INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Biological nitrogen fixation, the conversion of atmospheric nitrogen to ammonium, is catalyzed by the nitrogenase complex, which consists of two proteins: dinitrogenase (or MoFe protein) and dinitrogenase reductase (or Fe protein) (7). It is a very energy-demanding process and is thus tightly regulated at both transcriptional and posttranslational levels.

Transcriptional regulation of the nif genes has been found in all studied nitrogen-fixing bacteria and is best characterized in Klebsiella pneumoniae, a free-living nitrogen-fixing bacterium, where it involves the general nitrogen regulation (ntr) system (36). Analysis of the ntr regulatory system in K. pneumoniae and Escherichia coli (36, 39) has shown that it controls the transcription of many genes involved in nitrogen fixation and assimilation, such as glnA (encoding glutamine synthetase [GS]) and nifA (encoding the transcriptional activator for the other nif genes). The ntr system involves a number of gene products, including those of glnD, ntrA, ntrB, ntrC, and glnB, glnD encodes a bifunctional, uridylyltransferase-uridylyl-removing enzyme (UTase-UR) that is believed to be the sensor of the intracellular concentration of glutamine in the cell. UTase-UR reversibly controls the activity of the P_II protein (the gene product of glnB) by uridylylation or deuridylylation. P_II is responsible for sensing -ketoglutarate (-KG) in E. coli (24), and it controls NtrB (NRII) activity. NtrB and NtrC (the gene products of ntrB and ntrC) belong to the family of two-component regulators. NtrB is a histidine kinase that phosphorylates NtrC (NRI) under nitrogen-limiting conditions and also can act as a phosphatase to dephosphorylate NtrC under nitrogen excess conditions. Both kinase and phosphatase activities of NtrB are regulated by P_II in response to the level of -KG in the cell (21). At low -KG concentrations, the P_II trimer (bound to only one molecule of -KG) interacts with NtrB in vitro, thereby inhibiting its kinase activity and activating its phosphatase activity to dephosphorylate NtrC. However, at high -KG concentrations, the P_II trimer binds additional molecules of -KG and thereby is unable to interact with NtrB, so that NtrB acts as a kinase to phosphorylate NtrC (21). The phosphorylated form of NtrC acts as a transcriptional activator of nifA, glnA, and other operons involved in nitrogen assimilation. The activation involves ⁵⁴, encoded by ntrA (also referred to as rpoN). P_II, together with adenylytransferase (ATase), encoded by glnE, also controls GS activity by reversible adenylylation (22).

Besides the transcriptional regulation of nifA expression by the ntr system, NifA activity is also regulated. In K. pneumoniae and Azotobacter vinelandii, NifA activity is inhibited by a cotranscribed nif gene product, NifL, in response to NH₄⁺ and oxygen, probably through a direct interaction (29, 37). Recently, a P_II homolog (or paralog), GlnK, has been identified to be involved in the relief of NifL inhibition of NifA under N₂-fixing conditions in K. pneumoniae (18, 20).

GlnK has been found in E. coli, K. pneumoniae, and many other Bacteria, as well as Archaea (47). GlnK and P_II show very high sequence and structural similarity to one another (8, 34, 50). Each protein has been purified as a homotrimer (9, 34). Although these two proteins share some functions, such as the interaction with ATase to adenylylate GS (3, 48), they also have distinct functions in the cell. Only GlnK is involved in the relief of NifL inhibition in K. pneumoniae (18), although recent studies have shown that overexpressed P_II can substitute for GlnK to relieve the NifL inhibition (2). In E. coli and K. pneumoniae, the expression of glnK is regulated in response to NH₄⁺, but glnB is expressed constitutively (20, 48).

Nitrogen fixation is also regulated at the posttranslational level, which has been well characterized only in Rhodospirillum rubrum, Azospirillum brasilense, Azospirillum lipoferum, and Rhodobacter capsulatus, in which it involves reversible mono-ADP ribosylation of dinitrogenase reductase (33, 53). Dinitrogenase reductase ADP-ribosyl transferase (referred to here as DRAT, the gene product of draT) carries out the transfer of the ADP-ribose from NAD to the Arg-101 residue of one subunit of the dinitrogenase reductase homodimer, resulting in inactivation of that enzyme. Dinitrogenase reductase-activating glycohydrolase (referred to here as DRAG, the gene product of draG) removes the ADP-ribose group attached to dinitrogenase reductase, thus restoring nitrogenase activity. The DRAT-DRAG system negatively regulates nitrogenase activity in response to exogenous NH₄⁺ or energy limitation in the form of a shift to darkness (in the cases of R. rubrum and R. capsulatus) or to anaerobic conditions (in A. brasilense) (13, 16, 17, 25, 30, 35, 40, 51, 54).

As illustrated in Fig. 1, the regulation of the ADP-ribosylation of dinitrogenase reductase is effected through the posttranslational regulation of both DRAT and DRAG activities. Under nitrogen-fixing conditions, DRAT is inactive and DRAG is active, so that dinitrogenase reductase is in its active form. Following a negative stimulus, such as exogenous NH₄⁺ or energy depletion, DRAT is activated and DRAG becomes inactive, resulting in the loss of nitrogenase activity and the modification of dinitrogenase reductase. However, DRAT activation is only transient, and it becomes inactive again even in the continued presence of the negative stimulus. After removal of the negative stimulus, DRAG becomes active again, and it then reactivates dinitrogenase reductase by cleavage of the ADP-ribose group. Details of the mechanisms of regulation of DRAT and DRAG activities are still unknown. However, it has recently been found that the redox state of dinitrogenase reductase affects its ability to serve as substrate for DRAT and DRAG both in vitro and in vivo (14). DRAT can only modify oxidized dinitrogenase reductase and DRAG only removes ADP-ribosyl group from reduced dinitrogenase reductase. While the redox state of dinitrogenase reductase likely plays an important role in the regulation of DRAT and DRAG activities, it cannot be the only mode of regulation.

View larger version (27K): [in this window] [in a new window]   FIG. 1.   Scheme for the regulation of DRAT, DRAG, and dinitrogenase reductase activities in R. rubrum. The top panel shows the idealized time course of the regulation of nitrogenase activity following the treatment of negative stimuli, such as exogenous NH4+ and energy depletion. The bottom panel shows the regulation of DRAT, DRAG, and dinitrogenase reductase activities at different stages. Under nitrogen-fixing conditions (stage A), dinitrogenase reductase is in an active form without ADP-ribosylation, because DRAG is active and DRAT is inactive under these conditions. Following treatment with a negative stimulus (stage B), DRAT is activated and DRAG is inactivated, resulting in the loss of nitrogenase activity and the modification of dinitrogenase reductase. DRAT activity is only transient. In the period after the treatment with negative stimuli (stage C), both DRAT and DRAG have become inactive, and a steady residual nitrogenase activity remains. After cultures were returned to nitrogen-fixing conditions (stage D), DRAG becomes activated, and it removes the ADP-ribose from dinitrogenase reductase, thus restoring its activity.

draTG mutants have been constructed and physiologically characterized in R. rubrum, R. capsulatus, and A. brasilense (30, 35, 51, 54), confirming their functions in vivo. When draTG genes from A. lipoferum and R. rubrum were transferred into K. pneumoniae, which lacks this regulatory system, the nitrogenase activity was reversibly regulated in response to NH₄⁺ (12). These results suggest that the regulatory signals that control DRAT-DRAG are conserved among these organisms and that analysis of the DRAT-DRAG system in the well-studied K. pneumoniae would provide valuable insights into that regulation.

There are, however, significant differences in the ntr systems of the two organisms with respect to nitrogen fixation. In R. rubrum and A. brasilense, P_II is required for the activation of NifA activity (31, 55, 56); in R. rubrum, a strain with an alteration in glnB (P_II-Y51F, where the tyrosine that serves as the site of uridylylation was altered) had low nitrogenase activity, and only slight effects on the regulation of DRAT activity are seen (54). In R. rubrum and A. brasilense, ntrBC mutations have no effect on nif expression, whereas in K. pneumoniae, NtrBC proteins are required for nif expression (31, 36, 55). However, ntrBC mutations in A. brasilense do alter the regulation of DRAG activity and cause a slower inactivation of DRAG activity by NH₄⁺ (52), which indicates that some parts of the ntr system are involved in DRAG regulation.

Because of their key roles in the ntr system in response to an NH₄⁺ signal, P_II and GlnK are very reasonable candidates for serving as the factor in R. rubrum that actually causes the observed effects on DRAT-DRAG regulation. glnB and glnK of K. pneumoniae have therefore been examined for their effects on the heterologously expressed DRAT-DRAG system. The results demonstrate a striking effect of glnK mutations, strongly implying a role of P_II (or GlnK) homologs of R. rubrum in this regulation.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Bacterial strains and plasmids. The strains and plasmids used in this study are listed in Table 1. Antibiotics were used at the following concentrations (in milligrams/liter): streptomycin, 25; kanamycin, 12.5; and tetracycline, 10 (for K. pneumoniae); and streptomycin, 25; kanamycin, 25; and tetracycline, 12.5 (for E. coli).

Growth conditions and whole-cell nitrogenase activity assay. K. pneumoniae was first grown in rich LC solid medium (1% tryptone, 0.5% yeast extract, 0.5% NaCl, and 1.5% Bacto agar) at 30°C and then inoculated into 3 ml of minimal medium containing 0.2% ammonium acetate (KBSa) (38) and grown overnight at 30°C. A 0.5-ml portion of culture was inoculated into 25 ml of KBSa medium in a 125-ml flask and grown on a shaker at 250 rpm overnight. The cells were collected by centrifugation and resuspended in 125 ml of minimal medium containing 0.015% of L-serine to replace ammonium acetate (KBSser). To induce the expression of draTG, a low level (10 or 50 µM as indicated) of isopropyl--D-thiogalactopyranoside (IPTG) was added in KBSser medium. A 15 to 25-ml culture was transferred to a 60-ml serum vial with a rubber stopper, degassed and flushed with argon, and then derepressed for nitrogenase for 4 to 5 h under anaerobic conditions at 30°C. Assay of whole-cell nitrogenase activity has been described previously (15). To maintain plasmids, appropriate antibiotics were added in all media. Because all K. pneumoniae strains used in this study have the hisD2 mutation (19), L-histidine was added in all minimal media at a final concentration of 25 µg/ml.

DNA techniques. A freeze-thaw protocol was used for the transformation of plasmids into K. pneumoniae (46). pCH7 (P_tac-draTG) was constructed by digestion of pCH1 (P_tac-draTGB) (15) with NcoI and religated, resulting in the deletion of draB.

Immunoblotting of dinitrogenase reductase. A trichloroacetic acid precipitation method was used to extract protein quickly as described previously (51). Low-cross-linker sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) (the ratio of acrylamide to bisacrylamide was 172 to 1) was used for protein separation to obtain better resolution of the modified and unmodified subunits of dinitrogenase reductase, since the modified subunit migrates more slowly. Proteins from SDS-PAGE were electrophoretically transferred onto a nitrocellulose membrane, immunoblotted with polyclonal antibody against dinitrogenase reductase, and visualized with horseradish peroxidase color detection reagents (Bio-Rad, Richmond, Calif.).

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Effect of a glnB mutation on the DRAT-DRAG regulatory system. Plasmid pCH1, bearing R. rubrum draTGB, was transferred into UNF122 (wild type), UNF1537 (glnB insertion mutant which lacks P_II), and UNF3432 (glnK insertion mutant which lacks GlnK), yielding UN5507, UN5508, and UN5509, respectively. Similar to their parental strains lacking draTGB, UN5507 and UN5508 showed high initial nitrogenase activity, but glnK mutants (UNF1537 and UN5509) showed little nitrogenase activity (Table 2). The loss of nitrogenase activity in glnK mutants is caused by the failure to relieve NifL inhibition to NifA under N-limiting conditions (18, 20).

View larger version (17K): [in this window] [in a new window]   FIG. 2.   Regulation of nitrogenase activity by NH4Cl in K. pneumoniae UN5507 (wild type with draTGB) (), UN5508 (glnB mutant with draTGB) (), and UN5510 (wild type with draT) (). At t = 0, NH4Cl was added to a final concentration of 1 mM. Samples (1 ml) of the cells were withdrawn and assayed for nitrogenase activity for 2 min at the times indicated. The initial activities of UN5507, UN5508, and UN5510 were about 900, 700, and 700 nmol, respectively, of ethylene produced per h per ml of cells at an optical density at 600 nm of 1.0. IPTG was added to final concentration of 50 µM in the derepression medium.

The effect of a glnB mutation on DRAT regulation is dependent on the amount of DRAT in the cell. We routinely use 50 µM IPTG to induce the expression of draT or draTGB, which results in a low level of DRAT (and DRAG) protein accumulation, and the pattern of regulation of DRAT and DRAG activities in K. pneumoniae background is very similar to that seen in R. rubrum (15). In the glnB mutant, some DRAT escapes the regulation to become active under derepression conditions, resulting in a low initial nitrogenase activity in UN5511 (Table 2). It is possible that the loss of the regulation of DRAT activity was due to a decreased level of negative effector for DRAT activity in this mutant. We therefore tested different levels of IPTG to see if different levels of DRAT had any effect on the regulation of its activity.

View larger version (14K): [in this window] [in a new window]   FIG. 3.   Regulation of nitrogenase activity by NH4Cl in K. pneumoniae UN5510 (wild type with draT) () and UN5511 (glnB mutant with draT) (). At t = 0, NH4Cl was added to a final concentration of 10 mM. The initial nitrogenase activities of UN5510 and UN5511 were ca. 1,000 and 600 nmol, respectively, of ethylene produced per h per ml of cells at an optical density at 600 nm of 1.0. IPTG was added to final concentration of 10 µM in the derepression medium.

Effect of a glnK mutation on the regulation of nitrogenase activity. Because a very low nitrogenase activity was detected in K. pneumoniae glnK mutants (UNF3432 in Table 2), we are unable to study the modification of dinitrogenase reductase by the DRAT-DRAG regulatory system in these backgrounds. As reported previously, GlnK is necessary for the dissociation of NifL from NifA, which allows NifA activity under N-limiting conditions (18, 20). In a glnK mutant, NifL always inhibits NifA activity, probably forming a tight complex that cannot be dissociated under N-limiting conditions. We therefore introduced a multicopy plasmid with nifA (without nifL) into a glnK mutant background, so that excess NifA (relative to NifL) would allow NifA to be constitutively active. A plasmid, pCK3, which carries K. pneumoniae nifA expressed from a constitutive promoter (26), was transferred into these K. pneumoniae backgrounds, and the nitrogenase activities from these strains are shown in Table 4. As expected, in the presence of pCK3, glnK mutants (UN5515) showed a high nitrogenase activity, similar to that in wild-type and glnB backgrounds (UN5513 and UN5514). UN5515 showed no loss of nitrogenase activity after NH₄⁺ treatment, since it lacks the DRAT-DRAG regulatory system. The expression of K. pneumoniae nifA from pCK3 showed no significant effect on nitrogenase activity in wild-type and glnB backgrounds, and the nitrogenase activities in these strains (UN5513 and UN5514 in Table 4) are very similar to those in their parental strains (UNF122 and UNF1537 in Table 2). As seen before (Table 2), a higher nitrogenase activity (100 to 150% of the initial activity) was seen after NH₄⁺ treatment in these K. pneumoniae strains and that probably reflects de novo synthesis of nitrogenase enzymes.

The effect of a glnK mutation on the regulation of nitrogenase activity is due to altered regulation of DRAG activity but not of DRAT activity. pCH3, bearing R. rubrum draT alone, was transferred into a K. pneumoniae glnK background, which also contains K. pneumoniae nifA. This allowed us to study the effect of a glnK mutation on DRAT activity alone. Unlike the glnB mutants (UN5520 in Table 4 and UN5511 in Table 2), UN5521 (glnK with draT and nifA) showed a substantial initial nitrogenase activity at 50 µM IPTG, a level similar to that seen in UN5519 (wild type with draT and nifA). This suggested that most DRAT is inactive under N-limiting conditions in this glnK mutant. However, the nitrogenase activity in both UN5519 and UN5521 (both containing draT) is lower than that seen in UN5516 and UN5518 (both containing draTGB). As will be shown below, some ADP-ribosylation of dinitrogenase reductase exists in UN5519 and UN5521 but not in UN5516 and UN5518. We believe that this small amount of modification of dinitrogenase reductase is caused by the L-histidine used in the derepression medium. L-Histidine is a nitrogen source and could cause activation of some DRAT to modify dinitrogenase reductase even under derepression conditions, especially at 50 µM IPTG. Because of the lack of DRAG in UN5519 and UN5521, a small portion of active DRAT caused some modification of dinitrogenase reductase in these strains. In contrast, the presence of active DRAG in UN5516 and UN5518 can compensate for this low DRAT activity, so that no modification of dinitrogenase reductase was seen in these strains under derepression conditions (see below). Both UN5519 and UN5521 showed a normal NH₄⁺ response and a complete loss of nitrogenase activity after NH₄⁺ treatment (Table 4). To further analyze the DRAT regulation and its response to NH₄⁺, a time course of the loss of nitrogenase activity following NH₄⁺ treatment was also monitored in UN5519 and UN5521. Both UN5519 (wild type) and UN5521 (glnK) showed a fast rate of NH₄⁺ response similar to that in UN5510 (Fig. 2) (data not shown). These results indicate that the absence of glnK has no significant effect on DRAT regulation, so that a lack of NH₄⁺ response in UN5518 must be caused by altered DRAG regulation. While DRAG in the wild type becomes inactive after NH₄⁺ treatment, DRAG in this K. pneumoniae glnK background appears to stay active even in the presence of NH₄⁺. All of these results indicate that mutation of glnK showed no significant effect on the regulation of DRAT activity but has a profound effect on the regulation of DRAG activity.

Different levels of IPTG have no effect on DRAG regulation in the glnK mutant background. We also examined the regulation of DRAT and DRAG activities in these glnK mutants in the presence of pCK3 at a low IPTG level (10 µM) and to see if it was also dependent on the amount of DRAT and DRAG proteins in the cell. As shown in Table 4, in the presence of draTGB, a very similar pattern of NH₄⁺ response was seen in UN5516, UN5517, and UN5518 at both 10 and 50 µM IPTG levels, with only a small difference in residual activity in UN5516 and UN5517. No NH₄⁺ response was seen in UN5518 in either condition. In the presence of draT alone, all strains have a higher initial nitrogenase activity at 10 µM IPTG than that seen at 50 µM IPTG, especially UN5520. This is consistent with the results seen in UN5511 (without pCK3) in Tables 2 and 3 and might be due to less DRAT expression at 10 µM IPTG. The NH₄⁺ response in the wild type and the glnK mutant was exactly the same at these two different levels of IPTG. Unlike the altered regulation of DRAT activity in glnB mutant, the effect of a glnK mutation on the regulation of DRAG is independent of the level of DRAG protein in the cell. Together these results suggest that the regulation of DRAT and DRAG have different mechanisms.

Correlation of ADP-ribosylation of dinitrogenase reductase with the regulation of nitrogenase activity in K. pneumoniae strains containing K. pneumoniae nifA and R. rubrum draTGB or draT. In addition to the assay of nitrogenase activity to monitor DRAT and DRAG activities indirectly, a direct assay for DRAT and DRAG activities is the ADP-ribosylation of dinitrogenase reductase, which can be monitored directly by immunoblotting. In R. rubrum, the ADP-ribosylated (inactive) form of dinitrogenase reductase runs as two bands on SDS-PAGE, because the ADP-ribosylated subunit migrates slower, while active dinitrogenase reductase has two identical subunits that migrate as a single band (55). In contrast, active dinitrogenase reductase in K. pneumoniae migrated as two bands (labeled "U" in lanes 1, 3, and 5 in Fig. 4). These two bands were also found in UNF122 (wild type without draT) (data not shown) and were reported previously (43); the reason for this doublet is unknown. As we expected, UN5516, UN5517, and UR5518 showed no modification of dinitrogenase reductase before NH₄⁺ addition (Fig. 4). After NH₄⁺ treatment, both UN5516 and UN5517 showed a modified (ADP-ribosylated) upper band (labeled "M" in Fig. 4). However, very little modification of dinitrogenase reductase was found in UN5518 after NH₄⁺ treatment, a finding consistent with the lack of NH₄⁺ effect on nitrogenase activity in this glnK mutant (Table 4).

View larger version (21K): [in this window] [in a new window]   FIG. 4.   Western immunoblot of dinitrogenase reductase in K. pneumoniae UN5516 (wild type with R. rubrum draTGB and K. pneumoniae nifA) (lanes 1 and 2), UN5517 (glnB mutant with R. rubrum draTGB and K. pneumoniae nifA) (lanes 3 and 4), UN5518 (glnK mutant with R. rubrum draTGB and K. pneumoniae nifA) (lanes 5 and 6), UN5519 (wild type with R. rubrum draT and K. pneumoniae nifA) (lanes 7 and 8), UN5520 (glnB mutant with R. rubrum draT and K. pneumoniae nifA) (lanes 9 and 10), and UN5521 (glnK mutant with R. rubrum draT and K. pneumoniae nifA) (lanes 11 and 12). Samples of dinitrogenase reductase were collected by trichloroacetic acid precipitation from cultures grown in derepression medium in the presence of 50 µM IPTG in derepression medium before (lanes 1, 3, 5, 7, 9, and 11) and after (lanes 2, 4, 6, 8, 10, and 12) treatment of 10 mM NH4Cl for 40 min. Arrow "M" indicates the position of the modified (ADP-ribosylated) subunit, and arrow "U" indicates that of the unmodified subunits, as explained in the text.

DRAB plays little role in the effect of GlnK and P_II on the regulation of DRAG activity. draB is an open reading frame located in downstream of draTG in R. rubrum, and it is apparently cotranscribed with draTG (D. P. Lies and G. P. Roberts, unpublished data). draB shows high sequence similarity to nifO of A. vinelandii (44) and some similarity to arsC of E. coli (4, 10). While the function of nifO of A. vinelandii is unknown, it might be involved in molybdenum metabolism (44). arsC encodes an arsenate reductase, which catalyzes the reduction of arsenate to arsenite (10). In R. rubrum the mutation of draB caused a decrease in DRAG activity and protein level, and the total DRAG activity in draB mutant is approximately 35% of that of the wild type (30). Recently, DRAG has been shown to have a binuclear manganese center when it is treated with Mn²⁺ (1). Based on its sequence similarity to other metal processing proteins, therefore, DRAB might be involved in the processing of a metal center in DRAG. To investigate if DRAB is involved in the regulation of DRAG activity by P_II and GlnK, plasmid pCH7, carrying R. rubrum draTG (no draB), was transferred into different K. pneumoniae backgrounds. Nitrogenase activity in these strains and its response to NH₄⁺ was monitored, and the results are shown in Table 5. Overall, there was no significant difference in nitrogenase activity and its response to NH₄⁺ between K. pneumoniae strains with draTGB and those with draTG alone. A partial NH₄⁺ response was also seen in glnB mutants with draTG alone (UN5523 and UN5526), but the residual activity in these strains is lower than that seen in glnB mutants with R. rubrum draTGB (UN5508 in Table 2 and UN5517 in Table 4). This might due to the effect of draB on DRAG activity itself, as reported previously (30). The glnK mutant with R. rubrum draTG (UN5527) showed no NH₄⁺ response, a result similar to that seen in the glnK mutant with R. rubrum draTGB (UN5518 in Table 4). These results indicate that draB has no significant effect on the regulation of DRAG activity by P_II and GlnK under the conditions employed for these experiments.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

Although the general function of DRAT and DRAG in the regulation of nitrogenase activity has been characterized in R. rubrum, A. brasilense, and R. capsulatus, the mechanisms for the regulation of DRAT and DRAG activities are still unknown. When draTG genes from A. lipoferum or R. rubrum were transferred into K. pneumoniae, the nitrogenase activity was reversibly regulated in response to NH₄⁺ (12, 15), indicating that the regulatory effectors in the signal transduction pathways for the regulation of DRAT and DRAG should be present in K. pneumoniae. Because P_II and GlnK play very important roles in sensing NH₄⁺ status to regulate nif expression in K. pneumoniae, it seemed reasonable to consider that these proteins may also be involved in the regulation of the DRAT-DRAG system in response to NH₄⁺ as well. Indeed, the results presented here show significant effects on the DRAT-DRAG system by mutations in both glnB and glnK. These results provide a testable hypothesis that P_II and its homolog(s) probably have a similar function in the regulation of DRAT and DRAG activity in organisms that normally contain the DRAT-DRAG regulatory system.

The mutation of glnB in K. pneumoniae showed effects on the regulation of both DRAT and DRAG activities. DRAT escapes normal regulation and becomes active under N-limiting conditions in this glnB mutant, and this altered regulation of DRAT activity is probably due to the decreased level of some inhibitor for DRAT in this mutant, because elevated DRAT levels exacerbate this effect. It has been our hypothesis that both DRAT and DRAG activities are regulated by loosely binding negative inhibitors, since both DRAT and DRAG always are active in in vitro assays either in extracts or when purified. It is possible that the expression of the negative effector for DRAT is affected by P_II. The level of this negative effector might be significantly lower in a glnB mutant than in the wild type, resulting in the altered regulation of DRAT activity when more DRAT is induced in the cell. Using this glnB mutant as a control, it is possible to screen for potential candidates of the negative effector for DRAT in vitro.

The regulation of DRAG activity is also altered in a glnB mutant, with DRAG being inactivated more slowly by NH₄⁺ in this mutant than in the wild type, resulting in a high residual nitrogenase activity. A K. pneumoniae glnK mutation showed an even more profound effect on the regulation of DRAG activity but no significant effect on the regulation of DRAT activity. It has been previously shown that glnK is normally expressed in a glnB mutant under N-limiting conditions (20), and Western blots also showed that the level of GlnK in this glnB mutant is very similar to that seen in the wild type (data not shown), so that the effects of the glnB mutation are not due to direct or indirect effects on GlnK. These data indicate that both P_II and GlnK are required for the maximum inactivation of DRAG activity in response to NH₄⁺. Though it is unknown if the effects of P_II and GlnK on DRAG are direct, we have added purified P_II and GlnK to an in vitro DRAG activity assay (which measures the conversion of modified dinitrogenase reductase into active form), but no inhibition of DRAG activity was seen (data not shown). The effect of P_II and GlnK on DRAG might be indirect and may involve the regulation of an unknown negative effector that regulates DRAG activity. The lack of regulation of DRAG activity in a K. pneumoniae glnK mutant could be caused by either the absence of the negative effector for DRAG (transcriptional regulation) or the accumulation of the inactive form of this effector (posttranslational regulation). Regardless of which regulatory mechanism is involved, the lack of the active form of the negative effector in a glnK mutant should allow us to screen for such small molecules or protein effectors in vitro.

While the results presented above provide compelling evidence that P_II and GlnK might be central to DRAT-DRAG regulation in organisms that normally contain this regulatory system, the specific roles of these P_II homologs are probably not identical to that seen in K. pneumoniae. In R. rubrum, creation of P_II-Y51F (in which the tyrosine site for uridylylation of P_II is changed to phenylalanine) allows some DRAT to escape regulation under N-limiting conditions (56). Unlike the case in K. pneumoniae, however, a glnB mutation in R. rubrum showed only a small effect on the regulation of nitrogenase activity (Y. Zhang and G. P. Roberts, unpublished data). Some other aspect of the ntr system of R. rubrum and A. brasilense must be involved in the regulation of DRAT-DRAG system, since the mutation of ntrBC had a significant effect on the regulation of DRAG activity, causing a slower inactivation of DRAG activity by NH₄⁺ (52). One candidate for this regulator is the P_II homolog, GlnK. Consistent with this hypothesis, R. rubrum and R. sphaeroides cbbM mutants (defective in the Calvin-Benson-Bassham pathway) showed high nitrogenase activity even in the presence of NH₄⁺ (23), indicating an alteration in nif expression and DRAT-DRAG regulation in this R. rubrum cbbM mutant. The expression of glnB and glnK were also affected in the cbbM mutant in R. sphaeroides (41). Therefore, the GlnB and GlnK families, although falling into evolutionarily distinct groupings, might not have specific functions that are as consistently grouped. The eventual analysis of the effects on the DRAT-DRAG system of different P_II and GlnK homologs from different species should begin to clarify the sequence features that define the functional differences in these proteins.

The effect of P_II and its homologs on the posttranslational regulation of nitrogenase activity has been reported in the Archaea Methanococcus maripaludis (27). In M. maripaludis, a deletion mutant of two glnB homologs has no effect on nif expression but a significant effect on the posttranslational regulation of nitrogenase activity in response to NH₄⁺. Unfortunately, the mechanism for the posttranslational regulation of nitrogenase activity has not been characterized in M. maripaludis, but it may be also involved in the ADP-ribosylation of dinitrogenase reductase. DRAG homologs have also been found in other Bacteria and some Archaea, such as Archaeoglobus fulgidus, Methanococcus jannaschii, E. coli, Aquifex aeolicus, Deinococcus radiodurans, Listeria monocytogenes, and Streptomyces coelicolor (4-6, 11, 28, 42, 49).

In conclusion, the results presented here are that (i) P_II of K. pneumoniae affects DRAT regulation and this effect is dependent on levels of DRAT protein, suggesting that regulation is through a molecule (an inhibitor for DRAT) that is nonabundant; (ii) P_II also affects the regulation of DRAG activity, but less dramatically than does GlnK; (iii) the absence of GlnK alters DRAG, but not DRAT, and this effect is independent of the levels of DRAG; (iv) overexpression of nifA of K. pneumoniae overcomes the effect of glnK mutations on expression of other nif genes; (v) ADP-ribosylation of dinitrogenase reductase correlates with the loss of nitrogenase activity in K. pneumoniae, a result consistent with the view that there is no other posttranslational regulation of nitrogenase activity in this organism in response to NH₄⁺; (vi) DRAB, whose absence has modest effects on DRAG activity in R. rubrum, does not appear to be involved in the P_II or GlnK effects on DRAT-DRAG seen in K. pneumoniae; and (vii) the precise functions of P_II and GlnK appear to be different in different organisms.