Redox regulation of glycogen biosynthesis in the cyanobacterium synechocystis sp. PCC 6803: Analysis of the AGP and glycogen synthases

1" " Redox regulation of glycogen biosynthesis in the cyanobacterium Synechocystis sp. PCC 6803. Analysis of the AGP and glycogen synthases. Runing title: Cyanobacterial glycogen synthesis regulation Sandra Díaz-Troya†, Luis López-Maury†, Ana María Sánchez-Riego, Miguel Roldan and Francisco J. Florencio*. Instituto de Bioquímica Vegetal y Fotosíntesis, Universidad de Sevilla-CSIC, Américo Vespucio 49, E-41092 Sevilla, Spain. *To whom correspondence should be addressed: Francisco J. Florencio, Instituto de Bioquímica Vegetal y Fotosíntesis, Universidad de Sevilla-CSIC, Av Americo Vespucio 49, E 41092 Seville, Spain; Telephone: +34 954489509; FAX: +34 954460065; E-Mail: [email protected]; † These authors contributed equally to this work. Keywords: Cyanobacteria, glycogen, Synechocystis, thioredoxin, redox regulation 2" " Abstract Glycogen constitutes the major carbon storage source in cyanobacteria, as starch in algae and higher plants. Glycogen and starch synthesis is linked to active photosynthesis and both of them are degraded to glucose in the dark to maintain cell metabolism. Control of glycogen biosynthesis in cyanobacteria could be mediated by the regulation of the enzymes involved in this process, ADP-glucose pyrophosphorylase (AGP) and glycogen synthase, which were identified as putative thioredoxin targets. We have analyzed whether both enzymes were subjected to redox modification using purified recombinant enzymes or cell extracts in the model cyanobacterium Synechocystis sp. PCC 6803. Our results indicate that both AGP and glycogen synthases are sensitive to copper oxidation. However, only AGP exhibits a decrease in its enzymatic activity, which is recovered after reduction by DTT or reduced thioredoxin (TrxA), suggesting a redox control of AGP. In order to elucidate the role in redox control of the cysteine residues present on the AGP sequence (C45, C185, C320 and C337), they were replaced with serine. All AGP mutant proteins remained active when expressed in Synechocystis, although showed different electrophoretic mobility profiles after copper oxidation, reflecting a complex pattern of cysteines interaction. 3" " Introduction Cyanobacteria constitute a large group of oxygenic photosynthetic prokaryotes able to fix atmospheric CO2 utilizing the ATP and the reducing electron power generated by the photosynthetic process. Cyanobacteria are widespread on different environments including freshwater, oceans and terrestrial habitats. They are responsible for a significant part of the global biomass and the oxygen in the atmosphere (Blankenship and Hartman, 1998; García-Pichel, 2008). Furthermore, they are also evolutionary precursors of chloroplast by an endosymbiotic process (Criscuolo and Gribaldo, 2011; Price et al., 2012). Carbon assimilation in cyanobacteria is a typical photosynthetic process and most of the carbon fixed is usually combined with nitrogen, via the glutamine synthetase-glutamate synthase pathway, to produce amino acids and subsequently proteins (Muro-Pastor et al., 2005). In addition to proteins, carbon is also used to synthetize other compounds including sugars, fatty acids, nucleotides and polysaccharides (Machado and Atsumi, 2012). Polysaccharides represent an interesting way to storage carbon for its use under low-energy conditions, as occurs in darkness (Steuer et al., 2012; Suzuki et al., 2010; Wyman and Thom, 2012; Xu et al., 2013). In cyanobacteria, glycogen is the main polysaccharide accumulated inside the cell under certain metabolic situations, such as nitrogen deprivation (Ball and Morell, 2003; Grundel et al., 2012; Suzuki et al., 2010). Recently, the synthesis of a polysaccharide much more similar to starch than to glycogen have been reported in some cyanobacterial strains (Suzuki et al., 2013). This fact supports the view of a cyanobacterial origin of plastid starch in algae and plants (Ball et al., 2011; Ball and Morell, 2003). In addition, many cyanobacteria are able to produce exopolysaccharides (EPS) that are exported outside the cell in response to environmental stresses or unbalanced carbon/nitrogen status (Pereira et al., 2009). Other soluble sugars, such as trehalose, sucrose or glucosylglycerol are also produced as osmoprotectants by cyanobacteria in response to osmotic stress adaptation (Klahn and Hagemann, 2011). 4" " The biosynthesis of glycogen in cyanobacteria requires the participation of different enzymes. The first step is the conversion of glucose-6P into glucose-1P by phosphoglucomutase, followed by the action of ADP-glucose pyrophosphorylase (AGP) that synthesizes ADP-glucose, using glucose-1P and ATP. This reaction generates PPi, which is converted into phosphate by a soluble pyrophosphatase. The glucose moiety of the ADP-glucose is then transferred to the non-reducing end of a linear α-1,4 glucan by glycogen synthase. Finally, glycogen branches are introduced by the branching enzyme, which adds α-1,6 glycosidic bonds (Ball and Morell, 2003; Miao et al., 2003; Suzuki et al., 2010; Xu et al., 2013; Yoo et al., 2002). In the cyanobacterium Synechocystis sp. PCC 6803 (hereafter Synechocystis) glycogen biosynthesis requires the participation of phosphoglucomutase (encoded by sll0726), AGP (glgC: slr1176 gene), two different glycogen synthases (encoded by glgA1: sll0945 and glgA2: sll1393), and the branching enzyme (glgB: sll1058). Mutant strains lacking AGP protein or both glycogen synthases are unable to produce detectable glycogen (Carrieri et al., 2012; Grundel et al., 2012; Miao et al., 2003). In addition, the AGP mutant presents a pleitropic phenotype that includes, among others, impaired cell-viability under different stress conditions and the lack of bleaching, a process that allows the cells to degrade phycocyanin after being transferred to a nitrogen-free medium (Carrieri et al., 2012; Grundel et al., 2012; Miao et al., 2003). A similar phenotype has also been described for a double mutant lacking both glycogen synthases suggesting that glycogen is essential under particular environmental conditions (Grundel et al., 2012). The key control step in glycogen/starch synthesis in photosynthetic organisms seems to be the regulation of AGP enzyme. In this regard, it is well known that AGP is finely regulated by the allosteric effect of 3-phosphoglycerate in higher plants, algae and cyanobacteria (Ballicora et al., 2004; Crevillen et al., 2003; Iglesias et al., 1991; Ventriglia et al., 2007). In addition to allosteric regulation, redox regulation has been also documented in the case of higher plants, showing a stimulation of AGP activity by thioredoxins, preferentially thioredoxin f (Buchanan 5" " and Balmer, 2005; Hendriks et al., 2003; Sanz-Barrio et al., 2013; Thormahlen et al., 2013). Such redox regulation of AGP was related to the formation of a disulfide bridge between the two small subunits of AGP, APS1, with the reduced form being the active one. In this regard, replacement of Cys81 with Ser impaired the formation of the disulfide between the two APS1 subunits in Arabidopsis AGP (Hadrich et al., 2012). Transgenic plants containing the APS1C81S accumulated more starch during the day than plants expressing the WT version, which correlated with the absence of the dimeric form (Hadrich et al., 2012). However, other reports indicated that there were no differences in the amount of starch stored by these transgenic plants compared to that containing the wild-type APS1 version (Li et al., 2012; Thormahlen et al., 2013). By contrast, data concerning the redox regulation of cyanobacterial AGP is scarce and indicated a lack of this type of regulation (Fu et al., 1998; Iglesias et al., 2006). Up to date no data is available concerning a possible regulation of the glycogen synthase activity in cyanobacteria. Glycogen synthesis is active in the light, especially in cells subjected to nitrogen starvation. However, only glycogen content was determined but not the glycogen synthase activity (Ball and Morell, 2003; Grundel et al., 2012; Xu et al., 2013). In this study we analyzed whether glycogen synthesis is redox modulated in the cyanobacterium Synechocystis. Our results indicate that none of the two glycogen synthases present in Synechocystis undergo redox regulation. In contrast, AGP activity is regulated by oxidoreduction of its cysteines both in vitro and in cell extracts. Furthermore, AGP is able to form mixed disulfides with TrxAC35S (a mutant version of the most abundant thioredoxin in Synechocystis) and TrxA is also able to reduce the AGP both in vitro and in cell extracts. Mutations of the four cysteines present in AGP revealed a complex pattern of oxidation that probably involved all of them. All of these AGP mutant proteins, with the exception of AGPC45S, presented activities comparable to the wild type protein. However, when these different versions were reintroduced in the mutant background the resultant strains accumulated less glycogen than the WT strain. This suggests a deregulation of the whole pathway in the ΔAGP mutant, which could be due to an 6" " unknown secondary mutation. Finally, we analyzed whether thioredoxin mutant strains have a defect in glycogen synthesis regulation but none of the tested strains presented any significant differences in glycogen accumulation, after nitrogen deprivation, when compared to the WT. Results Glycogen synthases are not able to interact with TrxA. Most of the glycogen metabolism enzymes in Synechocystis have been identified as glutaredoxin or thioredoxin targets (Li et al., 2007; Lindahl and Florencio, 2003). These included phosphoglucomutase, AGP, glycogen synthases, glycogen branching enzyme and glycogen phosphorylase. Among these enzymes, only phosphoglucomutase has been shown to be redox regulated (Lindahl and Florencio, 2003). In the current study we have tested whether glycogen synthases and AGP were also redox regulated. Synechocystis genome contains two genes coding for putative glycogen synthases: sll0945, which encodes GlgA1, and sll1393, which encodes GlgA2. Of the two glycogen synthases, only GlgA2 was identified as TrxA target (Lindahl and Florencio, 2003), although both of them were identified as GrxA targets (Li et al., 2007). This suggests that these two proteins could be redox regulated. In order to elucidate if both proteins are able to interact with TrxA, TrxAC35S was used. This recombinant version of TrxA lacks the resolving cysteine in the active site, allowing the formation of stable mixed disulfides with its target proteins"(Lindahl and Florencio, 2004; Mata-Cabana et al., 2007). Soluble cell extracts were incubated with 50 µM TrxAC35S and analyzed by western blot using non-reducing gels and anti-GlgA1 or GlgA2 specific antibodies (Figures S1 and S2). None of the enzymes were able to form mixed disulfides with TrxAC35S, either in non-treated or oxidized cells extracts (Figure S2). Moreover, both purified recombinant proteins from E. coli (Figure S1) were incubated with a 4- fold excess of TrxAC35S for 1h and aliquots were analyzed by western blot. None of the glycogen synthases formed any detectable complexes with TrxAC35S (Figure S2) using anti-TrxA antibodies. These strongly suggest that none of the 7" " glycogen synthases are thioredoxin targets or, at least, not in these recombinant glycogen synthase preparations. Furthermore, to confirm that glycogen synthases were not redox regulated we have assayed their activities in oxidized and reduced conditions. Purified proteins were assayed as isolated, oxidized (after incubation with 10 µM Cu2+) or after reduction with 5 mM DTT. These conditions were chosen because oxidized and reduced forms of both GlgA1 and GlgA2 could be detected (Figure 1B and C). These two forms were inter-convertible, i.e. copper oxidized proteins were fully reduced by the addition of DTT. Most GlgA1 was oxidized when purified from E. coli and its oxidation was only slightly increased by the copper treatment. In contrast, GlgA2 was mostly reduced when purified from E. coli and was fully oxidized by copper (Figure 1B and C). Under these conditions, GlgA1 and GlgA2 had specific activities (measured as ADP formation) of 3.33±0.71 and 3.30±0.89 U/mg protein, respectively. Those activities values were not significantly reduced or enhanced by oxidation or reduction (Figure 1A). Taken together these results suggest that none of the glycogen synthases are redox regulated in Synechocystis. AGP is a thioredoxin target. AGP is another enzyme involved in glycogen synthesis that was identified as putative TrxA target (Lindahl and Florencio, 2003). In order to confirm this interaction, formation of mixed disulfides between AGP and TrxAC35S was analyzed in Synechocystis extracts. Soluble cell extracts were pre-oxidized with 25 µM CuCl2 and incubated with 50 µM TrxAC35S. The formation of stable disulfides with AGP was analyzed by western blot using anti-AGP antibodies. Since our antibodies detect two unspecific bands, a cell extract from a Synechocystis mutant strain lacking the AGP gene (∆AGP) was included as control (Figure 2A and Figure S1). Additional protein bands were detected in the lane corresponding to the WT extracts incubated with TrxAC35S that were absent in ΔAGP mutant extracts 8" " (Figure 2A). Thus, this result indicates that AGP is able to form a mixed disulfide with TrxAC35S in cell extracts. In order to confirm this interaction, purified recombinant AGP was incubated with a four-fold excess of TrxAC35S to obtain stable mixed disulfides. Complex formation was detected by western blot analysis using anti-TrxA antibodies. Two control lanes containing only TrxAC35S or AGP were also included, as these two proteins formed disulfides with themselves (Figure 2B). A protein band with an electrophoretic mobility of ~65 kDa was detected in the non-reducing gel in the lane in which both AGP and TrxAC35S were present but it was absent in all others. This result shows that TrxAC35S can also form a mixed disulfide with AGP in vitro. Plant and bacterial AGP enzymes have been shown to be tetramers (Ballicora et al., 2003; Ballicora et al., 2004) and we have investigated the quaternary structure of Synechocystis AGP protein using gel filtration. Extracts from WT cells were sizefractionated on a Superdex 200 column and the eluted fractions were tested for the presence of AGP by western blot. AGP fractionated as 200 kDa molecular mass complexes (Figure 2C). This confirms that native AGP protein forms a tetramer, as previously described for purified AGP from Synechocystis"(Iglesias et al., 1991). Since glycogen accumulation is dependent on the presence of light, changes in the quaternary structure of AGP were compared between light and dark conditions. For that, the oligomeric state of AGP was analyzed by gel filtration in extracts prepared from cells transferred to the dark for 8 h. In this experiment, AGP eluted as a tetramer in cells extract prepared from light grown cells or dark treated cells (Figure 2C). Moreover, the tetrameric structure of AGP was retained even when cell extracts were reduced with DTT prior to gel filtration (data not shown). This indicates that the tetramer does not require any intermolecular disulfide bridge between subunits for its stabilization. AGP activity is redox modulated. 9" " Once the interaction between AGP and TrxA was confirmed, we investigated whether this interaction played a role in the modulation of the AGP activity. Analysis of the enzyme activity was carried out with the recombinant purified protein treated in the same conditions as described for glycogen synthases. In the case of AGP, the purified recombinant protein was obtained partially oxidized from E. coli. Full enzyme oxidation was achieved by copper treatment (10 µM), which led to a 40% reduction of AGP activity (Figure 3A). In contrast, addition of DTT or reduced TrxA stimulated the activity 3 and 5 times, respectively, compared to the oxidized enzyme (Figure 3A). The treatment conditions used for the enzymatic determination were selected because they were able to produce changes in the redox state of the protein (Figure 3B). The completely oxidized protein forms oligomers and intramolecular disulfides, which were visualized, under non-reducing conditions, as slower and faster migrating bands, respectively (Figure 3B). All of these oxidized forms were removed by DTT or reduced TrxA treatments (Figure 3B). The specific activity of the recombinant protein (10 mU/mg prot) was very low compared to previously described activity for native purified protein from Synechocystis (Iglesias et al., 1991). This could be due to the tag used for protein purification in E. coli, as it occurs for Solanum AGP (Hwang et al., 2004), and/or because the protein would require a post-translational modification that occurs in Synechocystis but not in E. coli. For this reason, redox regulation of AGP activity was also assayed in Synechocystis extracts. Consistently with the in vitro data, AGP activity was inactivated by copper treatment and activated by of DTT or reduced TrxA additions (Figure 3C). The specific activity in Synechocystis extracts (104 mU/mg protein) was about two times higher than the previously published values (Iglesias et al., 1991), although in this report, enzyme activity was assayed in the opposite direction. Analysis of the AGP oxidation state by western blot showed the formation of oligomers and intramolecular disulfide bonds in the fulloxidized extracts. In contrast, the protein was reduced in non-treated extracts or after DTT or TrxA additions (Figure 3D). Therefore, AGP activity could be correlated with its redox state, being less active in its oxidized form and fully active in the reduced form both in vitro and in cell extracts. These results support the view 16" " this glycogen was consumed during dark periods and accumulated in the light like in the WT (Figure 4C). These mutants were generated in the ΔAGP strain, which carried a complete deletion of the glgC gene. Our results suggest that our ΔAGP strain could carry a secondary mutation, although the phenotype of our mutant was identical to previously published phenotypes for other AGP mutants, i.e. nonbleaching after nitrogen starvation or lack of glycogen (Figure 4, 6 and data not shown; (Carrieri et al., 2012; Grundel et al., 2012; Miao et al., 2003)). In fact, similar phenotypes have been recently reported in another complemented glgC mutant strain (Carrieri et al., 2012), which suggest that this mutation is required for full segregation of the glgC inactivation. This secondary mutation affects glycogen metabolism and accumulation because all these strains accumulated less glycogen than the WT. We have checked that this mutation is not a consequence of the antibiotic cassette introduced down-stream to the glgC gene, because WT strain transformation with pGLGC-Km, the plasmid used to generate WTc in the ΔAGP background, rendered a strain indistinguishable from the WT (data not shown). Furthermore, this mutation is not clone specific either, as different clones behaved in the same way when we re-introduced the wild-type AGP version (data not shown). Further work is under way in our lab to elucidate the gene(s) mutated, and its effect in glycogen metabolism. Finally, we have also analyzed the effects of the mutation of different thioredoxins genes in the accumulation of carbohydrates during nitrogen starvation, a condition that has been shown to induce glycogen accumulation (Ball and Morell, 2003; Grundel et al., 2012; Suzuki et al., 2010). Under this condition most of the carbohydrates accumulated corresponded to glycogen as reflected by the lack of carbohydrates accumulation in ΔAGP. None of the thioredoxin mutant strains tested are affected in carbohydrate accumulation. Unfortunately, a complete trxA mutant could not be tested since this is an essential gene in Synechocystis (Navarro and Florencio, 1996). Instead a mutant that expresses the trxA gene under gifB promoter was used. This promoter is induced by NH4, not induced in nitrate and repressed during nitrogen deprivation (Garcia-Dominguez et al., 2000). The STXAc strain has less TrxA than the WT strain under all conditions (Figure 17" " 6C), but despite this fact, it showed similar growth and carbohydrate accumulation to the WT (Figure 6A and data not shown). These suggest that even the reduced amounts of TrxA present in the STXAc strain are enough to support all the essential functions of TrxA. The lack of phenotype of these thioredoxin mutants is in agreement with the data which showed that TrxA is the most abundant thioredoxin in Synechocystis (Florencio et al., 2006), while TrxB and TrxQ has been proposed to be expressed only under specific stress conditions (Perez-Perez et al., 2009a; Perez-Perez et al., 2009b). Therefore, TrxA will probably be in charge of regulating most of the light modulated processes in Synechocystis. 18" " Methods. Strains and culture conditions. Synechocystis cells were grown photoautotrophically on BG11C (Rippka et al., 1979) at 30º C under continuous illumination (50 µE m-2 s-1) and bubbled with a stream of 1% (v/v) CO2 in air. Nitrogen source removal was carried out by washing cell cultures with BG110C (BG11C medium lacking NaNO3). 10 mM NH4Cl and 20 mM N-tris(hydroxymethyl)-methyl-2-aminoethane-sulphonic acid (TES) buffer pH 7.5 were used for ammonium supplementation of cultures. For plate cultures, medium were supplemented with 1% (wt/vol) agar. Kanamycin, chloramphenicol and spectinomycin were added to a final concentration of 50 µg mL-1, 20 µg mL-1 and 5 µg mL-1, respectively. Strains used in this work are described in Table 1. A description of the plasmids used to construct new strains is included in Supplementary methods. Primers used in this work and plasmids constructed to generate Synechocystis strains are described in Table S1. Expression and purification of recombinant proteins Using Synechocystis DNA as template, full coding sequences from glgC (slr1176), glgA1 (sll0945) and glgA2 (sll1393) were amplified and unique restriction sites BamHI/SalI added by PCR using the following pairs of primers: glgC-BamHI F/glgC-SalI R, glgA1-BamHI F/glgA1-SalI R and glgA2-BamHI F/glgA2-SalI R. The amplified products were cloned BamHI/SalI into pET45b vector and correct sequence confirmed by DNA sequencing. For 6xHis-tagged recombinant proteins expression, the plasmids were transformed into E. coli BL21. Under standard protocol for expression induction with IPTG, all three recombinant proteins were insoluble, however culture incubation for 48h at 25ºC, 100 rpm without IPTG induction allows synthesis of enough soluble recombinant protein to be purified. Cultures were harvested by centrifugation, resuspended in 50 mM Tris-HCl pH 8.0, 500 mM NaCl, 1 mM PMSF and disrupted (20 kHz, 75 W) on ice for 3 min (in 30-s periods) in a Branson sonifyer. Lysates were then centrifuged at 20000 g for 30 19" " min for clarification. Supernatants were supplemented with imidazole to a final concentration of 25 mM and loaded onto a Ni-NTA agarose (Invitrogen) column for affinity chromatography purification. After extensive washing with 50 mM Tris-HCl pH 8.0, 500 mM NaCl, 25 mM imidazole, bound recombinant proteins were eluted with increasing concentrations of imidazole in 50 mM Tris-HCl pH 8.0, 500 mM NaCl. The purified proteins were then desalted using PD-10 columns (GE Healthcare) pre-equilibrated with 50 mM Tris-HCl pH 8.0, 150 mM NaCl, 20% glycerol and concentrated in a 30K-molecular-weight cutoff Amicon Ultra centrifugal filter unit (Millipore). Wild type TrxA, and the modified TrxA35, in which Cys35 has been substituted by Ser, were expressed and purified as previously described (Lindahl and Florencio, 2003). Antiboby production, Western blotting and preparation of crude extracts from Synechocystis cells. Anti-AGP, anti-GlgA1 and anti-GlgA2 antisera were obtained according to standard immunization protocols by injecting purified AGP, GlgA1 and GlgA2 proteins in rabbits. For analysis of AGP, GlgA1 and GlgA2 in Synechocystis cells, crude extracts were prepared using glass beads and minibead-beater. Cells were resuspended in 50 mM Hepes-NaOH pH 8.0, 150 mM KCl and subjected to 2 cycles of 1’ vortexing separated by 5’ on ice. Cell extracts were recovered from the beads and samples were clarified by two sequential centrifugations: 5’ at 5000 g to eliminate cells debris and 15’ at 15000 g to remove membranes. Protein concentration in cell-free extracts or purified protein preparations were determined by the method of Bradford, using ovalbumin as a standard and the specified amounts of proteins were separated on SDS PAGE under reducing or nonreducing conditions. Gel were transferred, blocked in PBS containing 0.1% tween 20 and 5% of skimmed milk and incubated with antibodies at a 1:3000 dilution for TrxA (Navarro et al., 2000) and 1:5000 for GlgA1, GlgA2 and AGP. Gel filtration 20" " Soluble extract from Synechocystis wild type strain was prepared as indicated above and 100 µg in a volume of 200 µl was loaded onto a Superdex 200 column (GE Healthcare) at a flow rate of 0.5 ml/min. 0.5-ml fractions were collected and the elution profile of AGP was analyzed by western blot. The column was calibrated with known standards (Bio-Rad). Glycogen synthase activity determination. Glycogen synthase activity was assayed using glycogen as primer and purified GlgA1 and GlgA2. 2.5 micrograms of the proteins samples were incubated at 30°C for 30 min in 100 µL of the following buffer: 100 mM Tricine pH 8.0, 25 mM K-- acetate, 5 mM EDTA, 0.5 mg mL-1 BSA, 1 mM ADP-glucose and 1 mg mL-1 glycogen. The reaction was stopped by boiling for 10’ and samples were centrifuged 20’ at 15000 g. The amount of ADP production and ADP-glucose consumption was determined by HPLC fitted with a Partisil-10-SAX column as described in (Munoz et al., 2005). AGP activity determination. AGP activity was assayed in the ADP-glucose synthesis direction as described in (Fusari et al., 2006). Reaction mixtures contained 50 mM HEPES pH 8, 10 mM MgCl2, 2 mM ATP, 1 mM glucose 1-phosphate, 2 mM 3-phosphoglycerate and 1 U yeast inorganic pyrophosphatase. Assays were initiated by the addition of the purified recombinant AGP or 0.5 µg Synechocystis soluble extracts in a total volume of 50 µl. After 30 min incubation at 30ºC, reactions were stopped by the addition of Malachite green reagent and read at 650 nm. Activity was calculated by subtracting the amount of phosphate formed in the absence of glucose1-phosphate in the same samples (this background phosphate was equivalent to the amount released in ΔAGP extracts assayed in the same conditions and therefore is produced by other(s) enzyme(s)). Samples were assayed in triplicate for each experiment. Carbohydrates and glycogen content determinations. 21" " Total carbohydrates were determined using the phenol–sulfuric acid method as described in (Dubois et al., 1951). Total protein content was determined by the method of Lowry as described in (Markwell et al., 1978). Glycogen was determined using the method described in (Xu et al., 2013) with some modifications. Cells were disrupted in 300 µl of 30 mM Na-acetate pH 5.2 using a beadbeater and glass beads. Two aliquots of 100 µl of broken cells suspension were boiled for 20 min and one of them was treated with 10 U of amyloglucosidase (from Asperguillus niger; from Sigma) overnight while the other was not treated. Glucose released was determined in both aliquots using the glucose oxidase/peroxidase method (Sigma catalog #GAGO-20). A calibration curve was prepared with commercial glycogen. Glycogen content was determined by substracting glucose present in the samples that were not treated with amyloglucosidase. Acknowledgments. This work was supported by Ministerio de Economía y Competitividad (MINECO) and the European Regional Fund (FEDER) to F.J. Florencio (grant BFU2010– 15708, and plan E–CLPN09–001) and by Junta de Andalucía (group BIO–284). L.L.-M. was recipient of a Junta de Ampliación de Estudios-Doctor contract from Consejo Superior de Investigaciones Científicas (CSIC). We thank Angel Merida and Alejandro Mata for critical reading the manuscript. 22" " Table 1. Strains used in this work. Synechocystis strains Genotype Mutated ORF(s) Source or Study WT Synechocystis sp. PCC 6803 None ΔAGP ΔglgC::C.C1 slr1176 This study. WTc glgCWT ::C.K1 None This study. C45S glgCC45S ::C.K1 slr1176 This study C315S glgCC315S ::C.K1 slr1176 This study. C320S glgCC320S ::C.K1 slr1176 This study. C337S glgCC337S ::C.K1 slr1176 This study. ΔglgA1 ΔglgA1::C.K1 sll0945 This study. ΔglgA2 ΔglgA1::SpΩ sll1393 This study. STXB slr1139::SpΩ slr1139 (Perez-Perez et al., 2009a) STXQ slr1139::SpΩ; slr0233::C.K1 slr0233 (Perez-Perez et al., 2009a) 23" " STXBQ slr0233::C.K1 slr1139; slr0233 (Perez-Perez et al., 2009a) STXC sll1057::C.K1 sll1057 This study. STXAc slr0623:: SpΩ slr0796::PgifB:trxA::C.K1 slr0623; slr0796 This study. Figure Legends. Figure 1. Glycogen synthases are not redox regulated. A. Glycogen synthase activity in untreated, oxidized and reduced conditions. Glycogen synthase activity of purified GlgA1 or GlgA2 was measured in vitro by measuring ADP formation as described in the Methods section. Proteins were not treated (control), oxidized in the presence of 10 µM of CuCl2 or oxidized with copper and subsequently reduced with 5 mM DTT. Activity was normalized to untreated sample. Data represents average of three independent experiments and error bars represent SE. B. Western blot analysis of GlgA1 under the same conditions used for activity measurements in A. Proteins were separated on a 7% acrylamide non-reducing gel and proteins detected using GlgA1 specific antibodies. C. Western blot analysis of GlgA2 under the same conditions used for activity measurements in A. Proteins were separated on a 7% acrylamide non-reducing gel and proteins detected using GlgA2 specific antibodies Figure 2. AGP is able to interact with TrxA. A. AGP interacts with TrxAC35S in Synechocystis extracts. Soluble extracts from WT and ΔAGP strains were oxidized with 25 µM CuCl2 for 30 min and incubated with 50 µM TrxAC35S for 1h to allow the formation of mixed disulfides. Proteins 24" " were resolved in 7% reducing (left panel) or non-reducing acrylamide gels (right panel) and probed with AGP antibodies. B. AGP interacts with TrxAC35S in vitro. Purified AGP was incubated with a 4 fold excess of TrxAC35S for 1 h to allow mixed disulfide to be formed, proteins were resolved in 12% reducing (left panel) or non-reducing (right panel) acrylamide gels and probed with TrxA antibodies. C. AGP is a tetramer in Synechocystis cell extracts. Extracts from light grown cells (top panel) or cells maintained in the dark for 8 h (bottom panel) were separated on a Superdex 200 column and fractions were probed for the presence of AGP by western blot on 10% acrylamide gels. The column was previously calibrated with molecular mass protein standards which are indicated by arrows. Figure 3. AGP activity is redox regulated. A. AGP activity is redox regulated in vitro. AGP activity was measured using 1 µg of purified protein as described in the Methods section. Protein was not treated (control), oxidized in the presence of 10 µM of CuCl2, oxidized with copper and subsequently reduced with 5 mM DTT or oxidized with copper and reduced with 100 µM of TrxA. Activity was normalized to the untreated sample. Data represents average of three independent experiments and error bars represent SE. B. Analysis of the oligomerization state of purified recombinant AGP. 2 µg of AGP were oxidized with 10 µg of CuCl2 and incubated with 2 mM DTT or increasing amounts of reduced TrxA. Proteins were separated on a 7% acrylamide nonreducing gel and Coomassie stained. C. AGP activity was measured in soluble extracts (0.5 µg total proteins) as described in the Methods section. Extracts were not treated (control), oxidized in the presence of 25 µM of CuCl2 and subsequently reduced with 5 mM DTT or oxidized with copper and reduced with 100 µM of TrxA. Activity was normalized to 25" " untreated sample. Data represents average of three independent experiments and error bars represent SE. D. Western blot analysis of AGP oligomerization state under the same conditions used for activity measurements in C. Extracts were separated on a 7% acrylamide non-reducing gel and probed with AGP antibodies. Figure 4. All conserved cysteines are not essential for AGP activity. A. AGP specific activity of Synechocystis extracts. AGP activity was measured from extracts prepared from WT, WTc, C45S, C315S, C320S and C337S strains. Data represents average of three independent experiments and error bars represent SE. B. Glycogen content in WT cells grown in long day conditions. Glycogen was determined in late exponential phase (10-15 µg chl mL-1) in WT (white circles) and ΔAGP cells (black circles) grown in a 16 h-light/8-h dark photoperiod for at least 3 cycles. C. Glycogen content in WT, WTc, C45S, C315S, C320S and C337S strains. Glycogen was determined in late exponential phase cells (10-15 µg chl mL-1), grown in a 16 h-light/8-h dark photoperiod for at least 3 cycles, at the end of the day (white bars) or at the end of the night (black bars). Figure 5. All conserved cysteines are involved in oxidoreduction of AGP. A. AGP redox state in cysteine mutants. Soluble extracts from WT, C45S, C315S, C320S and C337S strains were oxidized in the presence of 25 µM of CuCl2, 5 µg of total proteins were separated in 7% non-reducing (top panel) or reducing acrylamide gels (bottom panel) and the redox state of AGP was analyzed by western blot.