Quantitative Analysis of mTRAQ-Labeled Proteome Using Full MS Scans Un-Beom Kang,† Jeonghun Yeom,†,§ Hoguen Kim,‡ and Cheolju Lee*,†,§ Life Sciences Division, Korea Institute of Science and Technology, Seoul 136-791, Korea, Department of Pathology, Yonsei University College of Medicine, Seoul 120-752, Korea, and Department of Biomolecular Science, University of Science and Technology, Daejeon 305-333, Korea Received December 1, 2009
Abstract: Proteomic techniques are mostly used these days to identify proteins in a biological sample. Quantification of the differences between two or more physiological conditions, such as disease or no disease, has become an increasingly challenging task in proteomics. Mass tags introducing stable isotopes into peptides or proteins provide means for quantification in mass spectrometry. The mass tags are recognized by mass spectrometry and at the same time provide quantitative information. In the current study, we introduce mTRAQ for the purpose of quantification by full MS scans. Although mTRAQ reagent was initially designed for multiple reaction monitoring, we verified the utility of mTRAQ for MS1-based relative quantification using standard protein mixtures and blood plasma samples. mTRAQlabeled peptides showed better quality MS2 spectra with increased XCorr values in a SEQUEST search output than corresponding unlabeled peptides. The improved spectral quality was due mostly to the enhanced matching of b-type ions. By combining mTRAQ with ICAT and applying them to colon cancer tissues, we identified and quantified a total of 3,320 proteins. mTRAQ covered a wider range of the proteome than did ICAT, and only 1053 proteins were shared by the two independent methods. Our results suggest the usefulness of mTRAQ, alone or in combination with ICAT, as a comparative profiling method in quantitative proteomics. Keywords: Quantitative analysis • mTRAQ • ICAT • MS1based quantification • XCorr • b-type ion
Introduction In recent times, proteomic techniques have been increasingly used in the biomarker field to simultaneously identify and quantify as many proteins as possible.1 Hence, acquiring more proteins with larger dynamic depth is the most important goal of proteomics. Despite the phenomenal impact of mass spec* Corresponding author: Cheolju Lee, Life Sciences Division, Korea Institute of Science and Technology, 39-1 Hawolgok, Seongbuk, Seoul 136791, Korea. E-mail:
[email protected]. Fax: +82-2-958-6919. † Korea Institute of Science and Technology. § University of Science and Technology. ‡ Yonsei University College of Medicine.
3750 Journal of Proteome Research 2010, 9, 3750–3758 Published on Web 05/13/2010
trometry and peptide separation techniques, the identification and quantification of all proteins in a biological system is still a technical challenge.2 Quantitative proteomic techniques utilizing stable isotopes are classified via isotope introduction. In 2002, Mann and colleagues devised a method for introducing stable isotopes using metabolic labeling during cell growth and division thusly named stable isotope labeling by amino acids in cell culture (SILAC).3 The cellular proteome can be labeled by any metabolic precursor from the inexpensive uniform 15N labeling using 15NH4Cl to specific labeling with isotopeenriched amino acids, such as 13C6-arginine and 13C6-lysine.4,5 Unlike SILAC, which is an in vivo labeling strategy, many in vitro labeling approaches have been developed and are employed when metabolic labeling is impossible, for example, human tissues. Representatively, these methods include isotopecoded affinity tag (ICAT) analysis, in which cysteine residues are specifically derivatized with reagent containing either 12C9 or 13C9 as well as a biotin group for subsequent affinity purification followed by MS,6 and more recently isobaric tags for relative and absolute quantification (iTRAQ), which targets the peptide N-terminus and ε-amino group of lysine residues.7 These two methods share the same basic principle of introducing stable isotopes at specific sites by chemical reaction. However, the noteworthy distinguishable feature is that relative quantification by iTRAQ is performed via MS2 while that of ICAT is done through MS1. It is sometimes required that the absolute amounts of proteins of interest in a given biological sample be known. This issue can be addressed by a method called multiple reaction monitoring (MRM). MRM is the current method of choice for quantifying small molecules, including drugs, metabolites, and peptides that are present at low levels.8 With MRM, the mass spectrometer is set up to monitor only specific mass-to-charge (m/z) values of interest; as a consequence, the probability of detecting a low-level peptide in the presence of a complex mixture of peptides will be increased. Quantification of the absolute amount of a target protein by MRM requires corresponding specific peptides as quantification standards that are isotopically different from the native peptides. A recently developed variation of iTRAQ, called mTRAQ, provides a means for performing absolute quantification by MRM. Specifically, there are two versions of the mTRAQ reagent available: the “heavy” version is identical to the iTRAQ 117 tag, while the “light” version is devoid of any intentional isotopic enrichment.9,10 Consequently, labeling of a peptide with the light version adds 10.1021/pr9011014
2010 American Chemical Society
Analysis of mTRAQ-Labeled Proteome from Full MS Scans 140 Da, whereas 144 Da is gained by labeling with the heavy version. This mass difference allows unique MRM transitions to be generated for any given peptide labeled with mTRAQ tags. In turn, this implies that, in a sample mixture where peptides originating from different sources are tagged separately with heavy and light labels, the two versions can be monitored independently and distinctly. Among the different techniques that employ chemical labeling, ICAT and iTRAQ are the two most commonly practiced in LC-shotgun-based quantitative proteomics. Although ICAT can detect low-abundance proteins by affinity purification, it would miss identification of proteins with few or no cysteine residues, and frequently lose information for post-translational modifications.11 These limitations have been somewhat solved by the iTRAQ method, which covers any peptides of a particular protein. However, there also are restrictions in the choice of the mass instrument used to analyze iTRAQ-labeled peptides. In this study, we noticed the mass difference between the two mTRAQ tags and exploited it for protein quantification in MS1 level using an LTQ-Orbitrap hybrid mass spectrometer. The labeling strategy was exactly the same as iTRAQ, with specificity to the primary amine. This alleviated the task of optimizing the labeling reaction. Although mTRAQ has been specifically developed for MRM quantification,9 we confirmed it can also be successfully applied to relative quantification using full MS scans. Especially, labeled peptides generally displayed enhanced MS2 spectral quality compared to corresponding unlabeled peptides. The mTRAQ-based quantification was applied to colon cancer tissues, and the resulting data were compared to those obtained by the ICAT labeling method. The range of the proteome covered by mTRAQ was wider than but complementary to ICAT. The introduction of mTRAQ, alone or in combination with ICAT, to proteome analysis will therefore advance quantitative proteomics.
Experimental Procedures Materials. Reagent grade chemicals and proteins were purchased from Sigma Aldrich (St. Louis, MO) and Fisher Scientific (Pittsburgh, PA). mTRAQ and ICAT reagents were obtained from Applied Biosystems (Framingham, MA). Standard Protein Mixtures. Two kinds of standard protein mixtures were prepared for mTRAQ quantification testing by mixing 6 bovine proteins. Each mixture consisted of R-lactalbumin, β-casein, serotransferrin, R-S1-casein, R-S2-casein, and pancreatic ribonuclease in 50 mM Tris pH 8.0 at different amounts: 10, 10, 20, 25, 25, and 10 µg for standard mixture 1 (Std1); 10, 20, 10, 5, 5, and 50 µg for standard mixture 2 (Std2). Human Plasma Sample. Blood samples were collected from normal healthy volunteers at Seoul National University Hospital (Seoul, Korea). The use of human samples for research purposes was authorized by the Institutional Review Board of Seoul National University Hospital, and all volunteers agreed to take part in the experiment by signing their name on an informed consent document. Plasma samples were depleted of the six most abundant serum proteins using a multiple-affinity MARS column (Agilent Technologies, Palo Alto, CA), and the unbound fraction was concentrated using a Microcon ultrafiltration system (3000 Da cutoff; Millipore, Billerica, MA). Protein concentration was determined by Bradford assay. Colon Cancer Tissue Sample. A total of 3 MSS type colorectal carcinomas and matched nontumorous colonic mucosal tissues from 3 patients were used. Detailed clinical information of the tissue samples was described previously.12 Authorization
technical notes for use of these tissues for research purposes was obtained from the Institutional Review Board of Yonsei University of College of Medicine. The frozen tissue samples (100 mg) were washed twice with chilled PBS and lysed by sonication on ice in 600 µL of lysis buffer consisting of 7 M urea, 2 M thiourea, 4% 3-[(3cholamidopropyl)-dimethylammonio]-1-propane sulfonate, and 30 mM Tris. The lysates were cleared by centrifugation at 14 000g for 5 min at 4 °C. All 3 lysates from cancer tissues and all 3 lysates from matched normal tissues were pooled separately and precipitated with acetone. The pellet was dissolved in 50 mM Tris buffer (pH 8.0) containing 3 M Urea for mTRAQ analysis or in 50 mM Tris-HCl (pH 8.3) containing 6 M urea, 0.05% SDS, and 5 mM EDTA for ICAT. mTRAQ Labeling. Standard protein mixtures (Std1 and Std2), depleted plasma samples, and tissue extracts were labeled with mTRAQ reagent according to the manufacturer’s protocol. Samples were reduced with 50 mM tris(2-carboxyethyl)phosphine (TCEP) for 1 h at 60 °C, treated with 200 mM methyl methane-thiosulfonate for 10 min at 25 °C, and then diluted 10-fold with 50 mM Tris (pH 8.0). Digestion was performed with sequencing-grade trypsin (Promega, Madison, WI) at 37 °C overnight at a protein/trypsin molar ratio of 10:1. Tryptic digests were desalted using a C18 SPE cartridge and dried in vacuo. The dried samples were reconstituted in 500 mM triethyl ammonium bicarbonate and incubated with mTRAQ reagent at 25 °C for 1 h. For standard protein labeling, each of the reconstituted tryptic digests of Std1 and Std2 were divided into two equal parts. One part of each standard was labeled with light mTRAQ reagent (referred to as Std1L and Std2L) while the other was labeled with heavy mTRAQ reagent (referred to as Std1H and Std2H). Std1L was mixed with Std2H, and Std1H with Std2L. Prior to MS analysis, 0.4 µg of the mTRAQ labeled standard protein mixture was added to 1 µg of the trypsin-digested plasma proteome in order to test performance under more realistic conditions. For labeling of the plasma proteome, 100 µg of trypsin-digested plasma sample was divided into two equal parts, each of which (50 µg) was labeled with light and heavy mTRAQ. After reaction for 1 h, samples were dried in vacuo, redissolved in 0.1% TFA, desalted with an MCX cartridge and dried again. For the colon cancer tissues, the tumor tissue extract (100 µg) was labeled with heavy reagent while the nontumor tissue extract (100 µg) was labeled with light reagent. After reaction for 1 h, the light and heavy mTRAQ-labeled tissue extracts were combined and applied to a Polysulfethyl A column (Western ¨ KTA Explorer system (GE Analytical, Murrieta, CA) using an A Healthcare Biosciences, Uppsala, Sweden). Samples were eluted with a 40-min gradient from zero to 0.4 M KCl, and were collected in 13 fractions. The SCX fractions were desalted using a C18 SPE cartridge and were dried in vacuo. ICAT Labeling. The colon cancer tissue samples used for mTRAQ labeling were also labeled with ICAT. The tumor tissue was labeled with heavy reagent, whereas the adjacent normal section was labeled with light reagent. ICAT labeling was performed in 50 mM Tris-HCl (pH 8.3) containing 6 M urea, 0.05% SDS, and 5 mM EDTA. The tissue extracts (100 µg each) were reduced with 250 mM TCEP for 30 min before ICAT labeling. The labeling reaction was performed using 350 nmol of ICAT reagent with gentle shaking for 2 h at 37 °C, and was terminated with 1.75 µmol DTT for an additional 5 min. The heavy and light ICAT-labeled samples were mixed, diluted 10fold with 50 mM Tris (pH 8.0), and digested with 5 µg of trypsin Journal of Proteome Research • Vol. 9, No. 7, 2010 3751
technical notes
Kang et al.
Table 1. Experimental Strategy for the Establishment MS1-Based Quantitative Method Using mTRAQ experiment
plasma proteome profiling
mTRAQ quantification test 1
mTRAQ quantification test 2
Sample Labeling Fractionation prior to LC-MS Gradient time for RPLC-MSa Database search Data validation Quantification
plasma unlabeled, mTRAQ-L, mTRAQ-H no fractionation
standard protein mixtures mTRAQ no fractionation
plasma mTRAQ no fractionation
40 min
40 min
40 min
SEQUEST TPP -
SEQUEST TPP XPRESS
SEQUEST TPP XPRESS
a
colon cancer tissue ICAT, mTRAQ SCX 16 fractions (ICAT) SCX 13 fractions (mTRAQ) 90 min (ICAT) 150 min (mTRAQ) SEQUEST TPP XPRESS
10-40% acetonitrile.
(Promega) for 16 h at 37 °C. The reaction was quenched at 0.5% phosphoric acid. The digest sample was applied to a Polysulfethyl A column (Western Analytical) equilibrated with 10 mM KH2PO4 in 25% acetonitrile (pH 3.0), eluted with a 40-min gradient from zero to 0.4 M KCl, and collected in 16 fractions. The SCX fractions were neutralized by the addition of 10 vol of 2× PBS, loaded on an avidin catridge (Applied Biosystems), and then washed with PBS and 50 mM ammonium bicarbonate in 20% methanol, pH 8.3. ICAT-labeled peptides were eluted with 0.4% TFA in 30% acetonitrile, dried in vacuo, redissolved in 90 µL of 95% TFA, incubated at 37 °C for 2 h to cleave off the biotin moiety, and finally dried again. Liquid Chromatography and Tandem Mass Analysis. Samples were reconstituted in 0.4% acetic acid. An aliquot (∼1 µg) was then injected into a reversed-phase Magic C18aq column (15 cm ×75 µm) on an Eksigent MDLC system at a flow rate of 300 nL/min. The column was equilibrated with 95% buffer A (0.1% formic acid in H2O) and 5% buffer B (0.1% formic acid in acetonitrile) prior to use. Several gradient conditions were adopted depending on the nature of each sample. The peptides from standard protein mixtures and plasmas were eluted with a linear gradient of 10-40% Buffer B over 40 min. The same gradient was applied over 90 min for ICAT samples and over 150 min for mTRAQ samples. An HPLC system was coupled to an LTQ XL-Orbitrap mass spectrometer (Thermo Scientific, San Jose, CA). The spray voltage was set to 1.9 kV, and the temperature of the heated capillary was set to 250 °C. Survey full-scan MS spectra (m/z 300-2000) were acquired in the Orbitrap with 1 microscan and a resolution of 100 000 allowing the preview mode for precursor selection and charge-state determination. MS/MS spectra of the five most intense ions from the preview survey scan were acquired in the ion-trap concurrently to the full-scan acquisition in the Orbitrap with the following options: isolation width, 10 ppm; normalized collision energy, 35%; dynamic exclusion duration, 30 s. Precursors with unmatched charge states were discarded during data-dependent acquisition. For the tissue sample, singly charged precursors were also excluded for the MS/MS scan. Data were acquired using Xcalibur software v2.0.7. Database Search and Data Analysis. The data for tandem mass spectra were generated by the Extract-msn program (v3) of Bioworks software (v3.2) with the following parameters: minimum ion count threshold, 15; minimum intensity, 100. The acquired MS/MS spectra were searched using SEQUEST (TurboSequest version 27, revision 12) against the human International Protein Index database, which includes 72 065 protein entries (IPI, versions 3.44, European Bioinformatics Institute, http://www.ebi.ac.uk/IPI) as well as known contaminants, with 3752
application experiment
Journal of Proteome Research • Vol. 9, No. 7, 2010
options of no enzyme, 0.5 Da mass tolerance for MS/MS and 15 ppm mass tolerance for MS. For ICAT-labeled samples, the ICAT option (227.1270 Da as fixed modification plus +9.0306 Da as variable modification) for cysteine residues was used, and for mTRAQ-labeled samples, the mTRAQ option (140.095 Da as fixed modification plus +4.0071 Da as variable modification) for the N-terminus and lysine residues was used along with a fixed modification of 45.9877 Da for cysteine residues. Variable modification of methionine oxidation (+15.9949 Da) was commonly allowed for all cases. For the MS/MS spectra of the standard protein mixtures, a compound database consisting of IPI human v3.44 and IPI bovine v3.30 was used with all options the same as the plasma data set. To evaluate the contribution of b- or y-ion fragments to the matching of experimental spectra to theoretical spectra, a SEQUEST search that considered only b or y ions was performed in parallel on the unlabeled and mTRAQ labeled plasma samples. Briefly, either the b or y check box was deselected in the ‘Ions and Ion Series Calculated’ section of the SEQUEST search parameters. All other options were maintained. Peptide assignment and quantification were performed using the Trans-Proteomic Pipeline (TPP, version 4.0, http://www. proteomecenter.org). The SEQUEST search output was used as an input for the pepXML module with trypsin restriction and ‘monoisotopic masses’ as options. The Peptide-Prophet was applied with the ‘accurate mass binning’ option. Peptides with probabilities greater than 0.05 were included in the subsequent Protein-Prophet, and proteins having probabilities more than 0.9 were gathered (the sensitivity and error rates are described in Supplementary Table 1). Quantification analysis was achieved using XPRESS during TPP analysis. For the ICAT-labeled sample, 9.0306 Da of mass difference and 0.05 Da of ‘XPRESS mass tolerance’ were used. For the mTRAQlabeled sample, 4.0071 Da was used as the XPRESS mass difference on N-terminus and lysine residues, and 0.05 Da as ‘XPRESS mass tolerance’.
Results In our attempts to apply mTRAQ labeling to the analysis of human plasma proteins, we repeatedly observed that the majority of the MS2 spectra of our target peptides was of high quality insofar that they were easy to manually interpret. Unlike iTRAQ, mTRAQ is a nonisobaric tag originally designed for MRM analysis. However, we designed a systematic study to evaluate the unconventional use of mTRAQ as an MS1quantification tag in comparative profiling where chemical isotopic labeling is needed (Table 1). Initially, we assessed the spectral quality of mTRAQ-labeled trypsin-digested plasma
Analysis of mTRAQ-Labeled Proteome from Full MS Scans
technical notes
Figure 1. Improved MS/MS spectral quality by mTRAQ labeling. (A) Venn diagram for peptides (left) and proteins (right) which were identified and quantified from unlabeled and mTRAQ-labeled plasma sample. The number of shared peptides with the same charge state appears in parentheses. (B) Log2-transformed fold change of SEQUEST XCorr values between unlabeled and mTRAQ-labeled peptides of human plasma (left), and the distribution pattern of XCorr fold change according to the charge state and the C-terminal residue (right).
proteins, tested quantification of standard protein mixtures and plasma proteins, and finally applied mTRAQ labeling in combination with ICAT to the analysis of colon cancer tissues. Enhanced Spectral Quality via mTRAQ Labeling. The tryptic peptides of unlabeled and mTRAQ-labeled plasma proteins were analyzed by a single 40-min LC-MS/MS run, leading to the identification of 1452 and 1256 unique peptides by SEQUEST and TPP (Figure 1A). From both data sets, 934 peptides were commonly identified, 625 of which were identified from precursors having the same charge state. We collected the MS2 spectra of these 625 peptides from both data sets. If there was more than one MS2 spectrum for a single peptide, we chose the spectrum with the highest XCorr value. About 80% of these 625 spectra showed increased XCorr values upon mTRAQ labeling (Figure 1B; left panel). The increase in XCorr values was more prominent for peptides with charge state +2 than those with charge state +3 or more and for peptides with lysine than those with arginine (Figure 1B; right panel). As for XCorr, another identification score deltCn was also increased in most cases (data not shown). Manual inspection of matched spectra found that mTRAQlabeled peptides frequently produced b ions, unlike normal tryptic peptides from which y ions were produced more than b ions due to the presence of basic residues at the C-terminus. One example is shown in Figure 2A. To determine whether the enhanced b ion fragmentation was a causal factor for the increase in XCorr, we checked the degree of b and y ion matching in each spectrum compared to the identified peptide sequence. To estimate the degree of b and y ion matching in the MS2 spectra, a SEQUEST search was performed in which only b or y fragments were considered. Figure 2B shows how the fragment ion matching varies depending on mTRAQ labeling. Peptides with increased XCorr values showed a greater degree of both b and y ion matching after mTRAQ labeling. This increment was larger for b ion than for y ion. In contrast,
peptides with decreased XCorr values displayed lower b and y ion matching, with the greatest effect on y series ions presumably as a compensation for the enhanced b ion score (Figure 2B). In a control experiment comparing the light and heavy mTRAQ-labeled samples, there was no difference in Xcorr values (data not shown). Quantification Using mTRAQ: Two Test Cases. To evaluate the utility of mTRAQ for relative quantification from full MS scans, we applied our labeling method to the standard protein mixtures and human blood plasma sample. Two standard mixtures (Std1 and Std2) contained the same six bovine proteins at known but different concentrations. The protein mixtures were labeled, combined, and then added to a trypsindigested human plasma proteome matrix in order to construct more realistic sample conditions. In the two quantification tests (Std1L vs Std2H and Std2L vs Std1H), all 6 proteins were unambiguously identified with 10 unique peptides for ribonuclease at the lowest case and 50 unique peptides for serotransferrin at the largest case (Supplementary Tables 4 and 5). All observed peptides were confirmed as fully derivatized at their N-termini and lysine side chains. The ratios of mTRAQ-labeled standard mixtures showed the same variability as the expected values. The mean differences between the observed and expected quantities for the six proteins ranged between 5% and 27.6% and CV values ranged between 0% and 15.5% (Table 2). We then attempted to quantify plasma proteins. The trypsindigested plasma proteome was split into two equal parts, one of which was labeled with light mTRAQ and the other with heavy mTRAQ prior to being combined and analyzed. Exactly 867 peptides (PeptideProphet P > 0.9) were quantified (Supplementary Table 6). Log2-ratios showed a bell-shaped distribution (mean ( SD ) -0.0009 ( 0.4537) with a mean light-to-heavy ratio of 0.9994 (Figure 3A). Ninety-five percent of the peptides were in the range of 0.67-1.56. When log2-ratios were plotted as a function of the molecular weight of identified peptides, Journal of Proteome Research • Vol. 9, No. 7, 2010 3753
technical notes
Kang et al.
Figure 2. Enhanced b ion matching by mTRAQ labeling. (A) Tandem mass spectrum of unlabeled and mTRAQ-labeled peptide (nEWFWDLATGTMK), where n* and K* denotes mTRAQ-labeled N-terminus and lysine residues, respectively. (B) Box-plots of log2transformed ratio of fragment ion matching score. The fragment ion matching score is the Xcorr value obtained by considering only b or y ions during SEQUEST search. Identified peptides were categorized into two groups based on the change direction of original XCorr value. The fragment ion matching scores are compared between unlabeled and mTRAQ-labeled peptides. Table 2. Quantification of Protein Standard Mixtures Using mTRAQ sample protein
R-Lactalbumin β-Casein Serotransferrin R-S1-casein R-S2-casein Ribonuclease pancreatic
Std1L vs Std2HQuantification Using mTRAQ
Std1H vs Std2LQuantification Using mTRAQ
observerd ratio mean ( SDb
expected ratio
error (%)
coverage (%)
XNPc
USPd
observerd ratio mean ( SDb
expected ratio
error (%)
coverage (%)
XNPc
USPd
0.83 ( 0.06 0.43 ( 0.00 1.87 ( 0.13 3.93 ( 0.37 3.62 ( 0.23 0.17 ( 0.02
1 0.5 2 5 5 0.2
-17 -14 -6.5 -21.4 -27.6 -15
39.4 33 48.3 45.3 42.3 52.7
24 2 52 34 23 21
22 14 50 26 20 13
0.94 ( 0.06 1.56 ( 0.06 0.42 ( 0.04 0.18 ( 0.01 0.21 ( 0.02 4.52 ( 0.7
1 2 0.5 0.2 0.2 5
-6 -22 -16 -10 5 -9.6
39.4 37.1 42.5 37.9 39.6 52.7
25 5 46 26 24 17
22 13 42 18 20 10
a Std1L, light-mTRAQ-labeled standard protein mixture 1; Std1H, heavy-mTRAQ-labeled standard protein mixture 1; Std2L, light-mTRAQ-labeled standard protein mixture 2; Std2H, heavy-mTRAQ-labeled standard protein mixture 2. b SD, standard deviation. c XNP, the number of a quantified peptide by XPRESS. d USP, the number of a unique stripped peptide.
the distribution was slightly skewed toward the heavy-labeled part at the higher molecular weight region (Figure 3B). Actually, the mean L/H ratio for larger peptides (>2000 Da) was 0.94 (mean ( SD of log2-ratios ) -0.0875 ( 0.4062). This reflects a spectral overlap between the fifth isotope peak of the lightlabeled part and the monoisotopic peak of the heavy-labeled part. However, the error originating from the spectral overlap was almost negligible considering the overall distribution of 3754
Journal of Proteome Research • Vol. 9, No. 7, 2010
L/H ratios, and moreover, would be corrected using the natural abundance of 13C and 15N isotopes. Quantification Using mTRAQ and ICAT: An Application. The mTRAQ quantification method was applied to colon cancer tissue extract. Colon cancer tissue and adjacent normal tissue samples were labeled with heavy and light mTRAQ, respectively, and pooled together into a sample set. In a parallel experiment, the same samples were labeled with ICAT. The
technical notes
Analysis of mTRAQ-Labeled Proteome from Full MS Scans
of the peptides showed increased XCorr values, which appeared at all charge states. This increment was larger for the peptides that ended with lysine rather than arginine (Figure 5A). Among the mTRAQ-unique peptides, 945 had a cysteine residue in their sequence (Figure 4A). About 55% (524 peptides) of the mTRAQunique cysteine-containing peptides ended with lysine while 40% (386 peptides) ended with arginine (Figure 5B). In contrast, ICAT-unique peptides and common peptides had similar preferences for C-terminal amino acids.
Discussion
Figure 3. Quantification of plasma samples using mTRAQ, and the effect of spectral overlaps on quantification. (A) Profile of protein ratios. The ratios are log2-transformed. Bin size: 0.1. (B) The log2-transformed light to heavy peptide ratio as a function of peptide molecular weight.
number of unique peptides identified was 13 250 for mTRAQ and 5147 for the ICAT experiment. Exactly 947 peptides were common between the two data sets. The number of proteins identified was 2418 (Supplementary Table 7) and 1955 (Supplementary Table 8), respectively (Figure 4A). Altogether, 3290 proteins were identified and quantified concurrently by the two independent labeling methods, 1083 of which were identified as common and only 180 of which were identified by shared peptides. Interestingly, 173 proteins were identified by cysteinecontaining peptides in mTRAQ which were not observed in ICAT. Therefore, although 18% of ICAT peptides were common in mTRAQ, as much as 55% of ICAT proteins were quantified by mTRAQ. In mTRAQ, 519 proteins (21%) were identified with single peptide match, whereas 455 proteins (23%) were identified with single peptide match in ICAT. In addition, 59% of the 2418 proteins were identified based on 1-4 unique peptides, and 44% were identified based on 1-4 scans in mTRAQ; these values were 72% and 65% in ICAT. Therefore, the spectral counts of the mTRAQ data were distributed over wider range than those of the ICAT data for each peptide count. For the common peptides and proteins, log2-transformed mTRAQ ratios were plotted against log2-transformed ICAT ratios. The pearson correlation coefficient between peptide ratios was 0.844 and 0.707 between protein ratios (Figure 4B). The mTRAQ ratios were compared between two technical replicates (Figure 4C). About 42% of the 8058 common peptides were less than 7% different (∆log2-ratio
