MSE for Global Analysis of

Absolute Protein Quantification by LC/MSE for Global Analysis of Salicylic Acid-Induced Plant Protein Secretion Responses Fang-yi Cheng,† Kevin Blackburn,‡ Yu-min Lin,§ Michael B. Goshe,*,‡ and John D. Williamson*,† Department of Horticultural Science, North Carolina State University, Raleigh, North Carolina 27695-7609, Department of Molecular and Structural Biochemistry, North Carolina State University, Raleigh, North Carolina 27695-7622, and Department of Industrial Engineering and Operations Research Program, North Carolina State University, Raleigh, North Carolina 27695-7906 Received August 15, 2008

The plant cell wall is a dynamic cellular compartment consisting of a complex matrix of components that can change dramatically in response to environmental stresses. During pathogen attack, for instance, a wide spectrum of proteins that participate in various sequential processes involved in plant defense is secreted into the cell wall. In this study, a mass spectrometry, data-independent acquisition approach known as LC/MSE was used to assess temporal changes in the cell wall proteome in response to different levels of an endogenous inducer of plant disease defense responses, salicylic acid (SA). LC/MSE was used as a label-free method that enabled simultaneous protein identification and absolute femtomole quantification of each protein secreted into the extracellular matrix. A total of 74 secreted proteins were identified, 63 of which showed increased specific secretion in response to SA. A majority of this induced secretion occurred within 2 h of treatment, indicating that many proteins are involved in the early stages of plant defenses. We also identified a number of apparently nonclassically secreted proteins, suggesting that, as in many nonplant systems, Golgi/ER-independent mechanisms exist for plant protein secretion. These results provide new insight into plant apoplastic defense mechanisms and demonstrate that LC/MSE is a powerful tool for obtaining both relative and absolute proteomescale quantification that can be applied to complex, time- and dose-dependent experimental designs. Keywords: Arabidopsis • secretome • liquid chromatography • mass spectrometry • MSE • dataindependent acquisition • salicylic acid • pathogen • nonclassical protein secretion

Introduction The plant extracellular matrix is a dynamic structure primarily composed of polysaccharides, phenolic compounds, proteins and other components and whose plasticity is essential for cell division, expansion and differentiation. Since the cell wall provides structural support and represents the outermost barrier of the cell, it also plays an important role in plant resistance to biotic and abiotic stresses. In response to pathogen attack, plant cells secrete a wide range of defense-related proteins into the cell wall, including antimicrobial peptides, detoxifying enzymes and pathogenesis-related proteins of presently unknown function.1-3 Despite the importance of this dynamic compartment, the proteome of the plant cell wall, or * To whom correspondence should be addressed. Dr. John D. Williamson, Department of Horticultural Science, North Carolina State University, Raleigh, NC 27695-7609. Phone: 919.515.5366. Fax: 919.515.2047. E-mail: [email protected]. Dr. Michael B. Goshe, Department of Molecular and Structural Biochemistry, North Carolina State University, Raleigh, NC 27695-7622. Phone: 919.513.7740. Fax: 919.515.2505. E-mail: [email protected]. † Department of Horticultural Science, North Carolina State University. ‡ Department of Molecular and Structural Biochemistry, North Carolina State University. § Department of Industrial Engineering and Operations Research Program, North Carolina State University.

82 Journal of Proteome Research 2009, 8, 82–93 Published on Web 11/11/2008

secretome, is considerably less characterized than that of the cytosol. Thus, investigation of secretome dynamics during pathogen attack will facilitate our understanding of the complex protein interaction network that regulates plant defense mechanisms in this unique compartment. The conventional approach to proteome analysis uses gel electrophoresis coupled with mass spectrometry. Protein mixtures are first separated by two-dimensional (2D) gel electrophoresis. Next, visualized protein spots of interest are individually excised from the gel, digested with trypsin and identified by mass spectrometric analyses. In this approach, relative quantification of individual proteins between samples is based on the staining intensity of protein spots on the 2D gel. Newer gel-based techniques, such as differential in-gel electrophoresis (DIGE),4 that use differential fluorescent dye tagging to allow visualization of multiple samples on a single gel, have enhanced the accuracy of relative protein quantification. Although these gel-based approaches are widely used, the entire process from gel electrophoresis, staining, protein spot selection to protein identification can be labor intensive and difficult to reproduce quantitatively. In addition, a given stained spot is often composed of several proteins, so a change in spot intensity cannot be unequivocally attributed to a single protein. 10.1021/pr800649s CCC: $40.75

 2009 American Chemical Society

research articles

Global Analysis of SA-Induced Plant Protein Secretion Responses Because of the limitations of gel-based methods, several liquid chromatography (LC)-based quantitative approaches have been developed, such as ICAT (isotope-coded affinity tags),5 SILAC (stable isotope labeling by amino acids in cell culture)6 and enzyme mediated stable isotope labeling.7 These quantification strategies are based on incorporation of stable isotopes into proteins or peptides by direct chemical or enzymatic modification or by in vivo labeling.8-10 Relative protein abundances between two or more treatments can then be determined by comparing the intensities of heavy isotopecoded peptides and their light isotope-coded counterparts in a combined sample. Isotope labeling, however, is not only expensive, but generally works only for pairwise comparisons. Recently introduced multiplexed isotope tagging reagents such as iTRAQ11 do allow the comparison of as many as 8 distinct samples due to unique reporter product ions produced by collision-induced dissociation (CID) of an isobaric tag. However, a reference sample must still be mixed with every other sample when comparing more than 8 treatment groups. These chemical labeling techniques require larger sample amounts and additional sample handling and manipulation which can lead to protein losses and quantification bias. Recently, label-free LC/MS quantitative techniques have emerged as an attractive alternative to isotope coding, because they do not require special labeling chemistries or growth conditions and are amenable to complex study designs with any number of treatment groups.12-16 This approach is predicated on the observation that electrospray ionization provides peak signal intensities that are linearly proportional to the analyte concentration.17,18 Unlike isotope labeling approaches, where control and treated samples are pooled for analysis, unlabeled protein samples are analyzed separately, and peptide peak areas directly compared between runs for relative quantification. As a result, label-free quantification does not require extra preparation for isotope-labeling and can be directly applied to any number of protein samples from virtually any source. More recently, a new variant of label-free quantification known as LC/MSE was introduced for quadrupole time-of-flight (Q-Tof) mass spectrometers. For this method, alternating scans of low collision energy and elevated collision energy during LC/ MS analysis are used to obtain both protein quantification and protein identification data in a single experiment.19-21 The lowenergy scan mode is used to obtain accurate precursor ion mass and intensity data for quantification, while the elevated collision energy mode generates multiplex peptide fragmentation of all peptide precursors with associated accurate mass product ion information for database searching and subsequent protein identification. In addition to relative quantification between samples based on electrospray intensity, absolute quantification (moles protein) for each identified protein is possible using the LC/MSE mode of acquisition.21 This is predicated on the observation that average signal intensity measured by LC/MSE of the three most intense tryptic peptides for any given protein is constant at a given concentration, regardless of protein size. In addition to tracking the relative changes in protein abundance in response to treatment, the ability to derive absolute quantities enables the stoichiometric comparison of proteins within and across samples. Moreover, LC/MSE provides substantial advantages for protein identification over conventional LC/MS/MS approaches. Unlike datadependent LC/MS/MS, where the most abundant precursors in an MS scan are sequentially subjected to MS/MS fragmenta-

E

tion, MS utilizes parallel, multiplex fragmentation where all peptide precursors are simultaneously fragmented throughout the chromatographic separation process regardless of intensity. This allows data-independent identification of lower abundance peptides and provides increased proteome coverage and dynamic range of protein identification compared to datadependent LC/MS/MS. In this study, absolute quantification by LC/MSE was used to perform a comprehensive, quantitative analysis of salicylic acid (SA)-induced changes in the proteome of the plant extracellular space (the secretome). In plants, pathogen attack initiates a complex signaling cascade that ultimately leads to the synthesis of the compound SA. SA in turn induces the expression and/or secretion of a large number of proteins collectively called pathogenesis response (PR) proteins that, acting together, lead to acquisition of pathogen resistance in the plant (Systemic Acquired Resistance). Given SA’s proximal role as an endogenous inducer of pathogen defense responses in plants, direct SA treatment provides a convenient means of assessing pathogen-induced responses. Arabidopsis suspension cultures were thus treated with increasing doses of SA for various periods of time. After treatment, proteins secreted into the culture medium were collected and temporal and dosedependent changes in the secretome were characterized both qualitatively and quantitatively using LC/MSE. Our results provide unique insight into the highly integrated mechanisms utilized by plants and illustrate the complexity underlying plant physiological responses yet to be explored.

Materials and Methods Materials. Acetonitrile (HPLC grade), dithiothreitol (DTT), formic acid (ACS reagent grade), salicylic acid (SA), 4-hydroxybenzoic acid (HBA) and iodoacetamide were from SigmaAldrich. Ammonium bicarbonate was from Fluka (Milwaukee, WI). Sequencing grade-modified trypsin was from Promega (Madison, WI). Water (18 MΩ) for LC/MSE analysis was distilled and purified using a High-Q 103S purification system (Wilmette, IL). All other chemicals were from Fisher Scientific or Sigma-Aldrich and used without further purification. Plant Materials, Growth Conditions, and Treatments. Cell suspension cultures derived from Arabidopsis thaliana ecotype Columbia were maintained on a 7 day transfer cycle in MS medium22 containing 0.5 mg/L 2,4-D (2,4-dichlorophenoxy acetic acid) and 3% sucrose. Cell cultures were shaken at 110 rpm at 25 °C with a 12 h light/12 h dark photoperiod at a light intensity of 5 (µE/m2)/s under mixed fluorescent lights. After 7 day subculture, cells for secretion experiments were washed thoroughly and incubated for the indicated time in MS medium supplemented with 0.5 or 1.0 mM SA. For control treatments, cell medium was supplemented with 1.0 mM of HBA, an analogue of SA that does not induce plant resistance responses.23 Three independent experiments were performed. Each experiment consisted of 12 treatments including three SA levels (control, 0.5 mM SA and 1.0 mM SA) at four time points (1, 2, 6, and 18 h). Preparation of Secreted Proteins. After treatment, suspension cultures were filtered to remove cells from the medium. Following collection, protein quantities secreted into the media in response to different treatments were determined by the method of Bradford,24 and the sample volume to be used in subsequent analysis was determined accordingly. To ensure that proteins in the media were not a result of cell lysis, an enzymatic assay for the presence of the cytosolic enzyme Journal of Proteome Research • Vol. 8, No. 1, 2009 83

research articles

Cheng et al. 25

glucose-6-phosphate dehydrogenase (G6PDH) was performed immediately upon media collection at the completion of a treatment. G6PDH enzyme activities were detected in cell lysates (0.028 units/mg protein), but not in the collected media. Media samples were then concentrated by freeze-drying and resuspension in water. Salts, small peptides, and other watersoluble medium components were then removed by dialysis against 50 mM Tris (pH 7.5) at 4 °C overnight, and final sample volumes adjusted by acetone precipitation and resuspension in 50 mM ammonium bicarbonate (pH 8.3). Because the isolated secretome is a collection of water soluble proteins, MScompatible detergents were not necessary. Each protein sample was further assessed by immunoblot analyses to confirm the absence of the cytosolic marker protein, hexokinase (HK).26 G6PDH activity and HK immunoblot analyses both indicated that cytoplasmic contamination of secretome samples was extremely low. Conversely, immunoblot analyses were also used to verify consistent, experiment-to-experiment induction and secretion of the classical pathogenesis response protein PR1a.27 Finally, visual inspection of Coomassie-stained 1D gels for each sample was used to monitor protein recovery and relative protein quantities secreted. After confirming experiment-to-experiment reproducibility, secretome samples from three replicates from equal amounts of cells for each of the 12 time and dose treatments were pooled. Although perhaps less than ideal statistically, this was deemed necessary to reduce samples to a manageable number without compromising our MS measurements since lack of experimental variation was verified by immunological and biochemical means. Given the biological uniformity of the cell cultures as opposed to individual organisms, this approach is not unprecedented.28 For each pooled sample, the proteins were reduced with 5 mM DTT for 1 h at 37 °C, alkylated with 5 mM iodoacetamide for 1 h at room temperature, and proteolytically digested with trypsin by adding a 1:20 (w/w) trypsin-to-protein ratio and incubating at 37 °C overnight. LC/MSE Analysis. Prior to LC/MSE analysis, each digested protein sample was spiked with a predigested rabbit phosphorylase B internal standard (Waters, Milford, MA) at a level of 50 fmol per 10 µL injection. For sample analysis, 10 µL aliquots of secretome tryptic digests were analyzed in duplicate (2 technical replicates per sample) by LC/MSE using a nanoACQUITY ultrapressure liquid chromatograph (UPLC) and Premier Q-Tof mass spectrometer equipped with a nanolockspray ion source (Waters). Samples were injected online onto a Waters Symmetry C18 trapping cartridge (300 µm i.d. × 1 cm length) at a flow rate of 10 µL/min. Next, peptides were separated by in-line gradient elution onto a 75 µm i.d. × 25 cm column packed with BEH C18 Stationary phase (Waters), 1.7 µm particle size, at a flow rate of 300 nL/min using a linear gradient from 2 to 40% B over 60 min (A ) 0.1% formic acid in water, B ) 0.1% formic acid in acetonitrile). The Q-Tof was operated in the LC/MSE mode of acquisition, where alternating 2 s scans of low (4 V) or high (10-32 V) collision energies are used to generate either intact peptide ions (low energy) or peptide product ions (high energy). Glu-fibrinopeptide at a concentration of 200 fmol/µL (m/z 785.8426) was infused via the nanolockspray ion source at a flow rate of 600 nL/min and sampled every 30 s as the external mass calibrant. Samples were injected as sets based on treatment (control, 0.5 mM SA and 1.0 mM SA) from earliest to latest time points. As protein content in the secretome typically increases over treatment dose and time, queuing samples in this manner resulted in 84

Journal of Proteome Research • Vol. 8, No. 1, 2009

samples with lower protein content being analyzed first. This sample order was designed to minimize protein carryover more effectively between consecutive injections than using randomized sample injection. A protein standard (tryptic digest of rabbit phosphorylase B) analyzed prior to the first sample injection and again following the last sample injection showed no significant loss in instrument sensitivity or performance during the course of the analysis. Data Processing and Database Searching. Each raw data file was processed using ProteinLynx Global Server V2.3 software (Waters) to generate charge state reduced and deisotoped precursor mass lists as well as associated product ion mass lists for subsequent protein identification and quantification. Each processed file was then searched against the TAIR7 protein database obtained from www.arabidopsis.org using the IDENTITYE database search algorithm within PLGS 2.3.29 Prior to searching, the internal standard rabbit phosphorylase B sequence was added to the database, and redundant entries (identical sequences reported more than once) were removed using an ad hoc C++ program. Except for the false positive rate, default search parameters were used including the “automatic” setting for mass accuracy (10 ppm for precursor ions and 15 ppm for product ions), a minimum of 1 peptide match per protein, a minimum of 3 consecutive product ion matches per protein, and a minimum of 7 total product ion matches per protein. The maximum false positive rate (FPR) against the randomized forward database was set to 2%, and the absolute protein quantification functionality was enabled using the phosphorylase B internal standard. Only 1 missed tryptic cleavage site was allowed during the search. A fixed carbamidomethyl-Cys modification was used, in addition to the following variable modifications: deamidation of Asn and Gln; oxidation of Met; and dehydration of Ser and Thr (insource modification). Data Preprocessing for Quantification. Following database searching, identified proteins and their amounts derived from absolute quantification were compared across injections to determine dose- and time-dependent secretion profiles. “Missing” data (proteins not identified in a given injection or treatment) were replaced with a value representing the limit of detection as determined by the smallest detected protein amount within the data set (3 fmol). Outlier absolute quantification values between technical replicates were manually corrected by normalizing peptide intensities of the protein of interest against the intensities of peptides from the spiked internal standard.

Results and Discussion Identification and Quantification of Secreted Proteins. Tryptic digests of secreted proteins were analyzed using an alternate scanning mode of data acquisition (LC/MSE) on a Q-Tof mass spectrometer as described above. LC/MSE is a recently introduced, novel mode of data-independent acquisition where alternating MS scans of lower and higher collision energy are used to simultaneously capture peptide precursor intensity data and peptide fragmentation data, allowing both protein quantification and identification in the same analysis. Figure 1A shows a chromatogram from one of the duplicate injections of a 6 h treatment with 1.0 mM SA. Each “stick” in the chromatogram corresponds to a unique peptide [M + H]+ at its retention time, and each peak’s height corresponds to the overall peptide intensity derived from combining the intensities of all charge states and associated isotopes for that

Global Analysis of SA-Induced Plant Protein Secretion Responses

research articles

Figure 1. Identification of a secreted Arabidopsis extracellular dermal glycoprotein (AT1G03220) after 6 h SA treatment. (A) Chromatogram of unique peptide [M + H]+ monoisotopic masses based on retention time and intensity following charge state reduction and deisotoping. A total of 15 unique peptides were detected for AT1G03220, which are indicated in blue and are observable for the more highly abundant peptides. (B-D) LC/MSE product ion spectra for the three most intense peptides detected for AT1G03220. Each peptide precursor is matched to its corresponding product ions by alignment of retention time peak profiles based on chromatographic coelution.

peptide (i.e., charge state reduction and deisotoping). Within this chromatogram are highlighted 15 unique peptides (covering 43% of the protein sequence) for the extracellular dermal glycoprotein AT1G03220, several of which cannot be seen clearly without selecting a smaller retention time window. This reflects the sample complexity and the dynamic range of the observed peptide intensities. Figure 1B-D shows product ion spectra for the three most intense peptides detected from AT1G03220 during this analysis. These LC/MSE spectra provide excellent product ion coverage across each peptide with relatively high mass measurement accuracy (2-fold increase in secretion in response to SA at one or more time points. Previous proteomic studies using standard 2D gel-based proteomic approaches identified 18 SA-responsive1 and 8 chitosan/elicitor-responsive30 proteins in the Arabidopsis secretome. As discussed above, a major advantage of a label-free LC/MS approach such as LC/MSE is that it does not require pairwise comparisons between samples, thus, accommodating more complex experiments. In turn, the type of factorial study reported here not only enabled us to assess time and dose effects of SA on protein secretion, but also led to identification of a larger number of SA-responsive proteins.

Figure 2. Absolute protein quantification using LC/MSE. (A) Intensities of the three most intense extracellular dermal glycoprotein (AT1G03220) peptides detected for duplicate injections of the 2 and 6 h, 1.0 mM SA treated samples. The peptide intensity is based on deconvoluting and deisotoping the peptide signal during its chromatographic elution and integration of the corresponding peak area. Absolute protein amounts for each replicate were reported by PLGS 2.3 and calculated by comparing the average intensity of the three most intense peptides for AT1G03220 in each injection with the average intensity of the three most intense peptides matched to the internal standard protein rabbit phosphorylase B (spiked into each sample at a level of 50 fmol per injection). (B) Scatter plot comparison of log2transformed absolute protein amounts obtained from two technical replicates produced a correlation coefficient of 0.988. The plot was created with SAS software, version 9.1 (SAS Institute, Cary, NC).

peptides in each injection with a response factor generated from the average intensity of the three most intense peptides matched to the 50 fmol phosphorylase B internal standard. Comparing the resulting protein quantification between duplicate technical replicates by calculating the correlation coefficient between replicate measurements plotted as log2 intensities in a scatter plot (Figure 2B) showed that reproducibility between replicate analyses was excellent (correlation coefficient ) 0.988). In addition, the percent coefficient of variation (% CV) was calculated across all proteins and technical replicates. Over 86% of the duplicate injections displayed % CVs of less than 20%, another verification of the excellent analytical reproducibility obtained with the LC/MSE approach. LC-MSE Reveals 63 Proteins with Increased Secretion in Response to SA. To identify proteins involved in plant defense responses, we assessed temporal changes in the Arabidopsis secretome in response to SA. Absolute quantities of individual proteins from samples treated with 1.0 mM SA and from uninduced, HBA-treated controls at 1, 2, 6, and 18 h were compared to determine potential levels of induced secretion 86


The identified SA-responsive proteins were assigned to one or more of the 11 functional groupings described by Schoof et al.31 (Figure 3). About three-quarters of the proteins are in the four “major” groups, that is, metabolism (34%), unclassified (15%), defense-related (13%), and proteins with binding functions (12%). Other “minor” group proteins include those involved in protein fate, interaction with the environment, metabolic regulation, cellular transport, protein synthesis, cell fate, and development. Temporal Patterns of SA-Induced Protein Secretion. Plant defense mechanisms are an interconnected set of transient as well as sustained processes that initiate immediately after pathogen attack and last for periods ranging from minutes to hours to days. These responses include the initial rapid, transient production of reactive oxygen species (ROS), with the subsequent induction of a localized hypersensitive response (HR), and ultimately the SA-mediated induction of proteins and processes comprising what is termed systemic acquired resistance. Each of these responses in turn requires the expression and function of a potentially wide range of proteins, hormones or secondary metabolites. Our longitudinal study across 18 h post SA treatment generated a picture of the distribution of proteins with significantly altered secretion over four points (induced), compared with time-matched uninduced controls (Figure 4). These results suggest that most induced protein secretion takes place within the first 2 h after SA treatment, with the number of proteins with induced secretion decreasing after 6 h. To further evaluate the temporal dynamics of secretome changes, SA-responsive proteins were categorized into four groups based on secretion patterns induced by 1.0 mM SA (Figure 5). (1) In pattern 1, induction of protein secretion reached a maximum 1 h post-treatment. There are 26 SAresponsive proteins that fall into this group, including calmodulin, glutathione transferase, peroxidase and jacalin lectin (Figure 5A). (2) In pattern 2, induction of protein secretion reached a maximum 2 h after SA treatment. The 14 proteins in this group include chitinase, Cu/Zn superoxide dismutase (SOD) and a glycosyl hydrolase (Figure 5B). (3) In pattern 3, induction of protein secretion reached a maximum 6 h after SA treatment. The 14 proteins in this group include a cyclase family protein, glycoside hydrolase and expansin-like proteins (Figure 5C). (4) In pattern 4, secreted proteins showed two maxima over the 18 h time course. These 9 proteins include a pectinesterase, an aminopeptidase, a TSK-associating protein and a lipid transfer protein (Figure 5D). The observation that a majority of proteins fell into the first two groups suggests

research articles

Global Analysis of SA-Induced Plant Protein Secretion Responses E

Table 1. Arabidopsis Secreted Proteins Identified Using LC/MS

induction levela (1.0 mM SA) AGI number

AT4G37530 AT1G02360 AT4G25810 AT2G38380 AT2G15220 AT3G56310 AT5G67360 AT2G15130 AT1G74790 AT5G57560 AT3G07390

protein name and annotation

2h

6h

18 h

SPb

secretome Pc NN-number

1.0 0.7 1.0 1.3 1.5 1.0 1.8 1.5 0.6 1.0 1.0

1.0 0.7 1.1 0.7 0.4 1.7 0.9 0.7 0.7 0.8 0.9

0.9 1.0 0.3 0.2 0.3 1.0 0.7 0.7 0.7 1.0 0.7

Yes Yes Yes Yes Yes Yes Yes No Yes Yes No

0.79 0.86 0.36 0.81 0.93 0.86 0.65 0.56 0.73 0.77 0.76

4.4 10.6 6.3 18.3 8.0 12.5 10.3 5.9 10.6 11.8 10.0 7.3 3.2 15.1 18.0 5.9 21.2 9.2 2.7 7.4 15.8 19.9 9.2 11.1 34.7 34.7

3.2 1.0 5.4 8.5 4.1 7.7 6.4 4.5 6.2 4.8 1.0 3.4 2.1 5.9 8.3 1.7 3.1 2.9 2.3 2.8 1.1 3.9 3.5 3.8 15.7 8.1

1.0 1.0 3.8 5.5 2.1 2.4 1.4 3.0 1.9 2.3 0.1 1.0 0.7 1.8 4.8 1.8 1.0 2.2 0.7 1.0 0.6 2.1 1.3 0.9 2.9 2.9

1.1 1.0 1.0 1.0 2.4 1.0 1.0 1.2 1.1 3.9 1.0 1.9 1.1 1.4 1.0 1.4 1.3 1.2 0.3 1.4 0.8 2.2 1.0 2.0 2.0 2.0

Yes Yes No No No No No No No No Yes No No Yes Yes No Yes No Yes Yes Yes No No No No No

0.84 0.71 0.73 0.51 0.51 0.46 0.47 0.51 0.35 0.49 0.69 0.46 0.85 0.70 0.74 0.64 0.93 0.76 0.72 0.74 0.85 0.25 0.79 0.33 0.57 0.56

1.1

5.3

5.1

1.0

Yes

0.56

0.8 1 0.8 1.1

3.7 2.4 6.8 2.2

0.8 0.4 3.0 1.2

1.5 0.5 1.0 0.2

No Yes No No

0.73 0.93 0.47 0.29

0.2 0.1 0.8 1.2 1.6 1.1 1.0 0.9 0.8

3.4 20.8 5.1 5.0 3.7 25.2 2.1 4.3 8.5

0.1 8.1 2.7 5.2 5.4 17.1 2.1 1.3 7.9

0.1 1.0 1.0 1.0 2.9 1.0 1.0 1.0 1.0

Yes No No Yes Yes No Yes Yes No

0.37 0.74 0.50 0.90 0.82 0.23 0.92 0.86 0.47

1.0 1.3 0.2 1.0 0.1

3.0 3.8 2.3 1.0 1.4

6.3 4.0 8.9 13.1 1.7

2.7 1.0 0.8 1.0 1.0

No No Yes Yes Yes

0.03 0.53 0.96 0.62 0.95

1h

Uninduced Proteins PER51, peroxidase 1.0 Chitinase 1.0 XTR6, xyloglucan endotransglycosylase 0.7 PER22, peroxidase 0.4 Putative secretory protein 1.0 R-galactosidase 1.0 ARA12, subtilisin-like protease 0.4 Basic secretory protein (BSP) family protein 1.0 HIPL1 hedgehog-interacting protein precursor 1.0 XTH22, xyloglucan endotransglucosylase 0.7 AIR12, extracellular matrix constituent 1.4 SA-Responsive Proteins

AT3G12500 AT1G77510 AT4G29680 AT5G18170 AT5G15650 AT2G36460 AT5G07440 AT3G08900 AT1G78380 AT3G02230 AT5G60360 AT2G24200 AT1G23410 AT5G06860 AT2G17420 AT3G43810 AT5G23820 AT5G53560 AT3G49120 AT4G08780 AT4G11650 AT3G16430 AT3G32980 AT5G55730 AT3G16410 AT3G1639

Pattern 1 PR3, basic Chitinase ATPDIL1-2, protein disulfide isomerase Phosphodiesterase GDH1, glutamate dehydrogenase R-1,4-glucan-protein synthase Fructose-bisphosphate aldolase Glutamate dehydrogenase RGP, R-1,4-glucan-protein synthase GST8, glutathione transferase ATRGP1, reversibly glycosylated polypeptide AALP, cysteine proteinase Leucine aminopeptidase Ubiquitin extension protein PGIP1, polygalacturonase inhibiting protein ATNTRA, NADPH-thioredoxin reductase CAM7, Calmodulin MD-2-related lipid recognition domain-containing protein Cytochrome b5 reductase PERX34, peroxidase PER38, peroxidase ATOSM34, osmotin-like protein Jacalin- related protein (lectin) PER32, peroxidase FLA1, fasciclin-like arabinogalactan protein jacalin- related protein (lectin) jacalin- related protein (lectin)

AT5G20230 AT5G18100 AT3G03780 AT4G34480 AT1G03220 AT1G12080 AT4G36430 AT2G22420 AT4G23590

Pattern 2 ATCWINV3, β-fructofuranosidase, hydrolyzing O-glycosyl compounds CSD1, Cu/Zn superoxide dismutase Chitinase ATMETS, methionine synthase PBP1, assists PYK10 (β-glucosidase complex) activity in pest damaged tissues. BCB, blue-copper-binding protein CSD3, Cu/Zn superoxide dismutase AtMS2, methionine synthase Putative glucan endo-1,3-β-glucosidase Extracellular dermal glycoprotein Similar to vacuolar calcium binding protein PER49, peroxidase PER17, peroxidase Aminotransferase

AT4G20260 AT3G03910 AT4G34180 AT5G08380 AT2G43535

DREPP, plasma membrane protein Glutamate dehydrogenase Cyclase family protein R-galactosidase Trypsin inhibitor, defensin-like protein

AT1G55120 AT1G08830 AT2G43570 AT5G17920 AT3G16420

Pattern 3

Journal of Proteome Research • Vol. 8, No. 1, 2009 87

research articles

Cheng et al.

Table 1. Continued induction levela (1.0 mM SA) AGI number

protein name and annotation


Glycosyl hydrolase Reversibly glycosylated polypeptide FSD1, iron superoxide dismutase EXPL1, expansin-like protein PER37, peroxidase BGLU46, Glycoside hydrolase Lipid transfer proteins, stress and pathogen-inducible motifs FAD-binding domain-containing protein Meprin and TRAF homology domain-containing protein


Pattern 4 LOS2 (low expression of osmotically responsive prot.), phosphopyruvate hydratase PER15, peroxidase Pectinesterase (methylesterase) Jacalin- related protein (lectin) Lactoylglutathione lyase Cytosolic aminopeptidase family Cytosolic aminopeptidase family TSK-associating protein LTP3, lipid binding/transfer protein

18 h

SPb

secretome Pc NN-number

5.3 9.5 11.5 15.6 0.9 9.2 3.1 13.6 4.3

1.0 1.0 1.0 1.0 2.8 1.0 1.0 1.0 1.0

Yes No No Yes Yes No Yes Yes Yes

0.79 0.51 0.20 0.93 0.65 0.44 0.81 0.73 0.74

1.0

15.7

7.2

No

0.60

2.7 4.7 3.9 5.9 0.3 1.8 7.2 4.1

0.7 1.2 1.4 8.2 1.0 1.0 1.4 1.5

4.3 7.6 13.2 1.0 5.6 6.6 8.8 2.1

Yes No No No No No Yes Yes

0.85 0.60 0.93 0.37 0.51 0.57 0.36 0.85

1h

2h

6h

1.0 1.5 9.8 1.0 0.7 1.0 0.8 1.0 0.4

1.0 1.0 2.4 5.5 1.1 0.5 0.6 1.8 2.9

6.9 1.2 6.5 13.1 7.2 12.1 12.7 7.9 5.0

a Fold change of secreted protein induced by 1.0 mM SA treatment compared to uninduced control based on the measured absolute quantity of each protein. b Result of Signal-P analysis for the prediction of a signal sequence within each protein. c Protein secretion predicted by SecretomeP (http:// www.cbs.dtu.dk/services/SecretomeP/). Neural network (NN) scores higher than 0.5 are considered to predict extracellular localization.

Figure 3. Functional classification of 63 SA-responsive proteins with >2 fold increase in secretion compared to uninduced controls. A functional classification was assigned to each unique protein, and the total number and percentage of proteins in each class were determined.

that many secreted proteins have key roles during the early stages of plant defense. Many Early Secreted Proteins Lack a Signal Peptide. Extracellular proteins are classically characterized by the presence of an amino-terminal signal peptide in the protein sequence that has three domains: a positively charged N-region, a central hydrophobic H-region, and a polar C-region. This peptide sequence constitutes a signal that directs proteins to the ER/Golgi secretory pathway. It has been traditionally believed that this N-terminal leader peptide is strictly required for targeting specific protein secretion in plants. However, when the program SignalP 3.0 (http://www.cbs.dtu.dk/services/ SignalP/) was employed to predict the presence of a signal peptide in the 74 observed secreted proteins, there were 37 proteins without a recognizable signal peptide (Table 1). Among the proteins showing maximal secretion within the first hour after SA treatment, 65% lacked a signal peptide. Among those showing SA-induced maxima at 2 h, 50% lacked a signal peptide. In contrast, for proteins secreted at later time-points, only 35% lacked a signal peptide. Therefore, many of the 88


Figure 4. Time course of SA responsive proteins. Curves represent the total number of proteins with >2-fold increased secretion in response to SA (0.5 or 1.0 mM) at each time point. The largest number were secreted 1 to 2 h after treatment.

proteins that were most rapidly secreted in response to SA might be secreted by leaderless, so-called nonclassical, Golgiindependent mechanisms. The virtual absence in our secretome samples of proteins normally abundant in the plant cytoplasm (e.g., ribosomal, actin and tubulin subunit proteins, and glycolytic enzymes) indicates that the appearance of these leaderless proteins is not a result of nonspecific cell lysis. Protein sequences were also analyzed on the neural network server SecretomeP (www.cbs.dtu.dk/Services/SecretomeP). This program is based on the observation by Bendtsen et al.32 that secreted proteins all share characteristic features associated with extracellular function, and that these features are independent of specific secretion mechanisms. Features assessed include specific post-translational modification consensus sequences and degradation signals, as well as protein structure, size, charge and amino acid composition. Interestingly, when we used SecrotomeP to predict secretion of the 37 leaderless proteins identified, a majority (60%) were indeed predicted to be secreted (NN-scores >0.5; Table 1). However, one must note

Global Analysis of SA-Induced Plant Protein Secretion Responses

research articles

Figure 5. SA-responsive proteins categorized into four groups based on their temporal secretion in response to 1.0 mM SA. Proteins with maximal secretion at (A) 1 h, (B) 2 h, (C) 6 h, and (D) any two time points during SA exposure are shown. The quantitative level of each protein at each time point is plotted as the log2 value (SA-induced/uninduced control).

that a number of classically secreted proteins (i.e., those with a signal peptide) had NN-scores

MSE for Global Analysis of

Recommend Documents