Jasmonate controls polypeptide patterning in undamaged tissue in wounded Arabidopsis leaves 1

Wounding initiates a strong and largely jasmonate-dependent remodelling of the transcriptome in the leaf blades of Arabidopsis thaliana . How much control do jasmonates exert on wound-induced protein repatterning in leaves? Replicated shotgun proteomic analyses of 2.5 mm-wide leaf strips adjacent to wounds revealed 106 differentially regulated proteins. Many of these gene products have not emerged as being wound regulated in transcriptomic studies. From experiments using the jasmonic acid (JA)-deficient allene oxide synthase ( aos ) mutant we estimated that approximately 95% of wound stimulated changes in protein levels were deregulated in the absence of JA. The levels of two tonoplast proteins already implicated in defense response regulation, TWO-PORE CHANNEL1 (TPC1) and the calcium-V-ATPase ACA4 increased on wounding, but their transcripts were not wound-inducible. The data suggest new roles for jasmonate in controlling the levels of calcium-regulated pumps and transporters, proteins involved in targeted proteolysis, a putative bacterial virulence factor target, a light-dependent catalyst, and a key redox-controlled enzyme in glutathione synthesis. Extending the latter observation we found that wounding increased the proportion of oxidized glutathione (GSSG) in leaves, but only in plants able to synthesize JA. The oxidising conditions generated through JA signaling near wounds help to define the cellular environment in which proteome remodelling occurs.


INTRODUCTION
Leaves, often having large surface-to-volume ratios and lacking thickened protective barriers, are particularly prone to wounding. However, since damage elicits strong defense responses extending beyond the wound site, leaves are some of the most common and resilient of living structures. These wound responses, first identified as damage-induced defense protein accumulation occurring in physically damaged leaves as well as in distal leaves (Green and Ryan, 1972), are now known to involve extensive transcriptional reprogramming (Reymond et al., 2004). Importantly, even strong wound responses in adult plants are not truly systemic.
Instead, they depend in large part on source-sink relationships, as shown for the expression of the gene WOUND-INDUCED3 (WIN3) in poplar (Davis et al., 1991). In adult-phase Arabidopsis, the wound-induced expression of the genes JASMONATE ZIM DOMAIN 10 (JAZ10) and LIPOXYGENASE2 (LOX2) follows intervascular connections termed parastichies (Glauser et al., 2009). While the maximum extent of wound response domains generated after a single severe wound is not yet established fully for the vegetative tissues of any plant, the approximate limits of wound-induced gene expression should, in theory, be predictable if a plant's vascular architecture is known. In summary, the regulation of transcription (and, potentially, translation) takes place within the first few h of wounding in reproducible patterns extending away from the wound into discrete distal tissue domains.
In terms of regulation, the wound response of plants involves the action of multiple signal pathways (e.g. Onkokesung et al., 2010;Walley and Dehesh, 2010). One of them, the jasmonate pathway, is based on the synthesis of the small fatty acid-derived regulator jasmonic acid (JA), and is now known to control and coordinate a particularly large number of responses to wounding (Koo and Howe, 2009). JA signaling relies on the production of amino acid conjugates such as jasmonoyl-L-isoleucine (JA-Ile) from JA, followed by the perception of JA-Ile which probably takes place in the nucleus (Browse, 2009;Fonseca et al., 2009;Gfeller et al., 2010). Significantly, the wound-induced accumulation of JA, like jasmonate-regulated transcripts, also occurs in domains closely reflecting interleaf vascular connections (Glauser et al., 2009).
Statistical analysis have revealed that the levels of 67-84% of transcripts were controlled via jasmonate signaling upon wounding Arabidopsis leaves (Reymond et al., 2004). This transcriptome study, however, used whole wounded leaves i.e. samples containing both crushed and intact tissue. Now, to improve the quantitative assessment of the impact of the jasmonate pathway, better spatial resolution is needed and it will be necessary to 6 focus increasingly on discrete, well defined tissue regions. Several proteomic studies of leaves in response to wounding or herbivory have been conducted (Shen et al., 2003 ;Giri et al 2006 ;Soares et al 2009;Collins et al., 2010;Thivierge et al., 2010), and a recent study has compared the proteomes of wild type (WT) leaves with those from a jasmonate perception mutant (Shan et al., 2011). Nevertheless, wound proteomic studies comparing the WT and jasmonate mutants have not been reported.
In this work we aimed to quantitate the role of jasmonate in early proteome repatterning after wounding. The selection of the tissue area to be studied proved to be challenging since many wound responses in leaves diminish with increasing distance from the site of damage. This has been shown to be the case for two proteinase inhibitor proteins in wounded tomato leaves (Graham et al., 1986). Similarly, this occurs with wound-stimulated JA synthesis in Arabidopsis leaves (Glauser et al., 2009). Furthermore, other studies indicate that the expression of multiple inducible transcripts decrease with distance from a wound (e.g. Chung et al., 2008;Farmer et al., 1992;Reymond et al., 2004). In summary, protein and transcript levels can form sectors or gradients within leaves or intervascular domains and these are not truly systemic responses in adult phase plants.
A specific goal of our study was to provide new information on the molecular environment in undamaged plant tissues in the vicinity of a wound, coupled to a quantitative estimation of the extent of control exerted by jasmonate signaling in reshaping the proteome in response to tissue damage. Through the analysis of nearly 6000 proteins we assessed proteome remodelling events that unfold in the 6 h following wounding and observed a larger than expected impact of jasmonate on the formation of the wound proteome. Among the results was the finding that the levels of proteins of glutathione (GSH) synthesis and deployment were affected by wounding. Based on this we tested whether the oxidation status of GSH pools was affected and this has provided new insights into the nature of the cellular environment near a wound.

Experimental design
To evolve an experimental design we first examined the possibility that the presence of crushed tissue from the wounding procedure might complicate the interpretation of protein pattern changes. For this, leaves were wounded successively 3 times (each time damaging approximately 10% of the total leaf area) with forceps at times 0, 2 and 4 h, starting at the 7 apex. Six h after the first wound, tissue strips of 2.5 +/-0.5 mm width were harvested ( Fig.   1A and B). Quantitative PCR was performed to assess the relative levels of wound-responsive transcripts. From data in Reymond et al. (2000) we used the RNS1 gene as a crush marker and found that RNS1 transcripts were more highly expressed in crushed tissue than in the proximal strip of undamaged tissue (Fig. 1C). In contrast, the expression a jasmonate-response reference gene JAZ10 (formerly JAS1; Yan et al., 2007) was higher in intact tissue than in the crushed tissue (Fig. 1D). These results suggested that the inclusion of crushed tissue could, in theory, introduce heterogeneity into proteomic studies. Only intact, undamaged tissue was harvested for subsequent proteomics experiments. We chose a 6 h harvest time since, at the protein level, this is likely to be an early timepoint and may reduce indirect effects where, for example, an induced protein might alter the level of a second protein. Since many herbivores wound plants in a repetitive manner three successive wounds were inflicted.

High-throughput proteomic experiments
With linear trap quadrupole (LTQ) ion-trap mass spectrometry (MS/MS) techniques described by Baerenfaller et al. (2008) we performed 240 individual ion-trap MS/MS runs of protein extracts from the 2.5 mm wound-proximal zone and its spatial equivalent in unwounded plants. From these replicated experiments a total of 6530 proteins were identified based on 220,484 peptide spectrum assignments with two search algorithms at a spectrum false discovery rate below 1%. As proteins are quantified by taking into account solely true tryptic peptides, the number of quantified proteins is 5936 originating from 219,463 spectra. This led to the identification of 106 wound-regulated proteins (69 induced, 37 repressed) in the WT ( Fig. 2A, Table S1). To assess the reliability of the data we used enzymes involved in JA synthesis as landmarks and finding of 6 of these proteins (LOX3, AOC1, OPR3, ACX1, ACX3, and JAR1) upregulated in the wounded WT suggested that we captured many proteins associated with wound responses. Extended analysis revealed that GO categories that were up-or down-regulated in wounded WT leaves (Fig. 2B, Table S2). Next, we compared control and wounded leaves of the aos mutant and found 114 wound-regulated proteins of which 51 are more abundant in wounded and 63 in control leaves (Table S3). The list of proteins that were more abundant in wounded aos leaves only shared 4 proteins with the list of proteins that are more abundant in wounded WT leaves, namely a sulfotransferase family protein, a putative peroxidase, AtEXO70H7 and ALDH6B2. ANNAT2 protein was repressed both in WT and aos. Remarkably, 95% of the wound-regulated proteins in the WT were not woundregulated in aos. However 109 proteins (47 induced, 62 repressed; Table S3)  (1 induced in WT; 5 induced, 2 repressed in aos; Tables S1 and S3). In order to investigate how the aos mutation affects the otherwise healthy plants, the protein levels in resting, unwounded WT and aos plants were then compared. The control unwounded WT and control aos showed 58 differences with 34 proteins being more abundant in aos (Table S4, summarized in Fig. S1). Almost all of the differentially regulated proteins in resting leaves were found in lower levels in aos relative to the WT. At this point we directly compared the ratios of protein induction/repression in the WT with the analogous ratios for aos (see Table   S5). This led to the identification of 116 proteins that were differentially regulated in the two genotypes where the levels of 62 proteins were lower in wounded aos than in wounded WT.
The functional classification of the proteins that are more abundant in control WT as compared to control aos gave jasmonic acid biosynthetic process as the most over-represented GO category (Table S2). Many proteins not known previously to be jasmonate-regulated were observed in this way and the major categories are summarized in Fig. 2C. Fig. 3 shows annotated proteins that were wound-induced in the WT, and indicates which of these were regulated in a jasmonate-independent manner. Finally, 60% of the wound-regulated proteins we found in WT tissues have also been recorded as wound-regulated at the transcript level in WT plants grown under similar conditions. However, we observed 15 downregulated and 26 upregulated proteins that did not show the same pattern of regulation at the transcript level.
These proteins are listed in Table S6. Using wounding conditions identical to those used for proteomics we assessed the mRNA levels of several of these genes chosen since they represent transport functions. Two proteins that we upregulated by wounding were chosen : TPC1 and ACA4. These were not induced at the mRNA level, and both mRNAs were lower in wounded than in unwounded leaves. ADNT1 and ANNAT2 were chosen for transcript abundance analysis since they were both wound-repressed at the protein level. ADNT1 9 transcripts were weakly downregulated by wounding whereas ANNAT2 transcripts were weakly upregulated by wounding in the WT (Fig.4).

Glutathione analyses
A number of proteins that were found to be wound-inducible participate in GSH synthesis or employ GSH in various reactions. We therefore analysed the oxidation state of GSH proximal to the wound. A significantly higher proportion of oxidized GSH (GSSG/total GSH ratio = 6.24 +/-0.68) was seen in the wounded WT relative to unwounded WT (ratio = 3.79 +/-0.35).
However, this was not observed in aos leaves where the GSSG/total GSH ratio in unwouded leaves was 4.29 +/-0.21 and this remained essentially constant at 3.76 +/-0.54 after wounding ( Fig. 5).

DISCUSSION
Unless very severe, signals emanating from wounds probably only rarely fill the potential boundaries of wound response domains: parastichies in adult-phase Arabidopsis. Instead, wound-generated signals are often restricted to smaller regions. These can take many forms depending on where the leaf is wounded, but for proteomic experiments we used staggered leaf tip wounding in keeping with the fact that many invertebrate herbivores feed at intervals on the same area of leaf and due to the fact that small wounds such as pin pricking do not activate the jasmonate pathway strongly in expanded leaves (Farmer et al., 1992). We chose to harvest tissue at 6 h after the first wound so that secondary effects on the proteome would, in theory, be mimimized.  (Table S1) differed greatly from those that were wound-regulated in aos (Table S3). Firstly, it was striking that only five of the wound-regulated proteins in WT leaves were still significantly wound-regulated in aos. The levels of these proteins may be regulated through signal pathways that are largely JA-independent. Secondly, the large number of protein changes (114) elicited in wounded aos was unexpected. It seems most likely to us that jasmonates otherwise repress the majority of these changes, probably through crossregulation of other, non-jasmonate-based signal pathways. It is also possible that in the absence of jasmonate-regulated increases in GSH oxidation in aos (Fig. 5)  environment near the wound leads to a strong deregulation of many genes. We also compared the protein complement of resting (undamaged) leaves of the WT and of aos. In apparently healthy leaves the aos mutation impacts the resting chloroplast proteome and, in addition, we noted the deregulation of proteins that were annotated as stress-related. In aos a novel set of 47 proteins becomes wound-inducible. The simplest explanation of this is that jasmonates normally suppress the accumulation of these proteins in the WT through modulating the activity of other signal pathways. Indeed, Table S5 shows that more proteins that are implicated in signaling (including protein kinases) and posttranslational modification were downreglated in wounded aos (8 proteins) than in wounded WT (1 protein). In order to generate results that would be comparable with previous transcriptome studies we compared the ratios of protein induction and repression in the WT and in aos. This was necessary because of the altered basal levels of proteins that were observed when we compared the two unwounded genotypes. The results of these experiments (Table S5)  that the bulk of early proteome remodelling in response to wounding is directed at improving survival and in many cases, protein level changes may be associated indirectly with defense.

Generalities from proteome analysis
For example, the downregulation of the porphyrin-metabolizing enzyme POR C suggests decreased nitrogen flux into porphyrin synthesis. This may ultimately be defense-related since nitrogen is needed for glucosinolate and defense protein synthesis.

Wounding affects transport functions
The tonoplast is emerging as a potentially important compartment for the regulation of defense pathways. Three tonoplast transporter proteins, two of which have been implicated previously in the regulation of defense signal pathways, accumulated after wounding. In each case, this required JA synthesis. First is the calcium-regulated two-pore channel 1 (TPC1;  et al., 2010). We found that ACA4 protein was wound inducible. The third tonoplast-localized transporter that was wound-induced in the WT was MRP2 (At2g34660). This ABCC-type ABC transporter protein is a known part of the organic anion pump capacity of the vacuole membrane (Frelet-Barrand et al., 2008). Several proteins that might directly regulate water flux (and therefore turgor) were also found to be upregulated in the wound-proximal region.
An example of this was PIP2A (At3g53420), a plasma membrane protein that facilitates the bidirectional flow of water and H 2 O 2 (Ludewig and Dynowski, 2009).
IMPa-4 (At1g09270), an α-importin, forms part of the nuclear pore-targeting complex responsible for the energy-dependent translocation of proteins into and out of the nucleus.
This wound-induced protein is utilized by the pathogen Agrobacterium tumefaciens in order to efficiently transform plant cells (Bhattacharjee et al., 2008). Another wound-induced protein that has been linked to pathogenesis is PEN3 (PDR8; At1g59870) a plasma membrane-localized ATP binding cassette transporter possibly involved in the export of antimicrobial agents into the apoplast (Stein et al., 2006). PXA1 (At4g39850), a peroxisomal ATP-binding cassette transporter implicated in the transport of fatty acyl CoAs (Nyathi et al., 2010), was upregulated in resting aos relative to the WT. Finally, we found that three predicted chloroplast transport proteins were all downregulated in response to wounding and, in each case, this required JA synthesis. One of these was cpSecY (SCY1; At2g18710), a nuclear-encoded thylakoid translocase component (Dalbey and Kuhn, 2000). To our knowledge this is the first component of a thylakoid protein import machine found to be jasmonate-regulated.

Chloroplast and light-related functions
Treatment of excised barley leaf segments with exogenous jasmonate is known to impair the translation of mRNAs for ribulose-1,5-bisphosphate carboxylase/oxygenase (Rubisco). In the case of the plastid-encoded large subunit, jasmonate treatment resulted in the production of abnormally long transcripts (Reinbothe et al., 1993). We did not observe changes in Rubisco protein levels in the 6 hour timeframe of our experiments, although we did observe downregulation of a putative Rubisco activase (At1g73110; Fig. 3 Our study also revealed that PHOTOSYSTEM 1 SUBUNIT P (PSI-P; PSAP, At2g46820) was strongly downregulated in the resting leaves of aos with respect to the WT meaning that both the dark and light reactions of photosynthesis are wound-and jasmonate-regulated at the protein level. A number of other proteins involved in various aspects of photobiology were wound-regulated (Tables S1 and S5). Interestingly, POR C (At1g03630), one of the rare enzymes requiring light for full catalytic activity (Frick et al., 2003), was downregulated following wounding.

Defense functions
LOX2 (At3g45140) levels increased significantly in the wounded WT compared to wounded aos but is not indicated in Fig. 3 since its induction fell just below the 2-fold threshold we set. was found to be deregulated in wounded aos relative to the wounded WT. Additionally, the levels of two jacalin lectins also increased in a JA-dependent manner after wounding (Table   S1 and S5) as did β-glucosidase BGL1 (At1g52400), a hydrolase known to hydrolyze glucose-conjugated ABA to release the active hormone (Lee et al., 2006).

Strong impact on proteins of vesicular transport, cell wall and cytoskeleton
Levels of dynamin-like protein 6 (ADL6; At1g10290) and V-Snare 11 (VTI11; At5g39510), both of which are annotated as being potentially involved in vesicle trafficking from Golgi to vacuoles, were altered by wounding. These candidates could now be employed as fusion proteins in order to probe endomembrane dynamics in the wound-proximal zone. Vesicle secretion (exocytosis) may modify the apoplast in response to damage and Soares et al. (2009) found that apoplastic glucanases, chitinases, and peroxidases accumulated in response to leaf wounding in Medicago truncatulata. We observed wound-regulated proteins involved in 13 realignment around wound sites and this peaked between 1.5 to 5.5 h after wounding, i.e.
within the general timeframe of our study.

Jasmonate-dependent remodelling of the protein synthesis and degradation machinery
We observed that INITIATION FACTOR 3 (IF-3; At5g39510) was down-regulated in leaves in response to wounding and 5 putative ribosomal proteins (both prokaryotic and eukaryotic) and one ribosome-associated protein were also downregulated in the wounded WT. It is known that a ribosome-inactivating protein, JIP60, is induced upon jasmonate treatment in barley (Reinbothe et al., 1994). While JIP60 activity was not itself affected by jasmonate treament, its substrates (ribosomes) themselves were probably modified in response to jasmonate treatment. Our results are also consistent with ribosome remodelling under control of the jasmonate pathway near the wound. Additionally, the comparison of resting WT and aos leaves also suggests that some of the protein synthesis machinery in resting leaves may be modeled in part due to the activity of the jasmonate pathway. For example, two 40S ribosomal proteins, RPS3aA (At3g04840) and RPS24B (At5g28060) were downregulated in the resting leaves of aos relative to the WT and a putative eukaryotic cytosolic elongation factor (At1g57720) was upregulated at the protein level in resting aos relative to WT. Another interesting aspect was the indication that wounding, via activity of the jasmonate pathway, may regulate components of ubiquitin-dependent proteolysis. One such example is proteasome-associated 200, PA200 (At3g13330), a regulatory subunit of the 26S proteosome (Book et al., 2010) that we now find to be upregulated by wounding. A second example, SKP1 (At1g75950), emerged from comparison of the proteomes of wounded WT and aos plants.

Translational and post-translational events
Some of the proteins identified in be regulated posttranscriptionally in the WT since the levels of their RNA and protein products behave oppositely on wounding. Only one of the genes tested (ADNT1) behaved similarly at both the protein and RNA levels although it was much more strongly repressed at the protein level than at the transcript level. It seems, then, that transcription and translation may not be well correlated for these proteins. A number of polypeptides we found to be JAregulated are known to be themselves involved in post-translational modification, potentially modifiying the activities of proteins rather than their levels. An example of this is TIP1 These considerations and the fact that GCL (RML1/PAD2; At4g23100), the major regulatory enzyme of GSH synthesis, was wound-inducible prompted us to assess GSH oxidation status and we found that the ratio of oxidized GSH (GSSG) over total GSH increased in the wounded WT, but not in wounded aos (Fig. 5). Therefore JA production actively signals the generation of a relatively oxidizing cellular environment near the wound border and this might potentiate the activities of many proteins. The jasmonate-controlled redox environment in wounded tissue may help to directly or indirectly regulate the activities of many genes and proteins including jasmonate-regulated genes such as JAZ10, and also some of the redox-sensitive proteins seen in our proteomics experiments.

Proteomics summary: an estimated 95% of protein repatterning in a zone near the wound is jasmonate-regulated
Many processes that result from wounding are known to be heavily JA-dependent and among these are transcriptome remodelling. Previous estimates of transcript accumulation after wounding in Arabidopsis leaves are within the 1 to 5 hour timeframe (Reymond et al., 2004).
Within this interval 67 -84% of transcriptome remodelling in leaves was controlled by jasmonate signalling. The present results show that, after a series of wounds inflicted at 2 to 6 h prior to harvest, 95% of the differentially expressed proteins recovered in a strip of healthy tissue 0.5 mm to 3.0 mm from a wound were regulated in a jasmonate-dependent manner.
This percentage should be taken only as a rough estimate since a) the proportion of low abundance proteins (not accessible to our analyses) regulated by the jasmonate pathway remains unknown and b) the results depend on the statistical criteria we used. This excluded many proteins upregulated or downregulated by < 2-fold after wounding.
Five expanded leaves (leaves 8 to 12) per plant were successively wounded (starting at the leaf tip, each time 10% of the leaf surface was wounded) at 0, 2 and 4 h with metal forceps.
The 2.5 +/-0.5 mm of tissue directly neighbouring the wound were harvested, excluding a border of approximately 0.5 mm of unwounded tissue. Eight plants were utilized for each proteomic sample and each data set was from three biologically independent replicates.
www.plantphysiol.org on September 2, 2017 -Published by Downloaded from Copyright © 2011 American Society of Plant Biologists. All rights reserved.

Real-Time PCR Quantitation
Total was ultracentrifuged at 100,000 g for 45 min at 25°C. The supernatants from the two centrifugation steps were combined. Protein concentration of the soluble fractions was determined with a Bradford assay. About 250 μg of each protein fraction were subjected to SDS-PAGE. After electrophoresis, the gels were cut into 5 equal parts and each gel slice was diced into small pieces. Gel pieces were de-stained in 50% (v/v) methanol containing 100 mM ammonium bicarbonate, then washed with water and stored at -20°C prior to in-gel tryptic digest according to Shevchenko et al. (1996). After tryptic digest, the peptides were purified using Sep-Pak reverse-phase cartridges (Waters, Milford, MA). Peptide ions were detected in a survey scan from 400 to 2,000 amu followed by 3 datadependent MS/MS scans (isolation width 3 amu, relative collision energy 35%, dynamic exclusion enabled, repeat count 1, repeat duration 30 seconds, followed by peak exclusion for 4 min). Each sample was injected and measured twice.

Interpretation of MS/MS spectra, data filtering and export to PRIDE
Peptide spectrum assignment and data filtering have been described (Baerenfaller et al., 2008;Baerenfaller et al., 2011). In brief, these consisted of searching the MS/MS spectra 1) with results, the cut-off was set to a minimum probability of 0.9, for PepSplice the false discovery rate was adjusted to < 0.01, assessed for each search space separately. All peptide spectrum assignments above the determined threshold, except those of known contaminants, were filtered for ambiguity. Peptides matching to several proteins were excluded from further analyses. This does not apply to different splice variants of the same protein or to different loci sharing exactly the same sequence. All remaining spectrum assignments were entered into the pep2pro database (Baerenfaller et al., 2011). After database upload, the following spectrum assignments were flagged and not taken into consideration for further analyses: 1) spectrum assignments to decoy database peptides, and 2) spectra for which PeptideProphet and PepSplice assign a different peptide to the same spectrum (including different posttranslational modifications and differently charged peptides). The spectrum false discovery rate was calculated by dividing the number of decoy database spectrum assignments by the number of spectrum assignments in the final dataset. PRIDE 2.1 XML files were created from the final fully integrated dataset and exported to the PRIDE database (Vizcaino et al., 2010: accessions 13327-13334). The dataset is also available at www.pep2pro.ethz.ch in assembly 'Arabidopsis thaliana leaf wound proteome'.

Protein quantification, Statistics and Gene Ontology (GO) classification
The normalized spectrum count (factor) for the proteins is determined by calculating the expected contribution of each individual protein to the sample's total peptide pool as differentially regulated if they fulfilled the following criteria: A) the p-value of the two-tailed paired Student's t-test between the normalised factors expressing the abundance of the protein had to be below 0.1; B) The fold induction or repression ratio had to be ≥ 2 or ≤ 0.5; C) There had to be at least 10 spectra per protein; D) The spectral counts and normalised factors had to show the same pattern of induction or repression in the three independent replicates.
Assignment of protein functions was based on the TAIR GO categories from the aspect biological process (download ATH_GO_GOSLIM_20080510.txt) (Berardini et al., 2004).
The assignment was performed in R (http://www.r-project.org) using the elim method from the topGO package (Alexa et al., 2006). Fisher's exact test was used for assessing the GO term significance.

Comparison of protein and transcript levels
Wound-induced protein levels were compared with transcriptome data in Yan et al. (2007); ArrayExpress accession E-ATMX-9. In this analysis only proteins and transcripts from genes covered in both types of analysis were considered. Wound-induced proteins were considered likely to be transcriptionally regulated if the average of the log-transformed ratio (log 2 ) of transcript level between unwounded and wounded leaves was below -0.59 or above 0.58 in the microarray dataset, corresponding to a 1.5 fold change in transcript levels.

GSH redox state
GSH redox state in 50 mg fresh weight samples from the 2.5 mm region adjacent to the wound was determined with an enzyme-cycling assay (Griffith, 1980).

Immunoblotting
LOX2 protein was detected as described in Glauser et al. (2009). Each lane of the 8% polyacrylamide gel was loaded with 30 μg protein extract. Figure S1: Predicted cellular locations of proteins differentially regulated in WT and aos. Table S1: Wound-regulated proteins in WT 6 h after wounding.    20 Table S6: Potentially posttranscriptionnally wound-regulated proteins.  A. Numbers of differentially regulated proteins from all proteomics experiments. Also indicated are the five wound-induced proteins in common between WT and aos. B. Overrepresented GO categories for proteins that were more abundant in wounded WT leaves (dark grey) and in control (unwounded) WT leaves (light grey). C. Overview of protein group regulation by jasmonates in response to wounding based on Table S5. Upward arrows indicate that proteins in the category are mostly upregulated via activity of the jasmonate pathway in response to wounding and downward arrows that they are mostly downregulated via activity of the jasmonate pathway in response to wounding. Contrasense arrows indicate differential protein regulation controlled by jasmonates. Arrows show the two-fold up-and down-regulation of proteins. Downward arrows mean that wounding leads to decreased protein level. Arabidopsis Gene Indentifier (AGI) codes are given for proteins lacking consistently used acronyms. Quantitative data for this experiment are given in Table S1. Black arrows = p-value < 0.05; grey arrows = p-value 0.05 -0.1.

Figure legends
Proteins indicated in red were regulated in a jasmonate-independent manner. PTM = proteins potentially involved in postranslational modification.