Targeted Proteomics Analysis of Protein Degradation in Plant

Aug 18, 2014 - †Proteome Analytics Research Group, ‡Independent Junior Research ... View: ACS ActiveView PDF | PDF | PDF w/ Links | Full Text HTML...
3 downloads 0 Views 2MB Size
Article pubs.acs.org/jpr

Targeted Proteomics Analysis of Protein Degradation in Plant Signaling on an LTQ-Orbitrap Mass Spectrometer Petra Majovsky,† Christin Naumann,‡,⊥ Chil-Woo Lee,§ Ines Lassowskat,∥ Marco Trujillo,§ Nico Dissmeyer,*,‡,⊥ and Wolfgang Hoehenwarter*,† †

Proteome Analytics Research Group, ‡Independent Junior Research Group on Protein Recognition and Degradation, §Independent Junior Research Group Ubiquitination in Immunity, and ∥Cellular Signaling Group, Department of Stress and Developmental Biology (SEB), Leibniz Institute of Plant Biochemistry (IPB), Halle (Saale) D-06120, Germany ⊥ ScienceCampus Halle − Plant-Based Bioeconomy (SCH), Halle (Saale) D-06120, Germany S Supporting Information *

ABSTRACT: Targeted proteomics has become increasingly popular recently because of its ability to precisely quantify selected proteins in complex cellular backgrounds. Here, we demonstrated the utility of an LTQ-Orbitrap Velos Pro mass spectrometer in targeted parallel reaction monitoring (PRM) despite its unconventional dual ion trap configuration. We evaluated absolute specificity (>99%) and sensitivity (100 amol on column in 1 μg of total cellular extract) using full and mass range scans as survey scans together with data-dependent (DDA) and targeted MS/MS acquisition. The instrument duty cycle was a critical parameter limiting sensitivity, necessitating peptide retention time scheduling. We assessed synthetic peptide and recombinant peptide standards to predict or experimentally determine target peptide retention times. We applied optimized PRM to protein degradation in signaling regulation, an area that is receiving increased attention in plant physiology. We quantified relative abundance of selected proteins in plants that are mutant for enzymatic components of the N-end rule degradation (NERD) pathway such as the two tRNA-arginyl-transferases ATE1 and ATE2 and the two E3 ubiquitin ligases PROTEOLYSIS1 and 6. We found a number of upregulated proteins, which might represent degradation targets. We also targeted FLAGELLIN SENSITIVE2 (FLS2), a pattern recognition receptor responsible for pathogen sensing, in ubiquitin ligase mutants to assay the attenuation of plant immunity by degradation of the receptor. KEYWORDS: Degradomics, N-terminal, targeted protein degradation, protein quantification, N-end rule pathway, Arabidopsis, plant−pathogen interaction, parallel reaction monitoring



INTRODUCTION

C18 reverse-phase LC and electrosprayed online into a mass spectrometer. Traditionally, a triple quadrupole configuration (QqQ), where three quadrupole mass analyzers are operated as selective mass filters in series, is employed. The direct current and radio frequency of the first quadrupole (Q1) are set to afford only ions with a certain m/z, a stable transitory trajectory. These ions, filtered from the total ion population, are fragmented in the second quadrupole (q2), usually with collision induced dissociation (CID), and selected product ions are again filtered from all product ions in the third quadrupole (Q3). The product ion signal is recorded in a chromatographic trace used to monitor peptide abundance as a proxy for protein abundance. Triple quadrupole mass spectrometers do not trap ions but detect a continuous ion beam using electron multipliers. Their duty cycle is very short, so multiple precursor/product

The quantification of protein abundance is a fundamental issue in cell biology. Over the past decade, liquid chromatography online mass spectrometry (LC−MS) has emerged as a powerful technique because of its accuracy, sensitivity, and ability to measure the abundance of many proteins in one analysis.1 Shotgun proteomics experiments try to quantify as many proteins as possible without any qualitative preselection. In many cases, however, a subset of the total proteins or proteoforms present in the sample may be of particular interest to researchers. To this end, targeted proteomics applications have become increasingly powerful and are enjoying unprecedented popularity.2 The gold standard in targeted proteomics is selected reaction monitoring (SRM).3,4 As in discovery-type shotgun proteomics experiments, the protein extract is digested into peptides either directly or following prefractionation using techniques such as SDS-PAGE, liquid chromatography (LC), or gel filtration to reduce complexity. The peptide mixture is then separated by © 2014 American Chemical Society

Received: February 23, 2014 Published: August 18, 2014 4246

dx.doi.org/10.1021/pr500164j | J. Proteome Res. 2014, 13, 4246−4258

Journal of Proteome Research

Article

Skoog (MS) medium (Duchefa, cat. no. M0221) containing 1% (w/v) sucrose and 8 g/L phytoagar (Duchefa, cat. no. P10031). Arabidopsis mutants used were the single T-DNA insertion allele prt6 (prt6−3, At5g02310, GABI_270G04, NASC stock ID: N425900)12,46 and the double mutant ate1 ate2 carrying TDNA alleles for both ate1 and ate2 (ate1, At5g05700, SAIL_023492; ate2, SAIL_040788)13 or the ethyl methanosulfonate allele prt1 (prt1-1, At3g24800).14−16 Homozygous pub12 pub13 double knockouts37 were grown in liquid 0.5 MS medium (Duchefa) with a vitamin supplement (Duchefa, cat. no. P10031) until seedling stage with supplemented 0.25% (w/v) sucrose and 0.05% (w/v) MES (pH 5.7) under short-day conditions (8 h light; 16 h dark) at 21 °C (20 °C in dark) on a rotary shaker for 12 days. For flg22 treatment, the culture medium was supplemented with flg22 to a concentration of 1 μM for 30 min before harvest.

transitions can be measured several times, even for lowabundance peptides, with very small chromatographic elution windows. Therefore, QqQ instruments have unprecedented sensitivity in the low-attomole range. However, mass accuracy and resolution of quadrupole mass analyzers is poor (∼0.2 Da), so great care must go into selecting the appropriate transitions to ensure adequate specificity to unambiguously identify the monitored peptides in the complex proteolytic background.5 Exchanging the third quadrupole with a time-of-flight (TOF) analyzer with a mass accuracy of 99%) with an optimized method using narrow MS scans followed by retention time scheduled MS/MS scanning of target m/z on a global mass list. This is the same as that reported for the QExactive QqOrbi8,9 and other state-of-the-art instruments. Evaluation of Full and Narrow Mass Range MS Scaning in DDA Analysis

The use of standard peptides as references to determine peptide retention times and ESI responses as well as for absolute quantification is preferred in targeted proteomics. In many cases, however, standard peptides may not be readily available for each and every protein of interest. Therefore, we explored the possibility of DDA analysis using an inclusion list without knowing peptides mass spectrometric behavior and retention times beforehand. For method development, we used a previously acquired shotgun proteomics data set comprising more than 700 000 MS/MS spectra as a reference map. This data was based on the proteome of several mutants defective for different enzymatic components of the NERD pathway in Arabidopsis and was used here to select target peptides. The mutants are described below (Figure 2). According to the current understanding, all protein sequences starting with a Met-Cys (MC) at the N-terminus can be co- or post-translationally cleaved by aminopeptidases and lose their Met1, exposing an N-terminal Cys residue. Therefore, these MC-starting proteins may be considered putative NERD pathway substrates. For our work, we have selected peptides from MC-starting proteins identified with four or more MS/ 4251

dx.doi.org/10.1021/pr500164j | J. Proteome Res. 2014, 13, 4246−4258

Journal of Proteome Research

Article

spent on survey scanning, thus reducing the overall duty cycle time (Table 1).

MS spectra in the reference data set. We also selected peptides from a small set of regulatory proteins and diverse enzyme classes to later investigate changes in their abundance under nondestabilizing conditions by NERD. This included a total of 136 peptides, each identified with at least four peptide spectral matches (PSMs) with 99% confidence (FDR q-value threshold = 0.01), in the reference data set, mapping to 22 proteins. These peptides were selected for an inclusion list containing 136 m/z; each peptide was targeted in one charge state. Four PSMs were chosen because this reflects a reasonably strong ion signal (for instance, the minimum of the accurate quantitative dynamic range of spectral counting).20 The respective MS/MS spectra of a selection of these peptides are shown in the Supporting Information. We developed two DDA methods to measure a tryptic digest of Arabidopsis total protein extract. One was similar to the method we used for BSA with 10 narrow MS scan windows framing the m/z on the inclusion list each followed by five DDA MS/MS scans targeting the five most abundant m/z on the inclusion list, and another partitioned the entire scan range into four broad MS scan windows analogous to gas-phase fractionation experiments, each also followed by five DDA MS/ MS events. More peptides on the inclusion list mapping to more of the target proteins were identified using four broad as opposed to 10 narrow MS scan windows (Table 1). Next, we examined the use of a full MS scan compared to the more advantageous four broad MS scans prior to DDA. We also extended the gradient from 90 to 120 min to better resolve the peptide mixture, which led to a substantial increase in identified peptides for both full and broad MS scan methods. Broad mass range scanning increased the S/N more than 3-fold for selected targeted ions. The normalized intensity was the same as or less than that in the full scan despite the higher AGC target value and max IT. It seems that although narrow or broad mass range scanning selectively accumulates ions it does not amplify absolute abundance because the Orbitrap can easily accommodate the total ion current and so the entire population of all ion species in a full scan. Indeed, AGC target values and max IT were seldom reached (Supporting Information Figure 1). Therefore, there is no signal gain of low-abundance peptide ions to trigger a DDA MS/MS event using mass range scans, an effect that we had hoped to achieve. The relative abundance of targeted ions, especially in narrow MS windows, however, will increase, translating into the observed increase in S/N. In PRM, peptides are identified by an acquired MS/MS spectrum (and not a priori by retention times and precursor/ product transitions as in SRM). Sensitivity is ultimately linked to peptide identification, i.e., a signal may be quantifiable, but it is of little use if it cannot be annotated with a peptide sequence. Therefore, peptide identification is essential, posing an inherent limit in DDA, where a signal threshold must be surpassed to trigger an MS/MS scan. Utilizing a single full scan instead of multiple mass range scans reduced duty cycle times, which were less than 1 s for the full scan, 3 s for the broad MS scan, and 6 s for the narrow MS scan methods, on average. We found the instrument duty cycle time to be a critical parameter because, in conjunction with dynamic exclusion, a short duty cycle allows the instrument to pick more low-abundance precursors that appear for only a few seconds in chromatographic elution. Therefore, more proteins were identified by employing a full scan, translating into increased sensitivity at comparable numbers of peptide identifications because less time was

Evaluation of Standard Peptides for Retention Time Scheduling and Targeted Proteomics Analysis

Targeted peptide analysis can potentially overcome the sensitivity limit of DDA because m/z on the mass list are fragmented independently and regardless of MS survey scans in the more sensitive LTQ. However, MS/MS analysis of every m/z on the mass list throughout the analysis is not feasible for long mass lists because cycle times are extended to minutes. Acquisition of the m/z on the mass list must be scheduled according to the peptides retention time in chromatography to keep cycle times short. To this end, we assayed the possibility of using a mixture of standard peptides to predict retention times of target peptides. Fifty femtomoles of the retention time calibration mixture consisting of 15 heavy isotope labeled peptides spanning a wide hydrophobicity range (hydrophobicity factor 7.56−46.66) were spiked into 1 μg of Arabidopsis total protein extract and measured with the DDA full scan method and a 120 min gradient with the isotope labeled peptides added to the mass list. The Krokhin hydrophobicity21 of the standard peptides was determined, and it showed an excellent linear relationship with retention (R2 = 0.98). The linear model was used to predict target peptide retention times. The mean difference between observed and predicted retention times for all PSMs of identified target peptides was 2.6 min, with a standard deviation of 1.75, experimentally underscoring the accuracy of peptide retention time prediction.22 The m/z on the mass list were then scheduled for acquisition using a ±5 min retention time window around their predicted retention times following either one full MS scan or a set of narrow MS scans employing a 120 min gradient (Table 1). More than half of the targeted peptides were identified. This is a substantial amount when considering that 1/3 of these peptides were identified with 10 or fewer PSMs in the original shotgun proteomics study comprising more than 700 000 PSMs, indicating that they are not abundant (∼10 ppm of the total spectral count). The number of identified peptides was comparable for DDA and targeted analyses using a 120 min gradient. The effect of reducing the time spent on survey scanning as well as the total duty cycle time by employing a full scan was even more pronounced when targeted peptide analysis was employed because maximum time is allocated to MS/MS acquisition in the LTQ independent of the survey scan and regardless of a signal threshold. Thus, combining a full scan with retention time scheduled targeted peptide analysis maximized sensitivity, confirming the results for BSA described above and allowing us to identify 19 of the 22 targeted proteins (86%). Selective Identification of NERD Target Protein Candidates

Protein degradation by the ubiquitin proteasome system is a key aspect of protein stability and is central to the regulation of nearly all cellular processes (Figure 2).23−25 One means of protein degradation is the N-end rule degradation pathway (NERD).26 Proteins bearing an N-degron, which consists of an N-terminal destabilizing amino acid residue (R, K, H, F, W, Y, L) and one or more proximal internal lysine residues, are potential substrates of the highly specific NERD E3 Ub protein ligases termed N-recognins. In Arabidopsis, only two of these, namely, PROTEOLYSIS1 (PRT1)14−16 for hydrophobic (type 4252

dx.doi.org/10.1021/pr500164j | J. Proteome Res. 2014, 13, 4246−4258

4253

AT4G00570.1d AT4G39260.1 AT5G25980.2 AT5G26000.1 AT5G26000.2

AT2G21660.1 AT3G26650.1

AT1G47128.1

GRANULIN REPEAT CYSTEINE PROTEASE FAMILY PROTEIN COLD, CIRCADIAN RHYTHM, AND RNA BINDING 2 GLYCERALDEHYDE 3-PHOSPHATE DEHYDROGENASE A SUBUNIT 1 NAD-DEPENDENT MALIC ENZYME 2 COLD, CIRCADIAN RHYTHM, AND RNA BINDING 1 GLUCOSIDE GLUCOHYDROLASE 2 THIOGLUCOSIDE GLUCOHYDROLASE 1 THIOGLUCOSIDE GLUCOHYDROLASE 1

350 447

(0)4 (12)14 (9)11c (1)1

(2)2 (2)2 (3)4 (0)1 (0)1

(1)1 (2)4

399

(0)4

607 169 547 541 456

176 396

462

288 169 578 577 326 578 579

length (aa)

(1)1 (2)2 (0)1 (0)1 (1)1 (0)2 (0)2

(unique) peptides for quantificationb

66.6 16.6 62.7 61.1 51.4

16.9 42.5

50.9

47.6

37.6

42.8

32.9 18.8 65.2 65 35.8 65 65.1

MW

7.06 5.68 7.44 5.92 5.88

6.15 7.75

5.41

6.8

7.77

8.05

6.04 8.27 6.57 6.8 5.19 6.46 6.46

pI

1.1 1.3 0.6 0.8 0.8

1.1 1.3

1.4 1.4c 1.2

1.3

1.3

1.1 1.4 1.2 1.2 0.7 1.6 1.6

ate1 ate2

0.9 1.2 0.6 0.7 0.7

1.3 1.1

1.2 1.1c 1.0

1.1

1.1

0.7 1.6 1.4 1.4 0.5 1.7 1.7

prt1

1.1 1.3 1.0 1.1 1.1

1.1 1.4

1.3 1.4c 1.6

1.4

1.4

1.3 1.2 1.2 1.2 0.9 1.2 1.2

prt6

1.0 1.0 1.0 1.0 1.0

1.0 1.0

1.0 1.0c 1.0

1.0

1.0

1.0 1.0 1.0 1.0 1.0 1.0 1.0

wildtype

normalized abundance

9.1 6.1 37.6 n.a. n.a.

n.a. 9.7

13.2 12.7c n.a.

9.7

9.7

n.a. 8.9 n.a. n.a. n.a. 33.8 33.8

ate1 ate2

17.7 13.6 15.4 n.a. n.a.

n.a. 8.0

17.4 15.7c n.a.

8.0

8.0

n.a. 7.6 n.a. n.a. n.a. 26.6 26.6

prt1

5.0 19.0 15.4 n.a. n.a.

n.a. 6.1

17.5 17.5c n.a.

6.1

6.1

n.a. 2.4 n.a. n.a. n.a. 7.2 7.2

prt6

0.0 0.0 0.0 n.a. n.a.

n.a. 0.0

0.0 0.0c n.a.

0.0

0.0

n.a. 0.0 n.a. n.a. n.a. 0.0 0.0

wildtype

coefficient of variance

The table contains protein information as well as quantitative information for each protein identified. bThe first value in parentheses is the number of peptides unique to this protein used for quantification, and the second value is the total number of peptides, including peptides also assigned to other proteins, used for quantification. cQuantification following outlier removal. dPrecursor quantification only.

a

protein description

MC-Starting Proteins AT3G50440.1 METHYLESTERASE 10 AT4G09320.1 NUCLEOSIDE DIPHOSPHATE KINASE FAMILY PROTEIN AT5G10240.1 ASPARAGINE SYNTHETASE 3 AT5G10240.2 ASPARAGINE SYNTHETASE 3 AT5G48180.1 NITRILE SPECIFIER PROTEIN 5 AT5G65010.1 ASPARAGINE SYNTHETASE 2 AT5G65010.2 ASPARAGINE SYNTHETASE 2 Diverse Enzyme and Protein Classes AT1G12900.1 GLYCERALDEHYDE 3-PHOSPHATE DEHYDROGENASE A SUBUNIT 2 AT1G12900.3 GLYCERALDEHYDE 3-PHOSPHATE DEHYDROGENASE A SUBUNIT 2 AT1G42970.1 GLYCERALDEHYDE-3-PHOSPHATE DEHYDROGENASE B SUBUNIT

protein accession

Table 2. Protein Abundance in Knockout Mutants of NERD Enzymatic Components Determined by Targeted PRMa

Journal of Proteome Research Article

dx.doi.org/10.1021/pr500164j | J. Proteome Res. 2014, 13, 4246−4258

Journal of Proteome Research

Article

consistent throughout the three experiments, with a mean coefficient of variance (CV) of 23.6%, which we find to be low for three independent experiments with different plants. Ten proteins were quantified with two or more peptides. Peptide abundance for these proteins was in good agreement (CV; Table 2 and Supporting Information Table 2). MS/MS spectra of selected peptides used for quantification are shown in the Supporting Information. From the group of MC-starting proteins, we found several diverse members upregulated in ate1 ate2 and prt6 mutants, where the NERD arginylation branch is completely interrupted (Table 2 and Figure 2C,D). The abundance of METHYL ESTERASE 10 (MES10), NUCLEOSIDE DIPHOSPHATE KINASE FAMILY PROTEIN (NDPK1), and the two ASPARAGINE SYNTHETASES (ASNs) was increased in the ate1 ate2 double mutant and also in the prt1 single mutant (Figure 2B). These proteins have N-terminal cysteine residues, possibly explaining the very prominent elevation in ate1 ate2. This suggests possible recognition and modification by ATEs followed by subsequent ubiquitination of their arginine conjugated form by PRT6. The increased abundance in prt1 is much more modest, leading us to speculate that some activity of PRT1 may be involved in the degradation of proteins that also have basic N-degrons (Figure 2B). From the members of diverse Arabidopsis enzyme (or protein) classes listed in Tables 1 and 2, we could properly detect all three GLYCERALDEHYDE 3-PHOSPHATE DEHYDROGENASE subunits, GAPA, GAPA2, and GAPB, in both ate1 ate2 and prt6 mutants (Table 2; unpaired, homoscedastic Student’s t test with assumed symmetric distribution of difference of the mean, α < 0.01). The GRANULIN REPEAT CYSTEINE PROTEASE FAMILY PROTEIN (RD21) and the three GLUCOSIDE and THIOGLUCOSIDE GLUCOHYDROLASES (TGG1, TGG2) were found in samples of ate1 ate2 and in the Nrecognin mutants prt1 and prt6 (Table 2). TGG1 and TGG2 show a significant decrease in abundance in ate1 ate2, suggesting that they are downstream of the ATEs and conceivably of NERD, although abundance was not decreased in the N-recognin mutant genotypes. Also, for RD21, significance testing was not possible because of single peptide quantification. Moreover, the two glycine-rich proteins (GRPs) COLD, CIRCADIAN RHYTHM, AND RNA BINDING 1 and 2 and the NICOTINAMIDE ADENINE DINUCLEOTIDE-DEPENDENT MALIC ENZYME 2 (NAD-ME2) were successfully detected in the mutant samples. Their abundance was not greatly changed in the nerd mutants, suggesting that they are not necessarily linked to NERD despite presenting N-terminal destabilizing residues after predicted proteolytic cleavage of Nterminal signal peptides from the prepro-proteins (Figure 2). In order to verify that the relatively small changes in protein abundance we measured are not the result of variability in our experimental procedure, we extracted total proteins three times from WT plants. We then performed three analyses of each of the digested total protein extracts targeting the same set of peptides as in the original experiments with the WT and mutant genotypes for a total of nine targeted PRM analyses of these peptides in the WT background. As expected, there was no significant change in protein abundance for any of the measured peptides, indicating that the measured changes in the mutant backgrounds are indeed of a physiological nature (Supporting Information Table 3).

2) residues (F, Y, W) and PROTEOLYSIS6 (PRT6) for charged (type 1) residues (R, H, K),12 are known. Furthermore, proteins with N-terminally oxidized cysteine or aspartate and glutamate (Cox, D, E) are thought to be modified by tRNA-arginyl-transferases ATE1 and ATE2, which conjugate Arg to their N-terminus, making them accessible to Nrecognins. NERD functions in plants are just now emerging, with documented roles in controlling multiple developmental processes throughout the life cycle. Its enzymatic components are associated with seed ripening, lipid breakdown, hormonal signaling and germination,27 leaf and shoot morphogenesis, flower induction and apical dominance,13 delay in senescence,28 and even cell division.13 The most recent and prominent insight is that NERD functions as an oxygen sensor and regulates stress response after hypoxia, e.g., after flooding and plant submergence via stabilizing stress-related transcription factors.29,30 Little is known about the biological significance of NERD, as true evidence for metabolization of proteins by the NERD cascade enzymes such as ATEs is still lacking. Indeed, only a handful of proteins affected by NERD have been described solely genetically, but neither on the biochemical nor on the proteomic level.29,30 The Arabidopsis proteins momentarily considered to be substrates of the NERD protein degradation pathway are RESISTANCE TO PSEUDOMONAS SYRINGAE 1 INTERACTING PROTEIN 4 (RIN4, The Arabidopsis Information Resource (TAIR) at arabidopsis.org), five members of the ethylene response factor (ERF) subfamily B-2 of the ERF/ APETALA2 transcription factor family (i.e., HYPOXIA RESPONSIVE ERF 1 and 2 (HRE1 and 2),29 RELATED TO AP2.2 (RAP2.2)29 and 12 (RAP2.12),29,30 ETHYLENE RESPONSE FACTOR 72 (EBP)),29 and VERNALIZATION 2 (VRN2).29 These may have been called bona fide substrates prematurely, as not a single one has yet been validated molecularly. The identity of true NERD substrates remains elusive, partly because, besides the N-terminal residue, structural factors such as flexibility and accessibility of the N-degron and overall protein conformation determine recognition by the Nrecognins. Besides that, the protein sequence needs to harbor accessible, i.e., surface exposed, lysine residues in order to serve as Ub acceptor sites. One strategy is to measure the abundance of proteins with N-terminal destabilizing residues in mutant backgrounds where NERD is perturbed, i.e., in prt1 and prt6 mutants and ate1 ate2 double mutants, where the substrate abundance is expected to increase. To assess whether MCstarting proteins are upregulated in nerd mutants, we targeted the MC-starting proteins that were identified in our reference data set of the proteome effected by NERD mutation as well as the set of regulatory proteins and enzymes used for method development in the NERD perturbed mutant backgrounds. Note that none of the currently known NERD substrates were included in our reference data set, so we did not target them. The entire experiment was performed three times (Table 2 and Supporting Information Table 2). Only peptides identified with 99% confidence and with a well-defined ion signal peak with a clear S/N were used for peak integration and quantification. In all but two cases, where only the precursor signal was quantified, PRM was employed. Eighteen of the 22 targeted proteins listed in Table 1 could be repeatedly quantified in the nerd mutant genotypes in the three experiments and are listed in Table 2. Peptide abundance was 4254

dx.doi.org/10.1021/pr500164j | J. Proteome Res. 2014, 13, 4246−4258

Journal of Proteome Research

Article

terminal peptide of flagellin. We also measured FLS2 levels in the f ls2 knockout mutant as a negative control. The microsomal protein fraction was separated by SDS-PAGE (Supporting Information Figure 2), and approximately 100 ng of tryptic peptides from the gel band with the approximate molecular weight of FLS2 of each genotype were analyzed, targeting FLS2. The intracellular FLS2 kinase domain was expressed as a recombinant protein and in-gel digested with trypsin to produce FLS2 standard peptides that were used to determine FLS2 peptide m/z and retention times under matrix conditions. Seventeen recombinant tryptic FLS2 kinase domain peptides were identified with 95% confidence (FDR q-value threshold 0.05), and 14, with 99% confidence, leading to a retention time scheduled mass list containing 24 m/z values (5 peptides were identified in two charge states, and two peptides were identified with peak apex retention times differing more than 15 min and so were included twice). The peptides were targeted using a scan strategy comprising six narrow MS scans with windows of approximately 100 framing the m/z values on the mass list followed by targeted MS/MS scans of all retention time scheduled peptides on the mass list. FLS2 abundance can be approximated at 1.2 pmol/mg of microsomal protein, as determined from the quantification of flg22 binding sites and equimolar binding of flg22 by FLS2.39,40 We estimated that separating 1−2 μg of microsomal protein by SDS-PAGE (Supporting Information Figure 2) would result in 1 fmol or less of tryptic FLS2 peptides on column in our experiments. We measured three independent preparations of protein from microsomal fractions in this manner and consistently identified four native FLS2 peptides with 99% confidence in the WT and pub12 pub 13 backgrounds, two of which gave product ion signal peaks suitable for quantification with minimal differences in retention time (Figure 3). Therefore, FLS2 was consistently detected in all samples with exception of the f ls2 negative control, as expected. In addition, we observed a decrease in the levels of both quantified FLS2 peptides in WT plants following flg22 treatment. This was consistent for all three repetitions of the experiment with the exception of one peptide (VAHVSDFGTAR), which once showed an inverse relationship. This outlier may be attributed to the low abundance of the FLS2 peptides on column (estimated at less than 1 fmol, see above), close to the limit of quantification of our experimental system, which we determined to be 100 amol using BSA as a standard in conjunction with the relative complexity of the sample preparation, i.e., protein extraction followed by preparation of microsomal fraction and SDS-PAGE. The fold changes in abundance following flg22 treatment for each peptide were also quite consistent with the exception mentioned above. The observed decrease in FLS2 could be the result of FLS2 degradation triggered by flg22 treatment, as previously reported.37,44 Genetic data obtained from pub12 pub13 suggested that polyubiquitination is involved in the degradation of the activated receptor in the course of the downregulation of the immune response.37 However, we also repeatedly observed a decrease in FLS2 levels in the pub12 pub13 background with consistent fold changes in abundance following flg22 treatment for both peptides with the exception of the second experiment, where the decrease was marginal for the peptide VAHVSDFGTAR. Nevertheless, further analyses are required to confirm the observed changes in protein levels and to more precisely determine fold changes in FLS2 abundance.

However, in order to truly link NERD with biological function and to unravel its requirement in a biological context, the here determined overabundant proteins need to be tested, e.g., for relatively short half-lives in WT compared to increased half-lives in nerd mutants or inhibition of their degradation by application of proteasomal inhibitors in cell culture or cell-free systems. The probability of being NERD substrates can be relatively easily assessed for MC-starting proteins such as was done for the five ERFs and VRN2 in Gibbs et al.29 and MCstarting recombinant β-glucoronidase (MC-GUS). It will be experimentally much more sophisticated to rule out NERD-mediated degradation of the second half of Table 2. Here, for the eight members of diverse enzyme and protein classes, the identity of their N-termini is still elusive. As they are not among the list of MC-starting proteins, their first amino acid is comprised from the list of stabilizing residues (Gly, Ala, Ser, Thr, Val, and possibly also Ile or Leu) or a shielding methionine (Met). The mechanistics of altering their sequence in order to become recognizable by NERD E3 ligases needs to be investigated as a next step. One predestined possible activation pathway is N-terminal cleavage by endopeptidases, generating C-terminal fragments with shorter half-lives, e.g., after targeting signals and within prepro-proteins. Suggested reasons for such protein activation/inactivation and breakdown cascades are the acquisition of different molecular and cell biological features compared to the full-length protein or safeguarding protein function in the correct cellular compartment in the framework of protein quality control. Quantification of FLS2 in flg22-Elicited Arabidopsis Cells

Plants have a very effective innate cellular immune system that accommodates their sessile lifestyle.31 FLAGELLIN SENSITIVE 2 (FLS2) is the best characterized pattern recognition receptor (PRR) that activates pathogen-associated molecular pattern (PAMP)-triggered immunity (PTI) upon perception of the N-terminal epitope of bacterial flagellin in several plant species.32,33 FLS2 has an extracellular leucine-rich repeat (LRR) receptor domain, a transmembrane domain, and an intracellular kinase domain. FLS2 signaling requires a complex that includes the coreceptor BAK1 and the cytoplasmic kinase BIK1.32,42,43 Activation of the immune response results in a growth penalty, and prolonged activation can be adverse to plant fitness and must thus be tightly controlled. The attenuation of signaling is achieved through different processes. The function of receptor ubiquitination to mediate their endocytosis and sorting into the vacuole for degradation is well-known in various non-plant systems. 34 FLS2 was shown to be endocytosed upon ligand binding and subsequently degraded.35 This process is proposed to mediate signal downregulation. The link between ubiquitination and vesicle trafficking to modulate signaling via protein abundance is becoming more and more evident in plants.36 PUB12, PUB13, PUB22, PUB23, and PUB24 have all been shown to contribute to the attenuation of PAMP-triggered signaling in Arabidopsis.11,37,38 PUB12 and PUB13 are phosphorylated by BAK1 and were shown to subsequently interact with and polyubiquitinate FLS2. The reduction of FLS2 levels was reported to be impaired in pub12 pub13 double mutants, suggesting a role in its degradation.37 In the course of an extended study of E3 ubiquitin ligase function in FLS2 degradation and control of PTI, we sought to detect FLS2 in Arabidopsis WT plants and pub12 pub13 double knockout mutants after 30 min of treatment with flg22, the N4255

dx.doi.org/10.1021/pr500164j | J. Proteome Res. 2014, 13, 4246−4258

Journal of Proteome Research

Article

MS/MS scanning is independent of survey scans, yet it necessitates peptide retention time scheduling. The use of the targeted peptides themselves as retention time standards is preferable; however, we found that a mixture of synthetic peptide standards to interpolate target peptide retention times works equally well. The absolute sensitivity of the LTQOrbitrap Velos Pro was comparable to that of other state-ofthe-art instruments. The main drawback remains multiplexing capacity; however, mass lists of more than 100 m/z values can be processed effectively. While the LTQ-Orbitrap may not be the first choice, we found it to be a serious option for targeted proteomics applications. Having quantified the relative abundance of several MCstarting proteins as well as a diverse set of test proteins from different enzyme classes in nerd mutants and having detected genotype-specific changes in their abundance, we found the presented method to be a versatile setup for targeted proteomics that can also be used for low-abundance proteins and/or proteins with possibly short half-lives such as NERD target candidates. Regarding FLS2, we confirmed that, as reported, FLS2 levels decrease after activation, but notably, we observed a comparable tendency in pub12 pub13 mutants.



ASSOCIATED CONTENT

S Supporting Information *

Figure 1: MS1 and MS2 fill times in the different scan methods show the maximum injection time (max IT). Figure 2: SDSPAGE of Arabidopsis microsomal protein fraction. Table 1: Mass spectrometric parameters for all methods used. Table 2: Target list of N-end rule mutant analysis. Table 3: Analysis of possible N-end rule target proteins in wild-type plants. This material is available free of charge via the Internet at http:// pubs.acs.org.

Figure 3. Targeted PRM quantification of FLS2. Arabidopsis WT (ecotype Col-0), pub12 pub13, and f ls2 knockout mutant plants treated with 1 μM flg22 for 30 min. Approximately 100 ng of microsomal protein fraction was separated by SDS-PAGE. Product ion signals of two peptides identified with more than 99% confidence and a single peak apex, symmetric inflection points, and baseline resolution as well as exact alignment in the retention time of the product ion signals and no interfering signals were used for quantification. The log2 fold change following flg22 treatment is shown on top of the bars.





CONCLUSIONS Targeted PRM protein quantification easily achieves high specificity (>99%) by PSM without the expertise required for SRM method development. However, because peptide identification is necessary, a suitable MS/MS spectrum must be acquired. We found survey scan speed and the total instrument duty cycle time to be critical parameters limiting MS/MS spectral acquisition of low-abundance peptides with ion signal peaks appearing for only a few seconds in chromatographic elution at the limit of detection (LOD) and hence also limiting sensitivity. Utilizing a single full scan as a survey scan before MS/MS acquisition minimizes the duty cycle time, whereas partitioning the full MS into mass range scanning increases the S/N, potentially improving precursorbased quantification; however, this comes with a trade-off of extended duty cycle time and reduced sensitivity. Mass range scanning is advantageous when one or few proteins are being targeted with many peptides per protein and one has the luxury to identify only a small number of peptides suitable for quantification such as in our measurements of BSA and FLS2. When peptide identification must be maximized to quantify more proteins such as in the analysis of NERD substrates, it is advisible to tightly constrain the instrument duty cycle and use full scans. DDA MS/MS scanning places a strong limit on sensitivity because a precursor signal threshold must be surpassed in the preceding survey scan and is not recommended. Targeted peptide analysis is preferred because

AUTHOR INFORMATION

Corresponding Authors

*(N.D.) E-mail: [email protected]. Tel.: 0049 345 5582 1710. Fax: 0049 345 5582 1409. *(W.H.) E-mail: [email protected]. Tel.: 0049 345 5582 1411. Fax: 0049 345 5582 1409. Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS This work was supported by a grant for a Junior Research Group of the ScienceCampus Halle − Plant-Based Bioeconomy awarded to N.D; by a Ph.D. fellowship of the Landesgraduiertenfö rderung Sachsen-Anhalt awarded to C.N. Support comes from the Deutsche Forschungsgemeinschaft (DFG) Graduate Training Center GRK1026 “Conformational Transitions in Macromolecular Interactions” at Halle, Germany, and the Leibniz Institute of Plant Biochemistry at Halle, Germany. M.T. is financed by the Leibniz Association and the State of Sachsen-Anhalt and C.W.L. by the DFG SPP1212. The prt1-1 Arabidopsis mutant was a kind gift of Andreas Bachmair (Max F. Perutz Laboratories, Vienna, Austria), the prt6-3 mutant (GABI_270G04) was kindly provided by the GABI-Kat team via NASC - The European Arabidopsis Stock Centre (arabidopsis.info) and ate1 ate2 double mutants, from Emmanuelle Graciet (National University of Ireland Maynooth, Maynooth, Ireland). The expression vector containing the 4256

dx.doi.org/10.1021/pr500164j | J. Proteome Res. 2014, 13, 4246−4258

Journal of Proteome Research

Article

(20) Old, W. M.; Meyer-Arendt, K.; Aveline-Wolf, L.; Pierce, K. G.; Mendoza, A.; Sevinsky, J. R.; Resing, K. A.; Ahn, N. G. Comparison of label-free methods for quantifying human proteins by shotgun proteomics. Mol. Cell. Proteomics 2005, 4, 1487−502. (21) Krokhin, O. V.; Spicer, V. Peptide retention standards and hydrophobicity indexes in reversed-phase high-performance liquid chromatography of peptides. Anal. Chem. 2009, 81, 9522−30. (22) Gallien, S.; Peterman, S.; Kiyonami, R.; Souady, J.; Duriez, E.; Schoen, A.; Domon, B. Highly multiplexed targeted proteomics using precise control of peptide retention time. Proteomics 2012, 12, 1122− 33. (23) Yen, H. C.; Xu, Q.; Chou, D. M.; Zhao, Z.; Elledge, S. J. Global protein stability profiling in mammalian cells. Science 2008, 322, 918− 23. (24) Yen, H. C.; Elledge, S. J. Identification of SCF ubiquitin ligase substrates by global protein stability profiling. Science 2008, 322, 923− 9. (25) Vierstra, R. D. The ubiquitin-26S proteasome system at the nexus of plant biology. Nat. Rev. Mol. Cell Biol. 2009, 10, 385−97. (26) Varshavsky, A. The N-end rule pathway and regulation by proteolysis. Protein Sci. 2011, 20, 1298−1345. (27) Holman, T. J.; Jones, P. D.; Russell, L.; Medhurst, A.; Ubeda Tomas, S.; Talloji, P.; Marquez, J.; Schmuths, H.; Tung, S. A.; Taylor, I.; Footitt, S.; Bachmair, A.; Theodoulou, F. L.; Holdsworth, M. J. The N-end rule pathway promotes seed germination and establishment through removal of ABA sensitivity in Arabidopsis. Proc. Natl. Acad. Sci. U. S. A. 2009, 106, 4549−54. (28) Yoshida, S.; Ito, M.; Callis, J.; Nishida, I.; Watanabe, A. A delayed leaf senescence mutant is defective in arginyl-tRNA:protein arginyltransferase, a component of the N-end rule pathway in Arabidopsis. Plant J. 2002, 32, 129−37. (29) Gibbs, D. J.; Lee, S. C.; Isa, N. M.; Gramuglia, S.; Fukao, T.; Bassel, G. W.; Correia, C. S.; Corbineau, F.; Theodoulou, F. L.; BaileySerres, J.; Holdsworth, M. J. Homeostatic response to hypoxia is regulated by the N-end rule pathway in plants. Nature 2011, 479, 415−8. (30) Licausi, F.; Kosmacz, M.; Weits, D. A.; Giuntoli, B.; Giorgi, F. M.; Voesenek, L. A.; Perata, P.; van Dongen, J. T. Oxygen sensing in plants is mediated by an N-end rule pathway for protein destabilization. Nature 2011, 479, 419−22. (31) Jones, J. D.; Dangl, J. L. The plant immune system. Nature 2006, 444 (7117), 323−329. (32) Chinchilla, D.; Zipfel, C.; Robatzek, S.; Kemmerling, B.; Nurnberger, T.; Jones, J. D.; Felix, G.; Boller, T. A flagellin-induced complex of the receptor FLS2 and BAK1 initiates plant defence. Nature 2007, 448, 497−500. (33) Robatzek, S.; Bittel, P.; Chinchilla, D.; Köchner, P.; Felix, G.; Shiu, S. H.; Boller, T. Molecular identification and characterization of the tomato flagellin receptor LeFLS2, an orthologue of Arabidopsis FLS2 exhibiting characteristically different perception specificities. Plant Mol Biol. 2007, 64 (5), 539−547. (34) MacGurn, J. A.; Hsu, P. C.; Emr, S. D. Ubiquitin and membrane protein turnover: from cradle to grave. Annu. Rev. Biochem. 2012, 81, 231−59. (35) Robatzek, S.; Chinchilla, D.; Boller, T. Ligand-induced endocytosis of the pattern recognition receptor FLS2 in Arabidopsis. Genes Dev. 2006, 20, 537−42. (36) Furlan, G.; Klinkenberg, J.; Trujillo, M. Regulation of plant immune receptors by ubiquitination. Front. Plant Sci. 2012, 3, 238. (37) Lu, D.; Lin, W.; Gao, X.; Wu, S.; Cheng, C.; Avila, J.; Heese, A.; Devarenne, T. P.; He, P.; Shan, L. Direct ubiquitination of pattern recognition receptor FLS2 attenuates plant innate immunity. Science 2011, 332, 1439−42. (38) Stegmann, M.; Anderson, R. G.; Ichimura, K.; Pecenkova, T.; Reuter, P.; Zarsky, V.; McDowell, J. M.; Shirasu, K.; Trujillo, M. The ubiquitin ligase PUB22 targets a subunit of the exocyst complex required for PAMP-triggered responses in Arabidopsis. Plant Cell 2012, 24, 4703−16.

MBP-FLS2 kinase domain fusion was a gift from Cyril Zipfel (Sainsbury Laboratory, Norwich, UK), and the pub12 pub13 double mutants were kindly provided by Libo Shan (Texas A&M University, College Station, TX, USA).



REFERENCES

(1) Schulze, W. X.; Usadel, B. Quantitation in mass-spectrometrybased proteomics. Annu. Rev. Plant Biol. 2010, 61, 491−516. (2) Marx, V. Targeted proteomics. Nat. Methods 2013, 10, 19−22. (3) Picotti, P.; Aebersold, R. Selected reaction monitoring-based proteomics: workflows, potential, pitfalls and future directions. Nat. Methods 2012, 9, 555−66. (4) Lange, V.; Picotti, P.; Domon, B.; Aebersold, R. Selected reaction monitoring for quantitative proteomics: a tutorial. Mol. Syst. Biol. 2008, 4, 222. (5) Sherman, J.; McKay, M. J.; Ashman, K.; Molloy, M. P. Unique ion signature mass spectrometry, a deterministic method to assign peptide identity. Mol. Cell. Proteomics 2009, 8, 2051−62. (6) Ramanathan, R.; Jemal, M.; Ramagiri, S.; Xia, Y. Q.; Humpreys, W. G.; Olah, T.; Korfmacher, W. A. It is time for a paradigm shift in drug discovery bioanalysis: from SRM to HRMS. J. Mass Spectrom. 2011, 46, 595−601. (7) Michalski, A.; Damoc, E.; Hauschild, J. P.; Lange, O.; Wieghaus, A.; Makarov, A.; Nagaraj, N.; Cox, J.; Mann, M.; Horning, S. Mass spectrometry-based proteomics using Q Exactive, a high-performance benchtop quadrupole Orbitrap mass spectrometer. Mol. Cell. Proteomics 2011, 10, M111.011015. (8) Peterson, A. C.; Russell, J. D.; Bailey, D. J.; Westphall, M. S.; Coon, J. J. Parallel reaction monitoring for high resolution and high mass accuracy quantitative, targeted proteomics. Mol. Cell. Proteomics 2012, 11, 1475−88. (9) Gallien, S.; Duriez, E.; Crone, C.; Kellmann, M.; Moehring, T.; Domon, B. Targeted proteomic quantification on quadrupole-orbitrap mass spectrometer. Mol. Cell. Proteomics 2012, 11, 1709−23. (10) Graciet, E.; Wellmer, F. The plant N-end rule pathway: structure and functions. Trends Plant Sci. 2010, 15, 447−53. (11) Trujillo, M.; Ichimura, K.; Casais, C.; Shirasu, K. Negative regulation of PAMP-triggered immunity by an E3 ubiquitin ligase triplet in Arabidopsis. Curr. Biol. 2008, 18, 1396−401. (12) Garzon, M.; Eifler, K.; Faust, A.; Scheel, H.; Hofmann, K.; Koncz, C.; Yephremov, A.; Bachmair, A. PRT6/At5g02310 encodes an Arabidopsis ubiquitin ligase of the N-end rule pathway with arginine specificity and is not the CER3 locus. FEBS Lett. 2007, 581, 3189−96. (13) Graciet, E.; Walter, F.; Maoileidigh, D. O.; Pollmann, S.; Meyerowitz, E. M.; Varshavsky, A.; Wellmer, F. The N-end rule pathway controls multiple functions during Arabidopsis shoot and leaf development. Proc. Natl. Acad. Sci. U.S.A. 2009, 106, 13618−23. (14) Bachmair, A.; Becker, F.; Schell, J. Use of a reporter transgene to generate arabidopsis mutants in ubiquitin-dependent protein degradation. Proc. Natl. Acad. Sci. U.S.A. 1993, 90, 418−21. (15) Potuschak, T.; Stary, S.; Schlogelhofer, P.; Becker, F.; Nejinskaia, V.; Bachmair, A. PRT1 of Arabidopsis thaliana encodes a component of the plant N-end rule pathway. Proc. Natl. Acad. Sci. U.S.A. 1998, 95, 7904−8. (16) Stary, S.; Yin, X. J.; Potuschak, T.; Schlogelhofer, P.; Nizhynska, V.; Bachmair, A. PRT1 of Arabidopsis is a ubiquitin protein ligase of the plant N-end rule pathway with specificity for aromatic aminoterminal residues. Plant Physiol. 2003, 133, 1360−6. (17) Laemmli, U. K. Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 1970, 227, 680−5. (18) Perkins, D. N.; Pappin, D. J.; Creasy, D. M.; Cottrell, J. S. Probability-based protein identification by searching sequence databases using mass spectrometry data. Electrophoresis 1999, 20, 3551−67. (19) MacLean, B.; Tomazela, D. M.; Shulman, N.; Chambers, M.; Finney, G. L.; Frewen, B.; Kern, R.; Tabb, D. L.; Liebler, D. C.; MacCoss, M. J. Skyline: an open source document editor for creating and analyzing targeted proteomics experiments. Bioinformatics 2010, 26, 966−8. 4257

dx.doi.org/10.1021/pr500164j | J. Proteome Res. 2014, 13, 4246−4258

Journal of Proteome Research

Article

(39) Bauer, Z.; Gomez-Gomez, L.; Boller, T.; Felix, G. Sensitivity of different ecotypes and mutants of Arabidopsis thaliana toward the bacterial elicitor flagellin correlates with the presence of receptorbinding sites. J. Biol. Chem. 2001, 276, 45669−76. (40) Robatzek, S.; Wirthmueller, L. Mapping FLS2 function to structure: LRRs, kinase and its working bits. Protoplasma 2013, 250, 671−81. (41) Schwessinger, B.; Roux, M.; Kadota, Y.; Ntoukakis, V.; Sklenar, J.; Jones, A.; Zipfel, C. Phosphorylation-dependent differential regulation of plant growth, cell death, and innate immunity by the regulatory receptor-like kinase BAK1. PLoS Genet. 2011, 7, e1002046. (42) Lu, D.; Wu, S.; Gao, X.; Zhang, Y.; Shan, L.; He, P. A receptorlike cytoplasmic kinase, BIK1, associates with a flagellin receptor complex to initiate plant innate immunity. Proc. Natl. Acad. Sci. U.S.A. 2010, 107, 496−501. (43) Zhang, J.; Li, W.; Xiang, T.; Liu, Z.; Laluk, K.; Ding, X.; Zou, Y.; Gao, M.; Zhang, X.; Chen, S.; Mengiste, T.; Zhang, Y.; Zhou, J. M. Receptor-like cytoplasmic kinases integrate signaling from multiple plant immune receptors and are targeted by a Pseudomonas syringae effector. Cell Host Microbe 2010, 7, 290−301. (44) Smith, J. M.; Salamango, D. J.; Leslie, M. E.; Collins, C. A.; Heese, A. Sensitivity to Flg22 is modulated by ligand-induced degradation and de novo synthesis of the endogenous flagellinreceptor FLAGELLIN-SENSING2. Plant Physiol. 2014, 164, 440−54. (45) Tasaki, T.; Sriram, S. M.; Park, K. S.; Kwon, Y. T. The N-end rule pathway. Annu. Rev. Biochem. 2012, 81, 261−89. (46) Kleinboelting, N.; Huep, G.; Kloetgen, A.; Viehoever, P.; Weisshaar, B. GABI-Kat SimpleSearch: new features of the Arabidopsis thaliana T-DNA mutant database. Nucleic Acids Res. 2012, 40, D1211− D1215.

4258

dx.doi.org/10.1021/pr500164j | J. Proteome Res. 2014, 13, 4246−4258