INTRODUCTION

Top Abstract Introduction Materials and Methods Results and Discussion References

In diazotrophic proteobacteria, the ⁵⁴-dependent activator NifA activates transcription of the nif (nitrogen fixation) genes by a conserved mechanism common to members of the enhancer binding protein family (5). Although NifA proteins have similar domain structures, both transcriptional regulation of nifA expression and posttranslational regulation of NifA activity by oxygen and fixed nitrogen vary significantly from one organism to another.

Signal transduction in response to fixed nitrogen status has been best studied in Escherichia coli at both the genetic and biochemical levels (26, 27). Cells respond to changes in nitrogen availability by modifying the activity of a key sensory regulatory protein, PII, the product of glnB. Under conditions of nitrogen excess, when the internal concentration of glutamine is high, PII is primarily unmodified; in conditions of nitrogen limitation, PII is mainly uridylylated. The enzyme responsible for this covalent modification is an uridylyltransferase/uridylyl-removing enzyme (UTase/UR) encoded by the glnD gene. The degree of uridylylation of PII has major physiological implications with respect to both the level and activity of glutamine synthetase (GS), the product of glnA. Under nitrogen-excess conditions, native PII protein prevents transcription from the ntr promoters by stimulating the histidine protein kinase NtrB to dephosphorylate its cognate response regulator, NtrC (16, 17), leading to decreased expression of glnA. In addition, PII stimulates the enzyme adenylyltransferase to adenylylate GS, which decreases its activity (15, 18). Under nitrogen limitation, uridylylation of PII prevents its interaction with NtrB, and thus NtrC is maintained primarily in its phosphorylated form. In addition, PII-UMP stimulates the deadenylation activity of adenyltransferase, which catalyzes the conversion of the inactive GS-AMP to active GS.

E. coli contains a second PII-like protein, which is encoded by the glnK gene (36, 37). This PII paralogue is also regulated by the UTase/UR in response to nitrogen availability and also plays a role in nitrogen regulation (2, 3). While expression of the glnB gene is constitutive with respect to the intracellular nitrogen status, glnK is encoded in the NtrC-dependent operon glnK-amtB, in which amtB encodes a high-affinity ammonium transporter (33). Thus, expression of glnK is subject to nitrogen regulation by the NtrB-NtrC two-component regulatory system, and essentially little GlnK is expressed under conditions of nitrogen excess.

Many proteobacteria express more than one homologue of the PII protein, and current evidence suggests that at least one of the PII-like proteins is involved in nitrogen control of NifA expression and/or activity in several diazotrophs. In Klebsiella pneumoniae and Azotobacter vinelandii, which are members of the gamma subdivision of Proteobacteria, nifA is coordinately transcribed with a second gene, nifL, whose product inhibits NifA activity in response to oxygen and fixed nitrogen (9, 10). Transcription of the K. pneumoniae nifLA operon is activated by the phosphorylated form of NtrC and is therefore influenced by nitrogen availability, whereas in A. vinelandii nifLA expression is constitutive. The identification of two PII-like proteins in enteric bacteria has facilitated analysis of nitrogen control of nitrogen fixation in K. pneumoniae. Recent evidence suggests that the alternative PII-like protein GlnK is required to relieve inhibition of NifA activity by K. pneumoniae NifL under nitrogen-limiting conditions (12, 14). This conclusion is based on the observation that NifL inhibits NifA activity irrespective of the nitrogen status in enteric glnK mutants. Furthermore, the uridylylation state of GlnK is apparently irrelevant for relief of inhibition by NifL in K. pneumoniae (11) and in E. coli (12). Although GlnK clearly has a role in maintaining NifL in an inactive state under derepressing conditions for nitrogen fixation, it is not obvious how the K. pneumoniae NifL-NifA system responds rapidly to changes in nitrogen status.

Although K. pneumoniae and A. vinelandii NifL have similar functions, there is some evidence that A. vinelandii NifL may use a different mechanism to respond to the cellular nitrogen status. Previous genetic studies in A. vinelandii revealed that a mutation in nfrX (now called glnD), which encodes a homologue of E. coli UTase/UR, gave rise to a Nif phenotype which could be suppressed by a secondary mutation in nifL (8). These results suggested that in contrast to K. pneumoniae, uridylylation of a regulatory component may be required to prevent inhibition by NifL. Current evidence indicates that A. vinelandii contains only a single PII-like protein encoded by a gene designated glnK, and in contrast to E. coli and K. pneumoniae, the glnK-amtB operon is expressed under all conditions regardless of fixed nitrogen supply (25). The PII-like protein encoded by A. vinelandii glnK (Av PII) could therefore be a good candidate for a signal transduction component which could interface between GlnD and NifL, thus controlling the activity of NifL in response to the uridylylation state of PII. However, determination of the role of A. vinelandii glnK gene product in nitrogen signaling has been thwarted by the essential nature of this gene since it has not been possible to isolate null mutations in glnK (25).

Since it has not been possible to study in vivo regulation of nitrogen fixation in A. vinelandii in the complete absence of the PII-like protein, we have resorted to a heterologous system to analyzse the response of the A. vinelandii NifL-NifA system in glnB and glnK mutants. We have shown previously that, as is the case with the K. pneumoniae NifL-NifA system, A. vinelandii NifL modulates NifA activity in E. coli in response to oxygen and fixed nitrogen (32). We have used this heterologous system to study the role of the PII-like regulatory proteins in nitrogen regulation of A. vinelandii NifL activity. We anticipated that both A. vinelandii and K. pneumoniae NifL would respond to a common signal transduction pathway for sensing the nitrogen status in E. coli. Surprisingly, we found that in contrast to K. pneumoniae NifL, neither PII nor GlnK is required to relieve inhibition by A. vinelandii NifL under nitrogen-limiting conditions. Moreover, in mutants lacking both the PII and NtrC proteins, NifL activity is no longer responsive to the nitrogen status. These results reveal striking differences in the mechanism of nitrogen regulation of NifL activity between A. vinelandii and K. pneumoniae.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results and Discussion References

E. coli mutant strain constructions. All strains used were derivatives of E. coli ET8000 (Table 1). To construct the glnB1 mutation, the 0.31-kb AgeI-EcoNI fragment from plasmid pAH5 which carries the E. coli glnB region was deleted, and the 5' extensions from the digested plasmid were trimmed by incubation with mung bean nuclease. The blunt-end extensions were ligated, producing plasmid pglnB19. A SalI-BglII fragment carrying the deleted glnB gene from this plasmid was cloned into the gene replacement vector pKO3 digested with SalI and BamHI, generating plasmid pglnBO3. E. coli ET8000 was transformed with pglnBO3, and homologous recombination between the cloned fragment and bacterial chromosome was carried out as described previously (19), resulting in strain PT8000. The absence of wild-type sequences in the recombinant was confirmed by PCR analysis using primers 5' TGAAACGCCTGATGAC 3' and 5' CTATTCCCGATGCCGTTG 3', which flank the glnB region.

-Galactosidase assays and growth conditions. To assay -galactosidase activity, E. coli strains were transformed with plasmid pRT22, which carries a nifH-lacZ translational fusion. NifA activity was measured by determining the level of expression from the nifH promoter. To monitor the ability of wild-type NifL or the truncated NifL_(147-519) protein (which lacks the redox response) to inhibit NifA activity, E. coli strains were transformed with plasmids pRT22 and pPR34 or pRT22 and pPR54, respectively. The activity of NifA alone was assayed by transforming E. coli strains with plasmids pRT22 and pPR39 (Table 1) (32).

1

1

1

1

Determination of intracellular 2-oxoglutarate. Extracts for measurements of 2-oxoglutarate were prepared as described previously (21), with minor amendments. Portions of cultures (8 ml) were filtered through 0.45-µm-pore-size membrane filters (25-mm diameter) with vacuum suction. As soon as the liquid was removed (~30 s), the filter was transferred to an Eppendorf centrifuge tube containing 1 ml of 0.3 M HClO₄ plus 1 mM EDTA at 0°C. Once the tube contents were mixed thoroughly, the filter was removed from the Eppendorf tube and the content was centrifuged to remove debris. Then 500 µl of the extract was removed and neutralized by the addition of 75 µl of 2 M K₂CO₃. The resulting KClO₄ precipitate was removed by centrifugation, and the supernatant fluid was stored at 80°C for further analysis.

Western blotting. Western blotting was performed as described previously (30), using polyclonal antisera against A. vinelandii NifL and NifA raised in rabbits and the Amersham enhanced chemiluminescense system for detection.

	RESULTS AND DISCUSSION

Top Abstract Introduction Materials and Methods Results and Discussion References

Activity of A. vinelandii NifL and NifA in E. coli. We have shown previously that when the A. vinelandii nifL and nifA genes are expressed from a constitutive promoter in E. coli, transcriptional activation of a nifHp-lacZ reporter is responsive to fixed nitrogen and oxygen in vivo. These results indicate that the Azotobacter NifL and NifA proteins are competent to interact with signal transduction components in E. coli and that this heterologous system can be used to analyze nitrogen regulation of NifA activity by NifL. Since some of the E. coli mutant strains studied below grew poorly on minimal medium unless supplemented with glutamine, we checked that the addition of glutamine in our standard culture conditions did not alter the level of NifA activity in the presence of NifL (Table 2). As observed previously, we found that when both nifL and nifA were expressed, NifA was active only under anaerobic and nitrogen-limiting conditions (32). We also confirmed that removal of the first 146 residues of NifL, which includes the redox-sensing PAS domain, rendered the protein unable to inhibit NifA under oxic conditions unless ammonia was present in the medium (Table 2). We use this truncated protein throughout the course of our experiments as a control on the specificity of the nitrogen response. It is notable that when the coding sequence of nifL is almost entirely deleted (producing a putative truncated peptide containing 65 C-terminal residues of NifL), NifA activity is 10-fold higher under anaerobic and nitrogen-limiting conditions than when wild-type NifL is present (Table 2). Similar high levels of NifA activity were observed with some mutant NifA proteins which escape inhibition in the presence of wild-type NifL (data not shown). This suggests that NifL retains some inhibitory function even when conditions are derepressing for nitrogen fixation. Thus, some factor(s) required to maintain A. vinelandii NifL in its inactive form may be limiting under our growth conditions in E. coli.

PII-like proteins are not required to prevent A. vinelandii NifL from inhibiting NifA under derepressing conditions for nitrogen fixation. Since the A. vinelandii NifL-NifA regulatory system is apparently responsive to the nitrogen status in E. coli, it was of interest to determine which nitrogen regulatory genes are required for the nitrogen response. Inhibition of NifA activity by NifL was assessed in various E. coli mutant strains defective in nitrogen regulation by determining the level of expression from a nifH-lacZ translational fusion carrying the reporter plasmid pRT22 (Table 3). Single null mutations in glnB and glnK had little influence on the activity of NifA in the absence of NifL. That is, the levels of NifA activity in the absence of NifL were similar under nitrogen-deficient and nitrogen-excess conditions, as previously observed with the wild-type strain. The glnB null mutation also had little apparent influence on the activity of either NifL or NifL_(147-519) under nitrogen-limiting anaerobic conditions. However, there was almost a threefold decrease in the inhibitory activity of NifL compared with the wild-type strain under nitrogen-excess conditions, resulting in a decrease in the nitrogen repression ratio (Table 3, PT8000). A similar pattern of activity was seen in the strain containing only the glnK mutation. We also observed an increased level of -galactosidase activity under nitrogen limitation when this strain carried exclusively the reporter plasmid pRT22 (Table 3, GT1002). This could be largely due to a high intracellular level of phosphorylated NtrC under these conditions, as suggested by the observation that this strain displayed an apparent increased growth rate in liquid glucose-arginine medium, a demanding test for NtrC activity (3). This can be explained in the context that GlnK is required for fine control of the level of phosphorylated NtrC as previously suggested (3). The observation that the activity of neither NifL nor NifL_(147-519) was significantly influenced by the glnK mutation is remarkably different from what is observed with the K. pneumoniae NifL-NifA system. In this case no activation of the nifH promoter is observed in glnK mutant strains when both NifL and NifA are present, indicating that glnK is absolutely required to prevent K. pneumoniae NifL from inhibiting NifA activity (12, 14).

Nitrogen regulation of NifL activity is absent in strains carrying both ntrC and glnB mutations. Since the data obtained with the glnB glnK mutant could be influenced by its slow growth and the presence of high levels of phosphorylated NtrC, we examined the effect of introducing a ntrC::Tn5 mutation into the double-mutant strain. As expected, this mutation rendered cells auxotrophic for glutamine, presumably as a consequence of the influence of these mutations on the expression and activity of GS. Surprisingly, under conditions of nitrogen excess, NifL inhibition was substantially relieved in the triple-mutant strain in the case of both native NifL and the truncated variant, and in both cases the repression ratio was close to 1 (Table 3, MT8000). Moreover, in this background, greater relief of NifL inhibition was seen under nitrogen-limiting conditions compared with the wild-type strain, regardless of whether native NifL or NifL_(147-519) was present (Table 3, compare ET8000 and MT8000). These data clearly demonstrate that neither of the PII paralogues are required to relieve inhibition by A. vinelandii NifL under N-limiting conditions. Since phosphorylated NtrC is required for activation of GlnK expression in E. coli, we suspected that the glnK mutation did not contribute to the phenotype of the glnB glnK ntrC strain. Accordingly, we constructed a glnB ntrC double-mutant strain. The nitrogen response of NifL in this strain was similar to that observed with the triple mutant, although the levels of NifA activity observed in the presence of NifL or NifL_(147-519) under nitrogen-limiting conditions were lower than those observed with the glnB glnK ntrC strain (Table 3, RT8000). Western blotting analysis suggested that there were no major differences in the levels of NifL and NifA expression in the mutant strains compared with the wild-type, and the ratios of NifL to NifA were similar in all cases (Fig. 1). Hence, the failure of NifL to inhibit NifA activity in the mutant strains under nitrogen-excess conditions is unlikely to be a consequence of decreases in the level of NifL protein or to the presence of excess NifA in these strains.

View larger version (40K): [in this window] [in a new window]   FIG. 1.   Immunodetection of NifA and NifL in E. coli wild-type and mutants strains grown under nitrogen-excess conditions. All strains carried the reporter plasmid pRT22 in addition to pPR34, which encodes wild-type NifL and NifA, and were grown in NFDM medium containing (NH4)2SO4 and glutamine as nitrogen sources. Samples were subjected to Western blotting as described in Materials and Methods with polyclonal antiserum against either NifA (A) or NifL (B). Lanes: 1, wild type (Wt; ET8000); lane 2, glnB glnK (FT8000); lane 3, glnB ntrC (RT8000); lane 4, glnB glnK ntrC (MT8000); lane 5, either purified NifA (A) or purified NifL (B) as control.

Relationship between in vivo 2-oxoglutarate pools and NifL activity. The above observation that the combination of glnB and ntrC mutations eliminates the nitrogen response of A. vinelandii NifL in E. coli suggests that more than one factor is involved in this response. In addition to the potential role of the PII paralogues, the additional factor could be a metabolite whose levels are influenced by the ntrC mutation or a gene product which is expressed under the control of NtrC. It is interesting to consider the former possibility in the light of our recent in vitro data, which demonstrates that the NifL-NifA system is directly responsive to 2-oxoglutarate within the physiological range (20). Under conditions of nitrogen excess, the reported 2-oxoglurate concentration in E. coli is ~100 µM, which increases to ~1 mM under conditions of nitrogen limitation (31). We therefore considered the possibility that 2-oxoglutarate could be an effector required for relief of inhibition by NifL in vivo. To investigate this hypothesis, we measured the intracellular accumulation of 2-oxoglutarate in E. coli mutant strains grown under conditions identical to those used to determine transcriptional activation by NifA using the reporter system. The 2-oxoglutarate concentration was below the level of detection (>50 µM) in the wild-type strain grown under conditions of nitrogen excess (glucose, ammonia, and glutamine) but increased by a factor of at least 30-fold to ~10 nmol/mg of protein under nitrogen-limiting conditions (Table 5, ET8000). This value corresponds to ~3 nmol/mg (dry weight), or an intracellular concentration of ~1.5 mM. Similar results were observed with the glnB mutant (Table 5, PT8000). Surprisingly, the level of 2-oxoglutarate increased substantially in the glnB glnK double mutant under nitrogen-excess conditions. This may result from overadenylylation of GS in this strain (3, 6) and consequent accumulation of 2-oxoglutarate in the absence of efficient nitrogen assimilation. The ntrC mutation did not appear to influence 2-oxoglutarate accumulation under conditions of nitrogen excess, but a small increase was observed under nitrogen-limiting conditions. This might be expected if nitrogen is poorly assimilated since the level of GS is significantly decreased in ntrC mutants under conditions of nitrogen limitation (28). In contrast, the combination of the ntrC with the glnB or glnB and glnK mutations gave rise to a high level of 2-oxoglutarate accumulation under conditions of nitrogen excess, resulting in levels similar to that observed with the wild-type strain under nitrogen-limiting conditions (Table 5, MT8000 and RT8000 compared with ET8000). We interpret this accumulation as a consequence of physiological nitrogen limitation in these strains even when excess external ammonium is present. Notably, the growth rates of these strains were similar in the presence and absence of ammonia. Overall, these results suggest that there is some correlation between the intracellular accumulation of 2-oxoglutarate and the nitrogen repression ratio observed with the nifH-lacZ reporter (Table 3), indicating that 2-oxoglutarate may have a role in modulating the activity of the A. vinelandii NifL-NifA system in vivo.

Conclusions. We have utilized defined E. coli mutant strains to analyze nitrogen regulation of A. vinelandii NifA activity mediated by NifL. This heterologous system has been used previously to examine the role of PII-like proteins in modulating the activity of K. pneumoniae NifL in vivo (1, 12) and hence allows direct comparison of the properties of the A. vinelandii and K. pneumoniae nitrogen fixation regulatory genes.