Intrinsic Basicity of Oligomeric Peptidesthat Contain Glycine, Alanine, andValine-The Effects of the Alkyl Side Chain onProton Transfer Reactions

Jiangyue Wu and Carlita B. LebrillaDepa rtment of Chemistr y, University of Ca lifornia, Davis, Ca lifornia , USA

The gas-phase bas icities of oligomers of alan ine and valine have been det ermined bybracketing measurements in an external source Fourier transform mass spectrometer. Theresults are compared to the oligomers of glycine, which were reported in an earlierpublication, to observe the effect of the alkyl group and the increasing gas-phase basicity ofthe monomer units on the rate s of proton transfer reactions. Molecular orbital calculationswere performed on protonated triglycine and trialan ine to determine how the alkyl groupsaffect intramolecular interactions . The results show that a high degree of ordering of thecarbony l groups is present in the protonated species. The carb onyl groups in turn order theside chain alkyl groups and de crease the rates of proton transfer react ions in, for example,the oligomers of valine.

92 WU AND LEBRILLA J Am Soc Mass Spectrorn 1995, 6, 91-101

1Q,----t-------t-----+-----+---_+_

9.5

9

8.5

7.5 -L----L----+----j-----f-----+2 3 4

amino acid units (n)

Figure 1. The plot of pK"2 versus II of glycine oligomers inaqueous solution. The acidity increases as the length of thepeptide chain increases. Values are taken from [17].

interaction is an important factor in the basicity of thecompound.

oH 3N.:Jl (aq)

J Am 50..- Mass Spcctrorn 1995, h, lJl IIlI INTRINSIC BASICITY 01' OLIGOMERIC I'EI'T1DES 93

mized. For each reported structur e, several other con-formations were analyzed to ensure that a local min i-mum of a specific proton binding site was found ,

Gas-Phase Basicity Scale

When this work began, the main sou rce of gas-phasebasicity values was a compilation by Lias, Liebmanand Levin (LLL) [6] of numerous gas-phase studies .Very recently, two other basicity scales have beenintrod uced that redefine the high end (more basic part)of the scale. A GB scale proposed by Moet-Ner andSieck (MNS) [24] increased the gas-phase basicity as-signments of many organic arn ines that have GB val-ues in the range of the compounds investigated for thisreport. Another scale presented by Szulejko andMcMahon (SM) [25] also reev aluated the high end andproduced values closer to the LLL scale, with many ofthe values supported by high level ab initio calcula-

tions. However, the SM basicity scale is not nearly ascomplete as the LLL bas icity scale. There are, forexample, onl y three compounds in the SM tabulationwith GB values with in the range of the peptides westud ied, Furthermore, direct comparison of the abso-lute values between the SM and the LLL scales pro-duces differences in absolute GB values that are typi-cally less than 1 kcaljmol. For consistency and conve-nience, therefore, we decided to use the LLL scaleexclus ive ly for this report, and all GB values are rela-tive to the LLL scale unless otherwise noted.

Results

Proton Transfer Reactions

The presence of rapid or abrupt changes in reactivity isused in this report to assign gas-phase basicity ofamino acids and peptides. After the protonated pep-

a) b)

0.8 0.8• •~

0.6~

0.6~

0.4

0.2

_(~I~~I'N!il ~_. . _.

4 6 8

tlme(s)10

0.4

0.2 n-PrNH·J ....• -.----1If

2 4 6t1me(a)

..8 10

.-- ' -... > I -BuNH·•

65

,,

.. -... -_. EtMeNHa'

234t1me(s)

d) 1

0.8

_ 0.6~

0.4 ,,

•0.2

0065

••

2 3 4time(s)

c)

0.8

0.6~

0.4

•0,2

00

Figure 2. Relative intens ity profile of protonated d ialan ine reacting with (a ) allylamine , G B = 207.9kcal yrnol 0 .4 x 10 -" torr), (b) n-propylam inc, 210.1 (1.3 x 10- " torr), (c) r-buty lam lnc. 211.10 .2 x 10-" torr), and (d) ethylmethylamine, 215.1 (1.1 x 10-" torr). The sy mbo l l lLI representsthe intensit y o f the ion over all ion intensities. No react ions or extremely slow reactions are observedin (a) and (b), whereas rapid reactions are observed in (c) and (d). The GB of d ialanine therefore liesbetween the GB of II-propylamine and r-butylarn inc .

94 WU AND LEBRILLA J Am Soc Mass Spectrorn 1995, 6, 91-10 1

12

•

1 .2 1.4

10

•

8

•

42

0.2 0.4 0 .6 0.8

tlme(e)

-.--- y = 0 .025422 + 0.2348x R== 0 .99186

00

3b

3

+:z::4( 2 .5

.;-2:z::

4(

-:;:+:z:: 1 .54(

:::.':I:::ID

-:;: 0.5-c0

0

2 Y = -0.061072 + 1.2215x R= 0.992253a

6

t1me(a)24

Figure 3. (a) Kinetic plot of d irect proton transfer reactioninvolving protonated valine and s-butylamine. The rate constantis obtained from the slope . (b) Kinetic plot of proton transferreactions when dimerization is involved for the reaction of proto-nated trivaline and cyd ohexylamine. The rate constant is ob-tained from the slope.

+:z::1.5

•>•-+ •r' 1z

:::IID..... •- 0.5 •c •

ylamine (R = 0.992; Figure 3b).

(2)

tide is transferred and isolated in the analyzer cell, it isallowed to undergo proton transfer reactions with thebackground pressure of a reference base. The behav-iors of the relative abundance of a protonated peptidein reaction with four different reference bases (at pres-sures between 1 and 7 x 10- 8 torr) of increasing basic-ity are shown in Figure 2 for dialanine. When the basesare n-propylamine (GB 210.1 kcal / mol) and allylamine(207.9 kcaly'mol), negligible proton transfer is observedover a period lasting up to 8 s (Figure 2a and b). Rapidproton transfer is observed over the same period whenthe bases are i-butylamine (211.1 kcaljmoD and ethyl-methylamine (215.1 kcalyrnol). The abrupt change inreactivity between n-propylamine and i-butylamine,which corresponds to a difference of only 1.0 kcaljmolchange in the GB values, allows simple assignment ofthe GB of this dipeptide. However, compounds withsteric hindrance complicate assignments because theirgeneral reactivities are decreased [26]. Alternatively,reaction efficiencies, defined as the ratio k exp/kADO'where kexp is the experimental rate constant and k ADOis the theoretical rate constant obtained from averagedipole orientation theory [27, 28], provide a more facileind icator of changes in reactivity. Theoretical rate con-stants are determined from the polarizability and thedipole moment of the reference bases, but do not takeinto account the structural features of the ions . Polariz-abilities are available or can be closely approximatedby using the method of Miller [21]. Dipole momentswere obtained from the literature or were calculatedby using AMI [22, 23]. Dipole moments obtained fromAMI were typically within 25% of the experimentaldipole moments and did not significantly affect thevalue of the theoretical rate constants.

Rate constants were obtained directly from protontransfer reactions (Reaction 2) by assuming a pseudo-first-order behavior for the protonated peptides.Derivation and manipulation of the rate equationsproduces a linear relationship between the expressionInO + (BH+)I(AH + » and reaction time (t), where(BH +)I(AH + ) is the ratio of the abundances of theprotonated reference base and peptide, respectively.The slope contains the rate constant of the reaction.The example of the reaction of protonated valine withs-butylamine is representative and shows the expectedlinear behavior (R = 0.992; Figure 3a).

Dirner formation is often observed when the gas-phase basicity of the protonated peptide and the refer-ence base are similar (Reaction 3). The presence of themixed dimer in the spectra produces an additionalcomplication in the rate analysis that is dealt with byusing the steady state approximation. Manipulation ofthe rate equations and plotting with respect to reactiontime similarly produce a linear behavior as illustratedby the reaction of protonated trivaline and cyclohex-

Rate constants (and the reaction efficiencies) for thereactions of protonated oligomers of alanine and valinewith selected groups of reference bases are provided inTables 1 and 2. Reaction efficiencies accurately predicttransitions between endergonic and exergonic reac-tions [29, 30]. Efficiencies greater than 50% have beenshown to characterize exergonic reactions. The samecriterion has been used in this report to assign GBvalues, that is, a reaction efficiency greater than 50%indicates that the reference base has a GB value greater

J Am Soc Mass Spectrom 1995. 6. 91-101

Table 1. Rate constants ( X 10-9 crn' molecule - I S - I) andefficiencies (in parentheses) for the reactions of alanine oligomerswith selected reference bases

Reference base(GB 300 K) Ala (Ala)2 (Ala)3 (Ala)4

Dimethylforma mide 0.07 0.15 0.17 0.09(203.6) (3) (7) (8) (5)

3-F-Pyridine 1.8 0.04 0.06 0.06(206.2) (122) (3) (5) (5)

Allylamine 0.64 0.03 0.03 0.09(207.9) (46) (2) (3) (7)

n-Propylamine 1.2 0.07 0.12 0.20(210.1) (90) (5) (10) (18)

i-Butylamine 1.1 1.1 0.82 0.75(211.1 ) (76) (85) (67) (61 )

s-Butylamine 0.96 1.5 0.91 2.1(211.7) (67) (117) (75) (178)

n-Ethylmethyla mine 1.6 1.6 0.72 0.68(215.1 ) (116) (128) (60) (58)

Table 2. Rate constants Lx l O"" crn' molecule T ' s·l)andefficiencies (in parentheses) for the reactions of valine oligomerswith selected reference bases

Reference base(GB 300 K) Val (Va1)2 (Va1)3 (Va1)4

DMF 0.09(203.6) (4)

c-Propylamine 0.03(206.6) (2)

Allylamine 0.80 0.13 0.14 0.08(207.9) (50) (10) (11 )

n-Propylamine 1.70 0.13 0.12 0.08(210.1) (134) (10) (11 ) (7)

s-Butylamine 0.73 0.22 0.22 0.22(211.7) (54) (19) (19) (19)

c-Hexylamine 2.10 0.49 0.55 0.41(213.4) (159) (42) (49) (37)

n-Ethyl methyla mine 1.10 0.89 0.56 0.40(215.1) (88) (89) (52) (37)

Di-ethylamine 1.20 0.99 0.55 0.36(217.7) (104) (90) (52) (35)

Di-n-propylamine 1.30 0.89 0.80 0.55(219.7) (100) (77) (73) (50)

than the peptide. For example, the efficiencies of proto-nated alanine reactions show an increase from dimethylformamide (3%) to 3-fluoropyridine (100%). Reactionsof protonated alanine with more basic compoundsproduce rate constants that are consistently at or higherthan 50%, with some fluctuations. The GB of alanine istherefore assigned to lie between 203.6 and 206.2, or204.9 ± 1.3 kcaljmol. Our assigned value is slightlylower than that reported in the literature, which rangesfrom 206.6 to 208.6 kcaljmol [4, 5, 7, 11]. The uncer-tainties listed in Table 3 correspond to the differencebetween the assigned GB and the GBs of the referencebases that bracket the peptide. One should keep inmind the uncertainties in the GBs of the referencebases when the uncertainty values are used.

INTRINSIC BASICITY OF OLIGOMERIC PEPTIDES 95

The oligomeric peptides (Ala)2' (Ala)3' and (Ala)4are all assigned the same GB-21O.6 kcaljmol-be-cause of the large change in efficiencies when theprotonated species are reacted with n-propylamineand i-butylamine. For protonated dialanine this corre-sponds to an increase in efficiency from 5 to 85%. Theefficiencies of reactions with protonated dialanine re-main consistently high as reference bases with GBgreater than i-butylamine are used. The increases inthe efficiencies of trialanine and tetra alanine with thetwo bases are not as large, but this is consistent withthe general behavior of these larger compounds. Thereactivity of trialanine indicates that it does not have aGB greater than dialanine as measured by this tech-nique. This behavior contrasts with the glycine serieswhere triglycine was found to have slightly greater GBthan diglycine. The high efficiencies of the reactions ofprotonated trialanine with both i-butylamine and s-butylamine, which are only 0.6 kcaljmol apart, con-firm the upper limit of its GB. The reactivity behaviorof tetravaline does not differ significantly from eitherdi- or trialanine.

Assignments of the GB of the valine oligomers wereperformed in a similar manner. Efficiencies of the reac-tions of protonated valine undergo a large changebetween cyclopropylamine and allylamine (2 and 59%,respectively). The assigned value of 207.2 kcaljmol isconsistent with the range of literature values, whichare between 207.4 and 209.2 kcaljmol. Divaline under-goes its largest change in efficiencies between n-butylamine (9%) and cyclohexylamine (41%). Thisdifference in efficiencies is relatively small, but consis-tent with the apparently poor reactivities of all thevaline oligomers (vide supra). The corresponding ef-ficiencies of the reaction of protonated trivaline withthese two bases increase from 19 to 50%, whereas forprotonated tetravaline the increase is even smaller; theefficiencies change from 17 to 37%. In any case, bothtrivaline and tetra valine are assigned the same GB asdivaline. The reactivity of both peptides appears toincrease only slightly with compounds more basic thancyclohexyl amine. The least reactive is clearly tetrava-line, where the efficiencies between cyclohexylamineand di-n-propylamine remain between 35 and 50%,despite the difference of 6.3 kcaljmol between the GBof the two reference bases.

Table 3. Gas-phase basicities of oligomers of glycine,alanine, and valine. Values are referenced by using theLLL basicity scale [6]

GB Ikcal/rnol)

n (GIY)n (Ala)n (Val)n

1 203.1 ± 0.6 204.9 ± 1.3 207.3 ± 0.72 208.5 ± 0.6 210.6 ± 0.5 212.6 ± 0.83 209.6 ± 0.6 210.6 ± 0.5 212.6 ± 0.84 209.6 ± 0.6 210.6 ± 0.5 212.6 ± 0.85 209.6 ± 0.6

96 WU AND LEBRILLA .1 Am Soc Mass Spectrom 1'195, n, '11- W]

Table 4. The effects of translational excitation on the rateconstants

Effects of Translational Excitation on ProtonTransfer Reactions

The GB values of linear glycine oligomers (Gly",II = 1-5) have been reported previously by using theMNS basicity scale: large changes observed in the rateconstants were used to assign gas-phase basicities (Ta-ble 3) [14). The data have been reevaluated for thisreport by using reaction efficiencies and the LLL basic-ity scale. The assigned absolute values of all the glycineoligomeric peptides change slightly with the use of theLLL basicity scale. The use of reaction efficiencies,instead of rate constants alone, does not alter therelative ordering. Earlier we assigned pentaglycine tobe slightly more basic than tetraglycine. Reanalysis ofthe data indicates no observable change in GB betweenthe two peptides.

Translational excitation is a concern in this techniquewhere ions are produced externally and transferred tothe analyzer region of the Fourier transform massspectrometry instrument. Translational excitation alsocan be obtained during the ion isolation step, when aseries of oscillating electric fields is applied to theanalyzer plates with frequencies near the cyclotronfrequency of the ion. The effect of this "off-resonance"excitation is particularly strong when the mass-to-charge ratio of the ions to be ejected is near that of thetrapped ion. Ion isolation is, however, necessary tominimize proton transfer reactions between the refer-ence base and other protonated species. The net effectof translational excitation is illustrated by the reactionof i-butylamine with triglycine (Table 4). The mass oftriglycine (111 /z 190) is close to the mass of the proto-nated dimer of glycerol (I11/Z 185); the latter is usuallyejected before the proton transfer reactions are allowedto proceed. Under typical experimental conditions

AM1 Calculations of Triglycineand Trialaninc

(ejection pulse 10 ms, amplitude of 4.0 V peak to peak)a rate of 1.4 X 10 'i crn ' molecule I s I is obtained.Pulsing nitrogen gas into the analyzer chamber afterion isolation increases the rate of the reaction to 1.6 x10 'i cm ' molecule 1 s 1. We commonly observe thattranslationally excited ions have slower reaction rates.For example, increasing the amplitude of the ejectionpulse to 7.6 V p-p decreases the reaction rate to1.1 X 10 - 'i crn ' molecule lSi. Increasing the ampli-tude further to 15.5 V decreases the rate constant to aslow as 6.8 X 10 III ern:' molecule 1 s 1. When theejection pulse is further away, as in experiments thatinvolve tetraglycine and Il-propylamine, the effect ofejection pulses is diminished (Table 4). For example,an increase in the amplitude by over a factor of 3 (from13 to 42 V p-p) does not significantly decrease therates.

Translational excitation can make a compound ap-pear more basic because it decreases the overall rate ofthe reaction. A stronger base would be needed toobserve proton transfer reactions, but would increasethe assigned GB value. However, under these experi-mental conditions, we find that translational excitation,at least that due to off-resonance excitation during ionejection, can be minimized and does not appear tointerfere with the assignment of the GB.

Molecular orbital calculations have been performed toevaluate the protonated structures of triglycine andtrialanine, Earlier we [14) reported the results of AMIcalculations performed on protonated diglycine. Thequalitative results recently have been supported byhigh level ab initio calculations [15). Both sets of calcu-lations predict that the terminal amine is the mostfavorable site of protonation in both glycine anddiglycine. Previously, we reported that the lowest en-ergy structure for protonated diglycine involved hy-drogen bonding interaction between the carbonylgroup amide and the protonated terminal amine(Structure La). The results from ab initio calculationssuggest a structure with hydrogen bonding interac-tions between the terminal amine and both carbonylamide and carboxylic acid groups that is only 0.47kcalyrnol less stable [15). Further search of the AMIhypersurface produces a similar type of interaction.Structure 1b shows distances of 2.498 and 2.106 A.between the terminal amine and the two respectivegroups. The ab initio calculations produced a moresymmetrical interaction with distances of 2.291 and2.247 A., respectively. The new structure 1b is, how-ever, lower than 1a by 5.2 kcaljmol.

Structures 2a-f are the most stable protonated con-formers of triglycine. These structures were chosenbecause they represent various protonation sites on thetripeptides. For example, Structure 2a involves proto-nation on the terminal amine with hydrogen bonding

1.1 X 10- 9

6.8 x 10- 10

0.39 X 10- 9

0.44 X 10 9

9.0x 10- 10

Reaction Rates (cm 3

conditions molecule - 1 S - 1)

Ejection 185, Vp p= 7.6V

Ejection 185, Vp p=12V

Ejection 185, Vp p= 15.5V

Ejection 185, Vp p=13V

Ejection 185. Vp p 1.6 x 10 ..9

= 4.0Vwith pulse of N2

Ejection 185, Vp p 1.4 x 10 9=4.0V

Ejection 185, Vp p=42 V

Reactions

[(GIY)4 + H'](m!z 247)with n propylamine

[(GIY)3 + H'](m!z 190)with i-butylamine

I Am Soc Mass Spcctrorn 1

98 WU AND LEBRILLA J Am Soc Mass Speclrom 1995.6,91-101

to the three carbonyl groups, whereas Structure 2binvolves hydrogen bonding between the terminalamine and the nearest carbonyl amide, and so forth.Hydrogen bonding interactions, often characterized bydistances of less than 2.6 Abetween the hydrogen andinteracting heteroatom (e.g., oxygen and nitrogen) arelabeled by their distances. For every structure shown,several other conformers were investigated but notincluded for having higher relative heats of formation.The lowest energy structure (2a) involves protonationon the terminal amine with extensive intramolecularhydrogen bonding interaction to possibly all three car-bonyl groups. The N - H ... O=C bond distances are2.673,2.110, and 2.054 Abetween the hydrogen on theterminal amine and the oxygen on the three carbonylgroups. Other hydrogen bonding interactions also oc-cur in the molecule, producing further stabilizationincluding the interaction within carboxylic acid thatforces this group to be cis-planar with respect to t,reCO-O bond (CO· .. H-0 distance of 2.281 A).Conformational microvariations also are observed inthe AMI calculations, even with peptides as small astriglycine. For example, a conformation similar to 2aexists (2a') that has the same number and type ofH-bonding interactions but is 1.1 kcaljmol higher inenergy and contains different hydrogen bond distancesfor the same interacting atoms. These structures fur-ther illustrate the difficulty in finding the global min-ima in molecules with extensive hydrogen bondinginteractions.

The other structures with protonated N-termini in-dicate that maximizing hydrogen bonding of the aminogroup with all the carbonyl groups generally providesthe most favorable interaction. For example, when thecarboxylic acid group does not interact with the proto-nated N-terminus but instead with a neighboring car-bonyl amide hydrogen (Structure 2a"), the heat offormation is slightly higher by 4.3 kcaljmol. Whenonly one carbonyl amide is involved in coordinatingwith the protonated terminal amine (2b and 2c), theheat of formation increases by about 10 kcaljmol rela-tive to the most stable structure. There is little differ-ence energetically when interaction occurs with eitherthe neighboring carbonyl group (2b) or one amino acidunit away (2c).

The most stable carbonyl protonated structure cor-responds to 2d, where the two carbonyl oxygens of theamides are linked by hydrogen bonding. This speciesis only 5.9 kcaljmol less stable than 2a. Extensivehydrogen bonding is also present in 2d, particularlybetween the first amide hydrogen and both the termi-nal amine and the acid carbonyl oxygen (2.533 and2.298 A, respectively). There appears to be a specificpreference, on the AMI surface, for the first carbonylamide as the site of protonation; the second is lessfavored by 6.1 kcaljmol (2d and 2e). This is somewhatunexpected because both carbonyl groups should havesimilar proton affinities.

Protonated trialanine structures analogous to the

protonated triglycine structures are found on the AMIsurface. To facilitate direct comparisons between thetwo sets, trialanine structures, which are closely analo-gous to triglycine structures in terms of protonationsites and intramolecular interactions, are given thesame letter designation. The trialanine analog (Struc-ture 3a) of the lowest energy protonated triglycinestructure (2a) is also the most stable protonated triala-nine structure. Extensive hydrogen bonding betweenthe protonated amine and the three carbonyl groups isevident from the N - H ... O=C distances of 2.606,2.111, and 2.065 A. Steric interactions due to the pres-ence of the methyl groups do not appear to preventhydrogen bonding interactions in this protonatedstructure. Instead, the alkyl groups are projected radi-ally from the site of protonation, minimizing stericinteractions. As with the protonated triglycine, we findconformers with similar energies that utilize the samecombination of interactions. For example, 3a" is also alocal minimum analogous to 2a". However, the differ-ence between 3a" and 3a is 7.4 kcaljmol compared to4.3 kcaljmol between 2a" and 2a. The structures 2band 2c also have analogs in trialanine as 3b and 3c,respectively. The difference in the heat of formationbetween 3b and 3a 02.3 kcaly'mol) is again slightlylarger than the difference between 2b and 2a (9.9kcalyrnol). There is evidently greater steric interactionin 3c because it is 17.8 kcaljmol less stable than 3acompared to 10.7 kcaljmol for 2c relative to 2a. Sev-eral attempts were made to locate a more stable struc-ture for 3c; however, steric interactions appear to pre-vent hydrogen bonding interactions between the amideN - H group and the carboxylic acid group as in 2c.In general we find that the difference in the heats offormation between 3a and 3b-f is greater than be-tween 2a and 2b-f. This may be the result of stericinteractions in structures 3b-f that destabilize thesestructures relative to 3a. It is likely that the differenceswould be further increased by the presence of largeralkyl groups such as isopropyl (in valine).

Protonation of the carbonyl amide remains unfavor-able compared to protonation on the terminal amine.The structure 3d is less stable than 3a by 8.5 kcaljmolcompared to 5.9 for triglycine. Other structures wherethe carbonyl amides are protonated (e.g., 3e and 3f) aresimilarly unstable relative to 3a.

Discussion

All three series show an increase of GB values betweenthe amino acid and the dipeptide (see Figure 4). Aslight increase is further observed between di- andtriglycine that is not observed in the alanine and valineseries. The contributions of intramolecular interactionsin the dipeptide appear constant. For the threeoligomers this means a similar increase in GB betweenthe amino acid and the dipeptide that corresponds to5.4, 5.7, and 5.3 kcaljmol for glycine, alanine, andvaline, respectively. An increase of the basicity of diva-

J Am Soc Mass Spectrom 1995, 6, 91-101

H ....H

2.4701L H

~H ~HN

1._ H

H 1.11yO H 0H / -H~ ;:;;;- H

H

3a6Hr=-44.0 kcal/mol

INTRINSIC BASICITY OF OLIGOMERIC PEPTIDES

3a"

6Hr=-36.6 kcaUmol

99

3c

6Hr=-26.2 kcal/mol

3b6Hr=-31.7 kcaUmol

KN

H

H

." fiIf H ~_o

Hr.H

H

N~.0,.o H

HH ·······H

H

H 0~'0H+-r"'N~ H H

H N1r Jt..~H ' , HH" N

H I H ~HO.......H_O

1.994

3d6Hr=-35.5 kcaUmol

3e

6Hr=-29.0 kcal/mol

3r

6Hr=-29.7 kcallmol

214 t----t----t---t----t----t---_t_

Figure 4. Plot of gas-phase basicity, assigned from bracketingmeasurements, versus peptide length for oligomers of glycine,alanine, and valine. The curve for each oligomeric series isshifted according to the difference in GB of the amino acids.

212

~ 210"u~'Ej~ 2080.. ...

~ ~ 206o

204

202o

II

I

1/I

I /

./

J

/," - - - .. - - -.... Valine

...-- - --- - ----

234Peptide Chain Length

(amino acids)

5

GlyciDe

6

line is consistent with observations reported by Gor-man and Amster [16], who report a 2.5-kcaljmol dif-ference between the two compounds. We observe ahigher increase, which is not unreasonable consideringthe error limits presented in this and the other authors'work.

The gas-phase basicity of the dipeptide and its rela-tive value compared to the amino acid is consistentwith the intramolecular interactions depicted in la andlb. The presence of two interacting base sites is remi-niscent of the protonated dimers that have been stud-ied [31]. In other words, a dipeptide can be seen as aprotonated amine solvated by a carboxylic acid or anyother carbonyl group that contains compounds. Linearcorrelations between proton affinities and hydrogenbond strength have been proposed [31] specifically forbinding of protonated amines with several other func-tional groups. For the dimers, the correlation followsthe relationship

100 WU AND LE13RILLA J Am Soc Mass Spectrum 1995, O. 91 -101

For -NH + " '0- bonds ~Hoo(O) = 30 ± 2 and theslope b = 0.26. Maximum binding occurs when theproton affinity difference (~PA) is zero. For meth-ylamine and a carboxylic acid the average value isabout 20 kcaljmol. If an estimate for entropy is in-cluded, that is, the entropy of formation of an eight-membered ring from cyclooctane, the expected ~Gwould be 9 kcaljmol-not too different from the in-crease of 5.4 kcaljmol observed between glycine anddiglycine. Correlation of proton affinity differences alsowould explain the identical increase in GB between theamino acid and the dipeptide for the three series, thatis, in all three the ~PA contribution should be similar.

The difference in GB between diglycine andtriglycine, though small, is supported by experimentsperformed by other groups [13, 15]. The results ofmolecular orbital (MO) calculations that involve thetriglycine and trialanine further support this behaviorand suggest that the increase in the experimental GBshould be even greater than that observed. The struc-tures 2b and 2c of protonated triglycine are analogousto the most stable structure of diglycine (La) and sharethe primary interactions found in the dipeptide. Thereappears to be little difference between the interactionof the protonated terminal amine with either the firstor second amide group because the energy differencebetween 2b and 2c is only 0.8 kcalyrnol. However, theadditional hydrogen bonding interaction provided bythe second amide group makes 2a about 10 kcaljmolmore stable than either 2b or 2c. The additional stabi-lization would be consistent with a greater protonaffinity ( - ~ H) for the tripeptide than the dipeptide.The increase in GB (- ~G) obtained experimentally(1.1 kcaljmoll is comparatively small. The comparisonof MO calculations of triglycine and trialanine suggeststhat a similar increase in proton affinity also should bepresent in the trialanine and the trivaline series. Thelowest structures found in both sets of calculations (2aand 3a) are similar because the methyl groups oftrialanine do not appear to disrupt the intramolecularhydrogen bonding interactions. Experimental GB val-ues, however, do not exhibit this behavior. The GBvalues of trialanine and trivaline are not observed toincrease relative to the respective dipeptides.

The small increase in the GB of triglycine relative tothe dipeptide and the absence of similar increases intrialanine and trivaline is possible if the net change inthe entropy is negative. MO calculations suggest ahigh degree of ordering due to the interaction of thecarbonyl groups with the protonated terminal amine.Furthermore, experimental deviation in this method ofassigning GB values may not readily allow observationof small increases in trialanine and trivaline. In thisregard, bracketing experiments that use equilibriumreactions would be more accurate than the kineticreactions used here, but are not possible because abackground pressure of neutral peptide would not besustainable. It is not likely that translational excitationwould be the reason for the apparent lack of a large

GB increase in the tripeptide. An increased transla-tional energy, as we have shown, would actually in-crease the apparent GB, slowing the rate and requiringstronger bases to produce proton transfer reactions.

A high degree of ordering of the carbonyl groupspresents a situation not unlike those found in thegas-phase solvation of ions or the formation of gas-phase ionic clusters [32]. The carbonyl groups in thesesmall peptides could act in the same manner as theinner solvation shell of the solvated ions . Moet-Ner hasrecently shown that alkylation does not affect the innershell binding energies in protons solvated by arninesand alcohols [32]. These results are consistent with thelack of a destabilizing effect of the alkyl groups on themost stable structures of protonated peptides. The alkylgroups in dialanine and divaline are positioned toallow intramolecular interactions similar to those foundin diglycine. In protonated trialanine, the other proto-nated structures appear to be destabilized relative tothe most stable structure (e.g ., compare 3a and 3a").This indicates that steric interactions play a greaterrole in destabilizing some structures and slightly favorthe most stable structure.

The most readily observed affect of the alkyl groupsis their interference in proton transfer reactions. This isreadily observed with the oligomers of valine and, to alesser extent, the oligomers of alanine. Proton transferreactions that involve divaline do not produce effi-ciencies of 100% with the reference bases chosen forthis study. The efficiencies of analogous reactions thatinvolve trivaline and tetravaline are lower. For proto-nated tetravaline, reaction efficiencies remain at orbelow 50% despite relatively large changes in GBs.These consistently low reaction efficiencies make as-signment of the GB values for these compounds lessaccurate. For tri- and tetravaline. the GB values shouldat least be as large as divaline, There are no apparentintramolecular interactions observed in the MO calcu-lation of triglycine and trialanine to suggest significantdestabilization in the protonated tri- and tetravalinestructures.

In general, steric interactions between the proto-nated peptide and the reference neutral base will cre-ate problems in the determination of gas-phase basici-ties of still larger peptides. The ability of alkyl groupinteractions to slow proton or hydride transfer reac-tions has been well established [33]. Steric interactionsat or near the site of protonation are already known tosignificantly decrease the rates of even highly exer-gonic reactions [26]. These experiments, however, in-volve molecules where the protonation sites areblocked directly. For example, the reactivity of proto-nated pyridinium compounds is significantly de-creased by the presence of large alkyl groups on theortho positions. In contrast, with protonated peptidessteric interactions are the result of ordering and not ofprimary structural effects. The protonation sites areblocked by the ordering of the alkyl groups that resultsfrom the ordering of the carbonyl groups as the latter

J Am Soc Mass Spectrom 1995, 6, 91-101 INTRINSIC BASICITY OF OLIGOMERIC PEPTIDES 101

try to maximize interactions with the protonated aminegroup. This behavior is supported by MO calculationsthat predict that the most stable arrangement for pro-tonated trialanine is structure 3a. The radial arrange-ment of the methyl groups in protonated trialanineappears to partially shield the site of protonation fromapproach by a neutral base. This effective shieldingwould greatly increase with amino acids that havelarger alkyl groups such as valine and leucine. Theproblem for those measuring gas-phase basicity be-comes particularly acute when the bases necessary fordeprotonation are themselves sterically hindered, suchas dialkyl amines or alkyl amines with large alkylgroups such as cyclohexyl and s-butyl, and even i-bu tylamines.

AcknowledgmentsFunding was provided by the National Science Foundation(CHE-9310092l. The computers for performing MO calculationswere provided by the Department of Chemistry, U.c. Davis.

References

1. Hipple, P. H. Y.; Schleich, T. Acc. CI1em. Res.1969,2,257-265.2. Robinson, D. R.; Jencks, W. P. f. Am. Cilem. Soc. 1965, 87,

2470-2479.3. Flanagan, M. A.; Ackers, G. K.; Matthew, J. B.; Hanania,

G. I. H.; Gurd, F. R. N. Biochemistry 1981, 20, 4739-4744.4. Locke, M. J.; Hunter, R. L.; Mciver, R. T., Jr. f. Am. Chem. Soc.

1979, 101, 272-273.5. Locke, M. J.; Mciver, R. T., Jr. f. Am. Chem. Soc. 1983, 105,

4226-4232.6. Lias, S. G.; Liebman, J. F.; Levin, R. D. ,. PI1ys. Chem. Ref Data

1984, 13, 695-808.7. Moet-Ner. M.; Hunter, E. P.; Field, F. H. f. Am. Chem. Soc.

1979, 101,686-689.8. Isa, K.; Omote, T.; Amaya, M. Drg. Mass spectrom. 1990, 25,

620-628.

9. Kinser, R. D.; Ridge, D. P.; Uggerud, E. Proceedings of theAmerican Society for Mass Spectrometry; Nashville, TN, May19-24, 1991.

10. Bojesen, G. ,. Am. Chem. Soc. 1987, 109, 5557-5558.11. Gorman, G. S.; Speir, J. P.; Turner, C. A.; Amster, I. J. f. Am.

Cilem. Soc. 1992, 114,3986-3988.12. Bojesen, G. ,. Chem. Soc. Chelll. Commrm 1986,244-245.13. Wu, Z.; Fenselau, C. ,. Am. Soc. Mass Spectrom. 1992, 3,

863-866.14. WU, J.; Lebrilla, C. B. ,. Am. Clrem. Soc. 1993, 115,3270-3275.15. Zhang, K.; Zimmerman, D. M.; Chung-Phillips, A.; Cassady,

C. J. f. Am. CI1em. Soc. 1993, 115, 10812-10822.16. Gorman, G. S.; Amster, I. J. f. Am. Chem. Soc. 1993, 115,

5729-5735.17. Burger, K Biocoordination Chemistry: Coordination Equilibria in

Biologically Active Systems; Ellis Horwood: London, 1990.18. Carroll, J. A.; Ngoka, L.; McCullough, S. M.; Gard, E.; Jones,

A. D.; Lebrilla, C. B. Anal. Clrem. 1991, 63, 2526-2529.19. McCullough, S. M.; Gard, E.; Lebrilla, C. B. Int. f. Mass

spectrom. Ion Processes 1991, 107, 91-102.20. Bartmess, J. E.; Georgiadis, R. M. Vacuum 1983, 33, 149-153.21. Miller, K J. f. Am. Clrem. Soc. 1990, 112,8533-8542.22. Dewar, M. J. S.; Zoebisch, E. G.; Healy, E. F.; Stewart, J. J. P. f.

Am. Chem. Soc. 1985, 107, 3902-3909.23. Dewar, M. J. S.; Dieter, K M. f. Am. Chem. Soc. 1986, 108,

8075-8086.24. Moet-Ner, M.; Sieck, L. W. f. Am. Chell/. Soc. 1991, 113,

4448-4460.25. Szulejko, J. E.; McMahon, T. B. ,. Am. CI1elll. Soc. 1993, 115,

7839-7848.26. Moet-Ner, M.; Smith, S. f. Am. Chem. Soc. 1991, 113,862-869.27. Su, T.; Bowers, M. T. t. Chem. Phys. 1973, 58, 3027-3037.28. Su, T.; Bowers, M. T. Int. f. Mass spectrom. Ion Phys. 1973, 12,

347-356.29. Bohme, D. K; MacKay, G. I.; Schiff, H. I. t. Chem. Phys. 1980,

73, 4976-4986.30. Buker, H. H.; Crutzmacher, H. F. Int. ,. Mass Specirom. Ion

Processes 1991, 109, 95-104.31. Meot-Ner, M. f. Am. Clrem. Soc. 1984, 106, 1257-1264.32. Meot-Ner, M. f. Am. Chem. Soc. 1992, 114,3312-3322.33. Sunner, J.; Hirao, K; Kebarle, P. ,. Phys. Clrem. 1989, 93,

4010-4016.