Proc. Natl. Acad. Sci. USAVol. 93, pp. 3143-3148, April 1996Chemistry

Conformations and folding of lysozyme ions in vacuo(electrospray ionization/proton transfer reactions/protein conformation/Fourier-transform mass spectrometry)

DEBORAH S. GROSS, PAUL D. SCHNIER, SANDRA E. RODRIGUEZ-CRUZ, CLIFTON K. FAGERQUIST,AND EVAN R. WILLIAMS*Department of Chemistry, University of California, Berkeley, CA 94720

Communicated by Fred W. McLafferty, Cornell University, Ithaca, NY, November 15, 1995 (received for review September 25, 1995)

ABSTRACT Proton transfer reactivity of isolated chargestates of the protein hen egg-white lysozyme shows that multipledistinct conformations ofthis protein are stable in the gas phase.The reactivities of the 9+ and 10+ charge state ions, formed byelectrospray ionization of "native" (disulfide-intact) and "dena-tured" (disulfide-reduced) solutions, are consistent with valuescalculated for ions in their crystal structure and fully denaturedconformations, respectively. Charge states below 8+ of bothforms, formed by proton stripping, have similar or indistinguish-able reactivities, indicating that the disulfide-reduced ions fold inthe gas phase to a more compact conformation.

The role of intramolecular interactions, such as electrostatic,hydrophobic, and hydrogen bonding, as well as solvent inter-actions, on the conformation and the dynamics of proteinfolding have been widely investigated (1). The development ofnew ionization techniques, including matrix-assisted laser de-sorption/ionization (2, 3) and electrospray ionization (ESI)(4), has made possible the formation of intact molecular ionsof large biomolecules, the structures of which can be investi-gated both in solution as well as in the complete absence ofsolvent using an impressive array of mass spectrometric tech-niques (5-22). In solution, information about protein confor-mation can be inferred from charge state distributions ob-served in electrospray mass spectra (9-13), as well as fromshifts in mass upon solution-phase deuterium exchange (14,15). In the gas phase, different protein conformations havebeen observed using deuterium exchange (16-18), collisionalcross-section measurements (19, 20), and proton transferreactions (21, 22). A key question that remains, however, ishow closely, if at all, the gas-phase conformations are relatedto those in solution. Accurate measurements of gas-phaseconformation should greatly improve our understanding of therole of solvent in protein conformation and folding. Althoughthe methods that have been used to date are able to differ-entiate gas-phase protein conformers, they have providedindirect evidence of their exact nature. We demonstrate herethat relatively unambiguous information about some of thesegas-phase structures can be obtained from differences in theirgas-phase proton transfer reactivity. Distinct gas-phase con-formations of disulfide-intact and disulfide-reduced hen egg-white lysozyme that have dramatically different proton trans-fer reactivities are observed; several of these ions have con-formations that are consistent with either a fully denaturedstructure or with a compact form that is consistent with thecrystal structure. In addition, clear evidence for protein foldingin the gas phase is presented.

Electrostatic interactions have a pronounced effect on thechemistry of multiply protonated ions, including proton trans-fer reactivity (21-28) and dissociation (29, 30). We have shownthat the proton transfer reactivity of doubly protonated 1,n-diaminoalkanes (n = 7-10, 12) (26) and the small cyclic

The publication costs of this article were defrayed in part by page chargepayment. This article must therefore be hereby marked "advertisement" inaccordance with 18 U.S.C. §1734 solely to indicate this fact.

peptide gramicidin S (27) can be fully accounted for by thecombined effects of the intrinsic reactivity of the site ofprotonation and a point-charge Coulomb model using aneffective dielectric polarizability of 1.01 ± 0.07 and <1.2,respectively, for these isolated gas-phase ions. We have ex-tended these measurements to isolated charge states of theprotein cytochrome c, the results for which are reportedelsewhere (28). Because of the long-range 1/r distance depen-dence (where r is the intercharge separation), accurate mea-surements of Coulomb interactions should provide a sensitiveprobe of ion conformation. Differences in gas-phase proteinion conformation have been inferred from differences inproton transfer reactivity. Smith and co-workers (21, 22) foundthat reactions of all charge states of several disulfide-intactproteins with different bases in the electrospray interface regionresulted in production of lower charge states than for the corre-sponding disulfide-reduced ions. More differences were observedfor higher charge states (21) than for lower charge states (22). Weshow here that, by comparison of the measured proton transferreactivity of individual charge states to those calculated using asimple model, relatively unambiguous information about theconformation of several gas-phase lysozyme ions can be obtained.

MATERIALS AND METHODSAll experiments were performed on an external electrosprayion source Fourier-transform mass spectrometer equippedwith a 2.7-T superconducting magnet (Fig. 1). Ions wereformed at atmospheric pressure by applying -4 kV to a 50-pami.d. aluminum-clad fused silica capillary (SGE, Inc. Austin,TX) through which analyte solution was continuously infusedat a rate of 2.0 ul/min. Ions, introduced to the first stage ofhigh vacuum through a heated metal capillary (180°C) and twodifferential mechanically pumped regions, were acceleratedpast the fringing fields of the 2.7-T superconducting magnetthrough a series of electrostatic lenses (-2.5 kV maximumvoltage) and decelerated prior to injection into the ion cell(base pressure of 5 x 10-9 torr; 1 torr = 133 Pa). Trapping (8V static, 5 s ion injection) and ion thermalization was enhancedthrough collisions with N2 introduced through a pulsed valve(K. J. Lesker, Co., Livermore, CA) to a pressure of -2 x 10-6torr. A mechanical shutter prevented additional ions fromreaching the ion cell during subsequent events. Charge stateswere isolated using stored waveform inverse Fourier transform(31) and allowed to react for times up to 60 s with neutralreference bases of known gas-phase basicity (GB) [fromMeot-Ner and Sieck (32) adjusted to the GB scale of Lias et al.(33); the bases DBN, DBU, and MTBD are from Decouzon etal. (34)] introduced through a sapphire leak valve (VarianVacuum Products, Lexington, MA) to a pressure of -1 X 10-7torr. Rates of proton transfer were obtained by fitting thesedata to pseudo-first-order kinetics. Spectra were acquired

Abbreviations: GB, gas-phase basicity (basicities); GBaPP, apparentgas-phase basicity (basicities).*To whom reprint requests should be addressed.

3143Dow

nloa

ded

by g

uest

on

Janu

ary

6, 2

021

Proc. Natl. Acad. Sci. USA 93 (1996)

2.7 Tesla SuperconductingMagnet

External SourceElectrospray Ionization

Pulsed Valve

Sapphire .-Leak Valve m

230 L/sion pump

5 x 10-9 Torr.. : : ::::... . : :: .300 Lsdiffusionpump

900 L/sdiffusionpump 1600 L/s

diffusionpump

FIG. 1. The Berkeley-Extrel external electrospray ion source Fourier-transform mass spectrometer. CFM, cubic feet per minute.

using an Odyssey Data System (Extrel FTMS, Madison, WI)using a rf sweep (120 V peak-to-peak; 1100 Hz/pus) and 256 x103 data points; trapping potentials were reduced to 1.0 V priorto detection.Hen egg-white lysozyme was obtained from ICN Biomedi-

cals (Costa Mesa, CA). Disulfide-intact lysozyme was electro-sprayed from a pure aqueous solution at room temperature.The disulfide bonds were reduced by boiling a solution oflysozyme in 0.02 M dithiothreitol for 30 min immediatelybefore it was electrosprayed. The solution was diluted withmethanol, acidified with acetic acid (4%), and infused througha syringe resistively heated to 60°C to ensure that the disulfidebonds did not re-form.

RESULTS AND DISCUSSIONESI mass spectra of hen egg-white lysozyme solutions elec-trosprayed from a pure aqueous solution (disulfide-intact) andfrom a heated methanol/water solution containing 4% aceticacid and 0.02M dithiothreitol (disulfide-reduced) are shown inFig. 2. Charge distributions of 8+ through 11 + and 9+ through18+ were produced from these solutions, respectively; the

disulfide-reduced ions were eight mass units higher, confirm-ing that each of the four disulfide bonds was reduced. Inaddition, these ions should be denatured under these solutionconditions. The isolated charge states were allowed to reactwith neutral reference bases of known GB. The apparentgas-phase basicities (GBaPP) of the (M + (n - 1)H)(n-l)+ ionswere assigned from the rates of proton transfer from the (M+ nH)n+ ions to these bases (28). The proton transfer rates ofthe disulfide-reduced and disulfide-intact ions are given inTables 1 and 2, respectively. The assigned values of GBaPP aregiven in Table 3.For several charge state/base combinations, an accurate fit

to the data requires more than one rate constant. A similarobservation has been reported for ubiquitin (23). As an

example, the data and best fit for the 9+ disulfide-intact ionreacting with triethylamine is shown in Fig. 3. These resultsindicate the existence of three distinct ion structures or groupsof structures that have different reactivity. The abundance offast, slow, and unreactive ion structures was 0.55, 0.35, and0.10, respectively. Doubling the number of collisions with N2had negligible effect on these proton transfer rates, indicatingthat the difference in ion reactivity is due to three different

100.0-

80.0-

60.0-

40.0-

20.0-

,0It

(M + 10H)10+

[ L1UU.U.........

(M + 15H)15+80.0

60.0(M + 11H)ll+

40.0

20.0-

0.0 ,','.7.-.-.7 r ,',-, ,I r ,- ^^.,̂ ,̂^ ~.A-^ ^

-- , ',

(M + 8H)8+

K,.'

200.0 400.0 600.0 800.0 1000.0 1200.0 1400.0 1000.0 1800.0 2000.0m/z

FIG. 2. Electrospray Fourier-transform mass spectrometer spectra of disulfide-intact (Upper) and disulfide-reduced (Lower) hen egg-whitelysozyme.

3144 Chmsr:Goseal

I-

Dow

nloa

ded

by g

uest

on

Janu

ary

6, 2

021

Proc. Natl. Acad. Sci. USA 93 (1996) 3145

Table 1. Rates of proton transfer (x 10- 12 cm3/mols) from individually isolated (M + nH)n+ charge states of disulfide-reduced henegg-white lysozyme to neutral reference bases

GB, nkcal/

Reference base mol 4+ 5+ 6+ 7+ 8+ 9+ 10+ 11+ 12+ 13+ 14+ 15+ 16+

Acetophenone 198.0 3.2 5.22-Fluoropyridine 203.9 1.7 2.4 2.03-Fluoropyridine 208.6 0.22 0.68Isopropylamine 213.7 3.7 9.0 22 > 100

0Pyridine 215.7 1.3 33 (0.28) >1000 >1000

190 (0.10)tert-Butylamine 216.7 5.7 (0.13) 4.8

80 (0.13)Diethylamine 221.4 7.4 42 96 (0.87) 160Dipropylamine 225.4 1.7 44 > 1000 >1000Triethylamine 227.5 * * 31 (0.78) 4.2 (0.15) 130

2600 (0.12) 110 (0.85)Tributylamine 231.1 * * * 2.7 19 (0.40) >1000 >1000 >1000

130 (0.30)DBN 237.4 26 (0.46) t t t tMTBD 243.3 31 (0.46)Lower charge states not produced directly by ESI are formed by proton transfer to the appropriate base. GBaPP of the (M + (n 1)H)(n-l)+

ions are assigned between the bases whose rates bracket the absolute rate of 1 x 10-1 cm3/mol.sec. Multiple rates indicate the presence of morethan one reactive conformer (abundances indicated in parentheses; unreactive conformers are not shown). DBN, 1,5-diazabicyclo[4.3.0]non-5-ene;MTBD, 7-methyl-1,5,7-triazabicyclo[4.4.0]dec-5-ene.*Charge stripping formed insufficient abundance of the ion to isolate for reaction, indicating that the GBaPP of this ion is greater than the GB ofthis neutral reference base.

tRates are significantly faster than 10-9 cm3/mol.s. This ion could not be isolated with this reference base.

structures and not due to a thermally excited population ofions. This is consistent with recent results showing that IRcooling is rapid for these large ions. These structures were

assigned GBaPP(8+) of 222.3 + 3.4, 226.5 + 2.5, and 229.3 ±

3.3 kcal/mol, respectively, based on their proton transfer ratesto this base, as well as to the other bases in Table 2.To obtain information about ion conformation, we com-

pared the measured GBaPP values to those calculated using a

model in which charges are localized to specific sites in theprotein. The calculated reactivity was obtained from thedifference in intrinsic reactivity of a site of protonation and the

Coulomb energy calculated using a point charge interaction,moderated by an effective dielectric polarizability (er),

n q2GBaPP(n, t) GBlntrinsic t i (47rTo)sri,t [1]

where GBIntrinsic, is the GB of a molecule protonated at site t,and the second term is the sum of all Coulomb energyexperienced by the proton at that site in the multiply chargedion with n charges. These calculations have been described in

Table 2. Rates of proton transfer (x 10-12 cm3/mol-s) from individually isolated (M + nH)n+ charge states of disulfide-intact hen egg-whitelysozyme to neutral reference bases

GB, nkcal/

Reference base mol 4+ 5+ 6+ 7+ 8+ 9+ 10+ 11+

Cyclohexanone 194.0 0.87Acetophenone 198.0 1.0 7.02-Fluoropyridine 203.9 0.83 28 (0.60)3-Fluoropyridine 208.6 0.18 45 110 (0.60)Isopropylamine 213.7 1.3 7.0 (0.56) >1000

51 (0.35)Pyridine 215.7 0.54 >1000tert-Butylamine 216.7 1.1Diethylamine 221.4 5.1Dipropylamine 225.4 0.84 5.5 (0.82)

100 (0.18)Triethylamine 227.5 * * 4.0 9.5 (0.35)

110 (0.55)Tributylamine 231.1 * * 0.0 32 (0.35) 4.1 (0.32) >1000

6700 (0.16) 120 (0.68)DBN 237.4 0.60 8.9 t t (0.59) tDBU 239.7 6.9 62MTBD 243.3 16 (0.30) t

380 (0.25)Footnote symbols are as in Table 1. DBN, 1,5-diazabicyclo[4.3.0]non-5-ene; DBU, 1,8-diazabicyclo[5.4.0]undec-7-ene; MTBD, 7-methyl-1,5,7-

triazabicyclo[4.4.0]dec-5-ene.

Chemistry: Gross et al.D

ownl

oade

d by

gue

st o

n Ja

nuar

y 6,

202

1

Proc. Natl. Acad. Sci. USA 93 (1996)

Table 3. Assigned GBaPP values for disulfide-reduced anddisulfide-intact lysozyme

GBaPP, kcal/molCharge state Disulfide reduced Disulfide intact

16+, GBaPP (15+) 211.2 (1.0)15+, GBaPP (14+) 211.2 (1.0)14+, GBaPp (13+) 214.7 (1.0)13+, GBaPP (12+) 214.7 (0.40)

219.1 (0.60)12+, GBaPP (11+) 216.2 (0.10)

219.1 (0.70)223.4 (0.20)

11+, GBaPP (10+) 219.1 (1.0) 201.0 (0.60)211.2 (0.40)

10+, GBaPP (9+) 223.3 (0.85) 206.3 (0.30)229.3 (0.15) 214.7 (0.70)

9+, GBaPP (8+) 226.4 (0.90) 223.3 (0.20)229.3 (0.10) 226.5 (0.40)

229.3 (0.40)8+, GBaPP (7+) 229.3 (0.70) 229.3 (0.70)

234.3 (0.30) 234.3 (0.30)7+, GBaPP (6+) 234.3 (1.0) 229.3 (0.51)

234.3 (0.49)6+, GBaPP (5+) 234.3 (1.0) 234.3 (1.0)5+, GBaPP (4+) 234.3 (1.0) 238.6 (1.0)4+, GBaPP (3+) 234.3 (0.46) 241.5 (0.55)

For charge states with more than one reactive conformer, therelative abundances are shown in parentheses.

detail elsewhere (35); a brief description follows. Each residuein the protein is assigned a GB,ntri,sic value (28). Two extremesin ion conformation are modeled: a fully denatured structureis modeled as an elongated one-dimensional string in whichamino acid residues are separated by 3.8 A; the native structureis modeled using the x-ray crystal coordinates (36) with chargesassigned to the coordinates for the side chain nitrogens of thebasic amino acids and the backbone carbonyl oxygen for allother amino acids. The GBaPP of each charge state is calculatedfrom the lowest free-energy charge configurations found usinga pseudo-random walk algorithm (35). We assume that allcharge configurations within -3 kcal/mol of the lowest energyconfiguration will be present in the ion cell. The calculatedGBaPP we report is an average value for all these chargeconfigurations. The number of charge configurations dependson ion conformation and charge state-e.g., for the 12+ ionin the denatured conformation, we find 11 ways to assigncharge that are within this 3 kcal/mol range. Results of thesecalculations are shown in Fig. 4. The calculated GBaPP for bothlinear and native conformations decreases with increasingcharge state. This is due to both fewer available basic sites andhigher Coulomb repulsion with increasing protonation. Themore rapid decrease for the native conformation is due to thecloser proximity of charges in the native structure.

For the 11+ through 15+ charge states of the disulfide-reduced ions formed from a denaturing solution, our experi-mentally measured values fit those calculated for a completelydenatured structure, indicating that these ions remain un-folded in the gas phase (Fig. 4). The 7+ through 9+ ions eachhave two different structures or groups of structures, one ofwhich is consistent with a denatured ion conformation. Chargestates below 7+ have GBaPP values lower than those calculatedfor denatured ions. For the disulfide-intact ions, the measuredGBaPP of the major structure (60% abundance) of the 10+charge state (from proton transfer from the 11+ ion) isconsistent with the value calculated using the crystal coordi-nates. This value is 19.1 kcal/mol lower than that of the 10+disulfide-reduced ion. We conclude that this dramatic differ-ence in GBaPP must reflect differences in Coulomb energy due

to different ion conformations. The minor conformer (30%abundance) of the 9+ ion is also consistent with the valuecalculated using the crystal structure. The GBaPP of this ion is23.0 kcal/mol lower than the fully denatured conformer of the9+ disulfide-reduced ion. The less reactive conformers of boththe 9+ and 10+ ions are consistent with structures that arepartially unfolded but still significantly more compact than thecorresponding 9+ and 10+ disulfide-reduced ions. Distinctstructural intermediates are known to exist along the foldingpathway for lysozyme in solution (37). The relationship be-tween these structures and those observed in the gas phase isnot currently known.The GBaPP of the 8+ and lower charge states of both

disulfide-intact and disulfide-reduced ions show smaller dif-ferences in reactivity. The GBaPP of some of these ion confor-mations is the same, within experimental error, indicatingconversion to a common structure or structures with nearly thesame reactivity. Because the disulfide-intact ions are notphysically able to become completely elongated due to thecovalent disulfide bonds, this similar reactivity indicates thatthe lower charge state disulfide-reduced ions must have aconformation that is at least partially folded. These lowercharge states are not formed directly from the ESI process butrather by proton transfer from higher charge states to neutralreference bases in the ion cell. Initially, these higher chargestate ions are denatured (vide supra). Thus, these ions must befolding in the gas phase to produce a more compact structure.

The high GBaPP of the 5+ through 8+ disulfide-intact ionsis consistent with ions that are partially unfolded. Both the 9+and 10+ ions are formed directly from the electrosprayprocess, whereas the 8+ and lower charge states are formedprimarily by charge stripping. This indicates that charge strip-ping denatures these intermediate charge state ions.t This is incontrast to the results for the disulfide-reduced ions wherecharge stripping resulted in ion folding.t McLafferty andco-workers (16, 17) reported that gas-phase cytochrome c ionsformed by charge stripping can yield levels of hydrogen/deuterium exchange both lower and higher than ions formeddirectly by ESI. These results also indicate that charge strippingcan result in either folding or unfolding of an ion.The GBaPP of the 3+ and 4+ disulfide-reduced ions are

approximately 6.9 and 4.3 kcal/mol lower, respectively, thantheir disulfide-intact counterparts, indicating a more compactstructure for the reduced ions.§ This could be the result ofincreased flexibility of the disulfide-reduced form enhancingthe ability of these ions to find more favorable folding inter-actions than their more constrained disulfide-intact counter-parts.

For an ion to fold in the gas phase, the enhanced intramo-lecular interaction associated with ion folding (e.g., hydrogenbonding and van der Waals attraction between residues) mustexceed the combined effects of the increased Coulomb energy

tThere are <8 lowest energy configurations calculated for the crystalstructure for the 11 + through 8+ charge states. In contrast, there are23 lowest energy configurations calculated for the 9+ ion modeled asan unfolded structure but with the disulfide linkages intact. Thus, theunfolded state is statistically favored, which may contribute to ionunfolding.tAn alternative possibility that we are investigating is that the value of8r may be greater for a more compact ion structure. A value of Er =2.0 was used for these calculations. This value was obtained from thebest fit to experimentally measured GBaPP values for cytochrome c (Er= 2.0 + 0.2) modeled as a completely denatured ion (28). Acomparison of experimentally measured maximum charge states ofelectrospray generated ions to those predicted using our modelindicates that this same value is applicable to a wide variety ofproteins, including both disulfide-intact and disulfide-reduced ly-sozyme (35).§The error bars on the GBaPP values reflect estimates of error of theabsolute value; relative differences of 3 kcal/mol are clearly distin-guishable.

3146 Chmsr:Goseal

Dow

nloa

ded

by g

uest

on

Janu

ary

6, 2

021

Proc. Natl. Acad. Sci. USA 93 (1996) 3147

1.0

0.5

0.0

0 10 20 30 40 50 60Time, sec

FIG. 3. Kinetic data for proton transfer from the isolated 9+charge state of disulfide-intact lysozyme to neutral triethylamine (GB= 227.5 kcal/mol). The decrease of 9+ ion abundance (0) with time(Upper), fit using rate constants of 1.1 x 10-10 cm3/mol-s (55%) and9.5 x 10-12 cm3/mol-s (35%) with 10% not reacting, and a semiloga-rithmic plot of these data (Lower) are shown.

of a more compact structure and the unfavorable entropy offolding. Neglecting the effects of entropy, our results for theGBaPP of the 9+ and 10+ charge states indicate that the foldingstabilization energy of these ions is between 19 and 23 kcal/mol. Thus, it appears promising that the role of intramolecularinteractions in native gas-phase protein ions and foldingintermediates can be obtained from accurate measurements ofgas-phase ion conformation as well as folding entropy.

This method provides information only about the shape ofthe ion, analogous to the information obtained from collisionalcross section measurements (19, 20, 38). Thus, we cannot statethat the folded conformation is directly related to that insolution. In fact, a key difference between the ion conforma-tion in vacuum and that in solution is that charges areintramolecularly solvated in the gas phase, even for the highcharge states (27, 28, 35). In solution, the charged side chainresidues extend into water, where they are efficiently solvated.It has been suggested that protein ions may turn "inside out"in the gas phase (39). This would expose hydrophobic residuesto the vacuum, which is an apolar medium, leaving hydrophilicor polar residues on the inside. However, such a structure isunlikely due to the unfavorable Coulomb interactions betweencharged polar residues, which would strongly favor keepingthese residues at the exterior surface of the protein.A second important difference between the gas-phase and

solution conformations is that the number of charges and thecharge distribution for a native structure in solution and a"native" structure formed by ESI in the gas phase are quitedifferent. In aqueous solution at pH 7, all arginine and lysineresidues are protonated (17 total) and all carboxylic acids aredeprotonated (10 total), resulting in a net charge of 7+ (40).In contrast, the ions in this study are formed by adding protons.In the gas phase, the absence of a high dielectric solvent resultsin enhanced stability for ions in which the charges are widelydispersed. The existence of zwitterionic forms of gas-phaseamino acids has been shown to be energetically unfavorable(41-43). However, our value of Er greater than the vacuumpermittivity is consistent with polarization occurring so thatthe contribution of partial negative charge on many sites mayresult in ions that have electrostatic distributions more similarto those in solution than our simple model indicates.

CONCLUSIONSThe distinct reactivity of the gas-phase folded and partiallyfolded ions suggests that each of these ion populations iskinetically stable and, within the resolution of this method,

Charge state

FIG. 4. GBaPP of (M + (n - 1)H)(n-1)+ [measured by proton transfer from (M + nH)n+ ions to neutral reference bases] of disulfide-intact(-) and disulfide-reduced (0) hen egg-white lysozyme. Where more than one reactive conformer is present, its percent abundance is indicated nextto the data point (percentages of the disulfide-intact and disulfide-reduced conformers are indicated on the left and right sides of the data points,respectively). Error bars include the uncertainty in the value of GB of the reference bases. Calculated values of GBaPP of these ions modeled usingthe x-ray crystal structure coordinates (e) and modeled as a fully denatured one-dimensional string (A) using an Sr = 2.0 are shown. The dashedline indicates the GB of methanol (174.1 kcal/mol).

Chemistry: Gross et al.D

ownl

oade

d by

gue

st o

n Ja

nuar

y 6,

202

1

Proc. Natl. Acad. Sci. USA 93 (1996)

unique. Given the structural complexity of these protein ions,minor unresolved conformational differences within these ionpopulations will likely be present under these experimentalconditions. The presence of ion populations of intermediatereactivity suggests the possibility of distinct folding interme-diates in the gas phase. Determining how closely the confor-mations of these folded and partially folded ion populationsresemble those in solution, as well as investigating the confor-mation of gas-phase protein ions with various numbers ofsolvent molecules attached, should provide valuable insightinto the role of solvent in protein conformation and folding.

The authors acknowledge Professors J. I. Brauman, K. Sauer, R. J.Saykally, and Mr. W. D. Price for helpful discussions. Generousfinancial support for this research was provided by the Arnold andMabel Beckman Foundation (M 1652), the National Science Foun-dation (CHE-9258178), the Exxon Foundation (13605), Extrel FTMS,and Finnigan MAT through sponsorship of the 1994 American Societyfor Mass Spectrometry Research Award (E.R.W.).

1. Creighton, T. E. (1993) Proteins (Freeman, New York).2. Hillenkamp, F., Karas, M., Beavis, R. C. & Chait, B. T. (1991)

Anal. Chem. 63, 1193A-1202A.3. Chait, B. T., Wang, R., Beavis, R. C. & Kent, S. B. H. (1993)

Science 262, 89-92.4. Fenn, J. B., Mann, M., Meng, C. K., Wong, S. F. & Whitehouse,

C. M. (1989) Science 246, 64-71.5. McLafferty, F. W. (1983) in Tandem Mass Spectrometry (Wiley,

New York).6. Busch, K. L., Glish, G. L. & McLuckey, S. A. (1988) Mass Spec-

trometry/Mass Spectrometry (VCH, Deerfield, FL).7. Loo, J. A., Edmonds, C. G., Udseth, H. R. & Smith, R. D. (1990)

Anal. Chim. Acta 241, 167-173.8. Loo, J. A., Quinn, J. P., Ryu, S., Henry, K. D. & McLafferty,

F. W. (1992) Proc. Natl. Acad. Sci. USA 89, 286-289.9. Chowdhury, S. K., Katta, V. & Chait, B. T. (1990) J. Am. Chem.

Soc. 112, 9012-9013.10. Mirza, U., Cohen, S. & Chait, B. T. (1993) Anal. Chem. 65, 1-6.11. Loo, J. A., Edmonds, C. G., Udseth, H. R. & Smith, R. D. (1990)

Anal. Chem. 62, 693-698.12. Guevremont, R., Siu, K. W. M., LeBlanc, J. C. Y. & Berman,

S. S. (1992) J. Am. Soc. Mass Spectrom. 3, 216-224.13. Loo, J. A., Ogorzalek, R. R., Udseth, H. R., Edmonds, C. G. &

Smith, R. D. (1991) Rapid Commun. Mass Spectrom. 5, 101-105.14. Katta, V. & Chait, B. T. (1991) Rapid Commun. Mass Spectrom.

5, 214-217.15. Miranker, A., Robinson, C. B., Radford, S. E., Aplin, R. T. &

Dobson, C. M. (1993) Science 262, 896-900.16. Suckau, D., Shi, Y., Beu, S. C., Senko, M. W., Quinn, J. P.,

Wampler, F. W. & McLafferty, F. W. (1993) Proc. Natl. Acad.Sci. USA 90, 790-793.

17. Wood, T. D., Chorush, R. A., Wampler, F. M., Little, D. P.,O'Connor, P. B. & McLafferty, F. W. (1995) Proc. Natl. Acad.Sci. USA 92, 2451-2454.

18. Winger, B. E., Light-Wahl, K. J., Rockwood, A. L. & Smith,R. D. (1992) J. Am. Chem. Soc. 114, 5897-5898.

19. Covey, T. & Douglas, D. J. (1993) J. Am. Soc. Mass Spectrom. 4,616-623.

20. Cox, K. A., Julian, R. K., Cooks, R. G. & Kaiser, R. E. (1994) J.Am. Soc. Mass Spectrom. 5, 127-136.

21. Loo, R. R. 0. & Smith, R. D. (1994) J. Am. Soc. Mass Spectrom.5, 207-220.

22. Loo, R. R. O., Winger, B. E. & Smith, R. D. (1994) J. Am. Soc.Mass Spectrom. 5, 1064-1071.

23. Cassady, C. J., Wronka, J., Kruppa, G. H. & Laukien, F. H.(1994) Rapid Commun. Mass Spectrom. 8, 394-400.

24. McLuckey, S. A., Van Berkel, G. J. & Glish, G. L. (1990) J. Am.Chem. Soc. 112, 5668-5670.

25. Ikonomou, M. G. & Kebarle, P. (1992) Int. J. Mass Spectrom. IonProcesses 117, 283-298.

26. Gross, D. S., Rodriguez-Cruz, S. E., Bock, S. & Williams, E. R.(1995) J. Chem. Phys. 99, 4034-4038.

27. Gross, D. S. & Williams, E. R. (1995) J. Am. Chem. Soc. 117,883-890.

28. Schnier, P. D., Gross, D. S. & Williams, E. R. (1995) J. Am.Chem. Soc. 117, 6747-6757.

29. Little, D. P., Speir, J. P., Senko, M.W., O'Connor, P. B. &McLafferty, F. W. (1994) Anal. Chem. 66, 2809-2815.

30. Rockwood, A. L., Busman, M. & Smith, R. D. (1991) Int. J. MassSpectrom. Ion Processes 111, 103-129.

31. Marshall, A. G., Wang, T. C., Ricca, T. L. (1985) J. Am. Chem.Soc. 107, 7893-7897.

32. Meot-Ner, M., Hamlet, P., Hunter, E. P. & Field, F. H. (1980) J.Am. Chem. Soc. 102, 6393-6399.

33. Lias, S. G., Liebman, J. F. & Levin, R. D. (1984) J. Phys. Chem.Ref. Data 13, 695-808.

34. Decouzon, M., Gal, J. F., Maria, P. C. & Raczynska, E. D. (1993)Rapid Commun. Mass Spectrom. 7, 599-602.

35. Schnier, P. D. & Gross, D. S. (1995) J. Am. Soc. Mass Spectrom.6, 1086-1097.

36. Blake, C. C. F., Koenig, D. F., Mair, G. A., North, A. C. T.,Phillips, D. C. & Sarma, V. R. (1965) Nature (London) 206,757-763.

37. Radford, S. E., Dobson, C. M., Evans, P. A. (1992) Nature (Lon-don) 358, 302-307.

38. von Helden, G., Wyttenbach, T. & Bowers, M. T. (1995) Science267, 1483-1485.

39. Wolynes, P. (1995) Proc. Natl. Acad. Sci. USA 92, 2426-2427.40. Honig, B. & Yang, A. S. (1993) J. Mol. Biol. 231, 459-474.41. Locke, M. J. & McIver, J. R. T. (1983) J. Am. Chem. Soc. 105,

4226-4232.42. Ding, Y. B. & Krogh-jespersen, K. (1992) Chem. Phys. Lett. 1992,

261-266.43. Yu, D., Armstrong, D. A. & Rauk, A. (1992) Can. J. Chem. 70,

1762-1772.

3148 Chmsr:GosealD

ownl

oade

d by

gue

st o

n Ja

nuar

y 6,

202

1