Collision-Induced Dissociation for Mass Spectrometric Analysis of

Weiqun Li, Christopher L. Hendrickson, Mark R. Emmett, and Alan G. Marshall ... Touradj Solouki, Mark R. Emmett, Shengheng Guan, and Alan G. Marshall...
0 downloads 0 Views 1MB Size
Download PDF
Accelerated Articles Anal. Chem. 1994, 66,4385-4389

Collision-Induced Dissociation for Mass Spectrometric Analysis of Biopolymers: High=ResolutionFourier Transform Ion Cyclotron Resonance MS4 Yulin Huang, Ljiljana Paia=ToliC.,Shenheng Guan, and Alan 0. Marshall*J National High Magnetic Field Laboratory, Florida State University, 1800 East Paul Dirac Drive, Tallahassee, Florida 32306-4005

Efficient collision-induceddissociation multistage tandem high-resolution mass spectrometry of peptide ions is demonstrated for the fist time. SpecXcally, four-stage Fourier transform ion cyclotron resonance collisioninduced dissociationtandem-in-timeMS4 is demonstrated for bradykinin quasimolecular ions, MH+, produced by matrix-assistedlaser desorptiodionhtion. We combine off-resonantexcitationand ion axialbationto improve the efficiency of parent ion dissociation and product ion collection and detection at every MS stage. We observe successive loss of water/ammonia from the C-terminus to leave an (MH - NHdHzO)+ ion in the second stage, followed by successive losses of the next two amino acids, arginine and phenylalanine. High mass resolving power is achieved throughout all four MS stages, in an experiment that consumes 10 pmol of peptide and takes only -5 min. We project that it should be possible to automate this experiment for high-speed sequencing of biopolymers.

-

Matrix-assisted laser desorption ionization (MALDI) invented by h a s , Hillenkamp, and co-workers in 1987, has become a powerful analyticalmethod for mass analysis of biomolecules, such as proteins, oligonucleotides, and synthetic polymers. Singlycharged positive ions (protonated molecules) or negative ions (deprotonated molecules) are created following irradiation with a short W laser pulse of a sample diluted in a solid W-absorbing matrix. Ions generated by pulsed ionization methods, such as MALDI, are conveniently mass-analyzed at low mass-resolving

' Present address: Department of Chemistry, Florida State University, Tallahassee, FL, 32306. (1) Karas, M.; Bahr, U.; Giessman, U. Mass Spectrom. Reu. 1991,10,335357. 0003-2700/94/0366-4385$04.50/0 0 1994 American Chemical Society

power (typically a few thousand or less) by a timeof-flight mass spectrometer. Interface of a MALDI source with magnetic sector and quadrupole ion trap mass spectrometers2 has also been demonstrated. Coupling a MALDI source with a Fourier transform ion cyclotron resonance mass spectrometer F-ICRMS or FTMS) is potentially highly advantageous, since FT-ICRMS is an inherently pulsed mass analyzer that provides ultrahigh massresolving power, precise mass measurement, multistage MS", simultaneous detection of a wide mass range of ions, ion remeasurement, and two-dimensional mass spectrometry in a single Ultrahigh FT-ICR mass-resolving power for laserdesorbed ions MALDI FT-ICR experiwas demonstrated about 10 years ments were conducted by Hettich and Buchananz0Sz1with a conventional LDI instrument configuration, and high-resolution mass spectra of peptides were obtained by introduction of fructose as co-matrix by Castoro, Wilkins, and c o - ~ o r k e r s Recently, .~~~~~ (2) Chambers, D. M.; Goeringer, D. E.; McLuckey, S. A; Glish, G. L. Anal. Chem. 1993,65, 14-20. (3) Marshall, A G. Adu. Mass Spectrom. 1989,1L4,651-668. (4) Wanczek, K-P. Int. J. Mass Spectrom. Ion Processes 1989,95, 1-38. (5) Wilkins, C. L.; Chowdhury, A K; Nuwaysir, L. M.; Coates, M. L. Mass Spectrom. Rev. 1989,8, 67-92. (6) Gord, J. R; Freiser, B. S. Anal. Chim. Acta 1989,225, 11-24. (7) Laude, D. A, Jr.; Hogan, J. D. Tech. Mess. 1990,57, 155-159. (8) Cody, R B., Jr.; Bjamason, A; Wed, D. A In Lasen in Mass Spectrometm Lubman, D. M., Ed.; Oxford University Press: New York, 1990; pp 316339. (9) Marshall, A G.; Grosshans, P. B. Anal. Chem. 1991,63, 215A-229A (10) Asamoto, B.; Dunbar, R C. Analytical A9plications of Fourier Transfotvn Ion Cyclotron Resonance Mass Spectrometry; VCH: New York, 1991. (11) Nuwaysir, L. M.; Wdkins, C. L. In Proceedings, SPIE Applied Spectroscopy in Material Science; Intemational Society for Optical Engineering, Bellingham, WA: 1991; pp 112-123. (12) Campana, J. E. In Proceedings,SPIEApPlied SpectroscoPy in Material Science; Intemational Society for Optical Engineering, Bellingham, WA: 1991; pp 138-149.

Analytical Chemistry, Vol. 66, No. 24, December 15, 1994 4385

Solouki and Russell detected transferrin dimer ions of m/z > ICRMS: parent ions are isolated by dipolar radial ejection with a 150 000 by use of an ion guide projecting into an ICR ion SWIFPG or a few frequency-sweep waveforms, the isolated parent McIver and Li obtained high-resolution FT-ICR mass spectra of ions are subjected to a brief dipolar excitation to increase the parent ion kinetic energy, collisions of the excited ions with small proteins (e.g., insulin) with their quadrupole extemal source 6.5 T FT-ICR in~trument.~~ neutrals deposit internal energy into the ions leading to fragmentation, and the product ions are detected by conventional FT-ICR High-resolution MALDI spectra could in principle provide accurate molecular weights for large biopolymers. With careful procedures. sample preparation, protonated molecules, MH+,are the dominant In order to detect ions more effectively and manipulate them species in a MALDI mass spectrum, although unwanted fragat will, we have developed a broad-band axialization ments, sodium and potassium cation adducts, and multimer spices based on quadrupolar excitation in the presence of collisional may also be present. Tandem mass spectrometry is the obvious cooling.44 In the axialization process, azimuthal quadrupolar choice for selection and structural analysis of ions of a given m/z excitation converts ion magnetron motion to cyclotron motion ratio.26 Although MALDI can generate abundant quasimolecular whose radius is damped by collisions of ions with neutrals. The ions, and although some of those ions can be captured and a x i h t i o n method is particularly useful for CID MS2 experiments, since collision gas is required for both processes. In a preliminary detected in an ICR ion trap, MS2of those singlycharged ions has paper,M we reported a method that provides simultaneous parent to date not been particularly successful. The high mass-to-charge ion dissociation by off-resonant excitation and product ion broadratio and large initial kinetic energy of ions created by MALDI make it difficult to deposit additional energy into parent ions and band axialization over an extended time period. Extensive to detect the (necessarily off-axis) product ions. fragmentation of potassiated gramicidin S ions was observed,and high mass-resolving power of CID MS2product ions was obtained. In an FT-ICR instrument, parent ions have been successfully fragmented by any of a variety of dissociation processes: collisionIn this paper, we extend that method to the successful efficient CID and observation of four MS stages starting from MALDE induced dissociation (0) ,27electron impact dissociation (EID) surfaceinduced dissociation (SID) ,29 and photodisso~iation.~~~~~ generated bradykinin ions. The three CID stages result in readily In identified losses of HzO or NH3 from the C-terminus, followed by fact, CID-based MS5 with m-ICR was achieved by Nibbering et the next two amino acid residues. With ongoing improvement in al. as early as 1984.32 (Several years later, multistage MS" experiments were realized in quadrupole ion trap mass spectromcontrol of the experimental event sequence, the current method e t e r ~for ; ~example, ~ ~ ~ ~Cooks et al. recently demonstrated MS'O promises to develop into a valuable tool for accurate and rapid for small organic hydrocarbon i0ns.3~) For relatively low-mass sequencing of singly-charged biopolymer ions. (m/z < 500 or so) organic ions and/or metal clusters of higher m/z, CID has proved particularly simple and effective by FTEXPERIMENTAL SECTION Sample Preparation and MALDI. The present experiments Marshall, A G.; Schweikhard, L. Znt. J. Mass Spectvom. Ion Processes 1992, 118/119, 37-70. were conducted with an Extrel FTMS 2000 instrument (Extrel Koster, C.; Kahr, M. S.; Castoro, J. A; Wilkins, C. L. Mass Specfrom. Rev. FTMS, Madison, WI) equipped with 1.875 in. dual cubic traps and 1992, 11,495-512. a 3 T superconductingmagnet. Near-UV photons (355 nm) of 7 Jacoby, C. B.; Holliman, C. L.; Gross, M. L In Mass Spectrometry in the Biological Sciences: A Tutorial; Gross, M. L., Ed.; Kluwer Academic ns pulse duration generated by frequency tripling of a Nd:YAG Publishers: Dordrechf The Netherlands, 1992; pp 93-116. laser (Model Surlite 1-10, Continuum, Santa Clara, CA) were Nibbering, N. M. M. Analyst 1992, 117,289-293. introduced from the analyzer end of the instrument, passing Brenna, J. T.; Creasy, W. R; Zimmerman, J. ACS Symp. Ser. 1993, No. through a 2 mm diameter conductance limit to impinge on the 236, 129-154. Buchanan, M. V.; Hettich, R L. Anal. Chem. 1993, 65, 245A-259A probe tip located behind the source trap plate. The third harmonic Cody, R B.; Kinsinger, J. A; Ghaderi, S.; Amster, I. J.; McLafferty, F. W. component (355 nm) of a Nd:YAG laser was separated from the Anal. Chim. Acta 1 9 8 5 , 178, 43-66. fundamental (1064 nm) and second harmonic (532 nm) by a WHettich, R L.;Buchanan, M. V. Int. J. Mass Spectrom. Ion Processes 1991, 111,365-380. reflective mirror and then loosely focused onto the probe tip by Deleted in proof. a 2:l telescope. Typical laser pulse energy was 1-2 mJ measured Castoro, J. A; Kaster, C.; W k s , C. Rapid Commun. Mass Spectrom. 1992, 6, 239-241. before the telescope. 2,SDihydroxybenzoic acid @HB)45was Castoro, J. A; Wilkins, C. L. Anal. Chem. 1993,65, 2621-2627. used as the MALDI matrix, with fructose as ~o-matrix.~~ BradySolouki, R; Gillig, K J.; Russell, D. H.Ana1. Chem. 1994,66,1583-1587. kinin (1 mM; Sigma Chemical Co., St. Louis, MO) was mixed McIver, R T.,Jr.; Li, Y.; Hunter, R L. Znt. J. Mass Spectrom. Zon Processes 1994, 132, Ll-L7. Tandem Mass Spectromety, McLafferty, F. W., Ed.; Wiley: New York, 1983. Cody, R B.; Bumier, R C.; Freiser, B. S. Anal. Chem. 1982,54,96-101. Cody, R B.; Freiser, B. S. Anal. Chem. 1 9 8 7 , 59, 1054. Ijames, C. F.; Wilkins, C. L. Anal. Chem. 1 9 9 0 , 62, 1295-1299. Hunt, D. F.; Shabanowitz, J.; Yates, J. R I.; Zhu, N.-Z.; Russell, D. H.; Castro, M. E. PYOC.Natl. Acad. Sci. U S A . 1987,84, 620-623. McLafferty, F. W.; Amster, I. J.; Furlong, J. J. P.; Loo, J. A; Wang, B. H.; Williams, E. R In Fourier Transform Mass Spectrometry: Evolution, Innovation, andApp1ication.s;Buchanan, M. V., Ed.; ACS Symposium Series 359; American Chemical Society: Washington, DC, 1987; pp 116-126. Kleingeld, J. C. Ph.D. Thesis, University of Amsterdam, 1984. Louris, J. N.; Brodbelt-Lustig, J. S.; Cooks, R G.; Glish, G. L;van Berkel, G. J.; McLuckey, S. A. Znt. J. Mass Spectrom. Ion Processes 1990,90, 117137. McLuckey, S. A; Glish, G. L.; van Berkel, G. J. Int. J. Mass Spectrom. Zon Processes 1991,106,213-235. Nourse, B. D.; Cox, K A; Morand, K L.; Cooks, R G. /. Am. Chem. SOC. 1992, 114, 2010-2016.

4386 Analytical Chemistry, Vol. 66, No. 24, December 15, 1994

(36) Marshall, A G.; Wang, T.-C. L.; Ricca, T. L. J. Am. Chem. SOC.1985,107, 7893-7897. (37) Guan, S.; Marshall, A G. J. Chem. Phys. 1993, 98, 4486-4493. (38) Guan, S.; Kim, H. S.; Marshall, A G.; Wahl, M. C.; Wood, T. D.; Xiang, X. Chem. Reu., in press. (39) Guan, S.; Wahl, M. C.; Marshall, A G. J. Chem. Phys. 1994, 100, 61376140. (40) Guan, S.; Marshall, A G.; Wahl, M. Anal. Chem. 1994, 66, 1363-1367. (41) Guan, S.; Marshall, A G . Rapid Common. Mass Spectvom. 1993, 7,857860. (42) Guan, S.; Wahl, M. C.; Wood, T. D.; Marshall, A G. Anal. Chem. 1993, 65, 1753-1757. (43) Wahl, M. C.; Kim, H. S.; Wood, T. D.; Guan, S.; Marshall,A G.Anal. Chem. 1993,65, 3669-3676. (44) Bollen, G.; Moore, R B.; Savard, G.; Stolzenberg, H. Appl. Phys. 1990,68, 4355-4374. (45) Strupat, K; Karas, M.; Hillenkamp, F. Znt. /. Mass Spectrom. Zon Processes 1991, 111, 89-102.

Homebuilt SWIFT Module

Pulse Train

Axialize

Analyzer Detect

486 PC I/O 4 On Interface

Data Station Trap Controller

7

17 Cl

Source Excite

B

I -4

Excitation Dipolar O+

II n

Conduciance Limit Figure 1. ICR ion dual-trap configurationfor simultaneous quadrupolar excitationand off-resonantdipolar excitationthrough a resistive network (left) and conventional dipolar excitation and dipolar detection (right). In the present experiments, parent ions are dissociated and the product ions are axialized in the presence of buffer gas in the source trap (left) and then passed to the analyzer trap (right) for conventional highresolution dipolar excitation/detectionat low pressure.

with DHB (1 M; Aldrich Chemical Co., Inc., Milwaukee, WI) and fructose (1 M; Aldrich) as matrix and co-matrix, respectively, in molar proportions, bradykinin:DHB:fructose = 1:1000:500. Each solid sample was dissolved in 0.5%trifluoroacetic acid (Aldrich) in methanol. About 10 pl of the mixture (-4 nmol of analyte) was applied to a thin stainless steel plate with a diameter of 3/4 in. and dried in air. However, the laser spot size is less than 1% of the area of the probe tip, and ions may be obtained from each of up to 20 laser shots at the same spot, so that the amount of sample consumed in an MS4experiment (i.e., four detected mass spectra, one for each MS stage) is -10 pmol. MSnTechniques. Figure 1shows the electrical connections between various experimental modules. An Extrel Odyssey data system provides the timing, trapping control, source dipolar excitation, and analyzer excitation and detection. A 'ITL signal was set during axialization periods and a train of TTL signals generated by a digital IO board in a 486 PC. Broad-band SWIFT' quadrupolar excitation39signals for axialization generated by a home-built SWIFT module36 were applied to the source trap through a resistive network which allows for simultaneous quadrupolar excitation and dipolar ex~itation.~~ The experimental event sequence is given in Figure 2. Laser-desorbed ions were decelerated by a 9.75 V bias of the conductance limit relative to the (grounded) probe tip and source trap plate. Bradykinin MH+ ions were then axialized for 20 s in the presence of nitrogen gas admitted continuously into the source trap chamber through a leak valve (Model 951-5100, Varian Vacuum Products, Palo Alto, CA). Pressure was maintained at 5 x lop7Torr in the source chamber and -0.3 x lo-* Torr in the analyzer chamber. Azimuthal dipolar SWIFTexcitation was then used to isolate MH+ ions of bradykinin by radial ejection of ions of other m/z ratios. Azimuthal dipolar off-resonantexcitation (6.25 Vp-pamplitude) at the frequency 44.2 kHz (i.e., 500 Hz below the cyclotron frequency of MH+ ions) was applied for 20 s to induce dissociation of MH+ with simultaneousrepeated SWIFT azimuthal quadrupolar excitations (6.5 Vp-p amplitude) over a cyclotron frequency range spanning 200 Im/z 5 1200 to axialize the product ions. The most abundant product ion was chosen to be the parent ion of the next stage, and the process was repeated to achieve multistage

Laser Desorptionllonlzation

Broadband Quadrupolar Excitation

_rz[JT]nJ+L Dipolar Excitation (Ion Isolation)

Off-resonance Dipolar Excitation

'

[

~ ~ l n Ion Transfer to Analyzer Trap

i

n I

I

Conventional ExcitationlDetection

Figure 2. Experimentalevent sequence for multistage MS", in which each stage comprises single-frequency off-resonant dipolar excitation to increase parent ion kinetic energy for collision-induced dissociation, with simultaneous broad-band axialization of product ions.

CID MS2, MS3, and MS4. For the parent MH+ collision-induced dissociation, both the duration and frequency offset of offresonance excitation were optimized. For subsequent MS stages, only the frequency offset was optimized. Excitation and Detection of FT-ICR Mass Spectra. To obtain a mass spectrum after a given CID stage, the conductance limit voltage was dropped to 0 V for 200 ,us to allow axialized product ions to transfer to the analyzer trap. The conductance limit and analyzer trap plate were then set to +3 V dc. Excitation was then performed by frequency sweep (chirp) excitation from1 to 500 kHz (500 Hz/,us sweep rate). Detection was performed in heterodyne mode (158.272 kHz bandwidth, 128K data points). Discrete Fourier transformation followed by magnitude calculation then yielded a mass spectrum. RESULTS AND DISCUSSION

A critical prerequisite for multistage MSn is to have a large number of parent ions at the first stage. In order to obtain a large Analytical Chemistry, Vol. 66,No. 24, December 15, 1994

4387

number of MH+ ions of bradykinin, we optimized the bradykinin concentration as well as the ratio of analyte to DHB matrix and fructose cc-matrix. We also used the deceleration potential method developed by Castor0 and W~lkins~~ to reduce parent ion kinetic energy and thereby increase the efficiency of capture of ions by the trap. Bradykinin MH+ ions could still be detected with excellent signal-to-noiseratio after 30 s in the source trap in the presence of 5 x Torr of Nz collision gas, indicating that unimolecular decomposition did not interfere with our CID experiments. Quadrupolar axialiiation was applied for 20 s after laser desorption, and the ions were transferred to the analyzer side for high-resolution heterodyne detection. Figure 3 (top) shows SWIFT-isolated quasimolecular bradykinin ions, MH+, detected in the analyzer trap. Effective collision-induceddissociation of parent ions generated by off-resonanceexcitation requires a lengthy excitation However, at a relatively high collisional gas pressure, the lower

m/z product ions are then quickly lost from the trap. Guan et al.40 demonstrated that simultaneous application of azimuthal dipolar off-resonance excitation of parent ions and azimuthal quadrupolar excitation of product ions results in high dissociation efficiencyand optimal mass analysis conditions for product ions. Here we extend that method to multistage CID MSNof MALDIgenerated peptide ions. Bradykinin quasimolecular ions, MH+, were first isolated from ions of all other m/z ratios by an azimuthal dipolar SWIFT ejection event. Application of off-resonant excitaTorr of Nz gas induced dissociation tion in the present of 5 x of the MH+ ions while axializing the product ions continuously. M e r 20 s of dipolar off-resonant excitation and broad-band axialization, ions were transferred to the analyzer trap. Highresolution detection was carried out and product ions in the mass range 600 5 m/z 5 1100 were detected. CID is highly efficient-essentially all parent MH+ ions are dissociated after the 20 s dipolar off-resonant excitation period. The primary CID fragment from the parent MH+ ions is (MH - NH$HzO)+, m/z = 1043/1044. (MH - NH3/H20)+ ions were then isolated from other fragments by application of an additional azimuthal dipolar SWIFTexcitation waveform (Figure 3, second from top), followed by a second stage of fragmentation by simultaneous azimuthal dipolar ofl-resonant excitation and azimuthal quadrupolar axialization of product ions. The most abundant secondary product ions, (MH - NH3 - Arg)+, as well as (MH - HzO - Arg)+, or be+, and the rearrangement product (MH - NH3 - Arg + Wt, were then isolated in the source trap as parent ions for the next MS stage. Figure 3 shows our CID tandem mass spectra of ions generated by MALDI of bradykinin up to four MS stages. Protonated bradykinin MHt ions lose HzO or NH3 from the C-terminus, followed by the next two consecutive amino acids, arginine and phenylalanine. The mass resolving power, m/Am, in which m is ionic mass and Am is the full width at half-maximum of the magnitudemode spectral peak, decreases from -50 000 to -2500 in proceeding from the first to fourth MS stage, due to the decreasing number of ions and inefficient trapping of product ions after multiple-stage CID, leading to loss of spatial coherence in the ion packet. Our mass measurement accuracies were generally in the low ppm range. The average error was -2.6 PPm. Although it is not easy to compute CID efficiency (because ICR signal strength depends on the spatial distribution as well as the number47 of trapped ions), we estimate that the efficiencies are -85%, -loo%, and -100% for MS2,MS3, and MS4,respectively. In contrast to gramicidin S, which distributes ion charge among many different fragments after a single CID stage, the fragmentation of bradykinin during CID is quite selective. The CID selectivities are -70%, -60%, and -66% for MS2, MS3, and MS4 for the fragment chosen as the parent ion for the next stage. Moreover, again in contrast to gramicidin S, the bradykinin polypeptide chain tends to cleave at the weakest bond. The different fragmentation pattern obtained with off-resonance collisional activation evidently reflects a difference in the amount or rate at which internal energy is deposited in the molecule during the off-resonance excitation process. In fact, it is interesting to note that the presently observed dissociation pattern of stepwise cleavage starting from the carboxyl terminus resembles stepwise

(46) Gauthier, J. W.; Trautman, T. R: Jacobson, D.B. Anal. Chim. Acta 1991,

(47) Limbach, P. A; Grosshans, P. B.; Marshall, A G. Anal. Chem. 1993, 65,

M = H-Arg-Pro-Pro-Gly-Phe-Ser-Pro-Phe-Arg-OH

II

MH+ I

AI/, 1065.0

B

A 1041.0

y7"+

1045.0

i 800.0

l

I..

I

760.0

1 MH+

b8+

li

(MH-NH3-Arg)+ (MH-NHrArg+H)+

-

.

I

Ye"+

1 1 .

900.0

,

~

~

-

1000.0

'

l

*

,

1100.0

-

b8+

(MH-NHrArg)+ (MH-NH3-Arg+H)+ (MH-H,O-NH2CN-NH,)* (MH-H*O-NH&N)+

700.0

I,

900.0

800.0

(MH-NH3-Arg-Phe-H20)+

7W.O

800.0

'

900.0

'

1obo.o

llbo.0

Figure 3. FT-ICR mass spectra following each stage of an MS4 experiment for bradykinin ions produced by matrix-assisted laser desorptionhonization: (A) protonated bradykinin ions, MH+, isolated in the source trap and detected in the analyzer trap. The inset is a mass scale-expanded spectral segment, in which carbon-13 isotopic species are well separated. The isotopic distributions in MS2and MS3 spectra in fact represent convolutions of two or more isotopic distributions corresponding to the loss of different neutrals (NH3 or H20 from the protonated molecular ion before further degradation). (B) product ions from collision-induced dissociation of protonated bradykinin. [MH - H*O/NH3]+ ion was selected to be the parent ion for the next CID stage. (C) product ions from CID of [MH - H20/ NH3]+ ions. (D) product ions isolated following CID of [MH - NH3 Arg]+ and its rearrangement derivative ions from the previous stage. Note that even after MS4, FT-ICR mass resolving power is still high enough to separate carbon-13 isotopes.

246, 211-225.

4388 Analytical Chemistiy, Vol. 66,No. 24, December 15, 1994

~

i

135-140.

Edman degradation starting from the amino terminus. Finally, since lower mass ions appear to require higher CID energy for comparably efficient fragmentation (see below), CID cleavage of higher mass peptides should be feasible with the present technique. The azimuthal dipolar off-resonance excitation frequency for CID obviously increases in successive MS stages, as each newly isolated parent ion has lower m/z ratio. One would nevertheless expect that the efficiency of off-resonant excitation should be conserved by maintaining a constant frequency offset between the excitation frequency and the reduced ICR frequency. However, we observed empirically that off-resonance CID was optimized by use of a smaller frequency offset at higher reduced ICR frequency: e.g., frequency offsets of 500,470,and 420 Hz for ions of reduced ICR frequencies, 44200, 44900, and 52900 Hz, respectively, according to an empirical rule that v(v+ - v) 2.2 x lo7 Hz2, in which v is the off-resonance excitation frequency and (v+ - v) is the frequency difference between the reduced cyclotron frequency and the off-resonance excitation frequency. Wang and Marshall have derived an expression for an ion trajectory during and after linearly polarized off-resonance single frequency excitation electric field, E@).@From the time derivative of their eq 24, we obtain the ion x-y velocity, v(t), and the instantaneous power absorption, qE(t)v, from which the kinetic energy absorbed by the ions at the end of an excitation period of T seconds is given by eq 1in which q is ion charge, EOand BOare

KE at time, T = JLTqE(t)vdt = qw+El sin[(w+ - o)T/21 2BO o+--0

-(

the radio frequency electric field amplitude and static magnetic field, respectively, and W + and w are the (reduced) ion cyclotron angular frequency and irradiation angular frequency, (SI units). Thus, ions are excited and then deexcited over a time period, 1/(v+ - v). Moreover, off-resonance excitation is typically conducted for an interval, T, which is long compared to one period of the frequency offset, W + - w. Therefore, since T>>1/(v+ v), we may compute the average kinetic energy of an ion during (48) Wang, M.; Marshall, A G. Int. J. Mass Spectrom. Ion Processes 1990,100, 323-346. (49) Heck, A J. R; de Koning, L. J.; F'inkse, F. A; Nibbering, N. M. M. Rapid Commun. Mass Spectrom. 1991,5 , 406-414. (50) Laukien, F. In Proceedings of the 35th American Society of Mass Spectrometry Conference on Mass Spectromety and Allied Topies; American Society of Mass Spectrometry: Denver, CO, 1987; pp 781-782. (51) Marzluff, E. M.; Beauchamp,J. L. In Proceedihgs ofthe 42ndAmerican Society of Mass Spectrometry Annual Conference on Mass Spectromety and Allied Topics;American Society of Mass Spectrometry: Chicago, IL, 1994; WF'222.

off-resonantexcitation as the time-average of eq 1over one period, l/v+ - v = 2n/(w+ - o),of the frequency offset, to obtain eq 2.

KE(av) = (v+ - v )

Equation 2 has been reported previously (except for a missing factor of q).49 Thus, for large frequency offset, the ion cyclotron radius oscillates with small amplitude; for smaller frequency offset, the ion cyclotron radius oscillates with larger amplitude, to give a larger average kinetic energy during excitation. [Note that because we average over one period of the frequency offset, eq 2 predicts infinite kinetic energy for on-resonant excitation (i.e., v v+), because the frequency offset period becomes infinite in that case. Therefore, in using eq 2,one must remember that it applies only to off-resonant excitation for a length of time, T >>

-

l/b+- $.I Equation 2 shows that the total kinetic energy of the ions absorbed during off-resonance excitation varies inversely with the square of the frequency offset.49 Our experimental results therefore suggest that more kinetic energy is needed to break apart ions of lower mass. For a hard-sphere model of the CID process (a reasonable approximation for ions following off-resonant excitation50), the ion-neutral collision frequency is proportional to ion average speed, vave(obtained by setting eq 2 equal to mvav,2/ 2), which is in turn proportional to v+/(v+ - v). (Again, note that the collision frequency does not become infinite as v v+, provided that T >> l/(v+ - v). Perhaps the efficiency of kinetic to internal energy conversion is reduced as the ion mass decreases, as suggested in a recent study by Marzluff and Bea~champ.~~ Tandem-in-spaceMSnbecomes impractical for more than two MS stages. Here, we demonstrate for the first time efficient MS4 of singly-charged peptide ions, with high-resolution detection of the product ions at each stage. We believe that the multistage MSncapability demonstrated here can be developed into a routine and robust analytical method for rapid and accurate sequencing of large molecules of biological interest.

-

ACKNOWLEDGMENT

This work was supported by NIH (Grant GM-31683) and the National High Magnetic Field laboratory at Florida State University. Received for review June 2, 1994. Accepted October 3, 1994. Abstract published in Advance ACS Abstracts, November 1, 1994.

Analytical Chemistry, Vol. 66, No. 24, December 15, 1994

4389