ARTICLE pubs.acs.org/ac
Narrow Mass Extraction of Time-of-Flight Data for Quantitative Analysis of Proteins: Determination of Insulin-Like Growth Factor-1 Cory E. Bystrom, Shijun Sheng, and Nigel J. Clarke* Quest Diagnostics, Nichols Institute, San Juan Capistrano, California 92694, United States ABSTRACT: Methods for quantitative analysis of proteins by mass spectrometry have progressed dramatically. While isotope-dilution approaches using selected reaction monitoring of tryptic peptides (also known as bottom up) have become common, the potential to use narrow mass extraction of high-resolution mass spectra provides a compelling alternative. We investigated the relationships between instrument performance and data processing with the aim of determining whether this approach can lead to robust bioanalytical assays for proteins. Our approach utilized off-line sample preparation combined with online sample extraction coupled to HPLC with the effluent from the analytical column directed to a high-resolution, high-mass accuracy quadrupole time-of-flight (qTOF) mass spectrometer operated in full scan mode. Narrow mass extraction of a single isotope from IGF-1 in the 7+ charge state (m/z 1093.5209) was used to generate extracted ion chromatograms. We found that with appropriate attention to instrument performance and data processing, quantitative protein assays with good sensitivity, high selectivity, and excellent analytical performance can be developed.
T
he quantitative analysis of proteins by mass spectrometry has received significant interest in the past decade as a result of the emergence of proteomics as a tool for biomarker discovery and systems biology research. While relative quantitation of proteins has been a key application of mass spectrometry enabled proteomics, the drive toward “absolute” quantitative methods has been motivated by interest in developing biomarkers for clinical diagnostics where relative quantitative measurements would be uncommon. These strategies have relied on the use of selected reaction monitoring (SRM) experiments carried out on triple quadrupole mass spectrometry to perform isotope dilution analyses of protein tryptic digests.1 3 Although there has been significant effort in elaborating platforms and workflows for the high-throughput implementation of SRM based protein analysis in the clinical lab, the approach remains challenging.4 Stoichiometric release and recovery of peptides, digest and stability kinetics,5 proper treatment of digest controls,6,7 and identification of appropriate SRM transitions are among the considerations necessary for successful assay design.8 In the laboratory, sample preparation is time-consuming and complex making the commercialization of liquid chromatography tandem mass spectrometry (LC MS/MS) based assays of proteins in the clinical lab slow. With the growing availability of high-precision mass spectrometers (time-of-flight, Fourier transform ion cyclotron resonance, and Orbitrap analyzers) that are capable of routine measurements with low part per million mass error and resolution >20 000 FWHM, groups have started to explore the use of accurate mass approaches to quantitative analysis of small molecules.9 In these experiments, assay specificity traditionally provided by specific molecular fragments derived from tandem r 2011 American Chemical Society
MS is replaced by molecular formula constraints afforded by accurate mass measurements. One advantage of these approaches is that very general analytical strategies can be employed without preconditions required when specific molecular fragments are used to develop an assay. Given the technical challenges and complexity of protein analysis using peptides, we became interested in the application of accurate mass approaches to quantitative analysis of proteins. In particular, we wished to gain an understanding of the selectivity afforded by precise mass measurements and how to perform data analysis consistent with instrument and desired assay performance. Insulin-like growth factor-1 (IGF-1) is a clinically relevant protein that is monitored in the diagnosis and treatment of growth disorders. Although traditionally measured by immunoassay, comparison of IGF-1 results between platforms is complicated by differences in antibodies, detection methodology, sample preparation, and sensitivity to protein interferences. For this reason, each assay has a unique set of normative ranges making direct comparison of results complicated. There have been efforts to elaborate IGF-1 assays using SRM approaches, but these assays have not yet been implemented in the clinical lab for the reasons mentioned earlier.10 13 In this light, the possibility of analysis of undigested IGF-1 extracted from human serum using high-resolution, high-mass accuracy analysis is compelling: sample preparation is dramatically simplified using automated equipment already found in many clinical laboratories Received: July 13, 2011 Accepted: October 5, 2011 Published: October 05, 2011 9005
dx.doi.org/10.1021/ac201800g | Anal. Chem. 2011, 83, 9005–9010
Analytical Chemistry and the method is amenable to rigorous standardization using reference materials.
’ MATERIALS AND METHODS HPLC grade acetonitrile, absolute ethanol, and water were purchased from Burdick and Jackson (Muskegon, MI). Highpurity formic and hydrochloric acids and methylethanethiol were obtained from Fluka (St. Louis, MO). High-quality recombinant human IGF-1 (Uniprot P05019|49-118; expected mass 7648.74 Da, observed mass 7648.68 Da) was obtained from Anjinomoto Science (Raleigh, NC). Rat IGF-1 (Uniprot P08025|49-118; expected mass 7702.87 Da, observed mass 7702.95 Da) was obtained from Cell Sciences (Canton, MA) or Prospec Tany (Rehovot, Israel). Synthetic peptides for qualifier ion experiments (DVLTQHNAPV, m/z 1093.5637 [M + H]1+; IEVEGGEDVF, m/z 1093.5048 [M + H]1+; PWGDEAYAAGG m/z 1093.4585 [M + H]1+) were obtained from New England Peptide (Gardner, MA). QC pools were either pooled stripped serum from Goldenwest Biologicals (Temecula, CA) or BioRad Immunochek pools (Hercules, CA). All protein components were fully characterized by gel electrophoresis, HPLC, high-resolution mass spectrometry, and amino acid analysis prior to use. Preparation of Calibrator Standard Solutions, Internal Standard, QC Pools, and Peptides. A high concentration stock
of pure IGF-1 at 10 000 ng/mL was diluted into stripped serum to prepare a high calibrator solution at 2000 ng/mL. Calibrator standard solutions covering the range of 15.6 2000 ng/mL were prepared immediately prior to assay. Pooled human serum and commercial QC pools were purchased and stored at 80 °C until use. Internal standard (IS) was prepared by oxidizing rat IGF-1 with 1% hydrogen peroxide. After 30 min, excess peroxide was quenched via the addition of 2 mM methylthioethanethiol. Complete conversion to the oxidized form was checked prior to assay by LC MS. A stock of oxidized rat IGF-1 at 10 ng/μL was prepared in an artificial matrix containing 30 mg/mL serum albumin in phosphate buffered saline (0.01 M phosphate, 2.7 mM KCl, 137 mM NaCl, pH 7.4). Peptides for infusion experiments were resuspended in water or DMSO to an initial concentration of 10 pmol/μL. Dilutions were prepared as needed in water containing 0.1% formic acid. Patient Samples. Normal control subjects provided written informed consent prior to blood draw. An institutional review board (IRB) waiver was obtained for collection of samples from control subjects and for utilization of residual deidentified clinical samples. In all cases, blood was collected into barrier-free serum preparation (red top) tubes, allowed to clot, then immediately processed to obtain serum that was kept frozen at 80 °C until analysis. Sample Preparation. Each patient sample, QC pool, or calibrator standard solution was thawed and vortexed before a 100 μL aliquot was mixed with 10 μL of internal standard. After 400 μL of acid ethanol (87.5% EtOH, 12.5% 1 N HCl) was added, the sample was vigorously mixed and allowed to incubate at room temperature for 30 min. The precipitate was removed by centrifugation, and 350 μL of clarified supernatant was then mixed with 60 μL of 1.5 M Tris base. The samples were chilled for 30 min at 20 °C generating additional protein precipitate and the extract reclarified by centrifugation. This extract was used for LC MS analysis. Automated Preparative and Analytical Chromatography. Chromatographic separation of IGF-1 from matrix components
ARTICLE
prior to MS was accomplished with an Aria TX-4 (ThermoFisher, San Jose, CA), a fully automated online chromatography system for sample preparation and separation. IGF-1 and internal standard were isolated from the serum extract using an Onyx guard column as an online SPE cartridge (5 mm 4.6 mm, C18) (Phenomenex). After injection of the extract, the cartridge was washed with 85% solvent A (water + 0.2% formic acid). The analytes were then back-flushed off the extraction cartridge onto the analytical column. Analytical separation of IGF-1 and internal standard was performed with an Onyx monolithic column (50 mm 2.1 mm, C18) (Phenomenex, Torrance, CA) using a fast, multiphase, linear gradient of increasing concentration of solvent B (acetonitrile + 0.2% formic acid) in solvent A. Data Acquisition and Processing. Mass spectrometry data acquisition was carried out using an Agilent 6530 qTOF instrument (Santa Clara, CA) operated in 4 GHz mode. A JetStream source was operated under the following conditions: capillary voltage, 5000 V; nozzle voltage, 500 V; nebulizer, 50 psi; sheath gas, 5 L/min; sheath gas temperature, 250 °C. Full scan data from m/z 900 1100 at 4 Hz was collected in combined centroid and profile mode. Reference ions (m/z 922.0908) were delivered to the source continuously during data acquisition and used to correct the TOF calibration continuously. For quantitative analysis, the most intense isotopic peak of the 7+ ion of IGF-1 (m/z 1093.5209) was extracted using a ( 10 ppm symmetric extraction window. Two ions (m/z 1093.3745, m/z 1093.6640) in the same isotopic cluster were simultaneously extracted for use in isotope ratio calculations. Narrow mass extraction of the internal standard was performed likewise. The analyte peak area to IS peak area ratio was plotted against the known concentration to obtain calibration curves, which were calculated using weighted (1/x) linear, least-squares regression. Reduction of raw MS peak areas to concentrations was performed using MassHunter Quant software (Agilent, Santa Clara, CA). Results were reported as the concentration of IGF-1 in nanogram/milliliter.
’ RESULTS AND DISCUSSION Sample Extraction and HPLC Method Development. Given the complexity of human serum and the overwhelming abundance of major protein components, fractionation prior to mass spectrometry analysis was required. In addition, IGF-1 circulates as a member of a protein complex assembled from insulin-like growth factor binding protein and an acid labile subunit. The quantitative release of IGF-1 from this complex in the presence of acid has been well characterized, and acid ethanol extraction is a standard technique in many IGF-1 assays.14 A modified procedure was reported in 1991 which relied on acid ethanol extraction followed by neutralization and cryoprecipitation.15 We found that the addition of the cryoprecipitation step was helpful in reducing background and reaching the highest levels of sensitivity in our assay. After extraction, additional fractionation was required to reduce suppression and background interference from matrix components. During selection of extraction media, we found that Onyx monolithic guard columns were well suited for use as an online extraction cartridge. Using traditional reverse phase extraction media, we observed that increasing the stringency of the column wash by increasing the percentage of acetonitrile in the mobile phase lead to a proportional reduction in IGF-1 recovery. In contrast, the monolithic material retained IGF-1 9006
dx.doi.org/10.1021/ac201800g |Anal. Chem. 2011, 83, 9005–9010
Analytical Chemistry
ARTICLE
Figure 1. Example chromatograms and averaged spectra for IGF-1 in low level calibrators (15.6 ng/mL) and patient samples (25 ng/mL). The isotopic resolution of the most abundant charge state (7+) for IGF-1 is shown with the quantifier (m/z 1093.5209) and two qualifier ions (m/z 1093.3778 and 1093.6641) indicated.
Table 1. Linear Regression Comparison of IGF-1 Results Determined at Various Extraction Widths Compared to 120 ppm Extraction Widtha centroid
Figure 2. Mean and standard deviation of mass determinations for IGF1 quantifier ion as a function of intensity across a chromatographic peak. The dashed lines represent a hypothetical chromatographic peak.
nearly quantitatively thus allowing aggressive washing of the extraction column, thereby removing interfering substances and excess protein that could not be removed otherwise. This allowed for exceptional analyte recovery and dramatic reduction of background both of which contribute directly to the high sensitivity achieved by our assay (LOQ 15 ng/mL). Gradient HPLC after online extraction was used to resolve IGF-1 from the residual protein interferences in the mixture. The Onyx monolithic column demonstrated high resolution with peak widths of 6 s at a flow rate of 0.8 mL/min. MS Detection Optimization. Detection of IGF-1 was established using a narrow mass extraction of full scan spectra. The highest peak in the [M + 7H]7+ isotopic cluster (m/z 1093.5209) was selected as a quantifier ion (Figure 1). The baseline resolved width of this peak is approximately 120 ppm. Extraction window selection to provide high selectivity while preserving sensitivity was undertaken Although the mass accuracy specification for the Agilent 6530 is
