Anal. Chem. 1995, 67,2180-2187
High-Performance Collision-Induced Dissociation of Peptide Ions Formed by Matrix-Assisted Laser Desorption/ionization in a Quadrupole Ion Trap Mass Spectrometer Vladlmir M. Doroshenko and Robert J. Cotter* Middle Atlantic Mass Spectrometry Laboratory, Department of Phamawlogy and Molecular Sciences, The Johns Hopkins University School of Medicine, Baltimore, Maryland 21205
A modified ion trap detector has been utilized to obtain high-performance collision-induced dissociation (CID) mass spectra of peptide ions formed by matrix-assisted laser desorption/ionization (MALDI). MALDI ions are trapped while increasingthe fundamental radio fiequency field, obviating the need for elevated helium gas pressures, Molecular ion isotopic clusters are then isolated by a reverse-forward-reverse scan sequence. A single species within the isotopic cluster (generally the monoisotopic mass) is then selected for activation. €ha&, modulation of the amplitude of the resonant excitation voltage on the end-cap electrodes, used previously to improve mass calibration in normal mass spectra, is now utilized to provide high mass accuracy for the product ions. The CID mass spectra of several protonated and sodium-cationized peptides are presented and are often characterized by a series of rearrangement ions that can be utilized in the determinationof amino acid sequences. Collision-induced dissociation (CID) enhances the structural information available from mass spectrometry by increasing the fragmentation of molecular and fragment ions through energetic collisions with an inert gas. CID is used most effectively on tandem instruments, which also provide the opportunity for selection of a single ionic species from a mixture in the first mass analyzer and recording of its CID mass spectrum in the second. Tandem mass spectrometers currently include four-sector? triplequadrupole? and hybrid4instruments, as well as Fourier transform mass spectrometers (FTMS)5 and quadrupole ion traps6 configured to carry out tandem-in-timemeasurements in a single mass analyzer. For the determination of complex biological structures, unit mass selection of precursor ions, high resolution, and accurate mass assignment of product ions may be required7 and can generally be achieved only on four-sector and FTMS instruments. These high-performance instruments are, of course, the most (1) Jennings, K. R Int. 1.Mass Spectrom. Ion Phys. 1968,1, 227. (2) Todd, P. J.; McGilvery. D. C.; Baldwin, M. A; McLafferty, F. W. In Tandem Mass Spectrometry; McIafferty, F. W., Ed.; Wiley: New York, 1983; p 271. (3) Yost, R A.; Enke, C. G. J. Am. Chem. SOC.1978,100, 2274. (4) Glish, G. L;Mchckey, S. A; Ridley,T. Y.; Cooks, R G. Int.J. Mass Spectrom. Ion Phys. 1982,41, 157. (5) Cody, R. B.; Burnier, R C.; Freiser. B. S. Anal. Chem. 1982,54, 96. (6)Louris, J. N.; Cooks, R. G.; Syka, J. E. P.; Kelley, P. E.; Stafford, G. C.;Todd, J. F. J. Anal. Chem. 1987,59, 1677. (7) Sato, K.; Asada, T.; Ishihara, M.; Kunihiro, F.; K a " e i , Y.; Kubota, E.; Costello, C. E.; Martin, S. A; Scoble, H. A; Biemann, K. Anal. Chem. 1987, 59, 1652.
2180 Analytical Chemisrry, Vol. 67,No. 73,July 7, 7995
expensive and unlikely to become the bench-top instruments that might be utilized routinely by biological scientists (for example) to determine amino acid sequences of peptides. Mass spectrometers based upon a quadrupole ion trap have considerable potential as inexpensive tandem instruments! An ion trap mass spectrometer was first produced commercially in 1983 by Finnigan MAT (San Jose, CA) as an ion trap detector (ITD) for gas chromatography with a mass range of 650 Da and limited resolution. Since that time, masses exceeding 70 OOO Da have been recordedgby ejecting ions from the trap through axial resonance excitation,'O while mass resolutions greater than one part in lo7 have been achieved at very slow scan speeds." Because resonance excitation increases ion velocities, it has also been used to induce fragmentation through energetic collisions for the recording of MS/MS spectra.6 Indeed, such processes can be carried out in multiple stages, MS", where n = 12 has been demonstrated for small ions.I2 In 1990, the first CID mass spectra of peptides were obtained on an ion trap mass spectrometer, using a secondary ion mass spectrometry (SIMS) ~0urce.l~ Because they are capable of desorbing and ionizing peptides and proteins with masses in excess of 100 kDa, electrospray ionization @SI) and matrix-assisted laser desorption/ionization (MALDI) have also been used as ion sources for the ion t r a ~ . ' ~ - ' ~ Because it is a pulsed technique, MALDI is particularly attractive as an ion source. Unlike in continuous sources, in MALDI, neutral species formed during the ionization process are rapidly cleared from the trap prior to the mass measurement cycle, thus reducing (8) March, R. E. Int. J. Mass Spectrom. Ion Processes 1992,118/119,71. (9) Kaiser, R. E.; Cooks, R G.; Stafford, G. C.; Syka, J. E. P.; Hemberger, P. H. Int. J. Mass Spectrom. Ion Processes 1991,106, 79. (10) Kaiser, R. E.; Louris, J. N.; Amy, J. W.; Cooks, R. G. Rapid Commun. Mass Spectrom. 1989,3, 225. (11) Londry, F. A; Wells, G. J.; March, R. E. Rapid Commun. Mass Spectrom. 1993,7, 43. (12) Louris, J. N.; Brodbelt-Lustig, J. S.; Cooks, R G.; Glish, G. L.; Van Berkel, G. J.; McLuckey, S. A Int. J. Mass Spectrom. Ion Processes 1990,96, 117. (13) Kaiser, R E.; Cooks, R G.; Syka, J. E. P.; Stafford, G. C. Rapid Commun. Mass Spectrom. 1990,4 , 30. (14) Van Berkel, G. J.; Glish, G. L.; McLuckey, S. A.Anal. Chem. 1990,62, 1284. (15) Cox, K. A; Williams, J. D.; Cooks, R G.; Kaiser, R. E. Biol. Mass Spectrom. 1992,21, 226. (16) Doroshenko, V. M.; Cornish, T.J.; Cotter, R. J. Rapid Commun. Mass Spectrom. 1992,6, 753. (17) Chambers, D. M.; Goeringer, D. E.; McLuckey, S. A; Glish, G. L. Anal. Chem. 1993,65, 14. (18) Jonscher, K.; Currie, G.; McCormack, A. L.; Yates. J. R Rapid Commun. Mass Spectrom. 1993,7, 20. (19) Schwartz, J. C.; Bier, M. E. Rapid Commun. Mass Spectrom. 1993,7, 27.
0003-2700/95/0367-2180$9.00/0 0 1995 American Chemical Society
ion-neutral interactions and increasing the lifetime of the ions in the trap. CID mass spectra have been obtained for ion traps I s o ~ r c e , ~but ~ J *not under the highequipped with a W performance conditions that include unit mass selection, high resolution, and high mass measurement accuracy. CID spectra obtained on a commercial Finnigan research grade ion trap mass spectrometer (TINS) were initially limited in mass resolution by the inability to utilize slow scan speeds.13J7J8Later modifications to the ITMS permitted slower scan speeds but reduced the mass recording window to only a few daltons?JS Isolation of the precursor ion in these experiments generally involved setting the direct current (dc) and radio frequency (rf) voltages so that ions of lower and higher masses are simultaneously ejected near an apex of the ion stability diagram.13J7J8 Because this approach does not permit unit mass selection, the entire isotopic cluster was used as the precursors for CID experiments. An alternative approach utilized dc and rf voltage ramping to remove lower and higher mass ions in two consecutive stages, but this has been demonstrated only for relatively low mass ions.20 Reverse-then-forward ejection techniques, utilizing rf-only scans in the axial resonance ejection mode, have been utilized for isolation of isotopic ion clusters13and (at slower scan rates) unit mass selection.2l In addition, broad-band excitation methods have also been ~ s e d . ~ ~ s ~ ~ We recently introduced two techniques, one for trapping MALDI ions and another for recording their mass spectra, that provide considerable improvement in mass resolution and accuracy and are therefore utilized in the present work to achieve high CID performance. First, we have improved the trapping efficiency for MALDI ions by increasing the rftrapping field (from zero to its full strength) during the time that the ion pulse travels toward the center of the trap.24925A similar method has been described by Nogar et alF6and obviates the need for using higher pressures of the helium buffer gas to stabilize and trap the ions. This, in turn, permits one to utilize pressures within the ion trap that can be optimized for high mass resolution and ion isolation. Second, while mass measurement accuracy is adversely affected by scan speeds, buffer gas pressures, and space charge effects due to large numbers of trapped ions, an apparent mass shift is often observed in the resonance ejection mode which arises from the fact that the secular frequencies of the ions are generally different from the excitation frequencies. We have previously addressed this problem by linking the ejection voltage amplitude to the fundamental rf voltage, in order to achieve a linear dependence of ejected mass on the scanned rf~oltage.2~ Both of these techniques are utilized in this work to achieve highperformance mass selection and recording of CID spectra. (20) Gronowska, J.; Paradisi, C.; Traldi, P.; Vettori, U. Rapid Commun. Mass Spectrom. 1990,4, 306. (21) Schwartz, J. C.; Jardine, I. Rapid Commun. Mass Spectrom. 1992,6, 313. (22) Kenny, D. V.; Callahan, P. J.; Gordon, S. M.; Stiller, S. W. Rapid Commun. Mass Spectrom. 1993,7, 1086. (23) Soni, M. H.; Cooks, R G. Anal. Chem. 1994,66, 2488. (24) Doroshenko, V. M.; Cotter, R J. In Loser Ablation: Mechanism and Applications It Proceedings of the 2nd International Conference on Laser Ablation, Knoxville, TN, April, 1993;Miller, J. C., Geohegan, D. B., Eds.; American Institute of Physics: New York, 1993; p 513. (25) Doroshenko, V. M.; Cotter, R J. Rapid Commun. Mass Spectrom. 1993,7, 822. (26) Eiden, G. C.; Cisper, M. E.;Alexander, M. L;Hemberger, P. H.; Nogar, N. S . J Am. SOC.Mass Spectrom. 1993,4, 706. (27) Doroshenko, V. M.; Cotter, R J. Rapid Commun. Mass Spectrom. 1994,8, 766.
EXPERIMENTAL SECTION Experiments were carried out on a Finnigan MAT ITD previously modified to accommodate MALDI ionization inside the t r a ~ . ~Ions ~ 3are desorbed from 0.04 in. diameter probe tip that is inserted into the ion trap cavity through the existing 0.05 in. diameter hole in the upper endcap electrode normally used for introduction of the electron beam F i r e 1) and centered by a cylindricalTeflon spacer. Ions were desorbed by the pulsed (10 ns) laser beam from a frequencyquadrupled Quantel International (Santa Clara, CA) model YG660-10 NdYAG laser attenuated by a Newport (Fountain Valley, CA) model 9355 attenuator and focused onto the sample probe through the gap between the ring and the bottom endcap electrode by a 50 cm focal length UV quartz lens. Ions formed by MAL,DI have initial kinetic energies of several electron voltsz8and, in our instrument, are trapped by increasing the fundamental rf voltage as the ions travel into the t1-ap.2~Thus, ions initially encounter and penetrate a weak trapping field as they enter the trap but are confined by the high trapping field present when they reach the center. The settling value of the rf trapping voltage during the storage period is generally 6 0 4 0 % of the maximum value of 15 kVp-p,corresponding to a mass cutoff level of 390-520 Da. This limits the number of stored ions and decreases possible effects due to space charging. Although collisions with helium are not required for trapping, a buffer gas is used to return the ions to the center of the trap and to improve sensitivity and mass re~olution.'~The buffer gas pressure, monitored by a CVC Products (Rochester, NY) GPT-450 vacuum gauge with a GIC-048-2Penning cold cathode discharge tube, was from 4 x to 5 x Torr. The mass range of the lTD was extended by isolating the endcaps from ground and applying a supplementary dipolar rf voltage across the endcaps to enable operation in the resonance ejection mode.I0 Voltages were supplied to the endcap electrodes by two controlled voltage regulators connected to the balanced outputs of a Wavetek (San Diego, CA) model 95 arbitrary/function generator. The controlled voltage regulators provide better performance than the transformers used p r e v i o u ~ l y ' and ~ ~ ~have ~*~~ summing amplisers at their inputs that permit the use of multiple signal sources and independent control of the voltages on each endcap electrode. Instrument Control and Data Acquisition. The modified ITD was controlled by an IBM AT-486 computer using a National Instruments (Austin, TX) LabPC+ multifunctional input/output plug-in board and Trapware, control software written in-house in the Turbo Pascal (Borland International, Scotts Valley, CA) lang~age.2~ Two LabPC+ 12-bit resolution analog outputs were used to control the trapping rf voltage via the ITD analog board. In addition, the LabPC+ digital output lines were used for other ITD operations,including multiplier and rf voltage supply on/off, multiplier voltage setup, and 12-bit resolution ITD digital-teanalog converter (DAC) output control. Other LabPC+ outputs were used to control the Nd:YAG laser, the digital oscilloscope, and multiplexer operation, using the LabPC+ timers to control laser firing and to trigger the oscilloscope. These operations were all controlled by Trapware, which was also used to set up all operational parameters, including output control voltages, excitation frequencies, and the duration between adjacent acquisition points (generally 120 ps), to transfer data (mass spectra and (28) Beavis, R C.; C h i t , B. T.Chem. Phys. Lett. 1991,181, 479.
Analytical Chemistry, Vol. 67, No. 13, July 1, 1995
2181
GPIB
II
Multlplsxsr control
I
Arbltrsry/ amnarstor
I
4
-
Buffer amp1 i f l e r
Analog
obtain clean excitation.) Such off-resonance excitation would be f =- I more important in this case if the losses of HzO and NH3 represent - 1 i fragmentation channels with lower activation energies than those resulting in the formation of bg and y"9 ions. Nonetheless, I S.+O b' excitation of the isotopic species at m/z 1297.7 contributesto some portion of the relative abundance of the ion at m/z 1279.6 and to the ion at m/z 1280.6. Figure 2, parts d and e, show the isolation of the isotopic cluster and excitation of m/z 1296.7, respectively, 800 850 900 950 for a 5 pmol sample, representing the current practical sensitivity of the technique. Ions are excited within a finite bandwidth of the excitation rf voltage applied at the endcaps, generally described as an absorp tion contour with a Lorentzian line shape. As described earlier?' h this gives rise to apparent mass shifts in recording spectra and was the rationale for utilizing excitation voltage amplitudes that are proportional to the fundamental rf voltage to provide a linear mass scale and improve mass calibration. This absorption contour is broadened by higher pressures of the helium gas used to return 400 500 500 7b0 800 mo 1000 ions to the center of the trap. Because excitation by the rf voltage m/ z applied to the end caps is involved in both ion ejection and ion Flgure 3. CID mass spectra of protonated molecular ions of activation, high pressures of helium gas will affect mass resolution, a-casein fragment 90-96 (MW = 912.5) at (a) 0.5 and (b) 2.5 mTorr mass isolation, and selective activation, as well as mass accuracy. of helium gas. In Figure 3, we compare the CID mass spectra of the monoisotopic 9ss4n0 n 1 2 3 4 5 8 7 protonated molecular ion of acasein fragment 90-96 (m/z 913.5) 136.1 267.1 414.2 551.2 884.3 795.4 910.4 2 i: 164.1 295.1 442.2 579.2 692.3 823.4 938.4 58 at different pressures of the helium gas. F i r e 3a shows the H - Tyr . Me1 . Pho . H N . LOU . Mol . Asp . NH, Y". 792.4 661.3 514.2 377.2 264.1 133.1 m results of excitation of the m/z 913.5 ion at a pressure of 0.5 mTorr. n 6 5 4 3 2 1 I Typical of peptides containing a basic amino acid residue on the n" C amino terminus, this spectrum reveals primarily a,, and b, ions and no y",, ions.31 Small neutral losses (H20 and NH3) are abundant, consistent with a quasiequilibrium model for peptide decomposition during prolonged (tens of milliseconds) activation and a low rate of energy deposition by multiple lowenergy collisions. This time frame also promotes rearrangement reactions, such as the b'g OH ions formed by the transfer of a 5w 760 860 000 m/z hydroxyl group from a terminal glutamic acid to an adjacent Figure 4. CID mass spectrum of protonated molecular ions of leucine and the formation of a stable ring structure. Such dermenkephalin (MW = 954.4). In this mass spectrum, the MH+ ions reactions have been observed only in low energy collisions31and at m/z 956.4 correspond to the 1 3 C isotopic peak, since the monoisoverified by l80labeling experiment^.^^ topic peak has been removed by excitation. Figure 3b shows the same experiment carried out at a helium gas pressure of 2.5 mTorr. The isotopic cluster is no longer which utilizes ramping of the fundamental rf field rather than high resolved, so selective isolation cannot be carried out, and the peaks pressures of helium to trap energetic ions. corresponding to losses of HzO and NH3 cannot be distinguished. Figure 4 shows the results of activation of the m/z 955.4 ion This result also underscores the advantages of the trapping of dermenkephalin, which has a basic residue in the middle of method for MALDI ions, which we have previously d e ~ c r i b e d , 2 ~ ~ ~the ~ molecule. In this case, a, and b, ions result from cleavages of the peptide bonds to the right of the basic residue, while the (31) Gross, M. L.; Hu, P. In Biological Mass Spectrometry: Present and Future: y"4 ion results from cleavage to the left of the basic residue, Matsuo, T., Caprioli, R M., Gross, M. L,Seyama, Y., Eds.: Wiley: New consistent with CID spectra of peptides with a charged center.3l York, 1994; p 255. Figure 5 shows the results of activation of the protonated (32) Thome, G . C.; Gaskell, S. J. Rapid Commun. Mass Spectrom. 1989,3, 217.
1
OD
7
"
T
$1
A
N
+
2184 Analytical Chemistry, Vol. 67,No. 13, July 1, 1995
A
5563DO
p’
(a)
%
4 H
v=
136.1 184.1
-
Tyr
n
-
103.1 221.1
250.1 278.1
.
GC
Ob
393.2 4
336.2 3
4
397.2 425.2
Pho
.
279.2 2
5 510.3 538.3
Leu. OH 132.1 1
m
-
2
i
0m; l a
sCC
h e2
m i
i n a. 574 p
oo
b, H
.
1 2 138.1 193.1 104.1 221.1 Tyr . G t 411.2
.
4
3 250.1 278.1
4 397.2 425.2
Gly
Phr
354.1
297.1
3
2
5
-
5282
N
556.2 Met - OH 150.1
2
1
i
si0 350
4 h
m/z
450
3dO
ai0
ldoo
llbo
109
1150
550
Flgure 5. CID mass spectra of protonated molecular ions of (a) Leu-enkephalin (MW = 555.3) and (b) Met-enkephalin (MW = 573.2). In these spectra, the MH+ ions at m/z 557.2 and 575.4 correspond to the 1 3 C isotopic peaks, since the monoisotopic peaks have been removed by excitation.
molecular ions at m/z 556.3 and 574.2 for the two similar peptides, Leuenkephalin and Met-enkephalin,respectively. These peptides do not contain any basic residues, so fragmentation, primarily of the a,, and b, types, is distributed along the peptide backbone. For Leu-enkephalin,the ratio of the y”2 to b3 ion has been shown to reflect the kinetic energy leading to collisional acti~ation.3~3~ Generally,the intensity of the y”2 ion becomes comparable to that of the b3 ion for high-energy c o l l i ~ i o n s . ~ ~ ~ ~ ~ Figure 6a shows the CID spectrum for the protonated molecular ion at m/z 1141.7 for gramicidin S. This cyclic peptide requires at least two cleavages of the peptide backbone to produce the fragment ions that are observed. Because the fist cleavage results only in opening the ring and not in a change in mass, the supplementary dipole field between the end-cap electrodes will continue to excite these ions until a second cleavage produces an ion of different mass, which then returns to the center of the trap. Gramicidin S is composed of two Leu-Phe-Pro-Val-Om sequences connected in a head-to-tailfashion, so the peak at m/z 571.4 could represent any of a set of fragment ions corresponding to onehalf of the molecule and is, therefore, particularly intense. Thus, in Figure 6b we have obtained the CID mass spectrum of ions with m/z 571.4, that is, an MS/MS/MS spectrum of gramicidin S. At the same time, the fragment ions observed in these spectra are consistent with a primary initial cleavage between ornithine and leucine, leading to the formation of an acylium ion. The b, fragments formed from a second cleavage are also consistent with the formation of acylium ions.35 The identity of these ions is then confumed by the facile losses of CO observed at m/z 429.2, 543.3, and 803.4 in the two CID spectra. (33) Baeten, W.; Claerboudt, J.; Van den Heuvel, H; Claeys, M. Biomed. Enuiron. Mass Spectrom. 1989,18, 727. (34) Alexander, A J.; Boyd, R IC Int. J. Mass Spectrom. Ion Processes 1989,90, 211.
(35) Eckart, K. M a s Spectrom. Rev. 1984,13, 23.
I
I . i
L
I
200
I
300
1
400
I . I
500
m/Z
Figure 6. (a) MSlMS spectrum of protonated molecular ions of gramicidin S (MW = 1140.7). (b) MSlMSlMS spectrum obtained by excitation of the fragment at mlz571.4. ions shown in the inset were observed in the experiment. In this mass spectrum, the MH+ ions at m/z 1142.7 correspond to the 1 3 C isotopic peak.
It is well known that the presence of proline residues can also direct the fragmentation of peptides.36 Proline has a fivemembered ring structure involving the amino nitrogen that reduces the bond strength of the amide bond on its N-terminal side. Thus, first cleavages at proline have often been observed for other cyclic peptides containing this residue.37 In our spectra, the peak at m/z 767.4 corresponds to an initial cleavage between the Phe and Pro residues. In addition, the peak at m/z 311.4 in the MS/MS/MS spectrum might as well result from first and second cleavages at both Phe-Pro bonds (which would also result in the formation of an ion at half the mass of the molecular ion), followed by cleavage of the Om-Leu bond through actbation of these ions at m/z 571.4. However, in most instances it appears that the more basic ornithine residue directs the fragmentation, although this pattern is undoubtably dependent upon the collision energy.38 Figure 7 shows the CID mass spectrum of the protonated molecular ion of bombesin, which is the highest mass peptide (MW = 1618.8) presented in this work. Despite the presence of (36) Schwartz, B. L.;Bursey, M. M. Biol. Mass Spectrom. 1992,21, 92. (37) Tomer, K. B.; Crow, F. W.; Gross, M. L.Anal. Chem. 1984,56, 880. (38) Bradley, C. D.; Curtis, J. M.; Demck, P. J.; Wright, B. Anal. Chem. 1992, 64, 2628.
Analytical Chemistty, Vol. 67, No. 13, July 1, 1995
2185
. n
1 2 3 4 5 6 7 8 8 1 0 1 1 1 2 1 3 1 4 84.0 zi2.i w . 2 481.3 w . 3 6u.4780.4 wo.5 i037.5ii98.t)ii83.ei330.7i443.01574.8 112.0 2 a . l 3W.2 509.3 E4.3 a . 3 991.5 lQ15511~4.81121.81958.71471.81802.8 H . pQlu. Gln. m . Lu. G i y - Am. Gln. Trp. Ala- Val. Gly. His. Lw. Met. NH, Yen 1 ~ . 8 1 ~ , 7 1 2 ? 4 . 6 1 1511 1W . 5 8 4 0 . 5 812.4 E S . 3 555.3 456.2 399.2 282.2 149.1 n i 3 1 2 i i i o 9 e 7 e 5 4 3 2 I b.
e
10
e
7
8
5
3
4
2
1
1 x40
-
I
I
x10
800
700
I
900
-
1000
xl
.-
I 1100
I 1200
I 1300
I
I
I
d z
I
1500
I
1600
l4O0
Figure 7. CID mass spectrum of protonated molecular ions of bombesin (MW = 1618.8). In this mass spectrum, the MH+ ions at m/z 1620.4 correspond to the 13C isotopic peak.
two basic residues, Arg and His, located near the amino and carboxy termini, respectively, the b, ions (which encompass the more basic Arg residue) are more often observed than f',, ions. a,, ions are also observed, as are extensive neutral losses (NHd and sidechain fragmentation. It is often the case for larger peptides that the most abundant fragments resulting from CID of the protonated molecular ion are those corresponding to small neutral losses. This, of course, limits the information that could be obtained from lower mass, sequencespecific fragmentation. In the CID mass spectrum of substance P (MW = 1346.7), for example, the most abundant fragment ion results from the loss of neutral NH3. In the experiment shown in Figure 8, we first isolated the protonated molecular ion cluster for substance P, activated the monoisotopic ion at m/z 1347.7, and then immediately activated the ion at m/z 1330.7 without isolating that ion. The composite CID spectrum shown in Figure 8 reveals ions resulting from the fragmentation of m/z 1347.7 and 1330.7. However, because of the similarity in the structures of these two precursors, the fragment ions can all be utilized to determine the amino acid structure. The presence of the basic residues, Arg and Lys, at the^ amino terminus results in a fragmentation pattern that consists primarily of a,, and b, ions, as well as a,, - NH3 and b, - NH3 ions arising from fragmentation of the ions at m/z 1330.7. In addition, y"~,y"9, and y"10 fragments reflect the presence of two proline residues. Figure 9 shows the CID mass spectra of the sodiumcationized "a+) molecules of Met-enkephalin (MW = 573.2) and Metenkephalinamide (MW = 572.2). In these spectra a,,,b,, and y", ions are replaced by their sodium cation analogs: a,,Na - H, b,,Na - H, and y',Na. Although these two peptides differ only in their Cterminal groups, there are considerable differences in their CID mass spectra. For example, the rearrangement ions b,Na + OH observed at m/z 261.3, 318.3, and 465.4 in the CID mass 2186 Analytical Chemistry, Vol. 67,No. 73,July 7, 7995
I
1000
I 1100
1050
I
lZb0
1150
1d50
L 1\ 1250
13bO
m/Z Figure 8. CID mass spectrum resulting from consecutive excitation of the monoisotopic protonated molecular ion at m/z 1347.7 of substance P (MW = 1346.7) and its most intense MS/MS fragment at m/z 1330.7. In this mass spectrum, the MH+ ions at m/z 1348.8 correspond to the 1 3 C isotopic peak. n 1 2 W a - H 158.1 215.1 b,NaH 188.1 243.1 H Tyr . Gly 433.2 n 4
596200
(a)
A
250
1
v,M
360
(b) 5% 2 on
.
Gly
.
Gty
.
319.1 2
5 550.2 578.2 Me1 . OH 172.1 1
4.b
2 215.1 243.1
.
432.2 4
.-a -
. Phe
378.1 3
460
350 n 1 M a - H 158.1 bJn-H 186.1 H Tyr
.
4 418.2 447.2
3 272.1 300.1
3 272.1 300.1
Gly
4 419.2 447.2
. Phe -
375.1 3
318.1 2
5bo 5 550.2 578.2 Me1 171.1 1
550
.
800
2g
NH,
'5
I
i
I
I
I
250
300
350
400
I
m/z
450
540
550
do
Figure 9. CID mass spectra of sodium-cationized molecules of (a) Met-enkephalin (MW = 573.2) and (b) Met-enkephalinamide (MW = 572.2). In these spectra, the MNa+ ions at m/z 597.6 and 596.7 correspond to the 1 3 C isotopic peaks, since the monoisotopic peaks have been removed by excitation.
spectrum of Metenkephalin (Figure 9a) are not observed in the CID mass spectrum of Met-enkephalinamide. These fragments are the formal equivalent to the b'6 + OH ion observed above for
825.4
h a
n 1 2 3 4 5 6 7 a W H 156.1 229.1 376.2 4332 696.3 693.3 780.3 b+-H 186.1 257.1 404.2 4812 624.3 721.3 808.3 H - T y r - A h - P h s - G l y - T y r - P r o - Ser-NH, 862.3 591.3 444.2 387.2 224.1 127.1 6 5 4 3 6 1
n
-c
. , 1 1 , I
I 500
I
4w
450
I
550
xl
I
I
600
650
700
r?%?
750
800
Figure 10. Tandem mass spectnrm of sodium-cationized molecules of dermorphin (MW = 802.4). In this mass spectrum, the MNa+ ions at m/z 826.4 correspond to the 13C isotopic peak.
acasein fragment 90-96. They have been observed in isolated cases on other mass analyzers,31~~~ while here they are observed as a series that can be used to recover the amino acid sequence. Also, in the CID mass spectrum of Metenkephalin, the intensities of fnNa fragments are considerably less than those of the b,Na OH ions. In the CID mass spectrum of Met-enkephalinamide (Figure 9b), the fragmentation pattern is considerably more complex and includes fragment ions formed by loss of small neutral species or side chains. Figure 10 shows that a similar result occurs in the CID mass spectrum of the MNa+ ion of Dermorphin (MW = 802.4), which also has a Gterminal amide group. In this spectrum, there is sufficient fragmentation information to recover the full amino acid structure.
+
SUMMARY AND CONCLUSIONS
The technique involving modulation of the resonant ejection voltage so that it is proportional to the amplitude of the fundamental rf voltage was reported previously as a means for improving (39) Mallis, L. M.; Russell, D. H. Anal. Chem. 1986,58, 1076. (40) Glish, G. L.;McLuckey, S. A; Goeringer, D. E Van Berkel, G. J.; Hart,IC J. Proceedings of the 39th ASMS Conference on Mass Spectrometry and Allied Topics: Nashville, TN, May 1991; p 536. (41) Orlando, R;Fenselau, C.; Cotter, R J.J. Am. SOC.Mass Spectrom. 1992.2, 189. (42) Julian, R IC; Reiser, H.:Cooks, R G. Int. J Mass Spectrom. Ion Processes 1993,123, 85. (43) Goeringer, D.E.;Asano, IC G.; McLuckey, S. A; Hoekman, D.; Stiller, S. W. Anal. Chem. 1994,66, 313. (44)Garrett, A W.; Cisper, M. E.; Nogar, N. S.; Hemberger, P. H. Rapid Commun. Mass Spectrom. 1994,8,174.
mass calibration by providing a linear dependence of the apparent mass shift on the scanned rf trapping v0ltage.2~In this work, we have extended the same modulation technique to provide accurate mass assignments (within 0.2-0.3 Da down to the lowest mass) in CID product ion mass spectra. Combined with the better than unit mass resolution observed in all CID spectra (with the exception of the high-pressure spectrum), these results form the elements of high-performance CID that is normally enjoyed only on tandem sector and Fourier transform mass spectrometers. Most signilicant is the fact that these results are obtained for ions formed by MALDI inside the ion trap. Such ions have a high initial kinetic energy distribution, but an approach using collisions with helium at elevated pressures to assist in trapping would also compromise both mass accuracy and resolution. Thus, the use of a ramped trapping field to improve trapping efficiencies at low helium pressuresz5is as essential for high activation selectivity and mass accuracy for CID mass spectra of MALDI ions as it has been for high performance in normal mass spectra.27 While the CID mass spectra show abundant amino acid sequence-specifk fragment ions, there are several problems in the general application of ion trap technology for peptide structure elucidation. Most prominent among these is the loss of the lower mass region. This problem is inherent in the ion trap, which is able to store only a finite region of mass for the product ions formed at the time of ion activation. In one case (gramicidin S), we addressed this problem by additional fragmentation of a prominent fragment ion of much lower mass. Alternatively, the CID mass range could be considerably extended using lower excitation frequencies. However, since the maximum energy associated with ion oscillatory motion is proportional to the square of the excitation this would substantially decrease CID efficiency, which might be increased by the use of heavier buffer molecules40or through chemically reactive collision^.^^ Additionally, there are a number of other techniques which, when combined with those described in this work, might signilicantly improve the performance of the ion trap mass spectrometer for peptide structural analysis using CID. For example, phase synchronization of the trapping and excitation voltages has been shown to improve the stability of mass calibration.42 Additionally, isolation of isotopic clusters using broad-band e x c i t a t i ~ nmight ~ ~ ~ be ~ ~combined with the selective activation and modulated ejection voltage methods described here. Finally, improvements in dynamic range might be achieved using selective injection of i0ns.2~9~3@ ACKNOWLEWMENT
This work was supported in part by grants from the National Science Foundation @IR 9016567) and National Institutes of Health (RO1 RR08912-01) and was carried out at the Middle Atlantic Mass Spectrometry Laboratory, an NSF Shared Instrumentation Facility. Received for review January 9, 1995. Accepted March 9, 1995.a AC950030Y @Abstractpublished in Advance ACS Abstracts, May 1, 1995.
Analytical Chemistty, Vol. 67, No. 13, July 1, 1995
2187
