Probing electrostatic effects: Formation and characterization of zwitterionic ions and their neutral counterparts in the gas phase

Article abstract of DOI:10.1021/ja0016708

A dipolar ion and its nonzwitterionic counterpart were generated and characterized in the gas phase. A single charge site is found to be sufficient to lower the energy of a zwitterion well below its neutral counterpart. Charge separation in a dipolar anion enables hydrogen-deuterium exchange to take place with acids over an extraordinarily large range. These results provide a basis for studying electrostatic effects and understanding mass spectroscopic processes involving large biological molecules.

Full text of DOI:10.1021/ja0016708

9014
J. Am. Chem. Soc. 2000, 122, 9014-9018
Probing Electrostatic Effects: Formation and Characterization of
Zwitterionic Ions and Their “Neutral” Counterparts in the Gas Phase
Katherine M. Broadus and Steven R. Kass*
Contribution from the Department of Chemistry, UniVersity of Minnesota, Minneapolis, Minnesota 55455
ReceiVed May 15, 2000
Abstract: A dipolar ion and its nonzwitterionic counterpart were generated and characterized in the gas phase.
A single charge site is found to be sufficient to lower the energy of a zwitterion well below its neutral counterpart.
Charge separation in a dipolar anion enables hydrogen-deuterium exchange to take place with acids over an
extraordinarily large range. These results provide a basis for studying electrostatic effects and understanding
mass spectroscopic processes involving large biological molecules.
Introduction
only limited thermochemical results on zwitterions are available;
basicities of betaine ((CH3)2N+CH2CO2-) and o-trimethylami-
nobenzoate (o-(CH3)3N+C6H4CO2-) have been measured by
kinetic methods.16,17 Many large peptide ions are believed to
possess salt bridges and exist as dipolar ions,18-22 but the
evidence is largely circumstantial since the independent prepara-
tion of the corresponding nonzwitterionic forms has not been
carried out, and the differences between the two are unknown.
Herein, we report the first preparation and characterization of a
gas-phase dipolar ion (1) and its corresponding nonzwitterionic
(“neutral”) ion (2). Differences in the dissociation, reactivity,
and thermochemistry of these species will be presented.
Zwitterions, also known as dipolar ions, are molecules with
oppositely charged sites. These species are used in a myriad of
practical applications from synthesis and the construction of
novel materials to buffers, enzyme inhibitors, and drugs.1-6 They
abound in nature as all 20 naturally occurring amino acids exist
in dipolar form over a wide pH range. The resulting electric
field plays a crucial role in determining the structure and
reactivity of many proteins and enzymes. As a result, the
consequences of electrostatics on molecular recognition and
catalysis have been studied in solution, but aggregation,
counterion, and solvent effects are difficult to factor out.7-11
The physical environment in a solution also can be quite
different from an enzyme’s active site. Computational and
experimental gas-phase studies on zwitterionic species, conse-
quently, are of considerable interest.
Calculations on glycine, the simplest amino acid, indicate that
the dipolar form does not exist in the gas phase as a stable
structure unless two or more water molecules are present.12
Larger and more basic amino acids are capable of supporting a
zwitterionic structure in some instances but the neutral (un-
charged) form is more stable, even for arginine, despite an earlier
claim to the contrary.13-15 Gas phase experiments also indicate
that amino acids exist in their neutral form, and consequently,
Experimental Section
Methods and Materials. 3,5-Dimethoxycarbonylpyridine and 3,5-
dimethoxycarbonyl-1-methylpyridinium methyl sulfate were prepared
according to literature procedures.23 All reagents were used as received
and solvents were dried via standard methods. NMR spectra were
collected on Varian VXR-300 or VAC-300 spectrometers and are
reported in ppm (δ). High-resolution mass spectra were obtained on a
Finnigan 2001 FTMS by electrospray ionization. Solutions were
prepared in methanol-water (35:65 v/v) with dilute ammonium
hydroxide (2% v/v) and poly(ethylene glycol) (MW ) ∼ 200 g/mol)
as the reference.
3-Carboxy-5-methoxycarbonylpyridine (4).24 3,5-Dimethoxycar-
bonylpyridine (0.4 g, 2.1 mmol) was dissolved in 22 mL of a 0.1 N
(1) Ramos, M. J.; Melo, A.; Henriques, E. S.; Gomes, J. A. N. F.; Reuter,
N.; Maigret, B.; Floriano, W. B.; Nascimento, M. A. C. Int. J. Quantum
Chem. 1999, 74, 299-314.
(13) Price, W. D.; Jockusch, R. A.; Williams, E. R. J. Am. Chem. Soc.
1997, 119, 11988-11989.
(2) Spencer, T. A.; Onofrey, T. J.; Cann, R. O.; Russel, J. S.; Lee, L. E.;
Blanchard, D. E.; Castro, A.; Gu, P.; Jiang, G.; Shechter, I. J. Org. Chem.
1999, 64, 807-818.
(14) Chapo, C. J.; Paul, J. B.; Provencal, R. A.; Roth, K.; Saykally, R.
J. J. Am. Chem. Soc. 1998, 120, 12956-12957.
(15) Maksic, Z. B.; Kovacevic, B. J. Chem. Soc., Perkin Trans. 2 1999,
2623-2629.
(3) Biegle, A.; Mathis, A.; Galin, J.-C. Macromol. Chem. Phys. 1999,
200, 1393-1406.
(16) Patrick, J. S.; Yang, S. S.; Cooks, R. G. J. Am. Chem. Soc. 1996,
118, 231-232.
(4) Gothelf, K. V.; Jorgensen, K. A. Chem. ReV. 1998, 98, 863-909.
(5) Theodore, T. R.; Van Zandt, R. L.; Carpenter, R. H. Cancer Biother.
Radiopharm. 1997, 12, 351-353.
(17) Strittmatter, E. F.; Wong, R. L.; Williams, E. R. J. Am. Chem. Soc.
2000, 122, 1247-1248.
(6) Caron, G.; Pagliara, A.; Gaillard, P.; Carrupt, P.-A.; Testa, B. HelV.
Chim. Acta 1996, 79, 1683-1695.
(18) Wyttenbach, T.; Witt, M.; Bowers, M. T. J. Am. Chem. Soc. 2000,
122, 3458-3464.
(7) Raposo, C.; Wilcox, C. S. Tetrahedron Lett. 1999, 40, 1285-1288.
(8) Kim, E.; Paliwal, S.; Wilcox, C. S. J. Am. Chem. Soc. 1998, 120,
11192-11193.
(19) Freitas, M. A.; Marshall, A. G. Int. J. Mass. Spectrom. 1999, 182/
183, 221-231.
(20) Ewing, N. P.; Cassady, C. J. J. Am. Soc. Mass Spectrom. 1999, 10,
928-940.
(9) Wilcox, C. S.; Kim, E.; Romano, D.; Kuo, L. H.; Burt, A. L.; Curran,
D. P. Tetrahedron 1995, 51, 621-634.
(21) Schnier, P. D.; Price, W. D.; Jockusch, R. A.; Williams, E. R. J.
Am. Chem. Soc. 1996, 118, 7178-7189.
(10) Smith, P. J.; Kim, E. I.; Wilcox, C. S. Angew. Chem., Int. Ed. Engl.
1993, 32, 1648-1650.
(22) Campbell, S.; Rodgers, M. T.; Marzluff, E. M.; Beauchamp, J. L.
J. Am. Chem. Soc. 1995, 117, 12840-12854.
(23) Kristiansen, M.; Eriksen, J.; Jorgensen K. A. J. Chem. Soc., Perkin
Trans. 1 1990, 101-103.
(11) Warshel, A. Acc. Chem. Res. 1981, 14, 284-290.
(12) Jensen, J. H.; Gordon, M. S. J. Am. Chem. Soc. 1995, 117, 8159-
8170.
10.1021/ja0016708 CCC: $19.00 © 2000 American Chemical Society
Published on Web 09/01/2000

Characterization of Zwitterionic Ions in the Gas Phase
J. Am. Chem. Soc., Vol. 122, No. 37, 2000 9015
potassium hydroxide solution in methanol. After the mixture was stirred
for 12 h at room temperature, any unreacted starting material was
extracted with ether and the aqueous layer was concentrated by rotary
evaporation. The product was recovered in 75% yield (0.28 g). 1H NMR
(300 MHz, DMSO-d6) δ 9.12 (d, J ) 2 Hz, 1H), 8.97 (d, J ) 2 Hz,
1H), 8.57 (t, J ) 2 Hz, 1H), 3.89 (s, 3H). 13C NMR (75 MHz, DMSO-
d6) δ 166.3, 166.2, 154.9, 150.0, 137.2, 136.2, 125.1, 52.8. HRMS ESI
(M - H)- calcd for C8H6NO4 180.0302, obsd 180.0310.
Table 1. Dissociation Efficiencies for 1 and 2 upon
Collision-Induced Dissociation (CID) and Sustained Off-Resonance
Irradiation (SORI)
1
2
energy
(eV)a
m/z 136 m/z 92 m/z 25 m/z 136 m/z 121 m/z 77
CID
2.9
3.7
3.6
5.8
0
0
0
0
1.0
2.8
0
0
0
0
1-Methylpyridinium-3,5-dicarboxylic Acid Hydrogen Chloride
(3). 3,5-Dimethoxycarbonyl-1-methylpyridinium (0.18 g, 0.9 mmol) was
dissolved in ∼1 mL of concentrated hydrochloric acid, and the mixture
was refluxed for 3 h. The precipitate that formed upon cooling was
filtered and washed with cold water. The crude product was recrystal-
5.8
6.5
11.7
SORI
3.9
9.3
9.8
21.2
2.4
3.3
8.4
3.5
3.3
10.1
7.1
9.6
15.8
0.5
0.7
1.6
0.5
0.9
1.7
1.5
15.3
0
4.8
0
5.2
0
11.2
0
1
0
1
1
lized from hot water to give 0.08 g (50%) of 3. H NMR (300 MHz,
9.8
DMSO-d6) δ 9.69 (s, 2H), 9.08 (s, 1H), 4.49 (3, 3H), (300 MHz, NaOD/
D2O, pH 10) δ 9.02 (s, 2H), 8.96 (s, 1H), 4.34 (s, 3H); 13C NMR (75
MHz, NaOD/D2O, pH 10) δ 167.5, 146.8, 144.0, 136.7, 48.3. HRMS-
ESI (M - H2)- calcd for C8H6NO4 180.0302, obsd 180.0309.
3,5-Dimethoxycarbonyl-1-methylpyridinium-d3 Methyl Sulfate.
The deuterated compound was prepared according to a procedure for
the protio material.23 To a mixture of 3,5-dimethoxycarbonyl pyridine
(0.22 g, 1.1 mmol) in benzene (5 mL) was added dropwise dimethyl
sulfate-d6 (0.7 mL, 3.8 mmol, Aldrich 99+% enriched). After refluxing
for 3.5 h, the reaction mixture was cooled and acetone (∼ 2 mL) was
added. The precipitated product was collected by suction filtration to
a The energies have not been corrected for the ion’s center of mass
and the given values represent the product percentage relative to the
amount of 1 (or 2) after excitation. See eqs 3 and 4 for ion structures.
Ions of interest were isolated by ejecting undesired species (frequen-
cies) using a SWIFT waveform or a broad band chirp excitation for
low masses.26 They were vibrationally cooled by collisions with argon
which was pulsed into the cell up to pressures of 5 × 10-5 Torr.
Reactions were carried out and monitored over time by introducing
neutral reagents into the FTMS via pulsed or slow leak valves.
1
afford 0.25 g (91%). H NMR (300 MHz, CDCl3) δ 9.64 (d, J ) 1.5
Hz, 2H), 9.32 (t, J ) 1.5 Hz, 1H), 4.07 (s, 6H).
Results and Discussion
1-Methylpyridinium-3,5-dicarboxylic Acid-d3 Hydrogen Chlo-
ride. 3,5-Dimethoxycarbonyl-1-methylpryidinium-d3 methyl sulfate
(0.24 g, 1.1 mmol) was hydrolyzed to its corresponding dicarboxylic
acid following the procedure for the unlabeled material. 1H NMR (300
MHz, DMSO-d6) δ 9.69 (d, J ) 1.5 Hz, 2H), 9.08 (t, J ) 1.5 Hz, 1H).
HRMS-ESI (M - H2)- calcd for C8H3D3NO4 183.04876, obsd
183.04879.
Gas-Phase Experiments. All experiments were carried out in a
Finnigan Fourier transform mass spectrometer (FTMS) equipped with
a 3.0 T superconducting magnet and interfaced with a custom-built
Analytica electrospray ionization source (ESI) which contains an
adjustable amplitude hexapole. Methylated pyridinium-3,5-dicarboxylate
(1) and the conjugate base of 5-methoxycarbonyl nicotinic acid (2)
were sprayed (3 µL/min) into the gas phase from 500 µM methanol-
water (35:65 v/v) solutions of the corresponding diacid (3, pH 6-7)
and monoacid (4, pH 8-9), respectively (eqs 1 and 2); dilute
The zwitterionic ion 1 can be distinguished from 2 by the
fragments observed when it is given energy in the presence of
an inert gas (argon). This process is referred to as collision-
induced dissociation (CID) or sustained off-resonance irradiation
(SORI) if energy is not directly applied to the resonant frequency
(mass-to-charge ratio) of interest.25 Under both CID and SORI
conditions, 1 primarily affords an ion corresponding to loss of
carbon dioxide (5, m/z 136), but double decarboxylation (6, m/z
92) and fragmentation to acetylide (HCtC-, m/z 25) also are
observed (eq 3).27 In contrast, 2 gives ions at m/z 121 (M -
CO2 - CH3)- and m/z 77 (M - CO2 - CO2CH3)- in addition
to a different species (see below) at m/z 136 (7, (M - CO2)-)
(eq 4). The energy dependence also indicates that it is easier to
break apart the dipolar ion than the “neutral” analogue under
nearly single collision conditions (Ar ∼ 1 × 10-7 Torr) using
both on- and off-resonance irradiation (Table 1). This observa-
tion is consistent with dissociation studies of multiply charged
ions and species that are thought to exist as salt bridges.28-30
ammonium hydroxide (2% v/v) in methanol-water (∼35:65) was used
to adjust the pH. The following settings gave optimal signals: needle
housing (cylinder) 2.6 kV, end plate 2.7 kV, capillary 5.9 kV (front)
and -57 V (back), skimmer cone -13 V, acceleration lenses 335 V,
deceleration lenses 21.4 V, and lens stack high voltage element 1.47
kV. Ions were accumulated in the hexapole for 0.80 s and transferred
to the source cell in 30 µsec (TOF). Ion 5-d3 was generated from 1-d3
by exciting the m/z 183 ion to 1.4 eV (sustained off-resonance
irradiation)25 for 0.02 s concurrent with a 10-5 Torr pulse of argon.
(26) Wang, T. C. L.; Ricca, T. L.; Marshall, A. G. Anal. Chem. 1986,
58, 2935-2938.
(27) A ring-opened isomer of 5 (CH3NdCHC(CO2-)dCHCtCH) also
is produced but it can be minimized and never is greater than ∼50%. The
structural assignment for this ion is based upon a reasonable pathway for
its formation (i.e. a ring-opening elimination of 5), and the fact that it does
not react with carbon dioxide in contrast to 5-7, which add 1 equiv of
CO2. Ion 6 only adds 1 molecule of carbon dioxide and presumably arises
from a ring-opening elimination.
(24) Misic-Vukovic, M.; Dimitrijevic, D. M.; Muskatirovic, M. D.;
Radojkovic-Velickovic, M.; Tadic, Z. D. J. Chem. Soc., Perkin Trans. 2
1978, 34-38.
(25) Gauthier, J. W.; Trautman, T. R.; Jacobson, D. B. Anal. Chim. Acta
1991, 246, 211-225.
(28) Summerfield, S. G.; Whiting, A.; Gaskell, S. J. Int. J. Mass
Spectrom. 1997, 162, 149-161.

9016 J. Am. Chem. Soc., Vol. 122, No. 37, 2000
Broadus and Kass
Table 2. Bracketing Results for the Proton Affinities of 1 and 2
To verify the structure of 1, the presence of two carboxylate
groups must be demonstrated. This was accomplished by taking
advantage of the reactivity of 5. It is an aryl anion and ions of
this type are known to react by nucleophilic addition with carbon
dioxide and sulfur dioxide, while sulfur-atom abstraction takes
place with carbon disulfide.31,32 Accordingly, 5 was found to
sequentially undergo these reactions twice as shown in eq 5.33
∆H°acid
a
acid
CH3CO2H
CH3CH2OCH2CO2H
HCl
(kcal mol-1
)
1
2
348.6 ( 2.9
342.2 ( 2.1
333.4 ( 0.1
no rxn
no rxn
adduct (85%)b no rxnc
(adduct-CO2)-
(15%)
HBr
323.4 ( 2.1
(adduct-CO2)- proton
transfer
HI
CF3SO3H
314.4 ( 0.1
305.4 ( 2.2
(adduct-CO2)-
proton transfer
a Acidity values come from ref 34 except for HI which comes from
Berkowitz et al.: Berkowitz, J.; Ellison, G. B.; Gutman, D. J. Phys.
Chem. 1994, 98, 2744-2765. b CID of the cluster affords 1 as the sole
product. c A cluster can be generated at high pressures, but not under
the same conditions employed for 1.
In the nonzwitterionic case, only one adduct or atom transfer
product can be generated. Moreover, aryl ion 7 does not afford
an (M-CO2)- ion upon CID; fragments arising from the loss
of carbon monoxide, methyl radical, and the methyl ester are
detected, whereas 5 does lose carbon dioxide to afford 6.
We have also shown that the two carboxylate groups in 1
are equivalent. A carbon-13 labeled species, 8b (m/z 181), was
synthesized as shown in eq 5. This ion was then isolated and
fragmented at 1 × 10-7 Torr of argon to afford a 1:1 mixture
of aryl ions arising from loss of labeled or unlabeled carbon
dioxide (eq 6). In a similar fashion, 2-13C was prepared by
addition of 13CO2 to 7 and in this case exclusive loss of the
label is observed upon dissociation.
sulfonic acid proton transfer is observed which suggests that
the proton affinity of 1 is between that of this acid and HI (i.e.,
PA(1) ) 310 ( 5 kcal mol-1). This value is in excellent accord
with a computed value of 311.0 kcal mol-1 (B3LYP/6-31+G-
(d)).36 In contrast, 2 does not react with any of the acids studied
to give an (adduct-CO2)- ion under the same conditions
employed for 1. Its proton affinity also is between that of
hydrochloric and hydrobromic acid (i.e., PA(2) ) 328 ( 5 kcal
mol-1), which agrees well with the computed B3LYP value of
327.4 kcal mol-1
.
The smaller proton affinity of 1 is consistent with the fact
that the remote positive charge stabilizes both carboxylates more
than their mutual repulsion destabilizes them. Using a point
charge model, where the nitrogen atom and the midpoint
between the carboxylate oxygens are taken as the charge centers,
leads to a significantly shorter distance for the attractive
interactions than for the repulsive one (i.e., 4.28 vs 6.18 Å).
Coulomb’s law thus indicates that the proton affinity of 1 should
be lower than that of 2 by 24 kcal mol-1 if a dielectric constant
of 1 is used. This crude model overestimates the stability of 1
relative to the experimental (18 ( 7 kcal mol-1) and computed
(16.4 kcal mol-1) differences, but is qualitatively useful.
Isomerization of 1 to 2 is computed to be exothermic by 9.4
kcal mol-1 (B3LYP), but no evidence for methyl group
migration was found in our work. It also is worth noting that
the calculated energy difference between 1 and 2 is considerably
smaller than that for 1H and 2H (25.8 kcal/mol) because of the
favorable additional electrostatic interaction, but that the zwit-
terionic species are less stable than their nonzwitterionic forms
in both of these cases.
The electrostatic interactions in 1 influence its thermochem-
istry relative to that of 2. We measured the proton affinity of 1
by the bracketing technique which involves monitoring the
occurrence or nonoccurrence of proton transfer with reference
acids. No reaction products are observed when 1 is allowed to
react with ethoxyacetic acid or weaker acids (Table 2).34 Upon
reaction with hydrochloric acid, an adduct ion is formed.
Dissociation of this product exclusively yields 1 which suggests
the ion’s proton affinity is less than that for Cl-. As one
progresses to stronger acids (HBr and HI) the adduct ion is no
longer observed, but interestingly, an (adduct-CO2)- ion (11)
is detected (eq 7).35 This product is also observed with HCl,
but it is small (∼15%) in comparison to the clustered product.
Presumably, 11 arises when the electrostatic interactions in the
zwitterion-ion complex 10 are sufficiently large to induce
fragmentation; collision-induced dissociation of 11 only gives
bromide or iodide ion. Upon reaction with trifluoromethane-
(29) Dongre, A. R.; Jones, J. L.; Somogyi, A.; Wysocki, V. H. J. Am.
Chem. Soc. 1996, 118, 8365-8374.
(30) Cox, K. A.; Gaskell, S. J.; Morris, M.; Whiting, A. J. Am. Soc.
Mass Spectrom. 1996, 7, 522-531.
The above experiments demonstrate that a zwitterionic ion
can be generated in the gas phase by electrospray ionization
and distinguished from its “neutral” analogue. As alluded to in
the previous section, the decarboxylated fragments of 1 and 2
(5 and 7, respectively) can be differentiated as well. The most
striking reactivity difference between the two aryl ions is the
former’s ability to undergo switching or ligand exchange
(31) DePuy, C. H. Org. Mass Spectrom. 1985, 20, 556-559.
(32) Bierbaum, V. M.; DePuy, C. H.; Shaprio, R. H. J. Am. Chem. Soc.
1977, 99, 5800-5801.
(33) Collision-induced dissociation of 8 gives a mixture of ions, and the
desired product, 9, was isolated and allowed to further react.
(34) All acidites, unless otherwise noted, come from the following:
Bartmess, J. E. NegatiVe Ion Energetics Data in Secondary NegatiVe Ion
Energetic Data, February 2000, National Institute of Standards and
Technology: Gaithersburg, MD 20899 (http://webbook.nist.gov).
(35) We presume that carbon dioxide is lost from the carboxylate and
then the acid proton migrates via a bimolecular process to afford the
observed ion. Such behavior is consistent with 13C-labeling studies which
(36) These calculations, which in all cases include a zero-point energy
correction using unscaled vibrational frequencies, will simply be referred
to as B3LYP hereafter. All of the computed acidities also include a
temperature correction to 298 K so as to provide a direct comparison with
experiment.
-
show CO2 is lost from the charged site in Ar(CO2R)CO2 species. Reed,
D. R.; Hare, M. C.; Kass, S. R. Submitted for publication.

Characterization of Zwitterionic Ions in the Gas Phase
J. Am. Chem. Soc., Vol. 122, No. 37, 2000 9017
reactions with sulfur dioxide, carbonyl sulfide, and 13C-labeled
carbon dioxide (eq 8). These reactions occur at low pressures
Table 3. Summary of Reactions Observed for the Zwitterionic Ion
5 with Reference Acids
∆H°acid
H/D
proton
a
acid
(kcal mol-1
)
exchangeb transfer SN2c
D2O
CH3OD
(CH3)3COD
CF3CH2OD
CH3CH2SH
CH3CO2H
392.9 ( 0.1 1, 2nd small
383.5 ( 0.7 3, 4th small
374.6 ( 2.1d 3, 4th small
no
no
no
no
no
trace
trace
minor
yes
yes
minor
361.8 ( 2.5d
355.2 ( 2.2
348.6 ( 2.9
4
e
f
-
-
-
no
trace
(∼3 × 10-8 to 2 × 10-7 Torr) of the probe reagents and without
kinetically exciting the starting ion. Moreover, these processes
are relatively facile as the intermediate is not necessarily
observed. This behavior strongly indicates the presence of a
second charge (reactive) site in the ion as the same reaction
has been noted with sulfur dioxide and the benzoate distonic
radical anions (C6H4CO2•-), but not with benzoate anion.37,38
Electrophile switching with the conjugate base of a Lewis acid-
base complex, -CH2S(CH3)BH3, also has been reported.39
In the reaction of 5-d3 with sulfur dioxide, oxygen-atom
transfer ((M + O)-, m/z 155) and an adduct ((M + SO2)-, m/z
203) also are detected. The exchange product (12a, m/z 159)
behaves similarly with SO2 to afford secondary product ions
corresponding to oxygen-atom transfer (m/z 175) and adduct
formation (m/z 223). The “neutral” equivalent (7) gives an
oxygen-atom transfer product and an adduct, but does not switch
CO2 for SO2. With high pressures of carbonyl sulfide (∼10-6
Torr), 5-d3 and 12c afford adducts at m/z 199 and 215,
respectively, while 7 yields a unique sulfur-atom transfer product
(m/z 168, 40%) as well as an adduct (m/z 196, 60%). Last, 5
and 7 show divergent chemistry with carbon disulfide. Ions
corresponding to abstraction of a sulfur atom are observed in
each case; however, the (5 + S)- product fragments to lose
carbon dioxide and undergo a second atom abstraction as shown
in eq 5, whereas this is not the case for 7.
CH3CH2OCH2CO2H 342.2 ( 2.1
yesg
a Acidity values come from ref 34. b See text for details. c Refers to
the acid’s conjugate base attacking the methyl group of 5H to give
nicotinic acid. See text for details. d Values for protio acid used. e If
the reaction is carried out with 5-d3, the deuteriums in the methyl group
are exchanged for hydrogen. f With 5-d3, some washing out of the label
is observed. g Proton transfer is the major reaction channel with this
acid although some (adduct-CO2)- is also generated.
-
Table 4. Relative Energies of C7H6NO2 Isomers Computed at the
B3LYP/6-31+G(d) Level of Theory (0 K) and Using Coulomb’s
Law (ꢀ ) 1) with a Point Charge Model
Erel
compda
B3LYP
Coulomb’s law
a (13)
b (14)
0.0
1.2
0.0
8.1
c
16.9
17.9
33.0
18.2
48.0
120.5
170.1
d (5)
e
7
a The letters represent the anionic site.
Reactions of 5 and 7 with acids further illustrate that the two
ions are unique and do not interconvert. The “neutral” ion reacts
with deuterium oxide and methanol-d4 to give the conjugate
base of nicotinic acid and a small amount of methoxide-d3 in
the latter case. The former product forms by initial deuteron
transfer followed by ester hydrolysis with the newly generated
DO- or CD3O- (eq 9). In addition, with D2O one hydrogen
additional deuteriums with methanol-OD when the d1 ion
generated from 5 and deuterium oxide is used as the substrate.
Alternatively, when 5-d3 is reacted with the deuterated alcohol
1 deuterium can be incorporated to a small extent, and when
protio reagent is used all of the label can be washed out. These
results indicate that methanol reacts with 5 and 14 to primarily
afford 13. In accord with these observations, we have found
that 5, 13, and 14 undergo characteristic reactions with carbon
disulfide. Ion 5-d3 reacts with the reagent largely via sulfur-
atom transfer, 14 (as the d4 ion produced by reacting 5-d3 with
D2O) is inert, and 13 (as the d2 ion produced by reacting 5-d3
with MeOH) affords the conjugate base of nicotinic acid (eq
11). The last of these reactions also occurs with carbonyl sulfide
and presumably proceeds via sulfur-atom transfer and subse-
quent expulsion of thioformaldehyde.
Zwitterion 5 incorporates three deuteriums with tert-butyl
alcohol-OD, and at long times (5-10 s at 5 × 10-8 Torr) a
fourth exchange is observed. With the stronger acid, trifluoro-
ethanol-OD, this fourth exchange is more prominent because
there is a closer correspondence between the acidities of the
conjugate acids of 13 and 14 and the exchange reagent. The
deuteriums in 5-d3 can easily be washed out with ethanethiol
and to some extent with acetic acid. Proton transfer is also
observed in the latter case and occurs readily with ethoxyacetic
acid; an (adduct-CO2)- ion is detected as a minor product with
the latter acid too. Likewise, 13-d2 generated from 5-d3 displays
can be exchanged for a deuterium to a minor (∼14%) extent
before the base attacks the ester group. Zwitterionic ion 5
undergoes acid-catalyzed isomerization with these reagents, and
rearranges to 13 and/or a second species depending on the acid
that is employed. We presume that 14 is the latter ion since it
is the most stable aryl anion isomer (see below, Table 4). Upon
reaction with deuterium oxide, 5 undergoes one hydrogen-
deuterium exchange and a small amount of a second deuterium
incorporation is observed at long (30 s at 4 × 10-8 Torr) reaction
times (eq 10). To better probe this process, 5-d3 was prepared
and reacted with H2O. Only a small amount of the label is
washed out to give a d2 ion (i.e., 13), suggesting that 14 is the
major product. In contrast, 5 undergoes 4 hydrogen-deuterium
exchanges with methanol-OD (Table 3) and incorporates 3
(37) Guo, Y.; Grabowski, J. J. J. Am. Chem. Soc. 1991, 113, 5923-
5931.
(38) Reed, D. R.; Kass, S. R. Unpublished results.
(39) Ren, J. H.; Workman, D. B.; Squires, R. R. J. Am. Chem. Soc. 1998,
120, 10511-10522.

9018 J. Am. Chem. Soc., Vol. 122, No. 37, 2000
Broadus and Kass
4). Introduction of a single charge stabilizes the zwitterionic
form relative to its “neutral” structure such that 5 and 7 are
predicted to be essentially the same energetically. Dipolar anions
13 and 14 differ by only 1.2 kcal mol-1, and both species are
significantly (17-18 kcal mol-1) more stable than the nonzwit-
terionic ion 7. This result has obvious implications when
considering the structures of large peptides and proteins in the
gas phase. It is also worth noting that a simple point charge
model using Coulomb’s law is qualitatively useful in that it
reproduces the relative DFT energies, and should be helpful in
designing superacids and bases by virtue of electrostatic
interactions.
An interesting reaction channel is observed when 5 is allowed
to react with strong acids. Ethanethiol and acetic acid give rise
to the conjugate base of nicotinic acid (15) via nucleophilic
attack at the newly formed methyl group (eq 13). This pathway
the same behavior with acetic and ethoxyacetic acids. These
results lead to a bracketed proton affinity of 345 ( 3 kcal mol-1
for ylide 13 which is in good accord with the B3LYP predicted
value of 343.8 kcal mol-1 and represents a remarkable acidifying
effect of 68 kcal mol-1 relative to trimethylamine (∆H°acid
)
412.8 kcal mol-1 (G2)).40 The bulk of this stabilization is the
result of the favorable electrostatic interaction between the
formal positive charge on nitrogen and the negative charge on
the dideuteriomethylene group.
also slowly occurs with ethoxyacetic acid, trifluoroethanol, and
methanol, which indicates that the proton affinity difference
between 13 and A- is a more important factor in determining
if the reaction occurs than the strength of the nucleophile. The
structure of the product was verified to be the conjugate base
of nicotinic acid upon comparison with an authentic sample. In
particular, under the same experimental conditions, both ions
fragment by loss of carbon dioxide, and the resulting pyridyl
ion (17) reacts with deuterium oxide to give -OD and two
hydrogen-deuterium exchanges. This reactivity is in accord
with that reported for the conjugate base of pyridine.44 In
addition to confirming the structure of 13, this derivatization
experiment supports the structure of our initial zwitterion 1.
Hydrogen-deuterium exchange between anions (HA-) and
deuterated acids (DX) has been extensively investigated.41 It is
well-known that the acidity of DX must be within 20 kcal mol-1
of that for H2A for the process to occur. This requirement is a
result of the mechanism (eq 12) that typically involves an
endothermic proton transfer between DX and HA-. This
thermodynamically unfavorable step can take place because it
is fueled by the ion-neutral complex energies of I and II which
usually are on the order of 20 kcal mol-1. When the initial
proton transfer is less favorable than this amount there is
insufficient energy in the system to drive this step and H/D
exchange does not occur. In the case of 13 hydrogen-deuterium
exchange is observed to slowly take place with deuterium oxide
even though the initial proton transfer is endothermic by 47
kcal mol-1 and exchange is rapid with methanol-OD where the
difference in acidity is still 37 kcal mol-1! These unprecedented
results at a carbon center are a consequence of the unusually
large electrostatic interactions present in the ion-zwitterion
complex (II, see 16 below).42 This finding may account, at least
in part, for the extraordinary deuterium incorporation that has
been observed in large peptide ions upon reaction with
deuterium oxide and methanol-OD where the proton affinity
Conclusions
We have provided the first evidence that a zwitterionic ion
and its “neutral” counterpart are unique species in the gas phase.
The isomer pairs, 1 and 2 as well as 5 and 7, exhibit
thermochemical differences as well as divergent reactivity with
probe reagents. Dipolar ion 5 undergoes acid-catalyzed isomer-
izations to remarkably stabilized ylide structures which also were
explored. A single charge site is found to be capable of lowering
the energy of a zwitterion relative to its neutral analogue such
that the former species becomes more stable than the latter.
Charge separation in a dipolar carbanion (A-) enables hydrogen-
deuterium exchange to occur with acids (DX) where the
difference in acidity between HA and DX is extraordinarily large
(i.e. 37-47 kcal mol-1) and would normally not take place.
This work provides a basis for studying electrostatic effects and
additional results will be reported soon.
difference is on the order of 40 to 50 kcal mol-1 22,43
.
Density functional theory calculations were carried out on 5,
7, 13, 14, and the other two dipolar aryl anion isomers (Table
(40) Curtiss, L. A.; Raghavachari, K.; Trucks, G. W.; Pople, J. A. J.
Chem. Phys. 1991, 94, 7221-7230.
Acknowledgment. Support from the National Science
Foundation, the donors of the Petroleum Research Foundation,
as administered by the American Chemical Society, the Min-
nesota Supercomputer Institute, and the University of Minne-
sota-IBM Shared Research Project are gratefully acknowledged.
(41) Squires, R. R.; Bierbaum, V. M.; Grabowski, J. J.; DePuy, C. H. J.
Am. Chem. Soc. 1983, 105, 5185-5192.
(42) Binding energies of zwitterionic ions with neutral molecules should
be similar to those for “normal” ions (i.e., ∼15-25 kcal/mol). In accord
with this expectation, preliminary computations at 0 K at the B3LYP/6-
31+G(d) (+ZPE) level indicate that the cluster energy of 5‚H2O is 15.9
kcal/mol.
JA0016708
(43) Gard, E.; Willard, D.; Bregar, J.; Green, M. K.; Lebrilla, C. B. Org.
Mass Spectrom. 1993, 28, 1632-1639.
(44) DePuy, C. H.; Kass, S. R.; Bean, G. P. J. Org. Chem. 1988, 53,
4427-4433.