The Global Nitrogen Regulator NtcA Regulates Transcription of the Signal Transducer P_II (GlnB) and Influences Its Phosphorylation Level in Response to Nitrogen and Carbon Supplies in the Cyanobacterium Synechococcus sp. Strain PCC 7942

Abstract

The P_II protein is encoded by a unique glnB gene in Synechococcus sp. strain PCC 7942. Its expression has been analyzed in the wild type and in NtcA-null mutant cells grown under different conditions of nitrogen and carbon supply. RNA-DNA hybridization experiments revealed the presence of one transcript species 680 nucleotides long, whatever the nutrient conditions tested. A second transcript species, 620 nucleotides long, absent in the NtcA null mutant, was observed in wild-type cells that were nitrogen starved for 2 h under both high and low CO₂ and in the presence of nitrate under a high CO₂ concentration. Primer extension analysis indicated that the two transcript species are generated from two tandem promoters, a ς⁷⁰ Escherichia coli-type promoter and an NtcA-dependent promoter, located 120 and 53 nucleotides, respectively, from the glnB initiation codon. The NtcA-dependent promoter is up-regulated under the conditions mentioned above, while the ς⁷⁰ E. coli-type promoter displays constitutive levels of transcripts in the NtcA null mutant and slightly different levels in the wild-type cells, depending on the nitrogen and carbon supplies. In general, a good correlation between the amounts of the two transcript species and that of the P_II protein was observed, as revealed by immunodetection with specific antibodies. The phosphorylation level of P_II in the wild type is inversely correlated with nitrogen availability and directly correlated with higher CO₂ concentration. This regulation is correspondingly less stringent in the NtcA null mutant cells. In contrast, the dephosphorylation of P_II is NtcA independent.

In cyanobacteria, nitrogen assimilation is a genuine photosynthetic process that requires ATP and reducing equivalents generated in the light. Both nitrate and nitrite are reduced to ammonium in the presence of photosynthetically reduced ferredoxin as the physiological electron donor. Ammonium is incorporated, through the glutamine synthetase-glutamate synthase pathway, into glutamate to yield glutamine by an ATP-dependent ligation reaction catalyzed by glutamine synthetase, and glutamate synthase transfers the amido group of glutamine to 2-oxoglutarate to regenerate glutamate in the presence of reduced ferredoxin. Nitrogen assimilation is tightly regulated in response to environmental cues. Nitrate and nitrite are taken up and reduced only in the absence of ammonium and under CO₂ fixation conditions, and the level of glutamine synthetase protein is severely reduced in the presence of ammonium (12).

In the unicellular Synechococcus sp. strain PCC 7942, which does not fix molecular nitrogen, the signal transduction protein P_II (the glnB gene product) is a key element in the coordination of nitrogen and carbon metabolism (15). This protein is a homotrimer of 36 kDa whose isomeric forms carry either zero, one, two, or three phosphorylated seryl residues (Ser49), depending on the carbon and nitrogen supply of the cells (13, 14), in contrast to the Escherichia coli P_II, which is uridylylated at a tyrosyl residue (Tyr51) (41). The highest degree of P_II phosphorylation is observed in cells incubated under a high CO₂ concentration in the presence of nitrate or under nitrogen-limiting conditions, while the protein is mainly dephosphorylated under low CO₂ in the presence of ammonium (14, 15). In vitro phosphorylation experiments revealed that both the P_II kinase and phosphatase activities depend on 2-oxoglutarate and ATP but not on glutamine or glutamate (16, 23). The modification of the Synechococcus sp. strain PCC 7942 P_II protein is facilitated by the binding of ATP and 2-oxoglutarate (13, 16). Since the intracellular concentration of ATP is high under physiological conditions, it has been proposed that P_II primarily functions as a sensor of 2-oxoglutarate (13). This metabolite not only serves as a source of carbon skeleton for nitrogen assimilation but would also be of particular importance as a small-molecule effector in the control of this metabolic process in Synechococcus sp. strain PCC 7942.

NtcA is a global nitrogen regulator that is widespread and highly conserved in cyanobacteria (12, 20). This DNA-binding protein, which belongs to the CRP family of bacterial transcriptional effectors, activates the expression of a number of genes in the absence of ammonium by recognizing the target consensus nucleotide sequence GTAN₈TAC in their promoter regions (30). In Synechococcus sp. strain PCC 7942, NtcA positively regulates its own expression and activates the transcription of the nir operon (encoding nitrite reductase, the ABC-type permease complex, and nitrate reductase) (12, 30, 39), the nirB ntcB gene cluster (which encodes, respectively, a protein required for expression of nitrite reductase activity and a transcriptional effector of the bacterial LysR-type family that activates the nir operon in the presence of nitrite) (1, 26, 44), the glnA gene (encoding glutamine synthetase) (6, 7, 30), and the cynBDS operon (encoding two proteins likely to be involved in the active transport of cyanate and cyanase) (22). In the N₂-fixing filamentous heterocystous cyanobacterium Anabaena sp. strain PCC 7120, NtcA acts as an activator for the expression of genes for the assimilation of nitrogen sources alternative to ammonium and as an activator for heterocyst development (18, 19, 49). It has also been proposed that it could behave as a repressor for the rbcL gene (encoding the large ribulose-1,5-bisphosphate carboxylase–oxygenase subunit) and the gor gene (encoding glutathione reductase) (24, 40). In the facultative photoheterotroph Synechocystis sp. strain PCC 6803, glnB appears to be regulated by both electron transport and nitrogen availability, and the increased level of glnB under nitrogen starvation might be under the control of NtcA (21). In the marine unicellular N₂-fixing Cyanothece sp. strain BH68K, NtcA is involved in nitrogen assimilation rather than nitrogen fixation, and the expression of the ntcA gene may be under the control of the circadian rhythm (4).

Analysis of the phenotype of MP2, a P_II null mutant of the obligate photoautotroph Synechococcus sp. strain PCC 7942, revealed that nitrate utilization no longer depended on CO₂ fixation (15). Moreover, in contrast to the wild-type cells, in which ammonium exerts a rapid and reversible inhibition of nitrate and nitrite uptake, no inhibition was observed in this mutant. It was thus concluded that the unphosphorylated form of P_II is involved in the short-term inhibition by ammonium of nitrate and nitrite uptake (27). In this mutant, the synthesis of nitrate and nitrite reductases and glutamine synthetase was still subject to control by ammonium, suggesting that there is no direct interaction between P_II and the activity of NtcA in the regulation of nitrogen assimilation (15). However, other relationships between P_II and NtcA were not excluded.

Here we present results demonstrating that NtcA regulates P_II synthesis at the transcriptional level and is required for a full control of the phosphorylation state of P_II in response to nitrogen and carbon availability in Synechococcus sp. strain PCC 7942.

MATERIALS AND METHODS

Strains and culture conditions.

Wild-type and mutant cells of Synechococcus sp. strain PCC 7942 were grown in liquid BG11₀ medium (43) containing 0.4 mM Na₂CO₃ and supplemented with 10 mM HEPES, pH 8.0. Either NaNO₃ (17.6 mM) or NH₄Cl (5 mM) was used as the N source. The NtcA⁻ mutant, constructed by using plasmid pMAV58 according to the method of Vega-Palas et al. (48), was grown in ammonium-containing BG11₀ medium with chloramphenicol (7 μg ml⁻¹). Precultures grown in the presence of ammonium were incubated for 3 to 4 days (optical density at 750 nm [OD₇₅₀], approximately 0.6) at 30°C in air and illuminated with fluorescent lamps (OSRAM L18W/25 universal white) providing a photosynthetic photon flux density of 50 μmol m⁻² s⁻¹ measured with a LI-COR LI-185B quantum radiometer-photometer equipped with an LI-190SB quantum sensor. Experimental cultures were incubated at 35°C under the same photosynthetic photon flux density in a culture medium containing ammonium and supplemented with NaHCO₃ (10 mM) and with constant bubbling with air–3% (vol/vol) CO₂. Cells from mid-exponential-phase cultures (OD₇₅₀, approximately 0.4) were collected by centrifugation at 5,000 × g for 10 min at 25°C. The cell pellets were washed twice with BG11₀ and resuspended at the same cell density in BG11₀ medium containing either ammonium, nitrate, or no nitrogen source. After 2 h of incubation of the cells either in air without bubbling or bubbled with air–3% (vol/vol) CO₂, samples were collected for analysis.

Nucleic acid methods.

Standard methods were used for E. coli plasmid DNA isolation. Restriction endonucleases (New England Biolabs or Pharmacia) and other DNA-modifying enzymes (New England Biolabs or Amersham) were used according to the manufacturers’ recommendations.

Extraction of total RNA, gel electrophoreses, blottings, and hybridizations were performed as described previously (27). DNA probes were labelled with [α-³²P]dATP (110 TBq mmol⁻¹) by using a Megaprime random-labelling kit (Amersham). A probe internal to glnB (241 bp) was obtained by PCR amplification of the corresponding fragment from plasmid pPM119 (45) with specific primers to which an EcoRI site was added at the 5′ extremity of the coding strand (5′ CGGAATTCCGTTCAAACTGGAC 3′) and a BamHI site was added at the 5′ extremity of the complementary strand (5′ CGGGATCCCGTCACCAATTTCGC 3′). The PCR product was cloned in the vector pTZ18R for further use in the DNA-binding assay. Plasmid DNA was extracted with the Nucleobond AX kit (Macherey-Nagel, Düren, Germany). A 0.6-kb KpnI-XhoI fragment containing the rnpB gene, encoding the RNA subunit of RNase P from Synechococcus sp. strain PCC 7942 (a gift from A. Vioque [3]), was used as a probe to quantify the amount of RNA loaded and transferred to the filters and to standardize the measurements.

Primer extension was performed as described by Liotenberg et al. (28) with 60 μg of total RNA and the 21-nucleotide-long primer 5′ GACTTCGTCCAGTTTGAACGG 3′. DNA sequencing was performed by using the sequencing dideoxynucleotide chain termination method (T7 sequencing kit; Pharmacia) with ³⁵S-dATP (37 TBq mmol⁻¹; Amersham) as the labelled nucleotide and the same primer mentioned above.

The relative transcript levels were quantified by scanning photoactivatable screens on a Molecular Dynamics 445SI PhosphoImager. All quantifications, data display, and analysis were performed with Molecular Dynamics Image Quant software.

Gel retardation assays.

Preparation of cell extracts from E. coli DH5α containing the expression vector pTrc99A and from isopropyl-β-d-thiogalactopyranoside (IPTG)-induced cells of the NtcA-overproducing strain DH5α(pCSI26), gel retardation assays, and labelling of the probes were performed as described previously (30). The DNA probes were a 440-bp NheI-AvaI fragment of the glnB promoter region from plasmid pPM119; the PCR product corresponding to a fragment internal to the glnB gene, obtained as described above in “Nucleic acid methods”; and a 350-bp EcoRI-XhoI fragment of the glnA promoter region from pCSI38 (30). These probes were labelled with [α-³²P]dCTP (110 TBq/mmol; Amersham). The protein-DNA complex was visualized with an Instant Imager (Packard).

Quantification of the P_II protein and determination of its modification state.

Cell extracts were prepared from cultures grown to an OD₇₅₀ of 0.4, and aliquots corresponding to 10 μg of total protein were separated by polyacrylamide gel electrophoresis under either denaturing or nondenaturing conditions as described by Forchhammer and Tandeau de Marsac (14). The protein content of cell extracts was estimated by the method of Lowry modified as described previously (32), with bovine serum albumin as a standard. The P_II protein was revealed by immunoblotting with a P_II-specific antiserum in an enhanced chemoluminescence detection system (ECL kit; Amersham). Quantification was done with the National Institutes of Health Image program.

RESULTS

Control of glnB expression by NtcA.

Examination of the nucleotide sequence of the glnB gene of Synechococcus sp. strain PCC 7942 revealed the presence of a consensus DNA-binding site for the transcriptional effector NtcA (Fig. 1) located upstream of the glnB initiation codon between nucleotides 88 and 102. Experiments were designed to analyze the conditions under which the expression of the glnB gene of Synechococcus sp. strain PCC 7942 was regulated and to determine the corresponding start sites for transcription. RNA-DNA hybridizations, using a probe internal to the glnB gene, and primer extension analysis were performed. Total RNA was extracted from ammonium-grown cultures of the wild type and the NtcA null mutant after transfer of the cells for 2 h to a medium containing either ammonium, nitrate, or no combined nitrogen and under either air or air enriched with 3% (vol/vol) CO₂. One or two transcript species of slightly different sizes (0.68 and 0.62 kb) were observed, depending on the nutrient conditions and the strain (Fig. 2). A particularly high level of transcripts was found in the wild-type cells incubated under nitrogen limitation and a high CO₂ concentration; their abundance was low under each of the other conditions tested.

FIG. 1.

Alignment of NtcA recognition sequences of different genes from Synechococcus sp. strain PCC 7942 (30).

FIG. 2.

RNA-DNA hybridization of total RNA from cells of wild-type Synechococcus sp. strain PCC 7942 (WT) and the NtcA-deficient mutant (NtcA⁻) in response to the nature of the nitrogen source and the availability of CO₂. Ammonium-grown cells were transferred for 2 h to BG-11₀ medium containing ammonium (lanes 1), nitrate (lanes 2), or no nitrogen source (lanes 3), under either air or air–3% (vol/vol) CO₂. The same RNA blots were hybridized with a DNA probe internal to the glnB gene encoding the P_II protein and with a DNA probe of the rnpB gene encoding the RNA subunit of RNase P to provide an estimate of the RNA loading.

Primer extension analysis revealed four extension products (Fig. 3). Two of them, which might correspond to the transcription start points designated tsp1 and tsp2, varied in abundance with the conditions of nitrogen and carbon availability in both the wild-type and the NtcA null mutant cells. The two additional small extension products most likely result from earlier pauses caused by a GC-rich stretch starting immediately downstream from tsp2. The transcription start point tsp1, situated 120 nucleotides from the glnB initiation codon, was found with RNA from both the wild-type and NtcA null mutant cells, whatever the conditions tested. The transcript species corresponding to tsp1 was preceded on the DNA sequence by a ς⁷⁰-like promoter (−10 TAAAAT; −35 TTGCCT). The second transcription start point, tsp2, localized 53 nucleotides from glnB, corresponded to extension products whose abundance increased under nitrogen limitation and high CO₂ concentration in the wild-type cells. This was accompanied by a decreased intensity of the extension products corresponding to tsp1. The transcription start point tsp2 was preceded on the DNA sequence by a −10 box, TACCAA, and a perfect consensus NtcA-binding sequence, GTAN₈TAC (30), between nucleotides −35 and −50. No extension products corresponding to tsp2 were detectable in the RNA from the NtcA⁻ mutant cells, whatever the conditions tested. These results indicated that NtcA controls the expression of the glnB gene at tsp2.

FIG. 3.

Primer extension with the glnB gene. Total RNA (60 μg) from wild-type Synechococcus sp. strain PCC 7942 (WT) and NtcA-deficient mutant (NtcA⁻) cells was annealed with an oligonucleotide specific to the glnB gene and extended with avian myeloblastosis virus reverse transcriptase as indicated in Materials and Methods. The cells were incubated as described in the legend to Fig. 2. Lanes A, C, G, and T contain a dideoxy sequencing ladder of the same DNA region used as a size control of the extension products. The sequences around the 5′ ends are listed on the right. tsp1 and tsp2 are putative transcription start points.

Binding of NtcA to the glnB promoter region.

Mobility shift assays of electrophoretically resolved DNA fragments carrying the upstream region of glnB were performed with cell extracts of an NtcA-overproducing E. coli strain. This strain harbors plasmid pCSI26, which carries the ntcA gene downstream from the synthetic IPTG-inducible promoter trc and thus overexpresses NtcA after IPTG induction (30). As a positive control, experiments were performed in parallel with the promoter region of the glnA gene (P_glnA). The DNA fragment containing the promoter region of the glnB gene (P_glnB) was retarded by the NtcA-containing extract from E. coli carrying pCSI26 but not by the extract from E. coli cells harboring pTrc99A, the vector used to construct pCSI26 (Fig. 4A). The DNA fragment corresponding to an internal part of the glnB gene did not display any mobility shift and did not compete with either P_glnB or P_glnA (Fig. 4A). These results confirmed that the NtcA protein might bind specifically to DNA upstream from the glnB gene. At least a 50-fold-higher concentration of the cell extract was required, however, to obtain an NtcA-promoted shift of P_glnB than to obtain a shift of P_glnA (Fig. 4B), and the retarded band was fuzzier, whatever the concentration of the cell extracts tested (Fig. 4A and B).

FIG. 4.

Gel retardation of DNA fragments from the glnB and glnA promoter regions by cell extracts of an NtcA-overproducing E. coli strain. (A) glnA promoter region, P_glnA (lanes 1 to 3), and glnB promoter region, P_glnB (lanes 4 to 6), incubated with a DNA fragment internal to the glnB gene as the competitor DNA. Lanes 1 and 4, no NtcA-containing extract added; lanes 2 and 5, 5 μg of extract from cells of E. coli DH5α(pTrc99A) added; lanes 3 and 6, 5 μg of NtcA-containing extract from IPTG-induced cells of E. coli DH5α(pCSI26) added. C_glnA and C_glnB are complexes formed after incubation of the DNA fragments carrying P_glnA and P_glnB, respectively, with the NtcA-containing extracts. (B) Same conditions as in panel A, lanes 1 and 4, with various amounts (0 [−] to 8.0 μg) of extract from IPTG-induced cells of E. coli DH5α(pCSI26) added.

Immunological detection of P_II in the wild type and in an NtcA null mutant.

The amount of P_II protein was estimated by immunoblotting with specific P_II antibodies. In general, the total amount of the protein was found to be in good correlation with the levels of the transcripts in both the wild-type and the NtcA⁻ cells incubated under the different nutrient conditions tested (Fig. 5). This indicated that, in addition to a basal level of expression of the glnB gene, which is NtcA independent, there is control by NtcA that occurs at a transcriptional level.

FIG. 5.

Immunoblot analysis of the P_II protein in cells of wild-type Synechococcus sp. strain PCC 7942 (WT) and of the NtcA-deficient mutant (NtcA⁻) in response to the nature of the nitrogen source and CO₂ availability. The cells were incubated as described in the legend to Fig. 2.

Modulation of phosphorylation levels of P_II isoforms by NtcA.

In the wild-type cells, the relative abundance of the four P_II isoforms, P_II⁰, P_II¹, P_II², and P_II³, which possess either zero, one, two or three phosphorylated monomers, respectively, varied with both the nitrogen supply and the concentration of CO₂ (Fig. 6). In the presence of ammonium, whatever the CO₂ level, or in the presence of nitrate under low CO₂, P_II was mainly unphosphorylated. Three to four isoforms were observed in the presence of nitrate under high CO₂ and in nitrogen-limiting conditions under both high and low CO₂. In the NtcA null mutant, whatever the availability of CO₂, P_II remained unphosphorylated in cells incubated in the presence of ammonium, as in the wild-type cells. In contrast, only two (under high CO₂) to three (under low CO₂) isoforms were observed when the cells were incubated with nitrate or under nitrogen limitation. As expected, the same results were obtained under both nitrogen conditions, since the NtcA-deficient mutant is impaired in the expression of a number of genes involved in nitrogen assimilation (48). Addition of ammonium to the nitrogen-starved cells led to a complete dephosphorylation of P_II in the NtcA null mutant, as in the wild type. Therefore, NtcA is not required for the dephosphorylation of P_II but only for the accumulation of phosphorylated forms in response to both the nitrogen and carbon sources.

FIG. 6.

In vivo phosphorylation of the P_II protein in cells of wild-type Synechococcus sp. strain PCC 7942 (WT) and the NtcA-deficient mutant (NtcA⁻) in response to the nature of the nitrogen source and CO₂ availability. Lanes 1 to 3, the cells were incubated as described in the legend to Fig. 2. Lanes 4, cells starved for nitrogen as in lanes 3 were further incubated for 2 h following addition of ammonium to the culture medium. P_II⁰, P_II¹, P_II², and P_II³ correspond to the four isoforms of the trimeric P_II protein containing zero, one, two, or three phosphorylated monomers.

DISCUSSION

In this study, a role for the global nitrogen regulator NtcA in the control of the transcription of the glnB gene, which encodes the signal transduction P_II protein, in response to changes in the availability of nitrogen and carbon in the environment has been demonstrated in the unicellular cyanobacterium Synechococcus sp. strain PCC 7942.

At least two distinct promoter-like structures have been identified that control the expression of the unique glnB gene in Synechococcus sp. strain PCC 7942, depending on the nitrogen supply and CO₂ concentration. The first, located upstream from tsp1, is of the common ς⁷⁰ E. coli type. It yields constitutive levels of glnB mRNA in both wild-type and NtcA null mutant cells. The second, located upstream from tsp2, is activated by the global nitrogen regulator NtcA. The putative NtcA-binding site from the Synechococcus glnB promoter exactly matches the consensus sequence (GTAN₈TAC), as is the case with other genes from this organism, such as glnA, ntcA, and the nir operon (Fig. 1). However, there is obviously a different binding affinity for the glnB and the glnA promoter regions (Fig. 4). Previous in vitro studies have shown that NtcA from the filamentous heterocystous cyanobacterium Anabaena sp. strain PCC 7120 displays different affinities, depending on the target genes (40). A similar type of control could occur in Synechococcus sp. strain PCC 7942.

The abundance of the NtcA-dependent transcripts is inversely correlated with the availability of nitrogen and increases with higher CO₂ concentrations (Fig. 3). This positive effect of CO₂ could be exerted by CO₂ itself or by a metabolite derived from its fixation. A corresponding decrease in the level of the transcripts that initiate at tsp1 is particularly evident in cells incubated with high CO₂. This would result if the binding of NtcA to the tsp2 transcript species prevents the RNA polymerase from initiating transcription from the more distal tsp1 promoter region, and it is not due to a specific recognition of the tsp1 promoter region by NtcA, since no NtcA consensus sequence is present in that region.

In Synechocystis sp. strain PCC 6803, nitrogen availability and electron transport have been shown to control the expression of the glnA and glnB genes (21). In this cyanobacterium, one putative promoter of the glnB gene is constitutively expressed at a very low level in the absence of nitrogen and in the presence of ammonium and/or nitrate. The other promoter is functional only in cells starved for nitrogen (21). Therefore, the glnB control system in Synechocystis sp. strain PCC 6803, as well as those of other genes (e.g., icd, encoding NADP⁺-isocitrate dehydrogenase, and glnA and glnN, encoding glutamine synthetases I and III) (38, 42), differs from the one that operates in Synechococcus sp. strain PCC 7942 for glnA (6, 30) and glnB (Fig. 3), in which the inducible promoter is active in cells grown both in the absence of nitrogen and in the presence of nitrate. Moreover, the physical organizations of the promoter regions in these two strains differ. In contrast to Synechocystis sp. strain PCC 6803, in Synechococcus sp. strain PCC 7942 the two promoter sequences do not overlap and the constitutively expressed promoter is upstream of the nitrogen-regulated one (Fig. 7).

FIG. 7.

Comparison of the physical organization of the promoter regions of the glnB genes from Synechococcus sp. strain PCC 7942 and Synechocystis sp. strain PCC 6803.

Two types of promoters for the genes encoding P_II and P_II-like proteins that recognize either ς⁷⁰ or a nitrogen-regulated RNA polymerase are commonly found in bacteria. Their physical organization and their levels of expression may vary, depending on the organism, but in most cases nitrogen control is exerted from a ς⁵⁴-dependent promoter through the NtrC protein (2, 5, 9–11, 17, 25, 29, 34, 36, 37, 46, 47, 50, 51).

In this study, we did not look at the kinetics of phosphorylation and dephosphorylation of P_II in the wild-type and mutant strains in response to nitrogen shifts. Therefore, we do not know whether the accumulation of the phosphorylated forms of P_II depends on a kinase or a phosphatase activity. Nevertheless, the dephosphorylation of P_II in the presence of ammonium appears to be NtcA independent, while the accumulation of the phosphorylated forms is at least in part NtcA dependent in Synechococcus sp. strain PCC 7942 (Fig. 6, compare lanes 3). The transcriptional activator NtcA might therefore modulate the phosphorylation state of P_II in response to the intracellular N-C balance by directly or indirectly controlling the synthesis, activity, and/or stability of the serine kinase-phosphatase enzyme system, which posttranslationally modifies P_II.

At present, in Synechococcus sp. strain PCC 7942, we do not know whether NtcA is a single protein or a complex of one or more proteins or if it needs to be liganded to some effector(s) or posttranslationally modified to be active or to modulate its activity, depending on environmental conditions. The effect of NtcA on the activation of transcription of the glnB gene in Synechococcus sp. strain PCC 7942 and on the increased phosphorylation state of the protein could result from a higher level of expression of the corresponding ntcA gene and/or from a greater affinity of this regulator for its target sites under a high CO₂ concentration. Whether this is due to an increased electron transport activity or to the presence of a specific CO₂ fixation product(s) remains to be elucidated. The fact that the concentration of 2-oxoglutarate varies greatly in cyanobacterial cells, depending on the supply of carbon and nitrogen (8, 33, 35, 38), and that this is a very important metabolite involved in the regulation of nitrogen assimilation pathways in cyanobacteria and other prokaryotes (15, 35, 31) favors the hypothesis that this compound could play a key role in the regulatory system that coordinates the corresponding metabolic processes via P_II and NtcA.