Comparative proteomic analysis provides insights into the complex responses to Pseudoperonospora cubensis infection of cucumber (Cucumis sativus L.)

Cucumber (Cucumis sativus L.) is an important crop distributed in many countries. Downy mildew (DM) caused by the obligate oomycete Pseudoperonospora cubensis is especially destructive in cucumber production. So far, few studies on the changes in proteomes during the P. cubensis infection have been performed. In the present study, the proteomes of DM-resistant variety ‘ZJ’ and DM-susceptible variety ‘SDG’ under the P. cubensis infection were investigated. In total, 6400 peptides were identified, 5629 of which were quantified. KEGG analysis showed that a number of metabolic pathways were significantly altered under P. cubensis infection, such as terpenoid backbone biosynthesis, and selenocompound metabolism in ZJ, and starch and sucrose metabolism in SDG. For terpenoid backbone synthesis, 1-deoxy-D-xylulose-5-phosphate synthase, 1-deoxy-D-xylulose 5-phosphate reductoisomerase, 2-C-methyl-D-erythritol 2,4-cyclodiphosphate synthase, 4-hydroxy-3-methylbut-2-en-1-yl diphosphate synthase, and geranylgeranyl pyrophosphate synthase were significantly accumulated in ZJ rather than in SDG, suggesting that pathogen-induced terpenoids accumulation might play an important role in the resistance against P. cubensis infection. Furthermore, a number of pathogenesis-related proteins, such as endochitinases, peroxidases, PR proteins and heat shock proteins were identified as DAPs, suggesting that DM resistance was controlled by a complex network. Our data allowed us to identify and screen more potential proteins related to the DM resistance.


Materials and Methods
Cucumber materials and sampling. A DM-resistant cultivated cucumber variety 'ZJ' and a DM-susceptible cultivated variety 'SDG' were used. Cucumber seedlings were planted in a growth chamber with photoperiod of 12 h light/12 h dark, relative humidity of 60%, and light intensity of 120 μmol m −2 s −1 . A solution containing P. cubensis sporangia (2 × 10 6 sporangia/mL), 5 mM glucose, and 2.5 mM KH 2 PO 4 was prepared as the inoculant. The same solution without pathogen was used as negative control. For inoculation, the third leaves of seedlings at the three-leaf stage were sprayed with P. cubensis sporangia or control solution. All inoculated seedlings (four groups × three replicates) were kept in the same condition and separately covered with plastic films. In total, 10 seedlings were contained in each group. The third leaves of seedlings at the three-leaf stage were harvested for protein isolation. Plant samples for each group were harvested at 48 h post inoculation and cleaned with deionized H 2 O, and then were immediately frozen in liquid N 2 until use. To get a comprehensive understanding of the proteins involved in the DM resistance, four sample groups, including DM-resistant variety treated with P. cubensis sporangia (ZJT), DM-resistant variety treated with control solution (ZJC), DM-susceptible variety treated with P. cubensis sporangia (SDGT), and DM-susceptible variety treated with control solution (SDGC), were used for comparative proteomic analysis. protein extraction. About 500 mg plant samples for each replicate were grinded by liquid N 2 into cell powder. Four volumes of pre-cooled lysis buffer containing urea (8 M), Triton-100 (1%), dithiothreitol (10 mM), and protease inhibitor (1%) was added, followed by sonication five times using an ultrasonic processor (Scientz, Ninbo, China) on ice. The remaining debris was discarded by centrifugation at 12,000 × g at 4 °C for 10 min. At last, the peptide samples were precipitated with 20% of cold TCA buffer for 3 h at −20 °C. The supernatant was removed by 12,000 × g centrifugation at 4 °C for 15 min. The remaining precipitate was cleaned with pre-cooled acetone for three times and was re-dissolved in urea solution (8 M). The proteins were quantified by a GE Healthcare 2-D Quant kit (Beijing, China). trypsin digestion. Sample solution was reduced with dithiothreitol (5 mM) for 30 min and alkylated with iodoacetamide (11 mM) for 15 min at 25°C in darkness. Then, TEAB (100 mM) was added to dilute the protein solution to concentration of 2 M. For trypsin digestion, trypsin was added at trypsin/protein mass ratio of 1:50 for the first overnight digestion and at mass ratio of 1:100 for the second 4 h digestion.
TMT/iTRAQ labeling. Tryptic peptide sample was desalted by a Strata X-C18 SPE column (Phenomenex, Torrance, USA). Then, the dried peptide samples were re-constituted in TEAB solution (0.5 M) and manipulated using TMT/TRAQ kit (Thermo-Scientific, Shanghai, China) according to the its protocol. One unit of TMT reagent was dissolved and reconstituted in acetonitrile. The resulting solution was incubated for 2 h at 25 °C. At last, the result peptides were pooled, desalted and dried.
Fractionation and LC-Ms/Ms analysis. The column was wetted with 0.1% formic acid in acetonitrile and conditioned with ddH 2 O. The labeled samples were loaded onto an Agilent 300Extend C18 column (5 μm particles, 4.6 mm ID, 250 mm length, Santa Clara, CA, USA) for high pH reverse-phase HPLC. A solution, containing 0.1% formic acid in 98% acetonitrile, at a flow rate of 400 nL/min was used for the EASY-nLC 1000 UPLC system (Thermo Scientific, Beijing, China). The gradient was increasing from 6% to 25% in 26 min, from 25% to 35% in 8 min, climbing to 80% in 3 min, and holding at 80% for 5 min.
The electrospray voltage was set at 2000 V. The range of m/z full scan was set from 350 to 1800. Intact peptides were identified in the Orbitrap at a high resolution of 70,000 and the fragments were identified in the Orbitrap at a low resolution of 17,500. Automatic gain control was set at 5E4. The first mass was fixed at 100 m/z. The MS proteomics data were deposited to the Proteome EXchange Consortium via the PRIDE partner repository with the identifier PXD010844.
www.nature.com/scientificreports www.nature.com/scientificreports/ Quantification of identified proteins. The ratios of the TMT reporter ion intensities in MS/MS spectra (m/z 126-131) were applied to compute changes in each protein between different samples. Data of each protein was mean-normalized to center the distribution of quantitative values. Then, protein quantitation was computed as the median ratio of corresponding peptides for a unique protein. Two-sided t-tests were applied to analyze the expression differences of proteins from two sample groups. A significance level at 0.05 was used for statistical testing.
searching and annotation of protein. The MS/MS data were searched against Cucumber Genome with an integrated MaxQuant engine (v.1.5.2.8) with default parameters. The mass tolerance of precursor ions was allowing up to 20 ppm in the first round searching and up to 5 ppm in the main searching. The other parameters were set according to the previous published work 22 .
Gene Ontology (GO) annotation of each protein was derived from the UniProt-GOA database. Firstly, all identified protein were uploaded onto UniProt database to get their IDs and then were searched against the GO database. Domain functional description of identified protein were annotated by the InterPro database. For Kyoto Encyclopedia of Genes and Genomes (KEGG) annotation, KEGG online service tool 'KAAS' was used to annotate the identified proteins. Then, the annotated proteins were searched against the KEGG database using another online tool 'KEGG mapper' .
The cells of eukaryotic organisms are elaborately subdivided into functionally distinct membrane bound compartments. In our study, Wolfpsort (http://wolfpsort.seq.cbrc.jp/), a subcellular localization predication software, was used to predict subcellular localization.
For enrichment analysis, a two-tailed Fisher's exact test was applied to analyze the GO, KEGG and domain enrichments of the differential abundance proteins (DAPs) against all the identified proteins. Correction for multiple hypothesis testing was performed using standard FDR control method 23 .

Validation of DAPs by parallel reaction monitoring (PRM).
To validate the DAPs under the P. cubensis infection, the changes in several key proteins were identified by an acquired MS/MS spectrum. Three independent biological replicates were used in this experiment. The peptides were treated with trypsin and the tryptic peptides were dissolved in 0.1% FA solution. The resulting peptide samples were loaded onto a 150 mm length, 75 μm, reversed-phase analytical column (Thermo, Shanghai, China). The gradient was comprised of an increase from 7% to 25% solvent B (0.1% formic acid in 98% acetonitrile) over 38 min, 25% to 36% in 14 min and climbing to 80% in 4 min then holding at 80% for the last 4 min, all at a constant flow rate of 700 nL/min on an EASY-nLC 1000 UPLC system.
The resulting MS data were processed using Skyline (v.3.6). Peptide settings: enzyme was set as Trypsin [KR/P], Max missed cleavage set as 2. The peptide length was set as 8-25, Variable modification was set as Carbamidomethyl on Cys and oxidation on Met, and max variable modifications was set as 3. Transition settings: precursor charges were set as 2, 3, ion charges were set as 1, 2, ion types were set as b, y, p. The product ions were set as from ion 3 to last ion, the ion match tolerance was set as 0.02 Da. statistical analysis method. Significant analysis was performed using a one-way analysis of variance with a Tukey's test. All values are showed using the averages of three replicates, and data are showed as the mean with the standard deviation.

Results
overview of Ms data. The proteomes of two cucumber varieties under the control and P. cubensis infection conditions were analyzed in our study. Correlation coefficients of 12 samples (three replicates × four groups) showed a well repeatability of our data (Fig. S1a). Quality validation showed that distribution of mass errors was lower than 0.02 Da and lengths of peptides varied from 7 to 20 amino acid residues, indicating a good quality of our sample preparation (Fig. S1b,c).

Impacts of the P. cubensis infection on proteome levels in cucumber.
To compare the DAPs between different sample groups, expression profiles of the quantified proteins were showed in a heatmap (Fig. 1a). To reflect the changing trends among different sample groups, all of the quantified proteins were assigned into six clusters using MeV software by K-means method. The proteins in the cluster I showed the highest levels in the ZJT and SDGC samples; the proteins in the cluster II highly accumulated in the ZJT sample; the proteins in the cluster III predominantly accumulated in the ZJC sample; the proteins in the cluster IV showed highest levels in SDG; the proteins in the cluster V showed the highest levels in ZJ; and the proteins in the cluster VI predominantly accumulated in the SDGC sample (Fig. 1b).
Protein annotation and classification. In our study, 6400 proteins were identified, 5629 of which were quantified. The detail information, including protein IDs, GO, KEGG and domain categories, subcellular localizations, and functional enrichments, of all identified proteins were listed Table S1.
DAPs related to terpenoid backbone biosynthesis. KEGG enrichment analysis showed that the 'terpenoid backbone biosynthesis' pathway was significantly enriched in the ZJT vs. ZJC comparison rather than in the SDGT vs. SDGC comparison. Our study identified most of the proteins participated in the essential steps of methylerythritol 4-phosphate (MEP) and mevalonate (MVA) pathways, which provide the precursors for terpenoid backbone biosynthesis 24 . In total, six key enzymes in the MEP pathway and our key enzymes in the MVA pathway were identified (Fig. 5a).
In the ZJ seedlings, the accumulation levels of DXS, DXR, MDS, HDS and GGPPS were up-regulated by the P. cubensis infection. In the SDG seedlings, only PMK was down-regulated by the P. cubensis infection (Fig. 5b). Under the control condition, no significantly changes in the accumulation levels of proteins related to terpenoid backbone biosynthesis were observed. After the P. cubensis infection, five enzymes, including PMK, DXS, DXR, DMK, and GGPPS, were significantly accumulated in ZJ than in SDG, and only IPPI_2 was significantly accumulated in SDG than in ZJ (Fig. 5c).

Differential accumulation of pathogenesis-related proteins.
Our data annotated a number of pathogenesis-related proteins, among which three endochitinases, one endoglucanase, seven peroxidases, one pathogenesis-related protein P2, and one lipid transfer protein were identified as DAPs under the P. cubensis infection ( Table 1). The protein levels have been analyzed to reveal the effects of P. cubensis infection on these pathogenesis-related proteins. Under the P. cubensis infection, three endochitinases were up-regulated in the ZJ seedlings and only two endochitinases were up-regulated in the SDG seedlings. The accumulation level of endoglucanase was reduced in the ZJ seedlings and no significantly change was observed in the SDG seedlings. For peroxidases, all of them were induced in the ZJ seedlings and only two were induced in the SDG seedlings under the P. cubensis infection. Pathogenesis-related P2 protein was up-regulated in both of the ZJ and SDG seedlings under the P. cubensis infection. Accumulation level of lipid transfer protein 1 was reduced in the ZJ seedlings and no significantly change was observed in the SDG seedlings.
A large number of HSPs were identified by the proteomes. In total, three HSPs were significantly induced and three HSPs were significantly reduced in the ZJ seedlings under the P. cubensis infection. In the SDG seedlings, 11 HSPs were significantly induced and only one HSP was reduced under the P. cubensis infection ( Table 2).

Discussion
DM is a major cucurbit foliar disease that defoliates cucumber crops and results to great economic losses 25 . Increasing works on the mechanisms of pathogen resistance explored many defense-related genes and proteins that might play a role in the DM resistance 26 . Therefore, more and more genetic resources, including potential genes and proteins associated with the DM resistance, must be explored to enhance the genetic improvement and resistance breeding of cucumber. www.nature.com/scientificreports www.nature.com/scientificreports/ Previous proteomic analyses of cucumber have been performed using traditional methods, such as 2-D, which could detect a limited number of proteins 27 . For example, 71 differential protein spots were confidently identified by Du et al. 16 , 28 Trichoderma asperellum induced proteins were identified in cucumber root colonization by Segarra et al. 28 , 21 protein spots in Cd-treated cucumber were identified by Sun et al. 29 , 33 salt stress responsive protein spots were successfully identified by Shao et al. 30 , 63 DAPs in NaCl treated roots of cucumber were identified by Yuan et al. 16 , and 22 up-and 12 down-regulated proteins in cucumber roots under hypoxic stress were identified by Li et al. 31 . In our study, 6400 proteins were identified, which was far more than the protein numbers in the previous studies. A large number of identified proteins gave us an opportunity to reach a deeper analysis of proteins responsive to the P. cubensis infection.
Recently, transcriptomes of cucumber have identified a number of genes involved in response to the P. cubensis infection. In Li's study, 58 unique ESTs belonging to several categories involved in plant defense, such as signal transduction, cell defense, cell cycle, protein binding and metabolism, were identified 5 . In our study, DAPs involving "defense response" and "protein binding" were identified in SDG, suggesting that the genes belonging to above two categories were controlled at both the transcriptional and posttranscriptional levels. In Adhikari's study, a number of defense related genes, including catalases, chitinases, lipoxygenases, and peroxidases, were rapidly induced by P. cubensis infection 20 . In our study, more defense related protein showed differential accumulation in ZJ (2 chitinases, 1 lipoxygenase and 5 peroxidases) than in SDG (1 chitinase and 1 peroxidase) ( Table 1), suggesting they were also controlled at both the transcriptional and posttranscriptional levels.
Comparative proteomic analysis identified 235 DAPs in the DM-resistant variety ZJ and 53 DAPs in the DM-susceptible variety SDG under the P. cubensis infection. Interestingly, at the early stage of P. cubensis infection, the number of differential expressed genes in the resistant line 'PI 197088' (4864 genes) was larger than that in the susceptible line 'Vlaspik' (1969 genes) 21 . The number of DAPs in the ZJ was larger than that in SDG, suggesting that a large-scale changes in proteomes occurred in resistant variety than susceptible variety. www.nature.com/scientificreports www.nature.com/scientificreports/ The Ribosome-related proteins should be involved in the translational control in plants 32 . Glutathione-related proteins are known to be involved in redox control in response to a variety of stresses including pathogen infection 33 . For example, several glutathione S-transferase genes accumulated in the Botryosphaeria dothidea-infected Populus and were potentially related to pathogen resistance 34 . Interestingly, In our study, a number of DAPs in the ZJT vs. ZJC comparison were enriched in the "Ribosome" and "Glutathione metabolism" pathways, suggesting that translational control and redox control played an important role in the resistance against P. cubensis infection.
Terpenoids, the most diverse class of chemicals produced by plants, are involved in the protection in various abiotic environments, particularly in pathogenic fungi infection 35 . For example, terpenoids were predicted to be important defense metabolites in Eucalyptus froggattii seedlings 36 . Arbuscular mycorrhizal fungi-plant interaction induces the accumulation of terpenoids in plants 37 . Diterpenoid core is derived by isoprenoid precursor isopentenyl diphosphate (IPP) and dimethylallyl diphosphate (DMAPP), which are the downstream metabolites of the MEP and MVA pathways 24 . Under the P. cubensis infection, five key enzymes, including DXS, DXR, MDS, HDS, and GGPPS, were significantly up-regulated in ZJ, indicating an accumulation of terpenoids the DM-resistance variety ZJ. In Norway spruce, three DXS enzymes were differentially expressed under MeJA treatment, suggesting an important role of DXS in the induced defense response 38 . Over-expression of a GGPPS in white spruce saplings enhanced the accumulation of terpenoids and played a role in plant defense 39 . In cucumber, pathogen-induced terpenoids accumulation might play an important role in the resistance against P. cubensis infection.
A number of pathogenesis-related proteins, such as glucanases, chitinases, and peroxidases, were reported to be involved in plant defense 40 . Endo-glucanases, which accelerate the degradation of fungal cell wall, are induced by pathogen infection 41 . Endo-chitinases, another important class of the PR family, play a role in the hydrolysis of the major component of fungal cell wall, chitin 42 . In cucumber, three identified endochitinases were significantly up-regulated in ZJ and only one was significantly up-regulated in SDG. More glucanases were induced in the resistant cultivated variety than in the susceptible cultivated variety, suggesting that degradation of fungal cell walls might play a role in the resistance to pathogen infection 43 . Furthermore, uncontrolled reactive oxygen species (ROS) accumulation occurs under pathogen infection and results to the susceptibility of plants to fungal pathogens 44 . Peroxidases play an essential role in scavenging ROS in plant cells under pathogen infections 45 . In our study, all identified peroxidases were siginificantly up-regulated in ZJ and only two peroxidases were significantly up-regulated in SDG. Accumulation of peroxidases might contribute to the DM resistance of ZJ.
HSPs are quickly up-regulated by heat stress, and increasing studies showed a close relationship between HSPs and disease resistance 46 . For example, HSP90 plays a role in regulating the structure and stability of R proteins in plant defense responses 47 . HSP90 associated chaperoning is essential for the defense responses to rice blast fungus 48 . In addition to HSP90, HSP70 is also essential for plant resistance to pathogen infections 49 . HSP70 was involved in the resistance to P. chicorii infection in tobacco and resistance to oomycete pathogens in Arabidopsis 50 . In our study, a number of HSPs were identified in both two cucumber varieties. Most of the identified HSPs were significantly up-regulated by the P. cubensis infection in SDG rather than in ZJ, suggesting strong communications between heat shock responses and susceptibility to DM disease.
In conclusion, proteomes of DM resistant variety ZJ and DM susceptible variety SDG under the control and P. cubensis infection conditions were obtained. In total, 6400 proteins were identified using a MS/MS-based integrating TMT analysis strategy. Furthermore, 236 DAPs and 53 DAPs were identified in ZJ and SDG under the P. cubensis infection, respectively. A number of terpenoid backbone synthesis enzymes, pathogenesis-related proteins, and HSPs showed significantly differences between ZJ and SDG under the P. cubensis infection, suggesting that DM resistance was conducted by a complex network.