Propensity of 4-methoxy-4-vinyl-2-cyclopentenones housed in tri- and Tetracyctic frameworks for deep-seated photochemical rearrangement

Article abstract of DOI:10.1021/ja001813q

The excited-state behavior of a select group of highly enantioenriched tri- and tetracyclic 2-cyclopentenones has been examined in two solvents. Prepared by a convergent pathway involving the coupling of cyclopentenyl bromide 12 to several chiral ketones followed by desilylation, perruthenate oxidation, and ring-closing metathesis, the reactants were conveniently irradiated through quartz. Following promotion to the triplet level, a cyclopropylcarbinyl biradical intermediate is presumably generated provided that the double bond linked to C4 is not saturated. Four pathways have been observed to operate, only one of which has been observed previously under rather different conditions. The new photorearrangements include non-Norrish fragmentation to a ketene, generation of a spirocyclic intermediate, and ring expansion to a cyclobutyl radical in advance of intramolecular fragmentation of the four-membered ring. In the latter instance, a hydroazulenone derivative is formed. In the majority of the examples, product structure and absolute stereochemistry were corroborated by means of X-ray diffraction analysis.

Full text of DOI:10.1021/ja001813q

9610
J. Am. Chem. Soc. 2000, 122, 9610-9620
Propensity of 4-Methoxy-4-vinyl-2-cyclopentenones Housed in Tri-
and Tetracyclic Frameworks for Deep-Seated Photochemical
Rearrangement
Leo A. Paquette,* Fabrice Gallou, Zhong Zhao, David G. Young, Jian Liu,
Jiong Yang, and Dirk Friedrich†
Contribution from the EVans Chemical Laboratories, The Ohio State UniVersity, Columbus, Ohio 43210
ReceiVed May 24, 2000
Abstract: The excited-state behavior of a select group of highly enantioenriched tri- and tetracyclic
2-cyclopentenones has been examined in two solvents. Prepared by a convergent pathway involving the coupling
of cyclopentenyl bromide 12 to several chiral ketones followed by desilylation, perruthenate oxidation, and
ring-closing metathesis, the reactants were conveniently irradiated through quartz. Following promotion to the
triplet level, a cyclopropylcarbinyl biradical intermediate is presumably generated provided that the double
bond linked to C4 is not saturated. Four pathways have been observed to operate, only one of which has been
observed previously under rather different conditions. The new photorearrangements include non-Norrish
fragmentation to a ketene, generation of a spirocyclic intermediate, and ring expansion to a cyclobutyl radical
in advance of intramolecular fragmentation of the four-membered ring. In the latter instance, a hydroazulenone
derivative is formed. In the majority of the examples, product structure and absolute stereochemistry were
corroborated by means of X-ray diffraction analysis.
In view of the co-occurrence of phorbol (1) and ingenol (2)
outside” manner. Accordingly, any plan that entails the potential
migration of C11 away from C9 (as found in phorbol) to C10
as in ingenol is certain to be endothermic in nature.6 Thus, a
photoisomerization was viewed to be an attractive option that
might surmount this hurdle. Described herein is a remarkable
array of deep-seated photorearrangements uncovered for the
substrates selected for study.7 While none has advanced a
successful approach to 2, the several excited-state pathways
defined during this investigation significantly embellish our
appreciation of the boundary limits of cyclopentenone photo-
chemistry.
as their fatty acid esters in a wide variety of Euphorbia
species1,2and the impressive tumor-promoting properties of these
substances,3 we thought it of considerable interest to determine
if 2 could be synthesized in possible biomimetic fashion from
a precursor structurally related to 1.4 Despite the efforts of
Strain Energy and Stereoalignment Considerations. The
relative stereochemistry at C8 and C10 in ingenol is primarily
responsible for locking 2 into near rigidity. The C-ring is notably
twisted. Consequently, closure of the B-C substructure has
proven to be difficult.8-12 Although molecular mechanics
(MM3) calculations of the heat of formation of various phorbol/
ingenol prototypes are likely to be only approximate because
of the divergent functionality involved, they establish that
several research groups, 2 has not yet yielded to synthesis.5 The
chief obstacle to this challenge appears to be the highly strained
nature of the bridge that spans C8 and C10 in an “inside-
† Present address: Aventis Pharmaceuticals, Inc., 1041 U.S. Route 202/
206, P.O. Box 6800, Bridgewater, NJ 08807-0800.
(1) Hecker, E. Carcinogenesis, Vol. 2. Mechanisms of Tumor Promotion
and Cocarcinogenesis; Slaga, T. J., Sivak, A., Boutwell, R. K., Eds.; Raven
Press: New York, 1978.
(6) Appendino, G.; Tron, G. C.; Cravotto, G.; Palmisano, G.; Annunziata,
R.; Baj, G.; Surico, N. Eur. J. Org. Chem. 1999, 3413.
(7) Preliminary communication: Paquette, L. A.; Zhao, Z.; Gallou, F.;
Liu, J. J. Am. Chem. Soc. 2000, 122, 1540.
(2) Hecker, E.; Adolf, W.; Hergenhahn, M.; Schmidt, R.; Sorg, B.
Cellular Interactions by EnVironmental Tumor Promoters; Fujuki, H., et
al., Eds.; Japan Sci. Soc. Press, Tokyo/VNU Science Press: Utrecht, 1984;
pp 3-36.
(8) (a) Winkler J. D.; Henegar K. E.; Williard, P. G. J. Am. Chem. Soc.
1987, 109, 2850. (b) Winkler J. D.; Hong B.-C.; Hey J. P.; Williard, P. G.
J. Am. Chem. Soc. 1991, 113, 8839. (c) Winkler, J. D.; Henegar, K. E.;
Hong, B.-C.; Williard, P. G. J. Am. Chem. Soc. 1994, 116, 4183. (d) Winkler,
J. D.; Kim, S.; Harrison, S.; Lewin, N. E.; Blumberg, P. M. J. Am. Chem.
Soc. 1999, 121, 296.
(9) (a) Funk, R. L.; Olmstead, T. A.; Parvez, M. J. Am. Chem. Soc. 1988,
110, 3298. (b) Funk, R. L.; Olmstead, T. A.; Parvez, M.; Stallman, J. B. J.
Org. Chem. 1993, 58, 5873.
(10) Rigby, J. H.; de Claire, V.; Cuisiat, S. V.; Heeg, M. J. J. Org. Chem.
1996, 61, 7992.
(11) Nakamura, T.; Matsui, T.; Tanino, K.; Kuwajima, I. J. Org. Chem.
1997, 62, 3032.
(3) Evans, F. J.; Soper, C. J. Lloydia 1978, 41, 193.
(4) (a) Rigby, J. H. Studies in Natural Products Chemistry; Atta-ur-
Rahman, Ed.; Elsevier Science Publ., Amsterdam, 1993; Vol. 12, p 233.
(b) Kim, S.; Winkler, J. D. Chem. Soc. ReV. 1997, 26, 387.
(5) For the earlier efforts of this group, consult: (a) Paquette, L. A.;
Nitz, T. J.; Ross, R. J.; Springer, J. P. J. Am. Chem. Soc. 1984, 106, 1446.
(b) Ross, R. J.; Paquette, L. A. J. Org. Chem. 1987, 52, 5497. (c) Paquette,
L. A.; Ross, R. J.; Springer, J. P. J. Am. Chem. Soc. 1988, 110, 6192. (d)
Paquette, L. A.; Ross, R. J.; Shi, Y.-J. J. Org. Chem. 1990, 55, 1589. (e)
Paquette, L. A.; Sauer, D. R.; Edmondson, S. D.; Friedrich, D. Tetrahedron
1994, 50, 4071.
(12) Molander, G. A.; Bessieres, B.; Eastwood, P. R.; Noll, B. C. J.
Org. Chem. 1999, 64, 4124.
10.1021/ja001813q CCC: $19.00 © 2000 American Chemical Society
Published on Web 09/21/2000

Excited-State BehaVior of 2-Cyclopentenones
J. Am. Chem. Soc., Vol. 122, No. 40, 2000 9611
ingenanes are the more strained isomers by a significant margin.
Scheme 1
The model structures 3 and 4 constitute an exemplary pair.
the synthetic pathway began with the readily available (4R)-
(+)-tert-butyldimethylsiloxy-2-cyclopenten-1-one (>95% ee),14
bromination-dehydrobromination of which gave rise to 1015
(Scheme 1). Exposure of 10 to vinylmagnesium bromide in the
presence of anhydrous cerium trichloride to deter unwanted
enolization16 resulted in preferred nucleophilic attack from the
face anti to the siloxy substituent with predominant generation
of 11. Subsequent O-methylation gave rise smoothly to 12 as
planned.
Despite the fact that 4 contains an additional carbonyl group,
the thermodynamic benefits of this linkage (179 kcal/mol)
relative to the energy of a CdC bond (146 kcal/mol) are
inadequate to render 4 more stable overall. Thus, 3a in its lowest
energy conformation is approximately 15.8 kcal/mol less
strained than 4a. For the 3b/4b pair, the ingenane is assessed
to possess 13.2 kcal/mol of additional energy.
The electrophilic partner was obtained by the sequence of
transformations outlined in Scheme 2. Following the efficient
conversion of commercially available (+)-2-carene into
(1R,3R,6S)-4-methylene-7,7-dimethylbicyclo[4.1.0]heptan-3-
ol,6d,17 the allylic hydroxyl was protected as the MOM ether.
Deprotonation of the resulting norcaranone 13 was seen to favor
the cyclopropylcarbinyl center. Although the direction of enoli-
zation of R-alkoxy ketones is sometimes difficult to predict,18
the situation in 13 is one in which the fused dimethylcyclopro-
pane ring blocks the top face from approach by the base.
Enolization toward the OMOM group is accordingly deterred
for steric reasons. Attempts to bring about the C-allylation of
this reactive intermediate gave 14 in a maximum yield of 28%.
Not only did it prove difficult to separate 14 from O-allylated,
bisallylated, and cyclopropane-cleaved byproducts, the allyl side
chain was projected in undesired â fashion (NOE analysis).
To circumvent this complication, the potassium enolate of
13 was O-acylated with allyl chloroformate to deliver 15
quantitatively. This intermediate served well as a precursor to
16 into which it was transformed stereo- and regiospecifically
with reasonable efficiency when exposed to 5 mol % of the
tris(dibenzylideneacetone)dipalladium‚chloroform complex19 in
1,2-dimethoxyethane containing 10 mol % Diphos.20 Advantage
was then taken of the ease with which bromide 12 undergoes
halogen-metal exchange in the presence of 2 equiv of tert-
butyllithium.21,22 Admixture of the resulting cycloalkenyl or-
ganometallic with 16 in the presence of anhydrous CeCl3 at
low temperature resulted in the formation of 17 in 95% yield.
In a two-step sequence involving fluoride ion-promoted desi-
lylation and perruthenate oxidation,23 17 was transformed via
18 into 19, thereby setting the stage for ring-closing metathesis
in the presence of Grubbs’ ruthenium-based catalyst.24 The
Entirely comparable computational analysis of 5 reveals that
its C9-C11 bond is particularly well stereoaligned with the
π-bond of the neighboring cyclopentenone moiety (see 7). A
conformation can be attained in which there is also good
stereoalignment of the p orbitals at C5 and C10. On the basis
of the reported ease with which vomifoliol acetate (8) and related
compounds undergo photoisomerization to 1,4-cyclohexanedi-
ones exemplified by 9,13 we were prompted to select 5 as a
prototypical substrate. At issue was the feasibility of effecting
its excited-state conversion into the ingenane 6 by virtue of a
1,2-shift from its photoexcited state. This process requires the
breaking of a σ bond at some stage (a “σ-route migration”).
Alternatively, a “π-route migration” involving a cyclopropyl-
carbinyl biradical is also possible. This investigation was
undertaken to make suitable distinction between these options.
Synthesis of the Reactants Leading to 5. The preparative
route to the target was designed to be enantioselective. Thus,
(14) Paquette, L. A.; Earle, M. J.; Smith, G. F. Org. Synth. 1996, 73,
36.
(15) Nakazawa, M.; Sakamoto, Y.; Takahashi, T.; Tomooka, K.; Ish-
ikawa, K.; Nakai, T. Tetrahedron Lett. 1993, 34, 5923.
(16) Imamato, T.; Takiyama, N.; Nakamura, K.; Hatajima, T.; Kamiya,
Y. J. Am. Chem. Soc. 1989, 111, 4392.
(17) (a) Gollnick, K.; Schroeter, S.; Ohloff, G.; Schade, G.; Schenck,
G. O. Liebigs Ann. 1965, 687, 14. (b) Gollnick, K.; Schade, G. Tetrahedron
1966, 22, 133.
(18) Paquette, L. A.; O’Neil, S. V.; Guillo, N.; Zeng, Q.; Young, D. G.
Synlett 1999, 1857.
(19) Ukai, T.; Kawazura, H.; Ishii, Y.; Bonnet, J. J.; Ibers, J. A. J.
Organomet. Chem. 1974, 65, 253.
(20) (a) Tsuji, J.; Minami, I.; Shimizu, I. Tetrahedron Lett. 1983, 24,
1793. (b) Tsuji, J.; Ohashi, Y.; Minami, I. Tetrahedron Lett. 1987, 28, 2397.
(21) Corey, E. J.; Beames, D. J. J. Am. Chem. Soc. 1972, 94, 7210.
(22) (a) Seebach, D.; Neumann, H. Chem. Ber. 1974, 107, 847. (b)
Neumann, H.; Seebach, D. Chem. Ber. 1978, 111, 2785.
(23) Ley, S. V.; Norman, J.; Griffith, W. P.; Marsden, S. P. Synthesis
1994, 639.
(13) Katsumura, S.; Isoe, S. HelV. Chim. Acta 1982, 65, 1927.

9612 J. Am. Chem. Soc., Vol. 122, No. 40, 2000
Paquette et al.
Scheme 2
Scheme 3
cyclization leading to generation of the cycloheptene ring was
notably efficient. No involvement of the double bond within
the cyclopentenone ring could be detected. Beyond this, chiral
HPLC analysis indicated 5 to be enantiomerically pure.
Exploratory Photochemistry and Product Structure Elu-
cidation. While four examples of the conversion of ingenane
congeners into tigliane systems (i.e., formally 2 f 1) have been
reported,6,25,26 bond migration in the reverse direction has not
been observed for reasons of strain buildup as outlined above.
The projected conversion of 5 to 6 was founded not only on
stereoelectronic grounds but as well on the realization that the
isomerization takes the form of an excited-state vinylogous
R-ketol rearrangement. Examples of the susceptibility of R-hy-
droxy ketones to base-promoted equilibration are many.27 The
most important consideration derived from the fact that ultra-
violet light has long been recognized28 to transform 2-cyclo-
pentenones readily into their corresponding triplet excited
states.29 The efficiency of this step was expected to contribute
heavily to initiation of the photochemical events in the A-ring
of 5.
Irradiation of 5 with a 450 W Hanovia lamp through quartz
in either dioxane or benzene solution led predominantly to the
two solid products 20 and 21 (Scheme 3), neither of which
exhibited any spectral evidence (viz., C/H correlation analysis)
that the integrity of ring C had been altered as required of 6.
Interestingly, the production of 20 was significantly favored
(>10:1) over 21 in the more polar reaction medium. The reverse
was true of the product distribution in benzene (1:7.5). The
infrared spectrum of 20 was characterized by an intense carbonyl
absorption at 1775 cm-1 suggesting that the ketone carbonyl
of the starting material had been transformed into an ester or a
lactone. This finding was corroborated by the appearance of a
carbon signal at δ 175.9. The 1H NMR spectrum also revealed
the presence of a terminal vinyl group (dCH2), an isolated
methylene group R to the carbonyl (as a widely separated AB
pair at δ 3.40 and 2.22), and a residual intracyclic double bond.
Firm evidence confirming the detailed structure and absolute
configuration of 20 was derived from an X-ray crystallographic
analysis (Figure 1).
The spectral features of 21 proved to be entirely different,
except for the identical way in which the B/C substructure could
be mapped. In the infrared, strong bands are seen at 3400 and
1684 cm-1 thereby signaling the unchanged nature of the tertiary
hydroxyl but significant electronic perturbation of the ketone
carbonyl relative to 5 (1714 cm-1). Although three olefinic
protons were noted, one of these had migrated significantly
upfield (to δ 5.32) relative to its position in the spectrum of 5
(δ 6.17). This proton was clearly no longer serving as the â-H
of an R,â-unsaturated carbonyl fragment. In addition, the
quaternary methoxyl-substituted carbon displayed by 5 (δ 82.9)
was replaced by a very different fully substituted carbon (δ 72.4)
that caused us to entertain a spirocyclic structure. This proposal
received ultimate corroboration following X-ray diffraction
measurements (Figure 2). The absolute configuration of the
newly generated stereogenic center holds particular mechanistic
significance.
(24) (a) Nguyen, S. T.; Johnson, L. K.; Grubbs, R. H.; Ziller, J. W. J.
Am. Chem. Soc. 1992, 114, 3974. (b) Schwab, P.; France, M. B.; Ziller, J.
W.; Grubbs, R. H. Angew. Chem., Int. Ed. Engl. 1995, 34, 2039.
(25) Hecker, E. Pure Appl. Chem. 1977, 49, 1423.
(26) Rigby, J. H.; Hu, J.; Heeg, M. J. Tetrahedron Lett. 1998, 39, 2265.
(27) Paquette, L. A.; Combrink, K. D.; Elmore, S. W.; Zhao, M. HelV.
Chim. Acta 1992, 75, 1772 and relevant references therein.
(28) Eaton, P. E. J. Am. Chem. Soc. 1962, 84, 2344.
(29) Eaton, P. E.; Hurt, W. S. J. Am. Chem. Soc. 1966, 88, 5038. (b)
For a recent review, consult: Schuster, D. I.; Lem, G.; Kaprinidis, N. A.
Chem. ReV. 1993, 93, 3.

Excited-State BehaVior of 2-Cyclopentenones
J. Am. Chem. Soc., Vol. 122, No. 40, 2000 9613
Scheme 4
Figure 1. Computer-generated perspective drawing of 20 as determined
by X-ray crystallography.
Figure 2. Computer-generated perspective drawing of 21 as determined
by X-ray crystallography.
Consequences of MOM Deprotection. A less than obvious
property of 20 and 21 is the close proximity of the MOM-
protected oxygen atom to the vinyl ether functionality in both
structures. Consequently, in an effort to ascertain whether a
comparable pattern of photoisomerization would be followed
and if further interaction of these functional groups would be
observed, 5 was selectively hydrolyzed to 22 with chlorotri-
methylsilane and tetra-n-butylammonium bromide30 (efficiency
of 66%). In photochemical experiments that were performed
identically at room temperature in dioxane and benzene as the
solvents of choice, 22 gave evidence of being transformed as
before into a pair of isomeric products. However, as the reaction
mixtures were allowed to stand longer and longer, additional
compounds made their appearance (Scheme 4). Direct evidence
that only 23 and 24 represent primary photoproducts was easily
gained by appropriate studies that monitored product appearance
same. Consequently, an additional ring had necessarily been
generated in all three structures. The presence of a lactone
carbonyl in 25 and 27 was deduced from an intense infrared
band at 1774-1772 cm-1. In 26, the carbonyl absorption was
shifted to lower wavenumber (1748 cm-1), in a region charac-
teristic of a cyclopentanone ring. Another major distinguishing
feature was the appearance of an added methyl singlet in the
1H NMR spectrum of 25 and 27, but not in 26. Although the
NOE analysis of 27 was somewhat complicated by the fortuitous
overlap of several signals, the stereochemistry of 25 was
revealed on this basis (see formula). The distinction between
these two epimers proved to be fully consistent with all the
available spectral parameters.
When similar attempts to deduce the absolute configuration
of 26 did not lead to unequivocal stereochemical assignments,
alternative recourse was made to X-ray diffraction analysis. The
ORTEP diagram depicted in Figure 3 reveals that 26 forms as
a consequence of the stereodirected intramolecular Michael
addition of the hydroxyl in 24 to the neighboring 3-methoxy-
cyclopentenone to afford a stable, stereoisomerically pure acetal
subunit.
1
versus time. The close similarities of the H and 13C NMR
spectra of 23 and 24 to those of 20 and 21 served to define
their structures to be as shown.
It could be directly surmised from COSY analyses of 25-
27 that the norcarane part structure had not been chemically
altered. Also firmly established by this technique was the
presence of two R-keto methylene groups in 26, but only one
in both 25 and 27. In all three photoisomers a double bond had
been lost, although the degree of unsaturation remained the
Saturation of the Cycloheptene Double Bond. As it was
becoming increasingly apparent that the mechanistic features
of the above photoisomerizations involved the cycloheptene
double bond, corresponding interest in the response of the
(30) Woodward, R. B.; et al. J. Am. Chem. Soc. 1981, 103, 3213.

9614 J. Am. Chem. Soc., Vol. 122, No. 40, 2000
Paquette et al.
Scheme 6
Figure 3. Computer-generated perspective drawing of 26 as determined
by X-ray crystallography.
Scheme 5
Scheme 7
dihydro derivative to photoexcitation grew. Experimentation
with 28, obtained by the controlled hydrogenation of 5, was
expected to help narrow the mechanistic possibilities in a
meaningful way. Similar exploratory irradiation of 28 led in
low yield to 29 (8%), 30 (9%), and a third photoisomer (<1%)
(Scheme 5). In the latter instance, the amount of material was
too small for rigorous identification.
ment. The synthetic route outlined in Scheme 6 began with
commercially available 2-allylcyclohexanone (31). Reaction of
this racemic ketone with the enantiopure lithiated form of 12
in the presence of cerium trichloride gave rise to a mixture of
allylic carbinols 32 and 33, which could be separated by careful
chromatography on silica gel. Although the individual desily-
lation and oxidation of these intermediates gave rise quite
efficiently to 36 and 38, a distinction between their absolute
stereostructures was not possible until arrival at a crystalline
photoproduct (see below).
The lactonic nature of both 29 and 30 was readily inferred
from their intense infrared bands at 1769 and 1776 cm-1
,
respectively. These data, combined with the observations of a
newly introduced methyl singlet at high field and a reduction
in the relative area of the -OCH2O- signal attributable to the
MOM group from two to one, suggested that these photoprod-
ucts were pentacyclic. Additional COSY, HMQC, and HMBC
experiments confirmed further that this pair of lactones shares
a common framework and differs uniquely in the configuration
of one center. The final necessary distinction was achieved by
NOE methods (see 29). The presence of two â-oriented
methoxyl groups made identification of this epimer fairly
straightforward. This comprehensive spectral analysis allowed
for unambiguous recognition of the fact that ring B in 28 had
not undergone ring contraction in a manner so characteristically
adopted by 5 and 22. These results point to the influential role
of B-ring unsaturation on product formation in the earlier studied
examples.
Examination of a Diastereomeric Structural Alternative.
The convergent assembly process for accessing 5 and its
congeners 22 and 28 relied entirely on the coupling of two
highly enantioenriched building blocks. As a consequence, we
became increasingly intrigued with the prospect of learning
whether the photochemical reactivity of a different diastereomer
would be matched or mismatched. To this end, we sought to
prepare 37 and 39 in order to gain added mechanistic enlighten-
Ring-closing metathesis of 36 and 38 led conveniently to 37
and 39, respectively. While the stereodisposition of the three
stereogenic centers across rings B and C in 37 parallels that in
hand earlier, the situation in 39 is reversed. Included in the
resultant modification of conformational features is inter alia
the absence of a feasible lactonization pathway. These changes
are nicely reflected in the photoconversion of 37 to 40 and 41,
and of 39 to 42 and 43 (Scheme 7). Both product mixtures were
readily separated into their pure components via chromatogra-
1
phy. Although the high-field H and 13C NMR spectra of 40
and 41 unmistakably identified them to directly parallel 23 and
24 in structure, specific knowledge of their absolute configu-
ration was still lacking. For this reason, the three-dimensional
structure of 40 was deduced by crystallographic methods (Figure
4). Definition of the absolute configuration of 40 in this manner
permitted reliable definition of the stereochemistry of all eight

Excited-State BehaVior of 2-Cyclopentenones
J. Am. Chem. Soc., Vol. 122, No. 40, 2000 9615
Scheme 8
Figure 4. Computer-generated perspective drawing of 40 as determined
by X-ray crystallography.
(44, Scheme 8).31 The stereochemical outcomes of the C-
allylation leading to 45 and the subsequent coupling with 12 to
generate 46 were controlled by equatorial approach to the enolate
anion such that the more stable chair develops, and by the steric
contributions of the endo methyl substituent on the substituted
methano bridge, respectively. These considerations hold im-
portance because the vinyl and allyl substituents become
proximally juxtaposed in trans fashion during the prelude to
the ring-closing metathesis that gives rise to 49. The identifica-
tion of 49 was made straightforward by the well-separated nature
of many key proton signals. Following their proper assignment
by making recourse to COSY and HMBC experiments, the NOE
correlations depicted on the illustrated structure were arrived
at quickly.
Whereas the photolysis of 49 in dioxane did not lead to the
formation of characterizable products, comparable treatment of
benzene solutions proved conducive to photodimerization.
Chromatography on silica gel led to the isolation of 50a (48%)
and 50b (3%), in addition to a third primary photoproduct (51)
formed to the extent of 15%. The closely related NMR and mass
spectral evidence recorded for 50a and 50b suggested that a
dimerization reaction course not previously encountered was
operational presently.
Figure 5. Computer-generated perspective drawing of 42 as determined
by X-ray crystallography.
precursors in Scheme 6 as well as the three companion
photoproducts.
The spectral features displayed by 42 are entirely different
than those of any photoproduct previously encountered in this
study. This photoisomer showed infrared bands at 3376, 1666,
and 1599 cm-1 consistent with retention of the hydroxyl and
1
of the conjugated enone chromophore. However, its H NMR
spectrum is characterized by only two olefinic proton signals,
the more deshielded appearing as a triplet at δ 5.97 and the
second as a doublet at higher field (δ 5.51). The presence of a
carbonyl 13C absorption at δ 190.9 and four olefinic carbon
peaks (δ 173.9, 165.1, 122.1, and 105.2) indicated that the
original cycloheptene double bond had become part of a cross-
conjugated dienone network. Final definition of its structure as
42 was achieved by means of X-ray crystallographic analysis
(Figure 5). The necessary diastereomeric relationship between
41 and 43 requires that the absolute configuration of the B/C-
ring junctures be as shown. This conclusion is further supported
by the data recorded for 42.
Not surprisingly, therefore, the stereochemical features that
distinguish 37 from 39 exert a primary effect on the excited-
state mechanistic pathways. While one eventuates in B-ring
contraction to the cyclohexene level with tandem lactonization,
the second culminates in twofold ring-A expansion with genera-
tion of a hydroazulenone part structure. Full explication of these
phenomena is provided in the sequel.
Like the starting material, products 50a and 50b also
displayed an intense infrared absorption at ca. 1724 cm-1
suggesting that the cyclopentenone had been minimally per-
1
turbed. Furthermore, their high-field H and 13C NMR spectra
revealed that whereas the number of CH2 and CH3 groups had
also not changed (DEPT analysis), only the conjugated enone
double bond remained. Consequently, the π-bond contained
within the cycloheptene ring of 49 had to have become involved
in the photoinduced bonding process. The full spectral analysis
of 50a and 50b is given in Supporting Information. The eight
formal possibilities given in Scheme 9 (presented in full in the
Supporting Information) remain after exclusion of less sym-
metric alternatives. Photodimers 50a and 50b constitute two
members of that group. The structural features of 51 continue
to elude us.
Mechanistic Discussion
Probe of a Pinanyl-Fused System. In an effort to change
the structural motif of ring C, we have also investigated the
consequences of the presence of a bicyclo[3.1.1]heptane subunit
in a photochemical precursor. The requisite cyclopentenone 49
was obtained in a comparable manner from (1R)-(+)-nopinone
Known Photochemical Reactions of 2-Cyclopentenones.
The ability of 2-cyclopentenones to undergo efficient inter- and
20
(31) [r]D +14.6 (neat).

9616 J. Am. Chem. Soc., Vol. 122, No. 40, 2000
Paquette et al.
Scheme 9
Scheme 10
intramolecular photochemical [2 + 2] cycloadditions to alkenes
has long been recognized to hold substantial synthetic poten-
tial.32 Cyclobutane ring formation occurs from the triplet excited
state of the enone and passes through one or more triplet 1,4-
biradical intermediates.29b,33 The major attractions of this
chemistry have been the ease of implementation, frequent high-
level regioselectivity, and capacity for delivering complex ring
systems.
In addition to the above, the sensitized addition of methanol
across the double bond (eq 1)34 and the photoisomerization of
On the other hand, the documented examples of Norrish type
1 cleavage as exemplified by eq 4 are few in number.38 Here,
the high level of R-substitution effectively redirects reaction such
that generation of the 1,5-biradical operates.
The 2-cyclopentenone systems studied earlier were, however,
not designed to elicit and elucidate the quite different processes
uncovered in the present investigation. Consequently, the
ensuing interpretative discussion is intended to rationalize at
the mechanistic level the course of these new and largely
unprecedented photoisomerization reactions.
3-substituted derivatives as typified by eq 235 have also received
significant attention. Furthermore, rearrangements akin to eq 2
are known in which an aryl group at C4 migrates to C3.36 These
processes, as well as intramolecular hydrogen abstraction
phenomena (eq 3),37 are clearly also triggered by activation of
Excited-State Behavior of 5. The first point to be noted is
that the original cyclopentenone A-ring in 5 is transformed into
either a fused lactone ring or a spirocyclic 3-methoxy-substituted
five-membered enone. Concurrently, the cycloheptene B-ring
has undergone one-carbon ring contraction in both examples.
We consider the most economical pathway to originate with
excitation of 5 to its π f π* state,29 this initial event being
followed by 3-exo cyclization to lead from 52 to 53 (Scheme
10). The occurrence of this bonding into the cycloheptene
π-linkage nicely sets the stage for concurrent A-ring fragmenta-
tion, rupture of the cyclopropane ring just formed, and evolution
of a ketene unit as depicted in 54. Not only are the stereoelec-
tronics of this process well aligned for stepwise fragmentation,
but prevailing stereochemistry also positions the ketene residue
syn to the hydroxyl. Intramolecular cyclization within 54 to give
20 presumably operates smoothly. Also, it seems likely that the
proximity of the ketene and vinyl ether functionalities in 54 is
notably conducive to thermal cyclization as in 55. Subsequent
prototropic shift within this dipolar species would result in
charge annihilation and the formation of 21.
The photochemistry of 22 and 37 can be rationalized in
identical terms. The product distribution where 22 is concerned
is made more complicated because of postphotochemical events.
These involve acid-promoted intramolecular acetalization pro-
cesses that bring the free hydroxyl group into heterocyclization
events (see 25-27).
Contrasting Biradical Reactivity in the Diastereomeric
Series. Although the distribution of products from 39 indicates
the ultimate operation of a diverted mechanistic pathway, the
the conjugated double bond.
(32) Crimmins, M. T. Chem. ReV. 1988, 88, 1453. (b) Crimmins, M. T.;
Reinhold, R. L. Org. React. (NY) 1993, 44, 297. (c) Crimmins, M. T.
ComprehensiVe Organic Synthesis; Trost, B. M., Fleming, I., Paquette, L.
A., Eds.: Pergamon: Oxford: 1991; Vol. 5, pp 123-150. (d) Winkler, J.
D.; Bowen, C. M.; Liotta, F. Chem. ReV. 1995, 95, 2003.
(33) For recent developments, see: (a) Andrew, D.; Weedon, A. C. J.
Am. Chem. Soc. 1995, 117, 5647. (b) Maradyn, D. J.; Weedon, A. C. J.
Am. Chem. Soc. 1995, 117, 5359. (c) Andrew, D.; Hastings, D. J.; Weedon,
A. C. J. Am. Chem. Soc. 1994, 116, 10870. (d) Hastings, D. J.; Weedon,
A. C. J. Am. Chem. Soc. 1991, 113, 8525.
(34) (a) Fraser-Reid, B.; Holder, N. L.; Hicks, D. R.; Walker, D. L. Can.
J. Chem. 1977, 55, 3978. (b) Parry, R. J.; Haridas, K.; De Jong, R.; Johnson,
C. R. Tetrahedron Lett. 1990, 31, 7549. (c) Buenger, G. S.; Marquez, V.
E. Tetrahedron Lett. 1992, 33, 3707.
(35) Zimmerman, H. E.; Zhu, Z. J. Am. Chem. Soc. 1995, 117, 5245.
(36) Zimmerman, H. E.; Little, R. D. J. Am. Chem. Soc. 1974, 96, 4623.
(37) (a) Schreiber, W. L.; Agosta, W. C. J. Am. Chem. Soc. 1971, 93,
3814. (b) Wolff, S.; Schreiber, W. L.; Smith, A. B., III; Agosta, W. C. J.
Am. Chem. Soc. 1972, 94, 7797.
(38) (a) Agosta, W. C.; Smith, A. B., III; Kende, A. S.; Eilerman, R. G.;
Benham, J. Tetrahedron Lett. 1969, 4517. (b) Agosta, W. C.; Smith, A. B.,
III; J. Am. Chem. Soc. 1971, 93, 5513.

Excited-State BehaVior of 2-Cyclopentenones
J. Am. Chem. Soc., Vol. 122, No. 40, 2000 9617
Scheme 11
Scheme 13
Scheme 12
excited-state events can again be accounted for in terms of π
f π* initiation and conversion to the cyclopropylcarbinyl
biradical 56 (Scheme 11).39 As in the earlier example, the
formation of this highly reactive species can be anticipated since
3-exo cyclizations have rate constants approximating 1000
s-1.40,41 The isolation of 42 leads to the conjecture that 56 is
capable of expansion to the cyclobutyl radical 57 en route to
42.
An additional consideration is relevant here. If one considers
the five-ring internal bond breaking in 56 to leave an odd-
electron center at the methoxyl-bearing carbon, one obtains the
stabilized 1,3-biradical 58 with the â-carbon of the cyclohex-
enone bonded to the middle carbon of this subunit (Scheme
12). 1,2-Shifts of π groups substituted at the central carbon of
a 1,3-biradical are known.42 Should 58 intervene and be subject
to such a migration, smooth conversion to 42 would occur. At
present, we cannot distinguish between the pair of mechanistic
options that result in hydroazulenone formation. It is clear,
however, that the stereochemical nature of the B/C-ring fusion
exerts a high level of control on the fate of intermediates such
as 53 and 56.
Fate of the Dihydro Derivative 28. Mechanistic Substan-
tiation. The mechanisms advanced above are certain to be
viewed by some as highly speculative with no substantiating
data. It is for this very reason that 28 was included in this
investigation. The expectation was that experiments involving
a substrate having a saturated seven-membered ring would go
far in narrowing the possibilities by eliminating the 3-exo
cyclization alternative for the π f π* biradical. In fact, 28
responds marginally well to photoactivation and leads with low
efficiency to products possessing structural features quite
divergent from those encountered heretofore.
this part structure is consistent with the notion that the π f π*
biradical 59 now has no productive outlet (Scheme 13). As a
consequence, r cleavage to deliver 60 becomes kinetically
significant. Resonance theory predicts that radical character will
also be manifested at the R-ketenyl carbon as in 61. The
stabilized intermediate must undergo several additional steps
prior to arrival at the products. The timing that we prefer
involves hydrogen atom transfer from the MOM substituent to
·CH2-, a process that is advantaged because of the close
proximity of these two groups and the obvious gain in stability.
It is imperative that the conversion to 62 occur prior to
lactonization. Molecular models provide clear indication that
ring closure in this fashion has the net consequence of distancing
these reaction centers. These divergent steps could be competi-
tive and account for the low efficiency with which 29 and 30
are formed.
Two ring closures lead from 62 to the observed products.
Should lactonization occur in advance of biradical coupling,
viz. 62 f 63 f 29/30, the stereoselective course of the
tetrahydrofuran ring-forming step is guaranteed. This reaction
pathway may therefore be favored. Since 62 is a conformation-
ally flexible entity, this penultimate intermediate has the
potential for C-C bond formation from either surface of the
cycloheptane ring. Further, it appears possible that closure from
the R face could occur reversibly since the ketene moiety then
would be unable to pursue any further facile intramolecular
processes.
One of the several noteworthy structural features of 29 and
30 is the presence of a newly generated methyl group in a
position geminal to the methoxy substituent. The evolution of
Whatever the actual timing of the individual steps, the data
reveal how critical the cycloheptene double bond is to the
unprecedented photochemical behavior of 5, 22, 37, and 39.
Triplet Fate of Pinanyl Dienone 49. The triplet 1,4-biradical
65, which is in principle formed upon photoexcitation of 49, is
not observed to participate in reactions comparable to those
previously encountered with closely related cyclopentenones
(39) Zimmerman, H. E.; Armesto, D. Chem. ReV. 1996, 96, 3065.
(40) Furxhi, E.; Horner, J. H.; Newcomb, M. J. Org. Chem. 1999, 64,
4064.
(41) Another important point is that a methoxy group only slightly
accelerates the opening of a cyclopropylcarbinyl radical: Martinez, F. N.;
Schlegel, H. B.; Newcomb, M. J. Org. Chem. 1998, 63, 3618.
(42) Zimmerman, H. E.; Heydinger, J. A. J. Org. Chem. 1991, 56, 1747.

9618 J. Am. Chem. Soc., Vol. 122, No. 40, 2000
Paquette et al.
δ 139.7, 136.2, 132.1, 114.1, 83.3, 73.0, 49.0, 25.7 (3C), 18.0, -4.8
Scheme 14
(2C); HRMS (EI) m/z (M+) calcd 319.0579, obsd 319.0565; [R]D
22
+131.2 (c 1.05, CHCl3).
Anal. Calcd for C13H23BrO2Si: C, 48.90; H, 7.26. Found: C, 49.01;
H, 7.31.
[[(1R,4R)-3-Bromo-4-methoxy-4-vinyl-2-cyclopenten-1-yl]oxy]-
tert-butyldimethylsilane (12). To a solution of 11 (1.90 g, 6.0 mmol)
in DMF (100 mL) cooled to 0 °C was added sodium hydride (230 mg,
7.5 mmol). The resulting mixture was stirred for 1 h, treated with freshly
distilled methyl iodide (0.60 mL, 9.0 mmol) at 0 °C, allowed to warm
slowly to room temperature overnight, quenched with water (100 mL),
and extracted with ether (3×). The combined organic phases were dried
and concentrated to leave a residue that was purified chromatographi-
cally (silica gel, elution with 5% ethyl acetate in hexanes) to provide
1
12 (1.90 g, 96%) as a colorless oil. IR (neat, cm-1) 1618; H NMR
(300 MHz, CDCl3) δ 6.11 (d, J ) 2.3 Hz, 1 H), 5.74 (dd, J ) 17.3,-
10.7 Hz, 1 H), 5.31 (dd, J ) 17.2, 1.1 Hz, 1 H), 5.17 (dd, J ) 10.7,
1.2 Hz, 1 H), 4.63 (ddd, J ) 3.1, 4.3, 7.1 Hz, 1 H), 3.28 (s, 3 H), 2.47
(dd, J ) 7.3, 14.0 Hz, 1 H), 2.04 (dd, J ) 14.0, 4.2 Hz, 1 H), 0.89 (s,
9 H), 0.08 (s, 6 H); 13C NMR (75 MHz, CDCl3) δ 139.5, 138.1, 129.6,
114.5, 88.3, 72.8, 50.8, 42.4, 25.8 (3C), 18.1, -4.7, -4.8; HRMS (EI)
m/z (M+) calcd 333.0675, obsd 333.0710; [R]D +102.8 (c 1.01,
25
(Scheme 14). In this instance, the fate of 65 could be formation
of a cyclopropylcarbinyl biradical, i.e., 66, as usual. This is
followed by regioselective attack on a second (ground-state)
molecule of 49 to generate 67, collapse of which leads to
cyclobutane ring formation. The stereochemical course of this
pathway would implicate the head-to-head dimers as the major
products. However, as discussed above, proof of this particular
point has eluded us to the present time.
Summary. The experiments detailed here demonstrate that
the positioning of a 4-vinyl substituent on a cyclopentenone
ring results in adoption of the “π-route migration” option when
such unsaturation is present. The range of photoisomerizations
available to these unsaturated ketones is thereby expanded
significantly. In the absence of this added unsaturation, the same
transformations are not possible. There exists no reason to expect
these phenomena to be limited to these systems. The involve-
ment of cyclopropylcarbinyl biradicals such as 53, 56, and 66
gives evidence of being easily accessible by a reaction mode
that holds considerable mechanistic diversity. In addition to their
intrinsic novelty, the observed photochemical isomerizations
constitute new, synthetically attractive methods for gaining entry
into polycyclic structures difficult to access by other methods.
It is important to recognize that while cyclopropylcarbinyl
biradicals themselves are not novel species, the structural
environments in which they are generated in this work gives
rise to much of the observed phenomena.
CHCl3).
Anal. Calcd for C14H25BrO2Si: C, 50.45; H, 7.56. Found: C, 50.59;
H, 7.58.
(1S,4R,6R)-4-(Methoxymethoxy)-7,7-dimethylbicyclo[4.1.0]heptan-
3-one (13). To a cold (0 °C) magnetically stirred solution of (-)-
(1R,3R,6S)-4-methylene-7,7-dimethylbicyclo[4.1.0]heptan-3-ol6d,17 (11.40
g, 75.0 mmol) in CH2Cl2 (0 °C) was added chloromethyl methyl ether
(6.60 g, 82.5 mmol) followed by diisopropylethylamine (10.7 g, 82.5
mmol). After reaction had been allowed to proceed for 3 h at room
temperature, CH2Cl2 (250 mL) was introduced prior to washing with
water (2×). The aqueous washes were back-extracted with CH2Cl2 (3×),
and the combined organic layers were washed with brine, dried, and
concentrated. Chromatography of the residue on silica gel (elution with
2% ethyl acetate in hexanes) afforded the MOM ether as a colorless
oil (9.70 g, 87%). IR (neat, cm-1) 1450; 1H NMR (300 MHz, CDCl3)
δ 4.88-4.83 (m, 2 H), 4.60 (d, J ) 6.7 Hz, 1 H), 4.56 (d, J ) 6.7 Hz,
1 H), 3.98 (t, J ) 3.1 Hz, 1 H), 3.37 (s, 3 H), 2.58 (ddt, J ) 16.2, 8.3,
2.7 Hz, 1 H), 2.35-2.25 (m, 2 H), 1.52 (dt, J ) 15.4, 3.2 Hz, 1 H),
0.98 (s, 3 H), 0.88 (s, 3 H), 0.77 (dd, J ) 16.9, 8.9 Hz, 1 H), 0.72
(ddd, J ) 12.8, 9.1, 3.5 Hz, 1 H); 13C NMR (75 MHz, CDCl3) δ 146.1,
111.9, 92.9, 73.9, 55.2, 28.7, 27.2, 25.1, 20.7, 18.1, 15.8, 14.3; HRMS
20
(EI) m/z (M+) calcd 196.1463, obsd 196.1463; [R]D +23.3 (c 2.15,
CHCl3).
Ozone was bubbled through a cold (-78 °C) solution of the above
acetal (11.0 g, 56.1 mmol) in 4:1 methanol-CH2Cl2 (150 mL) until a
blue color persisted (6 h). Dimethyl sulfide (100 mL) was added, and
the resulting mixture was stirred overnight, concentrated under reduced
pressure, and chromatographed over silica gel (elution with 15% ethyl
acetate in hexanes) to give 13 as a colorless oil (11.03 g, 99%). IR
Experimental Section
1
(neat, cm-1) 1720, 1451, 1407; H NMR (300 MHz, CDCl3) δ 4.64
(1R,4R)-2-Bromo-4-(tert-butyldimethylsiloxy)-1-vinyl-2-cyclopenten-
1-ol (11). To freshly dried cerium trichloride43 (46.4 g, 124.7 mmol)
under N2 was added 125 mL of 1.0 M vinylmagnesium bromide in
THF (125 mmol). The mixture was stirred at room temperature for 1
h and an additional hour at -78 °C. At this point, a solution of 1015
(36.22 g, 124.4 mmol) in dry THF (125 mL) was introduced dropwise,
and the mixture was allowed to warm to room temperature overnight
prior to quenching at 0 °C with saturated NH4Cl solution (300 mL)
and extraction with ether (3×). The combined organic layers were
washed with brine, dried, and concentrated to leave a residue that was
chromatographed on silica gel (elution with 5% ethyl acetate in hexanes)
to give 22.37 g (56%) of 11 as a colorless oil. IR (neat, cm-1) 3444,
1493, 1372, 1181; 1H NMR (300 MHz, CDCl3) δ 6.03 (d, J ) 2.0 Hz,
1 H), 5.73 (dd, J ) 17.2, 9.6 Hz,1 H), 5.31 (dd, J ) 17.2, 0.9 Hz,1 H),
5.15 (dd, J ) 10.6, 0.9 Hz, 1 H), 4.64-4.60 (m, 1 H), 2.69 (s, 1 H),
2.62 (dd, J ) 13.3, 6.8 Hz, 1 H), 1.98 (dd, J ) 13.3, 4.8 Hz, 1 H),
0.87 (s, 9 H), 0.07 (s, 3 H), 0.07 (s, 3 H); 13C NMR (75 MHz, CDCl3)
(d, J ) 6.8 Hz, 1 H), 4.60 (d, J ) 6.8 Hz, 1 H), 3.61 (t, J ) 3.6 Hz,
1 H), 3.35 (s, 3 H), 2.79 (dd, J ) 17.5, 8.5 Hz, 1 H), 2.47 (ddd, J )
15.7, 9.2, 3.7 Hz, 1 H), 2.23 (d, J ) 17.6 Hz, 1 H), 1.75 (dt, J ) 15.9,
5.4 Hz, 1 H), 1.15 (dt, J ) 0.6, 3.0 Hz, 1 H), 1.01 (s, 3 H), 0.91-0.74
(m, 1 H), 0.84 (s, 3 H); 13C NMR (75 MHz, CDCl3) δ 213.2, 95.7,
76.8, 55.7, 34.2, 28.1, 27.9, 23.6, 19.2, 16.0, 14.7; HRMS (EI) m/z
(M+) calcd 198.1247, obsd 198.1251; [R]D 40.4 (c 0.835, CHCl3).
20
Anal. Calcd for C11H18O3: C, 66.64; H, 9.15. Found: C, 66.72; H,
9.21.
(1R,2R,4R,6R)-2-Allyl-4-(methoxymethoxy)-7,7-dimethylbicyclo-
[4.1.0]-heptan-3-one (14). Ketone 13 (198 mg, 1.00 mmol) dissolved
in THF (4 mL) was added to a cold (-78 °C), magnetically stirred
solution of lithium hexamethyldisilazide in THF (11 mL of 1 M, 11
mmol). The reaction mixture was stirred for 45 min at this temperature,
maintained at 0 °C for 1 h, and heated at reflux for 30 min. To the hot
enolate solution was added allyl bromide (156 mg, 1.3 mmol) as a
solution in THF (2 mL) and HMPA (1 mL). After 8 h of heating and
subsequent cooling to room temperature, saturated NH4Cl solution (10
mL) was introduced and the aqueous phase was extracted with ether
(43) Paquette, L. A. In Encyclopedia of Reagents for Organic Synthesis,
Wiley: Chichester, UK, 1995; pp 1031-1034.

Excited-State BehaVior of 2-Cyclopentenones
J. Am. Chem. Soc., Vol. 122, No. 40, 2000 9619
(3×). The combined organic layers were dried and concentrated to leave
a residue that was purified by gradient elution chromatography on silica
gel (elution with 2-9% ethyl acetate in hexanes) to separate cyclo-
propyl-cleaved, polyallylated, and O-allylated products. There was
isolated 67 mg (28%) of 14. IR (neat, cm-1) 1710; 1H NMR (300 MHz,
CDCl3) δ 5.84 (dddd, J ) 17.2, 10.6, 9.4, 5.8 Hz, 1 H) 5.00 (m, 2 H),
4.65 (d, J ) 5.8 Hz, 1 H), 4.60 (d, J ) 5.8 Hz, 1 H), 3.68 (t, J ) 2.9
Hz, 1 H), 3.65 (s, 3 H), 3.19 (ddd, J ) 14.0, 8.6, 5.5 Hz, 1 H), 2.60
(ddd, J ) 12.9, 7.1, 5.8 Hz, 1 H), 2.52 (ddd, J ) 15.8, 9.3, 2.1 Hz, 1
H), 2.01 (dt, J ) 14.0, 8.6 Hz, 1 H), 1.90 (dt, J ) 16.1, 3.4 Hz, 1 H),
1.40 (t, J ) 8.6 Hz, 1 H), 1.01 (s, 3 H), 0.85 (s, 3 H), 0.80-0.65 (m,
1 H); 13C NMR (75 MHz, CDCl3) δ 215.1, 136.8, 116.1, 95.8, 78.7,
55.9, 43.5, 32.3, 31.1, 29.2, 28.8, 20.0, 16.8, 16.0; HRMS (EI) m/z
mixture was allowed to warm to room temperature overnight before
being extracted with ether (5×). The combined organic layers were
washed with brine, dried, and chromatographed on silica gel (elution
with 5% ethyl acetate in hexanes) to afford 17 as a colorless oil (40.43
1
g, 95%). IR (neat, cm-1) 3480 (br); H NMR (300 MHz, CDCl3) δ
6.15 (dd, J ) 17.6, 6.8 Hz, 1 H), 5.76-5.63 (m, 1 H), 5.36 (d, J ) 1.0
Hz, 1 H), 5.15-4.97 (m, 4 H), 4.58-4.53 (m, 2 H), 4.49 (d, J ) 6.9
Hz, 1 H), 4.37 (d, J ) 6.9 Hz, 1 H), 3.36-3.08 (m, 1 H), 3.24 (s, 3 H),
3.17 (s, 3 H), 2.75 (dd, J ) 11.9, 6.7 Hz, 1 H), 2.52 (br m, 1 H),
2.09-1.80 (m, 3 H), 1.80 (dd, J ) 11.6, 6.2 Hz, 1 H), 0.98 (s, 3 H),
0.94 (s, 3 H), 0.91 (s, 9 H), 0.82-0.77 (m, 1 H), 0.48 (dd, J ) 9.4, 4.6
Hz, 1 H), 0.06 (s, 6 H) (OH not observed); 13C NMR (75 MHz, CDCL3)
δ 151.6, 138.5, 137.9, 129.6, 115.2, 113.2, 96.1, 89.4, 81.5, 77.6, 73.1,
55.4, 51.0, 44.4, 43.6, 35.5, 28.4, 25.8 (3C), 24.8, 23.9, 22.2, 18.1,
17.6, 16.2, -4.6 (2C); HRMS (EI) m/z (M+) calcd 429.3361, obsd
22
(M+) calcd 238.1563, obsd 238.1566; [R]D +63.0 (c 0.90, CHCl3).
Allyl (1R,4R,6R)-4-(Methoxymethoxy)-7,7-dimethylbicyclo[4.1.0]-
hept-2-en-3-yl Carbonate (15). To a cold (-78 °C), magnetically
stirred solution of lithium hexamethyldisilazide (6.34 mL of 1.0 M,
6.34 mmol) in THF (30 mL) was added dropwise a solution of 13
(966 mg, 4.88 mmol) in THF (6 mL). After 1 h, neat allyl chloroformate
(0.67 mL, 6.34 mmol) was introduced dropwise at the same temperature
and allowed to react for another 30 min. The reaction mixture was
quenched with saturated NH4Cl solution (5 mL), diluted with ether
(30 mL), and extracted with ether (3×). The combined organic layers
were washed with brine, dried, and concentrated. Chromatography of
the residue on silica gel (elution with 10% ethyl acetate in hexanes)
25
429.3316; [R]D +65.6 (C 1.65, CHCl3).
Anal. Calcd for C28H48O5Si: C, 68.25; H, 9.82. Found: C, 68.27;
H, 9.74.
(1R,2S,3R,4R,6R)-2-Allyl-3-[(3R,5R)-3-hydroxy-5-methoxy-5-vinyl-
1-cyclop enten-1-yl]-4-(methoxymethoxy)-7,7-dimethylbicyclo[4.1.0]-
heptan-3-ol (18). A solution of 17 (15.53 g, 31.52 mmol) in dry THF
(200 mL) was cooled to 0 °C, treated with tetrabutylammonium fluoride
(63.0 mL of 1.0 M in THF, 63.0 mmol), stirred at 0 °C, and extracted
with ether. The combined ether layers were dried and concentrated,
and the crude product was redissolved in 2 L of CH2Cl2, redried, and
freed of solvent. Chromatography of the residue on silica gel (elution
with 50% ethyl acetate in hexanes) afforded 18 as a colorless oil (10.39
g, 87%). This compound proved not to be stable and was directly carried
afforded 15 (1.35 g, quantitative) as a colorless oil. IR (neat, cm-1
)
1628; 1H NMR (300 MHz, CDCl3) δ 5.99-5.90 (m,1 H), 5.75 (d, J )
3.9 Hz, 1 H), 5.42 (dd, J ) 17.2, 1.4 Hz, 1 H), 5.30 (dd, J ) 10.4, 1.2
Hz, 1 H), 4.70 (s, 2 H), 4.67 (dd, J ) 5.7, 0.8 Hz, 2 H), 3.98 (dd, J )
5.3, 3.6 Hz, 1 H), 3.40 (s, 3 H), 2.26 (ddd, J ) 15.2, 8.8, 5.2 Hz, 1 H),
2.02 (dt, J ) 15.1, 1.9 Hz, 1 H), 1.15 (dd, J ) 8.4, 3.9 Hz, 1 H), 1.01
(s, 3 H), 1.03-0.90 (m, 4 H); 13C NMR (75 MHz, CDCl3) δ 153.5,
149.4, 131.2, 119.1, 117.4, 95.3, 68.8 (2C), 55.3, 27.6, 27.1, 24.2, 22.9,
17.6, 14.9; HRMS (EI) m/z (M+ - OCH2OCH3) calcd 221.1083, obsd
1
into the next step. IR (neat, cm-1) 3447, 1634, 1443; H NMR (300
MHz, CDCl3) δ 6.19 (dd, J ) 17.6, 10.9 Hz, 1 H), 5.78-5.71 (m, 1
H), 5.50 (d, J ) 1.5 Hz, 1 H), 5.17-5.00 (m, 4 H), 4.64-4.62 (m, 1
H), 4.51 (d, J ) 6.9 Hz, 1 H), 4.44 (s, 1 H), 4.39 (d, J ) 6.9 Hz, 1 H),
3.42-3.22 (m, 1 H), 3.26 (s, 3H), 3.20 (s, 3 H), 2.93 (dd, J ) 12.2,
6.4 Hz, 1 H), 2.50-2.45 (m, 1 H), 2.17-1.97 (series of m, 3 H), 1.79
(dd, J ) 12.2, 7.1 Hz, 1 H), 1.63 (br s, 1 H), 1.36-1.26 (m, 1 H), 1.00
(s, 3 H), 0.95 (s, 3 H), 0.92-0.86 (m, 1 H), 0.38 (dd, J ) 9.4, 4.6 Hz,
1 H); 13C NMR (75 MHz, CDCl3) δ 152.9, 138.0, 137.5, 128.9, 115.2,
25
221.1130; [R]D +133.0 (c 0.91, CHCl3).
Anal. Calcd for C15H22O5: C, 63.81; H, 7.85. Found: C, 63.87; H,
7.89.
113.5, 96.0, 89.6, 81.4, 77.5, 72.4, 55.3, 51.0, 44.1, 43.0, 35.3, 28.3,
(1R,2S,4R,6R)-2-Allyl-4-(methoxymethoxy)-7,7-dimethylbicyclo-
[4.1.0]-heptan-3-one (16). To a solution of 15 (976 mg, 3.46 mmol)
and freshly prepared Pd2(dba)3‚CHCl3 (120 mg, 3 mol %) in dry
-
24.7, 24.0, 22.0, 17.4, 16.0; HRMS (EI) m/z (M+
H2O) calcd
19
360.2302, obsd 360.2317.
dimethoxyethane (8 mL) was added Diphos (276 mg, 20 mol %) in a
single portion. The reaction mixture was stirred overnight, concentrated,
and subjected directly to gradient elution chromatography on silica gel
containing 5% silver nitrate (3-50% ethyl acetate in hexanes) to furnish
563 mg (67%) of 16 and 32% of 13.
(5S)-5-Allyl-3-[(1R,2S,3R,4R,6R)-2-allyl-3-hydroxy-4-(meth-
oxymethoxy)-7, 7-dimethylbicyclo[4.1.0]hept-3-yl]-5-methoxy-2-cy-
clopenten-1-one (19). A solution of 18 (9.87 g, 26.01 mmol) in dry
CH2Cl2 (1 L) was treated with powdered 4 Å molecular sieves (8.02
g), N-methylmorpholine N-oxide (4.57 g, 39.0 mmol), and tetra-n-
propylammonium perruthenate (459 mg, 5 mol %) at 0 °C. The reaction
mixture was warmed to room temperature, stirred for 5 h, concentrated,
and directly chromatographed on silica gel. Elution with 15% ethyl
acetate in hexanes provided 19 as a faint yellow oil (9.29 g, 95%). IR
(neat, cm-1) 3479, 1722, 1698, 1631, 1595; 1H NMR (300 MHz, CDCl3)
δ 6.35 (dd, J ) 17.6, 10.9 Hz, 1 H), 5.86 (s, 1 H), 5.80-5.65 (m, 1
H), 5.31 (d, J ) 10.9 Hz, 1 H), 5.26 (d, J ) 17.6 Hz, 1 H), 5.04-4.98
(m, 2 H), 4.51 (d, J ) 7.0 Hz, 1 H), 4.34 (d, J ) 7.0 Hz, 1 H), 4.24
(s, 1 H), 3.51 (t, J ) 8.3 Hz, 1 H), 3.30 (s, 3 H), 3.21 (s, 3 H), 2.92 (d,
J ) 17.1 Hz, 1 H), 2.66 (d, J ) 17.1 Hz, 1 H), 2.30-2.23 (br s, 1 H),
2.12-2.02 (m, 3 H), 1.41-1.34 (m, 1 H), 1.01 (s, 3 H), 0.98-0.92
(m, 1 H), 0.95 (s, 3 H), 0.40 (dd, J ) 9.4, 4.8 Hz, 1 H); 13C NMR (75
MHz, CDCl3) δ 201.5, 184.1, 136.6, 136.3, 127.1, 115.9, 115.3, 95.2,
86.5, 80.5, 79.1, 55.1, 51.9, 44.7, 43.6, 35.1, 27.9, 24.0, 23.5, 21.9,
1
For 16: colorless oil. IR (neat, cm-1) 1724, 1442, 1377; H NMR
(300 MHz, CDCl3) δ 5.76 (dddd, J ) 17.0, 10.1, 6.9, 6.9 Hz, 1 H),
5.08-5.00 (m, 2 H), 4.74 (d, J ) 6.7 Hz, 1 H), 4.67 (d, J ) 6.7 Hz,
1 H), 3.85 (dd, J ) 4.9, 4.9 Hz, 1 H), 3.38 (s, 3 H), 2.56 (ddt, J )
12.9, 6.4, 5.1 Hz, 1 H), 2.33 (ddd, J ) 15.5, 8.2, 4.2 Hz, 1 H), 2.26-
2.15 (m, 1 H), 2.04 (dt, J ) 14.1, 5.2 Hz, 1 H), 1.88 (dt, J ) 15.9, 5.8
Hz, 1 H), 1.06 (s, 3 H), 1.04 (s, 3 H), 0.88 (dd, J ) 8.8, 6.4 Hz, 1 H),
0.66 (dd, J ) 9.0, 5.5 Hz, 1 H); 13C NMR (75 MHz, CDCl3) δ 211.5,
136.0, 116.3, 95.6, 74.9, 55.6, 44.1, 34.7, 27.9, 27.8 (2C), 19.7, 18.4,
25
15.1; HRMS (EI) m/z (M+) calcd 238.1551, obsd 238.1560; [R]D
+121.2 (c 1.17, CHCl3).
Anal. Calcd for C14H22O3: C, 70.56; H, 9.30. Found: C, 70.62; H,
9.27.
(1R,2S,3R,4R,6R)-2-Allyl-3-[(3R,5R)-3-(tert-butyldimethylsiloxy)-
5-methoxy-5-vinyl-1-cyclopenten-1-yl]-4-(methoxymethoxy)-7,7-
dimethylbicyclo[4.1.0]heptan-3-ol (17). Anhydrous cerium trichloride
was generated from 39.65 g (106.4 mmol) of the heptahydrate by an
accepted method.43 To this white powder was added ketone 16 (20.61
g, 86.6 mmol) dissolved in 335 mL of dry THF. Stirring was continued
at room temperature under N2 for 4 h. To a separate flask containing
bromide 12 (34.50 g, 103.9 mmol) dissolved in dry THF (265 mL)
cooled to -78 °C was introduced tert-butyllithium in pentane dropwise.
When the addition was completed, the mixture was stirred at -78 °C
for 1 h and transferred dropwise via cannula to the ketone‚CeCl3
complex in THF at -78 °C. After 5 h of agitation at this temperature,
saturated NH4Cl solution (100 mL) was slowly introduced and the
22
17.4, 15.7; HRMS (EI) m/z (M+) calcd 376.2251, obsd 376.2218; [R]D
+43.3 (c 1.07, CHCl3).
Anal. Calcd for C22H32O5: C, 70.19; H, 8.57. Found: C, 69.96; H,
8.59.
(1aR,1bS,4aR,7bR,8R,9aR)-1,1a,1b,2,4a,5,7b,8,9,9a-Decahydro-
7b-hydroxy-4a-methoxy-8-(methoxymethoxy)-1,1-dimethyl-6H-cy-
clopropa[3.4]-benz[1.2e]azulen-6-one (5). A 2 L round-bottomed flask
containing freshly distilled benzene (900 mL) was fitted by way of a
high-dilution trap to a reflux condenser. The benzene was heated at
reflux for 15 min, and a solution of phenylmethylene bis(tricyclohex-
ylphosphino)ruthenium dichloride (434 mg, 15 mol %) in freshly
distilled benzene (30 mL) was introduced via cannula. Subsequently,

9620 J. Am. Chem. Soc., Vol. 122, No. 40, 2000
Paquette et al.
a solution of 19 (1.32 g, 3.51 mmol) in the same medium (100 mL)
was added slowly overnight under argon via the high-dilution adapter
as a vigorous reflux rate was maintained. After 24 h of heating, an
additional 434 mg (15 mol %) of the ruthenium catalyst in benzene
(30 mL) was delivered via cannula, and the reflux period was extended
for an additional 24 h. Throughout the entire reaction period, the
solution maintained an orange-brown color. After being left open to
the air for 5 h, a black color was pervasive. The mixture was
concentrated and chromatographed on silica gel (elution with 4 f 10%
ethyl acetate in hexanes) to afford 5 as a colorless oil (1.19 g, 97%).
IR (neat, cm-1) 3496, 1714, 1608, 1455; 1H NMR (300 MHz, CDCl3)
δ 6.17 (s, 1 H), 5.99 (ddd, J ) 11.2, 7.3, 2.0 Hz, 1 H), 5.73 (dd J )
11.3, 2.6 Hz, 1 H), 4.43 (d, J ) 6.7 Hz, 1 H), 4.41 (d, J ) 6.7 Hz, 1
H), 3.69 (t, J ) 8.1 Hz, 1 H), 3.27 (s, 3 H), 3.26 (s, 3 H), 2.63 (s, 2 H),
2.61-2.53 (m, 1 H), 2.18 (dd, J ) 11.2, 5.8 Hz, 1 H), 2.09 (dd, J )
18.8, 7.7 Hz, 1 H), 2.02-1.81 (m, 2 H), 1.02 (s, 3 H), 0.94 (s, 3 H),
0.92-0.86 (m, 1 H), 0.25 (dd, J ) 9.1, 5.9 Hz, 1 H) (OH not observed);
13C NMR (75 MHz, CDCl3) δ 202.7, 181.0, 138.8, 132.2, 130.2, 96.3,
82.9, 79.2, 77.4, 55.5, 51.8, 51.7, 37.2, 32.8, 28.4, 24.3, 23.7, 21.4,
18.7, 15.8; HRMS (EI) m/z (M+) calcd 348.1937, obsd 348.1940;
(m, 1 H), 4.48 (d, J ) 7.0 Hz, 1 H), 4.43 (d, J ) 7.0 Hz, 1 H), 4.05
(d, J ) 3.2 Hz, 1 H), 4.01 (d, J ) 3.2 Hz, 1 H), 3.51 (s, 3 H), 3.53-
3.31 (m, 1 H), 3.40 (d, J ) 16.4 Hz, 1 H) 3.31 (s, 3 H), 2.34-1.98
(series of m, 5 H), 2.22 (d, J ) 16.5 Hz, 1 H), 1.00 (s, 3 H), 0.97 (s,
3 H), 0.87 (t, J ) 8.5 Hz, 1 H), 0.30 (dd, J ) 9.8, 3.5 Hz, 1 H); 13C
NMR (75 MHz, CDCl3) δ 175.9, 163.5, 132.6, 127.1, 98.1, 87.7, 82.4,
80.3, 56.2, 55.3, 51.6, 41.4, 32.4, 29.3, 29.2, 26.1, 24.6, 22.1, 17.5,
16.1; FAB MS m/z (M+ + H) calcd 349.20, obsd 349.23; [R]D25 -77.6
(c 0.41, CHCl3).
Anal. Calcd for C20H28O5: C, 68.94; H, 8.10. Found: C, 68.80; H,
8.37.
For 21: (6% with dioxane as solvent; 52% with benzene as solvent);
white solid, mp 146-149 °C. IR (CHCl3, cm-1) 3400, 1684; 1H NMR
(300 MHz, CDCl3) δ 5.94 (ddd, J ) 2.2, 5.1, 9.9 Hz, 1 H), 5.32 (s, 1
H), 5.25-5.17 (m, 1 H), 4.68 (d, J ) 7.0 Hz, 1 H), 4.53 (d, J ) 7.0
Hz, 1 H), 3.78 (s, 3 H), 3.45 (dd, J ) 4.8, 5.4 Hz, 1 H), 3.37 (s, 3 H),
3.12 (d, J ) 18.6 Hz, 1 H), 2.35-2.15 (m, 3 H), 2.22 (d, J ) 18.5 Hz,
1 H), 2.09-2.00 (m, 1 H), 1.91 (dt, J ) 6.1, 10.6 Hz, 1 H), 1.39 (dt,
J ) 5.2, 15.2 Hz, 1 H), 1.05 (s, 3 H), 1.02 (s, 3 H), 0.81 (ddd, J ) 5.8,
7.4, 9.0 Hz, 1 H), 0.34 (dd, J ) 6.6, 9.1 Hz, 1 H); 13C NMR (75 MHz,
CDCl3) δ 203.9, 190.3, 130.3, 129.0, 105.6, 96.7, 72.4, 59.0, 56.2, 55.1,
46.2, 32.9, 29.8, 29.7, 28.6, 24.5, 23.6, 19.3, 17.8, 15,2; HRMS (EI)
m/z (M+) calcd 348.1937, obsd 348.1956; [R]D20 +5.1 (c 0.70, CHCl3).
CHCl3
25
λmax
232 (ꢀ 2200), 250 (2525), 282 (310), 334 (80); [R]D +221
(c 0.70, CHCl3).
Anal. Calcd for C20H28O5: C, 68.94; H, 8.10. Found: C, 68.66; H,
8.13.
Acknowledgment. The authors are indebted to Prof. Robin
Rogers (The University of Alabama) and Dr. Judith Gallucci
(OSU) for the crystallographic analyses, John Hofferberth for
MM3 calculations, Dr. Kurt Loening for assistance with
nomenclature, and Prof. Martin Newcomb (Wayne State Uni-
versity) for a helpful exchange of information.
Photoisomerization of 5. Prototypical Procedure. Into a quartz
test tube containing the enone sample (20-60 mg) was introduced the
proper solvent (1-3 mL), and the solution was deoxygenated by
bubbling dry N2 through a serum cap seal for 5 min. The needle serving
as a gas outlet was capped, and the tube was positioned on the outside
wall of a quartz condenser fitted internally with a 450 W Hanovia lamp.
The irradiation was performed with careful monitoring of the reaction
progress by TLC. When Pyrex test tubes were used, only very little
reaction was observed in the same time frame. At the end of the process,
the solution was transferred into a round-bottomed flask and evaporated
under reduced pressure without warming. The residue was chromato-
graphed on silica gel (elution with ether-hexane mixtures) in order to
separate the isomeric photoproducts.
Supporting Information Available: Spectroscopic details
for 50a and 50b, experimental details for compounds 22 -51,
and tables giving the crystal data and structure refinement
information, bond lengths and bond angles, atomic and hydrogen
coordinates, and isotropic and anisotropic displacement coor-
dinates for 20, 21, 26, 40, and 42 (PDF). This information is
available free of charge via the Internet at http://pubs.acs.org.
For 20: (68% with dioxane as solvent; 7% with benzene as solvent);
white solid, mp 136.0-137.0 °C. IR (neat, cm-1) 1775, 1626, 1477,
1293; 1H NMR (300 MHz, CDCl3) δ 5.89-5.83 (m, 1 H), 5.18-5.14
JA001813Q