxalaticus" strain Ox1T (= DSM 1105T), which was described as an oxalate-decomposing bacterium, was reinvestigated to clarify its taxonomic position. 16S ribosomal DNA sequence comparisons demonstrated that this species is phylogenetically related to the species of the genus Ralstonia. and represents a new species. The result of the DNA-DNA hybridization value was supported in this placement. Strain Ox1T is closely related to Ralstonia eutropha with a less than 60% DNA-DNA hybridization value. The new name Ralstonia oxalatica comb. nov. is proposed to strain Ox1T, on the basis of these results and previously published data for the G+C content of the genomic DNA and the phenotypic characters.

In vivo recombination of two R68.45-primes, one carrying plasmid genes for autotrophy (Aut) from Alcaligenes hydrogenophilus and the other carrying a σ54-like chromosome gene from Pseudomonas oxalaticus

A cluster of aut genes on a plasmid of Alcaligenes hydrogenophilus and a σ 54-like gene on the chromosome of Pseudomonas oxalaticus have already been respectively cloned in vivo by R-plasmid R68.45. In this study, in vivo recombination of these two R68.45-primes was carried out. A new recombinant plasmid R68.45-prime carrying the aut and σ 54-like genes was obtained, both genes being stably maintained.

A sensitive determination of uric acid in serum using uricase/catalase/formaldehyde dehydrogenase coupled with formate dehydrogenase

We developed and evaluated an assay for serum uric acid based on the uricase (EC 1.7.3.3)-catalase (EC 1.11.1.6)-formaldehyde dehydrogenase (FADH, EC 1.2.1.46) method coupled with formate dehydrogenase (formate: NAD oxidoreductase, FDH, EC 1.2.1.2). Formate dehydrogenase from Pseudomonas oxalaticus catalyzes the formation of NADH from formate produced by FADH. Owing to the NADH and formate oxidase activity of the FDH itself, the full reaction curve is not linear, but gradually decreases. The formation of NADH is not stoichiometric with formate removal, but is strictly proportional to it. To overcome this decrease of extinction, we added hydroxylamine hydrochloride to the FDH. The sensitivity of the full reaction in the presence of FDH was about 1.8 times that without FDH. Analysis with a Cobas Bio centrifugal analyzer revealed a linearity of up to 3.56 mmol/L. The uricase-catalase-alcohol dehydrogenase method correlated well with the uricase-peroxidase-chromagen method. Our method is more sensitive than other methods.

In vivo cloning of a Pseudomonas oxalaticus σ54-like gene acting in expression of hydrogenase genes from Alcaligenes hydrogenophilus

Abstract In Pseudomonas oxalaticus OX1, which can express the hydrogenase genes from the H 2 -oxidizing bacterium Alcaligenes hydrogenophilus , a σ 54 -like factor-defective mutant was isolated by insertion of the kanamycin resistance marker. The P. oxalaticus OX1 chromosome gene complementing the mutation was cloned in vivo using the R-plasmid R68.45. The recombinant plasmids formed in strain OX1, the R68.45-prime carrying a σ 54 -like gene, were selected using intrastrain mating of strain OX1. Carrying one of the recombinant plasmids, the σ 54 -like factor-defective mutant recovered the abilities to express the hydrogenase genes and to assimilate nitrate as the nitrogen source and C 4 -dicarboxylates as the carbon sources for growth.

In vivo cloning of genes determining lithoautotrophy (Aut) on a plasmid from Alcaligenes hydrogenophilus

Hydrogenase ( hox) genes on the megaplasmid pHG21-a from Alcaligenes hydrogenophilus, whose lithoautotrophic growth (Aut) is supported by H 2-oxidation (Hox) and CO 2-fixation (Cfx), were cloned in vivo using a broad host range IncP1 plasmid R68.45. The recombinant plasmid was detected by the characteristic that it was transferred at a frequency 10 6-fold higher than pHG21-a in intrastrain mating of the Hox − Cfx + bacterium Pseudomonas oxalaticus OX1. All of six recombinant plasmids designated pFUs inherited all three resistance markers of R68.45. Four plasmids (pFU3, pFU8, pFU11, and pFU15) with a molecular size of 69 Md had only membrane-bound hydrogenase ( hoxP) genes, and two plasmids (pFU7 and pFU9) of 85 Md had both hoxP and soluble hydrogenase ( hoxS) genes. The Hox − Cfx − bacteria P. oxalaticus OX4 and OX6 gained Aut − phenotype by the possession of pHG21-a, pFU7 or pFU15. These results showed that Hox plasmid pHG21-a was an Aut plasmid and pFU7 and pFU15 inherited this phenotype, pFU7 was maintained stably in P. oxalaticus OX1 and had all of the lithoautotrophic phenotypes of pHG21-a. pFU7, rather than pHG21-a, is useful for further studies on the transfer of the Aut phenotype to a broad range of bacteria.

Formate dehydrogenase from the methane oxidizer Methylosinus trichosporium OB3b

Formate dehydrogenase (NAD+ dependent) was isolated from the obligate methanotroph Methylosinus trichosporium OB3b. When the enzyme was isolated anaerobically, two forms of the enzyme were seen on native polyacrylamide gels, DE-52 cellulose and Sephacryl S-300 columns; they were approximately 315,000 and 155,000 daltons. The enzyme showed two subunits on sodium dodecyl sulfate-polyacrylamide gels. The Mr of the alpha-subunit was 53,800 +/- 2,800, and that of the beta-subunit was 102,600 +/- 3,900. The enzyme (Mr 315,000) was composed of these subunits in an apparent alpha 2 beta 2 arrangement. Nonheme iron was present at a concentration ranging from 11 to 18 g-atoms per mol of enzyme (Mr 315,000). Similar levels of acid-labile sulfide were detected. No other metals were found in stoichiometric amounts. When the enzyme was isolated aerobically, there was no cofactor requirement for NAD reduction; however, when isolated anaerobically, activity was 80 to 90% dependent on the addition of flavin mononucleotide (FMN) to the reaction mixture. Furthermore, the addition of formate to an active, anoxic solution of formate dehydrogenase rapidly inactivated it in the absence of an electron acceptor; this activity could be reconstituted approximately 85% by 50 nM FMN. Flavin adenine dinucleotide could not replace FMN in reconstituting enzyme activity. The Kms of formate dehydrogenase for formate, NAD, and FMN were 146, 200, and 0.02 microM, respectively. "Pseudomonas oxalaticus" formate dehydrogenase, which has physical characteristics nearly identical to those of the M. trichosporium enzyme, was also shown to be inactivated under anoxic conditions by formate and reactivated by FMN. The evolutionary significance of this similarity is discussed.

Thermodynamic efficiency of bacterial growth calculated from growth yield of Pseudomonas oxalaticus OX1 in the chemostat

In order to determine the thermodynamic efficiency of bacterial growth, Pseudomonas oxalaticus OX1 was grown in carbon-limited continuous cultures. 11 different carbon sources, ranging from oxalate (most oxidised component) to ethanol (most reduced component), were used as limiting substrate in these experiments. From the experimental yield values (expressed as C-mol dry weight produced per C-mol carbon substrate consumed) the thermodynamic efficiencies were calculated. On substrates more reduced than biomass (such as ethanol and glycerol) the thermodynamic efficiency of growth of P. oxalaticus was negative but it reached a maximum of 23 +/- 3% with substrates with a degree of reduction of 3 (citrate) and lower. The actual concentrations of the components involved were incorporated into the calculations but this affected the overall thermodynamic efficiency only to a small extent. This result strengthens the conclusion of Westerhoff et al. (Westerhoff, H.V., Hellingwerf, K.J. and Van Dam, K. (1983) Proc. Natl. Acad. Sci. 80, 305-309) that bacteria have been optimised towards a theoretical thermodynamic efficiency of 24%, corresponding with maximisation of growth rate at optimal efficiency, with highly oxidised substrates.

Characterization of recombinant cosmids containing H 2-oxidation genes from Alcaligenes hydrogenophilus

Ten Pseudomonas oxalaticus OX1 harboring recombinant cosmids containing H 2-oxidation genes (hox) from Alcaligenes hydrogenophilus showed H 2-dependent autotrophic growth with different growth rates. Some strains grew autotrophically almost as fast as the strain containing the Hox plasmid pHG21-a. Ten strains had high activity of membrane-bound hydrogenase (MH), but the activity of soluble hydrogenase could not be detected. Sizes of DNA inserts in recombinant cosmids isolated from 10 strains were 26 to 29 kilobases (Kb), and there was a 22-Kb overlapped region. It is suggested that the DNA fragment required for conferring the ability of H 2-dependent autotrophic growth on CO 2-fixing P. oxalaticus appears to span a region of more than 20 Kb in A. hydrogenophilus. Structural genes of MH were found in the right end of the overlapped region by DNA hybridization analysis using oligonucleotide probes corresponding to the conserved sequences of MH genes in Bradyrhizobium japonicum and Rhodobacter capsulatus. The MH was induced by H 2 and repressed by the addition of fructose, as in A. hydrogenophilus. Therefore, the inserted DNA fragment contained the DNA region which determines the regulatory properties as well as structural genes of MH.

Regulation of autotrophic metabolism in Pseudomonas oxalaticus OX1 wild-type and an isocitrate-lyase-deficient mutant.

In Pseudomonas oxalaticus the activity and synthesis of the Calvin cycle enzyme ribulose-1,5-bisphosphate carboxylase/oxygenase (RuBisCO) are regulated by inactivation and endproduct repression, respectively. Phosphoenolpyruvate (PEP) has been suggested to function as a signal molecule for the latter control system. During growth of the organism in carbon-source-limited continuous cultures with various ratios of acetate and formate in the feed, the RuBisCO levels varied considerably, but no correlation was observed with the intracellular concentrations of PEP. To study whether the repression exerted by acetate utilization was dependent on the synthesis of glycolytic intermediates from this compound, an acetate-negative mutant defective in isocitrate lyase was isolated and characterized. Clear evidence was obtained that in this mutant acetate is as effective in repressing RuBisCO synthesis as in the wild-type. It therefore appears more likely that acetyl-CoA or a closely related metabolite functions as a signal molecule in the regulation of RuBisCO synthesis.

Taxonomy of non H 2-Lithotrophic, Oxalate-Oxidizing Bacteria Related to Alcaligenes eutrophus

The taxonomic relationship among strains of Alcaligenes eutrophus (H16, H20, TF93), strains tentatively placed in A. eutrophus (CH34, JMP134), Alcaligenes hydrogenophilus inv. nom . (M50), “Pseudomonas oxalaticus” inv. nom . (OX1), and non H 2 -lithotrophic, oxalate-utilizing new strains was established by numerical classification, DNA base composition and DNA-DNA hybridization studies. Strains CH34 and JMP134 show 25% of DNA homology with A. eutrophus H16 and a phenotypic similarity of more than 80% (S J ). “P. oxalaticus” OX1 shares 50% of DNA-DNA homology with A. eutrophus H20 and a phenotypic similarity of more than 90%. The non H- 2 -lithotrophic strains are wide-spread within the range of species variation of A. eutrophus , thus indicating that H 2 -lithotrophy cannot longer be considered as a predominant phenotypic criterium for A. eutrophus . Moreover, five strains have a subpolar flagellum (the others are peritrichous), suggesting that different flagellation types could exist in the same species. The presence of a cryptic megaplasmid was detected in strains TA07 and TA19.

Cloning of the Alcaligenes eutrophus alcohol dehydrogenase gene

Mutants of Alcaligenes eutrophus which are altered with respect to the utilization of 2,3-butanediol and acetoin were isolated after transposon mutagenesis. The suicide vehicle pSUP5011 was used to introduce the drug resistance transposable element Tn5 into A. eutrophus. Kanamycin-resistant transconjugants of the 2,3-butanediol-utilizing parent strains CF10141 and AS141 were screened for mutants impaired in the utilization of 2,3-butanediol or acetoin. Eleven mutants were negative for 2,3-butanediol but positive for acetoin; they were unable to synthesize active fermentative alcohol dehydrogenase protein (class 1). Forty mutants were negative for 2,3-butanediol and for acetoin (class 2). Tn5-mob was also introduced into a Smr derivative of the 2,3-butanediol-nonutilizing parent strain H16. Of about 35,000 transconjugants, 2 were able to grow on 2,3-butanediol. Both mutants synthesized the fermentative alcohol dehydrogenase constitutively (class 3). The Tn5-labeled EcoRI fragments of genomic DNA of four class 1 and two class 3 mutants were cloned from a cosmid library. They were biotinylated and used as probes for the detection of the corresponding wild-type fragments in a lambda L47 and a cosmid gene bank. The gene which encodes the fermentative alcohol dehydrogenase in A. eutrophus was cloned and localized to a 2.5-kilobase (kb) SalI fragment which is located within a 11.5-kb EcoRI-fragment. The gene was heterologously expressed in A. eutrophus JMP222 and in Pseudomonas oxalaticus. The insertion of Tn5-mob in class 3 mutants mapped near the structural gene for alcohol dehydrogenase on the same 2.5-kb SalI fragment.

Molecular evolution of the large subunit of ribulose-1,5-bisphosphate carboxylase/oxygenase (RuBisCO)

The evolutionary relationship of the RuBisCO large subunit gene(s) (rbcL) of several prokaryotes was examined using the technique of heterologous DNA hybridization. Restriction fragments of cloned rbcL from Anacystis nidulans 6301, Chlamydomonas reinhardtii, Rhodospirillum rubrum, and maize were nick-translated and used as probes. The C. reinhardtii and maize probes hybridized with restriction fragment(s) only from cyanobacteria: Agmenellum quadruplicatum, Fremyella diplosiphon, and Mastigocladus laminosus. In addition, the A. nidulans probe hybridized with restriction fragment(s) from Alcaligenes eutrophus, Chromatium vinosum, Nitrobacter hamburgensis, Paracoccus denitrificans, Pseudomonas oxalaticus, Rhodomicrobium vannielii, Rhodopseudomonas capsulata, Rhodopseudomonas palustris, Rhodopseudomonas sphaeroides, Thiobacillus intermedius, Thiobacillus neapolitanus, and Thiothrix nivea. The elucidated fragment of Rhodopseudomonas species is presumably for the Form I RuBisCO LSU of these organisms. The R. rubrum probe hybridized only to a restriction fragment(s) from R. capsulata, R. palustris, R. sphaeroides, T. neapolitanus, and T. nivea. The fragment(s) of Rhodopseudomonas species is the Form II rbcL of these organisms. The restriction fragments of T. neapolitanus and T. nivea were also different from those elucidated by the A. nidulans probe, suggesting the presence of a second (different) rbcL in these organisms. Positive hybridization was not obtained using any of the probes with DNA from Beggiatoa alba, Chlorobium vibrioforme or Chloroflexus aurantiacus. It appears that all rbcL have evolved from a common ancestor. Our data are consistent with and supportive of the evolutionary scheme for RuBisCO proposed by Akazawa, Takabe, and Kobayashi [1].

Molecular evolution of the large subunit of ribulose-1,5-bisphosphate carboxylase/oxygenase (RuBisCO)

The evolutionary relationship of the RuBisCO large subunit gene(s) (rbcL) of several prokaryotes was examined using the technique of heterologous DNA hybridization. Restriction fragments of cloned rbcL from Anacystis nidulans 6301, Chlamydomonas reinhardtii, Rhodospirillum rubrum, and maize were nick-translated and used as probes. The C. reinhardtii and maize probes hybridized with restriction fragment(s) only from cyanobacteria: Agmenellum quadruplicatum, Fremyella diplosiphon, and Mastigocladus laminosus. In addition, the A. nidulans probe hybridized with restriction fragment(s) from Alcaligenes eutrophus, Chromatium vinosum, Nitrobacter hamburgensis, Paracoccus denitrificans, Pseudomonas oxalaticus, Rhodomicrobium vannielii, Rhodopseudomonas capsulata, Rhodopseudomonas palustris, Rhodopseudomonas sphaeroides, Thiobacillus intermedius, Thiobacillus neapolitanus, and Thiothrix nivea. The elucidated fragment of Rhodopseudomonas species is presumably for the Form I RuBisCO LSU of these organisms. The R. rubrum probe hybridized only to a restriction fragment(s) from R. capsulata, R. palustris, R. sphaeroides, T. neapolitanus, and T. nivea. The fragment(s) of Rhodopseudomonas species is the Form II rbcL of these organisms. The restriction fragments of T. neapolitanus and T. nivea were also different from those elucidated by the A. nidulans probe, suggesting the presence of a second (different) rbcL in these organisms. Positive hybridization was not obtained using any of the probes with DNA from Beggiatoa alba, Chlorobium vibrioforme or Chloroflexus aurantiacus. It appears that all rbcL have evolved from a common ancestor. Our data are consistent with and supportive of the evolutionary scheme for RuBisCO proposed by Akazawa, Takabe, and Kobayashi [1].

Isolation of hydrogen-oxidation gene from Alcaligenes hydrogenophilus and its expression in Pseudomonas oxalaticus.

A gene bank of a megaplasmid encoding the hydrogen-oxidizing enzyme system (Hox) in Alcaligenes hydrogenophilus was constructed using a broad host range cosmid vector pVK102, and established in Escherichia coli. Hybrid cosmids containing hox genes were identified by transferring the bank into Pseudomonas oxalaticus OX1 and screening colonies for the ability of H2-dependent autotrophic growth. About 800 colonies were formed under autotrophic conditions. One of the Hox+ transconjugants was isolated and its hydrogenases activities were measured. Although soluble hydrogenase was not detected, the Hox+ transconjugant had four times the membrane-bound hydrogenase activity of A. hydrogenophilus.

The inhibition of ribulose 1,5-bisphosphate carboxylase oxygenase by several 2-carboxyhexitol 1- and 6-phosphates

The condensation of D-fructose 6-phosphate or 1-phosphate with cyanide has been used to synthesize 2-carboxyhexitol 6-phosphates and 1-phosphates. The products have been characterized in terms of their action on ribulose bisphosphate carboxylase/oxygenase. The reaction of D-fructose 6-phosphate with cyanide is four times as fast (at 22°C) at pH 7.5 than at pH 11.5 and the primary products of condensation are more easily isolated by anion exchange chromatography. Two minor chromatographic peaks (I and II) for diastereomeric 2-carboxyhexitol 6-phosphates are isolated in addition to two major peaks, III and IV, which are lactones. The lactones are those of 2-C-carboxy-D-glucitol 6-phosphate (CG6P) in peak III and 2-C-carboxy-D-mannitol 6-phosphate (CM6P) in peak IV, as established after dephosphorylation by the relative rates of oxidation by periodate and by gas chromatographic retention times of the acetates. Analogous methodology has been used to synthesize the diastereomeric 2-carboxy-hexitol 1-phosphates (CG1P and CM1P) and their lactones from D-fructose 1-phosphate. The four carboxylates inhibit ribulose bisphosphate carboxylase/oxygenase from spinach or Pseudomonas oxalaticus in the following decreasing order of potency: CG6P, CM6P, CG1P, CM1P. The inhibition pattern suggests that the binding of the 5-phosphate moiety of the intermediate in the reaction catalyzed by ribulose bisphosphate carboxylase/oxygenase may be stronger by an order of magnitude than the binding of the 1-phosphate group.

Potentiometric system for selective formate measurement and improvement of response characteristics by permeation of cells

A formate-selective biocatalytic potentiometric electrode system based on whole cells Pseudomonas oxalaticus has been developed. Permeation of the Gram-negative Pseudomonas oxalaticus microbial cells by EDTA is shown to improve the response slope from 38 mV/decade to 49 mV/decade, resulting in increased sensitivity. Out of 13 possibly interfering compounds tested, only pyruvate and lactate showed moderate response.

Regulation of Autotrophic and Heterotrophic Metabolism in Pseudomonas oxalaticus OX1. Growth on Fructose and on Mixtures of Fructose and Formate in Batch and Continuous Cultures

In Pseudomonas oxalaticus the synthesis of enzymes involved in autotrophic CO2 fixation via the Calvin cycle is regulated by repression/derepression. During growth of the organism on fructose alone, the synthesis of ribulosebisphosphate carboxylase (RuBPCase) remained fully repressed, both in batch culture and in fructose-limited continuous cultures at various dilution rates. Growth in batch culture on a mixture of fructose and formate resulted in the simultaneous utilization of both substrates. Under these conditions we observed synthesis of RuBPCase up to high levels, indicating that formate did not merely function as an ancillary energy source in the metabolism of fructose, but stimulated autotrophic CO2 fixation via the Calvin cycle. In subsequent experiments growth of P. oxalaticus on mixtures of fructose and formate was studied in carbon source-limited continuous cultures. In these experiments further evidence was obtained that fructose is a poor source of (co-)repressor molecules for the synthesis of RuBPCase in the presence of formate. Thus, addition of formate to the medium reservoir of a fructose-limited continuous culture resulted in derepression of RuBPCase synthesis at (relatively) high ratios of fructose over formate. In the reverse experiment the specific activity of RuBPCase decreased with increasing concentrations of fructose in the medium reservoir. However, it can be calculated that the total capacity of RuBPCase in the culture to fix CO2 remained constant. In these experiments the dry weight produced on the various mixtures equalled the sum of the dry weight values obtained during growth on the same amounts of the two substrates separately. This indicated that, once RuBPCase was present, autotrophic and heterotrophic carbon assimilation pathways functioned simultaneously and independently of each other. Possible explanations for the low repressing effect of fructose on autotrophic CO2 fixation in P. oxalaticus are discussed.

Transfer and expression of the herbicide-degrading plasmid pJP4 in aerobic autotrophic bacteria.

Plasmid pJP4 encoding the ability to degrade the herbicide 2,4-dichlorophenoxyacetic acid (Tfd+) was transferred by conjugation from Escherichia coli JMP397 to various lithoautotrophic strains of Alcaligenes eutrophus and to the autotrophic bacterium Pseudomonas oxalaticus. The herbicide-degrading function of the plasmid was phenotypically expressed in all of the recipients. The majority of Tfd+ transconjugants also exhibited additional plasmid-encoded properties such as 3-chlorobenzoate degradation, resistance to mercuric ions, and sensitivity to the male-specific bacteriophage PR11. Furthermore, Tfd+ transconjugants were able to act as donors of plasmid pJP4. Physical evidence is presented by agarose gel electrophoresis showing that plasmid pJP4 coexisted with the resident plasmids widely distributed in this group of bacteria. However, in some of the hosts plasmids pJP4 was not stably maintained, had a reduced size and tended to form multimers.

90] Formate dehydrogenase from Pseudomonas oxalaticus

This chapter describes the assay method, purification, and properties of formate dehydrogenase isolated from Pseudomonas oxalaticus . When Pseudomonas oxalaticus is grown on formate as the main carbon and energy source, formate dehydrogenase is the key enzyme that generates nicotinamide adenine dinucleotide dehydrogenase (NADH) and CO 2 . The CO 2 enters the ribulose diphosphate carboxylase reaction. The continuous spectrophotometric assay is based on formate oxidation by nicotinamide adenine dinucleotide kinase (NAD + ) or ferricyanide. The appearance of NADH is followed at 334 or 365 nm; ferricyanide decrease is followed at 405 nm. The NADH oxidase activity of the enzyme does not interfere in the presence of 5 m M sulfide. The steps involved in the purification of the enzyme are also discussed in the chapter. All steps are performed at 0°C. Stored enzyme solutions and all solutions for chromatographic steps are kept in an atmosphere of argon. Essential conditions of the purification procedure of the enzyme are (1) the exclusion of light and oxygen, (2) the absence of formate, (3) low temperature, and (3) rapid performance, with possible interruption only after precipitation by ammonium sulfate of the eluate of the hydroxyapatite chromatography.

Physical evidence for plasmids in autotrophic, especially hydrogen-oxidizing bacteria

Agarose gel electrophoresis of crude lysates from 23 species of autotrophic bacteria revealed plasmids of various sizes in 12 species. The plasmid pattern varied considerably. While the majority of the plasmid-bearing species harbored one or two plasmids, one species, Alcaligenes latus, exhibited more than six ccc-DNA bands. With one exception the molecular masses of the plasmids were 50×106 or higher. In Achromobacter carboxydus, Alcaligenes latus, Derxia gummosa and three strains of Paracoccus denitrificans large plasmids of molecular masses higher than 300×106 were resolved. The examination of Thiobacillus A2 resulted in the discovery of two plasmids while Pseudomonas oxalaticus was apparently free of resident plasmid DNA. So far these plasmids can only be characterized as cryptic. Future studies may allow to correlate them with specific metabolic activities of their hosts such as the ability to grow on carbon monoxide or thiosulfate, to fix molecular nitrogen and to form soluble NAD-reducing and/or membrane-bound hydrogenases.