Reflectron Time-of-Flight Mass

An ion trap storage (IT)/reflectron time-of-flight mass spectrometer (reTOFMS) has been coupled to capillary electrophoresis (CE) via a sheathless ...
0 downloads 0 Views 538KB Size
Download PDF
Anal. Chem. 1996, 68, 3388-3396

Use of an Ion Trap Storage/Reflectron Time-of-Flight Mass Spectrometer as a Rapid and Sensitive Detector for Capillary Electrophoresis in Protein Digest Analysis Jing-Tao Wu, Mark G. Qian, Michael X. Li, Lin Liu, and David M. Lubman*

Department of Chemistry, The University of Michigan, Ann Arbor, Michigan 48109

An ion trap storage (IT)/reflectron time-of-flight mass spectrometer (reTOFMS) has been coupled to capillary electrophoresis (CE) via a sheathless microelectrospray ionization method. This hybrid mass spectrometer has proved to be a rapid and sensitive detector for CE, where mass spectra could be acquired at a speed sufficient to maintain the high-resolution capabilities of CE separations. The nonscanning property of the time-of-flight mass analyzer can provide a full mass range spectral acquisition speed of up to 25 spectra/s with a data system developed in our laboratory. For the work reported herein, a spectral acquisition speed of 4 spectra/s was found to be optimal for maintaining the quality of the separation while achieving high sensitivity. Tryptic digests of bovine cytochrome c and β-lactoglobulin A were analyzed using the CE/IT/reTOFMS combination, resulting in total ion electropherograms similar to those obtained using UV absorption detection. Taking advantage of the ion storage capability of the ion trap, a detection limit in the lowfemtomole range was routinely obtained for these digests using the total ion electrophoretic mode and CE capillaries of typical dimensions (41 µm i.d.). This high sensitivity was achieved while maintaining a resolution of ∼1500 for mass identification using the capabilities of the IT/reTOF device. Due to the high acquisition speed and the mass discrimination capabilities of the mass detector, all the peaks in the total ion electropherograms, including some totally or partially unresolved peaks, could be unambiguously identified. Capillary electrophoresis (CE) has generated considerable interest in recent years because of its enormous capability for achieving high separation speed and efficiency in small sample volumes.1-3 In CE, analytes are separated on the basis of the differences of their electrophoretic mobilities, which are mainly determined by their mass-to-charge ratios. CE is therefore particularly suitable for the separation of charged species, such as proteins and peptides. Because of the small dimensions of the capillaries used in CE, on-line detection is usually required. A variety of on-line detection techniques have been developed for CE, such as UV absorbance, fluorescence, and electrochemical (1) Jorgenson, J. W.; Lukacs, K. D. Science 1983, 222, 266-72. (2) Gordon, M. J.; Huang, X.; Pentoney, S. L., Jr.; Zare, R. N. Science 1988, 242, 224-8. (3) Monnig, C. A.; Kennedy, R. T. Anal. Chem. 1994, 66, 280R-314R.

3388 Analytical Chemistry, Vol. 68, No. 19, October 1, 1996

methods.1,4,5 Although they have been shown to be useful for certain applications, these methods generally lack the universality for detection of different samples and have only exhibited limited capabilities for structure elucidation. It is therefore desirable to use mass spectrometric detection for providing both mass and structural information. Coupling CE with mass spectrometry (MS) was first performed by Smith et al. in the late 1980s.6,7 Following this work, several papers have been published in this field by various research groups, as recently reviewed by Cai and Henion.8 This hyphenated method has been proved to be best facilitated by electrospray ionization, which was developed by Fenn and co-workers in the mid-1980s as an ionization technique for mass spectrometry.9 More recently, new electrospray sources that can operate under much lower flow rates have also been developed, such as microelectrospray or nanoelectrospray.10-12 The microelectrospray source is ideal for CE/MS experiments, because it operates at a flow rate of less than 0.5 µL/min, which is compatible with the CE flow. Scanning mass spectrometers, such as quadrupole mass analyzers13,14 and magnetic sectors,15 as well as quadrupole ion traps16-18 and Fourier transform mass spectrometers (FT-ICR),19 have all been used as CE/MS detectors. (4) Hogan, B. L.; Yeung, E. S. Anal. Chem. 1992, 64, 2841-5. (5) Colon, L. A.; Dadoo, R.; Zare, R. N. Anal. Chem. 1993, 65, 476-81. (6) Olivares, J. A.; Nguyen, N. T.; Yonker, C. R.; Smith, R. D. Anal. Chem. 1987, 59, 1230-2. (7) Smith, R. D.; Wahl, J. H.; Goodlett, D. R.; Hofstadler, S. A. Anal. Chem. 1993, 65, 574A-84A. (8) Cai, J.; Henion, J. D. J. Chromatogr. 1995, 703, 667-92. (9) Fenn, J. B.; Mann, M.; Meng, C. K.; Wong, S. F.; Whitehouse, C. M. Science 1989, 246, 64-71. (10) Wilm, M. S.; Mann, M. Int. J. Mass Spectrom. Ion Processes 1994, 136, 16780. (11) Gale, D. C.; Smith, R. D. Rapid Commun. Mass Spectrom. 1993, 7, 101721. (12) Emmett, M. R.; Caprioli, R. M. J. Am. Soc. Mass Spectrom. 1994, 5, 60513. (13) Smith, R. D.; Olivares, J. A.; Nguyen, N. T.; Udseth, H. R. Anal. Chem. 1988, 60, 436-41. (14) Johansson, M.; Huang, E. C.; Henion, J. D. J. Chromatogr. 1991, 554, 31127. (15) Perkins, J. B.; Tomer, K. B. Anal. Chem. 1994, 66, 2835-40. (16) Ramsey, R. S.; Goeringer, D. E.; McLuckey, S. A. Anal. Chem. 1993, 65, 3521-4. (17) Stacey, C; Banks, J. F.; Schubert, M.; Hauser-Fang, A. The 44th ASMS Conference on Mass Spectrometry and Allied Topics, Portland, OR, 1996; WPH 143. (18) Teng, S.; Schnute, B.; Jardine, I.; Schwartz, J. The 44th ASMS Conference on Mass Spectrometry and Allied Topics, Portland, OR, 1996; WPH 144. (19) Hofstadler, S. A.; Wahl, J. H.; Bruce, J. E.; Smith, R. D. J. Am. Chem. Soc. 1993, 115, 6983-4. S0003-2700(96)00405-2 CCC: $12.00

© 1996 American Chemical Society

Recently, Zare and co-workers reported CE/MS results using a time-of-flight mass spectrometer (TOFMS).20 Banks and Dresch and Smith et al. also reported results using similar mass analyzers.21,22 Compared with other mass analyzers, TOFMS has several unique advantages, including its wide mass range and high acquisition speed. These features are especially important when TOFMS is used as an on-line detector for CE. Due to the speed and high efficiency of CE, the resulting peaks are very narrow, usually on the order of a couple of seconds. In these experiments, detection speed is thus critical in maintaining the high separation efficiency and reproducibility of CE. For a scanning mass spectrometer or FT-ICR, obtaining a full mass range spectrum requires ∼1 s or more, which is usually not sufficiently fast to cover the bandwidth of a CE separation as shown in recent work using ion trap mass spectrometers.17,18 Ion trap mass analyzers can reach a very high detection speed, but mass accuracy and resolution are often compromised. In addition, the ion transmission efficiency and duty cycle in the TOFMS are extremely high since most ions pulsed into the TOFMS are detected, which results in excellent sensitivity. Scanning mass spectrometers can only detect one mass at a time, resulting in a poor duty cycle. Also it should be noted that in a scanning mass spectrometer such as a quadrupole, mass resolution is usually compromised for sensitivity, while in TOFMS, sensitivity can be optimized without sacrificing mass resolution. In order to interface electrospray ionization with TOFMS, an orthogonal extraction method has been used.20-24 In this method, a dc extraction pulse is applied orthogonal to the incoming continuous ion beam from the electrospray source to convert the ions into pulsed ion packets that travel the flight tube to the detector. Using this method, high mass resolution and detection speed have been achieved. However, in order to obtain a high duty cycle with this method, a high pulsed extraction repetition rate must be used, which requires specially designed circuitry and software or the use of ion counting. In addition, this device is not capable of MS/MS for structural analysis. In recent work, our laboratory has demonstrated the use of a combination ion trap storage/reflectron time-of-flight mass spectrometer (IT/reTOFMS) as an on-line detector for HPLC.25 The IT/reTOF uses a quadrupole ion trap as a front-end storage device, prior to mass separation and identification by the reTOFMS. In this device, ions over a wide mass range are first accumulated in the ion trap by an rf-only voltage applied on the ring electrode and, after a delay, are simultaneously ejected by a dc pulse applied to the end cap into the reTOF for mass analysis. The ion trap thus serves as a means of converting a continuous electrospray beam into a pulsed beam for analysis by TOFMS. There are several advantages of this device compared with orthogonal extraction TOF when used as an on-line detector for CE. A high duty cycle (>99%) can be reached at a low pulse repetition rate, due to the storage property of the ion trap. In addition, the storage property of the ion trap provides ion integration of low-intensity (20) Fang, L.; Zhang, R.; Williams, E. R.; Zare, R. N. Anal. Chem. 1994, 66, 3696-701. (21) Banks, J. F.; Dresch, T. Anal. Chem. 1996, 68, 1480-5. (22) Muddiman, D. C.; Rockwood, A. L.; Gao, Q.; Severs, J. C., Udseth, H. R.; Smith, R. D. Anal. Chem. 1995, 67, 4371-5. (23) Mirgorodskaya, O. A.; Shevchenko, A. A.; Chernushevich, I. V.; Dodonov, A. F.; Miroshnikov, A. I. Anal. Chem. 1994, 66, 99-107. (24) Verentchikov, A. N.; Ens, W.; Standing, K. G. Anal. Chem. 1994, 66, 12633. (25) Qian, M. G.; Lubman, D. M. Anal. Chem. 1995, 67, 2870-7.

signals, thus enhancing the sensitivity for detection. The ion trap also has the capability for ejecting unwanted low-mass background ions for selective storage of the target ions and collisional breakup of cluster ions formed with the solvent. Also, the ion trap can operate at an elevated pressure (10-2-10-4 Torr), allowing for easy interfacing to atmospheric pressure ionization sources. In the work reported herein, the capability of the IT/reTOF as a rapid and sensitive detector for CE is demonstrated in the analysis of two protein digests. The analysis of complex peptide mixtures from protein digests is important in the characterization of posttranslational modifications of proteins where the results can provide critical information for a variety of cancer-related biomedical problems.26,27 In real biomedical applications, the sample components are very complex while sample amounts are usually extremely limited. This makes CE/MS potentially an ideal analytical tool in solving these problems. With the IT/reTOFMS, a fast detection speed can be achieved with a detection limit in the low-femtomole range while the high mass resolution achieved can provide mass identification for even closely related fragments in such protein digests. EXPERIMENTAL SECTION Mass Spectrometer. An ion trap storage/reflectron time-offlight mass spectrometer similar to that described in previous work was used in these experiments.28,29 This device consists of a reflectron time-of-flight mass analyzer (Model D1450) interfaced to a quadrupole ion trap (Model C-1251, R. M. Jordan Co., Grass Valley, CA). The ion trap chamber is pumped by a 450 L/min turbomolecular pump (Model V-450, Varian Vacuum Products., Lexington, MA) while the reflectron is pumped by a 250 L/min turbo pump (Model V-250) where the pressures are maintained at 1 × 10-5 and 2 × 10-7 Torr for the ion trap chamber and the time-of-flight tube, respectively. Ions generated from a microelectrospray source were introduced into an atmospheric pressure interface through a heated ss capillary (0.5 mm i.d., 140 °C). The interface was evacuated to a pressure of 0.5 Torr by two 700 L/min mechanical pumps (Model SD-700, Varian Vacuum Products). The ions traversing the interface were focused by a coaxial cylindrical lens (+100 V) and subsequently passed through a skimmer orifice (325 µm) into the high-vacuum chamber. The ions were then focused through an Einzel lens into the ion trap. The ions were stored in the ion trap under a preset rf voltage of 1250 V (Vpp) on the ring electrode (at a frequency of 1.1 MHz) for a period that varied from 125 to 1000 ms before being ejected by a dc pulse on the end cap of the ion trap into the time-of-flight device for analysis. The ions were then mass separated by the reflectron TOF device and detected by a 25 mm triple microchannel plate detector (Model C-2501, R. M. Jordan Co.). Capillary Electrophoresis. The capillary electrophoresis apparatus used in these studies was constructed in-house. The CE columns used in the experiments were prepared from 41 µm i.d. × 105 µm o.d. fused-silica capillaries (Polymicro Technologies Inc., Phoenix, AZ), whose inner walls were coated with (3aminopropyl)trimethoxysilane according to the method first (26) Patterson, S. D.; Aebersold, R. Electrophoresis 1995, 16, 1791-814. (27) Wahl, J. H.; Gale, D. C.; Smith, R. D. J. Chromatogr. 1994, 659, 217-22. (28) Michael, S. M.; Chien, B. M.; Lubman, D. M. Anal. Chem. 1993, 65, 261420. (29) Qian, M. G.; Lubman, D. M. Anal. Chem. 1995, 67, 234A-42A.

Analytical Chemistry, Vol. 68, No. 19, October 1, 1996

3389

described by Lukcas30 and then modified by Moseley et al.31 This coating generated positive charges on the inner surface of the capillary wall in acidic buffers and thus reduced the adsorption of peptides or proteins on the inner capillary surface. As a result, the direction of the electroosmotic flow was reversed. The length of the capillaries used in these experiments varied from 50 to 70 cm. A high-voltage power supply (Model CZE 1000R, Spellman High Voltage Electronics Corp., Plainview, NY) operating in the negative mode was used to provide the separation voltage for CE. The separation voltage used in the experiments was set to produce an electric field of -300 V/cm. Since nonvolatile buffers cause signal suppression and the formation of adduct ions during the electrospray process, especially for a microelectrospray source where no make-up liquid is added, only volatile buffers such as acetic acid or ammonium acetate were used as the running buffer, with a concentration of 10 mM and a pH ranging from 3 to 4.9. Samples of concentrations varying from 8 × 10-6 to 2 × 10-5 M were injected electrokinetically. For the amount of analyte i injected, Qi is calculated by the equation32

Qi ) [πr2(µapp)iVinjtinj/L][Ci] where r is the inner radius of the capillary, (µapp)i is the apparent mobility of analyte i, Vinj, and tinj are the injection voltage and injection time, respectively. L is the length of the capillary, and [Ci] is the concentration of analyte i. UV detection was performed on a variable-wavelength UV detector (Model SC100, Thermo Separation Products, Fremont, CA) at a wavelength of 198 nm with a sampling frequency of 7 Hz. Sheathless Microelectrospray CE/MS Interface. One end of the CE capillary was placed in a buffer vial while the other end was placed directly into a 2 cm long stainless steel (ss) needle (125 µm i.d., 250 µm o.d.), which was maintained at the electrospray voltage (+3 kV). The capillary protruded from the ss needle by ∼0.5 mm. The tip of the capillary was first burnt to remove the polymer coating and then etched by concentrated hydrofluoric acid for ∼10 min. After being washed by water and dried in air, the outer surface of the capillary tip was coated with silver using electroless plating, in which a chemical reducing agent instead of electricity is used to carry out the plating process. This method is a common technique for depositing metal onto nonconductive surfaces.33 Briefly, the etched capillary tip was first degreased in dichloromethane for ∼30 min and then placed in a sensitizer solution containing 0.1 M stannous chloride and 1 M hydrochloric acid for ∼2 min. It was then placed into an electroless plating solution which was prepared by adding 4% formaldehyde solution to a solution of 0.1 N silver nitrate and ammonia. The plating process required 20-30 min for completion. During the coating process, a small nitrogen pressure was maintained across the capillary in order to prevent coating solutions from entering the capillary. Even though the lifetime of this coating during the electrospray process was relatively short (several hours), it was very easy to regenerate by repeating the electroless plating (30) Lukacs, K. D. Diss. Abstr. Int. 1983, 44, 3766-7. (31) Moseley, M. A.; Jorgenson, J. W.; Shabanowitz, J.; Hunt, D. F.; Tomer, K. B. J. Am. Soc. Mass Spectrom. 1992, 3, 289-300. (32) Landers, J. P., Ed. Handbook of Capillary Electrophoresis; CRC: Boca Raton, FL, 1994. (33) Mallory, G. O.; Hajdu, J. B. Electroless Plating: Fundamentals and applications; American Electroplaters and Surface Finishers Society: Orlando, FL, 1990.

3390

Analytical Chemistry, Vol. 68, No. 19, October 1, 1996

procedure. This silver coating provided electric contact at the tip of the capillary and turned the capillary tip into a microelectrospray needle. The flow rate for this microelectrospray source was approximately between 0.3 and 0.5 µL/min. For a 50 cm long capillary, the CE power supply was operated at -12 kV, and the voltage ground was floated at + 3 kV on the microelectrospray needle, resulting in an effective voltage of -15 kV for CE separations. Data Acquisition. The data system for the CE/MS experiments has been described in detail elsewhere.34 Briefly, a 250 MHz high-speed transient recorder (Model 9846, Precision Instruments Inc., Knoxville, TN) was used to collect the signals from the TOF microchannel plate detector. This transient recorder was embedded in a Pentium-66 MHz PC compatible computer (Model P5-66, Gateway 2000, North Sioux City, SD), and data processing was also performed on this computer using a user-written program. In this program, data reduction was accomplished by converting averaged time-of-flight transients into indexed flighttime/intensity pairs. The method allows the storage of a full set of data in the RAM before being saved onto a hard disk at the end of every CE/MS run, thus eliminating the restraints of the disk storage speed. Using this program, a spectral acquisition speed of up to 25 Hz could be achieved for recording a CE separation of up to 30 min. Total ion electropherograms were generated by integrating each mass spectrum within a given time or mass frame. Selected ion electropherograms were generated by reprocessing the stored mass spectra during the separation run. In the display of the 3D map, the ion intensity values are scaled down to a 16-color spectrum, with each color corresponding to a certain range of values. A sampling time window width of 150 µs, which corresponds to a mass range from 0 to ∼1500 Da was used for all studies in this work. Mass calibration was carried out by measuring the flight time (T) of a few model peptides to find out the constants a and b in the empirical equation by linear regression analysis, (m/z)1/2 ) aT + b. Materials. All peptide and protein samples, ammonium acetate, and acetic acid were obtained from Sigma Chemical Co. (St. Louis, MO) and used without further purification. Trypsin was purchased from Promega (Madison, WI). Water used to prepare CE separation buffers was generated with a Milli-Q water purification system (Millipore Corp., Bedford, MA). For tryptic cleavages, a 100 µg aliquot of protein was incubated for 18-21 h at 37 °C with a protein-to-enzyme ratio of 50:1 (w/w) in 50 mM NH4HCO3 solution at pH 8.2. The digested materials were then spin-dried under vacuum to remove the salt and reconstituted in the CE running buffer to a concentration of 2 × 10-5 M original protein. RESULTS AND DISCUSSION CE/MS combines the high separation efficiency of CE with the identification capability of MS and thus can provide multidimensional information such as electrophoretic migration time, mass detection, and fragmentation patterns. For CE/MS experiments, detection speed is critical for providing accurate and reproducible peak shapes and mass information. The detection speed of conventional scanning mass spectrometers may result in distorted peak shapes or even missing peaks from the CE separations. IT/reTOFMS provides a method for accurately (34) Qian, M. G.; Wu, J.-T.; Parus, S.; Lubman, D. M. Rapid Commun. Mass Spectrom., in press.

Figure 1. Total ion electropherograms of a five-peptide mixture using CE/IT/reTOFMS. Spectral acquisition speeds in the electropherograms: (a) 2, (b) 1, (c) 0.5, and (d) 0.25 s/spectrum. Peaks: (1) neurotensin, (2) bradykinin, (3) angiotensin III, (4) angiotensin II, and (5) angiotensin I. Separation conditions: 70 cm long coated capillary, 10 mM acetic acid buffer (pH ∼3), and separation voltage -21 kV.

detecting the CE separation since it provides virtually 100% duty cycle; i.e., all the ions within a given mass range are stored, and the detection uses a pulsed dc ejection mode so that it is a nonscanning device. Our efforts have been to demonstrate the

Figure 2. UV trace (a) and total ion electropherograms (b-d) obtained at different spectral acquisition speeds of the separation of cytochrome c digest. Spectral acquisition speeds: (b) 1, (c) 0.5, and (d) 0.25 s/spectrum. Separation conditions: 50 cm long coated capillary, 10 mM ammonium acetate buffer (pH ∼4.4), and separation voltage -15 kV.

capabilities of the instrument as a rapid and sensitive detector for CE. Figure 1 shows the total ion electropherograms of a five-peptide mixture containing angiotensins I, II, and III, bradykinin, and neurotensin, obtained at different spectral acquisition speeds. Analytical Chemistry, Vol. 68, No. 19, October 1, 1996

3391

Table 1. Effect of Detection Speed on CE Separation detection speed (spectrum/s)

no. of data points to define the peak

migration time tm (min)

peak width at half-maximum W1/2 (s)

no. of theoretical platesa

S/N ratiob

0.5 1 2 4

5 8 11 17

4.84 4.80 4.76 4.82

6.0 4.9 3.1 2.8

1.3 × 104 1.9 × 104 4.7 × 104 5.9 × 104

12.9 14.2 16.3 17.2

a No. of theoretical plates was calculated by N ) 5.54(t /W 2 b m 1/2) . S/N was calculated by S/N ) X/(SD), where X is the peak height and SD is the standard deviation of the background noise.

These peptides were purposely chosen since they were closely related and difficult to separate, especially at a low detection speed. Acetic acid (10 mM; pH ∼3) was used as the running buffer with a sample concentration of ∼8 × 10-6 M for each peptide. The length of the capillary used in this experiment was ∼70 cm. Samples were injected electrokinetically at 1000 V for 7 s, corresponding to a sample amount of ∼8 fmol for each peptide. Panels a-d of Figure 1 were recorded using a spectral acquisition speed of 2, 1, 0.5, and 0.25 s/spectrum, respectively. The migration times of the five peptides in these four electropherograms were highly reproducible, although the separation efficiency differed. At a relatively low spectral acquisition speed of 2 s/spectrum (Figure 1a), the peaks of angiotensin I and II are only partially resolved, while the peaks of angiotensin III and bradykinin are not separated to the base line. When the spectral acquisition speed is increased to 1 s/spectrum (Figure 1b), the peaks of angiotensin III and bradykinin are separated to the base line level while the peaks of angiotensin I and II are still not completely resolved. At a spectral acquisition speed of 0.5 s/spectrum (Figure 1c), however, all the peaks are resolved. A spectral acquisition speed of 0.25 s/spectrum further improved the separation (Figure 1d). Table 1 lists the numbers of theoretical plates and signal-tonoise (S/N) ratios at different spectral acquisition speeds for the peak of bradykinin. As shown in Table 1, the number of theoretical plates and S/N ratios increased as the spectral acquisition speed increased. This can be explained by referring to the number of data points collected to define the peak as also shown in Table 1. There are only five data points collected to define the peak at 2 s/spectrum, which inevitably results in the failure to record the maximum of the actual peak, and thus the real shape of the peak is not maintained, resulting in a broader peak and lower signal-to-noise ratio. Nevertheless, when the spectral acquisition speed reached 0.25 s/spectra, 17 data points were collected for this peak, which is sufficient to define the true peak shape in this experiment. A further increase of the acquisition speed to 10 Hz did not improve the peak shape and separation efficiency. Indeed, sensitivity was sacrificed slightly due to the relatively short trapping time at such high sampling speed. Therefore, a spectral acquisition speed of 4 spectra/s was determined to be optimal for maintaining the quality of our current CE separation while achieving high sensitivity. In order to test the capabilities of this CE/IT/reTOFMS for analyzing complex peptide mixtures such as protein digests, a tryptic digest of bovine cytochrome c was analyzed. Figure 2 shows the UV trace (Figure 2a) and the total ion electropherograms (Figure 2b-d) of the separation of the digest at different mass spectral acquisition speeds. The UV and total ion electropherograms were obtained separately using two capillaries under 3392 Analytical Chemistry, Vol. 68, No. 19, October 1, 1996

Figure 3. Selected ion electropherograms of the unresolved peak marked 1 in Figure 2d.

the same separation conditions, except that the length from the inlet end of the capillary to the UV detector window was ∼7 cm shorter than that to the electrospray needle (50 cm), resulting in a slightly shorter migration time on the UV trace. Ammonium acetate with a concentration of 10 mM and a pH of 4.4 was used as the separation buffer. Injection was made at 1000 V for 4 s,

corresponding to ∼12 fmol of the original protein. As shown in Figure 2, the UV trace and the total ion electropherograms are qualitatively similar, except for peak 2 as marked in the UV trace, which is missing in the total ion electropherograms. This peak may result from a relatively large fragment due to the limited cleavage efficiency in the digestion process. The current m/z sampling range in these experiments is from 0 to ∼1500. Thus, fragments with m/z larger than 1500 were not detected, unless they could form multiply-charged ions. Since the CE separation was performed in a mildly acidic solution (pH 4.4), some large fragments would not be expected to form multiply-charged ions and thus would not be detected. Differences in the peak heights between the UV absorption and total ion electropherograms were mainly due to differences between the UV absorption efficiency and ionization efficiency for various peptides. Nevertheless, the separation efficiency recorded in the total ion electropherograms increases when the spectral acquisition speed is increased. One unique feature of coupling MS with CE experiments is that the mass detector can identify unresolved or partially resolved peaks, as illustrated by peak 1 marked in both the UV trace and the total ion electropherograms of cytochrome c digest shown in Figure 2. The mass spectrum reveals that the peak actually contains three partially resolved peaks, and they could be identified without ambiguity. Figure 3 shows the selected ion electropherograms for these three components corresponding to the separation recorded in Figure 2d. As seen in the selected ion electropherograms, these three components are clearly of different identities, even though their migration times were very close to each other. The mass spectra for these three components are shown in Figure 4. These mass spectra consist of only a single pulse-out with a trapping storage time of only 0.25 s. The result is that a relatively high intensity low-mass background can be observed in these spectra, as compared to spectra obtained over longer storage times and signal averaged over multiple pulses.28 Even though no mathematical smoothing or base line offset has been used, there is still a sufficient S/N ratio for mass identification. The mass resolution of the IT/reTOF is typically between 2500 and 3000. In these experiments, since unit resolution is sufficient to identify all the fragments, a time resolution of 10 ns was used instead of the maximum time resolution (4 ns) of the transient recorder resulting in an actual recorded mass resolution of ∼1500 in Figure 4. This was performed to reduce the size of the data file and to speed up the data transfer and storage process. Table 2 lists the calculated and measured molecular weights of the tryptic fragments of bovine cytochrome c from CE/MS analysis. The measured and calculated values are in good agreement. Because of the ion storage capability of the ion trap, sample ions can be accumulated in the ion trap for a period of time before they are ejected into the time-of-flight tube. This process results in improved sensitivity, especially when it is coupled with a continuous ionization source, such as electrospray. Figure 5 shows the total ion electropherograms for cytochrome c tryptic digest obtained with sample injections corresponding to about 50, 12, and 3 fmol of the original protein. A detection speed of 4 Hz was used in these experiments. At 12 fmol sample injection, the total ion electropherogram is still reproducible, and all the peaks that appear in the 50 fmol electropherogram can be observed with high S/N ratios. At 3 fmol injection, though, only four of the major peaks can be observed, and two of them with low S/N ratios.

Figure 4. Mass spectra corresponding to the selected ion electropherograms shown in Figure 3. All the mass spectra consist of a single pulse-out with a trapping storage time of 0.25 s. (a) m/z 849.5, doublycharged ion of fragment TGQAPGFSYTDANKNK, (b) m/z 729.1, doubly-charged ion of fragment TGQAPGFSYTDANK, and (c) m/z 792.7, doubly-charged ion of fragment KTGQAPGFSYTDANK.

However, under the selected ion electropherogram mode, these two peaks can still be identified with a high S/N ratio as shown in Figure 6. It is worthwhile to note that all these results were obtained with a 41 µm i.d. capillary, which is in the range of typical capillary dimensions for conventional CE experiments. Smith et al. reported attomole level detection of proteins using a 5 µm i.d. capillary for the CE/MS experiment and found that there was a sensitivity boost of 25-50 fold compared with the results obtained with 50 µm i.d. capillaries.35 Currently the detection limit of our CE/MS experiments is still not as low as that reported by Smith’s (35) Wahl, J. H.; Goodlett, D. R.; Udseth, H. R.; Smith, R. D. Electrophoresis 1993, 14, 448-57.

Analytical Chemistry, Vol. 68, No. 19, October 1, 1996

3393

Table 2. Comparison of Calculated and Measured Tryptic Fragments of Bovine Cytochrome c from CE/MS Analysis no.

fragment

calcd mass

measd massa

sequence

3, 4 8 8, 9 9, 10 10 10, 11 12 13, 14 14 15 15, 16 19 20, 21

8-13 28-38 28-39 39-53 40-53 40-55 56-72 73-79 74-79 80-86 80-87 92-99 100-104

762.5 1168.6 1296.7 1584.8 1456.7 1698.8 2010.0 806.5 678.4 779.5 907.5 964.5 562.3

762.9 1169.5 1296.3 1584.4 1457.1 1698.0 2009.2 806.2 678.8 779.7 906.9 964.8 562.5

KIFVQK TGPNLHGLFGR TGPNLHGLFGRK KTGQAPGFSYTDANK TGQAPGFSYTDANK TGQAPGFSYTDANKNK GITWGEETLMEYLENPK KYIPGTK YIPGTK MIFAGIK MIFAGIKK EDLIAYLK KATNE

a

Average mass of all charge states of the fragment observed.

group with the 5 µm i.d. capillary, but the potential for this instrument to reach a comparable detection limit is realistic if the dimensions of the capillaries used in our experiment could be reduced sufficiently. Also, if the ion transmission efficiency of our current instrument could be improved by using an rf-only multipole ion guide and a two-stage interface, a significant improvement in sensitivity might be expected. For CE/MS experiments using conventional scanning mass spectrometers, mass resolution is usually sacrificed for sensitivity. For quadrupole mass analyzers, for example, in the work reported by Smith’s group, a 2 mass unit window was used to obtain sufficient sensitivity.27 In the case of the IT/reTOF, mass resolution is independent of sensitivity since it is not a scanning device. Sensitivity and resolution can be optimized independently. At its maximum sensitivity, a mass resolution sufficient to discriminate 1 mass unit or even less is routinely achievable. Figure 7 shows the UV trace and total ion electropherogram of bovine β-lactoglobulin A digest. The separation was performed in ammonium acetate buffer at pH 4.9 with a spectral acquisition speed of 4 spectra/s. About 20 fmol of sample was injected for the UV and total ion electropherogram shown in Figure 7. Lactoglobulin A is a relatively large protein with several fragments of similar masses. One example are fragments 2 and 9 (see Table 3), which have calculated MH+ masses of 673.4 and 674.4, respectively. In the total ion electropherogram, these two fragments correspond to the peaks marked as 1 and 2. The high resolution of the mass detector can discriminate these two components unambiguously. It should also be noted that peak 1 contains three unresolved components, which were identified as fragments KIIAEK (m/z ) 701.6), IPAVFK (m/z ) 674.6), and VLVLDTDYKK (m/z ) 1194.1) (spectra not shown). Table 3 lists the calculated and measured tryptic fragments of bovine β-lactoglobulin A. One important observation in comparing the UV trace and the total ion electropherograms is that the separation efficiency with mass detection is slightly worse than that with UV detection. This was not caused by the detection speed, because even if the detection speed was further increased to 10 Hz (data not shown), the separation efficiency in the total ion electropherograms was still lower than that in the UV trace (the UV sampling frequency is 7 Hz). In addition, since we have used a microelectrospray 3394

Analytical Chemistry, Vol. 68, No. 19, October 1, 1996

Figure 5. Total ion electropherograms of cytochrome c digest with sample injection amount of (a) ∼50, (b) ∼12, and (c) ∼3 fmol. Separation conditions were the same as in Figure 2. Mass detection speed was 4 spectra/s.

source with sheathless coupling, there are no dead volumes or mixing effects which might otherwise affect the separation efficiency. One plausible explanation for the reduced separation efficiency observed in the total ion electropherograms is the electrostatic pressure caused by the strong electrical field between the microelectrospray needle and the grounded counter electrode. In fact, this electrostatic pressure, PE defined as10

PE ) 1/20E 2

(where 0 is the dielectric constant and E is the electrical field), was the driving force for a microelectrospray source reported by Wilm and Mann,10 in which the use of solvent delivery pumps

Figure 6. Selected ion electropherograms of peaks 1 and 2 as marked in Figure 5c: (a) 4.45 min at m/z 762.9 (singly-charged ion of fragment KIFVQK), (b) 4.82 min at m/z 678.8 (singly-charged ion of fragment YIPGTK).

was eliminated. During the CE/MS experiments, the microelectrospray needle, which was the outlet end of the CE separation capillary, was kept at 3 kV relative to the grounded ss capillary that was placed at ∼5 mm away from the needle. Thus, an electrostatic pressure caused by this electric field was applied at the CE capillary outlet and had an adverse effect on the flat profile of the electroosmotic flow, which is a major factor for high separation efficiency in CE. According to this explanation, the effect of the electrostatic pressure on the CE separation efficiency should be more prominent at smaller electroosmotic flow rates, where the flow caused by the electrostatic pressure contributes more in the total flow rate. This was confirmed by the above CE/ MS experiments on cytochrome c and β-lactoglobulin A digests as shown in Figure 2 and Figure 7. The separation of β-lactoglobulin A digest was performed in a less acidic buffer (pH 4.9) than that used in the separation of cytochrome c digest (pH 4.4), resulting in a smaller electroosmotic flow in coated capillaries. As shown in Figure 7, the difference in separation efficiency between the UV and the total ion electropherograms is more apparent than that of the cytochrome c digest (Figure 2). Another observation that also supports the above explanation is the difference in sample migration time between the UV and the total ion electropherograms. Compared with the results of cytochrome c digest in Figure 2, the difference between the migration times in the UV trace and in the total ion electropherogram for lactoglobulin A digest (Figure 7) is smaller. Since there is no electrostatic pressure with UV detection, a smaller difference

Figure 7. UV trace (a) and total ion electropherogram (b) of lactoglobulin A digest. Separation conditions: 50 cm long coated capillary, 10 mM ammonium acetate buffer (pH ∼4.9), and separation voltage -15 kV. Mass detection speed was 4 spectra/s. Table 3. Comparison of Calculated and Measured Tryptic Fragments of Bovine β-Lactoglobulin A no.

fragment

calcd mass

measd massa

sequence

1 1, 2 2 6, 7 7 8, 9 9 10 10, 11 11 11, 12 14, 15 15, 16 17

1-8 1-14 9-14 70-75 71-75 76-83 78-83 84-91 84-100 92-100 92-101 125-138 136-141 142-148

933.5 1587.9 673.4 701.5 573.4 903.6 674.4 916.5 1963.0 1065.6 1193.7 1635.8 721.4 837.5

933.8 1588.7 673.6 701.6 573.3 903.5 674.6 916.4 1962.5 1066.2 1194.1 1636.1 721.1 837.9

LIVTQTMK LIVTQTMKGLDIQK GLDIQK KIIAEK IIAEK TKIPAVFK IPAVFK LDAINENK LDAINENKVLVLDTDYK VLVLDTDYK VLVLDTDYKK TPEVDDEALEKFDK FDKALK ALPMHIR

a

Average mass of all charge states of the fragment observed.

between the migration times indicates that the flow caused by the electrostatic pressure contributes more in the total flow rate, resulting in a fast elution speed in the total ion electropherogram. CONCLUSIONS The IT/reTOF has proved to be a rapid and sensitive mass detector capable of maintaining the high separation efficiency of CE. The IT/reTOF can maintain the separation efficiency by achieving a high mass sampling speed without the loss of mass Analytical Chemistry, Vol. 68, No. 19, October 1, 1996

3395

accuracy and resolution. In comparison, scanning mass spectrometers such as quadrupoles or ion traps generally cannot achieve a sufficiently fast scan speed over the full mass range of a spectrum to maintain the resolution of CE separations. Also compared to the orthogonal extraction TOF device, the IT/reTOF can reach a high duty cycle without the use of ion counting or a specialized high repetition rate data collection system. In this work, total ion electropherograms of tryptic digests of cytochrome c and β-lactoglobulin A were obtained in the low-femtomole range without the miniaturization of the CE capillary. Selected ion electropherograms were generated by reprocessing the data using the software developed in our laboratory, resulting in an enhanced S/N ratio. Further, high mass resolution is not sacrificed for sensitivity in the CE/IT/reTOF configuration, which makes it

3396

Analytical Chemistry, Vol. 68, No. 19, October 1, 1996

capable of identifying unresolved peaks and discriminating between fragments with a mass difference of 1 Da or less. ACKNOWLEDGMENT We gratefully acknowledge support of this work by the National Institutes of Health under Grant 1R01GM49500 and the National Science Foundation under Grants BIR-9223677 and BIR9513878. Received for review April 23, 1996. 1996.X

Accepted July 8,

AC960405V X

Abstract published in Advance ACS Abstracts, August 15, 1996.