Esters of N-(4-Biphenylyl)hydroxylamine
J. Am. Chem. Soc., Vol. 119, No. 33, 1997 7663
-
4
-1
ion than 3c toward N3 (kaz/ks ) 5.4 × 10 M for 3c and 1.2
due to attack on both the aromatic ring and the N of the
5
-1
14b
20,21
×
10 M for 3g).
A linear extrapolation of a plot of log-
nitrenium ion.
The high selectivity of N-7 in basic purine
(
kdG/ks) vs log ks (not shown) for 3a-c provides an estimate of
nucleosides for N of the nitrenium ion is unprecedented and
not fully understood. Steric effects may play some role in this
selectivity since the initial product of N-7 attack on the ortho
or para carbons of 3a-c or 3g would be quite sterically
congested. Our data are in conflict with the recently published
scheme of Dipple meant to explain site selectivity of the
reactions of electrophilic carcinogens with purines.26 In this
scheme, hard electrophilic sites react at N-7 of G and A and
soft electrophilic sites tend to react at N-2 and O-6 of G and
4
-1
kdG/ks for 3g at 20 °C of 1.6 × 10 M . If the temperature
dependence of kdG/ks is similar for 3c and 3g, kdG/ks for 3g at
3
-1
3
7 °C will be ca. 8.6 × 10 M . We conclude that the low
yield of 4g observed by Kadlubar and co-workers is likely due
to trapping of 3g by the small amount of G present under the
reaction conditions.
Our results are consistent with mechanism b of Scheme 2.
The adducts 8a,b clearly show that deprotonation of C -H is
8
26
not necessary for C-8 adducts to form. We have observed
intermediates that lead to 8a,b. Since HPLC (Figure 8B) and
NMR (Table 4) data show that in each case there are two
intermediates leading to the final adducts, the intermediates we
observe are not consistent with structures similar to 19 that could
not exist as a pair of diastereomers. Our data are consistent
N-6 of A. Our data show the opposite selectivity for reactions
of purines with N-arylnitrenium ions. A general pattern of site
selectivity that can be applied across a wide range of electro-
philic carcinogens is not yet apparent.
The pattern that emerges from this study and others3 is that
,4
7
+
purine nucleosides containing a basic N-7 (pKa of N -H
g
8
with the proposed structure 16, the C -CH3 analogue of 23.
2.3) react preferentially and selectively with long-lived nitrenium
-
7
The hydrolysis products derived from the intermediates and the
ultimate stable adducts, 8, are also consistent with the proposed
structures.
ions (1/ks g 1.5 × 10 s) to form an N-7 adduct 21 that
subsequently undergoes rearrangement via mechanism b of
Scheme 2 to form the C-8 adducts. This reaction is not
competitive with solvent trapping of the ion for nitrenium ions
4
Kadlubar and co-workers did directly observe 19. This
compound was stable enough for H NMR measurements to be
1
-9
with significantly shorter lifetimes (1/ks < 5 × 10 s) under
made, but we see no sign of similar adducts building up to
detectable levels in our experiments with 8-MeG and 1a or 1b.
The most significant difference in the experimental conditions
employed lies in the choice of the purine nucleophile. The
reaction of nitrenium ion 3g with C ,N -dimethylguanine leads
to 19, in which the N-9 can stabilize the structure by a direct
accessible nucleoside concentrations (< 50 mM) because the
diffusion limit for the nucleoside trapping reaction appears to
9
-1 -1
be about 2 × 10 M s .
If the N-7 site of the purine nucleoside is less basic, alternative
nucleophilic sites on the purine can complete with N-7 to form
adducts of different structure. None of these reactions approach
the selectivity observed for purine nucleosides with a basic N-7.
Implications with Respect to Carcinogenesis. Our data
show why the major in vivo and in vitro DNA adducts of
carcinogens such as 1a-c and 17 are C-8 adducts formed at
8
9
9
resonance interaction. If the N -CH3 of 19 is replaced by an
electron-withdrawing ribofuranosyl moiety, this resonance
interaction is significantly curtailed and the resulting intermedi-
ate would be much less stable. This difference in stabilization
may be sufficient to prevent observation of analogous interme-
diates in our experiments. Despite our inability to directly detect
such intermediates in the present study, it seems clear, based
1
,27
the dG base sites.
The N-7 of the dG moiety is the most
reactive purine or pyrimidine base site toward nitrenium ions
on DNA, and the initial N-7 adduct rearranges to form the
observed C-8 adduct. The structure of the DNA does play some
role in the site selectivity of nitrenium ions, however. Treatment
of native DNA in vivo or in vitro with 1c or its precursors leads
to the formation of a minor (5-20% compared to C-8 adducts)
7
+
on the strong dependence of knuc on the pKa of N -H and on
the results of Kadlubar and co-workers, that the intermediates
1 (Scheme 2) are involved in these reactions.
Site Selectivity. For the more basic nucleosides, G, dG,
4
2
1
a-e,27
-
N-2 adduct, 24.
A similar material is also observed as a
8
-MeG, and X , reaction occurs exclusively at N-7 in aqueous
solution, but as the basicity of N-7 decreases, the rate constant
for trapping of nitrenium ions by N-7 decreases and other types
of adducts are formed, specifically the O-6 adducts derived from
I and the unique N-6 benzene imine adducts derived from A.
Our data show that these are also formed as the result of
nitrenium ion trapping (Figure 6). In these cases, the nucleo-
philic exocyclic O-6 or N-6 of the purines attack the nitrenium
ion at the ortho or para carbons of the aromatic ring proximal
to the nitrenium N. These are the normal sites of attack of
minor product of the reaction of 4-aminobiphenyl-derived
-
- 7,25
typical hard nucleophiles such as N3 and Cl .
28
carcinogens with native DNA. This type of adduct is not
-
1d,2,3
Only soft nucleophilic sites such as C, S , and neutral
aromatic amines show any tendency to react with nitrenium ions
at N, and even in these cases, a mixture of products is found
observed in solution studies with dG, G, or denatured DNA,
so its formation must be a consequence of the modification of
the site selectivity of the nitrenium ion by the three-dimensional
structure of duplex DNA.
Although the structure-induced modification of nitrenium ion
site selectivity may have a relatively minor effect on product
distributions, it can have a major effect on the physiological
role of the carcinogen. For example, the N-acylated C-8 adducts
are subject to relatively efficient excision and repair, but the
N-2 adduct is much more persistent in vivo and, for that reason,
(
19) Miller, J. A.; Miller, E. C. EnViron. Health Perspect. 1983, 49, 3-12.
Garner, R. C.; Martin, C. N.; Clayson, D. B. In Chemical Carcinogens,
nd ed.; Searle, C. E., Ed.; ACS Monograph 182; American Chemical
Society; Washington, DC, 1984; Vol. 1, pp 175-276.
20) Novak, M.; Rangappa, K. S. J. Org. Chem. 1992, 57, 1285-1290.
Novak, M.; Rangappa, K. S.; Manitsas, R. K. J. Org. Chem. 1993, 58,
2
(
7
813-7821.
(
(
(
(
(
21) Novak, M.; Lin, J. J. Am. Chem. Soc. 1996, 118, 1302-1308.
22) Novak, M.; Pelecanou, M. Unpublished results.
23) McClelland, R. A. Personal communication.
24) Novak, M.; Kennedy, S. A. J. Phys. Org. Chem. 1997, in press.
25) Novak, M.; Kahley, M. J.; Lin, J.; Kennedy, S. A.; Swanegan, L.
(26) Dipple, A. Carcinogenesis 1995, 16, 437-441.
(27) Kriek, E. Cancer Res. 1972, 32, 2042-2048. Westra, J. G.; Kriek,
E.; Hittenhausen, H. Chem.-Biol. Interact. 1976, 15, 149-164.
(28) Kriek, E.; Hengeveld, G. M. Chem.-Biol. Interact. 1978, 21, 179-
201.
A. J. Am. Chem. Soc. 1995, 117, 574-575. Novak, M.; Kahley, M. J.; Lin,
J.; Kennedy, S. A.; James, T. G. J. Org. Chem. 1995, 60, 8294-8304.