Glycogen Metabolism Supports Photosynthesis Startthrough the Oxidative Pentose Phosphate Pathwayin Cyanobacteria1[OPEN]

Shrameeta Shinde,a Xiaohui Zhang,a,b Sonali P. Singapuri,a Isha Kalra,a Xianhua Liu,b

Rachael M. Morgan-Kiss,a and Xin Wanga,2,3

aDepartment of Microbiology, Miami University, Oxford, Ohio 45056bSchool of Environmental Science and Engineering, Tianjin University, Tianjin 300354, China

ORCID IDs: 0000-0002-6185-9893 (I.K.); 0000-0001-5496-3011 (X.L.); 0000-0003-4783-5358 (R.M.M.-K.); 0000-0002-7174-3042 (X.W.).

Cyanobacteria experience drastic changes in their carbon metabolism under daily light/dark cycles. During the day, the Calvin-Benson cycle fixes CO2 and diverts excess carbon into glycogen storage. At night, glycogen is degraded to support cellularrespiration. The dark/light transition represents a universal environmental stress for cyanobacteria and other photosyntheticlifeforms. Recent studies revealed the essential genetic background necessary for the fitness of cyanobacteria during diurnalgrowth. However, the metabolic processes underlying the dark/light transition are not well understood. In this study, weobserved that glycogen metabolism supports photosynthesis in the cyanobacterium Synechococcus elongatus PCC 7942 whenphotosynthesis reactions start upon light exposure. Compared with the wild type, the glycogen mutant ΔglgC showed a reducedphotosynthetic efficiency and a slower P7001 rereduction rate when photosynthesis starts. Proteomic analyses indicated thatglycogen is degraded through the oxidative pentose phosphate (OPP) pathway during the dark/light transition. We confirmedthat the OPP pathway is essential for the initiation of photosynthesis and further showed that glycogen degradation through theOPP pathway contributes to the activation of key Calvin-Benson cycle enzymes by modulating NADPH levels. This strategystimulates photosynthesis in cyanobacteria following dark respiration and stabilizes the Calvin-Benson cycle under fluctuatingenvironmental conditions, thereby offering evolutionary advantages for photosynthetic organisms using the Calvin-Bensoncycle for carbon fixation.

Photosynthesis supports life on Earth by convertingsolar energy into chemical energy stored as organiccarbon. The Calvin-Benson cycle, the major carbonfixation pathway in cyanobacteria, algae, and plants, isemployed to reduce CO2 into organic carbonmolecules.Since its discovery in the 1950s, the enzymes of theCalvin-Benson cycle have been elucidated and theconsensus pathways established (Bassham et al., 1954).CO2 is reduced into the photosynthesis output glycer-aldehyde-3-phosphate (G3P) in three stages (i.e. carbon

fixation, carbon reduction, and carbon regeneration).The Calvin-Benson cycle is an elegant pathway inwhich the starting substrate ribulose-1,5-bisphosphate(RuBP) is continuously supplied by partially recyclingG3P back to RuBP through carbon rearrangement re-actions. This carbon regeneration stage resembles thenonoxidative portion of the pentose phosphate path-way. However, the Calvin-Benson cycle uses a keyenzyme, sedoheptulose-1,7-bisphosphatase (SBPase),to drive carbon regeneration rather than using trans-aldolase found in the pentose phosphate pathway(Sharkey andWeise, 2016). The activity of SBPase oftendetermines the photosynthetic capacity and carbonaccumulation in downstream metabolic processes(Harrison et al., 1997, 2001; Raines et al., 1999).In cyanobacteria, a major portion of the photosyn-

thetically fixed carbon is used to synthesize the storagepolymer glycogen. Two enzymes, ADP-Glc pyrophos-phorylase (AGPase) and glycogen synthase, catalyzethe sequential conversion of Glc-1-P (G1P) to 1,4-a-glucan (Preiss, 1984). Another enzyme encoded by a1,4-a-glucan-branching enzyme gene, glgB, can cata-lyze the formation of the 1,6-a branches of glycogen(Preiss, 1984). Glycogen synthesis is closely linkedto photosynthesis carbon output. The photosynthate3-phosphoglycerate (3PG) is an allosteric activator ofthe AGPase (Preiss, 1984; Gómez-Casati et al., 2003).

1This work was supported by Miami University startup funds toX.W. and by Department of Energy-BES grant DE-SC0019138 to X.W.and R.M.M.-K.

2Author for contact: xwang@miamioh.edu.3Senior author.The author responsible for distribution of materials integral to the

findings presented in this article in accordance with the policy de-scribed in the Instructions for Authors (www.plantphysiol.org) is:Xin Wang (xwang@miamioh.edu).

X.W. designed and performed the experiments, analyzed data, andwrote the article; S.S., X.Z., S.P.S., and I.K. performed the genetics,proteomics, and photochemistry experiments; R.M.M.-K. helped an-alyze the photochemistry data; S.S., X.L., and R.M.M.-K. edited thearticle.

[OPEN]Articles can be viewed without a subscription.www.plantphysiol.org/cgi/doi/10.1104/pp.19.01184

Plant Physiology�, January 2020, Vol. 182, pp. 507–517, www.plantphysiol.org � 2020 American Society of Plant Biologists. All Rights Reserved. 507 www.plantphysiol.orgon June 3, 2020 - Published by Downloaded from

Copyright © 2020 American Society of Plant Biologists. All rights reserved.

When photosynthetically fixed carbon is in excess, amajor carbon flux will be diverted to glycogen storage,serving as the carbon and electron sources for cellularrespiration in the dark (Preiss, 1984; Suzuki et al.,2010). More recently, glycogen metabolism has alsobeen recognized for additional roles played in cyano-bacterial carbon metabolism. These studies revealedthe involvement of glycogen metabolism as an energy-buffering system to maintain homeostasis (Cano et al.,2018) and as the carbon source for rapid resuscitationfrom nitrogen chlorosis (Doello et al., 2018).

Glycogen metabolism and cellular respiration arecritical for cell viability in the dark until light resumes(Lehmann and Wöber, 1976). Research in the past hasidentified the oxidative pentose phosphate (OPP)pathway (also known as the Glc-6-P [G6P] shunt) asthe major route for glycogen degradation in the dark(Smith, 1983; Broedel and Wolf, 1990). On the otherhand, the OPP pathway seems to be indispensableonly after long periods of darkness (over 24 h), whileglycolysis might have been able to compensate for theneed of a reducing equivalent during short incubationin the dark (Scanlan et al., 1995). However, it is alsoimportant to know that the OPP pathway is indis-pensable for cyanobacterial dark survival becausemany reactive oxygen species-detoxifying enzymesrequire NADPH as cofactors (Welkie et al., 2019). Incyanobacteria, cellular respiration is closely associatedwith photosynthesis. Many intermediates are sharedbetween the carbon fixation and respiration processes.When photosynthesis reactions start upon light, in-termediates in the Calvin-Benson cycle might be lim-ited after dark respiration, leading to stalled carbon

fixation reactions. Dynamic metabolic regulation toensure a smooth transition from dark respiration tophotosynthesis reactions is thus essential for the fitnessof cyanobacteria during diurnal growth (Puszynskaand O’Shea, 2017; Welkie et al., 2018). In this study,we discovered the active participation of glycogenmetabolism during the dark/light transition in cyano-bacteria. We found that glycogen degradation throughthe OPP pathway helps activate and stabilize theCalvin-Benson cycle reactions when photosynthesisreactions start upon light. We propose that this strategyhelps support photosynthesis during the dark/lighttransition and ensures the growth advantage of pho-tosynthetic organisms in their natural environment.

RESULTS

Proteomic Analysis Suggests a Supportive Role ofGlycogen Metabolism for Photosynthesis

Glycogen biosynthesis is a major carbon assimilatorypathway in the cyanobacterium Synechococcus elongatusPCC 7942. In this study, we generated a glycogensynthesis mutant (ΔglgC) by deleting the AGPase gene(glgC) in the S. elongatus genome (Fig. 1A). When wild-type and ΔglgC mutant cells were grown under con-tinuous light (50 mmol photons m22 s21), the ΔglgCmutant exhibited a longer lag phase relative to wild-type cells (Fig. 1B). This observation suggests that gly-cogen metabolism might play a supportive role for therapid start of photosynthesis (i.e. a short lag phase toquickly start the carbon fixation process).

Figure 1. Impact of glycogen metabo-lism on S. elongatus growth. A, Glyco-gen synthesis pathway in S. elongatus.The gene glgC was deleted from the S.elongatus genome to create the glyco-gen mutant (glgC::kan). B, Growthcurves of the S. elongatus wild type(WT) and DglgC mutant under contin-uous light. C, Comparative proteomicanalysis of the S. elongatus wild typeand DglgC mutant. Enzymes for bothglycogen biosynthesis (AGPase [glgC],glycogen synthase [glgA], and 1,4-a-glucan-branching enzyme [glgB])and degradation (a-1,4-glucan phos-phorylase [glgP]) were in higher abun-dance in wild-type compared withDglgCmutant cells. Bar graphs indicateenzyme expression levels based on thenormalized spectral abundance factor.Error bars indicate the SD; two-tailedStudent’s t test was used to compareprotein abundances: **, P , 0.01 and***, P , 0.001.

508 Plant Physiol. Vol. 182, 2020

Shinde et al.

www.plantphysiol.orgon June 3, 2020 - Published by Downloaded from Copyright © 2020 American Society of Plant Biologists. All rights reserved.

To understand how glycogen biosynthesis supportsphotosynthesis, we conducted a comparative proteo-mic analysis on the wild-type and ΔglgC mutant cells.The AGPase was not detected in the ΔglgC mutant,showing the success of constructing a clean mutant.All three enzymes related to glycogen synthesis (i.e.AGPase [glgC], glycogen synthase [glgA], and 1,4-a-glucan-branching enzyme [glgB]), were found to bein significantly higher abundance in the wild typecompared with the ΔglgC mutant (Fig. 1C). Interest-ingly, enzymes related to glycogen degradation werealso found to be more abundant in wild-type cells. Thea-1,4-glucan phosphorylase (glgP) that removes Glcfrom the glycogen chain was expressed over 4-foldhigher in wild-type cells (Fig. 1C; SupplementalTable S1). However, if glycogen degradation was tocontinue through the glycolytic Embden-Meyerhof-Parnas pathway, it would be a futile cycle to carbonfixation reactions. We hypothesize that the glycogenhydrolysis might proceed through other glycolyticpathways to support the rapid photosynthesis start inwild-type cells. We began to test our hypothesis byfirst validating the impact of abolished glycogen syn-thesis on photosynthetic functions.

The Glycogen Mutant Has Lower Photosynthetic EfficiencyWhen Photosynthesis First Starts

To test our hypothesis that glycogen metabolism cansupport a higher photosynthetic efficiency when pho-tosynthesis first starts, we measured the PSII efficiencyof wild-type and ΔglgC mutant cells using the WalzPhyto PAM II fluorometer. When cyanobacterial cellswere grown under continuous light, the maximumPSII efficiency of both the wild type and the ΔglgCmutant were relatively low in the first few hours ofgrowth but gradually reached above 0.4 (SupplementalFig. S1), consistent with values reported for S. elongatus(Campbell et al., 1998). The effective PSII efficiency (FII)exhibited a similar trend, with values ranging from 0.25to 0.35 in the first 20 h growing under continuous light

conditions (Fig. 2A). On the other hand, the FII of theΔglgCmutant was significantly lower than that of wild-type cells in the beginning of the growth phase and onlyrose to a comparable level at later stages of growth(Fig. 2A). Thus, the differential levels ofFII between thewild type and the ΔglgCmutant during the lag phase isconsistent with our hypothesis.To further validate the supportive role of glycogen

metabolism in starting photosynthesis, we monitoredFII following prolonged dark incubation. Cellular res-piration in the dark consumes the Calvin-Benson cycleintermediates, gradually lowering or depleting the C6,C5, and C3 sugar intermediates pool (Iijima et al., 2015;Diamond et al., 2017). Without the support from gly-cogen metabolism, we thus should observe a lowerphotosynthetic efficiency in ΔglgC cells when photo-synthesis restarts after the light cycle resumes. To en-sure the depletion of phosphate sugars, we incubatedcells in the dark for 30 h. Following dark incubation, theFII was significantly lower in the ΔglgC mutant com-pared with the wild type during the first few hours oflight (Fig. 2B).We further grew cyanobacterial cells under light/

dark cycles and measured the whole-cell oxygen evo-lution rate to further validate the photophysiology re-sults. Both wild-type and ΔglgC mutant cells weregrown under a light/dark (8 h/16 h) regime for a fewcycles, then the oxygen evolution rates were measuredat the end of the dark period and after light was turnedon for 2 h. When photosynthesis restarted followingdark incubation, the oxygen evolution rate was signif-icantly lower in the ΔglgC mutant compared with thewild type (Fig. 2C), supporting our hypothesis on therole of glycogen metabolism in supporting a rapid startof photosynthesis.

PSI-Mediated Cyclic Electron Flow Is Lower in theGlycogen Mutant

Cyclic electron flow (CEF) around PSI contributeslargely to additional ATP requirements in cyanobacterial

Figure 2. PSII efficiency of S. elongatuswild-type (WT) and glycogenmutant cells. A,FII measured under continuous light. PairedStudent’s t test showed a P value of 0.0193 between the wild type and the DglgCmutant. B, FII in dark-incubated cells. Both FII

following the dark incubation (0 h) and after light was turned on for 4 h were measured to monitor the FII during photosynthesisrestart. C, Oxygen evolution rates of cells grown in light/dark (8 h/16 h) cycles. After a few cycles of light/dark growth, the oxygenevolution ratesweremeasured at the end of the dark period (0 h) and after the light was turned on for 2 h. Error bars indicate the SD;two-tailed Student’s t tests were used to compare PSII efficiencies and oxygen evolution rates: *, P , 0.05 and ***, P , 0.001.

Plant Physiol. Vol. 182, 2020 509

Glycogen Metabolism Kick-Starts Photosynthesis

www.plantphysiol.orgon June 3, 2020 - Published by Downloaded from Copyright © 2020 American Society of Plant Biologists. All rights reserved.

carbon metabolism (Mullineaux, 2014). We measuredthe CEF rate through P700 photooxidation under far-red (FR) light (Klughammer and Schreiber, 1994; Cooket al., 2019). Oxidized PSI (P7001) absorbs strongly atwavelengths between 810 and 820 nm (A820), whilereduced P700 has minimal absorbance at this range.Both wild-type and ΔglgCmutant cells exhibited a rapidincrease in A820 after the FR light was turned on, indi-cating the oxidation of PSI (P700 to P7001; SupplementalFig. S2). Once a stableA820 was attained, the FR lightwasturned off and the rate of rereduction of P7001 to P700(t1/2red) was calculated as an estimate of P7001 rere-duction from alternative electron donors, of which CEFis the major pathway (Yu et al., 1993). We moni-tored CEF rates in the wild type and the ΔglgC mutantunder both dark and light conditions. P700 kineticswere monitored in the log-phase S. elongatus wild typeand ΔglgC mutant both at the end of dark incubationand after light was turned on for 2 and 24 h. The ratioDA820/A820 was comparable for both strains under allconditions (Fig. 3A), indicating that, unlike PSII, theamount of photooxidizable P700 was not affected by themutation in the glycogen synthesis pathway. After darkincubation for 16 h, both wild-type and ΔglgC mutantcells exhibited relatively slow t1/2red, suggesting thatCEF is low in the cyanobacterial cultures followingdark incubation (Fig. 3B; Supplemental Fig. S2).However, the wild-type cells exhibited more than atwofold faster t1/2red after 2 h of incubation in the light.After 24 h in the light, the t1/2red of both the wild typeand the ΔglgC mutant recovered to comparable levels(Fig. 3B; Supplemental Fig. S2).

Glycogen Metabolism Stimulates Photosynthesis throughthe OPP Pathway

To determine how glycogen metabolism can initiatephotosynthesis more rapidly, we conducted anothercomparative proteomics study on dark-incubated cya-nobacterial cells. Log-phase wild-type and ΔglgC mu-tant cells were dark incubated for 16 h before the light

was turned on. After 2 h in the light, cells were collectedfor proteomic analysis. Similar to log-phase cells col-lected under continuous light (Fig. 1), the enzymes forglycogen metabolism (i.e. AGPase [glgC], glycogensynthase [glgA], 1,4-a-glucan-branching enzyme [glgB],and a-1,4-glucan phosphorylase [glgP]) were found insignificantly lower abundances in the ΔglgC mutantcompared with wild-type cells (Fig. 4A). Interestingly,the enzyme 6-phosphogluconate dehydrogenase (gnd),a key enzyme of the OPP pathway, was also in signif-icantly lower abundance in ΔglgC cells (Fig. 4A). Theproteomic analysis indicates that the glycogen degra-dation could proceed through the OPP pathway.

To validate the involvement of the OPP pathway instimulating a rapid photosynthesis start, we generatedan OPP pathway mutant strain (Dgnd) by deleting thegnd gene in the S. elongatus genome. The selection of gndover the G6P dehydrogenase gene (zwf) for the deletionwas to minimize the potential impact on other meta-bolic processes such as the Entner-Doudoroff pathway,which depends on Zwf (Chen et al., 2016). When pho-tosynthesis starts after dark incubation, the Dgnd strainshowed impaired growth compared with the wild type(Fig. 4B), indicating a critical role of the OPP pathway insupporting photosynthesis. If the glycogen degradationproceeds through the OPP pathway, a higher ATP/NADPH ratio would be required during carbon fixa-tion (Sharkey and Weise, 2016). We thus measured theP7001 rereduction rate as a proxy for the CEF. Thewild-type and Dgnd cells were grown to log phase undermoderate light levels (50 mmol photons m22 s21) beforethe light was turned off for 16 h. Following the transi-tion to light, the P7001 rereduction rates of both wild-type and Dgnd cells were slow after dark incubation(Fig. 4B). However, the t1/2red of wild-type cells quicklyrecovered after 2 h of incubation in the light, whereasthe t1/2red of the Dgnd cells was significantly slower(Fig. 4B). Even after 24 h in light, the t1/2red of the Dgndcells was still relatively slow compared with wild-typecells (Fig. 4B). This observation further supports ourhypothesis that the OPP pathway is crucial for the rapidstart of photosynthesis after dark incubation.

Figure 3. P7001 reduction in S. elongatus wild-type and glycogen mutant cells following dark incubation. A, The oxidizableP700 pool (ΔA820/A820). B, t1/2

red. Cultureswere incubated in the dark for 16 h. The P7001 reductionwasmeasured right after darkadaption (0 h), 2 h into the light (2 h), and 24 h into the light (24 h). The data were plotted from the means of nine replicates (threebiological3 three technical replicates). Error bars indicate the SE; two-tailed Student’s t tests were used to compare the oxidizableP700 pool and P7001 rereduction rates: **, P , 0.01.

510 Plant Physiol. Vol. 182, 2020

Shinde et al.

www.plantphysiol.orgon June 3, 2020 - Published by Downloaded from Copyright © 2020 American Society of Plant Biologists. All rights reserved.

We further conducted light/dark cycle growth ex-periments to evaluate the role of glycogen metabolismand the OPP pathway in supporting the rapid start ofphotosynthesis. The S. elongatus wild type and twomutants (ΔglgC and Δgnd) were grown under twolight/dark cycle (12 h/12 h and 8 h/16 h) conditions.Compared with the wild type, ΔglgC and Δgndmutantshad much slower photosynthesis start after the darkperiod under both conditions (Fig. 4C). Under the 12-h/12-h light/dark condition, the photosynthesis restartrate of the ΔglgC mutant was lower in the first fewlight/dark cycles but gradually reached similar levelsto the wild type at later cycles of growth. The Δgndmutant showed similar growth to the ΔglgCmutant forthe first few light/dark cycles but completely abolishedits growth afterward, indicating the crucial role of theOPP pathway in the photosynthesis-starting processduring diurnal growth. The phenotype was even moreobvious in the 8-h/16-h light/dark cycle experiment.Throughout the growth cycle, the ΔglgC mutant hadmuch slower photosynthesis start after each dark pe-riod compared with the wild type, suggesting limitedavailability of intermediates needed for photosynthesisrestart after a 16-h dark incubation. On the other hand,the growth of the Δgnd mutant was completely abol-ished when growing under the 8-h/16-h light/darkcycles (Fig. 4C). The growth experiment strongly sup-ports our hypothesis that glycogen metabolism

stimulates photosynthesis through the OPP pathwayduring cyanobacterial diurnal growth.

The OPP Pathway Contributes to the Activation of theCalvin-Benson Cycle

Upon restarting of photosynthesis reactions follow-ing dark incubation, the intermediate pool needed toregenerate RuBP in the Calvin-Benson cycle may belimited or depleted (Iijima et al., 2015). Continuoussupply of RuBP during the dark/light transition islikely critical for maintaining homeostatic balance be-tween energy production by the light reactions andenergy consumption by the Calvin-Benson cycle. Toevaluate intermediate availability during cyanobacte-rial diurnal growth, we analyzed the intermediates ofcentral metabolism in the wild type and the ΔglgCmutant grown under light/dark (8 h/16 h) cycles. After16 h of dark incubation, several intermediates in theCalvin-Benson cycle, including Fru-6-P (F6P), ribose-5-phosphate (R5P), and Ru5P, were found in low abun-dances in both the wild type and the ΔglgC mutant;however, after short exposure in the light, these inter-mediates were quickly replenished in the wild type butnot in the ΔglgCmutant (Fig. 5A). In the mean time, theglycogen degradation products, G1P and G6P, in-creased to much higher levels in the wild type com-pared with the ΔglgCmutant after short exposure in the

Figure 4. Glycogen metabolism stimulates photosynthesis through the OPP pathway. A, Brown arrows indicate the proposedmechanism in which glycogen degrades through the OPP pathway to recharge photosynthesis during the start of photosynthesis.Bar graphs indicate the protein levels fromproteomic analysis. Error bars indicate the SD. B, Growth profile and P7001 rereductionanalysis of the wild type (WT) and theOPP pathwaymutant Δgnd. The bar graph indicates the calculated t1/2red. Error bars indicatethe SE. C, Light/dark cycle growth profile of the wild type, the ΔglgCmutant, and the Δgndmutant. Cells were grown in quadrupletat 30°C with 50 mmol photons m22 s21 illumination in the Multi-Cultivator MC-1000-OD-MULTI. Two-tailed Student’s ttests were used to compare protein abundances and P7001 rereduction rates: *, P, 0.05; **, P, 0.01; ***, P, 0.001; and ****,P , 0.0001. 6PGL, 6-Phosphogluconolactone; Ru5P, ribulose-5-phosphate.

Plant Physiol. Vol. 182, 2020 511

Glycogen Metabolism Kick-Starts Photosynthesis

www.plantphysiol.orgon June 3, 2020 - Published by Downloaded from Copyright © 2020 American Society of Plant Biologists. All rights reserved.

light (Fig. 5A), suggesting robust glycogen degradationin the wild type to replenish the Calvin-Benson cycleintermediates. Moreover, the level of the carbon fixa-tion product 3PG stayed at similar levels during thedark/light transition period in the wild type, whereasthe ΔglgCmutant had slightly decreased levels (Fig. 5A).It is also worth mentioning that 3PG was severely de-pleted in the ΔglgCmutant comparedwith thewild type,which is in line with the lower photosynthetic efficiencyin the ΔglgC mutant (Fig. 2).

During the dark/light transition, glycogen degrada-tion through the Embden-Meyerhof-Parnas pathwaywould form a futile cycle to carbon fixation reactions.The obvious route for glycogen degradation is thusthrough the OPP pathway to replenish C5 carbon pools.This hypothesis is supported by the observation of in-creased expression of 6-phosphogluconate dehydro-genase in the wild type relative to the ΔglgC mutantduring the start of photosynthesis (Fig. 4A). Ru5Pderived from the OPP pathway can then be phos-phorylated to RuBP by phosphoribulokinase (PRK),replenishing the starting substrate for Rubisco tocontinuously fix CO2 in the Calvin-Benson cycle.However, PRK is inactive in the dark and requiresredox regulation to become active in the light (Tamoiet al., 2005). To validate whether glycogen degrada-tion through the OPP pathway contributes to the ac-tivation of the Calvin-Benson cycle during the dark/light transition, we measured PRK activities in cellsgrown under light/dark (8 h/16 h) cycles. After a fewlight/dark cycles, both wild-type and ΔglgC cells werecollected at the end of the dark period and when lightwas turned on for 2 h. After light exposure, PRK

activity in the wild type increased significantly whereasthe increase was limited in the ΔglgC mutant (Fig. 5B).We further measured the NAD(P)H levels in both thewild type and the ΔglgC mutant. When photosynthesisreactions restart after the dark period, the NADPH levelwas significantly lower in the ΔglgC mutant comparedwith the wild type (Fig. 5C). Collectively, these resultssupport our hypothesis that glycogen metabolism helpsmodulate NADPH levels to activate the Calvin-Bensoncycle for a rapid start of photosynthesis.

DISCUSSION

Photosynthetic organisms experience drastic changesin their carbon metabolism during diurnal growth.Recent research has highlighted the contribution ofglycogen metabolism and the OPP pathway to the fit-ness of cyanobacteria during diurnal growth (Diamondet al., 2017; Puszynska and O’Shea, 2017; Welkie et al.,2018). In this study, we provided strong evidence of theactive participation of glycogen metabolism and theOPP pathway in supporting photosynthesis duringthe dark/light transition period of diurnal growth. Thissupportive role is likely essential for the survival ofcyanobacteria in their natural environment.

Glycogen Degradation through the OPP Pathway Serves asthe Perfect Buffering System for the StalledCalvin-Benson Cycle

Cellular respiration from glycogen degradation helpsmaintain cell viability in the dark. When light resumes,

Figure 5. Glycogen metabolism contrib-utes to activation of the Calvin-Bensoncycle during the dark/light transition. Af-ter a few light/dark (8 h/16 h) cycles, wild-type (WT) and ΔglgC mutant cells werecollected right after the dark period (0 h),30 min into the light (0.5 h), and 2 h intothe light (2 h) for the metabolomicsanalysis (A), PRK activities at 0 and 2 h(B), and NAD(P)H measurements at 0and 2 h (C). Error bars indicate the SD; *,P , 0.05; **, P , 0.01; ***, P , 0.001;and ****, P , 0.0001.

512 Plant Physiol. Vol. 182, 2020

Shinde et al.

www.plantphysiol.orgon June 3, 2020 - Published by Downloaded from Copyright © 2020 American Society of Plant Biologists. All rights reserved.

many enzymes for cellular respiration are inactivated,shifting cell metabolism to carbon fixation (Udvardyet al., 1984; Gleason, 1996; Sharkey and Weise, 2016).However, several recent studies suggest that glycogenmetabolism plays an additional supportive role for thestart of photosynthesis. A glycogen phosphorylase(DglgP) mutant of Synechocystis sp. PCC 6803 retardsdark respiration of glycogen and had a much lowerphotosynthetic oxygen evolution rate (Shimakawaet al., 2014). Another study conducted in the DglgCmutant of Synechocystis sp. PCC 6803 exhibited adelayed activation of the Calvin-Benson cycle and di-minished photochemical efficiency when cultured un-der high-carbon conditions (Holland et al., 2016). Thesestudies indicate a tight link between glycogen metab-olism and photosynthesis. In this study, we showedthat glycogen metabolism can stabilize photosynthesisthrough the OPP pathway in the unicellular cyano-bacterium S. elongatus when photosynthesis reactionsstart upon light (Fig. 4). When cells were grown underlight/dark cycles, we found that levels of several in-termediates (e.g. F6P, R5P, and Ru5P) in the Calvin-Benson cycle were low after dark respiration(Fig. 5A). The replenishment of these intermediates isessential for RuBP regeneration in the Calvin-Bensoncycle. However, due to limited RuBP availabilityupon photosynthesis restart, newly synthesized triosephosphate would be insufficient to replenish theCalvin-Benson cycle intermediates, leading to stalledcarbon fixation reactions. The alternative carbon flowfrom glycogen degradation through the OPP pathwayis thus essential for RuBP regeneration to ensure thecontinuous carbon fixation. The importance of this al-ternative carbon flow was shown in the ΔglgC mutantin that cells could not replenish these intermediatesefficiently without the glycogen synthesis pathway(Fig. 5A). The importance of this pathway is furthermanifested by the lack of efficiency to synthesize triosephosphates via phosphofructokinase and phosphoglu-coisomerase, both of which require high levels of sub-strate (Knowles and Plaxton, 2003; Preiser et al., 2018).Glycogen metabolism thus serves as a perfect bufferingsystem to supply G6P for the OPP pathway throughglycogen degradation (Gómez-Casati et al., 2003). Inaddition, our proteomics results showed that transal-dolase, a nonessential enzyme in the nonoxidativephase of the pentose phosphate pathway, was also insignificantly higher abundance in the wild type com-pared with the DglgC mutant when photosynthesis re-actions first start (Supplemental Table S1).Interestingly, a recent genome-wide fitness study in S.elongatus showed a similar result in that the transaldolasemutant was sensitive to light/dark cycles (Welkie et al.,2018). These results suggest an underappreciated rolefor transaldolase during the operation of the OPPpathway.Our findings echo with the proposed mechanism by

Sharkey and Weise (2016) for the involvement of theOPP pathway to stabilize photosynthesis in plants.They suggested that the OPP pathway (the G6P shunt)

could account for 10% to 20% of Rubisco activity andthat the enzyme SBPase plays an important role incontrolling the carbon flow between the Calvin-Bensoncycle and the shunt (Sharkey and Weise, 2016). Theirrecent evidence showed that the G6P shunt was acti-vated in a photorespiration mutant of Arabidopsis(Arabidopsis thaliana) in which the activity of triosephosphate isomerase (TPI) was inhibited (Li et al.,2019). TPI is responsible for interconverting G3P anddihydroxyacetone phosphate, a key reaction for RuBPregeneration in the Calvin-Benson cycle. During pho-tosynthesis, it is critical that RuBP can be continuouslyregenerated for CO2 fixation, which limits the carbonregeneration reactions to a small margin of error. Per-turbation to carbon regeneration reactions (e.g. TPIinhibition) would lead to a temporarily impairedCalvin-Benson cycle. An alternative mechanismsuch as the G6P shunt is thus essential to stabilizephotosynthesis.

Glycogen Metabolism Is Critical for the OptimalPhotosynthesis Performance

Glycogen metabolism and the OPP pathway arecritical for optimal photosynthetic performance duringthe start of photosynthesis. This can be observedthrough our photochemistry measurements of the wildtype and the ΔglgC mutant. Compared with the wildtype, the photosynthetic efficiency of the ΔglgC mutantwas significantly lower when photosynthesis starts(Fig. 2), suggesting that PSII is likely down-regulatedthrough a feedback mechanism in the ΔglgC mutant.The temporally impaired PSII in the ΔglgC mutantcould be partially explained by the proteomics data.Proteomic analysis of the dark-adapted wild type andΔglgCmutant showed differential expression of severalproteins associated with PSII (Supplemental Table S1).Most notably, the abundance of the ATP-dependentzinc metalloprotease FtsH was over 2-fold higher inthe ΔglgC mutant in comparison with the wild type. Inboth plants and cyanobacteria, the FtsH protease playsimportant roles in repairing damaged PSII by degrad-ing the D1 protein (Lindahl et al., 2000; Silva et al.,2003). Increased FtsH abundance in the ΔglgC mutantsuggests an active PSII repair cycle when light wasturned on. Without the support from glycogen metab-olism, the photosynthetic electron transport chain islikely overreduced due to an imbalance of energyproduction and consumption, leading to PSII photo-damage in the ΔglgC mutant. This hypothesis is sup-ported by the finding of additional key proteinsinvolved in PSII repair. Both carotene isomerase and theprotein Psb28 were in higher abundance in the ΔglgCmutant compared with the wild type (SupplementalTable S1). Carotene isomerase is an important enzymefor carotenoid synthesis. Past studies in cyanobacteriahave shown the active participation of carotenoidsduring PSII assembly (Masamoto et al., 2004; Tóth et al.,2015). Moreover, Psb28 was found to be involved in

Plant Physiol. Vol. 182, 2020 513

Glycogen Metabolism Kick-Starts Photosynthesis

www.plantphysiol.orgon June 3, 2020 - Published by Downloaded from Copyright © 2020 American Society of Plant Biologists. All rights reserved.

protecting the PSII RC47 assembly intermediate duringits conversion to functional PSII in Synechocystis sp.PCC 6803 (Dobáková et al., 2009; Weisz et al., 2017).

For PSI performance, we applied P7001 rereductionrates to evaluate the alternative electron flows return-ing to PSI as a proxy for CEF.When light was turned onfollowing dark incubation, we observed differentP7001 rereduction rates in thewild type comparedwithboth mutants (ΔglgC and Δgnd). Both wild-type andΔglgC mutant cells had comparably slow P7001 rere-duction rates after dark incubation; however, the wildtype exhibited rapid recovery within 2 h in the light,while it took 24 h for the ΔglgC mutant to reach acomparable level to the wild-type cells (Fig. 3). Whencomparing the wild type and the Δgnd mutant, theP7001 rereduction rate showed a similar trend (i.e. amuch faster rereduction rate in the wild type after ashort incubation in the light; Fig. 4B). This stronglysuggests a higher ATP/NADPH ratio requirementwhen photosynthesis first starts and is directly linked toglycogenmetabolism and the OPP pathway. The actionof the OPP pathway alongside the Calvin-Benson cycleexacerbates the deficit of ATP/NADPH ratio neededfor carbon fixation (Sharkey and Weise, 2016). It is wellaccepted that the additional ATP requirement in pho-tosynthetic carbon metabolism can be fulfilled throughthe CEF (Kramer and Evans, 2011). The higher CEF rateobserved in the wild-type cells fits well with the ad-ditional ATP requirements for operating the OPPpathway during the start of photosynthesis. Unlikethe DglgCmutant, the Δgndmutant did not recover toa comparable level of P7001 rereduction rate evenafter 24 h in the light, indicating that the OPP path-way is essential for the initiation of photosynthesisand the fitness of cells (Fig. 4B). The additional light/dark cycle growth experiments further showed theimportance of the OPP pathway in supporting thestart of photosynthesis. When growing under 12-h/12-h light/dark cycles, the Δgnd mutant showedsimilar growth to the ΔglgC mutant during the firstfew cycles but stopped growing afterward. On theother hand, growth of the Δgnd mutant was com-pletely abolished under 8-h/16-h light/dark growthcycles (Fig. 4C), suggesting depletion of the Calvin-Benson cycle intermediates during long periods ofdarkness and the necessity of the OPP pathway tostart photosynthesis.

Glycogen Metabolism Participates in Activation of theCalvin-Benson Cycle during CyanobacterialDiurnal Growth

Glycogen degradation through the OPP pathwaycould replenish RuBP to ensure the continuous opera-tion of carbon fixation during the start of photosyn-thesis. However, it is important to know that the OPPpathway does not lead to any net carbon fixation. CO2fixed in the Calvin-Benson cycle is quickly lost whenG6P is oxidized to Ru5P in the shunt. It is thus

intriguing to understand the role of the OPP pathway insupporting photosynthesis. We propose that the OPPpathway could generate additional NADPH requiredfor the redox regulation of the Calvin-Benson cycleduring the start of photosynthesis. Previous researchshowed that the activities of two key enzymes in theCalvin-Benson cycle, PRK and G3P dehydrogenase(GADPH), are suppressed in the dark through theformation of a CP12/PRK/GAPDH protein complex(Tamoi et al., 2005). The reversible dissociation of thecomplex was controlled by the pyrimidine nucleotidelevels in S. elongatus (Tamoi et al., 2005). The operationof the OPP pathway during the start of photosynthesisthus could contribute largely to the increase of NADPHlevels when light resumes, releasing PRK and GAPDHfrom the complex to activate the Calvin-Benson cycle.This hypothesis is supported by the measurements ofboth PRK activities and NADP(H) levels during thestart of photosynthesis. Following light/dark (8 h/16 h)cycles, both PRK activities and NADPH levels weresignificantly higher in the wild type relative to theΔglgCmutant after the light was turned on (Fig. 5, B andC). However, NAPDH generation from the OPP path-way relative to that from the light reactions is worthy offurther investigation.

Lastly, it is also important to understand the activa-tion mechanism of the OPP pathway. The expression ofmany genes in glycogen metabolism and the OPPpathway are controlled by the circadian clock outputprotein RpaA (Markson et al., 2013; Welkie et al., 2018).Our proteomic analysis confirmed that many enzymesin these two metabolic processes were in much higherabundance in wild-type cells in the beginning of pho-tosynthesis (Fig. 4A; Supplemental Table S1). Previousresearch also reported that the first enzyme in the OPPpathway, G6P dehydrogenase, is redox regulated andis inactive in its reduced form (Udvardy et al., 1984;Gleason, 1996). However, G6P could also stabilize andactivate G6P dehydrogenase to reverse the reductioninhibition in cyanobacteria (Cossar et al., 1984). It isthus likely that glycogen degradation would lead toincreased G6P levels and activate the G6P dehydro-genase. Our metabolite analysis supports this hypoth-esis in that G6P levels quickly increased during thedark/light transition period in the wild type comparedwith the ΔglgCmutant (Fig. 5A). Future studies towardunderstanding these mechanisms could help fully ap-preciate the roles played by the OPP pathway in sup-porting photosynthesis.

In conclusion, we found that glycogen metabolism incyanobacteria helps activate photosynthesis throughthe OPP pathway during the start of photosynthesisreactions. The OPP pathway is generally considered forits role to generate reducing equivalents for dark res-piration. However, evolution empowers cells with theability to leverage existing metabolic processes for theadaptation of constantly changing environments. Wehave witnessed an interesting strategy implemented bycyanobacteria to ensure their growth advantage in theirnatural environment.

514 Plant Physiol. Vol. 182, 2020

Shinde et al.

www.plantphysiol.orgon June 3, 2020 - Published by Downloaded from Copyright © 2020 American Society of Plant Biologists. All rights reserved.

MATERIALS AND METHODS

Growth Conditions

Seed cultures of Synechococcus elongatus wild-type and mutant strains weregrown in BG11 medium supplemented with 20 mM sodium bicarbonate and10 mM TES (pH 8.2) at 30°C under 30 mmol photons m22 s21 illumination. Thegrowth medium of the mutant strains was also supplemented with 5 mg L21

kanamycin. Cultures for measurements were grown at 30°C under 50 mmolphotons m22 s21 illumination in the Multi-Cultivator MC-1000-OD-MULTI(Photon Systems Instruments). Cyanobacterial growth was monitored usingthe built-in spectrophotometer of Multi-Cultivator by measuring OD720.S. elongatus wild-type and DglgC mutant cells used for oxygen evolution mea-surement, metabolite analysis, PRK enzyme activity, and NADP(H) levels weregrown in triplicate under light/dark (8 h/16 h) cycles at 30°C with 50 mmolphotons m22 s21 illumination in the Multi-Cultivator MC-1000-OD-MULTI.

Plasmid and Strain Construction

The gene fragments for plasmid construction were amplified using primersdesigned through the NEBuilder Assembly tool (New England Biolabs). TheglgC and gnd knockout plasmids were constructed using HiFi DNA Assemblymaster mix (New England Biolabs). S. elongatus wild-type cells were trans-formed with knockout plasmids to generate DglgC and Dgnd mutant strains.The strains and plasmids along with primers used in this study are listed inSupplemental Tables S2 and S3 and Supplemental Data S1 and S2.

Proteomics

Log-phase S. elongatus wild-type and DglgC mutant cells were used forproteomic analysis following our previously described method (Wang et al.,2016). For dark-incubated samples, log-phase cells were taken after light wasturned on for 2 h following dark incubation. Briefly, biological triplicates ofapproximately 10 OD of cells were used for protein extraction. Cell pellets wereresuspended in 1 mL of Tris buffer (50 mM Tris-HCl, 10 mM CaCl2, and 0.1%[w/v] n-dodecyl b-D-maltoside, pH 7.6) followed by cell lysis using a homog-enizer with bead-beating cycles of 10 s on and 2 min off on ice for a total of sixcycles. Protein extract in the supernatant was collected by centrifugation at13,000g for 30 min at 4°C, followed by protein concentration measurementusing the Bradford assay (Thermo Scientific).

For each sample, 100 mg of total protein was digested by Trypsin Gold(Promega) with a 1:100 (w/w) ratio at 37°C for 18 h. The digested peptides werecleaned up using a Sep-Pak C18 column (Waters) followed by peptide frac-tionation using the Pierce High pH Reverse-Phase Peptide Fractionation Kit(Thermo Scientific). Eight fractions of peptides from each sample were sub-jected to liquid chromatography-tandem mass spectrometry (MS/MS) analysisin a Thermo LTQ Orbitrap XL mass spectrometer. The mass spectrometer wasoperated under the data-dependent mode scanning the mass range of 350 to1,800mass-to-charge ratio at a resolution of 30,000. The 12most abundant peakswere subjected to MS/MS analysis by collision-induced dissociation fragmen-tation. The raw data collected from MS/MS analysis was searched and ana-lyzed using the pipeline programs integrated in the PatternLab for Proteomics(version 4.1.0.17; Carvalho et al., 2016). The normalized spectral abundancefactor was used to compare protein abundance in different groups (Zybailovet al., 2006). The mass spectrometry proteomics data have been deposited toboth theMassIVE repository with the data set identifierMSV000083575 and theProteomeXchange Consortium (Vizcaíno et al., 2016) with the data set identifierPXD013099.

PSII Analysis

Cyanobacterial cells grown under continuous light (50mmol photonsm22 s21)and dark-incubated cells were subjected to PSII measurement using the Phyto-PAMPhytoplanktonAnalyzer (HeinzWalz).WhenmeasuringPSII efficiency, thedark adaption that normally oxidizes the PSII center in green algae does notwork in S. elongatus. Instead, cells would transit to state II after dark adap-tation (Campbell et al., 1998). To get a proper measurement of PSII efficiency,cells were illuminated at moderate actinic light (25 mmol m22 s21) for 5 min tolock the photosystems in state I and to oxidize PSII reaction centers. The bluelight was used as the measuring light to minimize the background fluores-cence from phycobilisome light absorption. Standard induction curves by

exposing cells to saturation pulses were recorded to determine the maximumPSII efficiency and the FII.

Oxygen Evolution

S. elongatuswild-type and DglgCmutant cells were collected following a fewlight/dark cycles. Two milliliters of cells was harvested both at the end of thedark period and after light was turned on for 2 h. Fifty microliters of cells wereused for cell counting in an Attune NxT Flow Cytometer (Invitrogen). After ashort incubation in the dark, whole-cell oxygen evolution was measured for2 min at room temperature with saturation light illumination in a Clark-typeoxygen electrode (Hansatech Instruments). Since the growth medium wassupplemented with 20 mM NaHCO3, no additional bicarbonate was added inthe electrode chamber. The final oxygen evolution rate was corrected with thedark respiration rate and normalized to cell numbers.

PSI Analysis

Log-phase S. elongatus wild-type and mutant cells were incubated in thedark for 16 h before the PSI analysis. Five milliliters of the culture was vacuumfiltered through Whatman GF/C 25-mm filter discs (catalog no. 1822-025). FRlight-induced P700 oxidation-reduction kinetics measurements were per-formed using a dual-wavelength pulse amplitude-modulated fluorescencemonitoring system (Dual-PAM-100; Heinz Walz) with a leaf attachment. Theproportion of photooxidizable P700 (DA820/A820) was determined as the changein A820 after turning on the FR light (lmax 5 715 nm, 10 W m22; Scott filter RG715). The t1/2red was calculated after the FR light was turned off and used as anestimate of alternative electron flow around PSI.

Quantification of Central Metabolites

The metabolite analysis was conducted at the West Coast MetabolomicsCenter at the University of California, Davis. Ten milliliters of S. elongatuswild-type and DglgC cells was quenched in 15 mL of cold phosphate-buffered saline(PBS), followed by centrifugation at 7,800g for 15 min at 4°C to collect the cellpellet. Intermediates of the primary metabolism were analyzed by gaschromatography-time of flight-mass spectrometry and identified using theBinBase algorithm (Fiehn et al., 2008). The metabolite levels were normalized tocell numbers and used for comparison.

PRK Enzyme Assay

Following a few light/dark cycles, 10mL of S. elongatuswild-type andDglgCmutant cells was harvested by injecting into 15mL of cold PBS both at the end ofthe dark period and after light was on for 2 h. Cells pellets were collected bycentrifugation at 7,800g for 15min at 4°C, followed by resuspension in 500mL of50 mM Tris-HCl (pH 8). Cells were then disrupted in a bead beater for a total of1 min with 10 s on and 2 min off on ice. The supernatant was collected bycentrifugation at 15,000 rpm for 20 min at 4°C. An aliquot of the supernatantwas used to determine the protein concentration by the Bradford assay (ThermoScientific). The PRK assay was carried out by monitoring NADH A340 on a 96-well plate by modifying a previously described method (Kanno et al., 2017).Briefly, cell lysate was preincubated at 30°C for 30 min to obtain the baselineabsorbance. After the incubation, an enzyme mixture containing 0.5 mM Ru5P,2 mM ATP, 2.5 mM phosphoenolpyruvate, 10 mM MgCl2, 0.3 mM NADH, 4 unitsof pyruvate kinase, and 4 units of lactate dehydrogenase was added to the celllysate to make the final reaction mixture to 100 mL. NADH absorbance wasrecorded every 1min for a total of 30 min. The Δ340 nm/minwas not corrected tocuvette absorbance for enzyme unit calculation. Specific PRK activity wasobtained by normalizing the rate to total protein levels of each sample.

Measurement of NAD(P)H Levels

Following a few light/dark cycles, 10mL of S. elongatuswild-type andDglgCcells was collected both at the end of the dark period and after light was turnedon for 2 h. Cells were quickly transferred to 15 mL of cold PBS after collection,followed by centrifugation at 7,800g for 15 min at 4°C. Cell pellets wereresuspended in 500 mL of PBS and lysed by the addition of 500 mL of base so-lution (0.2 N NaOH) with 1% (w/v) dodecyltrimethylammonium bromide.Following centrifugation at 15,000 rpm for 20 min at 4°C, the supernatant wascollected and passed through a Pierce Concentrator spin column (catalog no.

Plant Physiol. Vol. 182, 2020 515

Glycogen Metabolism Kick-Starts Photosynthesis

www.plantphysiol.orgon June 3, 2020 - Published by Downloaded from Copyright © 2020 American Society of Plant Biologists. All rights reserved.

PI88513; Fisher Scientific) to remove enzymes that could affect the NADP(H)assay. The NADP(H) levels in the flow-throughs were determined by theNADP/NADPH-Glo assay (catalog no. G9071; Promega) following the man-ufacturer’s instructions. Fifty microliters of cells was also collected for cellcounting in an Attune NxT Flow Cytometer (Invitrogen). The amount ofNADP(H) in each sample was calculated from a NADPH standard curve andnormalized to cell numbers.

Statistical Analyses

Thestatistical analyses for all experimentswereperformedusing the softwarePrism. Two-tailed Student’s t tests were used to compare wild-type andmutantcells for differences in protein abundances, PSII and PSI efficiencies, oxygenevolution rates, PRK enzyme activities, and NADP(H) levels. One-wayANOVA was used to analyze the temporal changes of metabolite abundancein the wild type and the DglgC mutant during the start of photosynthesis. Avalue of P , 0.05 between comparing groups was considered statisticallydifferent.

Supplemental Data

The following supplemental materials are available.

Supplemental Figure S1. Maximum PSII efficiency of S. elongatus wild-type and ΔglgC mutant cells.

Supplemental Figure S2. P7001 reduction curve in S. elongatus wild-typeand glycogen mutant cells following dark incubation.

Supplemental Table S1. Comparative proteomics analysis of S. elongatuswide-type and ΔglgC cells.

Supplemental Table S2. Plasmids and strains used in this study.

Supplemental Table S3. Oligonucleotides used for assembly of deletionmutants.

Supplemental Data S1. Map and sequence of pXWK3_glgC.

Supplemental Data S2. Map and sequence of pXWK3_gnd.

ACKNOWLEDGMENTS

We thank Spencer Diamond (University of California, Berkeley), JamesGolden, and Susan Golden (University of California, San Diego) for the originalplasmid used to generate the glycogen mutant strain in the preliminary study.We thank Theresa Ramelot and Kundi Yang of the Chemistry Department atMiami University for their help in the preliminary metabolite analysis. Wethank the Center for Bioinformatics and Functional Genomics at Miami Uni-versity for providing access to several instruments used in this study. We thankJianping Yu (U.S. National Renewable Energy Laboratory) and Graham Peers(Colorado State University) for careful reading of the article.

Received September 27, 2019; accepted October 17, 2019; published October 24,2019.

LITERATURE CITED

Bassham JA, Benson AA, Kay LD, Harris AZ, Wilson AT, Calvin M (1954)The path of carbon in photosynthesis. XXI. The cyclic regeneration ofcarbon dioxide acceptor. J Am Chem Soc 76: 1760–1770

Broedel SE Jr, Wolf RE Jr (1990) Genetic tagging, cloning, and DNA se-quence of the Synechococcus sp. strain PCC 7942 gene (gnd) encoding6-phosphogluconate dehydrogenase. J Bacteriol 172: 4023–4031

Campbell D, Hurry V, Clarke AK, Gustafsson P, Öquist G (1998) Chlo-rophyll fluorescence analysis of cyanobacterial photosynthesis and ac-climation. Microbiol Mol Biol Rev 62: 667–683

Cano M, Holland SC, Artier J, Burnap RL, Ghirardi M, Morgan JA, Yu J(2018) Glycogen synthesis and metabolite overflow contribute to energybalancing in cyanobacteria. Cell Rep 23: 667–672

Carvalho PC, Lima DB, Leprevost FV, Santos MD, Fischer JS, Aquino PF,Moresco JJ, Yates JR III, Barbosa VC (2016) Integrated analysis ofshotgun proteomic data with PatternLab for proteomics 4.0. Nat Protoc11: 102–117

Chen X, Schreiber K, Appel J, Makowka A, Fähnrich B, Roettger M,Hajirezaei MR, Sönnichsen FD, Schönheit P, Martin WF, et al (2016)The Entner-Doudoroff pathway is an overlooked glycolytic route incyanobacteria and plants. Proc Natl Acad Sci USA 113: 5441–5446

Cook G, Teufel A, Kalra I, Li W, Wang X, Priscu J, Morgan-Kiss R (2019)The Antarctic psychrophiles Chlamydomonas spp. UWO241 and ICE-MDV exhibit differential restructuring of photosystem I in response toiron. Photosynth Res 141: 209–228

Cossar JD, Rowell P, Stewart WDP (1984) Thioredoxin as a modulator ofglucose-6-phosphate-dehydrogenase in a N2-fixing cyanobacterium.J Gen Microbiol 130: 991–998

Diamond S, Rubin BE, Shultzaberger RK, Chen Y, Barber CD, Golden SS(2017) Redox crisis underlies conditional light-dark lethality in cyano-bacterial mutants that lack the circadian regulator, RpaA. Proc NatlAcad Sci USA 114: E580–E589

Dobáková M, Sobotka R, Tichý M, Komenda J (2009) Psb28 protein isinvolved in the biogenesis of the photosystem II inner antenna CP47(PsbB) in the cyanobacterium Synechocystis sp. PCC 6803. Plant Physiol149: 1076–1086

Doello S, Klotz A, Makowka A, Gutekunst K, Forchhammer K (2018) Aspecific glycogen mobilization strategy enables rapid awakening ofdormant cyanobacteria from chlorosis. Plant Physiol 177: 594–603

Fiehn O, Wohlgemuth G, Scholz M, Kind T, Lee DY, Lu Y, Moon S,Nikolau B (2008) Quality control for plant metabolomics: ReportingMSI-compliant studies. Plant J 53: 691–704

Gleason FK (1996) Glucose-6-phosphate dehydrogenase from the cyano-bacterium, Anabaena sp. PCC 7120: Purification and kinetics of redoxmodulation. Arch Biochem Biophys 334: 277–283

Gómez-Casati DF, Cortassa S, Aon MA, Iglesias AA (2003) Ultrasensitivebehavior in the synthesis of storage polysaccharides in cyanobacteria.Planta 216: 969–975

Harrison EP, Olcer H, Lloyd JC, Long SP, Raines CA (2001) Small de-creases in SBPase cause a linear decline in the apparent RuBP regener-ation rate, but do not affect Rubisco carboxylation capacity. J Exp Bot 52:1779–1784

Harrison EP, Willingham NM, Lloyd JC, Raines CA (1997) Reducedsedoheptulose-1,7-bisphosphatase levels in transgenic tobacco lead todecreased photosynthetic capacity and altered carbohydrate accumula-tion. Planta 204: 27–36

Holland SC, Artier J, Miller NT, Cano M, Yu JP, Ghirardi ML, BurnapaRL (2016) Impacts of genetically engineered alterations in carbon sinkpathways on photosynthetic performance. Algal Res 20: 87–99

Iijima H, Shirai T, Okamoto M, Kondo A, Hirai MY, Osanai T (2015)Changes in primary metabolism under light and dark conditions in re-sponse to overproduction of a response regulator RpaA in the unicel-lular cyanobacterium Synechocystis sp. PCC 6803. Front Microbiol 6: 888

Kanno M, Carroll AL, Atsumi S (2017) Global metabolic rewiring for im-proved CO2 fixation and chemical production in cyanobacteria. NatCommun 8: 14724

Klughammer C, Schreiber U (1994) An improved method, using saturatinglight pulses, for the determination of photosystem I quantum yield viaP7001-absorbance changes at 830 nm. Planta 192: 261–268

Knowles VL, Plaxton WC (2003) From genome to enzyme: Analysis of keyglycolytic and oxidative pentose-phosphate pathway enzymes in thecyanobacterium Synechocystis sp. PCC 6803. Plant Cell Physiol 44:758–763

Kramer DM, Evans JR (2011) The importance of energy balance in im-proving photosynthetic productivity. Plant Physiol 155: 70–78

Lehmann M, Wöber G (1976) Accumulation, mobilization and turn-over ofglycogen in the blue-green bacterium Anacystis nidulans. Arch Microbiol111: 93–97

Li J, Weraduwage SM, Preiser AL, Tietz S, Weise SE, Strand DD,Froehlich JE, Kramer DM, Hu J, Sharkey TD (2019) A cytosolic bypassand G6P shunt in plants lacking peroxisomal hydroxypyruvate reduc-tase. Plant Physiol 180: 783–792

Lindahl M, Spetea C, Hundal T, Oppenheim AB, Adam Z, Andersson B(2000) The thylakoid FtsH protease plays a role in the light-inducedturnover of the photosystem II D1 protein. Plant Cell 12: 419–431

Markson JS, Piechura JR, Puszynska AM, O’Shea EK (2013) Circadiancontrol of global gene expression by the cyanobacterial master regulatorRpaA. Cell 155: 1396–1408

516 Plant Physiol. Vol. 182, 2020

Shinde et al.

www.plantphysiol.orgon June 3, 2020 - Published by Downloaded from Copyright © 2020 American Society of Plant Biologists. All rights reserved.

Masamoto K, Hisatomi S, Sakurai I, Gombos Z, Wada H (2004) Re-quirement of carotene isomerization for the assembly of photosystem IIin Synechocystis sp. PCC 6803. Plant Cell Physiol 45: 1325–1329

Mullineaux CW (2014) Co-existence of photosynthetic and respiratory ac-tivities in cyanobacterial thylakoid membranes. Biochim Biophys Acta1837: 503–511

Preiser AL, Banerjee A, Fisher N, Sharkey TD (2018) Supply and con-sumption of glucose 6-phosphate in the chloroplast stroma. bioRxiv442434

Preiss J (1984) Bacterial glycogen synthesis and its regulation. Annu RevMicrobiol 38: 419–458

Puszynska AM, O’Shea EK (2017) Switching of metabolic programs inresponse to light availability is an essential function of the cyanobacte-rial circadian output pathway. eLife 6: e23210

Raines CA, Lloyd JC, Dyer TA (1999) New insights into the structure andfunction of sedoheptulose-1,7-bisphosphatase: An important but ne-glected Calvin cycle enzyme. J Exp Bot 50: 1–8

Scanlan DJ, Sundaram S, Newman J, Mann NH, Carr NG (1995) Char-acterization of a zwf mutant of Synechococcus sp. strain PCC 7942.J Bacteriol 177: 2550–2553

Sharkey TD, Weise SE (2016) The glucose 6-phosphate shunt around theCalvin-Benson cycle. J Exp Bot 67: 4067–4077

Shimakawa G, Hasunuma T, Kondo A, Matsuda M, Makino A, Miyake C(2014) Respiration accumulates Calvin cycle intermediates for the rapidstart of photosynthesis in Synechocystis sp. PCC 6803. Biosci BiotechnolBiochem 78: 1997–2007

Silva P, Thompson E, Bailey S, Kruse O, Mullineaux CW, Robinson C,Mann NH, Nixon PJ (2003) FtsH is involved in the early stages of repairof photosystem II in Synechocystis sp PCC 6803. Plant Cell 15: 2152–2164

Smith AJ (1983) Modes of cyanobacterial carbon metabolism. Ann Micro-biol (Paris) 134B: 93–113

Suzuki E, Ohkawa H, Moriya K, Matsubara T, Nagaike Y, Iwasaki I,Fujiwara S, Tsuzuki M, Nakamura Y (2010) Carbohydrate metabolismin mutants of the cyanobacterium Synechococcus elongatus PCC 7942defective in glycogen synthesis. Appl Environ Microbiol 76: 3153–3159

Tamoi M, Miyazaki T, Fukamizo T, Shigeoka S (2005) The Calvin cycle incyanobacteria is regulated by CP12 via the NAD(H)/NADP(H) ratiounder light/dark conditions. Plant J 42: 504–513

Tóth TN, Chukhutsina V, Domonkos I, Knoppová J, Komenda J, Kis M,Lénárt Z, Garab G, Kovács L, Gombos Z, et al (2015) Carotenoids areessential for the assembly of cyanobacterial photosynthetic complexes.Biochim Biophys Acta 1847: 1153–1165

Udvardy J, Borbely G, Juhåsz A, Farkas GL (1984) Thioredoxins and theredox modulation of glucose-6-phosphate dehydrogenase in Anabaenasp. strain PCC 7120 vegetative cells and heterocysts. J Bacteriol 157:681–683

Vizcaíno JA, Csordas A, del-Toro N, Dianes JA, Griss J, Lavidas I, MayerG, Perez-Riverol Y, Reisinger F, Ternent T, et al (2016) 2016 update ofthe PRIDE database and its related tools. Nucleic Acids Res 44:D447–D456

Wang X, Liu W, Xin C, Zheng Y, Cheng Y, Sun S, Li R, Zhu XG, Dai SY,Rentzepis PM, et al (2016) Enhanced limonene production in cyano-bacteria reveals photosynthesis limitations. Proc Natl Acad Sci USA 113:14225–14230

Weisz DA, Liu H, Zhang H, Thangapandian S, Tajkhorshid E, Gross ML,Pakrasi HB (2017) Mass spectrometry-based cross-linking study showsthat the Psb28 protein binds to cytochrome b559 in photosystem II. ProcNatl Acad Sci USA 114: 2224–2229

Welkie DG, Rubin BE, Chang YG, Diamond S, Rifkin SA, LiWang A,Golden SS (2018) Genome-wide fitness assessment during diurnalgrowth reveals an expanded role of the cyanobacterial circadian clockprotein KaiA. Proc Natl Acad Sci USA 115: E7174–E7183

Welkie DG, Rubin BE, Diamond S, Hood RD, Savage DF, Golden SS(2019) A hard day’s night: Cyanobacteria in diel cycles. Trends Micro-biol 27: 231–242

Yu L, Zhao J, Muhlenhoff U, Bryant DA, Golbeck JH (1993) PsaE is re-quired for in vivo cyclic electron flow around photosystem I in the cy-anobacterium Synechococcus sp. PCC 7002. Plant Physiol 103: 171–180

Zybailov B, Mosley AL, Sardiu ME, Coleman MK, Florens L, WashburnMP (2006) Statistical analysis of membrane proteome expressionchanges in Saccharomyces cerevisiae. J Proteome Res 5: 2339–2347

Plant Physiol. Vol. 182, 2020 517

Glycogen Metabolism Kick-Starts Photosynthesis

www.plantphysiol.orgon June 3, 2020 - Published by Downloaded from Copyright © 2020 American Society of Plant Biologists. All rights reserved.