Scrambled Internal Standard Method for High-Throughput Protein

Mar 20, 2017 - Scrambled Internal Standard Method for High-Throughput Protein Quantification by Matrix-Assisted Laser Desorption Ionization Tandem Mas...
0 downloads 11 Views 2MB Size
Download PDF
Article pubs.acs.org/jpr

Scrambled Internal Standard Method for High-Throughput Protein Quantification by Matrix-Assisted Laser Desorption Ionization Tandem Mass Spectrometry Toshihiro Yoneyama,† Sumio Ohtsuki,*,‡,§ Masanori Tachikawa,† Yasuo Uchida,† and Tetsuya Terasaki† †

Division of Membrane Transport and Drug Targeting, Graduate School of Pharmaceutical Sciences, Tohoku University, Aoba, Aramaki, Aoba-ku, Sendai 980-8578, Japan ‡ Department of Pharmaceutical Microbiology, Faculty of Life Sciences, Kumamoto University, 5-1 Oe-honmachi, Kumamoto 862-0973, Japan § Japan Agency for Medical Research and Development (AMED) CREST, Tokyo 100-0004, Japan S Supporting Information *

ABSTRACT: Matrix-assisted laser desorption ionization (MALDI) could be advantageous for high-throughput MS acquisition but suffers from low signal reproducibility. The purpose of this study was to establish a reliable MALDItandem mass spectrometry (MS/MS)-based high-throughput quantification of tryptic peptides using our newly developed scrambled internal standard (sIS) method. The standard curves obtained with sIS peptides showed good linearity over a wide concentration range (5−1000 fmol/μL) compared to that with the IS-free method, and the coefficient of variation of data points at each concentration (5−1000 fmol/μL) was significantly reduced. Furthermore, the ion suppression effect of digested serum could be normalized with the sIS peptides. Differences of quantitative values obtained by MALDI−MS/MS and liquid chromatography-MS/MS with selected reaction monitoring were within 20% in the presence of 0.1−5 μL of immunoprecipitated model plasma. Furthermore, the effect of amino acid composition on peptide sensitivity was examined, and we found that sensitivity was significantly decreased if an aromatic amino acid was replaced with a nonaromatic amino acid. Thus, high sensitivity required the use of sIS peptides containing an aromatic amino acid. Finally, the sIS method enabled high-throughput quantification of tryptic peptides with high accuracy and a wide dynamic range. KEYWORDS: MALDI, MS/MS, high-throughput, scrambled internal standard, absolute quantification, plasma

■

INTRODUCTION Matrix-assisted laser desorption ionization (MALDI), which is mainly used for imaging mass spectrometry (IMS) and protein identification, is compatible with high-throughput analysis because mass spectrometry (MS) and tandom mass spectrometry (MS/MS) spectra can be recorded without the need for a liquid chromatography (LC) separation step.1 However, MALDI is not regarded as a suitable method for quantification because of poor spot-to-spot, region-to-region, and sample-tosample reproducibility.2 In an attempt to overcome this limitation, MALDI−MS (typically TOF/MS)-based quantification using an isotope-labeled peptide corresponding to the target peptide has been developed,3−5 but one of the problems in this approach is that the complexity of biological samples sometimes affects the signal intensity of target and/or isotopelabeled peptides because the quantification is only based on the precursor ion.6 MS/MS analysis is a better approach for reliable and accurate quantification because of the low noise level. Ng et al. developed a new protocol using MALDI-TOF/TOF MS, © XXXX American Chemical Society

named parallel fragmentation monitoring (PFM), which is analogous to the selected reaction monitoring (SRM) method.7 The main difference between PFM and SRM is that the target molecule and internal standard (IS) are isolated and fragmented simultaneously in PFM, whereas in SRM, the transitions are analyzed separately. Ng et al. performed PFM quantification of citrulline as a target with 13C-citruline, which is only 1 mass unit heavier than the target, as the IS. However, when a target and IS with different m/z values are used, a larger range of precursor ion filters is needed for quantification, which results in an increased noise level. Therefore, to apply this method for peptide quantification, precursor ions with the same m/z value for target and IS peptide should be adopted. They can then be distinguished based on product ions. One approach based on this idea uses tagging with isotopically labeled reagents, as in the isobaric tag for relative and absolute quantitation (iTRAQ)8−10 and tandem mass tags (TMT) Received: October 30, 2016 Published: March 20, 2017 A

DOI: 10.1021/acs.jproteome.6b00941 J. Proteome Res. XXXX, XXX, XXX−XXX

Article

Journal of Proteome Research Table 1. List of Peptides Used for Normalization of Ion Suppressiona m/z model peptide sequence

scrambled IS peptide sequence

Q1

LTIGEGQQHHLGGAK HPDEAAFFDTASTGK ESSSHHPGIAEFPSR YEASILTHDSSIR AHYGGFTVQNEANK LTYAYFAGGDAGDAFDGFDFGDDPSDK GLIDEVNQDFTNR VQHIQLLQK DLQSLEDILHQVENK EDGGGWWYNR HQLYIDETVNSNIPTNLR VELEDWAGNEAYAEYHFR NNSPYEIENGVVWVSFR VPPEWK IHLISTQSAIPYALR

ATIGEGQQHHLGGLK GPDEAAFFDTASTHK SISSHSPGHAEFPER IEASILTHDSSYR NHYGGFTVQNEAAK STYAYFAGGDAGDAFDGFDFGDDPDLK NLIDEVNQDFTGR QQHIQLLVK ELQSLEDILHQVNDK NDGGGWWYER LQLYIDETVNSNIPTNHR FELEDWAGNEAYAEYHVR FNSPYEIENGVVWVSNR WPPEVK AHLISTQSAIPYLIR

1587.8 1635.7 1637.8 1491.7 1577.7 2876.2 1520.7 1148.7 1822.9 1239.5 2127.1 2199.0 2010.0 797.4 1683.0

Q3_model (ion type) 1214.6 1401.6 1508.7 1199.6 1369.6 2273.9 1122.5 832.5 1707.9 995.4 1357.7 1613.7 947.4 601.3 1064.6

(b11) (y13) (y14) (y11) (y12) (b22) (y9) (b7) (y14) (y8) (y12) (y13) (b8) (y4) (b10)

Q3_IS (ion type) 1172.6 1366.6 1237.6 1249.6 1326.6 2362.9 1065.5 861.5 1634.8 1010.5 1381.7 1565.7 1748.9 514.3 1022.6

(b11) (y12) (b12) (y11) (y12) (b23) (y9) (b7) (b14) (y8) (y12) (y13) (y15) (y4) (b10)

hydropathy −0.692 −0.267 −0.257 −0.245 −0.206 −0.173 −0.155 −0.148 −0.139 −0.084 −0.012 0.016 0.090 0.185 0.320

a

List of peptides with intensity of more than 100 cps for 250 fmol/μL model peptides or 500 fmol/μL sIS peptides. Theoretical m/z values of singly charged ions of intact peptides (Q1) were assumed as precursor ions (Q1). High signal intensity product ions produced from precursor ions were selected as Q3_model and Q3_sIS. The ion type of Q3 ions is shown in parentheses. The hydropathy index of each peptide was calculated according to the SOSUI prediction system.

methods.11,12 Because the tags are isobaric, target peptide and IS peptide labeled with an isotopic variant of the tag can be isolated simultaneously in the precursor ion filter. Subsequent fragmentation of precursor ions generates reporter ions with different m/z values from the target and IS peptides. Then, quantitative values are calculated from the target-to-IS peptide intensity ratios of the reporter ions in the target samples using a calibration curve. However, these reagents are costly, and various problems, including sample loss, limited efficacy of labeling, and long sample preparation time, are associated with analytical schemes utilizing on-site chemical labeling.6,13 Furthermore, in this method, systematic error due to interference of overlapping precursor ions cannot be corrected because the quantification is only based on one reporter ion each for the target and IS peptides but not on multiple fragment ions.14 We have developed a new IS method, which we call the scrambled internal standard (sIS) method, to overcome these limitations. Thus, the purpose of this study was to develop a MALDI−MS/MS-based protein quantification method using sIS peptides.

■

include appropriate peptide length for detection by LC−MS/ MS, no post-translational modifications (PTMs), no single nucleotide polymorphisms (SNPs), no continuous sequence of arginine or lysine in digestion region, and no methionine or cysteine residues. The uniqueness of the amino acid sequences of targeted peptides was confirmed by BLAST search. For the peptide selection in the present study, the criteria regarding appropriate peptide length, no PTMs, and no SNPs were excluded to obtain a broad range of various model peptides for MALDI−MS/MS from fibrinogen protein. The hydropathy index of each peptide was calculated by the SOSUI prediction system (http://harrier.nagahama-i-bio.ac.jp/sosui/). For scrambled IS (sIS) peptides to be designed, the N-terminal amino acid and the second-from-C-terminal amino acid of the model peptides were inverted. When the difference of molecular weights between those 2 amino acids was within 2 Da, the second- and third-from-N-terminal amino acids were also inverted. SISSHSPGHAEFPER, the sIS peptide corresponding to ESSSHHPGIAEFPSR, is the only peptide that does not follow the method described above for designing sIS peptides. The model peptides and corresponding sIS peptides listed in Table 1 and Table S-1 were synthesized in the form of recombinant model protein and sIS protein, respectively. As shown in Figure S-1, the model protein was designed to contain all model peptides tandemly linked together, and the sIS protein contained all sIS peptides tandemly linked together. These recombinant proteins have a HAT-Tag (DHLIHNVHKEEHAHAHNK) inserted at the C-terminal to enable convenient purification. In addition, to allow determination of the absolute concentrations of the synthesized proteins, two monitoring peptide sequences for absolute quantification by LC−MS/MS, QIGDPTVPSGVK and NVAPAGPTLK (Table S-2), were inserted at the N- and C-terminals, respectively. The model and sIS proteins have the same molecular weight (54.2 kDa). The genes encoding these proteins were prepared by a gene synthesis service (FASMAC, Yokohama, Japan). They were transformed into Competent high DH5α (TOYOBO,

EXPERIMENTAL SECTION

Materials

The synthesized peptides were prepared at Thermoelectron Corporation (Sedantrabe, Germany). The concentrations of peptide stock solutions were determined by amino acid analysis in an HPLC-UV system with postcolumn ninhydrin derivatization (LaChrom Elite, Hitach, Tokyo, Japan). Standard human serum was purchased from Sigma Chemical Co. (St. Louis, MO). All other reagents were commercial products of analytical grade. Recombinant Protein Synthesis

Model peptides listed in Table 1 and Table S-1 were selected from human fibrinogen protein (Uniprot ID: P02671, P02675, and P02679) based on the in silico criteria15,16 reported previously with some modifications. The reported criteria B

DOI: 10.1021/acs.jproteome.6b00941 J. Proteome Res. XXXX, XXX, XXX−XXX

Article

Journal of Proteome Research Protein Quantification by LC−MS/MS in SRM Mode

Osaka, Japan) and amplified. As the original vectors for model protein and sIS protein were pGEM vector and pUC19 vector, respectively, the amplifed genes were inserted into pET17b vector for isopropyl-β-D-thiogalactopyranoside (IPTG, 1 mM) induction with BL 21-codon plus (DE3)-RIPL (Agilent Technology, Santa Clara, CA). The synthesized proteins were purified using HisPur Cobalt Spin Columns (Thermo Fisher Scientific, IL, USA) according to the manufacturer’s protocol with minor modifications of the buffers. Ten millimolar imidazole in 50 mM trizma base, 300 mM sodium chloride, pH 8.0 was used as the equilibration/wash buffer. Fifty millimolar, 150 mM, and 500 mM imidazole in 50 mM trizma base, 300 mM sodium chloride, pH 8.0 were used as elution buffers. Elution with each buffer was performed three times. Protein concentrations were measured by the Lowry method using the DC protein assay reagent (Bio-Rad, Hercules, CA).

LC−MS/MS analysis was performed with an electrospray ionization (ESI)-triple quadrupole mass spectrometer (API5000; AB SCIEX, Framingham, MA) coupled with an HPLC system (Agilent 1100 system; Agilent Technologies, Santa Clara, CA). LC−MS/MS analysis conditions were the same as described previously.17 C18 capillary columns (Waters XBridge BEH130 C18, 1.0 mm ID × 100 mm, 3.5 μm particles; Waters, Milford, MA) were used for HPLC. Linear gradients of 1−50% acetonitrile in 0.1% formic acid at a flow rate of 50 μL/ min for 50 min were applied to separate the peptides. The mass spectrometer was set up to run in multiplexed SRM mode for peptide detection with 10 ms per transition. For quantification of recombinant proteins, 500 fmol of stable isotope-labeled peptides (Table S-2) was added to 200-fold diluted tryptic digests of recombinant proteins. Each target peptide and corresponding isotope-labeled peptide were measured (Table S-2). For quantification of immunoprecipitated model plasma digests by LC−MS/MS, a half volume of sample digest was spiked with 500 fmol of stable isotope-labeled peptides. SRM transitions for GLIDEVNQDFTNR (GLI peptide) and the corresponding isotope-labeled peptide are shown in Table S2.17 Two-thirds (by volume) of the sample was subjected to LC−MS/MS analysis.

Enzyme Digestion of Recombinant Proteins and Standard Human Serum

Fifty micrograms of recombinant protein (0.922 nmol each for the model and sIS proteins) and 10 μL of 5-fold-diluted standard human serum corresponding to 2 μL of serum (100 μg) were solubilized in 7 M guanidine hydrochloride and 10 mM EDTA. The solubilized samples were S-carbamoylmethylated with dithiothreitol and iodoacetamide as described.16 The S-carbamoylmethylated samples were precipitated with a mixture of methanol and chloroform. The precipitates were dissolved in 6 M urea, diluted 5-fold with 100 mM Tris-HCl (pH 8.5), and treated with lysyl endopeptidase (Wako Pure Chemical Industries) at an enzyme/substrate ratio of 1:100 at 25 °C for 3 h. Subsequently, the samples were digested with sequence-grade modified trypsin (Promega, Madison, WI) at an enzyme/substrate ratio of 1:100 at 37 °C for 16 h.

Sample Preparation for MALDI−MS/MS Analysis (Product Ion Scan Mode)

For standard curves using GLI peptide and NLIDEVNQDFTGR (sGLI) as an sIS peptide, serial dilutions of GLIDEVNQDFTNR (5, 10, 25, 50, 100, 250, 500, and 1000 fmol/μL) were added to an equal volume of 500 fmol/μL sIS peptide. For immunoprecipitated samples, the digests were mixed with an equal volume of 500 fmol/μL sIS peptide. For peptides containing lysine at the C-terminal, which are reported to have a lower sensitivity in MALDI analysis compared with that in arginine-containing peptides,18 guanidination of lysine-containing tryptic peptides of recombinant proteins was performed according to a previous report with minor modifications.19 Briefly, a final concentration of 0.25 M of O-methylisourea (Sigma-Aldrich, St. Louis, MO) in 28−30% of ammonium hydroxide were mixed with peptides, and the reaction was allowed to proceed for 4 h at 30 °C and was then terminated by the addition of an equal volume of 1% aqueous trifluoroacetic acid. Mixtures of serial dilutions of guanidinated model peptides (final concentration: 50, 100, 250, 500, and 1000 fmol/μL), equal amounts of sIS peptides (final concentration: 500 fmol/μL), and serial dilutions of digested serum (final concentration: 0, 0.025, 0.05, 0.25, and 0.5 μg/μL) were analyzed by MALDI−MS/MS (total 25 samples). Before MALDI−MS/MS analysis (product ion scan mode), the peptides were purified with a Zip-Tip C18 (Millipore, Bedford, MA) according to the manufacturer’s protocol. Finally, the eluted solutions (∼1 μL) were directly spotted onto a polished steel target plate. Subsequently, 1 μL of 10 mg/mL of α-cyano4-hydroxycinnamic acid (CHCA) in solution A (0.003% TFA/ 13% ethanol/84% acetonitrile in Milli-Q water) was applied to the spotted samples. Subsequently, 1 μL of 10 mg/mL of CHCA in solution A was applied to the spotted samples. The samples were dried on a hot plate at 50 °C. For the accuracy of this method to be determined, the amounts of spiked peptide were determined from the standard curve using the peak intensity ratio of target peptide to sIS peptide, such that accuracy (%) = 100 × (determined amount

Immunoprecipitation of Serum Spiked with Fibrinogen (Model of Plasma) and Enzyme Digestion

Standard human serum spiked with fibrinogen protein was used as a model of plasma. Approximately 0.5, 1, 5, 10, and 50 pmol of fibrinogen protein were added to 0.05, 0.1, 0.5, 1, and 5 μL of standard human serum, respectively (0.05−5 μL model plasma). Antifibrinogen antibody (A0080) purchased from DAKO (Glostrup, Denmark) was used for immunoprecipitation along with Dynabeads Protein G and DynaMag-2 (Invitrogen, Carlsbad, CA) according to the manufacturer’s protocol with some modifications. The Dynabeads (1.5 mg) were washed and resuspended in 200 μL of phosphate-buffered saline (PBS) containing 0.02% Tween20. Six micrograms of antifibrinogen antibody was added, and the mixture was incubated at room temperature for 10 min with gentle agitation to form a Dynabead−antibody complex. After washing with PBS, 0.05−5 μL model plasma was added, and incubation was continued at room temperature for 1 h with gentle agitation. After three washes with 200 μL of PBS, the Dynabeads were resuspended in 20 μL of 6 M urea in 100 mM Tris-HCl (pH 8.5) and heated for 10 min at 70 °C. Extracted protein was measured by the Lowry method as described above. The extract was S-carbamoylmethylated with dithiothreitol and iodoacetamide. The S-carbamoylmethylated samples were diluted 5-fold with 100 mM Tris-HCl (pH8.5) and treated with lysyl endopeptidase and sequence-grade modified trypsin as described above. C

DOI: 10.1021/acs.jproteome.6b00941 J. Proteome Res. XXXX, XXX, XXX−XXX

Article

Journal of Proteome Research

Figure 1. Strategy for protein quantification with sIS peptide. Target peptide is selected from the target protein sequence and synthesized. The corresponding sIS peptide is also synthesized as a scrambled sequence of the target peptide. The target peptide and sIS peptide are isolated simultaneously at the first mass filter, as these peptides have the same m/z, leading to improvement of signal reproducibility. Because the target and sIS peptides are distinguished by restricted cleavage, the amount of peptide can be quantified.

Figure 2. MS and MS/MS spectra of the model peptide. Model protein digests (500 fmo/μL) spiked with 500 fmo/μL sIS protein digests were analyzed by MALDI−MS (A) and MS/MS (B). MS/MS analysis was performed for GLI and sGLI (B). Product ions of y9 (m/z 1122.5), y4 (m/z 537.3), and y8 (m/z 993.3) for GLI and y9 (m/z 1065.5) for sGLI are shown (B).

Sample Preparation for MALDI−MS Analysis (Precursor Ion Scan Mode)

of peptide/theoretical amount of peptide). The limit of detection (LOD) and lower limit of quantification (LLOQ)

For the MALDI−MS analysis (precursor scan mode) of the six peptides listed in Table 2, 1 μL of 1 μM each peptide spiked with 1 μM GLIDEVNQDFTNR peptide (nonlabeled form of GLI peptide) as an internal standard to correct signal reproducibility was spotted onto a polished steel target plate. Subsequently, 1 μL of 10 mg/mL of CHCA in solution A was

were calculated from the standard curve using N × noise level to the sIS peptide intensity. In this study, noise level was 10 cps, N = 3 for LOD, and N = 10 for LLOQ, such that LOD = (30 cps/sIS peptide intensity − intercept)/slope and LLOQ = (100 cps/sIS peptide intensity − intercept)/slope. D

DOI: 10.1021/acs.jproteome.6b00941 J. Proteome Res. XXXX, XXX, XXX−XXX

Article

Journal of Proteome Research

Figure 3. Standard curves for the IS-free and sIS methods. Serial dilution of GLI peptide (5, 10, 20, 50, 100, 200, 500, and 1000 fmol/μL) and 500 fmol/μL sGLI peptide as the sIS were analyzed by MALDI−MS/MS. Products ions were y9, y4, and y8 for the target peptide and y9 for the sIS peptide. Standard curves were prepared for the IS-free method (A−C) and the sIS method (D−F). Each data point shows the mean ± SD (n = 10 from different regions of a single spot). Closed and open circles represent the ranges of the standard curve within and outside the linear range, respectively.

■

applied onto the spotted samples. The samples were dried on a hot plate at 50 °C.

RESULTS

Strategy for Protein Quantification by MALDI−MS/MS

Peptide Measurement by MALDI−MS/MS or MS

The sIS method to overcome the low signal reproducibility in MALDI−MS/MS is illustrated schematically in Figure 1. Because sIS peptide has the same m/z value of the precursor ion as that of the corresponding target peptide, these peptides can be isolated at the same time, leading to improvement of shot-to-shot, region-to-region, and sample-to-sample reproducibility. Then, product ions of the target peptide and the corresponding sIS peptide are distinguished based on restricted cleavage using appropriate collision energy. Standard curves were prepared by calculating the ratios of the intensities from a dilution series of the target peptide to the intensity from a fixed amount of the corresponding sIS peptide.

All MS (precursor ion scan mode) and MS/MS (product ion scan mode) spectra were recorded on a MALDI-linear ion trap quadrupole (LTQ) mass spectrometer (Thermo Fisher Scientific) in positive ion mode with a 337 nm nitrogen laser (20 Hz firing rate). The MALDI laser energy was set to around 10 μJ for sample analysis. The parameters of MS and MS/MS analysis were crystal positioning system (CPS), 10 scans/step, and automated gain control (AGC) on. For MS/MS analysis, multiple scan events for each model and sIS peptide were set, and the product ions from each precursor ion were recorded. The MS/MS spectra (product ion scan) were acquired at 10 different regions in a single spot for each precursor ion (n = 10). For MS analysis (product ion scan), the MS spectrum was recorded at one region in four different spots (n = 4). Peak list extraction was done with Qual Browser 2.2 (Thermo Fisher Scientific). Peak intensities were determined as the highest peak in the range of theoretical m/z ± 0.5. Raw data files of MALDI−MS/MS analysis have been deposited in jPOST (http://jpostdb.org, jPOST ID: JPST000231).20

Recombinant Protein Synthesis

To demonstrate the usefulness of the sIS method, we selected 31 model peptides from the fibrinogen protein (15 peptides listed in Table 1 were finally used for the experiment, as the detection sensitivity was too low for the others; see the latter part of the Results), because model plasma could be prepared by adding fibrinogen protein or peptides to serum or digested serum. Moreover, we recently showed that hydroxylated αfibrinogen is a potential biomarker for pancreatic cancer,17 so if a quantification method for fibrinogen is developed, it could be applicable to hydroxylated α-fibrinogen. Each model peptide and the corresponding sIS peptide were synthesized in tandemly linked form with an affinity tag at the C-terminal for purification as model protein and sIS protein, respectively, as illustrated in Figure S-1. The concentration of each recombinant protein was determined by quantifying monitoring

Statistical Analysis

Student’s t test was used to determine the statistical significance of differences between two groups. A value of p < 0.05 was considered statistically significant. E

DOI: 10.1021/acs.jproteome.6b00941 J. Proteome Res. XXXX, XXX, XXX−XXX

Article

Journal of Proteome Research peptides located at the N-terminus and C-terminus (N- and Cmonitoring peptides, respectively) (Table S-3). Partially synthesized recombinant protein has N- but not C-monitoring peptide. Therefore, the ratio of C- to N-monitoring peptide can be quantitatively estimated as the amount ratio of full-length recombinant protein to all purified recombinant protein. The ratio was determined to be 0.910 and 0.930 for model and sIS proteins, respectively (Table S-3). This result suggested that more than 90% of purified recombinant protein was full-length protein on the assumption of uniform efficiency of digestion. Standard Curve for Quantification in the sIS Method

At first, an experiment using GLIDEVNQDFTNR (GLI peptide), which was used for absolute quantification of αfibrinogen in our previous report,17 was performed. GLI peptide and the corresponding scrambled IS peptide (sGLI peptide) have the same m/z value of the precursor ion (Figure 2A) but are detected as different signals of product ions in MALDI−MS/MS (Figure 2B). Three highly sensitive product ions for GLI peptide (y9, y4, and y8) and one product ion for the sGLI peptide (y9) were selected based on the signal intensities. The specificity of the product ion (y9) from the sGLI peptide was checked by adding serum digest (0.5 μg/μL) to the sample (Figure S-2). There was no difference of signal patterns between the two conditions, suggesting that digested serum did not interfere with the signals of product ions from GLI and sGLI peptides. Standard curves were prepared based on the GLI peptide intensities (IS-free method, Figure 3A−C) and the ratios of the GLI peptide intensity to the sGLI peptide intensity (sIS method, Figure 3D−F). As the linearity of the standard curves (5−1000 fmol/μL) using the IS-free method was poor (r2 values less than 0.7 for all selected product ions), the range of standard curves was changed to 5−100 fmol/μL. In this case, r2 values were more than 0.99 for y9 and less than 0.99 for y4 and y8. In contrast, the standard curves for the sIS method all showed good linearity (r2 values more than 0.99 for all selected product ions) over a wide range (5−1000 fmol/μL). Moreover, the % CV values (n = 10 from different regions of a single spot) of data points such as 1000, 500, 250, 100, and 50 fmol/μL were significantly reduced from 55.1, 41.6, 34.0, 26.1, and 46.1% in the IS-free method to 16.2, 18.6, 18.7, 13.4, and 22.8% in the sIS method, respectively.

Figure 4. Variation of GLI peptide intensities and GLI/sGLI peptide intensity ratios in the presence of five different concentrations of digested serum. Five different amounts (50−1000 fmol/μL) of GLI peptide spiked with 500 fmol/μL sGLI in the presence of five different concentrations of digested serum (0−0.5 μg/μL) were measured by MALDI−MS/MS. Each bar represents mean ± SEM (n = 10 from different regions of a single spot) of GLI peptide intensities (A) or GLI/sGLI peptide intensity ratios (B). The values above the bars are the % CV of intensities (A) and model/sIS intensity ratios (B) for the same GLI peptide amounts. y9 ions were used for analysis of GLI and sGLI peptides.

Normalization of the Ion Suppression Effect Using sIS Peptide

Tip because 9 of these 16 peptides are hydrophilic (hydropathy index < −0.35). The 14 model peptides (250 fmol/μL) spiked with 0, 0.025, 0.05, 0.25, and 0.5 μg/μL of digested serum were analyzed by MALDI−MS/MS. The % CV was reduced from more than 60% in the IS-free method to within 40% in the sIS method (Figure 5). Among the 10 peptides whose average intensities were over 300 cps, the % CV values of 9 were within 20% in the sIS method (Figure 5B). Similar results were obtained at 50, 100, 500, and 1000 fmol/μL (Figure S-3). These results suggest that the sIS method can normalize the ion suppression effect. For showing the accuracy and sensitivity of this method, the accuracy, LOD, and LLOQ of the 15 peptides listed in Table 1 are illustrated in Figures S-4 and S-5.

To demonstrate the usefulness of the sIS method for normalization of the ion suppression effect, we analyzed 50, 100, 250, 500, and 1000 fmol/μL GLI peptide spiked with 0, 0.025, 0.05, 0.25, and 0.5 μg/μL of digested serum by MALDI− MS/MS. As shown in Figure 4A, the intensities of GLI peptide were decreased in the IS-free method when the different amounts of digested serum were added, and the % CV ranged from 53.9 to 94.9%. However, when normalization using sIS peptide was conducted, the % CV at each concentration of peptides except for 50 fmol/μL was within 20%. For the sIS method to be generalized, similar experiments were performed using the 14 model peptides listed in Table 1 (except for the GLI peptide). The other 16 model peptides listed in Table S-1 could not be used because 250 fmol/μL model peptides or 500 fmol/μL sIS peptides in the absence of digested serum could not be detected (