Synthesis of linear and angular triquinane skeletons by O-stannyl ketyl-promoted fragmentation-cyclization reactions of α-keto cyclopropanes

Article abstract of DOI:10.1021/jo961655e

Several investigations of rigid α-keto cyclopropane cleavage by O-stannyl ketyls are summarized herein. Tricyclo[3.3.0.0^2,8]octan-3-one ring systems were treated with nBu₃SnH, which produced different ring-cleavage products depending on the location and type of substituent present. An examination of both radical-stabilizing substituents and stereoelectronic factors was initiated to understand what factors bias bond cleavage in a configurationally restricted α-ketocyclopropane via O-stannyl ketyls. A preference for cleavage of the cyclopropane bond with the best orbital overlap with the ketyl radical sp²-orbital even in the presence of radical stabilizing groups is indicated by these results. An O-stannyl ketyl ring scission-cyclization resulted in the novel synthesis of either a linear or an angular triquinane skeleton depending on the length and location an alkene tether on the tricyclo[3.3.0.0^2,8]octan-3-one precursor.

Full text of DOI:10.1021/jo961655e

174
J . Org. Chem. 1997, 62, 174-181
Syn th esis of Lin ea r a n d An gu la r Tr iqu in a n e Sk eleton s by
O-Sta n n yl Ketyl-P r om oted F r a gm en ta tion -Cycliza tion Rea ction s
of r-Keto Cyclop r op a n es
Eric J . Enholm* and Zhaozhong J . J ia
Department of Chemistry, University of Florida, Gainesville, Florida 32611
Received August 27, 1996X
Several investigations of rigid R-keto cyclopropane cleavage by O-stannyl ketyls are summarized
herein. Tricyclo[3.3.0.02,8]octan-3-one ring systems were treated with nBu3SnH, which produced
different ring-cleavage products depending on the location and type of substituent present. An
examination of both radical-stabilizing substituents and stereoelectronic factors was initiated to
understand what factors bias bond cleavage in a configurationally restricted R-ketocyclopropane
via O-stannyl ketyls. A preference for cleavage of the cyclopropane bond with the best orbital
overlap with the ketyl radical sp2-orbital even in the presence of radical stabilizing groups is
indicated by these results. An O-stannyl ketyl ring scission-cyclization resulted in the novel
synthesis of either a linear or an angular triquinane skeleton depending on the length and location
an alkene tether on the tricyclo[3.3.0.02,8]octan-3-one precursor.
Sch em e 1
In tr od u ction
Tributyltin hydride (nBu3SnH) reacts with a variety
of precursor functional groups to generate carbon-
centered radicals.1,2 A much less studied O-stannyl ketyl
radical is formed from the chemically neutral reaction
of nBu3SnH and a ketone or aldehyde.3,4 The electron-
rich radical character of this radical anion species should
promote bond scissions and rearrangement reactions.
However, the participation of an O-stannyl ketyl in
X Abstract published in Advance ACS Abstracts, December 15, 1996.
(1) (a) Giese, B. Radicals in Organic Synthesis: Formation of
Carbon-Carbon Bonds; Pergamon Press: New York, 1986. (b) Neu-
mann, W. P. Synthesis 1987, 665. (c) Curran, D. P. Synthesis 1988,
417, 489. (d) J asperse, C. P.; Curran, D. P.; Ferig, T. L. Chem. Rev.
1991, 91, 1237. (e) Motherwell, W. B.; Crich, D. Free Radical Chain
Reactions in Organic Synthesis; Academic Press: New York, 1992. (f)
Curran, D. P. In Comprehensive Organic Synthesis; Trost, B. M.,
Fleming, I., Eds.; Pergamon Press: New York, 1992; Vol. 4.
subsequent radical rearrangements is generally not well-
understood and has been the subject of only a few sparse
reports.2
An interesting application of an O-stannyl ketyl is to
promote the opening of a strained ring, a reaction
demonstrated in a seminal mechanistic study over 15
years ago, shown in Scheme 1.2,5 O-Stannyl ketyl 2 can
cleave the cyclopropane leading to either 3 or 4. An
R-radical scission of a cyclopropane bearing a radical-
stabilizing function, such as an ester attached to the ring,
generally cleaves to favor stabilization of the radical
center, leading to 3. It is also important to note that a
stereoelectronic requirement (vide infra) in correct orbital
overlap is concurrently essential as well. Interestingly,
except for the early mechanistic studies on simple
substrates, the modern synthetic potential of R-keto
cyclopropane scission by O-stannyl ketyls still remains
completely unexplored.2
The unique strained structure of tricyclo[3.3.0.02,8]-
octan-3-one (5), containing a rigid fused R-keto cyclopro-
pane component, is an interesting template for the study
of the tin ketyl-promoted opening of cyclopropane rings
(Figure 1). Rotation about the C3-C2 bond is not possible
in this tricycle, which should lead to preferred cleavage
of a single bond in the cyclopropane.
(2) Pereyre, M.; Quintard, J . -P.; Rahm, A. Tin in Organic Synthesis;
Butterworths: Boston, 1987.
(3) (a) Enholm, E. J .; Prasad, G. Tetrahedron Lett. 1989, 4939. (b)
Enholm, E. J .; Kinter, K. S. J . Am. Chem. Soc. 1991, 113, 7784. (c)
Enholm, E. J .; Burroff, J . A. Tetrahedron Lett. 1992, 1835. (d) Enholm,
E. J .; Xie, Y. J . Org. Chem. 1995, 60, 1112. (e) Enholm, E. J .; Kinter,
K. S. J . Org. Chem. 1995, 60, 4850. (f) Enholm, E. J .; Whitley, P. E.
Tetrahedron Lett. 1996, 559. (g) Enholm, E. J .; J ia, Z. J . Tetrahedron
Lett. 1996, 1177. (h) Enholm, E. J .; J ia, Z. J . Tetrahedron Lett. 1995,
6819. (i) Enholm, E. J .; Whitley, P. E. Tetrahedron Lett. 1995, 9157.
(j) Enholm, E. J .; J ia, Z. J . J . Chem. Soc., Chem. Commun. 1996, 1567.
(k) Enholm, E. J .; Whitley, P. E.; Xie, Y. J . Org. Chem. 1996, 61, 5384.
(4) (a) Rawal, V.; Krishnamurthy, V.; Fabre, A. Tetrahedron Lett.
1993, 2899. (b) Hagesawa, E.; Ishiyama, K.; Horaguchi, T.; Shimizu,
T. J . Chem. Soc., Chem. Commun. 1990, 550. (c) Hagesawa, E.;
Ishiyama, K.; Horaguchi, T.; Shimizu, T. Tetrahedron Lett. 1991, 2029.
(d) Hagesawa, E.; Ishiyama, K.; Kato, T.; Horaguchi, T.; Shimizu, T.
J . Org. Chem. 1992, 57, 5352. (e) Bowman, W. R.; Brown, D. S.; Burns,
C. A.; Marples, B. A.; Zaidi, N. A. Tetrahedron 1992, 48, 6883. (f)
Degueil-Castaing, M.; Rahm, A.; Dahan, N. J . Org. Chem. 1986, 51,
1672. (g) Neumann, W. P.; Hillgartner, H.; Baines, K. M.; Dicke, R.;
Vorspohl, K.; Kobs, U.; Nussbeutel, U. Tetrahedron 1989, 45, 951. (h)
Yoo, S.; Yi, K.; Lee, S.-H.; J eong, K. Chem. Lett. 1990, 575. (i) Lee, E.;
Tae, J . S.; Chong, Y. H.; Park, Y. C.; Yun, M.; Kim, S. Tetrahedron
Lett. 1994, 129. (j) Sugawara, T.; Otter, B. A.; Ueda, T. Tetrahedron
Lett. 1988, 75. (k) Beckwith, A. L. J .; Roberts, D. H. J . Am. Chem.
Soc. 1986, 5983. (l) Hays, D. H; Fu. G. C. J . Am. Chem. Soc. 1995,
117, 7283. (m) Hays, D. H; Fu. G. C. J . Org. Chem. 1996, 62, 4. (n)
Rawal, V. H.; Dufour, C.; Iwasa, S. Tetrahedron Lett. 1993, 2899. (o)
Newcomb, M. Tetrahedron 1993, 49, 1151. (p) Castellino, A. J .; Bruice,
T. C. J . Am. Chem. Soc. 1988, 110, 7512. (q) Harling, J . D.; Motherwell,
W. B. J . Chem. Soc. Chem. Comm. 1988, 1380. (r) Rawal, V. H.; Dufour,
C. J . Am. Chem. Soc. 1994, 116, 2613.
Cleavage of the R-keto cyclopropane in 5 bearing
appropriate alkene appendages at C1 or C7 can also lead
to new synthetic sequences that allow for the synthesis
of linear- and angular-fused triquinane skeletons.
(5) Pereyre, M.; Godet, J . Y. Tetrahedron Lett. 1970, 3653.
S0022-3263(96)01655-6 CCC: $14.00 © 1997 American Chemical Society

Synthesis of Linear and Angular Triquinane Skeletons
J . Org. Chem., Vol. 62, No. 1, 1997 175
Sch em e 3
F igu r e 1.
Sch em e 2
15, respectively. Thus, an investigation of the O-stannyl
ketyl-promoted cyclopropane scission in tricyclooctanone
systems 6 and 11 will hopefully lead to a powerful new
approach to the synthesis of either skeleton. It is worth
noting that because diquinanes 6 and 11 differ only in
the location of the alkene tether, each might arise from
a similar methodology. This paper will examine the
regiochemical control of the cyclopropane bond scission
in 5 and related molecules. The effects of both radical-
stabilizing substituents and stereoelectronic factors in
configurationally restricted R-keto cyclopropanes will be
considered. The application of those studies to the
triquinane synthetic protocol in Scheme 2 will also be
discussed. To begin the study of these reactions, a
general method to prepare the rigid tricyclic ketones
common to both skeletons was first examined.
Con figu r a tion a lly Restr icted r-Keto Cyclop r o-
p a n es. The construction of tricyclo[3.3.0.02,8]octan-3-one
skeletons has been achieved by different routes, including
metal carbene insertion reactions.8a Although not previ-
ously used in tin hydride reactions, they have been used
in the synthesis of various natural products.8 We decided
to use an oxadi-π-methane (ODPM) rearrangement to
construct this strained tricycle as shown in Scheme
3.8,10,11 Thus, trimethylsilyl chloride was used to trap out
the lithium enolate of cyclohexenone 16 in 71% yield.12
The
Diels-Alder
reaction
of
17
with
di-
methylacetylene dicarboxylate (DMAD) was carried out
in refluxing benzene at 80 °C. The desired cycloadduct
19 was isolated in 51% yield after acidic workup.9 The
photochemical ODPM rearrangement of 19 with a 450
W Hanovia lamp indeed afforded the desired tricyclic
product 20 using acetone as a triplet sensitizer.13
Triquinanes rank among the most important natural
carbon frameworks.6 These carbon skeletons possess
three five-membered rings that share one or two carbon-
carbon bonds. Natural products from a variety of biologi-
cal sources demonstrate that these compounds bear a
wide range of functionality. These interesting structures
have also been the target of many synthetic efforts.6,7
As shown in Scheme 2, if bond “a” is cleaved in the
ring-opening process of 7 or 12, the capture of the
generated radicals by suitably tethered olefins will afford
linear triquinane 10 or angular triquinane framework
O-Stannyl ketyl 25, generated from the diester-func-
tionalized tricyclic ketone 20, can cleave the cyclopropane
ring to afford 27a by “a” cleavage or 27b by “b” cleavage
(Scheme 4).14 We initially thought that the ester (R2 )
CO2R) in 20 might enhance “b” cleavage by intermediate
(8) (a) Demuth, M. In Comprehensive Organic Synthesis; Trost, B.
M., Fleming, I., Eds.; Pergamon Press: New York, 1992; Chapter 2.6.
(b) Demuth, M.; Raghavan, P. R.; Carter, C.; Nakano, K.; Schaffner,
K. Helv. Chim. Acta 1980, 63, 2434. (c) Demuth, M.; Chandrasekhar,
S.; Nakano, K.; Raghavan, P. R.; Schaffner, K. Helv. Chim. Acta 1980,
63, 2440. (d) Demuth, M.; Schaffner, K. Angew. Chem., Int. Ed. Engl.
1982, 21, 820.
(9) Rubottom, G. M.; Krueger, D. S. Tetrahedron Lett. 1977, 611.
(10) For general reviews of oxadi-π-methane rearrangement, see:
(a) Hixon, S. S.; Mariano, P. S.; Zimmerman, H. E. Chem. Rev. 1973,
73, 531. (b) Dauben, W. G.; Lodder, G.; Ipaktschi, J . Top. Curr. Chem.
1975, 54, 73.
(6) For reviews, see: Paquette, L. A.; Doherty, A. M. In Polyquinane
Chemistry. In Reactivity and Structure Concepts in Organic Chemistry;
Springer-Verlag: New York, 1987; Vol. 26. (b) Hudlicky, T.; Rulin, F.;
Lovelace, T. C.; Reed, J . W. In Studies in Natural Products Chemistry;
Atta-ur-Rahman, Ed.; Elsevier: Amsterdam, 1989; Vol. 3, pp 3-72.
(7) (a) Curran, D. P.; Rakiewicz, D. M. J . Am. Chem. Soc. 1985, 107,
1448. (b) Curran, D. P.; Chen, M. H. Tetrahedron Lett. 1985, 4991. (c)
Curran, D. P.; Kuo, S. -C. J . Am. Chem. Soc. 1986, 108, 1106. (d)
Curran, D. P.; Ferig, T. L.; Elliott, R. L. J . Am. Chem. Soc. 1988, 110,
5064. (e) Curran, D. P.; J asperse, C. P. J . Am. Chem. Soc. 1990, 112,
5601.
(11) Houk, K. N. Chem. Rev. 1976, 76, 1.
(12) Rubottom, G. M.; Grube, J . M. J . Org. Chem. 1977, 42, 1051.
(13) (a) Givens, R. S.; Oettle, W. F.; Coffin, R. L.; Carlson, R. G. J .
Am. Chem. Soc. 1971, 93, 3957. (b) Givens, R. S.; Oettle, W. F. J . Am.
Chem. Soc. 1971, 93, 3963.

176 J . Org. Chem., Vol. 62, No. 1, 1997
Enholm and J ia
Sch em e 4
F igu r e 2.
Sch em e 6
Sch em e 5
Sch em e 7
radical stabilization.1,15 Surprisingly, only the product
of “a” bond cleavage was observed in 59% isolated yield.
Spectroscopic evidence was not conclusive because the
diester product was complicated by an equilibrium
mixture with its tautomer, as shown in Scheme 5.16 In
order to clearly confirm the structure, we removed the
carbonyl at C3 using the three-step sequence shown in
Scheme 5, affording bicyclo[3.3.0]octene 31 as the sole
isolable product. Compound 31, confirmed by its 13C
NMR and the attached proton test (APT), offered clear
evidence that 28a was the product of “a” bond cleav-
age.17,18
A second example, using the monoester 24, gave
exclusively the symmetrical molecule 32 in 88% yield
under almost identical conditions with nBu3SnH in
refluxing benzene. This reaction was clearly different
from above because it resulted in bond “b” cleavage.
Symmetry simplified the structure and NMR spectra of
32, which was additionally confirmed by forming its (2,4-
dinitrophenyl)hydrazone derivative 33 (Scheme 6). In
this case, it first appeared that stabilization of the
intermediate radical predominated and bond “b” won out.
Although radical scissions of cyclopropanes bearing
radical-stabilizing functions, such as an ester or phenyl
group, usually favor stabilization of the radical center,
both 20 and 24 have ester functions capable of stabiliza-
tion. Each differs, however, only by the presence of an
additional ester (R1 ) CO2Me) in 20 that is not located
properly for radical stabilization. To explain the two
contrasting results from very similar structures, we
propose that stereoelectronic effects initially favor the
cleavage of bond “a” in both 20 and 24. The cyclopropane
“a” σ-bond has better orbital overlap with the sp2-like
orbital of the adjacent ketyl as seen in structures 34 and
35 (Figure 2).1,15,19 Bond “b,” conversely, is almost
orthogonal to the sp2-like orbital of the ketyl radical.
Another observation is that both processes involve
cyclopropylmethyl-type radicals and potentially can in-
volve a reversible ring-closure process; however, bond
cleavage is highly favored because of the release of ring
strain energy, as shown in Scheme 7.20,21 If R1 ) H in
36 (as in the case of 24), reclosure could become more
facile because no substituent is present and the center
is less hindered. At this point, cleavage of bond “b” likely
occurs, leading to the resonance-stabilized radical inter-
mediate 38b in which the reverse reaction is possible,
but less likely.
(14) Examples of the Li-NH3 opening: (a) Monti, S. A.; Bucheck,
D. J .; Shepard, J . C. J . Org. Chem. 1969, 34, 3080. (b) Yates, P.;
Stevens, K. E. Tetrahedron 1981, 37, 4401. (c) Dauben, W. G.; Deviny,
E. J . J . Org. Chem. 1966, 31, 3794.
If R1 ) CO2Me, the ester sterically blocks or slows the
radical in 38a toward reclosure to 37. Similar rate-
retarding effects by blocking 5-hexenyl radical cycliza-
tions at the internal position (C5) of the alkene are well-
(15) Deslongchamps, P. Stereoelectronic Effects in Organic Chem-
istry; Pergamon Press: Oxford, 1983.
(16) The IR spectrum of 28a /28b showed a broad peak at 3427 cm-1
indicating the presence of an enolized â-keto ester group. Its 1H NMR
spectrum was complex but included a singlet at 10.55 ppm for an enolic
proton. Its 13C NMR had signals at 119.0 and 103.9 ppm, attributed
to an enolic olefin.
(19) (a) Beckwith, A. L. J . Tetrahedron 1981, 37, 3073. (b) Beckwith,
A. L. J .; Easton, C. J .; Serelis, A. K. J . Chem. Soc., Chem. Commun.
1980, 482.
(20) For an example of a reversible cyclopropylmethyl radical, see:
Giese, B.; J ay, K. Chem. Ber. 1977, 110, 1364.
(21) (a) Beckwith, A. L. J .; Moad, G. J . Chem. Soc., Perkin Trans. 2
1980, 1473. (b) Mariano, P. S.; Bay, E. J . Org. Chem. 1980, 45, 1763.
(17) The attached proton test (APT) for structure of 31 featured two
CH3 units, four CH2 units, two CH units, and four quaternary carbons,
ruling out the product of “b” bond cleavage (27b, R1, R2 ) CO2Me).
(18) Hashimoto, S.; Shinoda, T.; Shimada, Y.; Honda, T.; Ikegami,
S. Tetrahedron Lett. 1987, 637.

Synthesis of Linear and Angular Triquinane Skeletons
J . Org. Chem., Vol. 62, No. 1, 1997 177
Sch em e 8
Sch em e 11
Sch em e 9
Sch em e 10
Sch em e 12
established1 but are not well-understood for 3-butenyl
radical cyclizations. It is also noteworthy that if R1 )
CO2Me (as in the case of 20) the reverse reaction of
reclosure (38a f 37) prevents conjugation of the ester
with the olefin, which is also energetically less favorable.
To lend support to the above argument we planned an
investigation of the O-stannyl ketyl-promoted cyclopro-
pane opening in the tricyclic template 39, which lacks
the radical-stabilizing ester function altogether.
Treatment with nBu3SnH and AIBN afforded 40, the
sole product of “a” bond cleavage, in 83% yield (Scheme
8). This result is consistent with the reversible ring
opening-reclosure process discussed above; however,
now the CH2OTBDPS substituent cannot stabilize the
radical by delocalization. The ring opening is now
controlled by stereoelectronic effects alone.
Two other examples that differ from the structures
above, but contain rigid R-keto cyclopropanes, also readily
cleave with O-stannyl ketyls. Schemes 9 and 10 show
their preparation and the results from the radical scis-
sion. Although ring-opened products 43 and 46 in these
reactions have been explained by traditional radical
stability arguments in the past, suitable orbital overlap
with the radical and reversibility should now be consid-
ered a viable pathway in the formation of 46 from 45.
Orbital overlap of the sp2-like ketyl radical formed from
45 is clearly superior for the cleavage of bond “b” rather
than orthogonal bond “a.” These results contrast direct
radical cleavage of bond “a” due to radical stability and
suggest that stereoelectronic factors perhaps play a much
greater role.
Ap p lica tion to th e Syn th esis of Tr iqu in a n es. Now
that cleavage of the cyclopropane can be reliably pre-
dicted, a route to an angular triquinane skeleton was
next studied. The successful implementation of the novel
tandem scission-cyclization approach postulated in
Scheme 2 could now be tested in a real system by the
synthetic approach shown in Scheme 11. Protection of
the ketone carbonyl in 24 (97%), followed by dibal
reduction (90%), readily afforded 47. Oxidation with
PDC smoothly gave an aldehyde (67%) in which an allyl
unit was next added with allylmagnesium bromide. The
ketal protecting group was removed in the standard
acidic workup (85%) for the Grignard reaction, producing
two diastereomers of 48 (ca. 1.3:1) by GC analysis that
were not separable by column chromatography. Treat-
ment of the mixture with nBu3SnH furnished the angular
triquinane 50 in 94% yield. Cleavage of the “a” bond was
the only pathway followed by the reaction.
Excellent stereochemical control was realized in the
5-exo-trig radical cyclization where the endo:exo stereo-
selectivity for the methyl was found to be >57:1.
A
Beckwith chairlike intermediate 49 readily explains the
stereochemistry of the endo-methyl in 50.22 Diketone 51
was obtained by direct oxidation with PCC, providing a
single diastereomer of triquinane diketone 51 in 78%
yield. The endo-methyl stereochemistry in 51 and its
precursor 50 was readily established from Whitesell’s 13C
NMR studies of closely related fused-cyclopentanes.23 The
entire synthetic sequence here is very efficient, producing
triquinane 50 in 46% overall yield from 24.
A model linear triquinane was also constructed using
a related methodology, as shown in Scheme 12.
A
cyclopropane ring would be first installed in commercial
diquinane 52 using a two-step method of monoiodination,
(22) (a) Beckwith, A. L. J . Tetrahedron 1981, 37, 3073. (b) Beckwith,
A. L. J .; Schiesser, C. H. Tetrahedron 1985, 41, 3925.
(23) (a) Whitesell, J . K.; Matthews, R. S. J . Org. Chem. 1977, 42,
3878. See also: (b) Hudlicky, T.; Koszyk, F. J .; Kutchan, T. M.; Sheth,
J . P. J . Org. Chem. 1980, 45, 5020.

178 J . Org. Chem., Vol. 62, No. 1, 1997
Enholm and J ia
followed by treatment with DBU.24 The labile iodide
intermediate was not characterized but was directly used
in the next step after workup. The dehydrohalogenation
constructed the symmetrical tricyclodione 53 in a 52%
overall yield. Next, reaction with the Grignard reagent
of 4-bromo-1-butene gave 54 as the sole stereoisomer,
which was isolated in 64% yield. Stereoselective addition
of the Grignard reagent to the most accessible face of the
carbonyl in 53 was important for later elaboration to the
normal cis,anti,cis configuration of the linear triquinane
skeleton.
Cleavage of the “a” bond by treatment with nBu3SnH
afforded linear triquinane 55 in 83% yield as the only
isolable product. Compared to the angular triquinane,
only modest stereochemical control was realized in this
5-exo-trig radical cyclization where the endo:exo stereo-
selectivity ratio, determined by capillary GC, was ca. 4:1.
As in the angular triquinane, a Beckwith chairlike
intermediate 54 readily explained the stereochemistry
of the endo-methyl in 55.22 The endo-methyl was also
established by comparison with previous 13C NMR stud-
ies of closely related fused-cyclopentanes.23 The endo-
methyl in 55 was observed at 14 ppm by 13C NMR, which
is in good agreement with Whitesell’s reported average
value of 15 ppm for an endo-methyl rather than 20 ppm
for an exo-methyl substituent.
on a capillary gas chromatograph using a fused silica capillary
column (30 m; film thickness 0.25 µm), unless otherwise noted.
2-[(Tr im eth ylsilyl)oxy]-1,3-cycloh exa d ien e (17). This
compound was prepared by the method of Rubottom and co-
workers.12
2,3-Bis(m e t h oxyca r b on yl)b icyclo[2.2.2]oct -3-e n -5-
on e (19). Diene 17 (10.7 g, 59.2 mmol) and dimethyl acety-
lenedicarboxylate (10.1 g, 71.1 mmol) were refluxed in benzene
(200 mL) at 80 °C overnight. After benzene was rotovaped,
to the residue was added ether (200 mL) and 6 M HCl (20
mL). After the mixture was stirred for 8 h, the ether phase
was separated. The aqueous layer was extracted with ether
and ethyl acetate. The organic layers were combined, dried,
and rotovaped. Chromatographic purification of the residue
afforded 19 (9.3 g, 61%) as a yellow thick oil: Rf (1:2 hexane/
ether) 0.30. IR (KBr) 2954, 1723, 1636; 1H NMR δ 3.72 (s,
3H), 3.70 (s, 3H), 3.63 (t, 1H), 3.43 (m, 1H), 1.68-1.88 (m, 4H);
13C NMR δ 208.6, 166.0, 164.7, 143.4, 134.6, 52.4, 49.6, 38.8,
34.9, 24.0, 22.7. Anal. Calcd for C12H14O5: C, 60.50; H, 5.92.
Found: C, 60.37; H, 6.08.
1,2-Bis(m et h oxyca r b on yl)t r icyclo[3.3.0.02,8]oct a n -3-
on e (20). Octenone 19 (180 mg, 0.76 mmol) was dissolved in
acetone (9.6 mL) in a Pyrex tube. The solution was degassed
with dry nitrogen stream for 30 min and then stirred under
direct irradiation of a 450 W Hanovia lamp for 24 h. The
reaction mixture was rotovaped and chromatographed to give
20 (150 mg, 83%) as a clear thick oil: Rf (1:2 hexane/ether)
0.37; IR (KBr) 2955, 1720 (broad),1637 (weak); 1H NMR δ 3.75
(s, 3H), 3.72 (s, 3H), 3.46 (m, 1H), 3.14 (d, 1H), 2.74 (m, 1H),
2.25 (m, 2H), 1.96 (d, 1H), 1.65 (m, 2H); 13C NMR δ 207.0 (s)
169.5 (s), 165.2 (s), 57.3 (s), 56.1 (s), 52.7 (q), 52.3 (q), 47.5 (t),
41.8 (d), 39.8 (t), 38.6 (d), 24.7 (t). Anal. Calcd for C12H11O5:
C, 60.50; H, 5.92. Found: C, 60.43, H, 6.08.
Con clu sion
Collectively, these examples show that the O-stannyl
ketyl-promoted R-ketocyclopropane opening is predomi-
nantly controlled by stereoelectronic effects in rigid ring
systems. A preference for cleavage of the cyclopropane
bond with the best orbital overlap with the ketyl radical
sp2-orbital even in the presence of radical-stabilizing
groups is indicated by these results. If the ring-scission
is reversible, ester substituents can eventually lead to
bond scissions that favor stabilization of the radical by
resonance delocalization. The cleavage reaction was
applied to the synthesis of angular and linear triquinane
carbon skeletons by a novel tandem scission-cyclization
approach.
1-(Eth oxyca r bon yl)tr icyclo[3.3.0.02,8]octa n -3-on e (24).
Diene 17 (15.0 g, 88.2 mmol), ethyl propiolate (13.0 g, 132
mmol), and a few BHT crystals (ca. 20 mgs) were heated in
benzene (44 mL) at 75 °C for 7 days. Then the reaction
mixture was rotovaped to remove benzene and acidified by 1
M HCl. After being stirred overnight, the aqueous mixture
was extracted with chloroform. The chloroform phase was
dried, rotovaped, and chromatographed to give 2-(ethoxy-
carbonyl)bicyclo[2.2.2]oct-2-en-5-one (15.1 g, 88%) as a colorless
oil: Rf (ether) 0.67; 1H NMR δ 7.20 (d, 1H), 4.24 (q, 2H), 3.62
(m, 1H), 3.35 (m, 1H), 2.07 (m, 2H), 1.84-1.56 (m, 4H), 1.33
(t, 3H); 13C NMR δ 210.5, 164.0, 140.0, 138.0, 60.6, 49.6, 39.5,
31.9, 24.1, 22.5, 14.1. Anal. Calcd for C11H14O3: C, 68.02; H,
7.27. Found: C, 67.83; H, 7.40.
The Diels-Alder adduct (316 mg, 1.61 mmol) was next
dissolved in acetone (20 mL) in a Pyrex tube. The solution
was degassed with a dry nitrogen stream for 30 min and
stirred under irradiation of a 450 W Hanovia lamp for 24 h.
The reaction mixture then was rotovaped and chromato-
graphed to produce 24 as a colorless oil (265 mg, 84%): Rf
(ether) 0.67; IR (KBr) 2957, 1724, 1624 (weak); 1H NMR δ 4.17
(q, 2H), 3.42 (m, 1H), 2.71-2.58 (m, 3H), 2.23 (m, 2H), 1.83
(d, 1H), 1.65 (m, 2H), 1.27 (t, 3H); 13C NMR δ 211.7 (s), 171.3
(s), 60.7 (t), 49.8 (s), 47.6 (d), 46.9 (t), 40.6 (t), 38.9 (d), 37.8
(d), 24.6 (t), 14.1 (q). Anal. Calcd for C11H14O3: C, 68.02; H,
7.27. Found: C, 67.88; H, 7.36.
Exp er im en ta l Section
Gen er a l Meth od s. Melting points were determined on a
capillary melting point apparatus and are uncorrected. All
reactions were run under an inert atmosphere of argon using
flame- or heat-dried apparatus. All reactions were monitored
by thin-layer chromatography (TLC) and judged complete
when starting material was no longer visible in the reaction
mixture. All yields reported refer to isolated material judged
to be homogeneous by thin-layer chromatography and NMR
spectroscopy. Temperatures above and below ambient tem-
perature refer to bath temperatures unless otherwise stated.
Solvents and anhydrous liquid reagents were dried according
to established procedures by distillation under nitrogen from
an appropriate drying agent: ether, benzene, and THF from
benzophenone ketyl; CH2Cl2 from CaH2. Other solvents were
used “as received” from the manufacturer.
1,2-Bis(m eth oxycar bon yl)bicyclo[3.3.0]octan -3-on e (28a/
b). To a solution of tricyclic octanone 20 (450 mg, 1.89 mmol)
in benzene (20 mL) were added tributyltin hydride (1.38 g,
4.74 mmol) and AIBN (100 mg, 0.61 mmol). The mixture was
degassed with an argon stream for 15 min and refluxed at 80
°C for 3 h. The reaction mixture was rotovaped and chro-
matographed to yield 28a /b (270 mg, 59%) as a clear thick
oil: Rf (1/2 hexane/ether) 0.48; IR (KBr) 3427 (broad), 1729,
Analytical TLC was performed using precoated silica gel
plates (0.25 mm) with phosphomolybdic acid in ethanol as an
indicator. Column chromatography was performed using
Kieselgel silica gel 60 (230-400 mesh) by standard flash
chromatographic techniques. Product ratios were determined
1
1664, 1626 (weak), 1286; H NMR δ 10.55 (s, 0.8H), 3.78 (s,
3.0H), 3.70 (s, 3.0H), 2.93 (m, 1.1 H), 2.65 (m, 1.0H), 2.39 (m,
1.0 H), 2.26 (d, J ) 18.3 Hz, 1.1H), 1.50-2.10 (m, 7.6H), 1.40
(m, 1.0H), 1.26 (m, 0.5H), 0.90 (m, 0.5H); 13C NMR δ 177.0
(s), 176.3 (s), 119.6 (s), 119.0 (s), 103.9 (s), 68.2 (s), 61.7 (s),
52.1 (q), 51.2 (q), 44.4 (d), 38.9 (t), 35.6 (t), 35.2 (t), 26.0 (t),
25.1 (d), 23.4 (d). Anal. Calcd for C12H16O5: C, 59.99; H, 6.71.
Found: C, 59.87; H, 6.82.
(24) (a) Barluenga, J ; Martinez-Gallo, J . M.; Najera, C.; Yus, M.
Synthesis 1986, 678. (b) Gleiter, R.; J ahne, G.; Muller, G.; Nixdorf,
M.; Irngartinger, H. Helv. Chim Acta 1986, 63, 71.

Synthesis of Linear and Angular Triquinane Skeletons
J . Org. Chem., Vol. 62, No. 1, 1997 179
1,2-Bis(m eth oxyca r bon yl)bicyclo[3.3.0]oct-2-en e (31).
Octanone 28a /b (220 mg, 0.92 mmol) was dissolved in metha-
nol (3 mL). At -78 °C, to the solution was added sodium
borohydride (140 mg, 3.67 mmol). After the mixture was
stirred for 3 h at -78 °C, the reduction reaction was quenched
with water and multiply extracted with ethyl acetate. The
organic layer was rotovaped to afford 29 (200 mg, 90%) as an
oil, Rf (ether) 0.31. The crude 29 (200 mg, 0.83 mmol) was
dissolved in CH2Cl2 (2 mL) with triethylamine (417 mg, 4.13
mmol). At -20 °C, to the mixture was added mesyl chloride
(188 mg, 1.65 mmol, in 2 mL of CH2Cl2) dropwise. After being
stirred for 1 h, the reaction mixture was allowed to warm to
room temperature and stirred for 2 h. The mixture was
rotovaped to afford crude solid mesylate 30. To the mesylate
were added DBU (628 mg, 4.13 mmol) and THF (3 mL). This
mixture was refluxed overnight. The reaction was quenched
with water and extracted with ether. The ether layer was
dried, rotovaped, and chromatographed to produce 31 (35 mg,
20%) as a clear oil: Rf (1:1 hexane/ether) 0.45; 1H NMR δ 6.82
(d, 1H), 3.74 (s, 3H), 3.65 (s, 3H), 2.83 (m, 1H), 2.24 (m, 1H),
2.12 (m, 1H), 1.94 (m, 2H), 1.75 (m, 2H), 1.68(m, 1H), 1.38 (m,
1H); 13C NMR δ 176.6 (s), 164.6 (s), 144.8 (d), 137.6 (s), 65.7
(s), 52.2 (q), 51.5 (q), 49.4 (d), 39.7 (t), 35.6 (t), 35.4 (t), 26.1
(t). Anal. Calcd for C12H16O4: C, 64.27; H, 7.19. Found: C,
64.12; H, 7.31.
methyl]tricyclo[3.3.0.02,8]octan-3-ol (393 mg, 37%) as a clear
thick oil: Rf (1:2 hexane/ether) 0.53; 1H NMR δ 7.71-7.65 (m,
4H), 7.39-7.32 (m, 6H), 4.78 (m, 1H), 3.78 (m, 2H), 2.64 (m,
1H), 2.51 (m, 1H), 2.31-1.79 (m, 4H), 1.57-1.45 (m, 1H), 1.38
(m, 1H), 1.27 (m, 1H), 1.06 (s, 9H); 13C NMR δ 135.5 (d), 134.0
(s), 129.5 (d), 127.5 (d), 76.1 (d) and 73.0 (d), 65.9 (t) and 65.8
(t), 50.5 (s) and 49.2 (s), 46.7 (t), 45.1 (d) and 43.7 (d), 39.8 (t)
and 39.7 (t), 38.6 (d) and 38.2 (d), 32.1 (d) and 29.1 (d), 26.9
(q) and 26.8 (q), 24.9 (t) and 23.6 (t), 19.2 (s); HRMS for
C25H32O2Si calcd 392.2172, found 392.2053.
Pyridinium chlorochromate (PCC) (420 mg, 1.94 mmol) was
finely ground with Kieselgel silica gel 60 (420 mg, 1 wt equiv),
and the light orange powder was suspended in CH2Cl2 (4 mL).
The tricyclic alcohol (320 mg, 0.816 mmol) was dissolved in
CH2Cl2 (2 mL) and added to the PCC suspension. After 2 h,
the reaction mixture was filtered through a Celite bed. The
brown cake was washed with copious amounts of ether. The
filtrate was rotovaped and chromatographed to yield 39 (222
mg, 70%, overall 26%) as a clear thick oil: Rf (1:1 hexane/
ether) 0.46; 1H NMR δ 7.67-7.63 (m, 4H), 7.42-7.37 (m, 6H),
3.92 (m, 2H), 2.87 (dd, 1H), 2.58 (dd, 1H), 2.07 (m, 2H), 1.86
(m, 1H), 1.75 (d, 1H), 1.63-1.51 (m, 2H), 1.27 (m, 1H), 1.06
(s, 9H); 13C NMR δ 214.8 (s), 135.4 (d), 133.5 (s), 129.7 (d),
127.6 (d), 64.3 (t), 51.9 (s), 47.2 (t), 43.2 (d), 39.9 (t), 39.5 (d),
34.4 (d), 26.8 (q), 25.3 (t), 19.2 (s); HRMS for C25H30O2Si M +
H calcd 391.2093, found 391.2040. Anal. Calcd for C25H30O2-
Si: C, 76.88; H, 7.74. Found: C, 76.73; H, 7.84.
8-(Eth oxyca r bon yl)bicyclo[3.2.1]octa n -3-on e (32).
A
mixture of tricyclic ketone 24 (112 mg, 0.571 mmol), tributyltin
hydride (384 mL, 1.43 mmol), and AIBN (30 mg, 0.171 mmol)
in benzene (5.7 mL) was degassed with argon for 15 min and
refluxed at 80 °C for 5 h. The reaction mixture was rotovaped
and chromatographed to afford 32 (106 mg, 94%) as a colorless
oil: Rf (1:2 hexane/ether) 0.53; IR (KBr) 2956, 1709, 1642; 1H
NMR δ 4.22 (q, J ) 7 Hz, 2H), 2.82-2.76 (m, 5H), 2.24 (d, J
) 15 Hz, 2H), 1.89 (d, J ) 9 Hz, 2H), 1.73-1.55 (m, 2H), 1.30
(t, J ) 7 Hz, 3H); 13C NMR δ 211.6 (s), 172.5 (s), 60.5 (t), 49.6
(d), 46.2 (2C, t), 36.6 (2C, d), 29.1 (2C, t), 14.3 (q). Anal. Calcd
for C11H16O3: C, 67.32; H, 8.22. Found: C, 67.18; H, 8.34.
(2,4-Din itr op h en yl)h yd r a zon e (33). In a dry flask was
dissolved (2,4-dinitrophenyl)hydrazine (125 mg, 0.631 mmol)
in methanol (4 mL) with concd H2SO4 (0.2 mL) and the mixture
warmed by a water bath. In another dry flask the bicyclic
ketone 32 (100 mg, 0.510 mmol) was dissolved in methanol (5
mL). To this ketone solution was added dropwise the clear
orange solution of (dinitrophenyl)hydrazine. An orange pre-
cipitate formed immediately. After the solution was chilled
in an ice bath, the precipitate was collected by vacuum
filtration. This dark orange crude product was multiply
recrystallized with ethanol-water. The clean product was
obtained as golden flakes (105 mg, 55%): mp 106-107.5 °C;
1H NMR δ 11.14 (s, 1H), 9.10 (s, 1H), 8.28 (d, 1H), 7.95 (d,
1H), 4.22 (q, 2H), 2.84 (m, 1H), 2.78 (m, 4H), 2.53 (m, 2H),
1.90 (m, 2H), 1.64 (m, 1H), 1.47 (m, 1H), 1.31 (t, 3H); 13C NMR
δ 196.1 (s), 172.3 (s), 158.1 (s), 145.1 (s), 139.6 (s), 129.8 (d),
123.4 (d), 116.3 (d), 60.4 (t), 49.8 (d), 38.3 (t), 36.1 (d), 35.4 (d),
1-[(ter t-Bu tyldiph en ylsiloxy)m eth yl]bicyclo[3.3.0]octan -
3-on e (40). To a solution of the tricyclic ketone 39 (120 mg,
0.308 mmol) in benzene (3.1 mL) was added tributyltin hydride
(250 mL, 0.923 mmol) and AIBN (50 mg, 0.308 mmol). The
mixture was degassed with argon for 15 min and refluxed at
80 °C for 4 h. The reaction mixture was rotovaped and
chromatographed to give 40 (89 mg, 83%) as a clear oil along
with unreacted starting ketone 39 (13 mg): Rf (1:1 hexane/
ether) 0.54; 1H NMR δ 7.66-7.63 (m, 4H), 7.43-7.38 (m, 6H),
3.52 (m, 2H), 2.64 (m, 1H), 2.59-2.52 (m, 2H), 2.11 (m, 2H),
1.95 (m, 1H), 1.75-1.49 (m, 4H), 1.39 (m, 1H), 1.05 (s, 9H);
13C NMR δ 219.6 (s), 135.6 (d), 135.6 (d), 133.4 (s), 133.3 (s),
129.7 (d), 129.7 (d), 127.7 (d), 70.4 (t), 52.5 (s), 48.2 (t), 45.4
(t), 42.7 (d), 36.3 (t), 34.1 (t), 26.9 (q), 24.8 (t), 19.3 (s); HRMS
for C25H32O2Si M + H calcd 393.2250, found 393.2201. Anal.
Calcd for C25H32O2Si: C, 76.48; H, 8.22. Found: C, 76.32; H,
8.30.
5,5-Dip h en ylbicyclo[4.1.0]h ep ta n -2-on e (42).25 Sodium
hydride (60% in oil, 39 mg, 0.968 mmol) was placed in a three-
necked flask, washed with n-pentane, and pumped to dryness.
Trimethyloxosulfonium iodide (213 mg, 0.968 mmol) was
added. DMSO (2 mL) was dripped to the solid mixture
through an addition funnel. After hydrogen evolution, a milky
solution turned clear and was stirred for 15 min. 4,4-Diphenyl-
2-cyclohexen-1-one (200 mg, 0.806 mmol) was added, and the
mixture was stirred overnight. The reaction was quenched
with water and extracted with ether. The ether layer was
dried, rotovaped, and chromatographed to give 42 as a white
31.8 (t), 29.6 (t), 28.3 (t), 14.3 (q). Anal. Calcd for C17H20
-
O6N4: C, 54.25; H, 5.36; N, 14.89. Found: C, 54.15; H, 5.37;
N, 14.93.
1
solid (171 mg, 81%): Rf (ether) 0.65; IR (KBr) 1681; H NMR
δ 7.38-7.17 (m, 10H), 2.55-2.42 (m, 1H), 2.33-2.11 (m, 4H),
1-[(ter t-Bu tyld ip h en ylsiloxy)m eth yl]tr icyclo[3.3.0.02,8]-
octa n -3-on e (39). In a flask the tricyclic ketoester 24 (560
mg, 2.86 mmol) was dissovled in CH2Cl2 (7.2 mL) and chilled
at -78 °C. To this flask was added Dibal (1.0 M solution in
hexane, 14.3 mL, 14.3 mmol) dropwise with a syringe. The
mixture was stirred at -78 °C for 2 h and then warmed to
room temperature. When the reaction was complete, it was
quenched with methanol at -78 °C. To this mixture was
added sodium potassium tartrate (aqueous saturated), and the
resulting mixture was stirred overnight to clear up the aqueous
layer. The mixture was then extracted with ethyl acetate. The
acetate phase was dried and rotovaped to afford a diol (440
mg, 99%), Rf (ethyl acetate) 0.28. Without further purification
this diol (420 mg, 2.73 mmol) was dissolved in pyridine (5.4
mL). At 0 °C, to this solution was added tert-butyldiphenylsilyl
chloride (851 mL, 3.27 mmol). The mixture was stirred for
10 h and quenched with water. The mixture then was
rotovaped and pumped to remove pyridine, and the residue
was chromatographed to give 1-[(tert-butyldiphenylsiloxy)-
1.89-1.76 (m, 1H), 1.41-1.35 (m, 1H), 1.21-1.13 (m, 1H); 13
C
NMR δ 207.3 (s), 148.3 (s), 146.3 (s), 128.3 (d), 128.2 (d), 127.7
(d), 126.7 (d), 126.4 (d), 126.2 (d), 44.5 (s), 33.5 (t), 28.5 (d),
27.9 (t), 27.6 (d), 10.0 (t); HRMS for C19H18O M + H calcd
263.1436, found 263.1441.
3-Meth yl-4,4-d ip h en ylcycloh exa n on e (43). The mixture
of 42 (84 mg, 0.321 mmol), tributyltin hydride (173 mL, 0.642
mmol), and AIBN (53 mg, 0.321 mmol) in benzene (3.2 mL)
was degassed by argon stream for 15 min. The mixture was
refluxed at 80 °C for 2.5 h. After being quenched with ethanol,
the reaction mixture was rotovaped and chromatographed to
give 43 as a white solid (73 mg, 86%): Rf (ether) 0.71; 1H NMR
δ 7.55-7.08 (m, 10H), 3.37-3.32 (m, 1H), 2.98-2.86 (m, 2H),
2.70 (m, 1H), 2.41-2.38 (m, 1H), 2.36-2.25 (m, 2H), 0.81 (d,
J ) 7 Hz, 3H); 13C NMR δ 210.9 (s), 146.8 (s), 145.0 (s), 128.9
(d), 128.3 (d), 126.8 (d), 126.5 (d), 126.2 (d), 125.7 (d), 48.2 (s),
(25) Corey, E. J .; Chaykovsky, M. J . Am. Chem. Soc. 1965, 87, 1353.

180 J . Org. Chem., Vol. 62, No. 1, 1997
Enholm and J ia
45.7 (t), 38.4 (t), 37.7 (d), 29.6 (t), 16.7 (q); HRMS for C19H20
M + H calcd 265.1592, found 265.1598. Anal. Calcd for
C19H20O: C, 86.31; H, 7.63. Found: C, 86.27; H, 7.85.
6-(Meth oxyca r bon yl)bicyclo[4.1.0]h ep ta n -2-on e (45).
This compound was identical to that prepared by Cossy and
co-workers.25,26
O
(d), 117.8 (t), 71.2 (d), 55.2 (s), 47.3 (t), 43.2 (d), 40.2 (t), 40.0
(t), 38.3 (d), 34.3 (d), 25.1 (t) for the major diastereomer and δ
215.3 (s), 134.3 (d), 117.9 (t), 70.9 (d), 54.8 (s), 47.1 (t), 42.7
(d), 40.0 (t), 39.8 (t), 38.2 (d), 34.6 (d), 25.1 (t) for the minor
diastereomer; HRMS for C12H16O2 M + H calcd 193.1229,
found 193.1226.
11-H yd r oxy-9-m e t h ylt r icyclo[6.3.0.01,5]u n d e ca n -3-
on e (50). Ketone 48 (120 mg, 0.625 mmol) was dissolved in
benzene (6.2 mL). To the solution were added tributyltin
hydride (336 mL, 1.25 mmol) and AIBN (102 mg, 0.625 mmol).
The mixture was degassed with argon for 15 min and refluxed
overnight. After 14 h ethanol was added to quench the
reaction. The mixture was rotovaped and chromatographed
to yield the cyclization product 50 as an oil (113 mg, 94%).
The GC ratio of the major (9-endo-methyl) and the minor (9-
exo-methyl) cyclization products was 57:1. The GC ratio of
the two C11 diastereomers of the 9-endo-methyl products was
still 1.3:1. The major and minor cyclization products had the
same Rf, and they were not separable from each other: Rf
4-(Meth oxyca r bon yl)cycloh ep ta n on e (46). This com-
pound was identical to that prepared by Cossy and co-
workers.25,26
1-(Hyd r oxym eth yl)tr icyclo[3.3.0.02,8]octa n -3-on e Eth -
ylen e Keta l (47). The mixture of 24 (2000 mg, 10.3 mmol),
anhydrous ethylene glycol (1.72 mL, 30.9 mmol), PPTS (170
mg, catalyst), and benzene (20 mL) was refluxed overnight in
a 100 mL flask fitted with a Dean-Stark tube. After 12 h
the reaction mixture was poured into cold water and extracted
with ether. The ether phase was dried, rotovaped, and
chromatographed to yield 1-(ethoxycarbonyl)tricyclo[3.3.0.02,8]-
octan-3-one ethylene ketal (2203 mg, 90%) as a colorless oil,
and a small amount of unreacted 24 (148 mg) was recovered:
1
(ether) 0.56; IR (KBr) 3432, 1727; H NMR δ 3.98-3.82 (m,
1
Rf (1:1 ether/hexane) 0.44; IR (KBr) 1720; H NMR δ 4.11 (q,
1H), 3.02-2.90 (m, 1H), 2.73-2.40 (m, 3H), 2.27-2.00 (m, 3H),
1.96-1.20 (m, 6H), 0.97-0.92 (3H: 0.96, d, J ) 7 Hz; 0.93, d,
J ) 7 Hz); 13C NMR δ 221.2 (s), 81.2 (d), 59.6 (s), 55.2 (d), 51.3
(t), 44.0 (t), 41.8 (d), 40.2 (t), 34.0 (t), 31.0 (d), 27.9 (t), 14.5 (q)
for the major C11 diastereomer and δ 220.7 (s), 77.3 (d), 63.0
(s), 54.9 (d), 46.7 (d), 44.9 (t), 43.1 (t), 41.9 (t), 34.0 (t), 33.2
(d), 27.0 (t), 14.6 (q) for the minor C11 diastereomer; HRMS
for C12H18O2 M + H calcd 195.1385, found 195.1383.
J ) 7 Hz, 2H), 4.00-3.85 (m, 4H), 3.18-3.13 (m, 1H), 2.53-
2.45 (m, 1H), 2.33-1.95 (m, 5H), 1.63 (d, J ) 14 Hz, 1H), 1.60-
1.54 (m, 1H), 1.24 (t, J ) 7 Hz, 3H); 13C NMR δ 173.1 (s), 117.5
(s), 64.7 (t), 63.8 (t), 60.2 (t), 48.2 (t), 48.2 (s), 45.0 (d), 40.6
(d), 40.4 (t), 37.7 (d), 23.6 (t), 14.1 (q). Anal. Calcd for
C13H18O4: C, 65.53; H, 7.61. Found: C, 65.38; H, 7.68.
This compound (3210 mg, 13.5 mmol) was dissolved in CH2-
Cl2 (25 mL) and chilled to -78 °C. To this solution was added
Dibal (1.0 M in hexane, 28.3 mL, 28.3 mmol) dropwise via a
syringe. After 1 h the reaction was quenched with ethanol,
warmed, and diluted with a large amount of ethyl acetate. The
aqueous solution of Rochelle’s salt was added, and the mixture
was vigorously stirred to clear up the aqueous phase. The
mixture was separated, and the aqueous layer was extracted
with ethyl acetate. The organic phase was dried, rotovaped,
and chromatographed to give 47 as a colorless oil (2.53 g,
96%): Rf (ether) 0.36; IR (KBr) 3406; 1H NMR δ 4.01-3.66
(m, 6H), 3.34-3.30 (m, 1H), 2.71-2.67 (m, 1H), 2.18-1.88 (m,
3H), 1.59 (d, J ) 14 Hz, 1H), 1.57-1.47 (m, 3H); 13C NMR δ
119.0 (s), 64.6 (t), 64.5 (t), 63.4 (t), 49.0 (s), 48.5 (t), 42.7 (d),
40.0 (t), 38.6 (d), 30.0 (d), 23.9 (t); HRMS for C11H16O3 M + H
calcd 197.1178, found 197.1226.
1-(1-H yd r oxy-3-b u t e n yl)t r icyclo[3.3.0.02,8]oct a n -3-
on e (48). The alcohol 47 (2050 mg, 10.5 mmol) was dissolved
in CH2Cl2 (20 mL). To the stirred solution were added Celite
(4.0 g) and PDC (7866 mg, 21.0 mmol). The reaction was
stirred overnight. The mixture was diluted with a large
amount of ether and vigorously stirred for 20 min. Then the
ether solution was decanted, and a new portion of ether was
added to the residue. This process was repeated five times.
The decanted ether solution was combined, rotovaped, and
chromatographed to give 1-formyltricyclo[3.3.0.02,8]octan-3-one
ethylene ketal as an oil (1368 mg, 67%): Rf (2:1 ether/hexane)
0.45; 1H NMR δ 9.09 (s, 1H), 4.03-3.85 (m, 4H), 3.25-3.20
(m, 1H), 2.50-2.23 (m, 4H), 2.13-1.99 (m, 2H), 1.69 (d, J )
15 Hz, 1H), 1.66-1.59 (m, 1H); 13C NMR δ 198.0 (d), 116.9
(s), 64.8 (t), 63.9 (t), 59.6 (s), 48.3 (t), 45.3 (d), 40.5 (t), 37.9
(d), 37.6 (d), 23.3 (t); HRMS for C11H14O3 calcd 194.0943, found
194.0958. The aldehyde (480 mg, 2.47 mmol) was dissolved
in THF (12 mL) and chilled to -78 °C. To the solution was
added dropwise allylmagnesium bromide (1.0 M in ether, 9.9
mL, 9.9 mmol) via a syringe. After 1 h the dry ice bath was
removed. After being stirred for another hour the reaction was
quenched with ethanol and acidified with 3 M HCl. After
being stirred for 30 min, the mixture was extracted with ethyl
acetate. The acetate phase was dried, rotovaped, and chro-
matographed to give 48 as an oil (403 mg, 85%). The GC ratio
of the two diastereomers of was found to be 1.3:1: Rf (ether)
0.50; IR (KBr) 3419, 1709; 1H NMR δ 5.93-5.78 (m, 1H), 5.20-
5.10 (m, 2H), 3.84-3.70 (m, 1H), 3.00-2.93 (m, 1H), 2.82 (s,
broad, 1H), 2.64-2.52 (m, 1H), 2.45-2.25 (m, 2H), 2.16-2.00
(m, 4H), 1.80-1.74 (1H; 1.77, d, J ) 18 Hz, 0.43H; 1.76, d, J
) 18 Hz, 0.57H), 1.69-1.53 (m, 2H); 13C NMR δ 215.5 (s), 134.5
11-Oxo-9-m eth yltr icyclo[6.3.0.01,5]u n d eca n -3-on e (51).
The triquinane alcohol 50 (31 mg, 0.16 mmol) was dissolved
in CH2Cl2 (0.5 mL). PCC (70 mg, 0.32 mmol) and silica gel
(70 mg) were finely ground and added to the CH2Cl2 solution.
After the solution was stirred for 1 h, a large amount of ether
was added and the dark brown suspension was filtered
through a Celite bed. The bed was thoroughly rinsed with
ether. The ether phase then was rotovaped and chromato-
graphed to give the oxidation product 51 as an oil (24 mg,
1
78%): Rf (2:1 ether/hexane) 0.47; IR (KBr) 1728; H NMR δ
2.70-2.66 (m, 1H), 2.64-2.58 (m, 2H), 2.46-2.04 (m, 6H),
1.93-1.83 (m, 1H), 1.56-1.45 (m, 1H), 1.39-1.21 (m, 2H), 1.14
(d, J ) 6 Hz, 3H); 13C NMR δ 222.1 (s), 216.5 (s), 64.4 (s), 54.4
(d), 46.6 (t), 45.6 (d), 43.8 (t), 43.0 (t), 34.8 (t), 30.1 (d), 26.9
(t), 16.2 (q); HRMS for C12H16O2 calcd 192.1150, found 192.1159.
1,5-Dim eth yltr icyclo[3.3.0.02,8]octa n e-3,7-d ion e (53).24
To the stirred solution of 52 (1000 mg, 6.02 mmol) and HgCl2
(820 mg, 3.01 mmol) in CH2Cl2 (12 mL) was added iodine (1530
mg, 6.02 mmol). HgCl2 was not initially soluble in CH2Cl2,
which turned dark purple after the iodine addition. A dark
orange powder was gradually formed and accumulated in
amount, which was HgI2. The reaction was stopped after 6 h,
and the insoluble orange HgI2 was removed by filtration. The
CH2Cl2 solution was washed with 0.5 M Na2S2O3 and saturated
KI aqueous solutions, dried, and rotovaped to give a crude light
brown solid (1.43 g). The solid was dissolved in MeCN (13
mL). To this stirred solution was added DBU (880 mL, 5.90
mmol, in 3 mL of MeCN) dropwise via an addition funnel. The
mixture was stirred overnight and rotovaped. The residue was
dissolved in CHCl3 and washed with water, 1 M HCl, and
saturated NaHCO3. The organic solution then was dried,
rotovaped, and chromatographed to give 53 (508 mg, 51%) as
a white needle crystal: Rf (ether) 0.39; 1H NMR δ 2.56 (d, J )
17 Hz, 2H), 2.36 (s, 2H), 2.16 (d, J ) 17 Hz, 2H), 1.51 (s, 3H),
1.47 (s, 3H); 13C NMR δ 208.3 (s), 56.2 (t), 47.5 (d), 41.8 (s),
22.8 (q), 13.6 (q). Anal. Calcd for C10H12O2: C, 73.15; H, 7.37.
Found: C, 72.98; H, 7.32.
7-(3-Bu ten yl)-1,5-dim eth yl-7-h ydr oxytr icyclo[3.3.0.02,8]-
octa n -3-on e (54). The dione 53 (280 mg, 1.71 mmol) was
dissolved in THF (5.7 mL) and chilled to -78 °C. To the stirred
solution was dropwise added 3-butenylmagnesium bromide
(0.5 M in THF, 4.44 mL, 2.22 mmol) through a syringe in 10
min. The reaction was quenched for 3 h with water, acidified
with 1 M HCl, and extracted with ethyl acetate. The acetate
phase was dried, rotovaped, and chromatographed. Unreacted
starting dione (46 mg) was recovered, and 54 was isolated as
1
a white solid (188 mg, 64%): Rf (ether) 0.55; H NMR δ 5.82
(26) Cossy, J .; Furet, N.; BouzBouz, S. Tetrahedron 1995, 51, 11751.
(m, 1H), 4.95 (m, 2H), 3.42 (s, 1H), 2.40-2.08 (m, 4H), 2.00-

Synthesis of Linear and Angular Triquinane Skeletons
J . Org. Chem., Vol. 62, No. 1, 1997 181
1.67 (m, 6H), 1.24 (s, 3H), 1.21 (s, 3H); 13C NMR δ 214.9 (s),
138.6 (d), 114.4 (t), 80.3 (s), 58.5 (t), 54.8 (t), 50.4 (d), 49.5 (s),
46.3 (s), 45.7 (d), 42.3 (t), 28.7 (t), 23.0 (q), 14.8 (q); HRMS for
C14H20O2 M + H calcd 221.1542, found 221.1544. Anal. Calcd
for C14H20O2: C, 76.33; H, 9.15. Found: C, 76.31; H, 9.16.
Lin ea r Tr iqu in a n e (55). The tricyclic ketone 54 (34 mg,
0.155 mmol), tributyltin hydride (125 mL, 0.465 mmol), and
AIBN (25 mg, 0.155 mg) were dissolved in benzene (0.6 mL).
The mixture was degassed with a steady argon stream for 15
min and refluxed overnight. The reaction was quenched with
ethanol and chromatographed to afford 55 as an oil (28 mg,
HRMS for C14H22O2 calcd 222.1620, found 222.1616. Anal.
Calcd for C14H22O2: C, 75.63; H, 9.97; Found: C, 75.62; H,
9.95.
Ack n ow led gm en t. We gratefully acknowledge sup-
port from the National Science Foundation (Grant CHE-
9414240) for this work.
Su p p or tin g In for m a tion Ava ila ble: Spectral data for
compounds 42, 47, 48, 50, and 51 (5 pages). This material is
contained in libraries on microfiche, immediately follows this
article in the microfilm version of the journal, and can be
ordered from the ACS; see any current masthead page for
ordering information.
1
83%): Rf (ether) 0.68; IR 4000 (broad), 2956, 1708, 1462; H
NMR δ 2.35 (q, J ) 18 Hz, 1H), 1.97-1.85 (m, 2H), 1.73-1.66
(m, 2H), 1.53-1.24 (m, 9H), 0.91-0.78 (m, 6H); 13C NMR δ
214.1 (s), 81.0 (s), 58.9 (s), 55.0 (t), 50.0 (d), 46.5 (s), 46.0 (d),
43.1 (t), 29.3 (t), 27.4 (t), 23.2 (q), 15.0 (q), 13.7 (q), 8.8 (t);
J O961655E