Proteomic analysis upon peach fruit infection with Monilinia fructicola and M. laxa identify responses contributing to brown rot resistance

Brown rot, caused by Monilinia spp., is a major peach disease worldwide. In this study, the response of peach cultivars Royal Glory (RG) and Rich Lady (RL) to infection by Monilinia fructicola or Monilinia laxa, was characterized. Phenotypic data, after artificial inoculations, revealed that ‘RL’ was relatively susceptible whereas ‘RG’ was moderately resistant to Monilinia spp. Comparative proteomic analysis identified mesocarp proteins of the 2 cultivars whose accumulation were altered by the 2 Monilinia species. Functional analysis indicated that pathogen-affected proteins in ‘RG’ were mainly involved in energy and metabolism, while, differentially accumulated proteins by the pathogen presence in ‘RL’ were involved in disease/defense and metabolism. A higher number of proteins was differentiated in ‘RG’ fruit compared to ‘RL’. Upon Monilinia spp. infection, various proteins were-down accumulated in ‘RL’ fruit. Protein identification by mass spectrometric analysis revealed that several defense-related proteins including thaumatin, formate dehydrogenase, S-formylglutathione hydrolase, CBS domain-containing protein, HSP70, and glutathione S-transferase were up-accumulated in ‘RG’ fruit following inoculation. The expression profile of selected defense-related genes, such as major latex allergen, 1-aminocyclopropane-1-carboxylate deaminase and UDP-glycoltransferase was assessed by RT-PCR. This is the first study deciphering differential regulations of peach fruit proteome upon Monilinia infection elucidating resistance responses.

control with microbial antagonists or use of resistance inducers, but none of these approaches has gained until now a commercial feasibility [12][13][14] .
Host resistance to brown rot could be an attractive, cost effective and environmentally friendly alternative to chemical control. However, most of the commercial peach cultivars are susceptible to brown rot, although some differences exist among them in the level of susceptibility to the disease [15][16][17] . The susceptibility of peach fruit is also associated with the stages of their development and appear more susceptible in the very early stages, during the development of the pericarp and later at the stage of ripeness of the fruit, after the endocarp and the mesocarp have been fully developed 8,11 . The mechanisms behind the different sensitivity levels at different stages are allegedly due to physical and chemical reasons 18 . For instance, the high levels of resistance in the peach cultivar Bolinha are associated with the presence of higher concentrations of phenolic compounds, such as caffeic acid and its quinate ester, chlorogenic acid, located in the epidermis and mesocarp 8,[11][12][13][14][15][16][17][18][19][20][21] . While these substances have no direct effect on mycelial growth and spore germination, they inhibit the production of the cell wall degrading enzymes polygalactorunase and cutinase, by down regulating their respective genes 8 . Additional factors, also playing an important role in resistance to the disease, are associated with different levels of the environmental pH of the host, regulated by the pathogen and affecting the expression of genes, responsible for pathogenesis 22 . Despite intensive investigation, the mechanism of peach fruit brown rot resistance remains unknown.
In recent years, proteomics has gained popularity in understanding host-pathogen interactions 23,24 . Comparative proteomics have been successfully used to study interactions of plants with a wide range of pathogens 25 . However, no proteomic studies have been published to date investigating the interactions of peach fruit with Monilinia species. To address this issue, the responses of two peach cultivars, 'RL' and 'RG' , differing in their level of resistance to infection by M. fructicola or M. laxa, was characterized. Using two-dimensional gel electrophoresis (2-DE) analysis, we compared the proteomic changes from the fruit mesocarp of the two cultivars following inoculation with either M. fructicola or M. laxa to identify proteins potentially involved in defense mechanisms.

Results
Disease incidence. The artificial inoculations of the fruit of the 2 different peach cultivars revealed a higher rate of disease incidence at 'RL' fruit compared to that observed at 'RG' fruit, either inoculated with M. fructicola or M. laxa isolates (Fig. 1a,b). However, the difference in the disease incidence between the 2 cultivars was lower, when the fruits were inoculated with M. fructicola isolates Mf 101 and Mf 167, with disease incidence values of 84.2 and 90.5% in 'RL' and 66.7 and 65.2% in 'RG' , respectively (Fig. 1b). Moreover, a higher difference in disease incidence was observed, when fruits were inoculated with M. laxa isolates Ml 1 and Ml 66, with infected fruit rates reaching 90 and 100% in 'RL' and only 18.2 and 8.7% in 'RG' , respectively (Fig. 1b). characterization of the peach fruit proteome. To obtain a better understanding of the mechanisms behind the disease incidence data, a two-dimensional gel electrophoresis  analysis was conducted to detect the protein changes in both cultivars 'RL' and 'RG' (Fig. 2a,b). In both cultivars a total of 250 protein spots, were detected, while 120 of them showed significant differences in their volume in the presence of either M. fructicola or M. laxa (Fig. 2c,d, Supplementary Data Figs. S1 and S2), according to the Student's t-test for 95% confidence further validated by the two-fold change threshold. Following mass spectrum analysis, 126 different peach proteins were identified, while 11 of them were detected in more than one spots. These proteins were aconitate hydratase (spots no.: 4828, 4829), adenylyl cyclase-associated protein (spots no: 8510, 8520), aminoacylase ( Table S1. Subsequently, proteins based on gene ontology and literature, were grouped into functional categories 26 . functional analysis of proteins modulated by the presence of the pathogen relative to the control fruit, within the same peach cultivar. In order to identify the peach proteins whose abundance was significantly changed (increased or decreased) by the presence of the pathogen in relation to the control, within the same peach cultivar, proteins were grouped into four sets: (1) a set of 33 proteins modulated in 'RG' by the presence of M. laxa (RGC vs RGL) (Fig. 3a), (2) a set of 32 proteins modulated in 'RG' by the presence of M. fructicola (RGC vs RGF) (Fig. 3b), (3) a set of 20 proteins modulated by the presence of M. laxa, in 'RL' (RLC vs RLL) (Fig. 3c) and (4) a set of 15 proteins modulated by the presence of M. fructicola, in 'RL' (RLC vs RLF) (Fig. 3d). Functional analysis revealed that the M. laxa affected several proteins (n = 33) in 'RG' that were mainly involved in energy (24.2%), metabolism (21.2%) and protein destination and storage (15.1%) (Fig. 3a). Similarly, the M. fructicola affected proteins in the same cultivar were mainly involved in metabolism (25%), energy (21.8%) and protein destination and storage (18.7%) (Fig. 3b). In contrast, M. laxa affected proteins in 'RL' participated mainly in disease/defense (30%), energy (25%) and metabolism (20%) (RLC vs RLL) (Fig. 3c) while, the M. fructicola affected proteins in the same cultivar participated in disease/defense (26.7%), metabolism (26.7%), energy (13.3%), protein destination and storage (13.3%) and secondary metabolism (13.3%) (RLC vs RLF) (Fig. 3d).
Proteins functional analysis: 'Royal Glory' and 'Rich Lady' inoculated with M. laxa or M. fructicola, versus not inoculated cultivars. Information concerning distinct or common proteins affected by M. fructicola or M. laxa in the 2 peach cultivars, in relation to control fruits, are given in Fig. 3f. Among the 2 sets of proteins modulated in 'RG' by the presence of M. fructicola and M. laxa, 16 proteins were common, while 10 proteins were specifically modulated in response to M. laxa and 9 proteins were specifically modulated in response to M. fructicola (Fig. 3f). In 'RL' fruits, 6 and 5 proteins were individually changed by the presence of M. laxa and M. fructicola, respectively, while 8 proteins overlapped between the 2 sets of proteins (Fig. 3f).
Information on common and distinct proteins between the 'RG' and 'RL' peach cultivars inoculated with the same pathogen are provided in Fig. 4e. The commonly affected proteins (n = 26) between the 3 sets of proteins modified by the presence of the same pathogen or in the absence of pathogen, between 'RL' and 'RG' respectively, were mainly involved in metabolism (34.6%), energy (23.1%) and protein destination and storage (11.5%) (Fig. 4d). Proteins targeted specifically by the presence of M. laxa (n = 17) between 'RL' and 'RG' were mainly involved in metabolism (17.6%), energy (17.6%), protein destination and storage (17.6%) and disease/defense (17.6%) (Fig. 4a). Proteins related with metabolism (33.3%) and cell structure (22.2%) were predominant in

Proteins functional analysis: comparison between 'Royal Glory' and 'Rich Lady' inoculated with different Monilinia species.
Comparisons performed to identify proteins modulated within the same peach cultivar in response to the 2 different pathogens, M. fructicola and M. laxa, revealed a set of 2 proteins in 'RG' differed between the fruits which were inoculated by M. fructicola and M. laxa (RGF vs RGL) (Fig. 5b) and a set of another 2 proteins in 'RL' differed between the fruits which were exposed to M. fructicola and M. laxa, respectively (RLF vs RLL) (Fig. 5c).
There were no common proteins among the 2 group of proteins modulated within the same cultivar by the presence of a different pathogen, M. fructicola and M. laxa, respectively (Fig. 5a). Only 2 proteins were specifically differentiated within 'RG' and 'RL' , respectively (Fig. 5b,c).
Peach proteins of the mesocarp of 'RL' and 'RG' cultivars differentiated (up-accumulated or down-accumulated) after inoculation with M. fructicola or M. laxa are given as heat map profile in Fig. 6 and discussed below.
Quantitative Rt-pcR assays. To investigate the correlation between proteins accumulation level and mRNA expression, a RT-qPCR analysis was performed for 6 representative defense-related genes, selected based on the data derived from the proteome analysis. Using this approach, we found that 4 of these 6 genes, namely thaumatin (Fig. 7a), glycosyltransferase_GTB_type (Fig. 7b), major latex allergen (Fig. 7c) and UDP-glycoltransferase (Fig. 7d) were in agreement with the corresponding protein levels as analyzed by proteomic analysis ( Fig. 7; Supplementary Data Table S1). However, no correlation between mRNA expression and protein abundance of formate dehydrogenase (Fig. 7e) and 1-aminocyclopropane-1-carboxylate deaminase (Fig. 7f) was observed.

Discussion
This study represents a first attempt to investigate the response of peach fruit at proteomic level following infection by either M. laxa or M. fructicola, the two most common brown rot agents, and may constitute a groundwork for future studies. The phenotypical data acquired by artificial inoculations, revealed significant differences in the levels of resistance to brown rot between the two different peach cultivars used making them an interesting experimental model to characterize proteome-based responses to the disease.
In this work, the comparison of the proteomic profile of artificially-inoculated and non-inoculated control fruit of 'RG' and 'RL' cultivars, showed that a higher number of proteins was changed in the 'RG' fruit exposed to either M. fructicola or M. laxa. Such results suggest that the relatively resistant cultivar displayed a stronger  www.nature.com/scientificreports www.nature.com/scientificreports/ proteome-associated response to brown rot infection in relation to the more susceptible cultivar. Despite the differences in the disease incidence between M. fructicola and M. laxa observed in both cultivars, the artificial inoculations with both Monilinia species induced a similar proteomic response in the two cultivars. Thus, it is tempting to speculate that the peach fruit's response to infection is presumably differentiated in terms of cultivar-specific protein changes that are involved in the resistance. On this basis, an interesting finding that emerged from this work is the overall accumulation of various defense-related proteins such as thaumatin, formate dehydrogenase, S-formylglutathione hydrolase, UDP-glycosyltransferase, 1-aminocyclopropane-1-carboxylate deaminase, CBS domain-containing protein, HSP70, L-galactono-1,4-lactone dehydrogenase, phosphoglycerate kinase, mitochondrial outer membrane protein porin, major latex allergen and glutathione S-transferase that observed in the relatively resistant 'RG' fruit following inoculation with either M. fructicola or M. laxa. For several of these proteins (thaumatin, major latex allergen, glutathione S-transferase or UDP-glycosyltransferase) their up-accumulation was confirmed by the over-expression of the respective genes using qPCR. However, there were also some of these proteins (1-aminocyclopropane-1-carboxylate deaminase, formate dehydrogenase) for which their over-accumulation was not confirmed by the qPCR data. This mismatch between the protein and gene expression can be due to various reasons such as post-transcriptional processes that may modify the active protein levels, such as synthesis, degradation, processing and post-translational modification, as shown in many studies 23,27 . In the following, we discuss the changes of several of these defense-related proteins and their respective encoding genes that presented in Fig. 6, in an effort, to determine their biological functions and their contribution in the defense of peach fruits against brown rot.
Plant cells have displayed specific defense-related protein reprograming to combat pathogen infection 23 . One of the acquired modes in response to pathogen attack is the production of the pathogenesis-related (PR) proteins, which play key roles in plant disease-resistance responses 28 . For example, thaumatins and thaumatin-like proteins with similar nucleotide sequences are considered pathogenesis-related (PR) proteins and are mainly involved in the induction of resistance against pathogens 29 . Members of the thaumatin family have been reported to exhibit strong antifungal activity, participating in inhibition of mycelial or germ tube growth and in reduction of sporulation of many fungal pathogens in several fruits 30 . The results of this study indicated that thaumatin was strongly over-accumulated in 'RG' fruits exposed to either M. fructicola or M. laxa as well as in 'RL' fruit treated with M. fructicola. In support to this observation we found that thaumatin expression was strongly induced in 'RG' fruits inoculated with M. laxa compared to control, suggesting that thaumatin plays an important anti-disease role in peach fruit against brown rot. Meanwhile, thaumatin proteins have been reported to be up-accumulated in chilled-injury resistant cultivars 31 . Thus, the accumulation of thaumatin might be a common response cascade during both abiotc and biotic stimuli in peach fruit.
Heat shock proteins (HSP) are a multi-gene family of stress proteins that protect the cell's functions in response to stress conditions 32 . In this study, the abundance of HSP70 was increased in 'RG' fruit that had been inoculated with either M. fructicola or M. laxa, providing evidence of their involvement in resistance to these pathogens. In plants, HSP70 exhibit a protective role against these stresses acting as a molecular chaperone and playing an important role  www.nature.com/scientificreports www.nature.com/scientificreports/ in protein folding, intracellular trafficking of proteins and helping to refold proteins that were denatured due to abiotc cell stress 33 . Overwhelming data support that HSP70 is also an important molecular chaperone during biotic stress responses. Cytoplasmic Capsicum annuum HSP70 (CaHSP70) significantly accumulates in pepper leaves, inducing the hypersensitive response (HR) by Xanthomonas campestris pv. vesicatoria (Xcv) infection. CaHSP70 silencing in pepper was shown to increase susceptibility to Xcv as well as alter the cell death response to Xcv infection 34 .
Another interesting finding that emerged from this work is the fact that the 'RG' proteome status was differently modified by the two Monilinia species. For example, several proteins, such as major latex allergen and glutathione S-transferases (GST) were up-accumulated only in 'RG' fruits inoculated with M. fructicola, signifying their involvement in the defense of this cultivar to brown rot. In support to the role of GST in brown rot, our data disclosed that 'RG' fruit exposed to M. fructicola induced GST expression. GST's are a large group of conserved enzymes, known for their ability to convert xenobiotic compounds to less reactive and more soluble substances, by facilitating their conjugation with glutathione and assisting to their discharge from the cells, via membrane-based glutathione conjugate pumps 35 . Several studies reviewed by Gullner et al. 36 , have shown that GSTs play a crucial role in plant-pathogen interactions, by suppressing ROS accumulation in the infected tissues of the host. In detail, it has been shown that infection of plant tissue by necrotrophic fungi leads to an up-regulation of GSTs that in subsequence, enhance the antioxidant plant system contributing to disease resistance to necrotrophic pathogens such as Botrytis cinerea, Alternaria brassisicola or Sclerotinia sclerotiorum 36,37 .
The present proteomic study also identified several proteins, such ascystathionine-β-synthase (CBS) domains containing protein (CDCP), S-formylglutathione hydrolase, subtilisin-like protease,4-hydroxy-3-methylbut-2-en-1-yl diphosphate synthase, inosine-5′-monophosphate dehydrogenase, peptidylprolyl isomerase, peptidylprolyl isomerase and dynamin-related protein that are specifically decreased by brown rot in the relatively susceptible 'RL' fruit. Particularly, the observation that CDCPs abundance was decreased in 'RL' fruits inoculated with both M. fructicola and M. laxa, which, given the role of this enzyme in the regulation of redox homeostasis under pathogenic infection [38][39][40] , prompts the question of whether this protein has an active function as a regulator of pathogen sensitivity in stressed plants. Our proteomic analysis also provide evidence that S-formylglutathione hydrolase (SFGH) may be involved in fruit response to pathogen attack since we observed that this protein was down-regulated in 'RL' fruits inoculated with M. laxa. SFGH is a glutathione thiol esterase that hydrolyzes S-formylglutathione to glutathione and formate and participates in the formaldehyde detoxification pathway in plants 41 . However, we cannot speculate on its role in resistance to brown rot as there is little information regarding this protein. It is widely recognized that pathogen attack promotes the site-specific generation of reactive oxygen species (ROS) that can exacerbate the oxidative stress at the local site 42 . One of the consequences of the oxidative stress is lipid per-oxidation that might generate formaldehyde and other reactive lipid peroxidation products that can be provoke oxidative damage to peach fruit 27 . We speculate that these differences in the SFGH abundance between cultivars reflect the differences in the ROS-originated oxidative damage contribution to brown rot sensitivity.
This study also documents that Monilinia species increased few proteins in 'RL' which are not affected in the 'RG' fruit, and involve, formate dehydrogenase (FDH), UDP-glycosyltransferase (UGT) and 1-aminocyclopropane-1-carboxylate. In this sense, FDH abundance was increased in the artificially-inoculated 'RL' fruit compared to the non-inoculated fruit, suggesting that it might be involved in fruit responses to brown rot agents.
In higher plants formate dehydrogenase is a mitochondrial, NAD-dependent enzyme that catalyzes the oxidation of formate to carbon dioxide 43 . Formate results from photorespiratory glyoxylate and play a metabolic role as a signal in response to different stressful conditions 44 . Formate dehydrogenase is induced by several abiotic stresses such as hypoxia, wounding, chilling and heat 43 . There is convincing evidence that FDH may be involved in host-pathogen interactions 45,46 ; however, the role of FDHs in defense response to pathogen attack remain unclear. It has been reported that FDH1 participates in the cell death signaling pathway, defense-related hormone and gene regulation, leading to hypersensitive cell death and defense response against pathogens 47 . Particularly, in peach fruits it was evidenced that formate may derive from the conversion of methionine to ethylene as it was expressed in high levels during fruit ripening, coinciding with the start of the climacteric period [48][49][50] . Therefore, FDH1 may regulate diverse peach fruit defense responses and ripening syndrome.
Glycosylation plays an important role in the phenylpropanoids by enhancing their solubility and stability, facilitating their storage and accumulation in living cells, and regulating their chemical properties and bioactivities 51 . Glycosylation is regulated by UDP-glycosyltransferase, which catalyzes the transfer of a saccharide moiety from an activated glycosyl donor to a nucleophilic glycosyl acceptor molecule, establishing natural glysodic linkages 52 . Several lines of evidence show that phenylpropanoid compounds (i.e. chlorogenic acid) play a crucial role in peach resistance to Monilinia spp. 20,21,53 while studies of UDP-glycosyl transferase in pathogen plant-pathogen interactions have revealed its function in disease resistance 54,55 . Despite the increase in UDP-glycotransferase abundance and expression in the tissues of the 'RL' fruit in the presence of M. laxa the fruit of this cultivar remained extremely susceptible to the pathogen, suggesting that other factors contribute to this increased susceptibility, remaining to be elucidated.
The abundance of the 1-aminocyclopropane-1-carboxylate deaminase was increased in fruits of 'RL' inoculated with M. laxa. Recent studies have shown that on climacteric fruits such as kiwifruit or apple, it degrades 1-aminocyclopropane-1-carboxylate (ACC) 56,57 . ACC constitutes the immediate precursor of ethylene and its degradation leads to a reduction in ethylene biosynthesis in these fruits. It is well established that ethylene plays a crucial role in the regulation of fruit defense against microbial pathogens 58 . Depending on the pathogen type (necrotrophic vs biotrophic pathogens) and the plant species, ethylene can promote disease development acting as a virulence factor, while, in other cases, ethylene modulates resistance responses acting as a signaling compound for resistance to the pathogens 59,60 . In several climacteric fruits has been shown that reduced ethylene production leads to an increase of their susceptibility to necrotrophic pathogens due to ethylene involvement into the regulation of defense-associated genes 60,61 . However, in a recent study aiming to explore the role of ethylene during Monilinia spp. infection has been proposed that brown rot agents try to suppress ethylene biosynthesis aiming to an inhibition of fruit defense reactions 62 . The different level of 1-aminocyclopropane-1-carboxylate deaminase in the present work might be due to the spatial and temporal action of ethylene that was produced both during fruit-pathogen interactions and peach fruit ripening process. An alternate action, that needs to be considered, is whether ethylene directly affects the growth of Monilinia spp. as speculated by Sharon et al. 63 for B. cinerea in other types of fruits. A previous report suggested that applications of exogenous ethylene did not affect the development of M. fructicola on peach, nectarine and plum fruit 64 . However, it has not been known yet what is the effect of endogenous ethylene produced within the host tissues. Thus, further research is required to highlight the precise role of the 1-aminocyclopropane-1-carboxylate deaminase, notably the ethylene metabolism, as a mechanism to account for fruit responses to Monilinia spp.
This study provides the first information concerning the peach fruit proteomic response to infection by brown rot pathogens such as M. fructicola and M. laxa. It is also illustrated the differential accumulation of proteins in response to both M. fructicola and M. laxa in peach cultivars with different resistance levels against brown rot. The significant up-accumulation of several proteins associated with defense in the disease-resistance cultivar and the down-accumulation of the defense proteins in the susceptible cultivar, may explain their differences in disease resistance levels. Further research should be conducted to determine their precise role and relationship of these proteins in peach fruit defense against brown rot, thereby enabling our ability to control this disorder using knowledge-based new peach breeding programs. Artificial inoculations. Two isolates of M. fructicola (Mf 101, Mf 160) and two isolates of M. laxa (Ml 1, Ml 66) were used for the artificial inoculations. Isolates were recovered from infected peaches of commercial orchards in Imathia region and identified to species level with a PCR assay developed by Hily et al. 65 . Artificial inoculation was conducted as described by Papavasileiou et al. 3 . Briefly, for conidia production, isolates were cultured in Petri dishes on V8 agar juice nutrient medium for 7 days at 25 °C, under continuous light. Five ml of sterile and distilled water were then added in each Petri dish, while the conidia were scraped off the cultures with a surgical scalpel. The conidial suspension was then filtered twice through a cheesecloth to withhold mycelia and adjusted to a concentration of 10 5 spores ml −1 with a Neubauer Haemotocytometer. Before the inoculation, fruit were disinfected by dipping them in a 10% sodium chloride solution for 1 min and then rinsed twice using distilled water and let to dry. Fruit were then placed in sterilized plastic containers lined with wet paper towels. The inoculation was carried out by pipetting a drop of 20 μl spore solution on the un-wounded surface of each fruit. In total 15 replicate fruits (3 × 5 fruit, in 3 independent replications) of each peach cultivar were inoculated with each isolate. The containers were then sealed airtight with a plastic lid to ensure 100% relative humidity necessary for the infection. The fruit were incubated for 6 days at 22 °C. After the incubation period the disease incidence was determined by measuring the number of fruit showing typical brown rot symptoms. The observed frequencies (%) of diseased fruits of the 2 cultivars were compared using the chi-square test. The statistical analysis tests were performed using SPSS Statistics (version 11.0; IBM, NY). plant material utilized for proteomic analysis. The experimental design included the following fruit classes: (i) fruit of the two cultivars artificially inoculated with M. fructicola isolates, (ii) fruit of the two cultivars artificially inoculated with M. laxa isolates, iii) fruit of the two cultivars inoculated with sterile water (control). For the proteomic analysis, samples were taken from the mesocarp of the artificially inoculated and control fruit after 6 days of incubation at 22 °C. Approximately 50-70 gr of skinless flesh of the mesocarp were carefully received from each fruit at 1 cm distance from the rotten tissue, to avoid the presence of mycelia in the proteomic analysis. Samples were cut into small pieces, placed in polyethylene bags (3 fruits per bag), frozen instantly in liquid nitrogen and stored at −80 °C until further analysis. Totally, 15 fruits (3 × 5 fruit in each replicate) were utilized for the analysis. protein extraction. The extraction was performed according to a previous report 50 following phenol-based extraction protocol. After pellet complete drying the proteins were solubilized in buffer containing 4% CHAPS (3-[(3-Cholamidopropyl) dimethylammonio]-1-propanesulfonate), 7 M urea, 2 M thiourea, 20 mM DTT, 0.5% ampholyte pH 3-10, 2% Triton X-100 and bromophenol blue. Protein concentrations were measured according to Bradford 66 using bovine serum albumin (BSA; Sigma) as a standard. Two-dimensional gel electrophoresis and protein quantification. Proteins (50 μg) were separated by 2D-PAGE as described by Minas et al. 44 . Isoelectric focusing was run onto immobilized pH gradient gel strips (IPG strip pH 3-10 NL, 11 cm; Biorad) and then SDS-PAGE onto 12.5% polyacrylamide gels (Criterion Tris-HCl Precast Gels-Biorad). For each treatment, 2-D gels were run in triplicate and for three independent extractions. Following silver-nitrate stained 2-D gels, were scanned with a Bio-Rad GS-800 Calibrated Densitometer equipped with PDQuest Advanced 2-D Gel Analysis Software. Spots were detected, background subtracted, matched and quantitative determination of the spot volumes was performed (mode: total quantity of valid spots normalization) as reported earlier 44 . Individual means were compared using Student's t-test (significance level 95%). To validate significant differences, Student's t-test was further combined by the quantitative 2-fold change of spot volume.

Methods
Protein identification by tryptic in-gel digestion, mass spectrometry and database searching. Cut silver-stained spots were destained and processed using classical tryptic-mediated in-gel digestion 67 .
The generated peptide mixture was solubilized in 10 μl solution A (2% acetonitrile in 0.1% formic acid) and pre-concentrated with a flow of 5 μl/min for 10 min on a C18 trap column (Acclaim PepMap, Thermo Scientific) and then loaded onto a 50 cm long C18 column (75 μm ID, particle size 2 μm, 100 Å, Acclaim PepMap RSLC, Thermo Scientific). The binary pumps of the HPLC (Ultimate 3000, RSLCnano, Thermo Scientific) consisted of solution A (2% acetonitrile in 0.1% formic acid) and solution B (80% acetonitrile in 0.1% formic acid). The peptides were separated using a linear gradient from 4% B up to 40% in 40 min for a 1-h gradient run with a flow rate of 300 nl/min. The column was placed in an oven operating at 35 °C. The HPLC was coupled online to a LTQ Orbitrap XL Mass spectrometer (Thermo Scientific). Full scan MS spectra were acquired in the orbitrap (m/z 350-2000) in a profile mode and data-dependent acquisition with the resolution set to 60,000 at m/z 400 and automatic gain control target at 10 6 . The six most intense parental ions were sequentially isolated for collision-induced MS/MS fragmentation (CID) and their daughter fragments were detected in the linear ion trap. Dynamic exclusion was set to 1 minute and activated for 90 s. Ions with single charge states were excluded. Lock mass of m/z 445,120025 was used for internal calibration. The Xcalibur software (Thermo Scientific) was used to control the system and acquire the raw files. The raw files were analyzed with Proteome discoverer 1.4 (Thermo Scientific) using the complete Uniprot database of Prunus persica (3760, 28650 entries). Search parameters were strict trypsin specificity, allowing up to two missed cleavage sites. Oxidation of methionines, cysteine caramidomethylation as well as deamidation of asparagines and glutamines were set as variable modifications. The false discovery rate (FDR) was set to 5%. In case the Prunus persica protein identified was not yet annotated its sequence was processed through a BLAST-mediated search against current plant databases in order to find homologous and annotated proteins suggesting its potential function.