Expression of nitrogen fixation genes in Rhodobacter capsulatus is repressed by ammonium at different regulatory levels including an NtrC-independent mechanism controlling NifA activity. In contrast to R. capsulatus NifA, heterologous NifA proteins of Klebsiella pneumoniae and Rhizobium meliloti, respectively, were not subjected to this posttranslational ammonium control in R. capsulatus. The characterization of ammonium-tolerant R. capsulatus NifA1 mutants indicated that the N-terminal domain of NifA was involved in posttranslational regulation. Analysis of a double mutant carrying amino acid substitutions in both the N-terminal domain and the C-terminal DNA-binding domain gave rise to the hypothesis that an interaction between these two domains might be involved in ammonium regulation of NifA activity. Western analysis demonstrated that both constitutively expressed wild-type and ammonium-tolerant NifA1 proteins exhibited high stability and accumulated to comparable levels in cells grown in the presence of ammonium excluding the possibility that proteolytic degradation was responsible for ammonium-dependent inactivation of NifA.
Biological nitrogen fixation in proteobacteria is catalyzed by the nitrogenase enzyme complex. Expression of the structural genes of nitrogenase nifHDK and other nif genes is controlled by the σ54(NtrA)-dependent activator NifA . The NifA proteins consist of at least three distinct domains, namely a regulatory N-terminal domain, a central ATP-binding activator domain, a DNA-binding C-terminal domain, and – in the case of the rhizobial type of NifA proteins – an interdomain linker connecting the central and the C-terminal domain, which is believed to confer oxygen sensitivity to these types of NifA proteins [6,8].
In Klebsiella pneumoniae, Azotobacter vinelandii, and Enterobacter agglomerans, NifL binds to NifA in cells grown in the presence of ammonium, and thereby inhibits NifA activity. A. vinelandii NifL directly interacts with the central catalytic domain of NifA [3,15,23]. However, for bacteria lacking NifL there is evidence that the N-terminal domain of NifA itself somehow senses the nitrogen status of the cell. In Azospirillum brasilense, NifA is synthesized both in conditions compatible and incompatible with nitrogen fixation (e.g. absence or presence of ammonium), and NifA is either present in an active or an inactive form in response to environmental conditions [1,7]. The activity of NifA strictly depends on the glnB product [5,16,28], but N-terminal deleted forms of NifA are active independently of the presence or absence of GlnB and ammonia . A mutation of a single tyrosine residue at position 18 of the N-terminal domain of NifA is sufficient to allow the protein to escape GlnB control but not ammonia control . Similar results have been described for Herbaspirillum seropedicae NifA, since a truncated NifA protein lacking the N-terminal domain is active in the presence of ammonium [21,27]. Furthermore, the N-terminal domain of H. seropedicae NifA can in-trans restore ammonium control of the N-truncated NifA activity .
In contrast to other diazotrophic bacteria, Rhodobacter capsulatus contains two functional copies of nifA– called nifA1 and nifA2– which encode NifA proteins of 579 and 582 amino acid residues, respectively. Both NifA proteins differ only in their N-terminal 19 or 22 amino acid residues, respectively, whereas the remainders of both proteins (560 amino acid residues) are identical to each other. Synthesis of NifA is controlled at the transcriptional level by an NtrB/NtrC-like system [4,10,17]. Analysis of an R. capsulatus nifH-lacZ reporter strain constitutively expressing nifA1 identified an NtrC-independent regulatory mechanism repressing NifA-mediated nif gene expression in the presence of high ammonium concentrations . To examine the molecular basis of ammonium sensitivity of NifA activity, in this work a random mutagenesis of the nifA1 gene was performed to identify mutant nifA1 genes mediating ammonium-tolerant nif gene expression. Analysis of mutant nifA1 genes indicated that not only mutations in the N-terminal domain but also mutations in the interdomain linker led to ammonium-tolerant NifA activity.
2Materials and methods
2.1Bacterial strains and growth conditions
The bacterial strains and plasmids used in this study are listed in Table 1. R. capsulatus nifA1 mutant strain R231A was constructed in a similar way as mutant strain R231 . The difference was that in strain R231 three BglII fragments encompassing most of nifA1 and nifB1 were replaced by a gentamicin cassette, whereas in strain R231A the same cassette was inserted into the BglII site at the 5′-end of nifA1 leaving nifB1 unaffected. Methods for conjugational plasmid transfer between Escherichia coli and R. capsulatus and the selection of mutants, growth media, growth conditions, and antibiotic concentrations were as previously described [14,18].
Table 1. Bacterial strains and plasmids used in this study
DNA isolation, restriction enzyme analysis, agarose gel electrophoresis, and cloning procedures were performed using standard methods . Restriction enzymes and T4 DNA ligase were purchased from Pharmacia LKB and MBI Fermentas, respectively, and used as recommended by the suppliers. DNA sequence analysis was carried out using the A.L.F. DNA sequencer (Pharmacia LKB) as instructed by the manufacturer. To create random mutations, PCR amplification was carried out in a Perkin-Elmer Cetus DNA Thermal Cycler 480 using Taq polymerase without proof-reading activity.
2.3Construction of plasmids carrying constitutively expressed nifA genes
In order to constitutively express wild-type and mutant nifA genes of R. capsulatus, K. pneumoniae, and Rhizobium meliloti, respectively, appropriate restriction fragments (Table 1) were cloned into derivatives of the mobilizable broad-host-range plasmids pPHU231 and pPHU234, respectively. In the resulting plasmids nifA expression was driven by the constitutively expressed aphII promoter located on the 1861-bp HindIII-BamHI fragment of transposon Tn5.
The R. capsulatus reporter strain R231-R279-R372 contains lesions in nifA1 and nifA2 in combination with a chromosomal nifH-lacZ fusion . Broad-host-range plasmids carrying constitutively expressed homologous or heterologous nifA genes (Table 1) were introduced into this reporter strain, and the resulting strains were grown phototrophically as batch cultures in RCV minimal medium with either 15 mM ammonium or 9.5 mM serine until late exponential phase . β-Galactosidase activities of these R. capsulatus strains were determined by the sodium dodecyl sulfate (SDS)–chloroform method .
Protein extracts of R. capsulatus were isolated as described earlier . Proteins were separated on 10% SDS–polyacrylamide gels and subsequently blotted to PVDF membranes (Bio-Rad). Detection of NifA1 wild-type and mutant proteins was performed using the ECL kit (Amersham Pharmacia Biotech). The NifA1-specific antiserum was raised against a synthetic oligopeptide (MTDQQSRPASPRRRSTQC) corresponding to amino acid residues 1–17 of R. capsulatus NifA1 plus an additional cysteine residue for KLH coupling (Eurogentec). Since NifA1 and NifA2 differed in their N-terminal 19 or 22 amino acid residues, respectively, the antiserum raised against the synthetic oligopeptide was specific for NifA1 (and mutant NifA1 proteins), but did not recognize NifA2 (data not shown).
3Results and discussion
3.1Complementation of an R. capsulatus nifA1-nifA2 double deletion strain with constitutively expressed homologous and heterologous nifA genes
As a prerequisite for the analysis of the NtrC-independent mechanism of ammonium repression of NifA activity , we analyzed nif gene expression mediated by different constitutively expressed homologous and heterologous nifA genes. For this purpose the nifA1 and nifA2 genes of R. capsulatus (Rc) as well as the nifA genes of K. pneumoniae (Kp) and R. meliloti (Rm) were cloned into mobilizable broad-host-range vector plasmids allowing constitutive expression of the nifA genes (nifAc) driven by the kanamycin resistance promoter (PaphII) of transposon Tn5 (Section 2). The resulting plasmids were designated pAPA1 (Rc-nifA1c), pAPA2 (Rc-nifA2c), pMAK5-Spc (Kp-nifAc), and pBK17-I (Rm-nifAc). These four plasmids were transferred into an R. capsulatus nifA1-nifA2 double mutant strain (Section 2; ), respectively, and the corresponding four strains were able to grow diazotrophically and exhibited comparable growth rates (data not shown). These results confirmed that either Rc-nifA1c or Rc-nifA2c was sufficient for diazotrophic growth, and demonstrated that the heterologous nifA genes from K. pneumoniae and R. meliloti could substitute for R. capsulatus nifA.
To quantify nif gene expression mediated by the different homologous and heterologous nifAc genes, hybrid plasmids pAPA1, pAPA2, pMAK5-Spc, and pBK17-I were introduced into the R. capsulatus reporter strain R231-R279-R372 (ΔnifA1, ΔnifA2, chromosomal nifH-lacZ reporter fusion). The resulting strains were grown either under nitrogenase derepressing (serine as sole N source) or repressing conditions (ammonium as N source), and β-galactosidase activities were measured in late-logarithmic cultures (Table 2). As expected, all nifAc genes mediated considerable nif gene expression under derepressing conditions. Despite the fact that transcription of all four nifA genes was driven by the same promoter (PaphII), the levels of expression of the nifH-lacZ fusion differed remarkably (Table 2). These differences may be explained as follows. (i) In the parental R. capsulatus strain two nifA genes are present and both copies are expressed under nitrogen-limiting conditions, whereas in strains carrying either Rc-nifA1c or Rc-nifA2c only one nifA copy was present. (ii) In addition to this copy number effect, promoter strengths of the native promoters of nifA1 and nifA2 will differ from PaphII. (iii) Last but not least, the translation efficiencies and promoter affinities of the heterologous NifA proteins might differ from the homologous proteins. However, the observed differences in absolute activities do not hinder the comparison of nifH-lacZ expression for a selected strain under different physiological conditions (e.g. presence and absence of fixed nitrogen). In contrast to Kp-nifAc and Rm-nifAc, the Rc-nifA1c and Rc-nifA2c genes did not lead to significant expression of the nifH-lacZ fusion in the presence of ammonium. Although both R. capsulatus and R. meliloti NifA belong to the class of oxygen-sensitive interdomain linker-containing NifA proteins , they differed in their ability to activate nif gene expression in the presence of ammonium. Therefore, ammonium sensitivity of NifA activity seems to be an intrinsic property of R. capsulatus NifA1 and NifA2 but may not be a general feature of rhizobial-type NifA proteins.
Table 2. Analysis of a chromosomal nifH-lacZ fusion in R. capsulatus strains carrying constitutively expressed homologous and heterologous nifA genes
aAbbreviations: Rc, R. capsulatus; Kp, K. pneumoniae; Rm, R. meliloti; nifAc, constitutive nifA expression driven by the aphII promoter of transposon Tn5.
R372 (no plasmid)
ΔnifA1-A2, nifH-lacZ, vector
ΔnifA1-A2, nifH-lacZ, Rc-nifA1c
ΔnifA1-A2, nifH-lacZ, Rc-nifA2c
ΔnifA1-A2, nifH-lacZ, Kp-nifAc
ΔnifA1-A2, nifH-lacZ, Rm-nifAc
3.2Isolation of R. capsulatus nifA1 mutant genes mediating nif gene expression in the presence of ammonium
In order to analyze the molecular basis for ammonium sensitivity of R. capsulatus NifA activity, we searched for ammonium-tolerant Rc-NifA1 mutant proteins. In a first approach, R. capsulatus nifA1 mutant genes were constructed coding for proteins carrying different deletions which either completely removed or partly eliminated the N-terminal domain (data not shown). However, in contrast to NifA of A. brasilense and H. seropedicae, in which deletions of the N-terminal domain lead to active ammonium-tolerant NifA proteins (1, 22, 27), all R. capsulatus NifA1 mutant proteins turned out to be completely inactive both in an R. capsulatus and also in an E. coli background (data not shown), indicating that the N-terminal domain somehow might be essential for activity of R. capsulatus NifA.
To create NifA mutants carrying only substitutions of single amino acid residues, a random PCR mutagenesis of the Rc-nifA1 gene was carried out making use of errors of Taq polymerase. For this purpose a 2440-bp SmaI-SacI fragment containing the entire nifA1 gene was PCR amplified, and subsequently cloned into vector plasmid pAPA3-Spc, in which expression of nifA1 was driven by PaphII. A pool of hybrid plasmids carrying wild-type and mutant nifA1 genes was conjugationally introduced into the R. capsulatus nifH-lacZ reporter strain R231-R279-R372. Three clones producing dark blue colonies on X-gal-containing plates in the presence of high ammonium concentrations were identified. These clones (carrying hybrid plasmids pAPA1-1, pAPA1-2, and pAPA1-5) were assumed to contain nifA1 mutant genes coding for ammonium-tolerant NifA1 proteins. Western analyses using NifA1-specific antibodies ruled out that this ammonium tolerance was simply due to increased amounts or stability of the mutant NifA proteins (see below; Fig. 2). The sites of mutations in plasmids pAPA1-1, pAPA1-2, and pAPA1-5 were identified by DNA sequence analysis (Fig. 1).
Since the deduced NifA1-1 protein (encoded by hybrid plasmid pAPA1-1) contained mutations in the N-terminal (L66Q) as well as in the C-terminal (K531T) domains, both mutations were separated in order to analyze the influence of the individual amino acid substitutions. For this purpose appropriate DNA fragments of the nifA1-1 mutant gene and the nifA1 wild-type gene were combined resulting in hybrid plasmids pAPA1-11 and pAPA1-12 coding for NifA1-11 (L66Q) and NifA1-12 (K531T), respectively.
In order to examine nif gene expression mediated by ammonium-tolerant NifA proteins under different physiological conditions, hybrid plasmids pAPA1-1, pAPA1-11, pAPA1-12, pAPA1-2, and pAPA1-5 were transferred into the R. capsulatus nifA1-nifA2 double mutant carrying a chromosomal nifH-lacZ reporter fusion (R231-R279-R372), and β-galactosidase activities were determined as described in Section 2 (Fig. 1). The results of these experiments can be summarized as follows. (i) A single mutation in the N-terminal domain (NifA1-11: L66Q) was sufficient for ammonium-tolerant NifA activity. (ii) A combination of this mutation (L66Q) with a second mutation in the C-terminal domain (K531T) resulted in a comparable phenotype, whereas mutant protein NifA1-12 (K531T) did not activate nif gene expression at all. Therefore, the N-terminal mutation conferred not only ammonium tolerance to NifA1-1 but also suppressed the negative phenotype of the C-terminal mutation indicating an intra- or inter-molecular protein–protein interaction between the N-terminal and the DNA-binding domain of NifA1. (iii) Another mutation in the N-terminal domain (NifA1-5: V42E) also resulted in ammonium-tolerant NifA activity albeit at lower levels confirming the regulatory role of the N-terminal domain. (iv) However, not only mutations in the N-terminal domain but also mutations in the interdomain linker (NifA1-2: I460F, E477Q) led to ammonium-tolerant NifA activity. It remains speculative whether these mutations affected the proposed intra- or inter-molecular interaction of the N-terminal and the DNA-binding domain, and thereby, indirectly mediated ammonium tolerance to NifA1-2.
3.3Accumulation of wild-type and mutant NifA1 proteins in the presence of ammonium
To analyze whether the differences in ammonium tolerance of NifA1 activities were due to differences in the amount and/or stability of wild-type and mutant NifA1 proteins, accumulation of NifA1 proteins was examined by Western analysis (Section 2). For this purpose plasmids carrying constitutively expressed nifA1 wild-type (pAPA1) and mutant genes (pAPA1-1, pAPA1-2, pAPA1-5, pAPA1-11, and pAPA1-12) were introduced into the nifA1-nifA2 double deletion strain R231-R279. The resulting R. capsulatus strains were grown photoheterotrophically either in the presence or absence of ammonium until late-logarithmic phase. After protein extraction, SDS–PAGE and blotting, wild-type and mutant NifA proteins were detected using a NifA1-specific antiserum (Section 2). The results shown in Fig. 2 can be summarized as follows. (i) In the wild-type strain B10S significant amounts of NifA1 could only be detected under derepressing conditions (Fig. 2A) reflecting the NtrC-dependent transcriptional control of nifA1 expression. (ii) In contrast, the constitutively expressed wild-type NifA1 protein in the nifA double deletion strain accumulated not only in cells grown under derepressing conditions but also in cells grown in the presence of ammonium. However, despite its presence in ammonium-grown cells, NifA was unable to activate nif gene expression under these conditions (Table 2). A similar situation has been described for Azorhizobium caulinodans NifA . At least in the gel system used in this study, no changes in electrophoretic mobility could be observed for ammonium-inactivated NifA indicating that neither proteolytic degradation nor covalent modification might be responsible for inactivation of NifA. The most likely mechanism for regulation of NifA activity, however, is expected to be an interaction with GlnB and/or GlnK, since a glnB/glnK double mutant showed no ammonium control of NifA activity . (iii) The ammonium-tolerant mutant proteins NifA1-1, NifA1-2, and NifA1-11 accumulated to levels comparable to the ammonium-sensitive wild-type NifA1 protein ruling out that ammonium tolerance was due to increased amounts of the mutant NifA proteins. (iv) Significantly reduced amounts of mutant protein NifA1-5 were found both under derepressing and repressing conditions. This result corresponded well to the observed reduction in the ability of NifA1-5 to activate a nifH-lacZ fusion. As shown in Fig. 1, NifA1-5 was able to activate nifH-lacZ equally well in the presence and absence of ammonium but exhibited only 20% of the wild-type activity. (v) Mutant protein NifA1-12 carrying a single amino acid exchange in the DNA-binding domain did not accumulate even in the absence of ammonium. Therefore, the failure of NifA1-12 to activate nif gene expression (Fig. 1) was due to instability of the mutant NifA1-12 protein. Interestingly, this instability of mutant protein NifA1-12 was suppressed by an amino acid exchange in the N-terminal domain (L66Q) leading to an ammonium-tolerant NifA protein. This result indicated long range changes in protein conformation or protein–protein interactions between the N-terminal domain and the DNA-binding domain of R. capsulatus NifA. This hypothesis is corroborated by data of Monteiro et al.  demonstrating that an independently expressed N-terminal domain of H. seropedicae NifA is able to restore the ammonium control of an N-truncated NifA protein.
In addition to these experiments measuring the amounts of NifA1 in ammonium-grown cells (Fig. 2), we analyzed the stability of NifA1 in a time course experiment. For this purpose R. capsulatus strains overexpressing wild-type and mutant NifA1 proteins were cultured in RCV minimal medium (containing serine as the sole N source) until early-logarithmic phase before ammonium (15 mM) and/or chloramphenicol (25 μg ml−1) were added. Aliquots of the cultures were taken at different times (0 min, 10 min, 1 h, 2 h, and 5 h after addition of ammonium) for preparation of protein extracts, which were analyzed by Western blotting using the NifA1-specific antiserum. The amounts of the wild-type and mutant NifA1 proteins did not decrease during the 5 h time range of the experiment (data not shown) confirming the high stability of NifA1 in the presence and absence of ammonium. This experiment also excluded the possibility that proteolytic degradation might be involved in ammonium regulation of the wild-type NifA1 protein. The high stability of NifA in the presence of ammonium indicated that the loss of NifA activity might be reversible, and might be caused by interaction with GlnB and/or GlnK  or with another yet unidentified protein.
We thank P.-B. Kamp for the construction of plasmids. This work was supported by financial grants from the Fonds der Chemischen Industrie and the Deutsche Forschungsgemeinschaft, Germany.