Charge-State Shifting of Individual Multiply-Charged Ions of Bovine

Charge State Reduction of Oligonucleotide Negative Ions from Electrospray Ionization. Xueheng. Cheng , David C. Gale , Harold R. Udseth , and Richard ...
1 downloads 0 Views 429KB Size
Anal. Chem. 1994,66, 2084-2087

Charge-State Shifting of Individual IVkrltiply-Charged Ions of Bovine Albumin Dimer and Molecular Weight Determination Using an Individual-Ion Approach Xueheng Cheng, Ray Bakhtiar, Steven Van Orden, and Richard D. Smith' Chemical Sciences Department, Pacific Northwest Laboratory, Richiand, Washington 99352

Ion-molecule reactions of individual multiply-protonated ions of bovine albumin dimer, formed from electrospray ionization, have been studied using a Fourier transform ion cyclotron resonance mass spectrometer. Upon reactionof ammoniawith a group of individual ions, charge-state shifting was observed due to proton transfer. Repeated additions of ammonia during remeasurements of the same ion population were observed to induce multiple-step charge-state shifts. Chargestate-dependent reactivity, as well as nonstatistical behavior in reactivity, was observed due to the small ion population. The molecular weights of individual ions whose charge state shifted during reaction were determined with an accuracy of 67 ppm, the first example of using an individual-ion approach to the determination of molecular weight for a large biopolymer. The molecular weight distribution of a group of ions can be determined with a precision related to the number of ions examined and the weight heterogeneity of the sample. We obtained the molecular weight for eight individual ions from which a molecular weight of 133 320 f 210 Da was calculated for bovine albumin dimer. Electrospray ionization (ESI) generates intact multiplycharged ions from solutions of a wide range of materials including large biological and synthetic polymers.' The multiple charging reduces the mass-to-charge ratio (m/z)of ions to within the range of conventional mass spectrometers, allowing molecular weight (MW) measurements with accuracies of 0.01%or better.' Recent developments in Fourier transform ion cyclotron resonance (FTICR)2mass spectrometry in conjunction with ESI have made possible ultrahighresolution measurements of smaller proteins (>2 X lo6 at a molecular weight of 8.6 kDa).3.4 The accurate molecular weight assignment from the distribution of multiply-charged ions conventionally requires measuring the m/z spacing between the mass spectrometrically resolved charge states, the 1-Da isotopiccontributions, or the m/z differences resulting from the adduction of known species. However, FTICR (1) (a) Fenn, J. B.; Mann, M.; Meng, C. K.; Wong, S.F.; Whitehouse. C. M. Science 1989,246,64.(b) Fenn, J. B.; Mann, M.; Meng, C. K.; Wong, S.F.; Whitehouse, C. M. Mass Spectrom. Reu. 1990,9,37.(c) Smith, R. D.; Loo, J. A.; Edmonds, C. G.; Barinaga, C. J.; Udseth, H. R. Anal. Chem. 1990,62, 882. (d) Smith, R. D.; Loo, J. A.; Ogorzalek Loo,R. R.;Busman, M.; Udscth, H. R. Mass Specfrom. Rev. 1991, 10, 359. (e) Kebarle, P.; Tang, L. AMI. Chem. 1993,65, 972A. (2) (a) Marshall,A.G.;Grosshans,P.B.Anal. Chem. l991,63,215Aandrefcrmces therein. (b) Buchanan, M. V.; Hettich, R. L. Anal. Chem. l993,65,245A and references therein. (3) Beu. S.C.; Senko, M. W.; Quinn, J. P.; Wampler, F. M., 111; McLafferty. F. W. J. Am. SOC.Mass Specfrom. 1993, 4, 557. (4) Bruce, J. E.; Anderson, G. A.; Hofstadler, S. A.; Winger, B. E.; Smith, R. D. Rapid Commun. Mass Spectrom. 1993, 7, 700.

2084

AnalyticalChemlstry, Vol. 66,No. 13, Ju& 1, 1994

performance is ultimately limited by space-charge effects2 (and the resulting limits on the number of charges in the FTICR cell necessary to avoid substantial electric field perturbationS ) and the effective isotopic heterogeneity (from naturally occurring isotopes, e.g., W ) , as well as the considerable heterogeneity of most macromolecules. We are developing an alternative approach for large-molecule characterization on the basis of the remeasurement of individual ions (Le., single ionic species) in FTICR that circumvents these difficulties. In this approach, the number of charges, and hence the molecular weight of an individual ion, can be determined by inducing an m/z shift through a reaction step involving charge transfer6 or the formation and/or dissociation of adducts (i.e., by methods that are analogous to those used for spectra from large ensembles of ions). Recently, we have successfully developed methods for trapping, detection, and remeasurement of individual ions of biological and synthetic polymers with MW's extending to >5 MDa.' In this correspondence, we report the application of the individualion approach to molecular weight determination of bovine albumin dimer, a globular protein with molecular weight of 133 kDa. The charge-state shifting was achieved by proton transfers upon reaction with ammonia (proton affinity (PA) = 204 kcal/mo18 ), allowing a molecular weight of 133 320 f 210 Da to be determined. EXPERIMENTAL SECTION This work was performed using a modified IonSpec (Irvine, CA) FTICR with a Oxford 7-T superconducting magnet and incorporates a modified Analytica (Branford, CT) ESI source. A detailed description of the ESI-FTICR instrumentation is given el~ewhere.~ Briefly, our instrument is equipped with an integral cryopump extending into the bore of a 7-T superconducting magnet and the vicinity of the trapped ion cell, greatly enhancing the pumping speed in the cell region. The pressure in the cell can be varied from =lo-" to I Torr in a few seconds, enabling both the collisional trapping of large ions and the rapid acquisition of high-resolution mass ~ p e c t r a . ~The trapping gas (nitrogen) and reagent gas ( 5 ) (a) Jeffries, J. B.; Barlow. S. E.; Dunn, J. H. h i . J. Mass Specrrom. Ion Processes 1983, 54, 169. (b) Wang, T.-C. L.; Marshall, A. G. Inf. J. Mass

Spectrom. Ion Processes 1986, 68, 287. (6) (a) McLuckcy, S.A.; Van Berkel, G. J.; Glish, G. L. J. Am. Chem. Soc. 1990, 112,5668.(b) McLuckcy, S. A.;Glish, 0.L.; Van Berkel, G. J. AMI. Chem. 1991, 63, 1971. (7) Smith, R. D.; Cheng, X.; Bruce, J. E.; Hofstadler, S.A.; Anderson, G. A. Nature, in press. (8) Auxiliary thermochemical data obtained from: Lias, S.G.; Bartmess, J. E.; Liebman, J. F.; Holmes, J. L.; Lcvin, R. D.; Mallard, W. G. J. Phys. Chem. ReJ Dara, Suppl. I 1988, 17. (9) Winger, B. E.; Hofstadler, S.A.; Bruce, J. E.; Udscth, H. R.; Smith, R. D. J . Am. Soc. Mass Specrrom. 1993, 4, 566.

0003-2700l94f 0366-2084504.50l0

0 1994 Amerlcan Chemical Soctety

(M + zH)'+ + N H ,

E

C

l ' " 3

1500

2000

2500

3000

3500

mlz Flgue 1. FTICR mass spectrashowing charge-state shifflngof bovine albumhdlmer: (a)isoiatron of a populatbn of individualions, (b) reactlon of the same population of individual ions (a) with ammonia yieiding chargestate shifting, and (c) background prior to ion injection.

(ammonia) were introduced via two separate magnetically compatible piezoelectric pulsed valves (Laser Techniques, Albuquerque, NM) into the vicinity of the ICR cell. A freshly prepared solution (0.5-1.Omg/mL in 25/75CH3OH/% CHsCOOH) of bovine albumin dimer (Sigma, St. Louis, MO), purified by filtration using an Amicon microfilter 100,was infused into the ESI source at a rate of 0.5 pL/min. The positive ion mass spectra exhibited the characteristic distribution of charge states (due to the different number of attached protons or residual of sodium ions) which are separated by m/z values consistent with the sample's molecular weight. By limiting the number of ions injected into the FTICR cell, using a variable length suspended trappingevent,1° and applying conventional radio-frequency (rf) excitation/ ejection methods, any number of discrete multiply charged ions can be trapped, isolated, detected, and remeasured over the course of hours. A typical spectrum after ion isolation of a limited ion population (Figure la) shows hundreds of peaks due to individual ions with signal-to-noise ratios (S/N) of 2-3. We also observe a small fraction of the peaks with a larger S/N of 4-6 that we attribute to tetramers (Le., dimers of the bovine albumin dimer, vide infra). RESULTS AND DISCUSSION When a population of trapped ions was reacted with ammonia, the m/z distribution of ions shifted to higher values (Figure la,b). Further additions of ammonia yielded additional m/z shifts, but the rate of shifting (i.e., reaction) decreased with each addition step. These observations are consistent with proton transfers from the protein ions to ammonia, yielding stepwise charge-state shifts (eq 1). (10) (a) Laude, D. A., Jr.; Bcu, S. C. Anal. Chem. 1989,61,2422.(b) Hogan, J. D.; Laude, D. A., Jr. Anal. Chem. 1990,62, 530.

-

[M+ ( z - l)H]('-')+

+ NH,'

(1)

The highly-charged ions (at lower m/z) generally react faster than ions with lower charge states (at higher m/z) due to the larger Coulombicrepulsion forces within more highly-charged ions. Coulombic repulsion increases the gas-phase acidity of the multiply-protonated ions,6J and therefore, the highlycharged ions show higher reactivity toward neutral bases.12 As shown in Figure 1, the ion population shifts from m/z 1800-3010 to m/z 1900-3500 upon multiple additions of ammonia. The average m/z shift of 300 corresponds to an average decrease of six in charge state. Under the conditions used here, the charge states for bovine albumin dimer ions can be shifted to as low as 40+ (m/z E 3400). In order to measure the m/z shift for each addition of neutral base (required for the molecular weight measurement of individual ions), a group of ions of limited number was isolated and reacted with ammonia. As shown in Figure 2,the products from such a charge-state shifting reaction can be identified unambiguously. The population of ions (part of which is illustrated in Figure 2) was reacted with ammonia for up to 12times and the behavior of the ions closely examined. Figure 3 shows plots of the m/z value vs the remeasurement experiment number for selected ions. In most cases, it was possible, if the ion population was not excessively large, to correlate individual ions between measurements. Upon each addition of ammonia, the majority of the ions undergo either a single charge-state shift or no shift at all. However, multiple charge-state shifts during a remeasurement also occurred for a small fraction of the ions, particularly for some of the highly charged ions, which is consistent with the charge-state-dependent reactivity discussed earlier. Someindividual ions with similar m/z values display pronounced differences in reactivity (i.e., the two ions at about m/z 2080 in Figures 2 and 3). The fraction of events (the total number of events = number of ions X number of remeasurements) that leads toonechargestate shift (multiple shifts during one addition of ammonia are counted as multiple shifting events) was determined to be 14% for ions with m/z 1800-3000. Assuming a cross section of ion-neutral collision equal to the physical cross section of the protein13 and the neutral pressure/residence time of 10-5 Torr/ 100ms,14the total number of collisionsbetween a protein ion and ammonia is calculated to be ca. 2.2 X 103 at room temperature during each remeasurement. Therefore, the (1 1) (a) McLuckey, S. A.; Glish, G. L.; Van Bcrkel, G. J. Proceedings of 39rh ASMS Conference on Muss Specrrom.und Allied Topics;l991; pp 901-902. (b) Pereie, S.; Javahery, G.; Wincel, H.; Bohme, D. K. J. Am. Chem. Soc. 1993, 115, 6290. (c) Pereie, S.;Javahery, G.; Bohme, D. K. Int. 1. Muss Spectrom. Ion Processes 1993. 124, 145. (12) Reactionsother than proton transfer (is., Na+ transfer) are unlikely because proton transfera are kinetically more facile. (13) The physical cross section of bovine albumin dimer is taken to be 1 X lW A', b a d on the dimension of serum albumins in the condensed phase (140 A x 40 A X 40 A): (a) Carter, D. C.; He, X.-M.; Munson, S. H.; Twigg, P. D.; Gernert, K. M.; Broom, M. B.; Miller, T. Y. Science 1989, 244, 1195. (b) Carter,D.C.;He,X.-M.Science1990,249,302.(c) Wright,A.K.;Thompson, M. R. Biophys. J. 1975,15,137.We also note a r a n t report of collision croas section of (1.1-1.4) X l W A 2 for bovine albumin: Covey, T.; Douglas, D. J. J. Am. Soc. Muss Spectrom. 1993, 4, 616. (14) The pressure profile in the cell region after pulsed addition of neutral gas was unknown. The peak pressure was measured using an ion gauge located away from the cell and outside the magnetic bore.

AnalyticalChembtry, Vd.66,No. 13, July 1, 1994

2085

a

DllW

MW *O

T a m 1. Molecular Wdghl Det.rminatlon of IndMdual Ions from Bovine Nbumln Dlmor mlz z no. of shiftsa MW (Da) AMW (Da)b

M5h

I

I

0

1154 1806 1908 208 1 2146 226 1 2564 2185 average fwhm 2155

16 14 70

64 62 59 52 48 119

133 222 133 538 133 435 133 104.1 132 911.3 133 341.9 133 260.1 133 631 133 315.1 2lof 256 322.1d

13 18 13 2.4 3.6 4.1 6.9 11

8.9 5.5

'Number of charge-state shifts observed during the course of 12 remeasurements. The accurac of molecular weight measurement for each individual ions. fwhm ofthe molecular weight distribution for bovine albumin dimer. Tetrameric spccics of bovine albumin.

2100

2150

2200

2250

m/z Flgure 2. Chargastate shifting mass spectra of bovine albumin dimer in a narrow m/z range. Consecutive addition of ammonia induces charge-state shifting of individual ions with charge state 64+ (one of the two, 0 , two shifts), 62+ (A,one shift), and 1I S + of tetrameric specks (X, two shifts). Symbols M and 0 denote bovinealbumin dimer and bovine albumin tetramer (Le., dimer of the dlmer), respectively. Individual ions with charge state 64+ (one of the two,+) and 594- (0) did not shift during the courseof the reaction. Spectra were taken after a 10-100-m~pulse of ammonia (peak pressure on the order of 10" Torr) and 10-s delay and with a 200kHz broad-band detection.

-!

L

.c! E

2800

-

2600

-

2400

-

2200

1

2000

-

1800

0

I

I

,

I

2

4

6

8

I

1

0

1

I

1

2

1

4

Remeasurement Experiment No. Flgure 3. Plots of m/z values of individual ions vs remeaswement experiment number. Duringeach remeasurement,a 10-100-ms pulse of ammonia was added (peak pressure on the order of Torr).

reaction efficiency's is estimated to be on the order of 0.6 X with a rate constant (kobervtd)ca. lo-" cm3 molecule-' s-1 for ions with m/z 1800-3000. This estimate can be compared with the results of McLuckey et a1.6awho reported 2086

rate constants for proton-transfer reactions of horse heart cytochrome c (MW = 12 360 Da) to dimethylamine (PA = 221 kcal/mo18), ranging from 9.3 X cm3 molecule-' s-1 (for charge state 15+, m/z 825) to