Highly Specific Enrichment of Glycopeptides Using Boronic Acid

Haojie Lu, Department of Chemistry and Institutes of Biomedical Sciences, Fudan University, Shanghai 200433 (P.R. China). Fax: 86-21-6564-2009. E-mail...
4 downloads 4 Views 675KB Size
Anal. Chem. 2009, 81, 503–508

Highly Specific Enrichment of Glycopeptides Using Boronic Acid-Functionalized Mesoporous Silica Yawei Xu,† Zhangxiong Wu,‡,§ Lijuan Zhang,† Haojie Lu,*,† Pengyuan Yang,† Paul A. Webley,§ and Dongyuan Zhao*,‡,§ Department of Chemistry and Institutes of Biomedical Sciences and Department of Chemistry and Shanghai Key Laboratory of Molecular Catalysis and Innovative Materials, Fudan University, Shanghai 200433 (P.R. China), and Department of Chemical Engineering, Monash University, Clayton, VIC 3800, Australia A novel boronic acid functionalized mesoporous silica, which holds the attractive features of high surface area and large accessible porosity, was developed to enrich glycopeptides. This is the first time that mesoporous material has been introduced into glycoproteome. In comparison to direct (traditional) analysis, this novel method enabled 2 orders of magnitude improvement in the detection limit of glycopeptides. The unbiased nature of organo-boronic acid groups also made this method applicable to all kinds of glycopeptides regardless of their sizes, structures, and hydrophilicities. Glycosylation, the most common post-translational modification, is of ultimate importance in developmental biology, cell division, tumor immunology, inflammation, and prion diseases. Currently, the methods applied to study glycopeptides or glycoproteins are more popularly based on mass spectrometry (MS) techniques. Because of their low abundances (2-5%), MS responses of glycopeptides are severely suppressed by nonglycosylated peptides. Moreover, most glycosylation sites carry a multitude of glycans, giving rise to different glycoforms. This phenomenon further reduces the relative amount of glycopeptides and makes them hard to be detected. As a result, it is almost impossible to analyze substoichiometric glycopeptides without specific enrichment steps. Among the glycopeptide enrichment techniques, lectin affinity chromatography is the most widely used one.1 Concanavalin A (Con A) has frequently been used for the enrichment of Nglycoproteins from diverse sources.2-4 It reveals good affinity * Haojie Lu, Department of Chemistry and Institutes of Biomedical Sciences, Fudan University, Shanghai 200433 (P.R. China). Fax: 86-21-6564-2009. E-mail: [email protected]. Dongyuan Zhao, Department of Chemistry and Shanghai Key Laboratory of Molecular Catalysis and Innovative Materials, Fudan University, Shanghai 200433 (P.R. China). Fax: 86-21-65641740. E-mail: dyzhao@ fudan.edu.cn. † Department of Chemistry and Institutes of Biomedical Sciences, Fudan University. ‡ Department of Chemistry and Shanghai Key Laboratory of Molecular Catalysis and Innovative Materials, Fudan University. § Department of Chemical Engineering, Monash University. (1) Hirabayashi, J. Glycoconjugate J. 2004, 21, 35–40. (2) Fan, X.; She, Y. M.; Bagshaw, R. D.; Callahan, J. W.; Schachter, H.; Mahuran, D. J. Anal. Biochem. 2004, 332, 178–186. (3) Bunkenborg, J.; Pilch, B. J.; Podtelejnikov, A. V.; Wisniewski, J. R. Proteomics 2004, 4, 454–465. (4) Wang, L.; Li, F.; Sun, W.; Wu, S.; Wang, X.; Zhang, L.; Zheng, D.; Wang, J.; Gao, Y. Mol. Cell. Proteomics 2006, 5, 560–562. 10.1021/ac801912t CCC: $40.75  2009 American Chemical Society Published on Web 12/08/2008

toward high mannose and hybrid N-glycans; slightly lower affinity toward diantennary N-glycans; no affinity toward tri- and tetraantennary complex-type glycans.5 As most tryptic glycopeptides have relatively high mass, size exclusion chromatography has also been applied to glycopeptide enrichment.6 Another enrichment method is based on the hydrophilicity of the glycan moiety. Gel matrixes such as cellulose or sepharose were used to extract the hydrophilic glycopeptides.7 However, many unglycosylated peptides, which contain several hydrophilic amino acids, reveal as strong hydrophilicities as glycopeptides do. Moreover, glycogroups attached with some hydrophobic peptides could also make them unable to be retained with hydrophilic groups. Hydrazide functionalized beads, another important matrix, can efficiently trap glycopeptides by covalent binding after oxidation of cis-diol groups of carbonhydrates with periodate, but it needs an additional oxidation step which clearly increases experiment time and sample complexity.8 Recently, a generally applicable matrix, diboronic acid functionalized magnetic beads, has been introduced for the unbiased enrichment of both N- and O-glycopeptides,9 because diboronic acid can form diesters with all glycans and glycoconjugates that contain cis-diol groups.10 According to this concept, herein, we carefully design to graft diboronic acid as a surface functional group to an ordered and high-surface-area mesoporous silica matrix, aiming at highly specific enrichment of glycopeptides. EXPERIMENTAL SECTION Materials and Chemicals. 3-Glycidyloxypropyltrimethoxysilane (GLYMO, 98%), 3-aminophenylboronic acid monohydrate (APB, 98%), bovine serum albumin (BSA, 95%), horseradish peroxidase (HRP, 98%), asialofetuin (98%), invertase (98%), fetuin (98%), 2,5-dihydroxybenzoic acid (DHB, 98%) ammonium bicarbonate (ABC, 99.5%), and PNGaseF (Proteomics grade, g95%) were obtained from Sigma (St. Louis, MO). H218O Water (isotope 18 O, 97%) was obtained from Cambridge Isotope Laboratories. Acetonitrile (ACN, 99.9%) and trifluoroacetic acid (TFA, 99.8%) were purchased from Merck (Darmstadt, Germany). All these (5) Cummings, R. D.; Kornfeld, S. J. Biol. Chem. 1982, 257, 11235–11240. (6) Alvarez-Manilla, G.; Atwood, J., III; Guo, Y.; Warren, L.; Orlando, R.; Pierce, M. J. Proteome Res. 2006, 5, 701–708. (7) Wada, Y.; Tajiri, M.; Yoshida, S. Anal. Chem. 2004, 76, 6560–6565. (8) Tian, Y.; Zhou, Y.; Elliott, S.; Aebersold, R.; Zhang, H. Nat. Protoc. 2007, 2, 334–339. (9) Sparbier, K.; Koch, S.; Kessler, I.; Wenzel, T.; Kostrzewa, M. J. Biomol. Tech. 2005, 16, 407–413. (10) Rawn, J. D.; Lienhard, G. E. Biochemistry 1974, 13, 3124–3130.

Analytical Chemistry, Vol. 81, No. 1, January 1, 2009

503

Scheme 1. Postsynthetic Steps (Left) of Ordered Mesoporous Di-Boronic Acid Functionalized FDU-12 (Denoted as FDU-12-GA)a

a GLYMO ) (3-glycidyloxypropyl)trimethoxysilane, APB ) 3-aminophenyl-boronic acid monohydrate. Specific enrichments (right) of glycopeptides from dilute peptides solution using the newly developed FDU-12-GA as a nanoreactor through the specific binding of di-boronic acid groups and glycopeptides.

reagents were used as received without further purification. Deionized water (18.4 MΩ cm) used for all experiments was obtained from a Milli-Q system (Millipore, Bedford, MA). Synthesis. Ordered mesoporous silica FDU-12 was synthesized according to the method reported by Fan et al.25,26 except that we removed the organic triblock copolymer templates by microwave digestion (MWD).27 For synthesizing the boronic-acid functionalized FDU-12 material, we adopted a two-step post graft method (Scheme 1). First of all, 3-glycidyloxypropyltrimethoxysilane (GLYMO) and 3-aminophenylboronic acid monohydrate (APB) were reacted to prepare boronic-acid bonded GLYMO (denoted as GA). Typically, APB (50 mg) was dissolved in 20 mL of deionized water. The solution was adjusted to pH of 9.18 by aqueous NaOH. To minimize the hydrolysis of GLYMO, 40 µL of GLYMO was slowly added into the APB solution with stirring when the temperature of solution was decreased to 0 °C at an ice-bath. The mixed solution was heated to 40 °C for reaction of 6 h with stirring, subsequently placed into an ice-bath for 5 min to decrease the temperature to 0 °C, and 40 µL of GLYMO was added and mixed again. Then the solution was raised to 65 °C for another 6 h with stirring. Finally, the prepared GA solution was stored in a refrigerator for usage. Second, 20 mg of FDU-12 after removal of the template by MWD were mixed with 5 mL of GA solution in a flask at 75 °C for 2 h with stirring, and then the solution was centrifugated. After the supernatant was removed, 5 mL of GA solution was added to the flask to restart the reaction. This procedure as described above was repeated two times to prepare the final product of boronic acid functionalized FDU-12, denoted as FDU-12-GA. Glycopeptides Enrichment. The specific enrichment of glycopeptides is shown in Scheme 1. To enrich the glycopeptides from tryptic glycoproteins in 50 mM of ammonium bicarbonate (ABC), 10 µL of FDU-12-GA suspension (20 µg/µL) was added into the mixture. Then it was incubated with a shaking for 15 min in room temperature, followed by centrifugation at 900 rcf for 3 min. After the supernatant was decanted, 100 µL of 50 mM ABC was used to wash the beads (15 min in room temperature), followed by centrifugation (900 rcf for 3 min) to remove supernatant. Then 10 µL or more of the elution buffer (1% trifluoroacetic 504

Analytical Chemistry, Vol. 81, No. 1, January 1, 2009

acid, 50% acetonitrle) was added to release the glycopeptides from FDU-12-GA (30 min at room temperature). MALDI Mass Spectrometry. After centrifugation, 0.5 µL of elute and 0.5 µL of matrix (12 mg/mL 2,5-dihydroxy benzoic acid dissolved in 1:4 (v/v) acetonitrile/water solution containing 0.1% trifluoroacetic acid) was spotted on the MALDI plate for MS analysis. All mass spectra were acquired by a MALDI AXIMA QIT (Shimadzu Biotech, Japan). Measurements. Small-angle X-ray scattering (SAXS) patterns were recorded on a Nanostar U small-angle X-ray scattering system (Bruker, Germany) using Cu KR radiation (40 mV, 35 mA). TEM images were obtained with a JEOL 2011 microscope operated at 200 kV. Samples for TEM measurements were suspended in ethanol and supported on a carbon-coated copper grid. FE-SEM (field emission SEM) images were obtained with a JEOL 6300F scanning electron microscope operated at 15 kV. N2 adsorption isotherms were measured using a Micromeritics Tristar 3000 analyzer at 77 K. Before measurements were taken, all samples were degassed at 393 K for more than 6 h. IR spectra were measured at a Nicolet Nexus 470FT-IR spectrometer. Thermogravimetric (TG) analyses were monitored using a Mettler Toledo TGA-SDTA851 analyzer (Switzerland) from 25 to 800 °C under oxygen. RESULTS AND DISCUSSION To fulfill the purpose of efficiently and specifically enriching glycopeptides, the structural material we chose needs to have some attractive features answering for the requirements of highly specific enrichment. First of all, it has comparatively large specific surface area so that it could absorb glycopeptides fast and efficiently. Second, it should be easily grafted with diboronic acid groups to achieve high specificity and unbiased affinity toward all kinds of glycopeptides. Third, it is easy to recover by centrifuge in order to extract glycopeptides from the peptide pool effectively. After centrifugation, the material could be easily redispersed for washing and eluting steps. Finally, captured glycopeptides could be released effortlessly. Mesoporous silica, with its prominent features of high surface area, large pore volume, and narrow distribution of regular pore

Figure 1. SAXS patterns (a), nitrogen sorption isotherms (b), and the corresponding pore size distributions (inset of part b, full symbols stand for the curve derived from the adsorption branch; open symbols stand for the curve derived from the desorption branch; circles stand for FDU12; triangles stand for FDU-12-GA) of mesoporous silica FDU-12 and FDU-12-GA materials.

size, is well believed to lead to considerable advantages in mass diffusion and transportation of adsorption processes, which lead to much shorter loading time, fast adsorption equilibrium, and larger adsorption capacity.11-17 Furthermore, the surface of the mesoporous silica, with a large amount of silanol groups, can be easily modified with chemical reactions, thus promising even greater potentials such as selective adsorbent in separation techniques.18-24 Herein, by employing the mesoporous silica FDU12 as the matrix, which has three-dimensional (3-D) pore system with large pore cavity and entrance sizes and thus promises even more advantages in adsorption processes,25,26 we synthesized a novel mesoporous boronic-acid-functionalized silica FDU-12 (denoted as FDU-12-GA) through postgraft method to meet the requirements for highly selective enrichment of glycopeptides. Our results show that the newly functional FDU-12-GA with high surface area, large pore volume, and well-defined large sizes, leads to much short loading time, dramatically high specificity, and a much lower limit of detection toward glycopeptides. Ordered mesoporous silica FDU-12 with very large pores was fabricated according to the method reported by Fan et al.25,26 (11) Kresge, C. T.; Leonowicz, M. E.; Roth, W. J.; Vartuli, J. C.; Beck, J. S. Nature 1992, 359, 710–712. (12) Yanagisawa, T.; Shimizu, T.; Kuroda, K.; Kato, C. Bull. Chem. Soc. Jpn. 1990, 63, 1535–1537. (13) Zhao, D. Y.; Huo, Q. S.; Stuky, G. D.; Feng, J.; Melosh, N.; Fredrickson, G. H. Science 1998, 279, 548–552. (14) Wan, Y.; Zhao, D. Y. Chem. Rev. 2007, 107, 2821–2860. (15) Lin, H. P.; Mou, C. Y. Acc. Chem. Res. 2002, 35, 927–935. (16) Hartmann, M. Chem. Mater. 2005, 17, 4577–4593. (17) Davis, M. E. Nature 2002, 417, 813–821. (18) Zhang, W.; Lu, X.; Xiu, J.; Hua, Z.; Zhang, L.; Robertson, M.; Shi, J.; Yan, D.; Holmes, J. Adv. Funct. Mater. 2004, 14, 544–552. (19) Liu, A. M.; Hidajat, K.; Kawi, S.; Zhao, D. Y. Chem. Commun. 2000, 1145– 1146. (20) Feng, X.; Fryxell, G. E.; Wang, L. Q.; Kim, A. Y.; Liu, J.; Kemner, K. M. Science 1997, 276, 923–926. (21) Stein, A.; Melde, B. J.; Schroden, R. C. Adv. Mater. 2000, 12, 1403–1419. (22) Yang, C. M.; Zibrowius, B.; Schu ¨ th, F. Chem. Commun. 2003, 1772–1773. (23) Antochshuk, V.; Olkhovyk, O.; Jaroniec, M.; Park, I. S.; Ryoo, R. Langmuir 2003, 19, 3031–3034. (24) Mercier, L.; Pinnavaia, T. I. Adv. Mater. 1997, 9, 500–503.

However, we employed the microwave digestion method to remove the organic structure-directing agent, which is wellbelieved to be very helpful to decrease the structural shrinkage and retain many more silanol groups compared to the calcination method.27 FDU-12 material thus-prepared has many silanol groups which act as active sites for the functionization. As shown in Scheme 1, we adopted a two-step method to covalently graft FDU12 with boronic acid. Unlike most direct postmodification, we first prepared a boronic-acid-bonded alkoxyorganosilane because of the lack of such commercial organosilanes. We selected (3-glycidyloxypropyl) trimethoxysilane (GLYMO) and 3-aminophenylboronic acid (APB) because of the easy reaction between epoxy groups and amino groups (Scheme 1). Then, the obtained alkoxyorganosilane (denoted as GA) was grafted to the silica walls by condensation reaction between the alkoxy and silanol groups, leading to the novel boronic-acid functionalized FDU-12 product (denoted as FDU-12-GA). Small-angle X-ray scattering (SAXS) patterns (Figure 1a) of both FDU-12 and FDU-12-GA specimens show five well-resolved diffraction peaks, which can be assigned to the 111, 220, 311, 331, and 442 reflections of a face-centered cubic (fcc) structure.25,26 The cell parameter (a0) of both the two materials are calculated to be about 40.8 nm, indicating the ordered mesostructure is well retained after the postgraft. Nitrogen sorption measurements were conducted to determine the surface areas and the cavity and entrance sizes of the FDU-12 and FDU-12-GA materials (Figure 1b). Both the two samples show type-IV isotherms with H1 hysteresis loops, which are unlike other mesoporous silicas with cagelike pores having H2 hysteresis loops, indicating large entrance sizes even after functionization. However, it is found that after the postsynthesis, the FDU-12-GA material has a much wider hysteresis loop than FDU-12, indicating smaller pore entrances. From the corresponding pore size distribution curves (inset Figure 1b), the size of the pore cavities of FDU-12GA calculated from the adsorption branch is centered at 18.2 nm, a little smaller than that of FDU-12 (18.3 nm), suggesting that Analytical Chemistry, Vol. 81, No. 1, January 1, 2009

505

Figure 2. FE-SEM image of FDU-12-GA at low magnification (a) and the high-resolution image of the corresponding square area in part a (b); TEM images along the [110] direction of FDU-12-GA (c,d).

the inner wall of the pore cavities are probably grafted with a thin layer of the alkoxyorganosilane. However, because the grafting density is probably low in the pore cavities, the measured average pore cavity sizes are still close to that of mother FDU-12. Moreover, the steep evaporation step of the desorption branch of FDU-12-GA shifts to much lower relative pressure than that of FDU-12, which means the pore entrances becomes smaller after the grafting. The pore entrance sizes calculated from the desorption branches shift from 5.1 nm for FDU-12 down to ∼3.5 nm for FDU-12-GA, suggesting that the grafting reaction mainly happens at the pore entrances. The surface area and pore volume of the FDU-12 matrix are 540 m2/g and 1.1 cm3/g, respectively. After functionalized with the boronic-acid-bonded GA, they fall down to 200 m2/g and 0.48 cm3/g. The large drop of the surface area and pore volume after grafting is attributed to the high percentage (up to 32 wt %) of the grafted organics in the material, further suggesting a successful pore modification of FDU-12. The FE-SEM image at low magnification (Figure 2a) shows that the FDU-12-GA material has hexagonal polyhedron morphology, the same as mesoporous silica FDU-12 without fictionalization (data not shown). At the surface of the polyhedron crystals, we found some small aggregates. Careful observations at high magnifications (Figure 2b) show that all of these aggregates are highly ordered mesoporous materials, without byproduct from condensation reaction of the functional GA organosilane itself. It further demonstrates that the GA molecules are covalently grafted to the silica walls of FDU-12. TEM images of FDU-12-GA material (typically taken along the [110] direction) further demonstrate that the excellent fcc mesostructure is retained after the func506

Analytical Chemistry, Vol. 81, No. 1, January 1, 2009

tionalization (Figure 2c). The d111-spacing value calculated from TEM images is about 23.6 nm, in good agreement with the value (23.5 nm) derived from SAXS data. In addition, 1-D-like large pores that are formed by the linear combination of many spherical cages can be further observed in the TEM images (Figure 2d), thus implying that even after functionalization, the FDU-12-GA material still has large pore entrances. The diameter of the mesopore cages can roughly be estimated from the thin edge of the particles to be about 18.0 nm (Figure 1d), in good accordance with that from N2 sorption results. IR spectra (Supporting Information, Figure 1s) show a band centered at 1636 cm-1 which is ascribed to the vibration of adsorbed water. Dramatic intensity loss of this band after the graft obviously demonstrates that the surface of the FDU-12GA material becomes more hydrophobic; correspondingly, the weight loss (Supporting Information, Figure 2s) of adsorbed water in FDU-12-GA is much less than that of the counterpart FDU-12; the total amount of the grafted organics in FDU-12-GA is about 32.2 wt %, suggesting a successful modification with the organic molecules. We estimate the C, N concentrations to be ∼16.7 mmol g-1 for C and ∼1.5 mmol g-1 for N, respectively. The vibrations (2930, 3040, 1601, 1577, and 1441 cm-1) contributed by methylene groups and benzene rings for FDU12-GA product further testifies the successful functionalization of the boronic-acid bonded molecules. Moreover, compared to FDU-12, the existence of bands at 1339, 1261, and 716 cm-1 due to vibrations of B-O bonds in FDU-12-GA clearly suggests that boronic-acid-functionalized mesoporous silica material is obtained.

Figure 3. MALDI mass spectra of the elution (a, b) and the flow-through (c, d) of 23 fmol of tryptic digest HRP. The upper (lower) two were acquired with the same spot but in a different mass range run: m/z > 750 (a,c) and m/z > 2000 (b,d). Asterisks represent glycopeptides of HRP. 2592 (HRP) ) PTLN#TTYLQTLR (proved by MSn); 3355 (HRP) ) SFAN#STQTFFNAFVEAMDR;29 3673 (HRP) ) GLIQSDQELFSSPN#ATDTIPLVR;29 3896 (HRP) ) LHFHDCFVNGCDASILLDN#TTSFR;29 4984 (HRP) ) LYNFSNTGLPDPTLN#TTYLQTLR.29

With the use of this new boronic acid grafted mesoporous silica (FDU-12-GA) for specific enrichment of glycopeptides, the loading time can be shortened to just 15 min, much shorter than those of conventional methods (1 h).9,28 Take commercially available boronic acid functionalized magnetic beads as an example, the absolute intensity of the mass spectrum obtained from a incubation time of 15 min is clearly poorer than that from 1 h incubation (Supporting Information, Figure 3s). In our experiments, it is comparably good to use a loading time of 15 min and 1 h when adopting FDU-12-GA as the matrix. It might be ascribed to the high surface area and large entrance pore sizes of the mesoporous silica matrix. Four tryptic glycoproteins were used as models to evaluate its enrichment selectivity toward glycopeptides. As shown in Figure 3a,b, FDU-12-GA can separate five glycopeptides from the 5 ng/µL HRP digest. While the glycopeptides can be recovered, all nonglycosylated peptides are left in the flow through (Figure 3c,d), suggesting specificity of binding. Then we mixed tryptic BSA with HRP in different mass ratios (from 1:1 to 10:1) to examine whether highly abundant nonglycosylated peptides would affect the specificity or recovery of glycopeptides (Supporting Information, Figure 4s). With the increasing amount of tryptic BSA, the intensities of glycopeptides decrease gradually. It is probably due to that lots of nonglycosylated peptides inhibit the binding process. Although the recovery is slightly influenced, the specificity is not; only glycopeptides are observed in the (25) Fan, J.; Yu, C. Z.; Gao, F.; Lei, J.; Tian, B. Z.; Wang, L. M.; Luo, Q.; Tu, B.; Zhou, W.; Zhao, D. Y. Angew. Chem., Int. Ed. 2003, 42, 3146–3150. (26) Fan, J.; Yu, C. Z.; Lei, J.; Zhang, Q.; Li, T.; Tu, B.; Zhou, W. Z.; Zhao, D. Y. J. Am. Chem. Soc. 2005, 127, 10794–10795. (27) Tian, B. Z.; Liu, X.; Yu, C. Z.; Gao, F.; Luo, Q.; Xie, S. H.; Tu, B.; Zhao, D. Y. Chem. Commun. 2002, 11, 1186–1187. (28) Sparbier, K.; Asperger, A.; Resemann, A.; Kessler, I.; Koch, S.; Wenzel, T.; Stein, G.; Vorwerg, L.; Suckau, D.; Kostrzewa, M. J. Biomol. Tech. 2007, 18, 252–258.

spectra of the elutent. It clearly reveals the excellent adsorption selectivity toward glycopeptides. Besides HRP, three more tryptic glycoproteins (fetuin, invertase, and asialofetuin) were applied to FDU-12-GA enrichment to find out whether it is applicable to various types of glycopeptides (Supporting Information, Figure 5s). From this figure, glycopeptides from different model proteins are successfully enriched. It could be safe to draw the conclusion that FDU-12-GA is highly specific and widely suitable to various types of N-glycopeptides. With this glycopeptide enrichment step, the limit of detection (LOD) of glycopeptide is also dramatically enhanced. Two aliquots of tryptic glycoprotein, with and without enrichment, respectively, were used to evaluate the sensitivities of glycopeptides under these conditions. In the enrichment procedure, the volume of elution buffer is equal to that of the loading buffer, so the molar concentrations of glycopeptides are nearly the same. As can be seen from Table 1, the originally undetectable glycopeptides (Supporting Information, Figure 6s parts a,c,e,g) could be easily detected after FDU-12-GA enrichment (Supporting Information, Figure 6s parts b,d,f,h). The LOD of glycopeptides is greatly improved. This result could be explained by the fact that the existence of prominent nonglycosylated peptides severely suppresses the ionization of glycopeptides. Once this suppression effect is eliminated by FDU-12-GA enrichment, the LOD of glycopeptides of these model proteins can be largely improved (Table 1, the third column). Moreover, the LOD of glycopeptides can be further improved by concentrating through FDU-12-GA enrichment. In the following experiments, 10 µL of elution buffer was used to detach glycopeptides which were dispersed in a bulky solution (500-1000 µL) (29) Wuhrer, M.; Hokke, C. H.; Deelder, A. M. Rapid Commun. Mass Spectrom. 2004, 18, 1741–1748.

Analytical Chemistry, Vol. 81, No. 1, January 1, 2009

507

Table 1. Limit of Detection (LOD) of Glycopeptides of Four Distinct Glycoproteins

protein

before enrichment (fmol/µL)

glycol-specific selectiona (fmol/µL)

glyco-specific enrichmentb (amol/µL)

HRP asialofetuin fetuin invertase

43.1 54.4 108 43.9

11.6 2.17 10.1 17.2

620 435 870 468

a The amount of loading buffer is the same as that of elution (100 µL). b The amount of loading buffer is 10 times larger than that of elution (10 µL).

in the beginning. After FDU-12-GA enrichment, the LOD could be decreased to the attomole level (Table 1, the fourth column). In summary, the detection limits of these four glycoproteins were improved by 70-125 times through our strategy. Finally, the recovery of glycopeptides from the FDU-12-GA after enrichment was investigated. A preprepared tryptic peptides sample was divided into two parts with equal amounts of glycoprotein asialofetuin. The release of the glycans with PNGaseF in H218O was the first part. The other part involved the trapping of the glycopeptides using the FDU-12-GA and eluting them from the silica and then releasing the glycans with PNGaseF in H2O. With the two parts mixed again, the product was profiled with MS for a comparative study of the abundances of the glycopeptides from different oxygen isotopes according to peak areas. As the MALDI spectra reveal (Supporting Information, Figure 7s), the recovery of glycopeptides is up to 83.5% which could be calculated according to the isotope distribution (based on peak area).30 In summary, a novel diboronic acid functionalized mesoporous silica (FDU-12-GA) was synthesized to specifically enrich glycopeptides. This developed FDU-12-GA matrix, with high surface (30) Johnson, K. L.; Muddiman, D. C. J. Am. Soc. Mass Spectrom. 2004, 15, 437–445.

508

Analytical Chemistry, Vol. 81, No. 1, January 1, 2009

area, large pore volume, and well-defined large sizes of pore cavities and entrances, leads to great advantages in glyco-specific enrichment. First, the synthesis of FDU-12-GA is quite simple under moderate temperature and neutral conditions. Second, the large specific surface area greatly increases its binding rate, and the entire loading time needs only 15 min. With the plentiful diboronic acid function groups on the surface, no nonspecific binding is observed in the presence of prominent tryptic BSA, demonstrating specificity of binding. The suppression effect from nonglycopeptides can be eliminated by capturing and concentrating target glycopeptides, the LOD of glycopeptides is enhanced by close to 2 orders of magnitude, and the recovery of the enriched glycopeptides is up to 83.5%. Accordingly, with our newly developed method, various kinds of N-glycopeptides could be specifically selected and enriched for further analysis. We also expect it could be further applied to complex biological samples and glycoproteomics. ACKNOWLEDGMENT Y.X. and Z.W. contributed equally to this work. This work was supported by the National Science and Technology Key Project of China (Grants 2007CB914100, 2006AA02Z134, 2009CB825607, 2006AA02A308, and 2008ZX10207), the National Natural Science Foundation of China (Grants 20735005, 20875016, 30672394, 20721063, and 20521140450), NCET, the Shanghai Leading Academic Discipline (Grants B108, B109), and the Shanghai Rising-Star Program (Grant 06QA14004). SUPPORTING INFORMATION AVAILABLE Additional information as noted in text. This material is available free of charge via the Internet at http://pubs.acs.org.

Received for review November 19, 2008. AC801912T

September

9,

2008.

Accepted