Consecutive Ion Activation for Top Down Mass Spectrometry

Huili Zhai, Xuemei Han, Kathrin Breuker, and Fred W. McLafferty* ... Udseth, and Smith13-16 (also “cone voltage fragmentation”, “source-CID”),...
0 downloads 0 Views 316KB Size
Anal. Chem. 2005, 77, 5777-5784

Accelerated Articles

Consecutive Ion Activation for Top Down Mass Spectrometry: Improved Protein Sequencing by Nozzle-Skimmer Dissociation Huili Zhai,† Xuemei Han,† Kathrin Breuker,‡ and Fred W. McLafferty*,†

Department of Chemistry and Chemical Biology, Baker Laboratory, Cornell University, Ithaca, New York 14853, and the Institute of Organic Chemistry and Center for Molecular Biosciences Innsbruck (CMBI), University of Innsbruck, Innrain 52a, 6020 Innsbruck, Austria

Mass spectra produced by nozzle-skimmer dissociation (NSD) have been little used in the past for structural characterization. NSD cannot be used on mass-separated ions (MS/MS), and for electrosprayed protein ions, previous NSD spectra showed backbone cleavages similar to those from energetic methods such as collisionally activated dissociation (CAD) or infrared multiphoton dissociation (IRMPD). However, our experimental configuration with Fourier transform (FT) MS makes possible three consecutive steps of NSD ion activation: thermal in the entrance capillary and collisional in both the nozzle-skimmer (N-S) region and the region after the skimmer before the quadrupole entrance lens (S-Q). In the high-pressure N-S region of adjustable path length, ions undergo high-frequency, low-energy collisions to rupture weak noncovalent or covalent bonds, with these “denatured” products then subjected to high-energy collisions in the low-pressure S-Q region to cleave strong backbone bonds. These NSD spectra, plus those from variable capillary thermal activation, of 8+ to 11+ ubiquitin ions electrosprayed from denatured solution show backbone cleavages between 74 of 75 amino acid pairs, vs 66 for CAD and 50 for IRMPD in the FTMS cell. Thermal activation by the inlet capillary of the newly desolvated 6+, 7+ ubiquitin ions from electrospraying the native conformer increases the NSD yield from 8% at 56 °C to 96% at 76 °C, but with little change in product branching ratios; this capillary heating has no effect on CAD or IRMPD of these ions collected in the FTMS cell. * Corresponding author. E-mail: [email protected]. † Cornell University. ‡ University of Innsbruck. 10.1021/ac0580215 CCC: $30.25 Published on Web 08/20/2005

© 2005 American Chemical Society

Ion desolvation with its concomitant H-bond strengthening appears to produce a transiently stable conformer whose formation can be prevented by capillary heating. The far more complex and stable noncovalent tertiary structures of large protein ions in the gas phase have made MS/MS difficult; initial inhibition of tertiary structure formation with immediate NSD (“prefolding dissociation”) appears promising for the top down characterization of a 200-kDa protein. “Top down” tandem mass spectrometry (MS) is a powerful new method for identification and structural characterization of proteins, especially when utilizing the unique resolving power of Fourier transform (FT) MS.1,2 This separates (M + nH)n+ molecular ions from electrospray ionization (ESI) of a protein mixture and produces accurate molecular weight (Mr) values; next, isolation of a protein’s (M + nH)n+ ions and their backbone dissociation (MS/MS) yields fragment ion masses that provide detailed information on that protein’s sequence and posttranslational modifications.1-6 The most common MS/MS techniques for ion fragmentation7 are collisionally activated dissociation (1) Kelleher, N. L.; Lin, H. Y.; Valaskovic, G. A.; Aaserud, D. J.; Fridriksson, E. K.; McLafferty, F. W. J. Am. Chem. Soc. 1999, 121, 806-812. (2) Kelleher, N. L. Anal. Chem. 2004, 76, 197A-203A. (3) Ge, Y.; Lawhorn, B. G.; ElNaggar, M.; Strauss, E.; Park, J.-H.; Begley, T. P.; McLafferty, F. W. J. Am. Chem. Soc. 2002, 124, 672-678. (4) Park, J.-H.; Dorrestein, P. C.; Zhai, H.; Kinsland, C.; McLafferty, F. W.; Begley, T. P. Biochemistry 2003, 42, 12430-12438. (5) Zabrouskov, V.; Giacomelli, L.; van Wijk, K. J.: McLafferty, F. W. Mol. Cell. Proteomics 2003, 2, 1253-1260. (6) Meng, F.; Du, Y.; Miller, L. M.; Petrie, S. M.; Robinson, D. E.; Kelleher, N. L. Anal. Chem. 2004, 76, 2852-2858. (7) Gabelica, V.; De Pauw, E. Mass Spectrom. Rev., 2005, 24, 566-587.

Analytical Chemistry, Vol. 77, No. 18, September 15, 2005 5777

Figure 1. Schematic diagram of the ESI/FTMS ion introduction region in which NSD takes place.

(CAD)8 using sustained off-resonance irradiation (SORI),9 infrared multiphoton dissociation (IRMPD),10 and electron capture dissociation (ECD).11,12 A special type of CAD, named “nozzleskimmer dissociation” (NSD) by its discoverers Loo, Udseth, and Smith13-16 (also “cone voltage fragmentation”, “source-CID”),7 accelerates the ions exiting the ESI capillary (Figure 1) to induce CAD in the partially reduced pressure region (a related method is “multipole storage-assisted dissociation”).17,18 However, NSD applications19-25 are rare, as CAD and IRMPD have generally been more informative, and molecular ions of a mixture cannot be MS separated before dissociation (MS/MS). Although energetic MS/ MS methods nominally cleave the weakest backbone bonds, here we show that NSD information can be greatly increased using three consecutive ion activation steps that cause separate primary and secondary dissociations of noncovalent or covalent bonds and that even prevent folding of the newly formed gaseous ions from ESI. Two consecutive excitation steps with gaseous ions reaching the FTMS cell were critical for “activated ion ECD”26 and “plasma ECD”;27 the first of these collisionally dissociates the already formed noncovalent tertiary structure of the ions, making possible ECD in the second step. (8) Senko, M. W.; Speir, J. P.; McLafferty, F. W. Anal. Chem. 1994, 66, 28012808. (9) Gauthier, J. W.; Trautman, T. R.; Jacobson, D. B. Anal. Chim. Acta 1991, 246, 211-225. (10) Little, D. P.; Speir, J. P.; Senko, M. W.; O’Connor, P. B.; McLafferty, F. W. Anal. Chem. 1994, 66, 2809-2815. (11) Zubarev, R. A.; Kelleher, N. L.; McLafferty, F. W. J. Am. Chem. Soc. 1998, 120, 3265-3266. (12) Zubarev, R. A.; Horn, D. M.; Fridriksson, E. K.; Kelleher, N. L.; Kruger, N. A.; Lewis, M. A.; Carpenter, B. K.; McLafferty, F. W. Anal. Chem. 2000, 72, 563-573. (13) Loo, J. A.; Udseth, H. R.; and Smith, R. D. Rapid Commun. Mass Spectrom. 1988, 2, 207-210. (14) Loo, J. A.; Edmonds, C. G.; Smith, R. D. Science 1990, 248, 201-204. (15) Loo, J. A.; Edmonds, C. G.; Smith, R. D. Anal. Chem. 1991, 63, 24882499. (16) Loo, J. A.; Edmonds, C. G.; Smith, R. D. Anal. Chem. 1993, 65, 425-438. (17) Sannes-Lowery, K.; Griffey, R. H.; Kruppa, G. H.; Speir, J. P.; Hofstadler, S. A. Rapid Commun. Mass Spectrom. 1998, 12, 1957-1961. (18) Keller, K. M.; Brodbelt, J. S.; Hettich, R. L.; Van Berkel, G. J. J. Mass Spectrom. 2004, 39, 402-411. (19) Loo, J. A.; Quinn, J. P.; Ryu, S. I.; Henry, K. D.; Senko, M. W.; McLafferty, F. W. Proc. Natl. Acad. Sci. U.S.A. 1992, 89, 286-289. (20) Senko, M. W.; Beu, S. C.; McLafferty, F. W. Anal. Chem. 1994, 66, 415417. (21) Speir, J. P.; Senko, M. W.; Little, D. P.; Loo, J. A.; McLafferty, F. W. J. Mass Spectrom. 1995, 30, 39-42. (22) Hakansson, K.; Zubarev, R.; Hakansson, P. Rapid Commun. Mass Spectrom. 1998, 12, 705-711. (23) Kelleher, N. L.; Taylor, S. V.; Grannis, D.; Kinsland, C.; Chiu, H.-J.; Begley, T. P.; McLafferty, F. W. Protein Sci. 1998, 7, 1796-1801. (24) Chen, H.; Tabei, K.; Siegel, M. M. J. Am. Soc. Mass Spectrom. 2001, 12, 846-852. (25) Ginter, J. M.; Zhou, F.; Johnston, M. V. J. Am. Soc. Mass Spectrom. 2004, 15, 1478-1486. (26) Horn, D. M.; Ge, Y.; McLafferty, F. W. Anal. Chem. 2000, 72, 4778-4784.

5778

Analytical Chemistry, Vol. 77, No. 18, September 15, 2005

As reviewed in detail recently by Gabelica and De Pauw,7 NSD takes advantage of the ESI apparatus requirement of transporting the ions from atmospheric pressure through a capillary “nozzle” into a first region of differential pumping in which the ions undergo many collisions before exiting through the central hole in a conical “skimmer” (Figure 1). An accelerating voltage gradient between the nozzle and skimmer (VN-S) provides kinetic energy for collisional activation of the incoming ions, with lower VN-S values dissociating noncovalent complexes (e.g., removing solvent molecules)28 not dissociated by capillary heating;29 higher VN-S values cause covalent bond dissociation.13-25 Here we compare the effect of VN-S and the additional parameters (Figure 1) skimmer-quadrupole potential (VS-Q), nozzle--skimmer distance (mmN-S), and capillary heating current (Acap) on NSD of both high- and low charge-state ions from the ESI of the protein ubiquitin, for which extensive MS/MS spectra have been reported from NSD,16,19 CAD,8,9,30 IRMPD,10 and ECD.31-33 These new NSD variables provide information sufficient to sequence de novo all except 1 pair of the 76 amino acids of ubiquitin. Also, folding of the newly formed gaseous 6+, 7+ ubiquitin ions can be delayed by capillary heating to increase the NSD yield from 8 to 96%; dissociation of these ions by CAD after trapping in the FTMS cell is unaffected by this capillary heating. EXPERIMENTAL SECTION Bovine ubiquitin and other chemicals were purchased from Sigma-Aldrich (St. Louis, MO) and used without further purification. Electrospray (∼+1 kV) with a nanospray emitter at 1-50 nL/min of 20 µM ubiquitin in MeOH/H2O/AcOH (50:45:5, denaturing) solution yielded “high-charge-state” (mainly 8+ to 11+) ions, and in Nanopure H2O (pH 5, stable native conformer) yielded “low-charge-state” (mainly 6+ and 7+) ions. Although the molecular ion abundances as a function of charge state could vary substantially,7 this had little qualitative effect on the spectra. The ions were guided through a heated capillary, skimmer, ion lens, and a 26-cm radio frequency-only quadrupole, exiting through a grounded orifice into the next differentially pumped region (Figure 1). The stainless steel capillary is resistively heated by a current Acap, with its temperature measured by a thermocouple on its outer surface. In the region between the capillary exit “nozzle” and skimmer (N-S region, adjustable length mmN-S), ions are accelerated by the potential difference VN-S, with ion focusing by an electrostatic mesh tube lens set at its maximum 250 V. The potential between the skimmer and the ion lens at the front of the quadrupole (S-Q region) accelerates and focuses the ions and screens them from the quadrupole rf field. The potential on (27) Sze, S. K.; Ge, Y.; Oh, H. B.; McLafferty, F. W. Anal. Chem. 2003, 75, 1599-1603. (28) Rogniaux, H.: Van Dorsselaer, A.; Barth, P.; Bjellmann, J. F.; Brabanton, J.; van Zandt, M.; Chevrier, B.; Howard, E.; Mitschler, A.; Potier, N.; Urzhumtseva, L.; Moras, D.; Podjarny, A. J. Am. Soc. Mass Spectrom. 1999, 10, 635-647. (29) Penn, S. G.; He, F.; Green, M. K.; Lebrilla, C. B. J. Am. Soc. Mass Spectrom. 1997, 8, 244-252. (30) Reid, G. E.; Wu, J.; Chrisman, P. A.; Wells, J. M.; McLuckey, S. A. Anal. Chem. 2001, 73, 3274-3281. (31) Horn, D. M.; Zubarev, R. A.; McLafferty, F. W. Proc. Natl. Acad. Sci. U.S.A. 2000, 97, 10313-10317. (32) Breuker, K.; Oh, H. B.; Horn, D. M.; Cerda, B. A.; McLafferty, F. W. J. Am. Chem. Soc. 2002, 124, 6407-6420. (33) Oh, H. B.; Breuker, K.; Sze, S. K.; Ge, Y.; Carpenter, B. K.; McLafferty, F. W. Proc. Natl. Acad. Sci. U.S.A. 2002, 99, 15863-15868.

this lens and the similar potential on the quadrupole are adjusted separately for maximum transmittance, but both values are 11 ( 1 V above the grounded quadrupole exit, so this value is called VS-Q. The N-S region is evacuated by a mechanical rotary pump and the S-Q-quadrupole region by a diffusion pump. After the first quadrupole exit, the ions pass through two separately pumped rf-only quadrupoles for gas pulse-assisted trapping in the ion cell of a 6-T Finnigan FTMS, resolving power 105.1,3 IRMPD utilized a 27-W Synrad (Bothell, WA) CO2 IR laser with a 60-80-ms pulse.10 CAD/SORI was applied 1.5-kHz off-resonance for 1.5 s at 28.5 (6+) to 34 (9+-11+) dB (8.4-4.5 V peak-to-peak).8 Spectra were recorded using the Odyssey data system and represent averages of 15-30 scans, with fragment ion masses and compositions assigned with the computer program THRASH and reported using its confidence level requirements.34 RESULTS AND DISCUSSION “Energetic” methods for MS/MS of protein ions, such as CAD (including NSD) and IRMPD, add internal energy to cleave a backbone amide bond and produce a complementary pair of b and y ions containing the N- and C-termini, respectively, of the protein. The further dissociation of either primary product ion at a CO-NH bond yields an internal i ion (in the spectra represented here, essentially all fragment ions identified by THRASH34 could be assigned by it as b, y, or i ions or their H2O loss products). In contrast to b or y mass values, an i mass value is based on two different bond cleavages, so that the value often corresponds to more than one internal sequence of a predicted protein (e.g., for Gly10-XXXX-Gly15, i10-14 will have the same mass value as i11-15). In this study, a mass value not predicted as a b or y ion is only assigned to an i ion if that ion is a possible dissociation product of an identified b or y ion (e.g., an appropriate mass is assigned as i18-52 only if one, not both, of the cleavage sites 18 or 52 has been identified; these are classified as ib or iy products, respectively). Most previous ubiquitin MS/MS studies have not reported internal i ions or have deemphasized them; note that individual ECD spectra of ubiquitin can identify all but 1 of its 75 cleavages using only terminal fragment ions.27 Comparison of spectra will be based on the number of interresidue cleavages shown of the 75 possible, as this directly reflects how closely, on average, a posttranslational modification can be located on the amino acid backbone.1-6 If a cleavage is indicated by both b and y ions, each is credited with 0.5 cleavages; only additional cleavages are credited for ib and iy ions. The relative degree of dissociation of the primary ions is indicated by the [y]/[b] value; b ions are generally less stable than y ions.8,10,35,36 The high (8+-11+) and low (6+, 7+) charge-state ions, those formed by ESI of denatured and native ubiquitin solutions, respectively, will be examined separately. Maximizing MS/MS of Stored High-Charge-State Ions. Separate CAD/SORI spectra measured for isolated ubiquitin ion charge states were averaged to give spectra for 10+ and 11+, 8+ and 9+, and 6+ and 7+ (Figure 2A-D). Similar b and y ion (34) Horn, D. M.; Zubarev, R. A.; McLafferty, F. W. J. Am. Soc. Mass Spectrom. 2000, 11, 320-332. (35) Jockush, R. A.; Schnier, P. D.; Price, W. D.; Strittmatter, E. F.; Demirev, P. A.; Williams, E. R. Anal. Chem. 1997, 69, 1119-1126. (36) Aaserud, D. J.; Little, D. P.; O’Connor, P. B.; McLafferty, F. W. Rapid Commun. Mass Spectrom. 1995, 9, 871-876.

Figure 2. CAD/SORI spectra of gaseous ubiquitin ions: (A) 10+, 11+ ions; (B) 8+, 9+; (C, D) 7+ (C, capillary 56 °C; D, 76 °C). Right columns: %Σions, % of ions in the spectrum that are not molecular ions; b (y), number of cleavages yielding b (y) ions (0.5 credit if cleavage yields both); ib (iy), number of additional cleavages yielding ib (iy) ions. Vertical bars: open, b ion; filled, y ion. Vertical lines: no top (bar top), ib (iy) ion from dissociation of an identified b (y) ion by a bond cleavage not yielding a b or y ion. Abscissa: relative ion abundance, with white space (b, striped for y) in the baseline indicating cleavages corresponding to smaller peaks predicted with satisfactory confidence by THRASH.34

identities are reported by McLuckey and co-workers,30 but they did not tabulate i ions. Radio frequency excitation in SORI/CAD is set close to the resonance frequency of the precursor ions8 (similarly, excitation for the ion trap is on-resonance),30 which should limit the excitation at other frequencies and thus limit the corresponding secondary b, y dissociation. However, a surprising number of internal i ions are found. These identified 20 of the 52 cleavages from the 10+, 11+ ions and 25 of the 55 cleavages from the 8+, 9+ ions (Figure 2A, B). Dissociation to produce the lowest energy products may be slow enough so that the repeated lowenergy collisions of SORI excite the ubiquitin ions sufficiently to produce i ions directly. IRMPD, which is not a resonant excitation, was used to determine the effects of excitation energy on the MS/MS dissociation of 8+ to 11+ ubiquitin ions stored in the FTMS cell. The spectra reflect (Figure 3) an early study10 that showed increased IR irradiation first producing y58, b52, and y24 ions and then with y58 and b52 decreasing while i18-52, y18, y37, b18, and b39 increase. Using laser pulse lengths of 60, 70, 75, and 80 ms, the number of cleavages represented by the products were 30, 45, 41, and 30, respectively, with 6, 8, 11, and 13 cleavages identified by i products. Although higher than values reported earlier,10 these values are surprisingly lower than the 52 and 55 values of the CAD spectra of Figure 2A, B, with a far smaller number of cleavages identified from i ions. It may be that the IR energy deposition is not fast enough versus radiative cooling to give highly excited ions that give i ions directly or that the kinetic energy of dissociation axially displaces b, y product ions out of Analytical Chemistry, Vol. 77, No. 18, September 15, 2005

5779

Figure 3. IRMPD spectra of ubiquitin ions vs irradiation time: (AD) 8+-11+; (E) 6+, 7+. Left columns, experimental settings (see text); other designations in Figure 2.

the IR beam (∼2-mm diameter) where they are not dissociated further.10 Such axial displacement would also be greater for secondary products, making their detection less efficient (note the dramatic decrease in yield with increasing irradiation time). In Figure 3B, three fragment ions account for 40% of the products, so that the number of i cleavages should be increased by increasing the laser beam diameter, aligning the beam accurately with the ion cloud, or quadrupolar axialization of the product ions before spectral measurement. Effect of Collisional Energy on Activation of 8+-11+ Ubiquitin Ions in the Nozzle-Skimmer Region. Although ions enter the capillary at ∼10+3 Torr, efficient regional pumping at ∼10-3 Torr should produce a supersonic expansion of the ion beam with variable collisional activation (VN-S) and scattering before reaching the skimmer.7 NSD was examined first with the settings VS-Q ) 2, VN-S ) 47, mmN-S ) 9.0, and Acap ) 5.0 that were found to give optimum transmission of (M + nH)n+ ions for the spectra of Figure 3. This ion acceleration of 47 V in the N-S region gave only a 4% yield of products (Figure 4A), primarily those from the lowest energy cleavage sites 18 and 52,10 although this excitation energy was sufficient to remove any noncovalent molecular ion adducts.13-15 Increasing the collision energy to VN-S ) 87 (Figure 4B) increased the NSD yield to 28%, with 16 cleavages forming b and y ions and 3 forming i ions. However, a further increase to VN-S ) 137 still gave only 19 cleavages and the yield was nearly halved, with the y58 ion (bond 18 cleavage) still dominant. Although past studies have shown that an increase in VN-S gives a corresponding increase in the ion internal energy,37,38 here the excitation achieved may be limited by the (37) Voyksner, R. D.; Pack, T. Rapid Commun. Mass Spectrom. 1991, 5, 263268. (38) Collette, C.; De Pauw, E. Rapid Commun. Mass Spectrom. 1998, 12, 165170.

5780 Analytical Chemistry, Vol. 77, No. 18, September 15, 2005

Figure 4. NSD mass spectra of 8+ to 11+ ubiquitin ions from a 50:45:5 solution of MeOH/H2O/AcOH, with designations as in Figures 2 and 3.

short mean free path between collisions, producing relatively little change in the NSD spectrum. The 137-V acceleration should give more off-axis scattering of both precursor and product ions; minimizing scattering loss by optimizing the tube lens potential was not tried. Effect of Collisional Energy in the Skimmer-Quadrupole Region. The skimmer conductance limit and diffusion pump efficiency in the S-Q region should subject the ions to an average pressure that is several orders of magnitude lower than in the N-S region. Increasing the voltage drop between the skimmer and the lens/quadrupole (VS-Q, Figure 1), presumed to adjust the focus of the product ions into quadrupole 1, gave unexpected results. With VS-Q raised from 2 to 9 and VN-S ) 40, the yield of

NSD products increased dramatically from 4 to 59% (Figure 4A versus 4D; other experiments showed that the decrease of VN-S from 47 to 40 V gave a minimal decrease in extent of dissociation). This cannot be due just to better focusing of product ions versus molecular ions; 41% of the latter are not dissociated (59% yield), but those transmitted still represent 24% of the total measured. Thus, much of this increased yield is due to increased CAD in the S-Q region. The number of cleavages also increased dramatically, from 5 to 41, while only an increase to 19 cleavages resulted from raising VN-S from 47 to 137. The far longer mean free path of the S-Q region makes possible ion acceleration before collision that approaches that of the full VS-Q value (9 eV), even if the ions exiting the skimmer have already been collisionally cooled to a low velocity. Scattering losses are also far less in the S-Q region. For ions leaving this region, little CAD occurs in quadrupole 1 before the grounded exit orifice despite an additional 11-V ion acceleration, showing that the pressure in the ion beam drops dramatically from the skimmer exit to the quadrupole 1 entrance. Thus, in our experimental configuration, NSD can be even more effective in the S-Q region than in the N-S region, despite the latter’s far higher pressure; this was not observed in another NSD configuration.22 Conditions for CAD in our S-Q region appear to resemble those in the triple quadrupole MS/MS instrument, although in other experimental configurations it is conventional NSD from VN-S excitation that gives triple quadrupole-type CAD spectra.39 For this study, “NSD” represents dissociations that occur near the nozzle and skimmer. Increasing VS-Q to 19 V gave nearly complete molecular ion dissociation (product %Σions ) 96) in the S-Q region (Figure 4E), despite decreasing VN-S from 40 to 30. The cleavages indicated by b, y products decreased from 24 to 13, but the i cleavages increased from 17 to 22. Note that both ib cleavages and [y]/[b] have increased dramatically, supporting previous observations that b ions are less stable than y ions.8,10,35,36 Offsetting this more extensive dissociation, however, the higher accelerating potential after the skimmer has seriously reduced the ion collection efficiency, with the NSD yield reduced from 59 to 12%. This suggests poorer focusing and increased scattering for both molecular and product ions; redesign for separate control of the ion acceleration and focusing parameters in this low-pressure region should improve its NSD efficiency. Advantages of Consecutive Ion Activation. Although high ion acceleration in the N-S region (VN-S ) 137 V, Figure 4C) was far less effective (19 cleavages) than that in the S-Q region (VS-Q ) 19 V, Figure 4E), the latter only indicated 35 cleavages. An intermediate degree of ion acceleration in both the N-S and S-Q regions (VN-S ) 80 V, VS-Q ) 9 V) gave an NSD spectrum (Figure 4F) indicating 58 of the 75 possible interresidue cleavages. This is the most found for any of the experimental conditions tried here and by far the highest reported for energetic MS/MS methods (although a single ECD spectrum can indicate 74 of 75 possible ubiquitin cleavages).27 By this increase of VN-S from 40 to 80 (Figure 4D-F), 9 of the i ion cleavages are new, and much of the increase from 41 to 58 cleavages represents new y ions produced from the dominant y24 and y58 product ions, reduced from 55 to 15% of the total. This suggests that these conditions produce y24 and y58 ions efficiently in the N-S region, and these (39) Harrison, A. G. Rapid Commun. Mass Spectrom. 1999, 13, 1663-1670.

ions are efficiently dissociated in the S-Q region. However, there are also product ions that appear in substantial part to represent new primary cleavages: b4-8, b12, b21, b24, b39-42, b60, y59, and y60. These dissociations could be due to excitation of molecular ions to higher internal energies at higher VN-S values.37,38,40 Partial or complete unfolding of the molecular ions’ noncovalent structure could also produce new cleavages in NSD spectra. Dissociation of a noncovalent bond has a far lower activation energy than that of a backbone covalent bond, so that the multiple collisions of low energy in the low mean free path N-S region can sequentially dissociate multiple noncovalent bonds, while further collisions that could cause higher energy backbone covalent bond cleavage could instead reduce ion kinetic energy. Refolding is entropically less favorable than unfolding, maximizing the proportion of denatured ions entering the S-Q region to undergo the far higher collisional activation needed for backbone cleavage. Separate activations with low, then high, amounts of energy also avoid the addition of an intermediate amount of energy in excess of that needed for noncovalent bond dissociation; blackbody infrared dissociation of carbonic anhydrase, 29 kDa, for 30 s at 90 °C gives mainly H2O loss, but significant backbone fragmentation occurs above 140 °C with little H2O loss.41 Effect of Nozzle-Skimmer Distance (mmN-S). For these most favorable parameters of VN-S ) 80 and VS-Q ) 9 (Figure 4F), ion activation in the N-S region can be decreased by decreasing the distance of ion travel. The 13% distance reduction of Figure 4G reduced the total number of cleavages from 58 to 51; however, 9 of these were not indicated by Figure 4F, and the shorter distance gave 7 b ions lost at the longer distance. Reducing the energy of the collisions (VN-S from 80 to 40) instead of the number of collisions decreased the cleavages to 41 (Figure 4D), but this spectrum indicates 6 and 10 cleavages not shown by Figure 4F and G, respectively. Increasing the collision path length by 13% (Figure 4F-H) gave complete molecular ion dissociation and increased the cleavages identified by i ions from 18 to 23, but lowered total cleavages from 58 to 46, of which only 4 were not indicated by Figure 4F. Although it is more convenient to change VN-S than mmN-S, both values can be raised when even higher activation is required (vide infra). Effect of Capillary Heating Current. For optimum parameters of VN-S ) 80 and VS-Q ) 9, decreasing the external capillary temperature from 56 to 38 °C, Acap ) 5.0-3.5 A (Figure 4F versus I), has produced the largest number of b, y cleavages (42), with y cleavages increasing from 22.5 to 27.5. Although the number of i cleavages decreased even more (18 to 10), this ion cooling made 10 new cleavages observable (although all of low abundance), 7 of which formed the less stable b ions. Increasing the capillary temperature from 56 to 76 °C (Acap ) 5.0-6.5 A, Figure 4F, J) gave the largest number of i cleavages, an increase from 18 to 25, with b, y cleavages reduced from 40 to 28. Adding energy by increased collisions, Figure 4H versus F, increased the number of ib cleavages impressively (7 to 17), as expected,8,10,35 but decreased those of other product ions (y, 22.5 to 13.5; iy, 11 to 6). In contrast, adding energy by capillary heating, (40) Gabelica, V.; De Pauw, E.; Karas, M. Int. J. Mass Spectrom. 2004, 231, 189-195. (41) Ge, Y.; Horn, D. M.; McLafferty, F. W. Int. J. Mass Spectrom. 2001, 210/ 211, 203-214.

Analytical Chemistry, Vol. 77, No. 18, September 15, 2005

5781

Figure 5. Summary of cleavages found in CAD, IRMPD, and NSD spectra. Open bars above (below) line: cleavages corresponding to b (y) ions; half bars, ib (and iy) ions. Top spectrum, black bars: additional cleavages from spectra in ref 30. NSD spectra: gray bars, cleavages shown in Figure 4F (8+ to 11+) or Figure 6D (6+, 7+); white bars, additional cleavages shown in other Figure 4 or Figure 6 spectra as indicated by the corresponding letters.

Figure 4J versus F, lowered y ion cleavages less (22.5 to 18.5) and even increased iy (11 to 15.5) as well as ib (7 to 9.5) cleavages. As mentioned above and further below, capillary heating can increase the rate of native conformer unfolding42 and reduce the extent of gaseous ion folding, as well as increase ion internal energy. Increasing Acap to 7.0 A, yielding an external capillary temperature of 100 °C (7.25 Acap gives 126 °C), reduced ion transmission by ∼90%. This appears to be due to solids deposited before the capillary; the boiling point of the 50:45:5 MeOH/H2O/AcOH solution used for ESI is 83 °C. Maximizing the molecular ion dissociation (96-100%) by high activation in each of the 3 regions (S-Q, Figure 4E; N-S, 4H; capillary, 4J) gave 35, 46, and 53 cleavages, respectfully. However, each of these regions showed 7, 5, and 7 cleavages not found from the maximum activation in the other two regions (totaling 67 different cleavages). Note that this 96-100% dissociation gave yields of 12, 15, and 33%, respectively; the latter favorable value from capillary heating is consistent with lower scattering losses and minimized folding of the newly formed gaseous ions. Maximizing the Number of NSD Cleavages with Consecutive Activations. The Figure 4F ion products represent cleavages between 58 of ubiquitin’s 75 amino acid pairs, while data from all of the Figure 4 spectra increase this to 74 of 75 (Figure 5). The missing cleavage after residue 72 was not found in either the IRMPD or CAD spectra. Of the 16 cleavages not in spectrum of Figure 4F, 10 can be added by Figure 4I (lowest capillary heating), 3 more by Figure 4G (shortest NS region), and 2 more by Figure (42) Breuker, K.; McLafferty, F. W. Angew. Chem., Int. Ed. 2003, 42, 49004904.

5782 Analytical Chemistry, Vol. 77, No. 18, September 15, 2005

4E (highest after-skimmer acceleration). CAD provides the next most complete sequence coverage (Figure 5) with 64 cleavages from the Figure 2 spectra and 2 more from ref 30, despite the supposed limitations of near-resonance excitation on product ion dissociations. These ib and iy dissociations in IRMPD spectra indicate far less cleavage sites (Figure 5), possibly due to off-axis displacement of products. However, for NSD in the capillary, N-S, and S-Q regions there appears to be far less discrimination in the activation and scattering of precursor versus product ions. Surprisingly, the problem of sequence coverage using energetic activation is that this should cleave just the weakest bonds. However, varying the degree of consecutive activation in all three regions can vary the branching ratios of noncovalent, and of primary and secondary covalent, bond dissociations, thereby maximizing the proportion of interresidue bonds cleaved. Dissociation of Low-Charge-State Ions. One or more different gaseous conformer structures have been identified for each of the 13+ to 6+ charge states of ubiquitin ions by H/D exchange,33,43-45 ion mobility cross sections,46-48 and ECD.32,33,49 The 6+ ions, and the 7+ ions up to 100 °C, exhibit an unfolding enthalpy of ∼31 kJ/mol, versus 11 and 6 kJ/mol for the 8+ and 9+ ions, respectively.32 For ubiquitin 6+ and 7+ ions produced by ESI of a native ubiquitin solution (pH 5), IRMPD gave only 11% product ions in the spectrum (Figure 3E) versus 74% for 8+11+ ions under the same conditions (Figure 3D). For CAD, the 6+, 7+ ions were also substantially more stable (Figure 2C, D versus 2A, B and ref 30). However, the 6+ and 7+ ubiquitin species subjected to NSD immediately after their formation appear to be far more stable than those trapped and equilibrated in the FTMS cell. With Acap ) 5.0, the 8+-11+ ions under NSD conditions of VS-Q, VN-S, and mmN-S of 19, 80, 9 or 9, 80, 10.2 (Figure 4E, H) gave 96-100% dissociation, yet for 6+, 7+ ions NSD with far higher values of 24, 165, 10.2 only yielded 8% product ions representing 9 cleavages (Figure 6A). However, capillary heating does have a dramatic effect; increasing capillary temperature from 56 to 62 to 68 to 76 °C increased product ions from 8 to 30 to 53 to 78% (Figure 6A-D), representing 9, 35, 43, and 46 cleavages (higher temperatures greatly reduced ion transmission, vide supra). For the 8+ to 11+ ions, in contrast, capillary heating from 56 to 76 °C actually reduced the number of cleavages from 58 to 53 (Figure 4F, J). Also the 38 to 76 °C heating of the high-charge-state ions caused dramatic changes in the NSD relative product yields (Figure 4I, F, J), indicating corresponding changes in the internal energy of those ions undergoing dissociation. On the other hand, the four 6+, 7+ NSD spectra (Figure 6) show almost the same major products in similar branching ratios; capillary heating of the 6+, 7+ ions appears instead to increase the proportion that are (43) Suckau, D.; Shi, Y.; Beu, S. C.; Senko, M. W.; Quinn, J. P.; Wampler, F. M., III; McLafferty, F. W. Proc. Natl. Acad. Sci. U.S.A. 1993, 90, 790-793. (44) Freitas, M. A.; Hendrickson, C. L.; Emmett, M. R.; Marshall, A. G. Int. J. Mass Spectrom. 1999, 185/186/187, 565-575. (45) Geller, O.; Lifshitz, C. J. Phys. Chem. A 2005, 109, 2217-2222. (46) Hoagland-Hyzer, C. S.; Counterman, A. E.; Clemmer, D. E. Chem. Rev. 1999, 99, 3037-3080. (47) Badman, E. R.; Hoagland-Hyzer, C. S.; Clemmer, D. E. J. Am. Soc. Mass Spectrom. 2002, 13, 719-723. (48) Myung, S.; Badman, E. R.; Lee, Y. J.; Clemmer, D. E. J. Phys. Chem. A 2002, 106, 9976-9982. (49) Breuker, K.; Oh, H. B.; Lin, C.; Carpenter, B. K.; McLafferty, F. W. Proc. Natl. Acad. Sci. U.S.A. 2004, 101, 14011-14016.

Figure 6. NSD mass spectra of 6+, 7+ ions from a 100% H2O solution of native-state ubiquitin, with designations as in Figures 2 and 3.

Figure 7. Representation of the native solution structure from NMR of human ubiquitin.50 The amino acid side chains with at least four hydrophobic contacts are shown as sticks, those of the β-sheet in darker shading. Not shown is the hydrogen bonding between the sheet and helix regions, Thr14-Lys33, Glu18-Asp21 (2 bonds), and Pro19-Ser57.

sufficiently unfolded to undergo NSD. Even the 8% dissociation products at 56 °C (Figure 6A) appear to arise mainly from the unfolded conformer, so that the initial conformer is stable even under the severe collisional NSD excitation in the N-S and S-Q regions. This unusual stability is transient, however; after storing the ions in the FTMS cell, CAD (Figure 2C, D) and IRMPD (not shown) spectra there show little effect of raising the capillary temperature from 56 to 76 °C (Figure 2C, D). In fact, these incell spectra are sufficiently different that they may result from an additional conformer formed after further equilibration, although Clemmer et al. have found closely similar CAD spectra for different conformers of a single charge state of ubiquitin ions.47 It seems unlikely that transient stable conformer is the native structure. In solution, its major stabilization is provided by hydrophobic bonding (Figure 7),50 which must be weakened by desolvation. Unfolding in the inlet capillary region during solvent evaporation was shown to cause “native ECD” cleavage of the solution dimer of native cytochrome c, with this unfolding complete with a 40 °C capillary.42,51 For native structures of both cytochrome c and ubiquitin, ECD of ions in the FTMS cell shows that the hydrophobic bonding is removed in the desolvation process. However, at the same time, desolvation greatly strengthens hydrogen bonding; stronger H-bonds between Thr14-Lys33, Glu18-Asp21 (2 bonds), and Pro19-Ser57 (Figure 7) or salt bridges could then preserve the compact (although expanded) general structure of the native conformer long enough to resist collisional dissociation in the N-S and S-Q regions. Such a transient gaseous conformer is consistent with Shelimov and Jarrold’s observation of collisionally stable 3+-5+ cytochrome c ions that fold further with collisional heating.52 Badman et al. observed a non-native gaseous conformer of 9+ cytochrome c ions that is formed in 10 ms after ESI; stored in an ion trap, this unfolds to

more open structures, which after ∼20 ms cooling refold, but not to the initial structure.53 Thus, the ubiquitin 6+, 7+ ion conformer that is formed almost simultaneously with unfolding of the native conformer appears to be transiently stabilized by strengthening of H-bonds, but with its formation prevented by heating the entrance capillary. NSD of Larger Proteins. Even though capillary heating causes the ions reaching the N-S region to be unfolded or less folded, this has little effect on their CAD (or IRMPD) spectra in the FTMS cell (Figure 2C vs D). Thus, NSD with capillary heating has unique advantages for MS/MS of gaseous ions that can form stable tertiary structures, as preventing any of this folding in the capillary saves the corresponding energy of unfolding (including the activation energy) that would have to be added by collision. For the 8+-11+ ubiquitin ions (Figure 4I, J) represented by several conformer structures,32,33,44-49 this could account, at least in part, for the surprising number of new cleavage sites identified by varying the capillary temperature. This NSD of minimally folded ions made possible by capillary heating (“prefolding dissociation”) should be of far greater advantage for larger proteins, as these will have far more complex tertiary structures in the gas phase, in addition to the higher dissociation energies required by their size.54 For example, collisional denaturing of the tertiary noncovalent strucures of gaseous protein ions followed by ECD gives the highest proportion of interresidue cleavages from a single spectrum for larger proteins, 183 of 258 possible cleavages for 29-kDa carbonic anhydrase.27 The NSD spectra obtained previously from 67-15,21 and 74-kDa23 proteins may have been made possible in part by incomplete formation of their gaseous noncovalent tertiary structure. Unfolding of a terminal portion of an ionized protein’s tertiary structure can lead to asymmetric charge partitioning during dissociations in either the ion cell55 or the inlet

(50) Cornilescu G.; Marquardt J. L.; Ottiger M.; Bax A. J. Am Chem. Soc. 1998, 120, 6836-6837. (51) Breuker, K.; McLafferty, F. W. Angew. Chem., Int. Ed., in press. (52) Shelimov, K. B.; Jarrold, M. F. J. Am. Chem. Soc. 1996, 118, 10313-10314.

(53) Badman, E. R.; Myung, S.; Clemmer, D. E. J. Am. Soc. Mass Spectrom., in press. (54) Griffin, L. L.; McAdoo, D. J. J. Am. Soc. Mass Spectrom. 1993, 4, 11-15. (55) Jurchen, J. C.; Williams, E. R. J. Am Chem. Soc. 2003, 125, 2817-2826.

Analytical Chemistry, Vol. 77, No. 18, September 15, 2005

5783

capillary;42,51 NSD of much larger proteins under variable denaturing conditions could maximize the charge state of specific fragment ions for their optimum dissociation after collection in the FTMS ion cell. CONCLUSIONS With its consecutive activation steps NSD is capable of identifying substantially more backbone cleavage sites in ubiquitin than either CAD or IRMPD. Immediately after electrospray formation, ions are subjected to by far their greatest changes in pressure and temperature, causing similarly unusual changes in their extent of thermal and collisional activation, loss by scattering, and degree of refolding into conformations stabilized at different sites. NSD should be a valuable complement to other MS/MS methods for structural characterization, especially for instruments without the extensive interresidue dissociation capabilities of ECD or similar techniques.56,57 If other methods have located a posttranslational modification of a fairly pure protein to within a specific backbone region, for further localization of the site, NSD conditions can be varied to search for new cleavages just in this region. If this region has a DNA-predicted sequence, the masses of the specific b, y, and i ions from these cleavages can be predicted, while real-time computer feedback control can be envisioned that could vary these instrument parameters to form and identify these key ions selectively.2,6 Although NSD can only be applied to electrosprayed ions that have not been subjected to mass separation, mixtures of proteins whose possible sequences are DNA predicted could still give NSD fragment ions whose mass (or MS/MS spectrum) can be assigned to a specific protein with confidence (the “shotgun” approach).58

Conformer ions can be collisionally unfolded in the N-S region for backbone dissociation in the S-Q region, and molecular ions can be activated immediately after they are desolvated, applying NSD before they fold into a stable gaseous tertiary structure(s). Capillary heating of the evaporating native ubiquitin solution eliminated the stable transient 6+, 7+ ion conformations, while low-energy collisions in the N-S region maintained this denaturation for NSD in the S-Q region. This consecutive ion activation is especially promising for extending the top down technique to larger proteins. Their gaseous molecular ions will have increasingly complex and stable tertiary noncovalent bonding; prefolding NSD should increase the yield of large fragment ions that can then be subjected to further top down MS/MS steps in the FTMS cell. In preliminary results, NSD of human complement C4 (reported as 200 kDa, with segments of 75, 93, and 33 kDa connected by S-S bonds)59 gives 36 b and 32 y ions (largest b186 and y184) that provide 48% sequence coverage. ACKNOWLEDGMENT We thank Ethan Badman, Barbara Baird, Tadhg Begley, Barry Carpenter, David Clemmer, David Holowka, Harold Hwang, Cheng Lin, Michal Steinberg, and Yun Xiang for helpful advice and discussions, and the National Institute of General Medical Sciences (NIH grant GM16609 to F.W.M.) and the Austrian BMBWK and FWF (grant T229 to K.B.) for generous financial support. Received for review April 11, 2005. Accepted June 28, 2005. AC0580215

(56) Syka, J. E. P.; Coon, J. J.; Schroeder, M. J.; Shabanowitz, J.; Hunt, D. F. Proc. Natl. Acad. Sci. U.S.A. 2004, 101, 9528-9533. (57) Silivra, O. A.; Kjeldsen, F.; Ivonin, I. A.; Zubarev, R. A. J. Am. Soc. Mass Spectrom. 2005, 16, 22-27.

5784

Analytical Chemistry, Vol. 77, No. 18, September 15, 2005

(58) McDonald, W. H.; Yates, J. R., III. Curr. Opin. Mol. Ther. 2003, 5, 302309. (59) Seya, T.; Nagasawa, S.; Atkinson, J. P. J. Immunol. 1986, 136, 4152-4156.