Synthesis of magnetic zwitterionic–hydrophilic material for the selective enrichment of N-linked glycopeptides

Journal of Chromatography A, 1482 (2017) 23–31

Contents lists available at ScienceDirect

Journal of Chromatography A journal homepage: www.elsevier.com/locate/chroma

Synthesis of magnetic zwitterionic–hydrophilic material for the selective enrichment of N-linked glycopeptides Yiman Zhao a , Yajing Chen a , Zhichao Xiong b , Xudong Sun a , Quanqing Zhang c , Yangyang Gan a , Lingyi Zhang a,∗ , Weibing Zhang a,∗ a Shanghai Key Laboratory of Functional Materials Chemistry, School of Chemistry and Molecular Engineering, East China University of Science and Technology, Shanghai, 200237, PR China b State Key Laboratory of High Performance Ceramics and Superfine Microstructure, Shanghai Institute of Ceramics, Chinese Academy of Sciences, Shanghai, 200050, PR China c Fudan University, Shanghai, 200433, PR China

a r t i c l e

i n f o

Article history: Received 17 September 2016 Received in revised form 17 November 2016 Accepted 17 December 2016 Available online 19 December 2016 Keywords: Magnetic nanoparticles Glycopeptides Zwitterionic-hydrophilic interaction liquid chromatography Enrichment Mass spectrum

a b s t r a c t In consideration of the close connection between glycopeptides and human diseases, the efficient method to separate and enrich glycopeptides from complex biological samples is urgently required. In the work, we developed a magnetic zwitterionic-hydrophilic material for highly effective separation and analysis of glycopeptides from complex samples. The Fe3 O4 particles were covered with a thick layer of polymer by one-step reflux-precipitation polymerization (RPP), subsequently decorated by Au nanoparticles (Au NPs) through in situ reduction and finally modified with zwitterionic groups. The abundant zwitterionic sites facilitate the selective enrichment of glycopeptides. Besides, the prepared Fe3 O4 @PGMA@Au-lcys showed high detection sensitivity (5 fmol IgG digest), approving enrichment capacity (75 mg g−1 ), satisfactory enrichment recovery (89.8%), and great performance in the analysis and profiling of lowabundance N-linked glycopeptides. Furthermore, the prepared material was employed in the enrichment of glycopeptides in intricate biological samples, and 774 unique N-glycosylation sites from 411 Nglycosylated proteins were reliably identified in three replicate analyses of a 75 ␮g protein sample extracted from mouse liver, suggesting wide application prospect in glycoproteomics. © 2016 Elsevier B.V. All rights reserved.

1. Introduction Protein glycosylation, one of the most important posttranslational modifications (PTMs) in the proteome, plays irreplaceable role in various biological processes, such as cell adhesion, signal transduction and immune response [1–3]. The degree of protein glycosylation and the changes of glycan chain structure are closely related to many diseases, such as cancer, immune deficiency, etc [4–6]. Therefore, the profiling of glycopeptides in the complex biology samples becomes intensely important. Nevertheless, vast non-glycopeptides in the sample severely disturb and restrain the detection of low abundant glycopeptides. It’s crucially significant to separate and purify glycopeptides from complex samples and identify their glycosylation sites. So far, various methods have been developed to effectively enrich glycopeptides, includ-

∗ Corresponding authors. E-mail addresses: [email protected] (L. Zhang), [email protected] (W. Zhang). http://dx.doi.org/10.1016/j.chroma.2016.12.054 0021-9673/© 2016 Elsevier B.V. All rights reserved.

ing lectin-based affinity chromatography (LAC) [7–9], hydrazine chemical reaction (HC) [10–12], boronic acid affinity chromatography (BAAC) [13,14] and hydrophilic interaction chromatography (HILIC) [15–19]. Among them, HILIC is the most widely utilized based on the stronger hydrophilic character of glycan in glycopeptides compared with other peptides. HILIC could enrich different glycopeptides unbiasedly without damaging the structure of glycan in glycopeptides, thus, the integrated structure information of glycopeptides could be obtained [20]. Based on the mild synthesis condition, high sensitivity and great reproducibility of HILIC materials, a large number of HILIC materials, such as maltose [21], zwitterionic materials [22], dopamine [23], hydrophilic dendrimer [24] have been applied to enriching glycopeptides from intricate samples. What’s more, with both positive and negative charges, the zwitterionic (ZIC)-HILIC material is a good candidate for glycopeptides enrichment based on the interactions of charge–charge and charge-dipole between material and analyte targets. Abundant zwitterionic groups could facilitate the concentrating of glycopeptides in the water layer formed by zwitterinoic groups. In the reported approaches, graphene oxide combined with L-cysteine

24

Y. Zhao et al. / J. Chromatogr. A 1482 (2017) 23–31

[25,26], zwitterionic polymer brushes hybrid silica nanoparticles [12], and zwitterionic resin [27] were synthesized and applied to enriching N-linked glycopeptides from biological samples. The ZIC HILIC materials show high selectivity and great enrichment recovery and are regarded as the super materials in enrichment of glycopeptides [28,29]. However, to further improve detection sensitivity of low abundant glycopeptides in the complex biological samples, it is necessary to design novel ZIC-HILIC materials with abundant acting sites by a facile synthesis method. Magnetic nanomaterials are developing rapidly in recent years because of their quantum effect, small size effect, surface effect and uniquely magnetic properties, such as, super paramagnetic and high coercive force [30]. At present, magnetic nanomaterials have been applied to various fields, such as, immobilization of enzyme [31], drug delivery [32], magnetic resonance imaging and protein and cell analysis [33]. Considering the biocompatibility, innocuity, easy surface modification, fast magnetic separation and other merits of functionalized magnetic materials, they were extensively applied to protein separation [34,35]. So far, zwitterionic polymer-coated core-shell magnetic nanoparticles [36], hydrophilic magnetic graphene oxide [21,37] and other functioned hydrophilic magnetic materials [38] were fabricated for rapid separation of glycopeptides from biologly samples. It is urgent to develop new high-loading hydrophilic magnetic material by simple approach for glycopeptides enrichment. Herein, we report a facile approach for the synthesis of a novel magnetic ZIC-HILIC material, Fe3 O4 @PGMA@Au-l-cys. A thick polymer layer was firstly introduced on the magnetic particle surface. After amination of the thick polymer layer, Au nanoparticles were anchored in situ and further grafted with hydrophilic l-cysteine. Benefit from the inherently reducing and stabilizing properies of enormous NH2 groups, large amount of Au NPs were steadily anchored. The abundant Au NPs possess highly effective surface for the immobilization of amphoteric l-cysteine through an Au–S bond. Furthermore, the l-cysteine provides great zwitterionic acting sites for the enrichment of glycopeptides. The final Fe3 O4 @PGMA@Au-l-cys was demonstrated as a promising material to separate glycopeptides from complex biology samples with high enrichment capacity, great selectivity and enrichment recovery for glycopeptides in complex biology samples.

2. Experiment section 2.1. Materials and chemicals Iron(III) chloride hexahydrate (FeCl3 ·6H2 O), ethanol, ethylene glycol (EG), sodium acetate (NaAc), ethylenediamine and aqueous ammonia solution (NH3 ·H2 O, 28–30 wt%) were obtained from Shanghai Chemical Reagents Company (Shanghai, China). Glycidyl methacrylate (GMA), ␥-Methacryloxypropyltrimethoxy-silane (␥MPS), N,N -methylenebis-acrylamide (MBA), trifluoroacetic acid (TFA), formaldehyde, HPLC grade acetonitrile (ACN), l-cysteine (lcys) and ammonium bicarbonate (NH4 HCO3 ) were purchased from Aladdin (Shanghai, China). HAuCl4 ·3H2 O was obtained from J&K (Beijing, China). Human serum immunoglobulin G (human IgG), chicken avidin, horseradish peroxidase (HRP), albumin from bovine serum (BSA), trypsin (TPCK-treated), PNGase F, dithiothreitol (DTT), iodoacetamide (IAA), urea, sodium bicarbonate (NaHCO3 ), formaldehyde-D2 (20% solution in D2 O) and 2,5-dihydroxyl benzoic acid (DHB) were obtained from Sigma-Aldrich (St Louis, MO, USA). 2,2-azobisisobutyronitrile (AIBN) was supplied by Sinopharm Chemical Reagent Company (Shanghai, China). All the chemical agents were used without further purification. Purified water (18.2 M cm) was obtained with a Milli-Q apparatus (Millipore, Bedford, MA, USA).

2.2. Synthesis of Fe3 O4 @PGMA@Au-l-cys nanoparticles Fe3 O4 nanoparticles were prepared by solvothermal reaction. The procedure was described in detail in supporting information. Then the Fe3 O4 nanoparticles were modified with ␥-MPS. 300 mg of Fe3 O4 nanoparticles were dispersed in a mixture solution containing 40 mL of ethanol, 10 mL of water and 1.5 mL of NH3 ·H2 O with ultrasonication for 30 min. After the mixture was stirred for 30 min, 800 ␮L ␥-MPS was dropped into the mixture and stirred for 24 h at 60 ◦ C. The obtained Fe3 O4 @MPS was washed with ethanol and water, respectively. Then obtained nanoparticles were dried at 50 ◦ C. Fe3 O4 @PGMA microspheres were synthesized in acetonitrile by a one-step reflux-precipitation polymerization of GMA, with MBA as the cross-linker and AIBN as the initiator. Specially, 60 mg Fe3 O4 @MPS nanoparticles were dispersed in 40 mL of acetonitrile under ultrasonic treatment, and 140 ␮L GMA, 150 mg MBA, 8 mg AIBN were added sequentially. The homogeneous mixture was heated in oil bath at 90 ◦ C for 1.5 h. The Fe3 O4 @PGMA was separated and washed with ethanol three times to eliminate excess reactants by external magnetic field. Finally, the obtained product was dried at 50 ◦ C for 12 h. 200 mg Fe3 O4 @PGMA nanoparticles were dispersed in 80 mL of anhydrous ethylenediamine and heated at 80 ◦ C for 3 h, then the product of Fe3 O4 @PGMA-NH2 with abundant NH2 was obtained. With vigorously stirred for 3 h in the water, plentiful NH2 in polymer layer interacted with HAuCl4 , and Au NPs were in situ anchored to the Fe3 O4 @PGMA-NH2 . Finally, excessive l-cysteine was added to 50 mg Fe3 O4 @PGMA@Au to chelate Au nanoparticles through the Au-S bond in the water for 24 h. 2.3. Material characterization Transmission electron microscopy (TEM) images were collected from a JEOL JEM-2100 EX transmission electron microscope (JEOL, Tokyo, Japan) and field emission scanning electron microscopy (FESEM) images were obtained by NOVA Nano SEM 450 (FEI USA). Energy dispersive spectrometer (EDS) analysis was acquired by EDAX TEAM (USA). Fourier-transform infrared spectroscopy (FT-IR) characterization was obtained on a Thermo Nicolet 380 spectrometer using KBr pellets (Nicolet, Wisconsin, USA). Thermo gravimetric (TG) analysis was performed under a nitrogen atmosphere at a heating rate of 10 ◦ C min−1 from room temperature to 900 ◦ C (NETZSCH, Selb, Germany). The saturation magnetization curve was performed on a Physical Property Measurement System 9T (Quantum Design, San Diego, USA) at room temperature. 2.4. Tryptic digestion of standard protein and proteins extracted from mouse liver 1.3 mg standard protein (IgG, HRP or chicken avidin) was dispersed in 150 ␮L 100 mmol L−1 NH4 HCO3 solution which contains 8 mol L−1 urea to degenerate the proteins. 26 ␮L 1 mol L−1 DTT was added and the mixture was heated in the water bath for 1 h at 60 ◦ C. After that, to alkylated proteins, 9.36 mg IAA was added at room temperature avoiding light for 45 min. Afterwards, the solution was diluted ten-fold with 50 mmol L−1 NH4 HCO3 to decrease the concentration of urea. The solution was digested with trypsin at an enzyme-protein ratio of 1: 25 (w/w), and then, the mixture was incubated in 37 ◦ C water bath for 16 h. The tryptic digests were stored at −20 ◦ C for further use. 2 mg BSA was dissolved in 1 mL buffer containing 8 mol L−1 urea and 50 mmol L−1 NH4 HCO3 . Then 20 ␮L DTT (1 mmol L−1 ) was added and the mixture was incubated for 1 h at 60 ◦ C. The protein was alkylated by 7.2 mg IAA at ambient temperature in darkness for 45 min. After the mixture was diluted ten-fold by 50 mmol L−1 NH4 HCO3 , trypsin was added at the ratio of 40: 1 (protein/enzyme, w/w) and incubcated at 37 ◦ C for 16 h.

Y. Zhao et al. / J. Chromatogr. A 1482 (2017) 23–31

The proteins of mouse liver were extracted following the literature with slight modification [39]. 2 mg rat liver protein was dissolved in 2 mL denaturing buffer containing 40 mmol L−1 TrisHCl and 8 mol L−1 urea. To open the disulfide bond in proteins, 60 ␮L 1 mol L−1 DTT was added and kept at 60 ◦ C for 1 h, then 16 mg IAA was superadded to alkylated the opened disulfide bond for 45 min at indoor temperature without light. Then 50 mmol L−1 NH4 HCO3 was added to the solution volume expanded ten times and the proteins were digested with trypsin (protein/enzyme = 25/1, w/w) at 37 ◦ C for 16 h. The mouse liver protein digests was desalted by C18 column then stored at −80 ◦ C for further use. 2.5. Selective enrichment of glycopeptides by Fe3 O4 @PGMA@Au-l-cys nanoparticles Fe3 O4 @PGMA@Au-l-cys nanoparticles were washed with 400 ␮L loading buffer (ACN-H2 O-TFA, 88: 11.9: 0.1, v/v/v) for three times to remove impurities and activate effective groups, then dispersed in loading buffer. Then the Fe3 O4 @PGMA@Au-lcys nanoparticles and 3 ␮g IgG digest or other standard protein digest were added into 400 ␮L loading buffer and the mixture was incubated gently for 30 min at room temperature. Then, to remove the non-glycopeptides and other impurities adsorbed by the zwitterionic-hydrophilic nanoparticles, 400 ␮L loading buffer was applied to wash the nanoparticles for three times with the assist of external magnet. Afterwards, to collect glycopeptides, 10 ␮L eluting buffer (ACN-H2 O-TFA, 30: 69.9: 0.1, v/v/v) was added into the nanoparticles and the mixture was vigorously incubated for 10 min twice for further MALDI-TOF MS analysis or deglycosylation for LC–MS/MS analysis. As for the enrichment of glycopeptides from complex samples, 3 mg material was incubated with digests of proteins extracted from mouse liver at 25 ◦ C for 40 min. After the material was washed with the loading buffer (ACN-H2 O-TFA, 88: 11.9: 0.1, v/v/v, 3*500 ␮L), the captured glycopeptides were eluted with eluting buffer (ACN-H2 O-TFA, 30: 69.9: 0.1, v/v/v, 3*50 ␮L). Finally, the eluate was lyophilized and deglycosylated for LC-MS/MS analysis. 2.6. Deglycosylation of N-linked glycopeptides by PNGase F The eluting buffer was lyophilized and re-dissolved in 40 ␮L 10 mmol L−1 NH4 HCO3 , then, 50 units of PNGase F were added and the mixture was incubated at 37 ◦ C overnight. 2.7. Mass spectrometry analysis The standard protein digest was analyzed by MALDI-TOF/MS in reflector positive mode with a pulsed Nd/YAG laser at 384 nm on AB Sciex 4800 plus MALDI TOF/TOF mass spectrometer (AB Sciex, CA). In order to ensure the effectiveness of the spectrum and reduce the error interference, including instrumental error and operation error, two duplicate samples were spotted on the MALDI plant (twice for each) to get average spectra (n = 4). Each spectrum was obtained by the superposition of 100 shots with relative standard deviation of S/N below 5%. 0.5 ␮L elution was dropped on the MALDI plant, and equivalent matrix was dropped onto dried plant subsequently. The 2,5-dihydroxybenzoic acid (DHB) was dissolved in ACN-H2 O-H3 PO4 (70: 29: 1, v/v/v, 25 mg mL−1 ) as the matrix. The enriched glycopeptides from mouse liver was deglycosylated and lyophilizated for further Nano LC–MS/MS analysis by using a LTQ Orbitrap Velos liner ion trap mass spectrometer (Thermal, San Jose, CA) combined with Thermal ACCELA 600 liquid chromatograph. The deglycosylated peptides were re-dissolved in FA/H2 O (0.1: 99.9, v/v) and loaded on trap column (200 ␮m i.d.) packed with C18 AQ beads (5 ␮m, 120 Å, Daison, Osaka, Japan). In order to separate the peptides, the capillary analysis column (75 ␮m i.d.)

25

filled with C18 AQ beads (3 ␮m, 120 Å, Daison, Osaka, Japan) was used. The gradient elution parameters of HPLC were set as follow: A phase and B phase were 0.1% FA-H2 O and 0.1% FA-ACN respectively, and B phase linearly raised from 5% to 35% within 150 min. The LTQ-Orbitrap Velos mass spectrometer adopted positive ion mode detection and the ESI needle injection voltage was 2.0 kV. The flow rate was adjusted to about 200 nL min−1 . The mass spectra were obtained in data-dependent collision induced dissociation (CID) mode, and the full scan mode at the mass range from 300 Da to 2000 Da with the resolution setting at 6000. 20 strongest ions were chosen for further CID with the collision energy value of 35% and the activation time of 10 ms. Dynamic exclusion is set as follows: repeat number 1, repeat duration of 30 s, exclusion list size 500 and exclusion duration 90 s. 2.8. Database searching All the raw files obtained from LTQ-Orbitrap Velos were analyzed by Mascot with the database Mouse Uniprot (updated on July 9, 2014). The parameters of data searching were set as follow: the trypsin was selected for the protein digest allowing two missed cleavages; the mass tolerance was 10 ppm for the precursor ions and 0.8 Da for the fragment ions. The cut off false discovery rate (FDR) for whole peptide identifications was limited below 1%. The amino acid sequences of peptides contains N#XS/T were regarded as N-linked glycopeptides. (N # is asparagine which presence + 0.98 Da variable modifications, X is any other amino acids except proline). 3. Results and discussion 3.1. Fabrication and characterization of Fe3 O4 @PGMA@Au-l-cys nanoparticles The protocol for the synthesis of Fe3 O4 @PGMA@Au-l-cys zwitterionic-hydrophilic nanoparticles is illustrated in Fig. 1. Firstly, Fe3 O4 nanoparticle was synthesized by a modified solvothermal reaction, and then modified with MPS to graft abundant active double bonds for the subsequent polymerization. Subsequently, a robust layer of PGMA was coated on Fe3 O4 @MPS core by one-step reflux-precipitation polymerization with GMA as the monomer and MBA as the cross-linker to form Fe3 O4 @PGMA. Thirdly, the epoxy groups of the polymer reacted with anhydrous ethylenediamine to obtain a hydrophilic material with amino groups through ring open reaction. Fourthly, the Au nanoparticles were in situ loaded onto the Fe3 O4 @PGMA polymer layer through the specific interaction between HAuCl4 and NH2 groups of the thick polymer layer. The thick polymer with vast hydrophilic groups acted as reducing and stabilizing reagent. The high specific surface area of Au NPs in the polymer provide large activing sites to bond zwitterionic l-cysteine through the Au-S bond to form zwitterionic-hydrophilic material (Fe3 O4 @PGMA@Au-l-cys). To observe the morphology and characteristics of the as-prepared magnetic ZIC HILIC microspheres, various characterization methods were used. The representative TEM and FE-SEM images of Fe3 O4 , Fe3 O4 @PGMA, and Fe3 O4 @PGMA@Aul-cys nanoparticles are shown in Fig. 2. The TEM image of Fe3 O4 nanoparticle (Fig. 2a) shows that Fe3 O4 nanoparticles are uniform in both shape and size with an average diameter of ca. 280 nm. The TEM image and FE-SEM image of Fe3 O4 @PGMA show a well-defined smoother core-shell structure of Fe3 O4 @PGMA nanoparticle (Fig. 2c, d), clearly indicating that a cross-linked PGMA layer has been successfully coated on the surface of Fe3 O4 @MPS. Besides, the thickness of the polymer shell is around 150 nm, which could provide a large number of acting sites. Most importantly, the

26

Y. Zhao et al. / J. Chromatogr. A 1482 (2017) 23–31

Fig. 1. Schematic illustration of the fabrication procedure of ZIC-HILIC magnetic Fe3 O4 @PGMA@Au-l-cys material.

TEM and FE-SEM image of Fe3 O4 @PGMA@Au-l-cys (Fig. 2e, f) with an obvious Au nanoparticles proves the successful in situ reduction of Au. further prove the successful manufacture of To Fe3 O4 @PGMA@Au-l-cys nanoparticles, the FT-IR spectra of each products were analyzed. Compared to the FT-IR spectrum of Fe3 O4 (572 cm−1 , corresponding to Fe–O) in Fig. 3a, new characteristics absorption peaks (1720 cm−1 , stretching vibration of C O, 1530 cm−1 , N H bending vibrations) in the spectrum of Fe3 O4 @PGMA could demonstrate the successful coating of PGMA polymer shell on the surface of Fe3 O4 . The new characteristic peak appearing at 908 cm−1 is ascribed to the stretching vibration of epoxy group of GMA. Furthermore, the l-cysteine fingerprint could be found in the FT-IR spectrum of Fe3 O4 @PGMA@Au-l-cys, which indirectly suggests the successful in situ Au reduction and greatly coordinating of l-cysteine (Fig. 3a). Energy Dispersive Spectroscopy (EDS) results, as shown in Fig. 3b, proved the successful ring-opening reaction of epoxy group according to the emergence of N element. The result could also indirectly indicate the successful coating of PGMA. Compared with Fe3 O4 @PGMA-NH2 , Au and S elements could be found in the EDS spectrum of Fe3 O4 @PGMA@Au-l-cys. To further intuitively observe the distribution of Au and S elements in Fe3 O4 @PGMA@Au-l-cys, the TEM and the energy dispersive X-ray (EDX) mapping of Fe3 O4 @PGMA@Au-l-cys are shown in Fig. S1 (see Supporting information). Au and S elements were homogeneously distributed in the polymer layer, suggesting the successful synthesis of Fe3 O4 @PGMA@Au-l-cys. Furthermore, thermogravimetric analysis (TGA) was uesed to confirm the successful preparation processes. Comparing to the naked Fe3 O4 , Fig. 3c shows the weight loss of Fe3 O4 @PGMA and Fe3 O4 @PGMA@Au-l-cys are 56.12% and 60.41% demonstrating the great coating amount of PGMA and successful bonding of abundant l-cys. As another proof, magnetic hysteresis loop curves show the saturation magnetization changes of different materials. Compared with the naked magnetic nanoparticles with saturation magnetization value of 64.21 emu g−1 , after modifying with PGMA, the saturation magnetization value decreases to 15.34 emu g−1 (Fig. 3d). Although the final microsphere has a low saturation magnetization value of 11.54 emu g−1 (Fig. 3d), it is strong enough for magnetic separation from the complex sample within 1 min.

3.2. Selective enrichment of glycopeptides from the tryptic digest of standard glycoproteins A typical standard protein digest (immunoglobulin G, IgG) was used to evaluate the enrichment capability of Fe3 O4 @PGMA@Aul-cys for glycopeptides. The procedure of enriching glycopeptides is shown in Fig. S2 (see Supporting information). According to the MALDI mass spectrum (Fig. 4), only 4 glycopeptides were detected with inferior S/N and weak MS signal intensities due to interference of abundant non-glycopeptides in the IgG tryptic digest, which severely suppressed the signal of low abundant glycopeptides. After being enriched by Fe3 O4 @PGMA@Au-l-cys nanoparticles, 28 glycopeptides were detected with S/N > 3 while the signals of non-glycosylated peptides were almost eliminated (Fig. 4b). The material just containing NH2 group (Fe3 O4 @PGMANH2 ) was also used for the enrichment of glycopeptides. Only 11 glycopeptides could be detected while some interference peaks of non-glycopeptides can still be observed in the mass spectrum, leading to baseline fluctuation and disturbance of glycopeptides. Apparently, Fe3 O4 @PGMA@Au-l-cys is more preeminent than Fe3 O4 @PGMA-NH2 for the glycopeptides enrichment owning to stronger hydrophilic interactions provided by plenty of zwitterionic groups anchored Au nanoparticles. More particular information of glycopeptides was listed in Tables S1 and S2, respectively (see Supporting information). In addition, after enriched by Fe3 O4 @PGMA@Au-l-cys, the glycopeptides in the eluent (shown in Fig. 4b) were deglycosylated by the PNGase F and detected by MALDI-TOF MS (Fig. 4d). There are only two primary peaks, EEQFNSTFR (m/z = 1158.54) and EEQYNSTYR (m/z = 1190.53), were found, which proved that all the peaks in Fig. 4b were N-linked glycopeptides. All the results manifest that great density of zwitterionic groups is crucially significant for enrichment of glycopeptides. To further appraise the universality of Fe3 O4 @PGMA@Au-l-cys for selective enrichment of glycopeptides, digests of horse radish peroxidase (HRP) and chicken avidin were used as standard protein digest samples. Only 3 glycopeptides could be detected in the direct analysis of HRP digest by MALDI-TOF MS while the signal of glycopeptides were greatly restrained by non-glycopeptides. After enrichment by Fe3 O4 @PGMA@Au-l-cys, 15 glycopeptides could be detected with the S/N > 10 (Fig. 5b). Furthermore, comparing to the direct analysis of chicken avidin digest, 16 glycopeptides with

Y. Zhao et al. / J. Chromatogr. A 1482 (2017) 23–31

27

Fig. 2. (a) TEM and (b) FE-SEM images of naked Fe3 O4 cores. (c) TEM and (d) FE-SEM images of Fe3 O4 @PGMA nanoparticles. (e) TEM and (f) FE-SEM images of Fe3 O4 @PGMA@Aul-cys nanoparticles.

strong MS signal intensities and high signal-to-noise ratio (S/N > 10) were detected. The results confirmed that Fe3 O4 @PGMA@Au-l-cys have the excellent performance in enrichment of glycopeptides. More particular information of glycopeptides were listed in Tables S3 and S4 (see Supporting information). 3.3. Evaluation of the enrichment sensitivity, enrichment capacity and enrichment recovery of Fe3 O4 @PGMA@Au-l-cys in glycopeptides enrichment On account of ultra low content of glycopeptides in the complicated biological samples, the detection sensitivity of Fe3 O4 @PGMA@Au-l-cys toward glyopeptides was evaluated. The lowest human IgG digest amount, which can still ensure detection of the six highest glycopeptides, was determined as sensitivity. 50, 10 and 5 fmol tryptic digests of IgG were applied to evaluate the enrichment limit of Fe3 O4 @PGMA@Au-l-cys. As presented in Fig. 6a, several high peaks with strong signal intensities were found in 50 fmol of IgG tryptic digest. When the amount of IgG digest was 10 fmol, 6 glycopeptides could be detected, though the signal slightly decreased. Even when the amount of IgG digest was as low as 5 fmol, six glycopeptides can still be detected with the

lowest S/N ratio of 10.21 (m/z = 2958.10). Mass spectrum with only one glycopeptide was shown in Fig. S3c (0.1 fmol, see Supporting information). The detection limit of Fe3 O4 @PGMA@Au-l-cys was lower than PEG brushes hydrophilic material (10 fmol) [20], zwitterionic polymer brushes hybrid silica nanoparticles (10 fmol) [12] and hydrophilic silica-based click maltose (30 fmol) [40]. The high detection sensitivity ascribes to the numerous zwitterionic groups. This result shows the great application prospect of Fe3 O4 @PGMA@Au-l-cys for the enrichment of glycopeptides from intricate samples. The enrichment capacity of Fe3 O4 @PGMA@Au-l-cys was evaluated by incubating different amount of materials with 3 ␮g IgG digest, and the eluates were detected by MALDI-TOF MS. Then the capacity was estimated by the signal intensities of six highest glycopeptides peaks. As shown in Fig. 7, with the decreasing amount of Fe3 O4 @PGMA@Au-l-cys, the signal strength of six highest glycopeptides were maintained at maximum value at first, then the signal intensities dropped sharply when the quantity of material was decrease to 40 mg. The binding capacity was calculated by the amount of human IgG digest (3 ␮g) divided by the amount of Fe3 O4 @PGMA@Au-l-cys. Therefore, to enrich total glycopeptides from 3 ␮g IgG digest, the minimum quantity of materials

28

Y. Zhao et al. / J. Chromatogr. A 1482 (2017) 23–31

Fig. 3. (a) FT-IR spectra and (c) TGA curves of (i) Fe3 O4 , (ii) Fe3 O4 @PGMA and (iii) Fe3 O4 @PGMA@Au-l-cys, (b) EDS spectrums of (i) Fe3 O4 @MPS, (ii) Fe3 O4 @PGMA-NH2 and Fe3 O4 @PGMA@Au-l-cys. (d) Magnetization hysteresis loops of (i) Fe3 O4 , (ii) Fe3 O4 @PGMA and (iii) Fe3 O4 @PGMA@Au-l-cys.

was 40 mg. One gram of Fe3 O4 @PGMA@Au-l-cys could process 75 mg IgG digest. Therefore, the Fe3 O4 @PGMA@Au-l-cys enrichment capacity was 75 mg g−1 . The great bonding capacity was attribute to the synergistic effect of hydrophilic polymer chain and amphoteric groups which could have an intense adsorption for the glycopeptides. The high enrichment selectivity of Fe3 O4 @PGMA@Au-l-cys was estimated by using IgG digest mixed with BSA tryptic digest acting as the high interferential samples at the mass ratio of 1:1, 1:50 and 1:100. As shown in Fig. 8, after enriched by Fe3 O4 @PGMA@Au-lcys, 26 glycopeptides were detected with high intensities whithout any non-glycosylated peptides interference when the ratio of IgG digest and BSA tryptic digest was 1:1. When the mass ratio was 1:50, the MALDI-TOF mass spectrum was extremely similar to that of 1:1. When the ratio was decreased to 1:100, the intensities of glycopeptides were not changed, and 23 glycopeptides detected at the ratio of 1:1 also appeared. The interference experiment illustrates that the Fe3 O4 @PGMA@Au-l-cys has decent performance in the enrichment of glycopeptides from the biology samples. The stable isotope dimethyl labeling was used to evaluate the enrichment recovery of Fe3 O4 @PGMA@Au-l-cys. According to the literature [41], 50 ␮g IgG digest was divided into two equal amounts, then labelled with HCHO (+28) and DCDO (+32), and heavy isotope marked IgG digest was processed according to the workflow in the supporting information. The eluent was mixed with light isotope marked digest. The mixture was re-enriched with Fe3 O4 @PGMA@Au-l-cys, and the final eluent was detected by MALDI-TOF MS. The recovery ratio was defined as the intensity proportion of heavy isotope marked glycopeptides divided by the light isotope marked glycopeptides. Six highest glycopeptides peaks in MS were used for calculation and the recovery ratios were all above 89.8% (Table 1), which suggests that the Fe3 O4 @PGMA@Au-l-cys

Table 1 Enrichment recoveries of N-linked glycopeptides from human IgG digest using Fe3 O4 @PGMA@Au-l-cys. No.

m/z

Recovery ± S.D (%, n = 3)

I6 I8 I13 I15 I21 I22

2601.02 2634.03 2764.08 2796.06 2926.09 2958.10

91.8 ± 4.6 94.0 ± 5.2 89.8 ± 2.9 107.0 ± 5.0 101.2 ± 3.5 101.0 ± 3.6

has satisfactory recovery for glycopeptides and is a decent material for glycopeptide enrichment. 3.4. Fe3 O4 @PGMA@Au-l-cys was applied in glycopeptides profiling from mouse liver To further confirm the practical enrichment performance of Fe3 O4 @PGMA@Au-l-cys in the enrichment of glycopeptides from complex biology sample, the digest of mouse liver proteins was enriched by this ZIC-HILIC material following the procedure in Fig. S2 (see Supporting information). The eluent was lyophilized, deglycosylated and analysed to Nano LC–MS/MS. As a result, 774 unique N-glycosylation sites from 411 N-glycosylated proteins were identified in three replicate analyses (Table S5, see Supporting information). The results were close to the literatures, such as magnetic nanoparticles coated with multilayer polysaccharide [17] and zwitterionic polymer-coated core–shell magnetic nanoparticles [36]. The concise data of three independent runs are shown in Fig. S4 (see Supporting information). The results show that the Fe3 O4 @PGMA@Au-l-cys has a great performance in the analysis

Y. Zhao et al. / J. Chromatogr. A 1482 (2017) 23–31

29

Fig. 4. MALDI-TOF MS spectra of (a) direct analysis of 0.5 pmol tryptic digest of human IgG; (b) after enrichment by Fe3 O4 @PGMA@Au-l-cys; (c) after enrichment by Fe3 O4 @PGMA-NH2 ; and (d) deglycosylated peptides of glycopetides shown in (b) processed by PNGase F.

Fig. 5. MALDI-TOF MS spectra of (a) direct analysis of 1 pmol tryptic digest of HRP; (b) after enrichment by Fe3 O4 @PGMA@Au-l-cys; (c) direct analysis of 1 pmol tryptic digest of chicken avidin and (d) after enrichment by Fe3 O4 @PGMA@Au-l-cys.

30

Y. Zhao et al. / J. Chromatogr. A 1482 (2017) 23–31

Fig. 6. MALDI-TOF MS spectra of (a) 50 fmol (0.5 ␮L), (b) 10 fmol (0.5 ␮L) and (c) 5 fmol (0.5 ␮L) human IgG digest after treatment by Fe3 O4 @PGMA@Au-l-cys nanoparticles.

Fig. 8. MALDI-TOF-MS analysis of tryptic IgG and BSA after enrichment. The mass ratio of digest IgG and BSA are 1:1 (a), 1:50 (b) and 1:100 (c).

endow the material with good enrichment ability for glycopeptides, including the high detection sensitivity, great enrichment capacity and satisfactory enrichment recovery. The practical application of Fe3 O4 @PGMA@Au-l-cys in enriching glycopeptides from mouse liver protein digest demonstrates the new ZIC-HILIC magnetic material has great practical value in the enrichment and identification of N-linked glycopeptides from complex biological sample. Acknowledgements We gratefully acknowledge the support of the National Natural Science Foundation of China (No.21475044), the National Key Scientific Instrument and Equipment Development Project (2012YQ120044), the State Key Program of National Science of China (21235005), the National Key Technology R&D Program(2015BAK44B00), and foundation of Shanghai Research Institute of Criminal Science and Technology.

Fig. 7. Intensities of six selected N-linked glycopeptides from human IgG digest after treatment with different amounts of Fe3 O4 @PGMA@Au-l-cys nanoparticles.

and profiling of low-abundance N-linked glycopeptides from complex biological sample. 4. Conclusion In summary, a new ZIC-HILCI magnetic material was synthesized by reflux precipitation polymerization and in situ reduction of Au. The thick polymer layer not only protect the inner core, but guarantee the large number of acting sites for in situ immobilization of Au and subsequently modification of hydrophilic l-cysteine. The excellent hydrophilicity and plentiful zwitterionic groups

Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.chroma.2016.12. 054. References [1] K. Ohtsubo, J.D. Marth, Glycosylation in cellular mechanisms of health and disease, Cell 126 (2006) 855–867. [2] G.W. Hart, R.J. Copeland, Glycomics hits the big time, Cell 143 (2010) 672–676. [3] D.N. Hebert, L. Lamriben, E.T. Powers, J.W. Kelly, The intrinsic and extrinsic effects of N-linked glycans on glycoproteostasis, Nat. Chem. Biol. 10 (2014) 902–910.

Y. Zhao et al. / J. Chromatogr. A 1482 (2017) 23–31 [4] P. Steffen, M. Kwiatkowski, W.D. Robertson, A. Zarrine-Afsar, D. Deterra, V. Richter, H. Schluter, Protein species as diagnostic markers, J. Proteomics 134 (2016) 5–18. [5] A.C. Kolbl, U. Andergassen, U. Jeschke, The role of glycosylation in Breast cancer metastasis and cancer control, Front. Oncol. 5 (2015) 1–5. [6] Y. Miura, T. Endo, Glycomics and glycoproteomics focused on aging and age-related diseases − Glycans as a potential biomarker for physiological alterations, Biochim. Biophys. Acta 1860 (2016) 1608–1614. [7] D.F. Zielinska, F. Gnad, J.R. Wisniewski, M. Mann, Precision mapping of an in vivo N-glycoproteome reveals rigid topological and sequence constraints, Cell 141 (2010) 897–907. [8] Y. Pan, H. Bai, C. Ma, Y. Deng, W. Qin, X. Qian, Brush polymer modified and lectin immobilized core-shell microparticle for highly efficient glycoprotein/glycopeptide enrichment, Talanta 115 (2013) 842–848. [9] Y. Liu, D. Fu, Y. Xiao, Z. Guo, L. Yu, X. Liang, Anal. Methods 7 (2015) 25. [10] L. Liu, M. Yu, Y. Zhang, C. Wang, H. Lu, Hydrazide functionalized core-shell magnetic nanocomposites for highly specific enrichment of N-glycopeptides, ACS Appl. Mater. Interfaces 6 (2014) 7823–7832. [11] Y. Li, X. Zhang, C. Deng, Functionalized magnetic nanoparticles for sample preparation in proteomics and peptidomics analysis, Chem. Soc. Rev. 42 (2013) 8517–8539. [12] G. Huang, Z. Xiong, H. Qin, J. Zhu, Z. Sun, Y. Zhang, X. Peng, J. ou, H. Zou, Synthesis of zwitterionic polymer brushes hybrid silica nanoparticles via controlled polymerization for highly efficient enrichment of glycopeptides, Anal. Chim. Acta 809 (2014) 61–68. [13] R. Ma, J. Hu, Z. Cai, H. Ju, Facile synthesis of boronic acid-functionalized magnetic carbon nanotubes for highly specific enrichment of glycopeptides, Nanoscale 6 (2014) 3150–3156. [14] M. Wang, X. Zhang, C. Deng, Facile synthesis of magnetic poly(styrene-co-4-vinylbenzene-boronic acid) microspheres for selective enrichment of glycopeptides, Proteomics 15 (2015) 2158–2165. [15] L. Yu, X. Li, Z. Guo, X. Zhang, X. Liang, Hydrophilic interaction chromatography based enrichment of glycopeptides by using click maltose: a matrix with high selectivity and glycosylation heterogeneity coverage, Chemistry 15 (2009) 12618–12626. [16] W.H. Wang, J.L. Dong, G.L. Baker, M.L. Bruening, Bifunctional polymer brushes for low-bias enrichment of mono- and multi-phosphorylated peptides prior to mass spectrometry analysis, Analyst 136 (2011) 3595–3598. [17] Z. Xiong, H. Qin, H. Wan, G. Huang, Z. Zhang, J. Dong, L. Zhang, W. Zhang, H. Zou, Layer-by-layer assembly of multilayer polysaccharide coated magnetic nanoparticles for the selective enrichment of glycopeptides, Chem. Commun. 49 (2013) 9284–9286. [18] C. Bi, R. Jiang, X. He, L. Chen, Y. Zhang, Synthesis of a hydrophilic maltose functionalized Au NP/PDA/Fe3 O4 -RGO magnetic nanocomposite for the highly specific enrichment of glycopeptides, RSC Adv. 5 (2015) 59408–59416. [19] J.B. Per Ha1gglund, Felix Elortza, Ole Nørregaard Jensen, Peter Roepstorff, A new strategy for identification of N-glycosylated proteins and unambiguous assignment of their glycosylation sites using HILIC enrichment and partial deglycosylation, J. Proteome Res. 3 (2004) 556. [20] Z. Xiong, L. Zhao, F. Wang, J. Zhu, H. Qin, R. Wu, W. Zhang, H. Zou, Synthesis of branched PEG brushes hybrid hydrophilic magnetic nanoparticles for the selective enrichment of N-linked glycopeptides, Chem. Commun. 48 (2012) 8138–8140. [21] H. Wan, J. Huang, Z. Liu, J. Li, W. Zhang, H. Zou, A dendrimer-assisted magnetic graphene-silica hydrophilic composite for efficient and selective enrichment of glycopeptides from the complex sample, Chem. Commun. 51 (2015) 9391–9394. [22] X. Zou, J. Jie, B. Yang, A facile and cheap synthesis of zwitterion coatings of the CS@PGMA@IDA nanomaterial for highly specific enrichment of glycopeptides, Chem. Commun. 52 (2016) 3251–3253. [23] C. Bi, Y. Zhao, L. Shen, K. Zhang, X. He, L. Chen, Y. Zhang, Click synthesis of hydrophilic maltose-Functionalized iron oxide magnetic nanoparticles based on dopamine anchors for highly selective enrichment of glycopeptides, ACS Appl. Mater. Interfaces 7 (2015) 24670–24678.

31

[24] Y. Wang, J. Wang, M. Gao, X. Zhang, An ultra hydrophilic dendrimer-modified magnetic graphene with a polydopamine coating for the selective enrichment of glycopeptides, J. Mater. Chem. B 3 (2015) 8711–8716. [25] B. Jiang, Y. Liang, Q. Wu, H. Jiang, K. Yang, L. Zhang, Z. Liang, X. Peng, Y. Zhang, New GO-PEI-Au-L-Cys ZIC-HILIC composites: synthesis and selective enrichment of glycopeptides, Nanoscale 6 (2014) 5616–5619. [26] R. Wu, L. Li, C. Deng, Highly efficient and selective enrichment of glycopeptides using easily synthesized magG/PDA/Au/L-Cys composites, Proteomics 16 (2016) 1311–1320. [27] P. Dedvisitsakul, S. Jacobsen, B. Svensson, J. Bunkenborg, C. Finnie, P. Hagglund, Glycopeptide enrichment using a combination of ZIC-HILIC and cotton wool for exploring the glycoproteome of wheat flour albumins, J. Proteome Res. 13 (2014) 2696–2703. [28] H. Huang, Y. Jin, M. Xue, L. Yu, Q. Fu, Y. Ke, C. Chu, X. Liang, A novel click chitooligosaccharide for hydrophilic interaction liquid chromatography, Chem. Commun. 45 (2009) 6973–6975. [29] A. Shen, Z. Guo, L. Yu, L. Cao, X. Liang, A novel zwitterionic HILIC stationary phase based on thiol-ene click chemistry between cysteine and vinyl silica, Chem. Commun. 47 (2011) 4550–4552. [30] A.H. Lu, E.L. Salabas, F. Schuth, Magnetic nanoparticles: synthesis, protection, functionalization, and application, Angew. Chem. Int. Ed. Engl. 46 (2007) 1222–1244. [31] Y. Cao, L. Wen, F. Svec, T. Tan, Y. Lv, Magnetic AuNP@Fe3 O4 nanoparticles as reusable carriers for reversible enzyme immobilization, Chem. Eng. J. 286 (2016) 272–281. [32] L. Agiotis, I. Theodorakos, S. Samothrakitis, S. Papazoglou, I. Zergioti, Y.S. Raptis, Magnetic manipulation of superparamagnetic nanoparticles in a microfluidic system for drug delivery applications, J. Magn. Magn. Mater. 401 (2016) 956–964. [33] D. Ling, N. Lee, T. Hyeon, Chemical synthesis and assembly of uniformly sized iron oxide nanoparticles for medical applications, Acc. Chem. Res. 48 (2015) 1276–1285. [34] Y. Pan, X. Du, F. Zhao, B. Xu, Magnetic nanoparticles for the manipulation of proteins and cells, Chem. Soc. Rev. 41 (2012) 2912–2942. [35] Z. Xiong, Y. Ji, C. Fang, Q. Zhang, L. Zhang, M. Ye, W. Zhang, H. Zou, Facile preparation of core-shell magnetic metal-organic framework nanospheres for the selective enrichment of endogenous peptides, Chemistry 20 (2014) 7389–7395. [36] Y. Chen, Z. Xiong, L. Zhang, J. Zhao, Q. Zhang, L. Peng, W. Zhang, M. Ye, H. Zou, Facile synthesis of zwitterionic polymer-coated core-shell magnetic nanoparticles for highly specific capture of N-linked glycopeptides, Nanoscale 7 (2015) 3100–3108. [37] B. Jiang, Q. Wu, N. Deng, Y. Chen, L. Zhang, Z. Liang, Y. Zhang, Hydrophilic GO/Fe3 O4 /Au/PEG nanocomposites for highly selective enrichment of glycopeptides, Nanoscale 8 (2016) 4894–4897. [38] J. Li, F. Wang, J. Liu, Z. Xiong, G. Huang, H. Wan, Z. Liu, K. Cheng, H. Zou, Functionalizing with glycopeptide dendrimers significantly enhances the hydrophilicity of the magnetic nanoparticles, Chem. Commun. 51 (2015) 4093–4096. [39] L. Shangguan, L. Zhang, Z. Xiong, J. Ren, R. Zhang, F. Gao, W. Zhang, Investigation of bi-enzymatic reactor based on hybrid monolith with nanoparticles embedded and its proteolytic characteristics, J. Chromatogr. A 1388 (2015) 158–166. [40] J. Zhu, F. Wang, R. Chen, K. Cheng, B. Xu, Z. Guo, X. Liang, M. Ye, H. Zou, Centrifugation assisted microreactor enables facile integration of trypsin digestion hydrophilic interaction chromatography enrichment, and on-column deglycosylation for rapid and sensitive N-glycoproteome analysis, Anal. Chem. 84 (2012) 5146–5153. [41] P.J. Boersema, R. Raijmakers, S. Lemeer, S. Mohammed, A.J.R. Heck, Multiplex peptide stable isotope dimethyl labeling for quantitative proteomics, Nat. Protoc. 4 (2009) 484–494.