Transcriptional regulator PrqR plays a negative role in glucose metabolism and oxidative stress acclimation in Synechocystis sp. PCC 6803

Plant and cyanobacteria can perceive signals from soluble sugar and reactive oxygen species (ROS) and then coordinate gene expression under stress acclimation, but the underlying mechanism remains unclear. In this study, we found that the transcriptional factor PrqR (Slr0895) in Synechocystis can perceive signals from ROS generated after shifting from prolonged darkness with glucose into high-light. The deletion mutant (DprqR) showed increased growth rate and decreased ROS content, whereas the complementary strain (CprqR) restored the growth characteristics, phenotypes and ROS status of WT, thereby establishing PrqR as a negative regulator of ROS.LC/GC-MS-based metabolic profiling also showed active ROS mitigation in DprqR mutant. Further study by qRT-PCR, ChIP-PCR and deletion of both prqR and prqA (DprqR-DprqA mutant) revealed that PrqR exerts this negative regulation of ROS removal by controlling the expression of sodB and prqA (slr0896). Furthermore, PrqR also found to control glucose metabolism by regulating a positive regulator of glucose metabolism, sigE, and its regulons. Results suggest that PrqR was involved in perceiving signals from ROS under physiological condition, as well as in regulating stress removal and glucose metabolism.


Results
Phenotypic analysis of Synechocystis wild type and prqR mutants. A previous study found that the Synechocystis mutant Prq20 (L17Q in prqR transcription factor) was resistant to methyl viologen, a strong ROSgenerating chemical 26,27 . In the current study, to further determine the roles of the prqR gene in ROS removal and glucose metabolism at the molecular level, a prqR-disrupted mutant (DprqR) was generated by inserting a kanamycin cassette under cpcB promoter into the Synechocystis (glucose-intolerant strain) chromosome. For complementation, a strain with prqR gene overexpressed under a psbA2 promoter along with the streptomycin cassette was also created in the DprqR mutant (CprqR) (the vectors and strains generated in this study are presented in Supplementary Fig. S1A). DprqR and CprqR were completely segregated on BG-11 plate supplemented with kanamycin and kanamycin + streptomycin (up to 100 μ g/mL), respectively. The segregation was confirmed by PCR (Supplementary Figs S1B and S1C). WT, DprqR and CprqR showed normal growth in BG-11 liquid media in the presence of light at 25 μ mol m −2 s −2 , with shaking at 130 rpm and temperature of 30 °C (Fig. 1A). To assess glucose-dependent growth, we added 5 mM D-glucose into the culture when its OD 730 was around 0.4 and then placed the culture into complete darkness [GLC(+ )DARK] with shaking for different durations. Through the growth courses, the DprqR mutant grew almost similarly to WT and CprqR under GLC(+ )DARK ; with subsequent placement of the cultures into high light, GLC(+ )DARK-LIGHT(at 80 μ mol m −2 s −2 ) led to approximately twofold increased growth of the DprqR mutant, whereas WT and CprqR showed significantly decreased growth as well as decolorization (Fig. 1B-D). To determine whether the phenotype attributed to the osmotic stress exerted by D-glucose, we repeated the experiments with D-sorbitol, which can produce an osmotic pressure similar to that D-glucose. However, all these strains grew equally normal ( Supplementary Fig. S2), suggesting that the phenotype was not ascribed to osmotic stress. To determine whether the cells are dying under GLC(+ )DARK and GLC(+ )DARK-LIGHT conditions, we monitored the percentages of dead cells at two time points: at third day under GLC(+ )DARK (72 h) condition and first day (24 h) under GLC(+ )DARK-LIGHT condition. However, the results showed no significant changes in terms of cell death rates between WT and DprqR ( Supplementary  Fig. S3). Notably, the growth and chlorophyll content were found remarkably different in between DprqR and WT (and CprqR) after shifting from GLC(+ )DARK to the GLC(+ )DARK-LIGHT conditions (Fig. 1D,E).

ROS decreased in DprqR mutant.
Assuming that a sudden change of metabolism from glucose-dependent [GLC(+ )DARK] condition to both glucose-and light-dependent [GLC(+ )DARK-LIGHT] conditions would generate ROS, we determined the total ROS contents at three time points: before placement under GLC(+ )DARK condition (normal cells), after 72 h under GLC(+ )DARK and after 48 h in GLC(+ )DARK-LIGHT condition. The total ROS contents were measured using membrane-permeant fluorescence indicator 5-(and-6)-chloromethyl-2′ , 7′ -dichlorodihydro fluorescein diacetate, CM-H2DCFDA (Invitrogen, Life technology). The results showed that under normal and GLC(+ )DARK conditions, the ROS contents of WT, DprqR and CprqR were at similar levels ( Fig. 1F,G), although a slightly higher accumulation of ROS in CprqR under normal condition was found and may be ascribed to the stronger psbA2 promoter. However, significantly (more than twofold) higher ROS contents in WT and CprqR were found than those in DprqR under GLC(+ )DARK-LIGHT condition after 48 h (Fig. 1H). In addition, no significant ROS accumulation difference was detected in WT, CprqR and DprqR in similar dark-light regime in the absence of glucose, that is, GLC(− )DARK and GLC(− )DARK-LIGHT ( Supplementary Fig. S4). This finding suggests that GLC(+ )DARK-LIGHT can cause ROS accumulation in Synechocystis cells, whereas the deletion of the prqR gene affected ROS mitigation.
Cyanobacteria exhibit diverse defense systems against oxidative stress, such as superoxide dismutase, which can remove superoxide radicals 4 . Synechocystis encompasses only one type of dismutase, that is, iron superoxide dismutase (FeSOD). Considering the role of FeSOD in ROS removal, we determined the expression order of sodB by qRT-PCR. The results showed that the sodB expression was upregulated in DprqR compared with WT at 24 and 48 h but equalized at 72 h under GLC(+ )DARK condition. Subsequent shifting of the culture under GLC(+ ) DARK-LIGHT condition caused about 100-fold increase of sodB in DprqR strain compared with WT ( Fig. 2A). To ensure that this fluctuation of the sodB expression in DprqR was ascribed to glucose rather than only high light, we maintained the GLC(− )DARK and GLC(− )DARK-LIGHT conditions but no significant upregulation or downregulation was observed (Fig. 2B).
PrqR binds to the promoter of prqA and negatively regulates the expression of prqA. In the Synechocystis chromosome, prqA located downstream of prqR encodes a hypothetical protein belonging to the multidrug resistance protein of the MATE family. Analysis of gene expression showed that prqA was upregulated in DprqR by threefold and fourfold than the WT under GLC(+ )DARK and GLC(+ )DARK-LIGHT conditions, respectively (Fig. 2C). However, these differential expressions was not obtained in the absence of glucose (Fig. 2D). To determine whether prqA was the regulatory target of PrqR, we overexpressed prqR with a His-tag at its C-terminus under cpcB promoter in the WT. The expression of prqR-6xHis was confirmed by Western blot (Supplementary Fig. S1E). Chromatin Immunoprecipitation (ChIP) was conducted in accordance with the manufacturer's manual (Beyotime, China). Before primers were designed for PCR, the 1000 bp upstream of the In (F-H), N = 5 ± SEM and "*" represents statistical significance as indicated by Student's t-test with maximum p-value of P < 0.05. start codon of prqA was subjected to analysis with BDGP online software and BPROM 28 for the transcription start site and transcription factor binding sites (Fig. 2E). Primers were then designed spanning the predicted binding site for an amplicon of 150 bp to 200 bp. After ChIP, DNA was used for PCR and the relative enrichment of DNA was analyzed against 1% input and IgG. The results showed that the putative prqA promoter was able to enrich approximately 50 and 10 fold of DNA, relative to control IgG, by primers 2 and primer 3, correspondingly (Fig. 2F). Furthermore, semi-quantitative PCR also found that primers 2 and 3 were able to amplify more DNA in the sample with anti-His antibody compared with the IgG control (Fig. 2G). This finding demonstrates that PrqR can bind directly with the promoter of prqA, which is immediately downstream of prqR. Similar analysis was performed on gap2, sigE and sodB genes (data are not shown). However, no significant binding was observed, suggesting that these genes were not regulated directly by PrqR.
We further deleted both prqR and prqA genes, resulting in the DprqR-DprqA mutant ( Supplementary  Fig. S1D). The DprqR-DprqA mutant exhibited similar growth of WT and DprqR under normal condition but we found significant growth retardation under GLC(+ )DARK-LIGHT condition (Fig. 2H), similar to that observed for WT. Thus, prqA may be negatively controlled by PrqR, which can be essential for the survival under high ROS conditions, such as GLC(+ )DARK-LIGHT. In all the cases from (A-D), N = 5 ± SEM, and p-value P < 0.05, and were normalized by rnpB expression level. In (E), the designed primer sets (listed in Supplementary table S4) surrounded the predicted binding site of prqA (ATTTGTTTTTTCATA) promoter. ATG is the start codon of prqA; the bold "T" is the start site of transcription. The prqR-His was expressed in OEprqR strain. After ChIP, samples were assessed by quantitative PCR, and the relative enrichment was calculated against 1% input and IgG (F). For clarity, normal PCR with those primer sets were performed with 25 cycles (G). The picture in (G) was cropped from the gel run under similar condition in electrophoresis (400 V, 125 amps) with the sample from 25 cycle of PCR. The experiment was repeated twice and found similar results. The deletion of prqR and prqA from WT (DprqR-DprqA) restored the phenotype of WT and CprqR (H). The culture was taken in after 48 h under GLC(+ )DARK-LIGHT condition.
Scientific RepoRts | 6:32507 | DOI: 10.1038/srep32507 Glucose uptake and glycogen content varied in DprqR mutant. The mutant DprqR showed remarkable difference from WT and CprqR in both glucose uptake and glycogen content in the entire course of GLC(+ ) DARK and subsequent GLC(+ )DARK-LIGHT condition. The glucose uptake in the first 24 h was found similar for WT, CprqR and DprqR mutants but declined more in WT and CprqR than in DprqR under GLC(+ )DARK condition. After shifting from GLC(+ )DARK to GLC(+ )DARK-LIGHT both WT and CprqR showed a sharp increase in glucose uptake compared with DprqR ( Fig. 3A) and after 48 h, DprqR consumed all the glucose added in the media, whereas WT stopped up-taking despite glucose remained in the media ( Supplementary Fig. S5). This result suggests that DprqR continues glucose catabolism under the GLC(+ )DARK condition, and more actively under GLC(+ )DARK-LIGHT condition, until all the glucose is consumed. Glycogen aggregation was found to follow similar trends in the first 48 h under GLC(+ )DARK but increased within 72 h in the WT, CprqR and DprqR mutants (Fig. 3B). The succeeding shift into the GLC(+ )DARK-LIGHT condition led to a decreased glycogen content in the DprqR mutant, whereas the WT and CprqR maintained the same amount as those under the GLC(+ )DARK condition.  Supplementary Fig. S2 ). The possibility of osmotic stress was excluded. In addition, we washed the WT cells after 24 h under GLC(+ )DARK-LIGHT. Afterward, we placed the cells into fresh media without glucose and found the cells resumed growth normally. These results led us to investigate the differential glucose metabolism abnormalities between WT and DprqR. Hence, we first targeted the following known regulators of the glucose metabolic pathway in Synechocystis through real-time qRT-PCR: a two-response regulatory protein Rre37 (sll1330) with HTH domain, which was previously found to control the expression of genes in glucose metabolic pathways in the presence of light and glucose 29 ; Hik31 (sll0790) with regulatory role in glucose metabolism under heterotrophic growth 30 ; a histidine kinase Hik8 (sll0750) involved in circadian rhythm and a regulator of glucose metabolism 31 ; and a group 2 sigma factor SigE, which regulates the expression of several genes in glucose metabolic pathways under normal growth condition 21,22 . The results showed no significant difference among the expression of hik8, hik31 and rre37 between WT and DprqR under GLC(+ )DARK condition ( Fig. 3C). Under GLC(+ )DARK-LIGHT condition, hik8 and hik31 exhibited a similar expression pattern in WT and DprqR, whereas rre37 displayed a relatively higher expression in DprqR than that in WT (Fig. 3C); this finding is consistent with a previous report 29 . Moreover, sigE had been found with differential expression in DprqR compared with WT ( In a previous study, the sigE mutant was found defective in glucose catabolism with decreased expression of phosphofructokinase, pfkB1 (sll1196), glyceraldehyde-3-phosphate dehydrogenase (gap1) and pyruvate kinase (pyk1) 21 . To investigate role of sigE in terms of regulating genes of glucose metabolism under prqR-deficient background, we investigated the expression of three genes in the glycolytic pathway via real-time qRT-PCR, including two target genes of sigE (i.e., pfkB1 and gap1). Although the results showed no significant change between WT and DprqR under either GLC(+ )DARK or GLC(+ )DARK-LIGHT conditions for two sigE target genes, except a slight increase of gap1 expression in DprqR under GLC(+ )DARK-LIGHT condition (Fig. 3F), the results showed upregulation of gap2, the gene encodes enzyme glyceraldyhyde-3-phosphate dehydrogenase in DprqR compared WT grown under GLC(+ )DARK. In addition, we determined the expression pattern of gap2 at different durations of GLC(+ )DARK and GLC(+ )DARK-LIGHT conditions (Fig. 3G). The results showed that the gap2 expression in DprqR was more than 10 fold higher than that in WT throughout the GLC(+ )DARK and GLC(+ ) DARK-LIGHT conditions. However, when we maintained a similar condition without glucose GLC(− )DARK for 72 h and then shifted to GLC(− )DARK-LIGHT for 24 h, no significant change of gap2 expression between DprqR and WT was found (Fig. 3H). This finding suggests that glucose caused the remarkable change of gap2 expression in D9prqR mutant under GLC(+ )DARK and GLC(+ )DARK-LIGHT conditions.

Dynamic expression of glucose metabolism regulators in
The OPP pathway is considered as the main route of carbon catabolism in Synechocystis and both zwf (slr1843 encoded enzyme glucose-6-phosphate-1 dehydrogenase, G6PDH) and gnd (sll0329 encoded enzyme 6-phosphogluconate dehydrogenase, 6PGD) play important roles in controlling the carbon flow regulated by sigE 21,22 . RT-PCR analysis showed that both zwf and gnd were slightly upregulated in DprqR under GLC(+ ) DARK condition compared with those in WT; a significant increase of zwf expression in DprqR was also observed under GLC(+ )DARK-LIGHT condition (Fig. 3F), and this finding was consistent with the increased glucose and glycogen metabolism immediately after shifting from GLC(+ )DARK to GLC(+ )DARK-LIGHT.   (Figs 5B and 6B), as well as D-gluconic acid, a derivatization degradation product of 6-phosphogluconate, increased by twofold in DprqR (Supplementary Table S1); this finding suggests the active role of the OPP pathway in DprqR under the above mentioned condition. By contrast, the intermediates of the TCA cycle, namely, alpha-ketoglutarate and acetyl-CoA, decreased in DprqR.

Changes of metabolites involved in oxidative stress mitigation.
The metabolomics under oxidative stress in Synechocystis are not so available till now. Given that Cyanobacteria is evolutionary ancestor of plant and they share similar environmental habitat, we compared changed of metabolites in our work with some well-studied metabolic profile of plant under stress. Among many of such studies, Noctor et al. 34 recently summarized a list of metabolites detected by GC/LC-MS that is more stably considered as marker to trace ROS status in Arabidopsis and Rice. The role and mechanism of many of these metabolites such as gluthathione, NADPH etc. in ROS acclimation are well-established. We compared the change of metabolites in our work with this information and listed. We found that under GLC(+ )DARK condition, 2-monopalmitin, cadaverine, and L-rhamnose increased by more than twofold, whereas sorbitol and trehalose decreased more than twofold in DprqR than in WT (Supplementary Table S2). The most distinct change of metabolites involved in oxidative stress removal  Table S1).
Glutathione (GSH) is considered as the most important defense metabolite against ROS damage in cells. GSH scavenges ROS and oxidize them to form glutathione disulfide (GSSG) by combining two GSH molecules with a sulfide bond. We were able to detect GSH by GC-TOF-MS, and the results showed that GSH was 1.75 fold higher in DprqR than in WT under GLC(+ )DARK-LIGHT condition; however, no such difference was detected under GLC(+ )DARK condition (Supplementary Table S1 ), and this finding was consistent with a previous report indicating that under the condition of oxidative stress, GSH declined and the redox state become more oxidized, thereby disrupting the cellular metabolic system 35 .
LC-MS based metabolite analyses revealed differential abundance of pyridine nucleotide and co-factors. Under GLC(+ )DARK condition, ATP, AMP, NAD + , NADH, NADPH and CoA significantly decreased in DprqR mutant compared with those in WT. By contrast, ADP and NADP + were remained unchanged (Fig. 7A). An almost similar trend was followed under GLC(+ )DARK-LIGHT condition, in which ATP, AMP, NAD + , NADH, NADPH, Co-A, ADP and NADP + decreased in DprqR mutant, compared with those in WT (Fig. 7B). The NADPH: NADP + ratio decreased by ~42% and ~26% when shifted from GLC(+ )DARK to GLC(+ ) DARK-LIGHT condition for WT and DprqR, respectively (Supplementary Table S5). Although the ratios of ATP:NADPH in WT and DprqR were similar under GLC(+ )DARK condition, the ratio was relatively higher

Discussion
Studies have progressively elucidated that soluble sugars, such as glucose and sucrose, are strongly related to stress-induced ROS signaling in plants, cyanobacteria, Saccharomyces, and even mammals 7,9,36,37 . Unlike the known pro-oxidants and antioxidants which are only involved in ROS removal, soluble sugars such as glucose play a role in both ROS production and ROS-induced signal for scavenging. To clarify the mechanisms controlling ROS signaling pathways under environmental stress and their interplay with glucose metabolism, the focus should be shifted from studying signals under single stress to a combination of abiotic stresses in the presence of glucose to accurately represent the natural habitat. In contrast to previous studies, that focused on glucose metabolism under different trophic conditions, as well as cellular regulation of glucose uptake and metabolism 21,22,[38][39][40] , the present study focused on GLC(+ )DARK-LIGHT condition, which generates ROS (Fig. 1C,H), and provided a platform to investigate the complicated interaction of glucose and ROS as well as the signaling network.
Given that PrqR is a negative regulator of ROS 26 , the knockout mutant of prqR (DprqR mutant) differed from WT and CprqR for its lower ROS content was found with an increased glucose metabolism under GLC(+ ) DARK-LIGHT condition (Figs 1H and 3A,B, Supplementary Fig. S5); these findings reflect the ROS removal capacity is functional in DprqR under above-mentioned condition which can be related to glucose metabolism. In this study, we will discuss first how PrqR plays a role in ROS mitigation followed by glucose metabolism. PCA analysis of metabolites identified by LC-MS and GC-MS (Fig. 4A,B) showed reprogramming of metabolism in DprqR, whereas metabolic profiling revealed several key marker metabolite changes, which further indicated an increased ROS removal in DprqR. First, DprqR showed superiority of breaking down glucose following the glycolytic pathway under GLC(+ )DARK condition (Figs 5A and 6A). This finding is in line with the ability of DprqR to enhance glucose uptake under GLC(+ )DARK condition (Fig. 3A). In addition, major changes in shifting the glucose metabolism from the glycolytic pathway into the OPP pathway were observed under GLC(+ )DARK-LIGHT condition (Fig. 6A,B). This switch is supposed to increase cellular NADPH given that the OPP pathway breaks down glucose to produce NADPH. On the contrary, we found decreased NADPH in DprqR cells under GLC(+ ) DARK-LIGHT condition (Fig. 7B). This finding explained that NADPH was used to reduce ROS-removing metabolites such as GSH. This NADPH-demand basis acceleration of the OPP pathway was also observed in photoheterotrophic condition in which NADPH production from photosynthesis was blocked and solely dependent on the OPP pathway 41 . Moreover, this re-routing of the central carbon metabolism, in favor of the OPP pathway to produce more NADPH, was previously reported in rice and yeast under oxidative stress removing cellular status 42,43 . Takahashi et al. found increased accumulation of glycolytic pathway intermediates such as fructose-6-phosphate, 3-phosphoglycerate, 2-phosphoglycerate, and phosphoenolpyruvate in mixotrophic 'DARK-LIGHT condition. "Green" represents increased abundance of the metabolites in DprqR compared with WT, "red" for decreased abundance and "black" was not significantly changed. The shadow is the pathway found to follow in particular conditions. Scientific RepoRts | 6:32507 | DOI: 10.1038/srep32507 condition in Synechocystis, similar to GLC(+ )DARK-LIGHT in current study, except the prior extension of GLC(+ )DARK incubation 38 . Overall, this study 38 suggested an active mode of glycolytic pathway under mixotrophic condition which is sharply contrasting with our findings obtained under GLC(+ )DARK-LIGHT condition. Thus, this particular information agrees with the understanding that glucose and ROS reprogrammed the central metabolic pathway favoring ROS removal in DprqR under GLC(+ )DARK-LIGHT condition. Second, the reduced GSH increased in DprqR under GLC(+ )DARK-LIGHT condition ( Supplementary Table S1). This observation is similar to previous report indicating that the response of GSH to ROS acclimation is a time-series event and a successful acclimation accumulate reduced GSH with unsuccessful acclimation increased GSSG/ GSH ratio 35 . Third, myo-inositol, a sugar alcohol which is a strong scavenger of ROS, was found with a sharply decreased concentration in DprqR mutant (Supplementary Table S1); this finding was consistent with a previous study in a Arabidopsis cat2 mutant, in which the ROS scavenger metabolites decreased in concentration during stress acclimation 44,45 . Finally, the results showed that the TCA cycle was perturbed in cells under GLC(+ ) DARK-LIGHT condition (Fig. 6B), which is within the expectation as a number of enzymes, including aconitase and 2-oxogluterate, are known to be affected during acclimation of oxidative stress 34,46 . Overall, metabolomics analysis demonstrated that the decreased ROS content and active ROS removal mechanism in DprqR mutant cell under GLC(+ )DARK-LIGHT condition confirmed the negative regulation of ROS by PrqR (Fig. 8).
These properties of the DprqR mutant led to the assumption that PrqR functions as a regulatory protein for ROS removal. Therefore, a search for possible targets of PrqR ended up establishing PrqR as a negative regulator of prqA and sodB; qRT-PCR showed an upsurge of the prqA and sodB expression in DprqR under both GLC(+ ) DARK and GLC(+ )DARK-LIGHT conditions ( Fig. 2A,C). Although very limited information is available about sodB regulation by sugars in cyanobacteria, the participation of glucose in antioxidant induction has been reported 20 . Studies suggested that sugar-regulated genes involved in anti-oxidative systems are under control of sugar sensors 14 , leading to a hypothesis that both sodB and prqA may be regulated directly by PrqR. In contrast to this initial expectation, the ChIP results showed that PrqR bound only to the prqA but not sodB promoter (Fig. 2F,G); thus the result suggests that PrqR regulates prqA directly and sodB indirectly. PrqA is a putative  (Fig. 2H), which was similar to WT, suggesting that prqA expression in DprqR mutant was crucial in ROS mitigation. However, the exact mechanism of how does prqA expression mitigate ROS is yet to be known. Altogether, the lower ROS content and elevated expression of ROS removing genes in DprqR implies that PrqR is a negative regulator of ROS under GLC(+ )DARK-LIGHT condition (Fig. 8).
Instead of feeding the OPP pathway by glucose and replenishing NADPH that helps to scavenge ROS in plants under stress 48,49 , glucose was observed to generate ROS and impaired its metabolism under GLC(+ ) DARK-LIGHT condition in WT and CprqR (Fig. 1C,H). The enhanced glucose metabolism is understandable with overexpression of sigE, which is the positive regulator of glucose metabolism in Synechocystis 21 (Fig. 3D), and its regulon in OPP pathway (i.e., zwf and gnd) (Fig. 3F) 21,50 . However, Chip-PCR measurement confirmed that PrqR does not regulate sigE directly (data not shown). Thus, an alternative explanation of glucose metabolism regulation in DprqR mutant through sigE expression under GLC(+ )DARK-LIGHT condition implies that DprqR mutant mitigated ROS under GLC(+ )DARK-LIGHT condition, thereby avoiding a negative consequence of ROS on sigE allowing sigE expression in the presence of glucose. Similarly, our result indicating that gap2 is overexpressed in DprqR under GLC(+ )DARK-LIGHT condition (Fig. 3G,H), can be fit with the suggestion relevent to a previous study 51 . Together, this suggestion justifyes our hypothesis that DprqR helps to remove stress from cells grown under high light at GLC(+ )DARK-LIGHT condition (Fig. 8). However, this suggestion cannot explain the increased expression of gap2 in DprqR at GLC(+ )DARK condition, during which no significant accumulation of ROS occured.
In summary, the established GLC(+ )DARK-LIGHT condition provides a platform to determine the ROS signaling and glucose metabolism in Synechocystis. Our results, as well as physiological analyses from a previous study 26 , demonstrated that the response regulator PrqR is involved in the regulation of glucose metabolism and oxidative stress acclimation in Synechocystis (Fig. 8).

Materials and Methods
Conditions of growth and maintenance for cyanobacteria. The Synechocystis sp. PCC 6803 strain used in this study was kindly provided by Prof. Weiwen Zhang (Tianjin University, China). Synechocystis was cultivated in BG-11 media (both solid and liquid) 52 . For normal growth, the culture was maintained in a shaker photo-incubator under continuous illumination at 25 μ mol m −2 s −2 , with shaking at 130 rpm and temperature of 30 °C. The mutant culture was supplemented with various antibiotics at different concentrations, as mentioned in the text (for DprqR mutant line, up to 100 μ g/mL kanamycin; for CprqR, 100 μ g/mL kanamycin + 100 mg/mL spectinomycin; for OEprqR, 100 μ g/mL kanamycin; and for DprqR-DprqA, 100 μ g/mL kanamycin). For heterotrophic [GLC(+ )DARK] growth, 5 mM glucose was added into the culture which was then maintained in the dark by wrapping with aluminum foil while shaking at 130 rpm and temperature of 30 °C. In this study, we maintained a special condition called GLC(+ )DARK-LIGHT in which cells were inoculated in liquid BG-11 media and kept for normal culture until the OD 730 was around 0.4; we then added 5 mM glucose and placed the samples under GLC(+ )DARK condition for 3 days with subsequent shifting into high light (at 80 μ mol m −2 s −2 ) for 2 or 3 days. For plate culture, the solid BG-11 on plate was always maintained in photo-incubator with continuous illumination at 25 μ mol m −2 s −2 and temperature of 30 °C. Antibiotics had been used during maintenance or selection of mutant strains but not when comparative study was conducted between WT and different mutants.
Cloning strategies, transformation and generation of stable strains. The constructed vectors and strains developed by transforming those vectors are listed in Fig. S1A. The vectors were constructed by combining the normal restriction-digestion-ligation system and fusion system, as described by Zhang et al. 53 . The two The "solid black" arrow represents the direct relation in contrast to the "interrupted" arrow is not. Overall, the prqR acts as a regulator and control oxidative stress and glucose metabolism in GLC(+ )DARK-LIGHT condition.
vector's backbones, namely, pUC19 and pBluescript KS (+ ), were used to construct all other vectors. The primers used in these constructions are listed in Supplementary Table S3. The constructed vectors transformed into Synechocystis following the method described by Liu et al. 54 . Colonies that appeared on the plate were streaked several times (at least 10 times) to obtain complete segregation. The mutant was confirmed by PCR, and positive clones were set in liquid culture supplemented with appropriate antibiotics.
In DprqR mutant ("D" stands for deletion), 245 bp out of 645 bp of prqR gene was deleted and a kanamycin cassette was inserted. The CprqR strain ("C" represents complementation) was generated by expressing the prqR gene under psbA2 promoter and spectinomycin under cpcB promoter in DprqR mutant by inserting in a neutral site (NS) at slr1311. For ChIP assay, we attached a 6XHis Tag at the C-terminus of the prqR gene, which was expressed under psbA2 promoter in WT strain at NS site of slr1311. To delete the two neighboring gene prqR and prqA, vector was designed using an upstream fragment of prqR and a downstream fragment of prqA flanked with kanamycin cassette under cpcB promoter.
ChIP-PCR. Chip-PCR was conducted using all reagents and protocols provided by Beyotime (Shanghai, China). Briefly, 40 mL of cell culture was collected from the normal culture of OEprqR strain, crosslinked using 1% formaldehyde and stabilized using 125 mM (final concentration) glycine. Sonication was performed as follows: 1 s ON, 2 s OFF for 18 min with 25% amplitude, to achieve the DNA fragment length of 400 bp to 800 bp. After centrifugation, the supernatant was transferred (about 0.2 mL) into a 2 mL ice-bathed tube. ChIP was performed in accordance with the protocol given by Beyotime (Shanghai, China).
The real time q-PCR was set with a 25 μ L system (12.5 μ L of SYBR GreeN, 0.5 × 2 Primers, 1 μ L of sample, 1 μ L of 1% input, 9.5 μ L of ddH2O). Three pairs of primers were used for each targeted set (Supplementary Table S4). The DNA sample eluted with normal unconjugated mouse IgG was used as control.
Semi-quantitative PCR was set using Taq 2X Master Mix Kit (New England Biolabs Inc). The reaction was set in a 25 μ L system (12.5 Taq 2X Master Mix, 0.5 × 2 Primers, 1 μ L each of sample and 1% input, and 10.5 ddH2O) for 25 cycles.

GC/LC-MS based metabolic profiling.
For GC-MS, approximately 3 × 10 9 cells were collected from the culture under GLC(+ )DARK and GLC(+ )DARK-LIGHT conditions and placed into − 80 °C for at least 2 min to stop further metabolic activity. Cells were collected by centrifugation, washed twice with ice-cold NaCl (0.9%) and placed in tubes in the presence of liquid nitrogen. Up to 1.5 mL of methanol-ddH 2 O (1:1 v/v) was added along with 10 μ L of L-2-chlorophenylalanine (0.3 mg/mL in dH2O). Sonication was conducted as follows: 5 s ON, 10 s OFF for 15 min at 90% amplitude, and then repeated thrice. The supernatant was dried up completely by freeze drying with continuous flow of nitrogen. Samples were derivatized and run under GC-TOF-MS following the method reported by Liu et al. 32 . The data were acquired in full-scan mode with the m/z range of 30-550 at a rate of 20 spectra per second following a solvent delay of 360 s. The acquired data were processed and analyzed as previously described 32,55 . Metabolite identification was performed by Chroma TOF linked with spectral library databases of National Institute of Standards and Technology (NIST) (2005). Other original standards were used to further validate the data. The dataset extracted in MS files from GC/TOFMS analysis was exported into NetCDF format by ChromaTOF software (v3.30, Leco Co., CA). CDF files were processed using custom scripts (revised MATLAB toolbox, hierarchical multivariate curve resolution (H-MCR), developed by Jonsson et al. 56 For LC-MS analysis, samples with equal cells number were collected at two time points: 72 h under GLC(+ ) DARK and 24 h under GLC(+ )DARK-LIGHT condition. The pellet was centrifuged and collected quickly, frozen in liquid nitrogen and then preserved it at − 80 °C before further use. Metabolites were extracted by quenching following the protocol described by Su et. al. 33 . In quenching metabolites, the frozen pellet was dissolved in 500 μ L of 100% LC-MS grade methanol and placed for thawing on dry ice. The supernatant was collected and stored at − 80 °C. The LC-MS experiment was conducted using an Agilent 1260 series binary HPLC system (Agilent Technologies, Waldbronn, Germany) attached with an Agilent 6410 triple quadrupole mass analyzer equipped with an electronic spray ion (ESI) source. Aqueous 10 mM tributylamine (pH 4.95, adjusted with acetic acid) and 100% HPLC-grade methanol were used as Mobile Phase A (MPA) and Mobile Phase B (MPB), respectively. Agilent Mass Hunter workstation LC/QQQ acquisition software (version B.04.01) was applied for data processing and Agilent Mass Hunter workstation qualitative analysis software version (B.04.00) applied for statistical analysis. The data obtained were analyzed using SIMCA 12.0 software (for PCA analysis), MeV4_9_0 (for Heat Map), and GraphPad Prism version 5.01 (t-test and others).
Growth and chlorophyll content measurements. The growth models of WT, DprqR and CprqR were studied under normal, GLC(+ )DARK and GLC(+ )DARK-LIGHT conditions by detecting the optical density (OD) at different wavelengths by using a spectrophotometer (METASH). OD 730 was used to determine growth curve. OD 678 , OD 720 and OD 750 were also measured, and the total chlorophyll content was quantified using following formula; 14.94 (OD 678 -OD 750 )−0.616 (OD 720 -OD 750 ), described by 58 . RNA isolation, cDNA synthesis, and qRT-PCR. Total RNA from 50 mL of culture had been isolated using TRIzol (Invitrogen) following the modified method described by Schwarzkopf et al. 59 . The collected RNA was then treated with DNase1 (RNase free), deactivated DNase1 to remove contaminated DNA following the provided protocol. Thus, the DNA-free RNA was subjected to convert cDNA applying kits, SuperScript ® III First-Strand Synthesis System (Invitrogen). qRT-PCR was performed by applying FastStart Universal SYBER Green (ROX) ( Roche) reagent in Applied Biosystems 7500 RealTime PCR System. The RNA subunit of Scientific RepoRts | 6:32507 | DOI: 10.1038/srep32507 ribonuclease P subunit B (rnpB) served as an internal control. The primers used to quantify gene expression are listed in Supplementary Table S4. ROS measurement. We measured the total ROS by using membrane-permeant fluorescence indicator 5-(and-6)-chloromethyl-2′ , 7′ -dichlorodihydro fluorescein diacetate, CM-H2DCFDA (Invitrogen, Life technology), following the protocols described by Hakkila, et al. 23 and provided by Life Technology. Samples from normal culture, GLC(+ )DARK culture (72 h) and GLC(+ )DARK-LIGHT (24 h) were taken to estimate the total ROS content. Up to 25 mM CM-H2DCFDA was added to 1 mL of culture in 2 mL tube and then mixed. About 1 mL of culture was also obtained without adding any CM-H2DCFDA and considered as control (auto-fluorescence). Samples were then incubated for 90 min. in complete dark at 32 °C. Cells were washed with BG-11 media and suspended with BG-11 media to reach the final volume of 0.5 mL. Up to 200 μ L of cell suspension was pipetted to a white 96-well microtiter plate. Fluorescence from CM-H2DCFDA treated and untreated cell (auto-fluorescence) was measured using Synergy Multi-Mode Reader (BioTek). The data measured immediately after dark incubation were considered as "Initial Reading" and then incubated in the corresponding condition for 1 h and measured the fluorescence for "Final Reading. " The strategy was followed both for treated and untreated (auto-fluorescence) samples. The ROS content was then calculated following the equation Glucose (extracellular) and glycogen measurements. Both glycogen and glucose were measured described elsewhere 60 . For glycogen quantification, 3 mL sample was obtained and centrifuged; the pellet was collected and washed with sterile ddH 2 O and then freeze dried. The weight of the dried cell was recorded. The cell was then suspended in 30% w/v KOH, mixed and incubated at 97 °C for 2 h. Ice-cold ethanol was added to the final concentration of 70% to 75% and incubated on ice for at least 2 h. The mixture was centrifuged, the supernatant was removed, and the pellet was washed with 98% ethanol. The glycogen pellet was dried at 60 °C for 10 min and then dissolved in 100 mM sodium acetate (pH 4.5). The dissolved glycogen was hydrolyzed enzymatically, by treating with 2 mg/mL amyloglucosidase (Sigma) at 60 °C for 2 h, into glucose. The glucose obtained from glycogen and extracellular (in media added externally) was measured using a Glucose (GO) Assay kit (Sigma-Aldrich) in accordance with the manufacture's protocol. For glycogen measurement, data were normalized by cell dry weight. To understand the glucose uptake ability of the cell, kinetics of the decrease of the glucose content in the growth medium under GLC(+ )DARK and GLC(+ )DARK-LIGHT conditions were calculated. The data was normalized by OD 730 according to time duration.