Species-specific roles of sulfolipid metabolism in acclimation of photosynthetic microbes to sulfur-starvation stress

Photosynthetic organisms utilize sulfate for the synthesis of sulfur-compounds including proteins and a sulfolipid, sulfoquinovosyl diacylglycerol. Upon ambient deficiency in sulfate, cells of a green alga, Chlamydomonas reinhardtii, degrade the chloroplast membrane sulfolipid to ensure an intracellular-sulfur source for necessary protein synthesis. Here, the effects of sulfate-starvation on the sulfolipid stability were investigated in another green alga, Chlorella kessleri, and two cyanobacteria, Synechocystis sp. PCC 6803 and Synechococcus elongatus PCC 7942. The results showed that sulfolipid degradation was induced only in C. kessleri, raising the possibility that this degradation ability was obtained not by cyanobacteria, but by eukaryotic algae during the evolution of photosynthetic organisms. Meanwhile, Synechococcus disruptants concerning sqdB and sqdX genes, which are involved in successive reactions in the sulfolipid synthesis pathway, were respectively characterized in cellular response to sulfate-starvation. Phycobilisome degradation intrinsic to Synechococcus, but not to Synechocystis, and cell growth under sulfate-starved conditions were repressed in the sqdB and sqdX disruptants, respectively, relative to in the wild type. Their distinct phenotypes, despite the common loss of the sulfolipid, inferred specific roles of sqdB and sqdX. This study demonstrated that sulfolipid metabolism might have been developed to enable species- or cyanobacterial-strain dependent processes for acclimation to sulfate-starvation.


Introduction
Sulfoquinovosyl diacylglycerol (SQDG), which includes sulfoquinovose as a negatively charged head group, is conserved widely among oxygenic photosynthetic organisms (reviewed in e.g., [1]). This anionic lipid, which accounts for ca. 10-20 mole% of total cellular glycerolipids, contributes to construction of the membranes of chloroplasts in plants and their postulated ancestor, cyanobacteria, with predominant localization in their thylakoid membranes. The SQDG synthesis system in oxygenic photosynthetic organisms includes UDP-sulfoquinovose synthase (encoded by SQD1 in plants or sqdB in cyanobacteria) and SQDG synthase (encoded by SQD2 in plants or sqdX in cyanobacteria). These two enzymes catalyze successive reactions, PLOS  i.e., the former combines sulfite and UDP-glucose for UDP-sulfoquinovose synthesis, and then the latter transfers the sulfoquinovose moiety of UDP-sulfoquinovose to diacylglycerol for SQDG synthesis. Interestingly, physiological characterization of SQDG-deficient mutants, including those disrupted in the above genes for SQDG synthesis, has thus far shown speciesdependent roles of SQDG under normal growth conditions or even strain-dependent ones within species: e.g., SQDG is essential for cell growth in a cyanobacterium, Synechocystis sp. PCC 6803 (herein referred to as Synechocystis, [2]), but not in other oxygenic photoysnthetic organisms including cyanobacteria Synechococcus elongatus PCC 7942 (herein referred to as Synechococcus, [3]) and Synechococcus sp. PCC 7002 [4], a green alga, Chlamydomonas reinhardtii [5], and a seed plant, Arabidopsis thaliana [6]. Meanwhile, SQDG is required for the normal functionality of photosystem II in Synechocystis [2] and Chlamydomonas reinhardtii [7], but not in Synechococcus [3], Synechococcus sp. PCC 7002 [4] or Arabidopsis thaliana [6]. Photosynthetic microbes, which could account for about a half of global net primary production [8], generally face deficiency in phosphate, the preferred environmental phosphorus (P)source (e.g., [9]). The SQDG content is elevated in photosynthetic microbes in response to deficiency in phosphate (e.g., [10]). The increase of SQDG is accompanied by a decrease in the content of another anionic lipid, phosphatidylglycerol (PG), whereby the summed content of these two anionic lipids is maintained at a certain level. The similar lipid-remodeling in response to phosphate-limitation occurs in seed plants like Arabidopsis thaliana, and also in anoxygenic photosynthetic and non-photosynthetic root nodule-forming bacteria that possess SQDG [11][12][13]. It therefore is generally accepted that SQDG substitutes for PG for proper functionality of membranes and reduction in P-quota (e.g., [1]). Consistent with this idea, sqdB/SQD1-or SQD2-disrupted mutants including those of Synechococcus, C. reinhardtii, and A. thaliana, which are devoid of SQDG that should otherwise be increased in quantity, were much more severely repressed in cell growth during acclimation to phosphate-limitation, relative to the wild types [3,6,14].
Sulfur (S) is one of macronutrients for photosynthetic organisms, similar to P, and is the essential component of proteins and sulfolipids. Distinct from phosphate, sulfate, the preferred S-source for photosynthetic microbes, however shows extremely uneven distribution in aquatic habitats (reviewed in [15]): although the oceans typically contain sulfate at around 29 mM, i.e., supra-optimal concentrations for photosynthetic microbes, freshwater environments were highly variable in sulfate concentration such that lake sulfate ranged from as low as 10 to 1000 μM, dependent on the level of industrialization around the surrounding regions. Industrial emission of SO 2 into the atmosphere could elevate the sulfate content of freshwater environments, which have however been decreased with the enforcement of legislations that limit SO 2 emission in North America and Europe [16][17][18]. Therefore, there might be a growing trend that photosynthetic microbes in freshwater face S-deficiency. S-starvation was experimentally found to affect SQDG metabolism in a freshwater microbe, C. reinhardtii. Upon sulfate-deficiency, SQDG is almost completely degraded, whereby a major intracellular S-source for protein synthesis is ensured at an early stage, accompanied by an increase in the PG content [19]. Simultaneously, SQDG synthesis activity is enhanced through up-regulation of SQD1 mRNA, which seems to be responsible for maintenance of SQDG at 5% of the normal level as a molecular tool to structurally stabilize the photosystem I complex [20]. However, behavior of SQDG metabolism and its physiological significance have never been examined under S-starved conditions in any freshwater microbes other than C. reinhardtii.
In this study, effects of S-starvation on SQDG metabolism was investigated in freshwater microbes such as cyanobacteria, Synechococcus and Synechocystis, and a green alga, Chlorella kessleri, to gain more insight into the taxonomical distribution of SQDG degradation ability. Also, disruptants of Synechococcus as to the sqdB and sqdX genes (ΔsqdB and ΔsqdX, respectively; [21]) were characterized as to the behavior of cells for acclimation to S-starvation stress to determine the physiological significance of SQDG synthesis. Obtained results will be discussed from an evolutionary aspect for oxygenic photosynthetic organisms.
When needed, C. kessleri or cyanobacterial cells were universally labeled with 35 S by culturing in the above [ 32 S]sulfate-replete medium in the presence of [ 35 S]sulfate (231 MbqÁnmol -1 , 3.7 MbqÁmL -1 ). The labeled cells were washed three times with the S-free medium, and then shifted to 32 S-replete or S-free conditions for further growth.

Determination of pigment and lipid contents
The contents of Chl a and b were measured in C. kessleri, whereas Chl a and PBS were measured in cyanobacterial strains, as we previously described [24,25]. Total lipids were extracted from the cells and subjected to TLC for separation into individual lipid classes, as previously described [7]. Total lipids and each lipid class were used for the synthesis of fatty acid methyl esters with 5% anhydrous methanolic HCl, and analyzed by capillary GLC, as described previously [7].

Measurement of radioactivity of [ 35 S]SQDG
During culturing of 35 S-labeled cells under 32 S-replete or S-free conditions, a portion of the culture was used for measurement of the radioactivity as total radioactivity. Another portion was used for extraction of total lipids, and subsequent separation of SQDG on a TLC plate and measurement of its radioactivity in the silica gel that was scraped off. The radioactivity of the culture and SQDG were measured with a liquid scintillation counter (LSC-6100 ALOKA, Tokyo, Japan), as described previously [20]. The radioactivity of SQDG was estimated relative to the total radioactivity.

Semi-quantitative reverse transcriptase (RT)-PCR
RT-PCR was performed for the Synechococcus WT for investigation of expression of sqdB, sqdX, and sqdBX operon at the transcript level, with the use of the rnpB gene as an internal control, as described previously [25]. The primer sets used were shown in supplementary S1 Table.

Results
Induction of SQDG degradation in a green alga, C. kessleri C. reinhardtii cells when starved of S exhibit almost complete degradation of preexisting SQDG within a day to ensure an internal S-source, which seems reasonable in view of dispensability of SQDG for its normal cell growth [5,19]. Here, this SQDG-degradation ability was investigated in another green alga, C. kessleri. S-starvation caused C. kessleri cells to reveal typical symptoms [19], i.e., repression of Chl a and b accumulation (-S, cf., +S, Fig 1A), which probably reflected a delay in the synthesis of global proteins including photosystem complexes. S-replete cells of C. kessleri contained chloroplast lipids, monogalactosyl diacylglycerol (MGDG), digalactosyl diacylglycerol (DGDG), SQDG, and PG, and extrachloroplast lipids, phosphatidylcholine (PC) and phosphatidylethanolamine (PE) (Fig 1B and 1C; [22]). S-starved cells, however, lost SQDG, and concomitantly increased the PG content by 2-fold as if to quantitatively complement the SQDG-loss (Fig 1B and 1C).
A radio-labeling experiment was then conducted to determine whether or not this defect in SQDG was due to positive degradation of it (Fig 2A). C. kessleri cells, which had been grown in the presence of [ 35 S] sulfate for universal labeling, were shifted to non-label medium with or without [ 32 S] sulfate, for further growth. The radioactivity of SQDG initially amounted to 5% of the total radioactivity in S-replete cells. S-replete conditions allowed the radioactivity of SQDG to be maintained at almost the same level during cell growth for 2 days, which indicated the high stability of SQDG. In contrast, the radioactivity gradually decreased under S-starved conditions, which proved that C. kessleri cells positively degrade SQDG in response to Sstarvation.

High stability of SQDG in cyanobacterial cells under S-starved conditions
The effect of S-starvation on the SQDG content was then examined in Synechocystis, which requires SQDG for cell growth under normal conditions [2]. S-starvation caused markedly retarded growth of cells, Chl being maintained at the initial level throughout S-starvation for 4 days (Fig 3A and 3B). Concerning lipids, the most abundant lipid in S-replete cells was MGDG, followed by SQDG, DGDG, and PG in that order (Fig 3C and 3D; [23]). It was of note that S-starvation caused no drastic decrease in the SQDG content in Synechocystis. Labeling experiments with [ 35 S] sulfate demonstrated that the radioactivity of SQDG initially corresponded to 6% of the total radioactivity in S-replete cells, and that the radioactivity of SQDG was maintained at this initial level irrespective of the presence or absence of an environmental S-source ( Fig 2B).
Synechococcus, in which SQDG is dispensable for cell growth [3], also showed delayed cell growth in response to S-starvation, concomitantly with repressed Chl accumulation (Fig 4A  and 4B). Synechococcus, similar to Synechocystis, exhibited no drastic decrease in the SQDG content (Fig 4C and 4D). Moreover, radio-labeling experiments demonstrated that the initial radioactivity of SQDG (8% of total one) was maintained at almost the same level throughout subsequent cell growth under S-starved conditions as well as under S-replete ones (Fig 2C). Collectively, SQDG is highly stable in cyanobacteria under S-starved conditions, irrespective of whether SQDG is essential or not.
Physiological significance of the genes for SQDG synthesis in Synechococcus during acclimation to S-starvation The ΔsqdB and ΔsqdX disruptants that we generated in Synechococcus retained sqdX and sqdB functionality, respectively, at some level [4], which would allow us to determine the independent roles, if any, of sqdB and sqdX, as well as the role of the final product, SQDG. Both disruptants, similar to the WT, showed relatively healthy growth under S-replete conditions such that the doubling time was 8.3 or 11.1h in ΔsqdB and ΔsqdX, respectively, c.f., 8.1h in the WT. Under S-starved conditions, the ΔsqdB disruptant showed retardation of cell growth and Chl accumulation in a manner similar to the WT (Fig 5A and 5B). However, phycobilisome (PBS) was retained in ΔsqdB at 30% of the initial level after 3-day starvation for S, in contrast to in the WT in which it was almost completely degraded (Fig 5C), as previously reported by Collier and Grossman [26]. The PBS degradation in WT markedly decreases the antenna size of photosynthesis to down-regulate photosynthetic electron transport activity, whereby the reducing potential of S-starved cells decreased to meet the low capacity of their cellular metabolism [27]. The PBS degradation in ΔsqdB was more seriously repressed such that >50% was retained with illumination with light at a milder intensity (Fig 5D, 14 μmol photons m -2 s -1 , cf., 24 μmol photons m -2 s -1 in Fig 5C). Meanwhile, the ΔsqdX disruptant showed more pronouncedly delayed cell growth than the WT or the ΔsqdB one, concomitantly with an abnormal quantitative decrease in Chl, i.e., that in the PSI and/or PSII complexes (Fig 5A and 5B). However, PBS degradation proceeded normally in the ΔsqdX disruptant as in the WT (Fig 5C). Overall, it was found that the ΔsqdB and ΔsqdX disruptants, despite their common lack of SQDG, were disturbed in the physiological response in quite distinct manners. We also found that the expression level of sqdB or sqdX was markedly up-regulated in the WT during S-starvation through RT-PCR, especially 24 and 48 h after the onset of S-starvation ( Fig 4E).
The WT, ΔsqdB or ΔsqdX cells starved of S were then cultured in the normal medium for investigation of recovery growth (Fig 5E). Despite defects in normal regulation of photosynthetic components and/or a concomitant defect in acclimating cell growth, ΔsqdB and even ΔsqdX cells (Fig 5E, right) exhibited normal recovery of cell growth with pigmentation. Sstarved cells of Synechococcus thus seemed to show growth recovery independently of the system of SQDG synthesis, and its final product, SQDG.

Evolutionary consideration of SQDG degradation under S-starvation
SQDG included S that accounted for 5-8% of the total cellular S in C. kessleri, Synechocystis, and Synechococcus, and was found to be stabilized in these respective species under S-replete conditions (Fig 2). These observations were compatible with the previous one for C. reinhardtii [19]. Through characterization of physiological defects in their SQDG-deficient mutants [1,4], the physiological roles of SQDG in photosynthesis and/or cell growth has been found to depend on microbial species. We consider that SQDG is bound to the PSII core complex for the structural and functional integrity of the complex in Synechocystis and C. reinhardtii [2,7]. In Synechocystis, the high stability of SQDG would therefore be useful for maintenance of the integrity of PSII. It will be a future subject to specify SQDG-dependent component(s) for DNA synthesis for progression of the cell cycle in Synechocystis [28], and evaluate the significance of the high stability of SQDG. Meanwhile, it seems that the high stability of SQDG demands properly regulated SQDG synthesis in Synechococcus, since an increase of the SQDG content through over-expression of the sqdBX operon disturbed its cell size determination [29].  Under S-starved conditions, SQDG degradation was induced probably for insurance of an intracellular S-source in C. kessleri as well as in C. reinhardtii [19,20]. In contrast, the stability of SQDG was maintained at a high level in S-depleted cells of either cyanobacterial strain (Fig  2). These lines of evidence would be consistent with the idea that the ability to induce SQDG degradation under S-starved conditions could have been obtained not by cyanobacteria, but by algae, in the course of the evolution of oxygenic photosynthetic organisms, and already by the phase of diversification of green algae. . (e) Cells starved of S for 3 days were shifted to S-replete conditions for recovery from the retarded growth. Left, S-starved cells of the WT (white bars) and ΔsqdB disruptant (grey bars) were initially adjusted to 0.1 in OD 730 value with S-replete medium. Right, S-starved cells of the WT (white bars) and ΔsqdX disruptant (grey bars) were initially adjusted to an OD 730 value of 0.2 with S-replete medium. The two-fold higher initial value of OD 730 for the ΔsqdX disruptant relative to that for ΔsqdB one was to avoid possible photodamage in the ΔsqdX one, owing to its severe damages to cell growth and Chl accumulation (Fig 5A  and 5B). Broken lines correspond to initial levels in the cultures. The inset in (e) shows photographs of 2-day cultures of the WT and ΔsqdX (ΔX). The values shown are the averages ± SD for three distinct groups of data. The significance of differences was evaluated by Student's t test. *P < 0.05. https://doi.org/10.1371/journal.pone.0186154.g005

Physiological significance of SQDG synthesis in Synechococcus cells during acclimation to S-starvation stress
Our observations as to the high stability of SQDG under S-starved conditions in Synechocystis and Synechococcus might reflect some crucial role of SQDG in cyanobacteria. In Synechococcus, however, it was definitely not SQDG, but the respective genes for SQDG synthesis, i.e., the sqdB and sqdX genes, that are necessary for regulatory rearrangement of the photosynthetic apparatus and/or acclimating cell growth (Fig 5). These requirements of sqdB and sqdX would be compatible with the high levels of their expression in S-starved cells (Fig 4E). PBS degradation proceeds through the nbl pathway including signaling components like NblR under Sstarved conditions [27]. The sqdB gene might be involved directly or indirectly in some process required for S-starvation induced PBS degradation, which, however, would be unnecessary for the ongoing cell growth during acclimation. Meanwhile, it is probable that the sqdX gene is required for some fundamental process of physiology for the ongoing cell growth during acclimation to S-starvation, and also for some mechanism that structurally stabilizes the PSI and/or PSII complexes. It thus seems that photosynthesis is affected in distinct manners between Sstarved cells of sqdB and sqdX. In an anoxygenic photosynthetic bacterium, Rhodobacter sphaeroides, a mutant impaired in the sqdD gene for SQDG synthase, which is structurally unrelated to that of oxygenic photosynthetic organisms, exhibited a much larger pool of UDPsulfoquinovose than in the WT [30]. It will be a future project to investigate the molecular basis of how the distinct phenotypes of ΔsqdB and ΔsqdX cells were caused, including their distinctly disturbed patterns of photosynthetic parameters and those of SQDG metabolism.
Mutants impaired in lipid metabolism in cyanobacteria or chloroplasts of plants have thus far been examined as to their impact on the structural and functional integrity of the oxygenic photosynthetic apparatus, whereby, the correlation of lipids, i.e., the final products of the lipid synthesis pathways, to photosynthesis has been discussed (reviewed in, e.g., [31]). Such studies include those concerning SQDG-and/or PG-deficient mutants of Synechocystis or C. reinhardtii, which successfully verified the roles of these lipids in photosynthesis, through chemical complementation of the photosynthetic defects in the mutant by supplementation of the deficient lipid [2,23,[32][33][34]. The causes of the defective phenotypes of photosynthesis that have so far been reported in other lipid mutants might have to be carefully re-evaluated, including possible accumulation of the substrates of enzymes encoded by the damaged genes.
In Synechocystis, its ΔsqdB cells, and probably ΔsqdX ones that have not been reported for its generation, have to be cultured in the presence of externally supplemented SQDG, owing to this cyanobacterial-strain specific essentiality of SQDG [2]. This particular feature becomes a hurdle to investigating the responsibility of sqdB and sqdX genes for acclimation to S-starved conditions. However, it is supposed that the role of sqdB under S-starved conditions in Synechocystis, if any, should exclude the responsibility for PBS degradation, since PBS is never subjected to positive degradation in S-starved cells of Synechocystis (supplementary S1 Fig, [35]). In other words, this study revealed a novel cyanobacterial-strain dependent role of SQDG synthesis or sqdB in Synechococcus, which might have been gained only by some group of cyanobacteria during their evolutionary diversification, or by cyanobacteria at an early phase of their evolution, followed by its persistent maintenance or loss during their evolutionary diversification. In either case, the role of sqdB in PBS degradation was not passed on to its counterpart, SQD1, in a green lineage where PBS was replaced completely by light-harvesting Chl-protein complexes [36]. At first, cyanobacteria could have integrated the SQDG synthesis system into essential physiological processes including acclimation to S-deficiency or DNA synthesis whereas green algae, after the evolutionary loss of such cyanobacterial roles of SQDG synthesis system, could have developed the system of SQDG degradation.

Conclusions
This study demonstrated that freshwater photosynthetic microbes, which might face S-deficiency, utilize the SQDG metabolism for acclimation to S-deficiency in species or even cyanobacterial-strain dependent manner. This situation, which is reminiscent of the speciesdependent physiological significance of SQDG under S-replete conditions, is quite distinct from the universal role of SQDG to substitute for PG under P-deficient conditions. Our findings will be a foundation for a fuller understanding of the species-dependent mechanisms by which photosynthetic microbes acclimate to S-deficiency stress, and for consideration of their changes through evolution of photosynthetic organisms. In line, with our study as a starting point, it is expected that research on lipid role will change to include consideration of their dependency on the organism, and also on the respective genes for lipid synthesis in an organism.
Supporting information S1 Table. Primer sets used for semi-quantitative PCR analysis of transcripts as to the genes for SQDG synthesis.