Chloride ion catalyzed conformational inversion of carbocation intermediates in the hydrolysis of a benzo[a]pyrene 7,8-diol 9,10-epoxide

Article abstract of DOI:10.1021/ja020436+

A highly efficient procedure for converting 7β,8α-dihydroxy-9α,10α-epoxy-7,8,9, 10-tetrahydrobenzo[a]pyrene (1) to its trans-9,10-chlorohydrin (5) with excellent yield and purity by the reaction of anhydrous HCI in THF has been developed. The rate of reac

Full text of DOI:10.1021/ja020436+

Published on Web 11/12/2002
Chloride Ion Catalyzed Conformational Inversion of
Carbocation Intermediates in the Hydrolysis of a
Benzo[a]pyrene 7,8-Diol 9,10-Epoxide
Lanxuan Doan,† Haruhiko Yagi,‡ Donald M. Jerina,‡ and Dale L. Whalen*,†
Contribution from the Department of Chemistry and Biochemistry, UniVersity of Maryland,
Baltimore County, Baltimore, Maryland 21250 and Laboratory of Bioorganic Chemistry,
NIDDK, National Institutes of Health, Bethesda, Maryland 20892
Received March 25, 2002
Abstract: A highly efficient procedure for converting 7â,8R-dihydroxy-9R,10R-epoxy-7,8,9,10-tetrahy-
drobenzo[a]pyrene (1) to its trans-9,10-chlorohydrin (5) with excellent yield and purity by the reaction of
anhydrous HCl in THF has been developed. The rate of reaction of 5 has been determined as a function
of sodium chloride concentration in 1:1 dioxane-water solutions. A large common ion rate depression for
the reaction of the chlorohydrin was observed, and the rate data are fit to a mechanism in which all of the
tetrol products are formed by the reaction of water with the C-10 carbocation intermediate. Yet, the cis/
trans ratio of tetrols from the reaction of the carbocation intermediate from the hydrolysis of chlorohydrin
5 is different than the cis/trans tetrol ratio from the acid-catalyzed hydrolysis of diol epoxide 1, which
hydrolyzes via a carbocation with the same connectivity as that formed in the hydrolysis of 5. To rationalize
these results, it is proposed that the SN1 reaction of chlorohydrin 5 yields a different distribution of carbocation
conformations than that formed from the reaction of 1 with H+. The energy barrier for the inversion of
these carbocation conformations must be large relative to the energy barriers for the reaction of each
carbocation conformation with water. In solutions containing sufficient concentrations of chloride ion, however,
a lower energy pathway via a halohydrin exists for the interconversion of the carbocation conformations.
Thus, chloride ion catalyzes the interconversion of these two carbocation conformations.
Introduction
To explain the different tetrol product ratios formed from the
reaction of diol epoxide 1 in solutions containing chloride ion,
The rates1 and products of the hydrolysis reactions1-4 of the
7,8-dihydroxy-9-10-epoxide metabolite (1) of the environmental
carcinogen benzo[a]pyrene have been thoroughly studied. At
pH < ∼7 in 10:90 dioxane-water, 1 reacts with H+ to form a
triol carbocation 2, which reacts with water to form trans 9,10-
and cis 9,10-tetrols 3 and 4 in a 94:6 ratio. In this reaction,
epoxide ring opening is rate-limiting, and carbocation 2 has a
sufficient lifetime in water solutions to react with external
nucleophiles such as azide5 and halide6,7 ions.
it was proposed that a trans chlorohydrin 5 is formed as a
reactive intermediate from the reaction of carbocation 2 with
chloride ion, and that 5 hydrolyzes to form cis and trans tetrols
in a ratio different than that formed from the reaction of 1 in
the absence of chloride ion (Scheme 1). This proposed mech-
anism was confirmed by the independent synthesis of chloro-
hydrin 5 and a study of its hydrolysis reactions.8,9 Chlorohydrin
5 is highly reactive toward hydrolysis and yields tetrols in the
same ratio as that formed from the reaction of diol epoxide 1
in solutions containing sufficient concentrations of chloride ion.
It has also been observed that the covalent adducts from
reaction of 5 with deoxyadenosine and with DNA have mostly
cis 9,10-stereochemistry,8 whereas the covalent adducts from
the reaction of diol epoxide 1 with DNA in the absence of
chloride ion have mostly trans 9,10-stereochemistry.6,10,11 There
are significant cellular concentrations of chloride ion (∼10 mM),
and therefore, chlorohydrin 5 may be formed from the reaction
of diol epoxide 1 in vivo; thus, 5 may play a role in chemical
The reaction of 1 in solutions containing chloride ion,
however, yields significantly more cis tetrol (∼35%) than that
formed from the reaction of 1 in the absence of chloride ion.6
* Corresponding author. Telephone: 410-455-2521. Fax: 410-455-2608.
E-mail: whalen@umbc.edu.
† University of Maryland.
‡ National Institutes of Health.
(1) Whalen, D. L.; Montemarano, J. A.; Thakker, D. R.; Yagi, H.; Jerina, D.
M. J. Am. Chem. Soc. 1977, 99, 5522-5524.
(2) Thakker, D. R.; Yagi, H.; Lu, A. Y. H.; Levin, W.; Conney, A. H.; Jerina,
D. M. Proc. Natl. Acad. Sci. U.S.A. 1976, 73, 3381-3385.
(3) Yang, S. K.; McCourt, D. W.; Gelboin, H. V. J. Am. Chem. Soc. 1977, 99,
5130-5134.
(8) Song, Q.; Negrete, G. R.; Wolfe, A. R.; Wang, K.; Meehan, T. Chem. Res.
Toxicol. 1998, 11, 1057-1066.
(4) Yagi, H.; Thakker, D. R.; Hernandez, O.; Koreeda, M.; Jerina, D. M. J.
Am. Chem. Soc. 1977, 99, 1604-1611.
(9) Meehan, T.; Wolfe, A. R.; Negrete, G. R.; Song, Q. Proc. Natl. Acad. Sci.
U.S.A. 1997, 94, 1749-1754.
(5) Lin, B.; Islam, N.; Friedman, S.; Yagi, H.; Jerina, D. M.; Whalen, D. L. J.
Am. Chem. Soc. 1998, 120, 4327-4333.
(10) Sayer, J. M.; Chadha, A.; Agarwal, S. K.; Yeh, H. J.; Yagi, H.; Jerina, D.
M. J. Org. Chem. 1991, 56, 20-29.
(6) Wolfe, A. R.; Meehan, T. Chem. Res. Toxicol. 1994, 7, 110-119.
(7) Lin, B.; Doan, L.; Yagi, H.; Jerina, D. M.; Whalen, D. L. Chem. Res.
Toxicol. 1998, 11, 630-638.
(11) Cheng, S. C.; Hilton, B. D.; Roman, J. M.; Dipple, A. Chem. Res. Toxicol.
1989, 2, 334-340.
9
14382
J. AM. CHEM. SOC. 2002, 124, 14382-14387
10.1021/ja020436+ CCC: $22.00 © 2002 American Chemical Society

Cl Ion Catalyzed Interconversion of Carbocation Conformations
A R T I C L E S
Scheme 1
exclusively via a discrete carbocation intermediate. A rationale
for the observation that the acid-catalyzed hydrolysis of diol
epoxide 1 and the hydrolysis of chlorohydrin 5 yield different
hydrolysis product mixtures is provided.
Experimental Section
Materials and Methods. (()-7â,8R-Dihydroxy-9R,10R-epoxy-
7,8,9,10-tetrahydrobenzo[a]pyrene (1), in which the benzylic 7-hydroxyl
group and epoxide oxygen are trans, was prepared by published
procedures.4,13 This material is carcinogenic and should be handled with
caution. Dioxane and THF were distilled from sodium prior to use.
All other reagents were purchased from commercial sources. NMR
spectra were recorded at 200 MHz. HPLC analyses of tetrols 3 and 4
were carried out as previously described.7 Quantum-chemical calcula-
tions were performed with the molecular modeling program Titan, from
Wavefunction, Inc.-Schro¨dinger, Inc.
Scheme 2
Synthesis of Chlorohydrin 5. A solution of 0.40 mL of 4 M HCl
in dioxane and 0.2 mL of anhydrous THF was added to a solution of
5.4 mg of diol epoxide 1 in 0.80 mL of anhydrous THF. The resulting
solution was swirled and allowed to stand at room temperature for 2
min. The solvent and excess HCl were removed by rotary evaporation
at reduced pressure to yield 7.0 mg of residue. The 1H NMR spectrum
of the product, supplied as Supporting Information, indicated that there
was a clean conversion of 1 to chlorohydrin 5: 1H NMR (THF-d8) δ
6.11 (d, J ) 2.8 Hz, H-C(10)), 5.09 (d, J ) 9.1 Hz, H-C(7)), 4.55
(m, H-C(9)), 4.29 (dd, J ) 9.1, 2.1 Hz, H-C(8)). These NMR data
are very similar to those reported8 for 5 in DMSO-d6 and support the
trans 9,10-stereochemistry. The H-C(7)-C(8)-H coupling constant
of 9.1 Hz is very similar in magnitude to the 7,8-coupling constant of
the corresponding 7,8,9,10-tetrahydro-7,8,9,10-tetraacetate with a trans
9,10-stereochemistry.4 All coupling constants for the C(7)-C(10)
hydrogens of the 7,8,9,10-tetrahydro-7,8,9,10-tetraacetate with a cis 9,-
10-stereochemistry are small (2-3 Hz).4 The H-C(7)-C(8)-H
coupling constant of 9.1 Hz and the H-C(9)-C(10)-H coupling
constant of 2.1 Hz suggest that the principal conformation of 5 is 5a
(cf Scheme 5), in which the hydrogens on C(7) and C(8) are in a trans,
pseudo-diaxial geometry and the hydrogens on C(9) and C(10) are
pseudo-diequatorial.4,8 This product was used without further purifica-
tion for kinetic and product studies. The rate of reaction of 5 in 1:1
dioxane-water is independent of pH at pH < 9. At pH > 10, the rate
of reaction of 5 increases with the increase of pH. The fitting of these
rate data to the equation kobsd ) ko + kOH[HO-] yielded values of ko )
(2.7 ( 0.1) × 10-2 s-1 and kOH ) (1.40 ( 0.05) × 102 M-1 s-1. The
magnitude of the second-order rate constant for the reaction of 5 with
hydroxide ion at pH > 10 is consistent with that expected for the
reaction of a trans chlorohydrin with hydroxide ion to form an epoxide,14
in this case diol epoxide 1.8 A plot of kobsd versus pH for the reaction
of 5 in 1:1 dioxane-water is provided as Supporting Information.
Kinetic Procedures. For each kinetic run, approximately 5 µL of a
stock solution of 5 in dioxane (1 mg/mL) was added to 2.0 mL of 1:1
dioxane-water solution in the thermostated cell compartment (25.0 (
0.2 °C) of a UV-vis spectrophotometer. Reactions were monitored at
344 nm, and pseudo-first-order rate constants were calculated by
nonlinear regression analysis of the absorbance versus time data.
Product Studies. The products from the reaction of 5 in 10:90
dioxane-water solutions were determined by reverse phase HPLC
analyses of the reaction solutions after >10 half-lives on a C18 column
with 60:40 methanol-water as eluting solvent.7 Products were moni-
tored by UV detection at 269 nm. The relative yields of cis and trans
tetrols from the reaction of 5 in 10:90 dioxane-water solution
containing NaClO4 (0-0.5 M) at pH 4 remained constant at 78-79%
trans and 21-22% cis. The relative yield of cis tetrol from the reaction
carcinogenesis.6 At [Cl-] ) 10 mM, it can be calculated that
∼10% of carbocation 2 is trapped by chloride ion in water
solution to form 5.7
The observation that trans chlorohydrin 5 undergoes hydroly-
sis to yield more cis tetrol than the acid-catalyzed hydrolysis
of diol epoxide 1 in the absence of chloride ion is particularly
intriguing. The reaction of 1 with H+ and the SN1 reaction of
chlorohydrin 5 are expected to yield the same carbocation
intermediate (Scheme 2). If tetrol products are formed from this
intermediate with water, then the acid-catalyzed hydrolysis of
diol epoxide 1 and the hydrolysis of chlorohydrin 5 by the SN1
mechanism would yield the same tetrol mixture, contrary to the
actual observation. The hydrolysis of 5 in solutions containing
chloride ion is reported to be slower than in solutions without
chloride ion,8 which suggests that 5 reacts at least in part via a
carbocation intermediate. To explain the greater yields of cis
tetrol and of cis nucleic acid adducts from reactions of diol
epoxide 1 in the presence of chloride ion, it was proposed that
the intermediate trans chlorohydrin 5 undergoes SN2 reactions
with water to yield cis tetrol and with nucleic acids to yield cis
adducts.6,9,12
From plots of the cis:trans tetrol ratio from the reaction of 1
versus the concentration of chloride ion, it can be concluded
that part or all of the tetrols from hydrolysis of chlorohydrin 5
must be formed by a mechanism distinct from that for the
hydrolysis of 1.7 Another pathway for the reaction of 5 to form
tetrol(s) must exist. In addition to the SN2 reaction of 5 with
water proposed as one possible pathway,6,9,12 other possible
mechanisms for the hydrolysis of 5 include a rate-limiting
reaction of a carbocation-chloride ion pair with solvent and
the ionization of 5 to give a discrete carbocation having a
conformation different from that formed when 1 reacts with
H+.7 In this paper, we summarize rate data for the reactions of
trans chlorohydrin 5 in dioxane-water solutions containing
chloride ion and conclude that 5, like diol epoxide 1, hydrolyzes
(13) Yagi, H.; Hernandez, O.; Jerina, D. M. J. Am. Chem. Soc. 1975, 97, 6881-
6883.
(14) Whalen, D. L.; Ross, A. M. J. Am. Chem. Soc. 1976, 98, 7859-7861.
(12) Wolfe, A. R.; Song, Q.; Meehan, T. Polycyclic Aromatic Hydrocarbons
1996, 10, 203-210.
9
J. AM. CHEM. SOC. VOL. 124, NO. 48, 2002 14383

A R T I C L E S
Doan et al.
Scheme 4
Figure 1. Plot of kobsd vs [NaCl] for the reaction of 5 in 1:1 dioxane-
water solution, µ ) 0.2 M (NaClO4).
Scheme 3
thus fit a mechanism in which all of the tetrol product is derived
from the capture of a discrete carbocation intermediate by water
and rule out both a mechanism in which the cis tetrol (35%) is
formed from an SN2 reaction of trans chlorohydrin 5 with water
and a mechanism involving the reaction of an ion pair with
water. The latter reaction pathways would have rates that are
not depressed by added chloride ion, and if these latter pathways
were significant, then the rate profile for the reaction of 5 in
solutions containing chloride ion (Figure 1) would be asymptotic
to a nonzero rate constant. In summary, the tetrol-forming steps
from both the acid-catalyzed hydrolysis of diol epoxide 1 and
the hydrolysis of chlorohydrin 5 must involve only the attack
of water on discrete carbocation intermediates. Mechanisms for
tetrol formation involving the reactions of “intimate” or “solvent-
separated” ion pairs with water are ruled out, since increasing
chloride ion concentration would not reverse the formation of
such ion pairs.
of 5 in 10:90 dioxane-water increased from ∼21% when NaCl was
absent to 35% in 1.0 M NaCl solution.
The relative yields of cis tetrol from the acid-catalyzed hydrolysis
of diol epoxide 1 in water,1 10:90 dioxane-water,1,7 and 25:75
dioxane-water1 are reported to be 14-19%, 6-8%, and 4-5%,
respectively. The relative yield of cis tetrol from the acid-catalyzed
hydrolysis of 1 in 10:90 dioxane-water increases from 6% in the
absence of NaCl to 35% in 1.0 M NaCl solution.7
Results
Trans chlorohydrin 5 has been synthesized by the reaction
of diol epoxide 1 with lithium chloride and acetic acid in THF.8
In our hands, this procedure yielded a mixture of products that
were not readily separated and purified. In an alternate approach,
we have synthesized 5 in high yield by the reaction of 1 with
anhydrous HCl in THF. The product from this reaction was
sufficiently pure for kinetic and product studies without further
purification.
Discussion
A number of intriguing observations in the hydrolysis
reactions of 1 and 5 need to be rationalized: (1) the acid-
catalyzed hydrolysis of diol epoxide 1 and the hydrolysis of
chlorohydrin 5 yield tetrols with different cis/trans ratios, even
though each reaction proceeds via a discrete carbocation
intermediate and both reactions might be expected to yield
identical cis/trans product ratios; (2) the reactions of both diol
epoxide 1 and chlorohydrin 5 in solutions of high chloride ion
concentrations yield the same cis/trans ratio of tetrols; and (3)
the cis/trans tetrol ratio from the hydrolysis of chlorohydrin 5
in solutions with high chloride ion concentrations is somewhat
different than that from the hydrolysis of 5 in the absence of
chloride ion (Experimental Section).
Reaction of Diol Epoxide 1. The acid-catalyzed hydrolysis
of 1 occurs by rate-limiting epoxide ring opening to form a
carbocation 2, which has a sufficient lifetime to react with
external nucleophiles at rates greater than that with which it
reacts with water.5 The reaction of 1 with H+ from its more
stable ground-state conformation yields carbocation 2a, which
can undergo ring inversion to yield a second carbocation
conformation, 2b (Scheme 4). Carbocation conformations 2a
and 2b are substituted cyclohexenyl carbocations, and it has
been shown that the axial attack of water on cyclohexenyl
carbocations is favored over equatorial attack.15,16 A benzylic
carbocation that is constrained to a geometry related to 2a, which
contains an axial hydroxyl group adjacent to the carbocationic
The rate of reaction of 5 in 1:1 dioxane-water was
determined as a function of increasing concentration of sodium
chloride at constant ionic strength maintained by the addition
of sodium perchlorate, and a plot of these data is provided in
Figure 1. These data show a marked common ion rate depression
for the reaction of 5, which is consistent with the mechanism
outlined in Scheme 3. In this mechanism carbocation formation
(k1) is rate-limiting in the absence of chloride ion, and capture
of the carbocation by solvent (ks) is rate-limiting at high chloride
ion concentrations. This “common ion effect” demonstrates that
there is an intermediate in the reaction that is captured by
chloride ion and converted back to a reactant, thus slowing the
reaction. The rate of reaction of 5 in 0.2 M sodium chloride
solution is 7.5 times slower than that of its reaction in the
absence of chloride ion. With the assumption that carbocation
2 is a steady-state intermediate in the hydrolysis of 5, the rate
expression for the mechanism of Scheme 3 is given by eq 1.
The rate data in Figure 1 fit eq 1 and yielded values of 0.19 (
0.01 s-1 and 23 ( 2 M-1 for k1 and k-1/ks, respectively. The
kinetic data
k1
kobsd
)
(1)
(1 + k-1[Cl-]/ks)
(15) Goering, H. L.; Josephson, R. R. J. Am. Chem. Soc. 1962, 84, 2779-2785.
(16) Goering, H. L.; Vlazny, J. C. J. Am. Chem. Soc. 1979, 101, 1801-1805.
9
14384 J. AM. CHEM. SOC. VOL. 124, NO. 48, 2002

Cl Ion Catalyzed Interconversion of Carbocation Conformations
A R T I C L E S
carbon, yields >99% of trans diol by an axial attack of water.17
An isomeric benzylic carbocation that is constrained to a
geometry related to that of 2b, which contains an equatorial
hydroxyl group adjacent to the carbocationic center, yields
∼75% of cis diol from an axial attack of water and ∼25% of
trans diol from an equatorial attack of water.17 By analogy, an
axial attack of water on 2a to yield trans tetrol 3 should be
strongly favored. An axial attack of water on 2b to yield cis
tetrol 4 should also be favored, but to a lesser extent. An
equatorial attack of water on 2b to yield trans tetrol 3 might
also be expected to occur as a minor reaction pathway. The
observation that the acid-catalyzed hydrolysis of 1 yields mostly
trans tetrol (∼94%) can, therefore, be explained by either of
the following two mechanisms in which most of the tetrol
product is derived from an axial attack of water on 2a: (1) 2a
and 2b are in rapid equilibrium, and an attack of water on 2a
is energetically favored over attack of water on 2b or (2) epoxide
ring opening favors the formation of 2a, and the rate of attack
of water on 2a (ks) is greater than the rate of conformational
inversion of 2a to 2b (kf). We had previously17 favored the first
interpretation, since we had assumed that the conformation 2a
was preferred at equilibrium because it has one less gauche
butane interaction than 2b. An axial attack by water on the
predominant 2a rather than on the higher-energy 2b could then
provide a more favorable pathway to products, resulting in trans
tetrol. However, our present results (see below) are inconsistent
with mechanism (1) and support the alternative mechanism (2);
namely that 2a and 2b are of comparable stabilities, but initially
formed 2a is not in rapid equilibrium with 2b relative to solvent
attack.
Reaction of Chlorohydrin 5. The common ion rate depres-
sion for the hydrolysis of chlorohydrin 5 (Experimental Section)
fits a rate equation that is asymptotic to zero at high chloride
ion concentration and, thus, requires that all of the tetrol product
must be formed from the reaction of water with an intermediate
(i.e., a carbocation) whose stoichiometry does not include
chloride ion and, thus, cannot involve any type of ion pair. If
the reactions of both 1 and 5 form the same carbocation and
tetrol products are formed from this carbocation, then the tetrol
products formed from the hydrolyses of 1 and 5 should be
identical. Yet, the ratio of cis 9,10- and trans 9,10-tetrols from
reactions of 1 and 5 are different. Thus, there must be two
distinct populations of the carbocation intermediate that inter-
convert more slowly than they react with solvent. These two
populations could consist of different solvation states or different
conformations. We first considered that the hydrolysis of
chlorohydrin 5 and the acid-catalyzed hydrolysis of diol epoxide
1 give carbocations in different solvation states that could
possibly yield tetrols with different cis:trans ratios. The rate of
reaction of carbocation 2 with water has been estimated to be
formed from the reaction of 1 with H+ and that the energy
barrier for the interconversion of these conformations must be
greater than the energy barrier for the reaction of each
carbocation conformation with water.19
From the partitioning of carbocation 2 between the reaction
with water and the addition of azide ion, with the assumption
that azide ion reacts with 2 at the diffusion-limited rate constant
of 5 × 109 s-1, a value of ks for the reaction of 2 with water is
estimated to be ∼2 × 107 s-1.5 From this rate constant, the
energy barrier for the reaction of 2 with water is calculated to
be 7.5 kcal/mol. In a preliminary calculational study, gas-phase
structures of carbocation conformations 2a and b and a transition
structure for their interconversion have been calculated at the
ab initio B3LYP/6-31G* level of theory.20 Conformation 2a is
calculated to be 0.03 kcal/mol more stable than conformation
2b, and the energy difference between the transition structure
and conformation 2a is calculated to be 7.92 kcal/mol.20
Although differences in solvation effects, zero point energy, and
temperature were not taken into account, the calculated struc-
tures are consistent with our proposal that carbocation confor-
mations 2a and b have similar energies and that the barrier to
their interconversion is greater than the energy barriers for their
reactions with solvent.
Mechanistic Considerations. At chloride ion concentrations
> ∼0.1 M, the rate of reaction of chloride ion with carbocation
2 exceeds the rate of reaction of water with 2.7 The reactions
of diol epoxide 1with HCl/THF or LiCl/HOAc yield only trans
chlorohydrin 5, and this chlorohydrin is most likely the major
product from the reaction of 2 with chloride ion in water solution
and may be the intermediate responsible for the different cis/
trans tetrol ratio from the hydrolysis of 1 when chloride ion is
present in solution. However, the intermediacy of a cis chlo-
rohydrin in the hydrolysis of 1 cannot be ruled out. Regardless
of the cis/trans composition of chlorohydrin intermediates in
the hydrolysis of diol epoxide 1 in the presence of chloride ion,
the relative amounts of carbocation conformations 2a and b from
chlorohydrin hydrolysis must be different than that formed from
the acid-catalyzed epoxide ring opening of diol epoxide 1.
One possible mechanism for the acid-catalyzed hydrolysis
of diol epoxide 1 in the presence of chloride ion involving only
trans chlorohydrin 5 as an intermediate is provided in Scheme
5. If cis chlorohydrin is also an intermediate, then the mechanism
will be more complicated. In the mechanism of Scheme 5, the
energy barrier for interconversion of carbocation conformations
2a and b is greater than the energy barriers for the reaction of
each carbocation conformation with water. Diol epoxide 1 reacts
with H+ primarily from its more stable ground-state conforma-
tion to form carbocation 2a. When chloride ion is not present
in solution, conformation 2a reacts with water to yield trans
tetrol faster than it undergoes conformational isomerization to
2b. However, it has been established that, if chloride ion is
present in sufficient concentration, then the carbocation inter-
mediate (i.e., conformation 2a) reacts with chloride ion to yield
trans chlorohydrin 5 faster than it reacts with solvent.7
∼2 × 107 s-1,5 whereas solvent relaxation occurs within 10-13
-
10-11 s,18 which is many orders of magnitude faster than the
rate at which 2 reacts with solvent. Thus, we conclude that
different solvation states of a carbocation would equilibrate too
rapidly to account for the observed product differences. We are,
therefore, left to conclude that the ionization of 5 yields a
different distribution of carbocation conformations than that
(19) In the acid-catalyzed methanolysis of K-region arene oxides, a mechanism
in which the rate of solvent capture of a carbocation intermediate is
comparable to that of its conformational inversion is proposed: Nashed,
N. T.; Bax, A.; Loncharich, R. J.; Sayer, J. M., Jerina, D. M. J. Am. Chem.
Soc. 1993, 115, 1711-1722.
(20) Calculated structures, energies,Cartesian coordinates of 2a and b and a
transition structure for their interconversion are provided as Supporting
Information.
(17) Sayer, J. M.; Yagi, H.; Silverton, J. V.; Friedman, S. L.; Whalen, D. L.;
Jerina, D. M. J. Am. Chem. Soc. 1982, 104, 1972-1978.
(18) (a) Billing, G. D.; Mikkelsen, K. V. Molecular Dynamics and Chemical
Kinetics; John Wiley & Sons: New York, 1996; p 108. (b) Castner, E.
W., Jr.; Maroncelli, M.; Fleming, R. J. Chem. Phys. 1987, 86, 1090-1097.
9
J. AM. CHEM. SOC. VOL. 124, NO. 48, 2002 14385

A R T I C L E S
Doan et al.
Scheme 5
Figure 2. Plots of free energy (E) vs Reaction Coordinate representing
the direct and chloride ion-catalyzed interconversions of carbocations 2a
and b and the reactions of 2a and b with water as outlined in Scheme 4
under conditions of high chloride ion concentrations. The reaction of diol
epoxide 1 with H+ to yield 2a is not shown.
Quantum chemical calculations indicate that there are two
principal conformations of chlorohydrin 5, one in which
hydroxyl groups occupy equatorial positions and the C-Cl bond
is oriented 74° out of the plane of the aromatic system (5a)
and a second in which the two hydroxyl groups occupy axial
positions and the C-Cl bond is oriented 59° out of the plane
of the aromatic system (5b).21 Therefore, ionization of 5a to
form carbocation 2a and ionization of 5b to form carbocation
2b can each occur without large changes in molecular geometry.
The capture of carbocation 2a by chloride ion, conformational
inversion of 5a to b, and ionization of 5b to form carbocation
2b, therefore, provide one of several possible indirect routes
for carbocation inversion.
NMR coupling constants indicate that 5 exists primarily as
conformation 5a (see Experimental Section), and this conclusion
is supported by quantum chemical calculations. Conformation
5a is calculated at the HF/6-31G* level of theory to be 4.66
kcal/mol more stable than conformation 5b, and a transition
structure for their interconversion has been calculated to be 11.52
kcal/mol less stable than that of 5a. This calculated energy
barrier for the conversion of 5a to b is much less than the free
energy for the hydrolysis of 5 (calculated to be 18.4 kcal/mol
from rate data), and therefore, these two conformations should
be at equilibrium during the hydrolysis of 5. The relative rates
of formation of carbocation conformations 2a and b from
chlorohydrin 5 do not depend on the energy difference between
5a and b, however, but rather on the energy difference between
the transition states for the formation of 2a and b from the
reaction of 5 (Curtin-Hammett principle).22
be in equilibrium because of the thermodynamic cycle outlined
in Scheme 5 or in a similar thermodynamic cycle if the cis
chlorohydrin is an intermediate. The cis/trans tetrol product ratio
will then be determined by the transition-state energy difference
for an attack of water on the two carbocation conformations 2a
and b (Curtin-Hammett principle)22 and will be independent
of the nature of the halide ion or the cis/trans stereochemistry
of chlorohydrin intermediates. The fact that the yields of cis
tetrol from the acid-catalyzed hydrolysis of 1 in solutions of
chloride, bromide, and iodide ions, extrapolated to infinite halide
ion concentrations, are identical within experimental error7
supports this conclusion and provides evidence that bromide
and iodide ions also catalyze the interconversion of 2a and b.
A free energy diagram that illustrates in a qualitative way
the mechanism by which chloride ion catalyzes the intercon-
version of carbocation conformations 2a and b by way of trans
chlorohydrin 5 is given in Figure 2. In this diagram, the energy
of the transition state for the interconversion of 2a and b (TS2)
must be greater than the transition-state energies for reactions
of 2a and b with water, TST(H2O), and TSC(H2O), respectively.
In the presence of sufficient concentrations of chloride ion,
however, the rate of reaction of 2a with chloride ion to form
5a by an axial attack of chloride on 2a exceeds that of 2a with
water. It has been determined from a previous study that the
minimum value of k-1/ks in 10:90 dioxane-water is 12 M-1 7
,
and the common ion rate depression for reaction of 5 in 50:50
dioxane-water (Experimental Section) yields a value of 23 M-1
for k-1/ks. From the analysis of the 1H NMR coupling constants
of 5 (Experimental Section), conformation 5a is more stable
than conformation 5b. The free energy for the conformational
inversion of 5a to 5b is expected to be equal to or greater than
that for the conformational inversion of carbocation 2a to
carbocation 2b. However, chlorohydrin 5 is much more stable
than carbocation 2 plus chloride ion under these conditions,23
and the lower energy pathway from 2a to 2b will be via
The observation that chlorohydrin 5 yields cis and trans tetrols
in a ratio different than that formed from the acid-catalyzed
hydrolysis of diol epoxide 1 requires that the rate of intercon-
version of 2a and b must be slow relative to the rates of their
reactions with water (ks and ks′, respectively). Therefore, during
the hydrolysis of 1 in the absence of chloride ion, the two
carbocation conformations, 2a and b, are not at equilibrium.
They will be at equilibrium during the hydrolyses of diol epoxide
1 and chlorohydrin 5 in solutions containing high concentrations
of chloride ion, however. Under this condition, 2a and b must
(21) Calculated structures, energies, Cartesian coordinates of 5a and b and a
transition structure for the interconversion of 5a and b are provided as
Supporting Information.
(22) (a) Curtin, D. Y. Rec. Chem. Prog. 1954, 15, 111. (b) Eliel, E. L.
Stereochemistry of Carbon Compounds; McGraw-Hill: New York, 1962;
pp 151-152.
(23) With values of 23 M-1 for k-1/ks (Scheme 3, this study) and 2 × 107 s-1
for ks (ref 5), the rate constant for the reaction of carbocation 2 with 1 M
Cl- is ∼108 s-1. The rate constant for the ionization of 5 to form carbocation
2 in 1:1 dioxane-water is ∼0.2 s-1. The equilibrium between 5 and
carbocation 2 under the conditions of this study, therefore, greatly favors
the chlorohydrin.
9
14386 J. AM. CHEM. SOC. VOL. 124, NO. 48, 2002

Cl Ion Catalyzed Interconversion of Carbocation Conformations
A R T I C L E S
chlorohydrin 5. Thus, chloride ion catalyzes the interconversion
of 2a and b.
istry of the reactions of nucleophiles other than water on
carbocation conformations 2a and b are in progress.
A second possible mechanism for the chloride ion-catalyzed
interconversion of carbocation conformations 2a and b involves
a cis chlorohydrin intermediate. This intermediate could be
formed by an “equatorial” attack of chloride ion on carbocation
2a, an “axial” attack of chloride ion on 2b, or by an SN2 reaction
of chloride ion with trans chlorohydrin 5. Although the cis
chlorohydrin may be sufficiently less stable than the trans
chlorohydrin so that it cannot be detected in the reaction mixture,
it cannot be ruled out as an intermediate in the conformational
isomerization of 2a to 2b.
The hydrolysis of chlorohydrin 5 in 10:90 dioxane-water
that does not contain external chloride ion yields trans and cis
tetrols in a 79:21 ratio. This ratio is somewhat different than
the ratio of trans and cis tetrols (65:35) formed from reaction
of either 5 or 1 in solutions of high chloride ion concentrations.
This observation is also consistent with the mechanism of
Scheme 5. At high chloride ion concentrations, the cis:trans
tetrol ratio depends on the transition-state energy difference for
an attack of water on carbocation conformations 2a and b.
However, in the absence of chloride ion, the ratio of tetrols will
depend on the transition-state energy difference for the formation
of carbocation conformations 2a and b. This will be true because
carbocation formation is rate-limiting for the reaction of 5 in
the absence of added chloride ion, and each carbocation
conformation reacts with water faster than it isomerizes to the
other conformation. The fact that a significant yield of cis tetrol
(20%) is formed from the hydrolysis of trans chlorohydrin 5 in
the absence of added external chloride ion, where carbocation
formation is rate-limiting and cis chlorohydrin is unlikely to
form by any of the mechanisms listed above, suggests that the
major pathway leading to carbocation conformation 2b is via
the trans chlorohydrin.
Summary
We propose that the carbocation intermediate formed from
the acid-catalyzed hydrolysis of diol epoxide 1 and from the
hydrolysis of chlorohydrin 5 exists in two conformations, with
an energy barrier to interconversion that is greater than the
energy barriers for their reactions with water. This is an example
where rate and product studies clearly demonstrate that two
discrete carbocation conformations are not at conformational
equilibrium in water solution. However, the interconversion of
these two carbocation conformations is catalyzed by chloride
ion, and in water solutions with high chloride ion concentrations,
the two conformations do interconvert rapidly relative to the
rates at which they react with water. The ratio of cis 9,10- and
trans 9,10-tetrols formed from both the reactions of 1 and 5
depends only on the transition-state energy difference for the
attack of water on the two carbocation conformations when the
concentration of chloride ion is sufficiently high, so that the
two carbocation conformations are at equilibrium. However, the
ratio of cis and trans tetrols formed from the hydrolysis of
chlorohydrin 5 in the absence of external chloride ion depends
on the transition-state energy difference for the formation of
the two carbocation conformations.
Acknowledgment. Helpful discussions with Dr. Jane Sayer
(NIH) are appreciated.
Supporting Information Available: Figure with plot of log
kobsd versus pH for the reaction of 7â,8R,9R-trihydroxy-10â-
chloro-7,8,9,10-tetrahydrobenzo[a]pyrene (chlorohydrin 5) in 1:1
dioxane-water, 0.2 M NaCl; H1 NMR spectrum of chlorohydrin
5; calculated structures and Cartesian coordinates of carbocation
conformations 2a and b (ab initio B3LYP/6-31G*); calculated
transition structure for the interconversion of 2a and b (ab initio
B3LYP/6-31G*); calculated structures and Cartesian coordinates
of chlorohydrin conformations 5a and b (ab initio HF/6-31G*);
calculated transition structure for the interconversion of 5a and
b (ab initio HF/6-31G*). This material is available free of charge
via the Internet at http://pubs.acs.org.
If carbocation conformation 2b undergoes energetically
favorable axial attack by water to yield cis tetrol as the major
product, then an axial attack of other similar nucleophiles on
2b leading to cis products should also be observed. The
observations that chlorohydrin 5 reacts with deoxyadenosine
and DNA to yield mostly cis adducts8 might, therefore, be due
to a mechanism involving an axial attack of the nucleic acid
base on conformation 2b. Studies to determine the stereochem-
JA020436+
9
J. AM. CHEM. SOC. VOL. 124, NO. 48, 2002 14387