Uracil and Thymine Reactivity in the Gas Phase: The SN2
Reaction and Implications for Electron Delocalization in
Anna Zhachkina and Jeehiun K. Lee*
Department of Chemistry and Chemical Biology, Rutgers, The State UniVersity of New Jersey,
New Brunswick, New Jersey 08901
Received August 11, 2009; E-mail: email@example.com
Abstract: The gas-phase substitution reactions of methyl chloride and 1,3-dimethyluracil (at the N1-CH3)
are examined computationally and experimentally. It is found that, although hydrochloric acid and
3-methyluracil are similar in acidity, the leaving group abilities of chloride and N1-deprotonated 3-methyluracil
are not: chloride is a slightly better leaving group. The reason for this difference is most likely related to the
electron delocalization in the N1-deprotonated 3-methyluracil anion, which we explore further herein. The
leaving group ability of the N1-deprotonated 3-methyluracil anion relative to the N1-deprotonated
3-methylthymine anion is also examined in the context of an enzymatic reaction that cleaves uracil but not
thymine from DNA.
Uracil and thymine are pyrimidine nucleobases that differ in
structure by only a methyl group at C5 (1a and 3a). Uracil
naturally occurs in RNA, while thymine is its DNA counterpart.
Although uracil and thymine are very similar in structure,
the presence of uracil in DNA is problematic.1-4Uracil can
arise in DNA from cytosine deamination, which is mutagenic;
uracil can also be misincorporated into DNA, leading to
cytotoxic uracil·adenine base pairs.1,2,4-6Uracil is removed
from the genome by the enzyme uracil DNA glycosylase
The mechanism of UDG has been shown to involve N1-
deprotonated uracil as the leaving group (LG).5,9,14-30Depro-
tonated uracil as a leaving group seems somewhat surprising
and begs the question: How good of a leaving group is
deprotonated uracil? The N1-H pKa in water is 9.8, which
would indicate a fair or mediocre leaving group ability. In the
enzyme, uracil has a depressed pKaof 6.4.19,20
Examining properties in the gas phase is useful for elucidating
inherent reactivity in the absence of solvent.3,17,31-40In previous
studies, we calculated and measured the gas-phase acidity of
uracil and found it to be as acidic as hydrochloric acid, indicating
that in the gas phase, deprotonated uracil might be, relatively
speaking, a good LG.3Furthermore, because enzyme environ-
ments are sometimes quite nonpolar, reactivity in the gas phase,
(1) Kavli, B.; Otterlei, M.; Slupphaug, G.; Krokan, H. E. DNA Repair
2007, 6, 505–516.
(2) Connolly, B. A.; Fogg, M. J.; Shuttleworth, G.; Wilson, B. T. Biochem.
Soc. Trans. 2003, 31, 699–702.
(3) Kurinovich, M. A.; Lee, J. K. J. Am. Chem. Soc. 2000, 122, 6258–
(4) Parker, J. B.; Bianchet, M. A.; Krosky, D. J.; Friedman, J. I.; Amzel,
L. M.; Stivers, J. T. Nature 2007, 449, 433–438.
(5) Stivers, J. T.; Jiang, Y. L. Chem. ReV. 2003, 103, 2729–2759.
(6) Seiple, L.; Jaruga, P.; Dizdaroglu, M.; Stivers, J. T. Nucleic Acids
Res. 2006, 34, 140–151.
(7) Bellamy, S. R. W.; Krusong, K.; Baldwin, G. S. Nucleic Acids Res.
2007, 35, 1478–1487.
(8) Stivers, J. T. Chem.-Eur. J. 2008, 14, 786–793.
(9) Savva, R.; McAuley-Hecht, K.; Brown, T.; Pearl, L. Nature 1995,
(10) Porecha, R. H.; Stivers, J. T. Proc. Natl. Acad. Sci. U.S.A. 2008, 105,
(11) Panayotou, G.; Brown, T.; Barlow, T.; Pearl, L. H.; Savva, R. J. Biol.
Chem. 1998, 273, 45–50.
(12) Krosky, D. J.; Schwarz, F. P.; Stivers, J. T. Biochemistry 2004, 43,
(13) Berti, P. J.; McCann, J. A. B. Chem. ReV. 2006, 106, 506–555.
(14) Slupphaug, G.; Mol, C. D.; Kavli, B.; Arvai, A. S.; Krokan, H. E.;
Tainer, J. A. Nature 1996, 384, 87–92.
(15) Kimura, E.; Kitamura, H.; Koike, T.; Shiro, M. J. Am. Chem. Soc.
1997, 119, 10909–10919.
(16) Parikh, S. S.; Walcher, G.; Jones, G. D.; Slupphaug, G.; Krokan, H. E.;
Blackburn, G. M.; Tainer, J. A. Proc. Natl Acad. Sci. U.S.A. 2000,
(17) Kurinovich, M. A.; Lee, J. K. J. Am. Soc. Mass Spectrom. 2002, 13,
(18) Dong, J.; Drohat, A. C.; Stivers, J. T.; Pankiewicz, K. W.; Carey,
P. R. Biochemistry 2000, 39, 13241–13250.
(19) Drohat, A. C.; Stivers, J. T. Biochemistry 2000, 39, 11865–11875.
(20) Drohat, A. C.; Stivers, J. T. J. Am. Chem. Soc. 2000, 122, 1840–
(21) Jiang, Y. L.; Drohat, A. C.; Ichikawa, Y.; Stivers, J. T. J. Biol. Chem.
2002, 277, 15385–15392.
(22) Drohat, A. C.; Jagadeesh, J.; Ferguson, E.; Stivers, J. T. Biochemistry
1999, 38, 11866–11875.
(23) Drohat, A. C.; Xiao, G.; Tordova, M.; Jagadeesh, J.; Pankiewicz,
K. W.; Watanabe, K. A.; Gilliland, G. L.; Stivers, J. T. Biochemistry
1999, 38, 11876–11886.
(24) Mol, C. D.; Arvai, A. S.; Slupphaug, G.; Kavli, B.; Alseth, I.; Krokan,
H. E.; Tainer, J. A. Cell 1995, 80, 869–878.
(25) Werner, R. M.; Stivers, J. T. Biochemistry 2000, 39, 14054–14064.
(26) Stivers, J. T.; Drohat, A. C. Arch. Biochem. Biophys. 2001, 396, 1–9.
(27) Guthrie, R. D.; Jencks, W. P. Acc. Chem. Res. 1989, 22, 343–349.
(28) Dinner, A. R.; Blackburn, G. M.; Karplus, M. Nature 2001, 413, 752–
(29) Shroyer, M. J. N.; Bennett, S. E.; Putnam, C. D.; Tainer, J. A.;
Mosbaugh, D. W. Biochemistry 1999, 38, 4834–4845.
(30) Parker, J. B.; Stivers, J. T. Biochemistry 2008, 47, 8614–8622.
(31) DePuy, C. H. Int. J. Mass Spectrom. 2000, 200, 79–96.
(32) DePuy, C. H.; Bierbaum, V. M. Acc. Chem. Res. 1981, 14, 146–153.
(33) Ervin, K. M. Chem. ReV. 2001, 101, 391–444.
Published on Web 11/24/2009
10.1021/ja906814d 2009 American Chemical Society
18376 9 J. AM. CHEM. SOC. 2009, 131, 18376–18385
a kind of “ultimate” nonpolar medium, can yield insight into
biological reactivity.3,17,35-43Uracil is particularly intriguing,
we found, because the acidities at the N1 and N3 sites are very
different in the gas phase, but coalesce in aqueous solution.3,17
Having established that uracil (1a) is quite acidic in the gas
phase, we now examine the leaving group ability of the
conjugate base, N1-deprotonated uracil (2a), in a substitution
reaction. Because uracil is as acidic as HCl, is deprotonated
uracil as good of a leaving group as chloride in the gas phase?
We are also interested in comparing the leaving group abilities
of deprotonated uracil versus deprotonated thymine. The oc-
currence of uracil in the human genome is typically one uracil
per >107normal DNA base pairs. The ability of the UDG
enzyme to find and excise the few uracils present while leaving
the structurally similar thymine untouched is of interest.44A
recent study indicates that the uracil versus thymine discrimina-
tion could be due in part to base-pair dynamics.4We wanted to
probe the possibility that another contribution to the favorable
excision of uracil over thymine could be due to the relative
leaving group abilities of the corresponding conjugate bases
(deprotonated at N1). If N1-deprotonated thymine (4a) is a
poorer leaving group than N1-deprotonated uracil (2a), this
could presumably contribute to the favorable excision of uracil
over thymine.3,17,37,40Leaving group ability has been implicated
SN2 reactions in the gas phase have been studied for more
than two decades. They follow the “classic” gas-phase double-
well potential energy surface, where an initial ion-molecule
complex is formed that can either dissociate back to reactants
or react to products.34,46- 60Previously measured second-order
reaction rate constants and efficiencies for SN2 reactions of
methyl chloride with different anionic nucleophiles are shown
in Table 1 (where reaction efficiency is defined as the ratio of
the observed rate constant to the estimated collision rate constant
calculated by parametrized trajectory calculations).34,61,62
Model Systems for Gas-Phase Study. The excision of uracil
from DNA involves nucleophilic attack of the C1′ of ribose
(Figure 1). Kinetic isotope effects point to a “dissociative SN2”
reaction mechanism (DN*AN).5,16,25Our interest is in testing
the leaving group ability of deprotonated uracil in a substitution
reaction. The simplest model would be to examine reactivity at
the N1-CH3group of 1-methyluracil (1c). However, experi-
ments with 1-methyluracil are limited by the acidity of the
N3-H, which has been measured to be 348 ( 3 kcal mol-1in
the gas phase.3,17Therefore, any nucleophile with a proton
affinity (PA) of 348 or greater will likely deprotonate the N3-H.
Proton transfers are enthalpically generally barrierless (∆Hq)
0) and will, if exothermic, compete with the SN2 reaction.33
Using nucleophiles with PAs less than 348 kcal mol-1would
be too limiting; for substitution reactions with CH3Cl, anions
with a PA at or below 348 kcal mol-1yield very low efficiencies
(Table 1). Therefore, to allow the use of more basic nucleophiles,
we chose to examine the 1,3-dimethyl substrates (1d, 3d)
wherein the methyl group at N3 acts as a sort of “protecting
Acidity Studies. With the chosen model systems, our leaving
groups are no longer the deprotonated uracil and deprotonated
thymine, but rather the 3-methyl derivatives (Figure 2).
As a starting point toward ascertaining whether N1-depro-
tonated-3-methyluracil is a better leaving group than N1-
(34) Gronert, S. Chem. ReV. 2001, 101, 329–360.
(35) Sharma, S.; Lee, J. K. J. Org. Chem. 2002, 67, 8360–8365.
(36) Kurinovich, M. A.; Lee, J. K. Chem. Commun. 2002, 2354–2355.
(37) Lee, J. K. Int. J. Mass Spectrom. 2005, 240, 261–272.
(38) Sun, X.; Lee, J. K. J. Org. Chem. 2007, 72, 6548–6555.
(39) Liu, M.; Xu, M.; Lee, J. K. J. Org. Chem. 2008, 73, 5907–5914.
(40) Liu, M.; Li, T.; Amegayibor, S.; Cardoso, D. S.; Fu, Y.; Lee, J. K. J.
Org. Chem. 2008, 73, 9283–9291.
(41) Gilson, M. K.; Honig, B. H. Biopolymers 1986, 25, 2097–2119.
(42) Simonson, T.; Brooks, C. L., III J. Am. Chem. Soc. 1996, 118, 8452–
(43) Sharma, S.; Lee, J. K. J. Org. Chem. 2004, 69, 7018–7025.
(44) Cao, C.; Jiang, Y. L.; Krosky, D. J.; Stivers, J. T. J. Am. Chem. Soc.
2006, 128, 13034–13035.
(45) Bennett, M. T.; Rodgers, M. T.; Hebert, A. S.; Ruslander, L. E.; Eisele,
L.; Drohat, A. C. J. Am. Chem. Soc. 2006, 128, 12510–12519.
(46) Olmstead, W. N.; Brauman, J. I. J. Am. Chem. Soc. 1977, 99, 4219–
(47) Riveros, J. M.; Jose, S. M.; Takashima, K. AdV. Phys. Org. Chem.
1985, 21, 197–240.
(48) Gronert, S.; Pratt, L. M.; Mogali, S. J. Am. Chem. Soc. 2001, 123,
(49) DePuy, C. H.; Gronert, S.; Mullin, A.; Bierbaum, V. M. J. Am. Chem.
Soc. 1990, 112, 8650–8655.
(50) Gronert, S. Acc. Chem. Res. 2003, 36, 848–857.
(51) Gronert, S.; Fong, L.-M. Int. J. Mass Spectrom. 1999, 192, 185–190.
(52) Gronert, S.; Flores, A. E. J. Am. Chem. Soc. 1999, 121, 2627–2628.
(53) Gronert, S. J. Mass Spectrom. 1999, 34, 787–796.
(54) Gronert, S.; Fagin, A. E.; Okamoto, K.; Mogali, S.; Pratt, L. M. J. Am.
Chem. Soc. 2004, 126, 12977–12983.
(55) Gronert, S.; Fagin, A. E.; Wong, L. J. Am. Chem. Soc. 2007, 129,
(56) Bohme, D. K.; Young, L. B. J. Am. Chem. Soc. 1970, 92, 7354–7358.
(57) Tanaka, K.; Mackay, G. I.; Payzant, J. D.; Bohme, D. K. Can. J. Chem.
1976, 54, 1643–1659.
(58) Pellerite, M. J.; Brauman, J. I. J. Am. Chem. Soc. 1983, 105, 2672–
(59) Wang, H.; Peslherbe, G. H.; Hase, W. L. J. Am. Chem. Soc. 1994,
(60) Ingemann, S.; Nibbering, N. M. M. Can. J. Chem. 1984, 62, 2273–
(61) Chesnavich, W. J.; Su, T.; Bowers, M. T. J. Chem. Phys. 1980, 72,
(62) Su, T.; Chesnavich, W. J. J. Chem. Phys. 1982, 76, 5183–5185.
Table 1. Second-Order Reaction Rate Constants and Efficiencies
for Reactions of Nucleophiles of Varying Proton Affinity with
PA, kcal mol-1
J. AM. CHEM. SOC. 9 VOL. 131, NO. 51, 2009
Uracil and Thymine Reactivity in the Gas Phase
deprotonated 3-methylthymine, we assessed the acidity of the
N1-H proton in 3-methyluracil (3-MeU, 1b) and compared it
to the acidity of the N1-H proton in 3-methylthymine (3-MeT,
3b). The acidity of 3-MeU has been calculated previously by
us (332.8 kcal mol-1, Figure 3); herein, we calculate the ∆Hacid
of 3-MeT to be 1.4 kcal mol-1less than the ∆Hacidof 3-MeU
(332.8 vs 334.2 kcal mol-1, Figure 3).17On the basis of these
acidities, we would expect deprotonated 3-methyluracil to be a
better leaving group than deprotonated 3-methylthymine.
We have also previously measured the gas-phase ∆Hacidof
3-MeU (1b) to be 333 ( 2 kcal mol-1.17We bracket the acidity
of 3-MeT herein (3b) (Table 2). We find that while the conjugate
base of 3-MeT deprotonates 2-chloropropanoic acid (∆Hacid)
337.0 ( 2.1 kcal mol-1) and acids with lower ∆Hacidvalues, it
cannot deprotonate trifluoro-m-cresol (∆Hacid) 339.3 ( 2.1
kcal mol-1) or reference acids with higher ∆Hacid values.
Consistent with this, 2-chloropropanoate cannot deprotonate
3-MeT, but trifluoro-m-cresolate can. We therefore bracket the
∆Hacidof 3-MeT to be 338 ( 3 kcal mol-1(∆Gacid(3-MeT) )
331 ( 3 kcal mol-1).
We also conducted Cooks kinetic method experiments to
measure the relative acidity of 3-MeT (3b) and 3-MeU
(1b).65-69We accomplished this by dissociating the [(3-
MeU)-·H+·(3-MeT)-] dimer. These experiments indicate that
3-MeU (1b) is 2-3 kcal mol-1more acidic than 3-MeT (3b).
The experiments therefore indicate a difference in acidity
between 3-MeU and 3-MeT that is on the order of 2-5 kcal
mol-1. Although this is a rather large range, what is consistent
is that 3-MeU is more acidic than 3-MeT. This could make the
N1-conjugate base of 3-MeU a potentially better leaving group
than that of 3-MeT.
Another important reason to probe acidities is to establish
the upper limit of proton affinity for the nucleophiles to be
studied experimentally. As we noted earlier, we are using the
N3-methyl substrates because the N3-H of 1-methyluracil has
a ∆Hacidof 348 kcal mol-1, and to avoid competition between
SN2 reaction at the N1-CH3and proton transfer at the N3-H,
we would be limited to using nucleophiles with proton affinities
(PA) under 348 kcal mol-1.3,17Now that we have established
that we will be using the 1,3-dimethyl derivatives 1d and 3d,
we need to assess the acidities of all of the sites of those
substrates (Figure 3). The C6 protons of both of these derivatives
are quite acidic, with values just below that of acetone (the C6
and C5 of 1,3-dMU have also been previously measured by
the Gronert lab and our lab).17,53Therefore, to avoid competition
from deprotonation, nucleophiles with proton affinities below
367 kcal mol-1will be utilized.
SN2 Reaction Studies: Calculations. a. 1,3-Dimethyluracil
(1,3-dMU). We next calculated the energetics associated with
SN2 reactions of 1,3-dMU. We chose formate and methyl
thiolate as the nucleophiles. Formate was chosen as a probable
slow reaction example, based on the acidity of formic acid
(∆Hacid ) 346.2 ( 1.2 kcal mol-1) and the known methyl
chloride data for nucleophiles in that acidity range (Table 1).
Methyl thiolate (∆Hacid) 357.6 ( 2.0 kcal mol-1) was chosen
as a faster (although still moderate) reaction example.
The reaction of formate with 1,3-dMU first forms the expected
reactant ion-molecule complex, a process that is 23.1 kcal
mol-1exothermic (Figure 4). The ∆Hqbarrier to the SN2 reaction
is 32.3 kcal mol-1from this complex (and 9.2 kcal mol-1from
the separated reactants (Figure 4)). The product ion-molecule
complex is 9.3 kcal mol-1more stable than the separated
reactants, but the separated products are 2.7 kcal mol-1higher
in energy than the separated reactants. Given that both the
transition state (TS) and the separated products are higher in
energy than the separated reactants, this reaction is not likely
to proceed significantly under our gas-phase conditions.
For the reaction of methyl thiolate with 1,3-dMU (figure in
Supporting Information), the transition state is calculated to be
just 4.3 kcal mol-1higher than the energy of the separated
reactants. The energy of the transition structure must be below
the energy of the separated reactants for reaction to be observed,
so we could see reaction depending on how accurate the
calculations are, and how entropically unfavorable the process
(63) Gronert, S.; Feng, W. Y.; Chew, F.; Wu, W. Int. J. Mass Spectrom.
2000, 195/196, 251–258.
(64) NIST Chemistry WebBook, NIST Standard Reference Database Number
69; Linstrom, P. J., Mallard, W. G., Eds.; National Institute of
Standards and Technology: Gaithersburg, MD, 2005; http://web-
(65) Green-Church, K. B.; Limbach, P. A. J. Am. Soc. Mass Spectrom.
2000, 11, 24–32.
(66) McLuckey, S. A.; Cooks, R. G.; Fulford, J. E. Int. J. Mass Spectrom.
Ion Phys. 1983, 52, 165–174.
(67) Brodbelt-Lustig, J. S.; Cooks, R. G. Talanta 1989, 36, 255–260.
(68) McLuckey, S. A.; Cameron, D.; Cooks, R. G. J. Am. Chem. Soc. 1981,
(69) Cooks, R. G.; Kruger, T. L. J. Am. Chem. Soc. 1977, 99, 1279–1281.
Figure 1. Nucleophilic attack at C1′ to excise uracil.5
Figure 2. SN2 reactions studied.
Figure 3. Calculated N1-H acidities of 3-methyluracil17(3-MeU, 1b) and
3-methylthymine (3-MeT, 3b) and all of the sites of 1,3-dimethyluracil (1d)
and 1,3-dimethylthymine (3d) (B3LYP/6-31+G*, ∆H298K, kcal mol-1).
Experimental values where known are in parentheses.17,63
18378J. AM. CHEM. SOC. 9 VOL. 131, NO. 51, 2009
Zhachkina and Lee
b. Methyl Chloride. We also calculated the energetics for
the reaction of formate (Figure 5) and methyl thiolate (figure
in Supporting Information) with methyl chloride, to provide a
benchmark for comparison. Interestingly, the methyl chloride
reactions are consistently more exothermic and have lower
barriers than the uracil reactions. With formate, the ∆Hqis 1.9
kcal mol-1below the separated reactants (Figure 5). For methyl
thiolate, the ∆Hqis 7.5 kcal mol-1below the separated reactants.
Our calculations therefore indicate that, although 3-MeU and
HCl have similar acidities, the 1,3-dMU SN2 reactions are
expected to be slower than those of CH3Cl.
c. 1,3-Dimethylthymine (1,3-dMT). Our calculations indicate
that in keeping with the fact that thymine is less acidic than
uracil, the SN2 reactions wherein deprotonated thymine is a
leaving group do have slightly higher barriers than the corre-
sponding reactions with deprotonated uracil. Using 1,3-dMT
as the model system, the energy surfaces for reaction with
formate and methyl thiolate were calculated (figures in Sup-
porting Information). For the formate reaction, the barrier is
1.2 kcal mol-1higher with 1,3-dMT than with 1,3-dMU (relative
to separated reactants) and is also more endothermic. For the
methyl thiolate reaction, the barrier for 1,3-dMT is 1 kcal mol-1
higher than that with 1,3-dMU. We also find computationally
that for a wide range of nucleophiles (formate, acetate, n-
pentylthiolate, methyl thiolate, and anilide), the exothermicities
of the 1,3-dMU versus the 1,3-dMT reactions are always roughly
1.4 kcal mol-1apart (with 1,3-dMU being more exothermic;
data in Supporting Information).
Table 2. Summary of Results of Acidity Bracketing of 3-Methylthymine (3-MeT, 3b)
343.8 ( 2.1
340.80 ( 0.60
339.3 ( 2.1
337.0 ( 2.1
335.8 ( 2.1
333.5 ( 2.9
331.0 ( 2.2
328.3 ( 2.9
336.7 ( 2.0
332.4 ( 2.0
330.4 ( 2.0
328.1 ( 2.0
326.5 ( 2.8
323.8 ( 2.0
322.0 ( 2.0
aAcidities are in kcal mol-1.64 bA “+” indicates the occurrence, and a “-” indicates the absence of proton transfer.
Figure 4. Calculated (B3LYP/6-31+G*) potential energy diagram of the reaction of 1,3-dMU (1d) with formate (∆H, kcal mol-1, 298 K).
J. AM. CHEM. SOC. 9 VOL. 131, NO. 51, 2009
Uracil and Thymine Reactivity in the Gas Phase
d. Methyl Chloride versus Pyrimidine Derivatives Acidity
Calculations Revisited. Experimentally, HCl and uracil have
similar acidities. However, our calculations indicate that,
although the N1-deprotonated uracil derivatives are still fairly
good leaving groups (for example, the reaction of methyl
thiolate, which is not that basic, with 1,3-dMU is exothermic
by 23.4 kcal mol-1and has a barrier of only 4.3 kcal mol-1
above the separated reactants), chloride is still better.
To assess whether this calculated difference in reactivity
might be a computational artifact, we compared the computed
acidities of 3-MeU and HCl. We know by experiment that the
two have comparable acidities.3,17However, we find that the
∆Hacid of HCl calculates to 325.1 kcal mol-1at B3LYP/6-
31+G*, 7.7 kcal mol-1more acidic than the calculated value
for 3-methyluracil, and 8.3 kcal mol-1more acidic than the
known measured ∆Hacid of HCl (333.4 ( 0.1 kcal mol-1).
Essentially, although 3-methyluracil and HCl have the same
experimental acidity, the B3LYP/6-31+G* calculations are not
accurate: HCl appears to be more acidic. To assess whether the
calculated differences in SN2 reaction barriers for the methyl
chloride and pyrimidine derivatives are due to a related
computational artifact or truly reflect the reactivity, we con-
ducted experiments, which are described later in this Article.
e. N1 versus N3 Attack. One possible complication for
examining these reactions experimentally is a potential competi-
tion between N1-CH3 attack (which we wish to see) and
N3-CH3attack (which is not of biological interest; Figure 6).
We computationally examined nucleophilic attack at the
N1-CH3versus N3-CH3for a series of nucleophiles (formate,
acetate, n-pentyl thiolate, n-propyl thiolate, methyl thiolate, and
anilide) and find that the N1-CH3SN2 reaction is consistently
on the order of 12 kcal mol-1more exothermic than the
N3-CH3path (Table 3).
We also calculated the activation enthalpies (∆Hq) for the
reactions with formate and methyl thiolate; for both reactions,
the barrier for attack at the N1-methyl group is 10 kcal mol-1
lower than that at the N3-methyl. These differences in barrier
are significant enough that, experimentally, SN2 reaction at
N1-CH3should be considerably favored over SN2 reaction at
N3-CH3. These results are as expected, given that the N1-H
is more than 10 kcal mol-1acidic than the N3-H.3,17
SN2 Reaction Studies: Experiments. The calculated barriers
for the reaction of formate with 1,3-dMU versus methyl chloride
indicate that the methyl chloride reaction should be faster
(Figures 4 and 5). Furthermore, Table 1 indicates that a
nucleophile with PA of 361.7 (CF3CH2O-) reacts with methyl
chloride with an efficiency of just 11%. We would therefore
expect that the SN2 reactions of nucleophiles whose PAs are
less than the ∆Hacidof C6-H (∆Hacid≈ 370 kcal mol-1) with
1,3-dMU and 1,3-dMT will all be relatively slow.
To establish that we can see SN2 reactivity under our
conditions, we repeated the known reactions of 2,2,2-trifluo-
roethoxide and methyl thiolate with methyl chloride. Our
efficiency values for these SN2 reactions are comparable to those
obtained previously (for 2,2,2-trifluoroethoxide, 13.3% (current
work) vs 11% (literature); for methyl thiolate, 8.0% (current
work) vs 4.7% (literature)).34,49,57
The reactions of a series of nucleophiles with 1,3-dMU and
1,3-dMT were studied (Table 4). The SN2 reaction product was
Figure 5. Calculated (B3LYP/6-31+G*) potential energy diagram of the reaction of methyl chloride with formate (∆H, kcal mol-1, 298 K).
Figure 6. N1 versus N3 attack of 1,3-dimethyluracil.
Table 3. Calculated Enthalpies of SN2 Reactions of 1,3-dMU (1d)
with a Series of Nucleophiles: N1-Methyl versus N3-Methyl
∆Hrxn, kcal mol-1
N1 attack N3 attack
18380J. AM. CHEM. SOC. 9 VOL. 131, NO. 51, 2009
Zhachkina and Lee
observed for the reaction of 1,3-dMU (1d) with nucleophiles
ranging from m-CF3PhO-(PA ) 339 kcal mol-1) to HO-(PA
) 390 kcal mol-1). However, when using nucleophiles with
PA greater than about 365 kcal mol-1, both proton transfer and
SN2 reactions were observed. Proton transfer is presumably the
result of deprotonation of the most acidic (C6-H) site of 1,3-
dMU (calculated ∆Hacid) 367.1 kcal mol-1; experimental ∆Hacid
) 369.9 ( 3.1 kcal mol-1).17,63The SN2 reaction efficiencies
of the reactions of nucleophiles with PA less than 365 kcal mol-1
with 1,3-dMU are all fairly low (less than 1%) (Table 4). For
nucleophiles with PA lower than 339 kcal mol-1, the efficiencies
were less than 0.01%.
We also examined the reactions of the same series of
nucleophiles with 1,3-dMT (3d). For nucleophiles with PA
higher than 365 kcal mol-1, both SN2 reactions and proton
transfer were observed (as we saw with 1,3-dMU). SN2 reaction
products were observed for nucleophiles with PAs as low as
∼352 kcal mol-1. Generally, the SN2 reaction efficiencies for
1,3-dMT (3d) are lower than those efficiencies for 1,3-dMU
(1d). Such small efficiencies are quite challenging to measure,
and therefore the precision for the 1,3-dMT (3d) measurements
varies from 0.001% to 0.02%. For the reaction of nucleophiles
with PA smaller than 352.5 kcal mol-1with 1,3-dMT (3d), no
SN2 reaction was observed (Table 4).
It therefore appears that the SN2 reaction proceeds for both
1,3-dMU (1d) and 1,3-dMT (3d) and that the efficiencies
observed for 1,3-dMT (3d) are lower than those for 1,3-dMU
Acidity. The experimental measurements of acidity indicate
that while 3-MeU and HCl have comparable acidities (around
333 kcal mol-1), 3-MeT is slightly less acidic, by 2-5 kcal
mol-1. The B3LYP/6-31+G* calculations involving HCl are
not quite in agreement with experiment. The computed (B3LYP/
6-31+G*) acidity for HCl is 325.1, while that for 3-MeU is
332.8 kcal mol-1. Thus, by calculation, the acidity difference
is more than 7 kcal mol-1, while by experiment that difference
is much less. The calculations are likely to be in error, because
the measured acidity of HCl is very well-known to be ∆Hacid
) 333.4 ( 0.1 kcal mol-1, and our previous bracketing studies
show that reaction of HCl and 3-MeU (i.e., the conjugate base
of one with the acid of the other and vice versa) proceeds in
both directions.17To establish that the discrepancy is due to a
computational artifact, we calculated the acidity of HCl and
3-MeU using the CBS-QB3 (complete basis set) model chem-
istry, which has been shown to accurately calculate thermo-
chemical values.70-74Using this method, we find that the
calculations and experimental data are in much better agreement:
for HCl, ∆Hacid) 332.2 kcal mol-1(calc) and 333.4 ( 0.1
kcal mol-1(expt); for 3-MeU, ∆Hacid) 335.5 kcal mol-1(calc)
and 333 ( 2 kcal mol-1(expt). Therefore, it does appear that
the discrepancy between calculations and experiments is a
computational issue. We therefore rely on our experimental data;
the experimental acidity studies indicate that Cl-and the
conjugate base of 3-MeU might be comparable leaving groups,
based on their acidities; 3-MeT would be a slightly worse
SN2 Reactions. Methyl Chloride versus Pyrimidines. The
calculated energy diagrams for the reactions of formate and
methyl thiolate with CH3Cl and 1,3-dMU are superimposed in
Figures 7 and 8. We are not indicating that the methyl chloride
and dimethyluracil systems start with the exact same total
energy; we plot them in such a way that the differences in
reactivity are easier to see. We leave off 1,3-dMT to keep the
diagrams uncluttered. In both the formate and the methyl thiolate
reactions, the methyl chloride energetics are more favorable than
the 1,3-dMU energetics (lower transition state energy and more
exothermic). The relevant values for discussion are the differ-
ences in the transition state energies for reaction of a given
nucleophile with 1,3-dMU versus methyl chloride (11.1 kcal
mol-1for formate (Figure 7); 11.8 kcal mol-1for methyl thiolate
(Figure 8)) and the differences in product energies for reaction
of a given nucleophile with 1,3-dMU versus methyl chloride
(7.3 kcal mol-1for formate (Figure 7); 7.4 kcal mol-1for methyl
thiolate (Figure 8)). We know from our acidity calculations that
the differences in the product energies are probably due to a
computational artifact: HCl calculates to be 7.7 kcal mol-1less
acidic than 3-MeU even though by experiment they have
comparable acidities (vide supra). This energy difference
between the acidity values of HCl and 3-MeU corresponds to
the ∼7 kcal mol-1difference in product energies for the
reactions in which chloride and deprotonated 3-MeU are the
leaving groups (Figures 7 and 8). The transition state energies,
however, show a larger difference for the 1,3-dMU versus
methyl chloride reactions (11.1 kcal mol-1for formate and 11.8
kcal mol-1for methyl thiolate, favoring the CH3Cl reaction).
The accuracy of these SN2 reaction transition state energies is
certainly called into question based on the failure of B3LYP/
6-31+G* to correctly predict the acidity of HCl. However, DFT
methods have been shown to be reasonable for SN2 reaction
energetics in degenerate reactions involving chloride as the
nucleophile and leaving group.71,75-78The consistent qualitative
conclusion from Figures 7 and 8 is that the methyl chloride
SN2 reaction appears to have a lower barrier than that of 1,3-
(70) Wheeler, S. E.; Moran, A.; Pieniazek, S. N.; Houk, K. N. J. Phys.
Chem. A 2009, 113, 10376–10384.
(71) Vayner, G.; Houk, K. N.; Jorgensen, W. L.; Brauman, J. I. J. Am.
Chem. Soc. 2004, 126, 9054–9058.
(72) Ochterski, J. W.; Petersson, G. A.; Montgomery, J. A., Jr. J. Chem.
Phys. 1996, 104, 2598–2619.
(73) Montgomery, J. A., Jr.; Frisch, M. J.; Ochterski, J. W.; Petersson,
G. A. J. Chem. Phys. 1999, 110, 2822–2827.
(74) Montgomery, J. A., Jr.; Frisch, M. J.; Ochterski, J. W.; Petersson,
G. A. J. Chem. Phys. 2000, 112, 6532–6542.
(75) Bento, A. P.; Sola, M.; Bickelhaupt, F. M. J. Comput. Chem. 2005,
(76) van Bochove, M. A.; Bickelhaupt, F. M. Eur. J. Org. Chem. 2008,
(77) Westaway, K. C.; Fang, Y.; MacMillar, S.; Matsson, O.; Poirier, R. A.;
Islam, S. M. J. Phys. Chem. A 2007, 111, 8110–8120.
(78) Swart, M.; Sola, M.; Bickelhaupt, F. M. J. Comput. Chem. 2007, 28,
Table 4. SN2 Reactions of 1,3-dMU (1d) and 1,3-dMT (3d)
substrate (efficiency of SN2 reaction, %)
0.38 ( 0.07
0.84 ( 0.38
0.35 ( 0.21
0.23 ( 0.14
0.22 ( 0.07
0.23 ( 0.21
0.28 ( 0.04
0.11 ( 0.09
0.08 ( 0.04
0.04 ( 0.02
0.006 ( 0.003
0.004 ( 0.001
0.033 ( 0.019
0.009 ( 0.003
0.014 ( 0.009
0.004 ( 0.001
0.008 ( 0.006
no SN2 reaction
no SN2 reaction
no SN2 reaction
J. AM. CHEM. SOC. 9 VOL. 131, NO. 51, 2009
Uracil and Thymine Reactivity in the Gas Phase
dMU. That is, although chloride and deprotonated 3-MeU have
similar basicities, chloride is the better leaving group.
The experimental SN2 reaction data qualitatively support the
calculations (Tables 1 and 4). For example, for the reaction of
1,1,1-trifluoroethoxide with methyl chloride, the efficiency is
11% (literature) to 13% (our lab). With 1,3-dMU, the efficiency
with the same nucleophile is less than 1% (a rate difference of
about 15×, corresponding to about a 1.5 kcal mol-1difference
in barrier). Thus, both calculations and experiments do indicate
that, although HCl and 3-MeU have comparable acidities in the
gas phase, the conjugate bases do not appear to be equivalently
good leaving groups. It appears that the calculations may
overestimate that difference (11 kcal mol-1by calculation versus
about 2 kcal mol-1by experiment). We understand that our
efficiency values for the pyrimidine reactions are incredibly low,
and it is in fact almost impossible to fully discount the possibility
that the 1,3-dMU sample (and the 1,3-dMT sample) are not
contaminated with some monomethyl substrate. Should, for
example, 3-methyluracil be present in the 1,3-dMU sample, then
deprotonation at N1-H of that contaminant 3-MeU would result
in the same product ion expected from the SN2 reaction. The
1,3-dMU purchased from Sigma Aldrich has a purity of 99%.
We synthesized the 1,3-dMT sample from thymine. The
compounds appear pure by NMR and also by our mass
spectrometric studies of reaction with H3O+, where we see only
the mass-to-charge ratio corresponding to protonated 1,3-dMU
and 1,3-dMT, and do not see the m/z ratio corresponding to
protonated monomethylated compound. We are therefore quite
confident that what we see are very slow SN2 reactions but
cannot discount the possibility that deprotonation of a trace
amount of monomethylated compound contributes to the
observed product ions. However, if there is such a contaminant,
that would mean that the contribution of the SN2 reaction to
the total reaction efficiency would be even lower than the
reported values. Therefore, the reported efficiency values for
the pyrimidine reactions are an upper limit. Our data indicate
that the methyl chloride reactions are on the order of 15 times
faster than the 1,3-dMU reactions; this is a lower limit in the
sense that our SN2 reaction efficiencies might be lower than
The reason for this is probably related to the nature of the
respective leaving groups: deprotonated uracil is a delocalized
ion while chloride is not. In the neutral uracil, the N1 electrons
can delocalize into the π system, but such delocalization
separates charge, and therefore those resonance structures
probably contribute less to the actual structure. Deprotonation
results in a negative charge that can be stabilized by pushing
that charge into the carbonyl oxygens (Figure 9). This is the
same argument that, for example, could be used to explain why
N-bromosuccinimide is an effective brominating agent.
Acidity is, of course, a thermodynamic property, while leaving
group ability is essentially a kinetic property: how fast is the
SN2 reaction? In the SN2 reaction transition state, because the
N1-deprotonated uracil anion is not yet fully formed, the stability
Figure 7. Superimposed energy diagrams for reactions of 1,3-dMU and methyl chloride with formate (B3LYP/6-31+G*, 298 K).
18382J. AM. CHEM. SOC. 9 VOL. 131, NO. 51, 2009
Zhachkina and Lee
provided by electron delocalization in the product is not yet
completely in place. Chloride, on the other hand, is a polarizable
entity whose good leaving group properties are likely present
in the transition state; little electron reorganization is required.
In other words, the special stability enjoyed by the deprotonated
uracil N1-anion, while making uracil as thermodynamically
acidic as HCl, helps only partially in the leaving group ability:
in the SN2 reaction transition state, the N1 anion is not fully
formed and therefore not fully delocalized.79-81(Presumably
the carbonyl groups also stabilize the anion by induction, but
one would expect that effect to be seen in both the acidity and
the SN2 reactivity.) In summary, delocalization in an anion may
make its conjugate acid acidic, but may not affect leaving group
ability as much as expected.79-82
We examined the bond lengths and charge distribution
(CHELPG) for the participants in the SN2 reaction of formate
and methyl thiolate with 1,3-dMU to probe these ideas further.
In Table 5, we list the distance for the breaking N1-CH3bond
(r1), the distance for the forming Nu-CH3bond (r2), and the
charges on the O2 and O4. For the reactant, negative charges
of -0.560 and -0.579 reside on the O2 and O4, respectively.
Those values increase (negatively) to -0.767 and -0.735 in
the product. For the reaction with formate, the transition state
is predictably a little late (based on the r1 length, relative to
(79) Gronert, S. J. Am. Chem. Soc. 1993, 115, 10258–10266.
(80) Gronert, S. Organometallics 1993, 12, 3805–3807.
(81) Gronert, S.; Glaser, R.; Streitwieser, A. J. Am. Chem. Soc. 1989, 111,
(82) Schreiner, P. R.; Schleyer, P. v. R.; Schaefer, H. F., III. J. Org. Chem.
1997, 62, 4216–4228.
Figure 8. Superimposed PES for reactions of 1,3-dMU and methyl chloride with methyl thiolate (B3LYP/6-31+G*, 298 K).
Figure 9. Resonance structures of neutral and N1-deprotonated uracil.
J. AM. CHEM. SOC. 9 VOL. 131, NO. 51, 2009
Uracil and Thymine Reactivity in the Gas Phase
the methyl thiolate reaction); the Nu-CH3bond is more formed
(1.94 Å), and the N1-CH3bond is quite elongated (2.03 Å).
The charge distributions are more negative than in the reactant,
but not nearly as negative as in the product (-0.670 and -0.686
for the O2 and O4, respectively). For the reaction with methyl
thiolate, the transition state is earlier than that for formate, which
would be expected because the reaction with methyl thiolate is
more exothermic (methyl thiolate is a more basic nucleophile).
The breaking N1-CH3bond is shorter (1.94 Å) than the forming
Nu-CH3bond (2.52 Å). Consistent with the earlier TS, the
oxygens are not as negative as they are in the formate transition
state (-0.641 and -0.683 for methyl thiolate, versus -0.670
and -0.686 for the O2 and O4, respectively, for formate). There
does seem to be an interesting paradox for reactions where the
leaving group gains stability from delocalization: as the nu-
cleophile becomes more basic, the reaction becomes more
exothermic, which is consistent with a slightly lowered barrier
(as we see with the formate reaction, which has a ∆Hqof 9.2
kcal mol-1above the separated reactants versus that of methyl
thiolate, which is only 4.3 kcal mol-1).83,84However, as the
reaction becomes more exothermic, the transition state may
move earlier, where charge might be less delocalized, which
would have an opposing effect on rate.
In summary, although uracil is as acidic as HCl, deprotonated
uracil is not as good of a leaving group as chloride due to
electron delocalization-related issues. This is not to say that
deprotonated uracil is necessarily a terrible leaving group; it is
clearly the species that leaves in the UDG reaction.5,18,20,25
In this study, we are also interested in comparing 1,3-dMU
to 1,3-dMT. Because UDG cleaves uracil but not thymine, we
wanted to probe whether there was an intrinsic reactivity
component that would favor uracil cleavage. The calculations
for the formate and methyl thiolate reactions with 1,3-dMU
versus 1,3-dMT indicate that the barriers for the 1,3-dMU
reactions are usually about 1 kcal mol-1lower (relative to the
separated reactants) than the reactions with 1,3-dMT. The
measured reaction efficiencies for the uracil reactions are
consistently higher than for the analogous thymine reactions
(Table 4). The efficiency values are very small, but the 1,3-
dMU reactions are always at least 10 times higher in efficiency
than the corresponding 1,3-dMT reactions. Assuming that the
efficiency results are not due in part to deprotonation reactions
of a monomethylated contaminant, it does appear that the uracil
reactions are faster than the thymine reactions, which is
consistent with the calculations and the measured acidities
(wherein 3-MeU is more acidic than 3-MeT, vide supra). Thus,
intrinsically, deprotonated uracil is more easily cleavable than
deprotonated thymine, which could be one factor (of many)
aiding in the discrimination between the two by UDG in DNA.
Last, although we cannot be sure that nucleophilic attack
occurs at N1 (rather than N3) of 1,3-dMU and 1,3-dMT, our
calculations indicate that the difference in reactivity of those
sites is so great (more than 10 kcal mol-1, vide supra) that attack
at N1 is likely. Future studies with appropriately deuterated
substrates will be conducted.
We find that although the acidities of HCl and 3-methyluracil
are comparable in the gas phase, the leaving group abilities of
chloride and deprotonated 3-methyluracil are different, with
chloride being slightly better. The basis of this difference lies
in the fact that deprotonated 3-methyluracil is thermodynami-
cally very acidic due to delocalization, which does not yield as
large of a beneficial effect in the SN2 reaction transition state.
Comparison of calculations to experiments indicates that the
B3LYP/6-31+G* estimates the acidity of the pyrimidine
derivatives well, but not that of HCl (the error is on the order
of 7 kcal mol-1, with the calculations yielding too low of a
value). In terms of the SN2 reactions, the calculations predict
that those with methyl chloride would be favored over those of
3-methyluracil. The experiments also indicate that the methyl
chloride reactions are faster than those of 1,3-dimethyluracil,
with rates of at least 15-20 times more.
We also compared the leaving group ability of 1,3-dimethy-
luracil versus 1,3-dimethylthymine. Our calculations and experi-
ments indicate that the N1-deprotonated uracil derivative is a
slightly better leaving group than the N1-deprotonated thymine
derivative, which is consistent with the cleavage of the former
but not the latter from DNA.
1,3-dMU (1d) and 3-MeT (3b) are commercially available and
were used as received. 1,3-dMT (3d) was synthesized from thymine
(3a). The procedure used is similar to that reported in the literature
for 1,3-dMU synthesis.85,86The product was purified, and the
identity was confirmed by1H and13C NMR.
(83) Hammond, G. S. J. Am. Chem. Soc. 1955, 77, 334–338.
(84) Evans, M. G.; Planyi, M. Trans. Faraday Soc. 1938, 34, 11–24.
(85) Newkome, G. R.; Nayak, A.; Otemaa, J.; Van, D. A.; Benton, W. H.
J. Org. Chem. 1978, 43, 3362–3367.
(86) Milligan, J. R.; Ward, J. F. Radiat. Res. 1994, 137, 295–299.
Table 5. Calculated (B3LYP/6-31+G*) Distances and Charges (CHELPG) for SN2 Reactions of 1,3-dMU
substrater1 (N1-CH3distance) (Å)
r2 (CH3-Nu distance) (Å)
N1-deprotonated 3-MeU (product)
TS with Nu-) formate
TS with Nu-) methyl thiolate
18384 J. AM. CHEM. SOC. 9 VOL. 131, NO. 51, 2009
Zhachkina and Lee
All nucleophiles except CH3S-were generated from com-
mercially available neutral reference acids by deprotonation with
hydroxide ions. CH3S-was generated via the elimination reaction
of hydroxide plus dimethyldisulfide, which is a well-known source
of the methyl thiolate anion:87-89
All SN2 reaction experiments and 3-MeT acidity bracketing were
conducted using a dual cell Fourier transform ion cyclotron
resonance mass spectrometer (FTMS), which has been described
previously.3,35,39The magnetic field is produced by a 3.3 T
superconducting magnet. Gas-phase reactions occur in a 1 in. cubic
cell located inside the instrument, which is pumped down to a
baseline pressure 1 × 10-9Torr. The solid nucleobases were
introduced to the cell via the solids probe. Nucleophile anions were
generated by deprotonation of commercially available reference
acids with hydroxide ions. To introduce the reference acids into
the cell, a leak valve or batch inlet was used. Hydroxide ions are
produced via electron impact of water (typically 8 eV, 6 µA, 0.5 s).
For our acidity bracketing experiments, we utilized a protocol
described previously.3,17,35,38-40Proton transfer reactions were
conducted in both directions (deprotonated 3-MeT with neutral
reference acids and the conjugate bases of neutral reference acids
plus 3-MeT). The occurrence or nonoccurrence of proton transfer
is denoted as a “+” or “-” in Table 2.
The reaction efficiency is the percentage ratio of collisions that
lead to product. We report the ratio of observed rate constant to
the estimated collisional rate constant calculated by parametrized
trajectory theory.61,62Dipole moments were calculated at B3LYP/
6-31+G*; polarizability was estimated using the method of Miller
and Savchik.90Each kinetics experiment was run at least three
times; reported values are the average and standard deviation.
The Cooks kinetics method39,40,65-69was used to conduct a
relative acidity study of 3-MeT (3b) and 3-MeU (1b). We use a
quadrupole ion trap (LCQ) mass spectrometer to conduct the
experiments. The detailed protocol for Cooks kinetic experiments
conducted in our lab has been described previously.38,39Briefly,
the “relative” Cooks kinetic method we use herein involves
formation of a proton-bound dimer of the two species of interest.
The dimer is isolated and dissociated via CID. The ratio of
intensities of the two deprotonated substrates yields the ratio of
rate constants of the two possible dissociation pathways, which
yields the relative acidities of the two substrates.
Proton-bound dimers were generated by electrospray (ESI) from
the 1:1 mixture of 250 µM 3-MeT (3b) and 250 µM 3-MeU (1b)
solutions in 20% methanol-water. A needle voltage of 4 kV,
capillary temperature 150 °C, and a flow rate of about 25 µL/min
were typically used. The proton-bound complex [(3-MeU)-·H+·(3-
MeT)-] was isolated and activated for about 30 ms. 40 scans were
averaged for the product ions, and the experiment was repeated
three times. A Teffof 420 K, obtained from a calibration experiment
with 3-MeT (3b), was used.
Calculations were conducted at B3LYP/6-31+G* using Gaussian
03;91-94the geometries were fully optimized and frequencies were
calculated. This method has been shown to be reasonable for
calculating SN2 reaction potential energy surfaces.71,75,77,78All of
the values reported are at 298 K. No scaling factor was applied.
All calculated TS structures have one negative frequency. Partial
charges were calculated using CHELPG as implemented in Gauss-
ian 03.95As described in this Article, we also for some of the
substrates utilized the CBS-QB3 model chemistry.73,74
Acknowledgment. We gratefully thank the NSF, the Alfred P.
Sloan Foundation, and the National Center for Supercomputer
Applications for support. We acknowledge Dr. Stephen Shi for
early, preliminary experiments.
Supporting Information Available: Cartesian coordinates for
all calculated species, additional ∆H and ∆G diagrams, other
additional data as noted in the manuscript, and full citations
for references with greater than 16 authors. This material is
available free of charge via the Internet at http://pubs.acs.org.
(87) Grabowski, J. J.; Zhang, L. J. Am. Chem. Soc. 1989, 111, 1193–1203.
(88) Lee, J. K.; Grabowski, J. J. J. Org. Chem. 1996, 61, 9422–9429.
(89) Schulze, S. M.; Santella, N.; Grabowski, J. J.; Lee, J. K. J. Org. Chem.
2001, 66, 7247–7253.
(90) Miller, K. J.; Savchik, J. A. J. Am. Chem. Soc. 1979, 101, 7206–
(91) Frisch, M. J.; Gaussian 03; Gaussian, Inc.: Wallingford, CT, 2004.
(92) Kohn, W.; Becke, A. D.; Parr, R. G. J. Chem. Phys. 1996, 100, 12974–
(93) Becke, A. D. J. Chem. Phys. 1993, 98, 5648–5652.
(94) Lee, C.; Yang, W.; Parr, R. G. Phys. ReV. B 1988, 37, 785–789.
(95) Breneman, C. M.; Wiberg, K. B. J. Comput. Chem. 1990, 11, 361–
J. AM. CHEM. SOC. 9 VOL. 131, NO. 51, 2009
Uracil and Thymine Reactivity in the Gas Phase