Anal. Chem. 1998, 70, 3553-3556
A Microfabricated Dialysis Device for Sample Cleanup in Electrospray Ionization Mass Spectrometry Naxing Xu,† Yuehe Lin,† Steven A. Hofstadler,† Dean Matson,‡ Charles J. Call,† and Richard D. Smith*,†
Environmental Molecular Sciences Laboratory and Materials Resources Department, Pacific Northwest National Laboratory, P.O. Box 999, Richland, Washington 99352
A laser microfabricated device was constructed for rapid microdialysis cleanup of biological samples for analysis by electrospray ionization mass spectrometry (ESI-MS) in both off-line and on-line modes. A microdialysis membrane was sandwiched between two chips having micromachined serpentine channels. The total volume of the sample serpentine channel used for the microdialysis was 1 µL. Efficient desalting was demonstrated for both DNA and protein samples using ESI with an ion trap mass spectrometer after microdialysis against a counter flow of ESI-compatible buffer. Signal-to-noise ratios were also greatly enhanced compared to direct infusion of the original nondialyzed samples. Importantly, the microfabricated device allowed use of sample flow rates 1 order of magnitude smaller than previous designs based on a microdialysis fiber, allowing reduced sample utilization and improved sensitivity with ESI-MS. The effectiveness of the cleanup was attributed to the size difference between the sample channel and the buffer channel and the fact that the sample is continuously refreshed by the buffer counterflow. The results indicate substantial potential for construction of highly compact and rugged devices enabling field applications of ESI-MS. Miniaturized analytical instrumentation is attracting growing interest due to the potential for enhanced sensitivity and speed as well as the large reductions in overall size and weight. Microfabricated devices have been utilized for chemical separations based on electrophoresis,1-2 open-channel chromatography,3 and micellar electrokinetic chromatography,4 as well as for performing chemical reactions such as solid-phase chemistry,5 polymerase chain reactions,6 and enzyme assays.7 Most of these * Corresponding author: (phone) (509) 376-0723; (fax) (509) 376-2303; (e-mail)
[email protected]. (1) Jacobson, S. C.; Hergenroder, R.; Koutny, L. B.; Ramsey, J. M. Anal. Chem. 1994, 66, 1114-1118. (2) Woolley, A. T.; Mathies, R. A. Anal. Chem. 1995, 67, 3676-3680. (3) Jacobson, S. C.; Hergenroder, R.; Koutny, L. B.; Ramsey, J. M. Anal. Chem. 1994, 66, 2369. (4) Moore, A. W.; Jacobson, S. C.; Ramsey, J. M. Anal. Chem. 1995, 67, 4184. (5) Fodor, S. P. A.; Read, J. L.; Pirrung, M. C.; Stryer, L.; Lu, A. T.; Solas, D. Science 1991, 251, 767. (6) Woolley, A. T.; Hadley, D.; Landre, P.; Demello, A. J.; Mathies, R. A.; Northrup, M. A. Anal. Chem. 1996, 68, 4081-4086. S0003-2700(98)00233-9 CCC: $15.00 Published on Web 08/06/1998
© 1998 American Chemical Society
devices reported to date have been fabricated from glass after photolithography, wet chemistry, and chemical-etching processes adapted from microelectronics manufacturing procedures. The combination of such “nanoscale” manipulations in small microfabricated devices with electrospray ionization mass spectrometry (ESI-MS) is particularly attractive because of the speed and sensitivity that can be achieved with MS as well as the amenability of ESI-MS to the low flow rates characteristic of such devices. It is well recognized that lower CE-MS flow rates with ESI-MS, for example, provide greater sensitivity.8 Recently, the potential for use of microchips in combination with ESI-MS has been reported by several groups.9-13 Multiple-channel microchips have been directly interfaced to ESI-MS, demonstrating the potential for high-throughput analysis.12-13 A microfabricated device was demonstrated for automating the continuous injection approach to protein identification using microelectrospray ion trap mass spectrometry.11 ESI-MS is amenable to the analysis of nearly the entire range of biomolecules, and the use of mild ionization conditions allows noncovalent associations to be directly detected. However, sample matrix interference (generally due to low-molecular-weight salts) presents a major limitation on the use of ESI-MS for analyzing large biopolymers. A variety of techniques have been explored for reducing sample matrix contributions and improving spectrum quality in ESI-MS.14-16 Recently we have shown that an on-line microdialysis approach is particularly effective for rapid sample cleanup, greatly enhancing the ability to address complex biologi(7) Hadd, A. G.; Raymond, D. E.; Halliwell, J. W.; Jacobson, S. C.; Ramsey, J. M. Anal. Chem. 1997, 69, 3407-3412. (8) Smith, R. D.; Wahl, J. H.; Goodlett, D. R.; Hofstadler, S. A. Anal. Chem. 1993, 65, 5, A574-A584. (9) Henry, C. Anal. Chem. 1997, 69, 359A-361A. (10) Ramsey, R. S.; Ramsey, J. M. Anal. Chem. 1997, 69, 1174-1178. (11) Figeys, D.; Ning, Y.; Aebersold, R. Anal. Chem. 1997, 69, 3153-3160. (12) Xue, Q. F.; Foret, F.; Dunayevskiy, Y. M.; Zavracky, P. M.; Mcgruer, N. E.; Karger, B. L. Anal. Chem. 1997, 69, 426-430. (13) Xue, Q. F.; Dunayevskiy, Y. M.; Foret, F.; Karger, B. L. Rapid Commun. Mass Spectrom. 1997, 11, 1253-1256. (14) Little, D. P.; Chorush, R. A.; Speir, J. P.; Senko, M. W.; Kelleher, N. L.; McLafferty, F. W. J. Am. Chem. Soc. 1994, 116, 4893-4897. (15) Stults, J. T.; Marsters, J. C. Rapid Commun. Mass Spectrom. 1991, 5, 359363. (16) Emmett, M. R.; Caprioli, R. M. J. Am. Soc. Mass Spectrom. 1994, 5, 605613.
Analytical Chemistry, Vol. 70, No. 17, September 1, 1998 3553
cal samples.17-19 This approach employs a small microdialysis tube through which the sample flows in conjunction with a countercurrent flow of a suitable buffer. The dialyzed samples can be directly interfaced with ESI source at a flow rate of approximately 2-5 µL/min or collected and used with low-flowrated, “constrained” variations of electrospray ionization sources (i.e., nanospray or microspray, 10-300 nL/min) to greatly reduce sample consumption and increase sensitivity.20 While the microdialysis approach has demonstrated great promise for rapid and efficient sample cleanup capability, dialysis fibers of less than 200-µm diameter (used in our earlier work) are not yet available and contributed to the minimum useful flow rates and sample-processing time achieved previously. It is clear that miniaturization of the approach will not only further increase its speed, flexibility, and robustness but also allow the use of lower rates and result in enhanced sensitivity. In this work, we report a microfabricated device constructed for this purpose and demonstrate its desalting efficiency for both DNA and protein samples. EXPERIMENTAL SECTION Construction of the Microfabricated Device. Serpentine flow channels were fabricated on 30 × 30 × 6 mm polycarbonate chips using a Potomac LMT-4000 excimer laser direct-write micromachining system (Potomac Photonics, Lanham, MD).21 The sample channel was machined directly into the polycarbonate chip by using multiple parallel laser passes to achieve a ∼160-µm channel width. The total length of the sample channel was 11 cm. The depth of the channel was limited to ∼60 µm by selecting the appropriate laser repetition rate (1 kHz) and machining rate (5 µm/s). A 500-µm-wide buffer channel was machined from a ∼250-µm-thick polyimide sheet having a 25-µm-thick silicone adhesive on one side. The buffer channel was produced by laser cutting around the channel perimeter followed by stripping the remaining polyimide and adhesive from the center. The adhesive was used to attach the buffer channel to a flat polycarbonate chip. Fluid access was provided by mechanically drilling through the chips and into headers at either end of the microchannels. In this preliminary work, holes through the backsides of the chips were counterdrilled and tapped to accept standard fittings for use with fused-silica capillary tubing. Figure 1 shows a presentation of the overlapping sample and buffer channels after alignment of the microchannel analysis chips. A Spectra/Por Biotech 1.1 dialysis membrane with a molecular weight cutoff (MWCO) of 8000 (Spectrum, Houston, TX) was sandwiched between the chips. The channels were connected to fused-silica capillary (200 µm o.d. × 100 µm i.d.) and sample and buffer solutions were connected so that they were in a counterflow configuration. The sample flow rate was 0.5 µL/min and the buffer flow rate was 100 µL/min. All dialysis experiments were carried out at room temperature (∼25 °C). (17) Wu, Q.; Liu, C. L.; Smith, R. D. Rapid Commun. Mass Spectrom. 1996, 10, 835-838. (18) Liu, C. L.; Wu, Q. Y.; Harms, A. C.; Smith, R. D. Anal. Chem. 1996, 68, 3295-3299. (19) Muddiman, D. C.; Wunschel, D. S.; Liu, C. L.; Pasa Tolic, L.; Fox, K. F.; Fox, A.; Anderson, G. A.; Smith, R. D. Anal. Chem. 1996, 68, 3705-3712. (20) Liu, C. L.; Muddiman, D. C.; Tang, K. Q.; Smith, R. D. J. Mass Spectrom. 1997, 32, 425-431. (21) Robers, M. A.; Rossier, J. S.; Bercier, P.; Girault, H. Anal. Chem. 1997, 69, 2035-2042.
3554 Analytical Chemistry, Vol. 70, No. 17, September 1, 1998
Figure 1. Schematic representation of the image of overlapping sample and buffer channels after alignment of the microchannel device.
Materials. The following samples were prepared: (1) 1.5 µM horse heart cytochrome c (Sigma, St. Louis, MO) in 10 mM NH4OAc and 100 mM NaCl; (2) 10 µM 30-mer oligonucleotide (5′AGCTAATTCGTTGCATCGATTAAGCAACGT-3′) (The Midland Certified Reagent Co.) in 10 mM NH4OAc; (3) 5 µM horse heart myoglobin (Sigma) in 500 mM NaCl, 100 mM Tris, and 10 mM EDTA. Mass Spectrometry. A Finnigan LCQ (San Jose, CA) quadrupole ion trap mass spectrometer with a modified micro-ESI source was used for all experiments. The heated inlet capillary was maintained at 160 °C. The sample flow rate for direct infusion was 0.3 µL/min. A coaxial SF6 gas flow around the ESI emitter was used to suppress corona discharge for DNA samples in negative ion mode. A typical electrospray voltage was 2.2 kV. The maximum sample injection time was 40 ms, and three microscans were summed for each scan with a m/z range of 1502000. RESULTS AND DISCUSSION To evaluate the efficiency of the microfabricated dialysis unit, a 10 µM 30-mer oligonucleotide was analyzed by ESI-MS before and after the dialysis. As shown in Figure 2A, direct infusion of the oligonucleotide in 10 mM NH4OAc produced a spectrum seriously compromised by sodium adduction and concomitant low signal-to-noise ratio (S/N), even though the inherent salt content of the oligonucleotide was expected to be low. After the off-line dialysis by the microfabricated device with 10 mM NH4OAc as countercurrent buffer, a dramatic improvement in spectrum quality was evident (Figure 2B), leading to more accurate molecular weight determination for the analyte (MWcalc ) 9205.0, MWobs ) 9204.1 ( 1.6). A more than 20-fold increase in S/N was achieved compared to Figure 2A. The microfabricated dialysis device was also evaluated for the desalting of protein samples. Figure 3A shows a mass spectrum of horse heart cytochrome c at a concentration of 1.5 µM in 10 mM ammonium acetate and 100 mM NaCl obtained by direct infusion with the microspray ESI source. No useful analytical signal was obtained. With the use of the microfabricated dialysis
Figure 2. (A) ESI-mass spectrum of a 10 µM 30-mer oligonucleotide (5′-AGCTAATTCGTTGCATCGATTAAGCAACGT-3′) in 10 mM NH4OAc from direct infusion; (B) ESI-mass spectrum of the 10 µM 30mer oligonucleotide in 10 mM NH4OAc after dialysis using the microfabricated device.
device in the off-line mode with 10 mM ammonium acetate as the dialysis buffer, effective desalting was achieved as shown in Figure 3B. Charge states at 7+, 8+, and 9+ for the analyte were clearly present in the spectrum, characteristic of the native protein electrosprayed in such buffer. With the current design, the total volume of the sample serpentine channel is only 1 µL and the actual dialysis time for the sample, i.e., the time the sample spent inside the microdialysis channels, was ∼2 min at a sample flow rate of 0.5 µL/min. The effectiveness of the cleanup can be ascribed to the size difference between the sample channel and the buffer channel and the fact that the cleanest sample solution is effectively exposed to the cleanest dialysis buffer, due to its countercurrent flow. For effective desalting, It has been found that air bubbles trapped inside the sample channels need to be forced out before the dialysis. The sample loss during the dialysis is minimal due to the short residence time in the microdialysis channel18 and the fact that the analytes are above the MWCO of the chosen membrane. The Spectra/Por Biotech 1.1 membrane used for this work has high tolerance to acidic solvents such as 90% acetic acid and organic solvents (e.g., 90% methanol), providing versatility in application of the microfabricated dialysis device. The microfabricated dialysis device can be used in the on-line mode for direct ESI-MS analysis and has been evaluated for this purpose using the arrangement shown in Figure 4. High voltage was applied to the metal syringe needle used for buffer delivery and provided electrical contact through the buffer and sample solution. In this arrangement, the ESI emitter was a short piece
Figure 3. (A) ESI-mass spectrum of 1.5 µM horse heart cytochrome c in 10 mM NH4OAc and 100 mM NaCl by direct infusion; (B) ESImass spectrum of 1.5 µM horse heart cytochrome c in 10 mM NH4OAc and 100 mM NaCl after dialysis using microchannel chips.
Figure 4. Schematic representation of the setup of the microfabricated device for on-line ESI-MS analysis. Top panel shows a photograph of the sample channel chip.
of silica capillary (4-6 cm, 200 µm o.d. × 100 µm i.d.) fixed to the chip through a standard fitting (1/32 in., Valco Instrument Co. Inc., Houston, TX). To ensure the safety of the operator, the connection capillary for the outlet of the buffer solution was contained inside a plastic vial. A 5 µM horse heart myoglobin solution in a complex matrix consisting of 500 mM NaCl, 100 mM Analytical Chemistry, Vol. 70, No. 17, September 1, 1998
3555
19+ to 10+ of denatured myoglobin were clearly resolved and the interference peaks from the Tris and EDTA greatly diminished. The S/N was increased by a factor of more than 40 compared to that obtained without desalting. The microdialysis device is readily fabricated on inexpensive polymer substrates such as polycarbonate or polyimide; therefore, there is a great potential for large-scale production of inexpensive and disposable device. Integration of this microfabricated dialysis device with other nanoscale manipulations is expected to lead to more powerful, sensitive, less expensive, and rugged instrumentation for chemical analysis while eliminating carry-over or crosscontamination problems in routine analysis.
Figure 5. (A) ESI-mass spectrum of 5 µM horse heart myoglobin in 500 mM NaCl, 100 mM Tris, and 10 mM EDTA by direct infusion; (B) ESI-mass spectrum of previous myoglobin sample after on-line microdialysis using 10 mM NH4OAc and 1% acetic acid as dialysis buffer.
Tris, and 10 mM EDTA was studied as an example. Figure 5A shows a mass spectrum obtained for this solution by direct infusion with the microspray ESI source. No protein charge-state envelope was evident, and the spectrum was complicated by undesired peaks due to Tris and EDTA (commonly used in protein sample preparation). On-line microdialysis was carried out by using dialysis buffer solution containing 10 mM NH4OAc and 1% (v/v) acetic acid at a flow rate of 100 µL/min, while the sample was introduced at a flow rate of 0.5 µL/min. As shown in Figure 5B, effective desalting was achieved. Multiple charge states from
3556 Analytical Chemistry, Vol. 70, No. 17, September 1, 1998
CONCLUSIONS We have demonstrated a microfabricated device for rapid desalting of DNA and protein samples prior to ESI-MS analysis in both off-line and on-line modes of operation. Initial results from both off-line and on-line experiments indicated the broad applicability of this device. This device could also facilitate the analysis of samples that are destablized (e.g., aggregate) by rapid removal of salts and other buffer constituents and also the analysis of higher-order structures of biopolymers such as noncovalent complexes. Importantly, the microfabricated device enable the use of lower flow rates for “on-line” ESI-MS, which provides decreased sample utilization (enhanced sensitivity). Significant potential for further size reduction and speed enhancement exists using the present approach and methods for its integration with other sample manipulations are presently being explored. ACKNOWLEDGMENT We thank Dr. Chuanliang Liu for helpful discussions. We thank the Human Genome Program, Office of Biological and Environmental Research, U.S. Department of Energy, for support of this work. Pacific Northwest National Laboratory is operated by Battelle Memorial Institute from the U.S. Department of Energy, through Contract DE-AC06-76RLO 1830. Received for review March 4, 1998. Accepted June 26, 1998. AC980233X
