Total synthesis of serofendic acids A and B employing tin-free homoallyl-homoallyl radical rearrangement

Article abstract of DOI:10.1021/ol051411t

(Chemical Equation Presented) Total syntheses of serofendic acids A (1a) and B (1b) are described. The key strategic element of the approach involves the novel tin-free homoallyl-homoallyl radical rearrangement of 5 for the construction of bicyclo[2.2.2]o

Full text of DOI:10.1021/ol051411t

ORGANIC
LETTERS
2005
Vol. 7, No. 18
3929-3932
Total Synthesis of Serofendic Acids
A and B Employing Tin-Free
Homoallyl−Homoallyl Radical
Rearrangement
Masahiro Toyota,*,† Takeshi Asano, and Masataka Ihara*
Department of Organic Chemistry, Graduate School of Pharmaceutical Sciences,
Tohoku UniVersity, Aobayama, Sendai 980-8578, Japan
toyota@c.s.osakafu-u.ac.jp; mihara@mail.pharm.tohoku.ac.jp
Received June 17, 2005
ABSTRACT
Total syntheses of serofendic acids A (1a) and B (1b) are described. The key strategic element of the approach involves the novel tin-free
homoallyl homoallyl radical rearrangement of 5 for the construction of bicyclo[2.2.2]octane ring system 4. In addition, the conversion of
methyl atisirenoate 2 to serofendic acids A (1a) and B (1b) was achieved on the basis of the Michael reaction of sodium thiomethoxide.
−
Serofendic acids A (1a) and B (1b) are neuroprotective
agents that were isolated from fetal calf serum in 2002 by
Akaike, Sugimoto, and co-workers.1 The unique structures
of 1a and 1b, determined by NMR analysis, were shown to
consist of bicyclo[2.2.2]octane ring systems bearing an
unusual sulfoxide side chain (Figure 1). The serofendic acids
are attractive targets for total synthesis, not only due to their
neuroprotective activity, but also because of the challenge
of their structures, which feature nine stereocenters and an
atisane framework. Prior to this report, the Eisai group’s
syntheses of 1a and 1b, starting from (-)-isosteviol, stood
as the only published synthetic work in this field.1
In a prior report, we demonstrated the utility of the
homoallyl-homoallyl radical rearrangement for construction
of the bicyclo[2.2.2]octane ring system, the CD ring part of
methyl atisirenoate 2.2 In an effort to further expand the
utility of this rearrangement, we have focused on the
development of tin-free conditions3 for promoting this useful
reaction. In this paper, we describe the total syntheses of
the serofendic acids (1), wherein the key step involves the
† Present address: Department of Chemistry, Graduate School of Science,
Osaka Prefecture University.
(1) (a) Kume, T.; Asai, N.; Nishikawa, H.; Mano, N.; Terauchi, T.;
Taguchi, R.; Shirakawa, H.; Osakada, F.; Mori, H.; Asakawa, N.; Yonaga,
M.; Nishizawa, Y.; Sugimoto, H.; Shimohama, S.; Katsuki, H.; Kaneko,
S.; Asaike, A. Proc. Nat. Acad. Sci. U.S.A. 2002, 99, 3288-3293. (b)
Terauchi, T.; Asai, N.; Yonaga, M.; Kume, T.; Akaike, A.; Sugimoto, H.
Tetrahedron Lett. 2002, 43, 3625-3628.
(2) (a) Toyota, M.; Yokota, M.; Ihara, M. J. Am. Chem. Soc. 2001, 123,
1856-1861. (b) Toyota, M.; Yokota, M.; Ihara, M. Org. Lett. 1999, 1,
1627-1629. (c) Toyota, M.; Yokota, M, Ihara, M. Tetrahedron Lett. 1999,
40, 1551-1554. (d) Toyota, M.; Wada, T.; Fukumoto, K.; Ihara, M. J. Am.
Chem. Soc. 1998, 120, 4916-4924.
Figure 1. Natural products possessing neuroprotective property.
10.1021/ol051411t CCC: $30.25
© 2005 American Chemical Society
Published on Web 08/02/2005

successful implementation of a tin-free homoallyl-homoallyl
radical rearrangement for construction of the bicyclo[2.2.2]-
octane ring system.
jugate addition of an isopropenyl group to 9 proceeded with
excellent diastereoselectivity and yielded the desired ketone
5 as the sole stereoisomer (Scheme 2).
Our plan for the preparation of serofendic acids A and B
(1) is based on three strategic disconnections: an intra-
molecular Diels-Alder reaction (3 f 2) to synthesize the
AB ring part of 1, a tin-free homoallyl-homoallyl radical
rearrangement (5 f 4) to generate the bicyclo[2.2.2]octane
ring system, and a palladium-catalyzed cycloalkenylation of
6, the absolute stereochemistry of which arises from an
asymmetric ene reaction4 to construct the bicyclo[3.2.1]-
octane moiety (Scheme 1).
With ketone 5 in hand, the crucial homoallyl-homoallyl
radical rearrangement was attempted under a variety of
conditions, a few of which are listed in Table 1. Among the
Table 1. Homoallyl-Homoallyl Radical Rearrangement of
Ketone 5 Assembling Bicyclo[2.2.2]octane 4
Scheme 1. Retrosynthesis of Serofendic Acids (1)
yield
entry
1
conditions
(1) NaBH4, (2) (imid)2CdS,
(%)
82a
(3) Bu3SnH, AIBN, toluene, reflux
2
3
(1) NaBH4, (2) PPh3, DEAD, NBSHb, THF
TsNHNH2, THF, reflux; NaBH3CN, ZnCl2, reflux 75
0
a Three-step yield. b o-Nitrobenzenesulfonylhydrazine.
The key transformation in our plan was the novel tin-free
homoallyl-homoallyl radical rearrangement; therefore, so
its feasibility was examined first. The requisite ketone 5 was
synthesized as shown in Scheme 2. Namely, regioselective
allylic oxidation of cyclohexene 75 was achieved by applying
Ishii’s protocol6 to give enone 8, which was transformed into
the bicyclo[3.2.1]octane 9 by means of silyl enolate formation
followed by palladium-catalyzed cycloalkenylation.7 Con-
traditional methods examined for triggering the homoallyl-
homoallyl radical rearrangement, the best result was obtained
by employing the Barton-McCombie deoxygenation pro-
tocol.8 Ketone 5 was reduced, and the resulting alcohol was
converted to the corresponding thioimidazolide. Subjection
of the thioimidazolide to standard tin hydride conditions
promoted the desired rearrangement and afforded the ther-
modynamically stable bicyclo[2.2.2]octane product in 82%
yield (entry 1). While this method proved effective, it
required the use of potentially toxic organotin compounds.
In an effort to avoid the use of such reagents, we examined
procedures for radical generation under tin-free conditions.
Myers has described a deoxygenation protocol9 that likely
Scheme 2. Preparation of Ketone 5
(3) A recent reView: Baguley, P. A.; Walton, J. C. Angew. Chem., Int.
Ed. 1998, 37, 3072-3082.
(4) A review: d’Angelo, J. D.; Desmaele, D.; Dumas, F.; Guingant, A.
Tetrahedron: Asymmetry 1992, 3, 459-505.
(5) Toyota, M.; Wada, T.; Ihara, M. J. Org. Chem. 2000, 65, 4565-
4570.
(6) Ishii, Y.; Nakayama, K.; Takeno, M.; Sakaguchi, S.; Iwahara, T.;
Nishiyama, Y. J. Org. Chem. 1995, 60, 3934-3935.
(7) Recent examples: (a) Toyota, M.; Sasaki, M.; Ihara. Org. Lett. 2003,
5, 1193-1195. (b) Toyota, M.; Odashima, T.; Wada, T.; Ihara, M. J. Am.
Chem. Soc. 2000, 122, 9036-9037. (c) A recent review: Toyota, M.; Ihara,
M. Synlett 2002, 1211-1222.
(8) Barton, D. H. R.; McCombie, S. W. J. Chem. Soc., Perkin Trans. 1
1975, 1574-1585.
3930
Org. Lett., Vol. 7, No. 18, 2005

proceeds through a radical intermediate. Disappointingly,
when ketone 5 was subjected to the Myers procedure (entry
2), it produced none of the expected rearrangement product.
On the other hand, the application of Taber’s technique10
for radical generation under tin-free conditions proved
successful for the desired rearrangement. Treatment of the
tosylhydrazone derived from 5 with NaBH3CN in the
presence of ZnCl2 effected homoallyl-homoallyl radical
rearrangement to produce the sought-after bicyclo[2.2.2]-
octane compound 4 in 75% yield (entry 3). It is worth noting
that whereas the mechanism of radical generation by the
Taber protocol has been investigated, it appears to have been
little utilized for natural product synthesis.11
transformation. Intramolecular Diels-Alder reaction of 3 was
conducted at 200 °C in a stainless steel autoclave to yield
the desired tetracyclic compound 14 (84%). Regioselective
reduction of 14 followed by stereoselective methylation gave
rise to (-)-methyl atisirenoate 2 as a single stereoisomer
(Scheme 3).
The complete stereoselectivity observed in the Diels-
Alder reaction of 3 can be rationalized by considering the
two likely pre-transition state conformations shown in Figure
2. The isopropenyl unit is expected to be oriented so as to
The conversion of diene 4 to tetraene 3, required for the
intramolecular Diels-Alder reaction, was achieved as de-
picted in Scheme 3. After reductive deprotection of pivaloyl
Scheme 3. Stereoselective Synthesis of (-)-Methyl
Atisirenoate 2
Figure 2. Plausible conformers 3A and 3B for intramolecular
Diels-Alder reaction.
minimize 1,3-allylic strain15 between the olefinic hydrogen
and the angular hydrogen as shown in 3A and 3B. While
conformer 3B is destabilized by nonbonding interactions
between the axial hydrogen and the ester group on the diene
Scheme 4. Transformation of (-)-Methyl Atisirenoate 2 into
(-)-Serofendic Acids A (1a) and B (1b)
group of 4, Parikh oxidation12 of the resulting alcohol gave
aldehyde 12, which was subjected to Still olefination13 to
afford E-olefin 13 as a single isomer. Suzuki cross-coupling
reaction of 13 was next accomplished under Molander’s
conditions14 to furnish tetraene 3, poised for the second key
(9) Myers, A. G.; Movassaghi, M.; Zheng, B. J. Am. Chem. Soc. 1997,
119, 8572-8573.
(10) Taber, D. F.; Wang, Y.; Stachel, S. J. Tetrahedron Lett. 1993, 34,
6209-6210.
(11) Miranda, L. D.; Zard, S. Z. Chem. Commun. 2001, 1068-1069.
(12) Parikh, J. R.; Doering, W. E. J. Am. Chem. Soc. 1967, 89, 5505-
5507.
(13) Still, W. C.; Gennari, C. Tetrahedron Lett. 1983, 24, 4405-4408.
Org. Lett., Vol. 7, No. 18, 2005
3931

part, the above interaction is absent in the conformer 3A,
which leads to the desired cycloadduct 14 (Figure 2).
The requisite side chain on the D ring system of 1 was
introduced as shown in Scheme 4. Allylic oxidation of 2
with benzeneseleninic anhydride produced enone 15 (80%).
The required thiomethyl group was introduced directly
through Michael addition of sodium thiomethoxide, which
gave a 1:1 mixture of the R- and â-thiomethyl ketones. Direct
reduction of this mixture of ketones afforded the four
separable thiomethyl alcohols shown, from which the desired
product, 16c, was isolated in 27% overall yield. Attempts to
improve the diastereoselectivity of this addition-reduction
sequence proved fruitless.
properties and specific rotations of both synthetic materials
were identical with those reported for the naturally occurring
compounds.1
In conclusion, a total synthesis of both (-)-serofendic acids
A (1a) and B (1b) has been accomplished. The synthetic
strategy involved the use of a tin-free homoallyl-homoallyl
radical rearrangement for the construction of bicyclo[2.2.2]-
octane ring system. The methodology developed here should
also provide access to synthetic analogues of serofendic acids.
Acknowledgment. This work is supported by a Grant-
in-Aid (No. 16390001) from the Ministry of Education,
Science, Sports and Culture, Japan. We thank Dr. Masahiro
Yonaga of Eisai Co., Ltd. for the generous supply of valuable
data about serofendic acids and kind discussion.
Finally, hydrolysis16 of 16c followed by Davis oxidation17
of the resulting sulfide provided serofendic acids A (1a) and
B (1b) as a 1:2 separable mixture. The spectroscopic
Supporting Information Available: Experimental condi-
tions and spectral data for compounds reported in this paper.
This material is available free of charge via the Internet at
http://pubs.acs.org.
(14) Molander, G. A.; Rivero, M. R. Org. Lett. 2002, 4, 107-109.
(15) (a) Hoffman, R. W. Chem. ReV. 1989, 89, 1841-1860. (b) Johnson,
F. Chem. ReV. 1968, 68, 375-413.
(16) Bartlett, P. A.; Johnson, W. S. Tetrahedron Lett. 1970, 11, 4459-
4462.
(17) Davis, F. A.; Stringer, O. D. J. Org. Chem. 1982, 47, 1774-1775.
OL051411T
3932
Org. Lett., Vol. 7, No. 18, 2005