Critical Role of COI1-Dependent Jasmonate Pathway in AAL toxin induced PCD in Tomato Revealed by Comparative Proteomics

Alternaria alternata f.sp. Lycopersici (AAL) toxin induces programmed cell death (PCD) in susceptible tomato (Solanum lycopersicum) leaves. Jasmonate (JA) promotes AAL toxin induced PCD in a COI1 (coronatine insensitive 1, JA receptor)-dependent manner by enhancement of reactive oxygen species (ROS) production. To further elucidate the underlying mechanisms of this process, we performed a comparative proteomic analysis using tomato jasmonic acid insensitive1 ( jai1), the receptor mutant of JA, and its wild type (WT) after AAL toxin treatment with or without JA treatment. A total of 10367 proteins were identified in tomato leaves using isobaric tags for relative and absolute quantitation (iTRAQ) quantitative proteomics approach. 2670 proteins were determined to be differentially expressed in response to AAL toxin and JA. Comparison between AAL toxin treated jai1 and its WT revealed the COI1-dependent JA pathway regulated proteins, including pathways related to redox response, ceramide synthesis, JA, ethylene (ET), salicylic acid (SA) and abscisic acid (ABA) signaling. Autophagy, PCD and DNA damage related proteins were also identified. Our data suggest that COI1-dependent JA pathway enhances AAL toxin induced PCD through regulating the redox status of the leaves, other phytohormone pathways and/or important PCD components.


COI1-dependent jasmonate pathway promotes TA induced cell death. The visible necrotic lesions
were observed at 36 h in leaves of WT plants after TA treatment (data not shown), and the lesions became typical at 48 h. The PCD symptom was enhanced dramatically in WT leaves when treated with TA and JA together in comparison with TA treatment alone (Fig. 1A), suggesting that JA can promote TA induced PCD. Neverthless, TA treated jai1 leaves displayed minor tissue damage at 48 h, and exogenous JA did not exert effect on it, indicating that COI1 is involved in the PCD process triggered by TA and the impaired JA perception in jai1 inhibited JA promoted PCD.
Trypan blue staining is an accepted method for cell viability assay, live cells or tissues with intact cell membranes are not coloured, and dead cells can be colored in light blue 19 . As can be seen in Fig. 1B, TA treated WT leaves accumulate more blue precipitate compared with the control, and JA enhanced the accumulation of blue precipitate, indicating that JA can promote TA induced PCD. Besides, TA treated jai1 leaves accumulated less blue precipitate compared with TA treated WT leaves, which was consistent with the visible phenotype (Fig. 1A), further proved that JA pathway is important for TA induced PCD.
Malondialdehyde (MDA) is the final product of membrane lipid peroxidation, which can be a marker for oxidative stress 20 . MDA content in TA treated WT leaves increased with the increase in JA concentration, and peaked at 100 μM JA. However, no visible change was found in jai1 leaves with the increase in JA concentration (Fig. 1C), indicating that JA enhanced membrane lipid peroxidation in a COI1-dependent way.
JA enhances ROS accumulation during TA induced PCD. Hydrogen peroxide (H 2 O 2 ) and superoxide (O 2 .− ) are the two key ROS molecules. We carried out histochemical 3,3′-diaminobenzidine (DAB) 21 and nitro blue tetrazolium (NBT) staining 22 to detect H 2 O 2 and O 2 .− in the leaves, respectively. As shown in Fig. 2A, after TA treatment for 48 h, large brown precipitation was observed to be around the leaf veins of WT plants by DAB staining, and the brown region was expanded by JA treatment. However, jai1 leaves showed little precipitation after TA treatment, suggesting less ROS production in jai1. The H 2 O 2 content in WT and jai1 leaves increased steadily during three days with TA treatment, whereas in jai1 it was significantly lower than WT leaves, except for 36 h after treatment (Fig. 2B). Similarly, much more blue precipitates in TA and JA treated WT leaves was observed than in jai1 leaves by NBT staining (Fig. 2C). O 2 .− production rate in WT and jai1 leaves both increased constantly in three days after TA treatment, but it was significantly lower in TA treated jai1 compared with WT leaves at the same time point (Fig. 2D). Based on these results, it is proposed that ROS burst occurs during TA induced PCD, and the impairment of COI1 suppresses the production of ROS, as well as the necrotic cell death. In Fig. 2B,D, H 2 O 2 content and O 2 .− production rate in the untreated leaves were also increased slightly, possibly caused by the minor stress brought by the treatment.
Proteome profiling of tomato leaves. To better understand the mechanisms in TA response of tomato leaves and how JA influences the process, a comparative proteomic analysis was performed in JA insensitive mutant jai1 and its WT after TA treatment with or without JA treatment. We employed high-resolution MS in combination with isobaric tags for relative and absolute quantitation (iTRAQ) proteomics approach to identify the proteome changes. In our workflow, there are four biological replicates for each treatment and one set of iTRAQ includes two biological replicates (Fig. 3A). We identified a total of 8501 tomato proteins with 1% global false discovery rate (FDR) (Supplementary Figure S1) and 10367 tomato proteins with 5% FDR (Fig. 3B), and the 10367 proteins were chosen for the subsequent analysis.
Functional classification of the 10367 proteins. The identified proteins were analyzed based on the assigned functions of the proteins using Uniprot, NCBI and phytozome databases. The biological process of the overall 10367 proteins were classified into 36 categories sorted by the Blast2Go level 4 (Fig. 3D), among them .− producing rate in WT and jai1 leaves after 0.2 μM TA treatment for 36, 48 and 72 h. Each data point represents the mean of three replicates. Error bars indicate standard deviation of three replicates, asterisks above the bars indicate statistically significant differences between different treatments, as determined by the Student t-tests (**P < 0.01, ***P < 0.001).

Functional annotation and KEGG pathway analysis of the 10367 proteins. The Kyoto
Encyclopedia of Genes and Genomes (KEGG) is a collection of pathways for understanding high-level functions and utilities of the biological system from genomic and molecular-level information 23 . We conducted KEGG pathway analysis of the identified proteins. The proteins were mapped to a total of 139 maps (pathways with more than 40 sequences were represented in Table 1, the others were represented in Supplementary Table S5). The proteins were annotated in various pathways relevant to the synthesis, metabolism or degradation of nucleotide (12%), amino acid (22%), carbohydrate (6%), lipid (14%), secondary metabolites (31%), energy (9%) and xenobiotics (6%). Especially, the pathways of the immune system included the T cell receptor signaling pathway and biotin metabolism pathway. In addition, phosphatidylinositol signaling system and glycosylphosphatidylinositol (GPI)-anchor biosynthesis were also identified in our study.   changes to tomato proteome. Interestingly, most DE proteins in WT after TA treatment or TA + JA treatment showed higher number of increasing patterns, while those in jai1 after TA treatment exhibited similar number of increasing and decreasing patterns compared with WT. The expression patterns of the differentially expressed proteins in each comparison groups were clustered and displayed as a heat map in Supplementary Figure S5. Among the increased proteins, 61 were shared by the three comparisons, whereas only one protein was found Table 2. Interesting and novel proteins changed by JA treatment in the process of TA induced PCD. a Accession numbers from GenBank database or Sol Genomics Network (SGN). b (WT + TA + JA/WT + TA) represented JA and TA treated WT vs TA treated WT. c Red color represents that the protein was up-regulated, green color represents that the protein was down-regulated. d t-test indicates significant difference in expression of these proteins between WT treated with JA and TA and WT treated with TA alone, with p-value threshold values (α = 0.05). to be decreased in all the comparisons (Fig. 3C), each comparison group also contained their own DE proteins. These proteins present a global view of WT and jai1 proteome level responses to TA and JA treatment. The expression patterns across each contrasts were classified into 24 categories, as was shown in Supplementary Table S1. The biological process classification of the 10367 proteins was used as the reference for the functional analysis of DE proteins. We have presented the comparison of functional classification between the 10367 overall detected and 2670 differentially expressed proteins in Supplementary Figure S3, and we found that the biological function processes of the 2670 DE proteins distributed similarly with the overall identified proteins.
JA-regulated proteins in TA induced PCD process. As shown by the previous physiological results that JA promoted PCD, we further investigated JA-regulated proteins in TA induced PCD process, by checking the DE proteins in WT + TA + JA vs WT + TA comparison. The DE proteins related to cell death, JA, ET, ABA, DNA repair and resistance were listed in Table 2. In animal systems, cysteinyl aspartate-specific proteinases (caspases), serine proteases and specific protease inhibitors play crucial roles during the regulation of PCD process 24 . We found that several proteases were increased by JA treatment, including a kunitz-like protease partial, a cysteine proteinase RD19a-like, an aspartic proteinase nepenthesin-1-like and a β-1, 3-glucanase. Previous study showed that some caspase-like activities are attributable to plant subtilisin-like proteases (SBTs), which are related to cell death 25 . We detected four SBT or SBT like proteins in tomato leaves which were increased by JA treatment. Furthermore, cysteine protease inhibitor (CPI) is known as one of the specific inhibitors of PCD in plant cells 26 . In our experiment, CPI was increased by 23 fold in the WT leaves after treatment with JA and TA together, while increased by two fold after singly TA treatment. These results indicated that TA treatment also induces the expression of some PCD supressors, which could be enhanced by JA treatment. In addition, two JA response proteins, αβ-hydrolases superfamily protein and biotin carboxylase chloroplastic-like, were found to be involved in TA induced PCD. An ET synthesis related protein, probable N-succinyldiaminopimelate aminotransferase-like, was up-regulated by TA + JA treatment compared with TA treatment alone. Four abscisic acid (ABA) response proteins (annexin D4-like, plastid-lipid-associated chloroplastic-like, stem-specific protein TSJT1-like, and PP2A 65 kda regulatory subunit αβ isoform-like) were also increased by JA treatment in response to TA, indicating the possible involvement of ET and ABA pathway in TA induced PCD.
Pathogenesis-related proteins (PRs) are general markers for basal defense response. Two PR10 proteins increased significantly after TA + JA treatment compared with TA treatment alone, while PR-sth2-like protein and PR1 were decreased by JA treatment. Besides, several late blight resistance proteins, a chitinase and a harpin binding protein were also increased by TA + JA treatment, implying that JA could promote the synthesis of resistance proteins. The DNA damage response (DDR) plays an important role against detrimental effects of stress 27 . Three DNA repair related proteins (DNA repair ATPase-related family protein, DNA damage-binding protein 1-like, and DNA mismatch repair protein), were increased by JA during TA induced PCD. COI1-dependent signaling pathways in response to TA. We were interested in COI1-dependent signaling pathways in response to TA, as previous results have shown that COI1 impairement inhibits TA induced PCD in tomato leaves. DE proteins between TA treated jai1 and WT were classified as COI1-dependent proteins in response to TA. The candidate proteins related to defense response and programmed cell death were listed in Table 3. Firstly, six reductase proteins (peroxidase 51-like, glutathione s-transferase T1-like, 2-cys peroxiredoxin, BAS1-chloroplastic-like, probable glutathione s-transferase, mannose-6-phosphate isomerase 1-like) were significantly higher in jai1, whereas an oxidase protein (monooxygenase 1) was decreased in jai1 compared to WT after TA treatment, suggesting that ROS may act downstream of COI1 to promote TA induced PCD process.
The levels of two ceramide synthesis related proteins, melibiase family protein and serine palmitoyltransferase, were higher in jai1 than in WT after TA treatment, indicating that COI1-dependent JA signaling might promote TA induced PCD by regulating the biosynthesis of ceramide. Proteins related to autophagy, which plays a role in maintaining the intracellular homeostasis, were also found among the list of DE proteins. Autophagy-related protein 11 (ATG 11) was decreased in jai1, whereas RIBONUCLEASE 2-LIKE as a negative regulator of autophagy, was increased in jai1 after TA treatment. The results indicated that autophagy might be promoted by COI1-dependent JA signaling, and related to inhibition of TA induced PCD. Kunitz Trypsin Inhibitor (KTI1) was previously proved to be an antagonist of cell death triggered by phytopathogens and Fumonisin B1 in Arabidopsis 28 . In the current survey, it was increased in jai1 by two folds. Metacaspase type II and the peptidase C14 caspase catalytic subunit P20, as potential PCD regulators, were increased slightly in jai1. This result was different from the former report that the type-II metacaspase from tomato (LeMCA1) was not increased during chemical induced PCD in suspension-cultured tomato cells 29 . As tomato genome contains at least two type-II metacaspases, we speculate that different metacaspases might have distinct functions during PCD process, and display different expression patterns. Most of the above results partly explained why PCD was lighter in jai1 than in WT.
Our previous studies have shown that JA, ET and SA are involved in the response of tomato leaves to AAL or AAL toxin 30 . In the current survey, we paid more attention to the proteins relevant to these three hormone pathways. The expression of two JA biosynthetic genes, 12-oxophytodienoate reductase 3 and allene oxide synthase were increased in jai1, while some of the JA response genes were decreased, due to the impaired perception of JA. ET biosynthesis and response were inhibited, whereas SA synthesis and response were enhanced in jai1 mutant, suggesting that COI1-dependent JA signaling acts synergistically with ET, and antagonistically with SA during response to TA. Several proteins in ABA response pathway, were decreased in jai1 after TA treatment (

Discussion
Previous studies have shown that AAL toxin can inhibit ceramide biosynthetic enzymes and lead to PCD in sensitive asc/asc tomato species, due to the reduced sphingolipids and accumulated dihydrosphingosine (DHS) and 3-ketodihydrosphingosine (3-KDHS) 31,32 . However, the underlying mechanism from toxin perception to PCD process is poorly understood in plants. Our research explored the potential pathways and regulators in AAL toxin induced PCD and the role of COI1-dependent JA signaling pathway in regulating the PCD process.
Proteases specifically the classical proteolytic enzymes called caspases, were reported to be participated in the regulation of animal PCD, implies that proteases may be involved in regulation of plant PCD 33 . Several reports also have proved the involvement of proteases in regulating plant PCD. Both protease activity and cell death were inhibited by soybean trypsin inhibitor, while exogenous application of another serine protease prematurely triggered cell death 34 . In tobacco, inhibition of the induced cysteine protease activity by ectopic expression of a cysteine protease inhibitor (CPI) gene, blocked the PCD triggered either by an avirulent pathogen or by ROS 26 . We identified several JA treatment induced proteases during AAL toxin induced PCD, including a kunitz-like protease, a cysteine proteinase RD19a-like, an aspartic proteinase nepenthesin-1-like. a β-1, 3-glucanase and four SBT or SBT like proteins (Table 2). Moreover, we also found that CPI was increased by two fold after toxin treatment alone, while it was increased by 23 fold after TA and JA treatment. The results suggest that TA treatment promoted the expression of some specific proteases to induce PCD, and JA treatment enhanced the response. However, two other caspase-like proteases, metacaspase type II and the peptidase C14 caspase catalytic subunit P20, were increased slightly in jai1 and decreased in WT after TA treatment, suggesting that these two proteases The DE proteins caused by JA treatment and COI1 mutation during AAL toxin induced PCD were used to construct the model, and the impairment of COI1 induced protein changes were presented with the contrary change patterns. The identified proteins were assigned to different organelles or groups according to their subcellular localization and molecular functions. Up-regulated proteins are highlighted red and down-regulated proteins are colored green. Arrows and bars represent positive and negative regulation respectively. Red solid lines indicate that the pathways are up-regulated or promoted by JA during TA response, green solid lines with bars or arrows indicate that the pathways are down-regulated by JA during TA response, black solid lines with arrows links the proteins within the same pathway. TA induce PCD by inhibiting the synthesis of ceramide and inducing the overproduction of ROS, COI1-dependent JA pathway may promote this PCD progress by influencing the ROS production and scavenging, other hormone signaling pathways or some possible PCD regulators such as caspase-like proteins, autophagy and DNA repair related proteins. AAO, L-ascorbic acid oxidase; ABI1, ABA insensitive1; ACO1, 1-aminocyclopropane-1-carboxylate oxidase homolog 1-like; AHA, plasma membrane H + -ATPase; AOS, allene oxide synthase; ATG11, autophagy related protein 11; CHS, chalcone synthase; CPI, cysteine protease inhibitor; Cyt, Cytochrome; DAP-AT, probable n-succinyldiaminopimelate aminotransferase-like; DDB, DNA damage binding proteins; DDR, DNA damage response; DRA, DNA repair ATPase-related family protein; ER, endoplasmic reticulum; GST, glutathione S-transferase; Golgi, golgi complex; KTI1, kunitz trypsin inhibitor; MMR, DNA mismatch repair; NAT, nucleobase-ascorbate transporter 6-like; Prx, Peroxiredoxin; RNase E, Ribonuclase 2-Like; SBTs, subtilisin-like proteases; TF, transcription factor. may function as negative regulators of plant PCD. Kunitz trypsin inhibitor gene (KTI1) was proved to play a regulatory role in PCD antagonizing pathogen and Fumonisin B1 induced cell death 28 . The KTI1 found in our study was increased in jai1 by two folds in comparison with WT after TA treatment (Table 3), implying that KTI1 is also a negative regulator of plant PCD induced by TA.
Mitochondrial quality control is important in maintaining proper cellular homeostasis, and selective mitochondrial degradation by autophagy (mitophagy) is suggested to play an important role in quality control 34 . Selective autophagy includes the cytoplasm to vacuole targeting (Cvt) pathway 35 and pexophagy 36 . To date, 31 autophagy-related genes have been identified, which function as the molecular machinery for autophagy. Among them, ATG 11 is essential for mitophagy, acting as an adaptor protein that is needed along with ATG 19 to recruit the Cvt complex to phagophore assembly site (PAS), where the sequestering cytosolic vesicles are generated 37 . In the present study, ATG 11 in tomato was was found to be decreased by COI1 impairment after AAL toxin treatment. Conversely, the autophagy negative regulator RIBONUCLEASE 2-LIKE was significantly increased by COI1 impairment during AAL toxin induced PCD, suggesting that autophagy in jai1 was suppressed, leading to reduced cell death.
In addition to the putative PCD regulators conserved throughout animal and plant, there also exist some plant-specific mediators of PCD. Various plant hormones are strong candidates, and supporting evidence began to accumulate 38 . Our previous studies showed that COI1-dependent JA pathway acts upstream of ET to promote TA-triggered PCD 11 , which was further verified by the proteomic evidence in this study. The modulator of AAL cell death 1 (MACD1), which is an AP2/ERF transcription factor acting downstream of ET signaling, has been reported to positively regulate AAL triggered cell death 29 . We found that an AP2/ERF transcription factor RAP2-7-like was decreased significantly in jai1 during response to TA, suggesting that COI1-mediated JA signaling promote TA induced PCD by enhancement of ET response via RAP2-7 like. The resistance of tomato plants to AAL and AAL toxin is enhanced by SA pathway 30 . The proteins related to SA biosynthesis and response were increased in jai1 mutant after TA treatment, indicating that COI1-dependent JA pathway interacts with SA pathway in an antagonistic way to enhance TA induced PCD.
ROS have emerged as important signals in activation of plant PCD. Studies on exogenous application of H 2 O 2 confirmed the role of H 2 O 2 as a cell death trigger 39 . Zhang 40 also proved that AAL toxin induced PCD is closely related to the production of ROS. In the present study, the decreased levels of H 2 O 2 and O 2 .− in jai1 is consistent with the reduced cell death in jai1 after AAL toxin treatment, suggesting that JA exert its effects on plant PCD through regulation of ROS accumulation. Inhibitors of ET biosynthesis or perception blocked H 2 O 2 production and cell death in tomato suspension cells 41 , therefore the decreased ROS production in jai1 might be owing to the inhibited ET pathway. In addition, JA also acted through regulation of ET pathway in many cases, for example, JA-promoted lycopene was correlated with JA-stimulated ET production 42 , and both ET biosynthesis and the signaling pathway are strongly decreased in jai1 leaves inoculated with AAL when compared with WT 30 . The Arabidopsis jasmonate-insensitive mutant jar1 shows enhanced cell death after exposure to O 3− , and wounding or treatment with JA has been shown to reduce O 3− induced cell death and ROS levels 16 , in contrast to our finding in TA induced PCD. Complex mechanisms exist among PCD processes induced by distinct factors, which may be species-specific. Besides JA and ET, ABA seems to act downstream of COI1-mediated JA signaling in promoting TA induced PCD. Virus induced gene silencing (VIGS) analyses proved that AAL toxin triggered cell death is dependent on the mitogen-activated protein kinase MEK2 in tobacco 43 . We observed that the mitogen-activated protein kinase 9-like (MAPK9-Like) was significantly decreased in jai1 after TA treatment, suggesting a possible role of MAPK9-Like in response to TA in tomato leaves.
The CCR4-NOT transcription complex has been well known as mRNA deadenylases in eukaryotic cells. Liang et al. 44 proved that the homologs of CCR4-associated factor 1 (CAF1) in Arabidopsis are involved in defence responses to pathogen infections. In our results, two subunits of CCR4-NOT transcription complex were upregulated in jai1 mutant compared with the wild type (Supplementary Table S4), indicating that the CCR4-NOT transcription complex is also involved in the defence responses to AAL toxin in tomato.
In plants, plasma membrane (PM) H + -ATPases are the primary pumps responsible for the establishment of cellular membrane potential, which are absolutely essential for normal plant growth and development 45 . A PM H + -ATPase was slightly decreased by TA treatment, suggesting its function in toxin response. Moreover, if DNA damage is left unrepaired or mis-repaired, it can be changed into a mutation. Three DNA damage response related proteins were found to be increased by JA, indicating that JA promotes the DNA repair process to cope with the occurrence of the PCD and mutation.
In summary, we explored the mechanisms in JA regulation of AAL toxin induced PCD using comparative proteomics. We identified a large number of DE proteins induced by TA + JA treatment in WT and jai1. The DE proteins revealed by iTRAQ quantitative proteomics approach in this study help to elucidate the molecular regulating network of COI1-dependent JA pathway in PCD. Numerous new components in cell death machinery were identified, which were summarized in several pathways and represented in Fig. 4. TA causes PCD by inhibiting the synthesis of ceramide and inducing the overproduction of ROS. COI1-dependent JA pathway may promote this PCD progress by influencing the ROS production and scavenging, other hormone signaling pathways or some possible PCD regulators such as caspase-like proteins, autophagy and DNA repair related proteins. Our findings have deepened the understanding of the mechanisms in fungal toxin induced PCD and JA mediated plant defense in response to the fungal toxin. Many of the proteins identified in the present study including their modifications are interesting targets for further genetic and molecular studies to establish the precise roles in cell death and defense regulatory networks. treatment with 1% sodium hypochlorite (NaOCl) for 10 minutes. The germinated seedlings were treated with 1 mM MeJA (Sigma, St Louis, MO, USA). Approximately 24 h or 36 h later, MeJA-insensitive seedlings (roots growth is not inhibited by MeJA) were selected by PCR using genomic DNA using three primers below: P1: 5′-GTGGAGACGATATGTTGAGACTAA-3′; P2: 5′-CCATGGAG TCCATCACCTAACAGT-3′; P3: 5′-GTGGTCAGATCAGAGCCCTCTATT-3′; PCR product with only a 777 bp band are homozygous jai1 mutant. Seedlings were grown in the growth chamber with day/night temperature of 26/18 °C (16/8h). All experiments were carried out using fully expanded leaflets from nodes 4-6 (except for the terminal leaflets) of 7-week-old tomato plants. jai1 homozygotes were screened according to Li et al. 14 . Detached leaflet treatment with JA and AAL Toxin. Treatment was performed as described earlier 46 were labeled with iTRAQ tags 113 and 117; TA treated WT lines were labeled with tags 114 and 118; TA and jasmonic acid treated WT lines were labeled with tags 115 and 119; TA treated jai1 lines were labeled with tags 116 and 121. The biological quadruplicates were analyzed to account for variation among individuals. The combined peptide mixtures were desalted and lyophilized. After labeling, samples were combined, desalted with solid phase extration (The Nest Group, Inc., Southborough, MA, USA), lyophilized and dissolved in strong cation exchange (SCX) solvent A (25% (v/v) acetonitrile, 10 mM ammonium formate and 0.1% (v/v) formic acid (pH 2.8)). The peptides were fractionated using an Agilent HPLC system 1260 with a polysulfoethyl column (2.1 mm × 100 mm, 5 μL, 300 Å; PolyLC, Columbia, MD, USA), flow rate of 0.2 mL/min. Peptides were eluted with a linear gradient of 0-20% solvent B (25% (v/v) acetonitrile and 500 mM ammonium formate (pH 6.8)) over 50 min, followed by ramping up to 100% solvent B in 5 min. The absorbance at 280 nm was monitored; 12 fractions were collected and lyophilized. The fractions were resuspended in LC solvent A (0.1% formic acid in 97% water, 3% acetonitrile). A hybrid quadrupole Orbitrap (Q Exactive) MS system (Thermo Fisher Scientific, Bremen, Germany) was used with high energy collisional dissociation (HCD) after each MS. The instrument was run in data dependent mode with a full MS (400-2000 m/z) resolution of 70,000 and five ms/ms (17500 resolution, HCD NCE = 28%, isolation width = 3 Th, first mass = 105 Th., 5% underfill ratio, peptide match set to 'preferred' , and an AGC target of 1e 6 ). Dynamic exclusion of 10 s was applied to prevent repeated analyses of the same peptides, and a lock mass of m/z 445.12003 (polysiloxane ion) was used for real-time internal calibration.). The MS system was interfaced with an automated Easy-nLC 1000 system (Thermo Fisher Scientific, Germerling, Germany). Each sample fraction was loaded on an Acclaim Pepmap 100 pre-column (20 cm × 75 μm; 3 μm-C18) and separated using a PepMap RSLC analytical column (250 cm × 75 μm; 2 μm-C18) with a flow rate at 350 nl/min. A linear gradient of solvent A (0.1% formic acid) to solvent B (0.1% formic acid, 99.9% acetonitrile) was run for 95 min, followed by a ramp to 98% B for 5 min. The MS proteomics data have been deposited in the ProteomeXchange Consortium 51 via the PRIDE partner repository with the data set identifier PXD002864 and 10.6019/PXD002864. iTRAQ LC-MS/MS data analysis. The raw MS/MS data files were searched against the specified non-redundant database (combined Uniprot, http://www.uniprot.org/uniprot/?query = solanum+lycopersi-cum&sort = score; NCBI, http://www.ncbi.nlm.nih.gov/gquery/?term=solanum+lycopersicum; Phytozome, http://www.phytozome.net/tomato.php) using the Fraglet and Taglet searches under the Paragon TM algorithm 52 of ProteinPilot v.4.5 software (AB Sciex, Inc.). Plant species, fixed modification of methylmethane thiosulfate-labeled cysteine, fixed iTRAQ modification of amine groups in the N-terminus and lysine, and variable iTRAQ modifications of tyrosine were considered. In addition, the iTRAQ data were analyzed using Proteome Discoverer v1.4 (Thermo Fisher Scientific, Bremen, Germany) and the following parameters: peptide tolerance at 10 ppm, tandem MS tolerance at ±0.01 Da, peptide charges of 2+ to 5+, trypsin as the enzyme, allowing one missed cleavage, iTRAQ label and methyl methanethiosulfonate (C) as fixed modifications, and oxidation (M) and phosphorylation (S, T, Y) as variable modifications. Peptides and proteins were filtered using ProteoIQ 2.7 (Premier Biosoft, Palo Alto, CA, USA) with strict peptide and protein probabilities, 0.8 and 0.95, respectively. For peptide confidence, we adopted the following cutoff values of Xcorr that are commonly used for the SEQUEST algorithm 53 : 2.31 for 2+, 2.41 for 3+, and 2.6 for 4+ peptides 54 . Peptide probability was applied to filter peptide assignments obtained from MS/MS database searching results using predictable false identification error rate (See Supplementary Table S2 for peptide information). To be differentially expressed with significance, a protein must have been quantified with a fold change >1.2 or <0.8, with at least three spectra in at least two of the biological quadruplicates, along with a Fisher's combined probability <0.05 55 . A protein is considered to be variable if it has an increase in expression in one component of the comparison and a decrease in the other; these variable proteins within biological replicates were excluded for further analysis. Functional analyses of proteins in biological process, molecular function, and cellular component were conducted using GO annotation (http://www.geneontology.org/) with Fisher's exact test based on false discovery rate (p-value ≤0.05) 56 . Blast2GO level 2-11 filtering was used to examine unique protein changes during comparison analysis. Accession numbers were compared between each genotype and treatment, and unique proteins were identified. Statistical analysis. Differences in the MDA content, H 2 O 2 content and O 2 .− production rate were analyzed by one-way analysis of variance (ANOVA); if the ANOVA analysis was significant (P < 0.05), Duncan's multiple range test was used to detect significant differences between groups.