Lowell Ericsson, University of Washington
and Ronald L. Niece, University of California at Irvine
Introduction
The ABRF Symposium held in conjunction with the Tenth Protein Society Meeting took place in San Jose, CA on Saturday, August 3, 1996. Approximately 200 people attended this symposium on Capillary Electrophoresis and Mass Spectrometry. The two speakers at the symposium have been pioneers in the field of capillary electrophoresis (CE) using mass spectrometry (MS) as the detection system. Dick Smith published the first paper describing on-line CE -MS in 1987 (1), and Curtis Monnig published one of the early papers on the off-line technique in 1992 (2).
Several reviews on this topic have recently appeared (3 -5); special attention should be paid to a recent article by Dick Smith (9), because the micro-dialysis membrane tubing junction he describes appears to reduce many problems encountered with on-line MS. Off-line MS may be attractive to many laboratories who do not wish to tie up an electrospray ionization mass spectrometer (ESI-MS) by coupling it to a capillary electrophoresis unit. At the present time the off -line mode is the only commercial option for matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF MS).
Richard D. Smith, Pacific Northwest National Laboratory, "On-line capillary electrophoresis and mass spectrometry".
Separations using capillary zone electrophoresis (CZE) are rapid because the small-diameter tubing allows effective heat dissipation, permitting use of high electric fields. Flow rates vary from zero to hundreds of nanoliters per minute. CZE is capable of producing separations that are very fast with extremely high resolution; as many as 300,000 theoretical plates per meter have been observed when CZE and MS are coupled. However, limitations on sample injection volumes create significant demands on detection technology. Often other versions of capillary electrophoresis, such as capillary isotachophoresis (CITP), are used to concentrate samples.
An advantage of on-line separations is that a large number of spectra can be used to detect weak signals. An attribute of CE is the absence of high pressure; instead, the
flow of liquid is easily manipulated by application of electric fields. Sensitivity in the MS can be increased by lowering the electric field on the capillary at the time the sample elutes to extend data acquisition times: separation efficiency is not sacrificed and signal intensities do not decrease. Ordinarily, the CE separation moves charged particles to the electrospray component of the MS faster than the electrospray unit can produce gas phase ions. Smaller diameter capillaries do not lead to increased sensitivity because analytes must form droplets to elute from the capillary.
There are three variations for interfacing CE and ESI -MS. Variation 1 is a direct interface that involves removing the polyimide coating from the surface of the capillary tip and coating the exposed capillary surface with a layer of silver for electrical contact (1, 6). Electroosmotic flow is much more stable if the tubing is etched to form a conical end. Variation 2 is a coaxial sheath-flow interface where liquid is added at the tip of the capillary (7) [in contrast to the liquid-junction approach where liquid is added a few cm before the end of the capillary (8)]. Variation 3 is a microdialysis junction made by butting a short piece of the electrospray emitter capillary to the CE capillary and inserting this joint into a microdialysis capillary immersed in a small amount of liquid (9). The microdialysis junction provides better signal to noise compared to the sheath-flow or liquid -junction, because little additional liquid is added, but like the liquid-junction interface there is some dead volume at the joint.
Among the MS technologies that can be interfaced with CE, the standard is triple quadrupole technology. Orthogonal time-of-flight mass spectrometry (TOF-MS) is fast. The quadrupole ion trap is easier to use than the Fourier transform ion cyclotron resonance mass spectrometer (FT-ICR). FT -ICR is slower and requires a superconducting magnet making it expensive; however, coupling CE to FT-ICR provides highly effective multistage mass spectrometry, extremely high resolution, and the best sensitivity available in mass spectrometry.
Isotachophoresis is a capillary format that can be coupled to MS. An advantage is that isotachophoresis focuses like isoelectrofocusing, using leading and trailing electrolytes to bracket the electrophoretic mobilities of the compound of interest. Capillary isoelectrofocusing can be combined with ESI-MS, because the ampholyte does not compete with the sample during the electrospray process.
A microdialysis junction can be used with a countercurrent flow of ammonium acetate to clean up a sample for electrospray. One trip through an on-line microdialysis capillary (about 10 cm long) is more effective than ten Centricon clean-up steps. Solution conditions can be modified to enhance or reduce non-covalent interactions before electrospray.
Sampling small volumes such as single red blood cells yield about 450 attomoles of hemoglobin. Both ends of the capillary are prepared, one for cell sampling to introduce only one cell and the other for "sheathless" electrospray, where the terminus of a 20-micron capillary is coated with gold. Compounds not separated during capillary electrophoresis can be distinguished by their mass.
Benefits of on-line CE-MS are sensitivity through less sample handling, the feasibility of tandem MS especially for large molecules (in part because of higher charge states), the preservation of separation quality, and the ability to study non-covalent associations in solution.
Curtis Monnig, University of California at Riverside, "Off -line capillary electrophoresis and mass spectrometry".
CE-MS can be approached from a separations viewpoint. On-line coupling is probably the more popular approach to interfacing CE and MS instruments. It permits minimum sample handling, which reduces potential sample loss and the labor involved, and uses the mass spectrometer as an extremely selective detector, collecting either total ion electropherograms or selected ion electropherograms.
However, there are several problems associated with on-line CE-MS. The number of ionization methods that can be used to couple the techniques is limited. From a separations point of view, a primary concern is that preferred CE buffers may be incompatible with the ion source; most popular CE buffers are not volatile and so often cause unstable behavior with the electrospray source and may contaminate the source. Timing can be an issue when attempting to collect MS data, especially in MS/MS experiments, from CE peaks of short duration. Dedicating MS instruments to CE is often not feasible, because of the location of the instruments and time -sharing arrangements in laboratories.
Off-line coupling offers a flexibility that minimizes several of these problems (10-12). In addition, it is simpler than on-line CE-MS, and the CE system can be optimized separately from the mass spectrometer. Less transfer of electrophoresis solvent further simplifies the operation of both instruments.
Off-line coupling offers the possibility for using different types of ion sources, such as MALDI, plasma desorption, and electrospray. Flexibility with the ion source offers the opportunity to generate simplified spectra and enhance ionization of different classes of molecules. Temporal
decoupling of CE and MS can allow more complicated experiments to be performed and can be convenient. Samples can be collected and held until the mass spectrometer becomes available. Post-separation treatment of samples before mass analysis can improve results. Cost savings result from the use of non-dedicated instrumentation with little modification of either instrument.
Off-line coupling is really fraction collecting. An ideal fraction collector for CE has the following characteristics: automatic, minimal voltage switching during the run, minimal dilution or contamination of sample, simple assembly, little sample carryover, and little loss in the efficiency of separation by placing each zone in a single spot. Most commercial CE instruments can collect fractions using built-in software for peak selection. However, using CE instruments with fraction collectors lowers eluted sample concentrations because the sample reservoirs must contain electrolyte, so peaks of a few nanoliters are diluted into tens-of-microliter volumes.
Alternatives to the fraction collection systems built into many commercial CE instruments exist. Alternative 1 uses a "cracked" capillary (13). An electrical connection is introduced to the central channel before the end of the capillary, permitting grounding of the capillary away from the terminus. Droplets form at the end where they can be collected for further analysis. However, reproducibly forming the crack is a difficult manufacturing problem, and the velocity of analyte zones changes after the crack. Losing the electroosmotic driving force at the crack can cause uncertainty about when the zone will elute from the end of the capillary for collection.
Alternative 2 is depositing the material on membrane surfaces. For example, the end of the capillary touches a rotating disk of PVDF on top of grounded filter paper, depositing the material directly on the membrane. However, it can be difficult to select a membrane that will bind all substances of interest, and removing sample from the membrane requires additional handling.
Alternative 3 uses coaxial-flow capillaries downstream of UV detectors. Electrolyte is pumped through a capillary that surrounds the separation capillary, constantly bathing it to maintain electrical contact without the need to immerse it in a buffer reservoir. Collection devices, in this case probes, are positioned under the capillary outlet. Samples are collected through computer control when peaks are detected by UV absorbance. By correcting for sample velocity (from electroosmotic flow or hydrostatic head when coated capillaries are used) and the length of the capillary after the detector, each peak can be spotted on a different probe for later analysis.
Alternative 4 replaces coaxial-flow, which diffuses the sample on the probe surface, with a gold-epoxy conductive coating on a beveled capillary outlet. Electroosmotic flow causes liquid to form small droplets, which are deposited on probes. Fractions as small as 50 nanoliters can be collected reproducibly.
MALDI is a useful technique for direct analysis of relatively simple mixtures, but for complex mixtures, it is necessary to first separate components. CE separation systems using a variety of buffers lead to usable spectra. However, some buffers work better than others for signal intensity and resolution. At concentrations of 30 mM, Tris-acetate, pH 5.0, is a better general purpose CE buffer than sodium phosphate, pH 2.5; and both sodium citrate, pH 4.0, and potassium borate, pH 9.0, are preferred over potassium carbonate, pH 5.0.
Sodium adduct peaks result in degradation of resolution of the instrument. Addition of formic acid at 4-6% in the matrix solution greatly suppresses the formation of sodium and potassium adducts and therefore enhances resolution and signal strength.
Alternative separation systems such as isotachophoresis or isoelectric focusing require changing the fraction collecting system because it is difficult to predict when components reach the end of the capillary. Two detectors on the column permits determination of sample velocity, and a computer can calculate when the sample will emerge.
Coaxial flow arrangements are versatile and collect proteins with high efficiencyon the order of 80%but dilute samples. The conductive capillary interface works only for separations with electroosmotic flow, but collection efficiency is good and dilution is minimal. Automated fraction collection is feasible and simplifies the operation of the CE and the mass spectrometer. This system is extremely tolerant of buffer components and additives. Miniaturization of the interface permits collection of small droplets on probe surfaces, enhances detection limits, and improves separation efficiency.
References
1. J. A. Olivares, N. T. Nguyen, C. R. Yonker, R. D. Smith (1987) "On-Line Mass Spectrometric Detection for CZE" Anal. Chem. 59: 1230-1232.
2. J. A. Castoror, R. W. Chiu, C. A. Monnig, and C. L. Wilkins (1992) "Matrix-Assisted Laser Desorption /Ionization of Capillary Electrophoresis Effluents by Fourier Transform Mass Spectrometry" J. Am. Chem. Soc. 114: 7571-7572.
3. J. Cai and J. Henion (1995) Review "Capillary Electrophoresis-Mass spectrometry" J. of Chroma -togr. A 703: 667-692.
4. R. L. St. Claire, III (1996) "Capillary Electrophoresis" Anal. Chem. 68: 579R-580R.
5. A. L. Burlingame, R. K. Boyd, and S. J. Gaskell (1996) "Mass Spectrometry" Anal. Chem. 68: 599R 651R.
6. H. Wahl, D. C. Gale, and R. D. Smith (1994) "Sheathless Capillary Electrophoresis-Electrospray Ionization Mass Spectrometry Using 10 mm I.D. Capillaries: Analyses of Tryptic Digest of Cytochrome c" J. Chromatogr. A. 659: 217-222.
7. R. D. Smith, J. A. Olivares, N. T. Nguyen, and H. R. Udseth (1988) "CZE-MS Using an Electrospray Ionization Interface" Anal. Chem. 60: 436.
8. E. D. Lee, W. Muck, J. D. Henion, and T. R. Covey (1989) "Liquid Junction Coupling for CZE/Ion Spray Mass Spectrometry" Biomed. Env. Mass Spectrom. 18: 844.
9. J. C. Severs, A. C. Harms, and R. D. Smith (1996) "A New High-performance Interface for Capillary Electrophoresis/Electrospray Ionization Mass Spectrometry" Rapid Communication in Mass Spectrometry 10: 1175-1178.
10. T. Keough, R. Takigiku, M. P. Lacey, and M. Purdon (1992) "Matrix-Assisted Laser Desorption Mass Spectrometry of Proteins Isolated by Capillary Zone Electrophoresis" Anal. Chem. 64 :1594-1600.
11. P. A. van Veelin, U. R. Tjaden, J. van der Greef, and F. Hillenkamp (1993) "Off-Line Coupling of Capillary Electrophoresis with Matrix-Assisted Laser Desorp-tion Mass Spectrometry" J. Chromatogr. 647: 367 -374.
12. K. L. Walker, R. W. Chiu, C. A. Monnig, and C. L. Wilkins (1995) "Off-line coupling of capillary electrophoresis and matrix-assisted laser desorption /ionization time-of-flight mass spectrometry" Anal. Chem. 67: 4197-4204.
13. R. A. Waliingford and A. G. Ewing (1987) "Capillary Zone Electrophoresis with Electrochemical Detection" Anal. Chem. 59: 1762-1770.
Lowell Ericsson may be contacted at the University of Washington, Department of Biochemistry, Seattle, WA 98195-0001, E-mail: ericsson@u.washington.edu; and Ronald L. Niece at the University of California, Biotechnology Resource Facility, Irvine, CA 92697-2526, E-mail: rniece@uci.edu.
Return to the
The ABRF Home
Page