PII-like signal transduction proteins are found in all three domains of life and have been shown to play key roles in the control of bacterial nitrogen assimilation. This communication reports the first target protein of an archaeal PII-like protein, representing a novel PII receptor. The GlnK1 protein of the methanogenic archaeon Methanosarcina mazei strain Gö1 interacts and forms stable complexes with glutamine synthetase (GlnA1). Complex formation with GlnK1 in the absence of metabolites inhibits the activity of GlnA1. On the other hand, the activity of this enzyme is directly stimulated by the effector molecule 2-oxoglutarate. Moreover, 2-oxoglutarate antagonized the inhibitory effects of GlnK1 on GlnA1 activity but did not prevent GlnK1/GlnA1 complex formation. On the basis of these findings, we hypothesize that besides the dominant effector molecule 2-oxoglutarate, the nitrogen sensor GlnK1 allows finetuning control of the glutamine synthetase activity under changing nitrogen availabilities and propose the following model. (i) Under nitrogen limitation, increasing concentrations of 2-oxoglutarate stimulate maximal GlnA1 activity and transform GlnA1 into an activated conformation, which prevents inhibition by GlnK1. (ii) Upon a shift to nitrogen sufficiency after a period of nitrogen limitation, GlnA1 activity is reduced by decreasing internal 2-oxoglutarate concentrations through diminished direct activation and by GlnK1 inhibition.
PII-like proteins GlnB and GlnK belong to the family of small signalling proteins identified in all three domains of life (Ninfa and Atkinson, 2000; Arcondeguy et al., 2001). They are known to play an important role in sensing and transducing cellular nitrogen signals and therefore being involved in the regulation of nitrogen metabolism (reviewed by Arcondeguy et al., 2001; Kessler et al., 2001; Ehlers et al., 2002; Forchhammer, 2004). Regulatory mechanisms mediated by PII-like proteins are well studied in bacteria, for which a variety of different receptor proteins have been identified and characterized (Kamberov et al., 1994; Jaggi et al., 1997; de Zamaroczy, 1998; Nolden et al., 2001; Coutts et al., 2002; Heinrich et al., 2004; Javelle et al., 2004). In general, bacterial PII-like proteins are covalently modified and demodified in response to changes in nitrogen availability; however, the modification differs: uridylylation of PII-like proteins has been demonstrated for enteric bacteria (Jiang et al., 1998a; Atkinson and Ninfa, 1999), adenylylation for the actinomycetes Streptomyces coelicolor and Corynebacterium glutamicum (Hesketh et al., 2002; Strösser et al., 2004) and phosphorylation for the cyanobacterium Synechococcus elongatus (Forchhammer and Tandeau de Marsac, 1995). Exception is the GlnY protein of Azoarcus sp. BH72, which exists exclusively in one modification state, in its uridylylated form, independently of the actual nitrogen availability (Martin et al., 2000). Furthermore, in various systems PII-like proteins seem not to be subject to covalent modification, such as in Prochlorophytes (Palinska et al., 2002), Bacillus subtilis (Detsch and Stülke, 2003), and in plant PII-like proteins (Smith et al., 2004). It has been recently shown that external nitrogen limitation is perceived as internal glutamine limitation in case of Escherichia coli (Ikeda et al., 1996; Jiang et al., 1998a). However, in unicellular cyanobacteria the cellular 2-oxoglutarate level is the internal nitrogen signal, which determines the modification state of the PII protein in response to nitrogen availability (Irmler et al., 1997; reviewed by Forchhammer, 1999). In addition to covalent modification, allosteric binding of small effector molecules, in particular ATP and 2-oxoglutarate to PII is an important signal input into the PII system, thus allowing the integration of various signals to generate a coherent response. Depending on their modification and ligand binding states, bacterial PII-like proteins modulate the activity of several receptor proteins involved in nitrogen assimilation. (i) E. coli GlnB effects the activity of the transcriptional activator NtrC by interacting with the histidine kinase NtrB (Jiang et al., 1998b) and regulates the activity of the adenylyltransferase (ATase), which controls the glutamine synthetase activity (Jiang et al., 1998c; Reitzer, 2003). (ii) The second PII-like protein in E. coli, GlnK, has been shown to act as a backup system and as a fine control regulator for the GlnB regulatory cascade (Atkinson et al., 2002). In addition, GlnK appears to regulate the activity of the ammonium transporter AmtB by direct protein interaction after a shift to nitrogen sufficiency (Coutts et al., 2002; Javelle et al., 2004). (iii) In nitrogen fixing bacteria, GlnK has been shown to transduce the internal nitrogen status towards the regulatory proteins NifA or NifL by direct interaction (Liang et al., 1992; Arsene et al., 1996; 1999; Little et al., 2000; Rudnick et al., 2002; Drepper et al., 2003; Stips et al., 2004). Besides those receptor proteins, a new target protein was recently discovered for the PII protein of S. elongatus. Under nitrogen excess conditions, the non-phosphorylated PII-protein forms stable complexes with the key enzyme of the arginine biosynthesis pathway, N-acetylglutamate kinase, and thereby enhances its enzyme activity by an order of magnitude (Heinrich et al., 2004).
In contrast to bacterial PII-like proteins, until now only one potential receptor protein has been identified for archaeal PII-like proteins. The gene products of the glnB-like genes nifI1 and nifI2 located within the nif gene operon in Methanococcus maripaludis have been shown to be essential for the ammonia switch-off of nitrogenase activity in response to a shift to nitrogen sufficiency (Kessler et al., 2001). Characterizing nifI mutant strains demonstrated that both NifI proteins are essential for modulating nitrogenase activity (Kessler et al., 2001). However, as no biochemical data are available, the regulatory mechanism is still not completely understood. Recently, we characterized the archaeal PII-like protein GlnK1 of Methanosarcina mazei strain Gö1 and demonstrated that the archaeal GlnK1 protein structurally differs significantly from bacterial PII-like proteins. Nevertheless, M. mazei GlnK1 was able to complement an E. coli ΔglnK mutant strain, strongly indicating that the archaeal GlnK protein is involved in nitrogen regulation (Ehlers et al., 2002). The goal of this work was to elucidate the regulatory function of GlnK1 in nitrogen metabolism of M. mazei by identifying potential interacting partners. During our studies, we identified glutamine synthetase (GlnA1) as the first receptor protein of GlnK1 and characterized the effect of complex formation between GlnK1 and GlnA1 on glutamine synthetase activity.
The goal of this work was to gain a deeper insight into the regulatory function of GlnK1 in nitrogen metabolism in M. mazei. In order to address the question, whether the archaeal PII-like protein interacts with other proteins involved in nitrogen metabolism, we examined potential complex formation between GlnK1 and cell extract proteins using different approaches.
Identification of potential interacting proteins by pull-down experiments
In order to identify potential receptor proteins that directly interact with GlnK1, we studied complex formation between GlnK1 N-terminally fused to a His6-tag (His6-GlnK1) and cell extract proteins by pull-down experiments using affinity chromatography on Ni-NTA agarose for detecting complexes. His6-GlnK1 was heterologously expressed, and purified to an apparent homogeneity of 98% by Ni-NTA affinity chromatography as described in Experimental procedures. Purified His6-GlnK1 protein was bound to Ni-NTA agarose and cell extracts of M. mazei cells, which were grown under nitrogen limitation (N2) but shifted to nitrogen sufficiency for 30 min (NH4+ shift), were applied to the immobilized His6-GlnK1. After extensively washing the chromatography material in order to remove all cell extract proteins, which bound unspecifically to GlnK1 or the Ni-NTA agarose, His6-GlnK1 and potential specifically interacting proteins were eluted in the presence of imidazole. The respective elution fractions were analysed by denaturating polyacrylamide gel electrophoresis (SDS-PAGE) and subsequent silver staining. Four additional protein bands corresponding to an approximate molecular mass of 30 kDa, 45 kDa, 62 kDa and 120 kDa were detected in the elution fractions besides the dominant protein band of GlnK1(Fig. 1A). The protein band of approximately 120 kDa was also detected in comparable amounts in control experiments, in which cell extract of ammonium-shifted cells was passed through unloaded Ni-NTA agarose; indicating that the protein binds to the agarose material in an unspecific way. Western blot analysis using antibodies directed against GlnK1 revealed that the co-eluting protein band with an approximate molecular mass of 30 kDa represents dimeric GlnK1 protein, which was not completely denatured (Fig. 1B). The remaining co-eluted protein bands with molecular masses of 45 and 62 kDa were further analysed by N-terminal sequencing as described in Experimental procedures. We did not succeed in identifying the 62 kDa protein, as a mixture of at least 2–3 overlapping sequences were obtained; however, a distinct amino acid sequence was obtained for the 45 kDa protein. Based on the N-terminal amino acid sequence, the 45 kDa protein was identified by genome-wide protein sequence analysis to be the gene product of the M. mazei open reading frame (orf) MM0964. This orf encodes for a homologous protein of a bacterial glutamine synthetase and thus was designated glnA1 (Deppenmeier et al., 2002). In order to exclude that the GlnA1 protein binds unspecifically to the Ni-NTA matrix, control experiments with M. mazei cell extracts from nitrogen-limited or ammonium-shifted cells and unloaded Ni-NTA agarose were performed. SDS-PAGE analysis followed by silver staining or Western blot analysis using antibodies directed against GlnA1 clearly demonstrated that GlnA1 was not detectable in these elution fractions.
In order to verify the complex formation between GlnK1 and GlnA1 a complementary pull-down experiment was performed. His6-GlnA1 was heterologously expressed and purified to an apparent homogeneity of 99% by fractionated ammonium precipitation followed by Ni-NTA affinity chromatography as described in Experimental procedures (see Fig. 2). The purified fraction was loaded on Ni-NTA columns and exposed to M. mazei cell extracts. His6-GlnA1 co-eluted from the affinity column together with three additional proteins present in cell extracts of nitrogen-limited or ammonium-shifted cells with approximate molecular masses of 14, 18 and 30 kDa (Fig. 3A). Western blot analysis using antibodies directed against GlnK1 confirmed that the prominent 14 and the faint 30 kDa protein bands detected in the silver stain represent monomeric and dimeric GlnK1 respectively (Fig. 3B). Control experiments in which M. mazei cell extracts were passed through non-loaded Ni-NTA columns ruled out unspecific binding of GlnK1 (Fig. 3C), further confirming complex formation between GlnK1 and GlnA1. Identification of the 18 kDa protein did not succeed because of insufficient amount of protein; the minute amounts of this protein suggest that it might be a contamination.
Interaction studies of GlnK1 and GlnA1 by gel filtration analysis
To confirm the observed protein–protein interaction, complex formation between GlnK1 and GlnA1 was studied by gel filtration analysis using independently expressed and purified proteins both fused to a His-tag. When applied separately to the BioSil gel filtration column, purified His6-GlnA1 eluted as a single peak (Fig. 4B, main fractions: fractions 11–13) corresponding to an apparent molecular mass of approximately 220 kDa as determined in comparison with standard protein markers. As monomeric His6-GlnA1 shows a molecular mass of 42 kDa, the native protein eluting from the gel filtration column appears to be in a higher oligomeric structure probably consisting of at least six subunits. In comparison, purified His6-GlnK1 eluted as a single peak (Fig. 4A, main fractions: fractions 14–16), corresponding to an apparent molecular mass of 40 kDa representing trimeric GlnK1, which is consistent with recent findings obtained by native PAGE analysis of M. mazei GlnK1 in cell extracts (Ehlers et al., 2002). In order to analyse complex formation between GlnK1 and GlnA1, His6-GlnK1 (0.95 µmol trimeric GlnK1) was preincubated together with His6-GlnA1 (0.3 µmol hexameric GlnA1) for 5 min at room temperature (RT) prior the application to the gel filtration column. The elution peak of free GlnK1 was still detectable in the elution profile as we used GlnK1 in excess (Fig. 4C). Further, no significant change of the elution volume of the His6-GlnA1 elution peak was detectable, as an increase of the molecular mass upon complex formation with trimeric GlnK1 is not resolved by the gel filtration column used here. Thus, the fractions of the GlnA1 elution peak and the elution peak of unbound GlnK1 (fractions 8–15) were analysed by Western blotting using antibodies directed against GlnK1. The quantification of GlnK1 in the various fractions clearly demonstrated the presence of significant amounts of GlnK1 in the main peak fractions of GlnA1 (Fig. 4C, fractions 11–13). Thus, in the presence of GlnA1, GlnK1 is eluting three fractions earlier compared to the elution profile with only GlnK1 applied to the column (compare Fig. 4A with Fig. 4C and E). The finding that GlnK1 co-elutes with GlnA1 from the gel filtration column, which was verified using several independent GlnA1 and GlnK1 preparations, strongly indicates that GlnA1 forms stable complexes with GlnK1 and confirms the protein–protein interaction between GlnK1 and GlnA1 demonstrated in the pull-down experiments.
For bacteria, it has been demonstrated that the small effector molecules 2-oxoglutarate and ATP can influence binding and interaction of PII-like proteins with interaction partners (Jiang et al., 1998c; Xu et al., 1998; Little et al., 2002; Ruppert et al., 2002). Therefore, we investigated the influence of 2-oxoglutarate and ATP on GlnA1/GlnK1 complex formation by gel filtration. Complex formation between GlnK1 and GlnA1 was carried out in the presence of either 1 mM 2-oxoglutarate, 1 mM ATP or both as described above, followed by gel filtration with the buffer system supplemented with the respective effector molecules (1 mM). Analysing GlnK1 in the respective elution fractions revealed that the presence of 2-oxoglutarate resulted in co-elution of GlnK1 with GlnA1 mostly in fractions 12 and 13, one fraction later than in the absence of 2-oxoglutarate (compare Fig. 4C and D). This finding, which was confirmed using different protein preparations, indicates that 2-oxoglutarate does not prevent complex formation between GlnA1 and GlnK1; however, it appears that in the presence of 2-oxoglutarate less stable GlnA1/GlnK1 complexes are formed (Fig. 4F). The presence of ATP or simultaneous presence of 2-oxoglutarate and ATP did not additionally affect complex formation (data not shown). In order to confirm that 2-oxoglutarate is not preventing complex formation between GlnK1 and GlnA1, we analysed the effect of 2-oxoglutarate on GlnK1/GlnA1 complex formation using pull-down experiments. As depicted in Fig. 5, GlnK1 present in ammonium-shifted cell extracts formed complexes and co-eluted with His6-GlnA1 from Ni-NTA agarose irrespectively, whether 2-oxoglutarate was present or not during complex formation (compare Fig. 5A with Fig. 5C) or whether the immobilized His6-GlnA1/GlnK1 complexes were extensively washed in the presence of 1 mM 2-oxoglutarate (Fig. 5B and C). This further supports the finding that 2-oxoglutarate does not prevent complex formation between GlnA1 and GlnK1; however, the method used does not allow to detect small differences in complex stabilities or differences in complex compositions.
GlnK1 affects glutamine synthetase activity of purified GlnA1
Glutamine synthetase activity of purified heterologously expressed His6-GlnA1 was analysed immediately after purification to avoid storage of the enzyme preparation (see below). The specific glutamine synthetase activity of several independently synthesized and purified His6-GlnA1 preparations was determined to be in the range of 0.4–0.6 U mg−1 using the coupled optical test assay developed by Shapiro (Shapiro and Stadtman, 1970) as described in Experimental procedures. Upon storage at either 4°C, −20°C or −70°C glutamine synthetase activity of purified His6-GlnA1 dramatically decreased, indicating that the purified protein is rather instable. This is consistent with results obtained by gel filtration analysis of purified GlnA1, revealing that GlnA1 elutes in a lower oligomeric structure (apparently in a hexameric structure) than expected based on its homology to other archaeal glutamine synthetases, which belong to the GSI-α subdivision of glutamine synthetases (Brown et al., 1994). Thus, the native potentially dodecameric structure of GlnA1 tends to rapidly dissociate into lower oligomeric structures. As those lower oligomeric GlnA1 structures are apparently inactive, this dissociation into lower oligomers is probably the cause of the instability of GlnA1 enzyme activity.
To analyse, whether GlnK1 affects GlnA1 activity, glutamine synthetase activity of GlnA1 was determined in the presence of different amounts of purified His6-GlnK1. Analysing at least four independent enzyme preparations of GlnA1 indicated that the presence of increasing GlnK1 amounts in the test assay resulted in a significant decrease of GlnA1 activity as depicted in Fig. 6A. In the presence of 1.75 µM trimeric GlnK1, glutamine synthetase activity (0.9 µM monomeric GlnA1) decreased to approximately 20 ± 10% of the activity determined in the absence of GlnK1. However, this inhibition of GlnA1 activity by GlnK1 was only detectable, when GlnA1 was preincubated for at least 5 min with GlnK1 before the assay was started, indicating that direct interaction and/or complex formation between the two proteins requires a certain period of time.
2-oxoglutarate stimulates glutamine synthetase activity and antagonizes inhibitory effects of GlnK1
As no 2-oxoglutarate dehydrogenase is present in M. mazei, internal 2-oxoglutarate concentrations may have a central role in the perception of nitrogen availabilities. Thus, we investigated, whether 2-oxoglutarate affects glutamine synthetase activity of GlnA1 directly. Purified His6-GlnA1 was preincubated for 5 min at RT in the presence of increasing 2-oxoglutarate concentrations (200 µM−2.5 mM) followed by determination of glutamine synthetase activity. Unexpectedly, the presence of 2-oxoglutarate resulted in a significant increase of glutamine synthetase activity (Fig. 6B). Up to 16-fold increase of glutamine synthetase activity of GlnA1 was obtained, when incubated in the presence of 2.5 mM 2-oxoglutarate. When GlnA1 (100 µg, equals 2.25 µM monomeric GlnA1) was preincubated for 5 min in the presence of both, the positive effector molecule 2-oxoglutarate (2.5 mM) and its inhibitor GlnK1 (1.75 µM trimeric GlnK1), glutamine synthetase activity was in the same range as the activity observed for GlnA1 preincubated exclusively with 2-oxoglutarate (4.9 U mg−1 versus 4.5 U mg−1) (Fig. 7A). This suggests that the effector 2-oxoglutarate is able to abolish the inhibitory effects of GlnK1. 2-oxoglutarate could antagonize inhibition by GlnK1 even after GlnK/GlnA complex formation has already occurred. As shown by the experiment depicted in Fig. 7B, preincubation of GlnA1 in the presence of GlnK1 inhibits glutamine synthetase activity (GlnA1 + GlnK1). However, only a few seconds after the addition of 2-oxoglutarate to the GlnA1/GlnK1 mixture, glutamine synthetase activity rapidly increased as visualized by the decrease in absorbance resulting from NADH oxidation [GlnA1 + GlnK1 (+OG)]. This suggests that 2-oxoglutarate appears to have a dominant positive effect on glutamine synthetase activity over GlnK1 inhibition. Together with the results obtained for GlnA1/GlnK1 complex formation by gel filtration and pull-down experiments in the presence of 2-oxoglutarate, these findings demonstrate that binding of 2-oxoglutarate prevents inhibition of glutamine synthetase activity by GlnK1 though GlnA1/GlnK1 complex formation per se is not affected.
GlnK1 inhibits glutamine synthetase activity present in M. mazei cells
In order to investigate, whether the GlnK1 effect can also be shown for glutamine synthetase activity of GlnA1 synthesized in the native background, the influence of GlnK1 on GlnA1 activity in cell-free extracts was studied. Cell extracts of M. mazei cells grown under nitrogen limitation were prepared in the presence or absence of 50 mM 2-oxoglutarate. Determination of glutamine synthetase activity demonstrated that an inhibitory effect of additional purified His6-GlnK1 was only obtained, when the cell extracts were prepared in the absence of 2-oxoglutarate. In this case glutamine synthetase activity of the nitrogen-limited cell extract was reduced from 25 mU mg−1 to 10 mU mg−1 upon the presence of GlnK1 (1.75 µM trimeric GlnK1), which is consistent with the findings observed with purified GlnA1. However, when 2-oxoglutarate was present during cell breakage, no change in enzyme activity was detectable in the presence of GlnK1, again demonstrating the positive dominant effect of 2-oxoglutarate over GlnK1 inhibition.
Characterization of GlnA1 in a GlnK1 mutant M. mazei strain
In M. mazei a second GlnK protein, GlnK2, is constitutively expressed independently of the nitrogen source (C. Ehlers and R.A. Schmitz, unpublished). In order to analyse, whether GlnK2 also affects GlnA1, we studied potential complex formation between GlnA1 and GlnK2 using a glnK1 mutant strain previously constructed (C. Ehlers and R.A. Schmitz, unpublished). Pull-down experiments with immobilized His6-GlnA1 protein using cell extracts of ammonium-shifted glnK1 mutant cells demonstrated that no additional protein was detectable in the elution fractions by Coomassie or silver staining (data not shown). This finding, which was confirmed in several independent experiments, indicates that GlnA1 appears not to interact with GlnK2. However, we cannot completely rule out that very low amounts of GlnK2 are present in the GlnA1 elution fractions, which are not detectable by silver staining.
We proposed that the archaeal PII-like protein GlnK1 of M. mazei is involved in nitrogen regulation (Ehlers et al., 2002). We now identified the first receptor protein of the archaeal GlnK protein, glutamine synthetase (GlnA1), and show that in the absence of small effector molecules (metabolites) interaction with GlnK1 directly affects enzyme activity of GlnA1, the key enzyme of ammonium assimilation.
GlnK1 inhibits GlnA1 activity by direct protein interaction
As ammonium assimilation via the glutamine synthetase/glutamate synthase (GS/GOGAT) pathway is one of the major intersections in central metabolism, the synthesis and activity of glutamine synthetase (GS) is strictly controlled by nitrogen availability in all organisms (Reitzer, 2003). Three major families of GS are described to date (Brown et al., 1994). However, until now, in methanogenic archaea exclusively glutamine synthetases of the GSI-α subdivision were identified including GlnA1 of M. mazei (Bhatnagar et al., 1986; Possot et al., 1989; Cohen-Kupiec et al., 1999; Deppenmeier et al., 2002). As they do not exhibit a potential adenylylation site nor have homologues of ATase been found in archaeal genomes, it is unlikely that the activity of those archaeal glutamine synthetases is regulated by adenylylation in response to nitrogen (Brown et al., 1994; Smith et al., 1997; Cohen-Kupiec et al., 1999; Deppenmeier et al., 2002). Thus, regulation of glutamine synthetase activity in methanogenic archaea in response to changes in nitrogen availability was not known.
We now obtained conclusive experimental evidence that the archaeal GlnK1 protein interacts and forms stable complexes with the glutamine synthetase GlnA1 in M. mazei thereby negatively affecting its enzyme activity in the absence of small effector molecules. (i) Pull-down experiments demonstrated that purified M. mazei His6-GlnK1 directly interacts with GlnA1 in cell extracts, which was confirmed by the reverse pull-down experiment with immobilized His6-GlnA1 (Figs 1 and 3). (ii) Complex analysis by gel filtration with independently purified proteins revealed that GlnK1 co-eluted with higher oligomeric GlnA1, when preincubated with the GlnA1 protein (Fig. 4). (iii) Incubation of purified GlnA1 with GlnK1 significantly reduces glutamine synthetase activity (Fig. 6A). This inhibitory effect of GlnK1 was further confirmed in cell extracts. Taking into account that GlnK1 is only synthesized under nitrogen limitation (Ehlers et al., 2002), we hypothesize that GlnK1 inhibits glutamine synthetase activity of GlnA1 by direct protein interaction in response to a shift to nitrogen sufficiency after a period of nitrogen limitation. This proposed mechanism for modulating glutamine synthetase activity in M. mazei in response to an ammonium upshift differs significantly from what is known for regulation of glutamine synthetases in bacteria. In most of the proteobacteria, after an ammonium upshift, PII-like proteins activate the ATase (GlnE), which then mediates the nitrogen signal to the glutamine synthetase by inactivating the enzyme via covalent modification (reviewed in Jiang et al., 1998c; Reitzer, 2003). The finding that in M. mazei GlnK1 interacts and effects glutamine synthetase directly, demonstrates that the small ubiquitous distributed PII-like sensory proteins evolved quite differently in various lines of descent taking over distinct functions in controlling nitrogen metabolism by protein–protein interactions with various receptor proteins, depending on the different metabolic requirements. In this respect, N-acetylglutamate kinase, the key enzyme of the arginine biosynthetic pathway, which is activated in response to nitrogen sufficiency by direct interaction with the non-modified PII protein in Synechococcus elongatus (Heinrich et al., 2004), is another recent example for a newly identified receiver protein of a PII-like protein.
Glutamine synthetase activity of M. mazei GlnA1 is directly affected by 2-oxoglutarate
Analysing glutamine synthetase activity of purified heterologously expressed GlnA1 demonstrated that purified M. mazei GlnA1 showed a low specific activity (approximately 0.4–0.6 U mg−1) compared to other archaeal dodecameric glutamine synthetases belonging to the GSI-α subdivision (Bhatnagar et al., 1986; Brown et al., 1994; Cohen-Kupiec et al., 1999). This might be due to the fact that GlnA1 is not correctly folded or assembled to its active dodecameric conformation, when expressed in an E. coli background. In addition, the purified active archaeal glutamine synthetase appeared not to be very stable, as over a period of 12 h incubation on ice the enzyme activity decreased dramatically, potentially based on the dissociation in lower oligomeric structures. This is further supported by gel filtration analysis of purified GlnA1, which demonstrated that the archaeal glutamine synthetase eluted in an oligomeric structure, which apparently consisted of six monomers (Fig. 4B). Most interestingly, we obtained conclusive evidence, that glutamine synthetase activity of purified GlnA1 was highly dependent on the presence of 2-oxoglutarate, which is known to be the major indicator of nitrogen status in cyanobacteria (Irmler et al., 1997; Forchhammer, 2004). The specific activity of GlnA1 increased up to 16-fold, when 2-oxoglutarate was present in the test assay at a final concentration of 2.5 mM. The minimal concentration, which showed an effect on the glutamine synthetase activity was determined to be 250 µM 2-oxoglutarate (Fig. 6B). This finding is consistent with the expected physiological concentration range of 2-oxoglutarate at different nitrogen availabilities, as internal 2-oxoglutarate concentrations of c. 100 µM under nitrogen sufficiency and of approximately 1.0 mM under nitrogen limitation have been determined for E. coli (Senior, 1975). A positive effect of 2-oxoglutarate was also observed for glutamine synthetase present in cell extracts of M. mazei cells grown under nitrogen limitation, confirming the regulatory role of the internal 2-oxoglutarate pool, which is highly diluted upon cell disruption. The activation of glutamine synthetase activity by 2-oxoglutarate in M. mazei is to our knowledge the first example for 2-oxoglutarate directly modulating glutamine synthetase activity. Taking together these findings indicate that M. mazei perceives external nitrogen limitation by changes in the internal 2-oxoglutarate pool, which is further supported by the finding that the TCA cycle is incomplete and no gene encoding for a 2-oxoglutarate dehydrogenase has been identified in M. mazei (Deppenmeier et al., 2002). This is in contrast to several proteobacteria and Bacillus subtilis, for which it has been shown that external nitrogen limitation is perceived as internal glutamine limitation (Ikeda et al., 1996; Hu et al., 1999; Yakunin et al., 1999; Schmitz, 2000). To date, a central role of 2-oxoglutarate for the perception of changes in nitrogen availabilities has been exclusively demonstrated for cyanobacteria, e.g. Synechococcus elongatus and Synechocystis sp. PCC6803 (Forchhammer, 1999; Muro-Pastor et al., 2001). In these organisms, the TCA cycle is also incomplete and synthesis of 2-oxoglutarate, as in methanogens, has merely anabolic functions as precursor for the synthesis of glutamate via the GS/GOGAT pathway. It is thus tempting to speculate that the similar metabolic organization with respect to nitrogen assimilation has led to similar regulatory solutions. Further it should be kept in mind that M. mazei has been shown to contain a high number of bacteria-like genes closely related to genes of cyanobacteria apparently acquired by lateral gene transfer (Deppenmeier et al., 2002).
2-oxoglutarate antagonizes inhibitory effects of GlnK1 on GlnA1 activity
Since 2-oxoglutarate directly affected glutamine synthetase activity it is very likely that binding of 2-oxoglutarate to GlnA1 induces a conformational change in the protein structure, which stimulates maximal glutamine synthetase activity. The finding that 2-oxoglutarate antagonizes the inhibitory effect of GlnK1 but did not prevent complex formation between GlnA1 and GlnK1, further indicates that this induced GlnA1 protein structure in the presence of 2-oxoglutarate (GlnA1*) prevents inhibition by GlnK1 although the complex formation per se seems not to be affected. Gel filtration analysis in the presence or absence of 2-oxoglutarate, however, indicated that complex stability appears to be reduced in the presence of 2-oxoglutarate (Fig. 4C and D), which was confirmed using several independent protein preparations. This may point out that GlnA1* has a lower binding affinity to GlnK1 or its induced structure results in a different complex composition, which has to be confirmed using more sensitive methods for quantification of binding affinities, e.g. by the method of surface plasmon resonance (SRP) spectroscopy. Whether 2-oxoglutarate binds in addition to GlnK1, as suggested by 2-oxoglutarate binding to its bacterial PII homologues, remains to be investigated.
Regulation of a glutamine synthetase activity by 2-oxoglutarate and a PII-like protein is novel. The only other example, in which glutamine synthetase activity is regulated by direct protein–protein interaction is known from a cyanobacterium. Glutamine synthetase activity in Synechocystis sp. PC6803 is inhibited upon increasing ammonium concentrations by protein–protein interaction with two small inactivating factors (IF3 and IF7) (Garcia-Dominguez et al., 1999; 2000). The IF encoding genes gifA and gifB are only transcribed under nitrogen excess conditions and are repressed under nitrogen-limiting conditions by the global transcription factor NtcA, which itself responds to 2-oxoglutarate (Vazquez-Bermudez et al., 2002).
Hypothetical model for post-translational regulation of glutamine synthetase in response to nitrogen availability in M. mazei
On the basis of our findings we propose the following working model (Fig. 8): M. mazei perceives external nitrogen limitation by sensing the internal 2-oxoglutarate pool, which increases because of reduced consumption by the ammonium-dependent GS/GOGAT way. Under those conditions, binding of 2-oxoglutarate induces a conformational change (GlnA1*), which enhances GlnA1 activity and simultaneously prevents inhibition by GlnK1, although complex formation with GlnK1 can occur potentially with reduced stability or an altered composition. After a shift to nitrogen sufficiency, however, the internal 2-oxoglutarate level decreases resulting in the not-activated state of GlnA1 (GlnA1), which forms more stabile (compact) complexes with GlnK1. Therefore, glutamine synthetase activity is reduced by diminished direct activation and the remaining activity is further inhibited by GlnK1 accumulated during nitrogen limitation. Because of the dodecameric structure of GlnA1 and the trimeric structure of GlnK1, it is tempting to speculate that one GlnK1 trimer interacts with one hexameric ring of GlnA1. Thus, the regulation by GlnK1 in addition to the dominant effector 2-oxoglutarate allows finetuning of glutamine synthetase activity under changing nitrogen availabilities after a long period of nitrogen limitation. At the current experimental status we do not know, whether GlnK1 additionally is controlled by the 2-oxoglutarate levels. When cells grow under nitrogen excess conditions for prolonged periods of time, GlnK1 levels strongly decrease and glutamine synthetase may solely be regulated by 2-oxoglutarate. Thus, the regulatory network in M. mazei allows for an efficient glutamine synthetase downregulation in cells that previously accumulated GlnK1 because of ammonium-limited conditions.
Strains and plasmids
Strains and plasmids used in this study are listed in Table 1. Plasmid DNA was transformed into E. coli according to the method of Inoue (Inoue et al., 1990).
Table 1. Strains and plasmids used in this study.
Strain or plasmid
Genotype or description
Source or reference
Methanosarcina mazei strain Gö1
DSM No. 3647
M. mazei ΔglnK
M. mazei, but glnK1::pac
C. Ehlers and R.A. Schmitz, unpublished
Escherichia coli DH5α BL21-CodonPlus®-RIL
General cloning strain, containing the pRIL plasmid (ileW, leuY, proL)
Stratagene, La Jolla, USA
General cloning vector, providing an N-terminal His6-tag
M. mazei glnA1 cloned into pET28a under the control of the T7 promoter coding for His6-GlnA1
M. mazei glnK1 cloned into pET28a under the control of the T7 promoter coding for His6-GlnK1
Construction of plasmids
Plasmid pRS203 was constructed as follows. M. mazei glnK1 gene was amplified by PCR using chromosomal M. mazei DNA and primers homologous to the glnK1 flanking 5′ and 3′ regions with additional synthetic restriction recognition sites for NdeI and HindIII respectively (underlined): glnKNde1 primer (5′-GTGGTCCATATGAAATACGTAATTGCAATG-3′ and glnKHindIII primer (5′-CGTTTTTACGCTGGACAAG CTTTCC-3′). The 472 bp PCR fragment was cloned into the NdeI and HindIII sites of the expression vector pET28a (Novagen), fusing six histidine codons in front of the glnK1 start codon (N-terminal His-tag). In order to construct plasmid pRS196 an approximately 1.3 kbp DNA fragment containing the M. mazei glnA1 gene was amplified by PCR using the glnA1 forward primer with an additional NdeI restriction site (5′-GGATGGAATCATATGGTGCAGATG-3′) and glnA1 reverse primer (5′-CTGGAGCGGATCCTTCCG-3′) with an additional BamH1 site. The obtained PCR product was cloned into the expression vector pET28a restricted with NdeI and BamHI fusing six histidine codons in front of the glnA1 start codon (N-terminal His-tag). The correct insertion and sequence of both genes were confirmed by DNA sequencing of both strands.
Methanosarcina mazei strains were routinely grown at 37°C under anaerobic conditions under a N2/CO2 (80:20) atmosphere in minimal medium supplemented with 150 mM methanol and 40 mM acetate as described (Deppenmeier et al., 1990; Ehlers et al., 2002). For nitrogen-limited growth, the ammonium was omitted from the media and molecular nitrogen in the gas phase served as sole nitrogen source. For ammonium upshift experiments, cells were initially grown under nitrogen limitation until they reached a turbidity of 0.35–0.4 at 600 nm (exponential growth phase) and were then supplemented with 15 mM ammonium followed by further incubation at 37°C for 30 min. Medium for growth of the M. mazei glnK1 mutant strain was in generally supplemented with 2.5 µg ml−1 puromycin (C. Ehlers and R.A. Schmitz, unpublished).
For heterologous expression and purification of M. mazei His6-GlnK1 and His6-GlnA1 the respective plasmids pRS203 (pET28a/glnK1) and pRS196 (pET28a/glnA1) were transformed into E. coli BL21-CodonPlus®-RIL (Stratagene, La Jolla), which provides additional tRNAs rare for E. coli. One litre of culture was grown aerobically in Luria–Bertani (LB) medium at 37°C and expression of His6-GlnK1 and His6-GlnA1 was induced with 100 µM and 1 µM IPTG respectively, when cells reached a turbidity of c. 0.6 at 600 nm. After 2 h induction at 37°C the cells were harvested and cell extracts were prepared by disruption in buffer A (50 mM NaH2PO4, 300 mM NaCl, pH 8.0) supplemented with the protease inhibitor cocktail for bacterial cell extracts (Sigma) using a French pressure cell followed by centrifugation at 20 000 g for 30 min. M. mazei His6-GlnK1 and His6-GlnA1 proteins were purified from the respective supernatant by Ni-affinity chromatography using Ni-NTA agarose (Qiagen) according to the manufacture's instructions. His6-GlnK1 was eluted from Ni-NTA agarose in the presence of 250 mM imidazole, dialysed into 50 mM Tris/HCl (pH 8.0) and stored at −70°C. Overexpressed His6-GlnA1 mainly accumulated in insoluble inclusion bodies, thus the expression conditions were varied and optimized to achieve maximal solubility of His6-GlnA1 and purification was performed from the remaining amounts of soluble GlnA1 protein present in the supernatant. Prior to affinity chromatography a 30% ammonium sulphate precipitation was performed leaving His6-GlnA1 in the soluble fraction, which was applied to the Ni-NTA agarose after dialysis. His6-GlnA1 protein was finally eluted in the presence of 100 mM imidazole, which was subsequently removed by dialysis into 50 mM Tris-HCl (pH 8.0). Samples of each purification step were analysed by 12.5% SDS-PAGE according to Laemmli (Laemmli, 1970) and protein concentrations were determined via the method of Bradford (Bradford, 1976) with the Bio-Rad protein assay using bovine serum albumin as standard.
Complex analysis by affinity cochromatography
One litre of culture of M. mazei wild-type and glnK1 mutant strain was grown under nitrogen limitation (N2) to a turbidity of 0.3 at 600 nm. If necessary, cultures were shifted to growth in the presence of 10 mM ammonium for 30 min prior harvest. After harvesting, the respective cells were resuspended in 50 mM Tris/HCl (pH 6.9) and cell extracts were prepared in the presence of the protease inhibitor cocktail for bacterial cell extracts (Sigma) using a French pressure cell. A total of 1 mg of purified His6-GlnK1 and 0.5 mg of His6-GlnA1, respectively, were bound to 500 µl Ni-NTA agarose for 1 h at 4°C. The matrix was filled into an empty column and 30 mg of protein of crude extracts (NH4+ shift) were applied. After washing with 2× 8 ml of buffer A supplemented with 20 mM imidazole, immobilized His6-GlnK1 and His6-GlnA1 and potentially interacting proteins were eluted with 5× 100 µl in the presence of 250 mM (His6-GlnK1) and 100 mM imidazole (His6-GlnA1). Elution fractions were analysed by 12.5% SDS-PAGE and silver staining or by Western blot analysis.
Complex analysis by gel filtration
A total of 80 µg of GlnA1 (1.8 µmol monomeric GlnA1) and 40 µg of GlnK1 (2.8 µmol monomeric GlnK1) were incubated in a total volume of 50 µl for 5 min at RT before loading on an analytic Bio-Sil® Sec 250-5 column (Bio-Rad Laboratories), which was equilibrated with 50 mM NaH2PO4 buffer pH 8.0 containing 300 mM NaCl. Protein was eluted from the column using a flow rate of 1.0 ml min−1 and 0.25 ml of fractions were collected. Calibration of the column was performed using the gel filtration mass standard (Bio-Rad Laboratories) containing thyroglobulin (670 kDa), IgG (150 kDa), myoglobulin (44 kDa), ovalbumin (17 kDa) and vitamin B12 (1.35 kDa). When analysing the effect of 2-oxoglutarate and ATP on complex formation, the effector molecules were added each to a final concentration of 1 mM followed by 5 min incubation at RT prior to gel filtration analysis and the buffer system was supplemented with 1 mM of the respective effector molecule. Fractions were assayed for the presence of GlnA1 by SDS-PAGE and for the presence of GlnK1 by Western blot analysis using specific antibodies directed against GlnK1.
Western blot analysis
Purified GlnK1 protein and purified His6-GlnA1 protein were used to generate polyclonal rabbit antibodies directed against GlnK1 and GlnA1 (Goetek Göttingen, Germany). Proteins from the respective elution fractions were separated on denaturating polyacrylamide gels and transferred to nitrocellulose membranes (BioTrace®NT, Pall Life Science) (Sambrook et al., 1989). Membranes were exposed to specific polyclonal rabbit antisera directed against GlnK1 and GlnA1. Protein bands were detected with secondary antibodies directed against rabbit immunoglobulinG coupled to horseradish peroxidase (Bio-Rad Laboratories) and visualized using the ECL plus system (Amersham/Pharmacia) with a fluoroimager (Storm, Molecular Dynamics). The GlnK1 protein bands were quantified using the ImageQuant v1.2 software (Molecular Dynamics) and known amounts of purified GlnK1 protein, which was simultaneously detected and quantified with the respective fractions.
Determination of N-terminal amino acid sequences
Elution fractions of GlnK1 obtained from co-elution experiments were separated by 12.5% SDS-PAGE and transferred onto a PVDF membrane (Immobilon-P from Millipore Cooperation Eschborn, Germany). The amino acid sequences of the co-eluting proteins were determined at the Max Planck Institute for experimental medicine by Dr B. Schmidt using a Procise491 Protein Sequencer from Applied Biosystems (USA).
Determination of glutamine synthetase activity
Glutamine synthetase activity was determined by using the coupled optical test assay described by Shapiro (Shapiro and Stadtman, 1970), which couples the consumption of ATP by the conversion of ammonium and glutamate to glutamine catalysed by glutamine synthetase to the oxidation of NADH by lactate dehydrogenase. The test assay was performed as follows: 30 µl of NADH (14 mM), 60 µl of ATP (60 mM), water and sample were added to 400 µl of freshly prepared assay buffer (0.125 M MOPS pH 7.0, 0.225 M KCl, 0.125 M MgCl2, 0.075 M sodium glutamate pH 7.0, 0.125 M NH4Cl) to give a final volume of 980 µl. When the absorbance at 340 nm was constant, 10 µl of the enzyme mixture consistent of pyruvate kinase and lactate dehydrogenase (Roche, Mannheim) was added and after additional 30 s the reaction was finally started with 10 µl 0.1 M phosphenolpyruvate (PEP). The absorbance was monitored over a time-course of 600 s. Assays in the presence of different concentrations of 2-oxoglutarate and purified His6-GlnK1 were either performed with or without preincubating GlnA1 with the respective supplement for 5 min at RT before the reaction was started with PEP. When cell extracts were analysed the absorbance at 340 nm was monitored until no further decrease in absorbance was detectable before adding pyruvate kinase and lactate dehydrogenase to exclude unspecific NADH oxidation by dehydrogenases present in the cell extracts.
We thank Gerhard Gottschalk for continuous support and helpful discussions and Bernhard Schmidt for N-terminal sequence analysis. This work was supported by the Deutsche Forschungsgemeinschaft (SCHM1052/6-1 and 6-2) and by a PhD fellowship to C.E. from the Fonds der Chemischen Industrie.