Kinetics of hydrolysis of 8-(arylamino)-2′-deoxyguanosines

Article abstract of DOI:10.1021/jo0163492

The 8-(arylamino)-2′-deoxyguanosines, or C-8 adducts, are the major adducts formed by reaction of N-arylnitrenium ions derived from carcinogenic and mutagenic amines with 2′-deoxyguanosine (d-G) and guanosine residues of DNA. The hydrolysis kinetics of three C-8 adducts 1a-c were determined by UV and HPLC methods at 20 °C under acidic, neutral, and mildly alkaline conditions. At pH < 2 the dominant hydrolysis process is spontaneous cleavage of the C-N bond of the doubly protonated substrate, 1H₂⁺² (Scheme 2). The C-8 adducts are 2- to 5-fold more reactive than d-G under these conditions. At 3 < pH < 6 the hydrolysis kinetics are dominated by cleavage of the C-N bond of the monoprotonated nucleoside 1H⁺. Under these conditions the hydrolysis kinetics are accelerated by 40- to 1300-fold over that of d-G. The rate increase appears to be caused by a combination of steric acceleration of C-N bond cleavage and a decrease in the ionization constant of 1H⁺, K_a1, due to the electron-donating properties of the arylamino C-8 substituent. Under neutral pH conditions a slow (k_obs ≈ 10^-8 s^-1 to 5 × 10^-7 s^-1) spontaneous cleavage of the C-N bond of the neutral nucleoside, 1, occurs that has not been previously reported for simple purine nucleosides. Finally, under mildly alkaline conditions a process consistent with spontaneous decomposition of the anion 1^- or OH^--induced decomposition of 1 is observed. The latter process has been observed for other purine nucleosides, including the closely related 1d, and involves nucleophilic attack of OH^- on C-8 to cleave the imidazole ring of the purine.

Full text of DOI:10.1021/jo0163492

J . Org. Chem. 2002, 67, 2303-2308
2303
Kin etics of Hyd r olysis of 8-(Ar yla m in o)-2′-d eoxygu a n osin es
Michael Novak,* Megan Ruenz, Shahrokh Kazerani, Krisztina Toth,
Thach-Mien Nguyen, and Brian Heinrich
Department of Chemistry and Biochemistry, Miami University, Oxford, Ohio 45056
novakm@muohio.edu
Received December 5, 2001
The 8-(arylamino)-2′-deoxyguanosines, or C-8 adducts, are the major adducts formed by reaction
of N-arylnitrenium ions derived from carcinogenic and mutagenic amines with 2′-deoxyguanosine
(d-G) and guanosine residues of DNA. The hydrolysis kinetics of three C-8 adducts 1a -c were
determined by UV and HPLC methods at 20 °C under acidic, neutral, and mildly alkaline conditions.
At pH < 2 the dominant hydrolysis process is spontaneous cleavage of the C-N bond of the doubly
+2
protonated substrate, 1H2 (Scheme 2). The C-8 adducts are 2- to 5-fold more reactive than d-G
under these conditions. At 3 < pH < 6 the hydrolysis kinetics are dominated by cleavage of the
C-N bond of the monoprotonated nucleoside 1H+. Under these conditions the hydrolysis kinetics
are accelerated by 40- to 1300-fold over that of d-G. The rate increase appears to be caused by a
combination of steric acceleration of C-N bond cleavage and a decrease in the ionization constant
of 1H+, Ka1, due to the electron-donating properties of the arylamino C-8 substituent. Under neutral
pH conditions a slow (kobs ≈ 10-8 s-1 to 5 × 10-7 s-1) spontaneous cleavage of the C-N bond of the
neutral nucleoside, 1, occurs that has not been previously reported for simple purine nucleosides.
Finally, under mildly alkaline conditions a process consistent with spontaneous decomposition of
the anion 1- or OH--induced decomposition of 1 is observed. The latter process has been observed
for other purine nucleosides, including the closely related 1d , and involves nucleophilic attack of
OH- on C-8 to cleave the imidazole ring of the purine.
Sch em e 1
In tr od u ction
During our studies of the trapping of N-arylnitrenium
ions by 2′-deoxyguanosine (d-G) we have isolated and
characterized a number of 8-(arylamino)-2′-deoxygua-
nosines commonly referred to as C-8 adducts (Scheme
1).1 These are the most common adducts isolated from
the reactions of mutagenic and carcinogenic metabolites
of carbocyclic and heterocyclic aromatic amines with d-G,
DNA oligomers, and DNA both in vitro and in vivo.1,2 We
noted that several of these C-8 adducts underwent
hydrolysis of the C-N glycosidic bond at appreciable
rates in moderately acidic to neutral aqueous buffers (3
e pH e 7 at 20 °C), conditions under which d-G shows
very little reactivity (kobs < 10-6 s-1).1
In DNA this hydrolysis would result in depurination
of guanine sites of the DNA polymer. It has been noted
that DNA modified with these C-8 adducts is subject to
relatively rapid spontaneous depurination of the modified
purinic sites.3 Spontaneous depurination is thought to
contribute to some of the mutagenic effects of several
classes of mutagens and to cellular death.4 Although the
hydrolysis reaction may be of some physiological impor-
tance and is also interesting from a mechanistic point of
view, it has never been investigated quantitatively.
(1) Novak, M.; Kahley, M. J .; Eiger, E.; Helmick, J . S.; Peters, H.
E. J . Am. Chem. Soc. 1993, 115, 9453-9460. Novak, M.; Kennedy, S.
A. J . Am. Chem. Soc. 1995, 117, 4173-4174. Kennedy, S. A.; Novak,
M.; Kolb, B. A. J . Am. Chem. Soc. 1997, 199, 7654-7664. Novak, M.;
Kazerani, S. J . Am. Chem. Soc. 2000, 122, 3606-3616.
(2) (a) Saris, C. P.; vanDijk, W. J .; Westra, J . G.; Hamzink, M. R.
J .; vandeWerken, G.; Zomer, G.; Stavenuiter, J . F. C. Chem.-Biol.
Interact. 1995, 95, 29-40. (b) Snyderwine, E. G.; Turtletaub, K. W. In
Food Borne Carcinogens: Heterocyclic Amines; Nagao, M., Sugimura,
T., Eds.; Wiley: New York, 2000; pp 131-161. (c) Kriek, E. Chem.-
Biol. Interact. 1969, 1, 3-17. (d) Kriek, E. Chem-Biol. Interact. 1971,
3, 19-28. (e) Meerman, J . H. N.; Beland, F. A.; Mulder, G. J .
Carcinogenesis 1981, 2, 413-416. (f) Fuchs, R. P. P. Anal. Biochem.
1978, 91, 663-673. (g) Beland, F. A.; Dooley, K. L.; J ackson, C. D.
Cancer Res. 1982, 42, 1348-1354. (h) Tamura, N.; King, C. M.
Carcinogenesis 1990, 11, 535-540. (i) Kriek, E.; Miller, J . A.; J uhl,
U.; Miller, E. C. Biochemistry 1967, 6, 177-182. (j) Nelson, J . H.;
Grunberger, D.; Cantor, C. R.; Weinstein, I. B. J . Mol. Biol. 1971, 62,
331-346. (k) Evans, F. E.; Miller, D. W.; Levine, R. A. J . Am. Chem.
Soc. 1984, 106, 396-401.
(3) Drinkwater, N. R.; Miller, E. C.; Miller, J . A. Biochemistry 1980,
19, 5087-5092.
(4) Foster, P. L.; Eisenstadt, E.; Miller, J . H. Proc. Natl. Acad. Sci.
U.S.A. 1983, 80, 2695-2698. Schaaper, R. M.; Glickman, B. W.; Loeb,
L. A. Cancer Res. 1982, 42, 3480-3485. Kunkel, T. A. Proc. Natl. Acad.
Sci. U.S.A. 1984, 81, 1494-1498. Loeb, L. A.; Preston, B. D. Annu.
Rev. Genet. 1986, 30, 201-230. Wang, Y.; Wang, Y.; Stoner, G.; You,
M. Cancer Res. 1993, 53, 1620-1624. Heflich, R. H.; Neft, R. E. Mutat.
Res. 1994, 318, 73-144.
10.1021/jo0163492 CCC: $22.00 © 2002 American Chemical Society
Published on Web 03/06/2002

2304 J . Org. Chem., Vol. 67, No. 7, 2002
Novak et al.
Sch em e 2
and adenine occurs at similar rates, and the extrapolated
rate constant for depurination of double stranded DNA
at 37 °C and pH 7.4 is 3 × 10-11 s-1 11
.
In contrast to their lability under acidic conditions
simple purine nucleosides such as adenosine (A), 2′-
deoxyadenosine (d-A), guanosine (G), and d-G are rela-
tively inert under neutral to weakly basic conditions
(7 e pH e 9).5-8,12 Under more strongly alkaline condi-
tions purine nucleosides are subject to a base-catalyzed
hydrolysis that involves OH- attack on C-8 with cleavage
of the imidazole ring of the nucleoside.12 Products include
pyrimidine and purine bases as well as some materials
that are not UV active.12 Some 8-(arylamino)-2′-deoxy-
guanosines are also subject to this alkaline decomposition
reaction.13
In this paper we present a kinetic study of the
decomposition of three representative C-8 adducts, 1a -
c, (Scheme 1) in the acidity range Ho ≈ -1 to pH 9. The
kinetics of hydrolysis of 1a -c have been compared and
contrasted with those of d-G. Kinetic analysis shows that
four different pathways contribute to the hydrolysis of
1a -c under these conditions. The N-arylamino-C-8 ad-
ducts 1a -c do exhibit significantly accelerated hydrolysis
compared to d-G at pH > 2, although under more strongly
acidic conditions they undergo hydrolysis at rates com-
parable to that of d-G. The reasons for the accelerated
hydrolysis of 1a -c at pH > 2 are described herein.
Resu lts a n d Discu ssion
Hydrolysis kinetics for 1a -c and d-G were monitored
in 20 vol % CH3CN-H2O at 20 °C for all compounds and
at 60 °C for d-G. In the pH range ionic strength was
maintained at 0.5 (NaClO4), but no attempt was made
to maintain constant ionic strength at [H+] > 0.5 M.
HClO4 was used to maintain pH or Ho at pH e 2.5.
Buffers of HCO2H/HCO2Na, AcOH/AcONa, NaH2PO4/
Na2HPO4, and trisH+/tris were used to maintain pH in
the range from ca. 2.5 to 9.5. Repetitive wavelength scans
showed that isobestic points held during the decomposi-
tion of all compounds at Ho g -1. Pseudo-first-order rate
constants were obtained by fitting UV absorbance vs time
data (kobs > 2 × 10-5 s-1) or HPLC peak area vs time
data (2 × 10-6 s-1 < kobs < 2 × 10-5 s-1) to the standard
first-order rate equation. For kobs < 2 × 10-6 s-1, rate
constants were obtained by initial rates methods from
HPLC peak area vs time data for the starting nucleosides
during the first 5% of the hydrolysis reaction. Figure 1
provides representative examples of these data for 1a .
Rate constants did depend on pH or Ho but were
independent of buffer concentrations in all buffers except
tris. In those solutions buffer-independent rate constants
were obtained by extrapolation to zero buffer concentra-
tion. In other buffers rate constants were typically
measured in solutions with total buffer concentration of
0.02 M. Tables of all buffer-independent rate constants
obtained in this study are provided in Supporting Infor-
Acid-catalyzed hydrolysis of monomeric purine nucleo-
sides has received a considerable amount of attention and
is a well understood reaction.5-8 A general mechanism
for the acid-catalyzed reaction that is consistent with
kinetic data, lack of buffer effects, isotope effects, entropy
of activation data, and the lack of anomerization of
unreacted substrate is shown in Scheme 2. Both mono-
protonated (1H+) and diprotonated (1H2+2) substrate
formed via preequilibrium protonation appear to be
subject to rate-limiting C-N bond cleavage. On the basis
of data from the 2-deoxyglucosyl oxocarbenium ion and
similar species, it is likely that the 2-deoxyribosyl oxo-
carbenium ion has a short but finite lifetime of ca. 10-12
to 10-11 s that requires all reactions of the cation to occur
in a complex in which the leaving group is present.9
Depurination of single and double stranded DNA
occurs via a similar acid-catalyzed pathway that pre-
dominates from moderately acidic conditions to at least
the physiologically relevant pH 7.4.10,11 Loss of guanine
(5) (a) Zoltewicz, J . A.; Clark, D. F.; Sharpless, T. W.; Grahe, G. J .
Am. Chem. Soc. 1970, 92, 1741-1750. (b) Zoltewicz, J . A.; Clark, D.
F. J . Org. Chem. 1972, 37, 1193-1197. (c) Panzica, R. P.; Rousseau,
R. J .; Robins, R. K.; Townsend, L. B. J . Am. Chem. Soc. 1972, 94, 4708-
4714. (d) Garrett, E. R.; Mehta, P. J . J . Am. Chem. Soc. 1972, 94, 8532-
8541.
(6) Romero, R.; Stein, R.; Bull, H. G.; Cordes, E. H. J . Am. Chem.
Soc. 1978, 100, 7620-7624.
(7) (a) Lo¨nnberg, H.; Lehikoinen, P. Nucleic Acids Res. 1982, 10,
4339-4349. (b) Oivanen, M.; Lo¨nnberg, H.; Zhou, X.-X.; Chatto-
padhyaya, J . Tetrahedron 1987, 43, 1133-1140. (c) Laayoun, A. Decout,
J .-L.; Lhomme, J . Tetrahedron Lett. 1994, 35, 4989-4990.
(8) (a) J ordan, F.; Niv, H. Nucleic Acids Res. 1977, 4, 697-709. (b)
Hovinen, J .; Glemarec, C.; Sandstro¨m, A.; Sund, C.; Chattopadhyaya,
J . Tetrahedron 1991, 47, 4693-4708.
(10) Greer, S.; Zamenhof, S. J . Mol. Biol. 1962, 4, 123-141. Lindahl,
T. Nature 1993, 709-715.
(11) Lindahl, T.; Nyberg, B. Biochemistry 1972, 11, 3610-3618.
(12) Lo¨nnberg, H.; Lehikoinen, P.; Neuvonen, K. Acta Chem. Scand.
B 1982, 36, 707-712. Lo¨nnberg, H.; Lukkari, J .; Lehikoinen, P. Acta
Chem. Scand. B 1984, 38, 573-578. Lo¨nnberg, H.; Lehikoinen, P. J .
Org. Chem. 1984, 49, 4964-4969. Lehikoinen, P.; Mattinen, J .;
Lo¨nnberg, H. J . Org. Chem. 1986, 51, 3819-3823.
(9) Aymes, T. L.; J encks, W. P. J . Am. Chem. Soc. 1989, 111, 7888-
7900. Huang, X.; Surry, C.; Hiebert, T.; Bennet, A. J . J . Am. Chem.
Soc. 1995, 117, 10614-10621. Zhu, J .; Bennet, A. J . J . Am. Chem. Soc.
1998, 120, 3887-3893.
(13) Kriek, E.; Westra, J . G. Carcinogenesis 1980, 1, 459-468.
Boitex, S.; Bichara, M.; Fuchs, R. P. P.; Laval, J . Carcinogenesis 1989,
10, 1905-1909.

Hydrolysis of 8-(Arylamino)-2′-deoxyguanosines
J . Org. Chem., Vol. 67, No. 7, 2002 2305
F igu r e 2. Log (kobs) vs pH or Ho for 1a (b), 1b (2), 1c (1), and
d-G (9 at 20 °C, 0 at 60 °C). Data were fit as described in the
text to obtain the kinetic parameters summarized in Table 1.
dependence of log kobs on Ho.5b,7a Figure 2 shows that the
C-8 adducts 1a -c hydrolyze at 20 °C at considerably
larger rates than does d-G in moderately acidic to neutral
pH solutions, although rate constants for hydrolysis at
pH < 2 are comparable (within a factor of 5) for all four
compounds. For all four compounds the plots show a
linear dependence of log kobs on pH or Ho at pH < 2 (kobs
≈ k2aH/Ka2, see Scheme 2). For d-G this linear dependence
continues at 20 °C to pH 4.0, beyond which the reaction
kinetics were not monitored because kobs < 10-7 s-1. At
60 °C the linear dependence of d-G hydrolysis rate
constants on pH can be shown to extend to at least pH
6.0 (Figure 2). Hydrolysis rate constants for d-G in acidic
solutions at 60 °C are 1.7 × 102 larger than at 20 °C.
Between pH 3.0 and 6.0, kobs for 1a -c appears to depend
on the concentration of the acid form of a substrate
species with a pKa in the range of 3.5-4.5 (kobs ≈ k1aH/
(Ka1 + aH)). At pH > 6.0 a pH-independent term is
observed (kobs ≈ ko), and at pH > 8.0 a term dependent
on the concentration of the basic form of a substrate
species with pKa ∼9.0 dominates the kinetics for 1a and
1b (kobs ≈ k-Kao/(Kao + aH)). Hydrolysis kinetics for 1c
were not followed beyond pH 8.
F igu r e 1. Kinetics of the decomposition of 1a . (A) Absorbance
at 308 nm vs time at Ho ) -0.67. (B) and (C) HPLC peak area
for 1a vs time at pH 2.52 (B) and 5.72 (C).
The kinetic behavior was analyzed in terms of the
mechanism of Scheme 2 that, in part, has been used to
describe the pH-dependent behavior of the hydrolysis
kinetics of other nucleosides.5-8,12,13 The lack of buffer
effects in formate, acetate, and phosphate buffers is
consistent with hydrolysis via rate-limiting unimolecular
decomposition of 1H2+2, 1H+, and 1. Buffer effects in tris
buffers indicate the reaction is more complicated in this
pH range, but the buffer-independent rate constants in
tris buffers do appear to follow this scheme. The expres-
sion for kobs derived from this Scheme (eq 1) can be
simplified to eq 2 if aH/Ka2 , 1. Under our conditions
there is no evidence for an ionization corresponding to
Ka2 in solutions as acidic as Ho ) -1.0, and pKa2
measured for other guanosine nucleosides is generally
in the range of -2.5.7b,14 The assumption used to derive
eq 2 appears to be valid for 1a -c in the acidity range of
this study. The kinetic data for 1a -b were fit to eq 2 to
produce the lines shown in Figure 2A. It was assumed
in performing these fits that pKao ) 9.4, the known value
for d-G under these conditions.15 This assumption was
mation. For 1a , 1b, and 1c it was not possible to obtain
reliable values of kobs at Ho < -1 because absorbance vs
time data no longer fit the first-order rate equation or
consecutive first-order rate equation well. No attempt
was made to examine d-G hydrolysis kinetics beyond
Ho ≈ -1.
Product isolation and characterization, combined with
HPLC monitoring of reaction mixtures, confirmed that
the major initial hydrolysis products observed throughout
the Ho and pH range up to ca. pH 7 were the expected
bases 2a -c and guanine (Scheme 1). In all solutions at
pH < 6 these products are formed quantitatively. In the
more basic buffers the yields of 2a and 2b are not
quantitative and decrease with increasing basicity, be-
coming negligible at pH g 9, although no other products
were isolated under these conditions. No attempt was
made to isolate or identify the sugar product generated
from the hydrolysis reactions. Under strongly acidic
conditions some of the initially formed bases, notably 2c,
did exhibit slow decomposition, but these reactions were
not examined further.
Plots of log kobs vs pH or Ho are provided in Figure 2.
The Ho acidity scale was chosen because previous studies
of purine nucleoside hydrolysis have shown a linear
(14) The ribose unit does lower pKa2: Benoit, R. L.; Frechette, M.
Can J . Chem. 1984, 62, 995-1000; 1985, 63, 3053-3056.
(15) Clauwaert, J .; Stockx, J . Z. Naturforsch. 1968, 23b, 25-30.

2306 J . Org. Chem., Vol. 67, No. 7, 2002
koka1/aH + k1 + k2aH/Ka2 + k-KaoKa1/aH2
Novak et al.
kobs
)
(1)
(2)
1 + Ka1/aH + aH/Ka2 + KaoKa1/aH2
koKa1 + k1aH + k2aH2 /Ka2 + k-KaoKa1/aH
aH + Ka1 + KaoKa1/aH
kobs
)
made because pKao is not likely to be strongly affected
by C-8 substituents and because we have insufficient
kinetic data to determine pKao with good certainty. The
kinetics in this pH region were not pursued further
because extrapolating the rate constants for such slow
reactions to zero tris buffer concentration is very time-
consuming and because the hydrolysis kinetics in this
pH range are not the main focus of this paper. The
kinetics of hydrolysis of 1c were fit to eq 2 with the last
term of both the denominator and numerator removed
because there is no evidence for reaction through 1-
(Scheme 2) for this compound in the pH range examined.
The rate constants and pKa’s derived from the fits are
summarized in Table 1.
F igu r e 3. Initial absorbance at 308 nm vs pH for 1a . Data
were fit to a standard titration curve to obtain pKa1.
Sch em e 3
Hydrolysis kinetics for d-G at both 20 and 60 °C
confirmed the previously reported observation that this
nucleoside does not exhibit a break in its pH-rate profile
around its known pKa1 of ca. 2.5.5a This has been
interpreted to mean that k1/Ka1 ≈ k2/Ka2 for this, and
several other purine nucleosides, that exhibit this
behavior.5a,7b The kinetic data for d-G were fit to eq 3,
where kH ≈ k1/Ka1 ≈ k2/Ka2:
been observed in the hydrolysis reaction of other C-8
substituted nucleosides.7c,8a This is not the case. The
effects are not large, but k1 for 1a -c are ca. 3.0- to 15.0-
fold larger than k1 for d-G (Table 1). The overall sub-
stituent effect on k1/Ka1 ranges from a 40-fold increase
for 1a to a 1.3 × 103-fold increase for 1c compared to d-G.
The combined substituent effects on Ka1 and k1 are
responsible for the rate accelerations of the hydrolysis
of 1a -c that are evident in Figure 2 in the pH range
from ca. 2.0 to 6.0.
A similar result has been obtained for the hydrolysis
of 8-dimethylaminoguanosine compared to the parent
nucleoside G.8a The 8-dimethylamino substituent in-
creases pKa1 and k1 compared to G even though the
8-amino and 8-methylamino substituents decrease k1.8a
The increase in k1 for the 8-dimethylamino substituent
was attributed to release of steric strain in the 8-di-
methylamino compound that was presumably locked into
a syn glycosyl conformation by the bulk of the C-8
substituent.8a The importance of the conformational
preference in establishing this rate acceleration has been
disputed, but the bulk of the substituent does appear to
be critical to the observed acceleration.8b
kobs ) kHaH
(3)
Spectrophotometric pKa1’s were measured for 1a -c and
d-G at 20 °C to confirm the magnitudes of the kinetically
determined pKa1’s and to provide pKa1 for d-G so that k1
could be estimated for this compound. A plot of initial
absorbance vs pH for 1a (Figure 3) is typical. The
spectrophotometrically determined pKa1’s are reported in
Table 1. For 1a -c they are identical, within experimental
error, to the kinetically determined values. The pKa1 of
2.7 ( 0.3 for d-G is in good agreement with the pKa1 of
2.3-2.5 for d-G measured previously in similar aqueous
solvent systems.5a,16
The 8-(arylamino) substituents of the C-8 adducts
1a -c are electron-donating groups that are expected to
increase the pKa of the N-7 protonated nucleosides. A
comparison of pKa1 for d-G and 1a -c (Table 1) shows that
this is the case. The stronger acid weakening effects of
the heterocyclic substituents of 1b and 1c may be caused
by an intramolecular H-bonding interaction (Scheme 3).
The electron-donating properties of the substituents
would be expected to decrease the magnitude of k1 as has
An alternative explanation for the rate accelerations
caused by the 8-(arylamino) substituents could be that
they differentially stabilize protonation at N-3 rather
Ta ble 1. Su m m a r y of Ra te Con sta n ts a n d Ion iza tion Con sta n ts for 1a -c a n d d -G a t 20 a n d 60 °C
1a a
1ba
1cb
d-Gc
k1 (s-1
)
(8.4 ( 2.9) × 10-6
(1.2 ( 0.3) × 10-5
(4.2 ( 1.1) × 10-5
2.8 ( 1.4 × 10-6
k1/Ka1 (M-1 s-1
)
)
(5.6 ( 3.3) × 10-2
(3.6 ( 1.5) × 10-1
1.8 ( 1.0
(1.4 ( 0.1) × 10-3
(2.2 ( 0.2) × 10-1
)
f
k2/Ka2 (M-1 s-1
(3.1 ( 0.4) × 10-3
(6.5 ( 0.8) × 10-3
(5.3 ( 0.9) × 10-3
(5.2 ( 1.6) × 10-7
(1.4 ( 0.1) × 10-3
(2.2 ( 0.2) × 10-1
2.7 ( 0.3
)
f
ko (s-1
)
(3.3 ( 1.6) × 10-8
(7.2 ( 1.7) × 10-6
3.8 ( 0.2
(1.2 ( 1.0) × 10-8
(1.3 ( 0.3) × 10-5
4.5 ( 0.2
k- (s-1
)
d
pKa1
4.6 ( 0.2
4.5 ( 0.1
e
pKa1
3.5 ( 0.1
4.4 ( 0.1
a
b
Derived from fits of log kobs to the logarithmic version of eq 2 with the assumption that pKao ) 9.4. Derived from fits of log kobs to
the logarithmic version of eq 2 with last terms in both the numerator and denominator removed. c Derived from fits of log kobs to the
logarithmic version of eq 3. Obtained from the kinetic fits. e Obtained from spectrophotometric titration. f Measured at 60 °C.
d

Hydrolysis of 8-(Arylamino)-2′-deoxyguanosines
J . Org. Chem., Vol. 67, No. 7, 2002 2307
Sch em e 4
Sch em e 5
than N-7, and the N-3 monoprotonated nucleosides are
subject to more rapid decomposition than N-7 monopro-
tonated nucleosides.17 This possibility cannot be entirely
ruled out, but 15N chemical shift data for protonation of
8-dimethylaminoguanosine, 8-methylaminoguanosine,
8-aminoguanosine, and G by TFA in DMSO are consis-
tent with exclusive monoprotonation at N-7 in all cases.8b
It is unlikely that the 8-(arylamino) substituents of this
study would have a substantially different effect on the
protonation site.
The 8-(arylamino) substituent effect on k2/Ka2 is se-
verely attenuated compared to k1/Ka1. It ranges from 2.2
for 1a to 4.6 for 1b compared to d-G. The 8-(arylamino)
substituents are expected to have little effect on Ka2. It
has been shown previously that this pKa is ca. -2.5 for
guanosine derivatives with quite different structures.7b,14
The steric accelerations previously observed for k1 appear
to also occur for k2.
The C-8 adducts 1a -c are subject to a unique hydroly-
sis pathway that is observed under neutral pH condi-
tions: the uncatalyzed hydrolysis of the neutral nucleo-
side 1 governed by ko. Significant neutral pH hydrolysis
of simple nucleosides such as d-G or d-A has not previ-
ously been reported.5-8 At 60 °C we can place an upper
limit on ko for d-G of ca. 2 × 10-7 s-1 based on the lack of
observable uncatalyzed hydrolysis for d-G at the highest
pH examined (6.0). If the rate constant ratio of 1.7 × 102
observed for acid-catalyzed hydrolysis of d-G at 60 and
20 °C applies to ko, this rate constant is no larger than
ca. 1.2 × 10-9 s-1 for d-G at 20 °C and is probably
considerably smaller.11 This suggests that ko for 1a -c is
accelerated by a minimum of between 10 (1b)- and 400
(1c)-fold over ko for d-G. Although we have insufficient
data to write a detailed mechanism for the uncatalyzed
reaction, we do know from product studies that the bases
2a -c appear to be the products of these reactions. This
is particularly true for 1c because the alkaline hydrolysis
(below) does not compete with the neutral hydrolysis of
1c in the pH range examined. Release of steric strain
may be an important factor in the acceleration of the
neutral hydrolysis of 1a -c. We currently do not under-
stand the reason for the accelerated neutral hydrolysis
of 1c compared to that of 1a and 1b.
processes, although the OH--induced decomposition of
1 is more chemically reasonable than spontaneous de-
composition of 1-.12,13
There is a direct precedent for the OH- pathway in a
closely related C-8 adduct.13 The product and kinetics of
the alkaline hydrolysis of 1d (Scheme 5) have been
reported in the pH range from 9.5 to 12.5.13 The product
is the pyrimidine derivative 3d formed by OH- attack
on C-8 of the nucleoside.13 Analysis of the kinetic data
at 37 °C according to the mechanism of Scheme 4
provides a value of kOH of ca. 1 M-1 s-1 and kOH′ of ca. 2
× 10-3 M-1 s-1 for 1d if all of the observed hydrolysis
occurs via these two paths. The ratio kOH/kOH′ of 5 × 102
is consistent with the expected deceleration of attack of
OH- on 1d - caused by a combination of electrostatic
effects and increased electron density at C-8 of 1d -.
Although the products of hydrolysis of 1a and 1b under
alkaline conditions were not isolated, it is very likely that
the same reaction occurs for these compounds as for 1d .
The reaction kinetics are consistent with this interpreta-
tion, and the magnitudes of kOH for 1a and 1b are
comparable to that for 1d if the temperature difference
is taken into account. The decreased yields of 2a and 2b,
the normal products of C-N glycosidic bond cleavage,
observed under moderately alkaline conditions are also
consistent with the reaction of Scheme 5.
In conclusion, the N-arylamino C-8 adducts 1a -c
undergo hydrolysis under acidic and weakly basic condi-
tions via processes that have been described for other
purine nucleosides.5-8,12,13 The rate accelerations com-
pared to that of d-G that are apparent in moderately
acidic solutions in which the observed rate constant is
k1/Ka1 are caused by an apparent steric increase in k1 and
a concurrent decrease in Ka1 caused by the electron-
donating effect of the N-arylamino substituent. A unique
uncatalyzed hydrolysis of the neutral nucleosides also
causes these compounds to be quite labile compared to
simple nucleosides under neutral pH conditions.
The apparent spontaneous hydrolysis of 1- for 1a and
1b is kinetically equivalent to OH--induced decomposi-
tion of 1 (Scheme 4). There is precedent for OH- attack
on C-8 of closely related nucleosides.12,13 If the assumed
rate constant for such a process is designated kOH and it
accounts for the entire reaction under these mildly basic
conditions, its value would be given by Kaok-/Kw, where
k- is the value shown in Table 1, Kao is assumed to be
the value for d-G,15 and Kw has its usual definition. For
1a , kOH would be ca. 3 × 10-1 M-1 s-1, whereas for 1b it
would be ca. 5 × 10-1 M-1 s-1. The magnitudes of k- and
kOH cannot be used to eliminate either of the two possible
Exp er im en ta l Section
Gen er a l. All salts used in the preparation of buffers were
reagent grade. The purifications of CH3CN and DMF have
been described elsewhere.18 Water for kinetic studies and
product studies was distilled, deionized, and distilled again.
All other reagents and solvents were reagent grade and
distilled. All pH measurements were made at 20 °C with an
Orion model 701 pH meter equipped with a Radiometer GK-
2402C electrode. UV kinetics and titration experiments were
performed on a Cary 3 UV-visible spectrophotometer equipped
with thermostated cell holders. Some of the kinetics were
performed on an Applied Photophysics stopped-flow spectro-
(16) Rauwitscher, M.; Sturtevant, J . M. J . Am. Chem. Soc. 1960,
82, 3739-3740.
(17) Remaud, G.; Zhou, X.-X.; Chattopadhyaya, J .; Oivanen, M.;
Lo¨nnberg, H. Tetrahedron 1987, 43, 4453-4461.
(18) Novak, M.; Brodeur, B. A. J . Org. Chem. 1984, 49, 1142-1144.

2308 J . Org. Chem., Vol. 67, No. 7, 2002
Novak et al.
photometer (SX.18MW). LC-MS experiments were carried out
on a Bruker ESQUIRE-LC-MS instrument with either EI or
ACPI. Deoxyguanosine was purchased from Fluka.
Kin etics a n d Titr a tion s. Reactions were performed in 20
vol % CH3CN-H2O solutions, µ ) 0.5 (NaClO4) at 20 or 60
°C. At pH e 2.5, HClO4 was used to maintain pH. For all other
solutions buffers of HCO2Na/HCO2H, NaOAc/AcOH, Na2HPO4/
NaH2PO4, and tris/trisH+ were used to maintain pH. Ionic
strength was not maintained for HClO4 solutions in the Ho
range. These solutions were prepared as described in the
literature.20 Initial concentrations of 1a -c or d-G of (1-2) ×
Syn th esis. The syntheses and characterization of 1a -c are
described in the literature.1,2a
N-(Gu a n in -8-yl)-4-a m in obip h en yl (2a ). This material
was obtained by allowing 25 mg (0.058 mmol) of 1a to
decompose in 250 mL of HClO4 solution (µ ) 0.5 M (NaClO4),
20% CH3CN-H2O, 0.316 M HClO4, 20 °C). Compound 1a was
dissolved in 10 mL of water and added in 0.9 mL aliquots to
the HClO4 solution at 10 min intervals. The reaction mixture
was allowed to sit for 100 min after the last addition. The
solution was neutralized with 5% NaHCO3 to pH 6.5 and
extracted with CH2Cl2. The CH2Cl2 extracts were dried over
Na2SO4, and the solvent was evaporated to dryness. The
impurities were removed by triturating the precipitate with
anhydrous ether to give 2a : 1H NMR (200 MHz, CDCl3) δ 8.01
(2H, d, J ) 8.4 Hz), 7.70 (4H, m), 7.44 (3H, m); 13C NMR (50.3
MHz, DMSO-d6) δ 153.8 (C), 151.8 (C), 148.4 (C), 140.0 (C),
139.6 (C), 138.0 (C), 134.9 (C), 129.0 (CH), 127.4 (CH), 127.1
(CH), 126.2 (CH), 120.1 (C), 118.1 (CH); LC-MS (ESI, positive
ion mode), C17H14N6OK (M + K) requires m/e 357.08, found
357.14; LC-MS (ESI, negative ion mode), C17H13N6O (M - H)
requires m/e 317.12, found 317.35; High-resolution MS (ES,
positive), C17H15N6O (M + H) requires m/e 319.1307, found
319.1325.
N -(Gu a n osin -8-yl)-2-a m in o-3-m e t h yl-5-p h e n ylp yr i-
d in e (2b). Compound 2b was obtained by allowing 50 mg of
1b to decompose in 250 mL of HClO4 solution (µ ) 0.5
(NaClO4)), 20% CH3CN-H2O, 0.316 M HClO4, 20 °C) as
described above for 2a . The CH2Cl2 extracts were dried over
Na2SO4, and the solvent was evaporated to dryness. The
impurities were removed by triturating the precipitate with
anhydrous ether to give 2b: 1H NMR (200 MHz, DMSO-d6) δ
13.48 (1H, bs), 8.24 (1H, s), 8.14 (1H, s), 7.80 (2H, bs), 7.67
(1H, d, J ) 6 Hz), 7.44 (4H, m); 13C NMR (75.5 MHz, DMSO-
d6) δ 152.2 (C), 141.5 (CH), 134.4 (C), 130.4 (CH), 129.1 (CH),
128.2 (CH), 127.1 (C), 126.7 (C), 126.0 (CH), 124.8 (C), 122.7
(C), 16.5 (CH3); LC-MS (ESI, positive ion mode), C17H16N7OH
(M + H) requires m/e 334.14, found 334.15; LC-MS (ESI,
negative ion mode), C17H14N7OH (M - H) requires m/e 332.13,
found 332.45.
10-5 M were obtained by injecting 15 µL of ca. (2-4) × 10-3
M
DMF stock solutions into 3.0 mL of the reaction solution that
had been incubated at the appropriate temperature for at least
15 min.
Kinetics were monitored by UV or HPLC methods that have
been described.21 HPLC was performed using C-8 reverse
phase analytical columns with MeOH/H2O eluents buffered
with 0.05 M 1/1 KOAc/HOAc. HPLC peaks were monitored
by UV spectroscopy. Absorbance or HPLC peak area vs time
data were fit to the standard first-order rate equation for kobs
> 2 × 10-6 s-1. For slower reactions initial rates were
determined from the slopes of HPLC peak area vs time data
for 1a -c and d-G for the first 5% of the decomposition. Initial
absorbances for spectrophotometric titrations were determined
by extrapolation to t ) 0 using either a linear extrapolation
or the first-order rate equation.
Wavelengths used in these studies are as follows:
1a
1b
1c
d-G
UV; 308 nm, HPLC; 280 nm
UV; 351 or 280 nm, HPLC; 280 nm
UV; 380 nm, HPLC; 340 nm
UV; 257 nm, HPLC; 280 nm
Ack n ow led gm en t. This work was supported, in
part, by a grant from the American Cancer Society
(RPG-96-078-03-CNE). NMR spectra were obtained on
equipment originally made available through an NSF
grant (CHE-9012532) and subsequently upgraded
through an Ohio Board of Regents Investment Fund
grant. MALDI-TOF and LC-MS mass spectra were
obtained on equipment funded by NSF (CHE-9413529)
and the Ohio Board of Regents Investment Fund,
respectively. The stopped-flow spectrophotometer was
obtained through an NSF equipment grant (CHE-
0076936). High resolution mass spectra were obtained
at the Ohio State University Chemical Instrumentation
Center.
N2-(Gu a n in -8-yl)-2-a m in o-9H-p yr id o[2,3-b]in d ole (2c).
This compound was synthesized by a literature procedure:19
1H NMR (300 MHz, DMSO-d6) δ (1H, s, exchangeable), 11.05
(1H, s, br, exchangeable), 10.73 (1H, s, br, exchangeable), 8.39
(1H, d, J ) 8.4 Hz), 7.99 (1H, d, J ) 7.7 Hz), 7.46 (1H, d, J )
8.0 Hz), 7.33 (1H, t, J ) 8.1 Hz), 7.17 (1H, t, J ) 7.5 Hz), 6.90
(1H, d, J ) 8.4 Hz), 6.33 (2H, s, br, exchangeable); 13C NMR
(75.5 MHz, DMSO-d6) δ 153.0 (3 x C), 152.0 (C), 150.0 (C),
137.6 (C), 131.0 (CH), 124.8 (CH), 121.1 (C), 119.7 (CH), 119.7
(C), 119.6 (CH), 111.1 (CH), 108.6 (C), 103.0 (CH); MALDI-
TOF-MS, C16H13N8O (M + H) requires m/e 333.12, found
333.00.
Su p p or tin g In for m a tion Ava ila ble: Table of rate con-
stants obtained for the hydrolysis of 1a -c and d-G. This
material is available free of charge via the Internet at
http://pubs.acs.org.
J O0163492
(20) Markham, A. E. J . Am. Chem. Soc. 1941, 63, 874-875. Yates,
K.; Wai, H. J . Am. Chem. Soc. 1964, 86, 5408-5413.
(21) Novak, M.; Pelecanou, M.; Roy, A. K.; Andronico, A. F.; Plourde,
F. M.; Olefirowicz, J . M.; Curtin, T. J . J . Am. Chem. Soc. 1984, 106,
5623-5631.
(19) Pfau, W.; Schulze, C.; Shiral, T.; Hasegawa, R.; Brockstedt, U.
Chem. Res. Toxicol. 1997, 10, 1192-1197. Pfau, W.; Brockstedt, U.;
Schulze, C.; Neurath, G.; Marquardt, H. Carcinogenesis 1996, 17,
2727-2732.