Fragmentation and Site-Specific Quantification of Core Fucosylated

TECHNICAL NOTE pubs.acs.org/ac

Fragmentation and Site-Specific Quantification of Core Fucosylated Glycoprotein by Multiple Reaction Monitoring-Mass Spectrometry Yan Zhao,† Wei Jia,† Jifeng Wang,†,‡ Wantao Ying,*,† Yangjun Zhang,† and Xiaohong Qian*,† †

State Key Laboratory of Proteomics, Beijing Proteome Research Center, Beijing Institute of Radiation Medicine, No. 33 Life Science Park Road, Changping District, Beijing 102206, P. R. China ‡ College of Life Science and Bioengineering, Beijing University of Technology, Beijing 100124, P. R. China

bS Supporting Information ABSTRACT: Glycosylation modifications of proteins have been attracting increasing attention due to their roles in the physiological and pathological processes of the cell. Core fucosylation (CF), one special type of glycan structure in glycoproteins, has been linked with tumorigenesis. The study of protein glycosylation has been hindered by the technical challenges caused by the microheterogeneity of glycan modifications. In commonly used methods, sugar chains on the peptide were released using endoglycosidase, and the glycan and peptides were analyzed separately with mass spectrometry. Although mass spectrometric analysis can be performed easily in this way, an increase in false positives when assigning glycosites was inevitable. Our earlier research demonstrated a strategy combining Endo F3-catalyzed partial deglycosylation with MS3 (MS/MS/MS) scanning triggered by the neutral loss of a fucose to precisely identify CF proteins on a large scale. In this research, fragmentations of partially deglycosylated glycopeptides were studied using a triple quadrupole mass spectrometer, and a quantification method that coupled our published identification strategy with multiple reaction monitoring-mass spectrometry (MRM-MS) analysis was developed to obtain site-specific quantification information of core fucosylated peptides. To illustrate the feasibility of the quantification method, the CF peptides of target proteins in clinical serum were quantified and compared as a preliminary demonstration.

G

lycosylation is a common post-translational modification, which has been estimated to occur on at least 50% of human proteins.1 Glycosylation plays an important biological role in a variety of cellular processes, such as cell recognition, signal transduction, tumor invasion, metastasis, and immune response. Altered glycosylation patterns can greatly affect the functions of glycoproteins and reflect pathological changes.2 Core fucosylation (CF), which has a structure of an α-1,6 fucose substitution on the innermost N-acetylglucosamine (GlcNAc) of the pentasaccharide core of N-linked glycans, has attracted attention because its variations have been reported to be closely related to cancers.3
10 As a practical example, the CF form of α feto-protein (AFP-L3) has been approved by the FDA as a reliable biomarker for early diagnosis of hepatocellular carcinoma (according to the accessdata published on www.accessdata.fda.gov). The study of protein glycosylation is based on the acquisition of comprehensive information, including glycosites and glycan structures.11 Unlike acetylation and phosphorylation, more than one type of glycan structure may exist on one glycosite. This microheterogeneity poses a great challenge for glycosylation analysis.12 To bypass the difficulty in identifying glycoproteins, in one commonly used strategy, the glycans were completely released by PNGase F treatment of the protein or peptide, followed by mass spectrometric analysis.13 N-Glycosites were r 2011 American Chemical Society

identified by monitoring the mass shift (0.98 Da) generated by the deamidation of asparagine residues to aspartic acid due to the action of PNGase F. Although the strategy has been widely used and has gained great success in identifying glycoproteins,14
16 false assignments were still a problem because of the spontaneous deamidation of asparagine residues. In our earlier research, an innovative strategy was developed for the precise and large scale identification of CF proteins, in which the innermost GlcNAc and fucose groups are kept as glycopeptides through Endo F3catalyzed partial deglycosylation.17 The preferential cleavage of the glycosidic bond between fucose and GlcNAc during tandem MS analysis generates significant neutral loss peak, through which the MS3 scan is triggered to acquire a uniform fragmentation spectrum of peptide bearing a GlcNAc group. Over 100 CF glycoproteins and CF glycosites were identified in plasma through this strategy. With the qualitative identification of CF proteins, quantitative methods also need to be developed to measure the change in glycoproteins under different pathological or physiological conditions. In traditional experimental procedures, CF proteins or Received: June 30, 2011 Accepted: October 4, 2011 Published: October 04, 2011 8802

dx.doi.org/10.1021/ac201676a | Anal. Chem. 2011, 83, 8802–8809

Analytical Chemistry peptides are enriched and separated with sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) or liquid chromatography (LC), followed by deglycosylation by PNGase F and identification by MS.18
21 The relative quantitative analysis of the glycoproteins in normal or diseased samples is performed by comparing gel-staining intensity or peak intensity during MS analysis. The power of multiple reaction monitoring-mass spectrometry (MRM-MS) based quantitation strategy advanced recently in proteome research is quickly acknowledged and used for targeted quantitative research of glycoproteins. For example, Li et al.22 established a method for quantitative measurement of glycosylated and sialylated prostate-specific antigen (PSA) in prostate cancer and noncancer tissues using solid-phase extraction of glycopeptides followed by PNGase F treatment and a high-throughput MRM-based assay. Because PNGase F deglycosylation was adopted in the study, false assignments by the spontaneous deamidation of asparagine residues might be caused. Ahn et al.23 quantified the aberrant GlcNAcylated TIMP1 from colon cancer serum by L-PHA-Enrichment and SISCAPA with MRM-MS. In this study, glycoprotein was quantified indirectly via MRM-MS analysis of nonglycosylated peptides. Kurogochi et al.24 described a novel strategy for quantitative determination of sialylated glycopeptides, which were pyridyl aminated-labeled, enriched, and then analyzed by LC
MRM-MS analyses. Their research made a promising strategy for the enrichment and quantitation of sialylated glycoproteins. In this study, our intention is to develop a novel quantification method for CF glycopeptides, which combines our previously developed strategy for CF peptide identification with an MRMMS-based quantification strategy. The fragmentation behavior of simplified CF peptides in triple quadrupole MS was investigated first, and the MS condition for the selection of the MRM transition was optimized. Next, a CF site-specific quantification based on the MRM-MS analysis was realized by introducing an 18 O stable isotope labeling technique to obtain the relative CF abundance levels between the samples and the pooled internal standard sample. Finally, the feasibility of the method was evaluated with clinical serum, and the result illustrates the potential of implementing an MRM-based strategy in the sitespecific quantification of protein core fucosylated glycosylation.

’ EXPERIMENTAL METHODS Chemicals and Reagents. Recombinant human erythropoietin (rhEPO) was purchased from the National Institute for the Control of Pharmaceutical and Biological Products. Human IgG standard protein was obtained from Beijing Chengwen Immunochemistry Laboratory. Apo-transferrin, bovine thyroglobulin, end glycosidase F3 (Endo F3), α-D-methylmannoside, and tris-(2carboxylethyl)-phosphine (TCEP) were purchased from SigmaAldrich; iodoacetamide, trifluoroacetic acid, and formic acid were obtained from ACROS; acetonitrile (HPLC grade) was purchased from J. T. Baker; and trypsin (sequencing grade) was obtained from Promega (Madison, WI). LCH-sepharose 4B was purchased from GE. Normal serum and HCC serum samples were supplied by the China Human Liver Proteome Project (CNHLPP) sample bank. The serum samples, containing four healthy cases and four HCC cases, were pooled as the internal standard sample. Enrichment of Core Fucosylated Proteins. Spin columns filled with LCH sepharose 4B were used to enrich the CF proteins in 50 μL serum samples. Serum samples were diluted with binding buffer (20 mM Tris, 0.3 M NaCl, 1 mM MnCl2, 1 mM CaCl2, pH 7.4)

TECHNICAL NOTE

and incubated overnight with LCH sepharose 4B at 4 °C. The unbound protein was removed using the binding buffer. When adding 0.2 M α-D-methylmannoside to the binding buffer, the bound proteins were eluted, and the proteins were further desalted using Microcon YM-3 ultrafiltration tubes. Preparation of Core Fucosylated Peptides. The core fucosylated proteins that had been resolved in 0.1% RapiGest SF (Waters) solution were reduced with 5 mM TCEP at 56 °C for 1 h and alkylated with 25 mM IAA at room temperature in the dark. Trypsin was added at a ratio of 1:50 (protease/protein, weight/ weight). The proteins were digested in 50 mM ammonium bicarbonate at 37 °C overnight, and the reaction was stopped by the addition of formic acid at a final concentration of 0.5%. Digested CF proteins were ultrafiltered for 3  3 h at 8000g using Microcon YM-3 tubes to remove nonglycosylated peptides. The retention components on the ultrafiltration membrane were combined and lyophilized. The enriched core fucosylated peptides were resuspended in 50 mM ammonium acetate (pH 4.5) and incubated with Endo F3 overnight at 37 °C to remove the outer sugar chain. Peptide 18O Labeling. The 18O labeling was performed according to our previously developed method.25 The peptide solutions, containing 50 mM ammonium bicarbonate, trypsin at a ratio of 1:50 (protease/protein, weight/weight), and 0.1% RapiGest SF, were freeze-dried and resuspended in H218O. The solution was further incubated in a microwave oven for 10 min, which could promote the label speed while guaranteed the label efficiency.25 Trypsin remaining in the solution was deactivated by the addition of 0.5 M TCEP and 1 M IAA to inhibit the 18O
16O back-exchange. Reversed-Phase Capillary Liquid Chromatography
Mass Spectrometric Analysis. NanoLC (Eksigent) was used to run LC analysis. Online desalting was performed on a C18 precolumn (pepMap 0.3 mm  5 mm trap column, Dionex), and washes were performed with 100% phase A (2% ACN in 0.1% formic acid) at a flow rate of 5 μL/min for 35 min. The CF peptides were separated on a homemade reversed-phase capillary column (75 μm  260 mm, GEAgel, SP-300-ODS-AP, 5 μm) at a constant flow rate of 300 nL/min. The elution gradient was 1
7% phase B (80% ACN in 0.1% formic acid) for 4 min, 7
20% B for 21 min, 20
30% B for 4 min, 30
95% B for 5 min, 95% B for 41 min, 95
5% B for 5 min, and 5% B for 5 min. The eluents were analyzed by the 4000 QTRAP with a NanoSpray II source operated in the positive ionization mode. The optimal acquisition parameters were as follows: curtain gas (20), ion spray voltage (2300 V), ion source gas (30), interface heater temperature (150 °C), collision gas (high), declustering potential (80), entrance potential (10), and collision cell exit potential (15). The resolution parameters of the first and the third quadrupoles were set as “unit”. The target ions were transmitted with a narrow window (0.7 Da). The dwell time was 50 ms for every transition.

’ RESULTS AND DISCUSSION Fragmentation of Simplified CF Peptides in Triple Quadrupole MS. We observed earlier17 that the MS2 mass spectrum of

simplified CF peptides (retaining only fucosyl-GlcNAc glycan residues on glycosites) had a high intensity neutral-loss peak caused by the preferential loss of fucosylated residues. The MS3 mass spectrum triggered by the neutral-loss peak could provide richer fragment ion information for CF peptide identification. 8803

dx.doi.org/10.1021/ac201676a |Anal. Chem. 2011, 83, 8802–8809

Analytical Chemistry

TECHNICAL NOTE

Figure 1. The MS2 spectra of EEQYNGFSTYR in a triple quadrupole and an ion trap with increasing collision energy. (A) MS2 spectrum acquired in a triple quadrupole, CE = 25; (B) MS2 spectrum acquired in an ion trap, CE = 60; (C) MS2 spectrum acquired in a triple quadrupole, CE = 40; and (D) MS2 spectrum acquired in an ion trap, CE = 120. Label definitions: The fragment parent ion attached to a GlcNAc residue (PG+2); (B) b-type or y-type ions attached to a GlcNAc or fucosyl-GlcNAc residues (bG+, bGF+, yG+, yGF+); (C) b-type or y-type ions (b+, y+); and (D) fragment ions from glycoresidues (Dn+).

However, the correlation between fragment ions in an ion trap and in a triple quadrupole and the reference values for MRM transition selection still need to be evaluated to develop a feasible MRM-MS-based quantification method. The fragmentations of simplified CF peptides in a triple quadrupole were first investigated to select suitable transition ions for the development of an MRM-MS-based quantification strategy. The MS2 spectra of simplified CF peptides derived from IgG (EEQYNGFSTYR, “GF” standed for the fucosyl-GlcNAc glycan residues retaining on glycosites) were acquired and compared using the quadrupole part and the ion trap part of the mass spectrometer with different collision energies (Figure 1). In the ion trap, the MS2 spectrum was dominated by parent ions attached to a GlcNAc residue (PG+2), while other ions appeared at very low abundance even at collision energies from 60 to 120. Collisions that occurred in the triple quadrupole were quite different. With an increase in collision energy from 25 to 40, the number of peaks increased, and the intensity of each peak became more and more uniform. The phenomena are inconsistent with the fragment principle of the two instruments.26 In the ion trap, fragmentation energy was obtained by m/z-dependent resonance excitation, which means that the fragmented ions do not reach the energy required for secondary fragmentation. Compared to the ion trap, the fragmented ions in the triple quadrupole had a higher chance of becoming refragmented in the second quadrupole, which was filled with low-pressure nitrogen. According to the results determined by pLabel (a software designed for mass peak assignment, http://pfind.ict.ac.cn), these fragmented ions can be broadly divided into the following categories: A ions, the parent peptide ion attached with a GlcNAc residue (PG+2); B ions, b-type or y-type ions attached with a GlcNAc or fucosylGlcNAc residues (bG+, bGF+, yG+, yGF+); C ions, b-type or y-type ions (b+, y+); or D ions, fragment ions from glycan residues (Dn+). Because of these results, further questions were raised, including the following. Will other CF peptides present similar

fragmentation patterns? Will different quantitative linear results be obtained if different MRM transitions are chosen? Which fragmentation ions can be selected as the preferred ions to constitute MRM transitions? To resolve these issues, experiments were performed as follows. First, the dissociation profiles of six standard CF peptides were observed, while the changes in the relative abundance of fragment ions were monitored with the collision energy varying from 20 to 60. As shown in Figure 2, with the increase of energy, the abundance curve of parent ions began to decline (Figure 2, yellow curves). At low energies (25 < CE < 30), the glycosidic bond that links fucose and GlcNAc residues was prone to breakage, which led to the PG+2 or PG+++ ions that dominate the MS/MS spectra (Figure 2A
E, red curves). Moreover, when the energy was increased to a medium level (30 < CE < 40), fragmentation occurred on the peptide main chain, forming a series of bG+, yG+, b+, and y+ ions (Figure 2A
E, blue curves). The relative intensity of the y-ions reached its peak at these energies. With a further increase in collision energy, CF peptides at a higher energy (CE > 50) were cleaved into smaller fragments, identified as the ions generated by the fragmentation of sugar chains (Figure 2A
E, black curves). For peptide DPAEAQANGFASCPGVTYDQDSR, more energy was needed to obtain the MS/MS cleavage for doublycharged CF peptide with a long sequence. In Figure 2F, when the value of the collision energy was set to more than 30, the intensity curve of the parent ion showed a downward trend. Meanwhile the collision energy at which the signal relative intensity of the PG+2 ion, y+ ions, and sugar fragmentation ions reaches the maximum, increased by 10 units accordingly. Similar fragmentation profiles with different collision energies were demonstrated for different CF peptides in Figure 2. Among these fragment ions, more attention was paid to y-ions due to their importance for developing relative quantification methods based on 18O labeling. Many tools such as MRMPilot and Pinpoint were designed to predict highly abundant y-ions of peptides. A high degree of consistency between predicted and 8804

dx.doi.org/10.1021/ac201676a |Anal. Chem. 2011, 83, 8802–8809

Analytical Chemistry

TECHNICAL NOTE

Figure 2. The change curves of the relative abundance of fragment ions with the variation of collision energy in the range of 20
60, for the following CF peptides: (A) APEHLNGFWTGSWEATKPR; (B) QQQHLFGSNGFVTDCSGNFCLFR; (C) GQALLVNGFSSQPWEPLQLHVDK; (D) EAEEIVTYSNGFSSR; (E) EEQYNGFSTYR; (F) DPAEAQANGFASCPGVTYDQDSR. The average values of three replicates per collision energy point were used to prepare the curves. Data are listed in the Supporting Information.

experimental results can be obtained for unmodified peptides. However, will the introduction of CF modification affect the consistency? It is a key issue that needs to be explored before these tools are applied to predict fragment ion intensity of CF peptides. In our further experiment, the experimental y-ions in the MS2 spectra of CF peptides were compared with those in the MS2 spectra of PNGase F-treated CF peptides to investigate the effect of CF modification on the peptide fragmentation behavior. Our results showed that introducing CF modification to peptides did not affect the fragmentation behavior of the main chain. The highly abundant y-ions appearing in both spectra had exactly the same m/ z value. Because of this feature, we can easily predict the y-ions of CF peptides to develop the MRM method using software. To further examine the feasibility of all four kinds of ions for the development of an MRM-MS-based quantification strategy, the quantitative linear results of A, B, C, and D ions at their optimal CE condition were all evaluated. First, the digested BSA samples spiked with 0.5, 1, 2, 3, 4, and 6 pmol of CF peptides

(EEQYNGFSTYR) were analyzed. The quantification curves of different transitions were parallel to each other with close correlation coefficients, which suggest that there is no significant difference in the linear response for different transitions (Figure 3). However, the linear regression slope derived from the PG+2 ion was significantly higher than those from the other ions, indicating that this transition offers higher sensitivity for quantification than other ions. We further diluted the sample solution to 6.25, 12.5, 25, 50, 75, 100, and 200 fmol and observed the same result. No signal was detected for other transitions, except for the PG+2 ion. The correlation coefficient of the curve was 0.999. Because the intensity and stability of the PG+2 ion is much better than that of any other ions, it can be used as a preferred ion to constitute the quantification transition. In addition, high-intensity y-series ions of peptides could also be obtained and used to constitute transitions with the following advantages: the y ions retained partial peptide sequence information to ensure the specificity of peptide detection; the y ions could be coupled with an isotope labeling method to develop an 8805

dx.doi.org/10.1021/ac201676a |Anal. Chem. 2011, 83, 8802–8809

Analytical Chemistry

TECHNICAL NOTE

Figure 3. Calibration curves of different transitions derived from EEQYNGFSTYR: (A) calibration curves of the gradually diluted solutions containing 0.5
6 pmol of CF peptide; (B) zoomed-in region of the curves; (C) calibration curves of transition 769.8/696.8 obtained by MRM analysis of the gradient dilution solution containing 6.25
200 fmol of CF peptide. The average values of three replicates per concentration point were used to prepare the curves. Data are listed in the Supporting Information.

Figure 4. Experimental flow chart of site-specific quantification of protein CF glycosylation using MRM-MS. Briefly, equimolar bovine thyroglobulin was added to each serum sample. Then the CF proteins enriched from the solution by LCH lectin were digested into peptides. After the addition of 18Olabeled normalized internal standard (CF peptides from pooled serum samples) and 18O-labeled CF peptides from bovine thyroglobulin, the CF peptides in the solution were extracted by ultrafiltration, partially deglycosylated by Endo F3 and finally analyzed by MRM-MS.

accurate quantification strategy; and the highly abundant y ions of deglycosylated versions of peptides could provide references for the MRM transition selection of CF peptides. Method Development for Site-Specific Quantification of CF Modifications in Serum. On the basis of the above results, a

site-specific quantification strategy for the CF modification of proteins by MRM-MS was developed. To evaluate the efficiency of this method, the CF proteins in clinical human serum were site-specifically quantified by combining two-step CF peptide enrichment and MRM-MS analysis (Figure 4). During the 8806

dx.doi.org/10.1021/ac201676a |Anal. Chem. 2011, 83, 8802–8809

Analytical Chemistry

TECHNICAL NOTE

Table 1. MRM Transition of CF Peptides in Serum Samplesa protein name

a

CF-peptide

Q1

Q31

Q32

Q33

Ig γ-1 chain C region

EEQYNGFSTYR

769.8

696.8

1189.5

1006.5

hemopexin precursor

SWPAVGNGFCSSALR

877.4

804.4

1067.5

1334.6

Ig γ-2 chain C region

EEQFNGFSTFR

753.8

1360.4

1157.3

974.4

Ig α-2 chain C region

TPLTANGFITK

654.3

581.3

647.4

760.3

α-2-macroglobulin precursor

GCVLLSYLNGFETVTVSASLESVR

916.1

867.4

848.5

1048.6

α-2-macroglobulin precursor

VSNGFQTLSLFFTVLQDVPVR

1256.7

1183.6

826.5

1320.6

ceruloplasmin precursor

ENGFLTAPGSDSAVFFEQGTTR

1238.6

1165.7

1598.5

985.2

bovine thyroglobulin (IS)

APEHLNGFWTGSWEATKPR

777.4

728.6

572.4

1132.6

bovine thyroglobulin (IS)

EAEEIVTYSNGFSSR

917.4

844.5

1017.5

1116.4

Q31,Q32 and Q33 were selected daughter ions of the CF peptide.

process, we considered the following: (a) CF protein enrichment by lectin columns. To capture enough glycoprotein so as not to discount accurate and reliable comparison between samples, we examined the enrichment efficiency of CF proteins from the total serum protein. A volume of 300 μL of LCH sepharose 4B columns were used to capture CF proteins from different volumes of serum (16.7, 100, 150, 300, and 450 μL). The saturation concentration of the lectin column was investigated to optimize the volume of serum sample handled. As shown Figure S1 in the Supporting Information, when the volume of serum reached 300 μL, the lectin column enriched the CF proteins to their maximal concentration. On the basis of the result, 50 μL of serum was used as the optimal volume to enrich CF proteins from human serum with lectin columns. (b) The selection of the MRM transition: The fragment study of simplified CF peptides by triple quadrupole MS suggests that the intact peptide ion attached to a GlcNAc residue and the y-series ions without glycan groups could be selected to constitute the MRM transitions with the advantages of high sensitivity and specificity. The MRM transitions of simplified CF peptides derived from six serum proteins are listed in Table 1. MS2 spectra were obtained by a MIDAS workflow and used to identify the targeted simplified CF peptides. Marked with pLabel software, the spectra of nine glycopeptides (see the Supporting Information, Figure S2) verified the specificity of MRM detection. (c) Correction of the relative quantitative results. In our study, the enrichment of the CF peptides was performed at two levels. At the protein level, LCH sepharose 4B was used to remove a large number of nonglycosylated proteins and capture the CF proteins. At the peptide level, peptide mixtures digested from glycoproteins were separated with a 3000 Da cutoff membrane from the unmodified peptides. In this two-step enrichment process, variability is hard to avoid. Thus, internal standard correction was taken into account to correct the quantitative results. 18O Labeling was a mild and efficient isotope incorporation method,27 which was used to prepare internal standards. In our study, 18Olabeled CF peptides of bovine thyroglobulin were added as the first internal standard to correct the deviation derived from the CF protein enrichment. The recovery was calculated by comparing the abundance before and after lectin enrichment. Bovine thyroglobulin is a highly core fucosylated protein, which is usually used to evaluate the activity of lentil lectin as a standard. The CF peptides derived from this protein can be distinguished from human thyroglobulin by MS because of the difference in protein sequence. Thus, in our study, two CF peptides from bovine thyroglobulin, APEHLNGFWTGSWEATKPR and EAEEIVTYSNGFSSR, were selected to calculate the recovery rates of

the enrichment method. Bovine thyroglobulin was added into human serum samples before lectin enrichment, while two 18O labeled CF peptides from bovine thyroglobulin were added after lectin enrichment. The peak areas of labeled and unlabeled standard peptides were compared to calculate the recovery rates. recovery rate of lectin enrichment ¼

A16 GF  100% A19 GF

16 19 A16 GF is the peak area of O CF peptides, and AGF is the peak area 18 of O labeled CF peptides. As the second internal standard, the 18O-incorporated peptide mixture of pooled serum, which was used to calibrate variation produced in the CF peptides enrichment process, was mixed with digested CF proteins from serum samples before filtration with a 3000 Da cutoff. Finally, the ratio of 18O and 16O labeled CF peptides was calculated to obtain the relative quantification results. Quantification of the CF-Modification of Proteins in HCC Serum. Hepatocellular carcinoma (HCC) is the fifth most common cancer and is rapidly fatal, making the development of biomarkers for diagnosis of great importance.28 The alteration of core fucosylation (CF) is closely related to the development of HCC, and some CF proteins have been reported as potential diagnostic markers.3,29 In the study of Comunale et al.3, 19 proteins were found to be HCC-associated CF proteins using multiple proteomic methodologies by comparison of pooled normal and cancer serum. To evaluate the feasibility of our developed strategy, we selected six CF proteins from the report of Comunale et al.3 and validated their CF alteration in individual serum by target quantitative analysis of seven CF peptides identified in our previous research17 (Table 1). According to the quantitative results of 18 samples (Table 2), the average level ratios of TPLTANGFITK (Ig α-2 chain C region levels), SWPAVGNGFCSSALR (hemopexinprecursor), and ENGFLTAPGSDSAVFFEQGTTR (Ceruloplasmin precursor) were increased in HCC patients (HCC/healthy ratio > 1.5), which is consistent with the published results with pooled serum.3,30 However, a significant difference between the healthy and HCC groups was not obtained for the CF levels of seven CF peptides. The large individual difference is probably one of the reasons that lead to the results. More clinical samples with more CF peptides are needed to validate the feasibility of the strategy. Our strategy has just demonstrated its potential in the analysis of the quantitative variation of CF peptide in HCC serum to find biomarker candidates. The results also indicate that, for discovery

8807

dx.doi.org/10.1021/ac201676a |Anal. Chem. 2011, 83, 8802–8809

Analytical Chemistry

TECHNICAL NOTE

Table 2. Relative Quantitative Results of Seven CF Peptides from Six Proteins in Clinical Seruma protein name Ig γ-1 chain C region

CF peptide EEQYNGFSTYR

hemopexin precursor

SWPAVGNGFCSSALR

Ig γ-2 chain C region

EEQFNGFSTFR

sample

1

2

3

4

5

6

7

8

9

average

HCC

0.735

1.317

0.674

1.114

0.694

0.895

0.458

1.570

1.588

1.005

healthy

0.686

0.947

1.265

1.262

1.228

1.304

0.951

0.856

0.456

0.995

HCC

1.048

1.998

0.924

0.945

1.394

1.418

0.468

1.668

1.142

1.223

healthy

0.491

0.504

1.080

1.009

0.584

0.588

0.272

1.776

0.690

0.777

HCC

1.369

1.042

0.584

0.950

0.446

1.037

1.151

1.099

1.219

0.989

healthy

0.590

0.739

1.235

1.492

1.335

1.238

0.806

0.728

0.940

1.011

Ig α-2 chain C region

TPLTANGFITK

HCC

1.052

0.920

1.138

2.201

0.913

0.409

0.818

1.846

2.418

1.302

α-2-macroglobulin

GCVLLSYLNGFETVTVSASLESVR

healthy HCC

0.162 1.041

0.373 0.990

0.636 1.073

1.518 1.067

0.936 0.973

0.324 0.764

0.352 0.967

1.191 1.072

0.792 0.812

0.698 0.973

healthy

0.749

0.988

1.049

1.043

1.279

0.841

0.834

1.460

0.998

1.027

HCC

1.144

1.231

1.155

0.879

0.965

0.670

0.961

1.582

0.564

1.017

healthy

0.933

0.997

1.045

0.617

1.300

0.868

0.947

1.031

1.111

0.983

precursor916 α-2-macroglobulin

VSNGFQTLSLFFTVLQDVPVR

precursor1256 ceruloplasmin precursor

ENGFLTAPGSDSAVFFEQGTTR

HCC

1.079

0.877

0.913

1.452

0.851

0.940

1.691

3.488

0.687

1.331

healthy

0.263

0.536

0.941

1.939

0.387

0.194

0.303

0.967

0.491

0.669

a

Clinical serum samples containing nine healthy cases and nine HCC cases were site-specifically quantified by MRM-MS. The data listed in the table are the ratio of each serum sample and pooled sera.

researches that look for biomarker candidates from pooled serum, further study is needed to validate the discovered molecules from individual samples in a large scale. One thing should be noted is that the partially deglycosylated peptides for target quantitation in our research may come from the same glycopeptides bearing different glycoforms, since Endo F3 preferentially cleaves core-fucosylated biantennary and triantennary glycans. Thus, currently our strategy could not be used to monitor the level of glycopeptides in a glycoform-specific manner. In addition, we think that further optimization and evaluation is needed before this method can be widely applied, especially the detection sensitivity of CF peptides in MRM-MS. Because of the hydrophilic nature of residues bearing fucosyl-GlcNAc modifications on the simplified peptide, the ionization efficiency of the CF peptide was compromised. Additionally, the MS signal of the CF peptides may be suppressed by coeluted nonglycopeptides. To improve the detection sensitivity and obtain more comprehensive information, more effective strategies should be adopted, including optimizing enrichment conditions or separation procedures and increasing the specificity of enrichment. Various techniques could be considered for this purpose. Online separation of glycopeptides by combining lectin microcolumns and nano C18 columns31 or the use of a ZIC-HILIC SPE,32 derivatization method using 1-pyrenyldiazomethane,33 etc., all of these methods may significantly improve the MS response of glycopeptides. When coupled with these techniques, we expect that MRM-MS will approach the quantification information of more CF sites, thus facilitating in-depth quantitative analysis of CF modification under pathological conditions.

’ CONCLUSIONS In this study, by analyzing the MS2 spectrum acquired by triple quadrupole mass spectrometry, we found that daughter ions of standard CF peptides, including PGn+, yG+, and y+, appeared regularly with high abundance. When these ions were used for MRM quantification, good linear relationships were obtained that supported the use of these ions as the optimal fragment ions to constitute MRM transitions. On the basis of the above results, the CF peptides of target proteins in clinical serum samples were quantified by MRM-MS combined with an enrichment procedure

at the protein and peptide levels and an isotope incorporation strategy. These methods offer a promising way to screen interesting core fucosylated glycoproteins that are related to different pathological processes.

’ ASSOCIATED CONTENT

bS

Supporting Information. Figures S1 and S2 and Tables S1
S8. This material is available free of charge via the Internet at http://pubs.acs.org.

’ AUTHOR INFORMATION Corresponding Author

*Xiaohong Qian: e-mail, [email protected]. Wantao Ying: e-mail, [email protected]; phone, 8610-80705055; fax, 8610-80705155.

’ ACKNOWLEDGMENT Yan Zhao and Wei Jia contributed equally to this work. We are grateful for financial support from the National Key Program for Basic Research of China (Grants 2011CB910603 and 2012CB910603), the National Natural Science Foundation of China (Grants 20735005, 20905077, 30900258, and 31100591), the International Scientific Cooperation Project of China (Grant 2011DFB30370), and the National Key Laboratory Foundation of China (Grants SKLP-Y200902). ’ REFERENCES (1) Wong, C.-H. J. Org. Chem. 2005, 70, 4219–4225. (2) Tian, Y.; Zhang, H. Proteomics Clin. Appl. 2010, 4, 124–132. (3) Comunale, M. A.; Lowman, M.; Long, R. E.; Krakover, J.; Philip, R.; Seeholzer, S.; Evans, A. A.; Hann, H.-W. L.; Block, T. M.; Mehta, A. S. J. Proteome Res. 2006, 5, 308–315. (4) Noda, K.; Miyoshi, E.; Gu, J.; Gao, C.-X.; Nakahara, S.; Kitada, T.; Honke, K.; Suzuki, K.; Yoshihara, H.; Yoshikawa, K.; Kawano, K.; Tonetti, M.; Kasahara, A.; Hori, M.; Hayashi, N.; Taniguchi, N. Cancer Res. 2003, 63, 6282–6289. (5) Noda, K.; Miyoshi, E.; Uozumi, N.; Yanagidani, S.; Ikeda, Y.; Gao, C.-x.; Suzuki, K.; Yoshihara, H.; Yoshikawa, M.; Kawano, K.; Hayashi, N.; Hori, M.; Taniguchi, N. Hepatology 1998, 28, 944–952. 8808

dx.doi.org/10.1021/ac201676a |Anal. Chem. 2011, 83, 8802–8809

Analytical Chemistry (6) Noda, K.; Miyoshi, E.; Uozumi, N.; Gao, C.-X.; Suzuki, K.; Hayashi, N.; Hori, M.; Taniguchi, N. Int. J. Cancer 1998, 75, 444–450. (7) Block, T. M.; Comunale, M. A.; Lowman, M.; Steel, L. F.; Romano, P. R.; Fimmel, C.; Tennant, B. C.; London, T.; Evans, A. A.; Blumberg, B. S.; Dwek, R. A.; Mattu, T. S.; Mehta, A. S. Proc. Natl. Acad. Sci. U.S.A. 2005, 102, 779–784. (8) Okuyama, N.; Ide, Y.; Nakano, M.; Nakagawa, T.; Yamanaka, K.; Moriwaki, K.; Murata, K.; Ohigashi, H.; Yokoyama, S.; Eguchi, H.; Ishikawa, O.; Ito, T.; Kato, M.; Kasahara, A.; Kawano, S.; Gu, J.; Taniguchi, N.; Miyoshi, E. Int. J. Cancer 2006, 118, 2803–2808. (9) Barrabes, S.; Pages-Pons, L.; Radcliffe, C. M.; Tabares, G.; Fort, E.; Royle, L.; Harvey, D. J.; Moenner, M.; Dwek, R. A.; Rudd, P. M.; Llorens, R. D.; Peracaula, R. Glycobiology. 2007, 17, 388–400. (10) Saldova, R.; Royle, Louise; Radcliffe, C. M.; Hamid, U. M. A.; Evans, R.; Arnold, J. N.; Banks, R. E.; Hutson, R.; Harvey, D. J; Antrobus, R.; Petrescu, S. M.; Dwek, R. A.; Rudd, P. M. Glycobiology 2007, 17, 1344–1356. (11) An, H. J.; Froehlich, J. W.; Lebrilla, C. B. Curr. Opin. Chem. Biol. 2009, 13, 421–426. (12) North, S. J.; Hitchen, P. G.; Haslam, S. M.; Dell, A. Curr. Opin. Struct. Biol. 2009, 19, 498–506. (13) Chaerkady, R.; Thuluvath, P. J.; Kim, M.-S.; Nalli, A.; Vivekanandan, P.; Simmers, J.; Torbenson, M.; Pandey, A. Clin. Proteomics 2008, 4, 137–155. (14) Zhang, H.; Li, X.-j.; Martin, D. B.; Aebersold, R. Nat. Biotechnol. 2003, 21, 660–666. (15) Parker, B. L.; Palmisano, G.; Edwards, A. V. G.; White, M. Y.; Engholm-Keller, K.; Lee, A.; Scott, N. E.; Kolarich, D.; Hambly, D; Packer, B.; Larsen, N. H.; Cordwell, M. R. Mol. Cell. Proteomics 2011, 10, M110.006833. (16) Liu, T.; Qian, W.-J.; Gritsenko, M. A.; Camp, D. G.; Monroe, M. E.; Moore, R. J.; Smith, R. D. J. Proteome Res. 2005, 4, 2070–2080. (17) Jia, W.; Lu, Z.; Fu, Y.; Wang, H.-P.; Wang, L.-H.; Chi, H.; Yuan, Z.-F.; Zheng, Z.-B.; Song, L.-N.; Han, H.-H.; Liang, Y.-M.; Wang, J.-L.; Cai, Y.; Zhang, Y.-K.; Deng, Y.-L.; Ying, W.-T.; He, S.-M.; Qian, X.-H. Mol. Cell. Proteomics 2009, 8, 913–923. (18) Ahn, Y. H.; Kim, Y.-S.; Ji, E. S.; Lee, J. Y.; Jung, J.-A.; Ko, J. H.; Yoo, J. S. Anal. Chem. 2010, 82, 4441–4447. (19) Lee, H.-J.; Na, K.; Choi, E.-Y.; Kim, K. S.; Kim, H.; Paik, Y.-K. J. Proteome Res. 2010, 9, 308–318. (20) Ueda, K.; Takami, S.; Saichi, N.; Daigo, Y.; Ishikawa, N.; Kohno, N.; Katsumata, M.; Yamane, A.; Ota, M.; Sato, T.-A.; Nakamura, Y.; Nakagawa, H. Mol. Cell. Proteomics 2010, 9, 1819–1828. (21) Wei, X.; Li, L. Brief. Funct. Genomics Proteomics 2010, 8, 104–113. (22) Li, Y.; Tian, Y.; Rezai, T.; Prakash, A.; Lopez, M. F.; Chan, D. W.; Zhang, H. Anal. Chem. 2011, 83, 240–245. (23) Ahn, Y. H.; Lee, J. Y.; Lee, J. Y.; Kim, Y.-S.; Ko, J. H.; Yoo, J. S. J. Proteome Res. 2009, 8, 4216–4224. (24) Kurogochi, M.; Matsushista, T.; Amano, M.; Furukawa, J.-i.; Shinohara, Y.; Aoshima, M.; Nishimura, S.-I. Mol. Cell. Proteomics 2010, 9, 2354–2368. (25) Zhao, Y.; Jia, W.; Sun, W.; Jin, W.; Guo, L.; Wei, J.; Ying, W.; Zhang, Y.; Xie, Y.; Jiang, Y.; He, F.; Qian, X. J. Proteome Res. 2010, 9, 3319–3327. (26) Sherwood, C. A.; Eastham, A.; Lee, L. W.; Risler, J.; Vitek, O.; Martin, D. B. J. Proteome Res. 2009, 8, 4243–4251. (27) Yao, X.; Freas, A.; Ramirez, J.; Demirev, P. A.; Fenselau, C. Anal. Chem. 2001, 73, 2836–2842. (28) Uto, H.; Kanmura, S.; Takami, Y.; Tsubouchi, H. Proteome Sci. 2010, 31, 70–80. (29) Wang, M.; Long, R. E.; Comunale, M. A.; Junaidi, O.; Marrero, J.; Bisceglie, A. M. D.; Block, T. M.; Mehta, A. S. Cancer Epidemiol. Biomarkers Prev. 2009, 18, 1914–1921. (30) Callewaert, N.; Vlierberghe, H. V.; Hecke, A. V.; Laroy, W.; Delanghe, J.; Contreras, R. Nat. Med. 2004, 10, 429–434. (31) Madera, M.; Mechref, Y.; Novotny, M. V. Anal. Chem. 2005, 77, 4081–4090.

TECHNICAL NOTE

(32) Mysling, S.; Palmisano, G.; Højrup, P.; Thaysen-Andersen, M. Anal. Chem. 2010, 82, 5598–5609. (33) Amano, J.; Nishikaze, T.; Tougasaki, F.; Jinmei, H.; Sugimoto, I.; Sugawara, S.-i.; Fujita, M.; Osumi, K.; Mizuno, M. Anal. Chem. 2010, 82, 8738–8743.

8809

dx.doi.org/10.1021/ac201676a |Anal. Chem. 2011, 83, 8802–8809