Characterization of Electrophoretic Behavior of Sugar Isomers by

Science and Technology Corporation (JST), Shomachi, Tokushima 770-8505, ... Nishinomiya 662-8580, Japan, and Single-Molecule Bioanalysis Laboratory, ...
0 downloads 0 Views 193KB Size
Download PDF
Anal. Chem. 2003, 75, 2433-2439

Characterization of Electrophoretic Behavior of Sugar Isomers by Microchip Electrophoresis Coupled with Videomicroscopy Fuquan Dang,*,† Lihua Zhang,†,‡ Mohammad Jabasini,† Noritada Kaji,† and Yoshinobu Baba†,§

Department of Medicinal Chemistry, Faculty of Pharmaceutical Sciences, The University of Tokushima, CREST, Japan Science and Technology Corporation (JST), Shomachi, Tokushima 770-8505, Japan, Furuno Electric Co., Ltd., Nishinomiya 662-8580, Japan, and Single-Molecule Bioanalysis Laboratory, National Institute of Advanced Industrial Science and Technology (AIST), Hayashi-cho, Takamatsu 761-0395, Japan

The electrophoretic behavior of oligosaccharide isomers was investigated by microchip electrophoresis (µ-CE) coupled with videomicroscopy using maltose, cellobiose, maltriose, and panose as oligosaccharide isomer models. The present study revealed for the first time that the formation of a carbohydrate-phosphate complex is a pHindependent rapid process, whereas the formation of a carbohydrate-borate complex is a highly pH-dependent slow process. As a result, phosphate buffer gave much better separation on oligosaccharide isomers than borate and borate-Tris buffers over a wide pH range in µ-CE. The imaging analysis of the complete process of sample loading and injection with field-amplified stacking (FAS) demonstrated that FAS could be used as an efficient method for manipulating the shape of injected sample plugs, and thus improving the performance of µ-CE in the absence of electroosmotic flow. However, once the ionic strength mismatch between sample and running buffer reached a critical threshold, a further increase in ionic strength mismatch deteriorated the effect of FAS, resulting in a surprising decrease in separation efficiency and peak distortion. Under optimal conditions, high-resolution separation of some oligosaccharide isomers and a complex oligosaccharide mixture released from ribonuclease B was achieved using PMMA microchips with an effective separation channel of 30 mm. Microchip electrophoresis (µ-CE) is a rapidly emerging analytical technology that has attracted much attention in the past decade because it possesses substantial advantages over conventional analytical technologies in terms of separation speed and cost, precision of data, and integration and miniaturization of the analytical instrument.1-6 µ-CE has been used for the analysis of * To whom correspondence should be addressed. Phone:+81-88-633-7285. Fax: +81-88-633-9507. E-mail: [email protected]. † University of Tokushima. ‡ Furuno Electric Co. § AIST. (1) Harrison, D. J.; Fluri, K.; Seiler, K.; Fan, Z.; Effenhauser, C. S.; Manz, A. Science 1993, 261, 895-897. (2) Zhang, B.; Foret, F.; Karger, B. L. Anal. Chem. 2001, 73, 2675-2681. (3) Shi, Y.-N.; Simpson, P. C.; Scherer, J. R.; Wexler, D.; Skibola, C.; Smith, M. T.; Mathies, R. A. Anal. Chem. 1999, 71, 5354-5361. 10.1021/ac034110a CCC: $25.00 Published on Web 04/18/2003

© 2003 American Chemical Society

various species, including small biomolecules,7,8 amino acids,9,10 oligonucleotides,11 carbohydrates,12,13 DNA fragments,14-18 peptides,19 and proteins.20-22 Separation performance of µ-CE is basically similar to that of conventional capillary electrophoresis (CE), but µ-CE can be performed on a time scale of seconds. This short analysis time is largely the result of integrated injectors on microchips, which allow injection of well-defined sample plugs that can be well resolved in short separation channels. These features make µ-CE an attractive technology for the next generation of CE instrumentation. Oligosaccharides from natural sources such as glycoproteins are complex compounds with subtle differences in structure, such as carbohydrate sequence, types of glycosidic bonds, and number and position of branched chains.23,24 Analysis of complex oligosac(4) Zhang, L.-H.; Dang, F.-Q.; Baba, Y. J. Pharm. Biomed. Anal. 2003, 30, 16451654. (5) Bruin, G. J. M. Electrophoresis 2000, 21, 3931-3951. (6) Henry, A. C.; Tutt, T. J.; Galloway, M.; Davidson, Y. Y.; McWhorter, C. S.; Soper, S. A.; McCarley, R. L. Anal. Chem. 2000, 72, 5331-5337. (7) Pasas, S. A.; Lacher, N. A.; Davies, M. I.; Lunte, S. M. Electrophoresis 2002, 23, 759-766. (8) Chiem, N.; Harrison, D. J. Anal. Chem. 1997, 69, 373-378. (9) Rodriguez, I.; Jin, L.-J.; Li, S.-F.-Y. Electrophoresis 2000, 21, 211-219. (10) Fister, J. C.; Jacobson, S. C.; Ramsey, J. M. Anal. Chem. 1999, 71, 44604464. (11) Effenhauser, C. S.; Manz, A.; Widmer, H. M. Anal. Chem. 1995, 67, 22842287. (12) Burggraf, N.; Krattiger, B.; de Mollo, A. J.; de Rovij, N. F.; Manz, A. Analyst 1998, 123, 1443-1447. (13) Suzuki, S.; Shimotsu, N.; Honda, S.; Arai, A.; Nakanishi, H. Electrophoresis 2001, 22, 4023-4031. (14) Lagally, E. T.; Medintz, I.; Mathies, R. A. Anal. Chem. 2001, 73, 565-570. (15) Khandurina, J.; McKnight, T. E.; Jacobson, S. C.; Waters, L. C.; Foote, R. S.; Ramsey, J. M. Anal. Chem. 2000, 72, 2995-3000. (16) Zhang, L.-H.; Dang, F.-Q.; Baba, Y. Electrophoresis 2002, 23, 2341-2346. (17) Ueda, M.; Endo, Y.; Abe, H.; Kuyama, H.; Nakanishi, H.; Arai, A.; Baba, Y. Electrophoresis 2001, 22, 217-221. (18) Mitnik, L.; Carey, L.; Burger, R.; Desmarais, S.; Koutny, L.; Wernet, O.; Matsudaira, P.; Ehrlich, D. Electrophoresis 2002, 23, 719-726. (19) Zhang, B.; Foret, F.; Karger, B. L. Anal. Chem. 2000, 72, 1015-1022. (20) Liu, Y.-J.; Foote, R. S.; Jacobson, S. C.; Ramsey, R. S.; Ramsey, J. M. Anal. Chem. 2000, 72, 4608-4613. (21) Jin, L. J.; Giordano, B. C.; Landers, J. P. Anal. Chem. 2001, 73, 49944999. (22) Bousse, L.; Mouradian, S.; Minalla, A.; Yee, H.; Williams, K.; Dubrow, R. Anal. Chem. 2001, 73, 1207-1212. (23) Rudd, P. M.; Elliott, T.; Cresswell, P.; Wilson, I. A.; Dwek, R. A. Science 2001, 291, 2370-2376. (24) Helenius, A.; Aebi, M. Science 2001, 291, 2364-2369.

Analytical Chemistry, Vol. 75, No. 10, May 15, 2003 2433

charides is generally considered a very challenging task and requires extremely high separation resolution and efficiency.25-27 Modern CE has demonstrated the necessary resolution for analysis of complex oligosaccharide mixtures labeled with suitable fluorescent reagents.29-31 The use of µ-CE in the analysis of complex oligosaccharides has recently been demonstrated by our group.32 However, the analysis of oligosaccharide isomers and complex oligosaccharides derived from glycoproteins by µ-CE remains challenging because of the much shorter separation channel on microchips and lower separation efficiency compared to conventional CE. In the present study, the resolution capacity of µ-CE for analysis of oligosaccharide isomers has been investigated using APTSlabeled maltose (G2), cellobiose (G2′), maltriose (G3), and panose (G3′) as model oligosaccharide isomers. A simple but effective procedure based on field-amplified stacking (FAS) was demonstrated for controlling the shape of injected sample plugs and, thus, improving separation performance of µ-CE when electroosmotic flow (EOF) is negligible. However, there was a critical threshold of ionic strength mismatch between the sample and running buffer. An increase in ionic strength mismatch beyond this critical threshold caused FAS to be ineffective, resulting in surprisingly poor separation efficiency and resolution. We found that formation of a carbohydrate-borate complex (borate complexation) is a highly pH-dependent slow process, whereas formation of a carbohydrate-phosphate complex (phosphate complexation) is a pH-independent rapid process. We obtained excellent separation of APTS-labeled oligosaccharide isomers and a complex oligosaccharide mixture released from ribonuclease B in phosphate buffer using PMMA chips with an effective separation channel of 30 mm. EXPERIMENTAL SECTION Reagents and Buffer Solution. Bovine ribonuclease B, maltose monohydrate, D-(+)-cellubiose, D-panose, and MC (viscosity of 2% aqueous solution at 20 °C, 4000 cP) were purchased from Sigma Chemical Co. (St. Louris, MO). Peptide N-glycosidase (PNGase F, EC.3.6.1.52 recombinant) was obtained from Roche Molecular Biochemicals (Minato-ku, Tokyo, Japan). APTS was acquired from Molecular Probes (Eugene OR). Dextrin 15 was obtained from Fluka (Buchs, Germany). Sodium cyanoborohydride, tetrahydrofuran, phosphoric acid (0.5 M), and all other chemicals were obtained as analytical grade reagents from Kanto Chemical Co. Inc. (Tokyo, Japan). Running buffers with cellulose polymer additives were prepared by adding MC to borate acid buffer (various concentrations) or phosphoric acid (20 mM) and stirring slowly until the solution appeared homogeneous and transparent. Buffers were then (25) Charlwood, J.; Birrell, H.; Bouvier, E. S. P.; Langridge, J.; Camilleri, P. Anal. Chem. 2000, 72, 1469-1474. (26) Kakehi, K.; Funakubo, T.; Suzuki, S.; Oda, Y.; Kitada, Y. J. Chromatogr., A 1999, 863, 205-218. (27) Frado, L.-L. Y.; Strickler, J. E. Electrophoresis 2000, 21, 2296-2308. (28) Osthoff, H. D.; Sujino, K.; Palcic, M. M.; Dovichi, N. J. J. Chromatogr., A 2000, 895, 285-290. (29) Chen, F.-T. A.; Evangelista, R. A. Electrophoresis 1998, 19, 2639-2644. (30) Oefner, P.; Chiesa, C.; Bonn, G.; Horva´th, Cs. J. Capillary Electrophor. 1994, 1, 3-12. (31) Liu, J.; Shirota, O.; Novotny, M. Anal. Chem. 1992, 63, 64-78. (32) Dang, F.-Q.; Zhang, L.-H.; Hagiwara, H.; Mishina, Y.; Baba, Y. Electrophoresis 2003, 24, 714-721.

2434 Analytical Chemistry, Vol. 75, No. 10, May 15, 2003

Figure 1. Schematic diagram of electrophoresis microchip and detection designs. The lengths from channel intersection to the sample, sample waste, buffer, and buffer waste wells are 5.25, 5.25, 5.75, and 37.5 mm, respectively. The LED fluorescence detection point is 30 mm from the center of the intersection.

adjusted to the designed pH using KOH. Double-deionized water was used for the preparation of buffers and samples. Release of N-Linked Oligosaccharides from Glycoprotein by the PNGase F. In a 500-µL screw-capped microcentrifuge tube, 200 µg of ribonuclease B was dissolved in 35 µL of 50 mM phosphate buffer (pH 7.5), followed by addition of 2 units of PNGase F solution. After incubation of 24 h at 37 °C, the mixture was heated at 100 °C for 3 min to terminate the reaction. The mixture was then mixed with 150 µL of cold ethanol to precipitate the protein, followed immediately by centrifugation at 14 000 rpm at 4 °C for 3 min. The supernatant was collected in a 500-µL screwcapped microcentrifuge tube and concentrated to dryness in a centrifugal evaporator at room temperature. The resulting residue was used in the derivatization reaction. Derivatization Procedure. In a 500-µL screw-capped microcentrifuge tube, a dried sample of hydrolysate of ribonuclease B or standard oligosaccharide (5 nmol) was mixed with 5 µL of 100 mM APTS in 15% acetic acid solution and 10 µL of 500 mM NaBH3CN in tetrahydrofuran. The mixture was kept in a water bath at 55 °C for 1.5 h, and then the reaction mixture was diluted with water to 200 µL and stored at - 20 °C. An aliquot of the above solution was diluted to the desired concentrations with doubledeionized water prior to analysis. Instrumentations. As shown in Figure 1, a Nikon Eclipse E800 microscope with a high-pressure Hg lamp (Nikon, Japan) and a LED confocal fluorescence detector were used to monitor the injection and separation processes, respectively. Two highvoltage power supplies (Hitachi Electronics Engineering Co., Ltd, Japan) were used to provide voltages from 0 to 2000 V, and from 0 to -2000 V for electrophoresis experiments. Two-dimensional fluorescence images of the injection process were recorded on a digital tape using an intensified charge-coupled camera (ICCD,

Hamamatsu Photonics, Hamamatsu, Japan) and an image processor (Hamamatsu Photonics, ARGUS 20) at an acquisition rate of 30 frames/s. Electropherograms were obtained using a Hitachi SV1100 microchip electrophoresis instrument with a LED detector. The median excitation wavelength of the LED detector was 470 nm, and the detection wavelength was longer than 580 nm. The PMMA microchips had a simple cross channel of 100 µm width and 30 µm depth. The distances from the channel intersection to the sample (S), sample waste (SW), buffer (B), and buffer waste (BW) wells were 5.25, 5.25, 5.75, and 37.5 mm, respectively. The effective separation channel length was 30 mm.32 Microchip Electrophoresis. Buffer solutions were introduced into the microchannels via syringe. All reservoirs on the microchips were filled with either matrix or sample by pipet prior to analysis. In the sample-loading procedure, 300 V was applied to the SW well for 30 s while the other three wells were grounded. During separation, the buffer well was grounded, and a fixed 130 V was applied to both the S and SW wells. High voltages were applied to the BW well to create the desired field strengths on the separation channel.16 RESULTS AND DISCUSSION The dynamic formation of charged complexes between sugar units and oxy acids in electrolyte is the basis of electrophoretic separation of sugar isomers with the same charge-to-mass ratio. The borate complexation is a well-studied example, and has produced very promising results in conventional CE.33,34 Although the separation performance of µ-CE is similar to that of conventional CE for most applications, such as the separation of DNA and proteins, unusual phenomena have been observed in separation of oligosaccharide isomers by µ-CE when separation time is on a time scale of seconds. In a previous study,32 it was found that the adsorption of APTS-labeled oligosaccharides to the surface of PMMA channels was too strong to allow separation in commonly used buffer systems. To eliminate the adsorption of APTS-labeled oligosaccharides and EOF in PMMA chips, all buffers in the present study contained 0.5% MC as an additive. Figure 2 shows that phosphate buffer allowed good separations of APTS-labeled G2 and G2′ by µ-CE, depending on whether the sample was dissolved in water or running buffer. The sample dissolved in water had better separation and a 10-fold enhancement of peak intensity. Conversely, borate and borate-Tris buffers produced very limited resolution of APTS-labeled G2 and G2′ only when the sample dissolved in running buffer was used. APTSlabeled G3 and G3′ (positional isomers) could not be separated at all in phosphate, borate, or borate-Tris buffers. Concentrations of borate as high as 400 mM did not have any obvious effects on resolution of APTS-labeled G2 and G2′ (data not shown), unlike the case with conventional CE.21,22 Although the reason for this unusual phenomenon is not clear, it is possible that the extensive complexation between MC and borate is partly responsible. Figure 3 depicts the effect of buffer pH on separation resolution and efficiency of APTS-labeled oligosaccharide isomers in the three buffer systems. In phosphate buffer, the separation efficiency (33) Stefansson, M.; Novotny, M. Anal. Chem. 1994, 66, 1134-1140. (34) Honda, S.; Suzuki, S.; Nose, K.; Yamamoto, K.; Kakehi, K. Carbhydr. Res. 1991, 215, 193-198.

Figure 2. Microchip electropherograms of APTS-G2, -G2′, -G3, and -G3′ in different buffers. Conditions: Esep ) 168 V/cm; (a) 0.5% MC in 50 mM borate-Tris buffer (pH 9.98), 2.1 × 10-6 M APTS-G2, -G2′, -G3, and -G3′ in corresponding running buffers; (b) 0.5% MC in 100 mM borate buffer (pH 8.90), 2.1 × 10-6 M APTS-G2, -G2′, -G3, and -G3′ in corresponding running buffers; (c) 0.5% MC in 20 mM phosphate buffer (pH 6.66), 2.1 × 10-6 M APTS-G2, -G2′, -G3, and -G3′ in corresponding running buffers; (d) 0.5% MC in 20 mM phosphate buffer (pH 6.66), 2.1 × 10-7 M APTS-G2, -G2′, -G3, and -G3′ in water.

and resolution obtained using a sample dissolved in water kept essentially constant until buffer pH reached 6.66 and then decreased drastically as buffer pH rose above 7.22, whereas separation efficiency and resolution obtained with samples dissolved in corresponding phosphate running buffers remained very constant throughout the pH range used. Clearly, as a sample solvent in µ-CE using phosphate buffer, water produces better resolution and higher sensitivity than the corresponding running buffer for APTS-labeled sugar isomers at buffer pH values below 7.22. In contrast, separation resolution and efficiency were extremely dependent on buffer pH in borate or borate-Tris buffer. In borate buffer, separation resolution and efficiency increased sharply as buffer pH rose to 7.60, reached a plateau at buffer pH between 8.49 and 9.58, and then dropped drastically as buffer pH rose above 9.90. A similar trend was observed in borate-Tris buffer with a plateau for separation efficiency and resolution at buffer pH between 8.98 and 10.26. It should be noted that the resolution of APTS-labeled G2 and G2′ obtained in borate and borate-Tris buffers was astonishingly lower than that obtained in phosphate buffer, considering that there was no evident difference in electrophoretic efficiency in any of the three buffer systems at the optimal pH range when samples dissolved in the corresponding running buffer were used. When samples in water were used, no resolution of APTS-labeled G2 and G2′ was obtained in borate and borate-Tris buffers (see later discussion). To better understand these usual phenomena, the imaging analysis of sample injection processes under different conditions was performed using a fluorescence microscope coupled with a highly sensitive ICCD camera. Still frames from a video of the sample plug formation process with FAS are shown in Figure 4A. The first frame started at the moment when a voltage of 300 V was applied to the SW well while all other wells were grounded for 30 s, during which the sample of APTS-labeled sugar molecules was loaded into the sample and Analytical Chemistry, Vol. 75, No. 10, May 15, 2003

2435

Figure 3. The theoretical plates of APTS-G2 (A) and resolution of APTS-G2 and -G2′ (B) as a function of buffer pH. Conditions as in Figure 2.

sample waste channels. Figure 4A clearly shows that, at t ) 0 ms, the intensity of the entire image was uniform and a little dark due to low sample concentration (2.1 × 10-8 M). The sample shape at the intersection was asymmetric, with a large part protruding into the separation channel. Then the separation mode was initiated, with potentials at the S, SW, and BW wells of 130, 130, and 750 V, respectively, and with the B well grounded. APTSlabeled oligosaccharide molecules started to migrate into the separation, sample, and sample waste channels, and the fluorescence intensity in the rear of the forming sample plug increased significantly due to the stacking effect at t ) 66 ms.35,36 The front edge of the forming sample plug gradually changed from a parabolic profile into flat profiles during this injection process, but the movement of the sample plug in the separation channel was virtually negligible until t ) 132 ms. At t ) 264 ms, a concentrated sample plug with nearly flat front and back ends was formed in the separation channel. However, as illustrated in Figure 4B, the process of sample plug formation without FAS was completed in a short time, and a larger sample plug with parabolic front and rear ends was formed in the separation channel at 165 ms. Surprisingly, there was a critical threshold of ionic strength mismatch between the sample zone and BGE. When ionic strength mismatch exceeded the critical threshold, FAS lost its (35) Quirino, J. P.; Terabe, S. Electrophoresis 2000, 21, 355-359. (36) Quirino, J. P.; Terabe, S. J. Chromatogr., A 1999, 850, 339-344.

2436 Analytical Chemistry, Vol. 75, No. 10, May 15, 2003

Figure 4. Fluorescence images of the sample plug formation with FAS (A and C) and without FAS (B). Arrows indicate the direction of flow in microchip channels. Conditions: Default pinched injection. In loading phase, potential at SW well was 300 V, and BW, B, and S wells were grounded. In separation phase, potentials at S, SW, and BW wells were 130, 130, and 750 V, respectively, and the B well was grounded. (A) 0.5% MC in 20 mM phosphate-KOH buffer (pH 6.66), 2.1 × 10-8 M APTS-G2, -G2′, -G3, and -G3′ in water; (B) 0.5% MC in 20 mM phosphate-KOH buffer (pH6.66), 2.1 × 10-7 M APTSG2, -G2′, -G3, and -G3′ in corresponding running buffers; (C) 0.5% MC in 20 mM phosphate-KOH buffer (pH 8.88), 2.1 × 10-8 M APTSG2, -G2′, -G3, and -G3′ in water.

stacking effect and yielded an even larger asymmetric sample plug in a short time than sample plug formation without FAS (Figure 4C). This novel phenomenon accounted for a dramatic decrease

Figure 6. Microchip electropherograms of APTS-G2, -G2′, -G3, and -G3′ dissolved in 100 mM borate buffer and water. Conditions as in Figure 2b. Running buffer, 0.5% MC in 100 mM borate buffer (pH 8.90). Samples: (a) 2.1 × 10-6 M APTS-G2, -G2′, -G3 ,and -G3′ in running buffer; (b) 2.1 × 10-7 M APTS-G2, -G2′, -G3, and -G3′ in water.

Figure 5. Microchip electropherograms of APTS isomers (A) and APTS-oligosaccharide ladder (B) in phosphate buffer at different pHs. Conditions as in Figure 2d. (A) 2.1 × 10-7 M APTS isomers in water, and (B) 1.56 µg/mL APTS-oligosaccharide ladder in water.

in separation efficiency and resolution in phosphate buffer at pH values of running buffer higher than pH 7.22. Figure 5 shows that in phosphate buffer at pH 8.88, in which the ionic strength mismatch was much greater than the critical threshold, the first peak of APTS was well-shaped, whereas the peaks of the other analytes were distorted with obvious fronting. Evaluation of the data shown in Figures 5 and 4C gives a clear picture of what happened when ionic strength mismatch exceeded the critical threshold. In such cases, the first analyte band of APTS in the sample zone was relatively well stacked, but it migrated too rapidly to slow as desired at the boundary between the sample zone and BGE as a result of the excessive field strength in the sample plug. As the sample band of APTS rapidly crossed the boundary, the abrupt disappearance of the boundary between the sample zone and BGE destroyed the desired field strength distribution between the sample zone and BGE and disrupted the progressing FAS of the following analyte bands. As a result, obvious fronting of peaks (except for the first peak) was observed. The present results suggest that FAS induced by ionic strength mismatch between sample and running buffer results in smaller symmetric sample plugs and, thus, higher efficiency and resolution in phosphate buffer when EOF is negligible. In addition, changing

the sample solvent from running buffer to water led to a significant increase of theoretical plates from 12 686 to 15 226, and ∼20-fold enhancement of peak intensity due to FAS effect in borate buffer. However, there was no resolution of APTS-labeled G2 and G2′ when the sample dissolved in water was used (Figure 6). These results strongly indicated that borate complexation is a slow process that is very highly dependent on buffer pH, whereas phosphate complexation is a rapid process that is virtually independent of buffer pH. That may explain why phosphorylation is one of most important posttranslational modifications of proteins, playing a pivotal role in signal transduction. Because of miniaturized channels fabricated on microchips and extremely low concentration of the real samples usually presented in very small volumes, it is always desirable to perform an on-line concentration to increase the detection sensitivity of µ-CE. Among the developed sample stacking techniques, FAS is a well-adopted technique on microfluidic devices.37-41 In the presence of EOF, Poiseuille flow effect is always induced by the higher EOF velocity in the sample zone, as compared to the rest of the column, resulting in a pressure-driven parabolic flow profile that can lead to considerable band-broadening of the sample plug, thereby deteriorating the separation performance.37-39 Using glass microchips with coated channels, Yang and Chien achieved several hundred signal enhancements by establishing a stationary boundary near the injection intersection via a multiple pressure controller.40 Vazquez et al.41 investigated the electrophoretic injection process of DNA fragments on coated glass microchips by digital images and found that DNA molecules redistributed according to mobility during the FAS process. As shown in Figure 4A, we established an ionic boundary in PMMA chips by applying a high voltage of 300 V at SW well for a period of 30 s. Still images in (37) Palmer, J.; Burgi, D. S.; Landers, J. P. Anal. Chem. 2001, 73, 725-731. (38) Kutter, J. P.; Ramsey, R. S.; Jacbson, S. C.; Ramsey, J. M. J. Microcolumn Sep. 1998, 10, 313-319. (39) Lichtenberg, J.; Verpoorte, E.; de Rooij, N. F. Electrophoresis 2001, 22, 258271. (40) Yang, H.; Chien, R. L. J. Chromatogr., A 2001, 924, 155-163. (41) Vazquez, M.; McKinley, G.; Mitnik, L.; Desmarais, S.; Matsudaira, P.; Ehrlich, D. Anal. Chem. 2002, 74, 1952-1961.

Analytical Chemistry, Vol. 75, No. 10, May 15, 2003

2437

Figure 8. Microchip electropherograms of APTS-G2, -G2′, -G3, and -G3′ at a field strength of 300 V/cm. Conditions: (A) Default pinched injection; 0.5% MC in 100 mM borate buffer (pH 8.90), 2.1 × 10-6 M APTS-G2, -G2′, -G3, and -G3′ in running buffer; (B) symmetric pinched injection; 0.5% MC in 20 mM phosphate buffer (pH 6.66), 2.1 × 10-7 M APTS-G2, -G2′, -G3, and -G3′ in water; (C) default pinched injection; 0.5% MC in 20 mM phosphate buffer (pH 6.66), 2.1 × 10-7 M APTSG2, -G2′, -G3, and -G3′ in water; (D) asymmetric pinched injection; 0.5% MC in 20 mM phosphate buffer (pH 6.66), 2.1 × 10-7 M APTSG2, -G2′, -G3, and -G3′ in water.

Figure 7. Manipulation of the injected sample plugs via fieldamplified sample stacking. Conditions as in Figure 4A. (A) Symmetric pinched injection. In loading phase, field strengths at S, SW, B, and BW channels were 187, 527, 171, and 171 V/cm, respectively. (B) Asymmetric pinched injection. In loading phase, field strengths at S, SW, B, and BW channels were 187, 385, 26, and 171 V/cm, respectively.

Figure 4A clearly show that the boundary is not stationary, even in the absence of EOF (only exited for a very short time), and the state of the injection sample plug appears to be determined by the part of the sample at the intersection that protrudes into the separation channel. Therefore, it may be possible to manipulate the shape of injected sample plugs by FAS on microchips in the absence of EOF. As shown in Figure 7A and B, more symmetric and concentrated sample plugs were obtained using two different pinched injection modes, in which the electric field strengths at the S, SW, B, and BW channels were 187, 527, 171, and 171 V/cm and 187, 385, 26, and 171 V/cm, respectively, for the loading phase. By combining Figures 4A and 7, it is evident that the shape of the injected sample plug is mainly determined by the part of the sample at the intersection protruding in the separation channel during the injection process with FAS in the absence of EOF. Although no further obvious increase in separation efficiency or resolution was observed for APTS-labeled oligosaccharide isomers with sample plugs shorter than 100 µm (consistent with theoretical prediction42), the present procedure proved to be a good method to achieve high performance separation of DNA and proteins on microchips with short separation channels, which will be discussed 2438 Analytical Chemistry, Vol. 75, No. 10, May 15, 2003

elsewhere. As seen in Figure 8, a significant improvement in sensitivity was achieved by using asymmetric pinched injection with a large amount of sample protruding into B channel. When compared to other proposed procedures for obtaining a sample plug narrower than the separation channel,43-45 the present procedure appears to be a more universal and effective method for obtaining well-shaped narrow sample plugs and, thus, improving the separation efficiency and sensitivity of µ-CE in the absence of EOF. The present results show that the contribution of the injection sample plug to plate height is negligible when the sample plug is shorter than 100 µm and separation is diffusion-dominated. In such cases, field strength in the separation channel (Esep) was a key factor in further improvement of the separation of APTS-labeled oligosaccharide isomers by µ-CE. The optimum Esep for resolution of APTS-labeled G2 and G2′ was ∼300 V/cm, at which the average resolution of APTS-labeled G2 and G2′ in phosphate buffer and borate buffer was 0.89 and 0.56, respectively; the corresponding electropherograms are shown in Figure 8. Because of high ionic strength, Joule heating became a concern when field strengths greater than 300 V/cm were used in phosphate or borate buffer, resulting in an obvious decrease in separation efficiency. Although samples dissolved in running buffer or water were used in the present study, it should be possible to achieve highperformance separation of the real samples usually containing salts using the present procedure. As shown in Figure 9, we achieved (42) Jacobson, S. C.; Hergenro¨der, R.; Koutny, L. B.; Warmack, R. J.; Ramsey, J. M. Anal. Chem. 1994, 66, 1107-1113. (43) Shultz-Lockyear, L. L.; Colyer, C. L.; Fan, Z. H.; Roy, K. I.; Harrison, D. J. Electrophoresis 1999, 21, 529-538. (44) Zhang, C. X.; Manz, A. Anal. Chem. 2001, 73, 2656-2662. (45) Deshpande, M.; Greiner, K. B.; West, J.; Gilbert, J. R.; Bousse, L.; Minalla, A. Proceeding of the µ-TAS 2000 Symposium; van den Berg, A., Olthuis, W., Bergveld, P., Eds; Kluwer Academic Publishers: The Netherlands, 2000; pp 415-418.

The figure shows that M5 (M5GlcNAc2, 7 sugar units) migrates slightly faster than maltoheptaose in the oligosaccharide ladder, and that M9 (M9GlcNAc2, 11 sugar units) migrates with a velocity similar to that of maltodecaose in the oligosaccharide ladder (Figure 9d). This evident difference in electrophoretic mobility between these sugars was also observed in other studies.46 It was noted that M7 was also resolved into three peaks by µ-CE, which was comparable to those obtained with CE-LIF.26,29 The results here show that the present µ-CE method is a promising alternative procedure for fast profiling of trace complex oligosaccharides from glycoproteins. CONCLUSIONS In the present study, the electrophoretic behavior of oligosaccharide isomers was systematically investigated using µ-CE coupled with videomicroscopy. The present results revealed for the first time that phosphate complexation is a pH-independent quick process, whereas borate complexation is a highly pHdependent slow process. We demonstrated a universal and effective procedure, based on FAS, for obtaining well-shaped narrow sample plugs and for improving separation efficiency and sensitivity of µ-CE in the absence of EOF. There was a critical threshold of ionic strength mismatch between the sample zone and BGE. Ionic strength mismatches exceeding the critical threshold resulted in poor separation with distorted peaks. Under optimal conditions, high-resolution separation of oligosaccharide isomers and a complex mixture of oligosaccharides released from ribonuclease B was successfully achieved using PMMA microchips with an effective separation channel of 30 mm.

Figure 9. Microchip electropherograms of APTS-labeled N-linked oligosaccharides from ribonuclease B and APTS-labeled oligosaccharide ladder at a field strength of 300 V/cm. Conditions as in Figure 2d. (A) Symmetric pinched injection; PNGase-catalyzed hydrolysis of bovine ribonuclease B. (B) Default pinched injection; PNGasecatalyzed hydrolysis of bovine ribonuclease B. (C) Asymmetric pinched injection; PNGase-catalyzed hydrolysis of bovine ribonuclease B. (D) Default pinched injection; 1.56 µg/mL APTS-oligosaccharide ladder in water.

ACKNOWLEDGMENT The present work is supported in part by the CREST program of the Japan Science and Technology Corporation (JST); a grant from the New Energy and Industrial Technology Development Organization (NEDO) of the Ministry of Economy, Trade and Industry, Japan; a Grant-in-Aid for Scientific Research from the Ministry of Health and Welfare, Japan; and a Grant-in-Aid for Scientific Research from the Ministry of Education, Science and Technology, Japan.

excellent resolution of a complex mixture of high-mannose-type oligosaccharides released from ribonuclease B (bovine pancrease).

Received for review February 4, 2003. Accepted March 25, 2003.

(46) Guttman, A.; Pritchett, T. Electrophoresis 1995, 16, 1906-1911.

AC034110A

Analytical Chemistry, Vol. 75, No. 10, May 15, 2003

2439