A. L. Lawrence et al. / Tetrahedron Letters 47 (2006) 8717–8720
8719
R
R
R
HO
Me
OH
O
29
Scheme 3. Required geometry of the enolic double bond for correct
orbital overlap.
γ
α
H
H
a
+
γ
H
O
OTMS
25
OTMS
26
16
Scheme 4. Possible c- versus a-deprotonation of 1,5,7-trien-4-one 16.
Reagents and conditions: (a) LDA, 0 °C, TMSCl, THF.
1
5
Figure 1. X-ray crystal structure of cyclic product 20.
was carried out, resulting in a 31% yield of the cyclic
product 21 with the double bond geometry determined
by a NOESY experiment (Table 1, entry 2). H NMR
1
DMSO as the solvent resulted in a yield of 9% of the
cyclized product 22, which was shown to have the same
relative stereochemistry as (À)-coprinolone 1 via NOE
experiments (Table 1, entry 4). The conditions were fur-
ther optimized for trienone 17, with MeOH resulting in
the best yield of 33% (Table 1, entry 5). Trienone 18 was
also cyclized to product 23 in a 24% isolated yield (Table
spectra of the crude TMS trapped enolate taken before
and after heating in benzene showed that a minor ole-
finic species was formed which was converted to 21 upon
heating. However, another major species was also pres-
ent with olefinic protons which did not undergo the elec-
trocyclization. We put forward the hypothesis that some
competing c-deprotonation may be taking place result-
ing in the formation of TMS enolate 26, which cannot
1
, entry 6).
1
3
A by-product in all the attempted cyclizations was the
corresponding trienone 31, where D-1,2 had moved to
form D-2,3, which is in conjugation with the ketone
undergo the electrocyclic ring closure (Scheme 4).
Mills and Beak reported in 1985 that the equilibrium
constant for keto-enol tautomerism is dominated by
the polarity and hydrogen bonding basicity of the sol-
(
Scheme 6). The cyclization of trienone 17 was carried
out in methanol-d , and carbon NMR showed that for-
4
1
4
mation of the conjugated trienone 31 (R = Me) was irre-
versible, as no deuterium was incorporated in the
cyclobutane ring of tricyclic ketone 22.
vent. It was envisaged that in a polar and/or hydrogen
bond acceptor solvent the enol form of trienone 16 may
directly undergo the cyclization in situ. Therefore, tri-
enone 16 was heated to 100 °C in DMSO and pleasingly
cyclic ketones 20 and 21 were isolated in 54% and 5%
yields, respectively (Table 1, entry 3). The exclusive for-
mation of the E-isomer 21 via cyclization of the TMS
trapped enolate (Table 1, entry 2) is in contrast to the
preferential formation of Z-isomer 22 by a simple heat-
ing in DMSO (Table 1, entry 3). We suggest that this
can be accounted for by kinetic formation of Z-isomer
To confirm that the formation of the conjugated tri-
enone 31 was irreversible we decided to attempt the
cyclization of a-phenyl trienone 19. Trienone 19 was
subjected to heating in MeOH and gave the cyclized ke-
tone 24 recovered in a mixture with other unidentifiable
products in a low yield of <8% by mass (Table 1, entry
2
0 by intramolecular keto-enol tautomerism of the cyc-
MeOH or
DMSO
Me
H
lized intermediate 28 (Scheme 5). A crystal structure of
Z-isomer 20 of the cyclic ketone was obtained (Fig. 1).
1
5
R
+
R
R
Δ
O
O
O
With trienone 16 successfully cyclized in a 59% yield it
was decided to attempt the cyclization of trienone 17,
which more closely resembles precursor 4, as it has a
methyl group a to the ketone. It was found that using
2
9
30
31
Scheme 6. Thermal electrocyclization of a 1,5,7-trien-4-one 29 to form
cyclic ketone 30, and the conjugated by-product 31.
MeOH or
DMSO
H
H
Δ
O
OH
27
O H
O
16
28
20
Scheme 5. Isomerizations leading to preferential formation of Z-isomer 20.