Signal Transduction Protein P_II Is Required for NtcA-Regulated Gene Expression during Nitrogen Deprivation in the Cyanobacterium Synechococcus elongatus Strain PCC 7942

Abstract

The transcription factor of the cyclic AMP receptor protein/FNR family, NtcA, and the P_II signaling protein play central roles in global nitrogen control in cyanobacteria. A dependence on P_II for NtcA-regulated transcription, however, has not been observed. In the present investigation, we examined alterations in gene expression following nitrogen deprivation in Synechococcus elongatus strain PCC 7942 and specifically the roles of NtcA and P_II. Global changes in de novo protein synthesis following combined-nitrogen deprivation were visualized by in vivo [³⁵S]methionine labeling and two-dimensional polyacrylamide gel electrophoresis analysis. Nearly all proteins whose synthesis responded specifically to combined-nitrogen deprivation in wild-type cells of S. elongatus failed to respond in P_II- and NtcA-deficient mutants. One of the proteins whose synthesis was down-regulated in a P_II- and NtcA-dependent manner was RbcS, the small subunit of RubisCO. Quantification of its mRNA revealed that the abundance of the rbcLS transcript following combined-nitrogen deprivation rapidly declined in wild-type cells but not in P_II and NtcA mutant cells. To investigate further the relationship between P_II and NtcA, fusions of the promotorless luxAB reporter genes to the NtcA-regulated glnB gene were constructed and these constructs were used to transform wild-type cells and P_II⁻ and NtcA⁻ mutants. Determination of bioluminescence under different growth conditions showed that NtcA represses gene expression in the presence of ammonium in a P_II-independent manner. By contrast, NtcA-dependent activation of glnB expression following combined-nitrogen deprivation was impaired in the absence of P_II. Together, these results suggest that under conditions of combined-nitrogen deprivation, the regulation of NtcA-dependent gene expression requires the P_II signal transduction protein.

The metabolism of cyanobacteria is based on oxygenic photosynthesis, which provides reducing equivalents and ATP to assimilate simple inorganic nutrients from the environment. All freshwater and terrestrial cyanobacteria are able to assimilate nitrogen from ammonium, nitrate, or nitrite. Ammonium, the preferred source of combined nitrogen, depresses the utilization of alternate nitrogen sources. It causes a general decrease in the abundance of nitrogen assimilatory enzymes and inhibits the activity of combined-nitrogen transport systems, a process known as global nitrogen control (reviewed in reference 14). In the absence of a combined-nitrogen source, diazotrophic strains are able to fix molecular dinitrogen, whereas nondiazotrophic strains respond by an acclimation process known as chlorosis (1, 33).

The transcription factor NtcA is a key element in global nitrogen control (14, 23). It belongs to the cyclic AMP receptor protein family of DNA-binding proteins and is required for the transcriptional activation of genes subject to ammonium repression. In addition to NtcA, the P_II signal transduction protein, known in proteobacteria to be a key regulator in nitrogen signaling (for reviews, see references 3 and 27), responds to the nitrogen status of the cells and is involved in balancing nitrogen and carbon assimilation. Previously, the cyanobacterial P_II protein was shown to be a 2-oxoglutarate sensor. 2-Oxoglutarate and ATP bind to P_II in a synergistic manner (8), stimulating the phosphorylation of Synechococcus elongatus strain PCC 7942 P_II at Ser 49 (7, 16). Conversely, the binding of ATP to P_II inhibits phospho-P_II dephosphorylation; this inhibition is greatly enhanced by 2-oxoglutarate and to a lesser extent by oxaloacetate (29). Dephosphorylation is catalyzed by a protein phosphatase of the PP2C family, which was recently identified in Synechocystis sp. strain PCC 6803 (17). The receptors of the P_II signal have yet to be identified at the molecular level, although physiological data indicate that nitrate-nitrite and bicarbonate transport systems are involved (15, 22).

Although the binding of NtcA to its DNA recognition sites is well established (14), the regulation of NtcA activity is less well understood. Despite the fact that the phosphorylation status of P_II parallels the activity of NtcA, P_II seemed not to be involved in NtcA regulation. In the presence of ammonium, a P_II-deficient mutant was still able to repress NtcA-regulated genes such as glnA, encoding glutamine synthetase I, and the nir operon, encoding components for nitrite and nitrate uptake and reduction (6, 22). Conversely, the expression of the glnB gene, encoding the P_II protein, was shown to be under NtcA control and the phosphorylation of P_II was impaired in an NtcA⁻ mutant (21, 31). These results implied that P_II regulation is subordinate to NtcA control. Recent studies showed that the binding of NtcA to the glnA promoter as well as to the promoter of its own gene, ntcA, was stimulated in the presence of 2-oxoglutarate (38). Furthermore, in vitro transcription studies revealed that the initiation of transcription from these promoters was completely dependent on the presence of 2-oxoglutarate (36). Measurements of metabolite pools in Synechocystis strain PCC 6803 showed that the intracellular 2-oxoglutarate levels corresponded well with the expression of NtcA-regulated genes (25). Therefore, the observed correlation between the phosphorylation state of P_II and NtcA activity may be due to the response of both proteins to the same regulatory metabolite, 2-oxoglutarate.

A previous study of the expression of the glnN gene (encoding a nitrogen starvation-specific glutamine synthetase of type III) in S. elongatus PCC 7942 (32) found that its induction following the depletion of combined-nitrogen sources was impaired in both NtcA- and P_II-deficient mutants. This observation contradicted the assumption that NtcA activity is autonomous from P_II signaling (see above), a conclusion that was derived from comparing the expression of NtcA-dependent genes in P_II-defective mutants incubated in the presence of nitrate to that in mutants incubated in the presence of ammonium. The expression of NtcA-dependent genes in a P_II mutant under conditions of combined-nitrogen deprivation has not been rigorously investigated. To address this issue in more detail, we initiated an analysis of global changes in de novo protein synthesis in S. elongatus wild-type cells and in P_II- and NtcA-deficient mutants following combined-nitrogen deprivation. The initial results, indicating a dependence of NtcA-regulated gene expression on the P_II signal protein, were corroborated by Northern blot analysis and fusions of the luxAB reporter to NtcA-regulated genes.

MATERIALS AND METHODS

Strains and culture conditions. (i) Cyanobacteria.

The following cyanobacterial strains were used in this study: S. elongatus strain PCC 7942 (formerly Synechococcus sp. strain PCC 7942) (13) and S. elongatus PCC 7942 mutants MP2, lacking the P_II signaling protein (glnB::aphII) (6), and MNtcA, lacking the NtcA protein (ntcA::aphII) (31). Cells were grown in liquid BG11 medium (28) at 30°C as described previously (10). When ammonium was used as a nitrogen source, the nitrate salt from BG11 was replaced by 5 mM NH₄Cl and the medium was buffered to pH 8.0 with 5 mM HEPES. Illumination was provided with standard fluorescent lamps at a photosynthetic photon flux density of 50 to 60 μmol of photons m⁻² s⁻¹, which resulted in generation times of 15 to 16 h. Nitrogen or sulfur step-down was initiated by filtration on polyvinylidene difluoride membranes and resuspension of the cells in medium deprived of a combined-nitrogen or sulfur source, as described previously (10). Cultures of MP2 and MNtcA were maintained in the presence of 25 μg of kanamycin/ml. Strains containing the luxAB reporter constructs were further supplemented with 5 μg of chloramphenicol/ml.

(ii) Escherichia coli.

E. coli strain DH5α (11) was used as the host strain for cloning experiments, and strain HB101 (4) was used for conjugal transfer of plasmids to S. elongatus. Cells were cultured according to standard procedures (30), and antibiotics were added to the following concentrations: ampicillin, 100 μg/ml, and chloramphenicol, 20 μg/ml.

Analytical and preparative 2-D PAGE.

To analyze the protein synthesis patterns by two-dimensional (2-D) polyacrylamide gel electrophoresis (PAGE), cells of wild-type S. elongatus and of strains MP2 and MNtcA were grown in BG11 medium with nitrate or ammonium as a nitrogen source. When the cultures reached an optical density at 750 nm (OD₇₅₀) of 0.5, 30-ml aliquots of the cultures were harvested and shifted to combined-nitrogen- or sulfur-deprived medium. As a control, cells were again resuspended in the nutrient-replete growth medium. After different incubation times as indicated in the text, 1.5-ml aliquots were removed into 25-ml Erlenmeyer flasks and labeled with 10 μCi of l-[³⁵S]Met (370 kBq). Incubation was continued under the same conditions as before for 2 h (this extended labeling time is required to incorporate sufficient radioactivity into the newly synthesized proteins, a consequence of the relatively slow growth rates of these cells). Then the cells were harvested, and proteins were extracted and separated by 2-D PAGE as described previously (33). The radioactive spots were visualized by exposing the dried gels to phosphorimager screens (K-Imaging Screen; Bio-Rad, Hercules, Calif.) which were recorded in a phosphorimager (Molecular Imager FX; Bio-Rad). Computer-assisted analysis of the radioactive signals was carried out with the aid of the ImageMaster 2-D Elite 2.00 software (Amersham Pharmacia Biotech, Freiburg, Germany). Preparative 2-D PAGE and identification of spots by N-terminal sequencing or by peptide mass fingerprinting were performed as described previously (33).

RNA isolation and Northern blot analysis.

Initiation of nitrogen deprivation, isolation of total RNA, separation of RNA in 1.2% (wt/vol) agarose gels containing 2.2 M formaldehyde, transfer of RNA to nylon membranes, and RNA-DNA hybridization experiments were carried out as described previously (31). The rbcL probe was a 0.71-kb DNA fragment generated by PCR amplification from S. elongatus genomic DNA by using the primer combination 5′-GATCGGTACAAAGGCAAGTG-3′ and 5′-AAGACACGGAAGTGAATCCC-3′. Detection of radioactivity was achieved by phosphorimaging (see above), and quantification was performed by using Quantity One software (Bio-Rad). The blots were rehybridized with a 16S rDNA probe (31), and the resulting signals were used as an internal standard to normalize for differences in total RNA loads in each lane.

Construction of luxAB reporter strains.

Manipulation and analysis of DNA were carried out according to standard protocols (30). The constructs used to create glnB::luxAB reporter strains were derived from the neutral site II targeting vector pAM1580 (2, 24). Plasmid pMP1A, which is identical to pMP1B (8) except that the 1.5-kb PstI fragment carrying the S. elongatus glnB gene is oriented in the opposite direction, was restricted with XhoI. The 659-bp fragment, containing the entire glnB upstream region and 183 bp downstream of the TTG initiation codon of glnB, was isolated and ligated into the XhoI site of pAM1580. Recombinant plasmids in which the glnB gene was oriented in the same direction as the luxAB reporter genes were identified by restriction analysis, resulting in plasmid pFAM1. To construct a reporter plasmid in which promoter 1 of glnB was deleted, a 360-bp fragment derived from pFAM1 was amplified by PCR. The following primers were used: glnBpr2 (5′-GGGGCTAGCGTAGACAGCGAATTTTCGATG-3′) hybridizes 24 bp upstream of the NtcA-binding site of the glnB promoter 2 and immediately downstream of the σ70-like recognition sequence of promoter 1 (21) and contains in addition a NheI restriction site; glnBpr1 (5′-GGCTCGATAAAGTCGACAGG-3′) overlaps the multiple cloning site in pFAM1 and contains a SalI restriction site. The amplification product was restricted with SalI and NheI, and the resulting fragment was ligated into SalI- and NheI-restricted pAM1580 to generate pFAM2. All constructs were checked by sequencing (dye terminator cycle sequencing reaction, ABI Prism 310 Genetic Analyzer; Applied Biosystems, Foster City, Calif.).

Transformation of S. elongatus wild-type cells with pFAM1 and pFAM2 was performed by using the natural competence of this organism (9). The construct integrates into a neutral site of the S. elongatus chromosome by homologous recombination through the neutral site II targeting sequences and confers chloramphenicol resistance on the transformants, which were selected on BG11 plates containing 5 μg of chloramphenicol/ml. The term neutral site designates a region in the S. elongatus chromosome in which an insertion causes no apparent phenotype; for details see http://www.bio.tamu.edu/users/sgolden/public/ns2.htm. The resulting transformants were termed WT-FAM1 and WT-FAM2 and were verified by PCR analysis. Transformation of MP2 and MNtcA cells with naked DNA of pFAM1 and pFAM2 did not yield chloramphenicol-resistant transformants. Therefore, pFAM1 and pFAM2 were transferred into these strains by conjugation (5) by using pRL542 (5) as a helper plasmid and pRL443 (5) as a conjugal plasmid. The transformants derived from MP2 were termed MP2-FAM1 and MP2-FAM2, and those derived from MNtcA were termed MNtcA-FAM1 and MNtcA-FAM2.

Determination of luciferase activity.

To determine the bioluminescence from the various reporter strains, 1 ml of cell culture, adjusted to an OD₇₅₀ of 0.45, was supplemented with decanal to a final concentration of 0.25 mM from a 50 mM stock solution made up in 10% (vol/vol) dimethyl sulfoxide in H₂O. Light emission by bioluminescence was recorded in a luminometer (Lumat LB9501; Berthold, Pforzheim, Germany). Bioluminescence values are given as relative light units (RLU), with 1 RLU corresponding to approximately 10 light impulses per s.

RESULTS

Differential gene expression in S. elongatus following nitrogen and sulfur deprivation.

To reveal specific alterations in protein synthesis following nitrogen step-down in S. elongatus PCC 7942, de novo protein synthesis was analyzed in cells grown in the presence of nitrate or ammonium as a nitrogen source and in cells from the same cultures after 5-h incubation in nitrogen-deprived medium. As a comparison, the same type of analysis was carried out with sulfur-deprived medium to distinguish between proteins that are specifically affected by nitrogen starvation and those that respond generally to nutrient deprivation. De novo protein synthesis was revealed by in vivo labeling of the cells with [³⁵S]methionine for 2 h, followed by 2-D PAGE analysis of the protein extracts and autoradiography on phosphorimager screens. Figure 1 shows the resulting autoradiograms for S. elongatus PCC 7942 wild-type cells labeled in the presence of ammonium-replete BG11 medium (A) and in the absence of a combined-nitrogen (B) or sulfur (C) source. Within the limits of detection for the method employed in this study, a total of 300 to 330 spots could be visualized in exponentially growing cells. In nitrogen-deprived cells, 28 spots that increased and 45 spots that decreased by a factor of at least two could be observed. Furthermore, over 50 spots appeared which could not be matched to any spot of the ammonium-replete control cells. In sulfur-starved cells, 34 spots increased and 39 decreased by a factor of at least two compared to those of the nonstarved control and more than 50 spots could not be matched. The spots could be classified into the following five categories. Spots that were induced or repressed only following nitrogen step-down (starting from either ammonium- or nitrate-supplemented medium) but did not respond to sulfur starvation were termed Nsi (nitrogen starvation induced) or Nsr (nitrogen starvation repressed). General starvation-induced spots (Gsi) increased in both combined-nitrogen- and sulfur-starved cells, and general starvation-repressed spots (Gsr) decreased under both conditions of starvation, whereas sulfur starvation-induced spots (Ssi) increased specifically under conditions of sulfur starvation. No spot was observed that was specifically repressed by sulfur starvation.

FIG. 1.

Protein synthesis patterns of S. elongatus PCC 7942 wild-type cells revealed by in vivo [³⁵S]Met labeling and 2-D PAGE analysis. Cells were grown in ammonium-replete medium until they reached the mid-exponential phase of growth (OD₇₅₀ = 0.5), then aliquots were shifted to the following media: BG11 with ammonium (as a control for nutrient-replete cells) (A), combined-nitrogen-deprived medium (B), and sulfur-deprived medium (C). After 5 h of preincubation, labeling was performed as described in Materials and Methods. Spots which could be classified into the categories Gsi, Gsr, Nsi, Nsr, and Ssi (for details, see the text) are indicated. In addition, the spots corresponding to GroEL, EF-Tu, thioredoxin M (TrxM), and plastocyanin (PC), which had been identified previously (33), are also shown. The pH of the isoelectric focusing gradient (pI) in the first dimension is indicated on the top of each gel, and the positions (in kilodaltons) of molecular size standards separated in the second dimension are indicated at the left.

To eliminate spots that responded to the different starvation conditions during only a very brief time period, labeling and 2-D PAGE analysis were also performed with cells starved for 12 and 25 h (data not shown). Data for those spots which were affected during all three time periods and which displayed reproducible alterations in triplicate labeling experiments after 5 h of starvation are shown in Tables 1 and 2, together with the quantified spot intensities (shown as percentages expressing the amount of the label in each spot relative to the total amount of label in a gel) under the given conditions. In contrast to that of most Gsi and Ssi spots, labeling of Nsi spots could not be detected in exponentially growing control cells, indicating that the synthesis of Nsi spots is strongly repressed in the presence of a combined-nitrogen source.

TABLE 1.

Relative intensities of Gsi and Gsr spots under ammonium- and sulfur-replete (Control), nitrogen-deprived (−N), or sulfur-deprived (−S) conditions^a

Spot	Control	−N	−S
Gsi1	0.06 ± 0.04	1.43 ± 0.29	0.34 ± 0.19
Gsi2	0.14 ± 0.04	0.84 ± 0.21	0.36 ± 0.21
Gsi3	0.12 ± 0.01	0.23 ± 0.09	0.24 ± 0.12
Gsi4	0.03 ± 0.02	0.30 ± 0.14	0.19 ± 0.09
Gsi5	0.18 ± 0.09	0.50 ± 0.14	0.78 ± 0.28
Gsi6	0.16 ± 0.02	0.42 ± 0.09	0.40 ± 0.08
Gsr1	11.90 ± 2.89	0.39 ± 0.49	6.4 ± 1.9
Gsr2	1.15 ± 0.05	ND	ND
Gsr3	0.93 ± 0.08	0.14 ± 0.06	0.13 ± 0.09
Gsr4	2.47 ± 0.75	0.88 ± 0.71	0.60 ± 0.07
Gsr5	1.05 ± 0.25	0.46 ± 0.21	0.61 ± 0.12

^aValues are percentages expressing the amount of the label in a particular spot relative to the total amount of label in a gel. Means ± standard deviations of results from triplicate experiments are shown. ND, not detectable.

TABLE 2.

Relative intensities of Nsi, Nsr, and Ssi spots under conditions of nitrogen or sulfur starvation and in the ammonium- and sulfur-replete control^a

Spot	Control	−N	−S
Nsi1	ND	0.47 ± 0.03
Nsi2	ND	1.14 ± 0.5
Nsi3	ND	0.91 ± 0.39
Nsi4	ND	0.75 ± 0.45
Nsi5	ND	0.53 ± 0.12
Nsi6	ND	0.31 ± 0.17
Nsr1	1.16 ± 0.30	0.47 ± 0.30
Nsr2	1.09 ± 0.40	0.40 ± 0.28
Nsr3	0.30 ± 0.03	0.10 ± 0.07
Nsr4	1.18 ± 0.16	0.69 ± 0.07
Ssi1	ND		0.6 ± 0.1
Ssi2	0.54 ± 0.06		1.06 ± 0.43
Ssi3	0.16 ± 0.04		0.98 ± 0.2

^aValues are given as explained in Table 1, footnote a. −N, nitrogen-deprived conditions; −S, sulfur-deprived conditions; ND, not detectable.

To identify the proteins corresponding to the various spots, those spots which contained a sufficient amount of protein to be visualized by Coomassie staining were cut out and the corresponding proteins were analyzed by N-terminal sequencing. Seven spots yielded N-terminal sequences of sufficient lengths for BLAST searches (Table 3). Of those, two sequences corresponded to those of known proteins from S. elongatus PCC 7942 (spot Nsi4 corresponded to the P_II protein and spot Gsr4 corresponded to the small subunit of RubisCO). Four sequences displayed significant homologies to those of proteins derived from the Synechocystis sp. strain PCC 6803 genome, whereas no protein with homology to the Gsi2 protein was found in the database. The large bulky label that we termed Gsr1 contains the phycobiliproteins CPC and APC, which do not separate clearly due to their high abundance. This was revealed by immunoblot analysis of the 2-D gels with antibodies which were raised against CPC and APC from S. elongatus (31) (data not shown). Furthermore, peptide mass fingerprints could be obtained for Nsr1, Ssi2, Ssi3, and Gsi6; however, the masses did not fit those of known proteins from S. elongatus. Therefore, the identification of these proteins must await the accessibility of the complete genome sequence of this organism.

TABLE 3.

Proteins identified from 2-D gels by N-terminal sequencing

Spot	Identical to/homologous to	N-terminal sequencea (reference)
Gsi2	No homology	AATSKTALLVQVDPFQSEVX
Gsi5	Thiol-specific antioxidant protein Sll0755	TEGALRVGQLAPDXEATA
Nsi4	PII protein	MKKIEAIIYPFKLDEVKIAL (37)
Nsi5	Hypothetical protein Sll0783	XEVTKPANQDGDFLV
Nsr2	Ribosomal protein S6 (Sll1767)	MKDFYYETMYILLADLTEEQV (33)
Gsr4	RubisCO small subunit	SMKTLPKERRFETFSY (33)
Gsr6	Inorganic pyrophosphatase Slr1622	MELSRIPAQPKPGLVN

^aHomologies to sequences from Synechocystis strain PCC 6803 (19) are underlined.

NtcA and P_II dependence of the response of protein synthesis to nitrogen starvation.

To determine which of the spots that responded to starvation were regulated by NtcA or P_II, we performed the same type of labeling and 2-D PAGE experiments as outlined above with P_II⁻ and NtcA⁻ mutants of S. elongatus PCC 7942. Cells were precultured to exponential phase in the presence of ammonium and were then shifted to the different starvation conditions. The resulting autoradiograms were matched to those of the wild type, and the spots which responded to starvation were compared (Fig. 2). All Nsi spots (Nsi1 to Nsi6) observed in the wild-type cells failed to accumulate in both the P_II- and NtcA-deficient mutants. Both mutants were also unable to decrease the amount of Nsr1 and Nsr2 (Rps6), whereas Nsr4 was unaffected. Of the spots that responded to both nitrogen starvation and sulfur starvation in the wild type, Gsi1, Gsi2, and Gsr4 (RbcS) showed different responses in the P_II⁻ and NtcA⁻ mutants, with the accumulation of Gsi1 and Gsi2 and the decrease of Gsr4 impaired; however, this was true only under the condition of nitrogen starvation. Upon sulfur starvation, these spots showed normal expression patterns in the mutants.

FIG. 2.

Comparison of spots which are differentially regulated by nitrogen starvation (−N) or sulfur starvation (−S) in S. elongatus wild-type cells (wt) and in mutants deficient in P_II (MP2) and NtcA (MNtcA). Labeling and 2-D PAGE analysis were carried out with MP2 and MNtcA cells as shown in Fig. 1 for wild-type cells. Sectors containing relevant spots were aligned to facilitate the comparison. C, control.

The repression of phycobiliprotein genes, which occurs rapidly following nitrogen step-down (20), is retarded in P_II- and NtcA-deficient mutants, as shown in a previous investigation (31). In verification of those data, we observed that the spots corresponding to the phycobiliproteins decreased more slowly in both mutants than in the wild type following combined-nitrogen deprivation (data not shown).

P_II and NtcA dependence of rbcLS expression at the level of transcript abundance.

Of the spots whose repression was impaired in NtcA- and P_II-mutant backgrounds, RbcS (Gsr4) was the only one whose gene sequence was recorded for S. elongatus (34). To reveal whether the observed repression of protein synthesis was reflected at the level of transcript abundance, we analyzed the expression of the rbcLS operon (encoding RbcL and RbcS) by Northern blot analysis. RNA was extracted from wild-type, P_II⁻, and NtcA⁻ cells of S. elongatus after they had been grown to mid-exponential phase in the presence of ammonium and from cells after they had been transferred to ammonium-depleted medium and incubated for a further 2 or 8 h. The blot was hybridized with a radioactively labeled rbcL probe, and transcript levels were evaluated by phosphorimaging. Figure 3 shows the quantified results of this experiment. In wild-type cells, the abundance of the rbcLS transcript rapidly declined following nitrogen step-down, whereas the P_II⁻ and NtcA⁻ mutants were apparently unable to repress rbcLS expression. This result suggests that the NtcA- and P_II-dependent reduction of RbcS, as detected by 2-D PAGE analysis, was based on repression at the transcript level.

FIG. 3.

Quantitative representation of a Northern blot analysis of rbcLS transcript abundance. RNA was extracted from wild-type, MP2, and MNtcA cells of S. elongatus prior to nitrogen step-down and from cells 2 and 8 h after transfer to combined-nitrogen-deprived medium. Twelve micrograms of RNA was loaded per lane, and the blot was hybridized with a radioactively labeled rbcL DNA probe. Visualization and quantification of the signals were carried out by phosphorimaging. Solid line, MNtcA cells; dashed line, MP2 cells; dotted line, wild-type cells.

Analysis of NtcA-regulated glnB::luxAB reporter fusions.

To analyze in greater detail the relationship between P_II regulation and NtcA activity with a selected NtcA-dependent promoter, fusions of the luxAB reporter (2) to the NtcA-dependent glnB gene were constructed. The glnB gene is transcribed from two start sites (21). Start site 1, located 120 bp upstream of the glnB initiation codon, is driven by a constitutive promoter (P1) that shows sequence similarities to σ70-like promoters. Downstream of start site 1 and 53 bp upstream from the initiation codon of glnB is a second start site, tsp2, which is under NtcA control and is preceded by a consensus NtcA-binding motif (23). This promoter (P2) was shown to be activated when the cells are nitrogen deprived or when they are growing with nitrate in a CO₂/HCO₃⁻-rich environment, and its regulation is dependent on NtcA (23). We constructed two fusions of the luxAB reporter genes to the glnB gene: construct pFAM1 contains the entire upstream region of glnB with both promoters and includes 183 bp of the glnB coding region, whereas pFAM2 has the sequences upstream of the NtcA-binding site deleted (Fig. 4). Therefore, in pFAM2, transcription depends only on the NtcA-activated promoter 2. After transformation of S. elongatus wild-type and P_II- and NtcA-deficient strains with the constructs and selection for double recombinants, in vivo bioluminescence from whole cells grown under different conditions was measured. Figure 4 shows the luciferase activity of the recombinant strains growing on nitrate- or ammonium-supplemented medium. In a wild-type background, luciferase activity from construct FAM1 largely exceeds that from FAM2, indicating that most transcription is initiated from the upstream promoter 1 in both ammonium- and nitrate-supplemented cells. In ammonium-grown cells, construct FAM2 showed a very tight repression; luciferase activity was significantly lower than that for the promoterless control strain WT-AM1580, indicating that active repression might eliminate background expression under these conditions. In the P_II mutant, repression by ammonium was similar to that in the wild type. The only significant difference between wild-type and P_II-deficient cells was an eightfold increase in luxAB expression in MP2-FAM2 grown in the presence of nitrate. The NtcA-deficient background supported growth only in the presence of ammonium, and under these conditions, luciferase activity of the full-length construct was approximately twofold lower than that in the wild-type background, whereas the tight repression of luxAB expression in construct FAM2 was impaired. Together, these results imply that ammonium repression of glnB expression requires NtcA but is independent of the P_II protein.

FIG. 4.

Expression of glnB::luxAB fusions which were recombined into S. elongatus wild-type (WT), P_II-deficient (MP2), and NtcA-deficient (MNtcA) cells. The bars represent the DNA inserts that contained the 5′ end of the glnB gene and upstream sequences and were cloned in front of the luxAB reporter genes of plasmid pAM1580 (2). For details of the constructs, see Materials and Methods. The first T from the TTG initiation codon of the glnB gene was defined as position +1. The locations of relevant regions in the glnB upstream region, promoter elements, and transcriptional start sites as determined in reference 21 are indicated above the bars. The numbers indicate the positions of the first 5′ nucleotides of the respective sequence elements relative to the TTG initiation codon. The reporter strains were grown for several generations in either nitrate- or ammonium-supplemented BG11 medium, and bioluminescence was recorded as described in Materials and Methods. Relative light units (RLU) from three to four independent determinations, together with the standard deviations, are shown. Bioluminescence values from the promoterless control strain WT-AM1580 are shown at the bottom.

To investigate the role of P_II in NtcA-activated gene expression following combined-nitrogen deprivation, the different recombinant strains were shifted to nitrogen-deprived medium and bioluminescence was monitored as a function of time. As shown in Fig. 5, nitrogen depletion in the wild-type background, initiated from either nitrate- or ammonium-supplemented cultures (A and B), caused a rapid and strong increase of bioluminescence. Maximum activities were obtained after 8 to 40 h of nitrogen starvation. Thereafter, the activities decreased again. Reporter strains WT-FAM1 and WT-FAM2 displayed similar activities during nitrogen deficiency, with the peak activities corresponding to an approximately 10- or 250-fold increase in induction for WT-FAM1 or WT-FAM2, respectively, compared to that for nitrate-grown cells. This indicates that under conditions of combined-nitrogen starvation, transcription from the NtcA-dependent promoter 2 is sufficient to yield the high expression of glnB. In contrast to the wild type, the P_II-deficient mutant was unable to elicit the high induction of glnB::luxAB expression when shifted from nitrate-supplemented medium to conditions of nitrogen starvation. When nitrogen step-down was performed with ammonium-grown cells, ammonium depression was relieved and the luciferase activities reached values corresponding to those of nitrate-grown cells. However, no induction above this intermediate value could be observed. This provides independent confirmation that the P_II-deficient strain has retained its ability to recognize ammonium. However, the mutant is unable to distinguish between the lack of combined nitrogen and nitrate-replete conditions, with respect to the activation of an NtcA-dependent promoter.

FIG. 5.

Time course of glnB::luxAB induction following nitrogen step-down in reporter strains of S. elongatus in wild-type (WT) (A and B), MP2 (C and D), and MNtcA (E) backgrounds. After the initiation of nitrogen deprivation at time point 0, bioluminescence from the reporter strains was recorded over a time period of 100 h. Experiments with shifts from nitrate-supplemented to nitrogen-deprived (−N) conditions (A and C) and from ammonium-supplemented to nitrogen-deprived media (B, D, and E) were performed. Dotted lines, FAM1 reporter strains; solid lines, FAM2 reporter strains. Independent experiments yielded similar results; for each condition, results from a representative time course experiment are shown.

As in the P_II⁻ mutant, induction of both glnB::luxAB reporter constructs was completely impaired in the NtcA⁻ background. Instead of the induction of glnB::luxAB, even a slight reduction of bioluminescence was observed upon prolonged incubation times in nitrogen-deprived medium.

DISCUSSION

The global analysis of protein synthesis patterns following nitrogen deprivation demonstrated that NtcA and P_II are the key players for regulating the acclimation of S. elongatus to nitrogen deprivation. For the sake of reliability, we applied a strict standard to define the various categories of induced or repressed spots, which meant that the actual number of proteins analyzed was a conservative underestimation. All those proteins for which synthesis was specifically induced by nitrogen starvation were dependent on NtcA and P_II. With one exception, the nitrogen-repressed spots were also NtcA and P_II controlled. This suggests that the P_II-NtcA regulatory pathway exerts a major level of control on the specific nitrogen starvation response. Interestingly, some proteins that responded to both nitrogen and sulfur starvation were under P_II-NtcA control, whereas others were not. This suggests that cellular responses to global nutritional changes are not controlled by a single signal. Apparently, there are global starvation signals that are sensed and transmitted in a P_II-NtcA-independent way. However, other genes, like rbcS, might be controlled by multiple signal transduction pathways, one of which responds to nitrogen limitation through the P_II-NtcA pathway. Therefore, the expression of these genes depends on P_II and NtcA only when cells are subjected to nitrogen deprivation.

Previous studies from our laboratory and from others suggested that NtcA is not controlled by P_II regulation (6, 22). This conclusion was derived from the observation that the repression of NtcA-activated genes such as glnA (encoding glutamine synthetase I) and the nir operon (encoding enzymes for nitrate uptake and reduction) by ammonium was unaffected in P_II-deficient mutants of S. elongatus. The P_II-independent response of NtcA towards ammonium was confirmed here by using glnB::luxAB fusions. The P_II-independent regulation of NtcA by ammonium might be based on the binding of 2-oxoglutarate to NtcA. 2-Oxoglutarate, the level of which was shown to be relatively low in ammonium-treated cells (25), affects the DNA-binding properties of NtcA (38) and is required for in vitro transcriptional activation by NtcA (36).

We conclude that under conditions of combined-nitrogen limitation, the regulation of gene expression by NtcA in S. elongatus also requires the P_II signal transduction protein. The evidence supporting this conclusion is derived from three complementary experimental approaches.

(i) 2-D PAGE analysis showed that nearly all proteins whose synthesis responded to combined-nitrogen deprivation were affected not only by NtcA but also by P_II. The similarity between P_II- and NtcA-deficient mutants suggested that under conditions of combined-nitrogen deprivation, P_II is involved in the synthesis of those proteins which are under NtcA control.

(ii) NtcA- and P_II-dependent repression of gene expression during nitrogen starvation was examined by quantification of the mRNA of one of the spots, RbcS, whose synthesis was down-regulated under conditions of nitrogen starvation in an NtcA- and P_II-dependent manner. NtcA-dependent repression of the rbcLS operon in the filamentous cyanobacterium Anabaena sp. strain PCC 7120 has previously been documented (39). Analysis of the nucleotide sequence of the rbcLS operon in S. elongatus revealed only half-sites of the NtcA consensus motif. Nevertheless, expression of that gene was undoubtedly repressed by combined-nitrogen deprivation, an effect that depended on both NtcA and P_II. Further studies are required to reveal the binding site on the DNA that is necessary for NtcA repression of the rbcLS operon.

(iii) Induction of glnB::luxAB fusions following combined-nitrogen deprivation was abolished in both NtcA⁻ and P_II⁻ mutants. The NtcA-dependent tsp2 of glnB is responsible for the high-level induction of glnB following combined-nitrogen depletion, as implied by the similar induction levels of constructs WT-FAM1 and WT-FAM2 and which is in agreement with previous primer extension analysis of the glnB promoter (21). The deficiency of the P_II mutant in activating this NtcA-dependent promoter is further evidence for the suggested requirement for the P_II protein in NtcA control during combined-nitrogen starvation.

Apart from showing that NtcA-activated glnB expression requires P_II, investigation of the glnB::luxAB fusions might point to another feature of NtcA. In addition to activating transcription under conditions of combined-nitrogen deprivation, NtcA might actively repress transcription of the same promoter in the presence of ammonium. This is suggested by comparing the FAM2 reporter strains. In the presence of ammonium, luxAB expression from WT-FAM2 and MP2-FAM2 was reduced to levels which are significantly lower than the background activity of the promoterless WT-AM1580 reporter strain. By contrast, in the NtcA-deficient background, lux activities were higher than background. We excluded the possibility of a reversion of the ntcA mutation, which might have generated an aberrant NtcA factor (data not shown). According to in vitro studies, when the 2-oxoglutarate levels are low (corresponding to growth in the presence of ammonium), NtcA is still able to bind to its DNA target sites, albeit with reduced affinity, but is unable to activate transcription (36, 38). In this conformation, NtcA might actively inhibit transcription. This would explain why those strains containing NtcA showed a very tight repression of the FAM2 reporter, well below the background level of the promoterless control. In the absence of NtcA, there could be unspecific transcriptional activation by NtcA-like transcription factors, such as CysR (26) and IdiB, whose recognition sequences closely resemble the NtcA binding site (24). Therefore, the binding of NtcA to its cognate sites in ammonium-grown cells might prevent cross talk by other transcription factors and might contribute to the repression of transcription from upstream promoters.

Whereas the mechanism of P_II-independent ammonium repression of NtcA activity probably involves direct interaction of NtcA with 2-oxoglutarate, the mechanistic basis of P_II-stimulated NtcA activation under conditions of nitrogen starvation is not understood so far. Regulation of NtcA by P_II ultimately will be governed by metabolites, which are a reflection of the carbon-nitrogen balances and energy statuses of the cells. The P_II protein is a sensitive 2-oxoglutarate sensor (8, 16), and recent studies have shown that S. elongatus P_II, albeit to a lesser extent, also responds to oxaloacetate and to the energy status of the cell (29). Analogous to the known mechanisms of P_II control in other bacteria, different strategies of P_II regulation are conceivable. P_II might regulate an antagonist of NtcA, analogous to NifA/NifL regulation in Klebsiella pneumoniae, where the P_II paralogue GlnK is required to relieve the inhibitory effect of NifL on the transcriptional activator NifA (12, 18). Alternatively, P_II, when signaling nitrogen starvation, might stimulate a factor which acts positively on NtcA, a situation opposite that found in NtrB/NtrC regulation. There, the histidine kinase NtrB, which activates the transcription factor NtrC by phosphorylation, is inhibited by P_II under conditions of nitrogen excess (reviewed in reference 27). Nitrogen starvation causes P_II uridylylation and the dissociation of P_II from NtrB, which then activates NtrC. P_II deficiency, therefore, causes constitutively active NtrC transcription. Finally, P_II might directly interact with the NtcA protein, analogous to the suggested regulation of NifA by P_II in certain members of the α- and β-subdivisions of Proteobacteria (3, 35). Currently, these issues are under investigation.