Integration of Microfabricated Devices to Capillary Electrophoresis

Anal. Chem. 1999, 71, 3036-3045

Integration of Microfabricated Devices to Capillary Electrophoresis-Electrospray Mass Spectrometry Using a Low Dead Volume Connection: Application to Rapid Analyses of Proteolytic Digests Jianjun Li,† Pierre Thibault,*,† Nicolas H. Bings,‡ Cameron D. Skinner,‡ Can Wang,‡ Christa Colyer,‡ and Jed Harrison‡

Institute for Biological Sciences, 100 Sussex Drive, Ottawa, Ontario, Canada K1A 0R6, and Department of Chemistry, University of Alberta, Edmonton, AB, Canada T6G 2G2

This report describes the development of a compact and versatile, micromachined chip device enabling the efficient coupling of capillary electrophoresis to electrospray mass spectrometry (CE-ESMS). On-chip separation provides a convenient means of achieving rapid sample cleanup and resolution of multicomponent samples (typically 2-5 min) prior to mass spectral analysis. A low dead volume connection facilitating the coupling of microfabricated devices to CE-ESMS was evaluated using two different interfaces. The first configuration used disposable nanoelectrospray emitters directly coupled to the chip device via this low dead volume junction, thereby providing rapid separation of complex protein digests. The performance of this interface was compared with that of more traditional configurations using a sheath flow CEESMS arrangement where a fused-silica capillary of varying length enabled further temporal resolution of the multicomponent samples. The sensitivity and analytical characteristics of these interfaces were investigated in both negative and positive ion modes using standard peptide mixtures. The separation performance for synthetic peptides using a chip coated with amine reagent ranged from 26 000 to 58 000 theoretical plates for a sheath flow CE-ESMS interface comprising a 15-cm CE column. Replicate injections of a dilution series of peptide standards provided detection limits of 45-400 nM without the use of on-line preconcentration devices. The reproducibility of migration time ranged from 0.9 to 1.5% RSD wheras RSDs of 5-10% were observed on peak areas. The application of these devices for the analysis of protein digests was further evaluated using on-line tandem mass spectrometry. Electrospray mass spectrometry (ESMS) has emerged as a sensitive technique, providing peptide sequence analysis on femtomole levels of protein digests using sequence tags and * Corresponding author: (phone) 613 998-0326; (fax) 613 941-1327; (e-mail) [email protected] . † Institute for Biological Sciences. ‡ University of Alberta.

3036 Analytical Chemistry, Vol. 71, No. 15, August 1, 1999

database searching.1,2 The combination of a high-resolution separation technique to the electrospray ion sources thus confers a unique advantage in situations where both sensitivity and selectivity are desired.3-8 Further improvement of sensitivity can be obtained using nanoelectrospray ionization operating at low nanoliter per minute flow rates. Such advances in analytical performance were also marked with the integration of highresolution separation techniques and led to the development of capillary electrophoresis-nanoelectrospray mass spectrometry methods for the analysis of protein and glycoprotein digests.9-15 Microchip technology has recently been applied to capillary electrophoresis (chip-CE), generating an extremely powerful separation and sample pretreatment tool for rapid analyses (typically less than 1 min).16,17 Separations have been combined on-chip with sample dilution, derivatization, enzyme digestion, and a set of independent manifolds all integrated onto a single chip (1) Gaskell, S. J. J. Mass Spectrom. 1997, 32, 677-688. (2) Mann, M.; Wilm, M. Anal. Chem. 1994, 66, 4390-4399. (3) Wilm, M.; Shevchenko, A.; Houthaeve, T.; Breit, S.; Schweiger, L.; Fotsis, T.; Mann, M. Nature 1996, 379, 466-469. (4) Figeys, D.; Ducret, A.; Aebersold, R. J. Chromatogr., A 1997, 763, 295306. (5) Figeys, D.; Ducret, A.; Oostveen, I. V.; Aebersold, R. Anal. Chem. 1996, 68, 1822-1828. (6) Licklider, L.; Kuhr, W. G.; Lacey, M. P.; Keough, T.; Purdon, M. P.; Takigiku, R. Anal. Chem. 1995, 67, 4170-4177. (7) Davis, M. T.; Lee, T. D.; Ronk, M.; Hefta, S. A. Anal. Biochem. 1995, 224, 235-244. (8) Yates, J. R., III; McCormack, A. L.; Link, A. J.; Schieltz, D.; Eng, J.; Hays, L. Analyst 1996, 121, R65-R76. (9) Wahl, J. H.; Gale, D. C.; Smith, R. D. J. Chromatogr., A 1994, 659, 217222. (10) Valaskovic, G. A.; Kelleher, N. L., McLafferty, F. W. Science 1996, 273, 1199-1202. (11) Figeys, D.; Ducret, A.; Yates, J. R. III; Aebersold, R. Nature Biotechnol. 1996, 14, 1579-1583. (12) Bateman, K. P.; White, R. L.; Thibault, P. Rapid Commun. Mass Spectrom. 1997, 11, 307-315. (13) Kelly, J. F.; Ramaley, L.; Thibault, P. Anal. Chem. 1997, 69, 51-60. (14) Bateman, K. P.; White, R. L.; Yaguchi, M.; Thibault, P. J. Chromatogr., A 1998, 794, 327-344. (15) Hannis, J. C.; Muddiman, D. C. Rapid Commun. Mass Spectrom. 1998, 12, 443-448. (16) Manz, A.; Harrison, D. J.; Verpoorte, E. M. J.; Fettinger, J. C.; Paulus, A.; Ludi, H.; Widmer, H. M. J. Chromatogr. 1992, 593, 253-258. (17) Harrison, D. J.; Fluri, K.; Seiler, K.; Fan, Z.; Effenhauser, C. S.; Manz, A. Science 1993, 261, 895-897. 10.1021/ac981420y CCC: $18.00

© 1999 American Chemical Society Published on Web 06/15/1999

leading to multiplexed analyses.18-27 Transport of biological cells using direct current electrokinetic effects in a chip-based capillary format was also demonstrated.28 A recent report has shown the separation and isolation of cultured cervical carcinoma (HeLa) cells from normal blood cells using electrophoresis on a bioelectronic chip.29 Thus, sample pretreatment can be automated within an integrated device, a feature that could offer significant advantages in sample preparation for mass spectrometry, particularly if the chip can be incorporated as part of the electrospray source. Among the detectors used for chip-based capillary electrophoresis, fluorescence detection has proven to be very effective.30 Recently, other detection techniques have been introduced to this field, such as planar absorbance,31 electrochemical detection,32 and electrospray mass spectrometry.33-35 The direct coupling of chip devices to ESMS systems would greatly expand the potential of both CE and ESMS for biotechnological applications requiring faster analysis time, enhanced sensitivity, and selectivity. Previous investigations using microchip devices coupled to ESMS have shown that the direct ionization of analytes from the chip is feasible.33,34 Multiple-channel glass chips can be interfaced to ESMS, whereby the sample is delivered to the mass spectrometer at a flow rate of 100-200 nL/min.33 Electroosmotic flow can also be used as a pump to supply the sample to the ESMS interface.34 To date, such devices were described for sample infusion experiments from multiple wells and have been of limited use for online separation. Alternate approaches have used a coupling junction whereby a capillary column is butted against the chip via a Teflon connector.35 This report is centered on the development of a compact and versatile micromachined chip device using a low dead volume connection that enables direct coupling to electrospray mass spectrometry with minimal loss in separation efficiency. A detailed description of this chip-capillary junction is described in a recent (18) Jacobson, S. C.; Koutny, L. B.; Hergenro der, R.; Moore, A. W., Jr.; Ramsey, J. M. Anal. Chem. 1994, 66, 3472-3476. (19) Chiem, N.; Harrison, J. Anal. Chem. 1997, 69, 373-378. (20) Fan, Z.; Harrison, D. J. Anal. Chem. 1994, 66, 177-184. (21) Jacobson, S. C.; Moore, A. W.; Ramsey, J. M. Anal. Chem. 1995, 67, 20592063. (22) Jacobson, S. C.; Ramsey, J. M. Anal. Chem. 1997, 69, 3212-3217. (23) van der Moolen, J. N.; Poppe, H.; Smit, H. C. Anal. Chem. 1997, 69, 42204225. (24) McCormick, R. M.; Nelson, R. J.; Alonso-Amigo, M. G.; Benvegnu, D. J.; Hooper, H. H. Anal. Chem. 1997, 69, 2626-2630. (25) Moore, Jr.; Jacobson, S. C.; Ramsey, J. M. Anal. Chem. 1995, 67, 41844189. (26) Jacobson, S. C.; Ramsey, J. M. Anal. Chem. 1996, 68, 720-723. (27) Woolley, A. T.; Sensabaugh, G. F.; Mathies, R. A. Anal. Chem. 1997, 69, 2181-2186. (28) Li, P. C. H.; Harrison, J. Anal. Chem. 1997, 69, 1564-1568. (29) Cheng, J.; Sheldon, E. L.; Wu, L.; Heller, M. J.; O’Connell, J. P. Anal. Chem. 1998, 70, 2321-2326. (30) Fister, J. C., III.; Jacobson, S. C.; Davis, L. M.; Ramsey, J. M. Anal. Chem. 1998, 70, 431-437. (31) Liang, Z.; Chiem, N.; Ocvirk, G.; Tang, T.; Fluri, K.; Harrison, J. Anal. Chem. 1996, 68, 1040-1046. (32) Woolley, A. T.; Lao, K.; Glazer, A. N.; Mathies, R. A. Anal. Chem. 1998, 70, 684-688. (33) Xue, Q.; Foret, F.; Dunayevskiy, Y. M.; Zavracky, P. M.; McGruer, N. E.; Karger, B. L. Anal. Chem. 1997, 69, 426-430. (34) Ramsey, R. S.; Ramsey, J. M. Anal. Chem. 1997, 69, 1174-1178. (35) Figeys, D.; Ning, Y.; Aebersold, R. Anal. Chem. 1997, 69, 3153-3160.

issue of this journal.36 The practical advantages of this low dead volume connection have been examined for the coupling of chip devices to CE-ESMS using two different interfaces: a nanoelectrospray configuration using disposable emitters and a more conventional sheath flow interface operating at a flow rate of 1-2 µL/min. The sensitivity and analytical characteristics of the respective interfaces were investigated for standard peptide mixtures and complex protein digests. The application of these devices for rapid protein identification is also evaluated using online tandem mass spectrometry. EXPERIMENTAL SECTION Chemicals and Materials. Fused-silica capillary was purchased from Polymicro Technologies (Phoenix, AZ) and Teflon tubing from LC Packing (San Francisco, CA). The lectins Dolichos biflorus and Pisum sativum and peptide standards used in this study were purchased from Sigma Chemical Co. (St. Louis, MO) and used without further purification. [(Acryloylamino)propyl]trimethylammonium chloride (BCQ) was obtained from Chemische Fabrik Stockhausen (Krefeld, Germany). 7-Oct-1-enyltrimethoxysilane was purchased from United Chemical Technologies Inc. (Bristol, PA). N,N,N′,N′-Tetramethylethylenediamine (TEMED) and morpholine were obtained from Aldrich Chemical Co. Inc. (Milwaukee, WI), and formic acid was from BDH Inc. (Toronto, ON, Canada). Methanol, acetonitrile, and isopropyl alcohol (IPA) were from Em Science (Gibbstown, NJ). Protein Digests. Approximately 0.2 mg of each lectin was dissolved in 200 µL of 0.1 M ammonium bicarbonate incubated with Promega trypsin (Fisher Scientific, Montre´al, PQ, Canada) for 10 h at 37 °C using a substrate-to-enzyme ratio of 50:1. The solution was freeze-dried using a SpeedVac preconcentrator, reconstituted once in 100 µL of deionized water, and subsequently evaporated to dryness. The residue was dissolved in deionized water to a final concentration of 0.5 mg/mL of the original mass of lectin. Device Fabrication. The microfluidic devices were fabricated at the Alberta Microelectronic Centre (Edmonton, AB, Canada), as described previously.28 The chip layout is presented in Figure 1 for the disposable nanoelectrospray emitters (PCRD2 design, Figure 1a) and for the conventional sheath flow configuration (UACEMS design, Figure 1b). To facilitate the coupling of the devices to the ESMS interface, a low dead volume junction enabling the direct insertion of the capillary column was developed. The procedure involves the precise drilling of small holes (typically 200 µm i.d., 600-800 µm depth) with a combination of pointed and flat carbide drills. A detailed description of the procedures used to prepare and construct these low dead volume connectors was presented recently.36 Fused-silica capillaries of varying lengths (1 cm for the disposable nanoelectrospray emitters and 10-40 cm length for the sheath flow interface, Figure 1) were prepared by tapering the outlet of the capillary (185 µm o.d. × 50 µm i.d.) to 60 µm (∼15 µm i.d.) using a laser puller (Sutter Instruments, Phoenix, AZ). The device was placed on a piece of glass and the inlet end of the capillary, carefully cut straight with a cleaving stone, was inserted into the hole. Crystal Bond was melted onto the face of the chip and allowed to flow into the hole (36) Bings, N. H.; Wang, C.; Skinner, C. D.; Colyer, K. L.; Li, J.; Thibault, P.; Harrison, J. D. Anal. Chem. 1999, 71, 3292-3296.

Analytical Chemistry, Vol. 71, No. 15, August 1, 1999

3037

Figure 1. Schematic representation of the chip-CE configuration using (a) a disposable nanoelectrospray emitter (device PCRD2) and (b) a sheath flow ESMS interface (device UACEMS).

until it nearly reached the end of the capillary. The assembly was removed and cooled with forced air. Coating Procedure for Chip. Chips were sequentially rinsed under 20 psi pressure with 1 M NaOH, deionized water for 1 h, and methanol, each for 1 h. A solution of 7-oct-1-enyltrimethoxysilane (25 µL) and glacial acetic acid (25 µL) in methanol (5 mL) was passed through the microchannels overnight at 20 psi. The chips were then sequentially rinsed with methanol and deionized water, each for 1 h. The 5-mL solution containing 10 µL of TEMED, 70 µL of 15% (w/v) ammonium persulfate, and 100 µL of BCQ was rinsed through for 12 h.12 Finally, the chip was flushed with deionized water for 1 h and purged with nitrogen. All aqueous solutions were filtered through a 0.45-µm filter (Millipore, Bedford, MA) before use. Separation Buffers. The separation buffers and sheath solution were prepared with aqueous 30 mM ammonium acetate adjusted to pH 8.8 using ammonium hydroxide. For negative mode, the separation buffers were prepared with 30 mM morpholine in deionized water adjusted to pH 9.0 using formic acid. A 0.1 M formic acid solution was used for separation of peptide standard mixture with BCQ-coated chip and capillary. All aqueous solutions were filtered through a 0.45-µm filter (Millipore, Bedford, MA) before use. Chip CE-ESMS. The thickness of the chip device was 3 mm in order to permit the use of septa (Thermogreen LB-1, Supelco) to ensure proper sealing of each channel. In this configuration, the chip rested on a Teflon support and a plexiglass top was used to compress the septa. Small plastic pipet tubes and/or Teflon tubing (180 µm i.d.) were inserted into the center of the septa before assembly. In this configuration, each well was tightly sealed with no risk of blockage or contamination from epoxy resin.35 This arrangement also permitted rapid changes of reservoirs and introduction of additional solutions using capillary column directly inserted in the Teflon tubing. Platinum wires insulated with plastic 3038 Analytical Chemistry, Vol. 71, No. 15, August 1, 1999

tubes were inserted into the reservoirs of the device. Caution should be exercised when high voltages are used. All chip-CE-ESMS analyses were carried out using a computercontrolled power supply and relay arrangement described previously.36 Chips were first conditioned with 0.1 M NaOH (uncoated chips) or 0.1 M HCOOH (BCQ-coated chips). All channels were filled with running buffer by applying vacuum to each well before the reservoirs were replenished with either buffer or sample. Electrokinetic sample injection was performed by applying a potential of typically 2.3 kV between wells A and C (PCRD2) or B and E (UACEMS). A separation voltage of either + 5 kV for reservoir B in PCRD2 or -3.8 kV for UACEMS was applied to reservoir A, and minimization of sample leakage was achieved using a push back voltage of typically 400 V.37 It is noteworthy that control of the sample plug shape with a concurrent focusing voltage applied to the separation channel22,38,39 was not practical for analysis conducted under anodal electrosmotic flow (UACEMS). Rather, the sample plug length was varied by controlling the injection time and potential. Previous investigations have shown that leakage at the injector can increase the sample loading over that calculated from a metered segment in the double-tee injector.20,31,37 The determination of sample loading was made by comparing the peak areas for standard peptide mixture with that obtained using hydrodynamic injection with a conventional sheath flow CZE-ESMS interface. To maintain high separation efficiency, the injection time was ∼5 s, which provided a sample injection of 0.7 ( 0.2 nL. However, in stacking experiments where the sample was prepared in 10-fold dilution of the separation buffer, the (37) Seiler, K.; Fan, Z. H.; Fluri, K.; Harrison, D. J. Anal. Chem. 1994, 66, 34853491. (38) Ermakov, S. V.; Jacobson, S. C.; Ramsey, J. M. Anal. Chem. 1998, 70, 44944505. (39) Patankar, N. A.; Hu, H. H. Anal. Chem. 1998, 70, 1870-1881.

electrokinetic injection resulted in a sample introduction of 10.4 ( 2.4 nL. Mass spectrometric experiments were conducted using an API 300 LC/MS/MS system (Perkin-Elmer/Sciex, Concord, ON, Canada). A MicroIonSpray interface configured to provide a concentric flow of 1.5 µL/min of makeup buffer solution to the tip of the electrospray needle was used for the sheath flow configuration (Figure 1b). The sheath buffer was composed of either a solution of 2:1 IPA/methanol (for negative ion mode) or 0.1 M formic acid containing 25% methanol (for positive ion mode). For the PCRD2 device (Figure 1a), addition of separation buffer was made using a Harvard syringe pump operating at a flow rate of 50 nL/min. Optimization of the interface was achieved by electromigration of a peptide standard (angiotensin I, 10 µg/mL) dissolved in the separation buffer (well A in Figure 1a and well E in Figure 1b). Mass spectra were acquired with dwell times of 2 ms per m/z step in full-mass scan mode or 100 ms per channel for selected ion monitoring (SIM) experiments. A Power Macintosh 9500/120 computer was used for instrument control, data acquisition, and data processing. Chip-CE-MS/MS. Tandem mass spectrometry experiments were also conducted using the API 300 triple-quadrupole instrument. Collision-induced dissociation of selected precursor ions was obtained using nitrogen as collision gas at collision energies of typically 50 eV (laboratory frame of reference). Fragment ions formed in the rf-only quadrupole were mass-analyzed by scanning the third quadupole using a dwell time of 1 ms per step of 0.5 m/z unit. Fragment ions are labeled according to peptide fragmentation nomenclature proposed by Roepstorff and Fohlman.40 RESULTS AND DISCUSSION A number of different approaches have been explored to provide efficient coupling of the chip to the ESMS interface. A very simple interface was first investigated by Karger and co-workers,33 whereby the electrospray voltage was applied to a side channel via a sheath solution. Application of this coupling method was demonstrated in sample infusion experiments for the monitoring of tryptic digestion products.41 However, such configuration was not found conducive to chip-CE-ESMS experiments using an internal gold electrode due to the release of gas upon application of the separation voltage. The gas formation observed here was associated with direct electrolysis of water as reported previously by Timperman et al.42 We have attempted to solve the problem associated with gas release by applying a gold layer at the exit of the chip device. Although successful analyte ionization was obtained following the formation of a Taylor cone directly from the device, a relatively large droplet was observed at the edge of the chip. This in turn gave rise to extensive peak broadening due to sample mixing and memory effects. Alternatively, a capillary column can be connected to the device via a Teflon tube so that conventional sheath flow interface can be employed.35 Attempts to use such a configuration resulted in (40) Roepstorff, P.; Fohlman, J. Biomed. Environ: Mass Spectrom. 1984, 11, 601. (41) Xue, Q.; Dunayevski, Y. M.; Foret, F.; Karger, B. L. Rapid Commun. Mass Spectrom. 1997, 11, 1253-1256. (42) Timperman, A.; Tracht, S. E.; Sweedler, J. V. Anal. Chem. 1996, 68, 26932698.

significant peak broadening due to the large dead volume associated with the Teflon connector. Ultimately, we directly inserted a capillary column into the chip device using a low dead volume connector which offered no significant loss in separation efficiency as determined with on-chip and on-capillary measurements.36 This chip-capillary connection also facilitated the coupling to the ESMS interface via a nanoelectrospray arrangement (chip-CE-nESMS) or a conventional sheath flow configuration (chip-CE-µESMS). A discussion of results obtained from these two interfaces is provided below. Chip-Nanoelectrospray Mass Spectrometry Using Disposable Emitters. Preliminary studies conducted using CE-ESMS with conventional fused-silica capillary indicated that some analytes, such as peptides, proteins, and even carbohydrates, can still be detected as positive ions while separated as anions. On the basis of this observation, we investigated a chip device that comprised a small electrospray emitter for which the running buffer B (Figure 1a) was used to initiate both the separation and the ionization of analytes. A schematic description of this device is shown in Figure 1a, using a 1-cm untreated fused-silica capillary tip (15 µm i.d, 60 µm o.d.) as electrospray emitter. The main separation channel on the chip was 4 cm in length, and neither the chip nor the electrospray emitter was coated. To improve the stability of the electrospray, and to supply solvent for proper analyte ionization, additional buffer was pumped from a side channel (well D in Figure 1a) to increase the flow rate. The configuration of the chip design required that the flow from the side channel be set to 150 nL/min or lower. For the particular layout of this device, an increase of flow rate above 150 nL/min resulted in improved signal stability, though the sensitivity was reduced as a result of peak broadening and counterflow during the electrophoresis separation. On the other hand, if the flow rate was too low, the electrospray was unstable. The support solution was applied from well D at a flow rate of 50 nL/min, and a voltage of +5 kV was applied to well B to effect for both separation and analyte ionization. A typical electrophoregram corresponding to the separation of a mixture of four peptides (64-180 fmol each) is presented in Figure 2 using ammonium acetate as an electrolyte. The total ion electropherogram (TIE) for m/z 500-1200 (Figure 2a) shows two resolved peaks at 1.49 and 1.92 min. Individual components present in this mixture were identified from their corresponding reconstructed ion electropherograms (RIE) as indicated in parts b-e of Figure 2 for the singly or doubly protonated molecule of angiotensin I (m/z 649), Leu-enkephalin (m/z 556), vasoactive intestinal peptide (m/z 714), and Glu1-fibrinopeptide B (m/z 786). The use of ammonium acetate enabled separation of individual analytes as anions though the mass spectrometer was set for positive ion detection. The peak width (half-height definition) for individual components ranged from 3.5 to 10 s, which provided separation efficiencies of 500-3500 theoretical plates for a total separation channel length of 5 cm. The device was further evaluated for the analysis of more complex samples such as those obtained from proteolytic digests where the added benefit of the second mass spectrometry separation enabled selective analyte detection even for comigrating components. This is illustrated in Figure 3 for the analysis of tryptic peptides from the digest of the seed lectin from D. biflorus. Analytical Chemistry, Vol. 71, No. 15, August 1, 1999

3039

Figure 2. Chip-CE-nESMS analysis of a mixture of peptide standards (positive ion mode). The device PCRD2 was coupled to a 1-cm untreated fused-silica capillary tip (15 µm i.d): (a) TIE (m/z 500-1200) and RIE for (b) m/z 649, (c) m/z 556, (d) m/z 714, and (e) m/z 786. Conditions: separation buffer, 30 mM ammonium acetate pH 8.8. +5 kV was applied to well B (see configuration) for both separation and ionization. Support solution was applied from well D with flow rate of 50 nL/min. A 4-s injection at 2.3 kV was made between wells A and C (Figure 1a) resulting in the injection of 0.7 nL of sample. A voltage of 4.9 kV was applied to A and C during the separation period.

The TIE (m/z 500-1200) of the corresponding Chip-CE-nESMS analysis is presented in the top panel of Figure 3. A number of singly and doubly charged tryptic peptides were identified from the same analysis as indicated from the corresponding RIE traces (Figure 3b-j). A list of tryptic peptide masses observed from the extracted mass spectra (not shown) was used to search a protein database (EMBL web site: http://www.mann.embl-heidelberg.de/ Services/PeptideSearch /FR_PeptideSearchForm.html). As expected, the first protein candidate encoding the highest number of peptide masses (eight hits out of nine masses) was a seed lectin (accession number Q39666). The assignment of tryptic peptides based on the sequence of this protein is presented in Table 1. Aside from the high-throughput potential of the chip-CE-nESMS technique for protein identification, an obvious benefit of this device is the capability of obtaining discrete peptide peaks separated from other sample interferences. Such difficulties are often encountered in infusion electrospray analyses of gelseparated proteins. Furthermore, it is expected that the resolution and separation efficiency can be improved with longer separation channels and/or using coated channel surfaces. An additional advantage of this device is the possibility of easily replacing the nanoelectrospray emitter by dissolving the adhesive at the junction with the microfabricated device, a process that normally takes less than 15 min. Chip-CE-µESMS Analyses Using a Sheath Flow Interface. The coupling of the chip device to ESMS can also be achieved 3040 Analytical Chemistry, Vol. 71, No. 15, August 1, 1999

using more traditional interfaces that combine either a sheath flow or a liquid junction arrangement.35 This can be accomplished using an extended fused-silica capillary (>10 cm) which is inserted in the microfabricated device via a low dead volume connector similar to that described in the previous section. This CE column also acts as an additional electrophoresis channel and provides further temporal resolution of the analyte bands. In this configuration, the electrospray voltage is independently applied to a stainless steel needle, which is also used to deliver the sheath solution at a flow rate of 1-2 µL/min via a tee located near the exit of the column (Figure 1b). This approach provides an independent means of modifying the composition of the electrospray buffer for enhanced sensitivity, while simultaneously maintaining continuity of the voltage gradient across the CE capillary. These devices were primarily intended for electrophoretic separation of peptides with mass spectrometric detection, and initial tests were conducted using negative ion mode for both analyte separation and ionization. A capillary column was attached to the uncoated chip and showed that peptide analysis was feasible using this chip-CE-µESMS configuration. A simple peptide mixture containing angiotensin I, Leu-enkephalin, vasoactive intestinal peptide, and Glu1-fibrinopeptide B could be easily separated and detected using an electrolyte buffer containing 30 mM morpholine, pH 9.0 (data not shown). The theoretical plate number for Leu-enkephalin reached 23 000 when 15 cm of uncoated capillary was coupled to the device. It is noteworthy that

Figure 3. Chip-CE-nESMS analysis of tryptic peptides from D. biflorus: (a) TIE (m/z 500-1000), (b) m/z 603, (c) m/z 757, (d) m/z 717, (e) m/z 986, (f) m/z 615, (g) m/z 969, (h) m/z 953, (i) m/z 549, and (j) m/z 989. Conditions as for Figure 2 except that a 30-s injection at 2.3 kV (10 nL) was used.

Table 1. Assignment of Tryptic Peptides from the Digest of a Seed Lectin from D. biflorus (Figure 3) mass migration time (min) obs calc 2.28

602

2.35 2.49 2.49 2.53 2.56 2.60 2.70 2.95

1511 1430 1968 1226 968 1902 1092 1974

position

601.4 29-33/ 75-80 1511.7 60-74 1430.7 48-59 1968.0 10-28 1226.6 36-47 967.5 186-193 1903.0 81-99 1094.6 138-147 1974.0 233-250

sequence (K)ISAPSK/(K)LQLTK (K)STGAVASWATSFTVK (R)AFYSPIQIYDK (K)NFNSPSFILQGDATVSSGK (K)ENGIPTSSLGR (R)TSYILSER (K)ASFADGIAFALVPVGSEPR (K)HIGIDVNSIK (K)LPDDSTAEPLDLASYLVR

an extension of the CE capillary from 15 to 40 cm provided an enhancement of separation efficiency of approximately 30-55% for most peptides studied. The range of performance of the chip-based system varied substantially with the nature of the sample as in conventional CE. The separation of small organic and inorganic ions, which do not significantly interact with the silica walls, can provide separation efficiencies in excess of 200 000 theoretical plates. However, the use of acidic buffers, often required when positive ion detection is used, can result in reduced EOF and severe adsorption of analytes to the untreated microchannel. This was also evidenced by the relatively large peak width (3.5-10 s) observed with the uncoated chip and capillary device (see above). Modification of the silanol groups present on the channel surface is often required

to enhance the analytical performance of the device. To this end, we coated both the chip and the transfer capillary with BCQ to impart an overall positive charge to the silica surface.12 The chipCE-µESMS analysis conducted using SIM is shown in Figure 4 for the separation of a 20 µg/mL mixture of each synthetic peptide. As indicated, the separation efficiencies ranged from 26 000 to 58 000 theoretical plates for the different analytes with peak widths extending from 1.7 to 3.2 s. Although higher separation efficiencies were expected for the longer separation channel used in the BCQcoated configuration (20 cm for BCQ device compared to 8 cm for the uncoated device), the narrower peak width observed here is attributed to reduced band diffusion and minimization of adsorption of these analytes on the silanol surface. However, evidence of analyte adsorption was noted for more basic peptides (substance P and bradykinin), which displayed broader peak width compared to earlier migrating peptide of comparable molecular weight (oxytocin, bombesin). The separation performance of the chip CE-ESMS was compared with that of conventional CE-ESMS, and the data obtained are summarized in Table 2. It is noteworthy that separation efficiencies obtained for the chip-CE-µESMS are typically a factor of 2 lower than that obtained for the conventional CE-µESMS interface. The field strengths were 330 and 360 V/cm for the conventional and chip-CE-µESMS configurations, and the enhancement of theoretical plate numbers observed in the conventional CE-ESMS experiment was associated with the extended capillary (90 cm compared 15 cm for the chip-CE) Analytical Chemistry, Vol. 71, No. 15, August 1, 1999

3041

Table 2. Limit of Detection and Peak Parameters (SIM Acquisition Mode) theoretical plate no.

migration timeb

peak areab

peptides

LODa (nM)

conventionalc

chip-CEd

time (min)

RSD (%)

mean (×105)

RSD (%)

oxytocin Met-enkephalin Leu-enkephalin bombsin LHRH Arg8-vasopressin bradykinin1-5 substance P bradykinin

149 87.2 45.0 105 127 231 437 297 472

102682 84422 111034 127037 112497 95393 77494 81969 93626

49300 52100 52500 58400 38000 38200 N/A 35700 26700

3.54 3.64 3.68 4.05 4.46 4.68 5.35 5.61 6.17

1.14 1.16 1.12 1.37 0.967 0.984 1.26 1.52 1.48

1.10 3.04 6.29 1.45 4.25 1.73 3.72 2.24 2.26

8.94 9.47 5.80 7.70 8.09 12.1 11.6 11.2 5.60

a

Based on a 10-nL injection. b Replicate injection, n ) 5. c Capillary length of 90 cm. d Capillary length of 15 cm, 0.7-nL injection.

Figure 4. Chip-CZE-µESMS analysis of a mixture of nine peptides at 20 µg/mL using a sheath flow configuration: (a) TIE (∑ 8 ions) and RIE for (b) oxytocin (m/z 504), (c) Met-enkephalin (m/z 575), (d) Leu-enkephalin (m/z 556), (e) bombesin (m/z 811), (f) LHRH (m/z 592), (g) Arg8-vasopressin (m/z 543), (h) substance P (m/z 675), and (i) bradykinin (m/z 531). Conditions: The microfabricated device was coated using BCQ and connected to a microsprayer through a 15cm BCQ-coated capillary (50 µm i.d., 185 µm o.d.). Separation conditions: 0.1 M formic acid, -3.8 kV at separation reservoir and +4.2 kV at microsprayer, -2.8 kV at both sample reservoir and sample waste reservoir during separation, and 4-s injection at 2.3 kV resulting in the introduction of 0.7 nL of sample. A sheath flow containing 0.1 M HCOOH in 25% methanol was introduced at a flow rate of 1.5 µL/min to the interface.

though progressive increase in band diffusion is expected for longer separation times. The analysis shown in Figure 4 provided relatively good separation efficiency for a sample loading of 0.7 nL but resulted in modest concentration detection limits (>1 µM). To improve the signal detectability in full-scan analysis, enhancement of sample loading (10-nL injection) was achieved through sample stacking by dissolving the peptide mixture in a 10-fold dilution of separation buffer. As a result of larger sample loadings, smaller separation efficiencies (∼8000-10 000 theoretical plates) than that typically observed for Figure 4 were obtained for a 15-cm capillary column. An example of separation obtained in full-mass scan acquisition is presented in Figure 5 for a mixture of nine peptides (20 µg/mL each). The TIE for this analysis is shown in Figure 5a along with reconstructed ion electropherograms and extracted mass spectra taken from the same analysis (Figure 5b-j). The 3042 Analytical Chemistry, Vol. 71, No. 15, August 1, 1999

mass spectrum for each peptide is dominated by singly and/or doubly protonated molecules of the corresponding peptides. The sensitivity of this chip-CE-µESMS interface was investigated using a dilution series of the same peptide mixture. The dependence of peak area on sample concentration was examined for replicate injections of serial dilutions of this peptide mixture using SIM acquisition mode, and the results obtained are summarized in Table 2. In all cases, good linearity, with a correlation coefficient r2 ) 0.997-0.999 was found over analyte concentrations ranging from 50 ng/mL to 5 µg/mL. The only peptides that appeared to deviate significantly from linearity at low concentrations were basic peptides such as bradykinin and its fragment 1-5. This lower dynamic range was attributed to analyte adsorption, which also has practical implications for limits of detection. As indicated in Table 2, the concentration detection limits for all the peptides ranged from 45 to 440 nM and were consistent with early reports using CE-µESMS.13 In is interesting to note that the limits of detection were also influenced by the nature of the analyte. Generally, low-nanomolar concentration detection limits were obtained for aliphatic peptides, whereas the detection limit is reached at concentrations of ∼10-7 M for peptides containing a larger number of basic amino acids such as bradykinin, LHRH, and substance P. This lower sensitivity is attributed not only to the larger peak widths associated with these later migrating components but also to analyte adsorption effects arising from underivatized silanol groups on the capillary walls. In an effort to enhance the sensitivity and to reduce the analysis time, we are currently investigating chip devices comprising short gold-coated nanoelectrospray emitters (