SciELO - Scientific Electronic Library Online

 
vol.31 issue11A chromatographic method for the production of a human immunoglobulin G solution for intravenous use author indexsubject indexarticles search
Home Pagealphabetic serial listing  

Services on Demand

Article

Indicators

Related links

Share


Brazilian Journal of Medical and Biological Research

On-line version ISSN 1414-431X

Braz J Med Biol Res vol. 31 n. 11 Ribeirão Preto Nov. 1998

http://dx.doi.org/10.1590/S0100-879X1998001100001 

Braz J Med Biol Res, November 1998, Volume 31(11) 1363-1374

Purification and binding analysis of the nitrogen fixation regulatory NifA protein from Azospirillum brasilense

L.M.P. Passaglia2, C. Van Soom3, A. Schrank1 and I.S. Schrank1

Departamentos de 1Biotecnologia and 2Genética, Centro de Biotecnologia, Universidade Federal do Rio Grande do Sul, Porto Alegre, RS, Brasil
3F.A. Janssens Laboratory of Genetics, Katholieke Universiteit Leuven, Heverlee, Belgium

down.gif (51 bytes) Abstract
Introduction
Material and Methods
Results and Discussion
References
Acknowledgments
Correspondence and Footnotes


Abstract  

NifA protein activates transcription of nitrogen fixation operons by the alternative s54 holoenzyme form of RNA polymerase. This protein binds to a well-defined upstream activator sequence (UAS) located at the -200/-100 position of nif promoters with the consensus motif TGT-N10-ACA. NifA of Azospirillum brasilense was purified in the form of a glutathione-S-transferase (GST)-NifA fusion protein and proteolytic release of GST yielded inactive and partially soluble NifA. However, the purified NifA was able to induce the production of specific anti-A. brasilense NifA-antiserum that recognized NifA from A. brasilense but not from K. pneumoniae. Both GST-NifA and NifA expressed from the E. coli tac promoter are able to activate transcription from the nifHDK promoter but only in an A. brasilense background. In order to investigate the mechanism that regulates NifA binding capacity we have used E. coli total protein extracts expressing A. brasilense nifA in mobility shift assays. DNA fragments carrying the two overlapping, wild-type or mutated UAS motifs present in the nifH promoter region revealed a retarded band of related size. These data show that the binding activity present in the C-terminal domain of A. brasilense NifA protein is still functional even in the presence of oxygen.

Key words: NifA protein, NifA antibodies, Azospirillum brasilense, gel shift assay, NifA activity, upstream activator sequence


Introduction

In many diazotroph organisms, the nitrogen fixation genes (nif) are transcribed by an alternative holoenzyme form of RNA polymerase and require an activator protein, NifA. Both activity and synthesis of NifA are regulated in response to environmental effectors. Two major signals regulate nitrogen fixation, oxygen and ammonia, and this regulation occurs mainly at the level of nif gene transcription mediated by NifA. NifA acts as a DNA-binding protein which recognizes typical upstream promoter sequences (TGT-N10-ACA) showing many enhancer element properties located upstream of promoters of the -24/-12 type (1). The integration host factor protein (IHF) stimulates K. pneumoniae NifA-mediated activation of nif promoters by facilitating DNA loop formation (2).

NifA activates nif transcription under oxygen and nitrogen-limiting conditions but becomes inactive when levels of fixed nitrogen or oxygen increase (3). This inactivation has been attributed to the NifL protein, which senses changes in the oxygen and/or nitrogen status of the cell in Klebsiella pneumoniae and Azotobacter vinelandii (4-7). In A. brasilense and members of the Rhizobiaceae family, where NifL has not been found, the NifA protein itself is inactivated by oxygen. The NifA protein contains a conserved motif of cysteine residues, which is absent in K. pneumoniae and A. vinelandii, involved in sensitivity to oxygen in Bradyrhizobium japonicum and possibly in A. brasilense (8,9).

NifA proteins are structurally similar to each other, and three independent domains, separated by two inter-linker-domains (IDL), were defined (8,10,11). The C-terminal domain contains a helix-turn-helix DNA binding motif (12,13); the central domain, involved in catalysis, contains the nucleotide binding site (10,14) and the N-terminal domain is not essential for K. pneumoniae NifA activity. In K. pneumoniae this latter domain, that has no homologue, appears to decrease the sensitivity to NifL (10,12,15). In A. brasilense, NifA is synthesized in an inactive form under conditions not compatible with nitrogen fixation (16). Moreover, its regulation differs considerably from that found in other free living diazotrophs. It has been proposed that, in A. brasilense, the N-terminal domain plays an inhibitory role which leads to the inactivation of NifA in the absence of PII protein (17,18). PII protein plays an essential role in the regulation of nitrogen metabolism in E. coli by controlling the activity of GS (19,20) and it appears that glnB is essential for nitrogen fixation in A. brasilense (18). Another important feature of A. brasilense is the regulation of nitrogenase activity at the post-translational level by a "switch-off" mechanism in response to environmental changes in ammonia concentration (21). Therefore, the plant growth-promoting A. brasilense, which associates with roots of economically important grasses, has some characteristic features at the level of nitrogen fixation regulation.

In this paper, we report the production of NifA from A. brasilense in E. coli cells, its purification, the production of specific antiserum and the analysis of the activity of the recombinant NifA protein. We also show that the recombinant NifA proteins were active only in A. brasilense cells and were able to induce transcription from the nifH promoter. The ability of A. brasilense NifA to bind to upstream activator sequences present on nif genes even in its inactive form in response to oxygen is also demonstrated.


Material and Methods

Bacterial strains, plasmids and growth conditions

The bacterial strains and plasmids used in the present study are listed in Tables 1 and 2. LB medium (22) was used for growing E. coli strains at 37oC. YT medium (22) was used to induce the glutathione-S-transferase (GST)-NifA fusion protein on E. coli BL21 strains at 30oC. A. brasilense strains were grown at 30oC in either LB medium or MMAb minimal medium (23) using 0.5% malate as carbon source. MMAb without ammonium chloride was used when nitrogen-free medium was required. The medium was supplemented with the antibiotics tetracycline (Tc), ampicillin (Ap) or kanamycin (Km) when necessary, at concentrations of 10, 100 or 30 µg/ml, respectively.

Conjugation experiments were performed on D-plates (8 g/l Bacto nutrient broth, 0.25 g/l magnesium sulfate, 1 g/l potassium chloride, 0.01 g/l manganese chloride). MMAb medium supplemented with Tc and Km was used to select A. brasilense transconjugants.

DNA manipulations, nitrogenase assays, ß-galactosidase assays and conjugation

Plasmid DNA preparation, restriction enzyme analysis, transformation and hybridization were performed as described in reference 22. Restriction endonucleases and other enzymes were purchased from Pharmacia (Uppsala, Sweden), Gibco/BRL (Gaithersburg, MD) or Promega (Madison, WI) and used according to the manufacturer's instructions.

Nitrogenase activity was measured by the acetylene reduction method described by Burris (24). Ethylene production was quantified on a Plot fused silica column (50 m x 0.32 nm, 5 µm Al2O3/KCl, Chrompack) installed in a Hewlett Packard 5890A gas chromatograph. Propane was used as internal standard.

Determination of ß-galactosidase activity of the nif-lacZ promoter fusions was carried out according to the standard procedure of Miller (as described in reference 22) at 25oC.

Plasmid mobilization from E. coli S17.1 to A. brasilense Sp7 by biparental conjugation was performed as previously described (25).

Plasmid constructions

The strategy used to obtain the pKKAbA plasmid is shown in Figure 1A. DNA from plasmid p10 was digested to generate the 2.2-kb SmaI/SalI DNA fragment carrying the coding region for NifA from A. brasilense. The SmaI site is located at 91 bp from the ATG start codon. In the pKKAbA (pKK233-3 vector) construct the A. brasilense nifA gene is under the control of the tac promoter.

The nifA gene from A. brasilense was also cloned into three other plasmid vectors. Plasmid pKKAbA was digested with SalI and the entire construction, ptac-nifA, was cloned into the SalI site of the pACYC184 vector (pAAbA plasmid). The pGEX 4T-1 vector was used to generate a GST-NifA fusion protein. The SmaI/SalI 2.2-kb DNA fragment from plasmid p10 was also cloned into the pGEX 4T-1 SmaI/SalI digested vector. In this way the nifA gene was cloned in frame with GST and the recombinant plasmid, pGEXAbA, was able to produce a GST-NifA fusion protein (Figure 2A).

For the complementation experiments, the nifA gene was cloned into the pLAFR3 cosmid. Figure 3A depicts the steps used to generate plasmid pLKSAbA carrying the ptac-nifA construction and Figure 3B shows the pLpucNifA derivative containing the GST-NifA fusion construction.

The nifA gene of K. pneumoniae was also cloned into the transcriptional vector pKK223-3 and into the translational fusion vector pGEX 4T-3. Plasmid pCRA37 (26) was digested with SalI, yielding a 2.9-kb DNA fragment carrying the K. pneumoniae nifLAB region. This fragment was cloned into the pBluescript KS+ vector generating the pKSKpLAB plasmid. To remove the nifL gene, pKSKpLAB was digested with XhoI and Bal 31 deletions were carried out. Two different deleted DNA fragments have been selected and confirmed by DNA sequencing. The first one, shown in Figure 1B, has only 91 bp upstream from the nifA ATG start codon and was cloned into the pKK223-3 vector under the control of the tac promoter to produce the recombinant plasmid pKKKpA. To construct plasmid pAKpA, the ptac-nifA region from pKKKpA was removed by SalI digestion and cloned into the pACYC184 SalI digested vector. The second Bal 31 deleted fragment was cloned into the pBluescript vector, named pKS3.1(5.1), and completely sequenced. Figure 2B shows the strategy used to clone the NifA as a fusion protein with GST into the pGEX 4T-3 vector to produce the plasmid pGEXKpA.


Figure 1 - Cloning strategy for the insertion of nifA, from A. brasilense (A) and from K. pneumoniae (B), into the pKK223-3 vector. The figure also shows the restriction map of plasmid p10 containing the nifA gene from A. brasilense (A) and plasmid pCRA37 containing the nifLAB genes from K. pneumoniae (B). The black boxes represent the pKK223-3 vector and the gray boxes represent the pUC18 vector in A and the pBluescript vector in B.

[View larger version of this image (34 K GIF file)]


Figure 2 - Cloning strategy for the insertion of nifA from A. brasilense (A) and from K. pneumoniae (B) into the pGEX 4T vector. The nucleotide sequence from the nifA gene is indicated in bold type. The first amino acids for the coding region of NifA are indicated in bold above the nucleotide sequence. Small capital letters represent the pUC18 polylinker. The region that represents the nifA gene/NifA protein is marked by an arrow.

[View larger version of this image (31 K GIF file)]


Figure 3 - Strategy for cloning the nifA gene from A. brasilense into cosmid pLAFR3. A shows the steps for cloning the transcriptional tac-nifA fusion construct, and B shows the steps for cloning the translational fusion GST-NifA. Black boxes indicate the pKK223-3 vector in A and the pGEX 4T-1 vector in B; striped boxes indicate the pBluescript vector in A and the pUC18 vector in B; gray boxes indicate the pLAFR3 vector. The stippled arrow indicates the sequence of the GST protein from vector pGEX 4T-1.

[View larger version of this image (14 K GIF file)]


Cell extracts and detection of NifA by Western blot

Coupled transcription/translation in S30 extracts (Promega), using either pKKAbA or pKKKpA plasmids, were prepared following the instructions provided by the manufacturers.

E. coli BL 21 extracts, carrying pGEXAbA or pGEXKpA plasmids, and DH5a, carrying plasmids pKKAbA or pAAbA, were prepared after growing the cells for 2 h at 30oC in YT medium supplemented with ampicillin. The fusion protein was induced in aerobically grown cultures by the addition of 0.1 mM IPTG and the cells were harvested after growing for another 3 h at the same temperature. Proteins (1/4 of the culture) were separated on 10% SDS/polyacrylamide gels and visualized by Coomassie blue staining. Western blotting was performed as described in reference 22. Detection was carried out using monoclonal antibodies against GST or A. brasilense NifA antiserum. The reaction was incubated with anti-mouse antibody labeled with alkaline phosphatase and developed with a goat anti-mouse (IgG) secondary antibody using the substrate B-CIP/NBT detection system (Sigma, St. Louis, MO).

Purification of NifA protein and immunological procedures

A. brasilense NifA protein was purified from E. coli harboring the pGEXAbA plasmid as described in the GST Gene Fusion System (Pharmacia 2nd edn.). The purified NifA protein was used as immunogen for the generation of mouse antiserum as follows. Five female BALB/c white mice were immunized by intradermal injection with NifA. The purified protein (10 mg) was mixed with an equal volume of Freund's complete adjuvant for the first injection, or with Freund's incomplete adjuvant for the second injection 12 days later. One week after the second injection, NifA antiserum was collected, diluted 1:100 and used for Western immunoblot experiments.

Gel shift motility assays

DNA-binding assays were performed as described before (27).

The oligonucleotides used in the PCR reactions were 5'TGCATCCCCCGTGCCAGTTCGCG3' and 5'CCTGACGCTGGCTCTGACGCTGG 3'. The first oligonucleotide was labeled with digoxigenin (DIG) and the detection system was that described in the DIG gel shift kit (Boehringer, Mannheim, Germany).


Results and Discussion

Detection of NifA activity on transcriptional fusions

To demonstrate the activity of A. brasilense NifA protein in vitro (binding assays), we directed the synthesis of this protein from an expression vector (pKK223-3) in a coupled transcription-translation system (S30 extracts). Two plasmids, pKKAbA and pKKKpA, carrying transcriptional fusions of A. brasilense and K. pneumoniae nifA genes with tac promoter were constructed (Figure 1) and tested. Induction of the expression from tac promoter in strains carrying the plasmids resulted in no detectable production of the 35S-labeled proteins on 10% SDS-polyacrylamide gels.

To analyze the capability of the nifA protein to activate the nifH promoter two other plasmid constructions were made. Plasmids pAAbA and pAKp6 carry the same transcriptional fusion present in pKKAbA and pKKKpA, but were cloned into a vector (pACYC184) compatible with pMCH (27), containing the A. brasilense nifH promoter fused to lacZ. Cell cultures were prepared from strain MC1061 carrying the pMCH plasmid and from the two different constructs which produced NifA either from A. brasilense or from K. pneumoniae. Transcriptional activation of the A. brasilense nifH promoter in vivo by K. pneumoniae NifA protein is shown in Table 3. The transcriptional fusion of A. brasilense nifA protein failed to activate the nifH promoter.

The transcription/translation coupled system was used previously to analyze the activity of NifA protein from R. meliloti (28) and K. pneumoniae (29,30). The activity present in the control construction (tac-nifA from K. pneumoniae) is in agreement with previous results which have shown that this protein is able to activate the nifH promoter from A. brasilense (27; Table 3). The failure of the A. brasilense tac-nifA fusion protein to activate the nifH promoter is probably due to its inactive form in the presence of oxygen (31).

Detection of NifA activity on translational fusions

In order to purify A. brasilense NifA and to induce the production of antibodies we cloned the coding region as a translational fusion with GST (pGEX vector). NifA protein from K. pneumoniae has been previously purified as active fusion protein with MBP using the whole sequence (13), only the catalytic domain (15), only the N-terminal domain or only the C-terminal domain (32). In the present study two plasmids were constructed, pGEXAbA carrying the A. brasilense NifA protein fused with GST and pGEXKpA carrying the GST-NifA fused protein from K. pneumoniae (Figure 2). After digestion with thrombin, the A. brasilense NifA protein has an addition of 31 amino acids from the NifA noncoding region of A. brasilense in the N-terminal domain. The GST-NifA fusion protein from K. pneumoniae has a NifA deleted for the 32 first amino acids and an insertion of 9 extra amino acids from the pUC18 polylinker (Figure 2).

To verify the constructions, ß-galactosidase activity was determined in E. coli cells carrying the pMCH plasmid and the activity of the truncated NifA from K. pneumoniae is shown in Table 3. The transcriptional activation shows low values due to the incompatibility between plasmids pMCH and pGEXKpA resulting in transient expression only. We demonstrated that the K. pneumoniae NifA fused with GST, partially deleted at the N-terminal domain, is in its active form. Similar results were reported by Berger et al. (32) using NifA fusion with MBP, showing the activity of NifA both in vivo and in vitro. However, the A. brasilense NifA fusion protein was unable to induce ß-galactosidase activity from the nifH promoter in E. coli extracts.

Immunological detection of A. brasilense NifA protein

A. brasilense NifA protein was purified from E. coli extracts as a GST-NifA fusion protein (plasmid pGEXAbA). NifA was released from GST by cleavage with thrombin (Figure 4, lanes 4 and 5). Approximately 50% of the NifA protein remained in the supernatant (Figure 4, lane 5) and was purified by standard procedures (Figure 4, lane 6). NifA polypeptide has an apparent molecular mass of 51 kDa (Figure 5), whereas the predicted NifA protein size, deduced from the nucleotide sequence, was 77 kDa (31). This discrepancy was also found by Arsène et al. (33) and they suggested that it could be due to the presence of a high proline residue content. Proteins rich in proline can cause anomalous migration in SDS-PAGE.


Figure 4 - Purification of NifA from E. coli BL21 (pGEXAbA). SDS-PAGE of denatured samples. Total E. coli extracts containing the GST-NifA fusion protein present in the supernatant (lane 2) and pellet (lane 3). Total E. coli extracts containing the A. brasilense NifA protein, after cleavage with thrombin, present in the supernatant (lane 4) and pellet (lane 5). Lane 6, 1 µg of purified NifA preparation and lane 1, molecular mass standards (Pharmacia). The position of NifA is indicated by the arrow.

[View larger version of this image (35 K GIF file)]


Figure 5 - Immunodetection of the A. brasilense NifA from E. coli cells using anti-A. brasilense NifA-antibodies. The E. coli strains contain the following plasmids: pGEXAbA (GST-NifA), lane 2; pKKAbA (ptac-NifA), lane 3; pAAbA (ptac-NifA), lane 4, and no plasmid, lane 5. Partially purified A. brasilense NifA protein (1 µg) was used as control (lane 1). The arrows indicate the position of the purified NifA or the GST-NifA fusion protein.

[View larger version of this image (38 K GIF file)]


The purified NifA protein was used to elicit antibodies. To detect the presence of NifA protein, Western blot analysis was performed using the anti-A. brasilense NifA antiserum. The antibodies reacted with the purified A. brasilense NifA protein (Figure 5, lane 1) and also with a related polypeptide band present in E. coli extracts harboring pKKAbA or pAAbA plasmids (ptac-nifA) (Figure 5, lanes 3 and 4). The fusion GST-NifA protein was also detected with antiserum (Figure 5, lane 2). The extra bands present in all extracts are cross-reactive nonspecific E. coli antibodies.

Immunodetection allowed us to verify the A. brasilense NifA constructs. We were able to demonstrate the expression of A. brasilense NifA from the tac promoter and also as a fusion GST-NifA protein (Figure 5). The antiserum is specific for A. brasilense NifA since it does not recognize NifA from K. pneumoniae (Figure 6, lanes 7-12). It is known that K. pneumoniae NifA protein shows an intrinsic tendency to aggregate and can only be purified as a fusion MBP-NifA protein (13,15,30,32). Although the NifA proteins from A. brasilense and K. pneumoniae have a high level of amino acid sequence identity (31), they differ in nifA expression patterns. We demonstrate here that A. brasilense NifA protein can be easily purified from aerobically grown E. coli cells, but is inactive. The present results also show that A. brasilense NifA can be synthesized in E. coli driven by the tac promoter and, although the protein is unable to activate the nifH promoter in E. coli cells, it is able to bind to the UAS present in this promoter (see Binding activity of the A. brasilense NifA protein).


Figure 6 - Immunodetection of the A. brasilense NifA from E. coli cells harboring the pGEXAbA plasmid. Lane 1, partially purified A. brasilense NifA protein; lanes 2-6, E. coli extracts carrying the pGEXAbA plasmid, and lanes 7-12, E. coli extracts carrying pGEXKpA plasmid (K. pneumoniae GST-NifA fusion protein).

[View larger version of this image (38 K GIF file)]


Activity of the A. brasilense NifA protein

The activity of the NifA protein was determined by its ability to restore nitrogen fixation to A. brasilense nifA-Tn5 mutant strains, under nitrogen fixation conditions. For this reason, we constructed the pLAFR3 derivative plasmids pLpucNifA and pLKSAbA (Figure 3). In this way, we were able to analyze both NifA constructs, the GST-NifA fused protein and the ptac-NifA expression system. Table 4 shows the nitrogenase activity of the wild type and mutant A. brasilense strains. The two constructs were able to restore a Nif+ phenotype to the nifA-Tn5 mutant strains. The reduced activity, as compared to the wild-type strain, could be due to the structure of the NifA protein found in these constructions containing the extra amino acids or even the GST protein. Transcriptional regulation should be ruled out because no nifA promoter was present.

Binding activity of the A. brasilense NifA protein

The A. brasilense nifH promoter presents two overlapping UASs and their role was examined by introducing base substitutions in both UAS sites, generating plasmids pKSUAS1 and pKSUAS2 (27). A 630-bp Sau3A DNA fragment from plasmids pKSwt, pKSUAS1 and pKSUAS2 (27) was used as target sequence to amplify by PCR a 200-bp fragment containing the two overlapping wild-type UASs, the UAS1 mutant and the UAS2 mutant, respectively. To assay the A. brasilense NifA binding capacity, total protein extracts isolated from E. coli MC1061 carrying plasmids p10, pAAbA or pGEXAbA were incubated with the DIG-labeled 200-bp DNA fragment (Figure 7A). Total protein extracts carrying K. pneumoniae NifA protein provided by plasmids pCK3, pAKpA or pGEXKpA were used as positive control. The binding reactions were carried out at 30°C for 20 min in order to maintain the NifA protein in an active form (13). A retarded band was visualized when extracts from MC1061/pCK3, MC1061/pAKpA, MC1061/pGEXKpA, MC1061/p10 and MC1061/pAAbA were incubated with the UAS-carrying fragments (Figure 7A, lanes 4-8). No shift was detected when an extract from MC1061 was used (Figure 7A, lane 2) or when an extract from MC1061/pCK3 was incubated with a 230-bp DNA fragment, with no UAS motif (Figure 7A, lane 3). Two retarded bands appeared when an extract from MC1061/pGEXAbA was used (Figure 7A, lane 9). The upper new extra band was visualized only with E. coli harboring this plasmid and was probably due to the high molecular weight GST-NifA fusion protein.


Figure 7 - Binding of NifA protein to the nifH UAS promoter region by gel-mobility shift assay. A, The 200-bp DIG labeled DNA fragment was incubated with crude cell supernatant from E. coli strains (20 µg total extract/reaction) containing the K. pneumoniae NifA protein synthesized by plasmids pCK3 (lane 4), pAKpA (lane 5) and pGEXKpA (lane 6) or A. brasilense NifA protein synthesized by plasmids p10 (lane 7), pAAbA (lane 8) and pGEXAbA (lane 9), alone or fused with GST protein. Lane 1, control of DNA fragment; lane 2, the control of the DNA fragment was incubated with MC1061 total extract, and lane 3, a 230-bp DNA fragment with no UAS motif, used as a probe, was incubated with MC1061/pCK3 total extract (negative controls). B, Lanes 1, 4 and 7 correspond to the added wild-type UAS region, UAS1 mutant and UAS2 mutant fragments, respectively. Lanes 2 and 3, wild-type fragment incubated with crude cell supernatant from E. coli strains (20 µg total extract/reaction): MC1061/pAAbA (lane 2) and MC1061/pGEXAbA (lane 3). Lanes 5 and 6, UAS1 mutant fragment incubated with total extracts from MC1061/pAAbA (lane 5) and from MC1061/pGEXAbA (lane 6). Lanes 8 and 9, UAS2 mutant fragment incubated with total extracts from MC1061/pAAbA (lane 8) and MC1061/pGEXAbA (lane 9). The full arrowheads represent the retarded bands and the open arrowhead represents the free-labeled fragment.

[View larger version of this image (51 K GIF file)]


When the 200-bp DNA fragments carrying the mutated UAS motifs were used as a probe a gel shift band was also visualized (Figure 7B). The extra band in gel shift assays was also detected when the UAS mutant DNA fragments were mixed with A. brasilense NifA fused with GST (MC1061/pGEXAbA). Previously, we have demonstrated that the two overlapping UAS sequences have differential effects on nifH promoter activity using K. pneumoniae NifA protein (27). However, the binding affinity of A. brasilense NifA protein, fused or not with GST, revealed no specificity when we compared wild-type or mutant UAS DNA fragments.

To further analyze the binding of NifA protein, mobility gel shift assays were carried out with NifA preparations (total extracts from MC1061/pAAbA and MC1061/pGEXAbA), wild-type UAS-carrying fragment, and incubated with anti-A. brasilense NifA-antibody (Figure 8). Increasing amounts of antibody led to the formation of a high molecular weight complex which was not visible in the control reaction (Figure 8, lane 2). Since this antibody was shown to be specific for A. brasilense NifA protein, the supershift effect detected when both extracts were used should be due to the formation of the DNA-NifA-antiNifA complex. A similar result was also reported by Zimmer et al. (34) using extracts enriched for HoxA protein in the presence of purified HoxA antibodies.


Figure 8 - Mobility shift assay of E. coli extracts carrying plasmids pAAbA (A) or pGEXAbA (B) in the presence of increasing amount of purified NifA antibody. Standard binding reactions were carried out with the total extracts previously incubated overnight with different antibody concentrations. Lane 1, Wild-type fragment only; lane 2, wild-type fragment and 20 µg of protein. Lanes 3 to 7, different amounts of antibody reacting with fragment/protein as in lane 2. Lane 3, 2 µl of anti-NifA; lane 4, 4 µl of anti-NifA; lane 5, 8 µl of anti-NifA; lane 6, 16 µl of anti-NifA and lane 7, 32 µl of anti-NifA. The supershift band is indicated by the full arrowhead.

[View larger version of this image (58 K GIF file)]


Up to now, only NifA from K. pneumoniae and A. vinelandii have been purified. The K. pneumoniae NifA protein has been purified as a fusion protein with MBP, providing a large N-terminal extension (13). The A. vinelandii NifA protein was purified as a soluble active protein and showed in vitro control functions predicted from in vivo functions (7). In these two bacteria the activity of NifA is modulated by the negative regulatory protein NifL in response to environmental oxygen and fixed nitrogen (6,7). The A. brasilense NifA protein belongs to a class of proteins that are not active in the presence of oxygen, based on the model of regulation proposed by Fischer et al. (8,9) that involves the cysteine residues found in the inter-linker domain. Another feature of A. brasilense NifA is that the N-terminal domain inhibits its own activity in the presence of ammonia (33). This was explained by the presence of two closely related PII proteins found in A. brasilense (18) that are involved in different regulatory steps of nitrogen metabolism. The PII protein, a product of the glnB gene, is required for modulating NifA activity in response to changes in cellular nitrogen levels (18,33).

Our work demonstrates that the NifA from A. brasilense can be expressed in E. coli extracts both as a fusion protein or as NifA itself but is still inactive in aerobically grown cells. This inactivation could be due to the lack of A. brasilense PII-like protein in E. coli cells. However, NifA activity was obtained in A. brasilense nifA mutants. Thus, even if a different mechanism is involved in the regulation of nitrogen fixation in the presence of ammonia, it is possible that one mechanism of control occurs at the level of NifA inactivation through its N-terminal domain, mediated by PII protein (18).

The modular structure of transcriptional activators suggests that inactive NifA would be able to bind to its target sequences (35). It has been shown, in the case of K. pneumoniae NifA, that DNA binding activity can be separated from the transcriptional activation capacity (32). The oxygen sensitivity of the A. brasilense NifA is attributed to the conserved motif of cysteine residues which is present in the interdomain linker between the central and carboxy-terminal domains (8,9). Upon oxygen inactivation, it can be hypothesized that the conformation of the central domain is altered in such a way that transcriptional activation capacity is lost, without affecting the DNA-binding capacity of the carboxy-terminal domain. In the case of NtrC-like transcriptional activators, where transcriptional activation capacity is regulated by phosphorylation, the inactive, unphosphorylated activator is still able to bind to its UAS (36). In this study we show that, although A. brasilense NifA protein is unable to activate nif promoters in aerobiosis due to its sensitivity to oxygen, the protein retains the ability to bind UAS sequences.


References

1. Buck M, Miller S, Drummond MH & Dixon RA (1986). Upstream activator sequences are present in the promoters of nitrogen fixation genes. Nature, 320: 374-378.        [ Links ]

2. Hoover TR, Santero E, Porter SC & Kustu S (1990). The integration host factor stimulates interaction of RNA polymerase with NIFA, the transcriptional activator for nitrogen fixation operons. Cell, 63: 11-22.        [ Links ]

3. Buchanan-Wollaston AV, Cannon MC, Beynon JL & Cannon FC (1981). Role of the nifA gene product in the regulation of nif expression in Klebsiella pneumoniae. Nature, 294: 776-778.        [ Links ]

4. Hill S, Kennedy C, Kavanagh EP, Goldberg R & Hanan R (1981). Nitrogen fixation gene (nifL) involved in oxygen regulation of nitrogenase synthesis in Klebsiella pneumoniae. Nature, 290: 424-426.        [ Links ]

5. Merrick M, Hill S, Hennecke H, Hahn M, Dixon R & Kennedy C (1982). Repressor properties of the nifL gene product in Klebsiella pneumoniae. Molecular and General Genetics, 185: 75-81.        [ Links ]

6. Lee H, Narberhaus F & Kustu S (1993). In vitro activity of NifL, a signal transduction protein for biological nitrogen fixation. Journal of Bacteriology, 175: 7683-7688.        [ Links ]

7. Austin S, Buck M, Cannon W, Eydmann T & Dixon RA (1994). Purification and in vitro activities of the native nitrogen fixation control proteins NifA and NifL. Journal of Bacteriology, 176: 3460-3465.        [ Links ]

8. Fischer HM, Bruderer T & Hennecke H (1988). Essential and non-essential domains in the Bradyrhizobium japonicum NifA protein: identification of indispensible cysteine residues potentially involved in redox reactivity and/or metal binding. Nucleic Acids Research, 16: 2207-2224.        [ Links ]

9. Fischer HM, Fritsche S, Herzog B & Hennecke H (1989). Critical spacing between two essential cysteine residues in the interdomain linker of the Bradyrhizobium japonicum NifA protein. FEBS Letters, 255: 167-171.        [ Links ]

10. Drummond MH, Whitty P & Wootton JC (1986). Sequence and domain relationships of ntrC and nifA from Klebsiella pneumoniae: homologies to other regulatory proteins. EMBO Journal, 5: 441-447.        [ Links ]

11. Drummond MH, Contreras A & Mitchenall LA (1990). The function of isolated domains and chimaeric proteins constructed from the transcriptional activators NifA and NtrC of Klebsiella pneumoniae. Molecular Microbiology, 4: 29-37.        [ Links ]

12. Morett E, Cannon W & Buck M (1988). The DNA-binding domain of the transcriptional activator protein NifA resides in its carboxy terminus, recognises the upstream activator sequences of nif promoters and can be separated from the positive control function of NifA. Nucleic Acids Research, 16: 11469-11488.        [ Links ]

13. Lee H, Berger DK & Kustu S (1993). Activity of purified NifA, a transcriptional activator of nitrogen fixation genes. Proceedings of the National Academy of Sciences, USA, 90: 2266-2270.        [ Links ]

14. Cannon W & Buck M (1992). Central domain of the positive control protein NifA and its role in transcriptional activation. Journal of Molecular Biology, 225: 271-286.        [ Links ]

15. Berger DK, Narberhaus F & Kustu S (1994). The isolated catalytic domain of NIFA, a bacterial-binding protein, activates transcription in vitro: Activation is inhibited by NIFL. Proceedings of the National Academy of Sciences, USA, 91: 103-107.        [ Links ]

16. Liang YY, de Zamaroczy M, Arsène F, Paquelin A & Elmerich C (1992). Regulation of nitrogen fixation genes in Azospirillum brasilense Sp7: Involvement of the nifA, glnB and glnA gene products. FEMS Microbiology Letters, 100: 113-120.        [ Links ]

17. de Zamaroczy M, Paquelin A & Elmerich C (1993). Functional organization of the glnB/glnA cluster of Azospirillum brasilense. Journal of Bacteriology, 175: 2507-2515.        [ Links ]

18. de Zamaroczy M, Paquelin A, Peltre G, Forchhammer K & Elmerich C (1996). Coexistence of two structurally similar but functionally different PII proteins in Azospirillum brasilense. Journal of Bacteriology, 178: 4143-4149.        [ Links ]

19. Magasanik B (1988). Reversible phosphorylation of an enhancer binding protein regulates the transcription of bacterial nitrogen fixation genes. Trends in Biochemical Sciences, 13: 475-479.        [ Links ]

20. Stock JB, Ninfa AJ & Stock AM (1989). Protein phosphorylation and regulation of adaptative responses in bacteria. Microbiological Reviews, 53: 450-490.        [ Links ]

21. Zhang Y, Burris RH, Ludden PW & Roberts GP (1994). Posttranslational regulation of nitrogenase activity in Azospirillum brasilense ntrBC mutants: Ammonium and anaerobic switch-off occurs through independent signal transduction pathways. Journal of Bacteriology, 176: 5780-5787.        [ Links ]

22. Sambrook J, Fritsch EF & Maniatis T (1989). Molecular Cloning: A Laboratory Manual. Cold Spring Harbor Laboratory Press, New York.        [ Links ]

23. Vanstockem M, Michiels K & Vanderleyden J (1987). Transposon mutagenesis of Azospirillum brasilense and Azospirillum lipoferum: physical analysis of Tn5 and Tn5-mob insertion mutants. Applied and Environmental Microbiology, 53: 410-415.        [ Links ]

24. Burris RH (1972). Nitrogen fixation - assay methods and techniques. Methods in Enzymology, 248: 415-431.        [ Links ]

25. Galimand M, Perroud B, Delorme F, Paquelin A, Vieille C, Bozouklian H & Elmerich C (1989). Identification of DNA regions homologous to nitrogen fixation genes nifE, nifUS and fixABC in Azospirillum brasilense Sp7. Journal of General Microbiology, 135: 1-13.        [ Links ]

26. Cannon FC, Riedel GE & Ausubel FM (1979). Overlapping sequences of Klebsiella pneumoniae nif DNA cloned and characterised. Molecular and General Genetics, 174: 59-66.        [ Links ]

27. Passaglia LMP, Schrank A & Schrank IS (1995). The two overlapping Azospirillum brasilense upstream activator sequences have differential effects on nifH promoter activity. Canadian Journal of Microbiology, 41: 849-854.        [ Links ]

28. Beynon JL, Williams MK & Cannon FC (1988). Expression and functional analysis of the Rhizobium meliloti nifA gene. EMBO Journal, 7: 7-14.        [ Links ]

29. Santero E, Hoover TR, Keener J & Kustu S (1989). In vitro activity of the nitrogen fixation regulatory protein NifA. Proceedings of the National Academy of Sciences, USA, 86: 7346-7350.        [ Links ]

30. Austin S, Henderson NC & Dixon RA (1990). Characterization of the Klebsiella pneumoniae nitrogen-fixation regulatory proteins NifA and NifL in vitro. European Journal of Biochemistry, 187: 353-360.        [ Links ]

31. Liang YY, Kaminski PA & Elmerich C (1991). Identification of a nifA-like regulatory gene of Azospirillum brasilense Sp7 expressed under conditions of nitrogen fixation and in the presence of air and ammonia. Molecular Microbiology, 5: 2735-2744.        [ Links ]

32. Berger DK, Narberhaus F, Lee H & Kustu S (1995). In vitro studies of the domains of the nitrogen fixation regulatory protein NIFA. Journal of Bacteriology, 177: 191-199.        [ Links ]

33. Arsène F, Kaminski PA & Elmerich C (1996). Modulation of NifA activity by PII in Azospirillum brasilense: evidence for a regulatory role of the NifA N-terminal domain. Journal of Bacteriology, 178: 4830-4838.        [ Links ]

34. Zimmer D, Schwartz E, Tran-Betcke A, Gewinner P & Friedrich B (1995). Temperature tolerance of hydrogenase expression in Alcaligenes eutrophus is conferred by a single amino acid exchange in the transcriptional activator HoxA. Journal of Bacteriology, 177: 2373-2380.        [ Links ]

35. North AK, Klose KE, Stedman KM & Kustu S (1993). Prokaryotic enhancer binding proteins reflect eukaryotic-like modularity: the puzzle of nitrogen regulatory protein C. Journal of Bacteriology, 175: 4267-4273.        [ Links ]

36. Van Soom C, de Wilde P & Vanderleyden J (1997). HoxA is a transcriptional regulator for expression of the hup structural genes in free-living Bradyrhizobium japonicum. Molecular Microbiology, 23: 967-977.        [ Links ]

37. Tarrand JJ & Krieg NR (1978). A taxonomic study of the Spirillum lipoferum group with descriptions of a new genus, Azospirillum gen. nov. and two species, Azospirillum lipoferum (Beijerinck) comb. nov. and Azospirillum brasilense sp. nov. Canadian Journal of Microbiology, 24: 967-980.        [ Links ]

38. Casadaban MJ, Chou J & Cohen SN (1980). In vitro gene fusions that join an enzymatically active beta-galactosidase segment to amino-terminal fragments of exogenous proteins: Escherichia coli plasmid vectors for the detection and cloning of translational initiation signals. Journal of Bacteriology, 143: 971-980.        [ Links ]

39. Studier FW & Moffatt BA (1986). Use of bacteriophage T7 RNA polymerase to direct selective high-level expression of cloned genes. Journal of Molecular Biology, 189: 113.        [ Links ]

40. Simon R, Priefer U & Pühler A (1983). A broad host range mobilization system for in vivo genetic engineering: transposon mutagenesis in Gram negative bacteria. Bio/Technology, 1: 784-790.        [ Links ]

41. Staskawicz B, Dahlbeck D, Keen N & Napoli C (1987). Molecular characterization of cloned avirulence genes from O and race 1 of Pseudomonas syringe p.v. glycinea. Journal of Bacteriology, 169: 5789-5794.        [ Links ]

42. Kennedy C & Drummond MH (1985). Use of cloned nif regulatory elements from Klebsiella pneumoniae to examine nif regulation in Azotobacter vinelandii. Journal of General Microbiology, 131: 1787-1795.        [ Links ]

43. Rose RE (1988). The nucleotide sequence of pACYC184. Nucleic Acids Research, 16: 355.        [ Links ]

Acknowledgments

The authors thank F. Pedrosa for the gift of plasmid p10 and H.B. Ferreira and S.S. Farias for the gift of plasmid pGEX and for help during the immunological experiments. We are indebted to J. Vanderleyden for providing the opportunity for L.M.P. Passaglia to develop part of the experiments in his laboratory and for helpful discussions.


Correspondence and Footnotes

Address for correspondence: I.S. Schrank, Centro de Biotecnologia, Bloco IV, Av. Bento Gonçalves, 9500, Caixa Postal 15005, 91501-970 Porto Alegre, RS, Brasil. Fax: +55-51-319-1079. E-mail: irene@dna.cbiot.ufrgs.br

Research supported by CNPq, FAPERGS and the Geconcerteerde Onderzoeksactie (to J. Vanderleyden). Received September 26, 1997. Accepted July 21, 1998.