Kinetic evidence for dihapto (η2) π-aryl participation in acid-catalyzed ring opening of diarylhomobenzoquinone epoxides

Article abstract of DOI:10.1021/jo902179x

(Chemical Equation Presented) The BF₃-catalyzed ring-opening reaction of variously endo/exo m- and p-substituted diarylhomobenzoquinone epoxides proceeded through a transannular S_E2-Ar cyclization of endo-aryl groups to give the tric

Full text of DOI:10.1021/jo902179x

pubs.acs.org/joc
Kinetic Evidence for Dihapto (η2) π-Aryl Participation in Acid-Catalyzed
Ring Opening of Diarylhomobenzoquinone Epoxides
Haruyasu Asahara, Kohta Saito, Naohiko Ikuma, and Takumi Oshima*
Department of Applied Chemistry, Graduate School of Engineering, Osaka University, 2-1 Yamadaoka,
Suita, Osaka 565-0871, Japan
oshima@chem.eng.osaka-u.ac.jp
Received October 19, 2009
The BF3-catalyzed ring-opening reaction of variously endo/exo m- and p-substituted diarylhomobenzoquinone
epoxides proceeded through a transannular SE2-Ar cyclization of endo-aryl groups to give the tricyclic diketo-
alcohols and cyclohexadienone spiro-linked tricyclic diketo-alcohols. Kinetics of these reactions has been
investigated in CDCl3 at 30 °C in order to elucidate the possible remote π-aryl participation. The rates were
significantly increased with increasing electron-donating ability of the endo-aryl substituents X (kp-MeO/kp-CF3
=
8200) but only negligibly influenced by the distal exo-aryl substituents Y (kp-MeO/kp-CF3 =2.1). For the endo-X
substituted series, an excellent linear free energy relationship, log krelendo=-2.49σipso - 1.62σortho - 0.108 (R2 =
0.98, n=8), was attained using two modified site-dependent substituent parameters σipso (usingσpþ for p-X and
σm for m-X) and σortho (using σm for p-X and σpþ for m-X). This means that the dihapto(η2) π-coordination
occurs in the π-aryl participation, with the ipso π-electron donation contributing 1.6 times more effectively than
the ortho one. On the other hand, the distal exo-Y substituted series gave an acceptable Yukawa-Tsuno equa-
tion with small polar and resonance contributions; log krelexo = -0.912(σ0 þ 0.237ΔσhRþ) (R2 = 0.96, n = 8).
These kinetic substituent effects were compared with those of the acid-catalyzed π-aryl assisted transannular SE2-
Ar cyclization of the cyclobutene-fused diarylhomobenzoquinones. It was found that the geometrical
characteristics of the vacant oxirane Walsh orbital and the cyclobutene antibonding orbital play a crucial role
in the topological features of η2 π-aryl participation.
Introduction
the contained aryl groups to the reaction center (usually but
not necessarily an incipient carbocation).2 A large number of
studies have been made of the π-aryl-assisted solvolyses of β-
aryltosylates and brosylates from the kinetic3 and stereo-
chemical4 point of view. In these adjacent β-aryl-substituted
substrates, the kinetic substituent effects as well as the
π-Aryl participation is one of the most sophisticated
physicochemical phenomena that control the reactivity of
substrates and sometimes govern the reaction mechanism.1
These effects are generally derived from (ascribed to) the
through-space electronic stabilization of the transition states
by the direct π-electronic donation (not by resonance) from
(2) Isaacs, N. S. Physical Organic Chemistry; Longman Science & Tech-
nical: Essex, 1995; Chapter 13, pp 643-700.
(1) (a) Lancelot, C. J.; Cram, D. J.; Scheyer, P. v. R. Carbenium Ions; Olah,
G. A., Scheyer, P. v. R., Eds.; Wiley-Interscience: New York, 1972; Vol. 3, Chapter
27, pp 1347-1483. (b) Smith, M. B.; March, J. March’s Advanced Organic
Chemistry; Wiley: New York, 2007; Chapter 11, pp 657-751. (c) Peeran, M. J.;
Wilt, W.; Subramanian, R.; Crumrine, D. S. J. Org. Chem. 1993, 58, 202–210.
(d) Kevil, D. N.; D'Souza, M. J. J. Chem. Soc., Perkin Trans. 2 1997, 257–263.
(e) Fujio, M.; Goto, N.; Dairokuno, T.; Goto, M.; Saeki, Y.; Okusaka, Y.; Tsuno, Y.
Bull. Chem. Soc. Jpn. 1992, 65, 3072–3079. (f) Nagumo, S.; Ono, M.; Kakimoto,
Y.; Furukawa, T.; Hisano, T.; Mizukami, M.; Kawahara, N.; Akita, H. J. Org.
(3) (a) Fujio, M.; Funatsu, K.; Goto, M.; Mishima, M.; Tsuno, Y.
Tetrahedron 1987, 43, 307–316. (b) Fujio, M.; Goto, M.; Mishima, M.;
Tsuno, Y. Bull. Chem. Soc. Jpn. 1990, 63, 1121–1128. (c) Okamura, M.;
Hazama, K.; Ohta, M.; Kato, K.; Horaguchi, T.; Ohno, A. Chem. Lett. 1997,
26, 973–974. (d) Tsuno, Y.; Fujino, M. Advances in Physical Organic Chemistry;
Bethell, D., Ed.; Academic Press: London, 1999; Vol. 32, pp 267-385.
(4) (a) Cram, D. J. J. Am. Chem. Soc. 1952, 74, 2129–2137. (b) Cram, D. J.
J. Am. Chem. Soc. 1952, 74, 2137–2148. (c) Streitwieser, A., Jr.; Walsh, T. D.;
Wolfe, J. R. J. Am. Chem. Soc. 1965, 87, 3686–3691. (d) Cram, D. J.;
Thompson, J. A. J. Am. Chem. Soc. 1967, 89, 6766–6768. (e) Thompson,
J. A.; Cram, D. J. J. Am. Chem. Soc. 1969, 91, 1778–1789.
ꢀ
ꢀ
Chem. 2002, 67, 6618–6622. (g) del Río, E.; Menendez, M. I.; Lopez, R.; Sordo,
T. J. Am. Chem. Soc. 2001, 123, 5064–5068.
DOI: 10.1021/jo902179x
r
Published on Web 12/21/2009
J. Org. Chem. 2010, 75, 733–740 733
2009 American Chemical Society

Asahara et al.
JOCArticle
SCHEME 1
aromatic ring and how the aryl group interacts with the
vacant oxirane Walsh orbital.7
Herein, aiming at obtaining physicochemical information
on the π-aryl participated transition state, we have con-
ducted kinetic studies of the BF3-catalyzed transannular
cyclization of variously endo/exo m- and p-substituted dia-
rylhomobenzoquinone epoxides 1a-k in CDCl3 at 30 °C. On
the basis of the kinetic substituent effects, we have found that
the dihapto (η2)-type coordination (through ipso and ortho)
of the endo-aromatic ring occurs in the rate-determining
oxirane ring cleavage. A modified Hammett treatment in-
dicated that the ipso/ortho contribution ratio (1.6) is con-
siderably higher than that (0.89) of the similar SE2-Ar
transannular cyclization of cyclobutene-fused diarylhomo-
benzoquinones 4 into the ortho-carbon bound tetracyclic
keto-alcohols 5 (Scheme 1b).8 These results were explained in
terms of the stereoelectronic effects in the orbital interaction
between the π-donor aromatic rings (HOMO) and the vacant
anti-bonding orbitals of the cleaved oxirane and cyclobutene
rings (LUMO).
Results and Discussion
Synthesis and Structural Characteristics of Epoxides. The
endo/exo-substituted diarylhomobenzoquinone epoxides
1a-k were prepared by the epoxidation of diarylhomoben-
zoquinones9 derived from 1,3-dipolar cycloaddition of the
corresponding m- and p-substituted diphenyldiazomethanes
with 2,5-dimethyl-1,4-benzoquinone.10 For the monosubsti-
tuted diphenyldiazomethanes, a mixture of endo/exo-substi-
tuted isomers was formed and separated by careful column
chromatography and recrystallization. The m-chloro-substi-
tuted endo/exo-1i, however, could not be isolated due to the
very similar endo/exo adsorbability on silicagel. The struc-
tures of 1a-k were deduced as the cis-transoid-cis tricyclic
dione frame from the representative X-ray crystal structural
analysis of bis(p-chlorophenyl)-substituted 1j (Figure 1).11
The original quinone frame was found to adopt an appreci-
able pseudoboat conformation with a slight deformation
opposite to the endo-aromatic ring. This deformation results
in an overhang of endo-aryl ring above the quinone plane and
a pseudoaxial flipping of the conjunct planar oxirane ring as
noted by the average dihedral angle (95.4°) between the
epoxide ring and the fused quinone plane. The very short
product analyses clearly indicated that the π-aryl participa-
tion occurs at the proximal ipso-carbon and hence the
phenonium ion (σ-complex) can be formed as an intermedi-
ate. By contrast, little is known for the anchimeric assistance
of the aryl group located in the γ-position or further away
from the reaction site.5 However, the remote π-aryl partici-
pation is of continuous theoretical interest because a detailed
understanding of the physical origin and scope of such
interactions has become one of the major goals of physical
organic chemistry. Compared to the more conventional
π-aryl participation at the ipso-position, the elucidation of
the steric and electronic factors controlling the remote
electron-donation and knowing which positions of the aro-
matic ring are responsible for such phenomenon are of prime
importance.
Very recently, we reported that the unsubstituted and
p-substituted diphenylhomobenzoquinone epoxides 1 (X,
Y=H, Me, and Cl) show a remote π-aryl participated trans-
annular SE2-Ar cyclization of endo-phenyl group associated
with the acid-catalyzed regioselective oxirane ring cleavage
to give the ortho-carbon bound tricyclic diketo-alcohols
2 and the ipso-carbon bound 2,5-cyclohexadien-4-one
spiro-linked tricyclic diketo-alcohol 3 (only for the π-reso-
nating p-Cl substituted entity) as shown in Scheme 1a.6 This
observation raises a question as to which positions of the
˚
spatial distance (3.21 A) between the endo-aromatic ipso-
carbon and the cleaved epoxide tertiary carbon seems to
allow the π-electron donation to the underlying acid-acti-
vated oxirane ring. However, the corresponding distance
˚
(4.74 A) with the outside exo-aromatic ipso-carbon is long
enough to inhibit the through-space π-electron donation.
Such π-electron participation becomes very important if the
(7) (a) Mollere, P. D.; Houk, K. N. J. Am. Chem. Soc. 1977, 99, 3226–
3333. (b) Iwamura, H.; Sugawara, T.; Kawada, Y.; Tori, K.; Muneyuki, R.;
Noyori, R. Tetrahedron Lett. 1979, 20, 3449–3452. (c) Bach, R. D.; Wolber,
G. J. J. Am. Chem. Soc. 1984, 106, 1410–1415.
(8) Koizumi, T.; Harada, K.; Mochizuki, E.; Kokubo, K.; Oshima, T.
Org. Lett. 2004, 6, 4081–4084.
(9) Miyashita, M.; Suzuki, T.; Yoshikoshi, A. Chem. Lett. 1987, 285–288.
(10) Oshima, T.; Nagai, T. Bull. Chem. Soc. Jpn. 1988, 61, 2507–2512.
(11) Crystallographic data for this compound have been deposited at the
Cambridge Crystallographic Data Centre. These data can be obtained free of
charge via www.ccdc.cam.ac.uk/data-request/cif, by emailing data_re-
quest@ccdc.cam.ac.uk, or by contacting The Cambridge Crystallographic
Data Centre, 12 Union Road, Cambridge CB2 1EZ, U.K.; fax: þ44
1223336033. CCDC 749514.
(5) (a) Jackman, L. M.; Haddon, V. R. J. Am. Chem. Soc. 1974, 96, 5130–
5138. (b) Gates, M.; Frank, D. L.; Felten, W. C. J. Am. Chem. Soc. 1974, 96,
5138–5143. (c) Ando, T.; Yamawaki, J.; Saito, Y. Bull. Chem. Soc. Jpn. 1978,
51, 219–224.
(6) Oshima, T.; Asahara, H.; Koizumi, T.; Miyamoto, S. Chem. Commun.
2008, 15, 1804–1806.
734 J. Org. Chem. Vol. 75, No. 3, 2010

Asahara et al.
JOCArticle
TABLE 1. Rate Constants k and Relative Rate Ratios krel(1) for BF3-
Catalyzed Rearrangements of 1 and Relative Rate Ratios krel(4) for the
Reaction of 4 in CDCl3 at 30°C
substituent
103 ꢀ k(1)a,
entry compound
X
Y
M
-1 s-1
krel(1) krel(4)b
37 17
11 5.3
8.0
1
endo-1a
1b
1c
p-MeO
p-Me
m-MeO m-MeO
H
p-Me
18.7
5.25
3.90
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
endo-1d
endo-1e
endo-1f
1g
endo-1h
endo-1i
1j
endo-1k
exo-1a
exo-1d
exo-1e
exo-1f
exo-1h
exo-1i
exo-1k
m-MeO
p-Me
m-Me
H
p-Cl
m-Cl
p-Cl
p-CF3
H
H
H
H
H
H
H
H
H
H
H
p-Cl
H
p-MeO
m-MeO
p-Me
m-Me
p-Cl
m-Cl
p-CF3
3.70
3.11
1.36
7.6
6.4
2.8
1.0
0.25
0.044
0.032
22
4.9
5.6
1.0c
0.017
0.078
0.015
0.486
0.122
0.0213
0.0154
0.00217
0.602
0.456
0.523
0.478
0.420
0.345
0.278
0.0045 0.0040
3.2
1.2
0.94
1.1
0.98
0.86
0.71
0.57
0.94
1.2
0.70
0.39
0.33
0.30
FIGURE 1. ORTEP drawing of bis( p-chlorophenyl) homobenzo-
quinone epoxide 1j with 50% ellipsoid probability.
H
H
present epoxide ring-opening reaction proceeds through an
SN2-type intramolecular transannular cyclization. As a re-
sult, the geometrical features of 1 can take advantage of
π-anchimeric assistance in the present acid-catalyzed ring
opening.
Kinetic Substituent Effects. The observed pseudo-first-
order rate constants kobs(1) were divided by the concentration
of BF3 to provide the second-order rate constants k(1). The
values for k(1) and the relative rate constants krel(1) (vs the
unsubstituted 1g) are collected in Table 1. For comparison,
Table 1 also includes the relative rate constants krel(4) (vs the
unsubstituted entity) for the previous BF3-catalyzed reactions
of similarly endo/exo m- and p-substituted analogous cyclo-
butene-fused diarylhomobenzoquinones 4a-k.8
aThe second-order rate constants k were obtained by dividing
the pseudo-first-order rate constant kobs by the concentration of BF3.
The k values are the average of at least two measurements. Error limit
of k is (2%. bThe relative rate constant of cyclobutene ring-opening
c
reaction of analogous homobenzoquinones 4. The second-order rate
constant k(4g) =1.25 ꢀ10-3 M-1 s-1 (cf ref 8).
to the acid-activated oxirane ring. In order to evaluate the
degree to which the present acid-catalyzed ring-opening
reactions respond to the meta and para substituent changes
of 1, we applied a linear free energy relationships analysis for
krel(1). As a consequence, it was found that the usual
Hammett treatment for the endo-X substituted series by
using σ or σþ single parameter13 provided rather poor
correlations; log krelendo(1) = -4.5σ þ 0.35 (r2 = 0.89, n = 8)
and -2.8σþ - 0.09 (r2 =0.77, n=8) respectively.14 These
unsatisfactory correlations with the normal single parameter
are suggestive of π-aryl participation of the endo-aromatic
ring not only at the ipso- but also at the ortho-carbon atom.
This simultaneous binding seems to be responsible for the
formation of the ortho- and the ipso-carbon bound products,
2j and 3j, from the bis(p-chlorophenyl)-substituted 1j
(Scheme 1a).6
On the basis of these findings, we thought that the
transannular SE2-Ar cyclization of endo-aryl groups of 1
involves a dihapto (η2) π-aryl participated transition state
through the ipso- and the ortho-carbon atom. Although the
m-substituted aromatic nucleus has two different ortho-
positions with respect to the ipso-carbon, the less congested
ortho-position, that is to say para to the introduced m-
substituent, is expected to participate in the product forma-
tion (vide infra). Accordingly, we attempted to correlate the
log krelendo(1) with the combination of newly defined two site-
dependent substituent parameters σipso and σortho, which
display the substituent effects on the ipso- and ortho-aro-
matic carbon atoms, respectively. Thus, the substituent
A perusal survey of Table 1 indicated some noticeable
differences between the epoxides 1a-k and the cyclobutene-
fused 4a-k: (1) for the endo-substituted series, the rates were
increased more effectively in 1 (k p-MeO/kp-CF3=8200) than
12
in 4 (5600); (2) by contrast, the same substituent change
brought about only a small increase for the exo-substituted
1 (2-fold) and for 4 (11-fold); (3) of particular interest is
that the endo-p-substituted 1 (X = MeO, Me, Cl) reacted
faster than the endo- m-substituted isomer, whereas the endo-
m-substituted 4 exhibited reactivity rather higher than that
of the endo-p-substituted isomer; (4) on the other hand, exo-
m/ p-substituted couples showed the usual substituent effects
for both 1 and 4.
Since the endo-aromatic ring is located at the δ-position
with respect to the epoxide carbon atom, the notable rate-
acceleration by electron-donating m- and p-substituents can
not be ascribed to direct induction or the resonance stabili-
zation of a rate-determining transition state. Indeed, the exo-
aromatic substituents lacked drastic effects on the rates.
Accordingly, the remarkable rate dependency on the endo-
substituents must be explained by considering the through-
space electron donation of the overhanging aromatic nucleus
(12) Our attempt to follow the rate of the reaction of di(p-anisyl)-
homobenzoquinone epoxide by NMR failed because of the very fast degrada-
tion. Interestingly, this reaction gave neither 2- nor 3-type product but
yielded di(anisyl)methyl-substituted benzoquinone derivative 6, which
seems to arise from the cyclopropane ring cleavage as well as the loss of
oxygen atom. The structure of 6 was determined by the X-ray crystal
structural analysis (see Supporting Information). CCDC 750112.
(13) Williams, A. Free Energy Relationships in Organic and Bio-organic
Chemistry; The Royal Society of Chemistry: Cambridge, 2003.
(14) We obtained rather worse correlations when we assumed the π-aryl
participation occurs only at the ortho-carbon; log krelexo(1) = -0.38σ - 0.01
(R2 = 0.96, n = 8) and -0.24σþ - 0.05 (R2 = 0.82, n = 8).
J. Org. Chem. Vol. 75, No. 3, 2010 735

Asahara et al.
JOCArticle
phenonium ion intermediates via the π-aryl participation
at the ipso-carbon.
exo
log krel
¼ -0:912ðσ0 þ 0:237ΔσRþÞ ðR2 ¼ 0:96, n ¼ 8Þ
ð2Þ
In the following section, we will discuss these kinetic
substituent effects from a mechanistic viewpoint in compar-
ison with those of the analogous SE2-Ar transannular cycli-
zation of 4a-k.
Mechanistic Considerations. (1). π-Aryl Participation in
SN2-Type Ring Opening of Epoxide. The acid-catalyzed
cleavage of the oxirane C-O bond preferentially occurs in
such a way that the substituent of the oxirane ring further
stabilizes the developing positive charge by both electron
donation and conjugation.17 Accordingly, the reaction fea-
tures are highly dependent on the substituents on the oxirane
ring as well as the nucleophiles. Mechanistically, the nucleo-
philic displacement of epoxides can be rationalized by either
concerted attack of a nucleophile on the acid-activated
oxirane ring (SN2 mechanism) or by stepwise attack of a
nucleophile on a preformed carbocation intermediate
(SN1 mechanism).17b,e,18 Most of the reactions are generally
known to proceed through the regio- and stereoselective ring
cleavage under SN2 anti nucleophilic assistance.19 However,
the SN1 mechanism is reported for the reaction of epoxides
bearing strongly electron-releasing and conjugated substit-
uents in highly polar solvents.20
Since the present epoxide ring is substituted by two
electron-withdrawing quinone carbonyl groups, the acid-
catalyzed ring cleavage of 1 seems to obey the SN2 process
in the less polar CDCl3.21 Under these conditions, the over-
hanging endo-aromatic ring is apt to behave as an effective
internal π-nucleophile. Indeed, we have obtained very poor
kinetic solvent effects (kdichloroethane/kbenzene =3.0) regard-
less of the appreciable change of solvent polarity (as indi-
cated by the polarity parameter; ET(30) = 41.3 for 1,2-
dichloroethane and 34.3 for benzene)22 for the same trans-
annular cyclization of 1g by protic acid MeSO3H.6 The
negligible solvent effects can be taken as a conclusive me-
chanistic criterion for the concerted process involving the less
polar transition state.21
FIGURE 2. Plots of the observed log krelendo(1)(obs) vs the calucu-
lated log krelendo(1)(calc) according to eq 1 for BF3-catalyzed
rearrangements of unsubstituted 1g and endo-substituted 1a, d-f,
h, i, and k in CDCl3 at 30 °C.
coþnstant σipso is related to the ipso-participation by using
σp for p-X and σm for m-X substituent.13 On the other
hand, the other σortho refers to the ortho-participation
þ
by using, specifically, σm for p-X and σp for the m-X
substituent, respectively. As expected, a very sufficient
regression was obtained by considering such η2-type co-
ordination of the endo-aromatic ring through the ipso- and
ortho-positions (eq 1 and Figure 2). Here, it is noteworthy
that the η2-binding of the aromatic ring is more effectively
performed at the ipso-carbon than the ortho as indicated
by the percent contribution of the parameter σipso (61%)
and σortho (39%).
endo
log krel ð1Þ ¼ -2:49σipso -1:62σortho
- 0:108 ðR2 ¼ 0:98, n ¼ 8Þ
ð1Þ
As to the exo-Y substituted series, we observed only 2-fold
rate acceleration irrespective of the wide range of p-MeO/
p-CF3 substituent variation. These exo-isomers provided a
good linear free energy relationship when we used the
Yukawa-Tsuno (Y-T) equation,15,3d where σ0 is the normal
substituent constant and ΔσhRþ is the resonance substituent
constant defined by (σþ - σ0) (eq 2). The present reaction
yielded a small negative reaction constant (F =-0.912) and
a small resonance parameter (r = 0.237). These very low
values are certainly due to the reduced exo-substituent effects
on the far more remote acid-activated oxirane ring (δ-posi-
tion through the intervened cyclopropane and carbonyl
function). Incidentally, the well-known acetolyses of β-ar-
yltosylates and brosylates at 75-115 °C also displayed the
Y-T equations with larger F values of -3.3 to -4.0 and
r values of 0.54-0.63.16 The enhanced substituent effects
of these reactions can be explained by the formation of
The mechanistic evidence for the concertedness was also
obtained for the above-mentioned dual ipso/ortho transan-
nular SE2-Ar reactions of bis(p-chlorophenyl)homobenz-
oquinone epoxide 1j.6 The common π-aryl participated
(17) (a) Parker, R. E.; Issaacs, N. S. Chem. Rev. 1959, 59, 737–799.
(b) Rickborn, B. In Comprehensive Organic Synthesis; Trost, B. M., Ed.;
Pergamon Press: Oxford, U.K., 1991; Vol. 3, pp 733-771. (c) Fujita, H.; Yoshida,
Y.; Kita, Y. Yuki Gosei Kagaku Kyokaishi. 2003, 61, 133–143. (d) Giner, J.-L.;
Li, X.; Mullins, J. J. J. Org. Chem. 2003, 68, 10079–10086. (e) Whalen, D. E.
Advances in Physical Organic Chemistry; Richard, J. P. Ed.; Elsevier Academic
Press: London, 2005; Vol. 40, pp 247-298.
(18) Aziridines and Epoxides in Organic Synthesis; Yudin, K. A. Ed.;
Wiley-VCH: Weinheim, 2006.
(19) (a) Posner, G. H.; Rogers, D. Z. J. Am. Chem. Soc. 1977, 99, 8214–
8218. (b) Bellucci, G.; Berti, G.; Ingrosso, G.; Mastrorilli, E. J. Org. Chem.
1980, 45, 299–303.
(20) (a) Pocker, Y.; Ronald, B. P. J. Am. Chem. Soc. 1980, 102, 5311–
5316. (b) Yagi, H.; Jerina, D. M. J. Org. Chem. 2007, 72, 9983–9990.
(c) Davis, C. E.; Bailey, J. L.; Lockner, J. W.; Coates, R. M. J. Org. Chem.
2003, 68, 75–82. (d) Blumenstein, J. J.; Ukachukwu, V. C.; Mohan, R. S.;
Whalen, D. L. J. Org. Chem. 1993, 58, 924–932.
(15) (a) Yukawa, Y.; Tsuno, Y. Bull. Chem. Soc. Jpn. 1959, 32, 971–981.
(b) Shorter, J. Correlation Analysis in Chemistry. Recent Advances; Chapman,
N. B., Shorter, J., Eds.; Plenum Press: New York, 1978; Chapter 4, pp 119-173.
(16) (a) Fujio, M.; Funatsu, K.; Goto, M.; Mishima, M.; Tsuno, Y.
Tetrahedron 1987, 43, 307–316. (b) Fujio, M.; Goto, M.; Mishima, M.;
Tsuno, Y. Bull. Chem. Soc. Jpn. 1990, 63, 1121–1128. (c) Fujio, M.; Goto, N.;
Dairokuno, T.; Goto, M.; Saeki, Y.; Okusako, Y.; Tsuno, Y. Bull. Chem.
Soc. Jpn. 1992, 65, 3072–3079.
(21) Reichart, C. Solvents and Solvent Effects in Organic Chemistry, 3rd
ed.; Wiley-VCH Verlag GmbH & Co. KGaA: Weinheim, 2003; Chapter 5, pp
162-199.
(22) Reichardt, C. Chem. Rev. 1994, 94, 2319–2358.
736 J. Org. Chem. Vol. 75, No. 3, 2010

Asahara et al.
JOCArticle
SCHEME 2
the initial epoxide ring opening with the aid of the endo-
aromatic ring.23 However, the aryl participation in the ring
opening of epoxides is scarcely reported but has been put
forward in the acid-induced ring opening of a particular case
of epoxides bearing aryl groups directly or indirectly linked
to the oxirane ring. For instance, the well-documented
phenonium ion intermediates via the aryl participated tran-
sition states are invoked in the reactions of stilbene oxides17a
and spiro-linked 2-phenyl-1,2-epoxide24 or 1-benzyl-1,2-ep-
oxides.25 In our previous study using the ethano-bridged
diphenylhomobenzoquinone epoxides, we have found that
the rate of the acid-catalyzed transannular cyclization is
highly dependent on the rigid conformation of the endo-aryl
ring.23 We rationalized the conformational effects on the
basis of the π-aryl participated orbital interaction with the
vacant oxirane Walsh orbital. Keeping this in mind, we will
later interpret the present kinetic substituent effects in terms
of the geometrically restricted orbital interaction in the
SN2-type epoxide ring opening.
We have previously found that the analogous cyclobu-
tene-fused endo/exo m- and p-substituted diarylhomobenzo-
quinones 4 exhibited the similar π-aryl participated trans-
annular SE2-Ar cyclization under BF3 acid conditions
(Scheme 1b).6 This cyclobutene ring-opening reaction in-
volves a formal vinyl-anion migration to the adjacent acid-
activated carbonyl carbon associated with the concerted SN2
transannular cyclization of endo-aryl group.26 Accordingly,
the vacant cyclobutene anti-bonding orbital activated by the
adjoining BF3-complexed carbonyl group will accept the π-
electrons of the endo-aryl group and facilitate the σ-bond
cleavage. Therefore, a comparison between the epoxide and
the cyclobutene ring-opening reactions will provide a very
useful insight into the stereoelectronic effects in the π-aryl
participated orbital interaction.
transition state leads to the ortho-linked product 2 via
σ-complex I and competitively to the ipso-linked product 3
via the π-resonating σ-complex II followed by the hydrolysis
with residual water (Scheme 2). The present p-anisyl- and
p-chloro-substituted endo-1a and endo-1h also provided 3a
(∼100%) and 3h (80%) along with 2h (20%), respectively.
Quantitative and high yield formation of compounds 3 for
the reaction of endo-1a and endo-1h (including p,p0-Cl sub-
stituted 1j) may be explained in view of the reaction pathways
on which the common transition state TS would undergo the
very facile degradation into the π-resonating and water-
sensitive intermediate II rather than the less stabilized I. By
contrast, other para-substituents brought about the exclu-
sive formation of 2 via the ortho-linked σ-complex I. This is
probably because the possible ipso-linked σ-complex II
cannot be transformed into the stable 3 via the nucleophilic
displacement by residual water but rather undergoes a facile
1,2-shift to the ortho-linked I leading to 2. Thus, it is not
suprising that the percent contribution of the σipso and σortho
in the transition state stabilization cannot be directly re-
flected on the product distributions, which are dependent on
the following channels to the respective I and II as well as the
contaminant water. We also found that the transannular
cyclization of the endo m-X-substituted aromatic ring occurs
exclusively at the less congested ortho-position (i.e., para-
position with respect to the m-X) as confirmed by the
representative reaction of endo m-MeO-substituted 1d. A
significant steric hindrance of m-MeO with the oxirane Me
group probably inhibits the alternative π-aryl participation
at the adjacent ortho-position.
Stereoelectronic Effects in Dihapto(η2) π-Aryl Participa-
tion. As previously reported, we have obtained an excellent
two parameter regression using the σipso and σortho para-
meters for the BF3-catalyzed transannular cyclization of
cyclobutene-fused endo m- and p-substituted diarylhomo-
benzoquinones 4 (eq 3).6 This equation also manifested
η2-coordination in the π-aryl participated transition state
(Figure 3b). The percent contribution of σipso (47%) is
slightly smaller than that of σ ortho (53%). This is in contrast
to the reaction of epoxides 1, where the σipso contributed
1.6 times more effectively than σortho (eq 1).
endo
log krel ð4Þ ¼ -2:07σipso -2:33σortho
- 0:172 ðR2 ¼ 0:99, n ¼ 8Þ
ð3Þ
Why does endo-4 undergo slightly enhanced substituent
effects on the ortho-carbon atom, but in endo-1 the more
(23) Oshima, T.; Asahara, H.; Kubo, E.; Miyamoto, S.; Togaya, K. Org.
Lett. 2008, 10, 2413–2416.
(24) Kita, Y.; Furukawa, A.; Futamura, J.; Higuchi, K.; Ueda, K.;
Fujioka, H. Tetrahedron 2001, 57, 815–825.
(25) (a) Costantino, P.; Crotti, P.; Ferretti, M.; Macchia, F. J. Org. Chem.
1982, 47, 2917. (b) Crotti, P.; Ferretti, M.; Macchia, F.; Stoppioni, A. J. Org.
Chem. 1984, 49, 4706–4711. (c) Crotti, P.; Ferretti, M.; Macchia, F.;
Stoppioni, A. J. Org. Chem. 1986, 51, 2759–2766.
(26) Recently, Wang and Tantillo suggested a stepwise reaction pathway
for the acid-catalyzed transannular cyclization of 4 based on DFT calcula-
tions: Wang, S. C.; Tantillo, D. J. J. Org. Chem. 2007, 72, 8394–8401.
However, the poor kinetic solvent effects for the reaction of 4 and especially
for the present 1 explicitly rule out the intervention of the carbocation
intermediate derived from the stepwise SN1 mechanism.
(2). Prominent Role of Oxirane Walsh Orbital. In the acid-
catalyzed intramolecular SE2-Ar reactions of 1a-k, the
vacant oxirane Walsh orbital seems to play a crucial role in
J. Org. Chem. Vol. 75, No. 3, 2010 737

Asahara et al.
JOCArticle
FIGURE 3. (a) Representative lowest unoccupied Walsh orbital of
epoxides 1. (b) Unoccupied orbital (in part) of cyclobutene ring of 4.
significant substituent effects are on the ipso-carbon atom? To
answer this question, we must recall that the acid-catalyzed
SN2-type nucleophilic displacement of 1 and 4 involves the
π-aryl participated transition states. The π-electron donation
should occur onto the vacant oxirane Walsh orbital of 1 and
the cyclobutene antibonding σ* orbital of 4, respectively.
Since the lowest unoccupied Walsh orbital (LUMO) is tri-
gonallybent, the geometrical orbital interaction withtheendo-
aromatic ring (HOMO) may be represented in Figure 3. Thus,
the dihapto π-aryl participation of 1 can be more effectively
performed at the ipso-carbon rather than the ortho-one
(Figure 3a). By contrast, such a stereoelectronic preference
for the ipso-position should be considerably reduced for 4
because the vector of the relevant four-membered cyclobutene
σ*-bond deviates from the vertical plane through the ipso- and
para-position of the aromatic ring (Figure 3b). Indeed, a
theoretical calculation by the DFT B3LYP/6-31G* method27
apparently showed the similar geometrical characteristics for
the relevant vacant orbitals of epoxide and cyclobutene ring
moieties (Figure 4). The vacant Walsh orbital of 1g at the
carbon corner is conjugated with the carbonyl π*-orbital but
israther directedtothe aromaticipso-carbon(asyellowdotted
arrow in Figure 4a). On the other hand, the vector of the
vacant cyclobutene orbital of 4g is essentially perpendicular
(yellow dotted arrow) to the quinone plane and is expected to
enjoy the ideal ipso/ortho η2 coordination with the faced endo-
aromatic nucleus (Figure 4b). This is because the epoxides 1g
display more effective ipso-contribution in the η2 π-aryl
participation than the cyclobutene-cleaved 4g.
As to the exo-Y substituted 4, we previously obtained the
Y-T equation (eq 4) with a larger r value of 0.904 compared
with that of the similar equation for exo-1 (r = 0.237, eq 2).6
The larger resonance parameter r means that the π-deloca-
lization of exo-aryl substituents of 4 can appreciably stabilize
the developing positive charge on the acid-activated γ-lo-
cated carbonyl carbon atom. Such an activation of the
carbonyl π*-orbital results in the acceleration of the con-
certed vinyl anion migration responsible for the SE2-Ar
transannular cyclization of 4. By contrast, as seen in Y-T
eq 2, the reduced resonance stabilization by the exo-aryl
substituents of 1 may be due to the change in the reaction
mechanism. The epoxide ring opening is substantially facili-
tated by the binding of BF3 not at the carbonyl oxygen but at
FIGURE 4. LUMO orbitals of (a) epoxide 1g and (b) cyclobutene-
fused analogue 4g optimized by B3LYP/6-31G* method (exo-
aromatic ring is omitted for clarity).
the epoxide oxygen atom. Even if the carbonyl function is
activated by BF3 as in 4, it would rather deactivate the ring
opening because of the destabilization of the developing
positive charge on the epoxide carbon atom.
exo
log krel
¼ -0:794ðσ0 þ 0:904ΔσRþÞ ðR2 ¼ 0:96, n ¼ 8Þ
ð4Þ
Conclusion
In summary, we have performed a kinetic study of the
BF3-catalyzed ring-opening reactions of endo/exo m- and p-
substituted diarylhomobenzoquinone epoxides 1. These re-
actions proceeded through two types of SE2-Ar transannular
cyclizations to give tricyclic diketo-alcohols and cyclohexa-
dienone spiro-linked tricyclic diketo-alcohols, respectively.
The rates were significantly accelerated by the through-space
π-aryl participation of electron-donating endo-aromatic
rings but negligibly influenced by the through-bond electro-
nic effects of exo-aromatic rings. The Hammett treatment
using modified site-dependent substituent parameters σipso
and σortho indicated that the ipso/ortho dihapto(η2) π-aryl
participation occurs for the endo-aryl groups. Kinetics sub-
stituent effects of these reactions were compared with those
of the analogous acid-catalyzed transannular cyclization of
the cyclobutene-fused diarylhomobenzoquinones 4. It was
found that the present epoxides 1 exhibit the η2 π-aryl
participation with 1.6-times more effective contribution at
the ipso-position, whereas the cyclobutene-fused homolo-
gues 4 show almost comparable ipso/ortho contribution.
These results were interpreted in terms of the geometrical
features of the π-electron accepting vacant orbital of oxirane
and the cyclobutene ring. The physicochemical information
obtained in the present study will provide very important
insight into the understanding of the π-aryl participation as
well as mechanistic aspects in the acid-catalyzed ring-open-
ing reaction of epoxides.
Experimental Section
(27) Kong, J.; White, C. A.; Krylov, A. I.; Sherrill, D.; Adamson, R. D.;
Furlani, T. R.; Lee, M. S.; Lee, A. M.; Gwaltney, S. R.; Adams, T. R.;
Ochsenfeld, C.; Gilbert, A. T. B.; Kedziora, G. S.; Rassolov, V. A.; Maurice,
D. R.; Nair, N.; Shao, Y. H.; Besley, N. A.; Maslen, P. E.; Dombroski, J. P.;
Daschel, H.; Zhang, W. M.; Korambath, P. P.; Baker, J.; Byrd, E. F. C.; Van,
Voorhis, T.; Oumi, M.; Hirata, S.; Hsu, C. P.; Ishikawa, N.; Florian, J.;
Warshel, A.; Johnson, B. G.; Gill, P. M. W.; Head-Gordon, M.; Pople, J. A.
J. Comput. Chem. 2000, 21, 1532–1548.
Materials. Deuterated chloroform (CDCl3) was purchased
from Aldrich Chemicals Ltd. and used without further purifica-
tion. All epoxides 1a-k were prepared according to the previous
methods by the epoxidation of the corresponding diarylhomo-
benzoquinones.6 The endo/exo isomeric mixtures were sepa-
rated by column chromatography on silica gel with a mixture
738 J. Org. Chem. Vol. 75, No. 3, 2010

Asahara et al.
JOCArticle
of hexane/ethyl ether as an eluent and purified by recrystalliza-
tion from chloroform/pentane. The analytical data are col-
lected below. Unfortunately, we could not obtain the pure
endo-1e, endo/exo-1i because of contamination of the respec-
tive isomers. The assignment of the endo/exo stereochemistry
of 1a, d-f, h, and k failed because of the complexity in their 1H
NMR spectra. Thus, the stereochemistry was tentatively de-
duced from their kinetic behaviors, i.e., the more reactive
isomers of 1a and d-f are endo, but the more reactive isomers
of 1h, i, and k are exo.
Kinetic Measurements. The kinetic data were obtained by at
least duplicate measurements according to the NMR spectro-
scopic methods. The solution of epoxide 1 (0.02 mmol) in CDCl3
(0.67 mL) for kinetic experiments was prepared in a stoppered
NMR tube and preheated at 30 °C ((0.1) in a thermostatted
bath. The reaction was initiated by the quick addition of the
129.6, 132.3, 138.7, 159.0, 198.5, 200.6; IR (KBr) 1705 (CdO) cm-1
.
Anal. Calcd for C22H20O4: C, 75.84; H, 5.79; O, 18.37. Found: C,
75.69; H, 5.72; O, 18.59.
exo-1,5-Dimethyl-8-(4-anisyl)-8-phenyl-4-oxa-tricyclo[5.1.0.03,5]-
octane-2,6-dione (exo-1a). Mp 105.8-106.8 °C, colorless prisms
(chloroform/pentane); 1H NMR (270 MHz, CDCl3) δ 0.99 (s, 3H),
1.17 (s, 3H), 2.79 (s, 1H), 2.86 (s, 1H), 3.73 (s, 3H), 6.76-6.79
(m, 2H), 7.16-7.37 (m, 7H); 13C NMR (67 MHz, CDCl3) δ 13.8,
16.8, 37.5, 39.5, 48.9, 55.3, 60.3, 60.5, 114.5, 127.5, 128.2, 129.0,
130.2, 130.8, 140.5, 159.1, 198.6, 200.8; IR (KBr) 1695 (CdO)
cm-1. Anal. Calcd for C22H20O4: C, 75.84; H, 5.79; O, 18.37.
Found: C, 75.69; H, 5.72; O, 18.59.
1,5-Dimethyl-8,8-ditolyl-4-oxa-tricyclo[5.1.0.03,5]octane-2,6-
dione (1b). Mp 139.9-140.8 °C, colorless prisms (chloroform/
pentane); 1H NMR(270MHz, CDCl3) δ 0.94 (s, 3H), 1.17 (s, 3H),
2.24 (s, 3H), 2.27 (s, 3H), 2.77 (s, 1H), 2.81 (s, 1H), 7.02-7.14
(m, 6H), 7.22-7.26 (m, 2H); 13C NMR (67 MHz, CDCl3) δ
13.8, 16.9, 21.1, 21.1, 37.7, 39.8, 49.3, 60.1, 60.4, 128.0, 129.2,
129.5, 129.7, 135.3, 137.3, 137.3, 137.6, 198.4, 200.5; IR (KBr)
1702 (CdO) cm-1. HRMS: calcd for C23H22O3 346.1569, found
346.1572.
requisite volume of catalyst BF3 OEt2 (0.30 M) by a microsyr-
3
inge and submitted for the NMR measurements in a thermo-
statted tube holder at 30 °C ((0.1). The introduced amount of
BF3 OEt2 was reduced to 1/10 and 1/5 for the reactive 1a and
3
increased to 3 times for the least reactive endo-1k. The NMR
tubes for the less reactive endo-1i, 1j, and for the least reactive
endo-1k were sealed with a gas burner. The progress of reactions
was followed at requisite time intervals by monitoring the
relative signal intensity of the diagnostic methy groups on
cyclopropane (0.91-1.01 ppm) and oxirane ring (1.17-1.22)
of 1 over 75% conversion with respect to the internal standard
TMS. For endo-1i, 1j, and endo-1k, the measurements were
performed by taking the sealed NMR tube out of the thermo-
statted bath (30 ( 0.1 °C). The logarithmic plots of the relative
amounts of 1 with respect to the amount at the measurement
starting point (t = 0) versus time gave a straight line. The obtai-
ned first-order rate constants kobs were divided by the concen-
tration of catalyst to provide the second-order rate constants k.
The first-order decay plots and the logarithmic plots for the
most reactive endo-1a, the reference 1g, and the least reactive
endo-1k are represented in Supporting Information.
1,5-Dimethyl-8,8-bis(3-anisyl)-4-oxa-tricyclo[5.1.0.03,5]octane-
2,6-dione (1c). Mp 194.5-195.1 °C, colorless prisms (chloroform/
pentane); 1H NMR (270 MHz, CDCl3) δ 0.97 (s, 3H), 1.20
(s, 3H), 2.78 (s, 1H), 2.84 (s, 1H), 3.74 (s, 3H), 3.78 (s, 3H),
6.71-7.36 (m, 8H); 13C NMR (67 MHz, CDCl3) δ 13.8, 16.7,
37.6, 49.5, 55.3, 60.1, 60.3, 112.7, 113.8, 114.6, 115.1, 120.7, 121.8,
130.0, 130.2, 139.5, 141.2, 159.8, 198.3, 200.5; IR (KBr) 1698
(CdO) cm-1. Anal. Calcd for C23H22O5: C, 73.00; H, 5.86; O,
21.14. Found: C, 72.97; H, 5.85; O, 21.18.
endo-1,5-Dimethyl-8-(3-anisyl)-8-phenyl-4-oxa-tricyclo[5.1.0.03,5]-
octane-2,6-dione (endo-1d). Mp 165.6-166.6 °C, colorless prisms
(chloroform/pentane); 1H NMR (270 MHz, CDCl3) δ 0.98 (s, 3H),
1.18 (s, 3H), 2.80 (s, 1H), 2.85 (s, 1H), 3.74 (s, 3H), 6.72-6.86
(m, 3H), 7.14-7.41 (m, 6H); 13C NMR (67 MHz, CDCl3) δ 13.8,
16.8, 37.6, 39.7, 49.6, 55.3, 60.1, 60.3, 113.8, 115.1, 121.8, 127.7,
128.4, 128.9, 130.3, 139.6, 139.8, 159.8, 198.4, 200.5; IR (KBr) 1699
(CdO) cm-1. Anal. Calcd for C22H20O4: C, 75.84; H, 5.79; O, 18.37.
Found: C, 75.72; H, 5.75; O, 18.53.
Acid-Catalyzed Reaction of 1a-k. The acid-induced reactions
of epoxides 1a-k (0.02 mmol) were carried out in the presence of
exo-1,5-Dimethyl-8-(3-anisyl)-8-phenyl-4-oxa-tricyclo[5.1.0.03,5]-
octane-2,6-dione (exo-1d). Mp 168.8-169.8 °C, colorless prisms
(chloroform/pentane); 1H NMR (270 MHz, CDCl3) δ 0.91 (s, 3H),
1.21 (s, 3H), 2.80 (s, 1H), 2.80 (s, 1H), 3.78 (s, 3H), 6.73-6.76
(m, 1H), 6.90-6.96 (m, 2H), 7.19-7.26 (m, 6H); 13C NMR (67
MHz, CDCl3) δ 13.6, 16.8, 37.6, 39.7, 49.5, 55.3, 60.1, 60.4, 112.7,
114.6, 120.6, 128.0, 129.2, 129.7, 130.0, 138.2, 141.5, 159.9, 198.3,
200.6; IR (KBr) 1699 (CdO) cm-1. Anal. Calcd for C22H20O4: C,
75.84; H, 5.79; O, 18.37. Found: C, 75.72; H, 5.75; O, 18.53.
endo-1,5-Dimethyl-8-phenyl-8-(4-tolyl)-4-oxa-tricyclo[5.1.0.03,5]-
octane-2,6-dione (endo-1e). (This compound could not be obtained
in pure form because of the contamination of small amount of
exo-1e). 1H NMR (270 MHz, CDCl3) δ 0.95 (s, 3H), 1.18 (s, 3H),
2.25 (s, 3H), 2.80 (s, 1H), 2.83 (s, 1H), 7.39-6.86 (m, 3H),
7.14-7.39 (m, 9H); 13C NMR (67 MHz, CDCl3) δ 13.7, 16.8,
21.0, 37.5, 39.6, 49.4, 60.1, 60.4, 127.6, 128.3, 128.9, 129.4, 129.8,
135.2, 137.9, 140.3, 198.5, 200.7. Anal. Calcd for C22H20O3: C,
79.50; H, 6.06; O, 14.44. Found: C, 79.50; H, 6.05; O, 14.45.
exo-1,5-Dimethyl-8-(4-tolyl)-8-phenyl-4-oxa-tricyclo[5.1.0.03,5]-
octane-2,6-dione (exo-1e). Mp 143.9-144.9 °C, colorless prisms
BF3 OEt 2 (0.40 mmol) in CDCl3 (0.67 mL) at room tempera-
3
ture. After completion of the reactions, the reaction solution
was transferred into a separate funnel, diluted with chloroform
(10 mL), and then washed with water (3 mLꢀ3). The aqueous
layer was extracted with chloroform (5 mLꢀ2). The combined
organic layer was washed with water (3 mLꢀ3) and then dried
over calcium chloride. After the evaporation of the solvent in
1
vacuo, the residue was submitted for the H NMR analysis to
determine the product distributions. The reactions were found
to give the ortho-carbon bound tricyclic diketo-alcohols 2b-k
together with the ipso-carbon bound 2,5-cyclohexadien-4-one
spiro-linked tricyclic diketo-alcohol 3a (∼100%), 3h (the same
as 3a, 80%), and 3j (81%) for endo- 1a, endo-1h, and 1j, respec-
tively, in almost quantitative total yields based on the consumed
1. Except for the representative reactions of endo-1a, 1b, endo-
1d, endo-1h, and 1j, no further detailed analyses of products were
carried out. The analytical data of the products 2b, 2g, 2j, and 3j
have been already described elsewhere.6,28 The new compounds
3a and endo-2d were isolated by column chromatography on
silica gel with a mixture of hexane/ethyl acetate as an eluent.
endo-1,5-Dimethyl-8-phenyl-8-(4-anisyl)-4-oxa-tricyclo[5.1.0.03,5]-
octane-2,6-dione (endo-1a). Mp 116.5-117.5 °C, colorless prisms
(chloroform/pentane); 1H NMR (270 MHz, CDCl3) δ 0.91 (s, 3H),
1.19 (s, 3H), 2.78 (s, 1H), 2.80 (s, 1H), 3.75 (s, 3H), 6.80-6.84
(m, 2H), 7.18-7.30 (m, 7H); 13C NMR (67 MHz, CDCl3) δ 13.6,
16.7, 37.9, 40.0, 49.2, 55.3, 60.1, 60.4, 114.4, 127.8, 129.2, 129.5,
1
(chloroform/pentane); H NMR (270 MHz, CDCl3) δ 0.91 (s,
3H), 1.19 (s, 3H), 2.28 (s, 3H), 2.79 (s, 1H), 2.80 (s, 1H), 7.10-7.28
(m, 9H); 13C NMR (67 MHz, CDCl3) δ 13.6, 16.8, 21.0, 37.8, 39.8,
49.5, 60.1, 60.3, 127.9, 128.2, 129.2, 129.6, 129.7, 137.1, 137.5,
138.5, 198.5, 200.7; IR (KBr) 1708 (CdO) cm-1. Anal. Calcd for
C22H20O3: C, 79.50; H, 6.06; O, 14.44. Found: C, 79.50; H, 6.05; O,
14.45.
endo-1,5-Dimethyl-8-phenyl-8-(3-tolyl)-4-oxa-tricyclo[5.1.0.03,5]-
octane-2,6-dione (endo-1f). Mp 137.1-138.1 °C, colorless prisms
(28) Asahara, H.; Kubo, E.; Koizumi, T.; Mochizuki, E.; Oshima, T. Org.
Lett. 2007, 9, 3421–3424.
J. Org. Chem. Vol. 75, No. 3, 2010 739

Asahara et al.
JOCArticle
(chloroform/pentane); 1H NMR (270 MHz, CDCl3) δ 0.93 (s, 3H),
1.18 (s, 3H), 2.27 (s, 3H), 2.79 (s, 1H), 2.80 (s, 1H), 6.99-7.41
(m, 9H); 13C NMR (67 MHz, CDCl3) δ 13.6, 16.8, 21.2, 37.6, 39.8,
49.7, 60.0, 60.2, 126.6, 127.6, 128.4, 128.7, 129.0, 129.1, 130.2, 138.2,
139.1, 140.1, 198.4, 200.5; IR (KBr) 1700 (CdO) cm-1. Anal. Calcd
for C22H20O3: C, 79.50; H, 6.06; O, 14.44. Found: C, 79.33; H, 5.97;
O, 14.70.
13.8, 17.0, 37.2, 39.2, 47.4, 60.4, 129.3, 129.4, 129.6, 130.8, 133.9,
.
134.3, 136.1, 138.0, 197.8, 200.0; IR (KBr): 1702 (CdO) cm-1
HRMS: calcd for C21H16Cl2O3 386.0476, found 386.0479.
endo-1,5-Dimethyl-8-phenyl-8-(4-trifluoromethylphenyl)-4-oxa-
tricyclo[5.1.0.03,5]octane-2,6-dione (endo-1k). Mp 158.8-159.8 °C,
colorless prisms (chloroform/pentane); 1H NMR (270 MHz,
CDCl3) δ 0.96 (s, 3H), 1.22 (s, 3H), 2.85 (s, 1H), 2.87 (s, 1H),
7.24-7.55 (m, 9H); 13C NMR (67 MHz, CDCl3) δ 13.5, 14.0, 16.9,
37.3, 39.3, 48.4, 60.2, 60.3, 123.6 (q, JCF = 272.1 Hz), 126.1 (q, JCF
= 3.91 Hz), 128.1, 128.4, 129.2, 130.1, 130.3 (q, JCF = 33.0 Hz),
139.1, 142.3, 198.1, 200.4; IR (KBr) 1704 (CdO) cm-1. Anal.
Calcd for C22H17F3O3: C, 68.39; H, 4.43; F, 14.75; O, 12.42.
Found: C, 68.27; H, 4.39; F, 14.84; O, 12.50.
exo-1,5-Dimethyl-8-(4-trifluoromethylphenyl)-8-phenyl-4-oxa-
tricyclo[5.1.0.03,5]octane-2,6-dione (exo-1k). Mp 157.8-158.8 °C,
colorless prisms (chloroform/pentane); 1H NMR (270 MHz,
CDCl3) δ 0.94 (s, 3H), 1.20 (s, 3H), 2.81 (s, 1H), 2.84 (s, 1H),
7.23-7.60 (m, 9H); 13C NMR (67 MHz, CDCl3) δ 13.6, 16.9,
37.1, 39.3, 48.6, 60.2, 60.3, 123.8 (q, JCF=272.1 Hz), 126.1 (q, JCF
=3.91 Hz), 128.4, 128.9, 129.4, 129.5 (q, JCF=33.0 Hz), 129.6,
137.3, 143.8, 197.9, 200.1; IR (KBr) 1701 (CdO) cm-1. Anal.
Calcd for C22H17F3O3: C, 68.39; H, 4.43; F, 14.75; O, 12.42.
Found: C, 68.27; H, 4.39; F, 14.84; O, 12.50.
exo-1,5-Dimethyl-8-(3-tolyl)-8-phenyl-4-oxa-tricyclo[5.1.0.03,5]-
octane-2,6-dione (exo-1f). Mp 116.2-117.2 °C, colorless prisms
1
(chloroform/pentane); H NMR (270 MHz, CDCl3) δ 0.92 (s,
3H), 1.19 (s, 3H), 2.32 (s, 3H), 2.80 (s, 1H), 2.81 (s, 1H), 7.02-7.27
(m, 9H); 13C NMR (67 MHz, CDCl3) δ 13.6, 16.8, 21.4, 37.6, 39.7,
49.7, 60.1, 60.3, 125.4, 127.9, 128.5, 128.8, 128.9, 129.2, 129.6,
138.4, 138.8, 139.9, 198.5, 200.7; IR (KBr) 1703 (CdO) cm-1
.
Anal. Calcd for C22H20O3: C, 79.50; H, 6.06; O, 14.44. Found: C,
79.33; H, 5.97; O, 14.70.
1,5-Dimethyl-8,8-diphenyl-4-oxa-tricyclo[5.1.0.03,5]octane-2,6-
dione (1g). Mp 122.5-123.5 °C colorless prisms (chloroform/
pentane); 1H NMR (270 MHz, CDCl3) δ 0.92 (s, 3H), 1.18
(s, 3H), 2.81 (s, 1H), 2.82 (s, 1H), 7.17-7.32 (m, 8H), 7.37-
7.40 (m, 2H); 13C NMR (67 MHz, CDCl3) δ 13.7, 16.9, 37.6, 39.7,
49.6, 60.1, 60.3, 127.5, 127.9, 128.2, 128.9, 129.1, 129.5, 138.1,
139.9, 198.1, 200.3. Anal. Calcd for C21H18O3: C, 79.22; H, 5.70;
O, 15.08. Found: C, 79.10; H, 5.86; O, 15.04.
(1R*,8R*,9S*,10R*,12S*)-12-Hydroxy-5-methoxy-1,10-di-
methyl-8-phenyl-tetracyclo-[6.4.11,9.02,7.08,13]trideca-2(7),3,5-
triene-11,13-dione (endo-2d). Mp 164.4-165.1 °C, colorless
endo-1,5-Dimethyl-8-phenyl-8-(4-chlorophenyl)-4-oxa-tricyclo-
[5.1.0.03,5]octane-2,6-dione (endo-1h). Mp 126.2-127.2 °C, color-
less prisms (chloroform/pentane); 1H NMR (270 MHz, CDCl3) δ
1.01 (s, 3H), 1.19 (s, 3H), 2.81 (s, 1H), 2.89 (s, 1H), 7.18-7.38 (m,
9H); 13C NMR (67 MHz, CDCl3) δ 13.6, 16.8, 37.4, 39.3, 48.3,
60.3, 60.4, 127.9, 128.3, 129.1, 129.4, 131.0, 134.2, 136.8, 139.7,
198.2, 200.5; IR (KBr) 1695 (CdO) cm-1. Anal. Calcd for
C21H17ClO3: C, 71.49; H, 4.86; Cl, 10.05; O, 13.60. Found: C,
71.20; H, 4.87; Cl, 10.23; O, 13.70.
exo-1,5-Dimethyl-8-(4-chlorophenyl)-8-phenyl-4-oxa-tricyclo-
[5.1.0.03,5]octane-2,6-dione (exo-1h). Mp 173.7-174.7 °C, color-
less prisms (chloroform/pentane); 1H NMR (270 MHz, CDCl3)
δ 0.92 (s, 3H), 1.19 (s, 3H), 2.77 (s, 1H), 2.82 (s, 1H), 7.21-7.34
(m, 9H); 13C NMR (67 MHz, CDCl3) δ 13.6, 16.8, 37.3, 39.5,
48.6, 60.2, 60.3, 128.2, 129.2, 129.3, 129.6, 129.7, 133.7, 137.8,
138.6, 198.1, 200.3; IR (KBr) 1705 (CdO) cm-1. Anal. Calcd for
C21H17ClO3: C, 71.49; H, 4.86; Cl, 10.05; O, 13.60. Found: C,
71.20; H, 4.87; Cl, 10.23; O, 13.70.
1
prisms (chloroform/pentane); H NMR (270 MHz, CDCl3) δ
1.14 (s, 3H), 1.24 (s, 3H), 2.63 (br, 1H), 3.05 (s, 1H), 3.58 (s, 3H),
3.65 (s, 1H), 6.06 (d, 1H, J =2.64Hz), 6.68(dd, 1H, J = 2.64, 8.24
Hz), 6.97 (d, 1H, J = 8.24 Hz), 7.14-7.16 (m, 1H), 7.38-7.42 (m,
2H), 7.48-7.50 (m, 2H); 13C NMR (67 MHz, CDCl3) δ 14.5, 21.1,
38.2, 49.7, 54.0, 55.1, 60.6, 81.3, 112.0, 115.8, 125.0, 128.5, 128.8,
129.3, 130.1, 131.1, 136.5, 138.3, 159.6, 203.2, 205.6; IR (KBr)
3330 (-OH), 1714, 1698 (CdO) cm-1. HRMS: calcd for
C22H20O4 348.1362, found 348.1356.
Spiro[cyclohexa-2,5-dienone-4,80-7-phenyl-1,4-dimethyl-3-hy-
droxy-tricyclo[2.2.2.06,7]octane-2,5-dione] 3a (same as 3h). Mp
97.3-98.1 °C, colorless prisms (chloroform/pentane); 1H NMR
(270 MHz, CDCl3) δ 1.00 (s, 3H), 1.07 (s, 3H), 2.79 (s, 1H), 3.06
(s, 1H), 4.01 (s, 1H), 6.13 (dd, J = 1.98, 10.22 Hz), 6.51 (dd, J =
1.98, 10.22 Hz), 6.58 (dd, J = 3.30, 10.22 Hz), 6.86 (dd, J = 3.30,
10.22 Hz), 7.07 (br, 2H), 7.29 (br, 3H); 13C NMR (67 MHz,
CDCl3) δ 10.8, 14.7, 43.2, 46.2, 53.0, 53.6, 56.0, 75.5, 128.6,
129.0, 129.2, 129.4, 131.2, 131.2, 134.2, 143.2, 148.1, 184.4,
203.7, 204.6; IR (KBr) 3411 (-OH), 1747, 1712, 1664 (CdO)
cm-1. HRMS: calcd for C23H22O3 334.1205, found 334.1197.
endo/exo-1,5-Dimethyl-8-phenyl-8-(3-chlorophenyl)-4-oxa-tri-
cyclo[5.1.0.03,5]octane-2,6-dione (endo/exo-1i). (These com-
pounds could not be separated); H NMR (270 MHz, CDCl3)
1
δ 0.93 (s, 3H), 1.01 (s, 3H), 1.19 (s, 3H), 1.20 (s, 3H), 2.78 (s, 1H),
2.81 (s, 1H), 2.81 (s, 1H), 2.87 (s, 1H), 7.14-7.40 (m, 18H); 13
C
NMR (67 MHz, CDCl3) δ 13.6, 13.6, 16.8, 16.8, 37.3, 37.5, 39.4,
39.5, 48.5, 48.7, 60.2, 60.2, 60.2, 60.3, 126.6, 127.8, 128.0, 128.0,
128.3, 128.3, 128.4, 128.5, 129.1, 129.3, 129.7, 129.7, 130.3, 130.5,
134.8, 135.0, 137.5, 139.3, 140.2, 141.9, 198.0, 198.0, 200.2, 200.2;
IR (KBr): 1703 (CdO) cm-1. Anal. Calcd for C21H17ClO3: C,
71.49; H, 4.86; Cl, 10.05; O, 13.60. Found: C, 71.43; H, 4.83; Cl,
10.00; O, 13.74.
Acknowledgment. This work was supported by Grant-in-
Aid for Scientific Research from Japan Society for the
Promotion of Science (JSPS). H.A. also expresses his special
thanks for The Global COE (center of excellence) program
“Global Education and Research Center for Bio-Environ-
mental Chemistry” of Osaka University.
1,5-Dimethyl-8,8-bis(4-chlorophenyl)-4-oxa-tricyclo[5.1.0.03,5]-
octane-2,6-dione (1j). Mp 142.7-143.5 °C, colorless prisms
Supporting Information Available: X-ray crystallographic
data, calculation procedure, tables of atom coordinates, and full
spectral data for all new compounds. This material is available
free of charge via the Internet at http://pubs.acs.org.
1
(chloroform/pentane); H NMR(270 MHz, CDCl3) δ 1.01 (s,
3H), 1.19 (s, 3H), 2.75 (s, 1H), 2.89 (s, 1H), 7.14-7.19 (m, 2H),
7.22-7.27 (m, 2H), 7.29 (s, 4H); 13C NMR (67 MHz, CDCl3) δ
740 J. Org. Chem. Vol. 75, No. 3, 2010