N-nitrosotolazoline: Decomposition studies of a typical N-nitrosoimidazoline

Article abstract of DOI:10.1021/tx700318k

N-Nitrosotolazoline (N-nitroso-2-benzylimidazoline), a N-nitrosated drug typical of N-nitrosoimidazolines, reacts readily with aqueous acid, nitrous acid, or N-acetylcysteine to produce highly electrophilic diazonium ions capable of alkylating cellular nucleophiles. The kinetics and mechanism of the acidic hydrolytic decomposition of N-nitrosotolazoline have been determined in mineral acids and buffers. The mechanism of decomposition in acidic buffer is proposed to involve the rapid reversible protonation of the imino nitrogen atom followed by slow general base-catalyzed addition of H₂O to the 2-carbon of the imidazoline ring to give a tetrahedral intermediate, which is also a α-hydroxynitrosamine. Rapid decomposition of this species gives rise to the diazonium from which the products are derived by nucleophilic attack, elimination, and rearrangement. The proposed mechanism is supported by the observations of general acid catalysis, a negligible deuterium solvent kinetic isotope effect (k_H/k_D - 1.15) and ΔS^≠ = -34 eu. In phosphate buffer at 30°C, the half-lives of N-nitrosotolazoline range from 5 min at pH 3.5 to 4 h at pH 6. The main reaction product of the hydrolytic decomposition is N-(2-hydroxyethyl)phenylacetamide. This and other products are consistent with the formation of a reactive diazonium ion intermediate. N-Nitrosotolazoline nitrosates 50 times more rapidly than tolazoline and results in a set of products derived from reactive diazonium ions but different from those produced from the hydrolytic decomposition of the substrate. N-Acetylcysteine increases the decomposition rate of N-nitrosotolazoline by 25 times at pH 7 and results in both N-denitrosation and induced decomposition to produce electrophiles. These data suggest that N-nitrosotolazoline shares the chemical properties of many known direct-acting mutagens and carcinogens.

Full text of DOI:10.1021/tx700318k

308
Chem. Res. Toxicol. 2008, 21, 308–318
N-Nitrosotolazoline: Decomposition Studies of a Typical
N-Nitrosoimidazoline
Richard N. Loeppky*,† and Jianzheng Shi
Department of Chemistry, UniVersity of Missouri, Columbia, Missouri 65211
ReceiVed September 6, 2007
N-Nitrosotolazoline (N-nitroso-2-benzylimidazoline), a N-nitrosated drug typical of N-nitrosoimid-
azolines, reacts readily with aqueous acid, nitrous acid, or N-acetylcysteine to produce highly electrophilic
diazonium ions capable of alkylating cellular nucleophiles. The kinetics and mechanism of the acidic
hydrolytic decomposition of N-nitrosotolazoline have been determined in mineral acids and buffers. The
mechanism of decomposition in acidic buffer is proposed to involve the rapid reversible protonation of
the imino nitrogen atom followed by slow general base-catalyzed addition of H2O to the 2-carbon of the
imidazoline ring to give a tetrahedral intermediate, which is also a R-hydroxynitrosamine. Rapid
decomposition of this species gives rise to the diazonium from which the products are derived by
nucleophilic attack, elimination, and rearrangement. The proposed mechanism is supported by the
observations of general acid catalysis, a negligible deuterium solvent kinetic isotope effect (kH/kD
)
1.15) and ∆Sq ) -34 eu. In phosphate buffer at 30 °C, the half-lives of N-nitrosotolazoline range from
5 min at pH 3.5 to 4 h at pH 6. The main reaction product of the hydrolytic decomposition is N-(2-
hydroxyethyl)phenylacetamide. This and other products are consistent with the formation of a reactive
diazonium ion intermediate. N-Nitrosotolazoline nitrosates 50 times more rapidly than tolazoline and
results in a set of products derived from reactive diazonium ions but different from those produced from
the hydrolytic decomposition of the substrate. N-Acetylcysteine increases the decomposition rate of
N-nitrosotolazoline by 25 times at pH 7 and results in both N-denitrosation and induced decomposition
to produce electrophiles. These data suggest that N-nitrosotolazoline shares the chemical properties of
many known direct-acting mutagens and carcinogens.
Introduction
Tolazoline 1, a nasal decongestant and antihypertensive agent,
is structurally typical of a growing number of imidazoline-based
drugs (1). These substances contain a 2-substituted imidazoline
ring and are secondary amidines, which can generate N-
nitrosoimidazolines upon nitrosation. In the previous publication,
we showed that N-nitrosotolazoline 2, and products derived from
it, are produced from 1 under nitrosation conditions, which vary
from those encountered endogenously in humans to those used
purely for the purposes of laboratory investigation (1). Because
N-nitrosotolazoline is representative of N-nitrosoimidazolines,
which could be produced from the endogenous nitrosation of
imidazoline-based drugs and other substances, we have inves-
tigated the chemical properties of 2 under several conditions
that it is likely to encounter upon endogenous production. We
report here on the kinetics and products of the acid-catalyzed
hydrolysis of N-nitrosotolazoline, on its facile nitrosation, and
on its denitrosation and other transformations resulting from
its reaction with N-acetylcysteine, a model thiol. These studies
have shown that the chemistry of N-nitrosotolazoline is not at
all typical of nitrosamines but is more similar to N-nitrosoam-
ides, N-nitrosoureas, and N-nitrosocarbamates, which are direct-
acting mutagens and carcinogens.
In addition to our prior work on amidine and imidazoline
nitrosation (2–4), there are two important publications by Iley
and co-workers that are relevant to the research being presented
here. In the first study, the kinetics of the acid-catalyzed
hydrolytic decomposition of several N-nitroso-2-arylimidazo-
lines, most particularly N-nitroso-2-phenylimidazoline 3, are
reported (5). While we compare their kinetic data and conclu-
sions with ours later in this paper, it is notable that 3 is reported
to denitrosate to 2-phenylimidazoline 4 and also produce
2-phenyl-1,3-oxazoline 5 from the intramolecular trapping of a
diazonium ion. These types of products are not significant ones
arising from the acid-catalyzed decomposition of N-nitrosoto-
lazoline. In the second study, the Iley group examined the acid-
catalyzed decomposition of N-nitrotolazoline 6, which arises
from the reaction of tolazoline with N2O4 (6). The aqueous acid-
induced decomposition of 2 or 6 results in the same major
product, N-2-hydroxyethylphenylacetamide.
Experimental Procedures
* To whom correspondence should be addressed. Tel: 425-493-1302.
E-mail: LoeppkyR@Missouri.edu.
† Visiting Scholar, Department of Medicinal Chemistry, University of
Washington, Seattle, Washington 98195.
Caution: Most nitrosamines are potent carcinogens. Con-
siderable care should be taken in their use so as to aVert
10.1021/tx700318k CCC: $40.75
2008 American Chemical Society
Published on Web 01/31/2008

N-Nitrosotolazoline Decomposition
Chem. Res. Toxicol., Vol. 21, No. 2, 2008 309
exposure to humans and to aVoid enVironmental contamination.
We routinely perform all operations with these substances,
except for dilute solutions thereof, in well-Ventilated fume hoods.
We rinse all nitrosamine-contaminated glassware with a solution
of concentrated HBr in glacial acetic acid (1:1), which is
effectiVe in cleaVing the NO group form the amine nitrogen
atom. The action of this agent in aprotic solutions is also
effectiVe in nitrosamine destruction. Aqueous solutions are
treated with either Ni(R) or Al in concentrated sodium hydrox-
ide. This process may produce hydrazines.
× 250 mm column employing 0.1% formic acid (A) and
acetonitrile (B) as eluents using a 40 min linear gradient of 90%
A to 90% at a flow rate of 1 mL/min. MS/MS and standards
were then utilized to identify compounds 1, 2, and 7-13.
Kinetics of N-Nitrosotolazoline Decomposition. The kinet-
ics of the decomposition of N-nitrosotolazoline 2 were deter-
mined on a HP diode array UV–visible spectrometer, using
various acids as discussed in the Results and Discussion section.
All reactants were brought to the desired temperature in a
constant temperature bath prior to mixing. Transformations were
conducted at the same temperature. In a typical experiment,
N-nitrosotolazoline solution (2 µL, 8 mM) in CH3OH was
transferred to a cuvette with a syringe and diluted with a
perchloric acid solution containing 1 M NaClO4 to maintain
ionic strength to an initial concentration of 80 µM. The kinetic
decay was monitored by the UV absorbance changes at 259
nm. The observed rate constant kobs was obtained by linear
regression of ln (At - A∞) vs time.
General. The instrumentation and reagents are the same as
noted in the previous paper (1).
Identification of Products from the Decomposition of
N-Nitrosotolazoline 2. N-Nitrosotolazoline (2, 200 mg, 1.06
mmol) was dissolved in 1.0 mL of acetonitrile. To this solution,
20 mL of 0.1 M HCl was added. The mixture was stirred at
room temperature for 12 h and then extracted with CH2Cl2 (15
mL × 3). The combined organic layers were washed with
saturated salt solution, dried over Na2SO4, and concentrated.
The residue was purified by flash chromatography using hexane/
EtOAc (from 4:1 to pure EtOAc) as an eluent. Three products
were obtained as white solids.
N-(2-Hydroxyethyl)phenylacetamide 7 (6). Yield, 70%; mp
59–60 °C. 1H NMR: δ 7.27 (m, 5 H), 6.57 (b, 1 H), 3.76 (b, 1
H), 3.56 (t, J ) 5.3, Hz, 2 H), 3.50 (s, 2H), 3.29 (t, J ) 5.3 Hz,
2 H). 13C NMR: δ 172.4, 134.7, 129.2, 128.7, 127.1, 61.4, 43.3,
42.4. EIMS (relative intensity): 179 (M•+ 8), 161(12), 136 (30),
91 (100).
N-(1-Hydroxyethyl)phenylacetamide 12. Yield, 13%; mp
99–100 °C. 1H NMR: δ 7.34 (m, 5 H), 5.97 (b, 1 H), 5.46 (m,
J ) 6.0 Hz, 1 H), 3.83 (b, 1 H), 3.57 (s, 2 H), 1.26 (d, J ) 6.0
Hz, 3 H). 13C NMR: δ 172.3, 134.1, 129.4, 129.1, 127.5, 71.7,
43.6, 20.8. IR (KBr): 3268 (s), 3064, 1654 (s), 1548 (s), 1135,
1078, 918 cm-1. HRMS calcd for C10H13NO2 + Na+, 202.0844;
found, 202.0839. (See the Supporting Information for the X-ray
crystal structure.) This compound was also prepared for
comparison as follows: A mixture of phenylacetamide (242 mg,
2 mmol), acetaldehyde (114 mg, 2.6 mmol), and K2CO3 (27.6
mg, 0.2 mmol) in 2 mL of toluene was stirred for 6 days. TLC
showed no difference between 2 and 6 days. After purification
by flash column (4:1, v/v EtOAc and hexane), the product was
isolated as a white solid (85 mg, 25%). Its properties were
identical to those of the substance isolated from the decomposi-
tion mixture.
Reaction of N-Acetylcysteine 33 and 2. A series of
phosphate buffers (0.5 M) with different concentrations of
N-acetylcysteine 33 (0, 0.01, 0.1, and 0.5 M, respectively) were
made, and the pH was adjusted to 4.0 or 5.0. Aliquots of
solutions of 2 (in CH3CN) were added to these buffers to bring
the concentration to 2.7 mM. The sample vials maintained at
room temperature (≈23 °C) were placed in an HPLC autosam-
pler holder and injected onto the HPLC every 25 min. The
substrate and product concentrations at each time were deter-
mined from calibration curves of authentic samples. HPLC
comparisons of decomposition mixtures with or without 33 gave
the retention volumes of the new products in the reactions with
33. N-Nitrosotolazoline 2 and 33 (0.5 M) were allowed to react
at pH 5 for 77 min. The new products were isolated by
preparative HPLC. Under these conditions, the yields were as
follows: 7, 36.5%; tolazoline 1, 13.2%; unreacted N-nitrosoto-
lazoline 2, 7%; S-phenylacetyl-N-acetylcysteine 34, 14%; and
(2-phenylacetylamino)ethyl N-acetylcysteine 35, 5%.
1
S-Phenylacetyl-N-acetylcysteine 34. H NMR (CDCl3): δ
7.9 (br s, 1 H), 7.28 (m, 5 H), 6.46 (d, 1 H), 4.67 (dd, 1H),
3.85 (s 2 H), 3.35 (m, 2 H), 1.90 (s, 3 H). 13C NMR: δ 198.02,
172.18, 171.85, 133.01, 129.46, 128.78, 127.64, 52.76, 50.30,
30.30, 22.55. HRMS calcd for C13H15NO4S + H+, 282.0800;
found, 282.0800.
2-Phenylacetylaminoethyl N-Acetylcysteine 35. N-Phenyl-
acetylaziridine 36 was prepared by the procedure reported by
1
Boggs et al. (9). For 36: H NMR (CDCl3): δ 7.22 (m, 5 H),
3.67 (s, 2 H), 2.10 (s, 4 H). 13C NMR: δ 183.3, 134.4, 129.2,
128.5, 126.9, 44.0, 24.2. N-Acetylcysteine 33 (0.163 g, 1 mmol)
and 36 (0.162 g, 1 mmol) were dissolved in 4 mL of DMF and
heated at 70 °C for 4 h. The reaction mixture was evaporated
to dryness and chromatographed to give 35 as a white solid
(0.23 g, 71% yield); mp 118 -119 °C. 1H NMR (in CDCl3): δ
7.32 (m, 5 H), 6.42 (b, J ) 4.5 Hz, 1 H), 5.88 (b, 1 H), 4.74
(td, J ) 8.8, 4.5 Hz, 1 H), 4.25 (m, 2 H), 3.58 (s, 2 H), 3.50
(m, 2 H), 2.90 (t d, J ) 8.8, 4.6 Hz, 2 H); 2.05 (s, 3 H), 1.34
(t, J ) 8.7 Hz, 1 H). 13C NMR: δ 171.4, 170.2, 170.0, 134.6,
129.5, 129.0, 127.5, 64.6, 53.8, 43.7, 38.5, 26.5, 23.1. HRMS
(FAB): m/z calcd for C15H20N2O4S + H, 325.1222; found,
325.1223.
N-Vinylphenylacetamide 10 (7). Yield, 2–3%; mp 81–82
1
°C. H NMR: δ 7.38 (m, 5 H), 6.96 (dd, J ) 15.6, 8.2 Hz, 1
H), 6.93 (b, 1 H), 4.48 (dd, J ) 15.6, 2.4 Hz, 1 H), 4.37 (dd,
J ) 8.2, 2.4 Hz, 1 H), 3.64 (s, 2 H). 13C NMR: δ 168.3, 133.9,
129.5, 129.2, 128.4, 127.7, 95.7, 43.6. IR (KBr): 3231, 3137,
3027, 1634 (s), 1511, 1262, 1192 cm-1
.
2-Aminoethyl Phenylacetate 8. In a separate experiment,
the acidic decomposition mixture was extracted with several
portions of ethyl acetate to remove the compounds described
above. The aqueous fraction was concentrated to 2 mL under
vacuum. Acetone was slowly added to this solution until the
formation of a white precipitate was complete. This compound
(8) was the hydrochloride of 8 (18% yield). 1H NMR (in
DMSO-d6): δ 8.33 (b, 3 H), 7.28 (m, 5 H), 4.24 (t, J ) 5.2 Hz,
2 H), 3.74 (s, 2 H), 3.07 (t, J ) 5.2 Hz, 2 H). 13C NMR: δ
171.1, 134.1, 129.5, 128.3, 126.8, 60.7, 40.0, 37.7.
Results and Discussion
LC-MS/MS of Reaction Mixture. In a separate set of
experiments when the decomposition was conducted either as
described above or in aqueous acetic acid, the reaction mixture
was subjected to LC-MS/MS using a Zorbak SB C-8 4.6 mm
Kinetics and Acid Catalysis. As can be seen in Figure 1,
the decomposition of N-nitrosotolazoline 2 is acid catalyzed by
HClO4. At 30 °C in phosphate buffer, the half-lives of 2 range
from 5 min at pH 3.5 to 4 h at pH 6 (data not shown). Rate

310 Chem. Res. Toxicol., Vol. 21, No. 2, 2008
Loeppky and Shi
Figure 3. Plot of kobs/T vs 1/T (K) over the temperature range 35–43
°C to obtain the ∆Hq and ∆Sq is shown.
Figure 1. Observed rate constants for the decomposition of N-
nitrosotolazoline in HClO4 containing 1.0 M NaClO4 at 37.1 °C are
plotted against [H+]. The slope as determined by linear regression gives
kH ) 10.1 ( 0.3 s-1 M-1
.
Figure 4. Observed rate constants for the decomposition of N-
nitrosotolazoline in either formic acid (solid lines) or acetic acid (dashed
lines) buffers containing 1.0 M NaClO4 at 37.1 °C are plotted against
[buffer]. The slopes and intercepts were used to determine kHA and
kH, respectively, as described.
Figure 2. Observed rate constants for the decomposition of N-
nitrosotolazoline in either HClO4 or DClO4 containing 1.0 M NaClO4
at 25 °C are plotted against [H+]. The slopes were used to calculate
kH/kD ) 1.15 ( 0.15.
We also determined reaction rates in acetate and formate
buffers at 37 °C over a pH range of 3–6 to further explore the
mechanistic details of the transformation. The data are presented
graphically in Figure 4. Plots and linear regression of kobs vs
[total buffer] at each pH gave the slopes (kA) listed in Table 1.
The values of kA so determined were then plotted against fHA
{fHA ) [H+]/(KA + [H+])} from which kHA can be determined
at fHA ) 1 for each buffer. The intercepts of the kobs vs [total
buffer] plots were, in turn, plotted against [H+] to give slopes
) kH. These data analyses show that there is buffer catalysis
and that the transformation is subject to some form of general
acid catalysis. While the buffer catalysis is somewhat modest
and could be attributed to media effects, we carefully controlled
the ionic strength. Accepting the buffer catalysis as a valid
experimental observation, we can express kobs as shown in eq
1, where BT is the [total buffer] and kw (in formic acid buffer,
constants, the average of at least two determinations, were
determined by following the transformation spectrophotometri-
cally and plotting the ln of the absorbance change vs time.
Isosbestic points were observed. As can be seen from Figure 1,
the rate constants increase linearly with [H+] in HClO4 at 37
°C giving kH (the slope of the plot) ) 10.1 ( 0.3 s-1 M-1. As
shown in Figure 2, experiments in DClO4 and its protio
counterpart revealed that this acid catalysis is subject to only a
modest to insignificant deuterium isotope effect (kH/kD ) 1.15
( 0.15), the meaning of which is discussed below. The
activation parameters for the decomposition of N-nitrosotola-
zoline in HClO4 (0.73 mM 2, 1 M NaClO4) were determined
by means of rate measurements over a modest temperature
range. A plot (Figure 3) of ln (k/T) vs 1/T gave a slope and
intercept from which ∆Hq (10.77 ( 0.8 kcal/mol) and ∆Sq
[-33.8 ( 2.5 cal/mol/deg (eu)] could be calculated.
0.8 ( 0.3 × 10-3 s-1; in acetic acid buffer, 0.0 ( 0.3 × 10-3
)
is the rate constant for H2O. The zero intercepts of the kA vs
fHA plots represent base catalysis and are not significant for either
acid (formic, 0.4 ( 0.4 × 10-3 s-1 M-1; acetic, -0.2 ( 0.1 ×

N-Nitrosotolazoline Decomposition
Chem. Res. Toxicol., Vol. 21, No. 2, 2008 311
Table 1
Scheme 2
s-1 M-1 × 10-3
a
pH
kA
error
formic acid
3.72
3.15
3.48
3.75
4.02
4.35
0.41
0.53
0.34
0.16
0.12
0.70
2.50
2.37
2.06
1.03
b
kHAc
4.3
kH
12930
825
acetic acid
0.71
0.69
0.42
0.098
1.41
4.56
4.74
4.92
5.34
0.039
0.046
0.011
0.003
0.25
b
kHAc
kH
27210
1130
a Determined from the slopes of linear regression of kobs at each pH
vs [total buffer]. b Determined from the linear regression equation of ka
vs fHA {[H+]/(KA + [H+])} at fHA ) 1. c Determined from the slope of
linear regression of intercepts of regression of footnote a vs [H+].
cation of the aqueous extract of the decomposition mixture and
extraction with CH2Cl2 gave 7, not 8 (by HPLC). In a separate
experiment, the acidic decomposition mixture (0.01 M HCl) was
extracted with ethyl acetate and the aqueous mixture was
concentrated under high vacuum. Trituration with acetone
produced a white solid (8 as the hydrochloride salt), which
cochromatographed with 8 in the acidic reaction mixture.
Basification of this salt even at pH 8 led to its conversion to 7
by O- to N-acyl migration, most likely through an intramolecular
rearrangement.
Scheme 1
Compounds 7-9, 11, and 13 are all known compounds and
1
were identified and characterized in our laboratory by H and
13C NMR, IR, and MS spectroscopic methods in comparison
with the authentic materials that we prepared. The independent
synthesis and characterization of the N-vinyl amide 10 have
been published (7) since the inception of our work. Our 1H and
13C NMR spectra are identical to those published for this
compound. The structure of the amide-hemiaminal, 12, was
initially deduced on the basis of its NMR and IR data, all of
which are completely consistent with its assigned structure.
However, because we anticipated that such a compound might
not survive our reaction and isolation procedures, we confirmed
our assignment by X-ray crystallography (see the Supporting
Information).
The relative yields of the products vary somewhat depending
upon the composition of the acid mixture. In additon to the
experiments in perchloric acid and phosphate buffer, we made
HPLC comparisons of the product profile from reactions in
0.0012-0.12 M HCl and 0.12 M HOAc. The yields of 7 and 8
decrease in comparison with other products, some of which were
not identified, when the hydrolytic decomposition is carried out
in acids with more nucleophilic or basic anions, but the
differences between the HCl and the HOAc decomposition
mixtures were small. In addition to 7, the yields of which are
always >50%, the other major products are 8 (8–10%), 10
(3–5%), 12 (5–6%), and unreacted N-nitrosotolazoline (5–6%).
In acetic acid, the acetate ester 9 formed in 5%, but the yield
of chloride 11 rarely exceeded 3% in the HCl-mediated
reactions. Regardless of the acid, the denitrosation product
tolazoline 1 yield stayed at about 2%. 2-Benzyl-1,3-oxazoline
(10) 13 was only a trace component of the mixture with yields
<0.7%.
10-3 s-1 M-1). The mechanistic significance of these data is
discussed below.
kobs ) kH[H+] + kHA fHABT + Kw
(1)
Decomposition Products. N-Nitrosotolazoline was allowed
to decompose in 0.1 M HCl at room temperature, and the
resulting aqueous mixture was analyzed by HPLC and extracted
with CH2Cl2 to isolate the products. HPLC chromatograms of
decomposition mixtures produced in HClO4, HCl, and phos-
phoric acid or acetic acid buffer solutions were also compared.
Flash chromatography of the CH2Cl2 extract led to the isolation
and characterization of products 7, 10, and 12, but not 8 (see
Scheme 1). N-2-Hydroxyethylphenylacetamide 7 is the major
product in all cases and was isolated in 70% yield from the
HClO4-catalyzed reaction. A comparison of the HPLC chro-
matograms of the CH2Cl2 extract and the original reaction
mixture showed that 8 was not extracted into CH2Cl2. Basifi-
The nature of the products, the relatively large number of
minor products, and prior work all permit the reasonable
conclusion that all of the products except tolazoline can be
rationalized as arising from the diazonium ion 14 (Scheme 2).

312 Chem. Res. Toxicol., Vol. 21, No. 2, 2008
Loeppky and Shi
Table 2. Comparison of Product Yields from the Hydrolysis
of 2 and 13a
is 10 min and is 4.5× faster than that of 2 under the same
conditions. Thus, rate studies do not rule out the formation of
7 through the hydrolysis of the oxazoline 13 in the hydrolytic
decomposition of 2, but the product studies do (see Scheme 3).
The hydrolysis of 13 gives only minor amounts of 7 (Table 2),
whereas 7 is the major product from 2. In phosphate buffer,
17-23% of the diazonium ion 14 undergoes intramolecular
trapping by path D (Scheme 2) as estimated from the yield of
8. However, the yields of 8 are much lower when the
decomposition occurs in HOAc, HCl, or HClO4.
substrate
N-nitrosotolazolineb
2
2-benzyl-1,3-oxazolinec 13
product/pH 3.5
4.5
5.5
3.5
4.5
5.5
7
8
69%
23%
71%
20%
70%
17%
1.8%
95%
1.8%
95%
2.2%
95%
a All hydrolyses were conducted in 0.5 M phosphate buffer at 37 °C.
b The initial concentration of 2 was 4 mM. c The initial concentration of
13 was 4.3 mM.
Scheme 3
The protonated oxazoline 16 (Scheme 3) will be attacked at
the C-2 to generate a tetrahedral intermediate 17, which is in
equilibrium with at least two other forms, 18 and 19. Additional
protonation states for these intermediates could affect the
outcome, but the important fact is that 8 is the major hydrolysis
product by far. The outcome cannot be easily rationalized of
+
the basis of leaving group ability, NH2 from 18 or one of its
conjugate bases vs OH+ from 19 (or a conjugate base), because
the latter is a far better leaving group. On the other hand, because
of the basicity of the nitrogen atom, 18 will be the most
populated of three tetrahedral intermediates (17–19) shown.
Thus, equilibria could drive the process toward the formation
of 8, as is observed. Consideration of the chemistry depicted in
Scheme 3 also gives us a frame of reference for understanding
why the hydrolytic decomposition of N-nitroso-2-phenylimida-
zoline results in the isolation of the oxazoline, but the hydrolysis
of N-nitrosotolazoline does not. 2-Benzyloxazoline 13 must
hydrolyze much more rapidly than 2-phenyloxazoline. This is
likely due to the conjugative stabilization of the protonated
2-phenyloxazoline by the aromatic ring adjacent to the iminium
ion. Thus, either the nucleophilic attack of H2O to form the
requisite tetrahedral intermediate is slower or its equilibrium
concentration is lower than what is produced from 16.
Mechanism of Diazonium Ion Formation. Our kinetic data
and literature precedent allow conclusions to be made regarding
the hydrolytic decomposition mechanism of N-nitrosotolazoline.
It is useful to compare our data with those of Iley et al. (5, 6)
and consider their mechanistic conclusions. The comparisons
are summarized in Table 3. In the case of 3, kH/kD is near what
is expected for a transformation that proceeds with equilibrium
protonation followed by slow, unassisted attack of H2O to form
a tetrahedral intermediate, which decomposes rapidly. Thus, the
isotope effect, the specific acid catalysis, and the negative
entropy of activation were cited by Iley et al. (5) to support
their mechanistic conclusion. The observation of curved plots
of kobs vs [H+] for the hydrolytic decomposition of 6 permitted
a separation of kH/kD for the first two steps (6). The nitro
substituent markedly decreases the basicity of 6, and the
observation of kH/kD ) 0.3 is consistent with its equilibrium
protonation. The conjugate acid so formed was judged to rapidly
react with H2O to generate a tetrahedral intermediate, which is
perceived to decompose by a slow cyclic intramolecular transfer
of the OH proton to the nitroso O atom with a kH/kD ) 1.7.
This conclusion was necessitated by the lack of general acid
catalysis and the observation of the large negative entropy of
activation associated with a very ordered transition state. This
conclusion is also reasonable from a structure-electronic
perspective. Because of the powerful electron-withdrawing
characteristics of the N-nitro group, the 2-carbon of the conjugate
acid of 6 should be very susceptible to nucleophilic attack by
H2O, although little of the conjugate acid will be present at
modest acidities.
Primary diazonium ions decompose rapidly by SN2 substitution,
elimination, and rearrangement pathways. Nucleophilic attack
of H2O on 14 (Scheme 2) gives the major product 7 (path A).
Iley et al. (6) reported that N-nitrotolazoline hydrolyzes in strong
acid to produce 7, which they envisioned to arise through the
displacement of N2O from an intermediate like 14. The acetate
9 and chloride 11 are produced by similar nucleophilic displace-
ments. Elimination by path B gives 10. Rearrangement by
hydride migration as nitrogen departs gives the resonance-
stabilized carbocation 15, which then gives 12 by attack of H2O
as depicted in path C (Scheme 2). Intramolecular attack of the
carbonyl oxygen atom on the diazonium ion carbon gives 16
and then 13 by deprotonation, but this compound is formed in
less than 1% from N-nitrosotolazoline. In contrast, Iley et al. found
that 2-phenyloxazoline and the denitrosation product 2-phenyimi-
dazoline were the major hydrolytic decomposition products of
N-nitroso-2-phenylimidazoline (5). These authors proposed that the
oxazoline formed by a pathway identical to path D of Scheme 2.
The amino ester 8 can be rationalized to form from the hydrolysis
of the protonated oxazoline 16 (Scheme 2); yet, this pathway (D)
is not a major competitor to path A in our system. Below, we
discuss another possible route to 8.
In phosphate buffer, the hydrolytic decomposition of N-
nitrosotolazoline gives only two significant products, 7 and 8
(see Table 2 for yields). If the major pathway involves the
intramolecular nucleophilic displacement of N2 from 14 (path
D, Scheme 2), it is possible that 7 as well as 8 arise from the
hydrolysis of the oxazoline 13 as shown in Scheme 3 rather
than by path A (Scheme 2). The distinction between these
pathways has toxicological significance because the intramo-
lecular trapping (path D, Scheme 2) can be viewed as the
disarming of a potent electrophile. Intramolecular displacement
of N2 from diazonium ions derived from some cyclic nitro-
samines is well-documented (11–17), but these transformations
yield cations (related to hemiacetals) that can covalently bind
DNA in contrast to the formation of the stable oxazoline 13
being considered here. To clarify this point, we prepared
2-benzyloxazoline 13 and compared both its rate of hydrolysis
in acid and its hydrolysis products with those arising from
N-nitrosotolazoline. The t1/2 for the hydrolysis of 13 at pH 5.5
We propose that N-nitrosotolazoline is undergoing hydrolytic
decomposition to a diazonium ion by the mechanism depicted

N-Nitrosotolazoline Decomposition
Chem. Res. Toxicol., Vol. 21, No. 2, 2008 313
Table 3. Comparison of Key Data and Acid Catalyzed Decomposition Mechanisms for 2, 3, and 6
Scheme 4
kH/kD values. This could be the case if the attack of H2O on 20
to give the uncharged tetrahedral intermediate 22 is assisted as
shown in 21 by proton transfer from the attacking water
molecule to the general base B. The modest nature of the buffer
catalysis shows that some, but minor, base assistance occurs in
the attack of the H2O molecule on the cation 20. Ample literature
precedent shows that this step will have a kH/kD value of 3 or
greater (18). Overall, the reaction is specific acid-general base-
catalyzed, which is kinetically equivalent to general acid
catalysis. This association mechanism involves the loss of a
number of degrees of freedom and will have a large negative
∆Sq (-34 eu), as is observed.
In addition to being a tetrahedral intermediate, so familiar in
carbonyl group transformations, 22 is also an R-hydroxynitro-
samine. Fishbein et al. (17, 19, 20) have carefully studied the
catalytic decomposition of these unstable intermediates, which
are also the immediate primary metabolic products of nitro-
samine carcinogens. The facile decomposition of R-hydroxyni-
trosamines is subject to acid and base catalysis. 2-Hydroxy-N-
nitrosopyrrolidine has a t1/2 of only 0.03 s at pH 7. General
acid-catalyzed hydrolysis is observed in the acid region. Careful
analysis of the experimental data led the Fishbein group to
conclude that the transformation occurred by equilibrium
protonation of the N-nitroso O-atom followed by general base-
assisted transfer of the R-hydroxy H -atom in the slow step to
generate a diazotic acid and an aldehyde (17, 19, 20). We show
this path in Scheme 4 for the conversion of 22 to 23 through
22-A. Although we cannot be certain, we believe that this step,
regardless of whether it proceeds as depicted or by specific
acid-general base catalysis, will not have a large negative ∆Sq.
The opening of the ring involves an increase in the degrees of
freedom. Transformation through the cyclic transition state
shown (22-T) may be associated with a significant negative ∆Sq,
but this process will not exhibit general acid catalysis. Thus,
for our system, we believe that the data best support specific
acid-general base formation of 22 in the rate-determining step
followed by its rapid decomposition to the reactive diazonium
ion 14 as shown or by one of the similar pathways just discussed.
in Scheme 4 where the transition state for the rate-determining
step is shown as 21. This pathway accounts for three notable
experimental observations: general acid catalysis, a very modest
to insignificant solvent kinetic deuterium isotope effect (SKIE),
and the very large negative value of ∆Sq. N-Nitrosotolazoline
2 is certainly much less basic than tolazoline but more basic
than 3 because the electron-donating benzylic alkyl group at
the 2-position enhances basicity as compared to the phenyl group
of 3. N-Nitrotolazoline 6 may be less basic than 2, but this
assessment is difficult because both the NO and the NO2 groups
are powerful electron-withdrawing moieties. Like 3, it is
probable that 2 undergoes reversible protonation to 20 (Scheme
4). This step should exhibit a kH/kD ≈ 0.3 because D3O+ can
be viewed as a stronger acid than H3O+ (18). Yet, we observe
no such SKIE. Conversely, if the protonation of 2 were rate
determining, then the SKIE should be ≈3. The observed modest
SKIE (1.15) must result from the product of at least two different

314 Chem. Res. Toxicol., Vol. 21, No. 2, 2008
Loeppky and Shi
Scheme 5
N-nitrosation may determine the reaction course, which we now
think more probable. The N-nitrosation of the N-CH3 moiety
(30, Y ) NO) will convert it into a much better leaving group,
the incipient nitrosamine, than the protonated amine group (30,
Y ) H). This could determine the course of the reaction as
shown. (The RNNO- anion has been shown to be a poor leaving
group (21), but we expect the RCH3NNO+ group of 30 (Y )
NO) to be a good leaving group based on the thermodynamic
stability of the incipient nitrosamine functional group.)
The partitioning of an intermediate akin to 22 (Scheme 5) or
29 (Scheme 6) with respect to decomposition paths b and c has
been the subject of a recent computational study (22). The
R-hydroxynitrosamine tetrahedral intermediates (various con-
formations and nitrosamine stereoisomers) were presumed to
arise through the nitrosation of imidazoline itself by a process
similar to that shown for the formation of 29 in Scheme 6. The
computations showed, of the two pathways considered (not
strictly b or c above), that the outcome was determined by the
stereochemistry of the N-NdO group in the R-hydroxynitro-
samine tetrahedral intermediate akin to 22 or 29. The retro-ene
decomposition route (22-T of Scheme 4) is the lowest energy
path (∆Gq ) 12.1 kcal/mol) and is only accessible to the
stereoisomer where the N-NdO group is syn to the OH. The
syn-anti interconversion barrier between the R-hydroxynitro-
samines is large (∆Gq ≈ 19 kcal/mol in either direction) as
compared to the energetics of the retro-ene reaction. Conversion
of the R-hydroxynitrosamine tetrahedral intermediate to the
N-nitrosoamide by C-N cleavage (bond c) was energetically
not feasible in the gas phase but did become within reach when
the “microsolvation” provided by one H2O molecule located
between the OH and the N of the amine bond being broken
was included in the transition state. The activation energy for
the conversion of the anti R-hydroxynitrosamine to the N-
nitrosoamide by this pathway was computed to have a ∆Gq )
17.3 kcal/mol, which is only 1.9 kcal/mol lower in energy than
the syn-anti isomerization barrier. Although there would be some
“leakage” across the syn-anti isomerization barrier, the anti
R-hydroxynitrosamine was predicted to give rise to the N-
nitrosoamide to the extent of 95% at 298 K. The activation
energy (∆Gq) for the similar conversion of the syn R-hydroxy-
nitrosamine to N-nitrosoamide was computed to be slightly more
favorable (16.5 kcal/mol) but large as compared to the retro-
ene cleavage of this isomer.
Scheme 6
The R-hydroxynitrosamine tetrahedral intermediate 22 can,
in theory, decompose by three competing paths, a, b, or c,
through acid catalysis as shown in Scheme 5. These paths
basically differ in the nature of the leaving group, which we
assume to be derived from a protonated intermediate: Path a,
H2O; path b, O-protonated R2NNdO; and path c, RNH2. Path
a converts 22 back to the conjugate acid of N-nitrosotolazoline.
In the case of imidazoline hydrolysis, this is the preferred path
as the equilibrium lies on the side of the imidazoline. The
cleavage of the other C-N bond (path c) will give the
N-nitrosoamide 24. As discussed above, we do find 8, which is
expected to be one of the decomposition products of 24 through
25, but if this pathway were operative, we would expect to
isolate 24 and the other decomposition products produced from
it, particularly phenylacetic acid. We previously reported that
the nitrosation of 1-methyl-2-phenylimidazoline 26 gives the
N-nitrosoamide 27 as shown in Scheme 6 (3). This work shows
that N-nitrosoamides akin to 24 can be isolated from acidic
nitrosating media. The conversion of 26 to 27 is logically
envisioned to occur through the R-hydroxynitrosamine tetra-
hedral intermediate 29 (Scheme 6), which is similar to 22
(Schemes 4 and 5). In our prior interpretation (3), we proposed
that the transformation occurs through reversible protonation
of 29 to give 30 (Y ) H, Scheme 6) that then decomposes as
shown by transformations akin to path c of Scheme 5 to give
27. We did not consider the possible generation of 30, Y )
NO. Although the conditions are different for the hydrolysis of
2 as compared to the acidic nitrosation of 26, it would
superficially appear that 22 (in acid) and 29 (in acid plus nitrous
acid) behave differently with respect to acid-catalyzed decom-
position, 22 taking path b and 29 proceeding by cleavage of
the other C-N bond (c). While this may be the case, our more
recent work on amidine (2), and particularly tolazoline nitro-
sation (1), show that we must also consider that a second
Although we do not know the stereochemistry of N-
nitrosotolazoline with certainty because only one isomer is
present, it is likely the anti (E) isomer. The N-NdO stereo-
chemistry of nitrosamines is principally determined by steric
factors and the NO group should prefer to be anti to the benzyl
group of tolazoline. Assuming that this stereochemical relation-
ship is retained during the addition of H2O across the CdN,
then the anti R-hydroxynitrosamine should be formed and a
superficial extension of computational results would appear to
predict the formation of the N-nitrosoamide 24 as the major
product of the hydrolytic decomposition of 2. We do not observe
this outcome. There are several probable explanations for the
difference. The protonation of N-nitrosotolazoline produces the
same intermediate 20 that arises from the nitrosation of
tolazoline. The computational modeling of the nitrosation of
imidazoline, referred to above, yields a K ) [Z]/[E] ≈ 0.01
and a kinetic barrier of ∆Gq < 10 kcal/mol, giving rise to a
very rapid interconversion at 25 °C (see Scheme 7). In the case
of 20, K ) [20-Z]/[20-E] is likely 0.001 or less, but the barrier
is probably not much different. Because the syn-anti (E-Z)
interconversion is rapid as compared to the additon of H2O to

N-Nitrosotolazoline Decomposition
Scheme 7
Chem. Res. Toxicol., Vol. 21, No. 2, 2008 315
generate the R-hydroxynitrosamines 22-E and 22-Z, application
of the Curtin-Hammett principle shows that the preference for
22-Z is given by KkZ/kE. Thus, if K ) 0.001, then kZ must be
1000 × kE to achieve a 1:1 mix of the two stereoisomeric
R-hydroxynitrosamines. In other words, ∆Gq for the generation
of 22-Z at 25 °C would have to be 4 kcal/mol less than that for
the formation of 22-E from 20-E. The computational study did
not calculate transition state energies for the H2O addition
reaction, so we have no computational model against which this
argument can be tested. While kZ may be less than kE because
of the repulsive effect of the lone electron pair on the nitroso
N, the experimental work of the Fishbein group (17, 19, 20)
provides us with a more probable explanation for the differences
between our experimental observations and the computational
predictions.
Figure 5. Yields of N-2-hydroxyethylphenylacetamide 7 produced from
N-nitrosotolazoline 2 are compared as a function of reaction conditions
in 0.5 M acetate buffer at pH 3.4 at 37 °C. Nitrosation significantly
reduces the yield of 7, the normal hydrolysis product, which is not a
nitrosation product.
-
faster than 1 in buffered HOAc containing 10 equiv of NO2
at 37 °C. The % yield of 7 as a function of time can be used to
probe the competition between the hydrolysis of 2 and its
nitrosation as is shown in Figure 5. Addition of 1 equiv of nitrite
to the acidic mixture reduces the final yield of 7 from 50 to
As discussed above, the Fishbein group has observed general
acid catalysis for all of the R-hydroxynitrosamine decomposi-
tions that they have investigated, which includes cyclic and
acyclic substrates (17, 19, 20). As shown in Scheme 4, they
have interpreted this to involve specific acid-general base
catalysis. The retro-ene process depicted in 22-T of Scheme 4,
which was assigned by the computation study (22) as the
decomposition pathway of the syn (Z) isomer, is not subject to
acid or base catalysis. Therefore, the catalyzed Fishbein pathway
22-Z must have a lower ∆Gq than the retro-ene process, else
acid/base catalysis would not be observed. It is also likely that
the anti or E isomer also decomposes by the Fishbein pathway.
These pathways were not subjected to computational modeling
(22). Thus, while it is possible that N-nitrosoamide formation
does compete with the Fishbein pathway (17, 19, 20), we believe
that 8 mainly arises in our decomposition mixture as shown in
Scheme 3. This pathway has precedent in the work of Iley et
al. (5), and we observe little to no phenylacetic acid, a coproduct
expected if 24 forms and decomposes. Still, the fact that the
nitrosation of 26 only gives the N-nitrosoamide 27 (Scheme
6), a process predicted by the computational work so far, shows
that subtle factors easily change the chemistry of this system.
Nitrosation of N-Nitrosotolazoline. As discussed in the
previous paper (1), we were unable to isolate significant
quantities of N-nitrosotolazoline 2 from the nitrosation of
tolazoline. The tolazoline nitrosation products were consistent
with their actually being products of the nitrosation of N-
nitrosotolazoline. Our preparative route to 2 involves its
continuous CH2Cl2 extraction from the nitrous acid solution.
Affirmation of our hypothesis that 2 nitrosates much more
rapidly than tolazoline 1 was aided by our finding that the amide
7 is the main product from the acid-catalyzed decomposition
of 2 but is not produced from the nitrosation of 1. The amide
7 is stable toward our nitrosation conditions. Although we found
it difficult to compare the respective nitrosation kinetics of 1
and 2, we estimated on the basis of half-lives, under the stated
conditions of each transformation, that 2 nitrosates at least 50×
-
35%, and 10 equiv of NO2 limit the final yield of 7 to 9%.
The much more facile nitrosation of 2 is undoubtedly due to its
much lower basicity than 1, since it is the free base that is
nitrosated.
Reaction of N-Acetylcysteine with N-Nitrosotolazoline.
With the exception of diarylnitrosamines, most nitrosamines do
not undergo denitrosation at a measurable rate unless the reaction
is carried out in very strong mineral acid (23, 24). Under these
conditions, halide ions, thiocyanate, and several other sulfur
compounds are effective catalysts of this transformation. On
the other hand, N-nitrosoamides, -ureas, and -urethanes, which
have a carbonyl bound adjacent to NNO group, denitrosate with
much greater facility (25, 26). As discussed above, Iley et al.
reported that N-nitroso-2-phenylimidazoline denitrosates sig-
nificantly under the conditions of its acid-catalyzed hydrolysis
(5). This is not true of 2, but glutathione and thiol-induced
denitrosation transformations can be modulators of the carci-
nogenicity and toxicity of some N-nitroso compounds. For
example, N-nitrosocimetidine is produced by the gastric nitro-
sation of the antiulcer drug cimetidine (27). However, studies
showed that this N-nitrosocyanoguanidine was rapidly denitro-
sated by thiols including those contained in blood-circulating
proteins (27). On the basis of these observations, we thought it
important to perform a preliminary examination of the interac-
tion between 2 and N-acetylcysteine 33.
In phosphate buffer at pH 5, 25 °C, N-acetylcysteine 33
diminishes the t1/2 of N-nitrosotolazoline 2. As the concentration
of 33 is increased from 0 to 0.01, 0.1, and 0.5 M, the respective
half-lives (min) of 2 are as follows: 36.5, 33.0, 28.7, and 19.6.
A more dramatic change is observed at pH 7 as shown in Figure
6. Under these conditions (pH 7, 25 °C), N-nitrosotolazoline in
a 0.1 M acetate containing solution (control) has a t1/2 of 213 h
but only a t1/2 of 8.15 h in a 0.1 M solution of 33. The major
decomposition product, even in the presence of 500 mM 33, is
the diazonium ion-derived amide 7 (see Scheme 3). At pH 5,
concentrations of 33 below 100 mM have little effect on the

316 Chem. Res. Toxicol., Vol. 21, No. 2, 2008
Loeppky and Shi
Table 4. Relative Product Yields from the Reaction of 2
with 33
compound
relative %a
1
8
21.5
5.1
7
38.8
4.2
20.7
4.2
1.7
3.7
12
34
35
10
2
a Determined by HPLC from the reaction of 2 (2.47 mM) in the
presence of 0.5 M 33 in pH 4 phosphate buffer at 25 °C after 75 min.
(34) was synthesized by reaction 33 with phenylacetyl chloride.
We anticipated that 33 would scavenge electrophiles produced
in the decomposition of 2, principally the diazonium ion 14,
and expected the sulfide 37 (Scheme 8) to be a product. The
NMR spectral properties of compound 35 are inconsistent with
it being assigned the structure 37. Both the 1H and the 13C NMR
spectra of the imidazoline-derived fragment exhibit chemical
shifts very close to those observed for 7, which has the same
atom connectivity. If the terminal CH2 were attached to S as in
structure 37, then this carbon and its attached protons would
Figure 6. Effect of 0.1 M N-acetylcysteine 33 at pH 7 on the
concentration of N-nitrosotolazoline 2 (initial, 1.8 mM) as a function
of time is shown. The control experiment shows the concentration of
2 under the same conditions except that the N-acetylcysteine was
replaced with sodium acetate. Tolazoline 1, the denitrosation product,
increases as 2 decreases but accounts for slightly greater than 50% of
the products derived from 2. The % of the others, arrived at by adding
the concentrations of 1 and 2 and subtracting the sum from 100,
increases at the same rate and shows that 33 induces the decomposition
of 2.
1
be found at higher field. As it is, the 13C and H for this CH2
are at δ 64.51 and δ 4.25, respectively, positions clearly
consistent with a CH2-O group. This compound also exhibits
at triplet at δ 1.34, which we have assigned to the SH. As shown
in Scheme 8, 35 was synthesized by the reaction of 33 with
N-phenylacetylaziridine 36. Although this is not an unambiguous
synthesis, the spectral characteristics of 35 are clearly in accord
with its structural assignment.
Scheme 8
The data shown in Figure 6 clearly show that 33 induces the
decomposition of 2 as compared to the control. The t1/2 drops
by an estimated factor of 25. The denitrosation product 1
represents only slightly more than 50% of the product mixture
under these conditions. The product studies at pH 4 and 5 are
also indicative of induced decomposition but, as compared to
the controls and the effects, are greater at higher pH, which
suggests the involvement of a base-catalyzed process. At pH 4,
the yield of 1 is only 21.5% but the yield of 34 is 20.7% (see
Table 4). Thus, 33 must be inducing the decomposition of 2 as
well as denitrosating it by thiol attack on the NO group.
Moreover, the structure of 34 suggests that it has been formed
by attack of the thiol group of 33 at the CdN of 2. Thiol
induction of the decomposition and denitrosation of N-methyl-
N′-nitro-N-nitrosoguanidine (MNNG) has been reported by
several groups. These processes are well-reviewed and cited in
the detailed mechanistic work of the Fishbein group (26). A
possible route, which could involve acid or base catalysis, is
shown in Scheme 9 and is broadly consistent with the Fishbein
mechanism (26), although they did not work in the acidic pH
range. A reactive diazonium ion 39 is formed by decomposition
of the tetrahedral intermediate 38. While it can participate in
various inter- and intramolecular transformations, the simplest
process envisions its hydrolytic conversion to 40, which then
hydrolyzes to give the thioester 34 and ethanolamine 41, which
would escape detection by our analytical method. While 35 can
also be rationalized to arrive from 39, it probably comes from
a reaction of the diazonium ion 14 produced by the acid-
catalyzed decomposition of 2 with 33. The thiol is the most
nucleophilic moiety in 33, but it will not be in its more
nucleophilic thiolate form at the pH of the product study, while
a significant fraction of the carboxyl groups will be present as
the carboxylate anions. Carboxylate attack on the diazonium
carbon of 14 provides some rationale for the surprising formation
composition of the product mixture. N-Acetylcysteine (500 mM,
pH 5, 25 °C, 77 min) reduces the yield of 7 from 47 (no 33) to
36%, while the yield of tolazoline increases from 0 to 13%. A
comparison of the HPLC chromatograms of the pH 5 reaction
mixtures with those from the acid-catalyzed hydrolysis of 2
showed that the diminished t1/2 of 2 was due to appearance of
three new reaction products. One of these was the denitrosation
product tolazoline 1. The other two compounds, 34 and 35 (see
Scheme 8), were isolated by chromatography and characterized
by NMR, MS, and ultimately synthesis. Relative product yields
at pH 4 are given in Table 4. The phenylacetyl thio-ester of 33

N-Nitrosotolazoline Decomposition
Scheme 9
Chem. Res. Toxicol., Vol. 21, No. 2, 2008 317
representative of N-nitrosoimidazolines, results in the production
of potent cell-damaging electrophiles. Even if these nitroso
compounds are not formed in the acidic stomach but, for
example, at a site of chronic inflammation where endogenous
nitrosation occurs, such as a chronically infected sinus, our
chemical data suggest the formation of reactive electrophiles.
The chemical properties of N-nitrosotolazoline are also char-
acteristic of other mutagenic and carcinogenic N-nitroso com-
pounds. On this basis, we conclude that N-nitrosoimidazolines
are likely direct-acting mutagens and carcinogens.
Acknowledgment. This research was supported by Grant
RO1 CA85538 from the National Cancer Institute, NIH. R.N.L.
thanks Profs. J. Iley and J. Fishbein for their helpful discussions
of acid catalysis. The X-ray structure of 12 was kindly provided
by Dr. Charles H. Barnes (University of Missouri).
of 35 rather than 37. There is ample precedent for thiol-induced
denitrosation reactions (28). Thus, assuming that it is the thiol
function, which is responsible for the denitrosation and the
induction of decomposition, biological thiols are expected to
have profound effect on the fate of 2.
1
Supporting Information Available: Selected H and 13C
NMR spectra, an Ortep drawing of the X-ray-determined
structure for 12, and crystallographic data for this structure. This
material is available free of charge via the Internet at http://
pubs.acs.org.
Conclusion
We have presented details of the chemistry of N-nitrosoto-
lazoline 2 under three conditions, which may be relevant to
human health: decomposition in acid such as what exists in the
stomach, nitrosation under conditions that may form N-
nitrosotolazoline endogenously (1), and N-acetylcysteine-
induced transformations that could be representative of the
reactions of biological thiols with 2 or similar N-nitrosoimid-
azolines. Although tolazoline, from which 2 is derived by
nitrosation, is an over-the-counter and prescription drug, the
research presented here is likely more significant because the
properties of 2 are logically expected to appropriately model
the chemistry of other N-nitrosoimidazolines. N-Nitrosotolazo-
line nitrosates much more rapidly than its precursor tolazoline.
As we have shown, this process directly produces a cascade of
electrophiles capable of damaging DNA, proteins, and other
important biological nucleophilic molecules.
References
(1) Loeppky, R. N., Shi, J., Barnes, C., and Geddam, S. (2008) A
diazonium cascade from the nitrosation of tolazoline, an imidazoline
containing drug. Chem. Res. Toxicol. 21, 295–307.
(2) Loeppky, R. N., and Yu, H. (2004) Amidine nitrosation. J. Org. Chem.
69, 3015–3024.
(3) Loeppky, R. N., and Cui, W. (1998) The nitrosation of imidazolines
and oxazolines. Tetrahedron Lett. 39, 1845–1848.
(4) Loeppky, R. N., and Yu, L. (1990) Nitrosamines, N-nitrosoamides,
and diazonium ions from tri-N-substituted amidines. Tetrahedron Lett.
31, 3263–3266.
(5) Iley, J., Norberto, F., and Rosa, E. (1989) N-Nitroso compounds. Part
1. Structure and decomposition of N-nitroso-2-arylimidazolines in
aqueous acidic media. J. Chem. Soc., Perkin Trans. 2 1471–1475.
(6) Iley, J., Norberto, F., and Sardinha, P. (1998) The acid-catalysed
decomposition of N-nitrotolazoline. J. Chem. Soc., Perkin Trans. 2
1998, 2207–2209.
(7) Hansen, A. L., and Skrydstrup, T. (2005) Fast and regioselective Heck
couplings with N-acyl-N-vinylamine derivatives. J. Org. Chem. 70,
5997–6003.
N-Nitrosotolazoline readily decomposes in acidic solution to
generate a highly reactive diazonium ion 14 from which the
alcohol 7 is derived. The amino ester 8 is a minor product of
this decomposition and is perceived to arise from intramolecular
trapping of the diazonium ion followed by the rapid hydrolysis
of the derived oxazoline 13. The extent of this trapping is a
function of the catalytic acid, is lowest in mineral acids, and is
always <23%. In this way, 2 behaves quite differently than the
extensive intramolecular trapping reported for the diazonium
ion derived from N-nitroso-2-phenylimidazoline 3. We also
observed very little (<2%) denitrosation of 2 under the
conditions of its acidic hydrolytic decomposition. The mecha-
nism of the acid-induced decomposition of 2, as justified by
physical chemical parameters and literature precedent, is
perceived to involve the rapid reversible protonation of the imino
nitrogen atom followed by slow general base-catalyzed addition
of H2O to the 2-carbon to give the resulting R-hydroxynitro-
samine 22, which is also a tetrahedral intermediate. Rapid
decomposition of this species gives rise to the diazonium 14
from which the products are derived by nucleophilic attack,
elimination, and rearrangement. In the following paper, we show
that the electrophiles so produced readily alkylate DNA (29).
N-Acetylcysteine increases the decomposition rate of 2 25-
fold at pH 7 and both denitrosates and induces its decomposition
to produce electrophiles. These processes occur to comparable
extents. Thus, the chemistry of 2, which we expect to be
(8) Behrens, O. K., Corse, J., Jones, R. G., Mann, M. J., Soper, Q. F.,
Van Abeele, F. R., and Chiang, M.-C. (1948) Biosynthesis of
penicillins. I. Biological precursors for benzylpenicillin (pencillin G).
J. Biol. Chem. 175, 751–764.
(9) Boggs, G. R., and Gerig, J. T. (1969) Nitrogen inversion in N-
benzoylaziridines. J. Org. Chem. 34, 1484–1486.
(10) Partridge, M. W., and Turner, H. A. (1949) Synthesis of 1-aryl-4,5-
dihydroglyoxalines. J. Chem. Soc. 1308–1310.
(11) Wang, M., Young-Sciame, R., Chung, F.-L., and Hecht, S. (1995)
Formation of N2-tetrahydrofuranyl and N2-tetrahydropyranyl adducts
in the reactions of R-acetoxy-N-nitrosopyrrolidine and R-acetoxy-N-
nitrosopiperidine with DNA. Chem. Res. Toxicol. 8, 617–624.
(12) Young-Sciame, R., Wang, M., Chung, F.-L., and Hecht, S. (1995)
Reactions of R-acetoxy-N-nitrosopyrrolidine and R-acetoxy-N-nitros-
opiperidine with deoxyguanosine: Formation of N2-tetrahydrofuranyl
and N2-tetrahydropyranyl adducts. Chem. Res. Toxicol. 8, 607–616.
(13) Wang, M., Chung, F. L., and Hecht, S. S. (1989) Formation of acyclic
and cyclic guanine adducts in DNA reacted with R-acetoxy-N-
nitrosopyrrolidine. Chem. Res. Toxicol. 2, 423–428.
(14) Chung, F. L., Wang, M., and Hecht, S. S. (1989) Detection of exocyclic
guanine adducts in hydrolysates of hepatic DNA of rats treated with
N-nitrosopyrrolidine and in calf thymus DNA reacted with R-acetoxy-
N-nitrosopyrrolidine. Cancer Res. 49, 2034–2041.
(15) Hecht, S. S., and Chen, C.-H. B. (1979) A study of chemical
carcinogenesis. 14. Hydrolysis of model compounds for R-hydroxy-
lation of the carcinogens N-nitrosopyrrolidine and N′-nitrosonornico-
tine. J. Org. Chem. 44, 1563–1566.
(16) Zink, C. N., Kim, H.-J., and Fishbein, J. C. (2006) Synthesis and
aqueous chemistry of R-acetoxy-N-nitrosomorpholine: Reactive in-
termediates and products. J. Org. Chem. 71, 202–209.
(17) Kim, H.-J., and Fishbein, J. C. (2003) Reexamination of the aqueous
chemistry of N-nitroso-3-hydroxymorpholine, a metabolite of the
carcinogen N-nitrosomorpholine. Chem. Res. Toxicol. 16, 715–720.

318 Chem. Res. Toxicol., Vol. 21, No. 2, 2008
Loeppky and Shi
(18) Jencks, W. P. (1987) Catalysis in Chemistry and Enzymology, Dover
(24) Dix, L. R., Oh, S. M. N. Y. F., and Williams, D. L. H. (1991)
Denitrosation of nitrosaminessA quantitative study. Reactions of
N-methyl-N-nitrosoaniline, N-nitrosoproline, dimethylnitrosamine and
N-nitrososarcosine. J. Chem. Soc., Perkin Trans. 2 1099–1104.
(25) Williams, D. L. H. (1988) Nitrosation, Cambridge University Press,
Cambridge, United Kingdom.
(26) Santala, T., and Fishbein, J. C. (1992) Thiolytic decomposition of the
carcinogen N-methyl-N′-nitro-N-nitrosoguanidine. A change in rate-
limiting step with nucleophile basicity controls alkylating activity.
J. Am. Chem. Soc. 114, 8852–8857.
(27) Jensen, D. E., and Magee, P. N. (1990) Nitrosation of cimetidine and
biological consequences: An overview. In The Significance of N-
Nitrosation of Drugs (Eisenbrand, G., Bozler, G., and Nicolai, H. v.,
Eds.) pp 235–250, Gustav Fischer Verlag, Stuttgart.
Publications, Inc., New York.
(19) Mesic, M., Peuralahti, J., Blans, P., and Fishbein, J. C. (2000)
Mechanisms of decomposition of R-hydroxydialkylnitrosamines in
aqueous solution. Chem. Res. Toxicol. 13, 983–992.
(20) 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.
(21) Wichems, D. N., Nag, S., Mills, J., and Fishbein, J. C. (1992) Evidence
for intermediates and a change in rate-limiting step in the aminolysis
of the carcinogen N-methyl-N′-nitro-N-nitrosoguanidine by cyclic
amines. J. Am. Chem. Soc. 114, 8846–8851.
(22) Wu, H., Loeppky, R. N., and Glaser, R. (2005) Nitrosation
chemistry of pyrroline, 2-imidazoline, and 2-oxazoline: Theoretical
Curtin-Hammett analysis of retro-ene and solvent-assisted C-X
cleavage reactions of R-hydroxy-N-nitrosamines. J. Org. Chem. 70,
6790–6801.
(28) Williams, D. L. H. (1985) S-Nitrosation and the reactions of S-nitroso
compounds. Chem. Soc. ReV. 14, 171–196.
(29) Loeppky, R. N., and Shi, J. (2008) Nucleoside and DNA adducts from
N-nitrosotolazoline. Chem. Res. Toxicol. 21, 319–329.
(23) Williams, D. L. H. (1975) Reaction of a nitrosamine with thiourea.
Direct sulfur nitrosation. J. Chem. Soc., Chem. Commun. 375–376.
TX700318K