Facile synthesis of hydrophilic magnetic graphene nanocomposites via dopamine self-polymerization and Michael addition for selective enrichment of N-linked glycopeptides

The development of methods to effectively capture N-glycopeptides from the complex biological samples is crucial to N-glycoproteome profiling. Herein, the hydrophilic chitosan–functionalized magnetic graphene nanocomposites (denoted as Fe3O4-GO@PDA-Chitosan) were designed and synthesized via a simple two-step modification (dopamine self-polymerization and Michael addition). The Fe3O4-GO@PDA-Chitosan nanocomposites exhibited good performances with low detection limit (0.4 fmol·μL−1), good selectivity (mixture of bovine serum albumin and horseradish peroxidase tryptic digests at a molar ration of 10:1), good repeatability (4 times), high binding capacity (75 mg·g−1). Moreover, Fe3O4-GO@PDA-Chitosan nanocomposites were further utilized to selectively enrich glycopeptides from human renal mesangial cell (HRMC, 200 μg) tryptic digest, and 393 N-linked glycopeptides, representing 195 different glycoproteins and 458 glycosylation sites were identified. This study provides a feasible strategy for the surface functionalized novel materials for isolation and enrichment of N-glycopeptides.

method [16][17][18] , hydrophilic interaction liquid chromatography (HILIC) [19][20][21] and so on. Among them, owning to good MS compatibility, excellent reproducibility, unbiased enrichment performance and simple operating process, HILIC approach, based on the hydrophilicity differences between glycopeptides and non-glycopeptides, is widely adopted in glycopeptides enrichment. Chitosan is a cationic polymer obtained by deacetylation of chitin, and has gained more attention as drug delivery carriers owning to its bio-safety, biocompatibility and biodegradability 22 . The polar groups (-OH, -NH 2 ) endowed chitosan with selectivity towards glycopeptides through hydrogen bond between the glycan moieties and -OH/-NH 2 . Chitosan microspheres as adsorbent for isolation need an inconvenient and time-consuming centrifugation separation 23 . In recent years, magnetic separation has received extensive attention because of its better separation efficiency in contrast to the traditional approach. The combination of HILIC enrichment strategy and magnetic separation technology realized the rapid and efficient enrichment and separation of glycoproteins/glycopeptides. Li et al. fabricated Fe 3 O 4 @G6P magnetic microspheres for specific capture of N-linked glycopeptides 24 . The hydrophilic glucose-6-phosphate immobilized on the Fe 3 O 4 nanoparticles endowed the microspheres with high selectivity and sensitivity. Fang et al. modified chitosan on magnetic Fe 3 O 4 nanoparticles 25 . The HILIC microspheres showed the strong ability of fast magnetic separation and recognition toward glycopeptides. Notwithstanding these excellent reports, the synthesis of novel magnetic materials is still attracting extensive attentions aimed at optimizing the enrichment selectivity and sensitivity towards glycoproteins/glycopeptides. Magnetic Fe 3 O 4 nanoparticle-decorated GO (Fe 3 O 4 -GO) has magnetic responsibility and large surface area, and has been widely used for glycoproteins/glycopeptides enrichment 26,27 . Although some efficient strategies (self-assembly, Au-S bond) were applied to fabricate hydrophilic materials, they suffered from tedious operation 28,29 . Based on these, the assembly of chitosan coated Fe 3 O 4 -GO nanocomposites by a convenient Michael addition reaction would be very attractive.
Herein, a new type of chitosan-functionalized hydrophilic magnetic nanocomposites, Fe 3 O 4 -GO@ PDA-Chitosan ( Fig. 1) was assembled via dopamine self-polymerization and Michael addition. Briefly, Fe 3 O 4 NPs were interspersed on the surface of graphene oxide nanocomposites by solvothermal reaction. Polydopamine (PDA)-coated magnetic graphene was prepared via dopamine self-polymerization. Chitosan was grafted on the surface of Fe 3 O 4 -GO@PDA nanocomposites via Michael addition. The dopamine self-polymerization and Michael addition reaction occurred simply under mechanical agitation in Tris-HCl buffer. The polydopamine layer and chitosan on the surface of magnetic graphene could effectively enhance hydrophilic enrichment performance of magnetic nanocomposites for N-glycopeptides from the complex biological samples under an external magnetic field. This novel nanocomposite was applied to achieve good selectivity (mixture of bovine serum albumin and horseradish tryptic digests at a molar ration of 10:1), sensitivity (0.4 fmol·μL −1 ), binding capacity (75 mg·g −1 ) for N-glycopeptides enrichment.

experimental Section
Materials. Chitosan (with a deacetylation, degree of 98%) was purchased from Aladdin (Shanghai, China).
Horseradish peroxidase (HRP), immunoglobulin G (IgG), peptide-N4-(N-acetyl-β-D-glucosaminyl) asparagine amidase F (PNGase F), bovine serum albumin (BSA), and HPLC-grade acetonitrile (ACN) were purchased from Sigma-Aldrich (USA) (Beijing, China). Dithiothreitol (DTT), urea, ammonium bicarbonate (NH 4 HCO 3 ) and iodoacetamide (IAA) were obtained purchased from Solarbio (China). Trifluoroacetic acid (TFA) was purchased www.nature.com/scientificreports www.nature.com/scientificreports/ from J&K (Beijing, China). 2,5-Dihydroxybenzoic acid (DHB) was obtained from TCI (Japan) (Shanghai, China Characterization. The morphology, structure and performance of the synthesized nanocomposites were evaluated according to a previous report 30 . Transmission electron microscope (TEM) images were obtained on a JEM-2100F (Japan) transmission electron microscope. Fourier transform infrared (FT-IR) spectra (4000-400 cm −1 ) in KBr were recorded using the BRUKER TENSOR 27 Fourier transform infrared spectrophotometer. The crystal structure of the nanocomposites was performed on a Rigaku (Japan) D/max/2500 v/pc with nickel-filtered Cu Kα source. The XRD patterns were collected in the range 3 < 2θ < 80° at a scan rate of 4.0°/min. The X-ray photoelectron spectra were obtained on a Thermo Fischer (USA) ESCALAB 250Xi X-ray photoelectron spectrometer (XPS) with an Mg Kα anode (15 kV, 400 W) at a takeoff angle of 45°. The source X-rays were not filtered and the instrument was calibrated against the C 1 s band at 285 eV. The magnetic properties were analyzed with a LDJ9600-1 (USA) vibrating sample magnetometer (VSM). The hydrophilicity of the nanocomposites was revealed with contact angle analyzer JY-82B (Dingsheng, China). Zeta potential of the nanocomposites was measured by Brook haven ZETAPALS/BI-200SM (USA) at room temperature. Thermogravimetric analysis (TGA) was carried out in nitrogen atmosphere at a heating rate of 10 °C·min Degradation of chitosan by hydrogen peroxide. 5% H 2 O 2 (100 mL) was added dropwise into the solution of Chitosan (10 g) in 2% acetic acid (200 mL). The mixture was stirred magnetically at 60 °C for 8 h. Then the reaction solution was cooled and passed through two layers of filter paper. The aqueous phase obtained was retained and the pH was adjusted to 10.0 with 1 N NaOH, set aside for 2 h, and again filtrated through two layers of filter paper. After concentration under reduced pressure, the solution was diluted with methanol, set aside for 5 h. The white precipitation was collected via filtration and washed with methanol. The white powder was obtained after drying under vacuum. Digestion of proteins. 1 mg HRP (human IgG or BSA) was dissolved in 1 mL NH 4 HCO 3 solution (50 mM, pH 8.3) under sonication and denatured by boiling for 15 min in water bath. And then, the denatured proteins were reduced with 3.1 mg DTT at 37 °C for 2 h in water bath and alkylated by 7.2 mg IAA using a shaking table at room temperature in the dark for 40 min. The mixture was incubated with trypsin at an enzyme-protein ratio of 1:25 (w/w) at 37 °C for 16 h. The tryptic digests were stored at −20 °C for later use.
Proteins extracted from human renal mesangial cells (HRMC) were precipitated by trichloroacetic acid and collected by centrifugation. The pellet was dissolved in 100 mM of NH 4 HCO 3 . The proteins underwent reduction, alkylation, and enzymolysis in sequence. The resulting digests were desalted using Sep-pak C18 cartridges (Water Ltd., Elstree, UK), evaporated to dryness, and stored at −20 °C for later use.  (Fig. 2b,e). The molecular weight of degraded chitosan was determined to be ~1471 g·mol −1 by Gel Permeation Chromatography, so the modification of degraded chitosan on the surface of Fe 3 O 4 -GO@PDA nanocomposites was not detected. When the chitosan shell was formed on the surface of Fe 3 O 4 -GO@PDA, the obtained Fe 3 O 4 -GO@PDA-Chitosan displayed an obvious core-shell structure (Fig. 2c,f).

N-Glycopeptides enrichment under hydrophilic mode.
The Zeta potentials of the obtained nanocomposites were detected in acid aqueous solution (H 2 O/ TFA = 99.5:0.5, v/v). The zeta potential of Fe 3 O 4 -GO@PDA was 25.70 mV (Fig. 3), which indicated that there were abundant phenolic group and amino groups onto the surface of PDA layer. And after graft of chitosan on the surface of PDA layer, the zeta potential increased to 29.01 mV due to the addition of amine and hydroxyl groups.
The    Fig. 4. It can be found that 13.9% weight loss occurred for Fe 3 O 4 -GO (Fig. 4a) corresponding to the content of GO. And there were 39.7 and 43.3% weight loss for Fe 3 O 4 -GO@PDA (Fig. 4b) and Fe 3 O 4 -GO@PDA-Chitosan (Fig. 4c), respectively. From the data of TGA curves, the amount of immobilized chitosan onto the    www.nature.com/scientificreports www.nature.com/scientificreports/

Glycopeptide enrichment from standard proteins by Fe 3 o 4 -GO@PDA-Chitosan nanocomposites.
To select the optimal enrichment condition to glycopeptides, three different kinds of loading buffer (89% ACN/ H 2 O, 0.1% TFA; 89% ACN/H 2 O, 0.5% TFA; 89% ACN/H 2 O, 1% TFA) were utilized for investigate the effect of capturing glycopeptides from HRP digestion by Fe 3 O 4 -GO@PDA-Chitosan nanocomposites and the results were displayed in Fig. S3 (Electronical Supporting Materials). The performance of Fe 3 O 4 -GO@PDA-Chitosan nanocomposites for enrichment of N-glycopeptides was the best in the loading buffer (89% ACN/H 2 O, 0.5% TFA). By taking advantage of the optimized enrichment condition, Fe 3 O 4 -GO@PDA-Chitosan nanocomposites were applied to capture N-glycopeptides from standard HRP tryptic digest (50 fmol·μL −1 ). As shown in Fig. 7a, the direct analysis of HRP digest (50 fmol·μL −1 ) without enrichment step, the signal peaks of low abundance of N-glycopeptides were completely suppressed, no target analyte was detected. After enrichment by Fe 3 O 4 -GO@PDA-Chitosan nanocomposites, the number and signal intensity of glycopeptides were distinctly enhanced, meanwhile, three N-glycopeptides were detected and non-glycopeptides were completely removed (Fig. 7b, detailed information is listed in Table S1, Electronical Supporting Materials). And no glycopeptides peaks were found in the supernatant (Fig. 7c). For comparison, three N-glycopeptides were also identified after enrichment with Fe 3 O 4 -GO@PDA nanocomposites, but the signal intensity was weaker and the one signal peak of non-glycopeptide appeared in MS spectra (Fig. 7d).
The detection sensitivity of Fe 3 O 4 -GO@PDA-Chitosan (or Fe 3 O 4 -GO@PDA) nanocomposites for N-glycopeptides enrichment was evaluated with lower concentrations of HRP tryptic digests. When the concentration of HRP digestion decreased to 0.5 fmol·μL −1 , five N-glycopeptides were detected after enrichment with Fe 3 O 4 -GO@PDA-Chitosan nanocomposites, and four N-glycopeptides and one non-glycopeptide were detected after enrichment with Fe 3 O 4 -GO@PDA nanocomposites (Fig. 8a,b). Even at the concentration of 0.4 fmol·μL −1 , two N-glycopeptides were still observed after treatment of Fe 3 O 4 -GO@PDA-Chitosan nanocomposites (Fig. 8c). The enrichment selectivity of Fe 3 O 4 -GO@PDA-Chitosan nanocomposites was further investigated, by evaluating the capture performance for N-glycopeptides from the mixture of HRP and BSA tryptic digest with different molar ratios. When the molar ratio of digests mixture of HRP and BSA was 1:1 or 1:10, the MS signals of N-glycopeptides were completely suppressed in direct analysis due to the strong signal suppression from an amount of non-glycopeptides (Fig. 9a,c). By contrast, after enrichment by Fe 3 O 4 -GO@PDA-Chitosan nanocomposites, two N-glycopeptides derived from HRP were detected with no non-glycopeptide signals (Fig. 9b). The molar ratio of HRP and BSA was further increased even 1:10, an overwhelming majority of non-glycopeptides were removed and the peaks of N-glycopeptides still absolutely dominated MS spectra (Fig. 9d). The enrichment selectivity of Fe 3 O 4 -GO@PDA-Chitosan nanocomposites was higher than some reported hydrophilic nanomaterials functionalized with saccharides, such as Au NP-maltose/PDA/Fe 3 O 4 -RGO 29  To demonstrate the unbiased enrichment performance of Fe 3 O 4 -GO@PDA-Chitosan nanocomposites towards different N-glycan, it was applied to capture N-glycopeptides from IgG digest, which contains different glycoforms from HRP. In direct MS analysis without enrichment, no N-glycopeptides were detected on account www.nature.com/scientificreports www.nature.com/scientificreports/ of the strong interference from non-glycopeptides (Fig. 10a). Seven N-glycopeptides were identified after enrichment with Fe 3 O 4 -GO@PDA nanocomposites (Fig. 10b). For comparison, fifteen N-glycopeptides were detected after enrichment by Fe 3 O 4 -GO@PDA-Chitosan nanocomposites (Fig. 10c, detailed information is listed in Table S2, Electronical Supporting Materials), and the signal intensity was stronger in MS spectra. The eluted glycopeptides from Fe 3 O 4 -GO@PDA-Chitosan nanocomposites enrichment were deglycosylated by PNGase F and two strong signals of deamidated peptides (at m/z 1158.4 (EEQFN#STFR), 1190.4 (EEQYN#STFR)) were detected (Fig. 10d), which demonstrated that the enriched peptide fragments were N-glycopeptides.