Cyanophycin Synthesis Optimizes Nitrogen Utilization in the Unicellular Cyanobacterium Synechocystis sp. Strain PCC 6803

To investigate the intracellular localization of CphA, we fused cphA to the gene encoding enhanced green fluorescent protein (eGFP) (yielding eGFP fused to the C terminus of CphA) and inserted cphA-eGFP under the control of the native cphA promoter into the Synechocystis sp. shuttle vector pVZ322 (34). The resulting pVZ322-cphA-eGFP plasmid was introduced into Synechocystis sp. by triparental mating (34). The resulting strain, termed Synechocystis sp. strain CphA-eGFP, was used to investigate the cellular localization of CphA using comprehensive statistical analysis of microscopy images.

In the majority of nutrient-replete cells (87% ± 3%; n = 486) in the mid-exponential phase of growth (optical density at 750 nm [OD₇₅₀], approximately 0.6), the GFP signal was uniformly distributed in the cytoplasm and did not colocalize with the fluorescence signal of the thylakoid membranes near the cell periphery (Fig. 1A to D). In the remaining cells (13% ± 3%), distinct foci formed. Accordingly, the cyanophycin content during exponential growth was usually less than 1% of the cell dry mass, and visible cyanophycin granules were absent (6, 17, 35).

To reveal the localization of CphA during periods of cyanophycin accumulation or degradation, we monitored CphA-eGFP localization after the addition of nitrate to Synechocystis sp. cells that had been starved for nitrogen for 4 days. These conditions are known to induce transient cyanophycin accumulation (7, 13, 29). The cphA gene is constitutively expressed in different nitrogen regimes (7). In agreement, a recent quantitative proteomics analysis of resuscitating nitrogen-starved Synechocystis sp. cells revealed minor changes in the abundance of CphA. It increased by about 10% from cells starved of nitrogen for 21 days to fully recovered exponentially growing cells (11). In nitrogen-starved cells, the GFP signal was almost exclusively found as a diffuse signal in the cytoplasmic space (98% ± 2% of the cells) (see Fig. S1 in the supplemental material). Three hours after the addition of nitrate, the CphA-eGFP strain formed distinct foci in almost half of the cells (Fig. 2). After 6 h and up to 27 h after nitrate addition, the majority of cells contained CphA-eGFP foci (83% ± 7%). As the foci appeared, their number per cell simultaneously increased, with a maximum between 6 and 12 h after nitrate addition (Fig. 2 and S2A); thereafter, the number of foci gradually decreased. The apparent diameter of the CphA-eGFP foci reached a maximum between 12 and 21 h after nitrate addition (Fig. S2B). To correlate the appearance of the CphA-eGFP foci with the occurrence of cyanophycin granules, we stained the granules with the arginine-specific Sakaguchi stain. In bright-field images, the granules are opaque; the Sakaguchi stain enables a sensitive and clear identification of cyanophycin granules (17). Compared to the appearance of CphA-eGFP foci, the appearance of the cyanophycin granules was delayed. Six hours after nitrate addition, the number of cells containing cyanophycin granules increased dramatically, reaching a plateau of 87.8% ± 4.2% of the cells at 15 h after nitrate addition. Eighteen hours after the addition of nitrate, both the number of cells with visible cyanophycin granules and the diameter of the foci decreased (Fig. 2 and S2B). As the cyanophycin granules degraded, the cells simultaneously started to grow again (Fig. 2).

During recovery from nitrogen starvation, cyanophycin is produced transiently, and the maximal level of accumulated cyanophycin was relatively low compared to other conditions that trigger cyanophycin accumulation (13, 17). Potassium starvation subjects cyanobacterial cells to an intense and immediate stress, which results in massive cyanophycin accumulation (17). This motivated us to analyze CphA localization under such conditions. Upon induction of potassium starvation, CphA relocalized from an initial diffuse cytoplasmic localization into distinct foci and later clearly localized on the surface of the cyanophycin granules, forming a halo-like structure (Fig. S3). Cyanophycin granules of larger diameter were amorphous and not spherical (Fig. S3). The proportion of cells with cyanophycin granules and CphA-eGFP foci increased during potassium starvation (Fig. S4A). Shifting the potassium-starved cells back to standard BG-11 medium restored the growth within approximately 24 h, concomitantly with the degradation of the cyanophycin granules. During cyanophycin degradation, it appeared in some cases that CphA-eGFP no longer colocalized with the granules (Fig. S3). The proportion of cells with cyanophycin granules and CphA-eGFP foci decreased accordingly. After 24 h, only about 20% of the cells contained cyanophycin granules. Simultaneously, 70% of the cells contained CphA-eGFP foci, while the remaining cells show the CphA-eGFP signal exclusively distributed in the cytoplasmic space (Fig. S4B). Furthermore, the average size of the granules decreased, while the number of granules per cell increased (Fig. S5), indicating that the granules disaggregate into smaller particles.

The above-mentioned results showed that CphA is homogenously distributed in the cytoplasm as long as no cyanophycin synthesis occurs, and when cyanophycin accumulation is triggered, CphA relocalizes into foci and subsequently attaches to the surface of growing cyanophycin granules. In vitro, the elongation reaction of CphA requires cyanophycin primers, arginine, aspartate, MgCl₂, ATP, and a sulfhydryl reagent (β-mercaptoethanol or DTT) (21). To determine how these components affect the localization of CphA-eGFP in cell extracts, we analyzed cell lysates using anti-eGFP antibodies and immunoblots. Cell extracts were prepared from potassium-starved Synechocystis sp. cells that produced CphA-eGFP and separated into soluble and insoluble fractions, with the insoluble fractions containing the cyanophycin granules. We determined whether CphA remained associated with granules in the presence of different buffer components.

In potassium-starved Synechocystis sp. cells, CphA-eGFP appeared to be localized on the surface of the cyanophycin granule in vivo (Fig. S3). When these cells were lysed in a buffer containing 50 mM Tris-HCl (pH 7.4) and 4 mM EDTA, CphA was detected only in the soluble fraction (Fig. 3A). In contrast, when cells were lysed in a buffer containing 50 mM Tris-HCl (pH 8.2), 20 mM MgCl₂, and 20 mM KCl, CphA appeared in the insoluble fraction, which indicated that it remained granule associated, because cyanophycin granules remain in the insoluble fraction, owing to their insolubility at neutral pH (Fig. 3B). To test whether MgCl₂ is responsible for the persistent association of CphA with granules in cell extracts, cells were lysed in the same buffer containing different MgCl₂ concentrations (Fig. 3C). In the absence of Mg²⁺, only residual amounts of CphA were detected in the insoluble fraction. Increasing amounts of Mg²⁺ led to increased amounts of CphA in the insoluble fraction, with an apparent optimal concentration of 10 mM Mg²⁺ for in vitro association with cyanophycin granules (Fig. 3C).

To gain further insights into the physiological role of cyanophycin, we generated a cphA deletion mutant in which the slr2002 (cphA) open reading frame was replaced with a kanamycin resistance cassette. Complete segregation of the mutant was confirmed by PCR (Fig. S6). The inability of the mutant to accumulate cyanophycin granule peptide (CGP) was confirmed by microscopy and cyanophycin extraction (data not shown). To investigate a possible influence of cyanophycin deficiency on the nitrogen starvation response, Synechocystis sp. wild-type and ΔcphA mutant cells were grown to an OD₇₅₀ of 0.4 to 0.6, washed, and resuspended in BG-11 lacking a combined nitrogen source. This treatment induces nitrogen chlorosis, which is visible as the degradation of the photosynthetic apparatus (9). We monitored the degradation of pigments during chlorosis by recording the UV-Vis spectra of the cells. In the course of chlorosis, the absorbance at 630 nm (phycobiliproteins; Fig. 4A) and at 680 nm (chlorophyll a/b; Fig. 4B) decreased in wild-type and ΔcphA mutant cultures similarly. After induction of nitrogen starvation, wild-type and ΔcphA mutant cells divided one more time, as indicated by the doubling of the OD₇₅₀ of the culture (Fig. 4C). After prolonged nitrogen starvation for 13 days, cells of both the wild type and the ΔcphA mutant completed chlorosis, and the cultures appeared yellowish (Fig. 4E). However, the pigment composition of long-term starved cultures of the ΔcphA mutant and wild type slightly differed; the peak heights at 440 nm and 680 nm in UV-Vis spectra in the ΔcphA mutant were higher, which indicated a higher residual amount of chlorophyll a/b (Fig. 4D).

As cyanophycin transiently accumulates during resuscitation from nitrogen starvation (7), we focused on this process to investigate the physiological role of cyanophycin. To determine whether cyanophycin synthesis influences the resuscitation process, Synechocystis sp. wild-type and ΔcphA mutant cells were first nitrogen starved for 2 weeks, and then resuscitation was initiated by the addition of 5 mM nitrate, adjusting the cultures to an OD₇₅₀ of 0.5. The cultures were incubated under continuous light with shaking. The time course of regreening was monitored by recording the UV-Vis spectra of the cultures.

Twelve hours after the addition of nitrate to the chlorotic cultures, the absorbance at 630 nm (phycobiliproteins; Fig. 5A) and at 680 nm (chlorophyll a/b; Fig. 5B) began to increase. The wild-type and ΔcphA mutant cultures regreened similarly (Fig. 5A to C). However, the ΔcphA mutant had a growth advantage over the wild type in liquid medium (as revealed by the optical density of the culture; see Fig. 5D). Thirty-six hours after the addition of nitrogen, the difference in the optical density amounts to 0.07 OD₇₅₀ units. After 46 h, the ΔcphA mutant reached an OD₇₅₀ of 0.747 ± 0.045 compared to the wild type, with 0.64 ± 0.036, which is a significant difference (P = 0.0329). This growth disadvantage of the wild type turned out even more clearly on solid medium (as revealed by the drop plate method; see Fig. 5E). In the drop plate method, the area of the ΔcphA mutant was green already after 2 days, which indicated a return to the vegetative cell cycle and growth; the wild type lagged behind in the regreening process. This growth advantage of the ΔcphA mutant over the wild type continued until day 8 of recovery, at which time the two strains showed a similar degree of recovery.

The growth disadvantage of the cyanophycin-forming wild-type cells during resuscitation from nitrogen chlorosis was puzzling. We questioned whether this disadvantage was due to unnatural laboratory conditions (nitrogen-rich medium and continuous light). Therefore, we nitrogen starved Synechocystis wild-type and ΔcphA mutant cells for 9 days and compared their resuscitation with a fluctuating and/or limiting nitrogen supplementation by a modification of the drop plate method (Fig. 6), which allows nitrogen supplementation to be changed without imposing additional stress to the cells that occurs with harvesting, washing, and resuspending cells in another medium.

When chlorotic cells were exposed under continuous light to 6-h periods of nitrogen excess (17.3 mM nitrate) per day (6 h on BG-11 with nitrate and 18 h on nitrogen-free BG-11), the ΔcphA mutant grew slightly faster than the wild type during the first 3 days (Fig. 6A). After the fourth day, the wild type and ΔcphA mutant reached the same culture density. After 7 days, the wild type displayed a slight growth advantage. After 9 days, the wild type clearly had a growth advantage over the ΔcphA mutant.

To test resuscitation with continuous but low nitrogen supplementation, chlorotic cells were exposed under continuous light to only 1.73 mM nitrate (10% of the standard concentration) and transferred to fresh plates containing 1.73 mM nitrate every day. Colonies of green cells of both strains appeared after 3 days (Fig. 6B), but the wild type showed a clear growth advantage over the ΔcphA mutant throughout the experiment (Fig. 6B).

In the next experiment, we combined nitrogen limitation and fluctuating nitrogen availability. A limiting amount of nitrogen (1.73 mM nitrate) was available for only 4 h per day under continuous light. Under these conditions, the wild type showed the clearest growth advantage over the ΔcphA mutant (Fig. 6C). Furthermore, a difference in the colors of the cultures of the two strains became apparent. Cultures of the ΔcphA mutant were more yellow than those of the wild type, which indicated that the ΔcphA cells could not fully recover their photosynthetic pigments and remained in a semichlorotic state.

To determine whether the growth advantage under fluctuating nitrogen concentrations of wild-type cells over the ΔcphA mutant could also be observed in liquid culture, we provided nitrogen-starved liquid cultures of Synechocystis sp. wild-type and the ΔcphA mutant with 10 mM nitrate for 4 h, followed by 20 h with no nitrogen source; this cycle was repeated for 4 days (Fig. 7). After 2 days, both strains began to regreen, but the wild type regreened faster than the ΔcphA mutant (Fig. 7A and B). After 4 days, the growth advantage of the wild type was clearly visible, since the optical density of the wild-type culture increased 3-fold, whereas that of the ΔcphA mutant increased only 2-fold (Fig. 7C).

Photosynthetic organisms in nature have to deal with day-night changes and a fluctuating and limiting nutrient supplementation. We hypothesized that cyanophycin that accumulates in nondiazotrophic cyanobacteria in the night could be used as an intracellular nitrogen source during the day. To test this hypothesis, we combined nitrogen limitation and fluctuating nitrogen supplementation with day/night periods.

Nitrogen-starved liquid cultures of Synechocystis sp. wild-type and the ΔcphA mutant were exposed to 5 mM nitrate for 20 h without light. After this dark period, the cells were washed and resuspended into fresh medium containing a small amount of nitrate (0.5 mM) and exposed to light. After the 20-h dark period, the wild type accumulated cyanophycin to 3.7% ± 0.6% of the cell dry mass, but the pigmentation of the wild type and ΔcphA mutant was still at its initial low level (Fig. 8A and B). Shortly after light exposure in low-nitrate medium, the wild type started to synthesize pigments, but the ΔcphA mutant was not able to regreen (Fig. 8A, B, and D). Forty-seven hours after the first addition of nitrate, wild-type cultures appeared green and cultures began to grow (Fig. 8C and D), whereas the ΔcphA mutant was still chlorotic. We further analyzed the growth behavior with a modified version of the drop plate method, in which Synechocystis sp. wild-type and ΔcphA mutant cells starved of nitrogen for 1 week were grown on plates with fluctuating nitrogen supply and with day/night cycles (16-h day/8-h night). A small amount of nitrate (1.7 mM) was provided only during the night. Under these conditions, the overall growth of both strains was very low. Nevertheless, the wild type showed a clear growth advantage over the ΔcphA mutant; growth of the wild type and mutant was detected after 8 and 12 days of cultivation, respectively (Fig. 8E).

Our results revealed different localizations of the inactive and active forms of CphA in cells of Synechocystis species. Under conditions in which no cyanophycin synthesis occurs, such as during exponential growth or nitrogen starvation, CphA is inactive and mainly located in the cytoplasm. Immediately after the induction of cyanophycin synthesis, CphA aggregated in foci that were randomly distributed in the cell. Formation of the CphA foci may be a consequence the priming of the CphA enzyme. The primers could induce self-aggregation.

CphA in its primed state is ready to elongate the cyanophycin primer. During this elongation process, the CphA foci increase in size and become visible as cyanophycin granules. As cyanophycin accumulates, the numbers of foci and granules per cell continuously decrease, possibly as initial granules fuse to form larger aggregates. This behavior would also explain the amorphous structure of the large granules that we observed when the cellular cyanophycin content was high and which we also observed previously in electron micrographs of a cyanobacterial cyanophycin-overproducer strain (17). The observed ring-like appearance of the fluorescence signal around large cyanophycin granules in our study suggests that during cyanophycin accumulation, CphA-eGFP covers the surface of the granule.

Previous studies have shown that CphA activity in vitro requires cyanophycin primers, arginine, aspartate, MgCl₂, ATP, and a sulfhydryl reagent (21). Our results indicate that Mg²⁺ is strictly required for the association of CphA to the cyanophycin granules. The primary structure of CphA is composed of two regions (23), both of which show sequence similarities to ATP-dependent ligases of two superfamilies. Sequence alignments of the d-alanine–d-alanine ligase DdIB with cyanobacterial CphA show that key residues involved in binding of the ATP-Mg²⁺ complex are conserved in CphA (22). Accordingly, Mg²⁺ may stabilize the active conformation of CphA which is able to bind the C terminus of the cyanophycin substrate. In the absence of Mg²⁺, CphA cannot maintain its enzymatic activity, which could lead to dissociation of the substrate and consequently to release from the granule surface.

Cyanophycin accumulation is triggered in the presence of combined nitrogen sources by several growth-limiting conditions that lead to growth arrest, and cyanophycin degradation can be triggered by restoring the growth of the arrested cells. During cyanophycin degradation, the localization of CphA-eGFP changed from the surface of the cyanophycin granule to the cytoplasm. This suggests that as cyanophycin degrades, CphA dissociates from the granule surface and distributes in an inactive form throughout the cytoplasm, where it remains silent until cyanophycin synthesis is triggered again by growth-arresting conditions. A model of the cyanophycin accumulation and degradation cycle during nitrogen-induced chlorosis and resuscitation is depicted in Fig. 9.

Our previous transcriptome study of nitrogen-depleted Synechocystis sp. cells has shown that the transcript levels of cphA in cells starved for nitrogen and of exponentially growing cells are very similar (1.2-fold upregulated during 14 days nitrogen starvation compared to nitrogen-supplemented cells). However, during resuscitation from nitrogen starvation, the cphA transcript level 24 h after the addition of nitrogen is approximately 1.7-fold higher than in exponentially growing cells (7). In agreement, quantitative proteomics revealed that the level of CphA after 21 days of nitrogen starvation is slightly reduced compared to nitrate-supplemented cells and goes back to initial values during recovery (11).

The synthesis of cyanophycin is probably tuned by the cellular level of arginine, which increases under growth-limiting conditions because of lowered protein biosynthesis. High concentrations of arginine surpass the K_m of CphA for arginine and could therefore trigger the biosynthesis of cyanophycin (29). Whether CphA is subjected to additional activity control remains to be elucidated.

The ability to synthesize cyanophycin is widespread among cyanobacteria and other eubacteria. With a C/N ratio of 2:1, cyanophycin is perfectly suited as a nitrogen storage compound. However, the physiological significance of cyanophycin for nondiazotrophic cyanobacteria remained largely unknown. Here, we showed that in the nondiazotrophic unicellular cyanobacterium Synechocystis sp., the deletion of cphA, which causes the inability to produce cyanophycin, has no impact on the nitrogen starvation response. The wild type and the ΔcphA mutant responded to nitrogen starvation similarly, namely, with chlorosis. However, the ΔcphA mutant reproducibly retained slightly higher residual pigment content in the chlorotic state. Merritt et al. (36) reported in nitrogen-limited cells of Synechocystis sp. the presence of a cyanophycin-like polymer that contains glutamic acid instead of arginine (36). Such a compound could be synthesized by a side-reaction of CphA and could affect the degradation of pigments in a so-far-unknown manner.

Several earlier studies have reported the transient accumulation of cyanophycin in cyanobacteria during resuscitation from nitrogen starvation (7, 13). Allen et al. (1980) reported an immediate synthesis of cyanophycin with a peak of 5 to 6% of the cell dry weight 8 to 12 h after the addition of NaNO₃ to nitrogen-starved Synechocystis sp. PCC 6308 cultures (13). Here, we show that under standard laboratory conditions with continuous light and nitrogen excess, the ΔcphA mutant showed a clear growth advantage over the wild type both in liquid medium and on agar plates. The synthesis of cyanophycin costs 1.3 ± 0.1 mol ATP per mol incorporated amino acid in vitro (37). Without using the benefit of cyanophycin as a storage compound, the synthesis of cyanophycin is only a burden to the cells, and consequently, the cyanophycin-deficient ΔcphA mutant has an advantage over the wild type. This is the case under artificial laboratory cultivation conditions, which provide an excess of nitrogen and light.

In nature, microorganisms have to deal with a fluctuating and limiting nitrogen availability (3). We hypothesized that cyanophycin accumulation is a strategy to overcome temporal limitations in nitrogen availability. Growth and resuscitation experiments of Synechocystis sp. wild-type and the ΔcphA mutant confirmed this hypothesis. The ability to store nitrogen in the form of cyanophycin was beneficial during resuscitation under a fluctuating or limiting nitrogen supplementation, and the cells have a growth advantage over the cyanophycin-deficient mutant. When nitrogen fluctuation and limitation are combined, the cyanophycin-deficient mutant was not able to fully recover from chlorosis.

To further mimic natural conditions, we tested day/night cycles in addition to a fluctuating and/or limiting nitrogen supply. In the unicellular diazotrophic cyanobacterium Cyanothece sp. ATCC 51142 and the filamentous cyanobacterium Trichodesmium sp., cyanophycin acts as temporary nitrogen storage to enable the coexistence of nitrogen fixation and photosynthesis in the same cell (19, 20). These cyanobacteria fix nitrogen in the night and store the fixed nitrogen in the form of cyanophycin. During the day, when cells carry out photosynthesis, cyanophycin is degraded to mobilize the fixed nitrogen. Based on this behavior in diazotrophic cyanobacteria, we hypothesized that cyanophycin is involved in nitrogen storage also in nondiazotrophic cyanobacteria at night. Indeed, when we combined fluctuating nitrogen supplementation and day/night cycles, the resuscitation of Synechocystis sp. wild-type cells proceeded as fast as in continuous light, with growth restoration after 48 h. In contrast, the cyanophycin-deficient mutant was not able to resuscitate in this time period.

The first phase of resuscitation from nitrogen chlorosis occurs in a light-independent mode, where the cells catabolize glycogen using the parallel operating Entner-Doudoroff and oxidative pentose phosphate pathways (10). Only when the photosynthetic apparatus is reinstalled after approximately 12 to 16 h does metabolism switch back from the heterotrophic mode to a mixotrophic mode and finally to the fully autotrophic mode. Our data suggest that most of the nitrogen that is required for regreening can be assimilated in the initial heterotrophic phase and stored as cyanophycin. The stored cyanophycin makes internal nitrogen available to the cells during the second phase of resuscitation, which can then proceed, even if the ambient nitrogen supply is again low.

Nitrate assimilation by cyanobacteria is tightly regulated at both the transcriptional and posttranslational levels (38, 39). Nitrate uptake is mediated by the bispecific nitrate/nitrite transporter encoded by the nrtABCD genes. This operon is only expressed in the absence of ammonium, and the nitrate/nitrite transporter is rapidly and reversibly inhibited by the addition of ammonia (38–40). The P_II signal transduction protein appears to be involved in this process, since a P_II-deficient mutant shows no ammonium-responsive inhibition of nitrate and nitrite uptake (41). In vegetative growing cells, in addition to the requirement for ammonium absence, the operation of the nitrate/nitrite transporter also requires CO₂ fixation, because the inhibition of photosynthetic CO₂ fixation results in the simultaneous inhibition of nitrate uptake (42). Newly assimilated nitrogen is incorporated in the central metabolism via the glutamine synthetase (GS) and glutamate synthase (GOGAT) cycle. GS activity is tightly regulated by light/dark transitions; induction of the inhibitory factor 7 (IF7) and IF17 in the dark causes GS to switch off (43–45). This leads to the inhibition of nitrogen assimilation in the dark, both at the level of nitrate transport and by inhibition of the ammonium-assimilating GS. Our results showed that cyanophycin accumulates upon nitrate addition to chlorotic cells in the dark, which requires active nitrate assimilation. This contradicts the dark switch-off nitrate assimilation in vegetative cells. It is possible that the P_II signaling protein is directly involved in this behavior. Nonphosphorylated P_II mediates a light-dependent inhibition of nitrate utilization (41, 46). However, during nitrogen starvation, P_II is highly phosphorylated and remains in the highly phosphorylated state during the first period of resuscitation (11, 47). In this state, P_II is probably unable to inhibit nitrate uptake. In agreement, Reyes et al. (48) have shown that the inactivation of GS in the dark is attenuated under nitrogen depletion, which allows the cells to utilize nitrogen in the dark after a nitrogen starvation period (48). The GS inhibitory factors IF7 and IF17 are regulated by the global nitrogen control transcriptional factor NtcA, which is in turn controlled by 2-oxoglutarate levels and the PII-PipX network (49–51). Apparently, during nitrogen starvation, where NtcA is activated by elevated levels of 2-oxoglutarate and by binding PipX, the nocturnal derepression of the gifA and gifB genes appears to be abrogated (52, 53).

The occurrence of cyanophycin in cyanobacteria has been known for more than 100 years. However, fundamental questions on its biological function or the benefit of cyanophycin accumulation remained largely uninvestigated. Altogether, our study sheds new light on the process of cyanophycin accumulation and reveals new insights into the biological function of cyanophycin in nondiazotrophic cyanobacteria. Interestingly, artificial laboratory conditions do not provide any fitness advantage for cyanophycin-producing cells, and in a fluctuating environment, CphA becomes beneficial. As the cphA gene is constitutively expressed in Synechocystis sp., the cells are programmed to overcome fluctuating nitrogen supply and transient periods of starvation in a constantly changing environment.

Standard procedures for cloning in Escherichia coli NEB 10-beta (NEB) and E. coli XL1-Blue (Stratagene) were followed. Strains were grown in LB medium at 37°C with constant shaking at 300 rpm.

Synechocystis sp. strains were cultivated photoautotrophically in BG-11 medium supplemented with 5 mM NaHCO₃ (54) at 27°C with constant shaking at 120 rpm and illumination at 40 to 50 microeinsteins. Growth was monitored by measuring the optical density at 750 nm. BG-11 agar plates contained 1.5% (wt/vol) Bacto agar (Difco), 0.3% (wt/vol) sodium thiosulfate pentahydrate, and 10 mM N-tris(hydroxymethyl)methyl-2-aminoethanesulfonic acid (TES)-NaOH (pH 8) (Roth). Antibiotics were added to the medium when required.

For the induction of starvation conditions, cells of exponentially growing cultures (OD₇₅₀, 0.4 to 0.6) were harvested, washed, and resuspended in BG-11 medium lacking a specific nutrient. For nitrogen starvation, the cells were resuspended in BG-11 medium without a nitrogen source. For resuscitation from nitrogen starvation, chlorotic cultures were adjusted to an OD₇₅₀ of 0.5, harvested, and resuspended in BG-11 containing a nitrogen source. To introduce potassium starvation, cells were harvested, washed, and resuspended in BG-11 medium containing Na₂HPO₄ in an amount equimolar to that of the K₂HPO₄ it replaced. Regeneration from potassium starvation was induced by harvesting, washing, and resuspending cells in BG-11 medium containing a potassium source.

For the drop plate method, Synechocystis sp. cultures were adjusted to an OD₇₅₀ of 1. The cultures were diluted 10-fold in series in BG-11 medium lacking nitrogen. Drops (5 μl) of five dilutions (10⁰ to 10⁻⁴) were placed on BG-11 agar plates containing specific nitrate concentrations. When cells were to be shifted to different solid media, the drops were placed on mixed-cellulose ester transfer membranes (pore size, 0.45 μm; HATF; Merck Millipore). The transfer membranes were placed on different BG-11 agar plates at the intervals specified in the figure legends. To minimize carryover of the residual nitrate from the initial plate, the transfer membranes were placed on nitrogen-free BG-11 plates for 10 min. The plates were incubated at 27°C with illumination at 40 to 50 microeinsteins.

PCR fragments were generated using high-fidelity Q5 polymerase (NEB) and oligonucleotides with overlapping regions. Genomic Synechocystis sp. DNA or plasmids served as the templates. The primers, plasmids, and strains used in this study are listed in Tables 1, 2, and 3, respectively.

To generate a ΔcphA mutant, 500-bp upstream and downstream genomic regions of slr2002 (cphA) were amplified using the oligonucleotides up_for/up_rev and down_for/down_rev. The kanamycin resistance gene (816 bp) was amplified from pVZ322 (34) using primers kan_for and kan_rev. Upstream and downstream fragments and the kanamycin resistance cassette were fused and incorporated into the linear pJET1.2/blunt cloning vector (Thermo Scientific) by isothermal single-reaction DNA assembly, according to Gibson et al. (55). The resulting construct (pJET ΔcphA) was introduced into competent E. coli NEB 10-beta by transformation. The correct length (1,816 bp) of the insert was confirmed via colony PCR using the primers provided in the CloneJET PCR cloning kit (pJet_seq_for and pJet_seq_rev) (Thermo Scientific). Synechocystis sp. was transformed with pJET ΔcphA via natural competence (56). Transformants were selected on BG-11 agar plates supplemented with 50 μg/ml kanamycin. Complete segregation was confirmed via PCR using primers seg_for, kan_for, and down_rev (Fig. S6).

For the cphA-egfp fusion, cphA including its promoter region was amplified using primers cphA_for and cphA_rev. Subsequently, the enhanced green fluorescent protein gene egfp derived from plasmid pCESL19 (57) was amplified using oligonucleotides gfp_for and gfp_rev. The cphA and egfp amplicons were fused and incorporated into XbaI-digested pVZ322 via DNA assembly according to Gibson et al. (55). The resulting plasmid, pVZ322 cphA-egfp, was introduced into competent E. coli XL1-Blue (Stratagene) cells by transformation. The sequence integrity of the plasmid was verified by sequencing using primers pVZ322_seq_for and pVZ322_seq_rev. Synechocystis sp. was transformed with pVZ322 cphA-egfp by triparental mating (34), and transformants were selected on BG-11 agar plates supplemented with 50 μg/ml kanamycin and 5 μg/ml gentamicin.

Cells were observed by fluorescence microscopy using a Leica DM5500B microscope with a ×100/1.3 oil objective. The GFP signal was detected with a BP470 40-nm excitation filter and a BP525 50-nm emission filter. Cyanobacterial autofluorescence was detected with a filter cube with excitation filter BP535/50 and suppression filter BP610/75. Images were acquired with a Leica DFC360FX black-and-white camera. Captured black-and-white pictures were colored using the Leica Application Suite Software (LAS AF) provided by Leica Microsystems. Bright-field images were exposed for 5 ms, and fluorescence images were exposed for 100 ms. Images, including those for counting fluorescent foci and measuring the diameter, were evaluated with the Leica Application Suite Software.

Cyanophycin granules were visualized using a staining method based on the arginine-selective Sakaguchi reaction, according to Watzer et al. (17). Photographs were taken with a Leica DM2500 microscope using a ×100/1.3 oil objective. Images were acquired with a Leica DFC420C color camera and Leica Application Suite Software.

During all microscopy studies, microscope slides covered with a dried 2% (wt/vol) agarose solution were used to immobilize the cells.

Potassium-starved Synechocystis sp. cells were harvested by centrifugation and resuspended in a buffer containing 50 mM Tris-HCl (pH 7.4), 4 mM EDTA, 1 mM DTT, and 0.5 mM benzamidine, or optionally in a buffer containing 50 mM Tris-HCl (pH 8.2), 20 mM MgCl₂, 20 mM KCl, and 1 mM DTT (21). Cells were lysed using FastPrep-24 (MP Biomedicals) with 0.1-mm glass beads at a speed of 6.0 m/s for 20 s with five repeats and 5 min of resting after every repeat. Soluble and insoluble fractions were separated by centrifugation at 25,000 × g for 25 min at 4°C. The protein concentration was determined using the Bradford assay (58).

Proteins were separated by SDS-PAGE on a 12% polyacrylamide gel according to Sambrook and Russell (59). Total protein (10 μg) was loaded on each lane. For immunoblot detection of CphA-eGFP, proteins were blotted onto a methanol-activated polyvinylidene difluoride (PVDF) membranes, as described previously (60). Membranes were blocked with 10% (wt/vol) milk powder in TBS buffer (50 mM Tris-HCl [pH 7.4], 75 mM NaCl) overnight. Afterwards, the membranes were transferred in 1% (wt/vol) milk powder in TBS buffer containing 1:2,500 diluted rabbit anti-GFP antibody (chromatin immunoprecipitation [ChIP] grade ab290; Abcam) and incubated for 2 h at ambient temperature. Unbound primary antibodies were removed by washing the membranes three times with TBS buffer. Anti-rabbit IgG secondary antibody conjugated to horseradish peroxidase (anti-rabbit polyclonal goat antibody; Sigma-Aldrich) diluted 1:1,000 in 1% (wt/vol) milk powder in TBS buffer was applied to the membranes and incubated for 30 min at ambient temperature. Afterwards, the membranes were washed three times with TBS buffer to remove unbound secondary antibodies. Bands were visualized using the Lumi-Light detection system (Roche Diagnostics). Luminograms were taken with the Gel Logic 1500 imaging system (Kodak) with the associated software.

Cyanophycin was extracted according to Watzer et al. (17) and quantified by determining the amount of arginine in the extracted sample using the modified Sakaguchi reaction, according to Messineo (61). The determined amount of cyanophycin was normalized to the cell dry mass. The cell dry mass was determined by centrifuging 10 ml of culture and washing and drying the pellet for 4 h at 60°C in a rotational vacuum concentrator. The dried pellets were weighed on an analytical balance.