Detection, Number, and Sequence Location of ... - ACS Publications

Center for Interdisciplinary Magnetic Resonance, National High Magnetic Field Laboratory, 1800 East Paul Dirac Drive,. Florida State University, Talla...
10 downloads 0 Views 166KB Size
Download PDF
Anal. Chem. 1997, 69, 1163-1168

Detection, Number, and Sequence Location of Sulfur-Containing Amino Acids and Disulfide Bridges in Peptides by Ultrahigh-Resolution MALDI FTICR Mass Spectrometry Touradj Solouki, Mark R. Emmett, Shengheng Guan,† and Alan G. Marshall*,†

Center for Interdisciplinary Magnetic Resonance, National High Magnetic Field Laboratory, 1800 East Paul Dirac Drive, Florida State University, Tallahassee, Florida 32310

Here, we present several strategies for determining the number of sulfur atoms and disulfide bridges in selected biologically active peptides, based on MALDI FTICR mass spectrometry at femtomole sample consumption level. First, based on the 2-Da mass increase per disulfide bridge reduction, we show that repeated laser shots on the same sample spot can reduce (and therefore reveal the presence of) the disulfide bridge in oxytocin. Second, we show that the primary sequence positions of the disulfide-bridged cystines can be inferred from the presence/absence of MALDI-induced reduction in cystinecontaining fragment ions. Third, we show that the presence and number of sulfur atoms as well as the degree of reduction in a peptide can all be determined directly from isotopic relative abundances of mass-resolved 34S, 13C2, and reduced all-12C species in a single ultrahigh-resolution MALDI FTICR mass spectrum. Methods for achieving such ultrahigh mass resolution of peptide ions of closely spaced m/z (m/∆m50% ≈ 950 000 at m/z ≈ 650) at modest magnetic field (3 T) are discussed. Disulfide bridge formation is a common posttranslational modification and a crucial step to achieve correct three-dimensional structure of peptides and proteins. To establish a structural basis for the biological activity of various sulfur-containing biomolecules, disulfide bonds must be identified and located in the primary amino acid sequence. Smith and Zhou described strategies for locating disulfide bonds in proteins;1 they pointed out that most traditional methods for detection of disulfide cross-linkages are inadequate to solve “most challenging problems”. Mass spectrometric instrumentation and techniques2 based on fast-atom bombardment (FAB) ionization,3 matrix-assisted laser desorption/ ionization (MALDI),4 and electrospray ionization (ESI)5 promise to provide the sensitivity and specificity needed for peptide and protein purification and detection at low (subpicomole) analyte levels. † Members of the Department of Chemistry, Florida State University. (1) Smith, D. L.; Zhou, Z. Methods Enzymol. 1990, 193, 374-389. (2) Burlingame, A. L.; Boyd, R. K.; Gaskell, S. J. Anal. Chem. 1996, 68, 599R651R. (3) Barber, M.; Bordoli, R. S.; Sedgwick, R. D.; Tyler, A. M. J. Chem. Soc., Chem. Commun. 1981, 325-327. (4) Karas, M.; Hillenkamp, F. Anal. Chem. 1988, 60, 2299-2301. (5) Fenn, J. B.; Mann, M.; Meng, C. K.; Wong, S. F.; Whitehouse, C. M. Science 1989, 246, 64-71.

S0003-2700(96)00885-2 CCC: $14.00

© 1997 American Chemical Society

Bortolini and Fogagnolo have surveyed mass spectral analysis of sulfur-containing compounds, including various naturally occurring and pharmaceutically significant products.6 Takao et al. demonstrated direct FAB mass analysis of disulfide-linked peptides in a complex mixture of a protein digest.7 Buko and Fraser examined factors that contribute to FAB mass spectra of disulfidecontaining peptides.8 Yazdanparast et al. showed in situ reduction of disulfide bonds during FAB-MS analysis of peptides9 and later demonstrated the utility of FAB-MS as a rapid and sensitive method for assignment of disulfide bonds in hen egg white lysozyme and bovine ribonuclease A.10 Svobda et al. used 2-mercaptoethanol as a disulfide bond reduction agent combined with ESI mass spectrometry to count the number of internal disulfide bridges in small proteins.11 Recently, Bencsath et al. demonstrated the advantages of FAB-MS to verify the existence of the disulfide bridge between Cys284 and Cys373 in irreversibly sickled cell β-actin.12 Prompt fragmentation of disulfide-linked peptides during MALDI-MS in time-of-flight (TOF) sources has been reported;13,14 however, confirmation of the peptide reduction and mass assignment with low mass resolving power MALDITOF instruments (m/∆m < 1000, in which m is ion mass and ∆m is the mass spectral peak width at a specified fraction of maximum peak height) required an additional analysis of the same peptides with ESI-MS. Coupling of a MALDI source with an FTICR mass analyzer15-25 offers potentially ultrahigh mass resolving power, precise mass measurement, and multistage MSn in a single instrument.22,26-37 (6) Bortolini, O.; Fogagnolo, M. Mass Spectrom. Rev. 1995, 14, 117-162. (7) Takao, T.; Yoshida, M.; Hong, Y.; Amioto, S.; Shimonishi, Y. Biomed. Mass Spectrom. 1984, 11, 549-556. (8) Buko, A. M.; Fraser, B. A. Biomed. Mass Spectrom. 1985, 12, 577-585. (9) Yazdanparast, R.; Andrews, P.; Smith, D. L.; Dixon, J. E. Anal. Biochem. 1985, 153, 348-353. (10) Yazdanparast, R.; Andrews, P. C.; Smith, D. L.; Dixon, J. E. J. Biol. Chem. 1987, 262, 2507-2513. (11) Svobda, M.; Meister, W.; Vetter, W. J. Mass Spectrom. 1995, 30, 15621566. (12) Bencsath, F. A.; Shartava, A.; Monteiro, C. A.; Goodman, S. R. Biochemistry 1996, 35, 4403-4408. (13) Patterson, S. D.; Katta, V. Anal. Chem. 1994, 66, 3727-3732. (14) Grimmins, D. L.; Saylor, M.; Rush, J.; Thoma, R. S. Anal. Biochem. 1995, 266, 355-361. (15) Hettich, R. L.; Buchanan, M. V. Int. J. Mass Spectrom. Ion Processes 1991, 111, 365-380. (16) Nuwaysir, L. M.; Wilkins, C. L. Proc. SPIE-Appl. Spectrosc. Mater. Sci. 1991, 112-123. (17) Solouki, T.; Russell, D. H. Proc. Natl. Acad. Sci. U.S.A. 1992, 89, 57015704. (18) Solouki, R.; Gillig, K. J.; Russell, D. H. Anal. Chem. 1994, 66, 1583-1587.

Analytical Chemistry, Vol. 69, No. 6, March 15, 1997 1163

FTICR mass spectrometers interfaced with MALDI,38,39 liquid secondary ion mass spectroscopy (SIMS) ion source,40 or ESI source41-44 have all been applied successfully to peptide sequence analysis. Simultaneous detection of all ions permits a significant reduction in sample consumption. Moreover, because a Penning trap is an ion storage device, it permits accumulation of ions produced in small abundance from a small amount of sample. In a previous publication, we used MALDI to generate peptide and protein ions for direct and rapid FTICR mass spectral accuratemass measurement and amino acid sequence analysis from relatively impure protein digest samples.45 In addition, we showed that “microsample” deposition onto a small indentation on the probe tip and ion multiple remeasurements combine to provide MALDI FTICR mass analysis from low attomole amounts of peptides and phospholipids.46 In this paper, we further optimize MALDI coupled with Fourier-transform ion cyclotron resonance mass spectrometry for rapid and sensitive determination of the presence and location of disulfide bonds in small peptides. Covalent disulfide bridges formed between cysteine residues in peptide/protein structure help to define tertiary conformation by cyclizing the molecule. This cyclization, in turn, directs (19) Castoro, J. A.; Ko ¨ster, C.; Wilkins, C. Rapid Commun. Mass Spectrom. 1992, 6, 239-241. (20) Buchanan, M. V.; Hettich, R. L. Anal. Chem. 1993, 65, 245A-259A. (21) Ko¨ster, C.; Castoro, J. A.; Wilkins, C. L. J. Am. Chem. Soc. 1992, 114, 75727574. (22) McIver, R. T., Jr.; Li, Y.; Hunter, R. L. Int. J. Mass Spectrom. Ion Processes 1994, 132, L1-7. (23) Yao, J.; Dey, M.; Pastor, S. J.; Wilkins, C. L. Anal. Chem. 1995, 67, 36383642. (24) Huang, Y.; Pasa-Tolic, L.; Guan, S.; Marshall, A. G. Anal. Chem. 1994, 66, 4385-4389. (25) Pasa-Tolic, L.; Huang, Y.; Guan, S.; Kim, H. S.; Marshall, A. G. J. Mass Spectrom. 1995, 30, 825-833. (26) Marshall, A. G. Adv. Mass Spectrom. 1989, 11A, 651-668. (27) Wanczek, K.-P. Int. J. Mass Spectrom. Ion Processes 1989, 95, 1-38. (28) Gord, J. R.; Freiser, B. S. Anal. Chim. Acta 1989, 225, 11-24. (29) Laude, D. A., Jr.; Hogan, J. D. Technisches Messen 1990, 57, 155-159. (30) Cody, R. B., Jr.; Bjarnason, A.; Weil, D. A. In Lasers in Mass Spectrometry; Lubman, D. M., Ed.; Oxford University Press: New York, 1990; pp 316339. (31) Marshall, A. G.; Grosshans, P. B. Anal. Chem. 1991, 63, 215A-229A. (32) Asamoto, B.; Dunbar, R. C. Analytical Applications of Fourier Transform Ion Cyclotron Resonance Mass Spectrometry; VCH: New York, 1991. (33) Campana, J. E. Proc. SPIE-Appl. Spectrosc. Mater. Sci. 1991, 138-149. (34) Marshall, A. G.; Schweikhard, L. Int. J. Mass Spectrom. Ion Processes 1992, 118/119, 37-70. (35) Ko ¨ster, C.; Kahr, M. S.; Castoro, J. A.; Wilkins, C. L. Mass Spectrom. Rev. 1992, 11, 495-512. (36) Jacoby, C. B.; Holliman, C. L.; Gross, M. L. In Mass Spectrometry in the Biological Sciences: A Tutorial; Gross, M. L., Ed.; Kluwer Academic Publishers: Dordrecht, 1992; pp 93-116. (37) Pastor, S. J.; Castoro, J. A.; Wilkins, C. L. Anal. Chem. 1995, 67, 379-384. (38) Solouki, T.; Russell, D. H. Appl. Spectrosc. 1993, 47, 211-217. (39) Castoro, A. C.; Wilkins, C. L.; Woods, A. S.; Cotter, R. J. J. Mass Spectrom. 1995, 30, 94-98. (40) Hunt, D. F.; Shabanowitz, J.; Yates, J. R.; Griffin, P. D.; Zhu, N. Z. In Mass Spectrometry of Biological Materials; McEwen, C. N., Larsen, B. S., Eds.; Marcel Dekker, Inc.: New York, 1990; pp 169-195. (41) Senko, M. W.; Speir, J. P.; McLafferty, F. W. Anal. Chem. 1994, 66, 28012808. (42) Beu, S. C.; Senko, M. W.; Quinn, J. P.; Wampler, F. M., III; McLafferty, F. W. J. Am. Soc. Mass Spectrom. 1993, 4, 557-565. (43) Wu, Q.; Van Orden, S.; Cheng, X.; Bakhtiar, R.; Smith, R. D. Anal. Chem. 1995, 67, 2498-2509. (44) Senko, M. W.; Hendrickson, C. L.; Pasa-Tolic, L.; Marto, J. A.; White, F. M.; Guan, S.; Marshall, A. G. Rapid Commun. Mass Spectrom. 1996, 10, 1824-1828. (45) Solouki, T.; Pasa-Tolic, L.; Jackson, G. S.; Guan, S.; Marshall, A. G. Anal. Chem. 1996, 68, 3718-3725. (46) Solouki, T.; Marto, J. A.; White, F. M.; Guan, S.; Marshall, A. G. Anal. Chem. 1995, 67, 4139-4144.

1164

Analytical Chemistry, Vol. 69, No. 6, March 15, 1997

ligand-receptor specificity and ultimate receptor response for many biologically active peptides. Here, we show that MALDIinduced fragmentation of disulfide-containing peptides provides a rapid means to establish the presence and primary amino acid sequence location of disulfide bridges in selected biologically active peptides. In addition, ultrahigh-resolution MALDI FTICR mass spectrometry makes possible the baseline resolution of species differing by 34S vs 13C2 for direct counting of sulfur atoms and disulfide bonds. In fact, we shall show that the degree of reduction and the number of sulfur atoms present in the molecule may be inferred from a single ultrahigh-resolution MALDI FTICR mass spectrum. Achievement of ultrahigh mass resolution requires elimination of space-charge-mediated coalescence of closely spaced cyclotron resonances47-49 and frequency drift during detection.50,51 We therefore demonstrate experimental procedures that minimize such effects. EXPERIMENTAL SECTION Sample Preparation. Peptides (oxytocin, isotocin, [D-Pen2,5]enkephalin) were purchased from Sigma Chemical Co. (St. Louis, MO) and used without further purification. Disulfide bonds were reduced with dithiothreitol (1% by volume) at pH ∼7 and 37 °C. 2,5-Dihydroxybenzoic acid (DHB) and D-fructose (Aldrich Chemical Co.) served as the matrix and comatrix, respectively, for all reported MALDI experiments. A 1 M stock solution of DHB matrix was prepared fresh daily in methanol acidified to a level of 0.1% (v/v) trifluoroacetic acid. MALDI FTICR mass spectra were obtained at a typical matrix:comatrix:analyte ratio of ∼1000: 500:1. Approximately 10 µL of the solution mixture containing analyte and matrix was applied to a solids insertion probe tip and allowed to dry in air before insertion into the mass spectrometer. Although MALDI FTICR-MS analysis in the attomole range is possible,46 here we applied a few hundred femtomoles of sulfurcontaining peptides to the solids probe tip so as to obtain several mass spectra from multiple laser irradiations of the same sample. FTICR Mass Spectrometer. Analyte solutions were massanalyzed with a 3-T Finnigan FTMS-2000 1.875-in.3 dual-trap FTICR mass spectrometer equipped with an Odyssey data system (Finnigan FTMS, Madison, WI). We performed desorption and ionization with a Laser Science: Inc., cartridge-type pulsed N2 laser (Model VSL-33ND, Laser Science, Inc., Newton, MA) operated at a wavelength of 337.1 nm with a pulse width of 3 ns. Additional experimental details and instrument configurations have been reported elsewhere.45,46 High-resolution and ultrahigh-resolution MALDI FTICR mass spectra were acquired in the analyzer compartment of the dual cubic ion trap. Other experiments, e.g., MALDI-induced ion fragmentation/disulfide bond reduction, were carried out in the source compartment. Ultrahigh-Resolution Mass Spectra. The experimental event sequence for ultrahigh-resolution experiments is shown in Figure 1. High kinetic energy positive ions are decelerated by biasing the sample probe, source trap plate, and conductance limit plate during the laser desorption event (60-150 ms) to 0, 0, and (47) Naito, Y.; Inoue, M. J. Mass Spectrom. Soc. Jpn. 1994, 42, 1-9. (48) Mitchell, D. W.; Smith, R. D. Phys. Rev. E 1995, 52, 4366-4386. (49) Peurrung, A. J.; Kouzes, R. T. Int. J. Mass Spectrom. Ion Processes 1995, 139-153. (50) Guan, S.; Wahl, M. C.; Marshall, A. G. Anal. Chem. 1993, 65, 3647-3653. (51) Bruce, J. E.; Anderson, G. A.; Hofstadler, S. A.; Winger, B. E.; Smith, R. D. Rapid Commun. Mass Spectrom. 1993, 7, 700-703.

Figure 1. Experimental event sequence for ultrahigh-resolution MALDI FTICR mass spectrometry from a single laser ionization event. First, laser-desorbed ions for the full m/z range pass into the source side compartment of the ICR dual cubic trap at an Ar background pressure of ∼2 × 10-7 Torr. After 60-150 µs to allow for ion transfer into the source trap, ion kinetic energy is nearly thermalized by ionneutral collisions. Following ion transfer (∼350 µs) into the lowpressure analyzer trap (∼1 × 10-9 Torr), the electrostatic trapping potential is lowered adiabatically to e+0.5 V. Selected parent ions are then isolated by radial dipolar ejection of of ions of undesired m/z values by high-resolution (124K data points) stored waveform inverse Fourier transform (SWIFT) broadband excitation. A long delay period (>100 s) before dipolar excitation/detection alleviates interaction of clouds of ions of closely spaced cyclotron frequencies (e.g., quasimolecular ions differing by 13C2 and 34S in oxytocin and [D-Pen2,5]enkephalin, (Figures 5 and 6). Conventional frequency-sweep excitation from 1 to 500 kHz at a 600 Hz/µs sweep rate is followed by heterodyne detection to yield 256K time domain data points. Each FTICR mass spectrum results from a single laser shot, based on direct FFT (and magnitude calculation) of a single time domain data acquisition.

+9 V, respectively. A relatively large number of parent ions can be generated and trapped by optimizing the matrix-to-analyte ratio and the deceleration potential52 that reduces parent ion kinetic energy and thereby increases trapping efficiency. After 60-150 µs to allow for ion transfer into the trap, the source trap and conductance limit plates are each restored to +2 V. A leak valve (Varian Associates, Walnut Creek, CA) maintains a constant neutral background pressure in the ion trap mass analyzer. A homebuilt relay circuit switches between dipolar excitation and detection modes to azimuthal quadrupolar axialization mode.53,54 Axialization is provided by a train of identical low-amplitude SWIFT55,56 excitations in the azimuthal quadrupolar mode at the cyclotron frequency of the ion(s) of interest in the presence of a collision gas (Ar at partial pressure of ∼2 × 10-7 in the source trap, corresponding to ∼2 × 10-9 Torr in the analyzer trap). Following ion transfer to the low-pressure analyzer trap, the electrostatic trapping potential is lowered adiabatically to e+0.5 V. The time delay (>100 s) before signal detection is lengthened to alleviate unwanted interactions between ion clouds with similar cyclotron frequencies (e.g., 13C2 and 34S isobars in oxytocin and [D-Pen2,5]enkephalin of Figures 5 and 6). Conventional frequency(52) Castoro, J. A.; Wilkins, C. L. Anal. Chem. 1993, 65, 2621-2627. (53) Speir, J. P.; Gorman, G. S.; Pitsenberger, C. C.; Turner, C. A.; Wang, P. P.; Amster, I. J. Anal. Chem. 1993, 65, 1746-1752. (54) Guan, S.; Marshall, A. G.; Wahl, M. Anal. Chem. 1994, 66, 1363-1367. (55) Marshall, A. G.; Wang, T.-C. L.; Ricca, T. L. J. Am. Chem. Soc. 1985, 107, 7893-7897. (56) Marshall, A. G.; Wang, T.-C. L.; Chen, L.; Ricca, T. L. In Fourier Transform Mass Spectrometry: Evolution, Innovation, and Applications; ACS Symposium Series 359; Buchanan, M. V., Ed.; American Chemical Society: Washington, DC, 1987; pp 21-33.

Figure 2. High-resolution analyzer trap MALDI FTICR mass spectra of quasimolecular [M + Na]+ oxytocin ions before (top) and after (bottom) reduction of the disulfide linkage by 1% dithiothreitol. The 2-u mass shift confirms the complete reduction of the disulfide bond. The disulfide bond also confers structural rigidity to the peptide, as can be seen from increased fragmentation (vertical bars) for the reduced peptide.

sweep excitation from 1 to 500 kHz at a 600 Hz/µs sweep rate is followed by heterodyne detection to yield 256K time domain data points, followed by Blackman-Harris apodization (no zero-filling). Each ultrahigh-resolution FTICR mass spectrum resulted from a single laser shot, based on direct FFT (and magnitude calculation) of the time domain data. RESULTS AND DISCUSSION Here we report MALDI FTICR mass spectral analyses of some selected disulfide-containing peptides. We chose oxytocin,57 a nonapeptide (CYIQNCPLGNH2), as a representative disulfidecontaining neurohypophyseal peptide hormone. Bis-penicillamine enkephalin, [D-Pen2,5]enkephalin, is a highly selective δ opioid receptor analog of enkephalin. The elevated δ opioid selectivity of [D-Pen2,5]enkephalin is attributed to the structure-rigidifying effect of the disulfide bond in penicillamine (β,β-dimethylcysteine) substituents (see below, Figure 6) that are substituted for cysteine in positions 2 and 5 of [Cys,Cys]enkephalin.58 A primary aim of this study is to develop a rapid and sensitive method to identify disulfide bridges in small peptides. MALDI-induced fragment ion mass spectra and ultrahigh-resolution mass spectra of selected disulfide-containing compounds in both reduced and native forms are presented. High-Resolution MALDI FTICR Positive-Ion Mass Spectra from Native and Reduced Forms of Oxytocin. Figure 2 shows high-resolution analyzer trap MALDI FTICR mass spectra of the quasimolecular ([M + Na]+) region of oxytocin (CYIQNCPLGamide, MW ) 1006.44) before (top) and after (bottom) reduction of the disulfide linkage by 1% dithiothreitol. We observed abundant [M + Na]+ signals for all of the disulfide-containing neuropeptides such as somatostatin, isotocin, oxytocin, and vasopressin that we (57) Inoue, T.; Kimura, T.; Azuma, C.; Inazawa, J.; Takemura, M.; Kikuchi, T.; Kubota, Y.; Ogita, K.; Saji, F. J. Biol. Chem 1994, 269, 32451-32456. (58) Mosberg, H. I.; Hurst, R.; Hruby, V. J.; Gee, K.; Yammamura, H. I.; Galligan, J. J.; Burk, T. F. Proc. Natl. Acad. Sci. U.S.A. 1983, 80, 5871-5874.

Analytical Chemistry, Vol. 69, No. 6, March 15, 1997

1165

Figure 3. MALDI FTICR mass spectra (analyzer trap) of quasimolecular [M + Na]+ oxytocin ions from a single laser shot (top) and from 100 laser shots averaged (bottom). FTICR MS resolves the 13C isotopic distribution of the quasimolecular ions of oxytocin, matching the calculated isotopic relative natural abundances for the single laser shot spectrum (top) to within ∼5%. However, subsequent laser shots on the same sample spot (bottom) induce disulfide bond reduction, as evidenced by increase abundance at 2 Da above the monoisotopic mass (see text).

analyzed in our study (only the oxytocin data are presented here); alkali metal adduct ions did not interfere with our experiments, and hence we did not attempt to purify the samples prior to MALDI analyses. However, when necessary, we selectively isolated [M + H]+, [M + Na]+, or [M + K]+ ions (e.g., see below, Figures 5 and 6). The 2-u mass shift in the bottom spectrum confirms the complete reduction of the disulfide bond. The disulfide bond also confers structural rigidity to the peptide, as demonstrated by peptide fragmentation (see below, Figures 3 and 4). For example, the vertical bars drawn across the peptide sequence denote the most prominent fragmentation sites resulting from MALDI for native (Figure 2, top) and reduced (Figure 2, bottom) forms of oxytocin. Briefly, loss of the NH2 end group and the PLG segment to produce B6 and Y3 fragment ions are the most prominent fragmentations observed in native oxytocin (Figure 2, top), whereas the fragment ions Y3, B5, B6, B7, and B8 are observed in the mass spectrum of reduced oxytocin. A true mass resolving power of at least m/∆m ≈ 95 000 is required in order to separate 34S (m/z ) 1031.421 472) and 13C2 (m/z ) 1031.432 386) peaks in Figure 2; however, peak coalescence47-49 and frequency drift during detection50,51 make peak separation of coupled “gyrators” difficult. For example, the 34S and 13C2 ion cyclotron resonances shown in Figure 2 exhibit m/∆m50% > 100 000 but are, nevertheless, coupled; hence, single resonances are observed at m/z ) 1031.4 (Figure 2, top) and 1033.4 (Figure 2, bottom). We shall present experimental parameters and considerations (see below, Figures 5 and 6) to reduce peak coalescence and to enhance the performance of FTICR for ultrahigh-resolution mass spectrometry. High-Resolution MALDI FTICR Positive-Ion Mass Spectra of Oxytocin from a Single Laser Shot vs 100 Laser Shots Averaged: Identification of Sulfur-Containing Residues on the Basis of Change in Isotopic Relative Abundances. Figure 3 shows high-resolution analyzer side positive-ion magnitude-mode MALDI FTICR molecular ion region mass spectra of native oxytocin from a single laser shot (top) and 100 laser shots averaged (bottom). The theoretically calculated isotopic pattern for the [M + Na]+ region of oxytocin is included for comparison 1166 Analytical Chemistry, Vol. 69, No. 6, March 15, 1997

Figure 4. Low-resolution MALDI FTICR mass spectrum (source trap) of native oxytocin from time domain data obtained by co-adding signals from 100 laser shots (bottom). Insets show m/z-expanded regions for B8, B6, and Y3 fragment ions for native (bottom inset) and reduced (top inset) forms of oxytocin. Clearly, some disulfide bonds are reduced in situ during the MALDI process (elevated relative abundance, deanoted by asterisk, at 2 Da above the monoisotopic mass). Comparison between reduced and nonreduced MALDI FTICR mass spectra serves to identify disulfide bonds and locate them in the primary amino acid sequence (see text).

(top). Within 5% experimental error, the observed pattern for isotopic distribution of ions produced from a single laser shot near m/z ≈ 1030 agrees with the theoretical isotopic distribution. However, after averaging signals from 100 subsequent laser shots (on the same laser spot), the relative abundance at the mass of the [M + Na]+ 13C2 isotopic species at m/z ) 1031.4 as well as for the [M + H]+ species (both denoted by asterisks) increases. (Yazdanparast et al. have previously reported in situ reduction of disulfide bonds in peptides by fast-atom bombardment ionization.9) Thus (see below, Figure 4), MALDI-induced in situ disulfide bond reduction as a rapid method provides a means to verify the presence and location of disulfide bridges in peptides. Figure 4 shows a low-resolution MALDI FTICR mass spectrum (source trap) of native oxytocin from time domain data consisting of co-added signals from 100 laser shots. Although higher quality mass spectra (in terms of high resolution) may be obtained in the low-pressure analyzer trap, mass discrimination due to ion transfer is minimized (and isotopic relative abundances are more accurately represented) in the source trap. The broadband mass spectrum (Figure 4, bottom) shows laser-induced metastable fragment ions of native oxytocin. Ions were detected following a 2-s delay from the laser irradiation event. Insets show m/zexpanded regions for B8-17, B6, and Y3 fragment ions for native (bottom inset) and corresponding reduced (top inset) forms of oxytocin. Metastable ion fragmentation is enhanced at long time delays (Solouki, 1994 #1162; Castoro, 1995 #1275); however, even at a very short time delay after the laser desorption event (∼100 µs), we observed ion fragmentation in multiple laser shot experiments. After a 2-s time delay, almost all of the ion fragmentation processes subside, and intact molecular and fragment ions relax to the center of the ICR cell, as is desirable for ICR signal detection. The top insets were obtained in a similar experiment in which the reduced form of oxytocin was analyzed. Clearly, some disulfide bonds are reduced in situ during the MALDI process, and relative abundance at 2 u higher than the monoisotopic peak (see asterisk in bottom insets) is elevated. For example, peaks at m/z ) 916.4 and 918.4 (lower inset) correspond

Figure 5. Ultrahigh-resolution MALDI FTICR mass spectrum (analyzer trap) of native oxytocin from a single laser shot (bottom). Inset shows m/z-expanded region for 34S and 13C2 isotopic peaks along with calculated ion relative natural abundances. Based on accurate mass measurement and comparison of observed and calculated ion relative abundances, the number of disulfide bonds can be determined (see text).

to all-12C and 13C2 species of B8-17; laser-induced in situ reduction of the disulfide bridge in oxytocin increases the ion relative abundance at 918.4. The B8-17 region of chemically reduced oxytocin (top inset) exhibits the theoretically expected isotopic pattern for its fragment ions. The elevated 13C2 abundance in nonreduced oxytocin (lower inset) not only verifies the presence of a disulfide bridge but also signifies that this fragment ion contains two sulfurs, i.e., the cysteines. Moreover, peaks at m/z ) 723.3 and 725.3 (lower inset) correspond to all-12C and 13C2 of the B6 fragment ion, respectively. Due to laser-induced in situ reduction of the disulfide bridge in oxytocin, the relative ion abundance at m/z ) 725.3 (bottom inset) is elevated, whereas the B6 fragment ion region of chemically reduced oxytocin (top inset) shows the theoretically expected isotopic pattern. The observed isotopic pattern for native (bottom inset) and chemically reduced (top inset) B6-17 fragment ion is consistent with the laserinduced in situ reduction of the disulfide bridge. Again, the elevated 13C2 abundance not only verifies the presence of a disulfide bridge but also signifies that this fragment ion (CYIQNC) contains two sulfurs, i.e., the cysteines. Finally, the top and bottom insets show Y3′′ fragment ions at m/z ) 285.19 of chemically reduced and native oxytocin. We may compare the experimentally observed ion abundance with theoretically calculated isotopic ratios to estimated the relative amount of disulfide bridge reduction during the MALDI process. The isotopic patterns for Y3′′ fragment ions from reduced (top inset) and native (bottom inset) oxytocin are identical; thus, the observed fragment ion (PLG) contains no sulfurs. Thus, comparison between the reduced and nonreduced MALDI FTICR mass spectra helps to identify and locate the disulfide bonds. Identification and Direct Counting of Sulfur Atoms from Ultrahigh-Resolution MALDI FTICR Positive-Ion Mass Spectra. Oxytocin. Figure 5 shows an ultrahigh-resolution, m/∆m50% ≈ 500 000, MALDI FTICR mass spectrum (analyzer trap) of native oxytocin from a single laser shot (bottom) at 3 T. The inset shows the m/z-expanded segment for 34S and 13C2 isotopic peaks along with the calculated ion isotopic relative abundances. Based on accurate mass measurement and relative ion abundances (theoretical versus experimental), the number of disulfide bonds can

Figure 6. Ultrahigh-resolution MALDI FTICR mass spectrum (analyzer trap) of native and reduced [D-Pen2,5]enkephalin from a single laser shot (top). Expanded m/z region (bottom) shows fully mass-resolved nonreduced 34S, 13C134S, and 13C2 and reduced all12C and 13C resonances. 1

be determined. For example, native oxytocin with two sulfur atoms should exhibit 34S at m/z ) 1031.421 472 (8.86%) and 13C2 at m/z ) 1031.432 386 (11.17%), and the experimental abundances agree to within 10%. The mass measurement accuracy with external calibration for all of the high-resolution mass spectra was within 10 ppm of theoretical values. In Figures 5 and 6, we used all-12C and 13C1 peaks as internal calibrants, and the mass measurement accuracy for all other peaks (including the massresolved 34S peaks) improved to within 1 ppm of their theoretical values. The presence of sulfur atoms in the molecule is established by the observation of mass-resolved peaks at m/z ) 1031.421 472 and 1031.432 386. The number of sulfur atoms may be determined by comparing the relative abundance ratio for 34S m/z ) 1031.421 472 to the monoisotopic abundance (32S) and to the theoretical natural abundance ratio of 34S to 32S (i.e., 4.215%/ 95.785% ) 0.044 per sulfur atom in the molecule). Alternatively, if the number of carbon atoms in the molecule is known (even approximately), then the number of sulfurs can be estimated from the relative abundance of 34S vs 13C2. Several authors47-49 have introduced expressions predicting the threshold for coupling of two packets of ions of nearly the same m/z value. Guided by those criteria, we have succeeded in obtaining ultrahigh mass resolution of closely spaced resonances by a combination of low trapping voltage (50 s) for ion equilibration before excitation/detection, and optimized ion excitation amplitude, duration, and bandwidth. The most important factor for reducing cyclotron frequency shift and ion coupling was the use of a long delay period at low trapping voltage (see below, Figure 6). Presumably, ions with high kinetic energy either leave the ICR trap or collisionally relax. Adiabatic lowering of the dc trapping potential effectively spreads ions over a longer axial region, thereby decreasing trapped-ion density and reducing “space-charge” effects (peak coalescence, frequency drift during detection). [D-Pen2,5]Enkephalin (DPDPE). Figure 6 shows an ultrahighresolution single laser shot positive-ion MALDI FTICR magnitudemode mass spectrum of partially reduced [D-Pen2,5]enkephalin (DPDPE). [D-Pen2,5]enkephalin is a highly selective δ opioid receptor analog of enkephalin, in which penicillamine (β,βdimethylcysteine) substituents have been substituted for cysteines Analytical Chemistry, Vol. 69, No. 6, March 15, 1997

1167

in positions 2 and 5 of [Cys,Cys]-enkephalin.58 In Figure 6 (bottom), m/z scale expansion shows mass-resolved nonreduced 34S, 13C 34S, and 13C and reduced all-12C and 13C peaks. It is worth 1 2 1 noting that less than 10 fmol of sample was consumed in the generation of Figure 6. Following ion transfer to the low-pressure analyzer trap (transfer time 260 µs), the electrostatic trapping potential was lowered adiabatically to 0.4 V (25, 10, 10, and 40 s at 2.0, 1.5, 1.0, and 0.4 V, respectively). A high-resolution SWIFT waveform served to isolate ions from 644 e m/z e 650. Conventional frequency-sweep excitation from 1 to 500 kHz at a 600 Hz/ µs sweep rate was followed by heterodyne detection to yield 256K time domain data points (60-s transient image current signal). The displayed FTICR mass spectrum resulted from a single laser shot, based on direct FFT (and magnitude calculation of the first half) of the time domain data. The resolving power (m/∆m50%) at m/z ≈ 648.232 is ∼936 000. The relative abundances of theoretically calculated 34S (m/z ) 648.232) and 13C2 (m/z ) 648.243) in DPDPE are 8.86% and 5.38%, respectively, and the observed ion relative abundances agree to within 10%. Note that two peaks, 13C134S from nonreduced (m/z ) 649.235) and 13C1 from reduced (m/z ) 649.255) forms of DPDPE are also separated! We calculate that ∼12% of the disulfide bridges are reduced by comparing the observed ion relative abundances of all-12C peaks of reduced and nonreduced [D-Pen2,5]enkephalin; we obtain the same value by comparing the reduced 13C and nonreduced 13C34S relative abundances. Figure 6 constitutes the highest mass resolution yet obtained for a peptide of this size by MALDI FTICR at 3 T. CONCLUSIONS: ADVANTAGES AND LIMITATIONS Here, we have shown how to analyze disulfide-containing peptides directly by MALDI and ultrahigh-resolution FTICR mass spectrometry. MALDI-induced fragmentation provides a rapid means to verify the presence and location of disulfide bridges in biologically active peptides. In addition, complementary ultrahighresolution MALDI FTICR mass spectra with mass-resolved 34S and 13C2 species establishes the presence of disulfide bonds as well as the number of sulfurs in the molecule (Figure 5). The

1168

Analytical Chemistry, Vol. 69, No. 6, March 15, 1997

number of sulfurs present in the molecule and the degree of reduction may be inferred from a single ultrahigh-resolution MALDI FTICR mass spectrum (Figure 6). We have not yet investigated the effects of MALDI matrix and other parameters, e.g., laser power, laser pulse duration, etc., on degree of in situ disulfide bridge reduction. The extent of laserinduced disulfide bond reduction appears to depend on the analyte; e.g., significant in situ disulfide bond reduction of isotocin was not observed, in contrast to relatively efficient laser-induced disulfide bridge reduction of oxytocin. (Similar variations have previously been noted for FAB by Yazdanparast et al.9) The extent of laser-induced fragmentation depends on laser power, and it is possible to adjust the laser power and control the degree of ion fragmentation: at low laser power (threshold laser power density for ion formation), the number and abundance of fragment ions are significantly reduced. In any case, our present results clearly establish the prospect of structural characterization of disulfidecontaining peptides and proteins by MALDI FTICR mass spectrometry. To our knowledge, the results in Figures 5 and 6 represent the highest resolution (i.e., actual separation of closely spaced resonances, rather than ultranarrow line width of isolated or coalesced resonances) obtained in a 3 T MALDI FTICR experiment. The present strategies should make possible rapid detection of posttranslational disulfide bridge formation from femtomole quantities of peptides, a capability of high relevance to in vitro and in vivo pharmacology of ligand-receptor interactions in biological systems. ACKNOWLEDGMENT This work was supported by NSF (CHE-93-22824), NIH (GM31683), Florida State University, and the National High Magnetic Field Laboratory in Tallahassee, FL. Received for review September 3, 1996. January 15, 1997.X

Accepted

AC960885Q X

Abstract published in Advance ACS Abstracts, February 15, 1997.