Fast Multiple Electron Capture Dissociation in a Linear Radio

ADVERTISEMENT .... Fast Multiple Electron Capture Dissociation in a Linear Radio Frequency Quadrupole Ion Trap ... We developed an ECD device using a ...
0 downloads 0 Views 294KB Size
Download PDF
Anal. Chem. 2007, 79, 8755-8761

Fast Multiple Electron Capture Dissociation in a Linear Radio Frequency Quadrupole Ion Trap Hiroyuki Satake, Hideki Hasegawa, Atsumu Hirabayashi, Yuichiro Hashimoto, and Takashi Baba*

Biosystem Research Department, Life Science Research Laboratory in Central Research Laboratory, Hitachi Ltd., 1-280, Higashi-Koigakubo, Kokubunji, Tokyo 185-8601, Japan Katsuyoshi Masuda

Suntory Institute for Bioorganic Research, 1-1-1Wakayamadai Shimamotocho Mishimagun Osaka 618-8503, Japan

We developed a fast electron capture dissociation (ECD) device using a linear radio frequency-quadrupole (RFQ) ion trap. The device dissociated peptides and proteins using a focused electron beam with an intensity of 0.5 µA and a diameter of 1 mm. The electron capture rate was 13%/ms for doubly charged peptides, and the total amount of ECD products was identical to the theoretical limit, i.e., 50% of incident precursor ions were observed as maximum ECD products by electron irradiation of 7 ms in a pulse counting detection scheme. Coupling this ECD device to a time-of-flight mass spectrometer, we applied multiple ECD. Protonated ubiquitin precursor ions with a charge state of 10 were repeatedly cleaved by ECD, i.e., charge-reduced species and their highly charged fragments were cleaved again and again, creating lower charged products, leaving only singly to triply charged states among the final products. Meanwhile with the amount of electron irradiated, lower charged products increased. Applying an electron beam for 8 ms, we obtained 96% of the total sequence coverage using a 40 fmol sample except at three proline sites. This fast ECD device should be widely applicable to proteomics including post-translational modification analysis and top-down analysis. Electron induced dissociations, electron capture dissociation (ECD),1,2 and electron-transfer dissociation (ETD)3-6 are unique sequence analyzing techniques for proteomics. They are powerful tools for obtaining wide coverage of backbone cleavage of * Corresponding author. E-mail: [email protected]. (1) Zubarev, R. A.; Kelleher, N. L.; McLafferty, F. W. J. Am. Chem. Soc. 1998, 120, 3265-3266. (2) Zubarev, R. A. Mass Spectrom. Rev. 2003, 22, 57-77. (3) Syka, J. E. P.; Coon, J. J.; Schroeder, M. J.; Shabanowitz, J.; Hunt, D. F. Proc. Natl. Acad. Sci. U.S.A. 2004, 101, 9528-9533. (4) Coon, J. J.; Ueberheide, B.; Syka, J. E. P.; Dryhust, D. D.; Ausio, J.; Shabanowitz, J.; Hunt, D. F. Proc. Natl. Acad. Sci. U.S.A. 2005, 102, 9463. (5) Coon, J. J.; Ueberheide, B.; Syka, J. E. P.; Geer, L. Y.; Bai, D. L.; Shabanowitz, J.; Hund, D. Proceedings of the 53rd ASMS Conference on Mass Spectrometry and Allied Topics, San Antonio, TX, June 5-9, 2005, A051532. (6) Hogan, J. M.; Pitteri, S. J.; Chrisman, P. A.; McLuckey, S. A. J. Proteome Res. 2005, 4, 628-632. 10.1021/ac071462z CCC: $37.00 Published on Web 09/29/2007

© 2007 American Chemical Society

protonated peptides and proteins,7 doing top-down analysis of long peptides and proteins,1,8-15 and analyzing post-translational modification sites6,16-19 and mono- and disulfide bond cleavage.20,21 Leucine/isoleucine identification by hot ECD has also been demonstrated.22-25 Before 2004, ECD was carried out by Fourier transform ion cyclotron resonance (FTICR) instruments because the low-energy electrons required for ECD had been obtained in such static electromagnetic fields for ion confinement.26 Since 2004, four methods of obtaining low-energy electrons in radio frequency (rf) ion traps have been reported: electron transfer from (7) Wu, S.-L.; Kim, J.; Hancock, W. S.; Karger, B. J. Proteome Res. 2005, 4, 1155-1170. (8) Jebanathirajah, J. A.; Pittman, J. L.; Thomson, B. A.; Budnik, B. A.; Kaur, P.; Rape, M.; Kirschner, M.; Costello, C. E.; O’Connor, P. B. J. Am. Soc. Mass Spectrom. 2005, 16, 1985-1999. (9) Zubarev, R. A.; Fridriksson, E. K.; Horn, D. M.; Kelleher, N. L.; Kruger, N. A.; Carpenter, B. K.; McLafferty, F. W. Anal. Chem. 2000, 72, 563-573. (10) Horn, D. M.; Zubarev, R. A.; McLafferty, F. W. Proc. Natl. Acad. Sci. U.S.A. 2000, 97, 10313-10317. (11) Sze, S. K.; Ge, Y.; Oh, HanBin; McLafferty, F. W. Proc. Natl. Acad. Sci. 2002, 99, 1774-1779. (12) Ge, Y.; ElNaggar, M.; Szea, S. K.; Oha, H. B.; Begley, T. P.; McLafferty, F. W.; Boshoffb, H.; Barry, E. H., III. J. Am. Soc. Mass Spectrom. 2003, 14, 253-261. (13) Patrie, S. M.; Charlebois, J. P.; Whipple, D.; Kelleher, N. L.; Hendrickson, C. L.; Quinn, J. P.; Marshall, A. G.; Mukhopadhyay, B. J. Am. Soc. Mass Spectrom. 2004, 15, 1099-1108. (14) Ge, Y.; Lawhorn, B. G.; ElNaggar, M.; Strauss, E.; Park, J.-H.; Begley, T. P.; McLafferty, F. W. J. Am. Chem. Soc. 2002, 124, 672-678. (15) Han, X.; Jin, M.; Breuker, K.; McLafferty, F.W. Science 2006, 314, 109. (16) Shi, S. D.-H.; Hemling, M. E.; Carr, S. A.; Horn, D. M.; Lindh, I.; McLafferty, F. W. Anal. Chem. 2001, 73, 1-22. (17) Håkansson, K.; Cooper, H. J.; Emmett, M. R.; Costello, C. E.; Marshall, A. G.; Nilsson, C. L. Anal. Chem. 2001, 73, 4530-4536. (18) Peterman, S. M.; Mulholland, J. J. J. Am. Soc. Mass Spectrom. 2006, 17, 168-179. (19) Adamson, J. T.; Håkansson, K. J. Proteome Res. 2006, 5, 493. (20) Zubarev, R. A.; Kruger, N. A.; Fridriksson, E. K.; Lewis, M. A.; Horn, D. M.; Carpenter, B. K.; McLafferty, F. W. J. Am. Chem. Soc. 1999, 121, 28572862. (21) Kleinnijenhuis, A. J.; Duursma, M. C.; Breukink, E.; Heeren, R. M. A.; Heck, A. J. R. Anal. Chem. 2003, 75, 3219-3225. (22) Kjeldsen, F.; Budnik, B. A.; Haselmann, K. F.; Jensen, F.; Zubarev, R. A. Chem. Phys. Lett. 2002, 356, 201-206. (23) Kjeldsen, F.; Sørensen, E.; Zubarev, R. A. Anal. Chem. 2003, 75, 12671274. (24) Cooper, H. J.; Hudgins, R. R.; Håkansson, K.; Marshall, A. G. Int. J. Mass Spectrom. 2003, 228, 723-728. (25) Kjeldsen, F.; Zubarev, R. J. Am. Chem. Soc. 2003, 125, 6628-6629. (26) Zubarev, R. A. Curr. Opin. Biotechnol. 2004, 15, 12-16.

Analytical Chemistry, Vol. 79, No. 22, November 15, 2007 8755

negatively charged reagent ions (i.e., ETD), capturing of focused electrons using a magnetic field,27,28 quasi-static electron injection using a digital ion trapping field,29 and electron transfer from neutral atoms.30 These techniques will offer opportunities to exploit the unique features of ECD revealed by FTICR instruments, and coupling the devices to an rf ion trap and a time-of-flight (TOF) mass spectrometer will offer extra benefits such as highthroughput performance. ECD was demonstrated in rf traps in 2004,27,28 but the techniques are still in development. Reported reaction speeds and signal-to-noise ratios were not good enough for the reported ECD to be applied to high-throughput proteomics. We developed an ECD device using a linear radio frequency quadrupole (RFQ) ion trap with a magnetic field to carry out highspeed spectrum acquisition. In this device, electron injection efficiency was improved by strengthening the magnetic field, and the pinpoint electron source and ion transfer efficiency were improved by separating electron/ion pathways.31-33 Post-translational modification analysis was also reported using this ECD device.34 We present an rf ion trap that produces ECD fast enough to be applied to high-throughput proteomics. Reaction speed and sensitivity were roughly equal to those of the conventional highthroughput dissociation technique, collision induced dissociation (CID), which is faster than ECD in FTICR and ETD. Though very similar to those observed by FTICR, features of spectra observed by our ECD are also slightly different. We also were able to conduct multiple ECD, a sensitive top-down analysis, by coupling the fast ECD device to a TOF mass spectrometer. Instrument. Figure 1 shows the ECD-TOF mass spectrometer used in this study. It is composed of a nanoelectrospray ionization (ESI) source, a linear RFQ ion trap that isolates precursor ions (referred to as a CID trap), the ECD device, and a TOF mass spectrometer. The configuration of the linear ion trap-TOF mass spectrometer was reported previously.35 The ECD device, which includes a quadrupole deflector, was installed between the CID trap and a quadrupole rod set to focus ions (thermalizer in Figure 1). Precursor ions isolated in the CID trap were injected into the ECD trap through the quadrupole deflector, the ion guide, and an end electrode of the linear ion trap (I wall in Figure 1). Product ions were ejected from the ECD trap to the TOF mass spectrometer through the I wall in Figure 1, the ion guide, the quadruple (27) Baba, T.; Hashimoto, Y.; Hasegawa, H.; Hirabayashi, A.; Waki, I. Anal. Chem. 2004, 76, 4263-4266. (28) Silivra, O. A.; Kjeldsen, F.; Ivonin, I. A.; Zubarev, R. A. J. Am. Soc. Mass Spectrom. 2005, 16, 22-27. (29) Ding, L.; Brancia, F. L. Anal. Chem. 2006, 78, 1995-2000. (30) Misharin, A. S.; Silivra, O. A.; Kjeldsen, F.; Zubarev, R. A. Proceedings of the 53rd ASMS Conference on Mass Spectrometry and Allied Topics, San Antonio, TX, June 5-9, 2005, A051572. (31) Satake, H.; Baba, T.; Hasegawa, H.; Hashimoto, Y.; Hirabayashi, A.; Waki, I. Proceedings of the 53rd ASMS Conference on Mass Spectrometry and Allied Topics, San Antonio, TX, June 5-9, 2005, A050564. (32) Baba, T.; Satake, H.; Hasegawa, H.; Hashimoto, Y.; Hirabayashi, A.; Waki, I. Proceedings of the 53rd ASMS Conference on Mass Spectrometry and Allied Topics, San Antonio, TX, June 5-9, 2005, A050571. (33) Baba, T.; Satake, H.; Yokosuka, T.; Hirabayashi, A.; Manri, N. H.; Hasegawa, H.; Hashimoto Yoshinari, K.; Kobayashi, K. Y.; Waki, I.; Deguchi, H. Proceedings of the 54th ASMS Conference on Mass Spectrometry and Allied Topics, Seattle, WA, May 28-June 1, 2006, A061594. (34) Deguchi, K.; Ito, H.; Baba, T.; Hirabayashi, A.; Nakagawa, H.; Fumoto, M.; Hinou, H.; Nishimura, S. Rapid Commun. Mass Spectrom. 2007, 21, 1-8. (35) Hashimoto, Y.; Hasegawa, H.; Waki, I. Rapid Commun. Mass Spectrom. 2005, 19, 1485-1491.

8756 Analytical Chemistry, Vol. 79, No. 22, November 15, 2007

Figure 1. ECD-TOF mass spectrometer.

deflector, and the thermalizer. The infused ESI source mainly used in this work had an infusion rate of 50-200 nL/min. The principle of the injection of low-energy electrons into the ECD device was presented in a previous paper.27 It is done to avoid heating the electrons in the rf field for ion trapping, and low-energy electrons are injected along the central axis of the linear RFQ. A magnetic field parallel to the central axis assists the electron traveling along the central axis. The quadrupole rods of ECD trap L were 50 mm long, and the radius of the trapping volume r0 was 6 mm. The rf voltage applied to the rods was 450 kHz at frequency f and 0-500 V in amplitude Vrf (zero-to-peak). Ions were trapped axially by a dc potential produced by two end electrodes (I and E walls). Helium gas was introduced into the ECD trap to cool precursor and fragment ions. The estimated pressure in the ECD device was 0.2 Pa. The magnetic field was generated by a neodymium cylinder magnet, and its flux density was 150 mT at the center of the trap. An edged thoriated tungsten filament (see Figure 1), whose diameter was 0.1 mm, produced thermal electrons that were extracted efficiently at the edge by the gate electrode and injected into the RFQ rods through the E wall. This shaped filament produces only minor dispersion of the kinetic energy of the extracted thermal electrons because the dc bias at the edge is defined uniquely. The density of the electron beam is high because the diameter of the beam is less than 1 mm, which is determined by the diameter of the hole in the E wall, as shown in Figure 1. The electrons passing through the I wall, are collected by the ion guide electrodes, which are a set of quadrupole rods coupling the ECD trap and the quadrupole deflector. The collected electron current indicates that there is an electron beam current traveling along the central axis that may or may not be slightly deflected by the rf field. The electron kinetic energy was controlled by the dc bias of the RFQ rods of the ECD trap, where the filament bias was fixed at constant voltage of 10 V. The electron current was controlled by the filament current, typically 1 A. Electron irradiation time was controlled by varying how long the gate electrode was open. Using separate inlets for the electrons and ions makes injection of both electrons and ions and ejection of product ions more efficient, speeding ECD. Electron transport from the electron source to the trapped ions was efficient because the electron source could be attached immediately adjacent to the linear RFQ.

Figure 2. (A) Spectrum of doubly charged substance P at maximum fragment intensity given by electron irradiation of 7 ms. (B) Electron energy intensity dependence of precursor ions before/after electron irradiation (indicated by 9/[, respectively), total fragments (s), and charge reduced precursors (- - - -). ECD and HECD bands are clearly observed using our RF-ECD device. Sample was doubly charged substance P.

In the previous configuration,27 in contrast, the electron beams generated by a dispenser cathode traveled through a set of static lens electrodes and a quadrupole deflector. There, electron and ion transmission were degraded because the electrons were charged up on the surface of the electrodes and the electric field along the particle paths was distorted, making the particles uncontrollable. In the TOF-MS, pulse signals produced by the microchannel plate detector (MCP) were selected based on pulse-height threshold and measured time interval between the TOF starting signal. The signal pulse was recorded by a time-to-digital converter. We confirmed that the electron yield of the MCP for a singly charged peptide was sufficient to cause the pulse height discriminator to produce a digital pulse. This means that all peptides were detected at the same level of efficiency irrespective of charge state when the counting rate was low enough to neglect the multiple injections of the same m/z species per TOF-MS operation. This contrasts with the charge dependent induced current measurement used in conventional FTICR measurements. We used three operation modes in measuring ECD efficiency. The ECD, trapping, and through modes were used, respectively, to induce ECD, to monitor the incident precursor ion intensity in the ECD trap, and to monitor the intensity of the ion source. In the ECD and trapping modes, precursor ions isolated by the CIDtrap (without CID) were injected into the ECD trap. The bias of the RFQ rods of the ECD trap was preset to a constant voltage to control the electron energy during ECD. In the ECD mode, the electron gate is opened for a preset duration, while in the trapping mode, the electron gate is kept closed. In the through mode, the quadrupole deflector is set to the straight trajectory from the CID trap to the TOF-MS. ECD efficiency was measured in consecutive combinations of “through”, ”trapped”, and “ECD” modes for each preset value for parameter scan measurements, such as dependence on electron energy, reaction time, and rf amplitude. EXPERIMENTAL SECTION AND RESULTS Fast Electron Capture Dissociation in rf Field. Figure 2A shows typical mass spectra observed when the trap bias was 9.8 V and the filament bias was 10 V. One micromolar water/ acetonitrile solution of substance P (Sigma) was infused. A series of c′ fragments and charge reduced precursor ions [M + 2H]+

were clearly observed. These are similar to spectra observed using ECD in FTICR,2,36,37 using our previous instruments,27 and using a digital ion trap.29 Figure 2B shows electron energy dependences of intensity of precursors and products of substance P that have been doubly charged by electron irradiation when the filament bias was 10 V. The bias of the linear RFQ trap of the ECD trap was scanned in 0.2 V steps, obtaining 100 spectra for each consecutive measurement in through, trapping, and ECD modes. During the scan, other parameters were kept constant, i.e., electron beam intensity was increased by elevating the trap bias. Accumulation for each spectrum was 4.7 s. In the figure, [ represent integrated intensities of the precursor ions without electron irradiation, which was measured in the trapping mode. The precursor ions were not trapped when the bias was set over 24.5 V because the dc potential walls, which were fixed at 25 V during the electron irradiation period, were not effective in axial trapping. In the ECD mode, spectra were observed at 7 ms electron irradiation. Integrated intensity of the remaining precursor [M + 2H]2+ ions after electron irradiation is indicated in Figure 2B by 9. Integrated intensity of charge-reduced product [M + 2H]+ ions and a small amount of [M + H] + ions is indicated by the dashed line. The total intensity of possible fragments observed by ECD, c2′, c4′-c10′, a2, and a4-a10, is indicated by the solid line. Both the electron capturing and the products of electron irradiation showed the same normal (or cool) ECD band at a trap bias of around 10 V and hot ECD bands over 11 V in FTICR instruments.22 It is noted that, unlike the incident precursor ions indicated by [, ECD products were obtained at over 24.5 V because the precursors and the fragment ions were trapped by an attractive potential produced by the intense electron beam, a phenomenon referred to as electron beam ion trapping (EBIT).38 (36) Chan, T. W. D.; Ip, W. H. H. J. Am. Soc. Mass Spectrom. 2002, 13, 13961406. (37) Håkansson, K.; Emmett, M. R.; Hendrickson, C. L.; Marshall, A. G. Anal. Chem. 2001, 73, 3605-3610. (38) Hendrickson, C. L.; Hadjarab, F.; Laude, D. A. Int. J. Mass Spectrom. Ion Processes 1995, 141 161, 0168-1176.

Analytical Chemistry, Vol. 79, No. 22, November 15, 2007

8757

Figure 3. rf amplitude dependence of (A) doubly charged substance P and (B) electron beam current. (C) ECD dependence on electron irradiation time of doubly charged substance P. Thin and thick lines represent theoretical intensity variation given by two consecutive decays at 5 ms for precursor ions and 10 ms for fragment ions, respectively. The dashed line indicates intensity of fragment ions at 20 ms. Maximum fragment intensity was given by 7 ms electron irradiation.

The two features, (1) spectra including c fragments and charge reduced precursors: [M + 2H]+ (Figure 2A) and (2) the electron energy dependence including the cool and hot bands, indicate that FTICR-like ECD occurred in our ECD device. Figure 3A shows ECD features above the rf amplitude applied to the ECD devices during the ECD period. The [, 9, and 2 indicate precursor intensity before ECD, remaining precursor ions, and total ECD products, respectively. Figure 3B shows electron beam intensity monitored by the ion guide. The high voltage boundary of the precursor trapping at 450 V represents the stability boundary of the precursor ions, i.e., m/z ) 674 for doubly charged substance P. Though the rf amplitude of the electron beam transmission decreased because of the diffraction caused by the rf field (Figure 3B), we obtained a good transmission for fast electron capture when we chose an rf amplitude lower than 50 V (indicated by 9 and 2 in Figure 3A). When the rf amplitude was set lower than 20 V, the precursor ions were not trapped in the trapping mode by a shallow ion trap potential. Although the precursor ions and fragment ions were trapped by EBIT during electron irradiation, such settings should be avoided to maintain stable operation. The results shown in Figure 3B suggest that the rf amplitude should be set to 20-50 V or to the rf amplitude normalized based on trap size: Vrf/r02 should be set up to 1.4 V/mm2 when magnetic flux density is 150 mT, which can be produced by a permanent magnet. These settings produce the best electron transmission and stable trapping. Figure 3C shows the remaining precursor ions and the total ECD products over electron irradiation time in the ECD mode. The total ECD products include the expected fragments, charge reduced precursors, and other products observed between c′10 and the charge reduced precursor [M + 2H]+ in Figure 2A. Each intensity is normalized by the trapped precursor intensity obtained in the trapping mode. The thin line is an exponential decay curve with a decay constant of 5 ms (1/e), which was fit to the remaining precursor intensity after ECD. This line indicates electron capture 8758

Analytical Chemistry, Vol. 79, No. 22, November 15, 2007

of doubly charged precursors. The thick line indicates a secondary decay with a constant decay rate, which indicates that electron capture caused fragment loss. These results indicate that the electron capture rate is 13%/ ms for doubly charged peptides and that the maximum fragment intensity is obtained for electron irradiation of 7 ms as a present performance. This reaction speed is comparable to conventional CID in ion trap mass spectrometers. This means the ECD device can be used in high-throughput proteomics. Multiple ECD for TOF Mass Spectrometry. Fast ECD was applied to the +10 charge states of ubiquitin (76 amino acid residues). Figure 4A-D shows a set of ECD spectra observed by changing the electron irradiation time from 1 to 8 ms. Applying the electron beam for 1-2 ms (Figure 4A,B) produced profiles of ECD spectra that were similar to those produced by conventional ECD in FTICR,1,9 i.e., most of the precursor remained, and charge reduced species and highly charged fragments were dominant. Very few fragments could be identified by TOF-MS because their charge state could not be identified by its mass resolution (Figure 4E). With application of the electron beam for 4 ms (Figure 4C), the global profile of fragments was shifted to larger m/z. It still contained highly charged fragments whose charge states could not be analyzed by the TOF-MS. Applying the electron beam for 6-8 ms, however, revealed a profile with one to three charged states (Figures 4D and 5) and, finally, profiles containing singly and doubly charged fragments for 8 ms irradiation of electrons (Figure 4F). Because the electron capture cross section is reported to be proportional to the square of the charge state,1,9 this is consistent with our observations, where the highly charged species vanished quickly and the lower charged ones remained and were dominant. Such fragment charge state profiles can be analyzed by TOF-MS. Intensities of lower charged fragments increased during electron irradiation, e.g., the peak height of the c2 fragment increased, as indicated in Figure 4A-D. Because the increase was still observed after the precursor ions vanished (Figure 4C,D), we thought that the fragments were produced from highly charged

Figure 4. (A-D) Multiple ECD applied to ubiquitin. (E, F) Focused spectra of ubiquitin as revealed by multiple ECD at (E) 2 ms and (F) 8 ms.

Figure 5. Spectra of multiple ECD for electron irradiation of 8 ms and spectrum accumulation of 13 s. Fragment indicators with thin numbers in top law, thick numbers in second law, and thin numbers in third law represent singly charged, doubly charged, and triply charged fragments, respectively.

fragments by multiple ECD as Horn et al. pointed out such mechanism.10 Similar spectra were reported using the sequential ion/ion reaction of ETD and a proton-transfer reaction (PTR) for charge reduction of fragment ions at a reaction time of 150 ms.4 However, the fragments produced by ETD decreased during PTR because PTR causes charge reduction instead of dissociation. This multiple ECD promises high sensitivity and high throughput. Coverage of sequencing by our ECD-TOF-MS was better than coverage reported for ETD. Figure 5 shows a typical multiple ECD spectrum of the +10 charged ubiquitin, which was obtained by electron irradiation for 8 ms and data accumulation of 13 s, where the sample consumption was 40 fmol. The spectrum contains 107 fragments, and 6 cleavage sites out of a total of 75 bonds between amino acids were missed. Three of the sites were on the N terminal side of prolines (indicated by

dark gray triangles in Figure 5) and two of them were c1 and z1 (3), which were out of mass range of the TOF-MS. Only one site, c13 (1), was not identified. Our ECD device coupled to a TOFMS, or unique multiple ECD, promises good coverage of de novo analysis including top-down analysis of small proteins. DISCUSSION Our ECD system has a fast electron capture rate, as shown in Figure 3C. We believe that the speed is due to the dense electron beam provided by the pinpoint electron source and the magnetic field. Because the electron capture rate (or probability of electron capture per unit time) is proportional to the electron capture crosssection and density of the electron beam when the ion is located in the electron beam, the dense beam is preferable for obtaining a high electron capture rate. Since trapped ions in a linear RFQ Analytical Chemistry, Vol. 79, No. 22, November 15, 2007

8759

field (both precursors and fragments) are strongly focused on the central axis by the rf field and by He bath gas cooling, we can use the finely focused electron beam generated by the edged filament. The pinpoint electron source shown in Figure 1 produced an electron beam with a current of 0.5 µA, a diameter of 1 mm, and a density of 1 µA/mm2. Because the ions in FTICR, on the other hand, have a circular orbit around the central axis, a broad electron beam generated by a dispenser cathode with a wide surface (typically 100 mm2) is often used to improve the electron capture rate.39 The typical electron current reported is similar to our electron current (1 µA), which produces an electron beam density of 0.005 µA/mm2. Electron density in our ECD device, which is 200 times stronger than in FTICR, can provide ECD performance faster than FTICR. In Figure 3C, the electron capture ratio of doubly charged to singly charged ions is 2 instead of 4, which is the widely recognized square of the charge state ratio in the point charge approximation for electron capturing at the zero energy limit.1,9 The dashed line in the figure indicates the ratio of 4. A further study showed that the reaction rate is proportional to Z(1.62(0.05) for a charge state Z between 1 and 13 (data not shown) in our ECD devices. We believe the point charge approximation is valid because our ECD was done using an ECD regime with nearly zero energy electrons (Figure 2B). We think the difference between the conventional knowledge and our results can be attributed to the uncertainty about electron density, which is proportionally related to the electron capture rate. The electron density at the center of the trap depended on the amount of the trapped charge. Evidence of the uncertainty appeared in the electron current monitored by the ion guide electrode. The current clearly depended on the amount of trapped ions and was stronger when the ions were trapped. This means that the electron beam is focused by the trapped ion cloud and that, due to the EBIT effect, the ions are in turn focused by the electron beams. The actual electron density in the ion cloud is difficult to measure. Another possible explanation of the difference is spatial distribution of the ions. When the same number of ions is trapped but the charge state of the ions is different, the lower charged ions are dense and the higher charged ions are thin in terms of ion number density. This suggests that lower charged ions overlap more on the finely focused electron beam, meaning that the electron capture rate has a tendency to be higher with lower charged ions than with highly charged ions. Spectra features obtained by our ECD device are similar to those obtained by FTICR,2,36,37 but a crucial difference was observed: there was no c′4 fragment in the spectrum obtained by low-energy electron injection (Figure 2A). This tendency is similar to one observed with ETD, for which no c′4 and only weak c′5 and c′6 were reported.5 When an energetic electron over 2 eV was used, c4′ fragments appeared (data not shown). This suggests that the gas collision in the trap cools down the internal state of the precursor, which has an effect that is the inverse of the activated ion ECD effect.2,19,41,42 Lack of internal energy for dissociation is also indicated by the branching ratio of fragments to charge-reduced species in Figure 2B, where the higher electron energy produced richer fragment ions than the charge-reduced (39) Tsybin Yu, O.; Håkansson, P.; Budnik, B. A.; Haselmann, K. F.; Kjeldsen, F.; Gorshkov, M.; Zubarev, R. A. Rapid Commun. Mass Spectrom. 2001, 15, 1849-1854.

8760

Analytical Chemistry, Vol. 79, No. 22, November 15, 2007

species in the cool ECD band. The energy feature of our ECD device is similar to ECD in FTICR in an ultrahigh vacuum, which is more energetic than ETD.43 The multiple ECD produced internal fragments, whose effect was pointed out by Zubarev et al.9 Figure 5 shows c′3-Met, c′4Met, etc., but their intensities were weak compared to normal c and z fragments. The probability of producing an internal fragment is roughly but not exactly inversely proportional to the square of the number of amino acid residues, as with normal c and z fragment generation. When ubiquitin with 76 residues was subjected to fast ECD, the intensity of the internal fragments was 1/ of that of the normal fragments, and the internal fragments 76 were negligible. This means de novo sequencing can be carried out using multiple ECD. By application of multiple ECD, the number of small m/z fragments was increased by the electron irradiation shown in Figure 4, which gives better sensitivity. However, this approach has the disadvantage: fragments showed a mass shift caused by neutralized protons that are not consumed for ECD. This may be caused by ECD of charge-reduced precursor ions such as [M + nH](n-1)+, [M + nH](n-2)+, and so on. Such fragments may cause confusion for de novo sequencing, but such fragments were observed only for large m/z fragments in our experiments (data not shown). Further studies of the mechanism for nondissociation ECD should be done to improve understanding of electron capturing by polypeptides as well as for de novo sequencing using TOF-MS with mass resolutions from 15 000 to 20 000. RF-ECD has a unique fast performance superior to that of ECD in FTICR and ETD in rf ion traps. This device can be connected both to TOF-MS, as in this study, and to other mass spectrometers, such as ion traps, FTICRs, and orbi-traps. This will provide a platform for scientific research on dissociation nature and a technological application for high-throughput proteomics. CONCLUSION We developed a uniquely fast ECD device using an RFQ ion trap with a magnetic field, whose speed is comparable to the conventional dissociation technique, collision induced dissociation. Coupling this ECD device to a TOF-MS, we demonstrated multiple ECD, where the charge reduced species and the highly charged fragments were cleaved again and again into lower charged products, leaving only products with singly to triply charged states among the final products. Lower charged products increased during electron irradiation. This fast ECD device using an RFQ ion trap coupled to a highthroughput, which allows coupling to an LC, high-mass resolution mass spectrometer should be widely applicable for high-throughput proteomics including post-translational modification analysis and top-down analysis including de novo sequencing. (40) O’Connor, P. B.; Lin, C.; Cournoyer, J. J.; Pittman, J. L.; Belyayev, M.; Budnik, B. A. J. Am. Soc. Mass. Spectrom. 2006, 17, 576-585. (41) Martin, S. E.; Shabanowitz, J.; Hunt, D. F.; Marto, J. A. Anal. Chem. 2000, 72, 4266-4274. (42) Håkansson, K.; Chalmers, M. J.; Quinn, J. P.; McFarland, M. A.; Hendrickson, C. L.; Marshall, A. G. Anal. Chem. 2003, 75, 3256-3262. (43) Anusiewicz, I.; Berdys-Kochanska, J.; Simons, J. J. Phys. Chem. A 2005, 109, 5801-5813.

ACKNOWLEDGMENT We are grateful to Prof. Kisaburo Deguchi of Hokkaido University, to Dr. Izumi Waki of Hitachi High Technologies, and to Dr. Toshiyuki Yokosuka, Dr. Kiyomi Yoshinari, and Dr. Kinya Kobayashi, of the Hitachi Research Laboratory at Hitachi, Ltd., for their valuable discussions. The ECD project was financially

supported by Hitachi High Technologies Co. and the New Energy and Industrial Technology Development Organization (NEDO), Japan. Received for review July 10, 2007. Accepted August 8, 2007. AC071462Z

Analytical Chemistry, Vol. 79, No. 22, November 15, 2007

8761