Synthesis and decomposition of an ester derivative of the procarcinogen and promutagen, PhIP, 2-amino-1-methyl-6-phenyl-1H-imidazo[4,5-b]pyridine: Unusual nitrenium ion chemistry

Article abstract of DOI:10.1021/jo070306p

(Chemical Equation Presented) The food-derived heterocyclic amine (HCA) carcinogen 2-amino-1-methyl-6-phenyl-1H-imidazo[4,5-b]-pyridine, PhIP, is often generated in the highest concentration of the HCAs formed during broiling and frying of meat and fish. Although it is considered to be an important contributor to human cancer risk from exposure to HCAs, the chemistry of PhIP metabolites that presumably react with DNA to initiate carcinogenesis has received only cursory attention. We have synthesized the ester derivative N-pivaloxy-2-amino-1-methyl-6-phenyl-1H-imidazo[4,5-b]pyridine, 1b, and investigated its chemistry in aqueous solution. Although 1b was too unstable to isolate, we could characterize it by NMR methods in DMF-d₇, a solvent in which it is stable at -40°C. It decomposed rapidly in aqueous solution, but its conjugate acid, 1bH⁺, is not reactive. The nitrenium ion, 2, was trapped by N₃^- to form the unusual tetrazole adduct, 16. In the absence of N₃^-, the expected hydration products of 2 were not detected, but the reduction product, 12, was detected. Although such products are often taken as evidence of triplet nitrenium ions, the efficient trapping of 2 by N₃^- indicates that it is a ground state singlet species. The product 12 appears to be generated by reduction of an initially formed hydration product of 2. An alternative addition-elimination mechanism for the formation of 12 does not fit the available kinetic data. The selectivity of 2, measured as k_az/k _s, the ratio of the second-order rate constant for its reaction with N₃^- and the first-order rate constant for its reaction with the aqueous solvent, is (2.3 ± 0.6) × 10⁴ M ^-1, a value that is in the middle of the range of k _az/k_s of 10-10⁶ M^-1 observed for nitrenium ions derived from other HCAs. The mutagenicity of aromatic amines (AAs) and HCAs, measured as the log of histidine revertants per nanomole of amine, log m, in Salmonella typhimurium TA 98 and TA 100 correlates with log(k_az/k_s) for a wide variety of carbocyclic and heterocyclic amine mutagens including PhIP. Previously developed linear regression models for mutagenicity that include log(k_az/k _s) as an independent variable predict log m for PhIP with good accuracy in both TA 98 and TA 100. Quantitative carcinogenicity data are less strongly correlated with log(k_az/k_s), so prediction of the carcinogenicity of PhIP and other HCAs or AAs based primarily on log(k _az/k_s) is less successful.

Full text of DOI:10.1021/jo070306p

Synthesis and Decomposition of an Ester Derivative of the
Procarcinogen and Promutagen, PhIP,
2-Amino-1-methyl-6-phenyl-1H-imidazo[4,5-b]pyridine: Unusual
Nitrenium Ion Chemistry
Thach-Mien Nguyen and Michael Novak*
Department of Chemistry and Biochemistry, Miami UniVersity, Oxford, Ohio 45056
noVakm@muohio.edu
ReceiVed February 14, 2007
The food-derived heterocyclic amine (HCA) carcinogen 2-amino-1-methyl-6-phenyl-1H-imidazo[4,5-b]-
pyridine, PhIP, is often generated in the highest concentration of the HCAs formed during broiling and
frying of meat and fish. Although it is considered to be an important contributor to human cancer risk
from exposure to HCAs, the chemistry of PhIP metabolites that presumably react with DNA to initiate
carcinogenesis has received only cursory attention. We have synthesized the ester derivative N-pivaloxy-
2-amino-1-methyl-6-phenyl-1H-imidazo[4,5-b]pyridine, 1b, and investigated its chemistry in aqueous
solution. Although 1b was too unstable to isolate, we could characterize it by NMR methods in DMF-d7,
a solvent in which it is stable at -40 °C. It decomposed rapidly in aqueous solution, but its conjugate
-
acid, 1bH+, is not reactive. The nitrenium ion, 2, was trapped by N3 to form the unusual tetrazole
-
adduct, 16. In the absence of N3 , the expected hydration products of 2 were not detected, but the reduction
product, 12, was detected. Although such products are often taken as evidence of triplet nitrenium ions,
-
the efficient trapping of 2 by N3 indicates that it is a ground state singlet species. The product 12
appears to be generated by reduction of an initially formed hydration product of 2. An alternative addition-
elimination mechanism for the formation of 12 does not fit the available kinetic data. The selectivity of
-
2, measured as kaz/ks, the ratio of the second-order rate constant for its reaction with N3 and the first-
order rate constant for its reaction with the aqueous solvent, is (2.3 ( 0.6) × 104 M-1, a value that is in
the middle of the range of kaz/ks of 10-106 M-1 observed for nitrenium ions derived from other HCAs.
The mutagenicity of aromatic amines (AAs) and HCAs, measured as the log of histidine revertants per
nanomole of amine, log m, in Salmonella typhimurium TA 98 and TA 100 correlates with log(kaz/ks) for
a wide variety of carbocyclic and heterocyclic amine mutagens including PhIP. Previously developed
linear regression models for mutagenicity that include log(kaz/ks) as an independent variable predict log
m for PhIP with good accuracy in both TA 98 and TA 100. Quantitative carcinogenicity data are less
strongly correlated with log(kaz/ks), so prediction of the carcinogenicity of PhIP and other HCAs or AAs
based primarily on log(kaz/ks) is less successful.
Introduction
sources.1 A list of selected carcinogenic HCAs is presented in
Scheme 1 along with the common metabolic pathway required
for activation into their ultimate carcinogenic forms, the sulfuric
or acetic acid esters of the corresponding hydroxylamines.1,2
The HCAs are generally classified into two structural groups:
It has been known since the mid 1970s that mutagenic and
carcinogenic heterocyclic amines (HCAs) are present in parts
per billion concentrations in broiled and fried meats and other
protein-containing foods, as well as in tobacco smoke and other
10.1021/jo070306p CCC: $37.00 © 2007 American Chemical Society
Published on Web 06/02/2007
4698
J. Org. Chem. 2007, 72, 4698-4706

Synthesis and Decomposition of an Ester DeriVatiVe of PhIP
heterocyclic nitrenium ions with N3-/solvent selectivity, mea-
SCHEME 1. Structure and Metabolic Activation of
Heterocyclic Amine Carcinogens
sured as kaz/ks, the ratio of the second-order rate constant for
-
N3 trapping of the nitrenium ion and the pseudo-first-order
rate constant for trapping the ion by the aqueous solvent, ranging
from ca. 10 to 106 M-1 3-7
We have also demonstrated that the
.
log of the mutagenicities (log m) of HCAs and aromatic amines
(AAs) in Salmonella typhimurium TA 98 and TA 100 correlate
with log(kaz/ks), and that two and three parameter linear
regression models that include log(kaz/ks) as an independent
variable (significant at the 95% confidence level for log(kaz/
ks)) predict log m for both Salmonella strains with r2 in the
adj
range of 0.85-0.90.8
PhIP was first isolated, but not identified, from a model
system containing creatinine, glucose, and glycine in 1984 by
Sugimura and co-workers.9 Subsequently, it was isolated and
identified by Felton and co-workers in 1986 in fried ground
beef.10 Later studies showed that PhIP is also present in beer,
wine, and cigarette smoke condensate.11,12 Recent studies
indicate that PhIP-DNA adducts are present in exfoliated ductal
epithelial cells in human breast milk, and that PhIP itself is
present in human breast milk.13,14 Quantitative studies show that
PhIP is usually present in the greatest concentration of the HCAs
detected in cooked meat products.15-17 Its mutagenicity to
Salmonella is less than that of other IQ type HCAs and several
of the other non-IQ type HCAs, including Glu-P-1, Trp-P-1,
and Trp-P-2, but significantly greater than that of AAs.17 Based
on its relative abundance in cooked meat products and its
substantial carcinogenicity in mammals, PhIP is considered to
be a significant contributor to human cancer risk from HCA
exposure.18
A synthetic acetic acid ester derivative of the hydroxylamine
metabolite of PhIP, 1a, apparently generates the nitrenium ion,
2, that reacts with 2′-deoxyguanosine (d-G) residues in DNA
the IQ type HCAs (also called aminoimidazoazaarenes) that are
characterized by an imidazoquinoline, imidazoquinoxaline, or
imidazopyridine structure, and the non-IQ type HCAs that are
characterized by a bicyclic or tricyclic ring system containing
a 2-aminopyridine structure.1 We have previously shown that
ester derivatives of Phe-P-1, ARC, Glu-P-1, Glu-P-2, IQx,
MeIQx, and Trp-P-2 decompose in aqueous solution to generate
(3) Novak, M.; Xu, L.; Wolf, R. A. J. Am. Chem. Soc. 1998, 120, 1643-
1644.
(4) Novak, M.; Kazerani, S. J. Am. Chem. Soc. 2000, 122, 3606-3616.
(5) Novak, M.; Toth, K.; Rajagopal, S.; Brooks, M.; Hott, L. L.;
Moslener, M. J. Am. Chem. Soc. 2002, 124, 7972-2981.
(6) Novak, M.; Nguyen, T.-M. J. Org. Chem. 2003, 68, 9875-9881.
(7) Rajagopal, S.; Brooks, M. E.; Nguyen, T.-M.; Novak, M. Tetrahedron
2003, 59, 8003-8010.
(1) (a) Eisenbrand, G.; Tang, W. Toxicology 1993, 84, 1-82. (b) Hatch,
F. T.; Knize, M. G.; Felton, J. S. EnViron. Mol. Mutagen. 1991, 17, 4-19.
(c) Layton, D. W.; Bogen, K. T.; Knize, M. G.; Hatch, F. T.; Johnson, V.
M.; Felton, J. S. Carcinogenesis 1995, 16, 39-52. (d) Sugimura, T. EnViron.
Health Perspect. 1986, 67, 5-10. (f) Sugimura, T. Mutat. Res. 1985, 150,
33-41.
(8) Novak, M.; Rajagopal, S. Chem. Res. Toxicol. 2002, 15, 1495-1503.
(9) Negishi, C.; Wakabayashi, K.; Tsuda, M.; Sato, S.; Sugimura, T.;
Saito, H.; Maeda, M.; Jagerstad, M. Mutat. Res. 1984, 140, 55-59.
(10) Felton, J. S.; Knize, M. G.: Shen, N. H.; Lewis, P. R.; Andreson,
B. D.; Happe, J.; Hatch, F. T. Carcinogenesis 1986, 7, 1081-1086.
(11) Manabe, S.; Tohyama, K.; Wada, O.; Aramaki, T. Carcinogenesis
1991, 12, 1945-1947.
(2) (a) Felton, J. S.; Knize, M. G.; Shen, N. H.; Anderson, B. D.;
Bjeldames, L. F.; Hatch, F. T. EnViron. Health Perspect. 1986, 67, 17-24.
(b) Hashimoto, Y.; Shudo, K.; Okamoto, T. Acc. Chem. Res. 1984, 17,
403-408. (c) Shinohara, A.; Saito, K.; Yamazoe, Y.; Kamataki, T.; Kato,
R. Cancer Res. 1986, 46, 4362-4367. (d) Aoyama, T.; Gonzalez, F. J.;
Gelboin, H. V. Mol. Carcinogen. 1989, 2, 192-198. (e) Snyderwine, E.
G.; Wirth, P. J.; Roller, P. P.; Adamson, R. H.; Sato, S.; Thorgiersson, S.
S. Carcinogenesis 1988, 9, 411-418. (f) Meerman, J. H. N.; Ringer, D.
P.; Coughtrie, M. W. H.; Bamforth, K. J.; Gilissen, R. A. H. J. Chem.-Biol.
Interact. 1994, 92, 321-328. (g) Raza, H.; King, R. S.; Squires, R. B.;
Guengerich, F. P.; Miller, D. W.; Freeman, J. P.; Lang, N. P.; Kadlubar, F.
F. Drug Metab. Dispos. 1996, 24, 385-400. (h) Watanabe, M.; Ishidate,
M.; Nobmi, T. Mutat. Res. 1990, 234, 337-348. (i) Turesky, R. J.; Bracco-
Hammer, I.; Markovic, J.; Richli, U.; Kappeler, A.-M.; Welti, D. H. Chem.
Res. Toxicol. 1990, 3, 524-535. (j) Turesky, R. J.; Markovic, J.; Bracco-
Hammer, I.; Fay, L. B. Carcinogenesis 1991, 12, 1847-1855. (k) Rich, K.
J.; Murray, B. P.; Lewis, I.; Rendell, N. B.; Davies, D. S.; Gooderham, N.
J.; Boobis, A. R. Carcinogenesis 1992, 13, 2221-2226. (l) Snyderwine, E.
G.; Schut, H. A. J.; Adamson, R. H.; Thorgeirsson, U. P.; Thorgeirsson, S.
S. Cancer Res. 1992, 52 (Suppl.) 2099s-2102s. (m) Schut, H. A. J.;
Snyderwine, E. G. Carcinogenesis 1999, 20, 353-368.
(12) Manabe, S.; Suzuki, H.; Wada, O.; Ueki, A. Carcinogenesis 1993,
14, 899-901.
(13) Gorlewska-Roberts, K.; Green, B.; Fares, M.; Ambrosone, C. B.;
Kadlubar, F. F. EnViron. Mol. Mutagen. 2002, 39, 184-192.
(14) (a) Thompson, P. A.; DeMarini, D. M.; Kadlubar, F. F; McClure,
G. Y.; Brooks, L. R.; Green, B. L.; Fares, M. Y.; Stone, A.; Josephy, P.
D.; Ambrosone, C. B. EnViron. Mol. Mutagen. 2002, 39, 134-142. (b)
DeBruin, L. S.; Martos, P. A.; Josephy, P. D. Chem. Res. Toxicol. 2001,
14, 1523-1528.
(15) (a) Becher, G.; Knize, M. G.; Nes, I.; Felton, J. Carcinogenesis
1988, 9, 247-253. (b) Felton, J. S.; Knize, M. G. Mutat. Res. 1991, 259,
205-217.
(16) Skog, K. Food Chem. Toxicol. 1993, 31, 655-675.
(17) Wakabayashi, K.; Nagao, M.; Esumi, H.; Sugimura, T. Cancer Res.
1992, 52 (Suppl.), 2092s-2098s.
(18) (a) Ito, N.; Hasegawa, R.; Sano, M.; Tamano, S.; Esumi, H.;
Takayama, S.; Sugimura, T. Carcinogenesis 1991, 12, 1503-1506. (b)
Malfatti, M. A.; Kulp, K. S.; Knize, M. G.; Davis, C.; Massengill, J. P.;
Williams, S.; Nowell, S.; MacLeod, S.; Dingley, K. H.; Turtletaub, K. W.;
Lang, N. B.; Felton, J. S. Carcinogenesis 1999, 20, 705-713.
J. Org. Chem, Vol. 72, No. 13, 2007 4699

Nguyen and Novak
SCHEME 2. Proposed Reaction Pathway for Generation of
SCHEME 3. Synthesis of 1b
the PhIP-d-G Adduct in DNA
No product was detected, and 5 decomposed into a tarry residue
during the reaction. Addition of 15-crown-5 to the reaction
mixture, restriction of NaNH2 to a 2-fold excess over 5, and
limitation of the reaction time to 1.5 h resulted in a 20% yield
of PhIP based on recovered unreacted 5. The heterogeneous
Chichibabin reaction is known to be very sensitive to the pKa
of the conjugate acid of the heterocycle undergoing amination.
Typically, the reaction works only for heterocycles with pKa
values in the narrow range of 5-8.25 Since the pKa of the
conjugate acid of 5 is expected to be <4.0, it is not surprising
that the original procedure failed.26 Addition of the crown ether
brings some of the NaNH2 into solution, and the homogeneous
Chichibabin reaction is known to be far less sensitive to the
pKa of the heterocycle.25 The nitro compound, 6, was generated
by treatment of PhIP with excess NaNO2 in HOAc/H2O using
an adaptation of a procedure of Tanga and co-workers.27
Reduction of 6 to generate the hydroxylamine, 7, was ac-
complished with N2H4‚H2O and Pd/C using a procedure
developed in our laboratory.28 Due to rapid decomposition, the
hydroxylamine was stored at -40 °C no longer than 24 h before
conversion to 1b. The procedure for the generation of 1b was
adapted from that previously used to generate 1a.19 Attempts
to isolate and purify 1b were unsuccessful since it rapidly
decomposes upon evaporation of DMF and also decomposes
on contact with silica gel. HPLC analysis of the DMF solution
of 1b generated from the synthetic procedure showed that only
one major peak was present (ca. 99% by peak area). DMF
solutions of 1b were stable at -40 °C for several months but
did decompose within 24-48 h at room temperature. The
compound was characterized by NMR methods by performing
the synthesis in DMF-d7 and collecting all NMR data at -20
°C. NMR data and details of all synthetic procedures are
presented in the Supporting Information.
to generate the C-8 adduct, 3 (Scheme 2).19 The same adduct is
detected in rats treated with PhIP.19 The C-8 adduct is a
characteristic product of the reaction of both carbocyclic and
heterocyclic nitrenium ions with d-G,4,5,20,21 but no additional
characterization of 2 was performed and the ester 1a was only
characterized by mass spectrometry.19 It was not known how
selectively 2 reacted in an aqueous environment or whether its
reactivity/selectivity correlated with the mutagenicity of PhIP.
The identity of products from reaction of 2 with simple
nucleophiles such as N3- was unknown. Typically, nucleophilic
-
attack of N3 on a trigonal CH that bears a positive charge in
one of the nitrenium ion resonance structures, followed by
tautomerization, leads to a simple aryl azide.3-7,20,22 There is
no obvious way for 2 to yield such a product. Considering the
abundance of PhIP in cooked meat products and its apparent
importance as a human carcinogen, greater knowledge of the
reactions of 1 and 2 would be advantageous.
In this paper, we present the synthesis and characterization
of the pivalic acid ester, 1b, and the characterization of its
reactions in aqueous media. The pivalic acid ester was chosen
because of the known tendency of acetic acid esters of some of
the hydroxylamine derivatives of HCAs to undergo acyl transfer
reactions rather than N-O bond cleavage.6,7 Although 1b
decomposes upon attempted isolation, it is readily characterized
in DMF solution and the kinetics of its decomposition are easily
measured by UV methods. The reaction kinetics, reaction
-
products, and nitrenium ion selectivity measured by N3
-
trapping methods are described herein. The correlation of
nitrenium ion selectivity with the mutagenicity and carcinoge-
nicity of PhIP is also examined.
Results
The decomposition kinetics of 1b were monitored at 20 °C
by UV spectroscopy. All kinetics were performed in 5 vol %
CH3CN-H2O, µ ) 0.5 (NaClO4). Buffers used to maintain pH
were NaH2PO4/Na2HPO4, AcOH/NaOAc, and HCO2H/NaHCO2
at pH > 2.5, and HClO4 solutions at pH e 2.5. Repetitive
wavelength scans were collected for at least 5 half-lives of the
decomposition reaction in the pH range of 0.5-7. The data
shown in Figure 1 are typical. Isosbestic points held throughout
the course of the reaction, indicating that there is no buildup of
The critical steps in the synthesis of 1b are shown in Scheme
3. The four-step synthesis of 5 from 4 followed the procedures
of Grivas and Lindstro¨m.23 The Chichibabin reaction of 5 with
excess NaNH2 in xylene at 130 °C was reported to generate
PhIP in 90% yield.24 We were unable to reproduce this result.
(19) Frandsen, H.; Grivas. S.; Anderson, R.; Andersson, R.; Dragsted,
L.; Larsen, J. C. Carcinogenesis 1992, 13, 629-635.
(20) Novak, M.; Rajagopal, S.; Xu, L.; Kazerani, S.; Toth, K.; Brooks,
M.; Nguyen, T.-M. J. Phys. Org. Chem. 2004, 17, 615-624.
(21) Novak, M.: Kennedy, S. A. J. Am. Chem. Soc. 1995, 117, 574-
575.
(22) Novak, M.; Kahley, M. J.; Eiger, E.; Helmick, J. S.; Peters, H. E.
J. Am. Chem. Soc. 1993, 115, 9453-9460.
(23) Grivas, S.; Lindstrom, S. J. Heterocycl. Chem. 1995, 32, 467-471.
(24) Choshi, T.; Tonari, A.; Yoshioka, H.; Harada, K.; Sugino, E.; Hibino,
S. J. Org. Chem. 1993, 58, 7952-7954.
(25) McGill, C. K.; Rappa, A. In AdVances in Heterocyclic Chemistry;
Katritzky, A. R., Ed.; Academic Press: San Diego, 1988; Vol. 44, pp 1-79.
(26) Perrin, D. D. J. Chem. Soc. 1965, 5590-5596.
(27) Tanga, M. J.; Bupp, J. E.; Bradford, W. W. J. Labelled Compd.
Radiopharm. 2001, 44, 405-411.
(28) Kazerani, S.; Novak, M. J. Org. Chem. 1998, 63, 895-897.
4700 J. Org. Chem., Vol. 72, No. 13, 2007

Synthesis and Decomposition of an Ester DeriVatiVe of PhIP
SCHEME 4. Kinetic Scheme Consistent with Equation 1
SCHEME 5. Possible Routes to 12
FIGURE 1. Repetitive wavelength scans for 1b at pH 6.81, in 0.02
M NaH2PO4/Na2HPO4 buffer at 20 °C, total time ) 10 min, cycle time
) 5 s.
either N-3 or N-4 of 1b. The pKa trends for the conjugate acids
of 1H-benzimidazole, 8H+ (5.53), and 1H-imidazo[4,5-b]-
pyridine, 9H+ (3.92), suggest that protonation of 1b occurs at
N-3.26
FIGURE 2. Log kobs versus pH for 1b. Data were fit to eq 1.
a detectable intermediate during the decomposition of 1b over
this pH range. In some solutions, precipitation occurred, but
only after at least 5 half-lives of the reaction. This did not
interfere with determination of rate constants. Absorbance versus
time data collected at 325 nm fit the first-order rate equation
well at all pH. Pseudo-first-order rate constants were determined
from the average of duplicate runs at all pH and buffer
concentrations. Tables of rate constants are provided in the
Supporting Information.
Substitution with the electron-withdrawing NHO pivaloxy group
at C-2 and the steric hindrance to solvation caused by that bulky
group account for the decreased pKa observed for 1bH+.4,5,29
Although 1b is protonated under acidic conditions, it is not
reactive at low pH. Some conjugate acids of esters of hydrox-
amic acids derived from the HCAs, such as 10 and 11, do
undergo either spontaneous or acid-catalyzed decomposition,
but no hydroxylamine esters derived from HCAs have exhibited
observable reactivity of their conjugate acids.3-7,20
Rate constants were independent of buffer concentrations, but
did exhibit pH dependence as shown in Figure 2. The rate data
fit the rate law of eq 1 well.
kobs ) Kako/(Ka + 10-pH
)
(1)
The apparent ionization constant, Ka, was also detected through
spectrophotometric titration of 1b performed by plotting initial
absorbance at 325 nm versus pH (data provided in the
Supporting Information). The pKa measured from the kinetic
fit (2.46 ( 0.04) is in good agreement with the spectrophoto-
metrically measured pKa (2.52 ( 0.05). The limiting rate
The only detectable products of the decomposition of 1b in
pH 6.8 phosphate buffer detected by HPLC were 12 (48 ( 2%)
and 5 (17 ( 1%). Neither of these are simple hydrolysis products
of 1b, although 12 is a 2e- reduction product of the anticipated,
but undetected, hydrolysis product, 13 (Scheme 5), and 5 is a
4e- reduction product of 13. Although the source of reducing
equivalents is not clear, reduction products are commonly
observed among the decomposition products of hydroxylamine
and hydroxamic ester derivatives of HCAs.4,6,7,20 Compound 12
is the major decomposition product of 1b upon long-term storage
constant at neutral pH, ko, is (1.85 ( 0.08) × 10-2 s-1
.
The rate law of eq 1 was previously observed for other ester
derivatives of heterocyclic hydroxylamines obtained from Phe-
P-1, ARC, IQx, and MeIQx.3-5 The rate law is consistent with
the uncatalyzed decomposition of the conjugate base of an
unreactive acid, 1bH+ (Scheme 4).3 Protonation could occur at
(29) Novak, M.; Martin, K. A. J. Org. Chem. 1991, 56, 1585-1590.
J. Org. Chem, Vol. 72, No. 13, 2007 4701

Nguyen and Novak
FIGURE 3. Absorbance at 335 nm versus time for the decomposition
of 1b in pH 6.8 phosphate buffer containing 10 mM N3-. Data were
fit to the consecutive first-order rate equation: At ) A1 exp(-k1t) +
A2 exp(-k2t) + A∞ by a five parameter nonlinear least-squares
procedure.
FIGURE 4. Comparison of k1 and k2 measured during the decomposi-
tion of 1b in the presence of N3- with kobs measured in the absence of
N3-: kobs (b), k1 (2), k2 (1). Average values of k1 and k2 are indicated
by the red and blue lines, respectively.
SCHEME 6. Tetrazole Products from 1b and 17
in DMF solutions at -40 °C. It has also previously been
observed as a byproduct of the formation of the nitro compound,
6, and can be made by refluxing 6 in 2/1 DMF/H2O for ca.
10-15 min.30,31 This reaction, and the formation of PhIP from
5 via the Chichibabin reaction (above), suggests an alternative
process for the formation of 12 from 1b via an addition-
elimination reaction involving the intermediate 15. Although
our solutions of 1b in DMF do slowly decompose into 12, we
can demonstrate by NMR and HPLC methods that there is
insufficient 12 present in freshly made solutions of 1b (typically
2-5%) to account for the results of our hydrolysis product study.
The precipitate encountered in some reactions appears to be
polymeric and was not further characterized.
Decomposition of 1b in pH 6.8 phosphate buffer in the
-
presence of N3 leads to a number of significant changes in
kinetics and products. The data in Figure 3 show that the reaction
in the presence of N3 no longer exhibits first-order behavior
but is characterized by a consecutive first-order process that is
observed over the entire range of [N3-] examined (0.5-10 mM).
Error limits for the two rate constants obtained in each least-
squares fit are relatively large as a consequence of the fact that
-
Product studies showed that at [N3-] g 1 mM, one new
reaction product, 16, is generated in quantitative yield (Scheme
6). No other products are detected under these conditions. The
tetrazole structure of 16 is unusual for an azide adduct of a
heterocyclic hydroxylamine or hydroxamic acid ester, but not
unprecedented. The ester 17 yields the tetrazole 18 in the
presence of N3-.3 At [N3-] e 0.5 mM, the yield of 16 becomes
dependent on [N3-]. The rate and product yield data are
consistent with trapping of an intermediate cation generated by
rate-limiting ionization. The data (Figure 5) are fit well by eq
2.
the two rate constants differ by less than a factor of 2. The
-
magnitudes of the rate constants are independent of [N3
]
(Figure 4), and the standard deviations of the averages of each
rate constant obtained from seven determinations are acceptably
small at ca. 10%. One of the two rate constants observed under
these conditions, k1, is of the same magnitude (average value
in the range of 0.5-10 mM N3- of (1.64 ( 0.18) × 10-2 s-1
s-1 at pH 6.8) (Figure 4). The other rate constant, k2, is
)
-
as kobs detected in the absence of N3 ((1.78 ( 0.08) × 10-2
-
-
[15] )[15]max((kaz[N3 ]/ks)/(1 + kaz[N3 ]/ks))
(2)
somewhat smaller (average value ) (0.99 ( 0.10) × 10-2 s-1
)
In eq 2, the azide/solvent trapping ratio, kaz/ks, is the ratio of
and appears to be associated with a process that only occurs in
the presence of N3-. The data indicate that k1 is kobs, the rate
constant for the disappearance of 1b, that is unaffected by the
presence of N3-, and k2 is a rate constant that describes the
decomposition of an intermediate produced by the reaction of
-
the second-order rate constant for trapping of 2 by N3 (kaz)
and the putative first-order rate constant for trapping by the
aqueous solvent (ks). This equation is commonly used to
characterize the yield of the N3--trapping products for hetero-
cyclic and carbocyclic nitrenium ions and carbenium ions.3-7,22,32
The azide/solvent trapping ratio, kaz/ks, for 2 of (2.3 ( 0.6) ×
104 M-1 is of intermediate magnitude compared to that of other
heterocyclic nitrenium ions.3-7,20
-
N3 with the initial product of the decomposition of 1b.
(30) Frandsen, H.; Rasmussen, E. S.; Nielsen, P. A.; Farmer, P.; Dragsted,
L.; Larsen, J. C. Mutagenesis 1990, 6, 93-98.
(31) (a) Alexander, J.; Wallin, H.; Rossland, O. J.; Solberg, K. E.; Holme,
J. A.; Becher, G.; Andersson, R; Grivas, S. Carcinogenesis 1991, 12, 2239-
2245. (b) Crofts, F. G.; Strickland, P. T.; Hayes, C. L.; Sutter, T. R.
Carcinogenesis 1997, 18, 1793-1798.
(32) (a) Richard, J. P.; Jencks, W. P. J. Am. Chem. Soc. 1982, 104, 4689-
4691; 1982, 104, 4691-4692; 1984, 106, 1383-1396. (b) Richard, J. P.;
Amyes, T. L.; Vontor, T. J. Am. Chem. Soc. 1991, 113, 5871-5873.
4702 J. Org. Chem., Vol. 72, No. 13, 2007

Synthesis and Decomposition of an Ester DeriVatiVe of PhIP
FIGURE 5. Yield of 16 as a function of [N3-]. Data were fit to eq 2.
Quantitative structure-activity relationships, QSARs, are used
to correlate quantitative bacterial mutagenicity data and carci-
nogenicity data for both AAs and HCAs with calculated or
observed properties of the amines or the corresponding nitrenium
ions.33 The dependent variable in most of the mutagenicity
correlations is the logarithm of histidine revertants in Salmonella
typhimurium strains TA 98 or TA 100 per nanomole of amine
(log m) taken under conditions in which the S-9 fraction of rat
liver homogenates, induced by Arachlor 1254 or other PCB
preparations, is added to the preparation.33,34 We recently
showed that log m for both Salmonella strains for a set of 18
AAs and HCAs with widely varying structure correlated with
log S, the log of the experimental azide/solvent product ratio at
-
2
1 M N3 for the corresponding nitrenium ions (radj ) 0.5491
for TA 98 and 0.6338 for TA 100).8 Log S is equivalent to
log(kaz/ks) for nitrenium ions with log S g 1.8 We developed
two parameter linear regression models, significant at the 95%
confidence level for both variables, that included log S and a
ring index variable, Irings, that is 0 for monocyclic AAs (MAAs)
and 1 for all other amines (radj2 ) 0.8448 for TA 98 and 0.8927
for TA 100). Inclusion of a third variable, C log P, increased
radj2 to 0.8913 for TA 98 and 0.9011 for TA 100, but the model
is only significant at the 80% confidence level for C log P in
TA 100.8
Figure 6 shows that the log m versus log S data for PhIP
appear to be unremarkable when compared to the same data
for the amines included in our previous correlation. Log m for
PhIP for both strains of Salmonella is found in the literature,17
while log(kaz/ks) (log S) is from the current study. The regression
lines for both Salmonella strains are shown with and without
the inclusion of the PhIP data. The significantly smaller
mutagenicity of MAAs compared to polycyclic amines with
nitrenium ions of similar log S was previously noted and is the
reason for the inclusion of Irings in the regression models.8
FIGURE 6. Log m for ArNH2 expressed as log(revertants/nmol) versus
log S for ArNH+ for S. typhimurium TA 98 (A) and TA 100 (B): IQ
type HCAs ([, magenta), non-IQ type HCAs (1, blue), PAAs (2,
green), MAAs (b, red), and PhIP (9, black). All data except PhIP are
from ref 8. PhIP data are from ref 17 and this work. Dashed lines are
least-squares regression lines excluding PhIP (r2 ) 0.5491 for TA
adj
98, 0.6338 for TA 100); solid lines are the least-squares regression
lines for the entire data set (r2 ) 0.5614 for TA 98, 0.6400 for TA
adj
100). The two lines nearly superimpose for TA 100.
In Figure 7, the values of log m calculated from the regression
models are plotted against the experimental log m for the amines
in the previous study and for PhIP. The regression model
equations for both Salmonella strains were not recalculated for
the PhIP data. Instead, the calculated log m for PhIP was
obtained from the previously determined regression equations.
The data for PhIP clearly fall into the trend observed for the
AAs and other HCAs.
(33) (a) Benigni, R.; Giuliani, A.; Franke, R.; Gruska, A. Chem. ReV.
2000, 100, 3697-3714. (b) Colvin, M. E.; Hatch, F. T.; Felton, J. S. Mutat.
Res. 1998, 400, 479-492. (c) Chung, K.-T.; Kirkovsky, L.; Kirkovsky,
A.; Purcell, W. P. Mutat. Res. 1997, 387, 1-16.
(34) (a) Maron, D. M.; Ames, B. N. Mutat. Res. 1983, 113, 173-215.
(b) Debnath, A. K.; Debnath, G.; Shusterman, A. J.; Hansch, C. EnViron.
Mol. Mutagen. 1992, 19, 37-52. (c) Hatch, F. T.; Knize, M. G.; Felton, J.
S. EnViron. Mol. Mutagen. 1991, 17, 4-19. (d) Hatch, F. T.; Colvin, M.
E.; Seidl, E. T. EnViron. Mol. Mutagen. 1996, 27, 314-330. (e) Hatch, F.
T.; Colvin, M. E. Mutat. Res. 1997, 376, 87-96. (f) Hatch, F. T.; Knize,
M. G.; Colvin, M. E. EnViron. Mol. Mutagen. 2001, 38, 268-291. (g)
Maran, U.; Karelson, M.; Katritzky, A. R. Quant. Struct.-Act. Relat. 1999,
18, 3-10.
Discussion
Although 1b could not be isolated, it is clear from the NMR
characterization in DMF-d7 and the observed reactions that this
compound was generated. The previously reported synthesis of
1a appears to be verified since it was made by an analogous
method from the hydroxylamine, 7.19 The pH rate profile for
J. Org. Chem, Vol. 72, No. 13, 2007 4703

Nguyen and Novak
SCHEME 7. Possible Mechanism for the Formation of 16
charge are 4° (C-6 and C-8, Scheme 7). An intriguing possibility
for the formation of the observed tetrazole product, 16, is initial
attack on C-6 to form 19, followed by an intramolecular dipolar
cycloaddition to form 20. Aziridene ring opening with elimina-
tion of the C-5 proton leads to 21, a tautomer of 16. Either 20
or 21 could be the intermediate responsible for the rate constant
k2 detected during the trapping experiments. An alternative
-
mechanism involving attack of N3 on pyridyl N-4 followed
by cyclization to C-5 coupled with net proton transfer from C-5
to the exocyclic N cannot be ruled out by the available data.
-
This mechanism is considered less likely because N3 always
attacks C rather than N in other heterocyclic and carbocyclic
nitrenium ions.3-7,22
The dipolar cycloaddition of alkyl azides with nitriles to form
tetrazoles is well-known.35 Alkyl azides also react with isocy-
anates to form tetrazolones.36 Dipolar cycloadditions of alkyl
azides with simple imines appear not to have been reported,
but the proposed intramolecular cycloaddition would be en-
tropically favored. Aziridine ring opening reactions are quite
common, but formation of a detectable enamine product as in
21 is rare.37 Aromatization of the system may provide sufficient
driving force to bring about this conversion.
FIGURE 7. Plots of calculated versus observed log m for TA 98 (A)
and TA 100 (B) for the two variable (log S and Irings) regression model
of ref 8 (closed symbols) and the three variable (log S, Irings, and C log
P) regression model of ref 8 (open symbols). Symbols are defined in
Figure 6. The models were not recalculated for PhIP. The calculated
values of log m for PhIP were determined from the previously calculated
regression equations, and the observed log m values were obtained from
ref 17.
1b (Figure 2) explains the greater stability of 1a that was
observed under acidic conditions.19 The N3--trapping kinetics
results indicate that trapping occurs on an intermediate generated
after the rate-limiting reaction step. Direct N3- attack on 1b is
ruled out because of the lack of dependence of the rate of
decomposition of 1b on [N3-] (Figure 4). The most likely
intermediate is the nitrenium ion, 2. The C-8 adduct, 3, reported
from reaction of 1a with DNA is also likely to be generated by
reaction with 2 based on the trapping results for 1b and those
we previously reported for the reaction of heterocyclic nitrenium
ions with d-G.4,5,19,20 The overall similarity in reaction kinetics
and trapping results with those of other ester derivatives of
heterocyclic hydroxylamines that have been demonstrated to
yield nitrenium ions also supports our conclusion.3-5 Unlike
other nitrenium ions that we have investigated, 2 cannot form
a stable aromatic N3- adduct by tautomerization of an initially
formed adduct because the carbon atoms in 2 that bear a positive
The origin of the unusual reaction product 12 and the minor
product 5 is of interest. As shown in Scheme 5, these products
cannot be obtained from the nitrenium ion 2 without a reduction
(35) (a) Smith, P. A. S.; Clegg, J. M.; Hall, J. H. J. Org. Chem. 1958,
23, 524-529. (b) Carpenter, W. R. J. Org. Chem. 1962, 27, 2085-2088.
(c) Kay, D. P.; Kennewell, P. D.; Westwood, R. J. Chem. Soc., Perkin
Trans. 1 1982, 1879-1884. (d) Klaubert, D. H.; Bell, S. C.; Pattison, T.
W. J. Heterocycl. Chem. 1985, 22, 333-336. (e) Demko, Z. P.; Sharpless,
K. B. Angew. Chem., Int. Ed. 2002, 41, 2113-2116.
(36) (a) Chung, J. Y. L.; Ho, G.-J.; Chartrain, M.; Roberge, C.; Zhao,
D.; Leazer, J.; Farr, R.; Robbins, M.; Emerson, K.; Mathre, D. J.;
McNamara, J. M.; Hughes, D. L.; Grabowski, E. J. J.; Reider, P. J.
Tetrahedron Lett. 1999, 40, 6739-6743. (b) Denecker, G.; Smets, G.;
L’Abbe, G. Tetrahedron 1975, 31, 765-775.
(37) Bellos, K.; Stamm, H. J. Org. Chem. 1995, 60, 5661-5666.
4704 J. Org. Chem., Vol. 72, No. 13, 2007

Synthesis and Decomposition of an Ester DeriVatiVe of PhIP
process. Reduction and dimerization products often dominate
the chemistry of other heterocyclic nitrenium ions, and the
sources of reducing equivalents are difficult to determine.20 The
suspected sources of reducing equivalents in many cases are
the initial hydration products.4,20 If 13 or 14 were the source of
reduction equivalents, oxidation products of 13 or 14 would be
expected. None were detected, but the combined yield of 12
and 5 is less than quantitative (ca. 65%), so there is a significant
amount of undetected reaction product. Reduction products
observed during the triplet sensitized photolysis of certain
carbocyclic nitrenium ion precursors have been taken as
evidence for the presence of triplet state nitrenium ions.38
Experimental and theoretical results suggest that carbocyclic
nitrenium ions with low energy triplet states that might be
accessible during thermal generation are highly unstable short-
lived species that are difficult to trap with nucleophilic trapping
agents.38,39 That is not the case for heterocyclic nitrenium ions
that generate reduction products. They are always species of
moderate to high azide/solvent selectivity that can be quanti-
concentrations of pivalate ion that must be present under the
reaction conditions (e2.5 × 10-5 M) for an ion that has
moderate stability in aqueous solution and is in diffusional
-
equilibrium with nonsolvent nucleophiles based on the N3
-
trapping data.43 Addition-elimination chemistry might be
possible for 1b in a solvent that does not favor ionization, but
in aqueous solution, ionization apparently overwhelms other
possible pathways. We intend to examine this possibility at a
future date.
Since our analysis indicates that 12 is ultimately generated
-
from the cation 2, the rate constant ratio determined by the N3
-
trapping experiments does appear to be kaz/ks for that ion. Rate
-
constants for the reaction of N3 with moderately selective
nitrenium ions (kaz/ks e 105 M-1) occur at the diffusion-
controlled limit of ca. 5 × 109 M-1 s-1 44
.
If this is true for 2,
the measured kaz/ks of (2.3 ( 0.6) × 104 M-1 leads to an estimate
of ks for this ion of ca. 2.2 × 105 s-1 or a lifetime (1/ks) of
4.6 × 10-6 s. Since kaz/ks varies from 10 to 106 M-1 for a wide
range of heterocyclic nitrenium ions, 2 is an ion of intermediate
selectivity.3-7,20 It is about as selective as the ion derived from
ARC but much less selective than ions derived from IQx,
MeIQx, Glu-P-1, and Glu-P-2.4,5 Whatever the mechanism for
the formation of 12 may be, the experimentally determined value
-
tatively trapped by N3 or other nucleophiles.3-7,20 A limited
number of calculations on heterocyclic nitrenium ion singlet-
triplet gaps have appeared, and the singlet-triplet gap for one
highly stabilized heterocyclic nitrenium ion has been experi-
mentally measured.40-42 The persistent 1,3-dimethylbenzotria-
zolium ion, 22, is a ground state singlet with a singlet-triplet
gap of -(66 ( 3) kcal/mol.40 The calculations indicate that
heteroatoms positioned to generate closed shell resonance
structures, such as N-1 in 2 or N-1 and N-3 in 22, stabilize the
singlet electronic state preferentially, while heteroatoms posi-
tioned to bear a positive charge without generation of a closed
shell resonance structure, such as N-3 and N-4 of 2, preferen-
tially destabilize the singlet state.40-42 The available experi-
mental data and the calculations suggest that 2 is unlikely to
have an accessible triplet state.
-
of kaz/ks does measure the selectivity for reaction of 2 with N3
compared to all other possible reactions in the absence of N3
-
,
so it is a valid experimental measure of the selectivity of 2.
Although 2 generates an unusual azide adduct, Figures 6 and
7 show that the overall selectivity of 2, measured as log(kaz/ks)
(log S), does predict the mutagenicity of PhIP in Salmonella
typhimurium TA 98 and TA 100 based on our previously
developed linear regression model for mutagenicity. That model
uses log(kaz/ks) as an independent variable based on the
underlying hypothesis that the nitrenium ion selectivity controls
the efficiency of reaction of the nitrenium ion with DNA in
vivo, and that the formation of the DNA adduct is a dominant
factor in mutagenesis.8 The success of the regression model
suggests that the hypothesis may be correct. Since 2 is of
intermediate azide/solvent selectivity compared to the variety
of both carbocyclic and heterocyclic nitrenium ions previously
examined, PhIP would be expected to be of intermediate
mutagenicity. Figure 6 shows that is the case, and Figure 7
shows that the previously developed regression model ad-
equately predicts the mutagenicity of PhIP in both Salmonella
strains. The azide/solvent selectivity is less successful at
correlating carcinogenicity of aromatic amines in mice and rats.8
Although there is a positive correlation, carcinogenicity data,
expressed as -log TD50, exhibit much less dependence on log-
(kaz/ks) than does log m with considerably more scatter.8 Since
log m for AAs and HCAs varies over a range of ca. 8 log units
while carcinogenicity data for the same amines vary over ca. 4
log units, slopes of the correlation lines for -log TD50 are about
half of those for log m. Plots of -log TD50 for mice (13 amines
An alternative explanation of the origin of 12 is that it is
formed via an addition-elimination mechanism that does not
involve 2 (Scheme 5). This attractive proposal has precedent
in the chemistry of this system, but it is at odds with the
available experimental evidence. At 20 °C the half-life of 1b
in aqueous solution is only 37 s. Compound 6 that has a
considerably better leaving group (NO2-) than 1b is stable in
aqueous solution at room temperature for several days. It must
be refluxed in DMF/H2O mixtures to be converted into 12.30,31
The pH dependence of the decomposition of 1b is difficult to
reconcile with this proposal. Protonation of 1b would be
expected to activate it toward nucleophilic substitution via the
addition-elimination mechanism, but experimentally, 1b is
-
stabilized by protonation. Finally, the results of N3 trapping
-
are inconsistent with this proposal. Trapping by N3 occurs
quantitatively at [N3-] g 1 mM without rate acceleration. This
could not occur if 12 was generated by a competing addition-
elimination process unless the formation of the nitrenium ion,
2, was reversible, and almost all of the nitrenium ion returned
to 1b in the absence of N3-. This is highly unlikely at the
(43) A reviewer has suggested that 12 could be formed by an intramo-
lecular rearrangement of 1b to a spiro intermediate that undergoes hydrolytic
decomposition to 12 and O-pivaloylhydroxylamine. This suggestion might
explain the rapid reaction of 1b but is still inconsistent with the N3--trapping
data.
(44) (a) Davidse, P. A.; Kahley, M. J.; McClelland, R. A.; Novak, M. J.
Am. Chem. Soc. 1994, 116, 4513-4514. (b) McClelland, R. A.; Gadosy,
T. A.; Ren, D. Can. J. Chem. 1998, 76, 1327-1337. (c) McClelland, R.
A.; Davidse, P. A.; Hadzialic, G. J. Am. Chem. Soc. 1995, 117, 4173-
4174. (d) Ren, D.; McClelland, R. A. Can. J. Chem. 1998, 76, 78-84. (e)
Bose, R.; Ahmad, A. R.; Dicks A. P.; Novak, M.; Kayser, K. J.; McClelland,
R. A. J. Chem. Soc., Perkin Trans. 2 1999, 1591-1599.
(38) Falvey, D. E. J. Phys. Org. Chem. 1999, 12, 589-596.
(39) (a) Sullivan, M. B.; Brown, K.; Cramer, C. J.; Truhlar, D. G. J.
Am. Chem. Soc. 1998, 120, 11778-11783. (b) Winter, A. H.; Falvey, D.
E.; Cramer, C. J. J. Am. Chem. Soc. 2004, 126, 9661-9668.
(40) McIlroy, S.; Cramer, C. J.; Falvey, D. E. Org. Lett. 2000, 2, 2451-
2454.
(41) Sullivan, M. B.; Cramer, C. J. J. Am. Chem. Soc. 2000, 122, 5588-
5596.
(42) Hilal, R.; Khalek, A. A. A.; Elroby, S. A. K. THEOCHEM 2005,
731, 115-121.
J. Org. Chem, Vol. 72, No. 13, 2007 4705

Nguyen and Novak
All kinetics of the decomposition of 1b in the presence of the
added N3 were performed in 0.02 M 1/2 NaH2PO4/Na2HPO4
including PhIP) and rats (11 amines including PhIP) versus log
S are shown in the Supporting Information. Of the variables
previously examined, C log P provided the strongest correlation
with the carcinogenicity data, in agreement with earlier observa-
tions made on a larger data set.8,33a The weaker correlation to
a single variable is not surprising since carcinogenesis in
mammals is a far more complicated process than mutagenesis
in bacteria. In light of these correlations, it is not surprising
that PhIP was demonstrated to be more effective than other
HCAs, including several that are more mutagenic than PhIP, in
transforming cultured mammalian cells.45
-
buffers at pH 6.8 under ionic strength and temperature conditions
-
identical to those of the kinetic studies done in the absence of N3
.
NaN3 solutions were prepared at high concentration (0.45 M) in a
0.02 M phosphate buffer at the appropriate pH and were diluted
into the working range with a 0.02 M phosphate buffer of the same
pH and ionic strength. Repetitive wavelength scans showed
consecutive first-order behavior. Absorbance versus time data were
collected at 335 nm and fit to a consecutive first-order rate equation
by nonlinear least-squares procedures.
Product Analysis. Reaction products obtained in the absence
-
of N3 were monitored by HPLC with UV detection at 308 nm.
HPLC conditions were C18 reverse-phase analytical column, 65:35
MeOH/H2O eluent, at a flow rate of 1.0 mL/min. For preparative
scale isolation, a ca. 9.3 × 10-3 M was prepared in 8 mL of DMF.
This solution was added to 50 mL of pH 7.0, 0.02 M phosphate
buffer at 20 °C in 270 µL aliquots at 40 s intervals. The reaction
was continued for 10 half-lives (400 s) after the last addition of
1b. HPLC indicated that 1b had decomposed. The solution was
extracted with CH2Cl2 (3 × 25 mL). The organic extracts were
dried over anhydrous Na2SO4, and the solvent was evaporated. The
residue was dissolved in MeOH, filtered to remove insoluble
materials, and purified by preparative HPLC. The purified products
were identified as 5 and 12 by comparison with authentic samples
(see Supporting Information).
Experimental Section
General procedures for purifying solvents, preparing solutions,
measuring rate constants by UV methods, and performing product
studies by HPLC have been published.22,46,47 All salts used in the
preparation of the buffers were reagent grade and commercially
available. H2O used in the kinetics studies was distilled, deionized,
and redistilled. MeOH used as an HPLC solvent was reagent grade
and was distilled before use. H2O used as an HPLC solvent was
distilled and deionized. Syntheses and characterization of all
compounds used in this study are presented in the Supporting
Information. The ester 1b could not be isolated but was prepared
in DMF solutions in which it was stable for several months if stored
at -40 °C.
Kinetic Studies. All kinetics were performed in 5 vol % CH3-
CN-H2O, µ ) 0.5 (NaClO4), at 20 °C. Buffers used to maintain
pH were 0.02 M NaH2PO4/Na2HPO4, 0.02 M AcOH/NaOAc, and
0.02 M HCO2H/NaHCO2 at pH > 2.5, and HClO4 solutions at pH
e 2.5. Higher concentrations of buffers were used to examine the
possibility of buffer catalysis. A stock solution of 1b was prepared
in DMF at a concentration of ca. 5 × 10-3 M as described in the
Supporting Information. Injection of 15 µL of this stock solution
into 3.0 mL of buffer at 20 °C generated reaction solutions of initial
concentration of ca. 2.5 × 10-5 M in 1b. Repetitive wavelength
scans in the range of 240-600 nm showed that the reaction followed
a first-order pattern. Absorbance versus time data were taken at
325 nm. The data were fit to a standard first-order rate equation to
The product studies of the decomposition of 1b in the presence
-
of N3 was performed in 0.02 M 1/2 NaH2PO4/Na2HPO4 buffers
at pH 6.8 under ionic strength and temperature conditions identical
to the kinetic studies. Products were monitored by HPLC with UV
detection at 335 nm. HPLC conditions were identical to those
described above. The initial concentration of 1b was adjusted so
-
that N3 was kept in at least a 10-fold excess in all experiments.
Product yields were determined by the average of two HPLC
injections taken after 8 half-lives of the slower of the two reactions
-
detected by kinetics. Isolation and characterization of the N3
adduct, 16, are described in the Supporting Information.
Acknowledgment. The 500 MHz NMR spectrometer and
the LC/MS were provided by grants to MU by the Hayes
Investment Fund of the Ohio Board of Regents.
provide kobs
.
Supporting Information Available: Synthesis and character-
ization of 6, 7, 1b, and 12, isolation and characterization of 16,
(45) (a) Thompson, L. H.; Wu, R. W.; Felton, J. S. Proc. Natl. Acad.
Sci. U.S.A. 1991, 88, 3827-3831. (b) Pfau, W.; Martin, F. L.; Cole, K. J.;
Venitt, S.; Phillips, D. H.; Grover, P. L.; Marquardt, H. Carcinogenesis
1999, 20, 545-551.
(46) Novak, M.; Pelecanou, M.; Roy, A. K.; Andronico, F. M.; Plourde,
F. M.; Olefirowicz, T. M.; Curtin, T. J. J. Am. Chem. Soc. 1984, 106, 5623-
5631.
spectrophotometric titration of 1b, kinetic data for 1b, 1H and 13
C
NMR spectra for 1b, 12, and 16, and plots of -log TD50 for mice
and rats, including PhIP, versus log S. This material is available
free of charge via the Internet at http://pubs.acs.org.
(47) Novak, M.; Brodeur, B. A. J. Org. Chem. 1984, 49, 1142-1144.
JO070306P
4706 J. Org. Chem., Vol. 72, No. 13, 2007