Studies toward the total synthesis of garsubellin A: Synthesis of 8-deprenyl-garsubellin A

Article abstract of DOI:10.1016/S0040-4039(02)00650-0

During studies directed toward the total synthesis of garsubellin A, synthesis of 8-deprenyl-garsubellin A, containing three (of four in the natural product) stereocenters with the proper configurations, was achieved. Key steps are a one-pot lactone forma

Full text of DOI:10.1016/S0040-4039(02)00650-0

TETRAHEDRON
LETTERS
Pergamon
Tetrahedron Letters 43 (2002) 3621–3624
Studies toward the total synthesis of garsubellin A:
synthesis of 8-deprenyl-garsubellin A
Hiroyuki Usuda, Motomu Kanai and Masakatsu Shibasaki*
Graduate School of Pharmaceutical Sciences, The University of Tokyo, Hongo, Bunkyo-ku, Tokyo 113-0033, Japan
Received 6 March 2002; accepted 29 March 2002
Abstract—During studies directed toward the total synthesis of garsubellin A, synthesis of 8-deprenyl-garsubellin A, containing
three (of four in the natural product) stereocenters with the proper configurations, was achieved. Key steps are a one-pot lactone
formation, a formal migration to the bicyclo[3.3.1]nonane-1,3,5-trione, and an intramolecular Wacker-type tetrahydrofuran
formation. © 2002 Elsevier Science Ltd. All rights reserved.
one-pot lactone formation (from 11a/12a to 13, see
below). We report herein a synthesis of 8-deprenyl-gar-
subellin A (2) in which all the stereocenters match the
natural configurations.
Garsubellin A (1, Fig. 1), isolated from Garcinia subel-
liptica by Fukuyama et al., is a potent inducer of
choline acetyltransferase and might thus be a promising
pharmaceutical agent for treating Alzheimer disease.1
Structurally, garsubellin
A
is
a
polyprenylated
Our retrosynthetic analysis guided by previous studies
for the synthesis of 3 is shown in Scheme 1. The prenyl
group at the C-2 position should be introduced at the
last stage via Stille coupling between tributylprenyltin
phloroglucin derivative characterized by a highly oxy-
genated and densely substituted bicyclo[3.3.1]nonane-
1,3,5-trione core fused to a tetrahydrofuran ring and
appended by a prenyl side chain. The structural com-
plexity and the biologic activity of garsubellin A
prompted us2 and the Nicolaou group3 to investigate its
total synthesis. We recently reported a concise synthesis
of the most advanced tricyclic core in a model system 3.
The obtained stereoisomer, however, contained an
unnatural configuration at C-18 resulting from the key
Figure 1.
Keywords: garsubellin A; cyclic carbonate; one-pot reaction; Tamao
oxidation; Wacker-type reaction; Stille coupling.
* Corresponding author. Tel.: +81-3-5841-4830; fax: +81-3-5684-5206;
e-mail: mshibasa@mol.f.u-tokyo.ac.jp
Scheme 1. Retrosynthetic analysis.
0040-4039/02/$ - see front matter © 2002 Elsevier Science Ltd. All rights reserved.
PII: S0040-4039(02)00650-0

3622
H. Usuda et al. / Tetrahedron Letters 43 (2002) 3621–3624
(5) and vinyl iodide 4. The tetrahydrofuran ring of 4
should be constructed via an intramolecular Wacker-
type reaction of enone 6. The bicyclo[3.3.1]nonane-
1,3,5-trione core of 6 should be constructed through a
formal migration of lactone 7. In this transformation,
the silyl group should act as a scaffold to produce the
enone olefin, as well as increase the flexibility of the
attached carbon chain (C-1–C-3) for the C-1 elec-
trophilic carbon to reach the C-6 nucleophilic carbon.
Compound 7 should be synthesized through a conju-
gate addition of a silyl group to lactone 8, which should
be synthesized by a one-pot lactone formation from 9
and 10. There are two major concerns in the present
synthesis. (1) How is the stereoselectivity of the key
lactone formation reversed? (2) Is the synthetic
sequence developed for the epi-form (3) applicable to
the isomer containing the natural configurations?
We started our investigations on point (1). Previously,
we developed a one-pot lactone formation from 11a/
12a and 10 to 13a (18-epi ) (Scheme 2).2 The overall
reaction proceeded through a regioselective potassium
enolate formation at C-4 of 11a/12a, transmetalation to
the corresponding lithium enolate, Michael addition to
10, followed by an elimination of the chloride, enolate
formation at C-6, and lactonization. The stereoselectiv-
ity of this reaction was high (5:1–25:1), giving unnatural
13a as the major isomer, irrespective of the configura-
tion at C-6 (Scheme 2, entries 1–3). Therefore, the
selectivity was mainly determined by the stereochem-
istry at C-18. Because it was impossible to invert the
configuration of C-18 at a later stage of the synthesis,4
we attempted to reverse the stereoselectivity of this key
lactone formation. The detailed structure of the corre-
sponding lithium enolate derived from 11a/12a was not
clear, however, we hypothesized that the high stereose-
lectivity of the lactone formation could be attributed to
the coordination of the oxygen atom at C-18 to the
lithium atom of the lithium enolate. Based on this
hypothesis, we expected that the stereoselectivity would
change if the coordination of the C-18 hydroxyl group
was reduced by choosing the appropriate protecting
group. Thus, we synthesized 12b and 11c/12c contain-
ing acetyl groups and a cyclic carbonate, respectively,
from the previously reported intermediate.5
Scheme 2. Key step one-pot lactone formation. Reagents and
conditions: (a) 1. KOtBu (3 equiv.), −78°C, 100 min; 2. LiClO4
(4.5 equiv.), −78°C, 50 min; 3. 10 (2 equiv.), −78°C, 5.5 h; 4.
DMAP (5 equiv.), 12-crown-4 (6 equiv.), −78°C to rt, 30 h.
to the lithium atom of the intermediate lithium enolate.
In the case of acetonide 11a/12a, the oxygen atom on
C-18 strongly coordinates to the lithium, which con-
strains the bulky acetal moiety at the position to block
the electrophile entry from the a-side of the electrophile
(Fig. 2). On the other hand, in the case of cyclic
carbonate 11c/12c, the electrophile should react
through the axial attack to the enolate in the most
probable conformation.8 Because the stereochemistry of
C-4 and C-6 disappears in this one-pot lactone forma-
tion, we performed the reaction with a mixture of
diastereomers 11c and 12c (1:1.3) to produce 8c and 13c
in a ratio of 1.2:1 (60% yield, Scheme 2, entry 7). This
procedure could greatly simplify the overall synthetic
route. These isomers were easily separable by silica gel
column chromatography.
Results of the lactone formation from 12b and 11c/12c
are summarized in Scheme 2 (entries 4–7).6 Although
the reaction from acetyl-protected 12b gave a low yield
of the target lactone (entry 4), the cyclic carbonate-pro-
tected 11c gave 8c containing the natural configurations
as the major isomer (4:1, entry 5). On the other hand,
the other diastereomer 12c gave the 18-epi form 13c as
the major isomer with a reduced selectivity (1:2.6, entry
6). These results indicated that in sharp contrast to the
acetonide-protected 11a/12a, the direction of the elec-
trophile entry was mainly determined by the configura-
tion at the C-6 carbon in the case of 11c/12c.7 The
dependency of the stereoselectivity on the different
stereocenters of the substrates (C-18 in the case of
acetonide 11a/12a versus C-6 in the case of cyclic
carbonate 11c/12c) might be explained by the difference
in the coordination ability of the oxygen atom on C-18
Figure 2. Working model to explain the stereoselectivity.

H. Usuda et al. / Tetrahedron Letters 43 (2002) 3621–3624
3623
Once the key intermediate lactone 8c was produced, the
next set of tasks were to introduce a silicon at C-3,
formal migration to the bicyclo[3.3.1] system, and re-
generation of the olefin through Tamao oxidation9 and
b-elimination to give enone 17. In previous studies, the
pentamethyldisilyl group10 was optimal for our purpose
in terms of its stability during the several conversions
and susceptibility to Tamao oxidation, when necessary.
In the present case (Scheme 3), however, although the
conjugate addition of a silylcuprate and the following
migration to the bicyclo[3.3.1] system 15 were success-
ful, Tamao oxidation using mCPBA and TBAF gave
epoxide 16 as the major product. Compound 16 was
produced through epoxidation of the concomitant
enone 17 that resulted from b-elimination of the target
alcohol under basic conditions of Tamao oxidation.
Scheme 3. Initial investigations.
Therefore, we conducted Tamao oxidation at the early
stage of lactone 7 (Scheme 1), in which b-elimination of
the resulting alcohol should be less probable compared
to the corresponding b-hydroxy ketone derived from
15. In this new strategy, we expected that the b-
hydroxyl group would remain intact during the forma-
tion of the bicyclo[3.3.1] system, and then re-generate
the olefin by treatment with a base. To avoid b-elimina-
tion in the Tamao oxidation step, a labile Et2NPh2Si
group11 that can be converted to the alcohol under
neutral conditions12 should be used. Thus, the synthesis
of 8-deprenyl- garsubellin A (2) was achieved, as shown
in Scheme 4.
Et2NPh2Si group13 was introduced through a conjugate
addition of a silylcuprate prepared from CuCN and
Et2NPh2SiLi in a ratio of 1:314 to give 18. Because this
silicon was unstable, crude 18 was directly subjected to
Tamao oxidation conditions, and the following protec-
tion with a TES group gave 20 in 33% yield (three steps
from 8c). Lactone 20 was then treated with Me2AlSEt
to give thioester 21, which was converted under
Fukuyama conditions15 to a mixture of aldehyde 22
and lactol 23.16 Treatment of the mixture with 0.2 M
HCl in THF converted 23 to 22, and crude 22 was
successfully converted to the desired bicyclo[3.3.1] sys-
tem 24 by treatment with Al2O3.17 Oxidation of the
secondary alcohol of 24 with Dess–Martin periodinane
gave trione 25, which was converted to enone 17
through b-elimination of TESOH by DBU. Basic
hydrolysis of the cyclic carbonate using LiOH resulted
in the deprotection and subsequent tetrahydrofuran
formation to give 26. Compound 26 was then subjected
to Wacker oxidation conditions and 27 was obtained in
54% yield (two steps from 17). Iodination followed by
Stille coupling with tributylprenyltin18 completed the
synthesis of 8-deprenyl-garsubellin A (2).19
Scheme 4. Reagents and conditions: (a) CuCN/Et2NPh2SiLi=
1/3, THF; (b) mCPBA, KF, DMF; (c) TESOTf, 2,6-lutidine,
CH2Cl2, 33% (three steps); (d) Me2AlSEt, CH2Cl2, 89%; (e)
Et3SiH, Pd/C, CH2Cl2; (f) 0.2 M HCl, THF; (g) Al2O3,
toluene; (h) Dess–Martin periodinane, CH2Cl2, 54% (four
steps); (i) DBU, CH2Cl2, 92%; (j) 0.1 M LiOH aq., THF; (k)
Na2PdCl4, TBHP, AcOH–H2O, 54% (two steps); (l) I2, CAN,
CH3CN, 100%; (m) PdCl2(dppf), tributylprenyltin, DMF,
25% (+13% from second cycle starting with recovered 27
(50%)).
In summary, we achieved synthesis of 8-deprenyl-gar-
subellin A (2), a model compound that contains proper
stereocenters. Key findings in the present study are as
follows. (1) The stereochemical course of the key one-
pot lactone formation was changed using a cyclic car-
bonate as a protecting group of the diol. (2) Mild
Tamao oxidation at the lactone stage was essential to
avoid b-elimination of the resulting hydroxyl group. (3)

3624
H. Usuda et al. / Tetrahedron Letters 43 (2002) 3621–3624
The b-hydroxyl group could be used to facilitate the
formal migration of the lactone to the bicyclo[3.3.1]
system and to re-generate the enone at the appropriate
stage. Based on these studies, efforts toward a total
synthesis of garsubellin A are currently in progress.
reactive for a conjugate addition to 10. The kinetically
formed trans isomers 11c and 12c appeared to contain a
conformation in which the C-6 proton existed in the
equatorial position, based on 1H NMR chemical shifts
[cis (di-equatorial) isomers (28 and 29): 3.14 and 3.20
ppm; trans isomers: 3.51 (11c) and 3.49 ppm (12c)].
Acknowledgements
Financial support was provided by RFTF of Japan
Society for the Promotion of Science and PRESTO of
Japan Science and Technology Corporation (JST).
i
8. 11c and 12c should contain axial PrCO groups to avoid
the repulsion between the solvated lithium atom of the
enolate.
References
9. Jones, G. R.; Landais, Y. Tetrahedron 1996, 52, 7599–
7622.
10. Krohn, K.; Khanbabaee, K. Angew. Chem., Int. Ed. Engl.
1. Fukuyama, Y.; Kuwayama, A.; Minami, H. Chem.
Pharm. Bull. 1997, 45, 947–949.
2. Usuda, H.; Kanai, M.; Shibasaki, M. Org. Lett. 2002, 4,
859–862.
3. (a) Nicolaou, K. C.; Pfefferkorn, J. A.; Kim, S.; Wei, H.
X. J. Am. Chem. Soc. 1999, 121, 4724–4725; (b) Nico-
laou, K. C.; Pfefferkorn, J. A.; Cao, G.-Q.; Kim, S.;
Kessabi, J. Org. Lett. 1999, 1, 807–810.
4. Deprotection of the acetonide was only possible at the
enone stage (acetonide-protected 18-epi-6). The free 18-
OH form (18-epi-6), however, could not be isolated due
to the rapid intramolecular conjugate addition to the
enone.
1994, 33, 99–100.
11. The formation of the bicyclo[3.3.1] system in the presence
of this silyl group was impossible due to the instability of
the silicon.
12. (a) Tamao, K.; Kawachi, A.; Ito, Y. J. Am. Chem. Soc.
1992, 114, 3989–3990; (b) Barret, A. G. M.; Head, J.;
Smith, M. L.; Stick, N. S.; White, A. L.; Williams, D. J.
J. Org. Chem. 1999, 64, 6005–6018.
13. Fleming, Y. In Organocopper Reagents; Taylor, R. J. K.,
Ed.; Oxford University Press, 1994.
14. Using Et2NPh2SiLi as a nucleophile did not produce any
product, and (Et2NPh2Si)2CuLi·LiCN gave unpredictable
results.
5. For example, 11c and 12c were prepared as follows:
15. Fukuyama, T.; Lin, S.-C.; Li, L. J. Am. Chem. Soc. 1990,
112, 7050–7051.
16. In the previous synthesis of 18-epi form, formation of the
lactol corresponding to 23 was not significant.
17. b-Elimination of TESOH did not occur during these
transformations. Previously, the formation of the bicy-
clo[3.3.1] system was performed using K2CO3 as a base.
In this case, however, these conditions gave a complex
mixture of products.
18. Undesired reduction of the vinyl iodide occurred in this
step and 27 was obtained in 50% yield, which was then
recycled to give an additional 13% of 2.
6. The relative configurations of lactones 8c and 13c were
determined by converting to 27 (and the epimer) and
comparing the NMR data with the compound obtained
in the previous studies. Previously, the relative configura-
tion of the compound was unequivocally determined
based on X-ray crystallography.
7. To produce the desired product 8c as the major isomer
from 12c, we planned to first epimerize C-6 to di-equator-
ial isomers 28 and 29. Although epimerization was suc-
cessful using KOtBu in tBuOH at 30°C, the one-pot
lactone formation did not proceed from the di-equatorial-
isomers. A possible reason might be that the deprotona-
tion in the first step of the lactone formation occurred at
C-6 due to the higher acidity of the proton in the axial
position, and the resulting anion was not sufficiently
19. Data of 2: 1H NMR (500 MHz, C6D6) l 5.42 (dd, J=7.5,
7.5 Hz, 1H), 3.88 (dd, J=5.8, 11.0 Hz, 1H), 3.38 (dd,
J=7.5, 14.3 Hz, 1H), 3.24 (dd, J=7.5, 14.3 Hz, 1H), 2.62
(dd, J=11.0, 12.9 Hz, 1H), 2.25 (m, 1H), 1.73 (s, 3H),
1.61 (s, 3H), 1.51–1.57 (m, 1H), 1.34–1.38 (m, 1H), 1.37
(s, 3H), 1.34 (d, J=6.4 Hz, 3H), 1.29 (d, J=6.4 Hz, 3H),
1.26 (s, 3H), 1.17–1.23 (m, 2H), 1.00 (s, 3H), 0.79–0.83
(m, 1H), 0.80 (s, 3H); 13C NMR (125 MHz, C6D6) l
208.4, 204.5, 192.7, 172.4, 132.4, 122.0, 116.7, 90.1, 82.3,
70.3, 59.5, 43.2, 42.3, 35.9, 32.1, 30.4, 30.2, 26.3, 25.7,
24.4, 22.7, 22.1, 21.7, 20.7, 17.9; IR (neat, cm−1): 3471,
1731, 1625; EI-MS m/z 416 (M+); EI-HRMS calcd for
C25H36O5 (M+): 416.2563, found: 416.2568.