Facile preparation of carbon-functionalized ordered magnetic mesoporous silica composites for highly selective enrichment of N-glycans

Abstract

Highly selective and efficient enrichment of glycans from complex biological samples is of great significance for the discovery and diagnosis of disease via identifying the related biomarkers. Mesoporous carbon materials were widely employed for the selective enrichment of glycans due to the strong interactions between carbon and glycans. In this study, novel carbon-functionalized ordered magnetic mesoporous silica composites (denoted as Fe₃O₄@3SiO₂@mSiO₂–C) with a core–shell structure, high carbon content, excellent hydrophilic property and unique magnetic character were designed and synthesized. A strategy involving CTAB as the mesoporous structure-directing agent and the carbon precursor during in situ carbonization was proposed. Besides, a compact silica layer with adequate thickness was essential to protect the magnetic core during the sulphonation process in further high-temperature calcination. As a result, the obtained Fe₃O₄@3SiO₂@mSiO₂–C composites exhibited a high carbon content (25%) with graphite structure, rapid magnetic separation (within 10 s), large pore volume (0.257 m³ g⁻¹), high BET surface area (269.14 m² g⁻¹) and a well-ordered mesostructure (3.39 nm). In addition, a strong magnetic response with a saturation magnetization value (59.8 emu g⁻¹) has been confirmed. By taking the advantage of the special interaction between the carbon and glycans, the size-exclusion ability and highly hydrophilic as well as unique magnetic properties, the Fe₃O₄@3SiO₂@mSiO₂–C composites exhibit satisfying enrichment ability in glycomic analysis. Better selectivity and efficiency than active carbon have been confirmed. Furthermore, 42 N-linked glycans with sufficient peak intensities were obtained from human serum after treatment with the Fe₃O₄@3SiO₂@mSiO₂–C composites, showing great potential as a tool for the enrichment and detection of glycans in biological samples.

---

1. Introduction

Since mesoporous silica materials were invented in the early 1990s,^1,2 mesoporous materials have been confirmed to have well-defined mesoporous structure, large surface area and narrow pore size distribution; as a result, they has been widely used in the fields of adsorbents, sensors, catalysts and nanodevices.^3–7 Enormous research enthusiasm has been triggered by carbon materials with regular mesoporous systems.^8–10 Due to their fascinating characteristics such as tunable pore size and mesostructure, exceptional thermal and chemical stability and strong hydrophobicity, great potential has been realised in a variety of applications such as drug delivery, energy storage and conversion, adsorption and separation.^11–13 It is particularly attractive to explore approaches to enhance and/or extend the properties of mesoporous carbon materials by forming nanocomposites. Therefore, various targeted applications has been designed based on cooperative and synergic effects between mesostructured carbon and other active nanoparticles.¹⁴

In recent years, magnetic separation has become an ideal separation technique by taking advantage of the strong magnetic response.^15,16 Functionalized magnetic materials with desirable building blocks or components have been commonly used in proteomic research.^17,18 Combining the superior properties of mesoporous carbon and the rapid magnetic responsivity of magnetic materials to build magnetic mesoporous carbon composites is a promising way for better separation and enrichment.^14,19

Several methods have been used to synthesize magnetic mesostructured carbon materials, such as thermal treatment of organometallic compounds, silica template etching processes and co-casting methods.^{1,2,6,8,20–23} The thermal treatment of organometallic compounds method is feasible for the majority metal catalyses, but the morphology and metal size are difficult to control. Meanwhile, most of the materials synthesized by this method lack a magnetic response.²⁴ The silica template etching method can solve the problems above, but the etching process may weaken mechanical strength. Moreover, the introduction of a carbon precursor involves multiple steps which are time-consuming and difficult to manipulate. The co-casting method is also faced with the problems of being cumbersome and having a weak magnetic response. As a result, the design and synthesis of magnetic mesoporous carbon materials with a facile, stable and convenient process is still attracting attention. Nevertheless, the complicated and time-consuming multistep procedures to obtain these materials is an ineluctable drawback to extensive use.²⁵ Therefore, a simpler approach to prepare ordered mesoporous carbon materials is highly desirable.^26,27

Glycosylation is a ubiquitous and essential protein post-translation modification which is involved in a mass of pathological and physiological process such as cell adhesion, cell growth and cell–cell recognition, etc.^28,29 Mass spectrometry is widely used in glycan profiling for glycosylation research. However, due to the interference of proteins and the low concentration of glycans, the glycan signals from complex biological samples are usually unacceptable. Therefore, it is of great significance to develop enrichment materials for the selective isolation of glycans digested from proteins.^30–32 Several media for selective enrichment of glycans have been reported. For instance, active carbon material is a promising medium based on the hydrophobic and polar interactions between carbon and glycans, however, complex proteins are still adsorbed due to the weak size-exclusion ability.^33–35 Thus functionalized mesoporous carbon materials triggered great interest for the selective enrichment of glycans.

Herein, a kind of magnetic core–shell composite with a carbon-functionalized mesoporous silica shell was synthesized using a system combining a surfactant-templated one-pot sol–gel method with the in situ carbonization strategy. A compact silica layer was introduced to protect the magnetic core during the sulfuric acid pretreatment which was indispensable for the further carbonization. Then a mesoporous shell was directly coated on the silica layer. The structure-directing agent CTAB was in situ carbonized on the inner surface of the mesoporous framework forming a graphitized carbon film. The as-prepared composites possessed several considerable merits for glycan enrichment, such as a well-ordered mesostructure with suitable pore size and high content of graphited carbon which can enrich glycan by polar interactions and strong hydrophilicity;³⁶ the high BET surface area can enhance the enrichment efficiency and the excellent magnetic response helps to achieve better isolation efficiency. By taking advantage of these characteristics, the novel composites were employed to enrich N-linked glycans from human serum samples with high efficiency and selectivity.

2. Experimental

2.1. Reagents and materials

Tetraethyl orthosilicate (TEOS, 99%), cetyltrimethyl ammonium bromide (CTAB), sulfuric acid (98%), sinapinic acid (SA), ammonium solution (25 wt%), 2,5-dihydroxybenzoic acid (2,5-DHB), bovine serum albumin (BSA) and standard glycoprotein (chicken ovalbumin) were purchased from Sigma (St. Louis, MO, USA). Urea, dithiothreitol (DTT) and iodoacetamide (IAA) were obtained from BioRad (Hercules, CA, USA). PNGase F was acquired from New England Biolab (Ipswich, MA). An ultrafiltration membrane with an MWCO of 10 kDa was acquired from Millipore (Bedford, MA). Human serum from healthy volunteers was provided by Dalian Medical University and stored at −80 °C before analysis. I confirm that all experiments were performed in compliance with the relevant laws and institutional guidelines of the Ethics Committee of the Hospital. And I confirm that I have received informed consent from the patients who provided the human samples. Acetonitrile (ACN) was purchased from Merck (Darmstadt, Germany). Deionized water used for all experiments was purified with a Milli-Q water system. All other chemicals were of analytical grade and purchased from Aladdin Corporation (Shanghai, China).

2.2. Apparatus and measurements

Transmission electron microscopy (TEM) was conducted on a JEOL 2000 EX electronic microscope with an accelerating voltage of 120 keV. Fourier transform infrared (FT-IR) spectroscopy characterization was conducted on a Thermo Nicolet 380 spectrometer using KBr pellets (Nicolet, Wisconsin, USA). The nitrogen adsorption–desorption measurement of Fe₃O₄@nSiO₂@mSiO₂–C was conducted at −196 °C (liquid nitrogen temperature) using a static-volumetric method on an ASAP 2010 (Micromeritics, USA). The pore diameter and distribution curves were calculated by the Barrett–Joyner–Halenda (BJH) method from the adsorption branch. The saturation magnetization curve was obtained at room temperature on a Physical Property Measurement System 9T (Quantum Design, San Diego, USA). The Raman spectra were obtained on a Via-Reflex with excitation from an argon ion laser (532 nm). Thermogravimetric analysis (TGA) was performed on a Netzsch STA 409 PC thermal analysis system (NETZSCH, Selb, Germany) under air flow. All MALDI-TOF-MS analysis results were achieved using an UltrafleXtreme MALDI-TOF/TOF System (Bruker Daltonics, German) equipped with a 1 kHz OptiBeamTM on-axis laser.

2.3. Preparation of compact silica layer coated magnetic composite

Firstly, magnetic composites were synthesized according to the method reported previously.³⁷ 200 mg of the prepared magnetic nanoparticles were dispersed in the solvents (160 mL ethanol, 40 mL deionized water, and 2 mL 25% ammonia solution) and sonicated for 0.5 h. Next, 1.5 mL TEOS was added into the flask drop-by-drop, and the mixture was mechanically stirred at room temperature for an additional 12 h. The above process was repeated three times to form a dense and thick shell encapsulating the magnetic core. Then, the obtained material was washed with deionized water and ethanol three times by magnetic separation and dried under vacuum at 60 °C for the next step of the experiment.

2.4. Synthesis of ordered mesoporous silica coated magnetic composite

This was conducted according to the reported methods of Deng’s group.^38,39 By using the obtained acidulated Fe₃O₄@nSiO₂ composites as supporter, a mesoporous silica layer was coated on the surface of the Fe₃O₄@nSiO₂ composites through a surfactant-templated one-pot sol–gel method. First of all, 57 mg acidulated Fe₃O₄@nSiO₂ composites and 370 mg CTAB were dispersed in the solvent (56 mL ethanol, 94 mL deionized water, and 1.2 mL 25% ammonia solution) and then mechanically stirred for 0.5 h in a three-necked bottle. Secondly, 800 μL TEOS were added into the stabilized dispersion–solution dropwise and mechanically stirring continued for more 20 h at room temperature. Then, the obtained material (Fe₃O₄@nSiO₂@mSiO₂–CTAB) was washed with deionized water and ethanol for three times by magnetic separation and dried under vacuum at 60 °C overnight.

Afterwards, 50 mg obtained Fe₃O₄@nSiO₂@mSiO₂–CTAB composites and 500 mg (NH₄)₃NO₃ were dispersed in 50 mL ethanol, and then the mixture was mechanically stirred for 24 h at room temperature to eliminate CTAB. The obtained Fe₃O₄@nSiO₂@mSiO₂ composites was washed with deionized water and ethanol three times by magnetic separation and dried under vacuum at 60 °C.

2.5. Synthesis of carbon-functionalized ordered mesoporous silica coated magnetic composites

The as-prepared Fe₃O₄@nSiO₂@mSiO₂–CTAB composites were dispersed in an acidic solution including 12 mL of deionized water and 500 μL of concentrated sulfuric acid (98 wt%), followed by mechanical stirring at room temperature for 30 min. Then, the mixed solution was heat-treated at 100 °C for 12 h; afterwards, the temperature was increased to 160 °C for an additional 12 h heat-treatment under an air atmosphere. In the end, the sulfuric acid pretreated compound was calcined at 300 °C (3 °C min⁻¹) for 3 h and then 700 °C (3 °C min⁻¹) for 3 h under a nitrogen atmosphere to obtain the final product.

2.6. Preparation of protein digests

1 mg of chicken ovalbumin (OVA) was dispersed in 1 mL 25 mM ammonium bicarbonate buffer at pH 7.5. Then, the mixed solution was boiled for 6 minutes to cause protein denaturation. Afterwards, the PNGase F (10 U) was added into 100 μL of the mixed solution which had been denatured and then incubated at 37 °C for 24 h.

Before enzymolysis, human serum was centrifuged at 12 000 r for 10 min. The obtained supernatant (50 μL) was mixed with ammonium bicarbonate (25 mM, pH 7.5, 450 μL) and denatured in a boiling water bath for 5 min. Then, an ultrafiltration membrane was used to filter out the endogenous peptides at 12 000 r for 20 min. The obtained deposition was washed by ammonium bicarbonate (200 μL) three times, and dissolved in ammonium bicarbonate (25 mM, pH 7.5, 500 μL). The enzymolytic processing of human serum was the same as for OVA.

2.7. Selective enrichment of glycans from biological samples

Fe₃O₄@nSiO₂@mSiO₂–C composites (10 mg mL⁻¹, 80 μL) were added into a 20 μL digest solution (ovalbumin or human serum), then a specified volume of deionized water was added to make a 200 μL total volume, which was incubated for 60 min. After removal of the supernatant by magnetic separation, the deposit was washed with deionized water (100 μL) for three times. Finally, 20 μL 80% ACN was selected as the eluent for MS analysis. As a comparison, the Fe₃O₄@nSiO₂@mSiO₂ composites and active carbon materials were studied under the same conditions.

2.8. Mass spectrometric analysis

DHB (10 mg mL⁻¹ 2,5-DHB, 50% ACN–H₂O, 10 mM NaCl) was used as the matrix for the analysis of glycans. A sinapinic acid (saturated in 50% ACN–H₂O solution containing 0.1% FA) was used for the analysis of proteins. Sample aliquots (0.5 μL) were first placed on a plate, and then desiccated at room temperature, and the SA matrix (0.5 μL) was then added prior to MALDI-TOF-MS analysis.

3. Results and discussion

The Fe₃O₄@nSiO₂@mSiO₂–C composites were obtained using a system combining a surfactant-templated one-pot sol–gel method with the in situ carbonization strategy. The detailed synthetic approach for the fabrication of Fe₃O₄@nSiO₂@mSiO₂–C composites is shown in Scheme 1. Typically, in order to acquire the non-porous silica-layer-coated Fe₃O₄ composites (denoted as Fe₃O₄@nSiO₂), the as-prepared mono-disperse magnetic Fe₃O₄ nanoparticles were treated with a Stöber procedure.⁴⁰ The outer mesoporous silica framework was synthesized by a sol–gel method with TEOS as the silicon precursor, and CTAB as the structure-directing agent. After that, a carbon film was directly in situ carbonized on the inner surface of the mesoporous silica framework with CTAB as the carbon precursor.⁴¹ Importantly, the carbonization rate of CTAB might decrease since it is possible to degrade under high temperature, so the pretreatment of Fe₃O₄@3SiO₂@mSiO₂–CTAB composites before calcination is necessary.^40,42,43 In this work, a pretreatment of H₂SO₄ at a lower temperature before carbonization was undertaken to improve the carbonization rate of CTAB. Compared with the conventional syntheses of mesoporous carbon materials, the novel method simplified the removal of the structure-directing agent and the carbon-source addition process, which were time-consuming, difficult to operate and could cause collapse of the structure.

Scheme 1 Illustration of the synthesis procedure for carbon-functionalized ordered magnetic mesoporous silica composite.

The sulfuric acid pretreatment is likely to destroy the magnetic core due to its strong corrosiveness. As a result, it is worth mentioning that the formation of the compact silica was the critical step in the preparation of core–shell structure. As shown in Fig. 1, materials coated with compact silica layers 1, 2 or 3 times, corresponding to thicknesses of 15 nm, 25 nm, 45 nm (roughly estimated by TEM images), were assembled separately. It was found that the material with a silica layer thickness of only 15 nm was spoiled severely when treated with sulfuric acid. The material with a silica layer thickness of 25 nm thick silica layer spoiled to a smaller extent when treated with sulfuric acid directly, but corroded almost completely after coating with the mesoporous shell and pretreatment with sulfuric acid before carbonization. Encouragingly, the composite with a silica layer thickness of 45 nm was seldom spoiled after the whole sulphonation processes. It could be speculated that during the sulphonation process, accompanied by water evaporation at high temperature, the concentration of H₂SO₄ solution became higher and higher. As a result, the condensed ionized H⁺ from H₂SO₄ and the active mobility of the H₂SO₄ solution might lead to destruction of the magnetic core. In this case, since the reaction between H₂SO₄ and the magnetic core is impeded by a silica layer with a proper thickness, the magnetic core could remain protected and the C₁₆-alky chain was sulphonated more completely.

Fig. 1 TEM images of the magnetic core (a); the products coated with compact silica layer 1 (b) or 2 (c) times after sulphonation; the magnetic core coated with a 45 nm thick compact layer (d); the final obtained core–shell magnetic mesoporous carbon composites (e and f).

In order to evaluate the surplus saturation magnetization values of the magnetic material in each operation, the room-temperature magnetization curve of the material was recorded as shown in Fig. 2. Magnetic measurement shows that the magnetization values of Fe₃O₄, Fe₃O₄@3SiO₂@mSiO₂–C and Fe₃O₄@1SiO₂@mSiO₂–C are 80.1, 59.8 and 13.8 emu g⁻¹, respectively. The magnetization value of the material coated only once with a compact silica layer is much weaker than that of the pure magnetic core. This result further indicates that a too-thin compact layer failed to protect the magnetic core. In striking contrast, after sulfonation treatment, the as-prepared Fe₃O₄@3SiO₂@mSiO₂–C composites still show a strong magnetization value which is extremely close to the magnetization value of the pure magnetic core. This property suggesting a good suitability for magnetic separation.³⁶ The desirable performance benefits from the compact silica layer with an appropriate thickness which protected the magnetic core from H₂SO₄ etching; meanwhile, the thickness of the layer is not too thick to interfere with the magnetization responses. In addition, the in situ carbonization method retained the silica template, and this strategy avoided a template removal process, which may impact the structure and magnetic responses of the composites.

Fig. 2 Saturation magnetization values of the pure magnetic core (a), Fe₃O₄@3SiO₂@mSiO₂–C (b) and Fe₃O₄@1SiO₂@mSiO₂–C (c).

Fe₃O₄@3SiO₂@mSiO₂–C still shows an excellent dispersibility in aqueous solution (Fig. 3) despite the hydrophobic carbon film which was coated on the inner surface of the mesopores. This benefits from the hydrophilic groups exposed on the surface of silica framework which was retained completely during the whole synthetic progress, and there are abundant hydrophilic hydroxyl groups exposed on it as well. Conventional mesoporous carbon materials are poorly soluble in aqueous solution,⁴⁴ which might result in serious agglomeration, slow interaction, and severe aggregation and adhesion on the tube. The excellent dispersibility in aqueous solutions of Fe₃O₄@3SiO₂@mSiO₂–C can eliminate the defects efficiently so that the field of application is enlarged.

Fig. 3 Fe₃O₄@3SiO₂@mSiO₂–C composites dispersed in water and magnetically separated.

Compared with carbon-based mesoporous materials, silica-based mesoporous materials have a much better mechanical strength. This may have some adverse effects during the separation processes, such as ultrasonic treatment or mechanical stirring. The retention of the silica framework can enhance the mechanical strength of the material efficiently which is crucial for application under high pressure or high strength.

To elucidate the form of the interstitial structure in Fe₃O₄@3SiO₂@mSiO₂–C, nitrogen sorption measurements were taken (Fig. 4). The abrupt increase of P/P₀ from 0.40 to 0.80, suggests well-ordered mesoporous pore size distribution. It is estimated that the Brunauer–Emmett–Teller (BET) surface area of the composites r is 269.14 m² g⁻¹.⁴⁵ The pore diameter aperture distribution curve according to the Barrett–Joyner–Halenda (BJH) model indicates that the composites have a large pore volume of 0.257 m³ g⁻¹. In addition, a well-ordered mesoporous pore structure with a narrow pore-size distribution centered at 3.39 nm was confirmed.

Fig. 4 Nitrogen adsorption–desorption isotherms and pore size distribution of Fe₃O₄@3SiO₂@mSiO₂–C composites.

The carbon content and the morphology of carbon were investigated by TGA, FT-IR spectra and Raman spectra. The TGA and DTG curves of Fe₃O₄@3SiO₂@mSiO₂ and Fe₃O₄@3SiO₂@mSiO₂–C are shown in Fig. 5. The maximum weight loss of Fe₃O₄@3SiO₂@mSiO₂–C (about 25%) can be observed between 400 and 500 °C. It indicates that the carbon content of Fe₃O₄@3SiO₂@mSiO₂–C reached 25%. The DTG curve shows maximum weight loss temperatures of Fe₃O₄@3SiO₂@mSiO₂ (180 °C) and Fe₃O₄@3SiO₂@mSiO₂–C (450 °C). It is interesting to note that the maximum weight loss temperature increased about 270 °C after in situ carbonization, and that might be due to the existing form of carbon changing substantially.

Fig. 5 The TGA and DTG curves of Fe₃O₄@3SiO₂@mSiO₂ (a) and Fe₃O₄@3SiO₂@mSiO₂–C (b).

The FT-IR spectra of Fe₃O₄@3SiO₂@mSiO₂–CTAB and Fe₃O₄@3SiO₂@mSiO₂–C were compared in Fig. 6. The strong peaks of C–H_x (2925 cm⁻¹, 2983 cm⁻¹) become much weaker after in situ carbonization. Meanwhile, the intensity of C=C (from 1620 cm⁻¹ to 1680 cm⁻¹) was enhanced, which means the C₁₆-alkyl chains of CTAB were transferred to an aromatic nucleus structure with π–π bonds after the pretreatment with sulphuric acid and carbonization.⁴⁶

Fig. 6 The FT-IR spectra of Fe₃O₄@3SiO₂@mSiO₂–CTAB (a) and Fe₃O₄@3SiO₂@mSiO₂–C (b).

As characterized by Raman spectroscopy (Fig. 7), two peaks around 1340 cm⁻¹ (D-mode) and 1580 cm⁻¹ (G-mode) were observed. The G-mode is characteristic of alkene stretching vibrations and suggests the formation of C=C linkages because the dehydration reaction is catalyzed by sulphuric acid.⁴⁶ Simultaneously, the peak of the G-mode is much higher than that of the D-mode, further reflecting the high graphitic crystallinity of the carbon content in Fe₃O₄@3SiO₂@mSiO₂–C.⁴⁰ The obvious change of the carbon from the chain to the graphitized form could result in highly enhanced hydrophobicity and thermostability, and coincide with the results of thermogravimetric analysis.

Fig. 7 Raman spectrum of Fe₃O₄@3SiO₂@mSiO₂–C.

The characteristic result indicates that the well-ordered mesoporous structure has a perfect cutoff size to exclude most highly abundant proteins (such as HSA; 67 kDa, 5 nm × 7 nm × 7 nm).⁴⁷ In addition, the high content of graphitized carbon signifies considerable efficiency in the enrichment of glycans by their hydrophobic and polar interactions with carbon.^33,34 So it is of great potential to enrich selectively the glycans and efficiently size-exclude the highly abundant large proteins in biological fluid digests.

On the basis of the high carbon content, the outstanding dispersibility in aqueous solution, the strong magnetic response and the well-ordered mesoporous structure, Fe₃O₄@3SiO₂@mSiO₂–C was adopted for the development of an effective separation and enrichment approach, aimed at enriching the low abundance, low molecular weight N-linked glycans in protein digests and human serum. The polar interactions between carbon and glycans, the size exclusion of high molecular weight proteins by the mesopores, and the rapid separation of the magnetic microsphere would be conducive to enhancing the enrichment efficiency.³⁶

In order to inspect the feasibility of N-linked glycan enrichment, the Fe₃O₄@3SiO₂@mSiO₂–C composites were applied to enrich the N-linked glycans in the ovalbumin digests (Fig. 8). Fig. 8a shows that a few N-linked glycans were detected with low intensities and signal-to-noise ratios before enrichment. After the enrichment with Fe₃O₄@3SiO₂@mSiO₂–C (Fig. 8c), no N-linked glycans can be detected. Furthermore, there was no protein signal in the eluant after enrichedment by Fe₃O₄@3SiO₂@mSiO₂–C (Fig. 8b). Compared with protein detection before enrichment (Fig. 8d), the signal intensity declined obviously. The above results suggest that Fe₃O₄@3SiO₂@mSiO₂–C composites are a possible medium for selective enrichment of N-linked glycans with excellent size-exclusion.

Fig. 8 MALDI-TOF MS analysis of N-glycans released from ovalbumin digests before enrichment (a) and the supernatant after (c) enrichment with Fe₃O₄@3SiO₂@mSiO₂–C; MAL-TOF MS analysis of proteins in ovalbumin digests before enrichment (d) and the elution after (b) enrichment with Fe₃O₄@3SiO₂@mSiO₂–C.

The high enrichment selectivity of Fe₃O₄@3SiO₂@mSiO₂–C was evaluated using a more complex sample containing a certain amount of ovalbumin digests with different amounts of BSA as the interference protein (Fig. 9). As shown in Fig. 9a, after enrichment by Fe₃O₄@3SiO₂@mSiO₂–C, there were 25 N-linked glycans detected with high signal intensities when the ratio of BSA to OVA was 0 : 1. After the extraction with active carbon (Fig. 9b), a total of 24 N-linked glycans were detected. The highest intensity of glycans enriched by Fe₃O₄@3SiO₂@mSiO₂–C reached 12 000, as much as twice of that enriched by active carbon. When the ratio increased to 10 : 1, only a few signal intensities decreased, and there were still 25 N-linked glycans that could be detected after enrichment by Fe₃O₄@3SiO₂@mSiO₂–C (Fig. 9c). In comparison, after being enriched by active carbon (Fig. 9d), only 21 N-linked glycan signals were observed in the MS spectrum (S/N > 3), and this might be due to the interference of protein. After enrichment with Fe₃O₄@3SiO₂@mSiO₂–C, there were still strong signal intensities of 25 N-linked glycans detected even though the ratio of BSA to ovalbumin increased to 50 : 1 (Fig. 9e). Activated carbon was also used for comparison (Fig. 9f); the signal intensity of the N-linked glycans decreased obviously, and only 19 N-linked glycans could be detected. The highest intensity of glycans enriched by Fe₃O₄@3SiO₂@mSiO₂–C is 7000, and less than 1200 of this was enriched by active carbon. As Fig. S1† shows, after enrichment with active carbon, there are still proteins detected in the eluent, which means the size-selectivity of active carbon is insufficient. Briefly, compared with active carbon materials, the intensity of N-linked glycans enriched by Fe₃O₄@3SiO₂@mSiO₂–C composites is much stronger, and since the content of interfering proteins increased, the more obvious superior size-exclusion is presented. This result illustrates that the Fe₃O₄@3SiO₂@mSiO₂–C composites have outstanding size-exclusion performance in eliminating the interference of large-size proteins, which is indispensable for glycan profiling and indicating the better size-selectivity of proteins.

Fig. 9 MALDI-TOF MS analysis of N-glycans released from ovalbumin digests of BSA (w/w) at 1 : 0, 1 : 10, 1 : 50 after enrichment with Fe₃O₄@3SiO₂@mSiO₂–C (a, c, e) and activated carbon (b, d, f). Marked peaks marked include the signals of N-linked glycans.

There are close connections between the glycan in human serum and many diseases;⁴⁸ therefore, the research to enrich selectively glycans in human serum for further analysis is of great scientific significance. As shown in Fig. S2,† almost no glycans can be detected before enrichment. Encouragingly, after the enrichment with Fe₃O₄@3SiO₂@mSiO₂–C composites, 41 N-linked glycans with obviously stronger signal intensities were detected (Fig. 10) in 200 μL human serum. The detailed structure of detected N-linked glycans was displayed in Table S1.† To evaluate the reusability of Fe₃O₄@3SiO₂@mSiO₂–C composites, the material was reused to enrich the glycans from human serum in the same way, and two groups were treated at the same time to evaluate the stability. As shown in Fig. S3,† the stability and reusability is excellent, the highest intensities are very close (4 × 10⁴) and also 41 glycans can be detected. All the above results suggest that the Fe₃O₄@3SiO₂@mSiO₂–C composites have excellent performance in N-linked glycan enrichment from complex bio-samples.

Fig. 10 MALDI-TOF MS analysis of N-linked glycan released from human serum mixture after enrichment by Fe₃O₄@3SiO₂@mSiO₂–C.

4. Conclusions

In summary, carbon-functionalized ordered magnetic mesoporous silica composites with a core–shell structure have been successful prepared by a surfactant-templated one-pot sol–gel method combined with a facile in situ carbonization strategy. A precisely controlled non-porous silica layer was coated previously to protect the magnetic core from being spoiled by H₂SO₄ and to help establish the outer mesoporous shell. A large pore volume (0.257 m³ g⁻¹), high BET surface area (269.14 m² g⁻¹) and a well-ordered mesostructure with a narrow pore-size distribution centered at 3.39 nm could be obtained. Moreover, a strong magnetic response with a saturation magnetization value of 59.8 emu g⁻¹ has been confirmed. The content of highly graphitized carbon structure reached 25% which is indispensable for glycan enrichment. Using these novel materials, 41 N-linked glycans from human serum were enriched with excellent selectivity and efficiency; also outstanding size-exclusion has been confirmed. It can be concluded that the obtained Fe₃O₄@3SiO₂@mSiO₂–C composites have great potential application in glycomics research.

Acknowledgements

The authors acknowledge funding support from the Scientific Instruments Special Projects (2012YQ09016703), the China State Key Basic Research Program Grant (2013CB-911203, 2012CB910601), the Creative Research Group Project of NSFC (21321064), and the Knowledge Innovation program of DICP to Hanfa Zou as well as the National Natural Sciences Foundation of China (No. 21175133, 21235006).

Notes and references

1. J. S. Beck, J. C. Vartuli, W. J. Roth, M. E. Leonowicz, C. T. Kresge, K. D. Schmitt, C. T. W. Chu, D. H. Olson, E. W. Sheppard, S. B. Mccullen, J. B. Higgins and J. L. Schlenker, J. Am. Chem. Soc., 1992, 114, 10834.
2. T. Yanagisawa, T. Shimizu, K. Kuroda and C. Kato, Bull. Chem. Soc. Jpn., 1990, 63, 988.
3. G. Yao, D. Qi, C. Deng and X. Zhang, J. Chromatogr. A, 2008, 1215, 82–91.
4. S. H. Joo, S. J. Choi, I. Oh, J. Kwak, Z. Liu, O. Terasaki and R. Ryoo, Nature, 2001, 414, 470.
5. P. Yang, T. Deng, D. Zhao, P. Feng, D. Pine, B. F. Chmelka, G. M. Whitesides and G. D. Stucky, Science, 1998, 282, 2244.
6. C. Liang, Z. Li and S. Dai, Angew. Chem., Int. Ed., 2008, 47, 3696.
7. A. Vinu, C. Streb, V. Murugesan and M. J. Hartmann, J. Phys. Chem. B, 2003, 107, 8297.
8. R. Ryoo, S. H. Joo and S. J. Jun, J. Phys. Chem. B, 1999, 103, 7743.
9. Y. Meng, D. Gu, F. Zhang, Y. Shi, H. Yang, Z. Li, C. Yu, B. Tu and D. Zhao, Angew. Chem., 2005, 117, 7215.
10. Y. Meng, D. Gu, F. Zhang, Y. Shi, L. Cheng, D. Feng, Z. Wu, Z. Chen, Y. Wan, A. Stein and D. Zhao, Chem. Mater., 2006, 18, 4447.
11. Z. Wu and D. Zhao, Chem. Commun., 2011, 47, 3332.
12. H. Zhou, S. Zhu, M. Hibino, I. Honma and M. Ichihara, Adv. Mater., 2003, 15, 2107.
13. H. Wan, H. Qin, Z. Xiong, W. Zhang and H. Zou, Nanoscale, 2014, 6, 8743.
14. Z. Wu, W. Li, P. A. Webley and D. Zhao, Adv. Mater., 2012, 24, 485.
15. W. Ma, L. Li, Y. Zhang, Q. An, L. You, J. Li, Y. Zhang, S. Xu, M. Yu, J. Guo, H. Lu and C. Wang, J. Mater. Chem., 2012, 22, 23981.
16. Z. Feng, S. Zhu, D. R. Martins de Godoi, A. C. Samia and D. Scherson, Anal. Chem., 2012, 84, 3764.
17. Z. Xiong, L. Zhang, C. Fang, Q. Zhang, Y. Ji, Z. Zhang, W. Zhang and H. Zou, J. Mater. Chem. B, 2014, 2, 4473.
18. Z. Xiong, Y. Ji, C. Fang, Q. Zhang, L. Zhang, M. Ye, W. Zhang and H. Zou, Chem.–Eur. J., 2014, 20, 7389.
19. J. Liu, S. Qiao, Q. Hu and G. Lu, Small, 2011, 7, 425.
20. T. Muraliganth, A. V. Murugan and A. Manthiram, Chem. Commun., 2009, 7360.
21. X. Dong, H. Chen, W. Zhao, X. Li and J. Shi, Chem. Mater., 2007, 19, 3484.
22. C. Wang, J. Wang and Z. Sheng, J. Phys. Chem. B, 2007, 111, 6303.
23. Q. Zhu, Q. Pan and F. Liu, J. Phys. Chem. C, 2011, 115, 17464.
24. J. Lee, S. Y. Lee, S. H. Park, H. S. Lee, J. H. Lee, B. Y. Jeong, S. E. Park and J. H. Chang, J. Mater. Chem. B, 2013, 1, 610.
25. M. Florent, C. Xue, D. Zhao and D. Goldfarb, Chem. Mater., 2012, 24, 383.
26. I. Moriguchi, A. Ozono, K. Mikuriya, Y. Teraoka, S. Kagawa and M. Kodama, Chem. Lett., 1999, 1171.
27. Z. Li, W. Yan and S. Dai, Carbon, 2004, 42, 767.
28. Y. Tian, Y. Zhou, S. Elliott, R. Aebersold and H. Zhang, Nat. Protoc., 2007, 2, 334.
29. A. Helenius and M. Aebi, Science, 2001, 291, 2364.
30. T. M. Block, M. A. Comunale, M. Lowman, L. F. Steel, P. R. Romano, C. Fimmel, B. C. Tennant, W. T. London, A. A. Evans, B. S. Blumberg, R. A. Dwek, T. S. Mattu and A. S. Mehta, Proc. Natl. Acad. Sci. U. S. A., 2005, 102, 779.
31. R. Goldman, H. W. Ressom, R. S. Varghese, L. Goldman, G. Bascug, C. A. Loffredo, M. Abdel-Hamid, I. Gouda, S. Ezzat, Z. Kyselova, Y. Mechref and M. V. Novotny, Clin. Cancer Res., 2009, 15, 1808.
32. P. A. Norton, M. A. Comunale, J. Krakover, L. Rodemich, N. Pirog, A. D’ Amelio, R. Philip, A. S. Mehta and T. M. Block, J. Cell. Biochem., 2008, 104, 136.
33. N. H. Packer, M. A. Lawson, D. R. Jardine and J. W. Redmond, Glycoconjugate J., 1998, 15, 737.
34. J. Q. Fan, A. Kondo, I. Kato and Y. C. Lee, Anal. Biochem., 1994, 219, 224.
35. M. L. A. de Leoz, H. J. An, S. Kronewitter, J. Kim, S. Beecroft, R. Vinall, S. Miyamoto, R. D. White, K. S. Lam and C. Lebrilla, Dis. Markers, 2008, 25, 243.
36. N. Sun, C. Deng, Y. Li and X. Zhang, Anal. Chem., 2014, 86, 2246.
37. L. Zhao, H. Qin, Z. Hu, Y. Zhang, R. Wu and H. Zou, Chem. Sci., 2012, 3, 2828.
38. P. Yin, Y. Wang, Y. Li, C. Deng, X. Zhang and P. Yang, Proteomics, 2012, 12, 2784.
39. P. Yin, N. Sun, C. Deng, Y. Li, X. Zhang and P. Yang, Proteomics, 2013, 13, 2243.
40. S. Gai, P. Yang, C. Li, W. Wang, Y. Dai, N. Niu and J. Lin, Adv. Funct. Mater., 2010, 20, 1166.
41. P. Valle-Vigón, M. Sevilla and A. B. Fuertes, Chem. Mater., 2010, 22, 2526.
42. S. Liu, H. Chen, X. Lu, C. Deng, X. Zhang and P. Yang, Angew. Chem., 2010, 122, 7719.
43. T. Ma, L. Liu and Z. Yuan, Chem. Soc. Rev., 2013, 42, 3977–4003.
44. H. Qin, L. Zhao, R. Li, R. Wu and H. Zou, Anal. Chem., 2011, 83, 7721–7728.
45. N. Sun, X. Zhang and C. Deng, Nanoscale, 2015, 7, 6487–6491.
46. G. Socrates, Infrared and Raman Characteristic Group Frequencies, Wiley, New York, 3rd edn, 2005.
47. H. Qin, P. Gao, F. Wang, L. Zhao, J. Zhu, A. Wang, T. Zhang, R. Wu and H. Zou, Angew. Chem., Int. Ed., 2011, 50, 12218.
48. K. Stumpo and V. Reinhold, J. Proteome Res., 2010, 9, 4823.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c5ra11998g

---