Targeted Quantitative Using Selected Reaction Monitoring (SRM)- Coupled with an 18O-labeled Reference as Internal Standards Jong-Seo Kim, Errol Robinson, Boyd L. Champion, Brianne O. Petritis, Thomas L. Fillmore, Ronald J. Moore, Liu Tao, David G. Camp II, Richard D. Smith, and Wei-Jun Qian Biological Sciences Division, Pacific Northwest National Laboratory, Richland, WA

Methods Results Effect of Q1 resolution Reproducibility 18O labeling efficiency Overview Patient Samples Conclusions Group A: • To address the need of stable isotope Group B: • The utility of 18O-labeled “universal” O area ratio O area

Trypsin 18 Red error bars: label-free peak area ratio labeled internal standards for accurate (= each 16O area / average 18O area) reference as internal standards for targeted O/ digestion 16 18 quantification, we introduce the use of an 16 Green error bars: each pair area ratio of O/ O has been Patient peptide 18 Pooled reference ratio O area O-labeled reference as comprehensive samples sample 18 successfully demonstrated 16 18 LC-SRM-MS O/ Concentration ratio ( O/ O) internal standards for accurate SRM-MS- 18O labeling 16 – A linear dynamic range of quantification Concen. based quantification instead of using labeled 18 0.01 0.02 0.04 0.1 0.2 1 5 10 25 50 100 4 O-labeled “universal” Ratio ~ 10 in relative concentration synthetic peptides. reference 16O/18O 6.1 12.5 13.7 13.8 5.5 7.7 16.8 2.8 4.6 4.2 1.6 pairs – Better reproducibility than label-free 18 Label • O-labeling efficiency for most peptides is 39.2 23.4 47.5 26.5 19.3 10.6 31.2 3.2 10.7 4.8 3.5 Fig. 3 LC-LTQ-Orbitrap MS analysis of 18O labeled peptides. (Da) free (%) approach sufficiently high to provide a residue of Patient samples mixed with Fig. 5 Effect of Q1 window width on y type ions. 16O / 18O y – Sufficient effectiveness for monitoring 16O/18O ratio of <1%. 18O-labeled reference Fig. 7 Comparison of reproducibility (% CV in table) between ion pair without interference under wide Q1 window. 16 18 16 18 16 18 O/ O pair and label free ( O/ Oavg) approach. O and O y-fragments at unit or even 4 654.36

• A linear dynamic range of quantification ~10 754.40 LC-SRM-MS analysis 689.88 higher resolution → → in relative concentration was obtained for → Linear dynamic range of Absolute quantification by most of the peptides being monitored based Quantitative information based on 18 • Endogenous concentration of CRP and 536.29 O residual transition 536.29 16 18 536.29 O labeled reference–based standard addition on the observed peak area ratios (16O/18O) O/ O peak area ratios 16 spiked PSA in human plasma successfully quantification C-Reactive Prostate Specific Antigen versus concentration ratios for each peptide. Fig. 1 SRM quantification strategy using 18O-labeled “universal” reference.1 determined by “standard protein (AFVFPK) (SVILLGR) addition/18O-labeled reference” • 16O/18O based quantification also provided (354.71→490.30)/ (379.25→458.31)/ Non-depleted mouse plasma digest (356.71→494.31) (381.25→464.32) 656.36 756.40 significantly better reproducibility in terms of (~40 mg/mL total protein) 691.88 → → →

the coefficient of variances among the labeled O Acknowledgements transitions 18 O area ratio O area O area ratio O area ratio 18 18 18 537.63 537.63 triplicate analyses compared to label-free Add 6 standard protein digests into plasma 537.63 Portions of this work were supported by the NIH Director's New O/ O/ at 0.001%, 0.01%, 0.1% and 1% (w/w) to total mouse plasma digest O/ Innovator Award Program 1-DP2OD006668-01 (to W.J.Q.) and NIH 16 16 16 quantification. National Center for Research Resources RR18522 (to R.D.S.) 18O labeling Experiment work was performed in the Environmental Molecular Fig. 4 LC-TSQ-SRM chromatogram showing 16O residual (top) Standard added CRP (μg/mL) Standard added PSA (μg/mL) Science Laboratory, a DOE/BER national scientific user facility at 16 18 18 O-samples O-labeled reference and O labeled (bottom) transitions of VDEDQPFPAVPK(3+) Pacific Northwest National Laboratory (PNNL) in Richland, peptide after 18O labeling. Fig. 8 Standard addition calibration curve for absolute Introduction Select reference level Washington. PNNL is operated for the DOE by Battelle under contract quantification of endogenous CRP and spiked PSA in non- DE-AC05-76RLO-1830. 16 depleted human female plasma. • Selected reaction monitoring (SRM)-MS coupled with Combine 16O-labeled samples and 18O-labeled reference Table 1. Percentage of O residual (unlabeled) peptide in an stable isotope dilution is becoming a powerful tool for (16O at 0, 5, 10, 25, 50, 100, 500 ng/mL per 18O at 500 ng/mL, 18O-labeled reference samples determined by SRM mode 16 18 targeted quantitative proteomics, e.g., biomarker O at 50, 250, 500, 1250, 2500, 5000 ng/mL per O at 50 ng/mL) % of 16O % CV References Protein Sequence residual (triplicate 354.71 379.25 verification. ratio O area peptides* analyses) 18 →490.30 →458.31 1. Qian W-J, et al. Large-Scale Multiplexed Quantitative Discovery

O/ Proteomics Enabled by the Use of an 18O-Labeled "Universal"

• Selective and sensitive detection of specific peptides LC-SRM-MS analysis and standard calibration curve Carbonic anhydrase DGPLTGTYR 0.35 10.7 16 Reference Sample. J Proteome Res 8:290-299 (2009). are achieved by selecting a specific parent ion and Carbonic anhydrase DFPIANGER 0.50 27.8 Fig. 2 Flow chart for preparing the non-depleted mouse plasma samples 2. Petritis BO et al. A Simple Procedure for Effective Quenching of monitoring specific fragment ions (See Fig.1). Carbonic anhydrase EPISVSSQQMLK 0.73 33.5 18 with 6 spiked standard protein digests for calibration curve. Area Area Trypsin Activity and Prevention of O-Labeling Back-Exchange. J ratio Proteome Res 8:2157-2163 (2009). Triple Quadrupole MS Carbonic anhydrase AVVQDPALKPLALVYGEATSR 0.15 10.1 356.71 ratio 381.25 Product ion →494.31 0.37 →462.32 0.33 (Selected Reaction Monitoring Mode) chromatogram β‐Lactoglobulin LIVTQTMK 0.29 35.8 • Bovine carbonic anhydrase, beta-lactoglobulin, cytochrome c, chicken Q1 Q2 Q3 ovalbumin, equine skeletal muscle myoglobin and E.coli beta- β‐Lactoglobulin VLVLDTDYKK 0.06 4.2 Detector galactosidase were used as 6 standard . β‐Lactoglobulin TPEVDDEALEK 2.27 13.2 LC-ESI Precursor Product • C-reactive protein (CRP) and prostate specific antigen (PSA) were β‐Lactoglobulin VYVEELKPTPEGDLEILLQK 0.20 13.0 ion ion 16 18 spiked in human female plasma (non-depleted) for absolute Cytochrome c EDLIAYLK 0.09 37.7 ratio O area Fig. 9 XIC of O/ O transition pair for endogenous CRP (left)

selection selection 18 Time (min) quantification using standard addition. and spiked PSA (right) at 500 ng/mL in non-depleted human Cytochrome c TGPNLHGLFGR 0.93 17.6 O/ 18 16 female plasma. CONTACT: Jong-Seo Kim, Ph.D. • To achieve accurate quantification, isotope labeled • O labeling was performed with the recently published protocol to Myoglobin LFTGHPETLEK 0.57 14.0 completely prevent 18O back-exchange.2 Biological Sciences Division, K8-98 synthetic peptides are often used as internal Myoglobin VEADIAGHGQEVLIR 0.48 3.0 Table 2. SRM Quantification results of endogenous CRP and • LC: Agilent 1100; 75 μm x 25 cm column (3 μm; C18); 60-min gradient; spiked PSA (500 ng/mL) in non-depleted human female Pacific Northwest National Laboratory standards, which limits the feasibility to monitor a Ovalbumin AFKDEDTQAMPFR 0.05 21.6 400 nL/min. Concentration ratio (16O/18O) plasma by standard addition method. P.O. Box 999, Richland, WA 99352 large number of candidates due to expense. Ovalbumin GGLEPINFQTAADQAR 0.25 34.5 • MS: Thermo Scientific TSQ vantage. (354.71→ 490.30) / (356.71→494.31) E-mail: [email protected] Fig. 6 Calibration curve showing linear dynamic range of CRP (ng/mL) • The 18O-labeled “universal” reference will serve as β‐Galactosidase VDEDQPFPAVPK 0.35 32.8 concentration CV • SRM: 10 ms dwell time; 0.002 Da scan width; unit resolution in Q1 and quantification ~104 in relative concentration. (endogenous) ideal internal standards for quantitative monitoring for Q3 (0.7 FWHM). β‐Galactosidase LWSAEIPNLYR 0.11 6.9 790 ng/mL 4.1% (379.25→458.31) / (381.25→464.32) any candidates of interest. * An average of ~0.5% 16O/18O ratio indicates the high percentage of heavy PSA (ng/mL) • Pinpoint software (Thermo Fisher Scientific) was used for data analysis. concentration CV label incorporation, a necessity for accurate quantification. (spiked) 800 ng/mL 0.8%