AmtB Is Necessary for NH₄⁺-Induced Nitrogenase Switch-Off and ADP-Ribosylation in Rhodobacter capsulatus

Bacterial strains and growth conditions. E. coli strains S17-1 (39) and S17-1/λpir (2) were used for conjugal transfer of plasmids to R. capsulatus by filter mating (27). E. coli strains were grown in Luria-Bertani (LB) medium (35) supplemented with the appropriate antibiotics. R. capsulatus strains, including the wild-type SB1003 (52), were routinely maintained in liquid YPS (yeast extract/peptone/salts) medium (46) supplemented with the appropriate antibiotics. For all experiments shown, cultures were inoculated and grown photosynthetically in modified (51) liquid RCV medium (46). Highly nitrogen limited cultures, obtained by growing cultures to early stationary phase on RCV medium without an added nitrogen source, were used for all experiments on fast effects on nitrogenase because these conditions have previously been shown to be optimal for the switch-off response (50). Under these conditions, growth is presumably supported by the dissolved N₂. For examining effects on long-term ADP-ribosylation, cultures were grown to early stationary phase on RCV medium with the indicated initial concentrations of ammonia as previously described (51)

Construction of an amtB::Km^r gene disruption.The R. capsulatus amtB gene was cloned from genomic DNA by PCR with oligonucleotide primers amtB-upstream (5′CCGTTCAAGCTCGAGGAGGTCCG-3′), which contains an XhoI site, and amtB-downstream (5′-TCAGCGGGAATGATGAGCGGCGC-3′), located downstream of a ClaI site). The PCR product (1,944 bp) was cloned as an XhoI/ClaI fragment into pBluescript KS− (Stratagene), and a kanamycin resistance cassette (Km^r;1.2-kb SphI fragment) from pBSL86 (1) was inserted into an NspI site in the coding region of amtB. Subsequently, an amtB::Km^r fragment was subcloned into the EcoRI site of the suicide vector pSUP202 (39), and the resulting plasmid was mobilized from E. coli S17-1 into R. capsulatus SB1003. Replacement of the wild-type amtB gene was confirmed by Southern hybridization.

Construction of an amtY::Gm^r gene disruption.The R. capsulatus amtY gene was subcloned as a 1.9-kb BamHI/EcoRI fragment from pRCN102 (3) into pBluescript KS(−), and a gentamicin resistance cassette (Gm^r; 0.9-kb BamHI fragment) from pUCGM (37) was inserted into the BclI site of amtY. amtY::Gm^r was subcloned as a BamHI/EcoRI fragment into the mobilizable vector pK18mobsacB (36), and the resulting plasmid was mobilized from E. coli S17-1 into R. capsulatus SB1003. Double recombinants were confirmed by Southern hybridization.

Construction of a double amtB amtY mutant.To create an R. capsulatus amtB amtY double mutant, the 0.7-kb BclI/HindIII fragment of the amtY gene was subcloned into the broad-host-range mobilizable suicide vector pKNOCK-gentamicin (2) which had been digested with BamHI/HindIII, and the resulting plasmid was mobilized from E. coli S17-1/λpir into the R. capsulatus amtB::Km^r strain. Recombinants were selected for kanamycin and gentamicin resistance. As a result of homologous recombination (confirmed by Southern hybridization), two mutant copies of the amtY gene, separated by the inserted pKNOCK, were formed, with one lacking the 5′ region (456 bp) and the other the 3′ region (233 bp).

Complementation studies.The entire glnK-amtB operon as well as 390 bp of upstream sequence (2.4 kb total) was cloned by PCR and inserted into the polylinker of the broad-host-range vector pJB3Tc20 (4). The resulting plasmid, pAY99, was mobilized from E. coli S17-1 into R. capsulatus amtB or amtB amtY mutant strains.

Determination of in vivo nitrogenase activity and Fe-protein modification state.In vivo nitrogenase activity was measured by the acetylene reduction method (13, 14) with 5-ml culture samples. Stimuli were applied as follows: light was removed by wrapping the vials in aluminum foil, and ammonium was added as an anaerobic solution. Solutions were made anaerobic by sparging with argon which had been passed through a heated copper catalyst to remove any traces of oxygen. At the indicated times, 50-μl aliquots of the gas phase and the culture were separately withdrawn from the vials for the analysis of ethylene by gas chromatography and Fe-protein modification state.

To monitor the modification state of Fe-protein, the entire culture samples without prior treatment were brought to a 1× sodium dodecyl sulfate sample buffer concentration (23) and immediately incubated in a boiling-water bath for 5 min. Equal amounts of total protein (2 μg/well) were analyzed by Laemmli sodium dodecyl sulfate-polyacrylamide gel electrophoresis (12.5% total acrylamide) with low-crosslinker gels (20) and subsequent immunoblotting with chemiluminescence detection as described previously (49). In this system, R. capsulatus Fe-protein shows an apparent molecular mass of 34.9 kDa for the unmodified subunit and 38.7 kDa for the ADP-ribosylated subunit. (On high-resolution gels, the unmodified Fe-protein usually produces a double band due to the transient formation of incompletely denatured intermediates.) Since only one of the two Fe-protein subunits of the Fe-protein dimer is modified, 100% modification of Fe-protein dimers would correspond to two bands of equal intensity.

Ammonium and methylammonium uptake.For measurement of ammonium uptake, early-stationary-phase cells grown photosynthetically on RCV medium without an added nitrogen source were incubated for 5 min at room temperature under light and anaerobic conditions. The reaction was initiated by the addition of ammonium (0.2 mM), and ammonium concentrations in culture supernatants from samples taken at 2.5-min intervals were measured with the phenol-hypochlorite method (40). [¹⁴C]methylammonium uptake was measured as previously described (34).

Ammonium excretion and intracellular ammonium concentrations.To measure ammonium excretion, highly nitrogen limited cultures in early stationary phase were sparged with N₂, and the appearance of ammonium in culture supernatants over time was measured with the phenol-hypochlorite method (40). Intracellular ammonium concentrations were also determined by the phenol-hypochlorite method after ice-cold methanol extraction of cell pellets. Intracellular concentrations were calculated based on an intracellular volume of 4 μl/mg of protein (7).

Glutamine synthetase assays.Glutamine synthetase activity (γ-glutamyltransferase) of whole cells and the relative state of glutamine synthetase adenylylation were measured as previously described (19). Protein concentrations were determined by the Bradford method (5) after 2 min of sonication.

R. capsulatus contains two amtB homologues.A BLAST search of the available genome sequence of R. capsulatus (http://wit.mcs.anl.gov/WIT/CGI/ ) revealed the presence of two genes potentially coding for ammonium transport proteins, which we call amtB and amtY. amtB is predicted to be part of a glnK-amtB operon, as in many other bacteria (43, 44), whereas the determined DNA sequence suggests that amtY is a monocistronic gene found a short distance (2.5 kb) from ntrBC and nifR3. The predicted AmtB protein (453 amino acids, 46.8 kDa) is highly similar to that of other bacterial and archaeal putative AmtBs, ranging from 59% identity to AmtB from Rhizobium etli to 30% for that from Thermotoga maritima. The predicted R. capsulatus AmtY protein (392 amino acids, 41.2 kDa) is less similar to putative AmtBs, with the highest similarity, 39% identity, to AmtB from Azospirillum brasilense. The R. capsulatus AmtB and AmtY proteins are only distantly related, 38% identity, and differ substantially in length, with the greatest difference being in the N-terminal region.

Disruption of amtB homologues in a number of bacteria has no effect on growth on various nitrogen sources, including ammonium, at neutral pH (28, 30, 31, 47). However, growth of an A. brasilense amtB mutant on low (0.1 mM) ammonium at pH 7.0 is affected (45), and an E. coli amtB mutant shows a growth defect on low concentrations of ammonium (≤1 mM) when growth is at pH 5 (41).

R. capsulatus amtB or amtY single mutants showed wild-type growth on rich organic medium (YPS) as well as on mineral medium (RCV) with NH₄⁺ (30 mM), N₂, or glutamate (7 mM) as a nitrogen source (data not shown). Upon transfer from YPS to RCV medium containing NH₄⁺ as the nitrogen source, amtB amtY double mutants also grew at wild-type rates after a somewhat longer lag phase (data not shown). Thus, disruption of amtB and/or amtY does not result in an observable growth defect at neutral pH.

Ammonium and methylammonium uptake.The ability of the mutant strains to grow on ammonium strongly suggests that ammonium is entering these cells, and several lines of evidence clearly corroborate this, as presented below. Physiological studies have previously shown that R. capsulatus possesses two ammonium transport systems, a system that appears to be constitutively synthesized, and a second system that appears to be Ntr regulated (34, 38) because it is abolished in the mutant J61, which carries a defect in ntrC (3). In contrast to the constitutively expressed system, the second system is competent for the uptake of methylammonium (34).

Here we measured the ammonium and methylammonium uptake activities of amt mutants of R. capsulatus. To test their uptake abilities, cultures were grown under nitrogen-limited conditions, thus permitting the expression of the Ntr-regulated system as well as the constitutively expressed system. Consistent with their ability to grow on ammonium, 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 (Fig. 1A). A mutation in amtY actually appeared to lead to a small increase in the apparent rate of ammonium uptake. This demonstrates that ammonium uptake and assimilation in R. capsulatus can proceed in the absence of AmtB and AmtY, which is consistent with the notion that, at least for the enteric bacteria Salmonella enterica serovar Typhimurium and Escherichia coli, unmediated diffusion of NH₃ is sufficient for optimal growth at reasonable concentrations of ammonium (greater than 50 μM) (41). Methylammonium uptake was largely unaffected in the amtY strain (in fact, levels somewhat higher than the wild-type level were seen), whereas the amtB mutant was no longer capable of transporting methylammonium (Fig. 1B). This suggests that AmtB corresponds to the Ntr-regulated system described previously (34). In support of this is the observation that the glnK-amtB operon is preceded by two DNA sequences 137 and 154 bp upstream of the translational start codon exhibiting strong similarity to NtrC binding sites (GC-N₇-T-N₃-GC) (11).

• Open in new tab
• Download powerpoint

FIG. 1.

Ammonium (A) and [¹⁴C]methylammonium (B) uptake. Cultures were grown under photoheterotrophic anaerobic conditions on RCV medium which lacked an added nitrogen source to the early stationary phase (0.07 mg of protein ml⁻¹), and uptake was measured after the addition of NH₄Cl to 200 μM for ammonium uptake assays or the addition of [¹⁴C]CH₃NH₃⁺ to 75 μM for methylammonium uptake assays. The values shown are the means of three independent replicates. Standard deviations are indicated as vertical bars.

Ammonium retention.One hypothesis concerning the possible role of AmtB has been that it functions as an ammonium transporter, regulating the concentration of ammonium in the interior of the cell (21). Since the uncharged species, NH₃, is in equilibrium with the charged species, NH₄⁺, NH₃ + H⁺ ⇔ NH₄⁺; pK_a = 9.25, there should be appreciable quantities of intracellular NH₃ at pHs close to neutral and NH₃ should pass rapidly across the cytoplasmic membrane by diffusion. However, the fact that free ammonium concentrations inside the cell have been determined to be much higher (in the millimolar range) than outside the cell (in the micromolar range) led to the postulation of active ammonium transporters (21).

According to this model, the NH₃ formed inside the cell must continuously diffuse out of the cell only to be taken back up as NH₄⁺ through a transport process, i.e., a futile cycle is created. According to Kleiner (21), who postulated the existence of this futile cycle, “Proof for this hypothesis would be the demonstration of constant NH₃ excretion upon loss of NH₄⁺ transport.” We checked for leakage of ammonium into the extracellular milieu by actively nitrogen-fixing cultures of wild-type and amt strains (Fig. 2). Under these conditions, measurements of intracellular ammonium concentrations showed that the concentration of intracellular ammonium was in the millimolar range (results not shown). The wild-type strain showed little ammonium in the external milieu, showing that the newly fixed ammonium was retained within the cells. However, analysis of the amtB and amtB amtY mutant strains showed that appreciable quantities of ammonium accumulated in the culture supernatants over time. Thus, AmtB is required for ammonium retention under these conditions.

• Open in new tab
• Download powerpoint

FIG. 2.

Ammonium retention by wild-type and mutant cells actively fixing dinitrogen. Cultures were grown as described in the legend to Fig. 1. At time zero, the cultures were sparged with oxygen-free dinitrogen, and extracellular ammonium was determined as described in Materials and Methods. The values shown are the means of duplicate assays. Standard deviations are indicated as vertical bars.

Nitrogenase switch-off and short-term Fe-protein ADP-ribosylation.Addition of ammonium to actively nitrogen-fixing cells of R. capsulatus causes quick, short-term-reversible responses. Recently, we have shown that these responses, nitrogenase switch-off and short-term Fe-protein ADP-ribosylation, are modulated by the cellular nitrogen status (50). Since methylammonium uptake activity was similarly affected by the cellular nitrogen status, we hypothesized that the nitrogenase switch-off effect observed upon addition of ammonium to actively nitrogen-fixing cells of R. capsulatus is mediated by the Ntr-regulated ammonium permease system (50). Here we examined the ability of amt strains to modulate nitrogenase activity and to ADP-ribosylate the Fe-protein of nitrogenase. Both the wild-type and amtY strains demonstrated a classical in vivo nitrogenase switch-off response to the addition of ammonium (200 μM final concentration) (Fig. 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 (Fig. 3).

• Open in new tab
• Download powerpoint

FIG. 3.

Effect of ammonium addition on in vivo nitrogenase activity and Fe-protein modification state. Cultures were grown as described in the legend to Fig. 1, and 5-ml aliquots were transferred by syringe to 25-ml argon-filled vials for the simultaneous analysis of nitrogenase activity by acetylene reduction, presented as graphs, and Fe-protein modification state, presented as immunoblots below the graphs. At the times indicated by the arrows, NH₄Cl was added. On all immunoblots, the lower bands correspond to the monomer form of unmodified Fe-protein, and the upper bands correspond to the ADP-ribosylated monomer. Since only one of the two Fe-protein subunits of the Fe-protein dimer is modified, 100% modification of Fe-protein dimers would correspond to two equal-intensity bands. Lane Std (immunoblots C to F) contains an ADP-ribosylated Fe-protein standard.

In striking contrast, the addition of the same amount of ammonium or more (50 mM final concentration) had no effect on the in vivo nitrogenase activity and Fe-protein modification state in either the amtB mutant or the amtB amtY double mutant (Fig. 3). To directly demonstrate that this effect was due to the introduced mutation, we studied the effects of complementation with pAY99, a plasmid carrying the cloned glnK-amtB operon in which expression of amtB is under the control of the native promoter upstream of glnK. 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 (Fig. 4).

• Open in new tab
• Download powerpoint

FIG. 4.

Effect of ammonium addition on in vivo nitrogenase activity and Fe-protein modification state in R. capsulatus amtB (A) and amtB amtY double (B) mutants complemented with amtB in trans. The broad-host-range replicon-based plasmid pAY99, carrying amtB under the control of the natural glnK promoter, was introduced by conjugation into the amtB and amtB amtY mutant strains. Experimental conditions were as described in the legend to Fig. 2.

Thus, disruption of amtB completely abolished the fast response to external ammonium addition. We wished to show that the effect occurred upstream of the regulatory processes and that therefore the nitrogenase modification system itself was intact. Apart from ammonium addition, short-term reversible ADP-ribosylation of Fe-protein is modulated by dark-light cycles (25). However, the timing and extent of Fe-protein ADP-ribosylation in response to darkness was unaltered in amtB and amtY mutants (Fig. 5). These results show that amtB mutants seem to be specifically affected in communicating the presence of extracellular ammonium to the nitrogenase switch-off and modification systems.

• Open in new tab
• Download powerpoint

FIG. 5.

Effect of light switch-off/switch-on on the modification state of Fe-protein. Cultures were grown as described in the legend to Fig. 1. Growth tubes were incubated for 15 min at room temperature under light, and 100-μl culture aliquots were withdrawn for the analysis of Fe-protein modification state (samples 1). After that, the tubes were incubated in darkness (sample 2, 5 min; sample 3, 10 min; sample 4, 20 min; and sample 5, 30 min). Sample 6, after 5 min of reincubation in the light.

Nitrogenase synthesis and long-term Fe-protein ADP-ribosylation.We wished to determine if cellular processes regulated with respect to the intracellular concentration of fixed nitrogen, such as nitrogenase expression (29) and long-term Fe-protein ADP-ribosylation (51), were affected. The sensitivity of nitrogenase synthesis to the presence of ammonium in amt strains was checked by varying the initial concentration of ammonium (0 to 25 mM) in the growth medium of batch cultures. Studies have previously shown that this produces cultures with different severities of nitrogen limitation, as assessed by the measurement of internal glutamine pools (51).

Within the range of ammonium concentrations used, the wild-type and amt strains demonstrated essentially the same relationship between the initial ammonium concentration in the medium and the final culture density as well as the presence of nitrogenase proteins. We assessed effects on synthesis by checking levels of nitrogenase protein by immunoblotting. The synthesis of nitrogenase in all strains was abolished with 21 to 22 mM ammonium (Fig. 6), although a slightly higher amount of initial ammonium appeared to be required to achieve complete inhibition of nitrogenase synthesis in the amtY strain. Thus, disruption of amtB or amtY does not appear to greatly affect the level of internal fixed nitrogen needed to suppress nif expression.

• Open in new tab
• Download powerpoint

FIG. 6.

Immunoblot analysis of nitrogenase Fe-protein in R. capsulatus strains grown with different initial concentrations of ammonium (indicated below the blots, in millimolar units). RCV medium containing the indicated initial concentrations of ammonium was inoculated with cells grown on RCV-ammonium excess (30 mM NH₄⁺) medium. On all immunoblots, the lower band represents unmodified Fe-protein monomer, while the upper, slower migrating band is an ADP-ribosylated subunit.

In addition to being sensitive to ammonium addition and light-dark cycles, the ADP-ribosylation status of the Fe-protein is regulated over the long term at a steady-state level by the internal pool of fixed nitrogen (measured as glutamine) (51). Immunoblot analysis also demonstrated that this long-term Fe-protein ADP-ribosylation was largely unaffected in the amtB mutant strain, where, as in the wild-type strain SB1003, constitutive Fe-protein ADP-ribosylation was observed in cells grown at moderate and higher initial ammonium concentrations (Fig. 6). This response appeared to be somewhat affected in the amtY mutants and the amtB amtY double mutant, where ADP-ribosylation was seen at slightly lower initial ammonium concentrations (13 and 14 mM). Thus, disruption of amtB or amtY does not appear to greatly affect the level of internal fixed nitrogen needed to repress nitrogenase expression or to induce long-term modification.

Regulation of glutamine synthetase activity.We wondered if the effect of the amtB gene disruption was specific to short-term nitrogenase regulation or if it extended to other enzymes whose activity is known to be regulated in the short term in response to exogenously added ammonium, such as glutamine synthetase. In R. capsulatus, as in many other bacteria, the addition of ammonium to cells results in a fast (within 5 to 10 min) decrease in glutamine synthetase specific activity as well as in the inactivation of this enzyme by adenylylation (9, 16, 32, 53). Therefore, we checked if the regulation of glutamine synthetase activity was affected in R. capsulatus amtB or amtY strains. Interestingly, based on the total activity as measured here, somewhat higher amounts of glutamine synthetase appeared to be present in the amtB and amtB amtY strains.

If total activity is a true reflection of the amount of protein present, this would suggest that AmtB might affect the synthesis of glutamine synthetase. Nevertheless, the addition of ammonium to cultures of the amt mutants caused a fast decrease in glutamine synthetase activity (Fig. 7) and an increase in adenylylation state (not shown), similar to what has been observed for wild-type cultures. Thus, a different mechanism is involved in the regulation of glutamine synthetase activity by ammonium.

• Open in new tab
• Download powerpoint

FIG. 7.

Effect of ammonium addition on glutamine synthetase (GS) activity. Cultures were grown as described in the legend to Fig. 1. Growth tubes were incubated for 15 min at room temperature under light, and 1-ml aliquots were withdrawn for the analysis of glutamine synthetase transferase activity (−NH₄⁺). A second set of samples (+NH₄⁺) were withdrawn 5 min after the addition of NH₄Cl (1 mM final concentration). Each bar represents an average of the results from at least three independent determinations, with standard deviations indicated by error bars.