Targeting Interferon Alpha Subtypes in Serum: A Comparison of

Apr 7, 2009 - Both triple quadrupole selected reaction monitoring and orbitrap selected ion monitoring produced linear calibration curves from 1 ng/mL...
0 downloads 0 Views 814KB Size
Download PDF
Targeting Interferon Alpha Subtypes in Serum: A Comparison of Analytical Approaches to the Detection and Quantitation of Proteins in Complex Biological Matrices Anita Izrael-Tomasevic, Lilian Phu, Qui T. Phung, Jennie R. Lill, and David Arnott* Protein Chemistry Department, Genentech Inc., South San Francisco, California 94080 Received January 30, 2009

The targeted detection and quantitation of proteins in complex biological fluids such as blood is as analytically challenging as it is crucial for biomedical research. Antibody-based techniques such as the ELISA are the current standards for such measurements, having in favorable cases high specificity and pg/mL detection limits. Long development timelines and susceptibility to cross reactivity have led researchers to investigate mass spectrometric alternatives. The literature contains diverse schemes for sample preparation and multiple platforms for mass spectrometric detection. Critical evaluations of competing technologies are, however, badly needed. Taking closely related subtypes of the proinflammatory cytokine interferon alpha as a test case, we compared a sample preparation workflow based on affinity enrichment to one based on generic multidimensional chromatography, and evaluated mass spectrometric techniques using tandem mass spectrometry on low resolution ion traps, high resolution “accurate mass tags,” and triple quadrupole selective reaction monitoring. Each workflow and detection method proved capable of detecting and discriminating between these proteins at or below the ng/mL level in human serum. Quantitation by isotope dilution was evaluated using full length protein as the internal standard. Both triple quadrupole selected reaction monitoring and orbitrap selected ion monitoring produced linear calibration curves from 1 ng/mL to 1 µg/mL, with lower limits of quantitation below 5 and 50 ng/mL, respectively. Keywords: mass spectrometry • LC-MSn • interferon alpha • quantitation • accurate mass tag • isotope dilution • serum

Introduction One of the major challenges in the proteomic analysis of biological fluids is the range of protein concentrations that is encountered. Well-known components of blood, for example, span 10 orders of magnitude, and the full range of concentrations is likely to be greater still.1 Highly abundant serum species such as albumin and the immunoglobulins typically mask the detection of low-level species by conventional proteome profiling methodologies. Immuno-assays such as the enzyme-linked immunosorbent assay (ELISA) have historically been employed, as these have very high sensitivity andsin many casesssufficient selectivity even in complex biological matrices.2 Cross reactivity of antibodies can nevertheless pose a problem, particularly for protein families possessing substantial sequence identity. Immuno-assays require that an antibody be raised against the antigen of interest and fully characterized, an undertaking that frequently requires months, and producing an antibody specific to just one family member or isoform is not always possible. The alpha interferons are a striking case in point. A family of cytokines with potent antiviral activity, the interferon alphas (IFNA) consist of 13 genes encoding 12 unique protein sub* To whom correspondence should be addressed. David Arnott, 1 DNA Way, M.S. 63, South San Francisco, CA 94080. Telephone: (650) 225-1240. FAX: (650) 225-5945. E-mail: [email protected].

3132 Journal of Proteome Research 2009, 8, 3132–3140 Published on Web 04/07/2009

types; IFNA13 and IFNA1 encode identical protein sequences. Each subtype has a distinct activity despite sharing high sequence identity.3 Essentially undetectable in the blood of healthy humans, IFNA levels rise in response to infection (primarily but not exclusively viral), and elevated levels of IFNalpha in serum correlate with the development of a host of autoimmune diseases including systemic lupus erythematosus (SLE) and insulin dependent diabetes mellitus (IDDM).4,5 Serum IFNA levels have been assayed according to biological activity, or by antibody-based methods including ELISA or biochip array technology.2,6 Although these assays have high sensitivity, they cannot currently measure IFNA at the level of individual subtypes, and since the subtypes of IFNA elevated in SLE and IDDM patient’s serum are not fully characterized, the need exists for an alternative that retains the sensitivity of immunoassays but has even higher specificity.7 Tandem mass spectrometric analysis should in principle provide the needed specificity. Despite high sequence identitys over 95% in some casessevery subtype contains at least one signature, or proteotypic, tryptic peptide, whose mass and products upon collision-induced dissociation (CID) distinguish it from any other human protein (Supplementary Figures 1 and 2, Supporting Information). These signature peptides have been identified in tryptic digests of each of the nine subtypes that were available to us (data not shown). Over the past decade, 10.1021/pr900076q CCC: $40.75

 2009 American Chemical Society

Detecting and Quantifying Interferon Alpha Subtypes in Serum researchers have used a variety of mass spectrometry based techniques to detect and sometimes quantify proteins in samples such as human serum, frequently in the pursuit of clinical biomarkers.8-10 Many such studies have been “unbiased” surveys that aimed to catalog large numbers of plasma proteins, either as an end in itself or as a comparison between sample sets. The dynamic range limitations imposed by the complexity of plasma have been dealt with by an array of separations techniques including depletion of abundant proteins and electrophoretic or chromatographic separations, applied to either the protein or peptide level or both.10-14 The range of mass spectrometry platforms is likewise diverse, ranging from relatively low resolution and inexpensive ion traps to quadrupole time-of-flight hybrids and ultra high resolution instruments like Fourier transform ion cyclotron or orbitrap mass spectrometers. Alternatively, targeted mass spectrometric experiments have also been performed, most commonly using triple quadrupole instruments in selective reaction monitoring (SRM) mode with quantitation as an objective. Several examples of serum protein quantitation have been published, with µg/mL limits of detection reported using relatively simple sample preparation schemes.15,16 Extending targeted experiments to the ng/mL level and below will be essential if low abundance proteins such as cytokines or tissue leakage proteins associated with disease are to be assayed. At least one report indicates that detection limits in the 10s of picograms are potentially achievable,17 but the optimal approaches to sample preparation and mass spectrometric analysis remains unclear. We therefore set out to compare sample preparation techniques based on either an affinity enrichment protocol or the more generally applicable combination of affinity depletion of high abundance proteins with proteolysis and two-dimensional chromatography at the peptide level. Mass spectrometric techniques evaluated included MS/MS and MS3 on ion trap instruments, triple quadrupole SRM, and “accurate mass tag” (AMT) selected ion monitoring (SIM) on FT-ICR or FT-orbitrap instruments. Two subtypes of interferon alpha (IFNA4 and IFNA2) were chosen as test cases. In addition to demonstrating the specificity of mass spectrometry in a case where antibody-based assays fail, the interferons have the advantage of being absent from normal serum so that well-characterized recombinant versions can be added without interference from the endogenous protein. Furthermore, the size and sequence identity among subtypes constrains the choice of signature peptide, forcing us to deal with nonideal peptide characteristics rather than choosing “best case” scenarios to showcase mass spectrometry. In addition to qualitative experiments, quantitation was performed using isotope dilution with stable isotope labeled (SIL) synthetic peptides and similarly labeled intact recombinant protein.

Experimental Section Affinity Enrichment. An IFNA antibody with broad specificity (in house number 1570 recognizing subtypes 1, 2, 4, 5, 8, 10, and 21) was coupled to an affinity resin (Affi Gel-10; Bio Rad Inc.). Coupling efficiency was determined to be over 90%, by bicinchoninic acid (BCA) assay. Three-hundred microliters of crude serum from healthy donor was spiked with different concentrations of recombinant IFNA subtypes 2 and 4, ranging from 833 pg to 333 ng/mL serum. Samples were mixed with 25 µL of affinity resin (in batch), and incubated for 2.5 h on a rotational shaker at RT. After incubation, the resin was washed with PBS (400 µL × 5) followed by 400 µL of deionized distilled

research articles

water and eluted with 50 µL of 50% propanol/0.1% TFA. The eluate was reduced to dryness by vacuum centrifugation and stored at room temperature. Affinity Depletion. Recombinant human IFN-alpha subtypes 2 and 4 were spiked into the serum (concentrations ranging from 1 ng to 100 ng/mL) of a healthy donor for the purpose of assay development. Removal of the six most abundant proteins from human serum was performed on a 4.6 × 100 mm multiple affinity removal system (MARS) column (Agilent Technologies) and high performance liquid chromatography (SMART system; Pharmacia). This column specifically removes albumin, IgG, antitrypsin, IgA, transferrin and haptoglobin in a single step. Eighty µL of crude serum was diluted 5-fold with the manufacturer’s equilibration buffer, filtered through a 0.22 µm microcentrifuge filter (Agilent Technologies), and injected onto the depletion column. The 1 mL flow-through fraction was collected, concentrated, and desalted (final buffer composition: 300 µL of 50 mM ammonium bicarbonate) by ultrafiltration (3KDa cutoff, YM3 spin concentrators; Millipore). Trypsin digestion was performed immediately following the concentration step. Trypsin Digestion. Fresh or reconstituted protein samples in 300 µL of 50 mM ammonium bicarbonate were digested with trypsin in solution. Disulfide bonds were reduced with tris (2carboxyethyl phosphine) (3 mM for 1 h at room temperature), and cysteine residues alkylated with iodoacetamide (25 mM in the dark at room temperature for 45 min). After alkylation, iodoacetamide was quenched with an excess of L-cysteine. Trypsin (Promega sequencing grade) was added at a 1:40 enzyme to substrate ratio by weight and incubated at 37 °C overnight. Samples were dried by vacuum centrifugation and stored at room temperature. Strong Cation Exchange Chromatography of Peptides. Affinity depleted serum digests were fractionated by strong cation exchange chromatography. Approximately 200 µg of depleted serum tryptic digest was reconstituted in 50 µL of HPLC solvent A and applied to a polysulfoethyl aspartamide column (The Nest Group) for HPLC at a flow rate of 200 µL/ min using an acetate buffer system. Solvent A was 100 mM acetic acid, 8.54 mM ammonium acetate in 25% acetonitrile, pH 3.8, and solvent B was 34.2 mM acetic acid, 600 mM ammonium acetate in 25% acetonitrile, pH 5.9. A gradient of 0 to 10% B in 35 min followed by a ramp to 20% B in 10 and 2.5 min at 100% B was used to elute peptides. Fractions were collected at 1 min intervals, dried by vacuum centrifugation, and stored at room temperature. Online LC and Ion Trap MS/MS, MS3, and FTICR-MS. Tryptic peptide mixtures reconstituted in 25 µL of 0.1% TFA were injected in 8 µL aliquots onto a 150 mm by 0.15 mm i.d. reverse-phase HPLC column (C18; Microtech Scientific) and eluted directly into the nanoelectrospray emitter (Picotip 50 µm by 30 µm; New Objective, Inc.) of a linear ion trap Fourier transform ion cyclotron resonance mass spectrometer (LTQFT; ThermoFisher). The HPLC gradient was evolved as follows: 2 to 10% B in 2 min, 10 to 35% B in 10 min, 35 to 70% B in 25 min, and 70% B for 4 min, all at 1.0 µL per minute generated by a capillary HPLC system (1100 Series, Agilent Inc.). Solvent A was 0.1% formic acid in water, and solvent B was 0.1% formic acid in acetonitrile. A potential of 1800 V was applied to the electrospray emitter via liquid junction and the heated capillary of the electrospray source was maintained at 180 °C. Tryptic peptides unique to IFN-alpha 2 and 4 subtypes as well as a peptide common to all but one subtype, were analyzed Journal of Proteome Research • Vol. 8, No. 6, 2009 3133

research articles

Izrael-Tomasevic et al. 3

by targeted ion trap LC-MS/MS and MS . The signature peptide for IFN-alpha 2, HDFGFPQEEFGNQFQK, was selected and the [M + 3H]3+ precursor (m/z 659.7 or 689.7 for the 15N and 15N13C labeled proteins, respectively) were fragmented in the linear ion trap. Product ions were analyzed directly (MS/ MS) or the y152+ products (m/z 915.7 and 957.0) were isolated and further fragmented to yield second generation product ions (MS3). Likewise, the signature peptide for IFN-alpha 4, HDFGFPEEEFDGHQFQK, with [M + 3H]3+ ) 698.6 was fragmented and its products analyzed (MS/MS) or the y152+ ion at m/z 921.6 was isolated and further fragmented (MS3). A peptide common to all but one subtype, YSPCAWEVVR, (carboxyamidomethyl cysteine; [M + 2H]2+ ) 633.8 unlabeled or 640.8 where 15N labeled) was similarly subjected to CID and its products analyzed directly or the y82+ product ion (m/z 508.5 or 514.8 for unlabeled or 15N labeled protein respectively) isolated and fragmented in turn. Selected ion monitoring (SIM) was performed in the ICR region of the LTQ-FT. Spectra were acquired over a 20 m/z window centered on m/z 698 at a resolution of 100 000 M/∆M at m/z 400. An automatic gain control target of 5 × 104 was selected. Sample injections and chromatographic conditions were identical to those used for the ion trap tandem mass spectrometry experiments. External calibration was used and no mass corrections were applied postacquisition. Quantitation by LC-Orbitrap Selected Ion Monitoring. Selected ion monitoring of the IFN-alpha 2 signature peptide HDFGFPQEEFGNQFQK was performed by LTQ-orbitrap mass spectrometer interfaced with a nanoAcquity UPLC (Waters Corp.). Samples were prepared by adding unlabeled IFNA2 to 300 µL aliquots of normal human serum to produce the following concentrations: 1, 5, 10, 50, 250, and 1000 ng/mL along with a constant 50 ng/mL of 15N labeled IFNA2 as an internal standard. Samples were prepared by affinity enrichment, trypsin digestion, and reconstitution in 25 µL of 0.1% TFA containing 5 fmol/uL of stable isotope labeled (SIL) synthetic peptide. Each sample was analyzed in triplicate 5 µL injections onto a C18 HPLC column maintained at 40 °C (BEH 130; 100 mm by 0.1 mm i.d. with 1.7 µm particles; Waters Corp.) and eluted using an ultra high pressure HPLC system (nanoAcquity, Waters) with the following gradient: 2% B for 9.5 min, 2-5% B for 0.5 min, 5-22% B for 25 min, 22-90% B for 1 min and 90% B for 5.9 min. Solvent A was 0.1% formic acid in 98% water/2% acetonitrile, and solvent B was 0.1% formic acid in 98% acetonitrile/2% water. Peptides were eluted into the nanoelectrospray emitter (PicoTip, as above) of the mass spectrometer at 1 µL/min with an electrospray potential of 3000 V and the heated capillary temperature at 190 °C. SIM scans were acquired throughout the LC-MS experiment covering the m/z range 649 to 664, which contains the [M + 3H]3+ ions of the IFNA2 tryptic peptide and its stable isotope labeled counterparts: 652.296 for the unlabeled protein, 655.636 for the SIL synthetic peptide, and 659.935 for the 15N labeled protein. The orbitrap resolution was set to 60 000 M/∆M at m/z 400, and an automatic gain control target of 1 × 105 was employed. Quantitation by Triple Quadrupole Selected Reaction Monitoring. SRM experiments were performed on a hybrid triple quadrupole/linear trap Mass Spectrometer (4000 Q-Trap; Applied Biosystems) coupled to a Shimadzu 10 AD LC system (Shimadzu Corp.). Samples were prepared by adding unlabeled IFNA2 to 300 µL aliquots of normal human serum to produce the following concentrations: 1, 5, 10, 50, and 1000 ng/mL along with a constant 250 ng/mL of 15N labeled IFNA2 as an internal 3134

Journal of Proteome Research • Vol. 8, No. 6, 2009

standard. Samples were prepared by affinity enrichment, trypsin digestion, and reconstituted in 25 µL of 0.1% TFA containing 5 fmol/µL of SIL synthetic peptide. Samples were run in triplicates; full loop injections of 5 µL of each sample were performed and peptides were separated by RP-HPLC on a 150 mm by 2.0 mm i.d. C18 capillary column (Aquasil 3 µm particles; ThermoFisher). Solvent A was 0.1% formic acid in water and solvent B was 0.1% formic acid in 98% acetonitrile, 2% water. Peptides were eluted into the mass spectrometer at 200 µL/min with a gradient of 5-20% solvent B for 3.5 min, 20-35% solvent B for 15 min, 35-65% solvent B for 0.5 min, 65-90% solvent B for 0.5 min, and 90% solvent B for 2 min. Samples were ionized by electrospray ionization with the spray voltage set at 5200 V, curtain gas of 20 psi, nebulizer gas of 50 psi and an interface heating temperature of 550 °C. The transitions from the [M + 3H]3+ precursor at m/z 652.3 to 604.3 (b5) and 721.4 (y6) were monitored, as were the corresponding 15 N peptide transitions from m/z 659.9 to 611.2 and 731.3 and the SIL synthetic peptide’s from m/z 655.6 to 604.3. SRM parameters were defined as follows: Declustering potential (DP) was 70 V, collision energy was 30 V, Dwell time was 250 ms per transition, for a total cycle time of 1.275 s. Q1 was set at unit resolution, Q3 was set at low resolution.

Results Qualitative Analysis by Ion Trap MS/MS and MS3. Inspection of the amino acid sequence of IFNA4 revealed a single unique tryptic peptide. Digestion of recombinant IFNA4 with trypsin produced this peptide (HDFGFPEEEFDGHFQK; [M + H]+ ) 2093.9) along with an incompletely digested form of the peptide (DRHDFDFPEEEFDGHFQK) in approximately equal abundance (IFNA2 signature peptide behaved similarly; data not shown). The fully cleaved peptide was chosen as a signature peptide and used to establish the capabilities of a linear ion trap mass spectrometer for the targeted detection of IFNA4 in serum. Samples were prepared by the addition of recombinant IFNA4 to human serum followed by immunoprecipitation with a pan-IFNA antibody and digestion of the resulting protein mixture with trypsin. Capillary LC-MS/MS or MS3 were performed on an LTQ linear ion trap mass spectrometer, targeting the [M + 3H]3+ ion of the signature peptide. For purposes of comparison, 1 fmol of a pure, synthetic peptide with the IFNA4 sequence was analyzed with the same 150 µm inner diameter reverse phase column and instrument method used for the serum samples. The resulting product ion spectrum (Figure 1a) is rich in sequence specific b- and y-ions, the most prominent of which is y152+. Serum spiked with the 4.5 pmol of the recombinant protein (100 ng in 300 µL; 333 ng/ mL), affinity enriched, and digested yielded a recognizably similar product ion spectrum that contained all of the major b- and y-ions observed in the synthetic peptide spectrum in their correct relative abundances (Figure 1c). The spectrum is markedly noisier than that of the pure peptide, however, showing the effect of interfering isobaric peptides even at this level. An extracted ion chromatogram for the y152+ ion contains multiple peaks, of which the correct one is not the most intense (Figure 1d). At 225 fmol in 300 µL of serum (5 ng; 16.7 ng/mL) all but a handful of the prominent b- and y- ions are obscured by interfering peaks and the peptide is now a minor peak in the y152+ extracted ion chromatogram (Figure 1e, f). The loss of signal-to-noise could be reversed, however, by performing MS3: [M + 3H]3+ f y152+ f products. A sample containing 11.25 fmol of IFNA4 in 300 µL of serum (250 pg; 833 pg/mL) produced

Detecting and Quantifying Interferon Alpha Subtypes in Serum

research articles

Figure 1. Detection of IFNA4 by targeted ion trap LC-MS/MS and MS3. Tandem mass spectra (left panels) and associated reconstructed ion chromatograms (right panels) for the following experiments. (a and b) 1 fmol of IFNA4 stable isotope labeled synthetic peptide HDFGFPEEEFDGHQF*QK, where F* indicates 13C915N-Phe. (c and d) 100 ng unlabeled IFNA4 in 300 µL serum (333 ng/mL); affinity enrichment followed by digestion and LC-MS/MS; products of [M + 3H]3+, m/z 698.5. (e and f) 5 ng unlabeled IFNA4 in 300 µL serum (16.7 ng/mL); affinity enrichment, digestion, and LC-MS/MS. (g and h) 250 pg IFNA4 in serum (833 pg/mL); affinity enrichment, digestion, and LC-MS3. The y152+ product of m/z 698.5 was isolated and further fragmented to yield the second generation products shown. Sequence specific b- and y-ions are labeled throughout. In the MS3 spectrum b- ions are labeled with respect to the truncated peptide sequence FGFPEEEFDGHQFQK isolated from CID of the precursor peptide. Spectra correspond to chromatographic peaks indicated by arrows.

a second generation product ion spectrum that is rich in sequence specific ions with very little background (Figure 1g). An extracted ion chromatogram for the y122+ fragment contains a single peak (Figure 1h).

Affinity Depletion and 2D Chromatography. Having established that linear ion trap MS3 can detect the spiked protein below the 1 ng/mL level from samples prepared by immunoprecipitation, the experiments were performed with a more Journal of Proteome Research • Vol. 8, No. 6, 2009 3135

research articles

Izrael-Tomasevic et al. 15

and was obtained in a N isotopically labeled form. Targeted LC-MS/MS and MS3 experiments were performed, yielding high quality product and second generation product ion spectra (Supplementary Figure 3a, b, Supporting Information). The MS3 experiment, [M + 3H]3+ f b152+ f products, produced abundant neutral losses and b-ion rearrangements that are characteristic of MS3 on b ions, but all the major peaks can be assigned.18 When a sample containing 3.6 fmol of IFNA2 in 80 µL of serum (80 pg; 1 ng/mL) was prepared by the depletion/ digestion/2D-LC procedure, clean extracted ion chromatograms were obtained (Supplementary Figure 3c, Supporting Information). Likewise, 90 fmol in 300 µL serum (2 ng; 6.7 ng/ mL) was detected using the affinity enrichment protocol (Supplementary Figure 3d, Supporting Information).

Figure 2. Detection of IFNA4 without affinity enrichment. Eighty picograms of recombinant IFNA4 in 80 µL serum (1 ng/mL) was treated as follows. (a) Depletion of six most abundant serum proteins; IFNA was present in the flow through fraction of the multiple affinity removal column as demonstrated by Western blot (inset). (b) Tryptic peptides generated by digestion of the flow through fraction above were applied to a strong cation exchange column; the indicated SCX fraction was (c) analyzed by ion trap LC-MS3 targeting the IFNA4 signature peptide. The left panel is the reconstructed ion chromatogram; the spectrum corresponding to the major peak is on the right, labeled as in Figure 1.

general sample preparation scheme involving depletion of the six most abundant serum proteins, digestion, and twodimensional liquid chromatography. Recombinant IFNA4 was spiked into normal human serum and applied to the affinity depletion step. Western blotting demonstrated that IFNA4 was present in the flow-through fraction of the multiple affinity removal (MARS) column and was not detectable in the bound fraction containing serum albumin, IgG, IgA, transferrin, haptoglobin, and alpha-1-antitrypsin (Figure 2a). When digested with trypsin and resolved by strong cation exchange chromatography, the signature peptide was recovered in the fraction eluting at 30 min (Figure 2b). When 3.6 fmol of IFNA4 was added to 80 µL of serum (80 pg; 1 ng/mL) and the relevant SCX fraction injected on a 150 µm reverse phase column for targeted MS3 as above, the spectrum revealed the expected fragment ions and the extracted ion chromatogram contained a single prominent peak (Figure 2c). The generality of this approach was further tested by analyzing a second interferon subtype, IFNA2 which has a signature peptide with the sequence HDFGFPQEEFGNQFQK 3136

Journal of Proteome Research • Vol. 8, No. 6, 2009

A third example is provided by the peptide YSPCAWEVVR, which is a tryptic peptide common to all subtypes of IFNA except IFNA8. As before, ion trap MS/MS and MS3 experiments were devised (Supplementary Figures 4a, b, Supporting Information). The peptide was then detected in samples prepared by the depletion/digestion/2D-LC method and the affinity enrichment method at 1 ng/mL in serum or 833 pg/mL in serum, respectively (Supplementary Figures 4c, d, Supporting Information). Because both sample preparation methods yielded similar degrees of sample complexity and limits of detection, the more convenient procedure of affinity enrichment was used in all subsequent experiments. Qualitative Analysis by Selected Ion Monitoring. The “accurate mass tag” (AMT) experiment is an alternative to tandem mass spectrometry in which the presence of a signature peptide is confirmed by observation of a peak with the correct retention time and mass, provided the mass can be measured with sufficient accuracy.19,20 A sample of 11.25 fmol IFNA4 was added to 300 µL of serum (250 pg; 833 pg/mL) and prepared by the affinity enrichment procedure. LC-MS on a linear ion trap-Fourier transform ICR mass spectrometer was performed in selected ion monitoring mode, whereby spectra covering a 20 m/z window centered on m/z 698 were acquired throughout. An extracted ion chromatogram for a 20 ppm window centered on the IFNA4 signature peptide’s [M + 3H]3+ ion at 698.638 produced a single, intense peak at the correct retention time (Figure 3a). The corresponding mass spectrum revealed ions spaced 0.33 Da apart agreeing with the calculated monoisotopic, 13C, and 13C2 masses of the IFNA4 signature peptide to within 4.3 ppm, 5.3 ppm, and 4.9 ppm respectively using the default external calibration (Figure 3b). Although multiple additional components are clearly present at the same retention time and having the same nominal mass as the IFNA4 peptide, the resolution of the FTICR (set to 100 000 M/∆M at m/z 400) is sufficient to resolve them. Quantitative Analysis. In order to evaluate the potential for different mass spectrometry platforms for quantitation, calibration curves for IFNA2 were established using the classical approach of isotope dilution mass spectrometry. Samples were prepared by adding to 300 µL serum aliquots unlabeled IFNA2 in amounts ranging from 1 ng/mL to 1 µg/mL along with constant amounts of both 15N labeled IFNA2 and a synthetic peptide corresponding to the tryptic peptide DHDFPQEEFGNQF*QK where F* is 13C915N labeled phenylalanine. Each sample was processed using the affinity enrichment protocol, digested, and analyzed by targeted LC-MS. Two methods were compared: SIM using a linear ion traporbitrap mass spectrometer, and selected reaction monitoring (SRM) on a Q-Trap hybrid mass spectrometer operated as a triple

Detecting and Quantifying Interferon Alpha Subtypes in Serum

Figure 3. “Accurate mass tag” detection of the IFNA4 signature peptide by FTICR selected ion monitoring. Two-hundred fifty picograms of IFNA4 in 300 µL serum (833 pg/mL) was affinity enriched, digested, and analyzed by LC-MS. (a) Reconstructed ion chromatogram for a 20 ppm window centered on m/z 698.638. (b) Spectrum corresponding to the chromatographic peak at 36.4 min. Peaks assigned to the triply charged signature peptide and its isotopes are indicated.

quadrupole. Results of both experiments are illustrated in Figure 4 and described below. SIM was performed on a linear ion trap-orbitrap mass spectrometer with 15 m/z scans encompassing the [M + 3H]3+ ions of the signature peptide and its stable isotope analogs. The orbitrap mass analyzer was operated at a resolution of 60 000 M/∆M at m/z 400 with external calibration. Microcapillary RP-LC-ESI-MS was carried out for each sample, with triplicate injections in order of increasing concentration. Peak heights of the monoisotopic forms of each peptide species were extracted and abundance ratios calculated with respect to the SIL internal standards: 25 fmol of the synthetic peptide, and the peptide derived from the 15 ng of 15N labeled protein added to the sample at the outset. These ratios were fit to a straight line by linear regression. Best results were obtained by reference to the peptide derived from 15N labeled protein, plotted against concentration in Figure 4a but similar results were obtained using the synthetic peptide (Pearson R2 values of 0.98 and 0.96 respectively). Interestingly, the number of counts recorded for the SIL synthetic peptide was higher than the unlabeled protein by a factor of 9.4 for equivalent loads, that is, the 25 fmol of synthetic peptide back-calculates to approximately 10 ng/mL of protein, and that concentration of unlabeled protein yielded an average ratio of 0.106 with respect to the peptide standard. This indicates that substantial losses were accrued during the sample preparation, as well as the known incomplete digestion of IFNA to the signature peptide. As expected from the qualitative experiments using the AMT approach, the signature peptides were detected at all concen-

research articles

trations down to 1 ng/mL. Deviations of the observed ratios from those calculated using the best-fit line (Figure 4b) were below 10% for concentrations of 50 ng/mL and above. Lower concentrations deviated from calculated values by 40% or more. A lower limit of quantitation in this experiment therefore lies between 10 and 50 ng/mL. SRM transitions were chosen and optimized on the Q-Trap. These all involved fragmentation of the [M + 3H]3+ ions of the IFNA2 signature peptide derived from unlabeled protein, 15N labeled protein, and the SIL synthetic peptide. The transitions of precursors to b5 and y6 were monitored. Each sample was analyzed in triplicate in order of increasing concentration. In contrast to the microcapillary LC-MS used in all of the previous experiments, a 2 mm inner diameter reverse phase column was employed at a flow rate of 200 µL/minute, these conditions being conducive to higher throughput and better quantitative precision. Peak heights and areas for the transitions to b5 were extracted for purposes of quantitation. Abundance ratios were calculated with respect to the stable isotope labeled internal standards (nominally 675 fmol from 15N labeled protein and 25 fmol synthetic peptide). Ratios were fit to a straight line by linear regression and plotted against concentration. Best results were obtained using peak heights and calculating ratios with respect to the 15N labeled protein added at the beginning of the sample preparation sequence, but comparable results were obtained using peak areas and the SIL synthetic peptide. As is apparent from Figure 4c, linearity is maintained across 3 orders of magnitude in concentration, with a Pearson R2 value greater than 0.99. An upper limit of linear response was not reached with the range of concentrations tested. Deviations of the observed ratios from those calculated using the best-fit line were below 10% for all concentrations down to 5 ng/mL. At the low end of the concentration range, the signature peptide was clearly detected at 1 ng/mL, but with an abundance ratio deviating by almost a factor of 2 from the calculated value (Figure 4d), indicating that the lower limit of quantitation is between 1 and 5 ng/mL for the assay. As with the AMT experiment on the orbitrap, the synthetic peptide yielded more counts compared to the equivalent load of unlabeled protein, in this case by a factor of 12.

Discussion The detection and quantification of individual proteins against a background of many other species is a common task in proteomics, with blood plasma or serum presenting the ultimate analytical challenge. Limitations of antibody-based techniques such as protein arrays or ELISA have spurred the development of mass spectrometric alternatives, which themselves have strengths and limitations. A multilaboratory study in which plasma and serum samples were subjected to analysis via multiple fractionation strategies and mass spectrometry platforms yielded slightly fewer than 900 proteins that met rigorous statistical criteria, and these were naturally biased toward the more abundant plasma proteins.21,22 The inherent sensitivity of mass spectrometers is not the limiting factor, as even 100 µL of a 10 pg/mL, 20 kDa cytokine works out to 50 attomoles, readily detectable by today’s more advanced mass spectrometers in the absence of matrix effects. Overcoming the dynamic range of complex mixtures is the greater challenge and critical evaluations of the sample preparation and instrumental options available are badly needed. Techniques for the targeted detection and quantitation of proteins in complex biological matrices were evaluated using Journal of Proteome Research • Vol. 8, No. 6, 2009 3137

research articles

Izrael-Tomasevic et al.

Figure 4. Calibration curves for quantifying IFNA2 by stable isotope dilution mass spectrometry. Samples were prepared by adding unlabeled IFNA2 to serum in amounts ranging from 1 ng/mL to 1 µg/mL, along with a constant amount of 15N labeled IFNA2 as an internal standard. Upper panels: Selected ion monitoring of the IFNA2 signature peptide HDFGFPQEEFGNQFQK by LTQ-orbitrap mass spectrometry. The abundance ratios of [M + 3H]3+ ions from three technical replicates of unlabeled and 15N labeled peptides were averaged, plotted and fitted to a straight line (a). The deviation of each measured value from the calculated best-fit line is charted (b). Lower panels: Selected reaction monitoring the of IFNA2 signature peptide on a triple quadrupole mass spectrometer. The transition from the [M + 3H]3+ at m/z 652.3 to 604.3 (b5) was monitored, as was the corresponding 15N peptide transition from m/z 659.9 to 611.2. Peak height ratios for technical replicates were averaged (n ) 3) and fitted to a straight line as above (c) as well as deviations of the measured values from those calculated (d).

as a test case the subtypes of human interferon alpha in serum. Affinity enrichment by immunoprecipitation was simple and effective in reducing sample complexity for mass spectrometric analysis, but requires the availability of a high affinity antibody against the class of proteins to be analyzed. Depletion of six abundant serum proteins followed by digestion and strong cation exchange chromatography yielded a similar degree of enrichment and is completely general although more involved and less amenable to high throughput analyses. Importantly, it was shown that the MARS column used to deplete high abundance serum proteins did not bind IFNA. It is known that serum albumin, for example, is a carrier protein for other plasma proteins and the depletion step thus has the potential to remove proteins of interest as well, confounding the analysis in the absence of proper controls.23 In principle, proteins of any concentration should be detectible given a large enough volume of plasma and the time and patience to apply multiple fractionation steps until the protein of interest is sufficiently enriched. Mindful of the sample volumes and throughput necessary for analyzing clinical samples, we have focused on relatively simple workflows. Mass spectrometry is unusual among instrumental techniques in that multiple physical principles can be applied to effect mass separation (e.g., time-of-flight, ion cyclotron resonance in a magnetic field, stability of motion in radio frequency electromagnetic fields), each of which has advantages and limitations for particular analyses. Likewise, mass spectrometers are extremely versatile instruments, capable of multiple 3138

Journal of Proteome Research • Vol. 8, No. 6, 2009

scan functions (e.g., molecular mass measurement, product ion, neutral loss, and precursor ion scans) and manipulations involving ion-molecule and ion-ion reactions. The choice of instrument for a particular analysis is thus often faced, along with the practical consideration that cost may constrain experiments to instruments on hand in a given laboratory. We found that multiple mass spectrometry platforms proved capable of detecting and discriminating the closely related proteins IFNA2 and IFNA4 at the 1 ng/mL level in serum. Linear ion trap mass spectrometry in MS3 mode provided the greatest sequence information for each peptide assayed, whereas high resolution instruments such as FTICR and orbitrap mass spectrometers are capable of confirming the presence of a signature peptide using the accurate mass tag approach without MS/MS. Duty cycle considerations mean that more peptides can be assayed by SRM or AMT experiments in a single LC-MS experiment than targeted MS/MS or MS3 on an ion trap. An advantage of the ion trap methods is that a more extensive set of product ions are measured compared to SRM (usually just a handful) or SIM (none at all). Less development time is therefore needed for each peptide assayed, and higher confidence in the identity of the peptide is obtained. This also means that precision in chromatographic retention times is less important than for SRM or AMT experiments. Although protein quantitation by stable isotope dilution has been successfully performed using ion traps,24 our attempts to quantify IFNA at low concentrations did not yield satisfactory results. The signal-to-noise ratios of precursor ions and even

Detecting and Quantifying Interferon Alpha Subtypes in Serum specific product ions were too low to be useful, and although signal-to-noise was much improved in the MS3 product ion scans, this came at a cost in absolute ion counts. Samples containing ng/mL concentrations of IFNA, for example, typically yielded tens to hundreds of counts in the MS3 spectra whereas the same sample yielded tens of thousands of counts in the FTICR spectrum (not perfectly comparable due to the different detectors). Ion statistics for quantitation in the ion trap were therefore unfavorable, causing us to shift our focus to SRM and AMT approaches. SRM on a triple quadrupole mass spectrometer remains the method of choice for quantitation, having a lower limit of quantitation below 5 ng/mL for the interferons, and a broad linear dynamic range. Quantitation using the orbitrap mass spectrometer in SIM mode was also feasible, albeit with less accuracy at the lower concentrations. The latter instrument and its microcapillary chromatography system are optimized in our laboratory for low level qualitative experiments, and could likely be improved upon for quantitative work. With either approach, it should be possible to expand the set of peptides quantified to include a signature peptide of every IFNA subtype, allowing total IFNA levelssvia the common peptidesand all individual subtypes to be measured in a single LC-MS experiment. The nanogram detection and quantitation limits demonstrated are still higher than those required for the analysis of endogenous IFNA but extend the practical range of serum proteins that can be quantified to include many important plasma proteins than have not been readily accessible to date.1 Several parameters in the sample preparation schemes (e.g., immunopreciptitation, digestion, and peptide fractionation) are open to optimization; no fundamental barrier to detection of lower abundance proteins has been encountered. Even better results should be attainable in samples less complex than serum. Absolute protein quantitation has been approached in at least two distinct ways by proteomics researchers. In the classical analytical chemistry experiment, which we have emulated in this study, a calibration curve is constructed using a dilution series of unlabeled standard with the constant addition of a stable isotope labeled internal standard. Samples of unknown concentration are spiked with the same amount of labeled standard, and the ratio of unknown to standard measured by the mass spectrometer is used to place the unknown onto the calibration line. The exact amount of labeled internal standard is not critical provided it is constant throughout and lies in the range of concentrations to be analyzed. In contrast, some proteomics researchers have dispensed with the calibration curve and performed quantitation by the addition of a precisely quantified synthetic SIL peptide as internal standard. The ratio of unknown peptide to the internal standard is multiplied by the concentration of the standard to yield a concentration of the unknown.25,26 A Proposal has been made to use this strategy in a massively multiplexed SRM experiment to simultaneously quantify thousands of serum proteins.27 Our results shed some light on the advantages and limitations of each approach. The greatest strength of the classical approach with calibration curve is that the resulting assay is well characterized; done properly, the limit of detection, upper and lower limits of quantitation, and departures from linearity across the dynamic range are documented. Our data showed that linear calibration could be achieved with very different mass spectrometry platforms and either a synthetic SIL peptide or labeled recombinant protein. The choice of peptide or protein as internal

research articles

standard is not critical in this setting. Had we chosen to proceed without establishing a calibration curve with unmodified recombinant protein and quantified based on a single point, the internal standard would have been far more important. The SIL peptide, introduced during or after protein digestion, cannot account for losses accrued throughout the sample preparation. These were substantial in the present study, and would lead to systematic underestimation of protein concentrations if the “one-point” workflow were followed. Even with high recovery in each step, multistage sample preparation protocols will rapidly incur significant losses, e.g. three procedures with 80% recovery each will deplete a sample of almost half its material. Introduction of labeled, intact protein prior to sample preparation is superior in that a loss of analyte through the procedure is experienced equally by endogenous protein and standard. Likewise, differences in the yield of individual proteolytic peptides are normalized, removing a primary source of variability in results for multiple peptides derived from the same protein.28,29 The IFNA2 and IFNA4 signature peptides were chosen by necessity, and are less than ideal in that partial digestion products were present and although somewhat variable, could not be eliminated. This has an inevitable impact on limit of detection, but does not affect accuracy if the full length protein is used as internal standard. The generation of calibration curves and development of well-characterized assays is not, however, easily implemented for large scale serum proteomic experiments. Apart from the time involved, it is impractical in the short term to acquire pure, well quantified standardsslabeled or unlabeledsfor every serum protein of interest. It may therefore be justifiable to sacrifice some analytical rigor in order to do rapid screens that will be followed by a small number of more accurate assays. The classical approach of isotope dilution with full calibration curve remains the benchmark, ideally with a full length isotopically labeled protein as the internal standard, but singlepoint reference to SIL synthetic peptides may be a necessary compromise to do relative quantitation of large numbers of proteins across large numbers of samplessat the cost of no longer being able to claim absolute quantitation.

Conclusions Techniques for the targeted detection and quantitation of proteins in complex biological matrices were evaluated using the subtypes of human interferon alpha in serum as a model system. Affinity enrichment by immunoprecipitation was simple and effective in reducing sample complexity for mass spectrometric analysis, but requires the use of a high affinity antibody. Depletion of six abundant serum proteins followed by digestion and strong cation exchange chromatography yielded a similar degree of enrichment and is completely general although more involved and less amenable to high throughput analyses. Mass spectrometry platforms including a low resolution ion trap, high resolution FT-ICR and orbitrap mass spectrometers, and a triple quadrupole instrument all proved capable of detecting and discriminating the closely related proteins IFNA2 and IFNA4 at the 1 ng/mL level in serum. Linear ion trap mass spectrometry in MS3 mode provided the most detailed sequence information, whereas high resolution instruments were capable of confirming the presence of a signature peptide using the accurate mass tag approach without MS/MS. SRM on a triple quadrupole mass spectrometer remains the method of choice for quantitation, having a lower limit of quantitation below 5 ng/mL for the interferons, and a broad linear dynamic Journal of Proteome Research • Vol. 8, No. 6, 2009 3139

research articles range. Quantitation was also achievable using the orbitrap mass spectrometer in SIM mode, albeit with less accuracy at the lower concentrations. Introduction of isotopically labeled protein at the outset of sample preparation was preferable to the use of stable isotope labeled synthetic peptides for quantitation by virtue of accounting for protein losses experienced through sample preparation, and compensating for any variability in the efficiency of digestion.

Acknowledgment. We are grateful to Donald Kirkpatrick, Richard Vandlen and Sahana Mollah for helpful advice on the project. Thanks also to the medicinal chemistry department for making synthetic peptides & Brigitte Maurer for the stable isotope labeled proteins.

Izrael-Tomasevic et al.

(16)

(17)

(18)

(19)

Supporting Information Available: Supplementary figures. This material is available free of charge via the Internet at http://pubs.acs.org. References (1) Anderson, N. L.; Anderson, N. G. The human plasma proteome: history, character, and diagnostic prospects. Mol. Cell. Proteomics 2002, 1 (11), 845–67. (2) Bienvenu, J. A.; Monneret, G.; Gutowski, M. C.; Fabien, N. Cytokine assays in human sera and tissues. Toxicology 1998, 129 (1), 55– 61. (3) Diaz, M. O.; Bohlander, S.; Allen, G. Nomenclature of the human interferon genes. J. Interferon Cytokine Res. 1996, 16 (2), 179–80. (4) Hooks, J. J.; Moutsopoulos, H. M.; Geis, S. A.; Stahl, N. I.; Decker, J. L.; Notkins, A. L. Immune interferon in the circulation of patients with autoimmune disease. N. Engl. J. Med. 1979, 301 (1), 5–8. (5) Stewart, T. A. Neutralizing interferon alpha as a therapeutic approach to autoimmune diseases. Cytokine Growth Factor Rev. 2003, 14 (2), 139–54. (6) Fitzgerald, S. P.; McConnell, R. I.; Huxley, A. Simultaneous analysis of circulating human cytokines using a high-sensitivity cytokine biochip array. J. Proteome Res. 2008, 7 (1), 450–5. (7) Wadhwa, M.; Thorpe, R. Cytokine immunoassays: recommendations for standardisation, calibration and validation. J. Immunol. Methods 1998, 219 (1-2), 1–5. (8) Lescuyer, P.; Hochstrasser, D.; Rabilloud, T. How shall we use the proteomics toolbox for biomarker discovery. J. Proteome Res. 2007, 6 (9), 3371–6. (9) Good, D. M.; Thongboonkerd, V.; Novak, J.; Bascands, J. L.; Schanstra, J. P.; Coon, J. J.; Dominiczak, A.; Mischak, H. Body fluid proteomics for biomarker discovery: lessons from the past hold the key to success in the future. J. Proteome Res. 2007, 6 (12), 4549– 55. (10) Hu, S.; Loo, J. A.; Wong, D. T. Human body fluid proteome analysis. Proteomics 2006, 6 (23), 6326–53. (11) Faca, V.; Pitteri, S. J.; Newcomb, L.; Glukhova, V.; Phanstiel, D.; Krasnoselsky, A.; Zhang, Q.; Struthers, J.; Wang, H.; Eng, J.; Fitzgibbon, M.; McIntosh, M.; Hanash, S. Contribution of protein fractionation to depth of analysis of the serum and plasma proteomes. J. Proteome Res. 2007, 6 (9), 3558–65. (12) Martosella, J.; Zolotarjova, N.; Liu, H.; Nicol, G.; Boyes, B. E. Reversed-phase high-performance liquid chromatographic prefractionation of immunodepleted human serum proteins to enhance mass spectrometry identification of lower-abundant proteins. J. Proteome Res. 2005, 4 (5), 1522–37. (13) Stalder, D.; Haeberli, A.; Heller, M. Evaluation of reproducibility of protein identification results after multidimensional human serum protein separation. Proteomics 2008, 8 (3), 414–424. (14) Tang, H. Y.; Ali-Khan, N.; Echan, L. A.; Levenkova, N.; Rux, J. J.; Speicher, D. W. A novel four-dimensional strategy combining protein and peptide separation methods enables detection of lowabundance proteins in human plasma and serum proteomes. Proteomics 2005, 5 (13), 3329–42. (15) Solassol, J.; Marin, P.; Demettre, E.; Rouanet, P.; Bockaert, J.; Maudelonde, T.; Mange, A. Proteomic detection of prostate-

3140

Journal of Proteome Research • Vol. 8, No. 6, 2009

(20)

(21)

(22)

(23)

(24)

(25)

(26)

(27)

(28)

(29)

specific antigen using a serum fractionation procedure: potential implication for new low-abundance cancer biomarkers detection. Anal. Biochem. 2005, 338 (1), 26–31. Barnidge, D. R.; Goodmanson, M. K.; Klee, G. G.; Muddiman, D. C. Absolute quantification of the model biomarker prostate-specific antigen in serum by LC-Ms/MS using protein cleavage and isotope dilution mass spectrometry. J. Proteome Res. 2004, 3 (3), 644–52. Wu, S. L.; Amato, H.; Biringer, R.; Choudhary, G.; Shieh, P.; Hancock, W. S. Targeted proteomics of low-level proteins in human plasma by LC/MSn: using human growth hormone as a model system. J. Proteome Res. 2002, 1 (5), 459–65. Yague, J.; Paradela, A.; Ramos, M.; Ogueta, S.; Marina, A.; Barahona, F.; Lopez de Castro, J. A.; Vazquez, J. Peptide rearrangement during quadrupole ion trap fragmentation: added complexity to MS/MS spectra. Anal. Chem. 2003, 75 (6), 1524–35. Lipton, M. S.; Pasa-Tolic, L.; Anderson, G. A.; Anderson, D. J.; Auberry, D. L.; Battista, J. R.; Daly, M. J.; Fredrickson, J.; Hixson, K. K.; Kostandarithes, H.; Masselon, C.; Markillie, L. M.; Moore, R. J.; Romine, M. F.; Shen, Y.; Stritmatter, E.; Tolic, N.; Udseth, H. R.; Venkateswaran, A.; Wong, K. K.; Zhao, R.; Smith, R. D. Global analysis of the Deinococcus radiodurans proteome by using accurate mass tags. Proc. Natl. Acad. Sci. U.S.A. 2002, 99 (17), 11049–54. Smith, R. D.; Anderson, G. A.; Lipton, M. S.; Pasa-Tolic, L.; Shen, Y.; Conrads, T. P.; Veenstra, T. D.; Udseth, H. R. An accurate mass tag strategy for quantitative and high-throughput proteome measurements. Proteomics 2002, 2 (5), 513–23. 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, 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 (13), 3226–45. States, D. J.; Omenn, G. S.; Blackwell, T. W.; Fermin, D.; Eng, J.; Speicher, D. W.; Hanash, S. M. Challenges in deriving highconfidence protein identifications from data gathered by a HUPO plasma proteome collaborative study. Nat. Biotechnol. 2006, 24 (3), 333–8. Zhou, M.; Lucas, D. A.; Chan, K. C.; Issaq, H. J.; Petricoin, E. F., 3rd.; Liotta, L. A.; Veenstra, T. D.; Conrads, T. P. An investigation into the human serum ”interactome. Electrophoresis 2004, 25 (9), 1289–98. Arnott, D.; Kishiyama, A.; Luis, E. A.; Ludlum, S. G.; Marsters, J. C., Jr.; Stults, J. T. Selective detection of membrane proteins without antibodies: a mass spectrometric version of the Western blot. Mol. Cell. Proteomics 2002, 1 (2), 148–56. Blankenship, J. W.; Varfolomeev, E.; Goncharov, T.; Fedorova, A. V.; Kirkpatrick, D. S.; Izrael-Tomasevic, A.; Phu, L.; Arnott, D.; Aghajan, M.; Zobel, K.; Bazan, J. F.; Fairbrother, W. J.; Deshayes, K.; Vucic, D. Ubiquitin binding modulates IAP antagonist-stimulated proteasomal degradation of c-IAP1 and c-IAP2(1). Biochem. J. 2009, 417 (1), 149–60. 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 (12), 6940–5. Stahl-Zeng, J.; Lange, V.; Ossola, R.; Eckhardt, K.; Krek, W.; Aebersold, R.; Domon, B. High sensitivity detection of plasma proteins by multiple reaction monitoring of N-glycosites. Mol. Cell. Proteomics 2007, 6 (10), 1809–17. Brun, V.; Dupuis, A.; Adrait, A.; Marcellin, M.; Thomas, D.; Court, M.; Vandenesch, F.; Garin, J. Isotope-labeled protein standards: toward absolute quantitative proteomics. Mol. Cell. Proteomics 2007, 6 (12), 2139–49. Hanke, S.; Besir, H.; Oesterhelt, D.; Mann, M. Absolute SILAC for accurate quantitation of proteins in complex mixtures down to the attomole level. J. Proteome Res. 2008, 7 (3), 1118–30.

PR900076Q