The Synechococcus Strain PCC 7942 glnN Product (Glutamine Synthetase III) Helps Recovery from Prolonged Nitrogen Chlorosis

Abstract

We report the cloning and sequencing of the glnN gene encoding a class III glutamine synthetase from the cyanobacterium Synechococcus strain PCC 7942. Mapping of the transcriptional start site revealed a DNA sequence in the promoter region that resembles an imperfect NtcA binding motif. Expression of glnN is impaired in NtcA- and P_II-deficient mutants. The only parameter which was negatively affected in the glnN mutant compared to the wild type was the recovery rate of prolonged nitrogen-starved cells with low concentrations of combined nitrogen.

When nondiazotrophic cyanobacteria are deprived of combined nitrogen sources, the intense blue-green cultures turn yellow, by a process which is known as chlorosis (1). Recently, it was shown that the complete depigmentation of cells of the obligate photoautotroph Synechococcus strain PCC 7942 is not a consequence of loss of cell viability but a specific acclimation process which enables the cells to survive for prolonged periods of nitrogen starvation (12). This acclimation process proceeds in three phases. The rapid trimming and degradation of the phycobilisomes (phase 1) (2, 12) is followed by a gradual loss of chlorophyll a (Chl a) and intracellular proteins (phase 2). Finally, the cells lose all visible pigmentation and enter a form of dormancy (phase 3), from which they are able to reinitiate growth within 3 to 4 days following the addition of a combined nitrogen source (12).

A previous study investigated factors involved in the initiation of nitrogen chlorosis. It was shown that the signaling protein P_II (the glnB gene product), which senses the cellular nitrogen and carbon status (7, 8, 13), was not involved in the process (18). However, P_II-deficient cells display an impaired capacity to rapidly reinitiate growth after nitrogen starvation (18). In contrast, NtcA, a global activator of ammonium-repressed genes (14), was required for sequential pigment degradation and cell survival (18). In addition, nbl genes mediate phycobiliprotein degradation during starvation for either sulfur or nitrogen (3, 5, 19). Mutants with mutations in these genes display a nonbleaching phenotype and an impaired capacity to recover from starvation. The nblB gene seems to be constitutively expressed, but its product is not active in phycobiliprotein degradation until nblA is expressed (5). The expression of nblA is under the control of the NblR regulator (19). In the ntcA-deficient mutant, induction of the nblA gene was not impaired (18), but nevertheless, the cells were unable to rapidly degrade the phycobiliproteins. This indicates that the response to nitrogen depletion involves the concerted action of the NblR-controlled global starvation program and an NtcA-controlled nitrogen-specific response.

A candidate for an NtcA-controlled factor that is required for nitrogen chlorosis is the key enzyme of ammonium assimilation, glutamine synthetase (GS). It was shown that inhibiting GS activity during chlorosis by the addition of l-methionine sulfoximine almost immediately stops phycobiliprotein degradation (18), suggesting that GS activity might be required to support nblA-mediated phycobiliprotein degradation. Interestingly, besides the primary GS (GSI), which is under the control of NtcA (14), a second GS isoenzyme (termed GlnN), belonging to the GSIII class, has been identified in nondiazotrophic cyanobacteria (9, 15). Although the subunit structures and the amino acid sequences of the GSI and GSIII classes differ markedly, similar enzymatic properties, such as K_m and V_max values, have been reported (9). In Synechocystis strain PCC 6803, the expression of glnN is strongly induced under conditions of nitrogen deprivation (9, 16). Together with the fact that glnN has so far only been observed in nondiazotrophic cyanobacteria (9, 15), this suggested a role for GlnN in nitrogen chlorosis. However, in nitrogen-depleted cells, despite being strongly induced, the GSIII class was reported to account for no more than 20% of the total GS activity (15). Since no phenotype has been reported for the glnN-deficient Synechocystis strain, it was unclear whether this relatively moderate contribution of the GSIII class to the total GS activity was sufficient to play any physiological role in the process of nitrogen chlorosis. The present study aimed to clarify the role of glnN in the cyanobacterium Synechococcus strain PCC 7942.

Characterization of the glnN gene from Synechococcus strain PCC 7942.

To identify the glnN gene from Synechococcus strain PCC 7942, we performed Southern blot hybridizations of digested Synechococcus strain PCC 7942 DNA using a PCR-generated Synechocystis strain PCC 6803 glnN probe. This, as well as other standard DNA manipulation and cloning techniques, was performed according to the method of Sambrook et al. (17). Genomic DNA was digested with HindIII, and DNA fragments in the size range of 9 kbp hybridizing with the glnN probe were isolated and redigested with EcoRI. Finally, DNA fragments in the size range of 4 to 4.5 kbp were isolated and cloned into pBluescript II KS(+). From 80 clones which were analyzed by dot blot hybridization of the isolated plasmid DNA, one plasmid (pUD3) contained the glnN hybridizing fragment. Sequencing of the 4,270-bp DNA insert revealed a 2,172-bp open reading frame corresponding to a protein of 723 amino acids with a calculated relative molecular weight of 78,941 and an isoelectric point at pH 5.02. Its deduced amino acid sequence is highly similar to those for GlnN from Synechocystis strain PCC 6803 and Pseudoanabaena strain PCC 6903 (4) (88 and 75% identity, respectively) and less similar to those for the glutamine synthetases from Bacteroides fragilis and Butyrivibrio fibrisolvens (42 and 38% identity, respectively).

The transcriptional start point (TSP) of the Synechococcus glnN gene was determined by primer extension analysis (Fig. 1A). For this purpose, the oligonucleotide 5′GGCGGATCCCGATTGGTGATCTGG3′ (from nucleotide +59 to nucleotide +36, with the first nucleotide of the translation start codon considered +1) was end labeled with [γ-³²P]ATP using T4 polynucleotide kinase, as described elsewhere (17). The primer (0.2 pmol; approximately 5 × 10⁵ cpm) was added to 12 μg of total RNA, and the mixture was heated in hybridization buffer (100 mM KCl, 50 mM Tris-HCl [pH 8.3]) for 30 s at 95°C and then for 20 min at 55°C. The extension reaction was performed for 1 h at 41°C with 10 U of avian myeloblastosis virus reverse transcriptase (Promega) in the presence of 5 U of RNase inhibitor (Roche). The extension product was run on a sequencing gel (8% polyacrylamide) adjacent to the DNA sequence obtained by using the same oligonucleotide as the primer.

FIG. 1.

(A) Primer extension analysis of the glnN promoter region from RNA isolated from nitrogen-deprived cells (−N). Lanes C, T, A, and G contain a dideoxy sequencing ladder obtained with the primer used for the extension reaction. The transcription start nucleotide is indicated by an arrow. (B) Alignment of the Synechococcus strain PCC 7942 glnN promoter region (glnN 7942) with the corresponding region for Synechocystis strain PCC 6803 (glnN 6803). A putative imperfect NtcA binding site is indicated in boldface, the TSPs are indicated by arrows, and the −10 sequence is boxed.

A single extension product was obtained when RNA from nitrogen-starved cells was used as a template for reverse transcription. A product of the same size could be derived from RNA of nitrate-replete cells, although it was less abundant (data not shown). The TSP, as deduced from this experiment, is located 22 bp upstream of the putative ATG translational start codon. A sequence with similarities to the canonical NtcA binding motif can be found upstream of the TSP, with a distance that corresponds well to that reported for NtcA binding motifs (14) (Fig. 1B).

Expression of glnN in mutants deficient in the nitrogen control genes ntcA or glnB.

As a first step to analyze the expression of glnN, Northern blot experiments were performed. Total RNA was prepared from Synechococcus strain PCC 7942 and Synechocystis strain PCC 6803 cells grown in modified BG11^N medium (12) and from cells of the same strains harvested 7 h after nitrogen deprivation, as previously described (18). Total RNA (15 μg) was separated on a 1% formaldehyde-agarose gel and transferred to a nylon membrane (Nytran NY 12 N; Schleicher & Schuell). As a Synechococcus strain PCC 7942-specific glnN probe, a 988-bp BglI DNA fragment from plasmid pUD3 (Fig. 2A) was used. Although the glnN probe revealed no distinct bands using RNA from Synechococcus strain PCC 7942 wild-type cells, a smear with hybridization signals of a maximal size of 2.3-kb was detected, which corresponded in size to a monocistronic glnN transcript. This signal was specifically increased in nitrogen-starved cells (Fig. 2B). As a positive control, RNA from nitrogen-depleted Synechocystis strain PCC 6803 cells gave a strong signal at 2.3 kb with a Synechocystis strain PCC 6803 glnN probe (Fig. 2B). We argue that the glnN transcript in Synechococcus strain PCC 7942 cells is even more unstable than that in Synechocystis strain PCC 6803 cells, for which a short half-life was reported previously (16). To get additional proof that glnN is transcribed monocistronically in Synechococcus strain PCC 7942 cells, Northern blot experiments were performed with a GlnN-deficient mutant (MGlnN), in which the Tn5 Kan^r Bleo^r cassette was inserted into the glnN gene in the same transcriptional orientation (Fig. 2A). Two transcripts with sizes of 2.5 and 1.65 kb could be detected that hybridized with a 3′ glnN probe but not with a glnN probe upstream of the inserted Tn5 Kan^r Bleo^r cartridge (Fig. 2). Control hybridization with a Tn5 Kan^r probe revealed that the longer transcript corresponds to an mRNA that encompasses the Tn5 Kan^r Bleo^r genes, as well as the 3′ half of glnN. The shorter transcript corresponds in size to a product of the Tn5 Bleo^r gene and the 3′ half of the glnN gene. No signals of longer transcripts could be observed, strongly indicating that the glnN transcript terminates at the 3′ end of this gene. To further analyze glnN expression, we examined GlnN synthesis by Western blot experiments using antibodies raised against Synechocystis strain PCC 6803 GlnN (9) (Fig. 3). As shown previously, in wild-type Synechococcus strain PCC 7942 cells the synthesis of GlnN is strongly induced under conditions of nitrogen starvation (9). In contrast, in cells of the P_II-deficient strain MP2 (7), GlnN abundance increased only slightly after 2 days of nitrogen starvation. Cells of the NtcA-deficient mutant MNtcA (18) were completely unable to increase the synthesis of GlnN beyond its very low basal level. This result indicates that the activated glnN expression following nitrogen deprivation depends on NtcA and at least partially on P_II. Recently, evidence was presented which suggests that glnN is under NtcA control in Synechocystis strain PCC 6803 (10). Together, these accumulated evidences suggest that NtcA recognizes the region upstream of glnN, which resembles the NtcA motif. The difference between the canonical NtcA binding motif (14) and the motif reported here is a change from GTAN8TACN22TANNNT to GTAN7TAGCN22TANGAT. A similar sequence (GTAN7TGTCN22TANGAT) can be found in the promoter region of glnN from Synechocystis strain PCC 6803 (16) (Fig. 1B). Despite the fact that no binding of the NtcA protein to this latter sequence could be observed by gel retardation experiments (16), because of the NtcA dependence of GlnN synthesis, we suggest that the DNA sequence upstream of glnN represents a weak NtcA binding motif. It is possible that strong binding of NtcA to this sequence requires additional factors. In this respect it is interesting that a P_II-deficient mutant is also impaired in glnN expression. Whether P_II regulates a factor that is involved in the expression of glnN remains to be elucidated.

FIG. 2.

(A) Structure of the glnN region in Synechococcus strain PCC 7942 wild-type cells and in a mutant in which a kanamycin-bleomycin resistance cassette was inserted into the glnN gene. To create the mutant MGlnN, plasmid pUD3 was restricted with BclI, which cleaved 948 bp downstream from the putative start codon of the glnN coding region, and the 1.6-kb BamHI-BamHI Tn5 Kan^r Belo^r cassette from plasmid pUC-KIXX (Pharmacia) was inserted into the BclI site, yielding plasmid pGLNN-KAN. Cells of Synechococcus strain PCC 7942 were transformed with plasmid pGLNN-KAN as described by Golden et al. (11) and tested by PCR and Southern blot analysis for complete segregation. The arrows indicate the transcriptional orientation. (B) Northern blot of total RNA isolated from Synechococcus strain PCC 7942 wild-type (7942) and GlnN-deficient cells (MGlnN) and from Synechocystis strain PCC 6803 wild-type cells (6803) grown in nitrate-containing medium (+) and incubated for 7 h in nitrogen-free medium (−). A 988-bp BglI DNA fragment from plasmid pUD3 was used as a Synechococcus strain PCC 7942-specific glnN probe, a 1,070-bp PCR-amplified internal glnN fragment was used as a Synechocystis strain PCC 6803-specific glnN probe, a 0.73-kb HindIII-XbaI fragment from plasmid pRN16A (R. Figge, unpublished data) containing a part of the 16S rRNA encoding gene from Synechocystis strain PCC 6803 was used as a 16S rRNA probe. Transcript size was estimated by comparison with 23S and 16S rRNAs. N, nitrogen.

FIG. 3.

Immunoblot analysis of the GlnN (GSIII) and GlnA (GSI) proteins. Cells of Synechococcus strain PCC 7942 (wt), of the P_II-deficient mutant (MP2), and of the NtcA-deficient mutant (MNtcA) grown in the presence of ammonium (+N) were transferred to nitrogen-free (−N) medium. Extracts were prepared (18) from cells harvested during exponential growth (+N) and from cells harvested 1 and 2 days after nitrogen step-down. Twenty micrograms of total protein from the cell extracts was separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis. GlnN was revealed by using antibodies raised against the GlnN protein of Synechocystis strain PCC 6803 (9) and GlnA was visualized by using antibodies raised against GlnA from Nostoc sp. strain UCD 7801.

Functional analysis of the glnN gene in Synechococcus strain PCC 7942.

The expression pattern of glnN, together with the fact that glnN has not yet been found in nitrogen-fixing strains, implies that this GS may be specifically designed for conditions of nitrogen deficiency. To determine the physiological function of GlnN, the GlnN-deficient mutant MGlnN was analyzed with respect to its growth rate, response to various stress conditions, and the capacity to survive nitrogen starvation. In the exponential growth phase, it exhibited the same growth rates as the wild type when supplemented with nitrate or ammonium at different concentrations (0.5 to 20 mM) (data not shown). To investigate the possibility that GlnN is a chlorosis-specific GS, MGlnN was deprived of nitrogen and its phenotype was analyzed. The level of total GS activity during starvation was determined by the GS transferase assay (7). The mutant displayed the same increase in GS activity as the wild type following nitrogen downshift (data not shown), implying that in Synechococcus strain PCC 7942, most GS is derived from glnA even under nitrogen-deprived conditions. With prolonged nitrogen starvation, the GS activity in both strains decreased progressively to a level below the limit of detection after about 18 days. The process of chlorosis was recorded by analyzing the levels of phycocyanin (PC), Chl a, and glycogen and measuring the optical density at 750 nm (OD₇₅₀) of the cultures following a transfer of cells from ammonium-replete to nitrogen-depleted conditions (Table 1). The experimental procedures were performed as previously described (12). A deficiency in GlnN caused no significant impairment in the chlorosis process. Up to the completion of the pigment degradation process (about 15 days into chlorosis), the glnN mutants recovered, following the addition of a combined nitrogen source, as fast as wild-type cells. This is in accord with the similar GS activities detected in the strains. Therefore, GlnN is not required to support the chlorosis process. However, upon prolonged incubation under nitrogen-deficient conditions (periods up to 16 weeks were investigated), the capacity of MGlnN to reinitiate growth was retarded compared to wild-type cells, and the mutant was not able to accumulate wild-type levels of phycobiliproteins. Figure 4 shows the amount of newly synthesized phycobiliproteins over a period of 3 days in Synechococcus strain PCC 7942 glnN mutants and wild-type cells that had been incubated for 5 weeks under nitrogen-deprived conditions and supplemented at the onset of the regeneration experiment with either 0.5 mM NaNO₃ or 0.5 mM NH₄Cl. One day after the addition of nitrogen to the chlorotic cultures no phycobiliproteins could be measured, but after 2 days the beginning of a transient repigmentation was observed. Maximal phycobiliprotein levels were obtained after 2 or 3 days, depending on the nitrogen source. After consumption of the added nitrogen source, phycobiliproteins were degraded again. Compared to the wild type, MGlnN accumulated significantly less phycobiliprotein and degraded it earlier. When cells recovery was induced with either 1 or 0.1 mM NaNO₃ or NH₄Cl, the transient pigmentation period was increased or decreased, respectively, but the difference between the wild type and MGlnN was consistent. Similar results were also obtained at different light intensities (data not shown). However, when chlorotic cultures were supplemented with excess nitrogen (20 mM), GlnN deficiency caused no phenotype. This shows that GlnN helps chlorotic cells to acquire combined nitrogen at low ambient concentrations, which are typical for natural environments (6). Therefore, wild-type cells are able to accumulate maximal levels of phycobiliproteins under these conditions which serve as a nitrogen reservoir for subsequent periods of nitrogen deprivation. Other physiological stresses, such as high light stress and sulfur or carbon starvation, had no effect on the phenotype of the glnN mutation. Moreover, the reinitiation of growth from lightlimited stationary-phase cultures was also unaffected. We conclude from this investigation that GlnN is specifically designed for nitrogen assimilation under conditions of severe nitrogen limitation. The fact that various nondiazotrophic cyanobacteria are equipped with the glnN gene emphasizes the selective advantage under natural conditions that results from the capacity to cope with prolonged periods of nitrogen starvation.

TABLE 1.

Change in the levels of several chlorosis indicators and in OD₇₅₀ per culture volume (12) in cultures of Synechococcus strain PCC 7942 wild-type and MGlnN cells following a transfer of cells from ammonium-replete conditions to nitrogen-depleted conditions

Time of nitrogen deprivation (days)	Wild-type cells				MGlnN cells
	OD750	PCa	Chl a	Glycogen	OD750	PC	Chl a	Glycogen
0	1	1	1	1	1	1	1	1
2	1.95	0.08	0.85	20.4	1.51	0.08	0.77	18.4
4	2.05	0.02	0.71	24.7	1.6	0.02	0.65	19.0
8	1.5	0	0.49	22.1	1.36	0.006	0.52	16.0

^aValues for PC, Chl a, and glycogen are expressed relative to the concentrations present after 0 days of nitrogen deprivation. 

FIG. 4.

Newly synthesized phycobiliproteins 2, 3, and 4 days (2d, 3d, and 4d) after addition of either 0.5 mM NaNO₃ (NO₃⁻) or 0.5 mM NH₄Cl (NH₄⁺) to chlorotic cultures of Synechococcus strain PCC 7942 (wt) and MGlnN that were incubated previously under nitrogen-deprived conditions for 36 days. The cultures were adjusted to the same OD₇₅₀ prior to the addition of nitrate. The amount of phycobiliproteins was estimated as described previously (12) and shown as the relative phycocyanin content per OD₇₅₀ (rel. PC-content/OD₇₅₀). nd, not detectable.

Nucleotide sequence accession number.

The nucleotide sequence of the glnN gene from Synechococcus strain PCC 7942 has been deposited in GenBank under accession number AF251806.