Different Reaction Modes for the Oxidative Dimerization of Epoxyquinols and Epoxyquinones. Importance of Intermolecular Hydrogen-Bonding

Article abstract of DOI:10.1021/jo0355303

An oxidative dimerization reaction, involving the three successive steps of oxidation, 6π-electrocyclization, and Diels-Alder reaction, has been experimentally and theoretically investigated for the three 2-alkenyl-3-hydroxymethyl-2-cyclohexen-1-one deriv

Full text of DOI:10.1021/jo0355303

Differ en t Rea ction Mod es for th e Oxid a tive Dim er iza tion of
Ep oxyqu in ols a n d Ep oxyqu in on es. Im p or ta n ce of In ter m olecu la r
Hyd r ogen -Bon d in g
Mitsuru Shoji,† Hiroki Imai,† Isamu Shiina,‡ Hideaki Kakeya,§ Hiroyuki Osada,§ and
Yujiro Hayashi*,†
Department of Industrial Chemistry, Faculty of Engineering, and Department of Applied Chemistry,
Faculty of Science, Tokyo University of Science, Kagurazaka, Shinjuku-ku, Tokyo 162-8601, and
Antibiotics Laboratory, Discovery Research Institute, RIKEN, 2-1 Hirosawa,
Wako, Saitama 351-0198, J apan
hayashi@ci.kagu.tus.ac.jp
Received October 16, 2003
An oxidative dimerization reaction, involving the three successive steps of oxidation, 6π-
electrocyclization, and Diels-Alder reaction, has been experimentally and theoretically investigated
for the three 2-alkenyl-3-hydroxymethyl-2-cyclohexen-1-one derivatives epoxyquinol 3, epoxyquinone
6, and cyclohexenone 10. Of the sixteen possible modes of the oxidation/6π-electrocylization/Diels-
Alder reaction cascade for the epoxyquinone 6, and eight for the cyclohexenone 10, only the endo-
anti(epoxide)-anti(Me)-hetero and endo-anti(Me)-hetero modes are, respectively, observed, while
both endo-anti(epoxide)-anti(Me)-hetero and exo-anti(epoxide)-anti(Me)-homo reaction modes occur
with the epoxyquinol 3. Intermolecular hydrogen-bonding is found to be the key cause of formation
of both epoxyquinols A and B with 3, although epoxyquinone 6 and cyclohexenone 10 both gave
selectively only the epoxyquinol A-type product. In the dimerization of epoxyquinol 3, two monomer
2H-pyrans 5 interact with each other to afford intermediate complex 28 or 29 stabilized by hydrogen-
bonding, from which Diels-Alder reaction proceeds. Theoretical calculations have also revealed
the differences in the reaction profiles of epoxyquinone 6 and cyclohexenone 10. Namely, the rate-
determining step of the former is the Diels-Alder reaction, while that of the latter is the
6π-electrocyclization.
SCHEME 1. Key Step s for th e Syn th esis of
Ep oxyqu in ols A a n d B
We have recently isolated and determined the struc-
tures of epoxyquinols A (1)1 and B (2),2 which are novel
angiogenesis inhibitors with a highly functionalized and
complicated heptacyclic ring system containing 12 ste-
reocenters. Although these structures are very complex,
we have proposed that they arise biosynthetically from
the rather simple epoxyquinol 3 via an oxidation/
6π-electrocyclization/Diels-Alder reaction cascade (Scheme
1). Recently we have accomplished the first asymmetric
total synthesis of these molecules and determined their
absolute stereochemistry.3 Key steps of this synthesis are
the HfCl4-mediated Diels-Alder reaction of furan with
the acrylate of Corey’s chiral auxiliary4 and a biomimetic,
oxidative Diels-Alder reaction of monomer 3. We have
also established a practical synthetic route to both
* To whom correspondence should be addressed. Phone: (+81)3-
5228-8318. Fax: (+81)3-5261-4631.
† Department of Industrial Chemistry, Faculty of Engineering.
‡ Department of Applied Chemistry, Faculty of Science.
§ RIKEN.
(1) Kakeya, H.; Onose, R.; Koshino, H.; Yoshida, A.; Kobayashi, K.;
Kageyama, S.-I.; Osada, H. J . Am. Chem. Soc. 2002, 124, 3496.
(2) Kakeya, H.; Onose, R.; Yoshida, A.; Koshino, H.; Osada, H. J .
Antibiot. 2002, 55, 829.
(3) Shoji, M.; Yamaguchi, J .; Kakeya, H.; Osada, H.; Hayashi, Y.
Angew. Chem., Int. Ed. 2002, 41, 3192.
(4) (a) Hayashi, Y.; Nakamura, M.; Nakao, S.; Inoue, T.; Shoji, M.
Angew. Chem., Int. Ed. 2002, 41, 4079. (b) Sarakinos, G.; Corey, E. J .
Org. Lett. 1999, 1, 1741.
enantiomers of epoxyquinols A and B using the chroma-
tography-free preparation of an iodolactone and a lipase-
mediated kinetic resolution as key steps.5 Soon after our
first synthesis, Porco et al. published a synthesis of
10.1021/jo0355303 CCC: $27.50 © 2004 American Chemical Society
Published on Web 01/27/2004
1548
J . Org. Chem. 2004, 69, 1548-1556

Reaction Modes for the Dimerization of Epoxyquinols
SCHEME 2. Oxid a tive Dim er iza tion of Hyd r oxyep oxyqu in on e 3 w ith DMP
epoxyquinols A and B,6 and just recently Mehta et al.
have also reported the synthesis of racemic epoxyquinols
A and B.7
While epoxyquinols A and B are epoxyquinol dimers,
torreyanic acid, isolated from an endophytic fungus,
Pestalotiopsis microspra, is an epoxyquinone dimer.8 The
proposed biosynthesis of torreyanic acid also involves an
oxidation/6π-electrocyclization/Diels-Alder reaction cas-
cade, and Porco et al. have developed an elegant bio-
mimetic total synthesis, backed up by theoretical calcula-
tions, in which only the epoxyquinol A-type dimer is
selectively formed.9
In the course of our total synthesis of epoxyquinols A
and B, we examined carefully the oxidation of monomer
3 and observed the following interesting phenomenon:
When monomer 3 was treated with the Dess-Martin
periodinane (DMP),10 both primary and secondary hy-
droxy groups were oxidized, affording cyclized 2H-pyran
derivatives 8a /b, which gave epoxyquinol A-type product
9 in 21% yield without formation of the epoxyquinol-B
type product (Scheme 2). The yield of 9 was increased to
70% when isolated epoxyquinone 6 was treated with
DMP (vide infra, Scheme 6).
The reaction modes for the Diels-Alder reaction initi-
ated by oxidation of epoxyquinol 3 can be classified as
follows: Consider the diene part first. The reacting face
of the diene is either anti or syn to the epoxide and anti
or syn to the methyl group, which we designate as
anti(epoxide) or syn(epoxide) and anti(Me) or syn(Me)
additions, respectively. When the diene and dienophile
molecules are the same, we call it homocoupling, while
heterocoupling is the reaction in which the diene and
dienophile components are different. Endo addition is an
addition in which the secondary orbital interaction
between the carbonyl of the dienophile and the diene is
present, while exo addition is an addition without such
an interaction. According to this classification, ep-
oxyquinols A and B are dimers of endo-anti(epoxide)-
anti(Me)-hetero addition and exo-anti(epoxide)-anti(Me)-
homo addition, respectively. In the dimerization of
epoxyquinol 3, the only two reaction modes observed are
the endo-anti(epoxide)-anti(Me)-hetero and exo-anti(ep-
F IGURE 1. Monomers 3, 6, 10, and 11.
oxide)-anti(Me)-homo modes, while the dimerization of
epoxyquinone 6 proceeds via a single mode, the endo-
anti(epoxide)-anti(Me)-hetero mode.
To understand the difference in reaction modes be-
tween epoxyquinol 3 and epoxyquinone 6, the oxidative
dimerization of a parent monomer, 10, without epoxide
and hydroxy groups, has been examined. The methoxy-
cyclohexenone 11 was also investigated to shed light on
the effect of the hydroxy group in 3. In this paper we
describe our investigation of the oxidation/6π-electro-
cyclization/Diels-Alder reaction by systematic compari-
son using the four monomers 3, 6, 10, and 11 together
with theoretical calculations (Figure 1).11
Resu lts a n d Discu ssion
Syn th esis of Mon om er s. Synthesis of the epoxyquinol
monomer 3 has already been described.3,5 The epoxyquino-
ne monomer 6 was prepared from epoxyquinol monomer
3 according to Scheme 3, by protection of the primary
alcohol with TBS, followed by oxidation with DMP and
deprotection. The methoxy monomer 11 was prepared
from 12 by methyl ether formation12 and deprotection of
the TBS group as shown in Scheme 4. The cyclohexenone
monomer 10 with neither epoxy nor hydroxy functional-
ities was prepared according to Scheme 5. 3-Ethoxy-2-
cyclohexene-1-one (15)13 was treated with LiCH2SPh14 to
afford 3-(phenylthiomethyl)-2-cyclohexen-1-one (16), which
was converted to 3-formyl-2-cyclohexen-1-one (18) via
Pummerer rearrangement.15 Reduction with CeCl3-
NaBH4,16 protection of the primary hydroxy group, and
(5) Shoji, M.; Kishida, S.; Takeda, M.; Kakeya, H.; Osada, H.;
Hayashi, Y. Tetrahedron Lett. 2002, 43, 9155.
(6) Li, C.; Bardhan, S.; Pace, E. A.; Liang, M.-C.; Gilmore. T. D.;
Porco, J . A., J r. Org. Lett. 2002, 4, 3267.
(7) Mehta, G.; Islam, K. Tetrahedron Lett. 2003, 44, 3569.
(8) (a) Lee, J . C.; Yang, X.; Schwartz, M.; Strobel, G. A.; Clardy, J .
Chem. Biol. 1995, 2, 721. (b) Lee, J . C.; Strobel, G. A.; Lobkovsky, E.;
Clardy, J . J . Org. Chem. 1996, 61, 3232. (c) J arvis, B. B. Chemtracts
1997, 10, 10.
(9) (a) Li, C.; Lobkovsky, E.; Porco, J . A., J r. J . Am. Chem. Soc. 2000,
122, 10484. (b) Li, C.; J ohnson, R. P.; Porco, J . A., J r. J . Am. Chem.
Soc. 2003, 125, 5095.
(11) A preliminary communication of this work has been reported;
see: Shoji, M.; Kishida, S.; Kodera, Y.; Shiina, I.; Kakeya, H.; Osada,
H.; Hayashi, Y. Tetrahedron Lett. 2003, 44, 7205.
(12) Greene, A. E.; Drian, C. L.; Crabbe, P. J . Am. Chem. Soc. 1980,
102, 7583.
(13) Gannon, W. F.; House, H. O. Organic Syntheses; Wiley &
Sons: New York, 1973; Collect. Vol. V, p 539.
(14) Corey, E. J .; Seebach, D. J . Org. Chem. 1966, 31, 4097.
(15) Ishibashi, H.; Kameoka, C.; Kodama, K.; Ikeda, M. Tetrahedron
1996, 52, 489.
(10) (a) Dess, D. B.; Martin, J . C. J . Org. Chem. 1983, 48, 4155. (b)
Dess, D. B.; Martin, J . C. J . Am. Chem. Soc. 1991, 113, 7277. (c)
Ireland, R. E.; Liu, L. J . Org. Chem. 1993, 58, 2899.
(16) Gemal, A. L.; Luche, J .-L. J . Am. Chem. Soc. 1981, 103, 5454.
J . Org. Chem, Vol. 69, No. 5, 2004 1549

Shoji et al.
SCHEME 3. P r ep a r a tion of Mon om er 6
SCHEME 4. P r ep a r a tion of Mon om er 11
without formation of epoxyquinol B-type product after
aldehyde 23 was allowed to stand neat for 10 h (Scheme
7). These results indicate that formation of the dimerized
product is slow, and that only the endo-anti(Me)-hetero
mode occurs. This phenomenon, namely, that the ob-
served intermediate (aldehyde or 2H-pyran) is completely
different for the reactions of epoxyquinone 6 and cyclo-
hexenone 10, is quite puzzling (vide infra).
17
oxidation with MnO2 gave 3-(tert-butyldimethylsiloxy-
The oxidation of epoxyquinol 3 was successfully carried
out using MnO2 without protection of the secondary
alcohol.3 The intermediate aldehyde 4 thus generated was
not detected but was rapidly and smoothly converted to
2H-pyran 5. The Diels-Alder dimerization takes 4 h to
go to completion, and epoxyquinols A and B are formed
in yields of 40% and 25%, respectively (Scheme 8). 6π-
Electrocyclization is a fast process, and the Diels-Alder
reaction is slower than that of the epoxyquinone 6. The
reaction proceeds via both endo-anti(epoxide)-anti(Me)-
hetero and exo-anti(epoxide)-anti(Me)-homo modes, gen-
erating both epoxyquinols A and B, while no other
diastereomers are formed.
methyl)-2-cyclohexen-1-one (21). The desired 3-hydroxy-
methyl-2-propenyl-2-cyclohexen-1-one (10) was obtained
by R-iodination of ketone 21, a Suzuki coupling reaction,
and deprotection via the method developed for the
synthesis of monomer 3.3,5
Exp er im en ta l Resu lts for th e Oxid a tion /6π-Elec-
tr ocycliza tion /Diels-Ald er Rea ction . The results for
the oxidation/6π-electrocyclization/Diels-Alder reaction
of the monomers 6, 10, 3, and 11 are summarized in
Schemes 6-9, respectively. Though oxidation of alcohol
6 with MnO2 did not proceed owing to the neighboring
electron-withdrawing substituent, 6 was smoothly oxi-
1
dized within 15 min by DMP in CDCl3. H NMR (400
The methoxy derivative 11, however, gave results quite
different from those of epoxyquinol 3. When 11 was
oxidized with MnO2, aldehyde 26, which was not detected
by 1H NMR, was smoothly and completely converted into
2H-pyran derivatives 27 in a 4.5:1 diastereomer ratio
after 1.5 h, although which isomer predominates has not
been determined. As the Diels-Alder reaction of meth-
oxy-2H-pyran 27 does not proceed even under more
forcing reaction conditions, the single process of 6π-
electrocyclization can be monitored in this system. The
diastereomer ratio of 27 changed from 4.5:1 to 1.2:1 after
10 h (Scheme 9). This result clearly indicates the exist-
ence of an equilibrium18 between anti- and syn-2H-pyrans
27.
The present oxidative dimerization is composed of the
three successive cascade reactions oxidation, 6π-electro-
cyclization, and Diels-Alder dimerization. Oxidation of
the primary alcohol proceeds smoothly for all the sub-
strates examined, while the next two reactions are
dependent on the substituents. The 6π-electrocycylization
and Diels-Alder reactions were separately investigated
using theoretical calculations.
MHz) showed that 2H-pyran 8 was formed while the
corresponding aldehyde 7 could not be detected, and that
epoxyquinol A-type product 9 had been formed in 50%
yield with some 2H-pyran 8 remaining. When the crude
reaction mixture was left neat for 1 h, 2H-pyran 8 was
completely converted to the epoxyquinol A-type product
9 in 70% yield without formation of any other diastere-
omers (Scheme 6). These results indicate that the 6π-
electrocyclization is fast, and that aldehyde 7 is readily
converted to 2H-pyran 8. The Diels-Alder reaction is also
a fast process, proceeding only via the endo-anti(epoxide)-
anti(Me)-hetero mode to afford epoxyquinol A-type adduct
9 in good yield. Although 2H-pyran 8 could be regarded
as a poor diene because the two electron-withdrawing
groups would decrease its HOMO energy, the dimeriza-
tion is fast, which indicates that 8 acts as a reactive
dienophile in the Diels-Alder reaction.
The reaction profile of cyclohexenone monomer 10 is
rather different from that of epoxyquinone 6. Unlike
epoxyquinone 6, the oxidation of 10 proceeds efficiently
1
with MnO2, and the H NMR spectrum of the reaction
suggests the presence of aldehyde 23, 2H-pyran 24 not
being observed. Generation of the Diels-Alder adduct 25
was slow, and aldehyde 23 was gradually converted into
epoxyquinol A-type product 25 without detection of the
2H-pyran intermediate 24. Eventually epoxyquinol A-
type product 25 was gradually formed in 70% yield
Th eor et ica l St u d y of t h e 6π-E lect r ocycliza t ion .
Theoretical calculations were carried out to understand
the reaction profile of the 6π-electrocyclization and
Diels-Alder dimerization. The geometries of all station-
ary points were fully optimized at the B3LYP/6-31G*
(17) Aoyama, T.; Sonoda, N.; Yamauchi, M.; Toriyama, K.; Anzai,
M.; Ando, A.; Shioiri, T. Synlett 1998, 35.
(18) Marvell. E. N. Thermal Electrocyclic Reactions; Academic
Press: New York, 1980.
1550 J . Org. Chem., Vol. 69, No. 5, 2004

Reaction Modes for the Dimerization of Epoxyquinols
SCHEME 5. P r ep a r a tion of Mon om er 10
SCHEME 6. Dim er iza tion of Ep oxyqu in on e 6
SCHEME 8. Dim er iza tion of Ep oxyqu in ol
Mon om er 3
SCHEME 7. Dim er iza tion of Cycloh exen ol 10
tional study indicates that this reaction is highly exo-
thermic and that the TS energy is 5.0 kcal/mol for the
syn-isomer and 10.2 kcal/mol for the anti-isomer at the
B3LYP/6-31G*//AM1 level.9b As there has been no sys-
tematic study of substituent effects on 6π-electrocycliza-
tions,18 the 6π-electrocyclizations of the methyl ether
epoxyquinol derivative 26, epoxyquinone 7, epoxyquinol
4, and cyclohexenone 23 have been investigated in detail.
Scheme 10 represents the transition structures leading
to both syn- and anti-2H-pyrans along with the TS
energies and distances of the newly formed O1-C2 bond
(2H-pyran numbering).
The methoxy derivative 26 is an ideal substrate as
there is no subsequent Diels-Alder reaction and hence
the electrocyclization itself can easily be monitored
experimentally. Therefore, the 6π-electrocyclization of
this compound was examined first. 6π-Electrocyclization
of 26 is exothermic, and both syn- and anti-2H-pyrans
27 are calculated to be more stable than the parent
aldehyde 26, by 4.18 and 4.16 kcal/mol, respectively. The
TS energies leading to the two diastereomers of the 2H-
pyran are low (15.50 and 17.74 kcal/mol), and the TS
energies of retro-6π-electrocyclization are under 22 kcal/
mol (19.68 and 21.90 kcal/mol). The TS energy for the
reaction leading to syn-2H-pyran 27a is lower than that
of the reaction leading to anti-isomer 27b, while the anti-
and syn-isomers have the same stability.
These calculations are in good agreement with the
experimental results as follows: (1) After oxidation, 2H-
pyran 27 was observed without detection of intermediate
aldehyde 26. This is because on formation aldehyde 26
was immediately converted into 2H-pyran 27 as the TS
energy is low and 6π-electrocyclization is exothermic. (2)
level, and the properties of the molecules were also
calculated at the same level.19 All points were character-
ized as minima or saddle points by calculation of the
harmonic vibrational frequencies, using analytical second
derivatives.
Rodriguez-Otero has studied a series of 6π-electro-
cyclizations of (Z)-hexa-1,3,5-triene and its hetero-
substituted analogues at various levels of theory and
found that the reaction is slightly endothermic, the
required TS energy for the 6π-electrocyclization of (2Z)-
2,4-pentadienal being calculated as 21.52 kcal/mol at the
B3LYP/6-31G**//B3LYP/6-31G** level.20 Porco et al.
studied the 6π-electrocyclization of epoxyquinone deriva-
tives in their torreyanic acid synthesis. Their computa-
(19) All calculations were performed with the program package
TITAN 1.0.5 of Schrodinger, Inc. (http://www.schrodinger.com) and
Wavefunction Inc. (http://www.wavefun.com).
(20) Rodriguez-Otero, J . J . Org. Chem. 1999, 64, 6842.
J . Org. Chem, Vol. 69, No. 5, 2004 1551

Shoji et al.
SCHEME 9. 6π-Electr ocycliza tion of 11
SCHEME 10. Th eor etica l Ca lcu la tion s on th e 6π-Electr ocycliza tion
Though the major isomer of 27 has not been determined,
the diastereomer ratio of 2H-pyrans 27a ,b changed from
4.5:1 (1.5 h) to 1.2:1 (10 h). We can surmise that this is
because of an equilibrium occurring between 2H-pyrans
27a ,b and aldehyde 26, a result of the low TS energies
of both the 6π-electrocyclization and its retro reaction.
As calculated results for the methoxy derivative 26
proved to be in good agreement with experimental
findings, the other 6π-electrocyclizations of 4, 7, and 23
have also been examined, and the following noteworthy
features have been found from comparison of the four
reactions: (1) The lone pair of the formyl oxygen is
involved to a great extent in the TS as shown by the loss
of planarity in the dihedral angles of C2O1C6C5 (2H-pyran
numbering), and this is consistent with the calculations
of Rodriguez-Otero on the parent 2,4-pentadienal.20 (2)
As the TS energies of both the 6π-electrocyclization and
retro-6π-electrocyclization are below 22 kcal/mol for all
the substrates, there is equilibrium among the anti- and
syn-2H-pyrans and aldehyde at rt. (3) As the substituent
becomes more electron-withdrawing, the TS energy be-
comes lower, indicating that 6π-electrocyclization be-
comes easier. (4) The order of the length of the newly
formed O1-C2 bond in the transition state is epoxyqui-
none 7 > epoxyquinol 4 ) methoxy derivative 26 >
cyclohexenone 23. As the substituent becomes more
electron-withdrawing, the length O1-C2 in the TS be-
comes longer, indicating that the new bond has formed
to a considerably lesser extent, and that the transition
state is closer to the starting material. (5) As the
substituent becomes more electron-withdrawing, the
reaction becomes more exothermic, and the stability of
the 2H-pyran over the aldehyde increases except in the
case of epoxyquinol 4. (6) The reaction of epoxyquinol 4
is a special case, in which aldehyde 4 and 2H-pyrans 5a ,b
are of almost the same energy, while for the correspond-
ing methoxy derivative 2H-pyrans 27a ,b are more stable
than aldehyde 26 by ca. 4 kcal/mol. This is because there
1552 J . Org. Chem., Vol. 69, No. 5, 2004

Reaction Modes for the Dimerization of Epoxyquinols
TABLE 1. TS En er gy of th e Rea ction Mod es of Dim er iza tion of Ep oxyqu in on e 8
entry
reaction mode
diene
dienophile
TS energy/kcalmol-1
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
endo-anti(epoxide)-anti(Me)-hetero
endo-syn(epoxide)-syn(Me)-hetero
exo-syn(epoxide)-syn(Me)-hetero
exo-anti(epoxide)-anti(Me)-hetero
exo-anti(epoxide)-anti(Me)-homo
exo-syn(epoxide)-syn(Me)-homo
endo-anti(epoxide)-anti(Me)-homo
endo-syn(epoxide)-syn(Me)-homo
exo-anti(epoxide)-syn(Me)-homo
exo-syn(epoxide)-anti(Me)-homo
endo-syn(epoxide)-anti(Me)-homo
endo-anti(epoxide)-syn(Me)-homo
endo-anti(epoxide)-syn(Me)-hetero
endo-syn(epoxide)-anti(Me)-hetero
exo-syn(epoxide)-anti(Me)-hetero
exo-anti(epoxide)-syn(Me)-hetero
8a
8a
8a
8a
8a
8a
8a
8a
8b
8b
8b
8b
8b
8b
8b
8b
8b
8b
8b
8b
8a
8a
8a
8a
8b
8b
8b
8b
8a
8a
8a
8a
13.52
27.55
18.76
17.24
15.43
20.56
25.04
18.38
19.04
15.93
20.40
21.79
23.19
18.60
18.25
20.98
TABLE 2. F r on tier Or bita l En er gies of 8, 24, a n d 5
entry
orbital
orbital energy/eV
entry
orbital
orbital energy/eV
1
2
3
4
5
HOMO of 8a
LUMO of 8a
HOMO of 8b
LUMO of 8b
HOMO of 24
-6.22730
-2.25359
-6.25759
-2.27970
-5.46902
6
7
8
9
10
LUMO of 24
HOMO of 5a
LUMO of 5a
HOMO of 5b
LUMO of 5b
-1.60257
-5.80672
-1.97206
-5.88086
-1.99471
TABLE 3. En er gy Ga p of F r on tier Or bita ls of 8
energy gap/
SCHEME 11. Rea ction Mod es of 2H-P yr a n
entry
frontier orbital
eV
1
2
3
4
HOMO of 8a -LUMO of 8a
HOMO of 8a -LUMO of 8b
HOMO of 8b-LUMO of 8b
HOMO of 8b-LUMO of 8a
3.9737
3.9476
3.9779
4.0040
is a hydrogen-bond interaction between the hydroxy
group and formyl group in 4 (2.159 Å) as shown in
Scheme 10, and this stabilizes the aldehyde 4, so a higher
TS energy is required for the conversion into 2H-pyrans
5 than for conversion of the corresponding methyl ether
26 into 27.
Th eor etica l Stu d y of th e Diels-Ald er Dim er iza -
tion . Theoretical calculations on the homo-Diels-Alder
reaction of 2H-pyran indicate the regiochemistry should
be one of those shown in Scheme 11 according to the
frontier orbital theory.21,22 For this regiochemistry, there
are 16 possible reaction modes23 of the Diels-Alder
reaction of both epoxyquinone 6 and epoxyquinol 3. Of
these, only the endo-anti(epoxide)-anti(Me)-hetero mode
is observed with epoxyquinone 6, while both the endo-
anti(epoxide)-anti(Me)-hetero and exo-anti(epoxide)-
anti(Me)-homo modes are detected with epoxyquinol 3.
In the case of cyclohexenone 10, of the eight possible
reaction modes24 only the endo-anti(Me)-hetero mode was
observed.
energies of these reaction modes are summarized in Table
1. The frontier orbital energies of 2H-pyran derivatives
8a ,b, 24, and 5a ,b and the HOMO-LUMO energy gap
of 8a ,b have been summarized in Tables 2 and 3,
respectively.
Calculations indicate that the orientation of the methyl
groups is very important. That is, the three reaction
modes with the lowest TS energy are the endo-anti-
(epoxide)-anti(Me)-hetero (Table 1, entry 1), exo-anti-
(epoxide)-anti(Me)-homo (entry 5), and exo-syn(epoxide)-
anti(Me)-homo (entry 10) modes, in which the two methyl
groups are oriented on opposite sides of the approaching
dienophile and diene. That is to say, the steric hindrance
caused by the methyl groups is so large that the methyl
group of the diene monomer should be oriented anti to
its reacting face, and that of the dienophile monomer
oriented anti to its reacting face; otherwise the TS
energies are over 16 kcal/mol.
All 16 reaction modes for epoxyquinone 6 (8a ,b) were
investigated by theoretical calculations, and the TS
Of the 16 reaction modes, the TS energy of the endo-
anti(epoxide)-anti(Me)-hetero mode is found to be the
lowest, 13.52 kcal/mol, indicating that this reaction is a
facile process (entry 1). Scheme 12 indicates that steric
hindrance is minimized through the two methyl groups,
occupying axial positions. This is orbitally preferable in
terms of the lowest HOMO-LUMO energy gap (Table
3, entry 2), and also because of its secondary orbital
interactions (endo rule). The next lowest reaction mode
is exo-anti(epoxide)-anti(Me)-homo, which would lead to
the epoxyquinol B-type product (Table 1, entry 5). As this
second lowest TS energy is 15.43 kcal/mol, which is 1.91
(21) Fleming, I. Frontier Orbitals and Organic Chemical Reactions;
J ohn Wiley & Sons: Chichester, U.K., 1976.
(22) The coefficients of the HOMO and LUMO energies of the 2H-
pyrans are described in the Supporting Information.
(23) The 16 possible reaction modes are as follows: endo-anti-
(epoxide)-anti(Me)-hetero, exo-anti(epoxide)-anti(Me)-homo, endo-
anti(epoxide)-anti(Me)-homo, exo-anti(epoxide)-anti(Me)-hetero, endo-
anti(epoxide)-syn(Me)-hetero, exo-anti(epoxide)-syn(Me)-homo, endo-anti-
(epoxide)-syn(Me)-homo, exo-anti(epoxide)-syn(Me)-hetero, endo-syn-
(epoxide)-anti(Me)-hetero, exo-syn(epoxide)-anti(Me)-homo, endo-syn-
(epoxide)-anti(Me)-homo, exo-syn(epoxide)-anti(Me)-hetero, endo-syn-
(epoxide)-syn(Me)-hetero, exo-syn(epoxide)-syn(Me)-homo, endo-syn-
(epoxide)-syn(Me)-homo, exo-syn(epoxide)-syn(Me)-hetero.
(24) Because there is no epoxide in 10, the number of possible
reaction modes is reduced from sixteen in 3 to eight.
J . Org. Chem, Vol. 69, No. 5, 2004 1553

Shoji et al.
SCHEME 12. Th eor etica l Ca lcu la tion of th e Diels-Ald er Rea ction
TABLE 4. TS En er gies of th e 6π-Electr ocycliza tion a n d
Diels-Ald er Rea ction
kcal/mol higher than that of the lowest, the reaction
should not proceed via this mode but the via endo-
anti(epoxide)-anti(Me)-hetero mode alone, which is con-
sistent with the experimental result that epoxyquinol
A-type dimer 9 was selectively obtained.
TS energy/kcalmol-1
Diels-Alder
reaction
substrate
6π-electrocyclization
10.53,a 13.37b
18.64
13.52
In the case of cyclohexenone 10 (24), two reaction
modes, which would lead to epoxyquinol A- and B-type
products, were investigated (Scheme 12). The TS energy
of the endo-anti(Me)-hetero mode (11.43 kcal/mol) is
lower than that of the exo-anti(Me)-homo mode (13.00
kcal/mol), which is consistent with the experimental
result that the endo-anti(Me)-hetero dimer 25 (ep-
oxyquinol A-type product) was selectively formed.
The TS energies of the 6π-electrocyclization and the
Diels-Alder reaction of 3, 6, and 10 are summarized in
Table 4. By comparing epoxyquinone 6 and cyclohexenone
10, the aforementioned puzzling experimental results can
be reasonably explained as follows: As described previ-
ously, only 2H-pyran 8, but not aldehyde 7, was detected
for epoxyquinone 6, while only aldehyde 23, but not 2H-
pyran 24, was observed for cyclohexenone 10. These
contrasting results are due to the different reaction
epoxyquinone 6 (8)
cyclohexenone 10 (24)
epoxyquinol 3 (5)
11.43
18.51,a 18.06b
12.87,c 15.66d
a
b
The TS energy to anti-2H-pyran. The TS energy to syn-2H-
pyran. c The TS energy to 1. The TS energy to 2.
d
profiles of these two reactions. That is, the rate-
determining step has been reversed: The rate-determin-
ing step for epoxyquinone 6 is the Diels-Alder reaction,
while that for cyclohexenone 10 is 6π-electrocyclization
(Table 4). For the 6π-electrocyclization, the TS energy
for epoxyquinone 6 is lower than that for cyclohexenone
10 because the TS energy becomes lower as the substitu-
ent becomes more electron-withdrawing, and there are
two electron-withdrawing groups in 6 compared with only
one in 10. On the other hand, the TS energy of the Diels-
Alder reaction of epoxyquinone 6 is higher than that of
1554 J . Org. Chem., Vol. 69, No. 5, 2004

Reaction Modes for the Dimerization of Epoxyquinols
TABLE 5. Solven t Effect on th e Oxid a tive Dim er iza tion
of 3
cyclohexenone 10 because there is steric hindrance
caused by the epoxide in 6, and also because the Diels-
Alder reaction of 10 is orbitally favorable. Namely, the
two electron-withdrawing groups reduce the reactivity
of the diene 8 by lowering its HOMO energy, and cause
a larger HOMO-LUMO energy gap for 8a /8b (3.9476 eV,
Table 3, entry 2) than for 24 (3.8665 eV, Table 2, entry
5 (HOMO of 24), entry 6 (LUMO of 24)). As a result, the
rate-determining step has been reversed.
yielda/%
entry
solvent
neat
LiClO4/Et2O
benzene
toluene
CH2Cl2
Et2O
time/h
1
2
1
2
3
4
5
6
7
8
4
2.5
12
12
33
46
94
140
40
46
39
25
21
25
21
14
25
25
32
45
38
21
21
21
Compared with the above two substrates, the ep-
oxyquinol 3 (5a ,b) had a different profile: Though the
two 2H-pyran monomers 8 and 24 react to afford dimer-
ized product, monomer 5 from epoxyquinol 3 was not
directly transformed into the Diels-Alder products 1 and
2. Calculations suggest that initially the two monomers
5a ,b preassociate to give intermediate complexes 28 and
29, which are more stable than the parent 5a + 5b and
5a + 5a by 9.22 and 9.92 kcal/mol, respectively. This
stabilization can be ascribed to a hydrogen-bond interac-
tion as shown in Scheme 12. The hydroxy group of 5a
coordinates the carbonyl lone pair of 5b in 28, at a
distance of 1.915 Å, while two OH groups of two different
5a molecules interact with each other in 29, at a distance
of 1.968 Å. From the intermediates 28 and 29, the
dimerization proceeds to afford Diels-Alder products 1
and 2. As the TS energies for the endo-anti(epoxide)-
anti(Me)-hetero and exo-anti(epoxide)-anti(Me)-homo
modes are 12.87 and 15.66 kcal/mol, respectively, the
former mode would be theoretically more favorable than
the latter, and this is consistent with the experimental
result that 1 was formed predominantly. However, the
large difference between the TS energies of the two modes
(2.79 kcal/mol) is not in good agreement with the experi-
ment, in which 2 was also formed in 25% yield. This
discrepancy could be the result of neglecting solvent
effects in the calculation, which would be detrimental
owing to the existence of hydrogen-bonding. The hydrogen-
bonding effect is found to be operative not only in the
ground state, but also in the transition state. As shown
in Scheme 12, the hydrogen-bond activates the ketone
function in the endo-anti(epoxide)-anti(Me)-hetero mode,
whereas there is hydrogen-bonding stabilization of the
TS in the exo-anti(epoxide)-anti(Me)-homo mode.
MeOH
CH3CN
a
Isolated yield.
TABLE 6. Solven t Effect on th e Oxid a tive Dim er iza tion
of 10
entry
solvent
time/h
yielda/%
1
2
3
4
5
neat
10
25
43
43
72
70
73
30
35
25
MeOH
benzene
toluene
CH2Cl2
a
Isolated yield of 25.
selectively obtained as the sole product, irrespective of
the solvent. Another interesting observation is that the
reaction of 10 in MeOH is much faster than that in
benzene, toluene, and CH2Cl2, while the reaction of
epoxyquinol 3 is slower in MeOH than that in benzene
and toluene. Hydrogen-bonding activation by MeOH,
which has been observed in the reaction of 10, cannot be
realized in the reaction of 3 owing to strong intermolecu-
lar hydrogen-bonding. This is further evidence for the
importance of hydrogen-bonding in the oxidative dimer-
ization of 3. Moreover, the predominant formation of
epoxyquinol B in toluene is synthetically useful, because
epoxyquinol B is a more potent angiogenesis inhibitor
than epoxyquinol A.2
The importance of the hydroxy group is also demon-
strated by the following experiment. Diels-Alder dimer-
ization of the methyl ether 11, in which there is no
hydrogen-bonding interaction, does not proceed, though
the 6π-electrocyclization does (vide supra, Scheme 9).
This is another piece of evidence supporting the impor-
tance of the hydroxy group in the dimerization of ep-
oxyquinol 3. The steric hindrance caused by the methoxy
groups of 27 would prevent the Diels-Alder reaction.
There are literature precedents in which intermolecu-
lar hydrogen-bonding can be successfully utilized for the
control of the stereochemistry of a Diels-Alder reaction.27
In this oxidative dimerization, nature has also success-
fully employed hydrogen-bonding in the Diels-Alder
reaction for the formation of epoxyquinol B.
This hydrogen-bonding interaction, which is found to
be important in the transition state in the theoretical
calculations, is also found in the crystal structure of the
final product, epoxyquinol B, in which a hydrogen-bond
between the two OH groups has in fact been observed.25
Moreover, if these transition-state hydrogen-bonds exist,
the distribution of epoxyquinols A and B should be
affected by the solvent, which is found to be the case. As
shown in Table 5, 1 was formed predominantly in neat
conditions or in benzene solution, while 2 was the major
product in toluene and CH2Cl2. Lewis acids such as
26
Con clu sion s
LiClO4 accelerate the reaction, affording epoxyquinol
A predominantly via the orbitally preferable endo mode
and in short reaction time. This is in a marked contrast
with the dimerization of cyclohexenone 10, in which the
product distribution is not affected at all by the solvent
(Table 6). That is, epoxyquinol A-type product 25 was
In the oxidative dimerization of epoxyquinone 6 and
cyclohexenone 10, the preferred reaction modes are endo-
(27) (a) Atherton, J . C. C.; J ones, S. Tetrahedron Lett. 2001, 42, 8239.
(b) Tripathy, R.; Carroll, P. J .; Thornton, E. R. J . Am. Chem. Soc. 1990,
112, 6743. (c) Fisher, M. J .; Hehre, W. J .; Kahn, S. D.; Overman, L. E.
J . Am. Chem. Soc. 1988, 110, 4625. (d) For excellent reviews of
biosynthetic Diels-Alder reactions, see: Stocking, E. M.; Williams, R.
M. Angew. Chem., Int. Ed. 2003, 42, 3078. (e) Ichihara, A.; Oikawa,
H. Curr. Org. Chem. 1998, 2, 365.
(25) See footnote 18 of ref 6.
(26) Grieco, P. A.; Nunes, J . J .; Gaul, M. D. J . Am. Chem. Soc. 1990,
112, 4595.
J . Org. Chem, Vol. 69, No. 5, 2004 1555

Shoji et al.
anti(epoxide)-anti(Me)-hetero and endo-anti(Me)-hetero,
respectively, while both epoxyquinols A and B are formed
via the endo-anti(epoxide)-anti(Me)-hetero and exo-anti-
(epoxide)-anti(Me)-homo modes in the dimerization of
epoxyquinol 3 because of intermolecular hydrogen-bond-
ing, which has been proved to exist by theoretical
calculations and several experimental results. Other
noteworthy features are as follows: The existence of an
equilibrium between the 2H-pyran and aldehyde has
been theoretically and experimentally demonstrated in
the case of the methoxy derivative 11. In the dimerization
of epoxyquinol 3, monomer 2H-pyrans 5 preassociate to
afford complexes 28 and 29, from which the Diels-Alder
reaction proceeds. Theoretical calculations have also
clarified the difference in reaction profiles between
epoxyquinone 6 and cyclohexenone 10. Namely, the rate-
determining step of the former is the Diels-Alder reac-
tion, while that of the latter is 6π-electrocyclization.
Ack n ow led gm en t. This work was supported by a
Grant-in-Aid for Scientific Research on Priority Areas
(A) “Exploitation of Multi-Element Cyclic Molecules”
from the Ministry of Education, Culture, Sports, Science
and Technology, J apan.
Su p p or tin g In for m a tion Ava ila ble: Complete experi-
mental procedures, full characterization, copies of 1H and 13C
NMR and IR of all new compounds, and Cartesian coordinates
for calculated transition states of 6π-electrocyclization and
Diels-Alder reaction (PDF). This material is available free of
charge via the Internet at http://pubs.acs.org.
J O0355303
1556 J . Org. Chem., Vol. 69, No. 5, 2004