Multiple reaction monitoring-based, multiplexed, absolute quantitation of 45 proteins in human plasma

Abstract

Mass spectrometry-based multiple reaction monitoring (MRM) quantitation of proteins can dramatically impact the discovery and quantitation of biomarkers via rapid, targeted, multiplexed protein expression profiling of clinical samples. A mixture of 45 peptide standards, easily adaptable to common plasma proteomics work flows, was created to permit absolute quantitation of 45 endogenous proteins in human plasma trypsin digests. All experiments were performed on simple tryptic digests of human EDTA-plasma without prior affinity depletion or enrichment. Stable isotope-labeled standard peptides were added immediately following tryptic digestion because addition of stable isotope-labeled standard peptides prior to trypsin digestion was found to generate elevated and unpredictable results. Proteotypic tryptic peptides containing isotopically coded amino acids ([(13)C(6)]Arg or [(13)C(6)]Lys) were synthesized for all 45 proteins. Peptide purity was assessed by capillary zone electrophoresis, and the peptide quantity was determined by amino acid analysis. For maximum sensitivity and specificity, instrumental parameters were empirically determined to generate the most abundant precursor ions and y ion fragments. Concentrations of individual peptide standards in the mixture were optimized to approximate endogenous concentrations of analytes and to ensure the maximum linear dynamic range of the MRM assays. Excellent linear responses (r > 0.99) were obtained for 43 of the 45 proteins with attomole level limits of quantitation (<20% coefficient of variation) for 27 of the 45 proteins. Analytical precision for 44 of the 45 assays varied by <10%. LC-MRM/MS analyses performed on 3 different days on different batches of plasma trypsin digests resulted in coefficients of variation of <20% for 42 of the 45 assays. Concentrations for 39 of the 45 proteins are within a factor of 2 of reported literature values. This mixture of internal standards has many uses and can be applied to the characterization of trypsin digestion kinetics and plasma protein expression profiling because 31 of the 45 proteins are putative biomarkers of cardiovascular disease.

Figures

Fig. 1.

Strategy for generation of highly sensitive and specific MRM protein assays. Isotopically labeled versions of proteotypic, endogenous, tryptic peptides representing each protein (selected from either Anderson and Hunter (14) or the Global Proteome Machine database) were synthesized and purified by reversed-phase HPLC. Concentrations of purified SIS peptides were determined by amino acid analysis, and the peptide concentration of each SIS peptide was corrected by the percent purity as determined by capillary zone electrophoresis. MRM Q1/Q3 ion pairs and parameters were determined empirically. Identities of natural peptide MRM signals were confirmed by co-elution with their isotopically labeled forms when analyzed by LC-MRM/MS. Isotopically labeled peptides were titrated to determine the concentration that produced a peak area ratio 1–10 times higher than the natural peptide peak area when added to a plasma tryptic digest.

Fig. 2.

Peptide solutions were analyzed by infusion. Nanoelectrospray was used for the selection of the most intense precursor and fragment ions for MRM assay development. Q1 scans with a DP voltage ramp were used to determine the dominant charge state of each ion and its optimal DP voltage. XICs of the double and triple charge states are presented for transferrin (a) and apolipoprotein A-I (b). MRM scanning with a CE voltage ramp was conducted using Q1/Q3 pairs for all possible singly and doubly charged b and y fragment ions and the most intense Q1 m/z. XICs for the most intense MRM Q1/Q3 pairs are presented for transferrin (c) and apolipoprotein A-I (d). The three most intense MRM pairs identified by nanoinfusion analysis are presented in addition to two MRM pairs (bold) containing Q3 fragment ions frequently used when generating MRM Q1/Q3 ion pairs in silico using MIDAS Workflow Designer (the b2 ion and one y ion are greater than the precursor). The vertical dashed lines represent the calculated CE voltage using a generic formula (CE = (precursor m/z) × 0.05 + 5) commonly used during MS/MS analysis and creation of MRM Q1/Q3 ion pairs using MIDAS.

Fig. 3.

Multiplexed MRM quantitation of 45 proteins in a single LC-MRM/MS analysis. a, each MRM assay contains two Q1/Q3 ion pairs to permit discrimination of co-eluting peaks for natural and isotopically labeled peptides. XICs of MRM ion pairs for natural (blue) and heavy peptides (red) reveal peak areas of heavy peptides within 10-fold of natural peptide levels. b, XICs of all 45 MRM protein assays in a single 60-min LC-MRM/MS analysis of 1 μg of plasma tryptic digest spiked with a concentration-balanced mixture of 45 SIS peptide internal standards. MRM ion pair XICs of natural peptides are blue, and SIS peptides are red. Signal intensity of the natural albumin peptide has been rescaled by a factor of 0.2. HC, heavy chain; cps, counts/s.

Fig. 3.

Multiplexed MRM quantitation of 45 proteins in a single LC-MRM/MS analysis. a, each MRM assay contains two Q1/Q3 ion pairs to permit discrimination of co-eluting peaks for natural and isotopically labeled peptides. XICs of MRM ion pairs for natural (blue) and heavy peptides (red) reveal peak areas of heavy peptides within 10-fold of natural peptide levels. b, XICs of all 45 MRM protein assays in a single 60-min LC-MRM/MS analysis of 1 μg of plasma tryptic digest spiked with a concentration-balanced mixture of 45 SIS peptide internal standards. MRM ion pair XICs of natural peptides are blue, and SIS peptides are red. Signal intensity of the natural albumin peptide has been rescaled by a factor of 0.2. HC, heavy chain; cps, counts/s.

Fig. 4.

Analytical reproducibility of MRM-based quantitation. CV frequencies of 45 peptide assays using raw peak areas (a), peak area ratios normalized to an equimolar SIS peptide mixture (b), and peak area ratios normalized to a concentration-balanced SIS peptide mixture (c) are shown.

Fig. 5.

Calibration curves and response factor plot for apolipoprotein A-I. a, linear regression analysis (1/x weighted) and response factor plot (blue line) for apoA-I using all analyte concentrations. b, final linear regression and response factor plot for apoA-I illustrating the linear dynamic range of the assay when analyte concentrations that respond non-linearly are excluded (◇).