Introduction

Cyanobacteria show flexible responses to environmental changes (Tandeau de Marsac and Houmard 1993), including orderly degradation of the light-harvesting phycobilisomes during starvation for a macronutrient, such as nitrogen or sulfur (Grossman et al. 1993), as part of a differentiation process that enables the cells to survive long periods of nutrient starvation (Sauer et al. 2001). Degradation of the pigmented phycobiliproteins is thought to free some of the limiting nutrient for synthesis of important proteins or down-regulate light absorption by the cyanobacterium in order to reduce the amount of photodamage when reductive anabolism is limited by low mineral nutrient sinks (Allen and Smith 1969; Görl et al. 1998). The degree and rapidity of phycocyanin loss through orderly phycobilisome dilution and degradation varies, however, even among studies of a single species, such as Synechococcus elongatus (Collier and Grossman 1992; Görl et al. 1998; Allen and Smith 1969; Wanner et al. 1986; Yamanaka and Glazer 1980; Lau et al. 1977). The effects of nutrient starvation on chlorophyll (Chl) content per cell also vary, with reports of decreases through net degradation (Wanner et al. 1986; Görl et al. 1998), dilution by cell division (Collier and Grossman 1992), and steady (Allen and Smith 1969) or even slightly increased Chl per cell (Lau et al. 1977). The culture conditions or procedures used to initiate nutrient deprivation may have contributed to these different patterns of pigment loss. Therefore, in this study, cultures of S. elongatus were systematically grown under different conditions, transferred to nitrogen-deplete media, and phycocyanin and Chl content were monitored. In S. elongatus, the protease(s) involved in the multiple steps of phycobilisome or Chl complex degradation are unknown, although phycobilisome degradation during nutrient deprivation is triggered by a 59-amino-acid protein, NblA (Collier and Grossman 1994).

The ClpP family of ATP-dependent, serine-type proteases is important in cyanobacteria such as S. elongatus (Clarke 1999; Porankiewicz et al. 1999; Adam et al. 2001). ClpP proteases are composed of two functionally distinct subunits: a proteolytic component, ClpP, that associates with either of two regulatory ATPase subunits, ClpX or ClpC which are members of the Hsp100/Clp family of molecular chaperones and likely confer substrate specificity to the Clp proteolytic complexes (Clarke 1999). In S. elongatus, there are three isozymes of ClpP encoded by three different genes, clpPI, clpPII and clpPIII (Clarke et al. 1998; Schelin et al. 2002) along with a homologous clpR gene that, however, lacks key residues characteristic of the ClpP proteolytic active site.

A Synechococcus mutant in which clpPI was inactivated grew more slowly and did not properly acclimate to low temperature, UV-B radiation or high light intensities (Clarke et al. 1998; Porankiewicz et al. 1998), and ClpP1 content increased under these same conditions in wild-type cells. All these conditions affect phycobilisome size and structure since excess light absorption can damage cells, particularly when nutrient limitation restricts anabolic processes in the cell. In order to investigate the roles of Clp proteases in pigment complex dynamics during nitrogen stress in cyanobacteria, ΔclpPI and ΔclpPII strains (Clarke et al. 1998; Schelin et al. 2002) were transferred to nitrogen-deficient media and pigment loss was followed spectrophotometrically. clpPIII and clpR are essential for viability in Synechococcus (Schelin et al. 2002) and were therefore not included.

Materials and methods

Cyanobacterial strains and growth conditions

Wild-type, ΔclpPI and ΔclpPII strains of S. elongatus were maintained in solid and liquid BG-11 media (Rippka et al. 1979; Clarke et al. 1995), with 5 μg of appropriate antibiotic ml−1 added to stock cultures of ΔclpPI (kanamycin) and ΔclpPII (chloramphenicol) to maintain selection. No antibiotics were added to the experimental cultures. Liquid cultures were grown in 80-ml culture tubes with temperature control provided by immersing the tubes in a water-bath at either 37 or 32°C (Eriksson et al. 2000), bubbled with 5% C02or ambient air, and illuminated from cool white fluorescent tubes at an intensity of 50 or 120 μmol photons m−2 s−1 measured with a LI250 quantum meter (LiCor, Lincoln, Neb., USA) and a micro-spherical quantum sensor (Walz, Effeltrich, Germany), placed within an empty 80-ml culture tube at the position to be occupied by the experimental culture tube. When measured with a micro-spherical quantum sensor, 120 μmol photons m−2 s−1 was equivalent to the 70 μmol photons m−2 s−1 described in Schelin et al. (2002), measured with a LI250 quantum meter fitted with a flat co-sine corrected quantum sensor in the same culture set up.

Initiation of nitrogen deprivation

Pre-cultures were grown in nitrate-replete BG-11 media under the appropriate experimental conditions. Exponentially growing cells were diluted into fresh BG-11 media and grown to a final Chl concentration of approximately 2.0 μg ml−1 . The cells were harvested by centrifugation at 3,000×g for 5 min at room temperature and washed twice in nitrogen-deficient BG-11 media (Collier and Grossman 1992). The cells were re-suspended to a Chl concentration of approximately 1.0 μg ml−1 in 50–80 ml of pre-warmed nitrogen-deficient BG-11 media that had been pre-bubbled in ambient air or 5% CO2. At selected time points, samples were collected with a sterile syringe and analyzed. Upon repeated microscopy inspections over several different trials, no difference in cell size was noted between wild-type, ΔclpPI, and ΔclpPII strains, nor did cell volume appear to increase after transfer to nutrient-deficient media in any of the strains. Chl and phycocyanin content were estimated from whole-cell spectra (400–750 nm) according to the method of Myers et al. (1980), while cellular growth rates were estimated by absorbance at 750 nm on a Shimadzu spectrophotometer. The absolute values of the exponential rates of change in Chl/A750, phycocyanin/A750 and growth for each replicate were estimated using the exponential curve fitting function of Excel (Microsoft); R2 values were above 0.85 for all the exponential fits of growth or decline in pigments for each replicate of each treatment and strain. The rates data were then analyzed with a two-way ANOVA, showing significant differences among the rates for the three parameters (Chl/A750, phycocyanin/A750 and growth) across the treatments for all strains, using Prism 4 (GraphPad Software), followed by Student t tests to evaluate the significance (P<0.05) of differences among rates within a particular treatment of a strain. An absolute value of the exponential rate of decline in pigment per A750 that is significantly larger than the exponential rate of growth in A750indicates active net degradation of the pigment pool, in addition to dilution by cell growth and division.

Results and discussion

Figure 1 shows the changes in Chl and phycocyanin content per cell and in cell numbers in cultures of Synechococcus deprived of nitrogen. Four growth conditions differing in carbon dioxide concentration, temperature, or light intensity were compared. Chl and phycocyanin (mol) per A750 were used to estimate the relative content of these pigments per cell over the course of nitrogen starvation.

Fig. 1
figure 1

Pigment changes in Synechococcus elongatus wild-type (filled squares), ΔclpPI (filled circles), and ΔclpPII (open triangles) cultures after transfer to nitrogen-deplete media.. Exponentially growing cultures were shifted to nitrogen-deplete media at a Chl concentration of approximately 1.0 μg ml−1 and grown under the following experimental conditions: a–c 120 μmol photons m−2 s−1, 37°C, 5% C02; d–f 50 μmol photons m−2 s−1, 37°C, 5% C02; g–i 50 μmol photons m−2 s−1, 32°C, 5% C02; j–l 50 μmol photons m−2 s−1, 32°C, ambient C02. The values are average ± SD, from three replicates. Chl Chlorophyll; PC phycocyanin. Absorbance at 750 nm is expressed as a percentage of the pre-shifted control, to account for small variations in the number of cells per ml at the start of each replicate

Full size image

When wild-type cultures were grown at 37°C, illuminated with 120 μmol photons m−2 s−1 and bubbled with 5% CO2, Chl/A750 declined exponentially to very low levels within 15 h after the onset of nitrogen starvation and was nearly undetectable after 24 h (Fig. 1a). The cultures sustained approximately one round of cell division within 15 h after the withdrawal of nitrogen (Fig. 1c), but the rapid concomitant drop in Chl/A750 to significantly less than 50% of control levels showed net degradation of the Chl complexes in addition to dilution through the single round of cell division. Phycocyanin/A750 also declined through cell division and some net degradation (Fig.1b), although the decrease was significantly slower than for Chl/A750.

Figure 1d–f shows the bleaching pattern in wild-type, ΔclpPI and ΔclpPII cultures in which the growth light was lowered to 50 μmol photons m−2 s−1 while retaining the 37°C and 5% CO2. Chl content again decreased rapidly in the wild-type and and in the ΔclpPII strain, again through both dilution over one round of cell division (Fig. 1f) and through net degradation to nearly undetectable levels after 24 h (Fig. 1d). Chl/A750 decreased significantly more slowly in the ΔclpPI strain, which retained Chl after 24 h, although there was still evidence for some net Chl degradation. Phycocyanin content per cell decreased mainly by pigment dilution with some net degradation in all three strains (Fig. 1b).

Figure 1g–i shows that at a lowered temperature of 32°C, again illuminated with 50 μmol photons m−2 s−1 and bubbled with 5% CO2, Chl losses in the wild-type and ΔclpPII strains were slower, although growth rates at 32°C were similar to those at 37°C. Both dilution and net degradation drove down the Chl/A750 to only about 10% of the initial level within 72 h of the shift to low nitrogen (Fig. 1g, i). In contrast, in the ΔclpPI strain the decrease in Chl/A750 could be fully accounted for by dilution of the pigment with continued cell division, and more than half of the Chl remained after 72 h. Therefore, the observed net degradation of Chl complexes at either temperature (37 or 32°C) with 5% C02 at 50 μmol photons m−2 s−1 is dependent upon an intact clpPI, and by inference, the ClpP1 protease. These growth conditions also resulted in both dilution and some net phycocyanin degradation to approximately 15% of the initial level after 72 h in all three strains (Fig. 1h, i).

Carbon dioxide concentration also had a pronounced effect on patterns of Chl and phycocyanin content under nitrogen deprivation. When wild-type cultures were bubbled with ambient CO2 instead of 5% CO2, at 32°C, with 50 μmol photons m−2 s−1, the resulting pattern of pigment loss was similar to that in previous findings for Synechococcus starved of nitrogen (Collier and Grossman 1992), in that Chl was not degraded, but rather, the moderate decrease in Chl per cell could be fully accounted for by dilution through cell division (Fig. 1j, l). These low CO2 growth conditions also resulted in both dilution and net phycocyanin degradation to approximately 15% of the initial level after 72 h in all three strains (Fig. 1k, l). Insertional inactivation of clpPII leads to small increases in ClpP1, ClpP3, ClpR, and ClpX (Schelin et al. 2002). If any of these proteins is involved in phycocyanin degradation during nitrogen stress, a slight increase in the rate of phycocyanin loss might occur in the ΔclpPII mutant, consistent with the non-significant but reproducibly slightly faster decline in phycocyanin/A750 in the ΔclpPII mutant (Fig. 1e, h, k). In any case, in contrast to the strong effects of clpPI on patterns of Chl loss under nitrogen deprivation, clpPII does not appear to be significantly involved in pigment dynamics under the early stages of nutrient deprivation, although our findings do not preclude a role for clpPII under more prolonged nutrient deprivation.

Thus, high concentrations of CO2 (5%) resulted in ClpP1-dependent net Chl degradation upon nitrogen deprivation, with Chl per cell dropping more than phycocyanin per cell in the wild-type and ClpP2 strains. In contrast, growth in ambient CO2 resulted in a decline in Chl per cell only through dilution by cell division. Collier and Grossman (1992) grew Synechococcus cultures with 3% CO2 and upon nutrient deprivation Chl content declined on a per cell basis through dilution by cell division. The achieved inorganic carbon status of a culture, however, depends on the balance between input through bubbling and diffusion and removal through light-dependent inorganic carbon uptake to support growth. Thus, although the CO2 inputs were similar in our study and in the study of Collier and Grossman (1992), the achieved inorganic carbon status may have differed significantly.

In S. elongatus under high inorganic carbon, moderate light and 37°C, most (ca. 90%) cellular Chl is associated with PSI (Mackenzie et al. 2004), so significant drops in Chl cell−1 reflect drops in PSI content. In our experiments, removing nitrogen from cells in high inorganic carbon therefore resulted in a ClpP1-dependent net degradation of PSI. Cells in low inorganic carbon, in contrast, did not degrade much PSI upon nitrogen deprivation, perhaps because their electron transport system is reorganized to drive cyclic electron transport around PSI to drive their carbon concentrating mechanisms (Kaplan et al. 2001; Price et al. 2002).

Synechococcus needs to modulate the photosynthetic apparatus to balance fluctuating light intensity, temperature and nutrient availabilities. ClpP1 is involved in this regulation. Furthermore, regulation of pigment dynamics under nutrient deprivation shows strong interactions with inorganic carbon status. For the first time, we have shown that removing nitrogen from cells in high inorganic carbon, moderate light and 32–37°C can result in a ClpP1-dependent loss of Chl, and by inference, loss of PSI.