Article pubs.acs.org/ac
Analysis of O‑Glycans as 9‑Fluorenylmethyl Derivatives and Its Application to the Studies on Glycan Array Keita Yamada,†,‡ Jun Hirabayashi,‡ and Kazuaki Kakehi*,† †
School of Pharmacy, Kinki University, Kowakae 3-4-1, Higashi-Osaka 577-8502, Japan Division of Glyco-Bioindustry, Life Science Research Center, Institute of Research Promotion, Kagawa University, Ikenobe 1750-1, Miki-cho, Kita-gun, Kagawa 761-0793, Japan
‡
S Supporting Information *
ABSTRACT: A method is proposed for the analysis of Oglycans as 9-fluorenylmethyl (Fmoc) derivatives. After releasing the O-glycans from the protein backbone in the presence of ammonia-based media, the glycosylamines thus formed are conveniently labeled with Fmoc-Cl and analyzed by HPLC and MALDI-TOF MS after easy purification. Fmoc labeled O-glycans showed 3.5 times higher sensitivities than those labeled with 2-aminobenzoic acid in fluorescent detection. Various types of O-glycans having sialic acids, fucose, and/or sulfate residues were successfully labeled with Fmoc and analyzed by HPLC and MALDI-TOF MS. The method was applied to the comprehensive analysis of O-glycans expressed on MKN45 cells (human gastric adenocarcinoma). In addition, Fmoc-derivatized O-glycans were easily converted to free hemiacetal or glycosylamine-form glycans that are available for fabrication of glycan array and neoglycoproteins. To demonstrate the availability of our methods, we fabricate the glycan array with Fmoc labeled glycans derived from mucin samples and cancer cells. The model studies using the glycan array showed clear interactions between immobilized glycans and some lectins.
O
Although various methods for releasing of O-glycans from the core protein have been developed, NaOH-catalyzed βelimination in the presence of sodium borohydride under mild conditions has been the most common method. However, the method cannot afford free aldose-form glycans, and the reduced form (i.e., alditol-form) glycans are obtained.13,14 Consequently, the original reducing terminal is no longer available for modification via the hemiacetal group, which is essential for derivatization to achieve sensitive and highresolution analysis. Some works on the release of O-glycans with intact reducing groups have been reported. Royle et al. employed mild hydrazinolysis to obtain O-glycans using microgram scale of glycoprotein samples.15 Huang et al. developed a method for releasing O-glycans in the presence of ammonia and successfully analyzed O-glycans from fetuin.16 Kakehi et al. developed an automatic O-glycan releasing apparatus for the release of O-glycans from mucin-type glycoproteins and proteoglycans17,18 and applied the apparatus to the comprehensive analysis of O-glycans expressed on cancer cells after labeling with 2-aminobenzoic acid (2AA).19,20 Furukawa and Wang also reported a novel one-pot procedure for the nonreductive release of O-linked glycans from glycoproteins
ne of the important posttranslational modifications of proteins is O-glycosylation at serine/threonine residues of mucin-type glycoproteins.1,2 Chen et al. revealed that inhibition of O-glycan synthesis resulted in suppression of CD8+ T cell activity both in vivo and in vitro.3 Tuboi et al. reported that O-glycans allow cancer cells to evade natural killer cells of the immune system and survive longer in the circulatory system, thereby promoting tumor metastasis.4,5 Currently, eight core structures for O-glycans (cores 1−8) are known,6 and a number of research works reported the aberrant profiles of O-glycans in relation to the development and metastasis of malignancy.7 For example, expression of core 3 and core 4 types of O-glycans are down regulated in gastric and colonic carcinoma.8 Highly sialylated structures or truncated structures such as Tn and sialyl-Tn antigens are observed in tumor tissues.9−11 Thus, O-glycans are believed to have potential as novel clinical biomarkers for the early detection of tumors or as a means of discriminating benign from malignant diseases. O-Glycan arrays and neoglycoproteins have been developed to analyze the functions of O-glycans.12 At present, O-glycans available for such purposes are relatively small molecules and have simple structures, because synthesis of large enough amounts of complex glycans is extremely difficult and they have to be provided from natural sources. Analysis of O-glycans is still challenging especially due to a lack of good O-glycan releasing methods from the core protein. © 2013 American Chemical Society
Received: December 26, 2012 Accepted: February 14, 2013 Published: February 14, 2013 3325
dx.doi.org/10.1021/ac303771q | Anal. Chem. 2013, 85, 3325−3333
Analytical Chemistry
Article
centrifugation at 18,000 g, an aqueous acetone solution (85%; 1.5 mL) containing 5% triethylamine and 5% acetic acid was added, and the mixture was incubated at −20 °C for 2 h. The precipitate was collected by centrifugation and washed with 70% ethanol (1.0 mL × 2) followed by drying in vacuo. The collected proteins were suspended in 0.1 M Tris-HCl buffer containing 2 M urea (pH 8.4, 250 μL), and the aqueous solution of trypsin (2.5 mg/mL, 160 μL) was added to the mixture. After digestion for 24 h at 37 °C, the mixture was passed through a centrifugal ultrafiltration membrane (100 kDa cutoff) to collect O-glycosylated glycopeptide fractions. After washing with the water (500 μL × 2), the concentrated solution on the membrane was collected with the water (100 μL × 3) and lyophilized to dryness. Analysis of Glycosylamine-Form O-Glycans after Labeling with Fmoc-Cl. The mixture of glycopeptides from cells obtained as described above or a glycoprotein sample (1− 1000 μg) was dissolved in 28% aqueous ammonia solution (1 mL) previously saturated with ammonium carbonate at room temperature, and the mixture was kept at 37 °C for 24 h. The mixture was evaporated to dryness to remove ammonia and ammonium carbonate by a Centra-Vap (Labconco Corp., Kansas City, MO). Phosphate buffer (30 mM, pH 8.4, 200 μL) was added to the dried materials to suppress hydrolysis of glycosylamine-form glycans to free-form glycans, and the mixture was evaporated to dryness. The dried mixture was dissolved in water (200 μL) and again evaporated to dryness, and the procedures were repeated several times. Finally, after the mixture was dissolved in water (300 μL), a freshly prepared solution of Fmoc-Cl in acetone (200 μL, 5 mg/mL) was added, and the mixture was kept at 37 °C for 1.5 h. Chloroform (300 μL) was added to the mixture, and the mixture was shaken vigorously and the chloroform layer was removed carefully. The same procedures for washing the reaction mixture with chloroform were repeated two times. The aqueous layer containing Fmoc labeled O-glycans was evaporated to dryness. The dried material was dissolved in water, and a portion was analyzed by HPLC. Free O-Glycans from Fmoc Derivatives. Free O-glycans were easily recovered from Fmoc-labeled glycans according to the method reported previously.25 Fmoc labeled O-glycans collected by HPLC were dissolved in water (20 μL). N,NDimethylformamide (DMF, 30 μL) and morpholine (20 μL) were added to the solution, and the mixture was kept at 37 °C for 30 min. Diethyl ether (500 μL) was added to the mixture and the mixture was shaken vigorously. After brief centrifugation, the aqueous phase was collected and evaporated to dryness by a centrifugal evaporator. Because the dried material is glycosylamine-form O-glycans, it is conveniently available for direct fabrication of glycan array. Alternatively, the dried material is dissolved in 20 μL of 0.5 M boric acid, and kept at 37 °C for 30 min. After drying the mixture, boric acid was removed by repeated evaporations with methanol. The procedure converts glycosylamine-form to freeform (i.e., hemiacetal-form). Fluorescent Labeling of Free-Form O-Glycans with 2AA. The sample of the free-form glycan obtained from Fmoclabeled glycans was dissolved in 2AA solution (100 μL) freshly prepared by dissolution of 2AA (30 mg) and sodium cyanoborohydride (30 mg) in methanol (1 mL) containing 4% sodium acetate and 2% boric acid. The mixture was kept at 80 °C for 1 h. After cooling, water (100 μL) was added, and the mixture was applied to a column of Sephadex LH-20 (1.0 cm
and the simultaneous derivatization of released glycans with pyrazolone analogs.21,22 Because these labeled glycans are converted to open-ring structure of the terminal monosaccharide, subsequent modification to neoglycoprotein or immobilization onto an array slide is difficult. Although various methods for immobilizing the labeled glycans to an array slide or removing the labeled reagent were reported,23,24 there have been no established methods. Recently, Kamoda et al. reported a novel labeling method for N-glycans with 9-fluorenylmethyl chloroformate (Fmoc-Cl).25 The method is based on direct derivatization of glycosylamineform N-glycans obtained after digestion of glycoproteins with PNGase F. The labeled glycans are analyzed by HPLC/CEfluorometry and MS with high sensitivity.25,26 It should be emphasized that the Fmoc labeled glycans are easily converted to free-form glycans after separation and collection of the peaks by HPLC. In the present sturdy, we developed a novel method for rapid and sensitive analysis of O-glycans using Fmoc labeling method, and applied the method to the analysis of O-glycans expressed on cancer cells. In addition, O-glycans obtained from some glycoproteins and cancer cells were employed to the model studies on O-glycan micro array, and the interaction between O-glycans and lectins was analyzed. The results indicate that the proposed method has great potential to make a breakthrough in the novel functional analysis of O-glycans.
■
EXPERIMENTAL SECTIONS Materials. 9-Fluorenylmethyl chloroformate (Fmoc-Cl) and 2-aminobenzoic acid (2AA) were obtained from Tokyo Kasei (Chuo-ku, Tokyo, Japan). Bovine submaxillary mucin (BSM), porcine stomach mucin (PSM), and bovine fetuin were obtained from Sigma-Aldrich (St. Louis, MO). 2,5-Dihydroxy benzoic acid (DHB) as the matrix material for MALDI MS measurement was from Bruker Daltonics (Bremen, Germany). Trypsin was purchased from Thermo Scientific (Yokohama, Kanagawa, Japan). Biotinylated lectins, Artocarpus integrifolia agglutinin (Jacalin), Datura stramonium agglutinin (DSA), Vicia Villosa agglutinin (VVAB4), and wheat germ agglutinin (WGA), were obtained from EY laboratories (San Mateo, CA). Model glycopeptide from caseinoglycomacropeptide (cow milk) was prepared according to the method reported previously.17 Other reagents and solvents were of the highest grade commercially available or HPLC grade. Cell Culture. MKN45 cells (human gastric adenocarcinoma cell line) were cultured in RPMI-1640 medium supplemented with 10% (v/v) fetal calf serum and 1% (v/v) penicillinstreptomycin mixed solution (10 000 u penicillin and 10 mg of streptomycin/ml: Nacalai Tesque). Fetal calf serum was previously kept at 56 °C for 30 min. The cells were cultured at 37 °C under 5% CO2 atmosphere and harvested at 80% confluent state. Collected cells (2.0 × 107 cells) were washed with phosphate buffered saline (PBS), and collected by centrifugation at 800 g for 20 min. Collection of O-Glycosylated Peptide Fractions from Cancer Cells. Cultured cancer cells (2.0 × 107 cells) were suspended in PBS (50 μL) containing 1 mM EDTA. After incubation on an ice bath for 10 min, 30 mM Tris-HCl buffer (pH8.5, 267 μL) containing 7 M urea, 2 M thiourea, and 4% CHAPS, and 1 M DTT (17 μL) were added to the suspension, and the mixture was shaken vigorously. Benzonase (5 μL; 25 units/μL) was added to the mixture, and incubated at room temperature for 30 min. To the supernatant obtained by 3326
dx.doi.org/10.1021/ac303771q | Anal. Chem. 2013, 85, 3325−3333
Analytical Chemistry
Article
Figure 1. Fmoc derivatization and fabrication of O-glycan array. (A) Ammonia-catalyzed β-elimination and Fmoc derivatization of O-glycans and (B) Fabrication of glycan array with glycosylamine-form O-glycans. The glycans having intact form (not open chain form) are immobilized on the array.
i.d., 30 cm length) previously equilibrated with 50% aqueous methanol. The earlier eluted fluorescent fractions were pooled and evaporated to dryness under reduced pressure. HPLC Analysis of Fluorescent Labeled O-Glycans. The HPLC system was composed from a SCL-10AVP system controller, two LC-10AD pumps, a DGU-20A3 degasser, a CTO-20AC column oven, and an RF-10AXL fluorescence detector (Shimadzu, Nakagyo-ku, Kyoto, Japan) connected with a data processor (SmartChrom, KYA Technologies, Hachi-oji, Tokyo, Japan). HPLC analysis of 2AA-labeled oligosaccharides was performed according to the method reported previously.19 A polymer-based Asahi Shodex NH2P50 4E column (Showa Denko, Hachi-oji, Tokyo, Japan; 4.6 mm i.d., 250 mm length) was employed. The flow rate was maintained at 1.0 mL/min, and the analysis was done at 50 °C. A linear gradient was formed by 2% acetic acid in acetonitrile (solvent A) and 5% acetic acid in water containing 3% triethylamine (solvent B). The column was initially equilibrated and eluted with 70% solvent A for 2 min, and then solvent B was increased to 95% over 80 min and kept at this composition for a further 100 min. Detection was performed by fluorometry at 350 nm for excitation and 425 nm for emission, respectively. For the analysis of Fmoc labeled oligosaccharides, similar conditions as those for the analysis of 2AA labeled oligosaccharides were used. Initial isocratic elution with 15% solvent B was performed for 2 min followed by a linear increase to 80% solvent B for 80 min. Detection wavelengths were 266 nm for excitation and 310 nm for emission, respectively. MALDI-TOF MS Analysis of Fmoc Labeled O-Glycans. MALDI-TOF MS was performed on an Ultraflextreme (Bruker Daltonics). 2,5-Dihydroxybenzoic acid (DHB) was used as a matrix material throughout the work. The sample solution (1 μL) was mixed with 1% DHB in 30% ethanol on a target plate and dried in atmosphere. The samples were analyzed in the positive or negative mode. Glycan Array Using Fmoc-Labeled O-Glycans. A sample of an Fmoc-labeled O-glycan (100 pmol) collected by HPLC was dissolved in water (20 μL). DMF (30 μL) and morpholine (20 μL) were added to the solution, and the mixture was kept at 37 °C for 30 min. Fmoc residue is released by this procedure, and the glycosylamine-form glycan is obtained with high-yield. After the reaction, diethyl ether (500 μL) was added to the mixture and shaken vigorously. After brief centrifugation, the
aqueous phase was collected and evaporated to dryness by a centrifugal evaporator. The glycosylamine-form glycan thus obtained was directly available for fabrication of glycan array, and was dissolved in 10 μL of basic spotting solution for DNA micro arrays (Matsunami Glass Ind., Ltd., Osaka, Japan). A portion (1 μL: 10 pmol/spot) was spotted on a microarraygrade epoxy-coated glass slide (Schott AG, Mainz, Germany) attached with a silicon rubber sheet to print the glycan spots with a diameter of approximately 1 mm. The glass slide was incubated in the incubation chamber (COSMO BIO Co. Ltd., Tokyo, Japan) containing wet paper at room temperature for 3 h to complete immobilization. After incubation, excess nonimmobilized materials were washed with the probing buffer (25 mM Tris-HCl (pH 7.4) containing 0.8% NaCl, 1% TritonX100, 1 mM MnCl2, and 1 mM CaCl2). And the unreacted epoxy groups were blocked with blocking buffer (25 mM TrisHCl, pH 7.4 containing 0.8% NaCl, 1% Triton-X100, and 4% BSA) at room temperature for 1 h. Binding Assay. The probing buffer (80 μL) containing a biotinylated lectin was applied to the array on which glycosylamine-form glycans were immobilized, and the slide was kept at room temperature for 1.5 h. After removing the biotinylated lectin solution, the probing buffer containing 1 μg/ mL of Cy3-conjugated streptavidin (80 μL) was added to the slide. Then, fluorescent images were acquired using an evanescent-field activated fluorescence scanner (Biorexscan: Rexxam, Takamatsu, Kagawa, Japan). An evanescent wave is created on the surface of the glass slide by internal reflection of the excitation light and excites the probes selectively only when they bind to the immobilized glycans on the surface. This surface-confined excitation, which has a limited penetration depth of about 200 nm into the adjacent probe solution, allows specific detection of bound probes under equilibrium conditions without washing steps.27 Scanning conditions of the cooled charge-coupled device camera were fixed at 5 μm and exposure time (1 s).
■
RESULTS AND DISCUSSION In the present study, O-glycans are released from the core peptides in the presence of ammonia (Figure 1A).16 The glycosylamine-form O-glycans thus released are derivatized with Fmoc-Cl, and analyzed by HPLC/MS techniques. The Fmoc 3327
dx.doi.org/10.1021/ac303771q | Anal. Chem. 2013, 85, 3325−3333
Analytical Chemistry
Article
reaction was performed at 4 °C for 48 h in 0.5 M LiOH, peeling reaction (i.e., formation of NeuAcα2−3Gal) was dominant, and the objective glycan, disialyl-T, could not be recovered (Figure 2; upper panel). When ammonia-catalyzed βelimination was performed at 60 °C for 48 h,16 the yields of the degradation product reached to 89% ± 9% (n = 5) (Figure 2; middle panel). The best yield of disialyl-T was 60% ± 5% (n = 5) under the conditions at 37 °C for 24 h (Figure 2; lower panel). Degradation was suppressed under the optimized conditions and the yields of the degradation product were less than 25%. We previously reported that peeling reaction occurs even when we use the alkali-borohydride method.17 The yields of O-glycans by the present method are equivalent with those observed by the alkaly borohydride method. Fmoc Derivatization of the Released O-Glycans. Because the released O-glycans are recovered as glycosylamine-form, the released O-glycans are easily in situ labeled with Fmoc-Cl. An example for the analysis of O-glycans from bovine submaxillary mucin (BSM) is shown in Figure 3. The released O-glycans was labeled with Fmoc-Cl in phosphate buffer (pH 8.4) and analyzed by MALDI-TOF MS after easy extraction of the excess reagent with chloroform (Figure 3A). Major ion peaks observed at m/z 936.50 and m/z 952.50 were confirmed as Fmoc labeled sialyl-core 3 structures, NeuAcα2−6(GlcNAcβ1−3)GalNAc-Fmoc and NeuGcα2−6(GlcNAcβ1−3)GalNAc-Fmoc, respectively. Two molecular ion peaks observed at m/z 733.42 and m/z 749.42 were due to Fmoc labeled sialyl-Tn, NeuAcα2−6GalNAc-Fmoc and NeuGcα2−6GalNAc-Fmoc, respectively. It was reported that these four O-glycans are the major O-glycans on BSM.17,30 In addition, Fmoc labeled monosialylated core 1 (m/z 895.48) and core 2 (m/z 1098.56 and m/z 1114.55) glycans were also observed. Because two sialic acid species (NeuAc and NeuGc) are present in the sialoglycans in BSM, a set of two molecular ions having (NeuAc or NeuGc) residue were observed. It should be noticed that the molecular ion peaks of the unlabeled O-glycans (i.e., free-form glycans) were not observed. These results clearly show that the released O-glycans by ammoniacatalyzed β-elimination were efficiently labeled with Fmoc-Cl without hydrolysis of glycosylamine-form to free-form. Fmoc derivatives were successfully analyzed by HPLC in the similar conditions as those employed for the separation of 2AA labeled oligosaccharides with an amine-bonded polymer column (Figure 3B; upper panel). Relative peak areas of Fmoc and 2AA derivatives of the O-glycans were compared (Table 1), and the ratios show similar values for both derivatives. We also examined the fluorescent intensities in Fmoc and 2-AA methods by comparing total peak areas after injecting the same amount of O-glycan mixture. The result showed that the total peak areas of Fmoc labeled O-glycans were 3.5 times bigger than those of 2AA O-glycans as reported previously.25 The lower detection limit for O-glycan analysis was determined using BSM as model (Figure 4). O-Glycans released from 10 μg of BSM were labeled with Fmoc-Cl, and a portion was analyzed by HPLC. Even at the 10 ng-level of the sample as the injected amount, we clearly observed the major O-glycans. Recovery of Free O-Glycans from Fmoc Derivatives. The one of the most important and useful points of the present method is that Fmoc labeled O-glycans are easily recovered as free-form under mild conditions. When Fmoc labeled sialyl-Tn (NeuAcα2−6GalNAc-Fmoc) derived from 1 μg of BSM was incubated in a mixture of morpholine and DMF, the peak of
derivative of O-glycan was collected by HPLC, and recovered as glycosylamine-form in aqueous DMF-morpholine solution. Because the amino group of the glycosylamine reacts readily with epoxy group at room temperatur,24 glycosylamine-form glycans can be immobilized onto the epoxy-coated slide. The strong point of the present method is that we can immobilize O-glycans having the intact reducing monosaccharide onto the epoxy-coated slide (Figure 1B). Optimization of Ammonia-Catalyzed β-Elimination. Releasing of O-glycans from the core protein with high efficiency without peeling reaction has been a quite big problem. Huang et al. reported ammonia-catalyzed βelimination for the first time. Although they did not evaluate degradation of O-glycans during elimination reaction, the method was applied to the analysis of O-glycans.16,28 Recently, Miura et al. evaluated the ammonia-catalyzed β-elimination by MS technique.29 However, optimization for ammonia-catalyzed β-elimination was not performed either.We optimized the conditions for ammonia-catalyzed β-elimination using a model glycopeptide derived from caseinoglycomacropeptide (cow milk) having the sequence GEPTSTPT, which has disialyl-T antigen (NeuAcα2−3Galβ1−3(NeuAcα2−6)GalNAc) at the fourth Thr residue (shown in bold face with underline) from the N-terminal (Figure 2).
Figure 2. Optimization for glycan releasing reaction with ammoniacatalyzed β-elimination. Analytical condition for HPLC: column, Asahi Shodex NH2P-50 4E (4.6 × 250 mm); eluent, solvent A, 2% CH3COOH in acetonitrile; solvent B, 5% CH3COOH/3% triethylamine in water, gradient condition, linear gradient (30−95% solvent B) from 2 to 82 min, maintained for 20 min. O-Glycans released from the model glycopeptide by the β-elimination were analyzed by HPLC after labeling with 2AA. The lower panel shows the results under the optimized conditions using ammonia-catalyzed β-elimination. The middle panel shows the results under the previously reported conditions using ammonia-catalyzed β-elimination.16 The upper panel shows the results in an aqueous LiOH solution which catalyzes β-elimination without the reducing reagents.
NeuAcα2−3Galβ1−3-residue of the disialyl-T antigen is labile under alkaline conditions. When disialyl-T antigen is released from the core peptide, the degradation product (NeuAcα2−3Gal) is spontaneously formed by β-elimination (i.e., peeling reaction). We calculated the yields of disialyl-T and degradation products using peak areas observed in Figure 2. To investigate the repeatability of these results, these examinations were repeated five times. When the releasing 3328
dx.doi.org/10.1021/ac303771q | Anal. Chem. 2013, 85, 3325−3333
Analytical Chemistry
Article
Figure 3. Analysis of O-glycans from BSM after derivatization with Fmoc. Fmoc-labeled O-glycans from BSM were analyzed by MALDI-TOF MS (A) and HPLC (B). Analytical conditions for HPLC: column, Asahi Shodex NH2P-50 4E (4.6 × 250 mm); eluent, solvent A, 2% CH3COOH in acetonitrile; solvent B, 5% CH3COOH/3% triethylamine in water, gradient condition for 2AA labeled O-glycans, linear gradient (30−95% solvent B) from 2 to 82 min, maintained for 20 min, gradient condition for Fmoc labeled O-glycans, linear gradient (15−80% solvent B) from 2 to 82 min, maintained for 20 min.
Table 1. Comparison of Relative Abundances of Mucin-Type Glycans and Total Peak Areas between Fmoc Labeling Method and 2AA Labeling Methoda NeuAcα2−6GalNAc NeuGcα2−6GalNAc NeuAcα2−6(GlcNAcβ1−3)GalNAc NeuGcα2−6(GlcNAcβ1−3)GalNAc other mucin-type glycansb total peak area a
Fmoc
2AA
ratio (Fmoc/2AA)
45% 25% 13% 6% 11% 43278937
47% 25% 15% 7% 6% 12299776
1.0 1.0 0.9 0.9 1.8 3.5
Fmoc-labeled glycans and 2AA labeled glycans (both 1 μg as BSM) were injected to HPLC. bMinor glycans observed in BSM (see Figure 3).
sialyl-Tn was completely disappeared after incubation for 30 min at 37 °C (Figure 5, panels A and B, respectively). The recovered sialyl-Tn has still glycosylamine-form, and is available for immobilization to glycan array. To recover sialyl-Tn with hemiacetal-form, the dried samples obtained by releasing reaction of Fmoc group has to be hydrolyzed by keeping in 0.5 M boric acid at 37 °C for 30 min. The free-form sialyl-Tn thus obtained was labeled with 2AA and analyzed by HPLC. As shown in Figure 5C, Fmoc-labeled sialyl-Tn (Figure 5A) was quantitatively converted to 2AA-
labeled sialyl-Tn. High response of Fmoc derivative (Figure 5A) is due to strong fluorescence of Fmoc group. Analysis of Fmoc-Labeled O-Glycans in Biological Samples. O-Glycans of bovine fetuin and porcine stomach mucin (PSM) were released by ammonia-catalyzed βelimination followed by derivatization with Fmoc-Cl, and analyzed by MALDI-TOF MS and NP-HPLC (Figure 6). Five glycans were observed in bovine fetuin (Figure 6A). The molecular ion peak at m/z 895.48 was assigned as sialyl-T (NeuAcα2−3Galβ1−3GalNAc-Fmoc). The molecular ion peak 3329
dx.doi.org/10.1021/ac303771q | Anal. Chem. 2013, 85, 3325−3333
Analytical Chemistry
Article
Figure 4. Detection limit of the Fmoc derivatization method as examined using BSM as a model sample. Fmoc-labeled O-glycans from BSM were analyzed by HPLC. Analytical conditions for HPLC were the same as those in Figure 3. As indicated in the figure, 10 ng of the sample as the injected amount could be successfully analyzed.
Figure 6. Analysis of Fmoc labeled O-glycans obtained from fetuin (A) and porcine stomach mucin (B). Fmoc-labeled O-glycans from fetuin (A) and PSM (B) were analyzed by MALDI-TOF MS and HPLC. Analytical conditions for HPLC were the same as those in Figure 3. Detailed structures of O-glycans observed in PSM are shown in the Supporting Information.
Figure 5. Conversion of Fmoc derivatives to 2AA labeled oligosaccharides after recovery of free-form O-glycans. Fmoc-labeled sialyl-Tn derived from BSM was converted to free-form oligosaccharides and analyzed by HPLC. (A) Fmoc labeled sialyl-Tn obtained from 1 μg of BSM. (B) Fmoc-labeled sialyl-Tn was converted to free-form. (C) Free-form sialyl-Tn (panel B) was labeled with 2AA. Analytical conditions for HPLC were the same as those in Figure 3. We could not find any ion peaks due to glycans from these peaks by MALDITOF MS analysis. Therefore, asterisks in C are due to artifacts during releasing reaction of Fmoc group and 2AA derivatization.
were assigned as described in the figure. The Fmoc labeled glycans released from PSM showed quite complex MS profiles, and 34 glycans in total were observed (Figure 6B), and their structures were confirmed by the previous reports17,30 and database searching results using Glycopeakfinder in EUROCarbDB (http://www.ebi.ac.uk/eurocarb/tools.action; Table S1). Asialo fucosylated core 2 and core 4 type mucin-type glycans were observed as the major components (m/z 752.43, m/z 977.51, m/z 1139.57, m/z 1342.65, m/z 1504.70, m/z 1545.73, m/z 1650.76, m/z 1707.78, m/z 1853.84, m/z 1869.83, and m/z 1910.86). In addition, sulfated mucin-type glycans were also observed (m/z 708.38, m/z 911.45, m/z 930.48, m/z 1219.57, m/z 1422.65, m/z 1584.70, m/z 1787.78, and m/z 1991.86). The complex chromatogram was observed by NP-HPLC analysis (Figure 6B; lower panel). Asialo Oglycans were observed from 5 to 35 min (Peaks 1−15) in the order of their molecular weights. Sulfated O-glycans were observed from 35 to 40 min (Peak 16--19). Although further studies are required, these results indicate that sulfated and
observed at m/z 1098.56 corresponds to monosialo core 2 structure, NeuAcα2−3Galβ1−3(GlcNAcβ1−6)GalNAc-Fmoc. Two molecular ion peaks at m/z 1186.57 and 1551.70 were due to disialo O-glycans, NeuAcα2−3Galβ1−3(NeuAcα2−6)GalNAc-Fmoc and NeuAcα2−3Galβ1−3(NeuAcα2−3Galβ1− 4GlcNAcβ1−6)GalNAc-Fmoc, respectively. The molecular ion peak observed at m/z 692.40 is due to NeuAcHexFmoc and is assigned as the degradation product, NeuAcα2−3Gal-Fmoc. These results are well consistent with those reported previously.17,30 Fmoc labeled O-glycans from fetuin were also analyzed by NP-HPLC (Figure 6A, lower panel). Five glycans were also observed with good resolution within 45 min and 3330
dx.doi.org/10.1021/ac303771q | Anal. Chem. 2013, 85, 3325−3333
Analytical Chemistry
Article
Figure 7. Analysis of Fmoc-labeled O-glycans from MKN45 cells. Fmoc-labeled O-glycans from MKN45 cells were analyzed by MALDI-TOF MS (A) and by HPLC (B). Analytical conditions for HPLC were the same as those in Figure 3. Detailed structures are shown in the Supporting Information.
polylactosamine-type O-glycans. In addition, MKN45 cells express the trisialo core1-type glycans having the biantennary polylactosamine unit. These O-glycans were specifically observed in poorly differentiated gastric cancer cell lines.19,20 We suggested that these O-glycans would be specific markers of malignancy and differentiation of gastric adenocarcinoma. Fabrication of Glycan Array Using Fmoc-Labeled OGlycans. Purified Fmoc-labeled glycans derived from BSM and fetuin were converted to glycosylamine-form and spotted on an epoxy group-coated slide as shown in Figure 1B. A polylactosamine-type glycan derived from MKN45 cells was also employed. After immobilizing the O-glycans onto the epoxide slide according to the procedures provided by the manufacturer, the interactions between some lectins and the immobilized glycans were examined by an evanescent-field fluorescence-assisted scanner (Figure 8).
fucosylated O-glycans in PSM can be successfully analyzed by the present method. We also applied the present method to the analysis of Oglycans expressed on MKN45 cells (a human stomach cancer cell line). Fmoc-labeled O-glycans in MKN45 cells were analyzed by MALDI-TOF MS and NP-HPLC (Figure 7). The list of the proposed structures based on our previously reported data is summarized in the Supporting Information (Table S2). Figure 7A indicated the presence of various asialo, monosialo, disialo, and trisialo polylactosaminyl O-glycans (Figure 7A). These Fmoc labeled polylactosamine-type glycans observed in MKN45 cells were also successfully separated by NP-HPLC (Figure 7B). Asialo, monosialo, disialo, and trisialo polylactosamine-type glycans were observed at 15−25, 25−40, 40−60, and 65−75 min, respectively. We previously reported that MKN45 cells express extremely large amount of 3331
dx.doi.org/10.1021/ac303771q | Anal. Chem. 2013, 85, 3325−3333
Analytical Chemistry
Article
We applied the present methods to the analysis of O-glycans expressed on MKN45 cells and found that large size polylactosamine-type O-glycans were present abundantly as reported previously.19 These results showed that our methods are useful for the analysis of O-glycans in various biological samples. It is one of the most strong points that purified Fmoc-labeled oligosaccharides are easily converted to free-form oligosaccharides after collection by HPLC. We succeeded in recovery of free-form O-glycans quantitatively. In addition, we achieved immobilization of Fmoc labeled O-glycans to epoxide slide via conversion of Fmoc derivatives to glycosylamine-form. Because the O-glycans have the intact reducing monosaccharide residues for glycan array, it should be emphasized that the present method precisely mimics the glycans attached to the proteins. The proposed methods allow construction of the library of O-glycans having large molecular sizes, which cannot be prepared by the present technologies. We believe that the present methods will be powerful tools for discovering the new disease markers and novel functions based on O-glycans in glycoproteins.
Figure 8. Glycan array analysis using Fmoc-labeled O-glycans isolated from mucin samples and MKN45 cells. Fmoc-labeled O-glycans derived from BSM, fetuin and MKN45 cells were convert to glycosylamine-form and immobilized to epoxy-coated glass slide. Cy3 labeled lectins were added, and interactions between Cy3 lectins and glycans on the array were analyzed by an evanescent-field fluorescence-assisted scanner.
■
ASSOCIATED CONTENT
S Supporting Information *
Wheat germ agglutinin (WGA) recognizes N-acetyl group on GlcNAc and NeuAc residues.31 The immobilized glycans, sialylTn (NeuAcα2−6GalNAc) and sialyl-core 3 (NeuAcα2− 6(GlcNAcβ1−3)GalNAc, were clearly recognized by WGA. Sialyl-T shows distinct interaction with Jacalin which recognizes T and Sialyl-T andtigens.32 We also succeeded in detection of the interaction between GalNAc and VVAB4 lectin which specifically recognizes Tn antigen.33 Tn antigen (i.e., GalNAc residue) was detected with extremely high sensitivity even at the pM level. Because immobilized glycans often have the open ring form by most conventional methods,24 it is difficult to detect the interaction with the monosaccharide and carbohydrate-binding protein. Our methods allow immobilization of glycans without reduction of the reducing terminal (i.e., the terminal monosaccharide keeps pyranose form). This is the most important and strong point of the present method. An Fmoc-labeled disialo polylactosamine-type O-glycan as indicated in Figure 8 (the lowest panel) was collected from MKN45 cells and immobilized onto the epoxide slide. DSA lectin which recognizes repeating units of Gal-GlcNAc34 can detect this glycan at the 200 pM level (16 fmol as the spotted amount). Preparation of such large glycan as pure state is quite difficult by the present technologies. Our methods allow usages of various types O-glycans from natural sources. We are constructing the library of Fmoc labeled O-glycans and fabricating glycan arrays, and the results will be reported elsewhere.
Additional information as noted in text. This material is available free of charge via the Internet at http://pubs.acs.org.
■
AUTHOR INFORMATION
Corresponding Author
*E-mail:
[email protected]. Tel: +80-6-6721-2332. Fax: +80-6-6721-2353. Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS We thank Mr. H. Sumida, and Mr. Y. Ushida, Rexxam Co., Ltd., for the assistance to use an evanescent-field activated fluorescence scanner (Biorexscan). This research is supported by the fund for Kagawa University Young Scientists 2011 and JSPS KAKENHI Grant No. 24790044.
■
REFERENCES
(1) Guzman-Aranguez, A.; Argueso, P. Ocul. Surf. 2010, 8, 8−17. (2) Fukuda, M. Biochim. Biophys. Acta 2002, 1573, 394−405. (3) Chen, H. D.; Zhou, X.; Yu, G.; Zhao, Y. L.; Ren, Y.; Zhou, Y. D.; Li, Q.; Zhang, X. L. J. Clin. Immunol. 2012, 32, 820−836. (4) Tsuboi, S.; Sutoh, M.; Hatakeyama, S.; Hiraoka, N.; Habuchi, T.; Horikawa, Y.; Hashimoto, Y.; Yoneyama, T.; Mori, K.; Koie, T.; Nakamura, T.; Saitoh, H.; Yamaya, K.; Funyu, T.; Fukuda, M.; Ohyama, C. EMBO J. 2011, 30, 3173−3185. (5) Suzuki, Y.; Sutoh, M.; Hatakeyama, S.; Mori, K.; Yamamoto, H.; Koie, T.; Saitoh, H.; Yamaya, K.; Funyu, T.; Habuchi, T.; Arai, Y.; Fukuda, M.; Ohyama, C.; Tsuboi, S. Int. J. Oncol. 2012, 40, 1831− 1838. (6) Jensen, P. H.; Kolarich, D.; Packer, N. H. FEBS. J. 2010, 277, 81− 94. (7) An, H. J.; Kronewitter, S. R.; de Leoz, M. L.; Lebrilla, C. B. Curr. Opin. Chem. Biol. 2009, 13, 601−607. (8) Iwai, T.; Kudo, T.; Kawamoto, R.; Kubota, T.; Togayachi, A.; Hiruma, T.; Okada, T.; Kawamoto, T.; Morozumi, K.; Narimatsu, H. Proc. Natl. Acad. Sci. U. S. A. 2005, 102, 4572−4577. (9) Brockhausen, I. Biochim. Biophys. Acta 1999, 1473, 67−95. (10) Brockhausen, I. Adv. Exp. Med. Biol. 2003, 535, 163−188. (11) Springer, G. F. Science 1984, 224, 1198−1206.
■
CONCLUSION We developed a novel method for the analysis of O-glycans using a combination of ammonia-catalyzed β-elimination and Fmoc derivatization. We revealed that higher than 80% of Oglycans were degraded by ammonia-catalyzed β-elimination under the conventional conditions, and optimized the ammonia-catalyzed β-elimination. The released O-glycans having glycosylamine-form were successfully derivatized with Fmoc-Cl. Fmoc-labeled O-glycans thus prepared showed 3.5 times higher sensitivities than those labeled with 2AA, and Oglycans obtained from 10 ng of bovine submaxillary mucin as the injected amount could be detected. 3332
dx.doi.org/10.1021/ac303771q | Anal. Chem. 2013, 85, 3325−3333
Analytical Chemistry
Article
(12) Tarp, M. A.; Sorensen, A. L.; Mandel, U.; Paulsen, H.; Burchell, J.; Taylor-Papadimitriou, J.; Clausen, H. Glycobiology 2007, 17, 197− 209. (13) Kakehi, K.; Susami, A.; Taga, A.; Suzuki, S.; Honda, S. J. Chromatogr. A 1994, 680, 209−215. (14) Schulz, B. L.; Packer, N. H.; Karlsson, N. G. Anal. Chem. 2002, 74, 6088−6097. (15) Royle, L.; Mattu, T. S.; Hart, E.; Langridge, J. I.; Merry, A. H.; Murphy, N.; Harvey, D. J.; Dwek, R. A.; Rudd, P. M. Anal. Biochem. 2002, 304, 70−90. (16) Huang, Y.; Mechref, Y.; Novotny, M. V. Anal. Chem. 2001, 73, 6063−6069. (17) Yamada, K.; Hyodo, S.; Matsuno, Y. K.; Kinoshita, M.; Maruyama, S. Z.; Osaka, Y. S.; Casal, E.; Lee, Y. C.; Kakehi, K. Anal. Biochem. 2007, 371, 52−61. (18) Yamada, K.; Hyodo, S.; Kinoshita, M.; Hayakawa, T.; Kakehi, K. Anal. Chem. 2010, 82, 7436−7443. (19) Yamada, K.; Kinoshita, M.; Hayakawa, T.; Nakaya, S.; Kakehi, K. J. Proteome Res. 2009, 8, 521−537. (20) Yamada, K.; Mitsui, Y.; Kakoi, N.; Kinoshita, M.; Hayakawa, T.; Kakehi, K. Anal. Biochem. 2012, 421, 595−606. (21) Wang, C.; Fan, W.; Zhang, P.; Wang, Z.; Huang, L. Proteomics 2011, 11, 4229−4242. (22) Furukawa, J.; Fujitani, N.; Araki, K.; Takegawa, Y.; Kodama, K.; Shinohara, Y. Anal. Chem. 2011, 83, 9060−9067. (23) Suzuki, S.; Fujimori, T.; Yodoshi, M. Anal. Biochem. 2006, 354, 94−103. (24) de Boer, A. R.; Hokke, C. H.; Deelder, A. M.; Wuhrer, M. Anal. Chem. 2007, 79, 8107−8113. (25) Kamoda, S.; Nakano, M.; Ishikawa, R.; Suzuki, S.; Kakehi, K. J. Proteome Res. 2005, 4, 146−152. (26) Nakano, M.; Higo, D.; Arai, E.; Nakagawa, T.; Kakehi, K.; Taniguchi, N.; Kondo, A. Glycobiology 2009, 19, 135−143. (27) Kuno, A.; Uchiyama, N.; Koseki-Kuno, S.; Ebe, Y.; Takashima, S.; Yamada, M.; Hirabayashi, J. Nat. Methods 2005, 2, 851−856. (28) Yu, G.; Zhang, Y.; Zhang, Z.; Song, L.; Wang, P.; Chai, W. Anal. Chem. 2010, 82, 9534−9542. (29) Miura, Y.; Kato, K.; Takegawa, Y.; Kurogochi, M.; Furukawa, J.; Shinohara, Y.; Nagahori, N.; Amano, M.; Hinou, H.; Nishimura, S. Anal. Chem. 2010, 82, 10021−10029. (30) Karlsson, N. G.; Packer, N. H. Anal. Biochem. 2002, 305, 173− 185. (31) Peters, B. P.; Ebisu, S.; Goldstein, I. J.; Flashner, M. Biochemistry 1979, 18, 5505−5511. (32) Tachibana, K.; Nakamura, S.; Wang, H.; Iwasaki, H.; Maebara, K.; Cheng, L.; Hirabayashi, J.; Narimatsu, H. Glycobiology 2006, 16, 46−53. (33) Konska, G.; Guerry, M.; Caldefie-Chezet, F.; De Latour, M.; Guillot, J. Oncol. Rep. 2006, 15, 305−310. (34) Yamashita, K.; Totani, K.; Ohkura, T.; Takasaki, S.; Goldstein, I. J.; Kobata, A. J. Biol. Chem. 1987, 262, 1602−1607.
3333
dx.doi.org/10.1021/ac303771q | Anal. Chem. 2013, 85, 3325−3333