Regulation of synthesis and degradation of a sulfolipid under sulfur-starved conditions and its physiological significance in Chlamydomonas reinhardtii

Authors


Author for correspondence:
Norihiro Sato
Tel: +81 42 6766716
Email: nsato@ls.toyaku.ac.jp

Summary

  • Regulation of synthesis and degradation of sulfoquinovosyl diacylglycerol (SQDG), one of the membrane lipids that construct thylakoids, under sulfur (S)-starved conditions and its physiological significance were explored in a green alga, Chlamydomonas reinhardtii.
  • Here, we used sac1 and sac3 mutants defective in response to ambient S-status to characterize the system of known induction of SQDG degradation by S starvation that ensures a major S source for protein synthesis. The SQDG synthesis system was monitored in the wild type during S starvation. An SQDG-deficient mutant, hf-2, was utilized to discover functions where SQDG metabolism participates during S starvation.
  • The induction of SQDG degradation was largely repressed in both sac1 and sac3 mutants. The SQDG synthesis capacity was increased by 40% after S starvation, with a sixfold elevation in the mRNA level of the SQD1 gene for SQDG synthesis. Compared with the wild type, hf-2 had decreased protein accumulation, photosystem (PS) I stability and growth rate.
  • A role of SQDG as an S storage lipid is fulfilled under the control of both SAC1 and SAC3 genes, and it is essential for proper protein synthesis in acclimatization of cells to S starvation. The enhancement in SQDG synthesis may reflect the importance of SQDG as the membrane lipid that stabilizes the PSI complex.

Introduction

Lipids can be classified into membrane and storage lipids by functional aspects. Membrane lipids form bilayers that maintain metabolic homeostasis inside and/or outside of the membranes by preventing solutes such as metabolites from freely crossing, or interacting with membrane proteins for proper exertion of their activities (Palsdottir & Hunte, 2004). Meanwhile, storage lipids are intracellular storage molecules, such as triacylglycerol for fatty acids, which are subject to β-oxidation for the production of reducing power and ATP, and can thus be seen as energy-storing lipids (Athenstaedt & Daum, 2006).

Sulfoquinovosyl diacylglycerol (SQDG) is found mainly as one of the membrane lipids that construct thylakoid membranes in the chloroplasts of plants and their postulated ancestors, cyanobacteria. It is synthesized through two successive reactions. First, UDP-sulfoquinovose is synthesized from UDP-glucose and sulfite by UDP-sulfoquinovose synthase, which are encoded by the SQD1 and sqdB genes, respectively, in plants and cyanobacteria. Then, sulfoquinovose is transferred from UDP-sulfoquinovose to diacylglycerol by SQDG synthase, which are encoded by the SQD2 and sqdX genes, respectively, in plants and cyanobacteria (Sato, 2004). Mutants defective in the synthesis of some particular lipid class are powerful tools for elucidation of its physiological significance. Characterization of the photosynthetic apparatus in mutants deficient in SQDG synthesis has thus far revealed that SQDG contributes to the structural and functional integrity of the photosystem II (PSII) complex in a green alga, Chlamydomonas reinhardtii, and a cyanobacterium, Synechocystis sp. PCC 6803, probably by associating with the complex (Sato et al., 1995a, 2003a; Minoda et al., 2002, 2003; Aoki et al., 2004). It is also of note that X-ray structural analysis of the PSII core complex of a thermophilic cyanobacterium, Thermosynechococcus elongatus, revealed three molecules of SQDG. One of these SQDG molecules, in particular, together with three molecules of other lipid classes, resided close to the binding pocket of QB, the secondary quinone, thereby providing a lipophilic environment and probably facilitating diffusion of the quinone (Loll et al., 2005), which is compatible with one of the structural defects specific to a region around the QB-binding site in the mutants mentioned (Minoda et al., 2002; Sato et al., 2003a; Aoki et al., 2004).

Sulfur (S), which is included in a variety of essential biomolecules, is one of macronutrients for plants (Saito, 2004). Apart from such physiological importance of S in plant cells, emission of S as dimethylsulfide from marine phytoplankton is considered as an important factor for global climate and the biogeochemical S cycle, since dimethylsulfide is the main natural source of reduced S to the atmosphere and is involved in formation of cloud condensation nuclei. (Giordano et al., 2005). Thus, the study of regulation of S metabolism in plants is very important. A greater part of the study has thus far been restricted within green plants, and especially to C. reinhardtii and a seed plant, Arabidopsis thaliana, where genetic tools have well been developed. In response to a limited ambient S-source, C. reinhardtii and A. thaliana exhibit similar upregulation of expression of the genes required for S acquisition, such as those for sulfate transporters and for primary S-assimilation, including cysteine synthesis (Nikiforova et al., 2003; Zhang et al., 2004). These responses can be interpreted as attempts of these organisms to satisfy the demand for cysteine and its derivative methionine for the synthesis of proteins to be expressed during acclimatization to S starvation. It is thus critical for plants to ensure intracellular S-sources for acclimatization to S-starved conditions. In addition to the role of SQDG in PSII as a membrane lipid, we recently demonstrated a novel role of SQDG as an internal S-source in C. reinhardtii under S-starved conditions: SQDG is almost completely broken down as early as within 6–12 h to provide a large proportion of the S for the concurrent synthesis of proteins (Sugimoto et al., 2007) and thus may be responsible for the construction of a basic system for acclimatization. The utilization of SQDG as an S-storage lipid is reasonable in C. reinhardtii, because SQDG accounts for as much as 13% of total S in cells grown under S-replete conditions and is dispensable for their growth (Sugimoto et al., 2007).

So far, two genes, SAC1 and SAC3, have been extensively characterized in C. reinhardtii, with the use of their corresponding disruptants, as encoding signaling components for acclimatization of cells to S starvation. The SAC1 protein, which is structurally homologous to the Na+/SO42− transporter, is postulated to sense the unavailability of external S and then to transduce a signal to upregulate the levels of expression of a special set of genes (Davies et al., 1996; Zhang et al., 2004). Meanwhile, SAC3, a putative serine/threonine kinase, seems to either positively or negatively regulate physiological processes related with the ambient S-status (Davies et al., 1999).

Here, the hypothesis was examined that regulatory systems of SQDG synthesis and its degradation function in acclimatization of the cells to S-starved conditions in C. reinhardtii. We investigated the relationship of a regulatory system of SQDG degradation in response to S starvation with SAC1 and SAC3, and with physiological processes associated with gene expression in general. Regulation of SQDG synthesis and the underlying system were also investigated. Meanwhile, hf-2 that lacks SQDG owing to a single nuclear mutation (Sato et al., 1995b) seems genetically clean except for the impaired gene for SQDG synthesis because of achievement of 5-time backcrossing with the WT (Sato et al., 2003a). We then examined the physiological significance of the regulation of SQDG metabolism with the use of hf-2.

Materials and Methods

Strains and growth conditions

The strains used were C. reinhardtii CC125 as the wild type (WT), hf-2, an SQDG-deficient mutant that has already been backcrossed five times with the WT, two mutants of sac1 and sac3 that are unable to normally respond to the ambient S-status, and complemented strains obtained by introduction of the WT SAC1 and SAC3 genes (Davies et al., 1996, 1999). The sac mutants and complemented strains were obtained from the Chlamydomonas Genetics Center (Duke University, Durham, NC, USA). Each strain was grown mixotrophically in a flask containing Tris–acetate–phosphate (TAP) medium (Harris, 1989) on a rotary shaker (3.2 × g) with continuous illumination (60 μmol photons m−2 s−1) at 30°C. For transfer to S-starved conditions, cells grown to the mid-logarithmic phase (1–5 × 106 cells ml−1) were harvested by centrifugation, washed twice and then resuspended in S-free TAP medium (TAP–S), which was prepared by replacing sulfate with chloride, as described in de Hostos et al. (1988). Growth of the cells was monitored by determination of the cell density or optical density at 730 nm. The chlorophyll (Chl) concentration was determined spectrophotometrically as described in Porra et al. (1989). When indicated, inhibitors such as rifampicin (Rif, 80 μg ml−1), actinomycin D (ActD, 20 μg ml−1), chloramphenicol (CAP, 100 μg ml−1), and cycloheximide (CHI, 8 μg ml−1) were added immediately after the medium change. Ethanol was used as a carrier of CAP (final concentration, 1%). Rif is an inhibitor of chloroplast and mitochondrial RNA polymerases whereas ActD inhibits transcription in general by interaction with double-stranded DNA. Both CAP and CHI are inhibitors of de novo protein synthesis on chloroplast and mitochondrial 70S ribosomes and cytoplasmic 80S ribosomes, respectively.

[35S] Labeling of the cells and fractionation of [35S]-labeled compounds

Chlamydomonas reinhardtii cells were grown for 2–3 d in TAP medium containing 37 kBq ml−1 of [35S]sulfate (55.28 MBq nmol−1) for universal labeling of S compounds. After three washings of the cells by centrifugation with fresh medium, they were shifted to non-radiolabeled TAP or TAP-S medium for further growth. The cell suspension was centrifuged to obtain a supernatant (medium fraction) just before the shift and after 6 h of the further growth. The pelleted cells were washed with fresh medium, and thereafter resuspended in CHCl3 for agitation with a vortex mixer and subsequent centrifugation to obtain a supernatant and a precipitate. The supernatant was mixed with an equal volume of methanol and a half volume of water for agitation, and then separated into upper (water-soluble fraction) and lower (lipid-soluble fraction) layers by centrifugation. The radioactivity in each fraction was measured with a liquid scintillation counter (LSC-6100; ALOKA, Tokyo, Japan). The radioactivity of the sulfolipid was quantified with a BAS imaging analyser (BAS2000; Fuji Film, Tokyo, Japan) on a thin-layer chromatography (TLC) plate, on which total lipids extracted from cells according to the method of Bligh & Dyer (1959) were separated into individual lipid classes as described previously (Sato et al., 1995b).

Protein analysis

Whole-cell extracts were prepared through disruption of cells in an extraction buffer (5 mm Tricine-NaOH, pH 7.8, 5 mm NaCl, 1 mm MgCl2) by sonication. The protein contents were determined with a BCA assay kit (Pierce, Rockford, IL, USA). Samples equivalent to 100 μl of culture were subjected to sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS--PAGE) and subsequent Western analysis with antibodies against PsaA/B or PsbA, as previously described (Sato et al., 2004). Thylakoid membranes were isolated and ones equivalent to 1.8 ml culture were used for analysis by SDS-PAGE as previously described (Sato et al., 1995a). The bands corresponding to LHCII subunits stained with CBB dye were quantified on the gel with ImageGauge software (Fuji Film).

Measurement of physiological activities

The rate of sulfate uptake was estimated through measurement of the radioactivity in the cells incubated in the TAP medium containing [35S]sulfate for 5 min and 15 min with a liquid scintillation counter. The rate of SQDG synthesis was estimated by measurement of the radioactivity of SQDG prepared from the cells after incubation in the medium containing [35S]sulfate for 10 min and 20 min. For arylsulfatase activity, the cells were disrupted by sonication in a buffer, 0.4 m glycine-NaOH, pH 9.0, and then collected by centrifugation at 100 000g. The supernatant was mixed with 10 mm imidazole and 0.3 mm 5-bromo 4-chloro 3-indolylsulfate (Ohresser et al., 1997), and then used for spectroscopic measurement of the activity at 650 nm with a microplate reader (Benchmark Plus; Bio-Rad Laboratories, Hercules, CA, USA).

Photosynthetic activities were measured with a Clerk-type electrode as described by Sato et al. (1995a). The electrode chamber was kept at 27°C, and illuminated with a tungsten projector lamp (550 μmol photons m−2 s−1). Photosynthetic O2 evolution coupled with CO2 fixation was measured in a reaction mixture containing cells equivalent to 12 μg Chl ml−1, 10 mm Tricine pH 7.5, and 1 mm NaHCO3, pH 8.5. Photosystem I (PSI) activity was measured with the reduced form of 2,6-dichloroindophenol and methylviologen as the electron donor and acceptor, respectively. The reaction mixture was composed of sonicated cell suspension equivalent to 2.4 μg Chl ml−1, 20 mm Mes/Tris pH 8.1, 21 μm 2,6-dichloroindophenol, 2 mm sodium ascorbate, 80 μm methylviologen, 6.0 μm 3-(3′,4′-dichlorophenyl)-1,1-dimethylurea (DCMU), 0.4 mm KCN and 2 mm NH4Cl. PSII activity was measured with p-benzoquinone as the electron acceptor. The reaction mixture consisted of cell suspension equivalent to 12 μg Chl ml−1, 20 mm Mes/Tris pH 8.1, 300 μmp-benzoquinone and 2 mm NH4CI.

RNA extraction and northern blotting

Total RNA was extracted from cells suspended in a buffer (50 mm Tris-HCl, pH 7.5, 50 mm EDTA, 1% SDS, 0.5 mg ml−1 proteinase K) through agitation with phenol–chloroform–isoamylalchol (25 : 24 : 1 v/v) at 70°C, and then precipitated by the addition of the same volume of isopropanol. The SQD1 mRNA was detected by Northern hybridization with a probe corresponding to a part of the SQD1 cDNA (Sato et al., 2003b) that was amplified by PCR with sense and antisense oligonucleotide primers, 5′-ATCGTGGACAACCTGGTGCG-3′ and 5′-CCATAGCAAGGGTTCCGA-3′, respectively. For determination of the rate of the degradation of mRNA, S-replete cells or ones starved of S for 2 h were treated with ActD for inhibition of transcription.

Results

Regulatory system that induces SQDG degradation on S-starvation

Cells of C. reinhardtii, when moved to S-starved conditions, showed reduced growth in cell number or optical density (OD730) with repressed synthesis of proteins, and, in particular, that of Chl which represents Chl–protein complexes (Fig. 1a–h, compare closed symbols for the WT, e.g. in a and b for growth). However, it should be emphasized that S-starvation still allowed proteins in total or Chl–protein complexes to quantitatively increase, although at much reduced rates compared with S-replete conditions. The role of SQDG in C. reinhardtii as a major internal S-source for the synthesis of proteins could be demonstrated under S-starved conditions simply as most of the radioactivity (ca. 80%) in cells labeled with 35S was initially found in the precipitate (see the Materials and Methods section), the remainder being in the lipid- and water-soluble fractions (13% and 8%, respectively). The radioactivity in the precipitate and lipid-soluble fractions accounted predominantly for proteins and SQDG, respectively (data not shown). There were no remarkable changes in the radioactivity in the precipitate and lipid-soluble fractions during further growth for 6 h under 32S-replete conditions (Fig. 1i, open bars). However, it should be noted that the cells, when exposed to S-starvation for 6 h, showed an increase in the radioactivity of precipitate of 10% (see closed bar for precipitate in Fig. 1i) from 80% to 90% of the total radioactivity. Meanwhile, the radioactivity of the lipid-soluble fraction showed a 10% decrease in radioactivity (closed bar for lipid in Fig. 1i) from 13% to 3% of the total radioactivity.

Figure 1.

 Effects of the sulfoquinovosyl diacylglycerol (SQDG)-deficiency mutation on growth, contents of proteins and chlorophyll (Chl), and sulfur (S) metabolism of Chlamydomonas reinhardtii. Cell number (a,b), optical density (OD730; c,d), and contents of proteins (e,f) and Chl (g,h) were measured during S starvation. The wild type (WT, closed symbols) and hf-2 (open symbols), which had been grown in Tris–acetate–phosphate (TAP) medium to about OD730 = 0.2, were further grown in the same medium (a,c,e,g) or were shifted to TAP–S (TAP without S) medium (b,d,f,h). The insets (b,f,h) show changes in the relative values during the early phase of S-starvation. (i,j) Cells universally labeled in advance with [35S]sulfate were further grown for 6 h under 32S-replete or S-starved conditions (open and closed bars, respectively). The precipitate, lipid-soluble and water-soluble fractions, and medium were obtained for measurement of radioactivity in the respective fractions as described in the materials and Methods section. The total radioactivity of cells plus medium stayed almost same throughout the experiments (data not shown). The radioactivity in the respective fractions that was gained (positive values) or lost (negative values) during 6 h is shown. The values are expressed relative to the total radioactivity. (i) WT, (j) hf-2. The values are the averages ± SE for three distinct groups of data. The significance of differences was evaluated using R 2.4.1 software with a Welch t-test (http://www.r-project.org/): *, < 0.05.

We first explored this induction system of SQDG degradation with the use of several metabolic inhibitors (Fig. 2a). Treatment of cells with ActD or CHI, inhibitors of the transcription and translation of nuclear genes, respectively, abolished induction of the SQDG degradation, while that with Rif or CAP, transcriptional and translational inhibitors, respectively, in chloroplasts and mitochondria, did not affect the induction pattern. These results suggested that induction of the SQDG degradation required de novo synthesis of mRNA and proteins encoded by nuclear genes only. To gain a deeper insight into the involvement of the transcription and translation of nuclear genes, the inhibitors were added 2 h or 4 h after the onset of S starvation (Fig. 2b,c). Intriguingly, exposure of the cells to S starvation in advance for at least 2 h abolished the inhibitory effect of ActD, but not that of CHI. The period when the transcription is required could thus be limited to the very initial stage of S-starvation within 2 h, whereas the demand for translation lasted throughout the induction period.

Figure 2.

 Regulation of sulfoquinovosyl diacylglycerol (SQDG) degradation during S-starvation in Chlamydomonas reinhardtii. Cells were cultured for 3 d in the presence of [35S]sulfate, and thereafter shifted to Tris–acetate–phosphate (TAP) medium containing [32S]sulfate or TAP-S (TAP without S) medium. Sulfolipids were then separated by TLC at the indicated times for measurement of their radioactivity. (a) Cells were shifted to the S-starved conditions in the presence of Rif (rifampicin), ActD (actinomycin D), CAP (chloramphenicol) or CHI (cycloheximide), as described in the Materials and Methods section. The radioactivity of SQDG at 12 h after the onset of S starvation is shown relative to that before the shift (pre). Note the remarkable decrease in 35S in the absence of the inhibitors and in the light (control). (b,c) Effects of the times of addition of inhibitors on induction of SQDG degradation are shown. The times when inhibitors were added were 0 (squares), 2 h (triangles), and 4 h (circles) after the onset of S starvation; diamonds, without an inhibitor (). (b) ActD; (c) CHI. (d,e) Effects of the sac1 (d) and sac3 (e) mutations on induction of SQDG degradation are shown. Mutants (diamonds) and genetically complemented strains (circles) under S-replete (closed symbols) and S-starved (open symbols) conditions. Mutants and complemented strains showed similar values throughout incubation under S-replete conditions so that corresponding symbols overlap each other. The values are the averages ± SE forthree to five distinct groups of data. The significance of differences were evaluated using R 2.4.1 software with a Tukey–Kramer multiple comparisons (http://www.r-project.org/): **, < 0.001.

We next examined whether or not the SAC1 and SAC3 genes are involved in the regulation of SQDG degradation. Shifting cells labeled with 35S to nonlabeling 32S-replete conditions only slightly decreased the radioactivity of SQDG in 12 h forboth the sac1 mutant and the complemented strain obtained by introduction of the WT SAC1 gene (Fig. 2d). These results indicated that SQDG in either strain, like that in the WT, is relatively stable under S-replete conditions. High stability of SQDG was also observed in cells of the sac3 mutant and in those of the complemented strain (Fig. 2e). It can thus be concluded that neither SAC1 nor SAC3 are involved in the high stability of SQDG under S-replete conditions. Conversely, the disruptant as to SAC1 or SAC3 was remarkably repressed in level of induction of SQDG degradation under S-starved conditions compared with the respective complemented strains showing rapid degradation of SQDG, like the WT (Fig. 2d,e). Both the SAC1 and SAC3 genes thus participate in triggering of SQDG degradation under S-starved conditions.

Upregulation of SQDG synthesis during S starvation

Despite extensive repression of the synthesis of proteins in total and Chl–protein complexes, the SQDG synthesis capacity was never decreased, but rather was increased by 40%, on average, after S starvation for 6 h (Fig. 3a). On Northern hybridization, we also observed that the level of mRNA of the SQD1 gene for UDP-sulfoquinovose synthase, which participates in SQDG synthesis, was upregulated sixfold in 3 h, thereafter remaining at almost the same level for a further 9 h (Fig. 3b). The SQDG synthesis system during S-starvation thus never seems to coordinate with the induction of SQDG degradation. Two putative isogenes for SQDG synthase, designated SQD2a and SQD2b, are found in the genome of C. reinhardtii. However, we could not detect a signal of either gene on Northern analysis or semi-quantitative reverse-transcription polymerase chain reaction (RT-PCR) even during S-starvation.

Figure 3.

 Regulation of sulfoquinovosyl diacylglycerol (SQDG) synthesis during S-starvation in Chlamydomonas reinhardtii. (a) The rate of incorporation of [35S]sulfate into SQDG after the onset of S-starvation was measured in cells as described in the Materials and Methods section. (b–f) Changes in the mRNA level of SQD1 were investigated by Northern hybridization. Total RNA was isolated from cells at the indicated times and loaded on a gel (5–20 μg depending on the experiment). The probe used was a fragment of SQD1 cDNA, as described in the Materials and Methods section. Cells of the WT (b) or hf-2 (c) that had been grown normally were shifted to TAP–S (Tris–acetate–phosphate without S) medium. Cells that had been grown normally were moved to fresh TAP or TAP–S medium (closed and open diamonds, respectively) for 2 h, and then incubated after the addition of ActD at 0 min (d). Cells grown normally were moved to fresh TAP or TAP–S medium (open and closed diamonds, respectively) in the presence of CAP (chloramphenicol) (e) or CHI (cycloheximide) (f). The values are the averages ± SE for three to seven distinct groups of data.

We then investigated whether or not the accumulation of SQD1 mRNA results from its enhanced stability by comparing the half-life of SQD1 mRNA after treatment with transcription inhibitor ActD between S-replete and S-starved cells. As S starvation for 2 h showed no significant effect on the half-life of the SQD1 mRNA (34 ± 3 min with S starvation vs 38 ± 8 min with S repletion), the increased amount of SQD1 mRNA appeared to be caused not by enhanced stability, but by an increase in the transcription rate (Fig. 3d). To understand the regulation system of the expression of SQD1 in more depth, we examined the effects of inhibitors concerning translation. The induction was significantly repressed by CAP or CHI, implying that de novo synthesis of proteins encoded on both the chloroplast and nucleus genomes is required for the upregulation of SQD1 mRNA (Fig. 3e,f).

We then determined the level of transcripts of SQD1 in hf-2, the nucleotide sequence of which was confirmed to be intact (data not shown), to explore the possibility that the decrease in SQDG itself (see control in Fig. 2a) is the trigger to activate expression of the SQD1 gene. The amount of SQD1 mRNA in hf-2 was normal under S-replete conditions (data not shown) but was enhanced by about sixfold under S-starved conditions (Fig. 3c). These results suggested that the elevated amounts of SQD1 mRNA in the WT is determined not by feedback regulation through sensing of the SQDG content. It was interesting that hf-2 exhibited a saturated level of SQD1 induction more rapidly (within 1 h) than the WT (Fig. 3b,c).

Physiological significance of regulation of SQDG synthesis and its degradation under S-starved conditions

The pattern of accumulation of total proteins or growth was then compared between the WT and hf-2 to obtain a deeper insight into the physiological significance of SQDG degradation under S-starved conditions. The two strains showed little difference in the accumulation pattern of proteins or Chl under S-replete conditions (Fig. 1e,g), suggesting that the biomass of the cells in the culture increased at a similar rate for the WT and hf-2. However, growth in cell number was slower in hf-2 than in the WT (Fig. 1a), with the cell size being larger in hf-2 than in the WT (data not shown). These results implied that SQDG-deficiency mutation has a deleterious effect on cell division, and would explain why there was little effect of the mutation on cell growth in biomass. This idea was consistent with similar growth curves in OD730 for the WT and hf-2 (Fig. 1c).

The two strains, despite above similarity under S-replete conditions, showed clear differences under S-starved conditions as early as 6–12 h after the onset of S starvation. The mutant had a lower accumulation of total proteins compared with the WT, (Fig. 1f). In accordance with the results, hf-2 increased the radioactivity of 35S in the precipitate by only 3% (the closed bar for precipitate in Fig. 1j) from 93% to 96% of the total radioactivity after S starvation of 6 h whereas the WT increased it by as much as 10% (Fig. 1i). The smaller increase in hf-2 resulted from its redistribution of 35S to the precipitate only from the water-soluble fraction (the closed bar of water in Fig. 1j, 3% from 5% to 2% of total radioactivity) and not from the lipid-soluble one (Fig. 1j). These observations indicate the essentiality of the induction of SQDG degradation for ensuring an S source for the proper synthesis of proteins. Accordingly, growth at the early stage was more severely retarded in hf-2 than in the WT (Fig. 1d). By contrast, no deleterious effects of SQDG-deficient mutation were observed on photosynthetic activities, including activities of photosystems PSI and PSII, which are the central functions of the photosynthetic electron transport system in thylakoid membranes where SQDG resides (Fig. 4a). The mutant and the WT exhibited a decrease in the PSII activity by c. 20% during 6 h of S starvation, with no decrease, but rather an increase in the PSI activity.

Figure 4.

 Effects of the sulfoquinovosyl diacylglycerol (SQDG)-deficiency mutation on photosystem (PS)I, PSII and rubisco in Chlamydomonas reinhardtii. (a) Changes in photosynthesis, and PSI and PSII activities on sulfur (S)-starvation for 6 h. The initial O2 evolution of photosynthesis coupled with CO2 fixation was 214 ± 5 μmol O2 mg Chl−1 h−1 in the wild type (WT) and 185 ± 7 in hf-2. The initial activity of PSI was 885 ± 112 μmol O2 mg Chl−1 h−1 in the WT and 913 ± 151 in hf-2, whereas that of PSII was 499 ± 6 in the WT and 292 ± 6 in hf-2. Changes in the contents of PsaA/B (b), PsbA (c) and rubisco large subunit (e) were monitored in total proteins by Western analysis with the respective antibodies, while those of apoproteins of LHCII (d) were analysed by sodium dodecyl sulfate–polyacrylamide gel electrophoresis. Changes in the content of the large subunit of rubisco were monitored by Western analysis with an antibody against the protein. The values are the mean ± SE for three distinct groups of data. The significance of differences were evaluated using R 2.4.1 software with a Tukey–Kramer multiple comparisons (http://www.r-project.org/): *, < 0.05.

Concerning the defect in the synthesis of proteins in hf-2, the almost completely repressed accumulation of Chl is of particular note, in contrast to its increase by c. 30% in the WT, during S-starvation for the first 6 h (Fig. 1h, inset). To identify the Chl–protein complex that is responsible for the difference in the behavior of Chl between these two strains, we performed a Western analysis with antibodies against PsaA/B (PSI subunits) or PsbA (PSII subunit) (Fig. 4b,c) and SDS-PAGE of thylakoid membranes (Fig. 4d). In contrast to the similar steady levels of PsaA/B in the two strains, we found 30% and 20% increases in the levels of PsbA and the light-harvesting complex of PSII (LHCII), respectively, in the WT, but not in hf-2, at this early phase of S starvation, which was consistent with the behavior of Chl. The prominent difference between the two strains in LHCII subunits allows the interpretation that hf-2, having no SQDG, failed to synthesize LHCII. Meanwhile, the induction level of sulfate uptake activity in hf-2 was 45–65% lower than that in the WT throughout the first 6 h of S-starvation whereas the enhancement of ARS activity was similar for the WT and hf-2 (Fig. 5), which suggested that S released from SQDG is essential for the full induction of sulfate transport activity, but not of ARS activity. Of further interest was a 21% decrease in the Chl content in hf-2 at later times (6–36 h) owing to a quantitative reduction in PsaA/B (i.e. probably in the PSI complex) down to 46% (6–24 h) when the level of either Chl or PSI was almost unaltered in the WT (Figs 1h and 4b). Concomitantly, hf-2 was repressed by 40% in the amount of extended cell growth during S starvation for 24 h, relative to the WT (Fig. 1d). It should be emphasized that hf-2 facilitated the degradation of proteins such as rubisco (Fig. 4e).

Figure 5.

 Effects of the sulfoquinovosyl diacylglycerol (SQDG)-deficiency mutation on the induction patterns of sulfate uptake and arylsulfatase activities in Chlamydomonas reinhardtii. Cells grown normally or starved of sulfur (S) for 3 h or 6 h were used for measurement of the following activities. (a) Sulfate transporter activity was measured as the rate of uptake of [35S] sulfate into the cells, as described in the Materials and Methods section. (b) Arylsulfatase activity was spectroscopically measured with the use of 5-bromo 4-chloro 3-indolylsulfate as a substrate, as described in the Materials and Methods section. The values are the averages ± SE for three distinct groups of data. The significance of differences were evaluated using R 2.4.1 software with a Welch t-test, (http://www.r-project.org): *, < 0.05.

Discussion

Characterization of the induction system of SQDG degradation under S-starved conditions

In a seed plant, Brassica napus, sulfate was considered to be a major internal S source, as it occupied 42% to > 70% of the total S in the leaves and disappeared almost completely when the external S supply was withdrawn (Blake-Kalff et al., 1998). In Lemna minor, rubisco, which is preferentially degraded among proteins on S-starvation, was proposed to become a S source (Ferreira & Teixeira, 1992). In both reports, however, no direct evidence was presented as to the redistribution of S from the proposed storage molecules to other compounds, and the changes were followed for a rather long time (at least 2 d after onset of the S stress).

We previously demonstrated that SQDG is a major S-source for protein synthesis at the early phase of S-starvation within 6 h through characterization of a S budget in the cells (Sugimoto et al., 2007). The initial S pool of the water-soluble fraction was much smaller than that of SQDG. We therefore considered that sulfate or glutathione never predominates as an S-storing molecule in cells of C. reinhardtii. Furthermore, individual proteins detectable on SDS-PAGE showed no remarkable decreases in quantity, excluding a possibility that some particular protein serves as a major S source at this early stage of S starvation (Sugimoto et al., 2007). Similar rapid degradation of SQDG was previously reported for another green alga, Chlorella ellipsoidea, on S starvation (Miyachi & Miyachi, 1966).

Here, the induction of SQDG degradation was found to require both the transcription and translation of gene(s) on the nuclear genome (Fig. 2a), which might encode the enzyme(s) that catalyse the degradation of SQDG and/or regulatory protein(s) for the activity. The absolute requirement of translation (Fig. 2c) allowed us to interpret that at least the enzyme activity that first attacks SQDG, which had been gained during the initial 2 h, was unstable and needed to be sustained through continued synthesis of the enzyme responsible or its activator protein. Meanwhile, the cancellation of the transcriptional requirement (Fig. 2b) might indicate that mRNA for the degrading enzyme and/or its activator that had accumulated during the initial phase of S-starvation was stable enough to support the entire SQDG degradation.

Nuclear genes, transcript levels of which are upregulated by SAC1, are highly responsive as early as 2 h after the onset of S-starvation (Zhang et al., 2004). SAC1 might be involved in upregulation of the levels of mRNA of genes responsible for SQDG degradation (Fig. 2d), and the action of SAC1 within the initial 2 h might lead to no further requirement of nuclear transcription (Fig. 2b). Conversely, the role of SAC3 as a negative regulator is improbable for SQDG degradation because the sac3 disruptant, compared with the WT, exhibited no stimulated degradation activity under S-replete conditions (Fig. 2e). Also of note was the lack of complete repression, but low induction of SQDG degradation in either the sac1 or sac3 disruptant. One signaling pathway at least, other than that including SAC1, may also be responsible for the transcriptional induction of the gene(s) for SQDG degradation: it is possible that SAC3 participates in this pathway. The identification of both SAC proteins as positive regulators would be a basis for elucidation of the mechanism for induction of SQDG degradation at the molecular level.

Physiological significance of the role of SQDG as an internal S-supplier for protein synthesis

It seems that the genome of hf-2 is intact except for the impaired gene for SQDG synthesis, as similar phenotypes such as growth comparative to that of the WT under normal conditions and enhanced sensitivity of photosystem (PS)II to DCMU were observed for hf-2 (Fig. 1c; Minoda et al., 2002; Sato et al., 2003a) and another SQDG-deficient mutant of C. reinhardtii produced through insertion mutagenesis in the SQD1 gene (Riekhof et al., 2003). In the present study, we used OD730, but not cell number, to compare cell growth between the WT and hf-2, in view of the role of SQDG in cell division (Fig. 3c), which will be studied in the future. By demonstrating defects in protein synthesis and thereby in accumulation of total proteins in hf-2 (Fig. 1f,i,j), this study reinforced validity of the notion of SQDG as a major S source for protein synthesis. In this context, the facilitated degradation of rubisco that accounts for a similar level of S to that in SQDG (Sugimoto et al., 2007) in hf-2 can be regarded as compensation for the mutational loss of SQDG as the S reservoir (Fig. 4e), but in vain (Figs 1d,f,h,4d). The results corroborated that the protein fraction that can be used as an S-supplier in place of SQDG is too small to support proper acclimatization to S-starvation. Meanwhile, more rapid induction of SQD1 mRNA in hf-2 than in the WT would reflect that hf-2, having no SQDG to degrade, underwent more severe S-starvation (Fig. 3c).

The use of hf-2 also enabled us to produce three lines of evidence that induction of SQDG degradation to ensure S-source is essential for the proper acclimation process in C. reinhardtii. First, hf-2 failed to synthesize Chl or LHCII (Figs 1h,4d). S-starved cells of C. reinhardtii were previously implied to utilize LHCII for dissipation of excess light energy as heat, and for the direction of more of the light energy away from PSII toward PSI via a transition to State 2, thereby maintaining cyclic electron flow at a proper rate for ATP synthesis (Wykoff et al., 1998). Second, hf-2, compared with the WT, had repressed induction of sulfate transport activity, which depends on the synthesis of proteins (Fig. 5a; Yildiz et al., 1994). This observation would reflect a defect in protein synthesis in hf-2 regarding sulfate transporters themselves and/or their regulators. No deleterious effects of SQDG-deficient mutation on the induction of ARS activity that also requires translation (Fig. 5b; Schreiner et al., 1975) would be exceptional, and interpretation of this observation awaits quantitative comparison of mRNA and product levels of the ARS genes with those of the genes for sulfate transporters. Third, hf-2 showed more severely repressed growth compared with the WT during S-starvation (Fig. 1d). We consider that SQDG is essential as an S-source for the proper synthesis of proteins encoded by house-keeping genes such as Chl–protein complexes, and of proteins that should be highly induced, such as sulfate transporters, for the construction of a basic system for acclimatization to S-starved conditions.

Induction of SQDG synthesis activity under S-starved conditions and its physiological importance

Our results concerning upregulation of the SQD1 mRNA and SQDG synthetic capacity (Fig. 3a,b) are consistent with a previous report of DNA microarray analysis (Zhang et al., 2004), but can be taken to be confirmative in view of our methodological exclusion of cross-hybridization. The DNA microarray analysis of C. reinhardtii also revealed upregulation of the level of SQD2a mRNA (Zhang et al., 2004). It is probable that the induced level of SQD2a was below the detection limit of our systems. Meanwhile, it seems likely that the upregulation of SQD1 mRNA is triggered by the unavailability of external S itself in view of no involvement of feedback regulation through sensing of the SQDG (Fig. 3c). Collectively, the upregulation of the SQDG synthetic capacity seemed to allow a metabolic equilibrium between SQDG synthesis and its degradation, thereby keeping SQDG at the minimal, but necessary level (3% of the initial level even after 24 h of S-starvation; Sugimoto et al., 2007) for some role in specific coping with the S-deficient stress.

Induction of SQD1 mRNA for upregulation of SQDG synthesis, distinct from that of SQDG degradation, required the synthesis of proteins encoded not only on the nuclear genome but also on the chloroplast genome (Fig. 3e,f). The proteins encoded on the nuclear genome, for example, might be factors for SQD1 transcription. Meanwhile, it was previously reported that enhancement of the levels of transcripts was partially repressed for SQD1 and almost completely repressed for SQD2a in cells of the sac1 disruptant, as seen on DNA microarray analysis (Zhang et al., 2004). As far as signaling components are concerned, there may be some overlap between the induction systems of SQDG synthesis and its degradation.

The similar behavior of PSI or PSII activity for the WT and hf-2 at the early phase of S starvation (Fig. 4a) coincided with the previously reported regulation of these activities in C. reinhardtii (Wykoff et al., 1998). The decrease in PSII activity would help prevent the production of active oxygen species through repression of the linear electron flow of photosynthesis whereas no deleterious effect on the PSI activity would allow the production of ATP necessary for the acclimatization process through maintenance of the cyclic electron flow at a certain level (Wykoff et al., 1998). Therefore, for regulation of these activities, we could obtain no evidence of an active role of SQDG as a membrane lipid, that is, there was no physiological significance of the induction of SQDG synthesis.

By contrast, hf-2 showed instability of the PSI complex at later times during S-starvation (6–36 h; Figs 1h, 4b). The low level of SQDG maintained in the WT through upregulation of SQDG synthesis may stabilize membrane proteins such as the PSI complex whereas the complete absence of SQDG in hf-2 may lead to destruction of the PSI complex at the later phase of S-starvation (i.e. when the damaged complex would hardly be repaired because of the harsh repression of regeneration of its subunits and/or Fe-S clusters). The aberrant feature of PSI in hf-2 might otherwise be caused indirectly by the failure of the cells to synthesize proteins for early responses. In any case, the instability of the PSI complex should result in reduced production of ATP and reducing power. Thus, in hf-2, the lesion in the PSI complex, together with the impaired synthesis of proteins, would lead to remarkable repression in cell growth during S starvation for 24 h, relative to in the WT (Fig. 1d).

Sulfoquinovosyl diacylglycerol has been co-purified with several protein complexes in a variety of photosynthetic organisms, such as chloroplast ATP synthase from a green alga, Dunaliella salina (Pick et al., 1985), PSII complex from C. reinhardtii (Sato et al., 1995a), LHCII from a seed plant, Nicotiana tabacum (Gasser et al., 1999), and cytochrome b6f complex from C. reinhardtii (Stroebel et al., 2003). However, detection of SQDG in these complexes in vitro is accompanied by possible artifacts, and, above all, such confirmation by itself has not uncovered any concrete roles of SQDG. This study, involving the use of hf-2, is novel to present in vivo evidence of a role of SQDG as a stabilizer of the PSI complex under S-starved conditions. In this context, it is notable that, in C. reinhardtii, SQDG was found to participate in stabilization of PSII activity against heat stress and in recovery from heat-induced inactivation of PSII, probably through characterization of hf-2 (Sato et al., 2003a). However, the synthesis of phosphatidylglycerol (PG), which, together with SQDG constitutes acidic lipids in thylakoid membranes, was enhanced in WT cells of C. reinhardtii under S-starved conditions, the content of PG being increased from 2.9% to 14.5% relative to total lipids in thylakoid membranes, as if to compensate for the loss of the negative charge on the membranes caused by SQDG degradation (7.9% to 0.5%; Sugimoto et al., 2008). Thylakoid membranes isolated from S-starved cells of C. reinhartdii exhibited functional damage to PSI when treated with phospholipase A2 for partial removal of PG, which allowed us to conclude that the increased content of PG is useful for stabilization of the PSI activity (Sugimoto et al., 2008).

Collectively, these two acidic lipids might support PSI both structurally and functionally for the duration of ATP synthesis under S-starved conditions, through interaction with PSI as membrane lipids, and, as regards SQDG, by supplying the S source for the synthesis of LHCII. In this context, it is of interest that a variety of photosynthetic organisms, with phosphorus-limiting stress, exhibit decreases in the contents of phospholipids, including PG, with an increase in SQDG content, in terms of the regulatory mechanism for lipid metabolism and its physiological relevance (Sato, 2004). The molecular mechanism by which SQDG together with PG supports PSI will be the subject of a future study.

Conclusion

Induction of SQDG degradation on S starvation was found to depend on SAC1 and SAC3, with the prerequisites of transcription and translation of the nuclear gene(s). Concomitantly, the level of the SQD1 mRNA was shown to increase to upregulate the SQDG synthesis capacity through a different induction system from that of SQDG degradation. Moreover, the use of hf-2 enabled us to show the importance of the regulation of SQDG metabolism for acclimatization of C. reinhardtii to S-deficient stress, from which we deduced the dual roles of SQDG: as a S-storage lipid that supports the synthesis of proteins during the early phase of S starvation, and as a membrane lipid that stabilizes the PSI complex at a later phase. This study will be the foundation for a fuller understanding of the mechanism by which photosynthetic organisms regulate S metabolism, and of how the system for acclimatization to S-deficient stress is established in thylakoid membranes. Such information will deepen the comprehension of S metabolism in terms not only of macronutrients, but also of energy production.

Acknowledgements

This work was supported in part by Grants-in-Aid from the Ministry of Education, Culture, Sports, Science, and Technology of Japan (MEXT), the Promotion and Mutual Aid Corporation for Private Schools of Japan, Scientific Research for Plant Graduate Students from the Nara Institute of Science and Technology supported by MEXT, and a Sasakawa Scientific Research Grant from the Japan Science Society. We are also indebted to Dr M. Ikeuchi (The University of Tokyo), and Drs K. Sonoike (The University of Tokyo) and I. Enami (Tokyo University of Science) for the gifts of the PsbA and PsaA/B antibodies, respectively.

Ancillary