Rapid and Sensitive Screening of N-Glycans as 9-Fluorenylmethyl

KIRIN BREWRY Co., Ltd., Hagiwara-machi 100-1, Takasaki 370-0013, Japan. Received September 29, 2004. There are a large number of labeling methods for ...
0 downloads 0 Views 180KB Size
Rapid and Sensitive Screening of N-Glycans as 9-Fluorenylmethyl Derivatives by High-Performance Liquid Chromatography: A Method Which Can Recover Free Oligosaccharides after Analysis Satoru Kamoda,†,‡ Miyako Nakano,† Rika Ishikawa,‡ Shigeo Suzuki,† and Kazuaki Kakehi*,† Faculty of Pharmaceutical Sciences, Kinki University, Kowakae 3-4-1, Higashi-osaka 577-8502, Japan and KIRIN BREWRY Co., Ltd., Hagiwara-machi 100-1, Takasaki 370-0013, Japan Received September 29, 2004

There are a large number of labeling methods for asparagine-type oligosaccharides with fluorogenic and chromophoric reagents. We have to choose the most appropriate labeling method based on the purposes such as mass spectrometry, high-performance liquid chromatography and capillary electrophoresis. Asparagine-type glycans are released from core proteins as N-glycosylamine at the initial step of the releasing reaction when glycoamidase F is employed as the enzyme. The N-glycosylaminetype oligosaccharides thus released by the enzyme are subjected to hydrolysis or mutarotation to form free-form oligosaccharides. In the detailed studies on the enzyme reaction, we found a condition in which the released N-glycosylamine-type oligosaccharides were exclusively present at least during the course of enzyme reaction, and developed a method for in situ derivatization of the glycosylaminetype oligosaccharides with 9-fluorenylmethyl chloroformate (Fmoc-Cl). The Fmoc labeled sialo- and asialo- (or high-mannose and hybrid) oligosaccharides were successfully analyzed on an amine-bonded polymer column and amide-silica column, respectively. The present method showed approximately 5 times higher sensitivities than that using 2-aminobenzoic acid (2-AA). The separation profile was similar to that observed using 2-AA method as examined by the analyses of carbohydrate chains derived from several glycoproteins including complex-type, high-mannose type and hybrid type of N-linked oligosaccharides. The labeled oligosaccharides were stable at least for several months when stored at -20 °C. Furthermore, it should be emphasized that the Fmoc-derivatized oligosaccharides could be easily recovered as free reducing oligosaccharides simply by incubation with morpholine in dimethylformamide solution. We obtained a pure triantennary oligosaccharide with 3 sialic acid residues as a free reducing form from fetuin in good yield after isolation of the corresponding Fmoc oligosaccharide followed by removing reaction of the Fmoc group. The proposed method will be useful for preparation of free oligosaccharides as standard samples at pmol-nmol scale from commercially available glycoproteins. Keywords: oligosaccharide mapping • glycoprotein • 9-fluorenylmethyl chloroformate (Fmoc-Cl)

Introduction Various techniques for oligosaccharide mapping have been developed to achieve higher sensitivity and resolution using high-performance liquid chromatography (HPLC), capillary electrophoresis (CE), and polyacrylamide gel electrophoresis (PAGE).1 High-pH anion exchange chromatography with pulsed amperometric detection (HPAEC-PAD) is also available for the analysis of free oligosaccharides.2 There are a large number of labeling methods with fluorogenic and chromophoric reagents for the analysis of asparagine-type oligosaccharides using the methods described above.3,4 The methods using 2-aminopyridine (2-AP) or 2-ami* To whom correspondence should be addressed. Tel: +81-6-67212332. Fax: +81-6-6721-2353. E-mail: [email protected]. † Faculty of Pharmaceutical Sciences, Kinki University. ‡ KIRIN BREWRY Co., Ltd..

146

Journal of Proteome Research 2005, 4, 146-152

Published on Web 01/06/2005

nobenzamide (2-AB) are most widely used for HPLC,5,6 and those using 8-aminopyrene-1,3,6-trisulfonate (APTS)7,8 and 2or 3-aminobenzoic acid (2-AA or 3-AA)9 are important for highresolution analysis of oligosaccharides by CE based on their negative charges. Some other naphthalene sulfonic acids are also employed for the analysis of oligosaccharides on slab gel format using fluorophore-assisted carbohydrate electrophoresis (FACE) technique.10 We have to choose the most appropriate labeling method based on the purposes such as mass spectrometry, HPLC and CE. These labeling methods usually involve attaching a fluorophore or a chromophore to the reducing terminus using such as 2-AP,5 4-aminobenzoic 2-(diethylamino)ethyl ester (ABDEAE),11,12 2-aminoacridone (AMAC),13 2-AB,6 2-AA,9,14 and 3-AA.9 All of these derivatizations are based on the reductive amination, and commonly require two-step reactions (i.e., 10.1021/pr049825o CCC: $30.25

 2005 American Chemical Society

Rapid and Sensitive Screening of N-Glycans

Schiff base formation and reduction of the Schiff base), and usually need tedious purification steps of the labeled oligosaccharides.4,15 Furthermore, since derivatization must be carried out under acidic conditions, the reaction is performed at relatively low temperature for a long time to avoid hydrolysis of sialic acid residues. Preparation of standard samples of highly purified free oligosaccharides is quite important, because these oligosaccharides are easily converted to various forms for the studies such as carbohydrate-protein interactions.16-18 If pure oligosaccharides can be obtained from natural source, then the library of oligosaccharides is useful for construction of sugar array, which offers a possibility for high-throughput analysis of interactions between carbohydrates and proteins (i.e., ‘glycome’).19 On the basis of these considerations, many challenges of chemical, chemoenzymatic, and biological synthesis of oligosaccharide have been reported.20-24 Although these strategies are useful for the synthesis of target carbohydrates in mgquantities or more, they are problematic in the preparation of diverse structures of oligosaccharides because they require multistep reactions which cause much labor. If we can obtain free oligosaccharides from natural sources such as, for example, commercially available glycoproteins, then use of a combination of glycosidases or glycosyltransferases allows preparation of wide variety of oligosaccharides for various purposes.25-27 However, there are only a few papers on regeneration of free oligosaccharides from the labeled carbohydrates,28,29 and it is generally difficult to recover free oligosaccharides from the corresponding fluorescent oligosaccharides. In the present study, we developed a new fluorescent derivatization of N-linked oligosaccharides released from glycoproteins using 9-fluorenylmethyl chloroformate (Fmoc-Cl), which is widely used as a protecting reagent for amino groups during peptide synthesis,30 and also used for derivatization of primary and secondary amines.31 We also show easy recovery of free oligosaccharides isolated from the complex mixture of oligosaccharides labeled with Fmoc.

Experimental Section Materials. Peptide N4-(acetyl-β-D-glucosaminyl)asparagine amidase (PNGase F; EC 3.2.2.18, recombinant) was obtained from Roche Diagnostics (Mannheim, Germany). 9-Fluorenylmethyl chloroformate (Fmoc-Cl) and 2-aminobenzoic acid (2AA) were obtained from Tokyo Kasei (Chuo-ku, Tokyo, Japan). Sodium cyanoborohydride was obtained from Aldrich (Milwaukee, WI). Fetuin (bovine) was obtained from Gibco (Invitrogen, Chuo-ku, Tokyo, Japan). R1-acid glycoprotein (human), transferrin (human), fibrinogen (human), ribonuclease B (bovine pancreas) and thyroglobulin (porcine) were obtained from Sigma (St. Louis, MO). Ovalbumin was purified from chicken eggs according to the method reported by Keckwick and Cannan.32 A pharmaceutical preparation of recombinant immunoglobulin (rIgG), trastuzumab, was kindly donated from Ms. Nishiura of Kinki University Nara Hospital. The solution of rIgG was dialyzed against distilled water for 3 days with changing water several times at 4 °C using cellulose membrane tubing (Sanko Junyaku, Chiyoda-ku, Tokyo, Japan), and then freeze-dried. Other reagents and solvents used in the present study were the reagent grade or HPLC grade and purchased from Wako (Dosho-machi, Osaka, Japan). Releasing of N-Linked Oligosaccharides and Fmoc Derivatization. Releasing of N-linked oligosaccharides from a glycoprotein sample followed by labeling with Fmoc was performed

research articles in one-pot reaction. Briefly, a sample of glycoprotein (1001000 µg) was dissolved in 100 µL of 20 mM phosphate buffer (pH 8.5) in a sample tube (1.5 mL). PNGase F (10 units, 10 µL) was added to the mixture, and incubated at 37 °C for 2 h. After the mixture was diluted with water (300 µL), freshly prepared solution of Fmoc-Cl in acetone (200 µL, 5 mg/mL) was added, and the mixture was incubated at 37 °C for 1 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. Finally, the aqueous layer containing Fmoc oligosaccharides was evaporated to dryness by a centrifugal evaporator (SpeedVac, Savant, Farmingdale, NY). The residue was dissolved in 100 µL of water and a portion (typically 10 µL) was used for the analysis by HPLC. The dried samples were stable at least for several months at -20 °C. Releasing of N-Linked Oligosaccharides and 2-AA-Derivatization. Fetuin (200 µg) was dissolved in 100 µL of 20 mM phosphate buffer (pH 7.0) in a sample tube (1.5 mL). PNGase F (10 units, 10 µL) was added to the mixture, and the mixture was incubated at 37 °C overnight. After digestion, acetic acid (1 µL) was added to the solution to change glycosylamine to reducing oligosaccharides. Derivatization of the released oligosaccharides with 2-AA and purification of the labeled oligosaccharides were carried out as described previously.14 Briefly, the enzyme reaction mixture was dried and dissolved in water (20 µL). A portion (100 µL) of the freshly prepared solution of 2-AA (30 mg) and sodium cyanoborohydride (30 mg) in methanol (1 mL) containing 4% (w/v) sodium acetate and 2% (w/v) boric acid was added and mixed with the oligosaccharide solution. The mixture was kept at 80 °C for 60 min. After cooling, the solution was diluted with 1.0 mL of acetonitrile-water (95:5) and mixed vigorously. The mixture was applied to an Oasis HLB cartridge (1 mL, Waters, Milford, MA) previously equilibrated with the same solvent (1 mL × 2). The cartridge was washed with acetonitrile-water (95:5, 1 mL × 2). Bound oligosaccharides were eluted with acetonitrilewater (20:80, 1 mL) and the eluate was evaporated to dryness by a centrifugal evaporator. The residue was dissolved in 200 µL of water and a portion (10 µL) was used for the analysis by HPLC. High-Performance Liquid Chromatography (HPLC) of Fluorescent Labeled Oligosaccharides. The HPLC system was composed from a SCL-10A system controller (Shimadzu, Nakagyo-ku, Kyoto, Japan), two LC-10AD pumps (Shimadzu), a DGU-12A degasser (Shimadzu), a 655A-52 column oven (Hitachi, Chiyoda-ku, Tokyo, Japan), and a FP-920 fluorescence detector (JASCO, Hachi-oji, Tokyo, Japan) connected with a data processor (SmartChrom, KYA Technologies, Hachi-oji, Tokyo, Japan). HPLC analysis of 2-AA labeled oligosaccharide was performed according to the method reported previously.14,33 Separation was done at 50 °C with an amine-bonded polymeric column (Shodex Asahipak-NH2P-50 4E, 4.6 × 250 mm, Showa denko, Minato-ku, Tokyo, Japan) using a linear gradient formed by 2% (v/v) 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 30% solvent B for 2 min, at which point solvent B was increased to 95% over 80 min and kept at this composition for further 100 min. The flow rate was maintained at 1.0 mL/min through the analysis. Detection was performed by fluorometry with 350 nm for excitation and 425 nm for emission, respectively. For the analysis of Fmoc labeled oligosaccharides, similar Journal of Proteome Research • Vol. 4, No. 1, 2005 147

research articles conditions as those for the analysis of 2-AA labeled oligosaccharides were used. Initial isocratic elution with 20% solvent B was performed for 10 min followed by a linear increase to 90% solvent B for 80 min. Then the column was washed with 95% B for 15 min. Detection wavelengths were 266 nm for excitation and 310 nm for emission, respectively. To achieve higher resolution among neutral oligosaccharides, we employed an amide-silica column (TSK-gel Amide-80, 4.6 × 250 mm, Tosoh, Minato-ku, Tokyo, Japan)14 using a linear gradient formed by 0.1% (v/v) acetic acid in acetonitrile (solvent A) and 0.2% acetic acid in water containing 0.2% triethylamine. Separations were done at 30 °C at a flow rate of 1.0 mL/min. The column was initially equilibrated with 23% solvent B for 5 min, at which point solvent B was increased to 44% over 75 min and kept at 95% B for 10 min. The column was equilibrated with initial conditions for 15 min prior to the next injection. Fmoc labeled oligosaccharides were detected in the similar manner as described above (266 nm excitation and 310 nm emission wavelength). Recovery of Free Oligosaccharides from Fmoc Labeled Oligosaccharide. Free oligosaccharides were recovered from Fmoc labeled oligosaccharides in the similar manner as used for release of Fmoc group from the amino terminal of synthetic glycopeptides as reported by Kajihara et al.34 Each of the Fmoc labeled oligosaccharides collected by HPLC was dissolved in 20 µL of water. Dimethylformamide (DMF, 30 µL) and morpholine (20 µL) were added, and incubated 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. The total efficiencies of the recovery of free oligosaccharides from released oligosaccharides were estimated according to the procedures as described in Figure 1. In the initial step, we prepared a mixture of oligosaccharides from fetuin by digestion with PNGase F under the optimized condition for Fmoc derivatization. A half of the mixture was fluorescently labeled with 2-AA after complete conversion of glycosylamines to reducing oligosaccharides by addition of acetic acid. Another half portion was labeled with Fmoc, and analyzed by HPLC. We collected the peak of trisialo-triantennary oligosaccharide (Peak 3 in Figure 3a, whose structure is A3S3 in Figure 5) and the Fmoc-A3S3 was converted to free form and then labeled to 2-AA derivative. We compared peak areas of A3S3 obtained by direct 2-AA derivatization with those from isolated Fmoc-A3S3 through conversion to free form followed by 2-AA labeling. Matrix-Assisted Laser Desorption/Ionization Time-of-Flight Mass Spectrometry (MALDI-TOF/MS) Analysis of Oligosaccharides. MALDI-TOF/MS analysis of Fmoc labeled oligosaccharides was performed on a Voyager DE-PRO apparatus (PE Biosystems, Framingham, MA). A nitrogen laser was used to irradiate samples at 337 nm, and an average of 50 shots was taken. The instrument was operated in linear mode using negative polarity at an acceleration voltage of 20 kV. Collected peaks by HPLC analysis of Fmoc oligosaccharide mixture derived from fetuin (150 µg) were dried and dissolved in water (20 µL). A portion of the solution (1 µL) was applied to a stainless steel target, to which was added a solution (1 µL) of 2,5-dihydroxybenzoic acid (DHB) in methanol-water (1:1). The target was dried in ambient temperature for several minutes. A mixture of 2-AA labeled glucose oligomers was used for calibration of molecular masses. 148

Journal of Proteome Research • Vol. 4, No. 1, 2005

Kamoda et al.

Figure 1. Evaluation of recovery of free oligosaccharide through Fmoc derivatization method.

Figure 2. Time course of Fmoc derivatization at 37 °C and ambient temperature. Recovery was calculated by peak height (mV) of the highest peak observed in fetuin oligosaccharides (peak 3 in Figure 3a).

Results and Discussion Principle. Asparagine-type oligosaccharides are linked to Asn-X-Ser/Thr motif through N-glycosidic linkage. Glycoamidase F (PNGase F) is generally used for releasing asparaginetype oligosaccharides from glycoproteins, although some oligosaccharides such as those containing fucose attached to N-acetylglucosamine residue through R(1-3)-linkage are resistant to this enzyme. When a glycoprotein containing asparagine-type oligosaccharides is incubated with PNGase F, the oligosaccharides attached to the protein core are released as β-glycosylamine.35,36 Stability of the glycosylamine is sensitive to the pH of the enzyme reaction solution.37 As previously reported by Tarentino et al., glycosylamine form of the oligosaccharide released by PNGase F are stable at high pH such as pH 8.8, and hydrolysis rate from the glycosylamine to free

Rapid and Sensitive Screening of N-Glycans

research articles

Figure 3. Analysis of fetuin oligosaccharides after derivatization with Fmoc (a) and 2-AA (b). Bars indicate the regions where glycans containing 0 to 4 sialic acid residues (SA) are eluted. The structures for peaks 1-4 were confirmed by MALDI-TOF mass spectrometry as shown in Figure 4.

form (i.e. reducing oligosaccharide) is very slow.38 Tarentino et al. also reported derivatization of the glycosylamines with 2-iminothiolane-Cl (2-IT) to introduce sulfhydryl groups which are then alkylated to mercapto-butyramido oligosaccharides.38 If we can catch the released oligosaccharides as glycosylamine form using a fluorescent tag directly, then we will achieve rapid analysis of asparagine-type oligosaccharides. Fmoc-Cl readily reacts with primary and secondary amines,31 and is widely used as blocking reagent for amino groups of amino acids and peptides.30 We successfully converted the glycosylamines to the corresponding Fmoc derivatives by in situ derivatization reaction. The Fmoc derivatives have following strong points: (1) Fmoc derivatives of glycosylamines have single configuration (i.e., β-form) at the reducing end, and have no isomers which complicate the chromatograms. (2) Fmoc residue has strong fluorescence at 310 nm irradiated at 266 nm. (3) The Fmoc group is easily removed by incubation with morpholine in dimethylformamide at mild conditions. We can easily obtain free oligosaccharides from the Fmoc labeled glycosylamines. The weak point is that the method is not appropriate for direct analysis of released oligosaccharides derived from glycopeptide mixture after digestion of the glycoprotein with proteases, because the contaminating peptides often give complex chromatograms. However, when we use glycoproteins, possibly contaminating proteins are easily removed as fluffy materials in extraction step after derivatization with Fmoc-Cl (see below). It should be also noted that oligosaccharides are efficiently released from glycoproteins when using 5-10 times larger amount of PNGase F than that used in the releasing reaction of glycopeptides. Optimization Studies for Labeling of Glycosylamine Oligosaccharides Released from Glycoprotein Samples. We applied Fmoc derivatization method to the analysis of N-linked oligosaccharides of fetuin. Stability of glycosylamine-form oligosaccharides released from glycoproteins depends on pH of the buffer employed in the enzyme reaction. Low pH of the solution causes hydrolysis of glycosylamines into reducing oligosaccharides, which are no longer derivatized with Fmoc-Cl. We tested N-glycan releasing reaction with PNGase F in some buffer solutions (Hepes, Bicine, acetate, and phosphate) at various pHs from pH 5 to pH 9.5. Recovery of Fmoc-glycans is dependent not only on the reactivity but also on glycan-releasing

Figure 4. MALDI-TOF mass spectra of some Fmoc labeled oligosaccharides from fetuin. Peaks 1-4 were indicated in Figure 3a. Mass spectra were observed in negative/linear mode using DHB as matrix as described in Experimental Section. Abbreviations of oligosaccharide structures and theoretical average masses are also shown. The list of oligosaccharides is shown in Figure 5. The peaks with asterisk show fragment ions formed by cleavage of one N-acetylneuraminic acid residue from parent ions.

Figure 5. List of oligosaccharides in fetuin as determined by MALDI-TOF/MS analysis. SA means N-acetylneuraminic acid.

rate and stability of glycosylamine. Because it was necessary to determine the optimized conditions for each step, we studied whole procedures in wide range of pHs, and found that phosphate buffer between pH 8-9 showed the best recovery of glyJournal of Proteome Research • Vol. 4, No. 1, 2005 149

research articles

Kamoda et al.

Table 1. Comparison of Relative Proportions of Sialyl Oligosaccharides and Total Peak Areas between Fmoc Labeling Method and 2-AA Labeling Methoda

(a) Fmoc (b) 2-AA ratio(a/b)

0SA%

1SA%

2SA%

3SA%

4SA%

total peak area

1.1 0.9 1.2

3.2 3.4 1.0

24.8 27.5 0.9

59.4 59.3 1.0

11.5 8.8 1.3

24041929 4665943 5.2

a Fmoc labeled oligosaccharides and 2-AA labeled oligosaccharides (both 10 µg as fetuin) were injected to HPLC. 0SA-4SA% means the percent ratios of total peak areas eluted in the regions of 0SA, 1SA, 2SA, 3SA, and 4SA in Figure 3.

cosylamines as Fmoc derivatives when evaluated by peak height using HPLC analysis. We also optimized reaction temperature and time for Fmoc derivatization. As shown in Figure 2, the Fmoc derivatization reaction was completed after ca. 4 h at room temperature, but was completed within 1 h at 37 °C. The data were calculated from the peak height of the biggest peak (see Figure 3a, peak 3) because peak area ratios observed at each point showed constant values. After PNGase F digestion followed by Fmoc derivatization, proteins were removed as fluffy materials during extraction of the excess reagent with chloroform. Under the optimized conditions, the present method provides rapid mapping of N-linked oligosaccharides of glycoproteins within 4 h, including enzyme releasing reaction of N-linked oligosaccharides, Fmoc derivatization and cleanup procedure of Fmoc oligosaccharides from the excess reagent and protein by extraction with chloroform. Most widely accepted methods for fluorescent labeling of N-glycans employ reductive amination using a large excess amount of reagents, and usually require at least 1 or 2 days for the sample preparation prior to the analysis by HPLC because the time-consuming steps are necessary including the elimination of excess reagent by solid-phase extraction or gel-filtration followed by drying steps.4 In contrast, Fmoc derivatization proceeds rapidly under mild conditions (Figure 2). Using the present method, the mixture of Fmoc derivatives of oligosaccharides derived from fetuin was analyzed (Figure 3a). Fmoc derivatives were successfully separated in the similar conditions employed for the separation of 2-AA labeled oligosaccharides using an amine-bonded polymer column.14 The analytical conditions were slightly modified as shown in Experimental Section. The separation profile quite resembled that of 2-AA labeled oligosaccharides (Figure 3b). We compared relative peak areas between Fmoc derivatization method and 2-AA method. Table 1 shows that the molar proportions of mono-, di-, tri-, and tetrasialyl oligosaccharides are almost equivalent between both methods. The results indicate that hydrolysis of sialic acid was hardly observed during derivatization. It has been demonstrated that there was no difference in reactivity among different oligosaccharides as reported for 2-AA method.14 Accordingly, similar ratios of oligosaccharides in Table 1 showed that Fmoc labeling also proceeds quantitatively and nonselectively. In addition, we also compared the fluorescent intensities in Fmoc and 2-AA methods by comparison of total peak areas after injecting the same amount of oligosaccharide mixture. The result showed that the total peak areas of Fmoc labeled oligosaccharides were ca. five times bigger than those of 2-AA oligosaccharides, indicating that Fmoc derivatization method shows five times higher sensitivity than that using 2-AA derivatization method. Furthermore, it should be noted that Fmoc method did not 150

Journal of Proteome Research • Vol. 4, No. 1, 2005

accompany partial hydrolysis of sialic acids, because derivatization is performed under slightly basic conditions. Separation of 2-AA labeled oligosaccharides derived from fetuin shows quite similar pattern to those observed for Fmoc derivatives. Therefore, Fmoc derivatives were mainly separated based on their charges (i.e., number of attached sialic acid residues), and five major groups of oligosaccharides having different number of sialic acid residues were observed. Oligosaccharides in each group were also resolved into multiple peaks according to their size and glycosidic linkages. To confirm the structures of some oligosaccharides, we collected a few peaks in Figure 3a and analyzed them by MALDI-TOF mass spectrometry (Figure 4). Peak 1 obtained from the region where monosialyl oligosaccharides were observed, showed a clear molecular ion at m/z 2517 (M-H)-. The molecular ion is due to monosialo-triantennary oligosaccharide (A3S1 in Figure 5). Peak 2 showed a molecular ion at m/z 2443 (M-H)corresponding to disialo-diantennary oligosaccharide (A2S2 in Figure 5). Peak 3 showed a molecular ion at m/z 3100 (M-H)indicating trisialo-triantennary oligosaccharide (A3S3 in Figure 5). Peak 4 showed a molecular ion at m/z 3391 (M-H)- from tetrasialo-triantennary oligosaccharide (A3S4 in Figure 5). Small peaks derived from the parent peaks were also observed at m/z 2153, 2809 and 3099 for Peak 2, 3 and 4, respectively. These peaks were obviously due to fragment ions formed by cleavage of one N-acetylneuraminic acid residue. Fmoc derivatization gave molecular ions of sialic acid containing oligosaccharides in high sensitivity, and may be a merit for sensitive analysis of minute amount of carbohydrates in glycoprotein samples. Analysis of Oligosaccharides Derived from Some Glycoproteins. We applied the present method to the analysis of oligosaccharides of some glycoproteins containing asparaginetype N-glycans (Figure 6). Alpha1-acid glycoprotein (AGP, human) has di-, tri-, and tetraantennary oligosaccharides. Some of the oligosaccharides are fucosylated at the nonreducing terminal lactosamine residue to form sialyl Lewis-X structure.39 In the analysis of AGP oligosaccharides, oligosaccharides containing mono-, di-, tri, and tetrasialic acid residues were clearly observed in the elution range of 1SA, 2SA, 3SA, and 4SA, respectively. Transferrin (human) contains diantennary oligosaccharides with one or two sialic acids and approximately 85% are substituted with disialyl residues.40-42 Our results indicated that disialo-diantennary carbohydrate chain was exclusively present and trace amount of trisialo-triantennary oligosaccharide and monosialo-diantennary oligosaccharide were observed in the elution range of 3SA and 1SA, respectively. Monosialo-diantennary oligosaccharide is a major oligosaccharide in fibrinogen. Disialodiantennary oligosaccharide occupies about 15-20% in fibrinogen from normal subjects. Presence of a considerable amount of asialo-diantennary oligosaccharide has also been reported.43-45 In contrast, we found mono- (61%) and disialo- (39%) oligosaccharides of total oligosaccharides. It should be noted that we could not detect asialo-diantennary oligosaccharides. Because the present method is performed at pH 8.5 for all procedures, sialic acids linked to the carbohydrate chains are not cleaved at all. The oligosaccharide map of thyroglobulin (porcine) also showed essentially the same pattern as reported previously by 2-AA derivatization method.14 Thyroglobulin contains not only sialylated complex-type oligosaccharides but also high-mannose type oligosaccharides.46,47 High-mannose type oligosaccharides were observed in the elution range of 0SA as an

Rapid and Sensitive Screening of N-Glycans

Figure 6. Analysis of oligosaccharides derived from some sialoglycoproteins using Fmoc labeling method. Separation was performed by normal phase HPLC using an amine-bonded polymer column as described in Experimental Section. Bars indicate the regions where glycans containing 0 to 4 sialic acid residues (SA) are eluted.

isolated group from those derived from sialyl oligosaccharides at later region (1SA-4SA). We also applied the present method to the analysis of carbohydrate chains derived from glycoprotein samples containing neutral oligosaccharides as the major component carbohydrate chains (Figure 7). In the separation of neutral oligosaccharides, we employed a silica-based amide-column and changed elution conditions according to the method reported previously.14 This mode of separation is especially useful for the analysis of complex-type asialo- and high-mannose type oligosaccharides. Ribonuclease B (bovine pancreas) has been reported to contain highmannose type oligosaccharides including Man5-9GlcNAc2.48 The oligosaccharide map derived from ribonuclease B showed excellent resolution. It should be noticed that positional isomers of M7 were resolved into three peaks. A pharmaceutical preparation of recombinant IgG, trastuzumab, which is used as a therapeutic pharmaceutical for breast cancer, has fucosylated diantennary complex-type oligosaccharides.49 Asialoagalacto diantennary oligosaccharide (G0) was observed as a major component and two isomers with a galactose residue at one of the two arms (G1) were resolved. We also successfully analyzed oligosaccharides derived from ovalbumin, although we could not confirm the peaks. Separations of Fmoc oligosaccharides by HPLC were performed in the similar elution program used for the analysis of 2-AA labeled oligosaccharides,14 and resolution among peaks was almost equivalent as shown in Figures 6 and 7. Recovery of Free Oligosaccharides from Fmoc Labeled Oligosaccharides. The one of the most important and useful points of the present method is that the Fmoc labeled oli-

research articles

Figure 7. Analysis of neutral oligosaccharides in some glycoproteins containing high-mannose, complex-type and hybrid-type oligosaccharides. Separation was performed by normal phase HPLC using an amide-silica column as described in Experimental Section. M5-M9 in ribonuclease B means high-mannose type oligosaccharide structures of Man5GlcNAc2, Man6GlcNAc2, Man7GlcNAc2, Man8GlcNAc2, and Man9GlcNAc2, respectively. G0-G2 in recombinant IgG (rIgG) means fucosylated complex diantennary structures having 0, 1, and 2 galactose residues, respectively.

gosaccharides can be recovered as free reducing oligosaccharides. Releasing reaction of Fmoc groups is completed within 30 min at 37 °C. When Fmoc oligosaccharides derived from fetuin were incubated in a mixture of morpholine and DMF, all oligosaccharide peaks were completely disappeared after incubation for 30 min at 37 °C (Figure 8a). These data suggest that all the Fmoc labeled oligosaccharides were converted to free oligosaccharides. To confirm that these converted oligosaccharides were recovered as free oligosaccharides, the mixture obtained after releasing reaction of Fmoc group was fluorescently labeled with 2-AA, and analyzed by HPLC. The results (Figure 8b) showed equivalent pattern as observed for direct derivatization with 2-AA (see Figure 3b). The largest peak in Fmoc oligosaccharides of fetuin (peak 3 in Figure 3a) was collected and converted to 2-AA derivative through free oligosaccharide form. As shown in Figure 8c, the converted and 2-AA labeled oligosaccharide was observed as single peak at ca. 62 min. We calculated the recovery of free oligosaccharides according to the procedures described in Figure 1, and found that the peak area observed for the isolated oligosaccharide (A3S3) after removal of Fmoc group and derivatization with 2-AA was 74.5 ( 3.0% of the area of corresponding A3S3 peak observed by direct derivatization. This means that we can obtain free asparagine-type oligosaccharides at least in 75% efficiency. The present method affords a versatile tool for preparation of free oligosaccharides from glycoproteins. The oligosaccharides thus obtained will be available for various analytical studies such as construction of diverse glycan library for sugar array in ‘Glycome’ studies. Journal of Proteome Research • Vol. 4, No. 1, 2005 151

research articles

Figure 8. Conversion of Fmoc derivatives to 2-AA labeled oligosaccharides after recovery of free form oligosaccharides. (a) The mixture of Fmoc oligosaccharides derived from fetuin was converted to free oligosaccharides and analyzed by HPLC fluorometrically at 310 nm with irradiating at 266 nm. (b) A half portion of the mixture analyzed in (a) was derivatized with 2-AA and analyzed fluorometrically using the conditions for 2-AA oligosaccharides. (c) An isolated peak of Fmoc labeled oligosaccharide (peak 3 observed in Figure 3a) was converted to free form followed by derivatization with 2-AA.

Conclusion The present method using Fmoc derivatization has two strong points: (1) High-speed, high-sensitive, and quantitative analysis of N-linked oligosaccharide mapping with extremely easy operations, which can be used for the analysis of oligosaccharides of various glycoproteins including glycoprotein pharmaceuticals and from natural sources; (2) Easy preparation of free-form oligosaccharides with high purity by removal of Fmoc group after HPLC separation, which will contribute to construction of a glycan library. In the present study, we developed a rapid method for the analysis of N-glycans from native glycoprotein samples. It may be better to digest glycoprotein samples with proteases prior to PNGase F digestion for preparation of free oligosaccharides. We reported a cleanup method for purification of glycopeptides from peptide mixture.50 By using this technique, we can analyze N-glycans even after protease digestion of the proteins (manuscript in preparation).

References (1) Davies, M. J.; Hounsell, E. F. Biomed. Chromatogr. 1996, 10, 285289. (2) Townsend, R. R. In Carbohydrate analysis: High Performance Liquid Chromatography and Capillary Electrophoresis; El Rassi, Z., Ed.; Elsevier: New York, 1995; pp 181-209. (3) Hase, S. J. Chromatogr. A 1996, 720, 173-182. (4) Lamari, F. N.; Kuhn, R.; Karamanos, N. K. J. Chromatogr. B 2003, 793, 15-36. (5) Hase, S. Methods Enzymol. 1994, 230, 225-237. (6) Guile, G. R.; Rudd, P. M.; Wing, D. R.; Prime, S. B.; Dwek, R. A. Anal. Biochem. 1996, 240, 210-226. (7) Evangelista, R. A.; Liu, M. S.; Chen, F. T. Anal. Chem. 1995, 67, 2239-2245.

152

Journal of Proteome Research • Vol. 4, No. 1, 2005

Kamoda et al. (8) Sei, K.; Nakano, M.; Kinoshita, M.; Masuko, T.; Kakehi, K. J. Chromatgr. A 2002, 958, 273-281. (9) Kakehi, K.; Funakubo, T.; Suzuki, S.; Oda, Y.; Kitada, Y. J. Chromatogr. A 1999, 863, 205-218. (10) Jackson, P. Biochem. J. 1990, 270, 705-713. (11) Yoshino, K.; Takao, T.; Murata, H.; Shimonishi, Y. Anal. Chem. 1995, 67, 4028-4031. (12) Wenjun, M.; Takao, T.; Sakamoto, H.; Shimonishi, Y. Anal. Chem. 1998, 70, 4520-4526. (13) Okafo, G.; Burrow, L.; Carr, S. A.; Roberts, G. D.; Johnson, W.; Camilleri, P. Anal. Chem. 1996, 68, 4424-4430. (14) Anumula, K. R.; Dhume, S. T. Glycobiology 1998, 8, 685-694. (15) Packer, N. H.; Lawson, M. A.; Jardine, D. R.; Redmond, J. W. Glycoconjugate J. 1998, 15, 737-747. (16) Nakajima, K.; Oda, Y.; Kinoshita, M.; Masuko, T.; Kakehi, K. Analyst 2002, 127, 972-976. (17) Nakajima, K.; Oda, Y.; Kinoshita, M.; Kakehi, K. J. Proteome Res. 2003, 2, 81-88. (18) Nakajima, K.; Kinoshita, M.; Oda, Y.; Masuko, T.; Kaku, H.; Shibuya, N.; Kakehi, K. Glycobiology 2004, 14, 793-804. (19) Sabine, L. F.; Rein, V. U. Nature 2003, 421, 219-220. (20) Toshima, K.; Tatsuta, K. Chem. Rev. 1993, 93, 1503-1531. (21) Wong, C. H.; Halcomb, R. L.; Ichikawa, Y.; Kajimoto, T. Angew. Chem., Int. Ed. Engl. 1995, 34, 521-546. (22) Unverzagt, C. Carbohydr. Res. 1997, 305, 423-431. (23) Palcic, M. M. Curr. Opin. Biotechnol. 1999, 10, 616-624. (24) Wymer, N.; Toone, E. J. Curr. Opin. Chem. Biol. 2000, 4, 110119. (25) Tamura, T.; Wadhwa, M. S.; Rice, K. G. Anal. Biochem. 1994, 216, 335-344. (26) Rice, K. G.; Wu, P.; Brand, L.; Lee, Y. C. Biochemistry 1993, 32, 7264-7270. (27) Crout, D. H. G.; Vic, G. Curr. Opin. Chem. Biol. 1998, 2, 98-111. (28) Kallin, E.; Loenn, H.; Norberg, T. Glycoconjugate J. 1988, 5, 145150. (29) Takahashi, C.; Nakakita, S.; Hase, S. J. Biochem. 2003, 134, 51-55. (30) Carpino, L. A.; Han, G. Y. J. Org. Chem. 1972, 37, 3404-3409. (31) Einarsson, S.; Josefsson, B.; Lagerkvist, S. J. Chromatogr. 1983, 282, 609-618. (32) Keckwick, R. A.; Cannan, R. K. Biochem. J. 1936, 30, 227-235. (33) Nakano, M.; Kakehi, K.; Tsai, M. H.; Lee, Y. C. Glycobiology 2004, 14, 431-441. (34) Kajihara, Y.; Suzuki, Y.; Sasaki, K.; Juneja, L. R. Methods Enzymol. 2003, 362, 44-64. (35) Plummer, T. H., Jr.; Elder, J. H.; Alexander, S.; Phelan, A. W.; Tarentino, A. L. J. Biol. Chem. 1984, 259, 10700-10704. (36) Trimble, R. B.; Tarentino, A. L. J. Biol. Chem. 1991, 266, 16461651. (37) Isbell, H. S.; Frush, H. L. J. Am. Chem. Soc. 1950, 72, 1043-1044. (38) Tarentino, A. L.; Phelan, A. W.; Plummer, T. H. Glycobiology 1993, 3, 279-285. (39) Treuheit, M. J.; Costello, C. E.; Halsall, H. B. Biochem. J. 1992, 283, 105-112. (40) Spik, G.; Bayard, B.; Fournet, B.; Strecker, G.; Bouquelet, S.; Moutreuil, J. FEBS lett. 1975, 50, 296-299. (41) Dorland, L.; Haverkamp, J.; Schut, B. L.; Vliegenthart, J. F. G.; Spik, G.; Strecker, G.; Fournet, B.; Montreuil, J. FEBS lett. 1977, 77, 15-20. (42) Spik, G.; Debruyne, V.; Montreuil, J.; Van Halbeek, H.; Vliegenthart, J. F. G. FEBS Lett. 1985, 183, 65-69. (43) Townsend, R. R.; Hilliker, E.; Li, Y. T.; Laine, R. A.; Bell, W. R.; Lee, Y. C. J. Biol. Chem. 1982, 257, 9704-9710. (44) Maekawa, H.; Yamazumi, K.; Muramatsu, S.; Kaneko, M.; Hirata, H.; Takahashi, N.; Arocha-Pinango, C. L.; Rodriguez, S.; Nagy, H.; Perez-Requejo, J. L.; Matsuda, M. J. Clin. Invest. 1992, 70, 6776. (45) Sugo, T.; Nakamikawa, C.; Takano, H.; Mimuro, J.; Yamaguchi, S.; Mosesson, M. W.; Meh, D. A.; DiOrio, J. P.; Takahashi, N.; Takahashi, H.; Nagai, K.; Matsuda, M. Blood 1999, 94, 3806-3813. (46) Tsuji, T.; Yamamoto, K.; Irimura, T.; Osawa, T. Biochem. J. 1981, 195, 691-699. (47) Yamamoto, K.; Tsuji, T.; Irimura, T.; Osawa, T. Biochem. J. 1981, 195, 701-713. (48) Fu, D.; Chen, L.; O’Neill, R. A. Carbohydr. Res. 1994, 261, 173186. (49) Kamoda, S.; Nomura, C.; Kinoshita, M.; Nishiura, S.; Ishikawa, R.; Kakehi, K.; Kawasaki, N.; Hayakawa, T. J. Chromatogr. A 2004, 1050, 211-216. (50) Nakano, M.; Kakehi, K.; Lee, Y. C. J. Chromatgr. A 2003, 1005, 13-21.

PR049825O