Gas-Phase Hydrogen/Deuterium Exchange asa Molecular Probe for the Interaction ofMethanol and Protonated Peptides

Eric Card, M. Kirk Green, Jennifer Bregar, and Carlita B. LebrillaDepartment of Chemistry, University of California, Davis, California, USA

The gas-phase hydrogen/deuterium (H/D) exchange kinetics of several protonated aminoacids and dipeptides under a background pressure of CH 30D were determined in anexternal source Fourier transform mass spectrometer. H/D exchange reactions occur evenwhen Ihe gas-phase basicity of the compound is significantly larger (> 20 kcal /mol) thanmeIhanol. In addition, greater deuterium incorporation is observed for compounds that havemultiple sites of similar basicities. A mechanism is proposed Ihat involves a structurallyspecific intermediate with extensive interaction between the protonated compound andmethanol. (J Am Soc Mass Spectrom 1994, 5, 623-631)

The production of ionic gas-phase macro­molecules allows the possibility of studying thesecomplex systems in Ihe absence of solvent. Al­

though the gas and solvated phases are intrinsicallydifferent, there is evidence to suggest Ihat conforma­tion may be retained by the molecule even in the gasphase [1-5]. Isolating the molecule makes it easier toinvestigate without the interference of solventmolecules and at significantly greater sensitivities.However, there is no effective probe for observingstructures of complicated gas-phase biomolecules. De­structive probes such as collisionally activated dissoci­ation (CAD) [6t which require vibrational excitationand subsequent fragmentation, have provided infor­mation on the structure of smaller ions. Unfortunately,not only is CAD ineffective wiIh large macro­molecules, it also provides essentially no informationon the conformation of the molecules. In this regard,hydrogen/deuterium (H/D) exchange offers a uniqueopportunity because it is both a nondestructive andlow energy molecular probe [7-17): Indeed, severalgroups have shown Ihat conformational effects may beobservable by using gas-phase H/D exchange [3-5,16).

Despite the utility of gas-phase H/D exchange andits potential as a tool for structural elucidation, itsmicroscopic mechanism is not well understood. Eluci­dating the mechanism is, however, important if H/Dexchange is to be a useful tool. In H/D exchangeinvolving protonated species, the role of gas-phasebasicity is apparent. H/D exchange has been shown tooccur when the gas-phase basicities of the reagent andthe conjugate base of Ihe protonated species are similar

Address correspondence to Carlito B. Lebrilla, Department of Chem­istry, University of California, Davis, CA 95616.

© 1994 American Society for Mass Spectrometry1044-0305 j94j$7.00

[10, 11, 18, 19). Investigations by Ausloos and lias [10)have shown Ihat for protonated compounds H/D ex­change reactions do not occur when the proton affinityof the neutral base is greater Ihan the deuteratedreagent by more Ihan 20 kcalyrnol. The correlationbetween basicity and reactivity has led many people toconclude that H/D exchange in protonated speciesproceeds by proton transfer reactions within the dimercomplex as outlined in Scheme I (7). The intermediatesare commonly thought to reside in a multiple-wellpotential where proton transfer reactions occur [20).For example, in the mechanism shown, the two speciesform an ion-dipole complex in which the proton mi­grates from the primary base site B to the deuteratedreagent A. Reverse transfer of the deuterium producesthe H/D exchange.

Recent exceptions to Ihe 20-kcal/mol limit havebeen shown wiIh pep tides. H/D exchange reactionshave been observed between protonated renin sub­strate and ND3 despite differences in proton affinitiesof greater Ihan 50 kcal /rnol [16). Similarly, the H/Dexchange reactions of several protonated amino acidsand CH 30D occur despite basicity differences of be­tween 30 and 40 kcalyrnol (21). We propose that Ihelimit on Ihe endothermicity of Ihe reaction observed byAusloos and Lias applies only to organic compoundsIhat contain a single functional group (monofunctionalcompounds). Protonated species with multiple func-

Scheme I. The commonly proposed mechanism for HjD ex­change involving proton transfer reactions between protonatedspecies and deuterated reagents such as CH 30D, O2°'and ND 3•

The mechanism involves the formation of a dimer (designated bydashed lines) and proton transfer within the complex.

Received January 17, 1994Revised February 21, 1994

Accepted February 22, 1994

624 GARDET AL. J Am Soc Mass Spectrom 1994, 5,623-631

tional groups (multifunctional compounds) undergoHID exchange with deuterated reagents even whendifferences in gas-phase basicities are significantlygreater than 20 kcaly'mol. We infer this to mean thatdimeric interactions that involve monofunctional andmultifunctional groups are different and that a uniqueinteraction exists between the protonated multifunc­tional compounds and deuterated reagents that allowsHID exchange to occur even when the differences inbasicity are large. Probing this interaction, particularlyfor compounds such as amino acids and peptides, isdifficult but necessary for determining the mechanismof HID exchange in complex gas-phase proteins.

In an earlier publication, we reported the HIDexchange behavior of amino acids containing alkyl sidechains (Structure 1) [21]. We noted that although pro­tonation occurs primarily on the terminal amine, themost rapid exchange occurs on the carboxylic acid.This observation was confirmed by comparing the rateconstants of the amino acids and their methyl esterderivatives. Exchange on the amino group is slow forglycine methyl ester and nearly nonexistent (at leastunder the same reaction conditions) for the esters ofamino acids with alkyl side chains such as alanine,valine, leucine, and proline. This situation already il­lustrates that the most favorable site of protonationdoes not necessarily become the sole or even primarysite of exchange. We further observed that the rate ofthe first exchange does not vary considerably betweenthe different amino acids in the group. The rate of thesecond exchange depends strongly on the basicity:increasing basicity results in decreasing exchange rates.Thus glycine, which is 3.2 kcal /mol less basic thanalanine, has a rate constant for the second exchangethat is three times greater.

HZ~CHO

IAR OH

In this article, we propose a mechanism for the HIDexchange reactions of CH 30D and protonated aminoacids and peptides that involves a structurally specificlong-lived ion-molecule complex responsible for theHID exchange both on the protonated basic sites andon less basic sites far removed.

Experimental

Experiments were performed in an external ion sourceFourier transform mass spectrometer equipped with aquadrupole ion guide that uses commercially availablecompounds [22). The experimental procedures usedfor HID exchange have been described in an earlierpublication [21]. Ions were produced via liquid sec­ondary ion mass spectrometry (LSIMS), isolated viaconventional ejection techniques, and allowed to react

with a background of CH300 (Aldrich Chemical Co.,Milwaukee, WI; 99.5% isotopic purity) maintained bya leak valve between 1 and 5 X 10-7 torr. The actualabundance of CH30D was diminished by HID ex­change that occurred on the chamber wall. It would beideal to "season" the chamber by allowing it to reactwith D20 before the experiment and HzO immediatelyafter. This step would allow the subsequent use of theinstrument for other experiments, for example, brack­eting measurements, which require that the system berelatively deuterium-free. However, seasoning the in­strument with DzO followed by a bakeout requires aminimum of several hours to be effective and may stillnot provide totally hydrogen-free conditions, in whichcase it would still be necessary to determine the HIDratio of the system. Furthermore, there is a desire todevelop this technique for routine use without inten­sive preparation. Pressure calibration was performedby using parametrized ion gauge response factors [23].All experiments were performed under ambient tem­peratures (300 K).

Analysis of the data and determination of the rateconstants have also been described earlier [21]. Ioninjection, detection, and data collection were carriedout under the control of Omega (Ionspec, Irvine, CA)software. The data consist of relative intensities of themass peaks for Do, D1,... , Dn, where D, represents aprotonatedydeuteronated species with n deuteriumatoms, measured at a number of time points (typically10) that correspond to different reaction times evenlyspaced over a preset interval. A program written inMathematica (Wolfram Research, Champaign, IL) firstnormalizes the intensities for each time point anddeletes mass peaks whose summed intensity falls be­Iowan arbitrary value (usually 0.03). A series of n + 1coupled differential equations, of the form of eqs (1)-(5)[for a system with four exchangeable hydrogen atomsassociated with the exchange reactions (6)-(9)], thatincorporate initial estimates of the rate constants, isthen solved numerically by using a built-in subroutineto yield n + 1 curves that describe the time behaviorof the different deuterated species:

d[Do]- -at = kl[Do][CHPO] - LdDtl[CHPH] (l)

d[DIl-~ = kl[Do][CHPO] - L1[D1][CHPH]

dt- kz[D1][CHPDl + L z[Dz][CH30H]

(2)

d[DJ-at ~ kz[DI][CHPD] - L Z[D2][CHPH]

- k3[D2][CHPD] + L 3[D3][CH30H)

(3)

J Am Soc MassSpectrom1994, 5, 623-631 GAS-PHASE HID EXCHANGE OF PEPTlDES 625

Rate parameters are optimized. by doing a least squaresfit of the curves to the experimental points using theMarquardt [24] algorithm as described by Bevingtonand Robinson [25]. The differential equations are re­solved at each iteration of the fitting procedure. Al­though reactions (6)-(9) are written as equilibriumreactions, in reality equilibrium is not obtained. Notethat the reverse reaction accounts for the constantpressure of background CH 30H during the experi­ment. The treatment, however, is strictly kinetic aswould be observed clearly if the CH 30H pressure inthe instrument was zero.

A significant abundance of CH 30H was present inthe background gas. This impurity varied to as muchas 30% of the total abundance and its contribution wastreated. as an additional unknown variable to be opti­mized during the curve fitting procedure. The d o-

methanol is believed to come from HID exchangereactions that occur on the surface of the vacuumchamber. This hypothesis is confirmed by the decreaseof the CH,,oH contribution in experiments performedlater in the day. As further evidence, reference gasesfor bracketing measurements to obtain gas-phase ba­sicity values subsequently introduced into the vacuumchamber also incorporated deuterium [26].

The deviation in the absolute values is approxi­mately 40% due to uncertainties in the pressure deter­mination. Deviations in the relative values, calculatedfrom multiple determinations, were found to be sig­nificantly smaller and are listed in Table l.

Isotope effects were assumed. to be negligible in thecalculation of the rate constants. This assumption issupported by earlier observations obtained in the re­verse reaction that involves deuteronated glycine andCH 30H [21]. Rate constants of the reverse reaction didnot vary greatly from the corresponding forward reac­tions. Similar conclusions concerning isotope effectshave been reported for other gas-phase HID exchangereactions [10, 12, 13].

Results

The time dependence behavior of the deuterated prod­ucts is shown for the two protonated amino acids,serine and histidine, that react with a backgroundpressure of CH 30D (Figures 1 and 2, respectively).The two compounds illustrate the contrasting behavior

Table 1. Measured rates for the HID exchange of selected protonated amino acids and peptides reacting with a backgroundpressure of CHgOD. For comparison, the ADO [28,291 rate for glycine with CHgOD is 1.59 X 10- 9 cm3 s~l.

All rates are k X 10- 11 crrr' molecule-I 5- 1. The letters "nrx" mean reactions not observed during reaction period

Compounds Labile H" k1 k 2 k 3 k 4 k. k.

Glycineb 4 10.6 ± 2.1 3.7 ± 0.7 2.6 ± 0.4 1.2 ± 0.1Glycine ester b 3 1.9 1.2 0.7

Alanine" 4 8.4±1.5 0.8 ± 0.3 nrx nrxAlanine ester b 3 1.3 nrx nrxThreonine 5 12.7 ± 2.5 0.4 ± 0.2 nrx nrx nrxSerine 5 11.7 ± 5.2 0.7 ± 0.2 nrx nrx nrx

Cysteine 5 18.7 ± 5.6 0.5 ± 0.1 nrx nrx nrxLysine 6 9.5 ± 1.8 8.1 ± 1.8 4.7 ± 1.5 4.7 ± 1.9 2.1 ± 0.5 1.5 ± 0.8Histidine 5 14.2 ± 3.2 9.2 ± 0.4 3.6 ± 0.1 2.0 ± 1.0 1.3 ± O.BHistidine ester 4 15.3 ± 5.4 13.0 ± 1.7 3.2 ± 1.0 1.9 ± 1.0Histamine 4 9.B ± 1.9 6.2 ± 0.5 2.3 ± 0.4 1.0 ± 0.6Diglycine 5 20.8 ± 1.2 20.4 ± 2.9 10.2 ± 3.5 4_8 ± 3.0 1.5 ± 0.4Dialanine 5 15.4 ± 4.9 13.7 ± 3.0 5.4 ± 1.3 4.1 ± 1.3 0.4 ± 0.2Gly-Ala 5 33.3 ± 5.0 45.8 ± 4.6 13.3 ± 1.6 6.3 ± 0.4 4.2 ± 3.0Ala·Gly 5 31.2 ± 3.2 36.6 ± 5.0 7.5 ± 1.0 4.4 ± 1.0 1.8 ± 0.8Arginine 8 nrx nrx nrx nrx nrx nrx4-Amino

benzoic acid 4 nrx nrx nrx nrxN-N-Dimethyl-4-

aminobenzoic acid 2 nrx nrx

aNumber of hydrogen atoms bound to hetero atoms including the proton.bValues obtained from ref. 21.

626 GARDET AL. J Am Soc Mass Spectrom 1994,5,623-631

" ,---~---~--_---r- --~--~De

D.4

1.0

,r

/0

I

I

I

I

I0.2 I

II

00'

D'

• D'

~ - _.- - . - .- --- ..: - ---.

__~ ~ _ ~...,... __ ~ -'i'" __ • _~_ - - - ~_.- - z,.- - -:- -_·0

a

1'04

,10g

i I f i104 toe

10

TimEl/S@C

Figure 1. The time dependence behavior of the relative intensi­ties of ions in the molecular ion region of protonated serinereacting with a background pressure of CH 30D (- 3 X 10-7

torr) up to a reaction time of 11 s. During this time period onlythe singly and the doubly deuterated products are observed. Thecurves represent the fitting functions used to determine thevalues of k.

C) ~

~~164 i I lJ~' I' 16.. i I f 1[~e '

0,

I i I104 108

Figure 3. Partial mass spectra of serine in the molecular ionregion under a pressure of 2.0 X 10- 7 torr CH 30D after a reac­tion time of (a) 0.20 s, (b) 1.29 s, (c) 2.38 s, (d) 3.47 S, and (e) 4.55s. Ion abundances are scaled relatively,

10-7 torr CH 30 D (Figures 3 and 4). Other ions pro­duced during LSIMS ionization were typically ejectedto prevent interference. Amino acids with polar weaklybasic side chains including serine [R = CH20 H ], cys­teine [R = CH 2SH], and threonine [R = CH(CH3)OH]have HID exchange behavior similar to amino acidswith alkyl side chains such as alanine, valine, leucine,etc., despite the presence of an additional exchange­able hydrogen. We have shown previously that proto­nated amino acids with alkyl side chains have thefastest exchange on the carboxylic acid, although pro­tonation occurs primarily on the amine. The behaviorof the rate constants for serine, cysteine, and threonineis similar to glycine, alanine, and other amine acidswith alkyl side chains. Although no direct evidencehas been found yet, we feel that the similarities inkinetics are indications of similar mechanistic behav­ior, that is, HID exchange also occurs initially on thecarboxylic acid.

Amino acids that contain strongly basic side chains,such as lysine and histidine, constitute a separategroup. Protonated lysine exchanges all six labile hy­drogen atoms whereas histidine (as already shown)rapidly exchanges four out of five labile hydrogens.These compounds do not necessarily exchange fasteston the carboxylic acid positions as shown by histidine

D'

o D'

- - - - I ,,,_ .:7.. '-"<>--'~- _- _

_,_,~. _' .!J._._- ~_.-' - ..-'-

e '0

-; -~.:.:.-;.~ ~.~.:::'{ .'- , .... - ..~ ..... " ..-....-...-- --

within the group of amino acids and peptides investi­gated for this report. Both contain the same number(five including the proton) of exchangeable hydrogenatoms, that is, hydrogen atoms bound to hetero atoms.Protonated serine undergoes rapid exchange of a sin­gle hydrogen atom followed by a slow exchange of asecond hydrogen atom with no further exchange ob­served during the 10-s reaction time. By comparison,protonated histidine rapidly exchanges a single hydro­gen followed by similarly fast second, third, and fourthexchanges. Exchange of the fifth hydrogen occurs butis not readily observed during this reaction period.

A series of partial mass spectra of the two com­pounds around the molecular ion region is shown aftervarious reaction times under pressures of about 3 X

MH'

o c'c'

• e'

Figure 2. The time dependence behavior of the relative intensi­ties of ions in the molecular ion region of protonated histidinereacting with a background pressure of CH 30D (-3 X 10- 7

torr) up to a reaction time of 10 s. Deuterated products corre­sponding to exchange of all labile hydrogen atoms are observed.The curves represent the fitting functions used to determine thevalues of k.

~ 0.6

i;

"I 0.4 I

II

•J, ,"

JAm Soc Mass SpeelTom 1994, 5, 623-631 GAS-PHASE HID EXCHANGE OF PEPTIDES 627

Figure 4. Partial mass spectra of histidine in the molecular ionregion under a pressure of 3.8 X 10-7 lorr CH 30D after a reac­tion time of (a) 0.30 s, (b) 1.34 s, (c) 2.39 S, (d) 3.96 s, and (e) 5.00s. Ion intensities are scaled rela tively.

tion, the number of deuterated products, and the indi­vidual deuterated products. For example, a better fit isobtained with serine, which has only three differentmass species, than histidine, which has six. It shouldbe noted that the fit for histidine is possibly the worst,for reasons yet unknown, and this compound wasselected only to illustrate the contrasting behavior. Thevalues of k reported for all compounds are averagedfrom both short and long reaction times. The deviationbetween the fit and the experimental points may origi­nate from several factors including the relaxation ofions in the cell during the reaction and the correlationof ion motion, particularly between ions with similarmasses and frequencies.

The pressure dependence behavior of HID ex­change reactions was determined for protonatedglycine between CH 30D pressures of 1 and 5 X 10-7

torr. This pressure range represents the practical exper­imental range with this instrument. At lower pres­sures, the reactions proceeded too slowly to be ob­served and at higher pressures, collisions in the cellincrease the magnetron radius of the ion cloud, whicheventually produces ion loss [27]. Within this pressurerange, the k values varied randomly by less than 20%,indicating the independence of the exchange to thepressure of CH 30D.

Within the group of the amino acids, HID ex­change reactions of the first hydrogen (kl ) occur withrate constants between 10 and 20 X 10-11 em3 mole­cule -1 s -1. When the values are compared to thetheoretical ADO (average dipole orientation) [28, 29]rate of 1.6 x 10-9 ern" s -I, they correspond to reactionefficiencies of 10-20%. The dipeptides have consis­tently higher k1 values compared to the amino acids, alikely consequence of higher collision rates and colli­sion cross sections due to the larger size of the ion.These factors are not considered in the calculation ofADO rates.

The major differences in HID exchange behaviorbetween the amino acids and dipeptides are observedin the higher order exchanges, that is, kn where n > 1.For compounds observed to undergo rapid multipleHID exchange, the values of k2 1 for example, areequivalent to kl • With protonated compounds that arenot observed to undergo rapid HID exchange, k2values are small and typically less than 0.05k 1. In thisregard, k21 k l is a general and simple measure ofmultiple HID exchange. For example, the k21kl ratiofor histidine is approximately 0.6, whereas for serinethe corresponding value is 0.04.

The reactions of glycine and alanine methyl estershave been presented earlier and are included for directcomparisons [21]. The similarities of the k2 values ofthe acids to the kl of the esters and the statisticalnature of k2 , k3, and k4 of glycine provide strongevidence that the first exchange in these amino acidsoccurs with the carboxylic acid hydrogen. The resultsalso indicate that exchange on basic and acidic posi­tions proceeds independently.

;iII I

~'~...J 0;~\~ _Iii I i I

158 16Z

~

II. ill!'! I it

~JlJ I~L-Ii 'I' .. "I f i j Ii

159 tb2

bl

OJ

I1&2

ii'l Iiiii I I159 1162

i I I I I...

and its derivatives, histidine ester and histamine. Allthree have comparable rates for the first exchangedespite the absence of carboxylic acid hydrogens in thelatter two compounds.

Rate constants (k n) associated with the incorpora­tion of the nth deuterium for selected amino acids anddipeptides are tabulated (Table 1). Statistical factorswere not included because the contributions of ex­change on specific sites to the overall rates are notknown. The deviations in the reported values corre­spond to 1 standard deviation from a set of multipledeterminations. For compounds that undergo rapidmultiple exchanges, rates of fourth or fifth deuteriumincorporation were often slow and could not be calcu­lated with high precision because of the diminishingintensity of the signal with time. Significantly higherpressures, which would allow slow reactions to bebetter represented, could not be used because theytended to degrade the quality of the spectra. Alterna­tively, increasing the trapping times at lower pressuredid not allow the reactions to be better representedbecause ion loss also eventually degraded the spectra.

The time plots of serine and histidine (Figure 1 and2) show the fitting curves from the calculated values ofk's, The fits, though satisfactory, are not exact andappear to vary depending on the degree of deutera-

628 GARD ET AL. J Am Soc Mass Spectrom 1994,5,623-631

This compound [Reaction (12)] does not undergo ex­change suggesting the importance of the amine­carboxylic acid combination:

H H OH" !> I(:yct'zV + CH,oD - No Reaction

These results are further supported by the HID ex­change behavior of serine (and to some extent cysteineand threonine), which has similar l-amino, 2-hydroxylfunctionalities and does not appear to undergo HIDexchange on the hydroxyl side chain.

Discussion

The rapid exchange of the carboxylic acid hydrogensuggests a single reaction site within the protonatedmolecule. Initial exchange of the carboxylic acid hydro­gen followed by rapid intramolecular HID exchangebetween the amine and carboxylic acid sites is a possi­ble mechanism, This mechanism, however, requiresthe carboxylic acid always to be present for the reac­tion to occur, inconsistent with the reactivity of theester analogs,

Alternatively, proton migration may occur from theN-terminus to the carboxylic acid where it undergoes

(11)Htf.... ~N OH

+ + CH30D - 1-2 H's Exchanged

nificant HID exchange. However, reaction conditionsin the triple quadruple instrument used by Ranasinghet al. differ from those in Fourier transform massspectrometry (FTMS). The ions in that study wereproduced by both electron impact and chemical ioniza­tion methods, which are believed to produce ions moreenergetically excited than fast-atom bombardment pro­duced ions. In addition, the pressures of CH 30D usedduring the reaction (3 mtorr) are nearly 4 orders ofmagnitude larger than in this investigation, producinga significantly greater number of two-body collisionsand even allowing three-body collisions. Mixed proto­nated dimers and in some cases trimers are observedin the spectra. In contrast, dimers are not observed inthe FTMS spectra. This means that conditions in theFTMS favored interactions of the protonated specieswith a single methanol molecule, whereas simultane­ous interactions with several methanol molecules werepossible in the high pressures used in the triplequadrupole instrument.

Other model compounds with features similar toamino acids were investigated to determine the impor­tance of the amine-carboxylic acid combination. Thecompound 2-piperidinemethanol is analogous to pro­line [Reaction (11)] without the carboxylic acid:

The reaction behavior of the dipeptides, lysine, andhistidine, which all undergo facile multiple exchange,suggests similarities in the interactions of these com­pounds with the methanol reagent. The dipeptidesglycylalanine and alanylglycine have k2 values greaterthan k 1• The significance of this is not known, but theresults are consistent with rapid multiple deuteriumexchange. The contrast between alanine and the ala­nine-containing pep tides is unexpected but providesinsight. Protonated alanine incorporates, at most, twodeuterium atoms, whereas protonated dialanine incor­porates five deuterium atoms-the total number oflabile hydrogen atoms-during the same reaction pe­riod. This trend is readily evident in the tabulated rateconstants. It further appears that the position of thealanine in the dipeptide is unimportant. Both alanyl­glycine and glycylalanine compounds undergo rapidmultiple exchange.

The lack of correlation between the extent of HIDexchange and molecular gas-phase basicity is in appar­ent contradiction to the results of Ausloos and Lias.For example, histidine and lysine are more basic thanmethanol by 50 and 48 kcaljmol [18], respectively, butthey also exchange a greater fraction of their labilehydrogens compared to serine (34.9 kcal z'mol). Sirni­lady, alanine, which is 32.5 kcaljmol [24, 25] morebasic than methanol, exchanges a smaller fraction of itslabile hydrogen compared to dialanine, which is 38.2kcaljmol more basic than methanol [26, 30]. Gas-phasebasicity again does not appropriately explain the trendsobserved with multifunctional compounds. The excep­tion is perhaps arginine, estimated to be over 60kcaljmol more basic than methanol [29], which doesnot undergo significant HID exchange [31]. The struc­tural similarities between arginine and lysine suggestthat structural factors may not be the primary reasonfor the lack of reactivity, but may instead represent anultimate limit for the differences in gas-phase basicitybetween the parent compound and the deuteratedreagent for HID exchange to occur.

To obtain further information on the importance ofindividual functional groups in HID exchange reac­tions, several model compounds were investigated.From the failure of 4-aminobenzoic acid and itsdimethyl derivative N,N-dimethyl-4-aminobenzoicacid to undergo HID exchange [Reaction (10»), it wasconcluded that neither an isolated protonated aminenor an isolated carboxylic acid group readily under­goes HID exchange:

H'NVCO'H + CIl 30D - NoReaction (10)These results also preclude a simple four-center mech­anism for the exchange. Close contact between theamine and the carboxylic acid is apparently necessaryfor HID exchange to occur. Ranasinghe [23] haveshown that protonated 4-aminobenzoic acid formedunder chemical ionization conditions undergoes sig-

J Am Soc Mass Spectrom 1994,5,623-631 GAS-PHASE HI D EXCHANGE OF PEPTIDES 629

exchange, Exchange rates, in this situation, would de­pend on the proton migration aptitude and, in tum, onthe gas-phase basicity difference between the amineand carboxylic acid group. For glycine, theory predictsa reaction that is about 12 kcalyrnol endothermic [26].Amino acids that contain alkyl side chains are ex­pected to increase this difference because the aminegroups are more susceptible to the inductive effects ofthe alkyl groups than the acid [18]. Consequently,basicity should play a stronger role by allowing lessbasic compounds to be more reactive than more basiccompounds-a trend not supported by experimentalresults. Proton migration could be extremely rapid,particularly with vibrationally hot ions, and Signifi­cantly faster than the collision rate. This scenario couldbe consistent with the observed independence of therates on pressure over the narrow range investigated,which would mean, however, that all labile hydrogensare equivalent, making the values of the rate constantsstatistical.

Although exchange can occur away from the site ofprotonation, results still point to the importance of therelative proximity of the functional groups. An iso­lated carboxylic acid and a protonated amine do notundergo HjD exchange with CH 30D as illustrated by4-aminobenzoic acid, its dimethyl derivative, and 2­piperidenemethanol. The least basic of these com­pounds (4-aminobenzoic acid) is estimated to be 26kcalyrnol more basic than methanol and would not beexpected to react according to the results of Lias andAusloos, Many of the weakly basic amino acids fall inthe same basicity range as the benzoic acids but aresignificantly more reactive. Similarly, there appears tobe no HjD exchange reactivity for an unprotonatedcarboxylic acid group under these reaction conditions,even when it is nearly 10 kcalyrnol more acidic thanglycine, for example, 4-aminobenzoic acid. Steric inter­ference also does not explain the low reactivity ofcompounds with alkyl side chains such as alanine,valine, isoleucine, etc. because the dipeptides alanyl­alanine, alanylglycine, and glycylalanine all undergoextensive HjD exchange in their protonated form.

The importance of the spatial relationship betweenthe carboxylic acid and amine functional groups hasalready been shown by Ranasinghe et al. [32] in thereaction of polyfunctional compounds with CH30Dand ND3• The authors observe greater multiple deuter­ation when the amine and carboxylic acid groups areortho rather than para. When the two groups areisolated, they are subject to the same constraints foundin the HjD exchange of compounds that contain asingle functional group. This suggests that HjD ex­change in compounds such as amino acids and pep­tides must involve multiple interactions between thedeuterated reagent and the protonated peptide.

An intermediate involving extensive hydrogenbonding and dipolar interactions between the proto­nated amino acid and CH 30D is proposed for glycine(Intermediate I); This intermediate takes advantage ofthe true amphoteric nature of methanol, that is, acting

both as a proton acceptor and a proton donor. Asimilar bicoordinated intermediate is predicted by abinitio calculations to occur in complexes of NHt and1,2~iols [33].With the amino acids, protonation occursprimarily on the terminal amine, which is predicted tobe 12 kcaljmol more basic than the carbonyl positionfor glycine [25]. Hydrogen bonding interactions canoccur between the protonated amine and the oxygenatom of the methanol (N- H ... 0) and between thedeuterium and the carbonyl group of the carboxylicacid (C=O ..· D). From Intermediate I, HjD ex­change can occur on the carboxylic acid and the termi­nal amine with independent rates. Replacing the car­boxylic acid with a methyl ester does not significantlyaffect the rates of exchange on the amine. This interme­diate further allows exchange to occur on the car­boxylic acid even though the site of protonation isthe terminal amine. A reasonable mechanism for ex­change on the carboxylic acid involves breakage of theCH30-D bond in Intermediate I to produce the ringopen Intermediate II (Scheme II). In this configuration,the CH 30 group can rock back and forth while at­tached to the protonated amine. Rotation along thebond a to the carbonyl allows exchange of the car­boxylic acid hydrogen to occur (Scheme II, pathway a).This mechanism explains the similarity in k, for all theamino acids, particularly those containing a singledominant base site.

Multiple HjD exchanges on the terminal amineproduce multiply deuterated products. We have al­ready shown for amino acids with R = H and alkylthat the rate of HjD exchange on the terminal aminedecreases with increasing amino acid basicity [21]. Forexample, the k2 value of protonated glycine is nearlyfive times greater than that of protonated alanine. Thisbehavior is possible given Intermediate I and the rela­tive strength of the N - H bonds and c=o ..·D

:)-~-D.... 'b) D-,, 'b'\

~"~C~--lOH )r-tc~-J--OH~ HS\~:LODI II

+CIi,oHll bCII;OH

D, + D 0

H, 'p = H~. + IIH'""""N<,C~~OH H/N............C~2-------0H

m

Scheme II. The proposed mechanism for the HID exchangereaction of CH 30D and protonated glycine.

630 GARD ET AL. JAm Soc Mass Spectrom 1994,5. 623-631

Table 2. Differences in proton affinities between the two mostbask sites estimated by the use of model compounds withknown values. Values for proton affinities are obtainedfrom ref. 18

affinities can be illustrated when the proton affinities ofthe actual sites are known. This is possible only byperforming time-consuming and expensive high levelmolecular-orbital calculations. Alternatively, protonaffinities of various base sites can be estimated byusing model compounds with known proton affinitiesand having functional groups similar to those foundon the amino acid or peptide. These values are listed inTable 2 along with the model compounds (denoted byItalic type) used to represent the peptides. For exam­ple, serine has a known proton affinity associated withprotonation on the amine. The two base sites are theamine and the carbonyl group of the acid. The twomodel compounds selected were serine and ethanoicacid, which correspond to a proton affinity differenceof 25.0 kcaljmol. Lysine is composed of two aminebase sites, one of which is similar to an amino acidcontaining a large alkyl side chain (valine) and theother is similar to an alkyl amine (n-propylaminc).giving a difference of only 0.9 kcaljmol. This treat­ment cannot accommodate intramolecular interactionsthat increase the basicity of individual sites. In­tramolecular interaction would decrease the differ­ences in basicity, because weakly basic sites becomesignificantly more basic through interaction withstrongly basic sites.

A plot of the ratio k2lk l , a measure of multipleexchange, versus proton affinity difference in the model

interactions. The weaker N - H bond of a less basiccompound allows movement of the proton from theamine to the methanol. The methanol then loses thedeuterium which result in exchange on the termjnalamine.

A strong c=o ... D interaction can similarlyweaken the N - H interaction through charge induc­tion, facilitating the breakage of the CH 30 - D bond.Loss of CH 30H and migration of the deuterium on thecarbonyl group to the more basic amine completes theprocess (pathway b). Indeed, this is the likely scenariopresent in the HID exchange of dipeptides and ex­plains the difference in exchange behavior betweenalanine and the dipeptides dialanine and alanyl­glycine. The carboxylic group is converted to a morebasic amide, thereby increasing the CO··· D interac­tion. Semiempirical calculations predict that the basic­ity of the carbonyl amide is only about 4 kcal /rnol lessthan that of the terminal amine [261.

Similar situations arise when at least two equallystrong base sites are present on the molecule, produc­ing a cyclic intermediate as shown with histidine (In­termediate IV). The strong N ... D interaction weakensthe CH30-D bond, allowing the CHp-H bondto form. This situation likely exists in the HID ex­change reactions of protonated lysine, histidine, andhistidine derivatives and accounts for the reactivity ofthe histidine derivatives lacking a carboxylic acidgroup. Additionally, protons can migrate between basesites in these compounds, producing, for example,Intermediate V,.which can also undergo exchange.

The relationship between multiple HID exchangeand the presence of base sites with similar proton

CompoundProton affinitiesofmodel compounds

Serine

Serine 216.8

CH3CHzCOzH 191.8

Cysteine

Cysteine 214.3

CH3 CHzCOzH 191.8

Alanine

Alanine 214.8

CH3CHzCOzH 191.8

Glycine

Glycine 211.6

CH3CHzCOzH 191.8

Histidine

Alanine 214.8

n-Imidazole 219.8

Dialanine

Alanine 214.8

n-(CH3}zNCH02".4

Lysine

Valine 217.0

n-Propylamine 217.9

Differencesin proton affinitybetween the two most basic

sites in the molecule

25.0 kcal /rnol

22.5 kcat/rnol

23.0 kcal/mol

19.8 kcal/rnol

5.0 kcat/rnol

3.4 kcat/mol

0.9 kcat/mo:

j Am Soc Mass Spectrom 1994, 5, 623-631 GAS-PHASE HID EXCHANGE OF PEPTIDES 631

Diglycine

• Dialanine

•0.8 Lysine:

.0.6 Histidine

0.4

• Glycine

0.2Alanine

Cysteine • Serine.

compounds is shown in Figure 5. It is readily apparentfrom Figure 5 that the smaller the basicity difference,the greater the degree of multiple exchange. Lysine,diglycine, and dialanine, which have two base siteswith nearly equal basicities, have significantly higherk21k1 ratios than serine, cysteine, and alanine, whichall have a single dominant base site. Glycine andhistidine provide intermediate points in which onebase site is slightly more basic than the other. 'TheHID exchange behavior of the two compounds diala­nine and lysine shows a large variation in through­bond distances over which one base site can facilitateHID exchange on another base site. In dialanine, thetwo base sites are separated by two carbon atoms,whereas in lysine they are separated by five carbonatoms. Histidine is an intermediate case with threecarbon atoms between the two base sites.

Strong intramolecular interactions such as thosefound in protonated lysine and histidine translate tostrong intermolecular interactions in the methanol­protonated peptide complexes. Presumably, the re­verse is also true so that strong intermolecular interac­tions in the methanol-protonated peptide dimer implythe capability for strong intramolecular interactions. Inthis regard, HID exchange can be a useful probe forstrong intramolecular interactions because sites thatundergo rapid HID exchange are also sites capable offorming strong intramolecular interactions. This situa­tion provides a unique method for determining whichbase sites within the peptide are able to "communi­cate" via strong intramolecular interactions.

Finally, the results presented provide evidence forunique and highly defined complexes in the interac­tion of methanol and protonated amino acids andpeptides. Complexes such as these could be extremelyimportant in other gas-phase ion-molecule reactions,and HID exchange provides a viable probe of theirstructures. The general presence of these complexes insimilar reactions will be the subject of future investiga­tions.

AcknowledgmentsFunding provided by the National Science Foundation (CHE­9310092) is gratefully acknowledged. M.K.G. thanks the Univer­sity of California for a Graduate Fellowship. The authors alsothank Professor T. Molinski and Professor M. Nantz for provid­ing some of the compounds and for their useful suggestions.

References1. Katta, v.;Chait, B. T. J. Am. Chern. Soc. 1993, 115, 6317.2. Katta, v ; Chait, B. T. Rapid Commun. Mass Spectrom. 1991, 5,

214.3. Winger, B. E.; Light-Wahl, K. J.; Rockwood, A. L.; Smith, R

D. J. Am. Chern. Soc. 1992, 114,5897.4. Suckau, D.; Shi, Y.; Beu, S. c, Senko, M. W.; Quinn, J. P.;

Wampler, F. M.; McLafferty, F. W. Proc. Nat!. Acad. Sci. USA1993, 90, 790.

5. Suckau, D.; Shi, Y.; Quinn, J. P.; Senko, M. W.; Zhang, M.-Y.;McLafferty, F. W.; The 40th ASMS Conference on Mass Spec­trometry and Allied Topics; Washington, DC, May 31~June 5,1992.

6. Mclafferty, F. W. Tandem Mass Spectrometry; Wiley: NewYork, 1983.

7. Freiser, B.S.; Woodin, R L.; Beauchamp, j. L. J. Am. Chern.SO(. 1975, 97, 6893.

8. Hunt, D. F.; Sethi, S. K. J. Am. Chern. Soc. 1980, 102,6953.9. Jasinski, J. M.; Brauman, J. r. J. Am. Chern. Soc. 1980, 102,

2906.10. Ausloos, P.; Lias, S. G. f. Am. Chern. Soc. 1981, 103, 3641.11. Lias, S. J. Phys. Chern. 1984, 88, 4401.12. Squires, R R; Bierbaum, V. M.; Grabowski, J. j.; DePuy, C. H.

J. Am. Chern. Soc. 1983, 105, 5185.13. Waugh, R. J.; Bowie, J. H.; Hayes, R. N. Int. f. Mass Spectrom.

Ion Processes 1991, 107, 333.14. Sepetov, N. F.; Issakova, O. 1.; Lebl, M.; Swiderek, K.; Stahl,

D. c, Lee, T. D. Rapid Commun. Mass Spectrom. 1993, 7, 58.15. Chi, H. T.; Baker, J. K. Org.Mass Spectrom. 1993, 28, 12.16. Cheng, X: Fenselau, C. Int. J. Mass Spectrom. Ion Processes

1992, 122, 109.17. Cushnir, J. R; Naylor, S.; Lamb,J. H.; Farmer, P. B. Rapid

Commun. Mass Spectrom. 1990, 4, 426.18. Lias, S. G.; Liebman, J. F.; Levin, R D. J. Phys. Chern. Ref Data

1984, 13, 695.19. Ausloos, P.; Lias, S. G. J. Am. Chern. Soc. 1977, 99, 4198.20. Brauman, J. L. In Kinetics of lon-Molecule Reactions; Ausloos,

P., Ed.; Plenum: New York, 1979; pp 153.21. Gard, E.; Willard, D.; Green, M. K; Bregar, J.; Lebrilla, C. B.

Org.Mass Spectrom., 1993, 28, 1632.22. McCullough, S. M.; Gard, E.; Lebrilla, C. B. Int. J. Mass

Spectrom. Ion Processes 1991, 107, 91.23. Bartmess, J. E.; Georgiadis, R M. Vacuum 1983, 33, 149.24. Marquardt, D. W. J. Soc. Ind. Appl. Math. 1963, 29,431.25. Bevington, B. R; Robinson, D. K. Data Reduction and Error

Analysis for the Physical Sciences; McGraw-Hill: New York,1992.

26. WU, J.; Lebrilla, C. B. J. Am. Chern. Soc. 1993, 115,3270.27. Guan, S.; Marshall, A G. J. Chern. Phys. 1993, 98, 4486.28. Su, T.; Bowers, M. T. Int. J. Mass Spectrom. IonPhys. 1973, 12,

347.29. Su, T.; Bowers, M. T. T. Chem. Phys. 1973, 58,3027.30. WU, J.; Lebrilla, C. B. Unpublished Manuscript.31. Gorman, G. 5.; Speir, J. P.; Turner, C. A; Amster, I. J. J. Am.

Chern, Soc. 1992, 114, 3986.32. Ranasinghe, A.; Cooks, R G.; Sethi, S. K. Org.Mass Spectrom.

1992, 27,77.33. Bouchoux, G.; Iezequel, 5.; Penaud-Berruyer, F. Org. Mass

Spectrom. 1993, 28, 421.