Synthesis of nitrate reductase in the unicellular cyanobacterium Synechococcus sp. strain PCC 7942 took place at a slow rate when the cells were incubated without a supply of inorganic carbon, but addition to these cells of CO2/bicarbonate or, in a Synechococcus strain transformed with a gene encoding a 2-oxoglutarate permease, 2-oxoglutarate stimulated expression of the enzyme. Induction by 2-oxoglutarate was also observed at the mRNA level for two nitrogen-regulated genes, nir and amt1, but not for the photosystem II D1 protein-encoding gene psbA1. Our results are consistent with a role of 2-oxoglutarate in nitrogen control in cyanobacteria.
Nitrogen control, a phenomenon of wide occurrence among microorganisms, consists of repression of some nitrogen assimilation genes when a preferred nitrogen source like ammonium is available to the cells. The best characterized nitrogen control system in bacteria is the NtrB–NtrC two-component regulatory system of the enterobacteria, in which phosphorylated NtrC is an activator of transcription from σ54-dependent promoters and NtrB is a kinase/phosphatase of NtrC whose activity is modulated by the PII (glnB gene product) protein . The action of PII is itself regulated by uridylylation catalyzed by the uridylyl transferase/uridylyl removing enzyme. The activity of this complex regulatory system responds to the nitrogen and carbon/energy status of the cell, which is signalled by glutamine, 2-oxoglutarate and ATP [2,3]. The Ntr system is however not universal in bacteria. Among the few other systems for nitrogen control that have recently been discovered is the one found in cyanobacteria. In these organisms, which carry out oxygenic photosynthesis and fix CO2 through the Calvin cycle, the ntcA gene encoding a transcriptional regulator of the CAP family has been identified as a key element for nitrogen control [4,5].
In the unicellular cyanobacterium Synechococcus sp. strain PCC 7942, NtcA is required for activation of transcription of, among others, the nir operon and the amt1 gene . The amt1 gene encodes an ammonium/methylammonium permease that is required for uptake of ammonium when it is present at low concentrations in the extracellular medium and for recapture of ammonium leaked out from the cells . The nir operon consists of the following genes: nir encoding nitrite reductase, nrtABCD encoding an ABC-type uptake permease for nitrate and nitrite, and narB encoding nitrate reductase [4,7–9]. Nitrate reductase activity can be readily determined and represents a convenient marker for proteins subjected to nitrogen control in Synechococcus sp. . Repression by ammonium of nitrate reductase is relieved by the glutamine synthetase inhibitor l-methionine-d,l-sulfoximine (MSX ). This observation led to the conclusion that ammonium has to be incorporated into carbon skeletons to exert its repressor role in Synechococcus sp. 2-Oxoglutarate is the carbon skeleton used for nitrogen assimilation, which in cyanobacteria takes place through the glutamine synthetase–glutamate synthase cycle . In fact, in these organisms, which lack 2-oxoglutarate dehydrogenase , incorporation of nitrogen is the main metabolic role for 2-oxoglutarate . Cellular levels of 2-oxoglutarate have been shown to be high under conditions leading to expression of NtcA-dependent genes , and NtcA binding to an NtcA-dependent promoter and in vitro NtcA-dependent transcription have been shown to be stimulated by 2-oxoglutarate [14,15]. These observations suggest that NtcA-dependent gene expression is influenced by 2-oxoglutarate.
In this work, we studied the effects of inorganic carbon supply and 2-oxoglutarate on the expression of nitrate reductase activity and of nitrogen-regulated genes in Synechococcus sp. strain PCC 7942. To analyze the 2-oxoglutarate effects, a Synechococcus strain transformed with the Escherichia coli kgtP gene encoding a 2-oxoglutarate permease, strain CSF70 , was used.
2Materials and methods
2.1Growth conditions and enzyme assays
Synechococcus sp. strain PCC 7942 and its derivative strain CSF70  were grown in BG110+NH4+ medium (BG11 medium  lacking nitrate and supplemented with 2.5 mM NH4Cl and 5 mM TES-NaOH buffer (pH 7.5)) at 30°C in the light (75 μE m−2 s−1) with bubbling with a stream of air. Growth medium for strain CSF70 was supplemented with 10 μg kanamycin ml−1. Cyanobacterial cell mass was estimated by measuring the concentration of chlorophyll a (Chl) of the cultures, determined in methanolic extracts of the cells . Nitrate reductase activity was determined in cells made permeable with mixed alkyltrimethylammonium bromide using dithionite-reduced methyl viologen as the reductant . Glutamine synthetase transferase activity was determined as described . For induction experiments, the ammonium-grown cells were washed with ammonium-free medium and transferred to BG11 medium (17.5 mM NaNO3 as the nitrogen source) buffered with 10 mM TES-NaOH (pH 7.5). When indicated, the cell suspensions were bubbled with N2 or with air freed from CO2 by passing the air through a saturated KOH solution.
2.2RNA isolation and analysis
Total RNA from Synechococcus sp. was isolated as described previously  and blotted to a filter (GeneScreen Plus™, Dupont). Internal probes of amt1 (334 bp), nir (310 bp), ntcA (485 bp), and psbA1 (889 bp) were generated by PCR by standard methods , and a 0.57-kb probe of rnpB was isolated after restriction with endonucleases. Probes were labeled with a DNA labeling kit (Ready to Go, Amersham Pharmacia Biotech) and [α-32P]dCTP, and hybridization was performed at 65°C in 5×SSPE, 5×Denhardt's solution, 1% SDS, and 100 μg herring sperm DNA ml−1 (for the contents of SSPE and Denhardt's solutions, see ). Filters were washed twice at room temperature with 2×SSPE for 15 min and, if necessary, once at 65°C with 2×SSPE and 2% SDS. Radioactive areas in the filter were visualized and quantified with a Cyclone storage phosphor system (Packard).
3.1Expression of nitrate reductase activity
Transfer of ammonium-grown cells of Synechococcus sp. strain PCC 7942 to a medium containing nitrate as the nitrogen source and air levels of CO2 resulted, as expected from previously available data , in development of nitrate reductase activity (Fig. 1). However, transfer to a medium lacking a supply of inorganic carbon, either by bubbling of N2 through the cell suspension (as in Fig. 1) or by bubbling with CO2-free air (not shown), notably inhibited the development of activity. When, after a 90-min incubation, the CO2-free cell suspension was supplemented with air or with 10 mM NaHCO3, development of nitrate reductase was observed (Fig. 1). Addition of 100 μg rifampicin ml−1 simultaneously to the bicarbonate abolished the bicarbonate-stimulated increase of activity (not shown). These results indicated that availability of CO2 is required for nitrate reductase development at a stage as early as transcription initiation and are consistent with the reported dependence of the nitrate reductase activity levels on the level of CO2 supplied for growth .
To test whether 2-oxoglutarate could reproduce the stimulatory effect of inorganic carbon on nitrate reductase synthesis, the Synechococcus sp. strain CSF70 carrying the kgtP gene was used. Strain CSF70 effectively takes up 2-oxoglutarate, which is accumulated within the cells and metabolized, mainly to glutamate and glutamine . In CSF70 cells treated with MSX, production of glutamine is abolished and 83–89% of the label from 2-[U-14C]oxoglutarate accumulates inside the cells as glutamate and 2-oxoglutarate itself . We therefore tested the effect of 2-oxoglutarate on nitrate reductase expression in ammonium-grown cells transferred to BG11 medium supplemented with 10 μM MSX and bubbled with N2. In contrast to the high induction that takes place in MSX-treated cells incubated with high levels of CO2, only a limited induction of nitrate reductase was observed for the MSX-treated cells incubated in the absence of inorganic carbon (Fig. 2A). Addition of 10 mM NaHCO3 to these cells stimulated nitrate reductase synthesis (Fig. 2A) and, in the CSF70 cells, addition of 1 mM 2-oxoglutarate had an effect similar to that of addition of 10 mM NaHCO3 (Fig. 2B). The 2-oxoglutarate-dependent increase in nitrate reductase expression was 2.5±0.5 (mean and standard deviation of four independent experiments). The 2-oxoglutarate effect was however not observed for strain PCC 7942 (Fig. 2A) indicating that KgtP-mediated 2-oxoglutarate uptake was required for stimulation of nitrate reductase synthesis. (Different levels of nitrate reductase activity in the wild-type and CSF70 strains were consistently observed, but we did not investigate the reason for this difference in these two independent bacterial clones.)
Addition of 1 mM l-glutamate instead of 2-oxoglutarate did not result in any increase of nitrate reductase expression (Fig. 2). To ensure that glutamate was incorporated into the cells, uptake of [14C]glutamate was measured and compared to that of 2-[U-14C]oxoglutarate (metabolite uptake was determined as previously described [12,24]). Uptake of 1 mM [14C]glutamate was about 8 nmol min−1 (mg of Chl)−1 for both strains, PCC 7942 and CSF70, whereas uptake of 1 mM 2-[U-14C]oxoglutarate was 2.5 and 10 nmol min−1 (mg of Chl)−1 for strains PCC 7942 and CSF70, respectively. Thus, glutamate uptake, probably mediated by the acidic amino acid permease , is comparable to that of 2-oxoglutarate in strain CSF70. It can be concluded, therefore, that in these short-term experiments of expression of nitrate reductase 2-oxoglutarate can substitute for inorganic carbon in a process that is dependent on 2-oxoglutarate uptake but independent of glutamate accumulation.
3.2Expression of nitrogen-regulated genes
The effect of 2-oxoglutarate on the expression of two nitrogen-regulated genes, amt1 and nir, in Synechococcus cells incubated without a supply of inorganic carbon was also tested. For comparison, expression of the psbA1 gene encoding a photosystem II D1 polypeptide  that appears not to be under nitrogen control in Synechococcus sp. strain PCC 7942  was also analyzed. As shown in Fig. 3, addition of 1 mM 2-oxoglutarate to MSX-treated cells incubated in the absence of CO2 resulted in an increase of the amt1 and nir transcript levels, but not in those of the psbA1 gene, specifically in strain CSF70. Normalization of the hybridization data with the data obtained from hybridization of the filters with a probe of the rnpB (ribonuclease P RNA) gene, used as an RNA loading control, rendered the figures presented in Fig. 3 for the ratio of transcript levels in the presence and absence of 2-oxoglutarate. In the kgtP-carrying CSF70 strain, this carbon compound stimulated about two-fold the expression of the nitrogen-regulated amt1 and nir genes, but not that of the psbA1 gene.
The physiological experiments described in this work show that expression of nitrate reductase after transferring cells of Synechococcus sp. strain PCC 7942 from ammonium- to nitrate-containing growth medium requires a supply of carbon. Addition of inorganic carbon, either as bicarbonate or air levels of CO2, induces the enzyme, and in Synechococcus sp. strain CSF70, which carries the KgtP 2-oxoglutarate permease, the inorganic carbon can be efficiently substituted by 2-oxoglutarate. Inhibition of nitrate reductase synthesis by rifampicin and analysis of mRNA levels of genes regulated (amt1, nir) or not regulated (psbA1) by nitrogen further indicated that the inorganic carbon or 2-oxoglutarate effect takes place at the level of transcription. The use of the NtcA-activated promoter of the glnB gene  and the expression of the amt1 gene  have also been shown to be stimulated under high-carbon conditions in Synechococcus sp. strain PCC 7942, consistent with a positive role of the carbon supply on the NtcA regulatory system. The relatively low (about 2–2.5-fold) inorganic carbon and 2-oxoglutarate effects in strain CSF70 may be due to background gene expression in the absence of added carbon, which in turn may reflect the presence of basal levels of 2-oxoglutarate in the cells. Unfortunately, attempts to isolate a Synechococcus mutant of the 2-oxoglutarate-producing enzyme, isocitrate dehydrogenase, have failed (our unpublished results).
Stimulation of nitrate reductase synthesis and gene expression by 2-oxoglutarate in strain CSF70 suggests that the positive effect of CO2 described above is exerted after CO2 fixation. Consistently, in experiments in which the Synechococcus cells were treated with d,l-glyceraldehyde, an inhibitor of CO2 fixation through the Calvin cycle effective in cyanobacteria , inorganic carbon failed to stimulate nitrate reductase synthesis (not shown). In MSX-treated cells of strain CSF70, a significant amount of label from the transported 2-[U-14C]oxoglutarate accumulates in glutamate , a result of metabolism of 2-oxoglutarate through glutamate synthase, which synthesizes two glutamate molecules from glutamine and 2-oxoglutarate. This production of glutamate from 2-oxoglutarate in the MSX-treated cells will additionally have the effect of lowering the intracellular pool of glutamine. Because glutamate cannot substitute for bicarbonate or 2-oxoglutarate in the induction of nitrate reductase, our results are consistent with 2-oxoglutarate (and/or glutamine) having a role in nitrogen-regulated gene expression in Synechococcus sp. strain PCC 7942, as is the case for other bacterial systems. These observations support the contention that 2-oxoglutarate may stimulate in vivo the activity of NtcA, as recently suggested by in vitro assays [14,15].
In addition to this putative transcriptional effect, 2-oxoglutarate has been shown to determine the phosphorylation degree of the signal transduction PII protein in Synechococcus sp. strain PCC 7942 [29,30]. In this cyanobacterium, the PII protein is essential for the regulation by ammonium of nitrate uptake , which is inhibited by the unphosphorylated PII protein or by other forms of the protein as long as 2-oxoglutarate levels are low . This carbon metabolite appears therefore to be essential for the regulation of nitrogen metabolism in cyanobacteria at different levels, such as membrane transport and gene transcription.
This work was supported by Grant BMC 2002-03902 from Ministerio de Ciencia y Tecnologı́a (Spain).