α-(Acyloxy)dialkylnitrosamines: Effects of structure on the formation of N-nitrosiminium ions and a predicted change in mechanism

Article abstract of DOI:10.1021/ja9833218

The decay of α-(acyloxy)dialkylnitrosamines in aqueous solutions has been studied with a view toward elucidating mechanistic details and effects of structure on mechanism and reactivity. Rate constants (k₁) for the pH-independent decay of 43 α-(acyloxy)dialkylnitrosamines have been determined. Observations from these and other experiments rule out decomposition via an anchimeric assistance mechanism involving the Z isomer that had previously been suggested. All of the reported data for most of the compounds is consistent with a mechanism involving the formation of N-nitrosiminium ions in or before the rate-limiting step. Structure -reactivity correlations indicate that the stability of α-(acyloxy)dialkylnitrosamines is determined by electronic properties of substituents at R_N and R_C as well as by the ability of substituents R_C to engage in hyperconjugative interactions of C-H bonds with the developing cationic center in the transition state for nitrosiminium ion formation. Attachment of substituents of sufficient electron-withdrawing power at R_N and R_C results in a predicted change in mechanism to what appears to be an acyl group attack mechanism.

Full text of DOI:10.1021/ja9833218

1826
J. Am. Chem. Soc. 1999, 121, 1826-1833
R-(Acyloxy)dialkylnitrosamines: Effects of Structure on the
Formation of N-Nitrosiminium Ions and a Predicted Change in
Mechanism
Hongliang Cai and James C. Fishbein*
Contribution from the Department of Chemistry, Wake Forest UniVersity,
Winston-Salem, North Carolina 27109
ReceiVed September 17, 1998
Abstract: The decay of R-(acyloxy)dialkylnitrosamines in aqueous solutions has been studied with a view
toward elucidating mechanistic details and effects of structure on mechanism and reactivity. Rate constants
(k1) for the pH-independent decay of 43 R-(acyloxy)dialkylnitrosamines have been determined. Observations
from these and other experiments rule out decomposition via an anchimeric assistance mechanism involving
the Z isomer that had previously been suggested. All of the reported data for most of the compounds is consistent
with a mechanism involving the formation of N-nitrosiminium ions in or before the rate-limiting step. Structure-
reactivity correlations indicate that the stability of R-(acyloxy)dialkylnitrosamines is determined by electronic
properties of substituents at RN and RC as well as by the ability of substituents RC to engage in hyperconjugative
interactions of C-H bonds with the developing cationic center in the transition state for nitrosiminium ion
formation. Attachment of substituents of sufficient electron-withdrawing power at RN and RC results in a
predicted change in mechanism to what appears to be an acyl group attack mechanism.
Introduction
had been observed in the case of â-(sulfatooxy)nitrosamines.5
Recently, data, in the form of product analysis and structure
activity effects, were summarized that indicated that nitroso
oxygen attack at the carbonyl carbon was not a likely mechanism
of reaction.6 It was concluded that the dominant mechanism
involved the formation of nitrosiminium ions (D) in, or prior
to, the rate-limiting step, as in eq 3.6-9 The onset of a carbonyl
Many dialkylnitrosamines (A) are enzymatically activated to
their relatively unstable R-hydroxy derivatives (B), which give
rise to DNA-alkylating diazonium ions and carbocations, as
illustrated in eq 1.1,2 R-(Acyloxy)dialkylnitrosamines (C) are
attack mechanism with electron-withdrawing substituents RN
and RC (C, eq 1) was predicted.6 An alternative mechanism or
mechanisms of anchimeric assistance were not addressed, and
the implications of the apparent correlation between isomeric
composition and stability have not been identified. To clarify
mechanistic aspects and the dominant structure-reactivity
effects in this biologically important class of compounds, we
have undertaken a more detailed study of the solvolysis of
R-(acyloxy)dialkylnitrosamines.
extensively used as precursors to R-hydroxydialkylnitrosamines
in studies of nitrosamine carcinogenesis.1,2 The esters vary
widely in their reactivity, some being more unstable than the
R-hydroxydialkylnitrosamine products.3
The effects of structure on the reactivity of the acyloxy
compounds has been a source of confusion. The limited
information available suggests that, for some substituents, the
effects of attachment to nitrogen (RN, eq 1) are opposite those
attached to carbon (RC, eq 1).4,5 An early observation indicated
that substituents that increase the relative proportion of the Z
form of the ester, eq 2, increase reactivity.5 This had been the
In summary, correlations of electron-withdrawing power of
substituents at RN with reactivity and experiments to determine
(1) Lawley, P. D. In Chemical Carcinogens; Searle, C. D., Ed.; ACS
Monograph 182; American Chemical Society; Washington, DC, 1984.
(2) Lijinsky, W. Chemistry and Biology of N-nitroso Compounds;
Cambridge University Press: Cambridge, U.K., 1992.
(3) Chahoua, L.; Mesic´, M.; Revis, C. L.; Vigroux, A.; Fishbein, J. C.
J. Org. Chem. 1997, 62, 2500.
(4) Pool, B. L.; Wiessler, M. Carcinogenisis 1981, 2, 991.
(5) Wiessler, M. In N-Nitrosamines; Anselme, J.-P., Ed.; American
Chemical Society: Washington, DC, 1979; p 57.
(6) Revis, C.; Rajama¨ki, M.; Fishbein, J. C. J. Org. Chem. 1995, 60,
7733.
(7) Rajama¨ki, M.; Vigroux, A.; Chahoua, L.; Fishbein, J. C. J. Org. Chem.
1995, 60, 2324.
(8) Cai, H.; Fishbein, J. C. Tetrahedron 1997, 53, 10671.
(9) Vigroux, A.; Kresge, A. J.; Fishbein, J. C. J. Am. Chem. Soc. 1995,
117, 4433.
basis of the hypothesis that there was anchimeric assistance by
the nitroso group of decomposition, possibly analogous to what
10.1021/ja9833218 CCC: $18.00 © 1999 American Chemical Society
Published on Web 02/18/1999

R-(Acyloxy)dialkynitrosamines
J. Am. Chem. Soc., Vol. 121, No. 9, 1999 1827
18O-incorporation into products from solvent rule out the two
conceivable mechanisms of anchimeric assistance. The stability
of R-(acyloxy)dialkylnitrosamines is determined by electronic
properties of substituents at RN and RC as well as by the ability
of substituents RC to engage in hyperconjugative interactions
of C-H bonds with the developing cation in the transition state
for formation of the nitrosiminium ion. The hyperconjugative
effect dominates the reactivity effects of simple alkyl substit-
uents and inverts the reactivity order expected on the basis of
electronic considerations alone. Attachment of substituents of
sufficient electron-withdrawing power at RN and RC results in
the predicted change in mechanism to what appears to be an
acyl group attack mechanism.
Kinetics. The kinetics of decomposition of nitrosamines were
observed using either a Milton Roy 1001+ or 3000 diode array
spectrophotometer, a Hewlett-Packard 8452A diode array spectropho-
tometer, or an Applied Photophysics DX.17MV stopped-flow spectro-
photometer. Temperature was maintained at 25 °C by means of attached
recirculating water baths.
The buffer solution in each reaction cell was prepared by diluting a
concentrated stock buffer solution. Ionic strengths of reaction solutions
were typically maintained at 1 M with NaClO4. The buffers that were
used include phosphate, ethanolamine, hydrazine, MOPS, MES, CHES,
acetic acid, cyanoacetic acid, methoxyacetic acid, cacodylic acid, and
morpholine.
The stock solution of the nitrosamine was prepared in freshly distilled
dry CH3CN. Reactions performed in the stopped-flow spectrophotom-
eter were carried out by mixing 25 parts of the aqueous buffer with 1
part of the nitrosamine in CH3CN solution using 2.5 and 0.1 mL
syringes, respectively. Final concentration of the nitrosamine in the
reactions was ∼10-4 M.
Experimental Section
Warning! Many R-(acyloxy)nitrosamines have been shown to be
direct-acting carcinogens. Handling precautions entail the use of a well-
ventilated fume hood and frequently changed double pairs of protective
gloves. All materials suspected of contact are treated in 50% aqueous
sulfuric acid baths containing the commercially available oxidant No-
Chromix (Aldrich).
Reagents. Water that was used as solvent in the kinetic studies was
redistilled from house-distilled water. Water for HPLC was redistilled
from house-distilled water and filtered. Methanol and ethanol were dried
and distilled from magnesium methoxide and ethoxide, respectively.
Ethyl ether was dried and distilled from sodium. Acetonitrile, dichlo-
romethane, chloroform, hexane, pentane, and ethyl acetate were dried
and distilled from calcium hydride. Reagents for syntheses and kinetics
were ACS grade or better, unless otherwise specified.
The pH values were obtained after the kinetic run using an Orion
model SA 720 pH meter equipped with a Corning combination
electrode. Prior to recording reaction pH values, we performed two-
point calibrations using commercially available standards or those
prescribed by ref 12.12
Product Analyses. Products were quantitated by 1H NMR. The
reaction solutions were buffered with either 0.05 M cacodylic acid (50%
anion) or 0.05 M chloroacetate (50% anion) and prepared in screw-
cap vials, which contained either 0.80 mL of D2O and 0.20 mL of
CD3CN or 0.90 mL of D2O and 0.10 mL of CD3CN. NaClO4 was used
to maintain the ionic strength of the reaction solutions. About 0.01 g
of a substrate was weighed out and transferred into the vial, and the
reaction was allowed to proceed for more than 10 half-times. The
1
reaction mixture was transferred into a 5 mm NMR tube, and the H
Syntheses. Imine precursors to the R-(acyloxy)nitrosamines were
synthesized by the method of Graymore10 and were typically purified
before use by distillation. Generally, the crude yield was 70%.
R-(Acyloxy)nitrosamines were synthesized from the corresponding
imines as described previously.6-9 The yields after final purification
were 20-50%. Assignment of E and Z configurations was based on
the method of Karabatsos.11 In all but two cases, assignment based on
the shift in R-protons was in agreement with that made by use of the
shift in â-protons. In the two remaining cases, the shift in â-protons
gave assignment in accord with those for other compounds with
sterically similar substituents. The E/Z ratios reported refer to the solvent
CDCl3. For two compounds, 13 and 18, the E/Z ratios in 60% D2O/
40% CD3CN were also determined and did not differ considerably from
the ratios determined in CDCl3.
NMR spectrum of the reaction mixture was recorded. The integration
values for signals of corresponding alcohols, ammonium ions in the
cases of (N-nitrosomethylamino)-p-chlorobenzyl and -p-(trifluoro-
methyl)benzyl chloroacetates, and acetaldehyde and methanol in the
case of (N-nitroso(methoxyethyl)amino)ethyl acetate were used to
calculate their percentages on the basis of integration of C-H proton
signals from acetate or chloroacetate leaving groups.
18O Incorporation. The extent of incorporation of 18O from solvent
water into benzaldehyde in the decomposition of 42 and 43 was
investigated. Decomposition of 42 was carried out at 25 °C in 0.1 mL
of 0.05 M 50% anion cacodylic acid buffer solution (µ ) 1 M (NaClO4))
which contained ∼50% H218O and ∼50% H216O. The reaction was
allowed to proceed for 3 min (∼8 half-lives). In the case of 43, the
reaction was carried out at 25 °C in 0.02 M 20% anion cacodylic acid
buffer solution (µ ) 1 M (NaClO4)) containing ∼50% H218O and ∼50%
H216O and was allowed to proceed for 20 min (∼6 half-lives). At the
end of the reaction times, in both cases, the product benzaldehyde was
extracted into ether and analyzed by GC-MS with an HP-1 (cross-
linked methylsiloxane), 12 m × 0.2 mm, capillary column. The relative
amounts of 18O and 16O in benzaldehydes were determined by
monitoring the signals at masses 107, 108 and 105, 106 for the
benzaldehydes, respectively. The relative intensities for a given sample
were the mean values of a minimum of four analyses in both cases.
Determination of the 16O/18O ratio of the solvent was accomplished by
recombination of the ether layers with the original reaction solvents
and evaporation of the ether under a stream of dry argon followed by
reincubation for 24 h. The 16O/18O content of the aldehydes was
subsequently redetermined by repeating the analysis method above.
Control experiments with 16O -benzaldehyde indicated that oxygen
exchange was complete after 24 h under these conditions.
For compounds 1, 4, 5, and 6, syntheses and analytical data have
been published previously.6-9
A typical procedure is given for (N-nitrosomethylamino)ethyl acetate
(7). N-Methylethyleneamine, 1.0 g, was placed in a 50 mL round-bottom
flask. A volume, ∼25 mL, of dry CH2Cl2 was added to the flask,
followed by cooling in an ethanol/water/(CO2)s bath (-10 °C). A
solution of 10 mL of dry CH2Cl2, 1.7 g of dry Et3N, and 1.1 g of acetic
acid was added to the flask dropwise. Then 2.0 g of NOBF4 was added
slowly to the reaction mixture, which bubbled and turned yellow. When
the bubbling ceased, the ice bath was removed, and the reaction mixture
was stirred at room temperature for 1/2 h. CH2Cl2 was evaporated under
a gentle stream of argon. The remaining material was extracted with
ether three times, and the combined ether solutions were evaporated
under a stream of argon. The product was purified by preparative silica
gel TLC using 1:4 ether/hexane as the eluting solvent. See the
Supporting Information for characterization of the compounds synthe-
sized.
Deuterated compounds 18-d1 and 4-d3, with the isotopes incorporated
at the carbon atom â to the ester oxygen, were synthesized starting
from the deuterated aldehydes. The final products were >97%
Results
Rate constants for the decay of R-(acyloxy)dialkylnitro-
samines were determined from absorbance decay curves that
exhibited good first-order behavior for 3-5 half-lives. Most of
the determinations were in the pH range between 2 and 9 at 25
1
deuterated by H NMR.
(10) Graymore, J. J. Chem. Soc. 1932, 1353.
(11) Karabatsos, G. J.; Vane, F. M.; Taller, R. A.; Hsi, N. J. Am. Chem.
Soc. 1964, 86, 3351. Karabatsos, G. L.; Taller, R. A. J. Am. Chem. Soc.
1964, 86, 4373.
(12) The Merck Index, 8th ed.; Stecher, P. G., Ed.; Merck & Co.:
Rahway, NJ, 1968.

1828 J. Am. Chem. Soc., Vol. 121, No. 9, 1999
Cai and Fishbein
Table 1. Kinetics Data for R-(Acyloxy)nitrosaminesa at 25 °C in Aqueous Solution (Ionic Strength 1 M (NaClO4))
compd
RN
RC
X
k1 , s-1
b
% Z isomerc
1
2
3
4
5
6
7
8
9
tert-butyl
methoxyethyl
dimethoxyethyl
tert-butyl
isopropyl
ethyl
methyl
methoxymethyl
benzyl
dimethoxyethyl
trifluoroethyl
R-methylbenzyl
tert-butyl
isopropyl
n-butyl
ethyl
methyl
tert-butyl
isopropyl
ethyl
methyl
methoxyethyl
ethyl
methyl
ethyl
methyl
ethyl
methyl
ethyl
methyl
methyl
ethyl
ethyl
CH3OCH2CH2
methyl
methyl
methyl
methyl
tert-butyl
tert-butyl
tert-butyl
methyl
hydrogen
hydrogen
hydrogen
methyl
methyl
methyl
methyl
methyl
methyl
methyl
methyl
methyl
ethyl
OAc
OAc
OAc
OAc
OAc
OAc
OAc
OAc
OAc
OAc
OAc
OAc
OAc
OAc
OAc
OAc
OAc
OAc
OAc
OAc
OAc
OAc
OAc
OAc
OAc
OAc
OAc
OAc
OAc
OAc
OAc
OAc
3.76((0.21)e-5
4.08((0.41)e-7
2.39((0.53)e-7
2.02e-3c
1.43e-3
1.03e-3
4
10
11
76
23
10
10
0
5
22
3
83
72 (64)e
23
8
6.86((0.50)e-4
3.66((0.10)e-4
1.85((0.12)e-4
1.72((0.17)e-4
2.31((0.14)e-5
2.84((0.02)e-4d
1.059((0.04)e-3
6.41((0.37)e-4
4.75((0.28)e-4
4.45e-4
2.52((0.28)e-4
2.99((0.12)e-4
1.582((0.07)e-4
8.67((0.39)e-5
6.39((0.74)e-5
1.976((0.12)e-5
1.29((0.08)e-5
6.65((0.43)e-6
1.24((0.01)e-3
8.36e-4
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
ethyl
ethyl
ethyl
ethyl
15
0
isopropyl
isopropyl
isopropyl
isopropyl
isopropyl
tert-butyl
tert-butyl
CH2C(CH3)3
CH2C(CH3)3
PhCH2CH2
PhCH2CH2
CH3OCH2CH2
2-methoxyethyl
2-methoxy-1-methylethyl
methoxymethyl
dichloromethyl
H
70 (69)e
29
0
16
4
6
0
0
0
9
0
12
0
2.03e-4
1.0e-4
1.34((0.054)e-4
7.94((0.88)e-5
2.85((0.33)e-5
7.15((0.93)e-6
2.93((0.31)e-6
8.55e-5
6.26((0.39)e-3
1.46((0.09)e-3
7.99((0.11)e-3
7.87e-4
0
19
17
13
0
<5
7
OAc
OAcCl
OAcOMe
OAcOMe
OAcCl
OAcCl
OcHex
OAc
O2CPh
OAcCl
OAcCl
ethyl
isopropyl
isopropyl
tert-butyl
0
methyl
7.92e-4
71
93
90
4
p-chlorophenyl
p-chlorophenyl
phenyl
1.31((0.21)e-4
5.6e-4d
4.6 × 10-2 d
3.30((0.09)e-3
tert-butyl
phenyl
90
a See structure C, eq 1, for general formula. b The pH-independent rate constant for decay. The average value of the buffer-independent rate
constant, k0, for four or more determinations in the range of pH 2-9. c Determined in CDCl3, except where noted. d Value from a single determination
of k0 at pH 6.1, cacodylic acid buffer. e Value in parentheses determined in 60% D2O/40% DMSO-d6.
°C and ionic strength 1 M. Rate constants were determined at
a given buffer ratio typically at three buffer concentrations. In
most cases, there was little effect on the magnitude of the rate
constant kobsd, over the range of concentration studied, 0.02-
0.2 M. Increases in kobsd of 20% or less were typical. In these
cases, the buffers included cacodylic, chloroacetic, and meth-
oxyacetic acids, CHES, and the ethylenediamine dication.
Reactions in morpholine and bisphosphate buffers showed
increases in kobsd of up to 50% above the extrapolated buffer-
independent value, k0. Effects of increasing concentrations of
acetate ions and acetate buffer solutions were variable and are
discussed in more detail below. Values of k0 were obtained from
linear extrapolations to the y-intercept of plots of kobsd against
buffer concentration. Plots of log k0 against pH were pH
independent for most compounds over this pH range. The value
of the rate constant k1, for the pH-independent reaction, was
taken as the average of all such experiments. For some
compounds, a hydroxide ion dependent reaction was also
observed in this pH range and the value of k1 was taken from
the best fit to the appropriate two-term rate law of eq 4. The
values of kOH will be reported elsewhere. In the case of the
four compounds 4, 12, 41, and 42, the rate constants k0 were
measured at a single pH value of ∼6.1, which was in the pH-
independent region in the case of all compounds for which the
pH dependence was more completely characterized. The value
of k0 in these cases was taken as equal to k1. All the values of
k1 are summarized in Table 1.
The secondary â-deuterium kinetic isotope effects for 18 and
4 were determined to be 1.02 ( 0.01 and 1.05 ( 0.02 per D,
respectively.
Rate constants, kH, for the acid-catalyzed decomposition of
7, 17, 21, and 24 were measured below pH 2. The rate constants
were determined as the slopes of plots of kobsd against acid
concentration. For 7, 17, 21, and 24, the values of kH were 6.9
× 10-4, 2.5 × 10-4, 6.4 × 10-5, and 6.7 × 10-6 M-1 s-1
,
respectively.
Products of the pH-independent decay of the several (acyl-
1
oxy)dialkylnitrosamines were characterized by H NMR after
decomposition of the compounds in D2O solutions containing
10% or 20% CD3CN (ionic strength 1 M, (NaClO4)). The
slightly higher concentration of acetonitrile in these studies
compared to that in the kinetic determinations was necessitated
k0 ) k1 + kOH[OH-]
(4)

R-(Acyloxy)dialkynitrosamines
J. Am. Chem. Soc., Vol. 121, No. 9, 1999 1829
Table 2. Product Yields from the Decomposition of
(Acyloxy)dialkylnitrosamines at 25 °C in Buffered D2O with 10 or
20% CD3CNa
On the basis of the observation of 6.6% incorporation in 20
min, an upper limit of 2 × 10-4 s-1 can be calculated for the
rate constant for exchange to equilibrium, 51.6% 18O in this
case, of the benzaldehyde oxygen with solvent water under these
conditions.
% CD3CN in
reacn medium
% yield
of RNOD
% yield
of RNND3
% yield
of RCCHO
compdb
+
7
8
24
38
43
10c
20d
10c
20f
20d
92
71e
92
99
89
nd
nd
nd
nd
10
99
98
100
91
Discussion
Mechanism of the pH Independent Reaction. (A) No
Anchimeric Assistance. An early study5 on the reactivity of
R-acetoxydialkylnitrosamines suggested a correlation between
increasing reactivity and the percent of Z-isomer. It was
concluded that the Z form decomposed with anchimeric as-
sistance, though the nature of this assistance was not stipulated.
It has been suggested6 that there is nitroso oxygen attack at the
acyl carbon to yield the ester products, formed via diazo ester
intermediates, that are observed in the decomposition reactions
of some R-(acyloxy)dialkylnitrosamines in nonpolar solvents.
Such a process is indicated in eq 5. It was concluded6 that such
99
1
a Yields determined by H NMR and based on the intergration of
the protons in the acetate anion leaving group. b See Table 1 for
compound structure. c Buffered with 0.025 M ClCH2CO2H buffer, 50%
anion. d Buffered with 0.019 M ClCH2CO2H buffer, 50% anion. e 28%
CH3OD, 27% CH3CHO, and 15% CH3OCH2CH2OD. f Buffered with
0.038 M cacodylic acid buffer, 50% anion.
a process was unlikely to be operative in aqueous solution due
to the absence of a decrease in reactivity upon placement of
carbonyl-stabilizing groups in the acyl function and the absence
of anticipated products for the reaction in the presence of azide
ion. The former point is buttressed in the comparison of
reactivities of 40 and 41. Both are >90% in the Z form, and
Figure 1. Plot of kobsd/k0 against acetate ion concentration for the
decomposition of 7 (squares), 33 (solid circles), 1 (solid triangles), and
3 (diamonds) or against trifluoroacetate ion concentration for the
decomposition of 1 (open circle) and 7 (inverted triangle) at 25 °C,
ionic strength 1 M (NaClO4).
by solubility considerations. Yields of the products were based
on the C-H proton signal from the acetate or chloroacetate
leaving groups. In cases where the yield of a product was low,
identity was confirmed by spiking the NMR tube after quan-
titation with the authentic compound and observing the increase
in signal in the appropriate frequency. The product yields are
summarized in Table 2.
the latter is 4-fold more reactivesopposite what might be
expected if carbonyl group attack is involved as in the
mechanism of eq 5.6
The effects of changing acetate ion or acetate buffer
concentrations on the values of kobsd were investigated in the
case of four R-acetoxydialkylnitrosamines, and the results are
depicted in Figure 1. Also included are data for control
experiments in which the effect of changing salt from 0.1 M
Na+CF3CO2-/0.9 M NaClO4 to 0.8 M Na+CF3CO2-/0.2 M
NaClO4 upon kobsd was determined for three R-acetoxydial-
kylnitrosamines.
An alternative possibility, involving nitroso oxygen attack
on the R-carbon to yield initially a four-membered-ring inter-
mediate that ultimately gives rise to products, as in eq 6, was
Activation parameters were determined for the pH-indepen-
dent solvolysis (k1) for two compounds, 3 and 33, by studying
the effect of temperature on kobsd over the temperature range
from 20 to 50 °C. Linear Eyring plots of ln(kobsdh/kBT) against
1/T(K) yielded values of ∆Hq )14 ( 1 and 8 ( 0.2 kcal/mol
and ∆Sq ) -37 ( 3 and -56 ( 3 eu for 33 and 3, respectively.
The results of experiments to measure the extent of 18O
incorporation into product benzaldehyde from the decomposition
of 42 and 43 (natural abundance in 18O) in ∼50% 18O-H2O are
summarized in Table 3. The compounds were decomposed for
6-8 half-lives, 3 and 20 min, for 42 and 43, respectively.
Control experiments with benzaldehyde (natural abundance in
18O) incubated under the same conditions indicated the incor-
poration of 18O from solvent into benzaldehyde to the extent of
0.8% and 6.6% after 3 and 20 min incubations, respectively.
not addressed. Transition states and/or intermediates that are
similar in structure to the four-membered ring in eq 6 are
precedented by the nitro- and nitrosoamide decomposition13,14
and by the nitrosative cleavage of imines to yield aldehydes.15
The observation in this study that the isotopic content of the
oxygen in the product aldehyde is essentially identical to that
of the solvent water (∼50% 18O)sand is very different from
(13) White, E. H.; Woodcock, D. J. In The Chemistry of the Amino
Group; Patai, S., Ed.; Wiley-Interscience, New York, 1968; p 440.
(14) White, E. H.; Field, K. W.; Hendrickson, W. H.; Dzadzic, P.;
Roswell, D. F.; Paik, S.; Mullen, P. W. J. Am. Chem. Soc. 1992, 114, 8023
and references within.
(15) Doyle, M. P.; Wierenga, W.; Zaleta, M. A. J. Org. Chem. 1972,
37, 1597. Doyle, M. P.; Zaleta, M. A.; DeBoer, J. E.; Wierenga, W. J.
Org. Chem. 1973, 38, 1963.

1830 J. Am. Chem. Soc., Vol. 121, No. 9, 1999
Cai and Fishbein
Table 3. Incorporation of 18O from 18O-Water in the Decay of
16O-Containing R-(Acyloxy)dialkylnitrosamines at 25 °C, (Ionic
Strength 1 M (NaClO4))
corresponding point for log k1. It can be seen that the points
for the values of log k1 for compounds which are substantially
or predominantly in the Z form do not noticeably deviate
positively from the correlation, and they exhibit similar behavior
for the correlation including only those compounds with 10%
or less of the Z form. There is a possibility that the extent of Z
conformer which we determined by 1H NMR in CDCl3 is
different in the much more polar solvolysis medium (H2O, ionic
strength 1 M), thus confounding the analysis above. However,
this seems unlikely because for two compounds, 13 and 18 (see
Table 1), the E/Z ratios determined in 60% D2O/40% CD3CN
were not substantially different from those determined in CDCl3
(see Table 1). The absence of a rate enhancement due to
increased Z conformation is a direct contradiction of the claim
of anchimeric assistance, which, by definition, requires enhanced
reactivity.
%18O-PhCHO
after reaction
compda
%Z form b
%
18O in solventd
42
43
4
90
50.6 ((0.3)
48.4 (0.5)
51.6 ((1.0)
46.4 ((1.2)
a Structure key in Table 1. Contained natural-abundance 18O in all
positions. b Determined in CDCl3. c Reaction time was 3 min for
compound 42 and 20 min for compound 43. d Determined by equilibra-
tion of the product benzaldehyde with the reaction solvent for an
additional 24 h.
(B) Nitrosiminium Ion Formation. 1. Hyperconjugative
Interactions of RC and Electronic Effects at RN. Figure 2
illustrates the dependence of log k1, the rate constant for the
pH-independent reaction, upon electron-withdrawing ability of
substituents attached to the amino nitrogen (RN). Inspection
shows that there are separate correlations, for each distinct RC
group. The order of reactivity, for a given RN group, is RC )
methyl > ethyl > isopropyl > tert-butyl > hydrogen. Among
the first four groups, the rate constant is enhanced by increasing
the number of hydrogen atoms â to the carbon atom undergoing
bond cleavage. This is due the commensurate increase in the
number of hyperconjugative interactions by C-H groups that
can stabilize the developing nitrosiminium ion in the transition
state.
There is evidence against the possibility that steric bulk,
through inhibition of transition state solvation or some other
means, is responsible for the observed order of reactivity:
First, steric bulk of the substituent attached to nitrogen (RN)
does not obscure the good linear correlation with σ* in Figure
2 even though the correlation includes the bulky tert-butyl,
isopropyl, and 1-phenylethyl groups. In a later section, it will
be shown that electronic substituent effects at RC are quanti-
tatively similar to those observed for RN substituents. This
suggests a similar change in charge, at the nitrogen and carbon
atoms of the developing iminium ion, in going from the ground
state to the transition state. Given the similarity, it is difficult
to see why steric hindrances of solvation should be so
dramatically different at the two sites. In contrast, a difference
in the involvement of hyperconjugative stabilization is indeed
expected (vide infra).
Second, steric hindrance to solvation of the leaving carboxylic
acid anion does not diminish reactivity beyond that expected
simply from a consideration of leaving-group basicity effects.
Both the (N-nitrosobutylamino)methyl benzoate and pivaloate
are more reactive than the acetate, despite the lesser bulk of
the acetate.6 In addition, while the (N-nitroso-tert-butylamino)-
ethyl acetate ester is more reactive, by less than a factor of 3,
than the bulkier cylcohexanecarboxylic acid ester, most of this
difference can be attributed to the greater basicity of the latter
carboxylate leaving group. The value of log k1 for this
compound, adjusted for the pKa difference, would only deviate
negatively from the appropriate correlation for acetates (solid
circles in Figure 2), by less than 0.2 log units.18
Figure 2. Plot of log k1 against σ* for substituents RN (structure C,
eq 1) in the solvolysis of R-acetoxydialkylnitrosamines at 25 °C,
ionic strength 1 M (NaClO4). RC: -CH3 (circles); -CH2CH3 (squares);
-CH(CH3)2 (diamonds); -C(CH3)3 (triangles); -H (inverted triangless
data from ref 6). Inset box applies only to circles: top row, percent of
Z isomer measured by 1H NMR; bottom row, number of R-hydrogens
in RN.
that in the starting R-(acyloxy)dialkylnitrosamine that contains
only natural-abundance 18Osrules out mechanisms of anchi-
meric assistance such as those in eqs 5 and 6. Table 3 indicates
the isotopic contents of benzaldehyde formed in the decay of
nitrosamines 42 and 43 (that contained only “natural-abundance”
levels of 18O) as well as the 18O content of the water in which
the reaction was carried out. The levels of 18O in the product
and solvent are the same within experimental error. If 10% of
the reaction occurred by anchimeric assistance involving
incorporation of oxygen from the starting material into the
product aldehyde, the measured 18O content of the benzaldehyde
from 42 and 43 would be expected to be 46.4% and 41.6%,
respectively. These values are different, outside 3 standard
deviations from the observed values. Some 16O-isotope retention
in the product could be obscured due to oxygen exchange of
the benzaldehyde with water, but allowance for this possibility
still confirms that less than 10% of the product from 43, and
less still from 42, can be due to anchimeric assistance pathways
such as those in eqs 5 and 6.16
The observation, summarized in Figure 2, that R-acetoxydi-
alkylnitrosamine reactivity toward solvolysis is a linear function
of electron-donating ability of the substituent RN, when RC is
held constant, and is independent of E/Z ratio is additional direct
evidence against anchimeric assistance. In Figure 2, the solid
circles are points for the log k1 values of (N-nitrosoalkylamino)-
ethyl acetates (RC ) methyl) plotted against σ*. There is a good
linear fit to the equation log k ) -3.173 - 1.603σ* (R2 )
0.995).17 The percent of Z isomer for each of these compounds
is listed in the top row of the inset box of Figure 2, above the
Third, steric hindrance to solvation of the anionic leaving
group as a cause for the reactivity order is an untenable ex-
(17) A value of F* ) -1.50 was published on the basis of three
substituents in ref 6.
(16) A scheme and details on the method of calculation are included in
the Supporting Information.
(18) This is based on the reported difference in pKa of 0.2 unit and the
average value for âlg of -1.09.

R-(Acyloxy)dialkynitrosamines
J. Am. Chem. Soc., Vol. 121, No. 9, 1999 1831
planation since the same order of reactivity (kHMe > kHEt > kH
The observed order of reactivity, Me > Et > i-Pr > t-Bu, is
the so-called “Baker-Nathan order”,27 which is widely observed
in organic reactivity.27,28 It was originally ascribed to a
hyperconjugative interaction,27 but this explanation has been
rejected.29-31 The notion that steric inhibition of solvation29
could account for the original observation was given credence
by the subsequent observation that the order was reversed in
the gas-phase equilibrium for protonation of alkylbenzenes.30
This observation indicates that the dominant equilibrium-
stabilizing factor in the gas phase is the well-known polariz-
ability effect. The observation led to the conclusion that the
Baker-Nathan order in solution was indeed due to hindrance
of solvation.30 This conclusion has since been generalized.31
i-Pr
> kHt-Bu) is observed in the acid-catalyzed decomposition
reaction that has been shown to involve loss of neutral acetic
acid with formation of the nitrosiminium ion. This mechanism,
illustrated in eq 9, is indicated by the acid-catalyzed acyloxy
This conclusion is not logically required by the observation
but rather is dependent on unverified assumptions. There is no
reason to expect that the relative strengths of various interactions
should be constant in going from the gas phase to polar liquids.
Group polarizability is strongly attenuated in polar liquid
media.32 Other types of stabilizing interactions, such as hyper-
conjugation, may have different dependencies on the influences
of solvent. Beyond this, it is currently unclear whether there
can be differential expressions of hyperconjugative and polar-
izability interactions in transition states compared to ground
states.
The experimental facts summarized above and elsewhere25,26
require that the transitions state stabilization by an adjacent C-H
bond is greater than that by an adjacent C-C bond. That this is
possible has been disputed.33 On theoretical grounds, with
optimal structure, the C-C bond interaction with an adjacent
carbocation is stronger.33,34 Again, however, the possibility of
differential expressions of â-H versus â-C hyperconjugative
interactions in transition states, due to intrinsic constraints on
heavy-atom motion or steric constraints, has not been addressed.
exchange and ether and hydroperoxide formation in alcoholic
solvents.19-21
Fourth, two observations are inconsistent with the notion that
steric bulk decreases reactivity by inhibition of solvation.
Compound 25 is 3-fold more reactive than 16 despite the greater
steric bulk at RC in the former. The Taft steric parameter of the
RC substituent is -1.74 for 25 and only -0.07 for 16. Similarly,
despite the greater steric bulk of the CH2C(CH3)3 (Es ) 1.74)
compared to the tert-butyl substituent (Es ) -1.54),22 the rate
constant for nitrosiminium ion formation is 100-fold greater with
the former substituent than with the latter in the case of the
N-nitrosoethylamino compounds (25 and 23).
The normal secondary â-deuterium kinetic isotope effects are
consistent with hyperconjugative stabilization of the developing
nitrosiminium ion character in the transition state. The values
of kH/kD )1.02 ( 0.01 and 1.05 ( 0.02 per D for the solvolysis
of 18 and 4, respectively, are in the direction expected for
loosening of the â-C-L bonds due to overlap in the transition
state with the adjacent electron-deficient center.23 The magnitude
of these effects is small compared to the largest effects reported
of kH/kD ) 1.14 per D. Small effects are expected in the case
of a developing cation that is strongly stabilized by electron
donation from the adjacent amino group.23 Secondary â-deu-
terium kinetic isotope effects of magnitude similar to those
observed here have been observed in the acid-catalyzed cleavage
of acetals in which electron donation by an adjacent oxygen
atom stabilizes the transition state for cleavage of the carbon
oxygen bond as in eq 10.24 The importance of C-H hypercon-
In the case of nitrosiminium ion formation from R-acetoxy-
dialkylnitrosamines, the extent of enhancement of the rate
constant by hyperconjugative stabilization is not a simple
function of the number of â-hydrogens, in contrast to what has
been reported in the case of acid-catalyzed acetal and ketal
hydrolysis.31 Thus, in Figure 2, at σ* ) 0, the increase in log
k1 as â-methyl groups are replaced by â-hydrogen atoms, going
from tert-butyl to methyl, is 0.96, 0.64, and 0.41 unit for each
successive replacement. These differences cannot be accounted
for by differences in polar effects upon each successive
replacement since each hydrogen-for-methyl replacement in-
(23) Westaway, K. C. In Isotopes in Organic Chemistry, Vol. 7;
Secondary and SolVent Isotope Effects; Buncel, E.; Lee, C. C., Eds.;
Elsevier: New York, 1987.
(24) Kresge, A. J.; Weeks, D. P. J. Am. Chem. Soc. 1984, 106, 7140.
Shiner, V. J.; Cross, S. J. Am. Chem. Soc. 1957, 79, 3599. Shiner, V. J.
Tetrahedron 1959 5, 243.
(25) Kreevoy, M. M.; Taft, R. W. J. Am. Chem. Soc. 1955, 77, 5590.
(26) Kreevoy, M. M. Tetrahedron 1959 5, 233.
(27) Baker, J. W.; Nathan, W. S. J. Chem. Soc. 1935, 1844.
(28) Berliner, E. Tetrahedron 1959 5, 202.
(29) Schubert, W. M.; Sweeney, W. A. J. Org. Chem. 1956, 21, 119.
(30) Herhe, W. J.; McIver, R. T.; Pople, J. A.; Schleyer, P. v. R. J. Am.
Chem. Soc. 1974, 96, 7162.
(31) March, J. AdVanced Organic Chemistry, 4th ed.; J. Wiley and
jugative stabilization of the transition state for this reaction has
been emphasized by the work of Kreevoy and Taft.25,26
Sons: New York, 1992; p 69.
(32) Fujio, M.; McIver, R. T.; Taft, R. W. J. Am. Chem. Soc. 1981, 103,
4017 and references within.
(19) Mochizuki, M.; Anjo, T.; Okada, M. Chem. Pharm. Bull. 1978, 26,
3905.
(20) Baldwin, J. E.; Scott, A.; Branz, S. E.; Tannenbaum, S. R.; Green,
L. J. Org. Chem. 1978, 43, 2427.
(21) Mochizuki, M.; Anjo, T.; Wakabayashi, Y. Tetrahedron Lett. 1980
21, 1761.
(33) Exner, O.; Bohm, S. J. Chem. Soc., Perkin Trans. 2 1997, 1235.
Glyde, E.; Taylor, R. J. Chem. Soc., Perkin Trans. 2 1977, 678. But see:
Cooney, B. T.; Happer, D. A. R. Aust. J. Chem. 1987, 40, 1537.
(34) Radom, L.; Pople, J. A.; Schleyer, P. v. R. J. Am. Chem. Soc. 1972,
94, 5935. Hoffmann, R.; Radom, L.; Pople, J. A.; Schleyer, P. v. R.; Hehre,
W. J.; Salem, L. J. Am. Chem. Soc. 1972, 94, 6221. Wierschke, S. G.;
Chandrasekhar, J.; Jorgensen, W. L. J. Am. Chem. Soc. 1985, 107, 1496.
Ibrahim, M. R.; Jorgensen, W. L. J. Am. Chem. Soc. 1989, 111, 819.
(22) Exner, O. In Correlation Analysis in Chemistry. Recent AdVances;
Chapman, N. B., Shorter, J., Eds.; Plenum: New York, 1978.

1832 J. Am. Chem. Soc., Vol. 121, No. 9, 1999
Cai and Fishbein
creases the value of σ* of the RC group by essentially the same
amount, ∆σ* ∼ +0.1. The differences may be the result of a
saturation effect similar to that observed in classical resonance
stabilization.35
The changes in slope in Figure 2 are consistent with either
(1) a change in transition state structure or (2) a change in charge
distribution in the transition state. The slopes of the lines, which
indicate positive charge buildup in the transition state, increase
in the order RC ) Me < Et < i-Pr < t-Bu with values of -1.60
(R2 ) 0.995), -2.05 (R2 ) 0.995), -2.23 (R2 ) 0.993), and
-2.88, respectively. (1) The change in transition state structure
would be consistent with a Hammond type effect35 in which
there is less carbon-oxygen bond cleavage as the reactivity of
the starting material is enhanced by increasing hyperconjugative
stabilization of the transition state in the order t-Bu < i-Pr <
Et < Me. This is not ruled out, but the existing evidence does
not suggest it. There is no indication of a significant change in
the extent of charge buildup on the leaving acetate on the basis
of the âlg values, the slopes of plots of log k1 against conjugate
acid pKa of the leaving acetate. The nearly constant slopes, )
-1.05,3 -1.13,3 -1.10, and 1.09, of plots containing points for
two or three different leaving groups for compounds in which
the parent acetate is 6, 17, 21, and 24, respectively, do not
suggest a changing transition state structure. The recently
reported secondary R-deuterium kinetic isotope effects on the
decay of methyl and propyl acetates average 1.15 ( 0.01 and
1.17 ( 0.02, respectively, for two compounds of each class,8
again suggesting no difference. (2) The differences in slope can
also be accounted for by a change in charge distribution in the
transition state due to a change in the extent of hyperconjugative
stabilization. Equation 11 illustrates three resonance forms that
Figure 3. Plot of log k1 against σ* for substituents RC (structure C,
eq 1) in the solvolysis of R-acetoxydialkylnitrosamines at 25 °C, ionic
strength 1 M (NaClO4). For open and closed circles, RN ) -CH2CH3.
For open and closed diamonds, RN ) -CH3. For solid symbols, RC )
-CH2X, X * H. For open symbols, RC ) -C(CH3)3, -CH(CH3)2,
and -CH3.
real differences. The negative values of F* are consistent with
the formation of a transition state that is electron deficient
relative to the ground state. The magnitudes indicate that the
change in effective charge in going from the ground to transition
state is similar to, in some cases slightly larger than, that
experienced by substituents attached to the amino nitrogen for
which the value of F* varies from -1.6 to -2.9 (above). The
relative unimportance of steric bulk (and possible consequent
hindrance of solvation) in determining the magnitude of rate
constants is emphasized by the fact that, for the solid circles in
Figure 3, proceeding from left to right, the value of the Taft
steric constant changes as follows: -1.74, -0.07, -0.38, -0.77,
and -0.19. This range of values is comparable with that
traversed in passing through the series t-Bu, i-Pr, Et, Me but
clearly does not obscure a reasonable correlation with σ* alone.
Rate constants for the Me, i-Pr, and t-Bu groups are plotted as
open symbols and indicate substantially larger deviations than
those observed for the solid points alone, indicative of the
aforementioned hyperconjugative interaction.
can be considered to describe the structure of the nitrosiminium
ion. Structure h represents that which describes hyperconjugative
stabilization involving â-hydrogen atoms.
(C) Change in Mechanism. The introduction of sufficiently
strong electron-withdrawing groups at either RN or RC positions
results in a change in mechanism, as previously predicted.6 The
expectation was based on the logic that such substituents would
destabilize the transition state for nitrosiminium ion formation
and simultaneously stabilize the transition state for nucleophilic
attack at the carbonyl group of the ester. As noted (Results),
the pH-independent reactions of 3 (RN ) dimethoxyethyl), 2
(RN ) methoxyethyl), and 33 (RC ) dichloromethyl) occur over
relatively narrow regions compared to those of most of the
R-acetoxydialkylnitrosamines. The rate constants for 3 and 2
lie 1.3 and 2.1 orders of magnitude above the line extrapolated
in Figure 1 for other methyl acetates. The rate constant for 33
lies 3 log units above the line extrapolated for (N-nitrosoeth-
ylamino)alkyl acetates in Figure 3.36 The positive deviations
from the correlations for compounds that decompose through
the formation of nitrosiminium ions are consistent with a change
in mechanism.
To the extent that structure h becomes more important in
the order t-Bu < i-Pr < Et < Me, this results in diminution of
the effective charge on nitrogensstructure f becomes a smaller
contributor to the overall energy.
Finally, the unimportance of hyperconjugative interactions
for substituents attached to the amino nitrogen, RN, is empha-
sized in Figure 2 for the ethyl acetates (filled circles). Given in
the second row of the inset frame is the number of R-hydrogens
(â to the nitrogen) for each RN substituent, and it can be seen
that the correlation is unperturbed as the number of such
hydrogens varies from 0 to 3. This is expected because in no
important resonance form of the nitrosiminium ion, eq 11, is
the valence shell of the amino nitrogen less than completely
occupied.
2. Substituent Effects at RC. Because of the importance of
hyperconjugation by hydrogen atoms on the RC substituent,
estimates of the polar effects of RC substituents must be made
by comparing substituents with identical numbers of R-hydrogen
atoms. Figure 3 (solid points) contains such correlations. The
slopes of plots of log k1 against σ* are -2.60 (R2 ) 0.974) and
-2.82 (R2 ) 0.970) for N-ethylamino and N-methylamino
compounds, respectively. Given the lower confidence in the
correlations, it is uncertain whether these values of F* represent
The onset of a new mechanism is also indicated by the change
in the effect on kobsd of added acetate anion. In the case of
reactions involving formation of N-nitrosiminium ions, the value
(36) As the above discussion indicates, this does not take account of the
fact that the CHCl2 substituent has a single hydrogen atom that can stabilize
the developing carbocation by hyperconjugation while those in Figure 3
have two; however, this means that the rate constant for 33 actually deviates
by more than +3 log units from the appropriate correlation.
(35) Lowry, T. H.; Richardson, K. S. Mechanism and Theory in Organic
Chemistry, 3rd ed.; Harper and Row: New York, 1987.

R-(Acyloxy)dialkynitrosamines
J. Am. Chem. Soc., Vol. 121, No. 9, 1999 1833
of kobsd decreases with increasing acetate ion concentration due
to diversion of the nitrosiminium ion intermediate to starting
materialssthe common ion inhibition.9 Common ion inhibition
is observed in Figure 1 for 1 and 7 (squares and triangles,
respectively). The opposite effect of increasing acetate ion is
observed in the case of 3 and 33 (diamonds and closed circles,
respectively). Control experiments suggest that this effect is not
likely due to a medium effect of replacing the NaClO4 coun-
terions with sodium acetate, which cancels a weak common ion
effect. Replacement of 0.8 M NaClO4 by 0.8 M sodium
trifluoroacetate (open circle (1) and inverted triangle (3) in
Figure 1) slightly diminishes the value of kobsdsopposite the
larger effect of sodium acetate upon decay of 1 and 7. The
enhancement of kobsd with increasing sodium acetate concentra-
tion is consistent with rate-limiting attack of water on the
carbonyl group that is assisted by general-base catalysis, as has
been observed in related systems.37
rate constants for formation of N-nitrosiminium ions are
generally enhanced for electron-donating substituents in both
RN and RC positions. In the case of simple alkyl groups in the
RC position, the number of hydrogens â to the cationic carbon
of the developing nitrosiminium ion appears to dominate
reactivity effects, increasing reactivity with increasing numbers
of â-hydrogens due to hyperconjugative stabilization of the
transition state. Attachment of substituents of sufficient electron-
withdrawing power at RN and RC results in a predicted change
in mechanism to what appears to be an acyl group attack
mechanism.
Acknowledgment. This work was supported by Grants RO1
CA52881 and KO4 CA62124 from the National Cancer Institute
(NIH). The authors wish to thank Professors M. Kreevoy
(Minnesota) and W. Jorgensen (Yale) for helpful discussions.
The large negative values of entropies of activation observed
for 3 and 33 are in contrast with what has been reported for
reactions involving formation of nitrosiminium ions, but are
consistent with what is known for ester hydrolysis involving
water attack on the carbonyl group. Reported values of ∆Sq are
near zero for the mechanism of eq 3. In contrast, ∆Sq ) -37
and -56 eu for 33 and 3, respectively. These values are in the
range reported for the mechanism of water attack on the carbonyl
group.38
Supporting Information Available: Text giving spectro-
scopic and other analytical data for compounds not previously
reported and a scheme and discussion of the analysis of
calculated limits for 18O incorporation experiments. This
information is available free of charge via the Internet at
http://pubs.acs.org.
JA9833218
(37) Jencks, W. P.; Carrioulo, J. J. Am. Chem. Soc. 1961, 83, 1743.
(38) Reference 37 and: Johnson, S. L. AdV. Phys. Org. Chem. 1967, 5,
236.
Summary. Experimental evidence has been adduced that
rules out possible mechanisms of anchimeric assistance. The