An Arabidopsis soluble chloroplast proteomic analysis reveals the participation of the Executer pathway in response to increased light conditions

Highlight Executer proteins have emerged as participants in singlet oxygen-mediated retrograde signalling in plants. This study relates the function of Executer in chloroplasts with the acclimation response to increased light.


Introduction
During evolution, plants have developed an intricate network of signalling pathways to trigger physiological responses as a consequence of diverse environmental stimuli. Recent observations have demonstrated that retrograde communication coordinates the expression of nuclear genes with the metabolic and developmental state of the cell through signals emitted from plastids and mitochondria (Woodson and Chory, 2008;Ng et al., 2014). Particularly, chloroplasts─photosynthetic organelles in plant cells─are the source of specific signalling molecules that relay information to the nucleus (Nott et al., 2006;Pfannschmidt, 2010;Pogson and Albrecht, 2011;Barajas-López et al., 2013;Chi et al., 2013). The plastid signals identified so far can be linked to specific stress conditions, and the best characterized signals are intermediates of the tetrapyrrole biosynthesis pathway, the redox state of the thylakoid membrane, and reactive oxygen species (ROS) (Karpinski et al., 1999;Strand et al., 2003;Gadjev et al., 2006;Koussevitzky et al., 2007;Foyer and Noctor, 2009;Galvez-Valdivieso et al., 2009;Pesaresi et al., 2009;Sun et al., 2011;Woodson et al., 2011;Zhang et al., 2011;Pfalz et al., 2012;Petrillo et al., 2014). Also, the metabolic state of chloroplasts can be sensed by exported metabolites such as carbohydrates, reductive power in the form of NADPH, isoprenoid intermediates, xantophyll derivatives or a phosphonucleotide (Rolland et al., 2006;Pfannschmidt, 2010;Zhang et al., 2010;Estavillo et al., 2011;Ramel et al., 2012;Xiao et al., 2012).
In chloroplasts, ROS are produced as an unavoidable side effect of the photosynthetic light reactions, as a consequence of the spatial and temporal concurrence of electron and energy transfer reactions with photosynthetically generated molecular oxygen (Asada, 2006;Galvez-Valdivieso and Mullineaux, 2010). Whereas electron transfer reactions to molecular oxygen in its ground triplet state lead to the formation of superoxide, hydrogen peroxide and hydroxyl radicals, energy transfer reactions result in the formation of the highly reactive singlet oxygen. ROS generation in chloroplasts is enhanced during environmental stress, and particularly singlet oxygen is the major ROS produced in leaves exposed to high light (Fryer et al., 2002;Krieger-Liszkay et al., 2008). Genetic approaches have shown that the production of singlet oxygen under controlled conditions in the chloroplast plays a significant role in signalling the oxidative response of plants (Apel, 2004). Subsequent work in C. reinhardtii showed that earlier exposure of the cells to low levels of singlet oxygen elicits an acclimation response that protects cells from photooxidative damage (Ledford et al., 2007).
However, the mechanism which transduces the signal from the chloroplast to the nucleus is far from delineated and, in higher plants, appears to undergo significant crosstalk with other signalling pathways controlling such responses as plastid differentiation, plant development and general stress responses (Laloi et al., 2007;Baruah et al., 2009). With the aim of studying the biological activity of singlet oxygen and the mechanisms involved in transduction of redox signalling, the conditional flu mutant of Arabidopsis thaliana was identified (Meskauskiene et al., 2001). The flu mutant accumulates free protochlorophyllide (Pchlide), the immediate precursor of chlorophyllide, in the dark; upon illumination free Pchlide acts as a potent photosensitizer that generates singlet oxygen by transferring light energy to molecular oxygen (Gollnick, 2007). Two stress responses in flu were triggered by the release of singlet oxygen within the plastid compartment during reillumination of dark-adapted plants: seedling lethality, and cell death and growth inhibition in mature plants (op den Camp et al., 2003). Global gene-expression studies showed that these stress responses were not primarily due to physicochemical damage caused by singlet oxygen during oxidative stress but were attributed to the activation of genetic stress response programs (op den Camp et al., 2003;Gadjev et al., 2006;Laloi et al., 2006;Przybyla et al., 2008). An extensive second-site mutant screen performed in flu identified a group of suppressor mutants named executer (ex). Plants with mutations in Executer1 (Ex1), a plastid protein of unknown function, lost the ability to perceive the presence of singlet oxygen in flu chloroplasts and, consequently, the activation of singlet-oxygen mediated response programs was suppressed (Wagner et al., 2004). Subsequently, a protein homolog to Ex1, named Executer2 (Ex2), was also identified and suggested to be involved in the singlet-oxygen-responsive gene network (Lee et al., 2007).
How the Executer proteins are involved and function in the signalling pathway remains elusive. The loss of function of either of the Executer protein showed no obvious phenotype compared to WT Arabidopsis (Lee et al., 2007;Wagner et al., 2004;Kim and Apel, 2013a). It was found however that ex1 plants were more resistant than WT to damage upon treatment with low concentrations of 3-(3, 4-dichlorphenyl)-1,1-dimethylurea (DCMU) together with high light intensities (Wagner et al., 2004). Also, analysis of the response of WT and ex1 plants to β-cyclocitral (a β-carotene oxidation product) treatments suggested that ex1 plants were more resistant to photooxidative stress than WT (Ramel et al., 2012). Studies on the hypersensitive response to pathogen infection pointed out that ex2 plants are slightly more resistant to low amounts of pathogens than WT (Mur et al., 2010). Furthermore, a double mutant ex1/ex2 was affected in chloroplast development in cotyledons (Kim et al., 2009), and the seedlings of the double mutant ex1/ex2 were less susceptible than WT when exposed to a combined low-temperature/highlight treatment (Kim et al., 2012).
The role of Executer proteins in singlet-oxygen mediated signalling is unclear. Because light availability is one of the key factors that modulates acclimation strategies and defence reactions in plants, we aim to analyse how plants adapt to their environment by studying the chloroplast proteome response to a perturbation in light intensity, which promotes ROS production but would not result in oxidative stress or cell death. Here, a differential-expression proteomics approach was used to analyse the impact of light on chloroplast protein abundance in two T-DNA insertional knockout lines (EX1 and EX2). Our study showed changes in abundance of several photosynthesis-and carbon metabolism-related proteins as well as proteins involved in plastid mRNA processing, among others. A good correlation between executer mutants and the changes occurring after exposure of WT plants at a moderate light intensity in the time frame of hours was inferred. It is suggested that Executer proteins participate in signalling in Arabidopsis under growth light conditions, and in the regulation of the response to environmental cues such as light acclimation, likely to avoid the misexpression of defence programs.

Light treatment
Plants were grown on soil in a growth chamber at 8 h-light/16 h-darkness (20 ºC) for 7-8 weeks under a photon flux density (PFD) of 120 μmol m -2 s -1 and relative humidity of 70 %. For the high light (HL) experiments, plants were transferred 1 h after the onset of the light period to a growth chamber under PFD of 600-700 μmol m -2 s -1 . As control material one set of plants was maintained at 120 μmol quanta m -2 s -1 , representing normal light (NL) conditions.

Protein extraction
After 6 h light treatment (HL or NL), plant leaves were harvested for chloroplast isolation and purification according to Hall et al. (2011). Briefly, 20 g of plant material was homogenized using a blender in ice-cold extraction buffer (20 mM Tricine-NaOH pH 8.4, 300 mM sorbitol, 10 mM KCl, 10 mM Na-EDTA, 0.25 % BSA, 4.5 mM sodium ascorbate and 5 mM L-cysteine). Cell debris was removed by a nylon mesh (22 μm), and chloroplasts were pelleted by centrifugation for 2 min at 1000 ×g. Chloroplasts were washed and ruptured by osmotic shock in 10 mM Na-pyrophosphate-NaOH pH 7.8 buffer. Following centrifugation at 100 000 ×g for 1 h at 4 ºC, the supernatant containing the soluble stromal proteins was concentrated using an Amicon Ultra-15 10 K ultrafiltration device. Protein concentration was determined using the Bradford assay (Bradford, 1976) and bovine serum albumin as reference. Seppro® Rubisco Spin Columns (Sigma) were used to reduce Rubisco abundance in the stroma samples.

Two-dimensional differential gel electrophoresis (2D-DIGE)
Chloroplast stroma samples were precipitated with ice cold acetone. Protein pellets were solubilized in DIGE labelling buffer (30 mM Tris-HCl pH 8.5, 2 M thiourea, 7 M urea, 2 % (w/v) CHAPS). Remaining insoluble material was removed by centrifugation for 10 min at 21 000 ×g. The final protein concentration for G-Dye labelling was 5 μg/μl. Solubilized protein samples were separately labelled with G-Dye100, G-Dye200 and G-Dye300 dyes (DyeAGNOSTICS) at a ratio of 400 pmol dye/50 µg protein extract for 30 min in darkness on ice. Labelling was quenched by addition of lysine. In general, samples were labelled using G-Dye200 and G-Dye300 dyes while an internal standard (consisting of a pooled sample comprising an equal amount of all samples in the experiment) was labelled with G-Dye100. Details of labelling and the subsequent combination of differentially labelled samples used are shown in the experimental design presented in Table 1. For each immobilized pH gradient (IPG) strip, equal amounts of G-Dye100, G-Dye200 and G-Dye300 labelled samples were combined, typically 50 µg protein per sample. Prior to iso-electric focusing the mixed samples were diluted with rehydration solution containing 2 M thiourea, 7 M urea, 2% (w/v) CHAPS, 20 mM DTT, 0.002% (w/v) bromophenol blue and 0.5% (v/v) IPG buffer pH 3-11 NL (GE Healthcare, Uppsala, Sweden). Samples were thereafter applied to 24 cm Immobiline Dry Strips pH 3-11 NL (GE Healthcare, Uppsala, Sweden) by passive rehydration for 2 h followed by active rehydration for 10 h at 30 V. Isoelectric focusing (IEF) was performed on an IPGphor II (GE Healthcare, Uppsala, Sweden). Prior to second dimension SDS-PAGE, strips were equilibrated first for 15 min in 75 mM Tris-HCl pH 8.8, 6 M urea, 2% (w/v) SDS, 30% (v/v) glycerol, 0.002% (w/v) bromophenol blue and 1% (w/v) DTT and secondly for 15 min in 75 mM Tris-HCl pH 8.8, 6 M urea, 2% (w/v) SDS, 30 % (v/v) glycerol, 0.002% (w/v) bromophenol blue and 2.5% (w/v) iodoacetamide. Strips were applied on top of 12% SDS-polyacrylamide gels and sealed in place using 0.5% (w/v) agarose. Second dimension separation was performed using an Ettan Daltsix electrophoresis unit (GE Healthcare). Preparative samples for spot picking and subsequent mass spectrometry analysis, corresponding to 500 µg of protein per IPG strip (comprising an equal mixture of all samples in the experiment), were precipitated in acetone. Proteins were solubilised in 2 M thiourea, 7 M urea, 2% CHAPS, 20 mM DTT, 0.002% (w/v) bromophenol blue and 0.5% (v/v) IPG buffer pH 3-11 NL (GE Healthcare, Uppsala, Sweden) before being separated by 2D-gel electrophoresis as described for DIGE gels above. Preparative gels were fixed in 30% ethanol, 10% acetic acid and were stained by Coomasie Brilliant Blue (CBB).
Gel image analysis G-Dye labelled samples were visualized using a Typhoon™ 9400 Variable Mode Imager (GE Healthcare, Uppsala, Sweden). All gel images were scanned at 100 µm resolution using a photomultiplier tube (PMT) voltage optimal for maximal pixel intensity without spot saturation. Prior to image analysis the gel images were cropped using ImageQuant™ v.5.2 (GE Healthcare, Uppsala, Sweden) in order to remove extraneous areas. DIGE analysis was performed using Redfin 3 software (Ludesi) as was matching of preparative CBB stained gels to DIGE gels. Images of CBB-stained gels were acquired using an image scanner and the Labscan software (GE Healthcare, Uppsala, Sweden). Spot detection, matching and statistical analysis was performed using the Redfin 3 program (www.ludesi.com). A principal component analysis (PCA) was performed to separate the gel samples according to their expression variation. One-way analysis of variance (ANOVA; p<0.001) and MannWhitney (p<0.05) tests were conducted to assess differential expression of protein abundance between the different groups. Minimum protein volume was set at 200 and differentially expressed proteins with a change in average spot volume of at least 2.0-fold were selected.

In-gel digestion and protein identification
Spots of interest were excised from preparative gels using an Ettan Spotpicker™ spot picking station fitted with a 1.4 mm picker head. Gel plugs were dehydrated and destained by incubation with a solution containing 20 mM ammonium hydrogen carbonate in 35% acetonitrile. The solution was removed and gel pieces were dried by addition and removal of neat acetonitrile twice. Dried gel plugs were rehydrated on ice with 20 mM ammonium hydrogen carbonate and 10% acetonitrile containing 2 ng/µl trypsin (Promega).
Pooled standard WTHL ex1HL 7 Pooled standard ex1HL WTHL 8 Pooled standard ex1HL WTHL 9 Pooled standard ex2NL ex2HL 10 Pooled standard ex2NL ex2HL 11 Pooled standard ex2HL ex2NL 12 Pooled standard ex2HL ex2NL In-gel digestion was performed overnight at 37 °C. Mass spectrometry analysis of in-gel digests was performed on a MALDI-TOF Voyager-DE™ STR Bio Spectrometry Workstation (Applied Biosystems). Database searches were performed on a local Mascot server licensed to Umeå University by Matrixscience (www.matrixscience.com), using Arabidopsis TAIR9 and Swiss-Prot databases. For searches a peptide mass error tolerance of 50 ppm was accepted and carbamidomethylation of cysteine and oxidation of methionine were specified as variable modifications.

Chlorophyll fluorescence measurements
In vivo chlorophyll a fluorescence was measured using a Dual-PAM-100 chlorophyll fluorescence photosynthesis analyser (Heinz Walz) on attached rosette leaves. After dark acclimation of the plants (15 min), the measuring light (9 µmol photons m -2 s -1 ) was turned on, and minimal fluorescence (F o ) was determined. Leaves were exposed to a pulse of saturating light (3000 µmol photons m -2 s -1 , 0.8 s) to determine the maximum fluorescence in the dark-adapted state (F m ). Subsequently, leaves were illuminated with actinic red light at 660 µmol of photons m -2 s -1 determining the steady-state level of fluorescence in the light (F s ) and subjected to saturating pulses (0.8 s) of 3000 µmol photons m -2 s -1 for 20 min to measure the maximal fluorescence in the light-adapted state (F m '). The maximum PSII efficiency was expressed as (Genty et al., 1989). The NPQ coefficient was calculated using the Stern-Volmer equation ) (Bilger and Björkman, 1990).

Results
The aim of the present study was to examine the collective response of the soluble chloroplast proteome of 8-week-old Arabidopsis plants following a transition from normal light (NL) to moderate high light (HL, 5-fold increase in PFD) in the time scale of hours. Furthermore, using Arabidopsis mutant plants, the link between the Executer pathway and the acclimation response upon exposure of plants to high light was investigated.
Photosynthetic performance in ex1 and ex2 single mutants is comparable to WT The role of Executer in chloroplast light response was analysed using two independent T-DNA insertion lines inactivating EXECUTER1 (EX1) and EXECUTER2 (EX2), respectively. Absence of Executer transcripts was confirmed in the T-DNA insertion lines indicating them to represent true knockout lines ( Supplementary Fig. S1). As previously reported (Wagner et al., 2004;Lee et al., 2007), ex1 and ex2 plants showed no obvious alterations of growth and development. Because Executer proteins have been associated with PSII (Kim and Apel, 2013b), PSII performance was assessed by measurement of chlorophyll-a fluorescence at room temperature in intact leaves from WT and executer mutant plants treated at the two different PFD (NL and HL) for 6 h. No differences could be detected between WT and mutant plants, as deduced from the ratio of variable to maximum fluorescence (F v /F m ), the quantum yield of PSII (Φ PSII ), the degree of non-photochemical quenching (NPQ) and the excitation pressure of PSII (1-qP) ( Table 2). HL treatment resulted in a 5% decline of F v /F m for both WT and mutant plants.
2D-DIGE analysis of the soluble chloroplast proteome highlights molecular differences between executer mutants and WT plants To examine whether chloroplasts are affected by the loss of Executer, the effect of light was investigated on the soluble chloroplast proteome of Arabidopsis. First, plants were grown at 120 μmol m -2 s -1 . After 8 weeks, plants were exposed to high light intensity (700 μmol m -2 s -1 ) during 6 h and chloroplasts were isolated from rosette leaves. The soluble protein extract from chloroplasts was analysed by 2D-DIGE, following a strategy as depicted in Table 1. The DIGE analysis revealed significant differences between WT and mutant plants, both under normal light (WTNL, ex1NL and ex2NL, respectively) and in response to high light (WTHL, ex1HL and ex2HL, respectively). The differences in the stroma proteome between WT, ex1 and ex2 are supported by PCA (Fig. 1). Based on the first principal component (PC1)-that represents the direction of highest variability after gel data dimensionality reduction-there is a large variation between WT in normal conditions versus high light treatment. Differences between executer mutants and WT under normal conditions are also evident (Fig. 1).

Proteome remodelling as a consequence of executer loss of function resembles light treatment in WT
Comparative analysis of WT and ex1 and ex2 Arabidopsis plants in the two light conditions (NL and HL) showed a number of soluble chloroplast proteins that significantly altered their expression level (Supplementary Table S1; foldchange >2, p<0.001). The majority of the observed changes in the proteome were upregulations, irrespectively of genotype or treatments, with the largest changes detected in ex2 plants.
In NL conditions, changes in 54 and 94 protein spots were detected in the ex1NL versus WTNL and ex2NL versus WTNL comparative groups, respectively (Supplementary  Table S1). When the responsive spots of the two mutant plants in NL versus WTNL were compared, 41 spots (37+4) were exclusively detected in ex2NL plants, as shown in a Venn diagram analysis in Fig. 2A. Upon exposure of WT Arabidopsis plants to high light (WTHL), 28 protein spots (5+1+18+4) displayed significant expression changes in response to the treatment (Supplementary Table S1). Actually, 23 light-responsive spots, which represent more than 82%, were detected in the ex1NL vs. WTNL and/or ex2NL vs. WTNL groups (1+18+4 in Fig. 2A). When ex1 plants were exposed to high light for 6 h (ex1HL), 27 protein spots (23+4 in Fig. 2B) were found to be differentially expressed relative to ex1NL. Interestingly, the behaviour of 23 spots (not present in ex1NL) was shared between ex1HL and ex2NL groups (Fig. 2B). Few differences were detected between ex2NL versus WTNL and ex2HL versus WTNL groups, with more than 91% common protein spots (Fig. 2C). In general, it was observed that a higher number of unique spots were significantly altered in ex2 plants compared to WT, and the changes were already significant in normal light conditions (37 and 18 spots in Figs 2A and 2B, respectively). Table 2. PSII performance of intact leaves from NL-or HL-treated WT and executer mutant plants was determined by chlorophyll-a fluorescence measurements using a PAM-fluorometer. The maximum quantum yield of PS II (F v /F m ), the effective quantum yield of PSII (Φ PSII ), the degree of non-photochemical quenching (NPQ) and the excitation pressure of PSII (1-qP) was deduced. Data represents mean ± standard error. No significant differences were found between plants in the same light condition (ANOVA).

Identification and analysis of differentially expressed proteins
Based on quantitative image analysis, 107 spots were detected to be significantly altered in their accumulation among genotypes and treatments and were selected as highly responsive proteins. Ninety differentially expressed protein spots were successfully excised from colloidal Coomassie stained gels (a representative gel is shown in Fig. 3) and analysed by mass spectrometry. Gel spots corresponding to protein mixtures or albumin were discarded. Several proteins were presented as multiple spots that likely represent different isoforms or proteins differentially modified in the protein extracts or degradation products. Finally, a total of 47 spots corresponding to 29 proteins were unambiguously identified (Table 3, Fig. 3 and Supplementary Table S2). The identified proteins fell into several functional groups: photosynthesis and carbon fixation, energy generation, RNA metabolism, protein folding, and other metabolic processes, such as nitrogen and tetrapyrrole metabolism (Fig. 4). There were also some proteins of unknown function.

Photosynthetic electron transport chain
Ferredoxin-NADP reductase isoform 2 (FNR-2), which mediates the transferring of electrons from ferredoxin to NADP, and two proteins of the oxygen evolving complex (OEC) from photosystem II-PsbO isoforms (PsbO1 and PsbO-2) and PsbP-were more abundant in the chloroplast soluble fraction of executer mutant plants compared to wild type Arabidopsis. Except for OEC proteins, a similar trend was observed in WTHL. Two peripheral ATP synthase subunits (alpha and beta) were significantly affected in executer mutants: an increase abundance of the alpha subunit (ATP-A) in the stromal compartment was detected, but a multiexpression pattern (spots 44, 62 and 69) was observed for the beta subunit (ATP-B) and therefore it could not be classified simply into the up-or down-regulated group. On the other hand, the stromal ATP synthase delta subunit (ATP-C) accumulated in a light-dependent manner.

Chloroplast protein synthesis and homeostasis
Alpha and beta subunits of Cpn60-a molecular chaperone that participates in protein folding in chloroplasts-increased in executer mutants plants and, to a lesser extent, in WTHL as compared with WTNL. Interestingly, a group of proteins Principal components 1 (PC1) and 2 (PC2)-where PC1 indicates the direction of the highest variability followed by PC2 with diminishing variability orthogonal to PC1-accounted for 51% and 11% of the study variance, respectively. PCA suggested a modification of the soluble chloroplast proteome in Arabidopsis upon light treatment. The genotype caused a large shift in the PC1 dimension to more positive values and significantly reduced the differences between NL and HL samples in PC2.
related to mRNA metabolism-the chloroplast stem-loop binding proteins of 41 kDa (CSP41A, CSP41B), and plastidspecific ribosomal protein 2 (PSRP-2)-were significantly upregulated in mutant plants. The accumulation of two chloroplast ribonucleoproteins, namely CP29B' and CP31A, was affected both by genotype and light treatment.

Other
The abundance of two chloroplast proteins of unknown function (Table 3) varied in a genotype-dependent manner. Spots 122 and 183 corresponded to a protein containing a tetratricopeptide repeat (TPR) region based on protein sequence analysis (Ishikawa et al., 2005). Spot 200 was identified as a thylakoid lumen protein of 19 kDa (TL19) (Schubert et al., 2002).

Discussion
Plants are exposed to a variety of environmental changes, such as light availability, that can compromise their metabolism, growth and development. Therefore, plants have developed a variety of mechanisms for sensing environmental fluctuations and, accordingly, adjust their developmental programs, metabolic processes and defence reactions. Upon changing light conditions, plants exhibit adaptation and acclimation strategies in order to optimize their photosynthetic performance and to avoid imbalance between energy absorption and utilization, which could promote ROS production in the chloroplasts (Apel, 2004;Kangasjärvi et al., 2009;Karpinski et al., 2013). The comparison of a fivefold increase in PFD over growth intensity, not enough for   Table 3. light-saturated photosynthesis, was chosen to challenge the acclimation response of WT and executer mutant Arabidopsis plants to irradiance. Chlorophyll fluorescence measurements performed under different conditions showed that the maximum quantum efficiency of PSII photochemistry remained unchanged between WT and mutants, with a moderate decrease in plants under moderate light intensity. Similar effects were detected in the light-dependent thermal dissipation component of NPQ and photochemical efficiency. It was concluded that the photochemical activity of executer plants was not affected relative to WT plants. Moreover, there was no significant loss of PSII efficiency during the chosen exposure time to increased light. Therefore, our experimental conditions would not promote photo-oxidative stress, but likely rapid adjustments in photosynthesis and chloroplast composition for acclimation to a change in the light environment. A double purpose is envisaged, that is, oxidative stress avoidance and molecular adjustments that would facilitate plants utilizing the additional light to improve their photosynthetic performance. The acclimation response would involve changes in the relative abundance of a number of proteins (Kosová et al., 2011). Therefore, a subcellular fractionation approach in combination with 2D-DIGE was followed to analyse the dynamics of the soluble chloroplast proteome of Arabidopsis to changing light. It has been shown that Executer proteins are necessary to transmit the signal produced by singlet oxygen from the plastid to the nucleus. Yet little is known about the role that Executers play under normal circumstances. In order to investigate the putative contribution of the Executer pathway to the plant response, a comparative proteome analysis of soluble chloroplast extracts of the two executer mutants was carried out. Our results showed very similar molecular phenotypes between ex1 and ex2 under growth light conditions, though the change was more significant in the latter. Interestingly, a clear separation among the WT and executer mutant groups was observed Table 3. Summary of differentially expressed proteins (light treatment and/or genotype effect versus WTNL) identified by MS. The functional classification (GROUP) and gene accession number (GENE ID) are shown. The fold change in protein abundance is indicated in colour code; values >1 or <1 indicate an increase or decrease in protein abundance, respectively. A colour version of this figure is available at JXB online when plants were challenged with increased light. As expected, light treatment significantly affected the soluble chloroplast proteome in WT, but the effect was less pronounced in mutant plants, that otherwise show more similarities to WTHL. The consistent phenotype observed in our proteomic experiments for the two executer mutant plants indicates that Ex1 and Ex2 might participate in the same signalling pathway. Moreover, as previously proposed from genetic studies, their functions might not be redundant as the presence of one protein cannot compensate the absence of the other.
Examination of differentially accumulated proteins revealed that the proteins that underwent differential expression upon light treatment in WT Arabidopsis were related to metabolic pathways (mainly carbon metabolism), protein synthesis and energy production (Fig. 4). We detected an increase in abundance of FNR-2, which has a critical role in the redistribution of photosynthetically derived electrons to various reducing pathways, such as carbon fixation, nitrogen metabolism and chlorophyll biosynthesis (Lintala et al., 2009). Also, the abundance of three peripheral thylakoid ATP synthase subunits was responsive to light; it has been shown that ATP synthase in chloroplasts is regulated by light and metabolite factors, and particularly ROS showed a direct influence in its activity (Buchert and Forreiter, 2010;Buchert et al., 2012;Kohzuma et al., 2013). Our results showed the alteration of a number of CB enzymes that catalyse readily reversible reactions and are not susceptible to 'fine' regulation, such as aldolase, transketolase, epimerase and isomerase (Raines, 2003;Michelet et al., 2013). In our experiments, a different expression pattern for two aldolase isoforms (FBA1 and FBA2) was detected, which could indicate some functional specialization. The importance of aldolase and transketolase activities in photosynthetic carbon flux control and for the acclimation of photosynthesis to changing environmental conditions has been reported (Haake et al., 1998(Haake et al., , 1999Henkes et al., 2001;Raines, 2003;Uematsu et al., Fig. 4. Schematic presentation of chloroplast proteins affected in executer mutants and WT plants after light treatment compared to WT plants under normal or growth light conditions, as shown in Table 3. Arrows indicate increased or decreased abundance compared to WTNL. An asterisk indicates that the protein is not differentially expressed in WTHL; 1 protein is not differentially expressed in ex1NL, 2 protein is not differentially expressed in ex2NL; 3 multi-expression pattern. Abbreviations: ALA, 5-aminolevulinic acid; ATPα, ATPβ and ATPγ, ATPase alpha, beta and gamma subunits, respectively; CA1, carbonic anhydrase 1; CP29B', chloroplast ribonucleoprotein 29 kDa; CP31A, 31 kDA RNA-binding protein; CPN60α and CPN60β, chaperonin 60 alpha and beta subunits, respectively; CSP41A and CSP41B, chloroplast stem-loop binding protein of 41 kDa A and B subunits, respectively; Cytb 6 f, cytochrome b 6 f complex; FBA1 and FBA2, ructose-bisphosphate aldolase subunits 1 and 2, respectively; Fd, ferredoxin; FNR2, ferredoxin-NADP(+)oxidoreductase 2; GAPB, glyceraldehyde 3-phosphate dehydrogenase B subunit; Glu, glutamate; Gln, glutamine; GlyK, D-glycerate 3-kinase; GS2, glutamine synthetase-2; GSA2, glutamate-1-semialdehyde 2 1-aminomutase); PC, plastocyanin; PGK, phosphoglycerate kinase; PQH 2 , plastoquinol; PRK, phosphoribulokinase; PsbO and PsbP, photosystem II subunits O and P, respectively; PSI, photosystem I; PSII, photosystem II; PSRP-2, plastidspecific ribosomal protein 2; RCA, rubisco activase; RPE, ribulose-phosphate-3-epimerase; RPI, ribose 5-phosphate isomerase; TKL, transketolase; TL19, thylakoid lumen 19 kDa; 2-OG, 2-oxoglutarate. 2012). Furthermore, their substrates and products act as precursors of associated metabolic processes, such as amino acid and fatty acid synthesis, and therefore variations in the abundance would affect other chloroplast pathways (Henkes et al., 2001;Tetlow et al., 2005). On the other hand, epimerase and isomerase decreased significantly in response to light. Our study showed that phosphoribulokinase (PRK), which plays an important role in regulating the flow of sugar through the Calvin cycle, was upregulated by light. PRK can become limiting when plants grown under low irradiance are exposed to high light (Paul et al., 2000). Two enzymes with a defined role in carbon fixation modulation, rubisco activase (Portis, 2003) and beta-carbonic anhydrase 1 (CA1) (Fett and Coleman, 1994), also increased upon light treatment. Other metabolic proteins such as glutamine synthetase-2-a central enzyme in nitrogen metabolism with a role in maintaining the balance of carbon and nitrogen (Miflin and Habash, 2002)-and a GSA aminotransferase-that participates in tetrapyrrole metabolism (Tanaka et al., 2011)-were induced. Increased level of proteins related to protein folding and RNA is in accordance with increased rate of protein synthesis in the light, when the translational machinery of the chloroplast is most active (Marín-Navarro et al., 2007). All these changes are likely associated with an increase in the efficiency of photosynthesis under non-saturating light conditions that modify the photosynthetic capacity of the plant with an impact in carbon fixation, chlorophyll synthesis, nitrogen metabolism and other processes, and would result in new ATP requirements.
Expression of most of the proteins discussed above also changed in executer mutants under normal growth conditions. Further, some spots exhibited abundance variations only linked to the Arabidopsis genotype and were not regulated by light. Two CB enzymes (GapB and phosphoglycerate kinase) significantly increased in the mutant plants, particularly in ex2 plants, in parallel to the alterations of CB enzymes detected upon light treatment. A subset of proteins of OEC accumulated in the soluble fraction of executer chloroplasts compared to WT (Ettinger and Theg, 1991;Bricker et al., 2012). Two ribonucleases from the CSP41 family (Qi et al., 2012) and a plastid-specific ribosomal protein (PSRP-2) (Yamaguchi and Subramanian, 2003) were more abundant in executer plants, which might indicate a modified transcriptional and/or translational activity compared to WT under normal growth light conditions. Two chloroplast proteins of unknown function-a protein with a predicted TPR motif and a thylakoid lumen protein that belongs to PsbPsuperfamily-showed higher expression levels in mutant plants.
Our results revealed significant differences between WT and mutant plants, but an interesting overlap was found between WTHL and executer plants. Although Executer proteins are dispensable for normal growth, our work has detected a molecular perturbation at the basal level in executer mutant plants. It seems that the loss of function of Executer results in a reorientation in chloroplast central metabolism that resembles the activation response of moderate light acclimation. A plausible interpretation of these data is that the Executer proteins set the light intensity threshold that triggers the high light response. Therefore, in the executer mutants, this threshold is lower than in wild type plants. It is thus proposed that Executer form part of a regulatory network for the coordination between environmental stimuli and metabolic adapattion and determine the acclimation response in chloroplasts, although the exact role of the Executer proteins is yet to be defined. Interestingly, our results showed that the absence of Ex2 has a stronger impact in chloroplasts than Ex1. Our findings are consistent with other studies that observed an Executer-dependent stress acclimation in green leaves of mature plants, such as suppression of cell death in ex1 plants treated with DCMU (Wagner et al., 2004), the resistance to photoxidative stress of ex1 plants (Ramel et al., 2012), or the slight increase to pathogen resistance in ex2 plants (Mur et al., 2010).
In conclusion, despite the demonstration that Executer proteins are necessary to transmit the signal produced by singlet oxygen from the plastid to the nucleus, little is known about the role that Executers play under normal circumstances. In an effort to elucidate the biological activity of Executer proteins as putative mediators of the singlet oxygen response in chloroplasts, we used proteomics to analyse the role of Executer in chloroplasts of Arabidopsis exposed to different light regimes. The analysis of the soluble chloroplast protein profile in WT has provided a deeper insight into changes associated with the acclimation response to light. Six hours of high light exposure triggered responses in chloroplast of WT Arabidopsis; not surprisingly, many proteins involved in photosynthesis and carbon metabolism were affected, which reflects an increase in the efficiency of photosynthesis. Under normal light growth conditions, our experiments detected significant changes in the soluble chloroplast proteome as a consequence of the loss of function of Executers that, interestingly, resembled the acclimation response of the plant to increased light. Our results suggest that Executer proteins form part of the signalling network for the perception of environmental perturbation in plants, and might participate in the basal repression of defence responses in chloroplasts under normal irradiance.

Supplementary material
Supplementary material is available at JXB online. Supplementary Fig. S1. Verification of two A. thaliana T-DNA lines homozygous for an insertion within EX1 and EX2, respectively.
Supplementary Table S1. Number of spots differentially expressed (up or down-regulated) in response to genotype (WT, ex1 and ex2 plants) and light treatment (NL and HL).
Supplementary Table S2. List of differentially expressed proteins identified by mass spectrometry in the SwissProt/ TAIR9 databases.