Anal. Chem. 1997, 69, 4554-4559
Analysis of 1-Aminopyrene-3,6,8-trisulfonate-Derivatized Oligosaccharides by Capillary Electrophoresis with Matrix-Assisted Laser Desorption/Ionization Time-of-Flight Mass Spectrometry Hirofumi Suzuki,† Odilo Mu 1 ller,‡ Andra´s Guttman,§ and Barry L. Karger*
Barnett Institute and Department of Chemistry, Northeastern University, Boston, Massachusetts 02115
Matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF MS), has been used for structure characterization of 1-aminopyrene-3,6,8-trisulfonate (APTS)-derivatized oligosaccharides previously separated by capillary electrophoresis (CE). The resolved components were first isolated by employing an automated high-resolution fraction collector. Using on-probe sample cleanup with a cation-exchange resin and a matrix mixture of (1:1) 6-hydroxypicolinic acid and 3-hydroxypicolinic acid, APTS-labeled oligosaccharides were successfully detected by MALDI-TOF MS in the negative ionization mode. Each APTS-labeled oligosaccharide produced single [M - H]-1 peaks. Detection limits for standard APTS-derivatized oligosaccharide (maltoheptaose) were down to 30 fmol. APTS labeled maltooligosaccharides and various standard carbohydrates were separated and collected by CE, followed by molecular weight determination with MALDI-TOF MS. The mass spectral characterization enhances the power of the CE analysis of oligosaccharides. The characterization of complex oligosaccharides in glycoproteins has recently become of increased importance, given the developing knowledge that such carbohydrates play a biological role well beyond simple aqueous solubility enhancement,1 including a direct effect on biological function.2 The analysis of such complex carbohydrates, as mono-, oligo-, and polysaccharides, as well as glycoproteins and glycopeptides, differing in size, cleavage, and even isomeric linkage position, requires high-resolution separation methods. Capillary electrophoresis (CE) is playing an important role in the analysis of such complex carbohydrate mixtures because of its intrinsically high resolving power.3 Due to the lack of a chromophores on the oligosaccharides, derivatization has proven to be necessary for successful detection in CE. Reductive amination and derivatization with fluorescently active dyes for high-resolution laser-induced fluorescence (LIF) CE analysis was first introduced by Novotny and co-workers.4 A popular dye, particularly for slab gel analysis,5 but also for CE
analysis6,7 is 8-aminonaphthalene-1,3,6-trisulfonate (ANTS). While effective, this dye requires the use of a He/Cd laser, which is more expensive and less stable than an argon ion laser. To accommodate the latter, a new dye, 1-aminopyrene-3,6,8-trisulfonate (APTS) was introduced for CE analysis8 and shown to yield quite attractive detection levels (low attomole).9 The approach can be applied to the analysis of glycans of fetuin10 and can be used to resolve linkage isomers.11 While the above CE approach with APTS is powerful, leading to sharp bands even for complex electropherograms, the identification of individual, closely related oligosaccharides derived from glycoproteins remains a necessary, but demanding challenge. Mass spectrometry has already been shown to be useful for obtaining molecular weights of individual oligosaccharide variants12 and, with appropriate enzymatic cleavage, detailed structure.13 Matrix-assisted laser desorption time-of-flight mass spectrometry (MALDI-TOF) has been shown to be useful for underivatized carbohydrate analysis;13,14 however, to the best of our knowledge no work on the MALDI-TOF analysis of sulfonated dye-derivatized oligosaccharides has been shown, a necessary methodology for identification of CE peaks of APTS-labeled carbohydrates. In this work, we have used a recently introduced highresolution automated CE fraction collector developed in our laboratory15 to isolate APTS-derivatized sample components, followed by MALDI-TOF. Due to the highly acidic sulfonate groups, most matrixes and sample cleanup procedures for MALDITOF were initially found to yield multiple peaks and low sensitivity due to cation adducts similar to the case of oligonucleotides.16,17 We have used a new matrix, (1:1) 3- and 6-hydroxypicolinic acid,
On leave from Kanebo Pharmaceuticals Research Center, Osaka, Japan. On leave from the Universita¨t des Saarlandes, Saarbru ¨ cken, Germany. § Beckman Instruments, Inc., Fullerton, CA 92634. Present address: Genetic BioSystems, Inc., San Diego, CA 92121. (1) Varki, A. Glycobiology 1993, 3, 97. (2) Sairam, M. R. FASEB J. 1989, 3, 1915. (3) El Rassi, Z. Electrophoresis 1996, 17, 275.
(4) Liu, J.; Shirota, O.; Wiesler, D.; Novotny, M. Proc. Natl. Acad. Sci. U.S.A. 1991, 88, 2302. (5) Jackson, P. Anal. Biochem. 1994, 216, 243. (6) Hase, S.; Ikenaka, T.; Matsushima, Y. Biochem. Biophys. Res. Commun. 1978, 85, 257. (7) Klockow, A.; Amado, R.; Widmer, H. M.; Paulus, A. J. Chromatogr., A 1995, 716, 245. (8) Evangelista, R. A.; Liu, M. S., Chen; F.-T. A. Anal. Chem. 1995, 67, 2239. (9) Guttman, A.; Chen, F.-T. A.; Evangelista, R. A.; Cooke, N. Anal. Biochem. 1996, 233, 234. (10) Guttman, A.; Chen, F.-T. A.; Evangelista, R. A. Electrophoresis 1996, 17, 412. (11) Guttman, A.; Herrick, S. Anal. Biochem. 1996, 235, 236. (12) Burlingame, A. L.; Boyd, R. K.; Gaskell, S. J. Anal. Chem. 1994, 66, 634R. (13) Papac, D. I.; Wong, A.; Jones, A. J. S. Anal. Chem. 1996, 68, 3215. (14) Stahl, B.; Steup, M.; Karas, M.; Hillenkamp, F. Anal. Chem. 1991, 63,1463. (15) Mu ¨ ller, O.; Foret, F.; Karger, B. L. Anal. Chem. 1995, 67, 2974.
4554 Analytical Chemistry, Vol. 69, No. 22, November 15, 1997
S0003-2700(97)00090-5 CCC: $14.00
† ‡
© 1997 American Chemical Society
along with on-probe sample cleanup by NH4+-loaded cationexchange resin18,19 to yield successful analysis down to 30 fmol for maltoheptaose. This paper demonstrates the feasibility of this approach with a mixture of standard oligosaccharides. Among other applications, structural characterization of the CE-separated APTS-derivatized oligosaccharides should prove useful in the analysis of glycoproteins being synthesized by fermentation in biotechnology. EXPERIMENTAL SECTION Instrumentation. For analytical CE, experiments were performed on a P/ACE 5000 (Beckman Instruments, Inc., Fullerton, CA) in the reversed polarity mode (cathode at the injection side). The separations were monitored on-column with a Beckman LIF detection system, employing a 4-mW argon ion laser with an excitation wavelength of 488 nm and an emission band-pass filter of 520 nm. The temperature of the capillary column was controlled at 20.0 ( 0.1 °C. The electropherograms were acquired and stored on an IBM 486/66 computer using the System Gold software package (Beckman Instruments). A detailed description of the micropreparative CE instrument and fraction collector using a sheath liquid has been discussed elsewhere.15 Briefly, the instrument was constructed in-house, using a high-voltage power supply (CZE 1000R; Spellman, Plainview, NY) and a fiber optic-based UV detector (LC 90; PerkinElmer, Cupertino, CA) operating at 254 nm with detection 11 mm before the exit of the capillary, a sheath liquid collection interface, and a stepper motor-driven/computer-controlled rotor holding 60 collection capillaries (20 µL volume, Idaho Technologies, Moscow, ID). A syringe pump (Model 341B; Sage Instruments, Boston, MA) was used to supply the 10 mM ammonium acetate buffer (pH 4.75) for the sheath liquid at a flow rate of 12 µL/min. The detector signal was digitized by an A/D converter, and the collection procedure was controlled by a 486 IBM-compatible computer. The software was written in house in LabView (National Instruments, Austin, TX). All MALDI-TOF mass spectrometry experiments were performed using a Bruker Protein TOF mass spectrometer (Bruker Instruments, Billerica, MA). Spectra for typically 20-30 shots of the 337-nm nitrogen laser were averaged for each spectrum, and linear, negative ion TOF detection was performed using an ion extraction voltage of -18 kV. For micropreparative CE, a capillary column 75 µm i.d. ( 30 cm (Polymicro Technologies, Phoenix, AZ) coated with covalently attached poly(vinyl alcohol) (available from Beckman Instruments) was used. Ammonium acetate buffer (10 mM, pH 4.75) was employed as running buffer. For analytical CE, poly(vinyl alcohol)coated capillary columns, 50 µm i.d. with 40 cm effective length (47 cm total length) was used. Ammonium acetate (25 mM, pH 4.75) containing 0.4% PEO was employed as running buffer. Materials. Maltoheptaose was purchased from Sigma (St. Louis, MO), and a ladder of maltooligosacharides and trisodium APTS were obtained from Beckman Instruments. Standard oligosaccharides (see Figure 5) were from Oxford GlycoSystems (Bedford, MA). The APTS-labeled oligosaccharide samples were used directly after derivatization or stored at -20 °C. (16) Nordhoff, E.; Ingendoh, A.; Cramer, R.; Overberg, A.; Stahl, B.; Karas, M.; Hillenkamp, F; Crain, P. F. Rapid Commun. Mass Spectrom. 1992, 6, 771. (17) Wu, K. J.; Steding, A.; Becker, C. H. Rapid Commun. Mass Spectrom. 1993, 7, 142. (18) Juhasz, P.; Biemann, K. Proc. Natl. Acad. Sci. U.S.A. 1994, 91, 4333. (19) Rouse, J. C.; Vath, J. E. Anal. Biochem. 1996, 238, 82.
The MALDI matrix, 2,4,6-trihydroxyacetophenone (THAP) was from Fluka (Buchs, Switzerland). 2,5-Dihydroxybenzoic acid (DHB) and R-cyano-4-hydroxycinnamic acid (CHCA) were from Sigma. All other matrixes were from Aldrich (Milwaukee, WI). Cation-exchange resin in the NH4+ form was prepared from the H+ form (200-400 mesh AG 50W-X8, Bio-Rad, Rockville Center, NY) according to a published procedure.20 Derivatization of Oligosaccharides with APTS. The procedure for the reductive amination of oligosaccharides with APTS followed a protocol described elsewhere.21 For standard derivatization, 2 µL of 0.2 M APTS in 15% acetic acid and 2 µL of 1 M NaBH3CN in tetrahydrofuran were added to approximately 1050 nmol of dried oligosaccharide. The reaction mixture was incubated for 16 h at 37 °C in order to avoid any desialylation of the samples. The resulting mixture was diluted in water or passed through a Centrispin-10 column (Princeton Separations, Adelphia, NJ) and injected into the CE column. Sample Preparation of CE-Isolated Fractions for MALDITOF MS. All samples were prepared for MALDI analysis using the on-probe cleanup with cation-exchange resin. The procedure of preparation of 6-hydroxypicolinic acid (6-HPA)/3-hydroxypicolinic acid (3-HPA) matrix solution was as follows: 6-HPA solution was prepared as a saturated solution (∼4 mg /mL) and the 3-HPA solution at a concentration of (40 mg /mL), both in acetonitrile/water (7:3 v/v). The solutions were vortexed for 1 min, sonicated for 5 min, and then centrifuged to remove undissolved crystals. A 90-µL aliquot of 6-HPA solution and 10 µL of 3-HPA solution were combined, and then 5-10 mg of the NH4+ form cation-exchange resin was added. An aliquot of 50 µL of the supernatant was removed, and 5-10 mg of the NH4+ form cation-exchange resin was added into the supernatant. This matrix solution was freshly prepared daily. Typically, 0.5 µL of aqueous APTS-derivatized oligosaccharide sample and 0.5 µL of matrix solution with cation-exchange resin were deposited on a metallic probe and mixed with a micropipet. After drying at room temperature, the resin were removed. The sample solution collected by CE was evaporated on the probe for 5 min at 45 °C before the addition of matrix solution with cation-exchange resin, unless otherwise noted. RESULTS AND DISCUSSION For this study, we used APTS-maltoheptaose to optimize MALDI-TOF procedures. We initially selected DHB as the matrix to examine, since it is widely used for oligosaccharides.22 The standard solution of APTS-maltoheptaose was desalted with a Centrispin-10 column; the expected [M - H]-1 peak was not observed in the negative ion mode using DHB (10 mg/µL in water/ethanol (1:1 v/v) as matrix. Cation adducts were assumed to be present, decreasing the signal. We then turned to on-probe sample cleanup using a cation-exchange resin in the NH4+ form (see Experimental Section), and a peak, relatively free of adducts, was observed for APTS-maltoheptaose (m/z ) 1592.3, [M - H]-1). The exchange of alkali metal cations by ammonium ions allows the formation of the free acid of the sulfonate groups by dissociation.23,24 Furthermore, we found the on-probe sample (20) Wang, B. H.; Biemann, K. Anal. Chem. 1994, 66, 1918. (21) Guttman, A.; Pritchett, T. Electrophoresis 1995, 16, 1906. (22) Strupat, K.; Karas, M.; Hillenkamp, F. Int. J. Mass Spectrom. Ion Processes 1995, 111, 89. (23) Nordhoff, E.; Cramer, R.; Karas, M.; Hillenkamp, F.; Kirpekar, F.; Kristiansen, K.; Roepstorff, P. Nucleic Aids Res. 1993, 21, 3347.
Analytical Chemistry, Vol. 69, No. 22, November 15, 1997
4555
Figure 1. MALDI mass spectra of APTS-labeled maltoheptaose (1 pmol) using the matrix of (A) a mixture of DHB/MSA (9:1 v/v) in water/ ethanol (1:1 v/v) with on-probe NH4+ cation exchange and (B) a mixture of 6-HPA/3-HPA (1:1) in acetonitrile/water (7:3 v/v) with onprobe NH4+ cation exchange. The accelerating voltages were -18 kV.
Figure 3. Electropherogram of the APTS-labeled maltooligosaccharide mixture in the micropreparative separation mode. Capillary, 75 µm i.d. × 30 cm (effective length, 28.9 cm); buffer, 10 mM ammonium acetate buffer, pH 4.75, E ) 333 V/cm; detection, UV 254 nm; injections, maltohexaose (3.1 pmol), maltoheptaose (3.6 pmol), maltooctaose (2.8 pmol). Peaks 6, 7, and 8 represent maltohexaose, maltoheptaose, and maltooctaose, respectively.
Figure 2. MALDI spectra of APTS-labeled maltoheptaose (120 fmol) (A) and (30 fmol) (B) using a mixture of 6-HPA/3-HPA (1:1, v/v) in acetonitrile/water (7:3 v/v) as the matrix with on-probe NH4+ cation exchange. The accelerating voltages were -18 kV.
cleanup procedure to be more efficient than the off-probe approach in minimizing sample losses. While some signal was obtained in the high-femtomole to lowpicomole range, we screened over 20 possible matrixes using the on-probe cleanup procedure for improved sensitivity. Comparable or slightly improved signal for APTS-maltoheptaose was obtained for 6-hydroxypicolinic acid and 2,4,6-trihydroxyacetophenone; however, our best results in terms of signal, resolution, and ruggedness were obtained with comatrixes, see below. Figure 1 shows the MALDI-TOF spectrum obtained for 1 pmol of the APTS-derivatized standard using a 9:1 (w/w) mixture of DHB and 5-methoxysalicylic acid (MSA).25,26 (Similar results were obtained for DHB with 1-hydroxyisoquinoline.27) Two major peaks (24) Stults, J. T.; Marsters, J. C. Rapid Commun. Mass Spectrom. 1991, 5, 359. (25) Karas, M.; Ehring, H.; Norfhoff, E.; Stahl, B.; Strupat, K.; Hillenkamp, F.; Grehl, M.; Krebs, B. Org. Mass Spectrom. 1993, 28, 1476. (26) Gusev, A. I.; Wilkinson, W. R.; Proctor, A.; Hercules, D. M. Anal. Chem. 1995, 67, 1034. (27) Mohr, M. D.; Bo¨rnsen, K. O.; Widmer, M. Rapid Commun. Mass Spectrom. 1995, 9, 809.
4556 Analytical Chemistry, Vol. 69, No. 22, November 15, 1997
Figure 4. MALDI-TOF mass spectra of the CE fractions from Figure 3 of APTS-labeled 6-mer (maltohexaose, 1430.3), 7-mer (maltoheptaose, 1592.3), and 8-mer (maltooctaose, 1754.4). (A) Each fraction was directly deposited on the probe of MALDI, without evaporation, before mixing with the matrix. (B) Each fraction was deposited on the probe after evaporation in vacuum and dissolving in the same amount of water. The numbers in parentheses represent the theoretical [M - H]-1 m/z value. See Figure 1B for MALDI-TOF conditions.
are observed in Figure 1A, m/z ) 1592.9 for [M - H]-1 and m/z ) 1512. The latter peak is likely cleavage of one SO3- group, similar to heparin-derived oligosaccharides.18
Figure 5. Structures of APTS-labeled standard oligosaccharides tested by CE-MALDI-TOF. (A) Trisialo-triantennary, (B) asialo-agalactotriantennary, (C) hybrid-type bisecting GlcNAc, (D) asialo-biantennary bisecting GlcNAc, (E) high mannose (M9), and (F) asialo-triantennary. Man, mannose; GlcNAc, N-acetylglucosamine; Neu5Ac, N-acetylneuraminic acid; Gal, galactose.
We then tested mixtures of 3-HPA and 6-HPA at different composition levels and settled on a 1:1 (w/w) mixture, leading to the result shown in Figure 1B. A single peak for [M - H]-1 was observed with no fragmentation or cation adducts. Figure 2 shows the MALDI-TOF of 120 and 30 fmol of APTS-maltoheptaose with this matrix and on-probe cleanup procedure. The detection level of less than 30 fmol is comparable to that recently obtained for underivatized oligosaccharides.13 With this success, we then analyzed the collected fractions of APTS-labeled maltooligosaccharides. The use of CE-LIF to achieve high-resolution separation of APTS-labeled complex carbohydrates can be seen in the open-tube separation of a ladder of maltooligosaccharides.9 Whereas other workers added 0.4% polymer to the buffer,9 we have chosen not to do this here because of potential contamination of isolated fragments in the MALDITOF analysis. The buffer employed was 10 mM ammonium acetate (pH 4.75), in order to minimize alkali cation adducts for MALDI-TOF. We next coupled the sheath flow fraction collector to the 75 µm column and employed UV detection ∼1 cm before the exit using an optical fiber.15 Figure 3 shows operation in this mode using hydrodynamic injection. Based on calibration curves,
the amounts of maltohexaose, -heptaose, and -octaose (6, 7, and 8 in the figure) were 3.1, 3.6, and 2.8 pmol, respectively. The separation is complete in Figure 3; however, the bands are a little broader compared to analytical CE.9 Two possible causes of the loss in efficiency are overloading and back pressure into the capillary caused by the sheath liquid (leading to laminar flow). In our previous fraction collection work, a polymer matrix-filled capillary was used15 or positive pressure applied.28 In this work, neither of these was available to overcome the back pressure leading to broadening. Recently, we have eliminated this broadening effect by using a valve on the injection side, allowing fraction collection in open tube capillaries of 200 µm i.d.29 In spite of the small increase in broadness, we were nevertheless successful in collecting isolated fractions of compounds 6, 7, and 8 in Figure 3. Since the APTS-derivatized species are highly negative (three sulfonate groups on APTS), collection of the separated bands in capillaries, as described previously,15 did not lead to adsorption losses. (28) Foret, F.; Mu ¨ ller, O.; Thorne, J.; Go¨tzinger, W.; Karger, B. L. J. Chromatogr., A 1995, 716, 157. (29) Minarek, M.; Foret, F.; Karger, B. L., unpublished result.
Analytical Chemistry, Vol. 69, No. 22, November 15, 1997
4557
Figure 6. (A) CE separation for isolation of the collection of APTSlabeled oligosaccharides shown in Figure 5. Capillary, 7 5µm i.d. × 30 cm (effective length, 28.9 cm); buffer, 10 mM ammonium acetate buffer, pH 4.75; E ) 333 V/cm; detection, UV 254 nm; electrokinetic injection (10 kV, 2 s). (A) trisialo-triantennary, (B) asialo-agalactotriantennary, (C) hybrid-type bisecting GlcNAc, (D) asialo-biantennary bisecting GlcNAc, (E) high mannose (Man-9), and (F) asialotriantennary. (B) MALDI-TOF mass spectra of the isolated fragments collected during the run in (A). (A) trisialo-triantennary (3319.0), (B) asialo-agalacto-triantennary (1959.5), (C) hybrid-type bisecting GlcNAc (2080.5), (D) asialo-biantennary bisecting GlcNAc (2283.6), (E) high mannose (Man-9) (2322.6), and (F) asialo-triantennary (2445.7). The numbers in parentheses represent the theoretical [M - H]-1 m/z value. The loaded amounts of each carbohydrates were 280-420 fmol except for trisialo-triantennary. See Figure 1B for MALDI-TOF conditions.
Figure 4A shows the MALDI-TOF of the three fractions, each deposited from the collection capillaries on the probe, followed by the procedures described above for Figure 1B. In Figure 4B, the sample was deposited in a microcentrifuge tube, evaporated to dryness under reduced pressure, and redissolved in the same amount of water, followed by mixing of the matrix and on-probe ion-exchange cleanup. The signal intensity is found to be greater in Figure 4B than in Figure 4A, as the amount of buffer (ammonium acetate) that reduces ionization has been decreased. In both (A) and (B), fragments corresponding to [M - H]-1 for APTS-maltohexaose (expected m/z ) 1430.3), -maltoheptaose (expected m/z ) 1592.3), and -maltooctaose (expected m/z ) 1754.4) were observed. In (B), trace amounts of 7-mer contaminated the 8-mer fraction. In summary, the results in Figure 4 demonstrate that specific fragments of APTS-labeled oligosaccha4558 Analytical Chemistry, Vol. 69, No. 22, November 15, 1997
Figure 7. (A) CE separation of the APTS-labeled oligosaccharides derived from ribonuclease B and (B) the MALDI-TOF mass spectra of the collected fragments. The peaks at m/z (A) 1673.6 (1674.4), (B) 1835.3 (1836.4), (C) 1995.4 (1998.5), (D) 2161.2 (2160.5), and (E) 2320.7 (2322.6) correspond to the [M - H]-1 APTS-labeled Man5, Man-6, Man-7, Man-8, and Man-9, respectively. CE conditions as in Figure 6A, and MALDI-TOF conditions as in Figure 1B.
rides can be isolated after CE separation and their molelcular weights determined by MALDI-TOF. A series of complex carbohydrates was next analyzed to explore the generality of this approach. Figure 5 lists the structures of the carbohydrates that were derivatized by APTS. Successful CE separation was obtained using a solution of 120140 pmol/µL and UV detection (Figure 6A). Each peak was collected with the automated fraction collector in individual capillaries. Using argon ion laser-induced fluorescence detection, single peaks were observed for each fraction except trisialotriantennary where overlapped peaks were found, likely due to the anomericity of the trisialo-triantennary structure. (Note in Figure 6A that trisialo-triantennary is well separated from the other peaks.) Next, 0.5-1 µL of each fraction were placed on the MALDI target. After evaporation for 5 min at 45 °C in order to remove ammonium acetate, 0.5 µL of the matrix solution with the NH4+ cation-exchange resin was added. Figure 6B shows the MALDI/ TOF analysis of the individual fractions, all spectra except trisialotriantennary being obtained from collection of a single injection. The loaded amounts of each carbohydrates were 280-420 fmol except trisialo-triantennary. As seen in the figure caption, the measured [M - H]-1 molecular weight values agreed well with the expected values.
In the case of trisialo-triantennary, the labeled carboyhydrate was collected from four runs and concentrated prior to MALDITOF analysis. As can be seen, while the correct molecular weight is obtained, a broad peak of low intensity was also obtained. As the number of sialic acids increase, difficulties in MALDI-TOF spectra are observed for underivatized oligosaccharides30 (however, see ref 13). Fragmentation of the analyte may be part of the loss in sensitivity,31 and it would be interesting to test this compound using delayed extraction with the MALDI-TOF.32 Nevertheless, all the other highly complex carbohydrates were successfully analyzed, as seen in Figure 6B. At this stage, we can handle ∼5 µL as the sample for this micropreparative CE. Therefore, ∼600 pmoL of oligosaccharides is sufficient for this analysis by single collection. Finally, the oligosaccharide pool derived from ribonuclease B was analyzed by CE and MALDI-TOF MS. The complete structural characterization of the oligosaccharides from ribonuclease B has recently been reported.33 Ribonuclease B (M 15 500) has high-mannose-type carbohydrate structures containing five to nine mannose residues. Figure 7A shows the micropreparative electropherogram for the oligosaccharides with concentrations of each component sample (determined by the analytical CE) as Man-5 (126 pmol/µL), Man-6 (71 pmol/µL), Man-7 (38 pmol/ µL), Man-8 (61 pmol/µL), and Man-9 (33 pmol/µL), respectively. The APTS-labeled oligosaccharides were individually collected using the automated fraction collector. The MALDI mass spectra (Figure 7B) reveal that the collections were successful and an accurate estimate of molecular masses obtained.
CONCLUSIONS This paper has demonstrated successful micropreparative collection of CE-separated APTS-derivatized oligosaccharides, followed by MALDI-TOF mass spectrometric analysis. Low-level detection (∼30 fmol) was achieved in the negative ion mode by careful selection of the matrix and on-probe ion-exchange cleanup using an ammoniated cation exchanger. This work opens up the possibility of the identification and characterization of labeled carbohydrate species separated by CE. This would be especially valuable in the biotechnology industry, where the required amount of sample material for fluorophore labeling does not represent a limitation. Indeed, full characterization of a released glycan pool is possible from subnanomolar amounts, including oligosaccharide profiling,9-11 exoglycosidase digestion-based sequencing34 and MALDI-TOF analysis for structure elucidation. As noted, for multisialo (tri-, tetra-, pentasialo and higher) species, the MALDI-TOF procedure is not as sensitive as for non-, mono-, or bisialo APTS-derivatized carbohydrates; however, delayed extraction may prove useful here as it has for oligonucleotides32 or an improved matrix may be found. Finally, APTS may represent a typical fluorescent dye used in CE (e.g., conjugated ring structure with sulfonate groups), and therefore, the MALDI-TOF conditions developed in this work may prove useful for other derivatized components besides oligosaccharides.
(30) Harvey, D. J. Rapid Commun. Mass Spectrom. 1993, 7, 614. (31) Juhasz, P.; Biemann, K. Carbohydr. Res. 1995, 270, 131. (32) Vestal, M. L.; Juhasz, P.; Martin, S. A. Rapid Commun. Mass Spectrom. 1995, 9, 1044. (33) Fu, D.; Chen, L.; O’Neill, R. A. Carbohydr. Res. 1994, 261, 173. (34) Guttman, A. Electrophoresis 1997, 18, 1136.
Received for review September 12, 1997.X
ACKNOWLEDGMENT The authors gratefully acknowledge NIH Grant GM15847 and Kanebo Ltd. for support of this work. Contribution 689 from the Barnett Institute. January
23,
1997.
Accepted
AC970090Z X
Abstract published in Advance ACS Abstracts, October 15, 1997.
Analytical Chemistry, Vol. 69, No. 22, November 15, 1997
4559
