Mechanisms of decomposition of α-hydroxydialkylnitrosamines in aqueous solution

Article abstract of DOI:10.1021/tx000084p

A study of the decomposition of α-hydroxydialkylnitrosamines in aqueous 9% acetonitrile, with an ionic strength of 1 M (NaClO₄), at 25 °C is reported. Plots of the logarithm of the buffer-independent rate constant, k(o), against pH are concave up and indicate a three-term rate law for the solvent reaction, including acid (k(H⁺))-, base (k(OH))-, and pH-independent (k(HOH)) terms. Secondary α-deuterium isotope effects for compound 1a, (N-nitrosomethylamino)-phenylmethanol, are as follows: k(α)(H)/k(α)(D) = 1.12 ± 0.03 and 1.19 ± 0.02 for k(H⁺) and k(OH), respectively. General acid (k(HA)) and general base (k(A^-)) catalysis by more acidic carboxylic acid buffers is also observed. Structure reactivity and other parameters obtained in this study, and their changes with substrate and catalyst structure, permit the assignment of mechanisms for the k(H⁺), k(OH), k(HA), and k(A^-) processes.

Full text of DOI:10.1021/tx000084p

Chem. Res. Toxicol. 2000, 13, 983-992
983
Mech a n ism s of Decom p osition of
r-Hyd r oxyd ia lk yln itr osa m in es in Aqu eou s Solu tion
Milan Mesic', J ari Peuralahti, Patrick Blans, and J ames C. Fishbein*
Department of Chemistry and Biochemistry, University of Maryland Baltimore County,
1000 Hilltop Circle, Baltimore, Maryland 21250
Received April 11, 2000
A study of the decomposition of R-hydroxydialkylnitrosamines in aqueous 9% acetonitrile,
with an ionic strength of 1 M (NaClO4), at 25 °C is reported. Plots of the logarithm of the
buffer-independent rate constant, ko, against pH are concave up and indicate a three-term
+
rate law for the solvent reaction, including acid (kH )-, base (kOH)-, and pH-independent (kHOH
)
terms. Secondary R-deuterium isotope effects for compound 1a , (N-nitrosomethylamino)-
phenylmethanol, are as follows: kRH/kR ) 1.12 ( 0.03 and 1.19 ( 0.02 for kH and kOH
,
+
D
-
respectively. General acid (kHA) and general base (kA ) catalysis by more acidic carboxylic acid
buffers is also observed. Structure reactivity and other parameters obtained in this study,
and their changes with substrate and catalyst structure, permit the assignment of mechanisms
+
-
for the kH , kOH, kHA, and kA processes.
decomposition of R-hydroxydialkylnitrosamines.
In tr od u ction
R-Hydroxydialkylnitrosamines are reactive intermedi-
ates that are believed to be formed by the action of P450
enzymes on the parent dialkylnitrosamines (1-3). They
decompose to form diazonium ions and in some cases
carbocations that can alkylate DNA, the reaction that is
believed to be responsible for the mutagenicity and
carcinogenicity of dialkylnitrosamines.
These compounds are generated by reduction of R-hy-
droperoxy precursors in acetonitrile and are studied in
largely aqueous media by means of stopped-flow tech-
niques. Previous spectroscopic, kinetic, and product
analyses have demonstrated the utility of this approach
to examining the chemistry of these reactive intermedi-
ates (8-10). This study has permitted the first descrip-
tion of the hydronium ion- and hydroxide ion-catalyzed
mechanisms, the latter being dominant in the physiologi-
cal pH region for the substrates that were studied.
Further, the mechanisms of heretofore unknown general
acid- and general base-catalyzed reactions are also
deduced.
At one time, it appeared that such species might be too
unstable to be characterized, but through elegant work
nearly 20 years ago, a number of R-hydroxymethyl
compounds 1 (R′ is H) were synthesized and their
stabilities were characterized in aqueous media (4, 5).
Some variation in R (R can be unbranched and branched
alkyl groups) indicated little change in reactivity as a
function of structure. In the intervening years, two
reports of highly stable R-hydroxynitrosamines have
appeared (6, 7). Work from this group has shown that
elaboration of the R′ group in eq 1 (R′ is CH3 or Ph, for
example) engenders significant instability (8, 9). Simi-
larly, carbocyclic R-hydroxynitrosamines appear to be
significantly more unstable than the R-hydroxymethyl
compounds originally reported (R′ is H, above) (10).
With a single exception (8), there has been little focus
on the detailed mechanisms by which R-hydroxydialkyl-
nitrosamines decompose and the way in which these
mechanisms manifest effects of structure on reactivity.
Such mechanisms and structure-reactivity relations
impact on the lifetimes and thus diffusibility of alkylating
equivalents and, with some caveats, the possibility of
targeting of alkylation events. This work summarizes
experiments with the compounds below that were de-
signed to give a detailed picture of the mechanisms of
Exp er im en ta l Section
Wa r n in g: All N-nitroso compounds must be regarded as
suspect carcinogens and handled appropriately using double
pairs of frequently changed gloves as well as standard laboratory
garments. All materials suspected of contact with such com-
pounds were treated with 50% aqueous sulfuric acid containing
the commercially available oxidant NoChromix.
Ma ter ia ls. The R-hydroperoxynitrosamine precursors to
R-hydroxynitrosamines were prepared by minor modification of
the original procedure (4, 5, 11, 12). This involved group
exchange starting with the R-chloroacetate esters in a mixture
of aqueous hydrogen peroxide and acetonitrile, followed by
extraction and purification by preparative thin-layer chomatog-
raphy. Synthesis of the esters was previously published (13).
The synthesis of the R-deuterated substrate, [R-2H]-1a , for the
measurement of secondary R-deuterium kinetic isotope effects
was carried out as described for the protio compound, starting
with deuterated benzaldehyde that was generated by the Neff
10.1021/tx000084p CCC: $19.00 © 2000 American Chemical Society
Published on Web 09/23/2000

984 Chem. Res. Toxicol., Vol. 13, No. 10, 2000
Mesic' et al.
1
Ta ble 1. Su m m a r y of P r od u cts for Rea ction s in D2O
reaction, as described previously (14). H NMR indicated <3%
Solu tion s (15 vol % CD3CN) Qu a n tified by 1H NMRa
product (yieldb)
contamination by protium in the benzaldehyde prior to synthesis
of the R-hydroperoxy compound and after decomposition in D2O
as described below.
reaction in 0.1 M CD3COOD
A typical synthesis for the R-hydroperoxynitrosamine precur-
sors of R-hydroxynitrosamines is given for compound 2. Spectral
data for the other R-hydroperoxynitrosamines follow. The 1H
NMR spectra, samples of which are included in the Supporting
Information, indicate that the purity of the R-hydroperoxyni-
trosamines is on the order of g95%, with most significant
common contaminants being solvents of purification or the
substituted benzaldehydes of decomposition, the latter increas-
ing with storage time at -20 °C.
r-H yd r op e r oxy(3-ch lor op h e n ylm e t h yl)-N -m e t h yln i-
tr osa m in e (2). N-Nitroso-N-methyl (3-chlorophenyl)methyl-
chloroacetate (5.1 mmol) was dissolved in 14 mL of acetonitrile.
A 25 mL volume of 30% H2O2 was added dropwise to the
mixture. Some precipitation appeared, and 10 mL of acetonitrile
was added to the mixture which was then stirred overnight. The
resulting mixture was clear, and 15 mL of water and 15 mL of
CH2Cl2 were added. The product was extracted into the CH2-
Cl2 which was washed with water. The CH2Cl2 solution was
dried with Na2SO4. The product was purified by preparative
compd reaction in 0.1 M DClO4
and 50% anion
2
CH3-xDxOD (83 ( 9%)
3-ClPhCHO (99 ( 5%)
CH3-xDxOD (81 ( 9%)
4-ClPhCHO (99 ( 9%)
NCCH2-xDxCH2-yDy-
OD (83 ( 5%)
CH3-xDxOD (79 ( 5%)c
3-ClPhCHO (105 ( 6%)c
CH3-xDxOD (75 ( 3%)
4-ClPhCHO (106 ( 5%)
NCCH2-xDxCH2-yDy-
OD (78 ( 4%)
3
1d
PhCHO (97 ( 6%)
PhCHO (95 ( 5%)
a
b
Reactions at room temperature. Relative to the internal
standard tert-butyl alcohol. Unless otherwise noted, values are
based on single experiments, and means and standard deviations
are based on quadruplicate analysis of the single experiment. For
the product alcohol from 1d , integration of the hydroxymethylene
proton was used. c Experiment carried out in duplicate.
reactions in D2O buffered with 0.1 M d4-acetic acid
buffers (DA/A ) 1) or 0.1 M DClO4 solutions. A solution
of the R-hydroperoxy precursor in CD3CN was reacted
with 1.1 molar equiv of tributylphosphine, and a volume
of this was reacted with a 5.6-fold excess of aqueous
media. The final concentration of the nitrosamine was
typically 0.01 M, and the yields were determined by
comparison with the internal standard tert-butyl alcohol
(∼0.001 M). The yields are summarized in Table 1.
Inspection of the spectra indicated no signal at 2.41 ppm
(for CH3ND3+). Based on the maximum height of the
1
TLC using 20% ethyl acetate in hexane as the eluent. H NMR
(CDCl3): δ 9.96 (1H, s), 7.41-7.21 (5H, m), 2.85 (3H, s). 13C
NMR (CDCl3): δ 134.6, 134.1, 129.7, 129.6, 129.3, 125.8, 123.7,
95.1, 26.5. In certain cases, there was evidence of both E and Z
rotamers, but due to overlaps and the preponderance of the E
form, it was not always possible to ditinguish signals for all
hydrogens in the Z forms. 1H NMR for (E)-3 (CDCl3, E/Z ∼ 40):
δ 9.47 (1H, s), 7.43-7.21 (5H, m), 2.87 (3H, s). 1H NMR for (Z)-3
(CDCl3, E/Z ∼ 40): δ 3.72 (3H, s). 13C NMR for 3 (CDCl3): δ
136.1, 131.5, 129.5, 127.8, 96.0, 27.1. 4 was crystallized from
pentane. 1H NMR (CDCl3): δ 10.17 (1H, s), 7.71 (2H, d), 7.49
(2H, d), 7.421 (1H, s), 2.87 (3H, s). 13C NMR (CDCl3): δ 136.1,
+
noise, the upper limit for CH3ND3 was 5% of that for
methanol.
An experiment with 1a (concentration of ∼1 × 10-4 M)
was carried out in 0.1 M HClO4 and 20 vol % CH3OH,
and the products were analyzed by HPLC. Based on the
retention time and integration factor, determined by a
three-point standard curve, of the authentic R-methoxy
compound (13), an upper limit for the yield of the
R-methoxy compound, which was not detected, was less
than 3%.
1
131.5, 129.5, 127.8, 96.0, 27.1. H NMR for 1b (CDCl3): δ 9.63
(1H, s), 8.06 (1H, d), 7.38 (5H, m), 3.72 (2H, t), 3.56 (2H, t),
3.35 (3H, s). 1H NMR for (E)-1c (CDCl3, E/Z ∼ 20): δ 10.74 (1H,
s), 7.42-7.32 (6H, m), 4.90 (1H, dd), 3.92 (1H, dd), 3.40 (6H, s),
3.00 (1H, dd). 1H NMR for (Z)-1c (CDCl3, E/Z ∼ 20): δ 3.08
1
(6H, s). H NMR for (E)-1d (CDCl3, E/Z ∼ 14.5): δ 9.15 (1H, s),
7.65-7.25 (6H, m), 3.86 (1H, m), 3.35 (1H, m), 2.70 (1H, m). 1H
NMR for (Z)-1d (CDCl3, E/Z ∼ 14.5): δ 4.4 (1H, m), 4.2 (1H,
m), 2.9 (1H, m). 13C NMR (CDCl3): δ 131.8, 129.4, 128.6, 125.4,
95.6, 36.2, 14.6.
Kin etics. (1) Solven t Rea ction s. For all compounds
studied at all pH values, the kinetics of aldehyde product
formation, monitored at 245, 255, or 265 nm, exhibited
good first-order behavior for more than four half-lives of
reaction. Plots of kobsd against buffer concentration,
containing at least three points spanning the concentra-
tion range of 0.05-0.3 M, were generally linear, maximal
increases being less than 100% of the intercept value
extrapolated to zero buffer concentration. Typical plots
are shown in Figure 1. Included in Figure 1 are results
from control experiments in which the addends trifluo-
roacetate ion and acetamide were varied over a similar
concentration range. The intercept values of the buffer
dilution plots were taken as buffer-independent rate
constants, ko. The values of log ko for all compounds,
except 1a (8), that were studied are plotted against pH
in Figure 2.
Meth od s. (1) P r od u cts. Product analysis was carried out
by 1H NMR. The hydroperoxide in CD3CN was reduced with
1.1 equiv of tributylphosphine. Immediately after reduction and
mixing, the CD3CN solution was rapidly mixed at room tem-
perature with aqueous buffer. Aqueous reaction solutions were
made up with D2O containing DClO4 or solutions buffered with
CD3COOD and its base form and contained tert-butyl alcohol
as an internal standard. Control experiments with authentic
aldehydes indicated that the signal intensity of the aldehydic
proton was 79% of that of the CH protons of tert-butyl alcohol
at equimolar concentrations. The aldehydic proton was used to
quantitate yields, so the reported yields are normalized by a
factor of 0.79 for the difference in response.
(2) Kin etics. Procedures have been previously described (8).
Kinetics were monitored using an Applied Photophysic DX17MV
stopped-flow spectrophotometer by monitoring the increase in
absorbance of the aldehyde at 245, 255, or 265 nm. Reaction
was initiated by mixing a 10-fold excess of aqueous buffer with
1 volume of hydroperoxide in acetonitrile that had been previ-
ously reduced by addition of 1 equiv of tributylphosphine.
Reduced stock solutions of hydroperoxide were typically gener-
ated and stored at -40 °C and discarded within 1 h of the
reduction.
The solvent deuterium isotope effects on decomposition
of compounds 1a , 2, and 3 were determined in 0.1 M
LClO4 (L is H or D). For these determinations, the H2O
and D2O reactions were carried out on the same day in
sequence for a given compound and the values of kobsd
for each compound in each solution were based on
Resu lts
replicates of not less than 12 runs. The values of kH
/
2O
P r od u cts. The products of decomposition were inves-
tigated by H NMR by carrying out the decomposition
kD O for compounds 1a , 2, and 3 are 0.82, 1.15, and 1.31
((2<8%), respectively.
1

R-Hydroxynitrosamines in Aqueous Solution
Chem. Res. Toxicol., Vol. 13, No. 10, 2000 985
Ta ble 2. Ra te Con sta n ts for Gen er a l Acid a n d Ba se
Ca ta lysis of Decom p osition of
r-Hyd r oxyd ia lk yln itr osa m in es a t 25 °C, w ith a n Ion ic
Str en gth of 1 M (Na ClO4), in 9 vol % Aceton itr ilea
1a (R ) CH3, 1d (R ) CH3, 3 [R ) (CH3O)2-
X ) H)
X ) 4-CF3) CHCH2, X ) H]
CH3OCH2COOH
k
HA′ (M-1 s-1
)
)
)
0.36
0.52
0.13
0.61
0.06b
0.90
-1
kA (M s-1
)
-
ClCH2COOH
k
HA′ (M-1 s-1
0.81
0.18
0.27
0.31
0.32
0.25
kA (M s-1
)
-1
-
NCCH2COOH
k
HA′ (M-1 s-1
1.05
0.062
0.46
0.1
0.89
kA (M s-1
)
0.05b
-1
F igu r e 1. Plot of kobsd divided by ko, the buffer- or addend-
independent rate constant, vs buffer or addend for the decay of
R-hydroxynitrosamines, aqueous 9% acetonitrile, with an ionic
strength of 1 M (NaClO4), at 25 °C: (b) 1a , varying acetamide
at [HClO4] ) 0.01 M; (0) 1a , varying sodium trifluoroacetate,
[dichloroacetate buffer, 50% anion] ) 0.05 M; (9) 1a , varying
methoxyacetic acid buffer, 50% anion; ([) 4, varying chloroacetic
acid buffer, 50% anion; (1) 1a , varying cacodylic acid buffer,
50% anion; and (4) 1a , varying acetate buffer, 65% anion.
-
a
From the intercepts of least-squares lines of plots of kcat
against the fraction base form of the catalyst. Standard errors are
less than (10% except where indicated. Standard errors are
(25%.
b
at no fewer than three different buffer ratios. Plots of
kcat against fraction buffer base were linear, and the
intercepts of such plots at fraction base ) 0 and 1 were
taken as the buffer acid and base catalytic constants, kHA
-
and kA , respectively. Values of these constants for three
carboxylic acid buffers are summarized in Table 2.
Discu ssion
P r od u cts. As has been reported previously (4, 5, 8,
12), the products quantitated, and summarized in Table
1, indicate that the principle transformation in the
extremes of pH studied here proceeds with cleavage of
the amino nitrogen (R-hydroxy)-carbon bond to generate
aldehyde and presumably the diazoate (or diazoic acid)
that gives rise to electrophilic fragments, as in eq 2 (15-
17). The yields of alcohol determined (Table 1) reflect
lower limits due to the likely incursion of proton-for-
deuterium exchange of the diazonium ion intermediates
(16-18).
F igu r e 2. Plot of the log of ko, the buffer-independent rate
constant for decay of R-hydroxynitrosamines in aqueous 9%
acetonitrile, with an ionic strength of 1 M (NaClO4), at 25 °C,
vs pH: (0) 1b, ([) 1c, (]) 1d , (b) 4, (O) 3, and (9) 2. Solid lines
+
are fits to the equation ko ) kH + kHOH + kOH (see the text).
The inset is the plot for 1b, where the dashed line is the fit for
+
ko ) kH + kOH (see the text).
The secondary R-deuterium isotope effect on decom-
position of 1a was determined under two conditions at
25 °C, an ionic strength of 1 M, and 9 vol % acetonitrile.
Reaction solutions contained either 0.1 M HClO4 or 0.05
M acetic acid buffer, 90% base form. In the latter case,
control experiments indicated for the protio compound
that the value of kobsd was less than 5% greater than the
value of ko, the buffer-independent rate constant. Reac-
tions with the protio and deuterio compounds were
carried out under one set of conditions on the same day
in sequence, with replicate determinations of kobsd for
each compound being comprised of not less than 11 runs.
The value of kRH/kRD ) 1.12 ( 0.03 and 1.19 ( 0.02 in 0.1
M HCl and 0.05 M acetic acid buffer (90% base form),
respectively.
All of the product determinations reported here, with the
one exception described in detail below, were carried out
in deuterium oxide solutions.
Two observations rule out substantial decomposition
via a nitrosiminium ion (10, 19-21) in acidic media. Such
a mechanism, below, involving initial CO bond cleavage
and subsequent denitrosation of the nitrosiminium ion
would yield benzaldehyde and methylammonium ion as
products (21).
(2) Bu ffer -Ca ta lyzed Rea ction s. For compounds 1a ,
1c, and 4, a systematic study of the effects of buffer
concentration on kobsd was undertaken in the pH region
between pH 2 and 5. Figure 1 illustrates the fact that
increasing concentrations of more acidic carboxylic acid
buffers catalyzed the decay of the R-hydroxynitrosamines.
The slopes of the lines in Figure 1 were taken as the
catalytic constants kcat, and values of kcat were determined
Experiments carried out in D2O (0.1 M DClO4) with NMR
analysis gave an 81% yield (relative to 4-chlorobenzal-
dehyde) of CH3OH and failed to detect methylammonium
ion, checked by subsequent addition of the hydrochloride

986 Chem. Res. Toxicol., Vol. 13, No. 10, 2000
Mesic' et al.
Ta ble 3. Ra te Con sta n ts for th e Deca y of
salt, setting an upper limit on the yield of 5%. In a second
experiment with 1a in 20% methanolic H2O (0.1 M
HClO4), HPLC analysis failed to detect the R-methoxy
compound that would result from capture of a putative
N-nitrosiminium ion by methanol. Based on the known
partitioning of the cation between capture by water and
by methanol (13), a yield of 47% is predicted assuming
quantitative conversion to the nitrosiminium ion. The
upper limit yield of 3% indicates that the nitrosiminium
ion pathway accounts for less than 6% of the total
reaction. On these bases, the contribution of a nitros-
iminum ion pathway in acidic media appears to be
minimal. Such a pathway has been previously proposed
for decomposition of R-hydroxydialkylnitrosamines in
neutral media (22), but there are alternative explanations
for the experimental observations that gave rise to this
conclusion.
r-Hyd r oxyd ia lk yln itr osa m in es in Aqu eou s Solu tion s, a t
25 °C, w ith a n Ion ic Str en gth of 1 M (Na ClO4), in 9 vol %
Aceton itr ilea
kH
kHOH
10-9kOH
compd
R1
X
(M-1 s-1
)
(s-1
)
(M-1 s-1
)
1a
1b
1c
1d
2
CH3
CH3
CH3
CH3
CH3OCH2CH2
(CH3O)2CHCH2
NCCH2CH2
H
56
34
16
10.5
72
0.12
1.7
2.1
2.9
3.0
3.5
5.8
8.5
4-Cl
3-Cl
4-CF3
H
H
H
0.13b
0.14b
0.16b
0.093c
0.055c
0.084c
3
4
73
24
a
Derived by nonlinear least-squares fitting with proportional
weighting from the data depicted in Figure 1. Standard errors are
b
less than (10% except where indicated. The standard error is
less than (15%. c The standard error is less than (25%.
Kin etics. (1) Gen er a l Ra te La w . In toto, the experi-
mental evidence described in detail below requires the
minimal rate law defined by eqs 3 and 4. The terms
accounting for catalysis by buffer acids (kHA) and bases
anism for catalysis by acids, both the hydronium ion and
the stronger carboxylic acids (Table 2), involves concerted
proton transfer and leaving group expulsion.
Two possible mechanisms for this concerted general
acid catalysis are simple general acid catalysis, repre-
sented in eq 5, and the alternative specific acid general
base catalysis, represented in eq 6.
-
(kA ) that were observed for some carboxylic acid buffers
over a narrow pH range have been represented slightly
differently in eqs 3 and 4, for reasons that will be
described in detail later in the discussion. At present, it
is sufficient to indicate that the mechanism of the
-
reaction of certain carboxylic acid anions (kA ) appears
to be distinct from that for the reaction of hydroxide ion
+
(kOH), while the catalysis by the proton (kH ) and some
carboxylic acids (kHA) is believed to be identical so that
they have been summed to a single constant, kHA′, in eq
3. Three buffer-independent rate constants
kobsd ) kHA′ + kHOH + kOH + kA
(3)
(4)
-
k
HA′ ) kH+ + kHA
Evidence for the concerted nature of the reaction is
presented in the first subsection, while the analysis that
leads to the conclusion that the mechanism is as indi-
cated in eq 6 is presented in the second subsection.
(i) Evid en ce for Con cer ted P r oton Tr a n sfer a n d
Lea vin g Gr ou p Exp u lsion . (1) The involvement of
proton transfer in the rate-limiting step is indicated by
the solvent deuterium isotope effect on catalysis by the
+
for hydrogen ion-catalyzed (kH ), hydroxide ion-catalyzed
(kOH), and water-catalyzed (kHOH) decomposition are
required by the pH-rate profiles of Figure 2. This is
consistent with what has been reported previously (8).
The water-catalyzed term contributes to the overall
buffer-independent term only over a narrow pH range,
in contrast to simple methylene R-hydroxynitrosamines
(4, 5, 8). The fact that such a reaction nonetheless
contributes is indicated by the fit in the inset plot of
Figure 2, which compares the best fit of the data for 3 to
a three-term rate equation (solid line) and a two-term
equation (dashed line) containing only terms for the
hydrogen ion- and hydroxide ion-catalyzed reactions.
Such fits for all compounds underestimated, by factors
of 2-4, the values of ko in the pH regions of the minima
of the profiles in Figure 2. The precision of the rate
constant determinations is in most cases better than
(10%, this error being too small to be consistent with
the fit to the two-term rate equation. Values for the rate
+
+
hydronium ion where kH /kD ) 0.82-1.3, depending on
X, eq 5 or 6, when R is CH3. This range of values is
inconsistent with equilibrium protonation followed by a
rate-limiting step that does not involve proton transfer
(23). Such mechanisms, typified by the generic example
of eq 7, are characterized by substantially smaller inverse
+
+
solvent isotope effects kH /kD of 0.3-0.5 that arise from
the loss of the differences in zero-point energy difference
for hydrogens of the lyonium ion.
+
constants kH , kHOH, and kOH derived from the best fits
are summarized in Table 3. In general, the standard
errors are less than (10%, but this is not the case for
most of the values of kHOH, for which some of the standard
errors are (25% due to the lack of dominance of this
reaction pathway.
In the present case, an upper limit for such a mechanism
might be as large as 0.69 (kH /kD ) if the reaction involved
equilibrium protonation of the hydroxyl oxygen, as in eq
8.
+
+
-
Rate constants kA and kHA for catalysis by carboxylic
acids are summarized in Table 2, and were determined
as previously described (see Results).
Mech a n ism s. (1) kHA′ Rea ction (eq 6). A number of
pieces of experimental evidence indicate that the mech-

R-Hydroxynitrosamines in Aqueous Solution
Chem. Res. Toxicol., Vol. 13, No. 10, 2000 987
+
+
Isotope effects kH /kD of ∼1 are typical of reactions in
which the inverse isotope effect due to the disappearance
of the lyonium ion is offset by a normal isotope effect that
arises from the loss of zero-point energy between H and
D in a transition state in which the hydrogen is “in flight”
(23).
(2) The observation of general acid catalysis for fairly
acidic carboxylic acids is also consistent with a mecha-
nism in which a proton is transferred in the rate-limiting
step (24, 25). Reactions such as those in eq 7 do not
exhibit catalysis by buffer acids, and the rates of such
reactions are only affected by the proton concentration
of the media. The observed general acid catalysis is
modest; Figure 1 indicates typical rate increases at the
highest buffer concentrations of less than 100%. As will
be discussed below, this effect is due to both contributions
of the acid as well as base form of the buffer, but the
constants summarized in Table 2 indicate a significant
contribution of the carboxylic acid in all cases. Despite
the modest effects, control experiments (Figure 1) confirm
that the increases do reflect catalysis and not medium
effects. The data in Figure 1 illustrate that the addition
of comparable concentrations of the very weakly basic
trifluoroacetate anion and acetamide, which mimic the
base and acid forms of the carboxylic acids, do not
similarly increase the values of kobsd. This indicates that
the observed increases with cyanoacetic, chloroacetic, and
methoxyacetic acid buffers are not the result of specific
salt or medium effects on the reaction.
The limited structure-reactivity correlations in this
study do not permit a precise determination of the
position of the proton in the transition state, but are in
two cases consistent with a proton in flight. The slopes
of the three point Bronsted plots of log kHA against pKa
for the three carboxylic acid catalysts are R ) -0.38 and
-0.42 (r2 ) 0.999 and 0.968, respectively) for compounds
1a and 4. These values of R are intermediate between
normally limiting values of 0 and -1 that are observed
for mechanisms of catalysis in which the rate-limiting
step involves diffusional encounter/separation or as-
sistance through hydrogen bonding (25); this leads to the
conclusion that in the present case, the proton is in flight.
The R value of -0.98 for compound 1d (r2 ) 0.989)
suggests that the addition of electron-withdrawing groups
to the leaving group results in a change to a transition
state that is highly asymmetric (product-like with respect
to proton transfer). The nature and implication of this
change will be described in detail in the further discus-
sion.
F igu r e 3. Plot of the logarithm of kHA, the second-order rate
constant for catalysis by the indicated acids, vs the pKa
calculated (17) for the diazoic acid leaving groups for compounds
1a (pKa ) 8.68) and 1c (pKa ) 8.07).
The observed effect on the rate constant for the present
reaction suggests that the bond to the leaving diazoate
anion is 40-30% [)(0.12/0.3-0.4) × 100] cleaved.
(ii) An a lysis In d ica tin g th e Mech a n ism of eq 6 for
k
HA′. The mechanisms of eqs 5 and 6 are an example of
kinetic ambiguity, and it is not possible on the basis of
the values of structure-reactivity correlations to distin-
guish between them. For example, both mechanisms are
consistent with the positive charge buildup on the leaving
group that is indicated by the positive slopes of the
dependence of the logarithm of the catalytic coefficients
for carboxylic acid catalysts upon the calculated leaving
group pKa (17) that is observed in Figure 3. A distinction
is however possible because the two mechanisms make
opposite predictions regarding the changes in such cor-
relations with catalyst pKa that are predicted by these
two mechanisms (29-31). Figure 3 shows that the slopes
of the lines decrease as the acid catalyst strength
increases on going from methoxyacetic acid to the hy-
dronium ion.
The changes that are predicted by the two mechanisms
can be understood by examination of the three-dimen-
sional reaction coordinate diagrams in Figure 4A for the
mechanism of eq 6 and Figure 4B for the mechanism of
eq 5. The “unseen” third dimension is a free energy axis
that is orthogonal to the axes in Figure 4 that indicate
progress along a coordinate for proton transfer and
leaving group departure, respectively. As mentioned
earlier, extensive structure-reactivity correlations are
not available to precisely place the transition state on
the map for either mechanism. However, the important
point for the present purposes is that the discussion in
the previous subsection establishes that the transition
states for some of the reactants that have been studied
are oriented diagonally with respect to the two axes for
bonding changes, indicating a coupled concerted mech-
anism.
(3) The secondary R-deuterium kinetic isotope effect
+R-D
kH R-H/kH
of 1.12, determined in 0.1 M HClO4, re-
+
quires that there is significant rehybridization at the
R-hydroxy carbon in the transition state compared to the
ground state, indicative of leaving group departure and
the onset of sp2 hybridization that typifies the product
aldehyde. The maximum effect, for a transition state with
essentially complete N-C bond cleavage, would be
expected to be 1.3-1.4, based on a number of equilibrium
isotope effects determined for reactions at the carbonyl
group of benzaldehyde (26-28) generally indicated in eq
9.
There is a detectable change in transition state struc-
ture that is consistent with the changes expected from
the reaction coordinate diagram in Figure 4A drawn for
the mechanism of eq 6 (31). The change in transition
state structure is indicated in Figure 3 which illustrates
a change in sensitivity to the diazoic acid leaving group
pKa of the second-order rate constants for catalysis by
the hydronium ion and carboxylic acids. This is a change
in the parameter known as âlg, the magnitude of which
corresponds to the slopes of the lines in Figure 3. The
sensitivity to leaving group increases, with decreasing
acid strength, from H3O+ (top) to methoxyacetic acid

988 Chem. Res. Toxicol., Vol. 13, No. 10, 2000
Mesic' et al.
catalyst acidity, is toward the left axis, where the leaving
group has a full positive charge, and away from the right
axis where the leaving group has no charge. This
predicted increase in positive charge on the leaving group
in changing from the catalyst H3O+ to methoxyacetic acid
is what is indicated by the experimental data in Figure
3.
The change in transition state structure evident in
Figure 3 is the opposite of that expected for a transition
state for the mechanism of eq 5 that has a diagonal
orientation, as indicated by the double-headed arrow in
Figure 4B. The change from a stronger to a weaker acid
catalyst lowers the bottom edge of the diagram in Figure
4B relative to the upper edge. The corresponding move-
ment parallel and perpendicular to the reaction coordi-
nate is indicated by the dashed arrows. The net move-
ment is indicated by the bold arrow in the diagram. The
projection of this movement onto the diagonal axis (open
arrow) indicated by the dashed line, along which diagonal
the charge on the leaving group varies from +1 (upper
left corner) to -1 (lower right corner), shows that a
decrease in positive charge on the leaving group is
predicted with decreasing acid strength, the opposite of
what is observed experimentally and illustrated in Figure
3.
-
(2) kA Rea ction (eq 10). It is shown below that the
observed changes in structure-reactivity correlations for
these rate constants can only be accommodated by a
concerted transition state for the mechanism of eq 10,
the specific base or general acid variant of the observed
general base catalysis.
F igu r e 4. Reaction coordinate diagrams for the two concerted
general acid-catalyzed reactions of eq 6 (A) and eq 5 (B). The
orientation of the reaction path at the transition state is
indicated by the double-headed arrows. The perturbations of the
parallel and perpendicular components of the reaction coordi-
nate at the transition state that result from decreasing acid
strength are indicated by the dashed arrows. The resultant
movement predicted for the transition state is indicated by the
bold arrow.
(bottom). A positive value of the slope indicates that the
more basic leaving group is more reactive and, for the
least acidic catalysts, in fact bears some positive charge
in the rate-limiting step. This observation requires that
in the transition state for the reactions of the least acidic
catalysts, proton transfer to the leaving group is more
advanced than nitrogen-carbon bond cleavage. With
increasing catalyst acidity, there is increasing balance
between the two processes. With the hydronium ion, the
near-zero slope of the line indicates little or no charge
on the leaving group; proton transfer and C-N bond
cleavage are more nearly balanced.
This type of change is exactly what is predicted by the
reaction coordinate diagram of Figure 4A (31). The
change from H3O+, a stronger acid catalyst, to methoxy-
acetic acid, a weaker acid, raises the lower edge of the
diagram in Figure 4A relative to the upper edge. The
response of the transition state to this perturbation will
be to move downhill on the perpendicular axis of the
reaction coordinate at the transition state and uphill
along the parallel axis, as indicated by the dashed arrows
in Figure 4A. The net movement is indicated by the heavy
arrow in the Figure 4A. The movement, with decreasing
It is further concluded below that the alternative con-
certed general base catalysis mechanism of eq 11 is
inconsistent with the observed changes.
Three types of changes in structure-reactivity rela-
tionships, which indicate changes in transition state
structure, are observed in the general base-catalyzed
reaction.
Figure 5 indicates (a) that the values of âlg, the
-
dependence of the catalytic constants kA against the
calculated pKa values of the conjugate acids of the
diazoate leaving groups for compounds 1a and 1c, change
with catalyst basicity. The value of âlg is slightly positive
for cyanoacetate (âlg ) 0.15) but changes to negative
values for chloroacetate and methoxyacetate (âlg ) -0.25
and -0.46, respectively). The slightly positive value for
cyanoacetate is most consistent with the mechanism of
eq 10, in which H+ donation in the second step is slightly
ahead of N-C bond cleavage. This imbalance would lead
to a net positive charge development on the leaving
group. The negative values in the cases of the other
catalysts indicate a change in transition state structure
to one in which there is negative charge buildup on the
leaving group in the transition state.

R-Hydroxynitrosamines in Aqueous Solution
Chem. Res. Toxicol., Vol. 13, No. 10, 2000 989
-
F igu r e 5. Plot of the logarithm of kA , the second-order rate
constant for catalysis by the indicated bases, vs the pKa
calculated (17) for the diazoic acid leaving groups for compounds
1a (pKa ) 8.68) and 1c (pKa ) 8.07).
F igu r e 7. Reaction coordinate diagram for the proposed
concerted general base-catalyzed mechanism of eq 10. The
orientation of the reaction path at the transition state is
indicated by the double-headed arrow. The bold arrows indicate
the expected movement of the transiton state for (a) an increase
in the strength of the base catalyst, (b) an increase in leaving
group ability, and (c) introduction of electron-withdrawing
groups into the benzene ring of compound 1a (see the text).
rate-limiting step is measured along the two axes. The
extent of charge development on the leaving group varies
along the diagonal from the upper left corner (ca. +1) to
the lower right corner (ca. -1). The transition state for
reaction of compound 1a is located in the lower right
quadrant, consistent with a âlg of ca. -0.25 (for chloro-
acetate) and the extent of proton transfer indicated by a
Bronsted â of 0.7 and the relationship for specific base
or general acid catalysis, R(for proton donation) ) 1 -
â(observed). The double-headed arrow indicates the
orientation of the reaction coordinate at the transition
state. The movements of the transition state in response
to changes in catalyst strength (a, above), improved
leaving group ability (b, above), and introduction of an
electron-withdrawing group (c, above) are indicated by
the bold arrows in Figure 7, labeled with the correspond-
ing letters a-c, respectively. The movements are consis-
tent with the observed effects as indicated below.
Ch a n ge a . The change from the weaker catalyst
chloroacetate to the more basic methoxyacetate raises the
energy of the upper edge relative to the lower edge of
the diagram in Figure 7. The transition state will slide
downhill along the axis perpendicular to the reaction
coordinate and uphill (Hammond effect) on the axis
parallel to the reaction coordinate with the net movement
indicated by vector a in Figure 7. This movement,
projected onto the diagonal (from the upper left to the
lower right), that indicates charge status of the leaving
group predicts movement toward the lower right corner
and a consequent increase in negative charge buildup on
the leaving group. The increase in negative charge on
the leaving group as the catalyst is changed from
chloroacetate to methoxyacetate is consistent with what
is observed in Figure 5, an increase in the negative slope
in Figure 5 on changing from chloroacetate to methoxy-
acetate.
F igu r e 6. Plot of the logarithm of kA- for 1c (9) and 4 (b) vs
those for 1a . The slopes of the lines are 1.4 (9) and 0.85 (b).
Additionally, there is (b) an increase in the parameter
Bronsted â with increasing leaving ability of the leaving
group and (c) a decrease in Bronsted â with introduction
of an electron-withdrawing group in the benzene ring.
The values of Bronsted â, reflecting to some degree the
change in charge experienced by substituents in the
catalyst due to proton transfer in the transition state,
are large to intermediate with â ) 0.7 (r2 ) 0.925), 0.57
(r2 ) 0.904), and 0.96 (r2 ) 0.942) for 1a , 4, and 1c,
respectively. The value near 1 for compound 1c indicates
that there is nearly unit negative charge buildup on going
from the ground to the transition state, indicating the
proton is almost completely transferred. The intermedi-
ate values for the other compounds suggest that the
proton transfer is less complete. There is imprecision in
these determinations, largely due to the systematic
scatter that is sometimes observed even in a “homoge-
neous” series such as substituted acetic acids (32). The
systematic nature of the scatter is indicated by the plots
in Figure 6 in which the log values of the rate constants
for 1c and 4 are plotted against those for compound 1a
and show appreciably better correlations (r2 ) 0.999 and
0.998, respectively). They indicate that (1) there is an
increase in Bronsted â with an increase in leaving group
ability, changing from 1a to 1c [the slope (9) ) 1.4], and
(2) there is a decrease in Bronsted â with the introduction
of an electron-withdrawing group in the phenyl ring,
changing from 1a to 4 [the slope (b) ) 0.85].
The changes in structure-reactivity correlations de-
picted in Figures 5 and 6 are consistent with the
concerted mechanism of eq 10. This can be understood
with reference to the three-dimensional reaction coordi-
nate diagram of Figure 7 in which progress with respect
to proton transfer and leaving group departure in the
Ch a n ge b. Increase in leaving group ability, changing
from 1a to 1c, lowers the lower right corner relative to

990 Chem. Res. Toxicol., Vol. 13, No. 10, 2000
Mesic' et al.
the upper left, the two corners at which the leaving group
has opposite charges. The resultant transition state
movement is as indicated by the bold arrow labeled b.
This movement, projected onto the vertical axis, predicts
a decrease in the extent of proton donation, which is what
is measured along the vertical axis. The net result is an
increase in Bronsted â, consistent with the slope of >1
for the squares in Figure 6.
by J encks (25, 34, 35), the driving force for concerted
catalysis in the second step of the mechanism of eq 10
derives from the change from a thermodynamically
unfavorable protonation of the nitroso oxygen in the
R-hydroxynitrosamine anion in the reactants to a ther-
modynamically favorable protonation of the diazoate
leaving group in the products. A significant driving force
is required to pay for the entropically unfavorable as-
sociation and proper orientation in the concerted transi-
tion state.
Ch a n ge c. Introduction of an electron-withdrawing
-
CF3 group into the benzene ring in compound 1a
(changing from compound 1a to 4) raises the energy of
the right side of the diagram, relative to the left. This is
because the carbonyl group is electron deficient relative
to the tetrahedral carbon atom on the left edge of Figure
7. The resultant transition state movement is indicated
by the bold arrow labeled c. The net movement relative
to the (vertical) proton transfer axis is an increase in the
extent of proton donation for 1a compared to 4. The net
result is a decrease in Bronsted â, consistent with the
slope of <1 for the circles in Figure 6.
In contrast to the above discussion, the appropriate
energy surface containing a transition state with an
appreciable diagonal orientation at the transition state
for the mechanism of eq 11 predicts transition state
movements in response to the above perturbations that
are opposite to what is observed, in all cases.
(3) kOH Rea ction (Mech a n ism of eq 12). With one
exception, the hydroxide ion-catalyzed reaction likely
involves rate-limiting leaving group expulsion from the
anionic conjugate base of the R-hydroxynitrosamine, as
in eq 12 (8).
This driving force disappears as the proton transfer from
the conjugate acid to the diazoate leaving group (conju-
gate acid pKa ∼ 8-9) becomes less thermodynamically
favorable. The absence of driving force in the case of the
hydroxide ion reaction is illustrated in eq 13 in which
the rate-limiting step from eq 10 (horizontal reaction) is
accompanied by analysis of the thermodynamics of the
relevant proton transfers in the reactants and products
(vertical equilibria) (17, 36). In both reactants and
products, comparison of the pKa values indicates thermo-
dynamically unfavorable proton transfers. The absence
of thermodynamic driving force makes the concerted
pathway entropically disfavored, so the lowest-energy
pathway is the mechanism of eq 12.
The large secondary R-deuterium kinetic isotope effect
R-D
(kOHR-H/kOH
) 1.19) indicates a significant degree of
sp3-sp2 rehybridization at the central carbon, consistent
with leaving group bond cleavage in the transition state,
as discussed above for the hydronium ion-catalyzed
reaction. A plot of log kOH against the calculated conjugate
acid pKa of the diazoate leaving group (not shown) has a
slope of -0.88 (r2 ) 0.98), again suggesting that there is
a significant amount of negative charge buildup on the
leaving group in the transition state. Neither of these
observations is consistent with rate-limiting proton trans-
fer with hydroxide ion, the anionic intermediate which
then forms products more rapidly than it reverts to
reactants. The rate-limiting step for such a reaction
would involve diffusion-limited encounter of hydroxide
ion with the R-hydroxynitrosamine (33), with no rehy-
bridization at the central carbon or significant charge
buildup on the leaving group.
With the best leaving group studied here, compound
1d , a change in the rate-limiting step is likely occurring.
The second-order rate constant for the hydroxide ion-
catalyzed reaction of compound 1d is ∼9 × 109 M-1 s-1
,
which approaches the value expected for a diffusion-
limited proton transfer involving hydroxide of ∼2 × 1010
M-1 s-1. This indicates that the barrier for leaving group
expulsion from the conjugate base of the R-hydroxyni-
trosamine is quite small; slightly less than half of all
diffusional encounters with hydroxide ion result in
product formation. Thus, both proton transfer and leav-
ing group expulsion are partly rate-limiting in this case.
It is likely that with better leaving groups, the rate-
limiting step would be simple proton transfer and might
well be accompanied by general base catalysis (34, 35),
according to the mechanism of eq 11, in which the rate-
limiting step for proton abstraction by bases weaker than
the anion of the R-hydroxynitrosamine would be either
rate-limiting diffusional separation of the catalyst con-
jugate acid from the R-hydroxynitrosamine anion or,
more likely, leaving group expulsion in the conjugate
acid-base encounter complex.
There is no evidence for a reaction mechanism for
hydroxide ion (kOH) analogous to the mechanism of eq
10, which is observed for the less basic carboxylate bases
-
(kA ), and such a mechanism is not expected. The absence
of catalysis by more strongly basic catalysts such as
acetate and cacodylate anions (Figure 1) argues against
the mechanism of eq 10 (and also eq 11) for the more
basic hydroxide ion. The disappearance, with increasing
catalyst basicity, of catalysis by the mechanism of eq 10
is expected because the driving force for catalysis de-
creases with increasing catalyst basicity. As formulated
(4) kHOH Rea ction . Due to the fact that this reaction
is never dominant and the rate constants could only be
determined with limited accuracy, the mechanism of this
reaction remains the subject of ongoing inquiry.

R-Hydroxynitrosamines in Aqueous Solution
Chem. Res. Toxicol., Vol. 13, No. 10, 2000 991
Gen er a l Im p lica tion s. (1) For most R-hydroxynitros-
amines studied by this group to date, the dominant
mechanism at neutral pH is that of eq 12, involving
equilibrium deprotonation to form the oxyanion and rate-
limiting expulsion of the diazoate anion (this work and
refs 8-10). The rate constant for this latter process
correlates quite strongly with the conjugate acid pKa of
the diazoate (âlg ) -0.88). The ultimate electrophilic
species arising from this process is the diazonium ion.
Alternative electrophilic species such as the N-nitros-
iminium ion, and nitrosonium ion equivalents that might
arise from it, are not encountered in this dominant
reaction mode or any of the other mechanisms delineated
in the study presented here. There is some evidence that
R-hydroxynitrosamines of differing structure may in fact
decompose by other modes in which alternative electro-
philic species may be encountered (6, 7). Both hydroxy-
nitrosoindoles and purported R-hydroxynitrosomorpho-
line exhibit surprising stability compared to the com-
pounds that we have studied here and in other reports,
and exhibit products of decomposition that are not
consistent with any reaction pathways elucidated in this
report.
gate base of the R-hydroxydialkylnitrosamines as in eq
12. A significant amount of cleavage of the carbon-leaving
group bond cleavage in the rate-limiting step is indicated
by the secondary R-deuterium kinetic isotope effect
(kOHR-H/kOHR-D ) 1.19) and the large value of âlg (-0.88).
Ack n ow led gm en t. This work was supported by
Grants RO1 CA52881 and KO4 CA62124 from the
National Institutes of Health.
Su p p or tin g In for m a tion Ava ila ble: 1H NMR spectra of
1d , 2, and 3 and 13C NMR spectra of 2 and 3. This material is
available free of charge via the Internet at http://pubs.acs.org.
Refer en ces
(1) Lawley, P. D. (1984) Carcinogenesis by Alkylating Agents. In
Chemical Carcinogens, Vol. 1 (Searle, C. E., Ed.) Vol. 182, pp 325-
484, American Chemical Society, Washington, DC.
(2) Loeppky, R. N., and Michejda, C. J . (1994) Nitrosamines and
Related N-Nitroso Compounds, American Chemical Society,
Washington, DC.
(3) Lijinsky, W. (1992) Chemistry and biology of N-nitroso compounds,
Cambridge University Press, Cambridge, U.K.
(4) Mochizuki, M., Anjo, T., and Okada, M. (1980) Isolation and
Characterisation of N-Alkyl-N-(hydroxymethyl)nitrosamines from
N-Alkyl-N-(hydroperoxymethyl)nitrosamines by Deoxygenation.
Tetrahedron Lett. 21, 3693-3696.
(2) This report indicates that general acid and base
moieties can catalyze the decomposition of R-hydroxyni-
trosamines and thus establishes a chemical precedent for
the involvement of biomolecules in their decay. This,
combined with a further consideration, presents a pos-
sible chemical origin for the well-known “hot spots” in
the mutation spectra of nitrosamine alkylating agents
(37). The work of Gold has established that nitrosoureas
covalently appended to various DNA binding agents
“deliver” alkylating equivalents at the preferred sites of
binding (38-41). This is surprising in light of the fact
that the diazoic acids and primary diazonium ion inter-
mediates that are involved are relatively long-lived
intermediates, at least relative to diffusion in bulk
solution (42). The observation requires that these inter-
mediates be to some degree confined to the site of
delivery. To the extent that DNA sequence contexts can
activate, by exposure or by changes in pKa, the acid and
base properties of DNA bases or the phosphate backbone,
the observations of acid and base catalysis in this report
suggest that such sites might stimulate R-hydroxynitros-
amine decomposition and subsequently be sites of en-
hanced alkylation. The numerous other chemical and
biological factors that likely contribute to the appearance
of hot spots have been discussed (37).
(5) Mochizuki, M., Anjo, T., Takeda, K., Suzuki, E., Sekiguchi, N.,
Huang, G. F., and Okada, M. (1980) Chemistry and Mutagenicity
of R-Hydroxy Nitrosamines. IARC Sci. Publ. 41, 553-559.
(6) J arman, M., and Manson, D. (1986) The metabolism of N-
nitrosomorpholine by rat liver microsomes and its oxidation by
the Fenton system. Carcinogenesis 7, 559-565.
(7) Buchi, G., Lee, G. C. M., Yang, D., and Tannenbaum, S. R. (1986)
Direct acting, highly mutagenic, R-hydroxy nitrosamines from
4-chloroindoles. J . Am. Chem. Soc. 108, 4115-4119.
(8) Mesic’, M., Revis, C., and Fishbein, J . C. (1996) Effects of
Structure on the Reactivity of R-Hydroxydialkylnitrosamines in
Aqueous Solutions. J . Am. Chem. Soc. 118, 7412-7413.
(9) Chahoua, L., Mesic’, M., Revis, C., and Fishbein, J . C. (1997)
Evidence for the formation of R-hydroxydialkylnitrosamines in
the pH-independent solvolysis of R-acetoxydialkylnitrosamines.
J . Org. Chem. 62, 2500-2504.
(10) Chahoua, L., Cai, H., and Fishbein, J . C. (1999) Cyclic R-acetoxy-
nitrosamines: Mechanisms of decomposition and stability of
R-hydroxynitrosamine and nitrosiminium ion reactive intermedi-
ates. J . Am. Chem. Soc. 121, 5161-5169.
(11) Mochizuki, M., Anjo, T., and Wakabayashi, Y. (1980) Formation
of N-Alkyl-N-(1-hydroperoxyalkyl)nitrosamines from N-Alkyl-N-
(1-acetoxy)nitrosamines. Tetrahedron Lett. 21, 1761-1764.
(12) Okada, M., Mochizuki, M., Anjo, T., Sone, T., Wakabayashi, Y.,
and Suzuki, E. (1980) Formation, deoxygenation and mutagenicity
of R-hydroperoxydialkylnitrosamines. IARC Sci. Publ. 31, 71-
82.
(13) Chahoua, L., Vigroux, A., Chiang, Y., and Fishbein, J . C. (1999)
Formation and nucleophilic capture of N-nitrosiminium ions. Can.
J . Chem. 77, 1148-1161.
(14) Cai, H., and Fishbein, J . C. (1997) Secondary R-Deuterium Kinetic
Isotope Effects in the Decomposition of Simple R-Acetoxydialkyl-
nitrosamines: Nitrosiminium Ion Intermediates. Tetrahedron 53,
10671-10676.
(15) Hovinen, J ., and Fishbein, J . C. (1992) Rate Constants for the
Decomposition of a Simple Alkanediazoate at Physiological pH.
J . Am. Chem. Soc. 114, 366-367.
(16) Hovinen, J ., Finneman, J . I., Satapathy, S. N., Ho, J ., and
Fishbein, J . C. (1992) Mechanisms of Decomposition of (E)-
Methanediazoate in Aqueous Solutions. J . Am. Chem. Soc. 114,
10322-10328.
Su m m a r y
The mechanism of hydronium ion- and general acid-
catalyzed decay (kHA′, eqs 3 and 4) of the R-hydroxydi-
alkylnitrosamines in the study presented here is as in
eq 6. Secondary R-deuterium kinetic isotope effects
(kH R-H/kH R-D ) 1.12), solvent deuterium isotope effecets
+
+
(17) Ho, J ., and Fishbein, J . C. (1994) Rate-Limiting Formation of
Diazonium Ions in Aqueous Decomposition of Primary Alkanedi-
azoates. J . Am. Chem. Soc. 116, 6611-6621.
(18) Smith, R. H., J r., Koepke, S. R., Tondeur, Y., Denlinger, C. L.,
and Michejda, C. J . (1985) The Methyldiazonium Ion in Water:
Competition Between Hydrolysis and Proton Exchange. J . Chem.
Soc., Chem. Commun., 936-937.
+
on the kH reaction, and general acid catalysis support a
concerted mechanism. The mechanism of eq 6 is indicated
by changes in structure-reactivity correlations which
dictate the nature of the transition state for this reaction.
-
The mechanism of general base catalysis (kA , eq 3) is
as indicated in the mechanism of eq 10. This is indicated
again by changes in structure-reactivity correlations.
In most cases, the mechanism of the hydroxide ion-
catalyzed reaction (kOH, eq 3) involves rate-limiting
leaving group expulsion of the diazoate from the conju-
(19) Revis, C. L., Rajamaki, M., and Fishbein, J . C. (1995) Reexamina-
tion of the Mechanisms of Decomposition of Simple R-Acetoxyni-
trosamines in the Physiological pH Range. J . Org. Chem. 60,
7733-7738.
(20) Cai, H., and Fishbein, J . C. (1999) R-(Acyloxy)dialkylnitros-
amines: Effects of Structure on the Formation of N-Nitros-

992 Chem. Res. Toxicol., Vol. 13, No. 10, 2000
Mesic' et al.
iminium Ions and a Predicted Change in Mechanism. J . Am.
Chem. Soc. 121, 1826-1833.
(33) Eigen, M. (1964) Proton transfer, acid-base catalysis, and
enzymatic hydrolysis. Angew. Chem., Int. Ed. 3, 1-72.
(21) Rajamaki, M., Vigroux, A., Chahoua, L., and Fishbein, J . C. (1995)
N-Nitrosiminium Ions are Ambident Electrophiles. J . Org. Chem.
60, 2324-2325.
(22) Frank, N., and Wiessler, M. (1986) N-Nitroso-hydroxyalkylamine
phosphate esters: a new class of N-nitroso compounds. Carcino-
genesis 7, 365-369.
(34) J encks, W. P. (1972) General acid-base catalysis of complex
reactions in water. Chem. Rev. 72, 705-718.
(35) J encks, W. P. (1972) Requirements for general acid-base catalysis
of complex reactions. J . Am. Chem. Soc. 94, 4731-4732.
(36) Layne, W. S., J affe, H. F., and Zimmer, H. (1963) Basicity of
N-Nitrosamines. II. Aqueous Sulfuric Acid Solutions. J . Am.
Chem. Soc. 85, 1816-1820.
(23) Schowen, R. L. (1972) Mechanistic deductions from solvent isotope
effects. Prog. Phys. Org. Chem. 9, 275-330.
(24) J encks, W. P. (1969) Catalysis in chemistry and enzymology,
McGraw-Hill, New York.
(25) J encks, W. P. (1976) Enforced general acid-base catalysis of
complex reactions and its limitations. Acc. Chem. Res. 9, 425-
432.
(26) Amaral, L. d., Bull, H. G., and Cordes, E. H. (1972) Secondary
deuterium isotope effects for carbonyl addition reactions. J . Am.
Chem. Soc. 94, 7579-7580.
(27) Amaral, L. d., Bastos, M. P., Bull, H. G., and Cordes, E. H. (1973)
Secondary deuterium isotope effects for addition of nitrogen
nucleophiles to substituted benzaldehydes. J . Am. Chem. Soc. 95,
7369-7374.
(28) Young, P. R., and McMahon, P. E. (1979) Separation of polar and
resonance substituent effects in the reaction of benzaldehydes
with HCN. A correlation cetween Fr/Freq ratios and central atom
rehybridization. J . Am. Chem. Soc. 101, 4678-4681.
(29) Funderburk, L. H., Aldwin, L., and J encks, W. P. (1978) The
mechanisms of general acid and base catalysis of the reactions of
water and alcohols with formaldehyde. J . Am. Chem. Soc. 100,
5444-5459.
(30) Sorensen, P. E., and J encks, W. P. (1987) The acid- and base-
catalyzed decomposition of acetaldehyde hydrate and hemiacetals
in aqueous solution. J . Am. Chem. Soc. 109, 4675-4690.
(31) J encks, W. P. (1985) A primer for the Bema Hapothle. An
empirical approach to the characterisation of changing transition-
state structures. Chem. Rev. 85, 511-527.
(37) Horsfall, M. J ., Gordon, A. J ., Burns, P. A., Zielenska, M., and
Glickman, B. W. (1990) Mutational Specificity by Alkylating
Agents and the Influence of DNA Repair. Environ. Mol. Mutag.
15, 107-122.
(38) Konakahara, T., Wurdeman, R. L., and Gold, B. (1988) Synthesis
of N-methyl-N-nitrosourea linked to a methidium chloride ana-
logue and its reactions with 32P-end-labeled DNA. Biochemistry
27, 8606-8613.
(39) Church, K. M., Wurdeman, R. L., Zhang, Y., Chen, F. X., and Gold,
B. (1990) N-(2-Chloroethyl)-N-nitrosoureas covalently bound to
nonionic and monocationic lexitropsin dipeptides. Synthesis, DNA
affinity binding characteristics, and reactions with 32P-end-labeled
DNA. Biochemistry 29, 6827-6838.
(40) Mehta, P., Church, K., Williams, J ., Chen, F. X., Encell, L.,
Shuker, D. E., and Gold, B. (1996) The design of agents to control
DNA methylation adducts. Enhanced major groove methylation
of DNA by an N-methyl-N-nitrosourea functionalized phenyl
neutral red intercalator. Chem. Res. Toxicol. 9, 939-948.
(41) Inga, A., Chen, F. X., Monti, P., Aprile, A., Campomenosi, P.,
Menichini, P., Ottaggio, L., Viaggi, S., Abbondandolo, A., Gold,
B., and Fronza, G. (1999) N-(2-Chlorethyl)-N-nitrosourea tethered
to lexitropsin induces minor groove lesions at the p53 cDNA that
are more cytotoxic than mutagenic. Cancer Res. 59, 689-695.
(42) Finneman, J . I., and Fishbein, J . C. (1995) Mechanism of Benzyl
Transfer in the Decay of (E)-Arylmethanediazoates and Aryldi-
azomethanes in Aqueous Solutions. J . Am. Chem. Soc. 117,
4228-4239.
(32) Kresge, A. J ., Chen, H. L., Chiang, Y., Murrill, E., Payne, M. A.,
and Sagatys, D. S. (1971) Vinyl Ether Hydrolysis. III. Some
Bronsted Relations and Transition State Structure. J . Am. Chem.
Soc. 93, 413-423.
TX000084P