Chloroplasts Require Glutathione Reductase to Balance Reactive Oxygen Species and Maintain Efficient Photosynthesis

Thiol-based redox-regulation is vital to coordinate chloroplast functions depending on illumination. Yet, how the redox-cascades of the thioredoxin and glutathione redox machineries integrate metabolic regulation and reactive oxygen species (ROS) detoxification remains largely unresolved. We investigate if maintaining a highly reducing stromal glutathione redox potential (EGSH) via glutathione reductase (GR) is necessary for functional photosynthesis and plant growth. Since absence of the plastid/mitochondrial GR is embryo-lethal in Arabidopsis thaliana, we used the model moss Physcomitrella patens to create knock-out lines. We dissect the role of GR in chloroplasts by in vivo monitoring stromal EGSH dynamics, and reveal changes in protein abundances by metabolic labelling. Whereas stromal EGSH is highly reducing in wildtype and clearly responsive to light, the absence of GR leads to a partial oxidation, which is not rescued by light. Photosynthetic performance and plant growth are decreased with increasing light intensities, while ascorbate and zeaxanthin levels are elevated. An adjustment of chloroplast proteostasis is pinpointed by the induction of plastid protein repair and degradation machineries. Our results indicate that the plastid thioredoxin and glutathione redox systems operate largely independently. They reveal a critical role of GR in maintaining efficient photosynthesis.


Summary
(1) Thiol-based redox-regulation is vital to coordinate chloroplast functions depending on illumination. Yet, how the redox-cascades of the thioredoxin and glutathione redox machineries integrate metabolic regulation and reactive oxygen species (ROS) detoxification remains largely unresolved. We investigate if maintaining a highly reducing stromal glutathione redox potential (E GSH ) via glutathione reductase (GR) is necessary for functional photosynthesis and plant growth.
(2) Since absence of the plastid/mitochondrial GR is embryo-lethal in Arabidopsis thaliana, we used the model moss Physcomitrella patens to create knock-out lines.
We dissect the role of GR in chloroplasts by in vivo monitoring stromal E GSH dynamics, and reveal changes in protein abundances by metabolic labelling.
(3) Whereas stromal E GSH is highly reducing in wildtype and clearly responsive to light, the absence of GR leads to a partial oxidation, which is not rescued by light.
Photosynthetic performance and plant growth are decreased with increasing light intensities, while ascorbate and zeaxanthin levels are elevated. An adjustment of chloroplast proteostasis is pinpointed by the induction of plastid protein repair and degradation machineries.

Introduction
In photosynthetic eukaryotes, changes in environmental conditions, such as light intensity or temperature, provoke changes in electron flow both in the chloroplasts and in the mitochondria. Several mechanisms serve to rapidly modulate or redirect electron flow to minimise over-reduction of the two electron transport chains (ETCs), which can otherwise give rise to excessive formation of reactive oxygen species (ROS) (Schwarzländer & Finkemeier, 2013;Schöttler & Toth, 2014). Moreover, ROS serve as important signalling molecules in stress acclimation (Suzuki et al., 2012;Dietz et al., 2016). This implies that the rates of ROS generation and scavenging must be precisely balanced in these organelles. The maintenance of cellular redox pools for metabolism, antioxidant defence, and thiol-based redox switching requires the constant influx of electrons via light-and NADPH-powered redox cascades, involving the oxidation and reduction of cysteines in thioredoxins (Trx) and glutathione (Meyer et al., 2012;Yoshida & Hisabori, 2016;Geigenberger et al., 2017;Gütle et al., 2017).
Reduced glutathione (GSH) is present in cells at low millimolar concentrations (Meyer et al., 2001). Glutathione functions include ascorbate regeneration via the ascorbateglutathione-cycle and detoxification of potentially toxic organic electrophils and heavy metals, as well as acting as a cofactor of monothiol glutaredoxins (Grx) for coordination of iron-sulfur clusters (Foyer & Noctor, 2011;Moseler et al., 2015). In addition, glutathione is also used as a substrate of dithiol Grx-catalysed protein (de)glutathionylation (Meyer et al., 2012;Zaffagnini et al., 2019). For the latter functions it is essential that glutathione can reversibly switch between its reduced form GSH and the oxidised form glutathione disulfide (GSSG) which involves the transfer of 2 electrons. The glutathione redox potential (E GSH ) is dependent on the GSH concentration, as well as on the balance between GSH and GSSG. The E GSH can vary drastically between subcellular compartments (Meyer, 2008;Kojer et al., 2012). In unstressed plant cells, the E GSH of cytosol, peroxisomes, mitochondrial matrix and plastid stroma is highly reducing between -310 to -360 mV (Meyer et al., 2007;Schwarzländer et al., 2008). In these compartments, GSSG is efficiently regenerated to GSH by the action of glutathione reductase (GR) using NADPH as electron donor. In the cytosol and the mitochondria of Arabidopsis thaliana the loss of GR is partially compensated for by the presence of the NADPH-dependent Trx reductases A and B (NTRA,B) (Marty et al., 2009). Nevertheless, decreased GR activity of the plastid/mitochondria-targeted isoform leads to reduced root growth in seedlings (Yu et al., 2013). However, a complete loss of GR in plastids causes embryo-lethality (Marty et al., 2009;L. Marty & A.J. Meyer, unpublished). dynamics of plastid E GSH by redox-sensitive GFP (roGFP)-based in vivo imaging and compare protein abundances between wildtype (WT) and GR mutants in response to a shift from low light to high light by quantitative proteomics. phenotypic and pigment analyses, P. patens was grown on KNOP ME agar plates (12 g l -1 purified agar, Oxoid) at the indicated light intensity.

Plant materials and growth conditions
For measurements of photosynthetic parameters and preparation of proteins samples for MS/MS, P. patens protonema tissue was propagated under axenic conditions either on 9 cm or 4.5 cm Petri dishes overlaid with a cellophane disk on solidified PpNO 3 medium(0.8 % (w/v) agar), or in glass flasks with PpNO 3 medium (Gerotto et al., 2016) in a growth chamber under low light conditions (LL, 15 µmol photons m -2 s -1 ) at 25°C with a 16 h:8 h light:dark photoperiod. For control and high light assays, 10-day-old protonema plates were moved from LL to 50 µmol photons m -2 s -1 (CL) and 450 µmol photons m -2 s -1 (HL) respectively, maintaining temperature and photocycle.  (Schwarzländer et al., 2008;Speiser et al., 2018) stably integrated at the PTA2 locus under the control of the PpActin5 promoter (Kubo et al., 2013;Mueller & Reski, 2015). In regenerated plants that survived hygromycin selection, integration of the construct into the target locus was verified using primer pairs spanning the 5' integration site (5P_F and H3b_R, Fig. S1b) and the 3' integration site (NosT_F and 3P_R) ( Table S1). Absence of PpGR1 transcript for independent knock-out lines was confirmed using the primer pair PpGR1_RT_F and PpGR1_RT_R in a reverse transcription PCR (Fig. S1c,

Transmission electron microscopy
Transmission electron microscopy (TEM) was performed as described in Schuessele et al. (2016).

NBT staining
Gametophores were stained in a 0.1 mg/ml nitro blue tetrazolium (NBT, Duchefa) solution in 75 mM potassium phosphate buffer (pH 7.0) for 1.5 h in the dark or in the light (120 µmol photons m -2 s -1 ). Chlorophyll was subsequently removed by incubation in 80 % ethanol at 70°C (Lee et al., 2002).

Ascorbate assay
Total and reduced ascorbate in P. patens samples was quantified according to Gillespie & Ainsworth (2007) with the modification that 2-2'-bipyridyl was dissolved in 95 % ethanol. Five to 100 mg material was flash frozen in liquid nitrogen, homogenised in a bead mill (TissueLyser II, Qiagen, 30 Hz for 2x 1.5 min) and processed immediately.

Pigment and glutathione analysis by HPLC
Photosynthetic pigments were quantified by high-performance liquid chromatography (HPLC) according to Thayer & Björkman (1990), see also Supplemental Information.
Glutathione was extracted from c.

Shotgun Quantification
Peptides were pre-concentrated and desalted for 3 min on a trap column (Acclaim PepMap 100, 300 µM x 5 mm, 5 µm particle size, 100 Å pore size, Thermo Fisher Scientific) using 2 % (v/v) acetonitrile/0.05 % (v/v) trifluoroacetic acid in ultrapure water at a flow rate of 10 µl/min. Gradient separation of peptides was performed on a reversed phase column (C18, Acclaim Pepmap, 75 µm x 50 cm, 2 µm particle size, 100 Å pore size, Thermo Fisher Scientific) at a flow rate of 300 nl/min using the respectively. The intensity threshold for MS2 was set to 1x10 4 . Maximum fill times were 50 ms (MS1) and 55 ms (MS2). Unassigned charge states, charged state 1 and ions with charge state 5 and higher were rejected.

Bioinformatic analyses
LC-MS/MS data was processed with Proteome Discoverer (PD, version 2.2, Thermo Fisher Scientific). Raw files were searched using the SequestHT algorithm against a P. patens protein database based on the V1.6 gene models (Zimmer et al., 2013) supplemented with common contaminant proteins (cRAP, www.thegpm.org/crap/) with the following settings: Precursor and fragment mass tolerances 10 ppm and 0.02 Da, respectively; minimum peptide length: 6; maximum of missed cleavages: 2; variable modifications: Oxidation of methionine, N-acetylation of protein N-termini.
For the identification of 15 N-labelled peptides, a second database search was performed with 14 N to 15 N substitution(s) set as static modifications for all amino acids. Peptide-spectrum-matches (PSMs) were filtered using the Percolator node to satisfy a false discovery rate of 0.01 (based on q-values). Identifications were filtered to achieve a peptide and protein level FDR of 0.01. LC-MS/MS runs were chromatographically aligned with a maximum retention time drift of 10 min. Precursor ion quantification was performed using unique and razor peptides. Abundances were normalized to the maximum total peptide abundance in all files. Protein ratios (Δgr1 vs. WT) were calculated using the 'pairwise ratio based' approach with subsequent hypothesis testing (background based t-test) for the calculation of p-values.

NPQ measurements
In

Spectroscopic measurements of photosynthetic parameters
Protonema from LL, CL and HL treated plates was measured with cellophane in buffer (Hepes 20 mM pH 7.5, KCl 10 mM). LEF+CEF (linear plus cyclic electron flow) and CEF (cyclic electron flow) were measured by following the relaxation kinetics of the carotenoid electrochromic band shift at 520 nm (corrected by subtracting the band shift at 546 nm) in the absence or presence of 10 μM DCMU and hydroxylamine, respectively. CEF and LEF+CEF were calculated as e −1 s −1 PSI −1 upon normalization to the PSI amount. The electrochromic shift signal upon excitation with a single saturating turnover flash (5 ns laser pulse) in the presence or absence of 10 μM DCMU and 1 mM hydroxylamine were used to estimate the PSI and PSI+PSII amount. DCMU and hydroxylamine in this measurement were used to fully block PSII photochemistry to facilitate the determination of the PSI amount (Terashima et al., 2012;Gerotto et al., 2016).
To measure the proton motive force (pmf), which consists of ∆pH (trans-thylakoid proton gradient) and membrane potential (∆ψ), 5-day-old protonema tissue from liquid cultures was harvested, dark-adapted for 15 min before analysis and exposed for 1:30 h to 300 µmol photons m -2 s -1 to obtain the steady-state of the ECS signal.
Afterwards, a 1 min dark-phase was recorded (5 min of illumination in between measurements) to obtain at least 3 technical replicates at 520 nm and 546 nm respectively. g H + , which reflects the proton conductivity of the ATP synthase, was estimated by fitting the first 300 ms of the decay curve with a first-order exponential decay kinetic, ∆pH and ∆ψ were calculated as described previously (Wang et al., 2015). photosynthetically active vegetative cells, circumventing embryogenesis and nonphotosynthetic tissue (e.g. Schween et al. (2005)). We hypothesized that null mutants of organellar GR might be viable in plants that maintain green plastids throughout their life cycle. We therefore generated knock-out constructs replacing exons two to five containing the translation start site and the active site in the gene encoding the previously identified dual-targeted mitochondria and plastid-localised glutathione reductase Pp1s13_127V6.1 (named GR1 in P. patens, Xu et al., (2013)) with a hygromycin resistance cassette via homologous recombination (Fig. S1a). As genetic background, we utilised a newly generated P. patens line expressing the plastid-targeted E GSH biosensor Grx1-roGFP2 under the control of the P. patens Actin 5 promoter (Weise et al., 2006;Mueller & Reski, 2015 (Fig. 1b). Thus, GR1 is not necessary for embryo development in P. patens.

Moss lacking PpGR1 is viable and displays reduced growth
However, spore germination of Δgr1 #48 and Δgr1 #88 was delayed by several days, with spores being able to germinate eventually (Fig. 1b). Apart from the dwarfed appearance, the mutant lines grew fewer caulonema filaments and rhizoids than the WT (Fig. S2).
The ultrastructure of chloroplasts lacking GR1 was investigated using transmission electron microscopy (TEM), which revealed normally packed grana stacks and stroma lamellae, undistinguishable from WT (Fig. S3).

Plastid redox state is dynamic in WT, shifted to less reducing values in Δgr1 plants and not rescued via Trx reduction under light
Taking advantage of the stromal-targeted Grx1-roGFP2, the steady state of the chloroplast E GSH was determined by confocal in vivo imaging of roGFP2 redox state Δgr1 #88 (Fig. 2a,b), indicating that the stromal E GSH is less reducing in Δgr1 lines. To test for stability of stromal E GSH after exogenous reduction, plants were incubated with 2 mM dithiothreitol (DTT) and then exposed to continuous laser scanning under the confocal microscope. As a result, plastid Grx1-roGFP2 was rapidly re-oxidised in the Δgr1 plants, but not in the WT background (Fig. 2c). This result prompted us to investigate the dynamics of E GSH in dark-to-light and light-to-dark transitions (Fig. 2d).
Dark-adapted plants were exposed to a dark-to-light transition ( Fig. S4b). These experiments revealed that Δgr1 #48 and Δgr1 #88 were not sensitive to incubation in darkness, suggesting that the FTR-dependent redox cascades are not required for the survival of the P. patens mutants.

Δgr1 plants show altered responses of reactive oxygen species dynamics and photosynthetic function
As Δgr1 knock-out plants were not sensitive to dark-incubation (Fig. S4) and showed an oxidative response to the laser light used for microscopic imaging (Fig. 2c), their growth habit under different light intensities was tested (Fig. S5a). Mutants did not profit from increasing light fluencies (30 to 130 µmol photons m -2 s -1 ), as their fresh weight did not increase, in contrast to WT (Fig. S5b). In order to investigate the impact of different light intensities on ROS scavenging, the levels of ascorbate and dehydroascorbate were determined. The measurements revealed a higher total level of ascorbate in Δgr1 #48 and Δgr1 #88 that reached 500 % to 700 % of WT levels in different light intensities (Fig. 3a). Interestingly, dehydroascorbate levels were increased as well, but were present at the same ratio to reduced ascorbate as in WT (on average 8.1+/-2.5 % of total ascorbate, Fig. 3b). We incubated gametophores with nitro blue tetrazolium (NBT) in the dark or in light as a means to detect superoxide (O 2 •-), and found increased staining intensity in light-incubated Δgr1 plants (Fig. 3c) Under exposure to higher light fluencies (HL, 450 µmol photons m -2 s -1 ), white sectors appeared in leaflets of Δgr1 plants (Fig. 4a). When HL was additionally combined with elevated temperature Δgr1 #48 and Δgr1 #88 plants died, whereas WT and Δgr1 mutants were able to recover from temperature stress only (Fig. 4a). An examination of photosynthetic parameters revealed a light intensity-dependent decrease of non-photochemical quenching (NPQ), linear and cyclic photosynthetic electron flow (LEF+CEF) and the photosystem I to photosystem II ratio (PSI/PSII) ( Fig. 4b). While the NPQ amplitude decreased with increasing light intensity in Δgr1, the NPQ dark relaxation was slower under all light conditions tested. CEF was not increased under different light intensities, but showed a slight increase after 6 h induction via anoxia (Fig. S6a,b). The Fv/Fm ratio decreased in Δgr1 with increasing light, indicating decreased photosynthetic efficiency (Fig. S6a). At the same time, total chlorophyll (a+b) levels were decreased (Fig. S6c).
To study the proton motive force (pmf) in mutant chloroplasts in comparison to WT, electrochromic shift assays (Kramer & Crofts, 1989) were conducted and revealed a decreased pH gradient and increased membrane potential ΔΨ in Δgr1, resulting in a similar pmf. The proton conductivity of the ATP synthase was not significantly altered ( Fig. 4c).
As NPQ relaxation was slower under all light conditions tested, similar to Arabidopsis npq2 (zeaxanthin epoxidase) mutants (Niyogi et al., 1998), we measured levels of photosynthesis pigments and found that zeaxanthin levels were increased up to 2-fold relative to total pigment content in the Δgr1 mutant in comparison to WT (Fig.   4d). Total glutathione levels (GSH+GSSG) were not significantly different to WT (Fig.   4e).

Quantitative proteomics reveals light intensity-dependent protein level changes in Δgr1 plants
To assess the significance of GR function for proteostasis under changing light intensities we used metabolic labelling with the stable isotope 15 N in combination with quantitative proteomics. Changes in protein abundances in WT and Δgr1 plants upon a shift from low light (LL) to high light (HL) were investigated using quantitative proteomics (Fig. S7). The sensor expressing line cpGrx1roGFP2 #40 (WT) and Δgr1#48 were labelled in vivo by growth on medium containing Ca (  were down-regulated and 59 up-regulated (four were either up-or downregulated, depending on the light conditions). The overlap between differentially regulated proteins in LL in comparison to HL was 8.6 % for the down-regulated proteins and 20.3 % for the up-regulated proteins (Fig. 5b). Following a manual annotation of subcellular localisation (Table S2) Arabidopsis homologs, the largest fraction of regulated proteins was attributed to plastids (38), followed by the cytosol (20) and proteins with unclear localisation (13) (Fig. 5c). In addition, differentially regulated proteins were sorted into 29 functional categories (Table S2) and categories containing more than two proteins plotted to visualise category-specific down-or up-regulation in the different light conditions (Fig.   5d). Here, more proteins with unknown function were down-regulated specifically in LL or HL whereas several proteins with unknown function were up-regulated in both light conditions. In the categories "protein homeostasis" and "photosynthesis light reactions" more proteins were down-regulated in LL, but more proteins were up- As PpGR1 (Pp1s13_127V6.1) was also identified in the proteomics analysis as differentially abundant (Table S2), we confirmed the absence of PpGR1 additionally by a targeted proteomics approach (Fig. S8).
Further, we screened the dataset for known redox-regulated proteins and found one of the three isoforms of gamma subunit of chloroplast ATP synthase

Stromal E GSH responds to photosynthetic status
We generated viable null mutants of P. patens GR1, and found that absence of GR1 leads to a shift in the stromal E GSH . After Grx1-roGFP2 calibration, sensor 405/488 nm excitation ratio measurements can be translated into degree of sensor oxidation.
As the redox potential of roGFP2 equilibrates with the redox potential of glutathione, E GSH can be calculated, with the limitation that compartment pH has to be estimated (Meyer et al., 2007;Schwarzländer et al., 2008). In Δgr1 plants, the degree of oxidation of the plastid-targeted E GSH sensor Grx1-roGFP2 was severely shifted.
Based on the change of the 405/488 nm ratio and the in vivo sensor calibration (Fig.   2b), we calculated a shift from c. 48 % oxidation in the WT background to c. 92 % oxidation in the Δgr1 lines. This would correspond to a 33 mV shift in the redox potential (calculated for pH8: -311 mV in WT vs. -278 mV in Δgr1). In comparison, the redox potential in Arabidopsis epidermal plastids was determined as c. -361 mV at pH 8 (Schwarzländer et al., 2008). A shift of 30 mV would mean an increase of the relative amount of GSSG from 0.01 % to 0.1 %, calculated for a total concentration of 2.5 mM GSH (Meyer et al., 2007).
As in the same compartments glutathione-and Trx-dependent thiol switching fuelled by distinct reductases co-exists (Buchanan & Balmer, 2005), the cross-talk of these systems has been dissected for the cytosol and the mitochondria. Thus, cytosolic Trx redox-state can be rescued via the glutathione system (Reichheld et al., 2007). Vice  et al., 2018). However, we did not find increased glutathione levels in the absence of GR1 (Fig. 4e). In dynamic measurements in ectopically reduced Δgr1 plants, the shifted stromal E GSH was rapidly re-established by exposure to laser light, suggesting that GSSG rapidly accumulates in mutants upon illumination. As the regeneration of GSSG via GR is lacking in the mutant plastids, this likely represents GSSG formed by the ascorbate-GSH cycle, i.e. dehydroascorbate reductase (DHAR) activity. While the relative contributions of monodehydroascorbate reductase (MDHAR) and DHAR to the plastid ascorbate regeneration were debated (Asada, 1999;Polle, 2001), plastid-targeted AtDHAR3 was shown to contribute to ascorbate recycling with mutants being sensitive to high light (Noshi et al., 2016). The increase in total ascorbate and dehydroascorbate levels in Δgr1 mutants indicates impaired function of the ascorbate-GSH cycle consistent with a substantial contribution of plastid GR, including non-stress conditions. Light was not necessary for the survival of Δgr1 plants, indicating that light-dependent reduction of Trxs was not a prerequisite for viability. Further, cross-talk between the Trx-and glutathione redox cascades in plastids is limited, which is in contrast to previous findings in the cytosol and mitochondria (Reichheld et al., 2007;Marty et al., 2009).
Notably, in WT plants exposed to a successive dark/light and light/dark transition, the stromal E GSH responded dynamically, showing that E GSH is rapidly light-responsive in the presence of GR. In addition, after the transition from light to darkness, we observed a rapid rise of the stromal E GSH pointing to oxidative processes in consequence of a light/dark transition. This oxidation is analogous to Trx oxidation that is required to deactivate redox-regulated Calvin Benson cycle enzymes (Wolosiuk & Buchanan, 1977;Yoshida et al., 2018). Peroxiredoxins have been reported to act as possible electron sinks (Pérez-Ruiz et al., 2017;Vaseghi et al., 2018).
At the measured stromal E GSH in Δgr1, still most of the stromal glutathione is present in the reduced state (99.9 %). Nevertheless, a shifted E GSH is likely to affect downstream redox-cascades, as well as GSH-dependent enzymatic reactions. This includes glutaredoxins (Grx) and dehydroascorbate reductase (DHAR). While the involvement of plastid Grxs in iron-sulfur cluster coordination and protein (de)glutathionylation has been shown (Zaffagnini et al., 2012;Moseler et al., 2015;Rey et al., 2017;Zannini et al., 2019), only very few target proteins are currently known. Thus, the future challenge is to identify specific target cysteines affected by a shifting E GSH in order to appraise its potential physiological role in redox regulation and signalling.

Consequences of a lack in plastid/mitochondrial GR
The absence of plastid/mitochondrial GR resulted in a pronounced dwarfism as well as in light sensitivity of the mutant plants. As dual targeting of one GR isoform is evolutionarily conserved (Xu et al., 2013), the contribution of the lack of mitochondrial GR to the overall phenotype of the mutants is not resolved yet. However, in Arabidopsis it is the lack of GR in plastids, and not in mitochondria, that results in embryo-lethality (L. Marty & A.J. Meyer, unpublished;Nietzel et al., 2019). We found that in P. patens, the plastid/mitochondrial isoform of GR is not necessary for embryo development. This suggests that either (1) the process that causes embryo-lethality in Arabidopsis is not important for P. patens embryo development, or (2) that the lack of GR is compensated for in P. patens allowing embryo development to proceed. For instance, in contrast to flowering plant gr mutants under stress (Ding et al., 2012), P. patens Δgr1 mutants showed a high increase in total ascorbate levels.
In the green haploid moss gametophyte, lack of GR1 caused slow growth as well as   et al., 2012).
In Δgr1  can increase PSII photodamage (Davis et al., 2016). In addition, we found increased sensitivity to ROS, as well as increased superoxide tissue staining in the Δgr1 Only a low percentage of quantified protein differed in protein abundance between against toxic electrophiles via acidification (Ferguson, 1999) and is negatively regulated by GSH, but activated by glutathione conjugates (Roosild et al., 2010) (DTT)).
(c) WT and Δgr1 mutant were pre-treated with 2 mM DTT to achieve Grx1-roGFP2 reduction and then exposed to continuous laser scanning in water (1 scan/timepoint; WT background n=4, Δgr1 n=9).
(d) In vivo measurement of light-dependent chloroplast E GSH dynamics. Dark-adapted WT (cpGrx1roGFP2 #40) and Δgr1 plants were exposed to a dark/light/dark transition (100 µmol photons m -2 s -1 for 10 min). Incubation with the electron transport inhibitor 10 µM DCMU blocked the observed E GSH dynamics; n=3. plot depicts mean +/-SD as boxes, individual data points, and data point density.

Fig. 5: Quantification of protein abundances by metabolic labelling.
(a) Protein abundances of Δgr1/WT in low light (LL) and after a shift to high light (HL) for 1 h.
(c) Comparison of manually annotated subcellular localisations (Table S2) of proteins with differential abundance.
(d) Cleveland dot plot showing functional categories (Table S2)  The following Supporting Information is available for this article: Fig. S1 Identification of PpGR1 knock-out mutants.