Multiple Reaction Monitoring of mTRAQ-Labeled Peptides Enables Absolute Quantification of Endogenous Levels of a Potential Cancer Marker in Cancerous and Normal Endometrial Tissues Leroi V. DeSouza,† Adrian M. Taylor,†,‡ Wei Li,† Marjorie S. Minkoff,§ Alexander D. Romaschin,|,⊥ Terence J. Colgan,⊥,# and K. W. Michael Siu*,† Department of Chemistry and Centre for Research in Mass Spectrometry, York University, Toronto, Ontario, Canada M3J 1P3, Applied Biosystems/MDS SCIEX, 71 Four Valley Drive, Concord, Ontario, Canada L4K 4V8, Applied Biosystems, 500 Old Connecticut Path, Framingham, Massachusetts 01701, Division of Clinical Biochemistry, St. Michael’s Hospital, 30 Bond Street, Toronto, Ontario, Canada M5B 1W8, Department of Laboratory Medicine and Pathobiology, University of Toronto, Toronto, Ontario, Canada M5G 1L5, and Pathology and Laboratory Medicine, Mount Sinai Hospital, 600 University Avenue, Toronto, Ontario, Canada M5G 1X5 Received April 23, 2008
While iTRAQ analyses have proved invaluable for the discovery of potential cancer markers, two outstanding issues that remained were its ineffectiveness to consistently detect specific proteins of interest in a complex sample and to determine the absolute abundance of those proteins. These have been addressed by availability of the mTRAQ reagents (Applied Biosystems, Inc., Foster City, CA) a nonisobaric variant of iTRAQ. We have applied this newly emerging technique to quantify one of our potential markers for endometrial cancer, viz. pyruvate kinase M1/M2. The mTRAQ methodolgy relies on multiple reaction monitoring (MRM) to target tryptic peptides from the protein of interest, thus, ensuring maximal opportunity for detection, while the nonisobaric tags enable specific quantification of each version of the labeled peptides through unique MRM transitions conferred by the labels. Known amounts of synthetic peptides tagged with one of the two available mTRAQ labels, when used as quantification standards in a mixture with the oppositely labeled tryptically digested sample, permit determination of the absolute amounts of the corresponding protein in the sample. The ability to label the sample and reference peptides with either one of the two possible combinations is an inherent advantage of this method, as it provides a means for verification of the reported ratios. In this study, we determined that the amount of pyruvate kinase present in the homogenate from a biopsied EmCa tissue sample was 85 nmol/g of total proteins, while the equivalent concentration in the nonmalignant controls was 21-26 nmol/g of total proteins. This approximately 4-fold higher amount of pyruvate kinase in the cancer sample was further confirmed not only by a direct comparison between the cancer sample and one of the nonmalignant controls, but also independently by an enzyme-linked immunosorbant assay (ELISA). Additionally, the 4-fold higher level of pyruvate kinase amount in the cancer homogenate reported in this study is considerably higher than the 2-fold higher ratio reported across 20 cancer samples in the discovery phase with the iTRAQ technique, suggesting that there exists a possibility that the dynamic range of ratios determined by the iTRAQ technique may have been compressed. Keywords: Multiple Reaction Monitoring • mTRAQ • Absolute quantification • Endometrial Cancer • Biomarkers • Isotope-dilution mass spectrometry
Introduction In recent years, biomarker discovery using isotopically labeled protein digests of homogenates of biopsied tissue, * To whom correspondence should be addressed. E-mail: kwmsiu@ yorku.ca. † York University. ‡ Applied Biosystems/MDS SCIEX. § Applied Biosystems. | St. Michael’s Hospital. ⊥ University of Toronto. # Mount Sinai Hospital. 10.1021/pr800312m CCC: $40.75
2008 American Chemical Society
followed by liquid chromatographic separation and mass spectrometric analyses, has yielded promising putative cancer markers. In particular, we have applied the iTRAQ methodology1 with considerable success in the discovery of potential endometrial cancer (EmCa) markers.2,3 Some of these markers were subsequently verified using tissue microarray analysis on an independent and larger cohort of samples.4 However, in the discovery phase, a number of potentially interesting proteins were not detected in all sample sets analyzed, due in part to sample complexity and the competitive nature in peak selection Journal of Proteome Research 2008, 7, 3525–3534 3525 Published on Web 07/17/2008
research articles 3
for MS/MS analysis. In this current study, we have begun to address those limitations through the use of multiple reaction monitoring (MRM), an approach that has been used since the 1980s for the detection and quantification of low levels of drugs and metabolites in biological samples, and which more recently has been adapted to targeting specific peptides from proteins of interest.5–8 MRM has been successfully used to detect low levels of proteins in plasma.7 The specific and targeting nature of MRM makes it a suitable candidate method for consistent detection of the proteins of interest in all the samples analyzed. A point of note in the quantification performed in the earlier discovery-phase analyses using iTRAQ-tagging is that the quantification was relative to individual or pooled control samples.2,3,9 A recently developed variation of iTRAQ, mTRAQ (Applied Biosystems), provides a means of performing absolute quantification based on the isotope-dilution mass spectrometry (IDMS) principle and is specifically designed to take advantage of the MRM mode. IDMS is a time-honored technology that has been widely applied for quantitative analysis of a large variety of analytes10–12 It was first introduced to determine the amount of calcium in blood in the early 1970s and soon became an established methodology for the quantification of a number of metabolites, including glucose and cholesterol and more recently of intact proteins in blood.13–16 Unlike the iTRAQ labels, the mTRAQ labels are designed to be nonisobaric in nature to maximize possible differences in the MRM transitions.1 The mTRAQ labels come in two chemically identical versions. The lighter version is lower in mass than the iTRAQ labels by 4 Da; specifically, no 13C or 15N was deliberately introduced. The heavy version is identical to the iTRAQ 117 label.1 Thus, where the mass of the heavy version of mTRAQ and iTRAQ labels is 145 Da, that of the light version of the mTRAQ label is only 141 Da. Consequently, the MRM transitions chosen for a peptide that is labeled with the two different versions are specific for each version by virtue of the differences in precursor and product ion masses (in instances where the product ion retains the mTRAQ tag). For absolute quantification, application of the mTRAQ technology involves the use of one of the two versions to label a known quantity of a synthetic peptide whose sequence is identical to that of a selected tryptic peptide from the protein of interest, while the other version is used to tag the tryptically digested sample homogenate. The labeled synthetic peptide is then mixed with the oppositely labeled sample at a known amount, and the combined sample is then analyzed by means of one-dimensional or two-dimensional liquid chromatography (LC)-MRM analysis, the selection of the LC mode being dependent on the complexity of the samples. As mentioned above, each of the labeled versions of the peptides is monitored by a unique MRM transition ensuing from the different masses of the tags. Also, as these versions of the peptides differ only in the isotopic content of their labels, they coelute during the course of the LC separation. The respective areas under the resulting MRM traces, which are superimposing, are then calculated, thereby permitting quantification of the peptide in the digested sample, and in turn, that of the protein in the original sample (Figure 1). The protein of interest in this proof-of-principle study is one of the potential biomarkers that we have previously reported for endometrial cancer: pyruvate kinase (PK)-M1/M2 (Accession no. P14618).2,3 Pyruvate kinase is an enzyme best known for its function in the glycolytic pathway, where it catalyzes the conversion of phosphoenol pyruvate (PEP) to pyruvate plus 3526
Journal of Proteome Research • Vol. 7, No. 8, 2008
DeSouza et al. the generation of ATP from ADP. Unlike energy production by mitochondrial respiration, the ATP production by pyruvate kinase is independent of the oxygen level. It was, therefore, initially suggested that PK plays a role in the survival of cells under the hypoxic conditions typically prevalent in tumors (reviewed in ref 17). Different PK isoforms have been found to predominate in different tissues and organs: isoform L is found in tissues with gluconeogenesis, including the liver and kidney; erythrocytes express isoform R. The L and R isoforms are encoded by the same gene, but are controlled by different promotors. Isoforms M1 and M2 are products of a second gene; they are splice variants of exons 9 and 10, which encode a stretch of 56 residues that have differences in 22 positions.17,18 The M1 isoform is expressed in skeletal muscle and brain tissue, while the M2 isoform in fetal tissue and proliferating cells, including adult stem cells, early embryonic cells, and cancer cells.17–19 A relatively recent, and perhaps paradoxical, finding is that the M2 isoform (PK-M2) in tumors negatively regulates the production of ATP in order to provide glycolytic intermediates that are vital precursors for the syntheses of cell components.17 A noted difference between the PK-M2 present in normal tissue and that present in tumors is that in the lung tissue as well as normal proliferating cells PK-M2 is in a tetrameric form, while in tumors, PK-M2 is dimeric.17 In terms of activity, the tetrameric form has a high affinity for PEP and is, therefore, highly active when this substrate is present at physiological levels. Conversely, the dimeric form with low affinity for PEP is nearly inactive, thereby permitting accumulation of the various intermediates upstream of PEP and providing the necessary precursors for cell-component syntheses.17 The tetramer-to-dimer ratio is not stationary, but rather oscillatory, and is regulated allosterically by intracellular fructose1,6-biphosphate (FBP): a high FBP concentration induces association of dimeric PK-M2 into the active tetrameric form, which enables the conversion of glucose into lactate.17 As the dimeric form of PK-M2 is only prevalent in tumors, it is often referred to as tumor PK-M2. Here, we report the absolute quantification of pyruvate kinase-M1/M2, using mTRAQ-labeling with two-dimensional LC-MRM analysis, in a few representative tissue homogenates examined previously in the discovery phase with iTRAQlabeling.3 The MRM analysis was performed on three samples: two nonmalignant normal proliferative endometrial samples (EmNo1 and EmNo2) and one endometrial cancer (Type I) sample (EmCa1). In each of these cases, the concentration of PK was determined in an aliquot of the same preparation of tissue homogenate used in the previous study that had been stored at -80 °C.3 For enhanced analytical confidence in the quantification of PK, two of the tryptic peptides that were most frequently detected in the discovery-phase study were chosen as targets for MRM analysis.3 These peptides were custom synthesized by a commercial source (Sigma Genosys, TX) for use as the standards for quantification. Last, as a means of independent verification, we also performed an enzyme-linked immunosorbant assay (ELISA) for tumor PK-M2 using the homogenate from EmCa1 and two other independent endometrial cancer samples (EmCa2 and EmCa3) in comparison with three independent normal proliferative samples (EmNo3, EmNo4 and EmNo5).
Methods Sample Preparation. As described above, the homogenates used for this study were aliquots of those used in the discovery
Multiple Reaction Monitoring of mTRAQ-Labeled Peptides
research articles
Figure 1. Representative MRM traces for one set of transitions for each peptide. PK1, 639.7/524.5 and 642.4/528.5; PK2, 669.4/597.4 and 671.4/601.4. PK1 transitions displayed are for a triply charged doubly labeled precursor peptide ion with an mTRAQ label at both the N-terminus as well as the C-terminal lysine side-chain. In this case, Q3 was set to select the singly charged singly labeled b4 ion. PK2 transitions are for a doubly charged singly labeled precursor peptide with the label at the N-terminus. Here too, Q3 was set to transmit the singly charged and labeled b4 ion. The traces are from Mixture 2 and show more intense peaks for the reference peptides labeled with the light version (shown in red and blue) than the endogenous peptides labeled with the heavy version (gray and green). Peak pairs coelute while the two peptides are temporally separated.
phase that had been archived by storing at -80 °C. Briefly, resected tissues that had been flash-frozen within 20 min of devitalization were sectioned and classified by a pathologist (T.J.C.). A piece of the mirror face of the section used for classification, 0.5-1.0 cm3, was washed three times in 1.0 mL of cold phosphate-buffered saline (PBS), before homogenizing in 0.5 mL of PBS with a cocktail of protease inhibitors (1 mM AEBSF (4-(2-aminoethyl) benzenesulfonyl fluoride), 10 µM leupeptin, 1 µg/mL aprotinin, and 1 µM pepstatin) using a hand-held homogenizer. The homogenate was then centrifuged for 30 min in a microfuge at 4 °C at 21 000g and the supernatant transferred to a fresh tube and stored at -80 °C until required. Synthetic Peptides. The two peptides chosen for this study were PK1 (GVNLPGAAVDLPAVSEK) and PK2 (LDIDSPPITAR) for pyruvate kinase M1/M2 (see Figure S1 in the Supporting Information for the locations of the peptides in PK-M2). The reported amino acid content as determined by amino acid analysis and the proportion of the desired peptide within that amino acid content in the respective stock solutions after
resolubilizing the supplied peptide pellet were determined (Sigma Genosys, TX) to be 91.2% and 91% (PK1); 86.8% and 86% (PK2). The proportion of the peptide was determined by direct infusion nanospray mass spectrometric analysis followed by a calculation of the peptide peak intensity relative to the total areas of all peaks detected. The working concentrations of the resuspended peptides prepared nominally as 10 nM in each case were therefore adjusted for purity of the peptides and resulted in a calculated concentration of 8.30 nM of PK1 and 7.46 nM of PK2. These were labeled individually with the light and heavy mTRAQ labels as were 75 µg aliquots of trypsin digests of the three tissue homogenates (EmCa1, EmNo1, and EmNo2). Trypsin digestion and labeling with the mTRAQ reagents were performed as described previously for the iTRAQ reagents.1,2 After labeling, the two heavy- and light-labeled synthetic peptides were mixed in a 1:1 ratio, to form one heavy (H)- and one light (L)-labeled “synthetic peptide pool” (SPP). Aliquots of these pools were combined with their corresponding oppositely labeled samples as shown in Table 6. Journal of Proteome Research • Vol. 7, No. 8, 2008 3527
research articles
DeSouza et al.
Table 1. Mixtures for mTRAQ-Labeling and Analysis
designation
sample (label)
Mixture 1 Mixture 2 Mixture 3 Mixture 4 Mixture 5 Mixture 6 Mixture 7 Mixture 8 Mixture 9
EmCa1 (L) EmCa1 (H) EmNo1 (L) EmNo1 (H) EmCa1 (H) EmCa1 (L) EmNo2 (L) EmCa1 (H) SPP (L)
sample (label)
amount of each peptide (based on the initially calculated nominal amounts)
SPP (H) SPP (L) SPP (H) SPP (L) EmNo1 (L) EmNo1 (H) SPP (H) SPP (L) SPP (H)
20 pmol 20 pmol 20 pmol 20 pmol 75 µg 75 µg 20 pmol 10 pmol 10 pmol
amount
75 75 75 75 75 75 75 75 10
µg µg µg µg µg µg µg µg pmol
Table 2. Reverse-Phase NanoLC Gradient time (min)
0
10
20
50
58
70
72
89
%B
5
5
30
65
80
80
5
5
Strong Cation Exchange Separation. After mixing the labeled samples as specified above, the resulting samples were vacuum-centrifuged to dryness and resuspended in Eluent A (1.0 mL of 10 mM KH2PO4 in 25% acetonitrile and adjusted with phosphoric acid to pH 3.0). Each of these resuspended mixtures was then fractionated by manual injection onto a 0.2 mL capacity strong cation exchange (SCX) cartridge supplied as part of an ICAT kit (Applied Biosystems, Foster City, CA) followed by a wash with 1.0 mL of Eluent A and step-elutions using 0.5 mL each of Eluent A with increasing concentrations of KCl. In preliminary runs, the three peptides of interest were established to elute in the 100 mM KCl fraction. The salt concentrations used, therefore, were 50 mM, 100 mM, 150 mM, and 1 M KCl. The 100 mM salt fraction was the only one examined for the remainder of the study. This fraction from each of the mixtures was vacuum-centrifuged to dryness and resuspended in 30 µL of 1.0% formic acid. Reverse-Phase NanoLC-MS/MS Analysis. A Tempo nano MDLC system with autosampler (Applied Biosystems/MDS SCIEX) was used to perform a reverse-phase (RP) nanoLC separation online with a QTRAP 4000 linear ion trap (Applied Biosystems/MDS SCIEX) tandem mass spectrometer. Except where specified, 1/10 of the resuspended 100 mM eluate fraction was loaded onto a 5-µm particle, 5-mm length × 300-µm i.d., 100-Å pore size, RP-C18 trapping cartridge (LC Packings, Amsterdam, The Netherlands) fitted onto a switching valve (VICI, Houston, TX). The sample was loaded and desalted using an aqueous solution of 5% acetonitrile in 0.1% formic acid (Solvent A). Desalting was performed at a flow rate of 20 µL/ min for 5 min before the precolumn was switched inline with the analytical column (150-mm length × 75-µm i.d., packed in-house with 3-µm, 100-Å pore, Kromasil beads). Peptides were eluted at a flow rate of 200 nL/min using a nonlinear binary gradient of Solvent A and Solvent B (95% acetonitrile in 0.1% formic acid) described in Table 7. MRM Transitions. Preliminary tests using the labeled and unlabeled synthetic peptides were run with an acquisition method that included an enhanced MS scan followed by enhanced product ion (EPI) scans with a rolling collision energy (see Figure S2 in Supporting Information for MS/MS spectra of the peptides). These scans provided information on the charge states of the peptide ions, the most-prominent fragment ions detected under the run conditions, as well as the elution times of the peptides. This information was then used to build 3528
Journal of Proteome Research • Vol. 7, No. 8, 2008
a second acquisition method which incorporated an MRM scan followed by two EPI scans triggered by a positive detection of the specific MRM transition event. Dwell times for all transitions were set at 29 ms. As the EPI scans in this method were used merely for purposes of verifying the peptide identity, the dynamic exclusion in the method was set to exclude the precursor ion for a period of 60 s after the initial scan. The version of the Analyst software (version 1.4.1) used to control data acquisition permitted the monitoring of a maximum of 100 MRM transitions over the course of the run. With the dwell times for each transition set for 29 ms, the time for the total cycle including the two EPI scans was 6.79 s. In turn, as the peptide peaks typically eluted over approximately 120 s, and in the unlikely event that all the transitions being monitored eluted simultaneously and, therefore, triggered EPI scans in every cycle, this represented a minimum of 17 data points across the entire chromatographic peak. In reality, as the EPI scans constituted more than half of the total scan cycle time, and as they were only triggered occasionally, the number of data points recorded across the chromatographic peak were far in excess of 17 and resulted in very well-defined peaks. MRM transition data were acquired using unit-resolution settings for both Q1 and Q3 to ensure maximum contribution from the monoisotopic peak, while minimizing any possibility of contribution from the oppositely labeled peptide. In addition, product ions monitored by Q3 were selected from both the yand b-ion series in the case of the labeled peptides. This was done as a precaution against instances where the m/z values of the multiply charged heavy- and light-labeled peptides were too close. In such a situation, if the product ion mass-selected by Q3 was an arginine-terminating y-ion without the label, the detection of a transition intended to monitor a heavy-labeled peptide could potentially have some contribution from an isotopic peak from the light-labeled version. The transitions used to detect the two peptides of interest are listed in Table 1. All three versions (heavy-labeled, light-labeled, and unlabeled) of the two peptides were monitored in all runs. In addition, for the last runs of Mixtures 5 and 6, transitions for the heavy- and light-labeled versions of the M2-specific isoform of pyruvate kinase LAPITSDPTEATAVGAVEASFK (designated PK3 in Table 1) were also monitored. This peptide had not been selected for synthesis as a reference peptide, as it had not been consistently detected in the discovery phase, suggesting in turn that its ionization efficiency was not as good as the other two peptides. Abundances were calculated on the basis of the peak area after integration, using the IntelliQuan algorithm provided in the Analyst 1.4.1 software. Only peaks with areas greater than 10 000 counts were considered in the calculation. ELISA Test. An ELISA test using a tumor pyruvate kinase M2specific kit (ScheBo Biotech AG Giessen, Germany) was also performed on three endometrial cancer tissue homogenates, including the one used for the MRM analysis (EmCa1) as well as three normal proliferative sample homogenates (EmNo3, EmNo4 and EmNo5). The protocol used was as recommended by the manufacturer, except for the addition of 4 mM EDTA to the sample prior to dilution with the supplied wash and binding buffer. The addition of EDTA was to ensure that the sample conditions were as close as possible to that of EDTA-plasma, for which this kit was designed. The protocol involved assaying 50 µL aliquots of the samples and the supplied standards, diluted 100-fold in the sample washing buffer provided. The standard supplied was aliquots of human serum containing differing amounts of tumor M2-PK. Each sample and standard
Multiple Reaction Monitoring of mTRAQ-Labeled Peptides a
Table 3. Transitions Used for MRM Analysis peptide
label type
charge state
Q1 m/z
Q3 m/z
CE used (V)
PK1
Unlabeled
2+
819.0
61.50
3+
546.3
2+
959.3
3+
639.7
2+
963.3
3+
642.2
Unlabeled
3+
599.4
Light
2+
669.4
3+
446.6
2+
671.4
3+
447.9
Light
3+
819.1
Heavy
3+
821.8
630.3 (y6) 363.2 (y3) 743.4 (y7) 630.3 (y6) 858.4 (y8) 680.3 (b8) 524.2 (b4) 883.3 (y7) 770.6 (y6) 820.5 (b8) 524.5 (b4) 528.10 (b4) 887.10 (y7) 774.4 (y6) 824.4 (b8) 528.5 (b4) 654.5 (y6) 856.6 (y8) 741.5 (y7) 741.4 (y7) 597.4 (b4) 684.5 (b5) 557.4 (y5) 597.4 (b4) 741.4 (y7) 601.4 (b4) 688.5 (b5) 557.4 (y5) 601.4 (b4) 325.2 (b2) 434.2 (y2) 838.5 (b7) 329.2 (b2) 438.3 (y2) 842.5 (b7)
Light
Heavy
PK2
Heavy
PK3
23.57
51.70 37.60
51.70 37.60
28.00
38.96 29.18
38.96 29.18
40.0
40.0
a All Q1 m/z values of the three peptides that lay within the 400-1000 Th range were chosen, while Q3 m/z values were chosen based on their abundances in an EPI scan performed in preliminary runs. CE was chosen based on the rolling collision energy used in the preliminary runs in each instance.
was analyzed in duplicate. After binding, washing, and probing with the monoclonal biotin-labeled secondary antibody, the wells were incubated with the peroxidase-streptavidin complex. Last, the wells were washed, and the substrate solution was added and incubated for 15 min, after which the reaction was stopped and the optical density (at 450 nm) was read using a plate reader (Victor3 1420 Multilabel counter, Perkin-Elmer Inc.).
Results A few pilot studies were first performed in order to test the MRM approach. These included nanoLC-MS MRM runs on aliquots of sample fractions that had been archived after the discovery phase. We initially performed a first-pass analysis on samples in which the potential markers of interest had been detected in the discovery phase in order to establish run and parameter conditions. This was followed by a second-pass experiment performed on samples where some of those proteins had previously not been detected, and in a number of instances, this targeted approach succeeded in detecting those proteins (data not shown). Another set of test runs was performed to establish detection limits using this approach. In these instances, we loaded, ran, and successfully detected a
research articles tryptic digest of 30 amol of β-casein in a 1000-fold excess of bovine serum albumin (see Figure S3 in Supporting Information). Additionally, we also spiked in 2 ng of prostate-specifc antigen (PSA) into a 100-µg aliquot of a homogenate of one of the endometrial cancer control samples, digested the pooled sample, and analyzed by means of nanoLC-MS MRM. The actual amount of sample loaded onto the precolumn was about 2 µg of total protein digest, in which we detected two of the three PSA peptides targeted (see Figure S4 in Supporting Information). The 2-ng amount had been chosen as a realistic approximation of a potential marker in blood based on the established normal amount of PSA, which is reported to be in the 0.04-6.8 ng/mL range in blood.20 Having ensured that the MRM approach was sufficiently sensitive for attempting the absolute quantification measurements, we proceeded with an experiment on the pure mTRAQ-labeled synthetic peptides. The first MRM analysis showed that, in both the heavy- and lightlabeled pools, the unlabeled peptide transitions detected were minimal (Figure 2), thereby demonstrating the near quantitative efficiency of the labeling. The actual proportion of unlabeled peptides in each of the instances was almost always below the threshold for quantification set at 10 000 counts. An exception to this was the unlabeled PK2 peptide which was detected as a doubly charged ion at 599.4 Th in the light-labeled reference pool (Figure 2K). A calculation of the relative area of the most intense of the three MRM transitions relative to that of the most-intense transition for the light-labeled version of the peptide (Figure 2J; 447.9/557.4 transition) showed that the unlabeled peptide was present at 1.06% of the light-labeled version, assuming that the ionization efficiencies of the unlabeled and labeled peptides were comparable. This assumption was verified by performing a separate MRM analysis on an equimolar mixture of heavy, light and unlabeled versions of both peptides (see Figure S5 in Supporting Information). The two pools of labeled peptides were also mixed in a 1:1 ratio (Mixture 9) and analyzed in order to assess the relative efficiency of labeling and ionization in each case, as well as to provide a means of normalizing the data when calculating absolute quantities. To ensure consistency with subsequent analyses, the 1:1 mixture was also SCX-fractionated and the 100 mM KCl fraction was injected and analyzed in triplicate by RP-nanoLC-MS MRM as described earlier. The Light/Heavy ratios in each case were found to be as follows: PK1, 1.10 ( 0.05 (mean ( standard deviation (SD)): PK2, 0.98 ( 0.01 (see Supporting Information). Although MRM transitions for both the doubly and triply charged states of the precursor peptides were monitored, the ratio for PK1 was calculated solely on the basis of the triply charged precursor transitions, as the doubly charged precursor transitions were below the threshold for reliable quantification as mentioned above (this threshold was empirically determined). The ratio for PK2, however, was calculated on all the transitions that were monitored for both the doubly and triply charged precursor ions, as these were all above the threshold for reliable quantification. Labeling efficiency of the endogenous peptides in the sample homogenates was also evaluated by calculating the ratio of the peak intensity of the most intense transitions for each of the unlabeled peptides relative to the combined intensities of the most intense transition of each of the labeled versions of the respective peptides in Mixtures 5 and 6. The rationale for this was based on the fact that the unlabeled peptides in these mixtures would be from both samples in the mixture. Hence, the unlabeled proportion was best evaluated against the total of the respective heavy- and Journal of Proteome Research • Vol. 7, No. 8, 2008 3529
research articles
DeSouza et al.
Figure 2. Transitions of the labeled and unlabeled version of PK1 and PK2. Transitions in each panel are in the color of the corresponding MRM trace. Panels A-F are transitions for PK1, while panels G-L are for PK2. (A) Triply charged unlabeled PK1 in light-labeled SPP; (B) triply charged light-labeled precursor of PK1 in SPP (L); (C) doubly charged unlabeled precursor of PK1 in SPP (L); (D) triply charged unlabeled PK1 in SPP (H); (E) triply charged heavy-labeled PK1 in SPP (H); (F) doubly charged unlabeled PK1 in SPP (H); (G) triply charged light-labeled PK2 in SPP (L); (H) doubly charged unlabeled PK2 in SPP (L); (I) doubly charged light-labeled PK2 in SPP (L); (J) triply charged heavy-labeled PK2 in SPP (H); (K) doubly charged unlabeled PK2 in SPP (H); (L) doubly charged heavy-labeled PK2 in SPP (H). Of all the negative control panels, the only ones with significant peaks are seen in panels H and K where the peak intensity is 100-fold lower than that in the panels displaying the labeled versions.
light-labeled peptides. The calculation was applied to all three replicate injections of the two mixtures and the proportion of unlabeled peptides was determined to be 7.4% ( 3.1 (proportion of unlabeled peptides ( SD) for PK1 and less than 1% for each of the six analyses for PK2. As alluded to above, the 100 mM KCl fractions from all nine sample mixtures listed earlier were injected in triplicate and analyzed by nanoLC-MS MRM. The ratios reported in each case were calculated using the same threshold specified above and showed good precision with low standard deviation in each case (Table 2). As the elution fraction used in these analyses contained a complex pool of peptides from the tissue homogenates being tested, which could amount to a total peptide content that exceeded the capacity of the precolumn, it was deemed necessary to verify that there be no significant variations in reported ratios resulting from possible saturation effects in either of the columns. We, therefore, also performed an analysis on a series of dilutions of the 100-mM KCl fraction from Mixture 1. Dilutions used for this analysis were 1:10 (run 3530
Journal of Proteome Research • Vol. 7, No. 8, 2008
in triplicate), 1:5, 1:2, and a 2-fold larger amount used for the standard quantification runs described earlier (Table 3). Likewise, Mixture 8 was included to ensure that the ratios did not vary as a result of different amounts of reference peptides used. On the basis of the ratios reported in Table 2 and the amounts of the reference peptides used, the absolute amounts of PK were determined and shown in Table 4: 85 nmol/g in EmCa1, the cancer tissue homogenate, and 22-26 nmol/g in EmNo1 and EmNo2, the normal tissue homogenates, representing an overexpression of PK in 3-4 folds. The results of the ELISA test are shown in Table 5. The abundance of PK-M2 in these samples is reported in units/mL as designated by the ELISA kit. The ratios reported are the average of the readings performed in duplicate in each case, and normalized by the total protein concentration in the sample. Perhaps significantly, while the individual relative abundances among the tumor and normal homogenates varied considerably, the ratio of PK-M2 in the EmCa1 sample relative to the median EmNo sample abundance (EmNo3) was 3.7,
research articles
Multiple Reaction Monitoring of mTRAQ-Labeled Peptides a
Table 4. Ratios of Endogenous Peptides to Reference Peptides sample
peptide
R1
R2
R3
average
norm. avg.c
Mixture 1
PK1 PK2 PK1 PK2 PK1 PK2 PK1 PK2 PK1 PK2 PK3 PK1 PK2 PK3 PK1 PK2 PK1 PK2
0.29 ( 0.01b 0.50 ( 0.04 0.28 ( 0.02 0.52 ( 0.02 0.071 ( 0.011 0.115 ( 0.006 0.092 ( 0.003 0.127 ( 0.006 3.78 ( 0.71 4.28 ( 0.23
0.25 ( 0.01 0.49 ( 0.03 0.30 ( 0.02 0.52 ( 0.05 0.090 ( 0.002 0.134 ( 0.004 0.090 ( 0.004 0.132 ( 0.016 3.91 ( 0.36 3.83 ( 0.81
0.27 0.50 0.30 0.53 0.083 0.124 0.090 0.134 3.89 4.09
0.27 0.50 0.27 0.54 0.083 0.124 0.082 0.136 3.89 4.09
4.12 ( 0.23 4.11 ( 0.54
4.68 ( 1.18 3.81 ( 0.83
4.32 4.12
3.93 4.21
0.094 ( 0.011 0.162 ( 0.012 0.68 ( 0.06 1.04 ( 0.14
0.106 ( 0.007 0.148 ( 0.008 0.72 ( 0.06 1.15 ( 0.15
0.27 ( 0.02 0.51 ( 0.03 0.32 ( 0.01 0.55 ( 0.02 0.087 ( 0.006 0.122 ( 0.016 0.089 ( 0.003 0.142 ( 0.015 3.99 ( 0.13 4.15 ( 0.34 4.12 ( 0.29 4.17 ( 0.44 4.45 ( 0.33 3.35 ( 0.90 0.098 ( 0.005 0.148 ( 0.008 0.74 ( 0.14 1.08 ( 0.11
0.099 0.153 0.71 1.09
0.099 0.153 0.65 1.11
Mixture 2 Mixture 3 Mixture 4 Mixture 5
Mixture 6
Mixture 7 Mixture 8
a Ratios reported in Mixtures 5 and 6 correspond to the abundance of those peptides in the cancer sample relative to the nonmalignant sample. Standard deviation (SD) was calculated across all transitions used to calculate the ratio reported; for PK1, this was across 3 transitions; for PK2, this was across 5 transitions. SD for PK3 was across ratios reported for 3 transitions. c Normalized average: ratios normalized using the average L/H ratio for each peptide as determined in the 1:1 SPP comparison. For PK1, the L/H bias was 1.1; for PK2, the L/H bias was 0.98. Only ratios relative to heavy-labeled peptides were adjusted using these adjustment factors. This adjustment was not applicable in Mixtures 5 and 6. b
Table 5. Ratios of Synthetic Peptide:Endogenous Peptide When Loaded at Different Dilutions
PK1 PK2
1:10a
1:5a
1:2a
1×b
2×b
avg.
SD
CVc
3.50 2.08
3.39 2.02
3.48 2.05
3.39 1.98
3.81 2.02
3.51 2.03
0.17 0.04
4.95 1.97
a Dilution. variation (%).
b
Fold of increase in concentration.
c
Coefficient of
approximating the ratio obtained with the mTRAQ analysis for the same sample relative to EmNo1 and EmNo2.
Discussion As the electrospray ionization efficiencies of the unlabeled and labeled versions of the peptides were not necessarily comparable, we tested the signal intensities of all three versions of both peptides using an equimolar mixture and found them to be comparable (Supporting Information Figure S5). As a precaution, we also chose to monitor the transitions for the unlabeled peptides in all subsequent analytical runs as a matter of quality assurance: a similar proportion of unlabeled peptides was noted in the sample mixtures containing tissue homogenates, suggesting that the labeling efficiency with respect to these peptides from the endogenous protein in the homogenates was similar to that of the peptides in the reference pool, and that the subsequent quantification was, therefore, reliable. In a further attempt to increase the confidence in quantification, in addition to the use of unit resolution for both Q1 and Q3 during the MRM scan, we monitored two or more transitions for each peptide and more than one charge state, whenever possible. It is evident from the results of this study that MRM with mTRAQ-labeling gives very consistent measurements not only across the replicate injections, but also with respect to both combinations of labeling schemes. The only obvious discrepancy lies in the ratios obtained for the individual peptides relative to the SPP standards and, thereby, the final concentrations of the two individual PK peptides calculated in all the mixtures tested. A closer inspection of the context of the two
peptides used suggests that, other than the possibility that the initial concentration of one of the two peptides was skewed, a possible explanation for this discrepancy might lie with the calculation of the concentration of PK1 in particular. An inaccurate amount of a standard peptide being the source of the difference is less likely, as the purity determination was performed on an aliquot of the same resuspended sample used in these labeling reactions, and also because the two labeling reactions performed independently of each other show similar relative levels. The second possibility is more probable as, in addition to the 7.4% of unlabeled PK1 peptide (as determined in the samples in Mixtures 5 and 6) that could also be a contributing factor, PK1 is immediately preceded by two consecutive lysine residues in intact pyruvate kinase. Missed cleavages are not unusual in peptides that contain consecutive lysine or arginine residues during trypsin digestion. If this has occurred in trypsinization of pyruvate kinase, a proportion of the endogenous PK1 peptides would contain an N-terminal lysine residue and would not have been detected by the MRM scan. Consequently, the ratio and pyruvate kinase amount reported based on PK1 would be lower. This discrepancy between the relative abundances of the two peptides when assessed in the two direct comparisons (Mixtures 5 and 6) between the two homogenates would not have been apparent as both samples might have contained a similar proportion of missed cleavages. Thus, while this hypothesis would suggest that the determination with PK2 would be more accurate, we are opting to report herein the average pyruvate kinase concentrations as determined with both peptides. A retrospective review of the archived data from our discovery phase3 shows that this missed cleavage at the N-terminus of PK1 peptide is, in fact, a common occurrence with many such instances being detected. This lends credence to our hypothesis and, in turn, suggests that peptides beginning or terminating in consecutive cleavage site residues are nonideal as standards for absolute quantification. The major advantage of isotope dilution is that the chemistries of the samples are identical after labeling and mixing, which means that these processes should Journal of Proteome Research • Vol. 7, No. 8, 2008 3531
research articles
DeSouza et al. a
Table 6. Quantification of Pyruvate Kinase
sample
average ( SD
normalized averageb
Mixture 1
0.27 ( 0.02 0.5 ( 0.01 0.3 ( 0.02 0.53 ( 0.02 0.71 ( 0.03 1.09 ( 0.06 0.08 ( 0.01 0.12 ( 0.01 0.090 ( 0.002 0.134 ( 0.008 0.10 ( 0.01 0.15 ( 0.01
0.27 0.50 0.27 0.54 0.65 1.11 0.08 0.12 0.08 0.14 0.10 0.15
Mixture 2 Mixture 8 Mixture 3 Mixture 4 Mixture 7
calc. conc.c (pmol/75 µg)
conc. corrected for peptide purity (pmol/75 µg)
avg. of both peptides (pmol/75 µg)
final conc. (nmol/g)
avg. conc. (nmol/g)
5.40 10.00 5.45 10.82 6.48 11.12 1.65 2.47 1.64 2.73 1.99 3.05
4.48 7.46 4.53 8.07 5.38 8.30 1.37 1.85 1.36 2.04 1.65 2.28
5.97
79.6
84.9
6.30
84.0
6.84
91.2
1.61
21.4
1.70
22.7
1.96
26.2
22.1
26.2
a Average ratios in the first column were the same as those reported in Table 2. SDs were calculated across the three runs reported in Table 2. Reference peptide pool used for Mixture 8 was half that used in all other mixtures. b Normalized according to the L/H ratio obtained for the SPP mixture. c Calculated concentration based on original SPP concentrations.
Table 7. Ratios of PK-M2 by ELISAa sample
concentration (U/mL)
CV
EmNo3 EmNo4 EmNo5 EmCa1 EmCa2 EmCa3
13.1 29.8 11.4 48.5 99.0 95.5
0.86 0.80 5.08 0.62 0.45 0.46
avg.
EmCa:EmNo3
EmCa (avg): EmNo (avg)
18.1
81.0
3.71 7.56 7.29
4.48
a Abundance is calculated in units (U)/mL based on optical density readings calibrated against standard samples supplied by the kit manufacturer. EmNo3 was the median value against which the three cancer samples were compared individually.
be carried out as early as possible in the work flow. The implicit assumption in using tryptic peptides as quantification standards for proteins is that the tryptic digestion efficiency for these proteins is 100%. One other factor that would merit consideration for explaining the discrepancy between the concentrations determined for the two peptides would be the possibility of the presence of a post-translational modification on the endogenous peptides. This possibility while conceivable, particularly in light of the presence of a serine residue in both peptides, is less likely to be the cause of the discrepancy, as it is the serine residue in PK2 that is known to be a phosphorylation site; there is, however, no evidence of phosphorylation at the serine contained in PK1, which we believe to be the source of the discrepancy.21–23 The ability to perform labeling in both possible combinations for absolute quantification is an advantage of the mTRAQ technology over the AQUA and SISCAPA approaches,24–26 as it provides an additional means for verification of the ratios, while eliminating any potential for systematic bias as a result of differing labeling or detection efficiencies. The ratios reported for Mixtures 5 and 6 (Table 2) are a good example of this verification, as they both show that the ratios of the two peptides from PK monitored were approximately 4-fold higher in the EmCa1 sample. Further, the ratios of the absolute amounts of PK in the EmCa1 to the EmNo1 samples based on the results obtained with Mixtures 1-4 (Table 4) are also consistent with the ratios based on the direct comparison between the two samples. The coefficients of variance (relative SDs) for both peptides were found to be lower than 5% in the experiments with different dilutions of the KCl fraction from 3532
Journal of Proteome Research • Vol. 7, No. 8, 2008
Mixture 1 (Table 3), demonstrating that there were no saturation effects in quantification within the range of concentrations involved in this study. Most importantly, although the areas of the MRM traces for the M2-specific peptide were below the threshold for quantification, the ratios obtained were in a range similar to those reported for the other two more-abundant peptides (Table 2), thereby strongly suggesting that the 4-fold larger amount reported for PK in the EmCa1 sample is a result of specific differential expression of the M2 isoform. The PKM2 detected in the normal, proliferative endometrium could be the tetrameric form that is prevalent in proliferating cells,17–19 which we expect would be true of a significant proportion of the cells in the proliferative endometrium. Finally, the amounts of pyruvate kinase detected in both the normal samples appear to be similar and are 4-fold lower than that in EmCa1, in validation of the results obtained in the iTRAQ analyses performed previously. Notably, the ratios reported in our earlier discovery-phase study were generally lower than those reported here for the comparisons between the EmCa samples and the EmNo controls. Should the 4-fold higher results for the EmCa samples reported in this study be duplicated and validated in a larger cohort of samples, it would strongly suggest that the dynamic range of iTRAQ-labeling is compressed. The ELISA results with antibodies specific to the M2 isoform further corroborates the mTRAQ result and supports the suggestion that the pyruvate kinase isoform present in the endometrium is PK-M2. In addition, it also appears to further strengthen the suggestion that the iTRAQ dynamic range is compressed. As mentioned earlier, while the expression of PK-M2 has long been known to be elevated in various tumors,17,18,27 its mechanism of action was ascribed to a role in ATP synthesis under anaerobic conditions prevalent in tumors.28 However, recently PK-M2 has been demonstrated to be the switch responsible for a shift in metabolism to aerobic glycolysis, a process termed the Warburg effect, which is an irreversible step that leads to tumorigenesis.19,29 There is compelling evidence that it has a role in the generation of metabolic intermediates essential for rapidly proliferating cells17,19 The explicit mechanism by which PK-M2 promotes tumorigenesis, however, has not yet been elucidated. In what could be a further complication in attempts to gain a better understanding into its mechanism of action, it has also been proposed that, in addition to a role in glycolysis, PK-M2 could also play a second, independent role via phos-
Multiple Reaction Monitoring of mTRAQ-Labeled Peptides 17,30
photyrosine signaling. Thus, while many studies including our own have suggested PK-M2 plays a significant role in tumor growth, the exact mechanism awaits elucidation in further investigations. The ability to target and quantify PK-M2 at low levels in both micro environments, such as cancer nodes, and macro environments, such as blood; as well as in response to various stimuli will be invaluable in such investigations. This study, therefore, represents a vital advancement in available technology that can be brought to bear in attempts to address this question. In conclusion, using the new mTRAQ technique, we have successfully quantified the amount of pyruvate kinase in both nonmalignant endometrial tissue in the proliferative phase, as well as an endometrial tumor sample. The performance of mTRAQ for quantification was validated by means of ELISA. Future studies will focus on detecting and quantifying the amount of PK-M2 in plasma. Pyruvate kinase has been detected in blood31,32 and found to be released extracellularly by endometrial cancer cells.33 To this end, we expect that, even if the levels of PK-M2 are too low for direct quantification in plasma, an approach similar to that in SISCAPA26 involving the use of antibody capture would enable the application of this technology to the quantification of proteins in bodily fluids.
Acknowledgment. This work was supported by Canadian Cancer Society Research Grant No. 016172 of the National Cancer Institute of Canada. Collaboration and support by Applied Biosystems/MDS SCIEX is gratefully acknowledged. Supporting Information Available: Pyruvate kinase peptides and their MS/MS spectra; MRM transitions for β-caesin and PSA; and data for light/heavy label ratio determinations. These materials are available free of charge via the Internet at http://pubs.acs.org. References (1) Ross, P. L.; Huang, Y. N.; Marchese, J. N.; Williamson, B.; Parker, K.; Hattan, S.; Khainovski, N.; Pillai, S.; Dey, S.; Daniels, S.; Purkayastha, S.; Juhasz, P.; Martin, S.; Bartlet-Jones, M.; He, F.; Jacobson, A.; Pappin, D. J. Multiplexed protein quantitation in Saccharomyces cerevisiae using amine-reactive isobaric tagging reagents. Mol. Cell. Proteomics 2004, 3, 1154–1169. (2) DeSouza, L.; Diehl, G.; Rodrigues, M. J.; Guo, J.; Romaschin, A. D.; Colgan, T. J.; Siu, K. W. Search for cancer markers from endometrial tissues using differentially labeled tags iTRAQ and cICAT with multidimensional liquid chromatography and tandem mass spectrometry. J. Proteome Res. 2005, 4, 377–386. (3) DeSouza, L. V.; Grigull, J.; Ghanny, S.; Dube´, V.; Romaschin, A. D.; Colgan, T. J.; Siu, K. W. Endometrial carcinoma biomarker discovery and verification using differentially tagged clinical samples with multidimensional liquid chromatography and tandem mass spectrometry. Mol. Cell. Proteomics 2007, 6, 1170–1182. (4) Dube´, V.; Grigull, J.; DeSouza, L. V.; Ghanny, S.; Colgan, T. J.; Romaschin, A. D.; Siu, K. W. Verification of endometrial tissue biomarkers previously discovered using mass spectrometry-based proteomics by means of immunohistochemistry in a tissue microarray format. J. Proteome Res. 2007, 6, 2648–2655. (5) Finlay, E. M.; Games, D. E.; Startin, J. R.; Gilbert, J. Screening, confirmation, and quantification of sulphonamide residues in pig kidney by tandem mass spectrometry of crude extracts. Biomed. Environ. Mass Spectrom. 1986, 13, 633–639. (6) Phillips, W. H., Jr; Ota, K.; Wade, N. A. Tandem mass spectrometry (MS/MS) utilizing electron impact ionization and multiple reaction monitoring for the rapid, sensitive, and specific identification and quantification of morphine in whole blood. J. Anal. Toxicol. 1989, 13, 349–355. (7) Anderson, L.; Hunter, C. L. Quantitative mass spectrometric multiple reaction monitoring assays for major plasma proteins. Mol. Cell. Proteomics 2006, 5, 573–588.
research articles (8) Luna, L. G.; Williams, T. L.; Pirkle, J. L.; Barr, J. R. Ultra performance liquid chromatography isotope dilution tandem mass spectrometry for the absolute quantification of proteins and peptides. Anal. Chem. 2008, 80, 2688–2693. (9) Ralhan, R.; Desouza, L. V.; Matta, A.; Tripathi, S. C.; Ghanny, S.; Gupta, S. D.; Bahadur, S.; Siu, K. W. Discovery and verification of head-and-neck cancer biomarkers by differential protein expression analysis using iTRAQ-labeling and multidimensional liquid chromatography and tandem mass spectrometry. Mol. Cell. Proteomics 2008, 7, 1162–1173. (10) Siu, K. W. M.; Bednas, M. E.; Berman, S. S. Determination of chromium in seawater by isotope dilution gas chromatography/ mass spectrometry. Anal. Chem. 1983, 55, 473–476. (11) Colby, B. N.; McCaman, M. W. A comparison of calculation procedures for isotope dilution determinations using gas chromatography mass spectrometry. Biomed. Mass Spectrom. 1979, 6, 225–230. (12) Mykytluk, A. P.; Russell, D. S.; Sturgeon, R. E. Simultaneous determination of iron, cadmium, zinc, copper, nickel, lead, and uranium in sea water by stable isotope dilution spark source mass spectrometry. Anal. Chem. 1980, 52, 1281–1283. (13) Moore, L. J.; Machlan, L. A. High accuracy determination of calcium in blood serum by isotope dilution mass spectrometry. Anal. Chem. 1972, 44, 2291–2296. (14) Cohen, A.; Hertz, H. S.; Mandel, J.; Paule, R. C.; Schaffer, R.; Sniegoski, L. T.; Sun, T.; Welch, M. J.; White, E. 5th. Total serum cholesterol by isotope dilution/mass spectrometry: a candidate definitive method. Clin. Chem. 1980, 26, 854–860. (15) White, E.; Welch, V. M.; Sun, T.; Sniegoski, L. T.; Schaffer, R.; Hertz, H. S.; Cohen, A. The accurate determination of serum glucose by isotope dilution mass spectrometry-two methods. Biomed. Mass Spectrom. 1982, 9, 395–405. (16) Barr, J. R.; Maggio, V. L.; Patterson, D. G., Jr.; Cooper, G. R.; Henderson, L. O.; Turner, W. E.; Smith, S. J.; Hannon, W. H.; Needham, L. L.; Sampson, E. J. Isotope dilution-mass spectrometric quantification of specific proteins: model application with apolipoprotein A-I. Clin. Chem. 1996, 42, 1676–1682. (17) Mazurek, S.; Boschek, C. B.; Hugo, F.; Eigenbrodt, E. Pyruvate kinase type M2 and its role in tumor growth and spreading. Semin. Cancer Biol. 2005, 15, 300–308. (18) Ugurel, S.; Bell, N.; Sucker, A.; Zimpfer, A.; Rittgen, W.; Schadendorf, D. Tumor type M2 pyruvate kinase (TuM2-PK) as a novel plasma tumor marker in melanoma. Int. J. Cancer 2005, 117, 825– 830. (19) Christofk, H. R.; Vander Heiden, M. G.; Harris, M. H.; Ramanathan, A.; Gerszten, R. E.; Wei, R.; Fleming, M. D.; Schreiber, S. L.; Cantley, L. C. The M2 splice isoform of pyruvate kinase is important for cancer metabolism and tumour growth. Nature 2008, 452, 230– 233. (20) Sa¨vblom, C.; Malm, J.; Giwercman, A.; Nilsson, J. A.; Berglund, G.; Lilja, H. Blood levels of free-PSA but not complex-PSA significantly correlates to prostate release of PSA in semen in young men, while blood levels of complex-PSA, but not free-PSA increase with age. Prostate 2005, 65, 66–72. (21) Olsen, J. V.; Blagoev, B.; Gnad, F.; Macek, B.; Kumar, C.; Mortensen, P.; Mann, M. Global, in vivo, and site-specific phosphorylation dynamics in signaling networks. Cell 2006, 127, 635–648. (22) Beausoleil, S. A.; Ville´n, J.; Gerber, S. A.; Rush, J.; Gygi, S. P. A probability-based approach for high-throughput protein phosphorylation analysis and site localization. Nat. Biotechnol. 2006, 24, 1285–1292. (23) Molina, H.; Horn, D. M.; Tang, N.; Mathivanan, S.; Pandey, A. Global proteomic profiling of phosphopeptides using electron transfer dissociation tandem mass spectrometry. Proc. Natl. Acad. Sci. U.S.A. 2007, 104, 2199–2204. (24) Stemmann, O.; Zou, H.; Gerber, S. A.; Gygi, S. P.; Kirschner, M. W. Dual inhibition of sister chromatid separation at metaphase. Cell. 2001, 107, 715–726. (25) Gerber, S. A.; Rush, J.; Stemman, O.; Kirschner, M. W.; Gygi, S. P. Absolute quantification of proteins and phosphoproteins from cell lysates by tandem MS. Proc. Natl. Acad. Sci. U.S.A. 2003, 100, 6940– 6945. (26) Anderson, N. L.; Anderson, N. G.; Haines, L. R.; Hardie, D. B.; Olafson, R. W.; Pearson, T. W. Mass spectrometric quantitation of peptides and proteins using Stable Isotope Standards and Capture by Anti-Peptide Antibodies (SISCAPA). J. Proteome Res. 2004, 3, 235–244. (27) Schneider, J.; Bitterlich, N.; Schulze, G. Improved sensitivity in the diagnosis of gastro-intestinal tumors by fuzzy logic-based tumor marker profiles including the tumor M2-PK. Anticancer Res. 2005, 25, 1507–1515.
Journal of Proteome Research • Vol. 7, No. 8, 2008 3533
research articles (28) Dombrauckas, J. D.; Santarsiero, B. D.; Mesecar, A. D. Structural basis for tumor pyruvate kinase M2 allosteric regulation and catalysis. Biochemistry 2005, 44, 9417–9429. (29) Warburg, O. On the origin of cancer cells. Science 1956, 123, 309– 314. (30) Christofk, H. R.; Vander Heiden, M. G.; Wu, N.; Asara, J. M.; Cantley, L. C. Pyruvate kinase M2 is a phosphotyrosine-binding protein. Nature 2008, 452, 181–186. (31) Shen, Y.; Kim, J.; Strittmatter, E. F.; Jacobs, J. M.; Fang, R.; Tolie, N.; Moore, R. J.; Smith, R. D. Characterization of the human blood plasma proteome. Proteomics 2005, 5, 4034–4045. (32) Omenn, G. S.; States, D. J.; Adamski, M.; Blackwell, T. W.; Menon, R.; Hermjakob, H.; Apweiler, R.; Haab, B. B.; Simpson, R. J.; Eddes, J. S.; Kapp, E. A.; Moritz, R. L.; Chan, D. W.; Rai, A. J.; Admon, A.; Aebersold, R.; Eng, J.; Hancock, W. S.; Hefta, S. A.; Meyer, H.; Paik,
3534
Journal of Proteome Research • Vol. 7, No. 8, 2008
DeSouza et al. Y. K.; Yoo, J. S.; Ping, P.; Pounds, J.; Adkins, J.; Qian, X.; Wang, R.; Wasinger, V.; Wu, C. Y.; Zhao, X.; Zeng, R.; Archakov, A.; Tsugita, A.; Beer, I.; Pandey, A.; Pisano, M.; Andrews, P.; Tammen, H.; Speicher, D. W.; Hanash, S. M. Overview of the HUPO Plasma Proteome Project: results from the pilot phase with 35 collaborating laboratories and multiple analytical groups, generating a core dataset of 3020 proteins and a publicly-available database. Proteomics 2005, 5, 3226–3245. (33) Li, H.; DeSouza, L. V.; Ghanny, S.; Li, W.; Romaschin, A. D.; Colgan, T. J.; Siu, K. W. M. Identification of candidate biomarker proteins released by human endometrial and cervical cancer cells using two-dimensional liquid chromatography/tandem mass spectrometry. J. Proteome Res. 2007, 6, 2615–2622.
PR800312M
