Results and Discussion

GlnB and GlnK are involved in the regulation of NifAl synthesis and activity. The influence of glnB and glnK single and double mutations on the nif-encoded nitrogenase system was analyzed at different levels of regulation. First, we examined the accumulation of the transcriptional activator NifAl in R. capsulatus wild-type and mutant strains using an antiserum specific for this protein (results not shown). In the wild-type strain, NifAl accumulated only in N-limited cells, which is consistent with NtrC-mediated transcriptional control of the nifAl gene. Inactivation of glnB resulted in the partial escape from NH3 control of the regulation of nifAl expression, and the NifAl protein accumulated in both N-limited, and to a lesser extent in ammonium-grown cells. Similar results were found for a glnK mutant that was constructed to allow the constitutive expression of amtB. In a glnB-glnK double mutant, synthesis of NifAl was greatly enhanced both under N-limiting and N-sufficient conditions.

In order to ascertain the effects of mutations in glnB or glnK on the post-translational control of NifA activity by NH3, a strain that constitutively expresses nifAl was constructed. NifAl activity was indirectly assayed by measuring levels of Fe-protein using Western blotting with Fe-protein specific antiserum (results not shown). As described elsewhere, overexpression of NifAl in the wild type did not lead to the significant synthesis of nitrogenase under nitrogen-replete conditions (Paschen et al, submitted). These data further corroborate the hypothesis that NifAl activity is posttranslationally controlled by the nitrogen status of the cells. Ammonium suppression of NifAl activity is mainly mediated by GlnK as demonstrated by Fe-protein synthesis in the glnK mutant strain. The level of Fe-protein, but not of NifAl, varied drastically in strains lacking glnK depending on slight variations in culturing conditions including culture volume, growth phase, temperature and light intensity, indicating that GlnK might act in the "fine-tuning" of NifAl activity in response to subtle changes in the environmental conditions. While a single glnB mutation had no effect on the regulation of NifAl activity, extremely high amounts of Fe-protein accumulated in the glnB-glnK double mutant independently of whether or not fixed nitrogen was available. Thus GlnB as well as GlnK seems to be involved in the ammonium-dependent control of NifA activity.

These results were confirmed and amplified by examining the effects of these mutations on the concentrations of NH3 needed to repress the synthesis of Fe-protein in the various strains. Batch cultures were grown on various initial concentrations of NH3 and the quantities of Fe-protein produced were visualized by Western analysis. In the wild-type strain (Figure 1A), nitrogenase synthesis was repressed by 18 mM NH3. The glnB strain was less sensitive to fixed nitrogen and 22 mM NH3 was required to repress nitrogenase synthesis (Figure IB). However, nitrogenase synthesis was completely derepressed in the glnK and glnB-glnK mutants where nitrogenase synthesis was observed at the highest levels of NH3 examined (Figure 1C, D).

A: wild-type

Figure 1. Batch cultures were grown on the indicated concentrations of NIT, (mM).


KNA|ralv. Iiounr ff


3.1. The role of GlnB and GlnK in regulating nitrogenase activity. In addition to the effects observed on the ammonium control of nitrogenase synthesis, the immunoblot shown in Figure 1 also demonstrates that there are effects of mutations in glnB and glnK on nitrogenase modification by ADP-ribosylation. This long-term ADP-ribosylation previously observed with wild-type cultures under a variety of growth conditions is responsive to the cellular nitrogen status (Yakunin et al. 1999). In the wild-type cultures, modification is not seen until the initial NH3 concentration used for growth is raised to 16 mM (Figure 1A). In the glnB strain, even higher initial NH3 concentrations (19 mM) are required before modification is observed (Figure IB). This suggests that GlnB normally communicates, possibly through direct interactions, the nitrogen status to DraT, the enzyme responsible for adding the ADP-ribose to the Fe-protein, and that higher levels of fixed nitrogen are needed to overcome its absence. Contrary to what is observed with the glnB strain, in the glnK strain ADP-ribosylation is seen at much lower initial NH3 concentrations, 10 mM and higher (Figure 1C). This suggests that GlnK communicates primarily with DraG, the enzyme responsible for removing the modifying group.

Thus the accompanying model can be proposed for the various GlnB, GlnK functions. At one level (I), GlnB and GlnK interact with NtrB to control the synthesis of the NifA proteins with the respect to the nitrogen status. Both proteins appear to be necessary for the complete repression of the nifAs in nitrogen replete media. At a second level (II), it is primarily GlnK that controls the activity of NifA, probably by directly interacting with it. At a third level (III), GlnB and GlnK communicate with the enzymes that carry out the Fe-protein modification/demodification cycle, DraT and DraG, respectively. It is likely that many of these relationships represent direct protein-protein interactions and this is being tested using a yeast two-hybrid system. So far protein-protein interactions between GlnB and NtrB, NifAl, NifA2, GlnK, and DraT as well as between GlnK and NifAl and NifA2 have been detected (Masepohl et al. 2001).

Figure 2. Model for GlnB, GlnK actions.

3.2. The role of AmtB and AmtY in regulating nitrogenase activity. What factors might control the activities of GlnK and GlnB in mediating the nitrogenase responses to NH3 addition? Previously we had suggested that the high-affmity ammonium transport system, which could be encoded by amtB, might be involved since we found that its activity varied under the same culture conditions that caused a variation in the nitrogenase responses (Yakunin, Hallenbeck 1999).

The AmtBs (ammonium transport B, TC no 2.A.49.3.2) are ubiquitous membrane proteins found in all three domains of life. They are thought to either transport ammonium or to facilitate the equilibration of ammonia across the cytoplasmic membrane. It has been proposed that the yeast high affinity ammonium permease Mep2p, an AmtB homolog, functions as an ammonium sensor, generating a signal which activates signal transduction cascades that regulate filamentous growth in response to ammonium starvation (Lorenz, Heitman 1998).

R. capsulatus contains two AmtB homologs: AmtB, which is organized in a glnKamtB operon as in many other bacteria (Thomas et al. 2000), and AmtY, which is monocistronic and found a short distance, 2.5kb, from ntrBC. The predicted AmtB protein is highly similar to other putative bacterial and archaeal AmtBs. However, the R. capsulatus AmtB and AmtY proteins are only distantly related. Both glnKamtB and amtY are Ntr regulated (Gross et al. 2001). We created knockout insertions in these genes and studied their effects. Strains carrying amt insertions showed wild-type growth on all media tested, including RCV-NH3. When their ammonium uptake activities were measured, nitrogen-limited cultures of both the amtB and amtY mutants as well as the amtB-amtY double mutant demonstrated very nearly wild-type rates of ammonium uptake, which is consistent with the observed growth characteristics and shows that AmtB and AmtY are not required for this activity in R. capsulatus. We then measured methylammonium uptake, which is diagnostic for the activity of the high-affinity ammonium transport Figure 3. Nitrogenase modification m glnB, glnK and system. The amtB, but not the amtY strain, glnB-glnK mutant strains, was completely deficient in this activity.

The ability of amt strains to modulate nitrogenase activity and to ADP-ribosylate the Fe-protein of nitrogenase was examined. Both the wild type and amtY strains demonstrated a classical in vivo nitrogenase switch-off response to the addition of ammonium (Figure 3). Almost complete inhibition of acetylene reduction was evident within 5 min of ammonium addition, and full recovery of the initial rate of nitrogenase activity was obtained within a relatively short period of time. In both strains the addition of ammonium also induced fast, short-term, ADP-ribosylation of Fe-protein (Figure 3). The amtY mutant showed higher levels of Fe-protein ADP-ribosylation than the wild-type strain in response to the addition of the same amount (200 |xM) of ammonium. Moreover, in this mutant ADP-ribosylation response was faster than in wild-type strain, and the modified Fe-protein can be detected within 5 min after ammonium addition, while it takes a longer time for the wild-type strain (Figure 3). In striking contrast, the addition of the same amount of ammonium, or higher (50 mM), had no effect on the in vivo nitrogenase activity and Fe-protein modification state

in both amtB and amtB-amt Y double mutants (Figure 3). To directly demonstrate that this effect was due to the introduced mutation, we studied the effects of complementation with a plasmid-borne wild-type copy of amtB. Analysis of the effects of ammonium addition on nitrogenase activity and Fe-protein modification showed that the wild-type response had been restored in the plasmid-carrying strains (results not shown). This shows that the defect is caused by the mutation in amtB.

We have shown that the DraT/DraG system is functionally intact by examining cultures that had gone through a period of light deprivation followed by reillumination. Previous studies have shown that this treatment results in a rapid ADP-ribosylation of the Fe-protein upon removal of the light followed by fast demodification upon light shift-up. All the mutant strains gave a wild-type response in this analysis. Thus the modification system is intact in the amtB mutants but is apparently not receiving the proper signal upon the addition of NH3. Several lines of evidence indicate that NH3 is entering the cell under these conditions and that therefore the lack of response is not merely due to a defect in transport. First, gross ammonium uptake activity was unaffected. Second, we examined the regulation of glutamine synthetase. The activity of glutamine synthetase is subject to two forms of fast short-term inhibition that respond to ammonium additions, covalent modification (adenylylation) and feedback inhibition. The activity of glutamine synthetase in amt mutants showed the same fold decrease as the wild type upon ammonium addition. Thus, AmtB appears to specifically signal the presence of NH3 in the external medium to the systems that control nitrogenase activity.

In conclusion, we have shown that: (1) AmtB is necessary for sensing external ammonium and initiating the switch-off and ADP-ribosylation responses; (2) GlnB and GlnK both have a role in the control of NifA synthesis; (3) it is primarily GlnK that controls NifA activity; and (4) GlnB and GlnK are involved in modulating the switch-off and ADP-ribosylation responses.

Was this article helpful?

0 0

Post a comment