Synthetic study of azaspiracid-1: Synthesis of the EFGHI-ring fragment

Article abstract of DOI:10.1021/ol0613766

Here, we report a synthesis of the lower half C₂₁-C₄₀ fragment of the shellfish toxin, azaspiracid-1. The C₂₈-C ₄₀ fragment was synthesized by a coupling between the C ₂₈-C₃₅ epoxide and the C₃₆-C₄₀ dithioacetal anion, followed by the HI-ring spiroaminal formation. An aldehyde corresponding to the C₂₈-C₄₀ fragment was then coupled with the C₂₁-C₂₇ allylic stannane by using InCl ₃. Finally, the FG-ring was constructed by HF-pyridine to accomplish the synthesis of the suitably protected C₂₁-C₄₀ fragment.

Full text of DOI:10.1021/ol0613766

ORGANIC
LETTERS
2006
Vol. 8, No. 18
3943-3946
Synthetic Study of Azaspiracid-1:
Synthesis of the EFGHI-Ring Fragment
Masato Oikawa,* Tomoko Uehara, Taizo Iwayama, and Makoto Sasaki*
Laboratory of Biostructural Chemistry, Graduate School of Life Sciences, Tohoku
UniVersity, Tsutsumidori-amamiya, Aoba-ku, Sendai 981-8555, Japan
mao@bios.tohoku.ac.jp
Received June 6, 2006
ABSTRACT
Here, we report a synthesis of the lower half C21
a coupling between the C28 C35 epoxide and the C36
corresponding to the C28 C40 fragment was then coupled with the C21
by HF pyridine to accomplish the synthesis of the suitably protected C21
−
C40 fragment of the shellfish toxin, azaspiracid-1. The C28
−
C40 fragment was synthesized by
−
−C40 dithioacetal anion, followed by the HI-ring spiroaminal formation. An aldehyde
−
−C27 allylic stannane by using InCl3. Finally, the FG-ring was constructed
‚
−C40 fragment.
Azaspiracid-1 (1) is a causative toxin for a new type of
shellfish poisoning syndrome named azaspiracid poisoning
(AZP), which has prevailed since November 1995 at a coastal
region in Europe.1 The structure was first elucidated spec-
troscopically by a group led by Yasumoto and Satake at
Tohoku University in 1998 by using contaminated Irish
mussels, Mytilus edulis.2 The structural characteristics are
(1) a C6-C17 bisspiroketal fused to a C17-C20 tetrahydrofuran
and (2) an unusual C33-C40 azaspiro ring fused with C28-
Since the first publication on 1,2 synthetic studies have
been actively developed by many groups, and the first total
synthesis and the structural revision of 1 were made by the
Nicolaou group in 2004 by a strategy synthesizing possible
combinations of partial structures.5 Recently, the Nicolaou
group also synthesized azaspiracid-2 and -3, confirming the
revised structure for these congeners.6 We have been working
toward the total synthesis of 1 to clarify the molecular
mechanism of AZP toward humans.7 Here, we describe our
C
40 2,9-dioxabicyclo[3.3.1]nonane. Four congeners, azaspi-
(4) Ofuji, K.; Satake, M.; Oshima, Y.; Naoki, H.; Yasumoto, T. In 43rd
Symposium on the Chemistry of Natural Products; Osaka, 2001; p 353.
(5) For the whole story for the structural revision of 1, see: (a)
Freemantle, M. Chem. Eng. News 2004, 82, 27. (b) Nicolaou, K. C.;
Vyskocil, S.; Koftis, T. V.; Yamada, Y. M. A.; Ling, T. T.; Chen, D. Y.
K.; Tang, W. J.; Petrovic, G.; Frederick, M. O.; Li, Y. W.; Satake, M.
Angew. Chem., Int. Ed. 2004, 43, 4312. (c) Nicolaou, K. C.; Koftis, T. V.;
Vyskocil, S.; Petrovic, G.; Ling, T. T.; Yamada, Y. M. A.; Tang, W. J.;
Frederick, M. O. Angew. Chem., Int. Ed. 2004, 43, 4318. (d) Nicolaou, K.
C.; Pihko, P. M.; Bernal, F.; Frederick, M. O.; Qian, W. Y.; Uesaka, N.;
Diedrichs, N.; Hinrichs, J.; Koftis, T. V.; Loizidou, E.; Petrovic, G.;
Rodriquez, M.; Sarlah, D.; Zou, N. J. Am. Chem. Soc. 2006, 128, 2244. (e)
Nicolaou, K. C.; Chen, D. Y. K.; Li, Y. W.; Uesaka, N.; Petrovic, G.; Koftis,
T. V.; Bernal, F.; Frederick, M. O.; Govindasamy, M.; Ling, T. T.; Pihko,
P. M.; Tang, W. J.; Vyskocil, S. J. Am. Chem. Soc. 2006, 128, 2258. (f)
Nicolaou, K. C.; Koftis, T. V.; Vyskocil, S.; Petrovic, G.; Tang, W. J.;
Frederick, M. O.; Chen, D. Y. K.; Li, Y. W.; Ling, T. T.; Yamada, Y. M.
A. J. Am. Chem. Soc. 2006, 128, 2859.
racids-2-5, were found by the same group thereafter,3 and
the lethality against mice (LD50) was found to be nearly
comparable to that of 1 (0.2 mg/kg) for azaspiracids-2-4
(0.11-0.47 mg/kg) and less toxic for azaspiracid-5 (>1 mg/
kg).4
(1) (a) McMahon, T.; Silke, J. Harmful Algae News 1996, 14, 2. (b)
James, K. J.; Furey, A.; Lehane, M.; Ramstad, H.; Aune, T.; Hovgaard, P.;
Morris, S.; Higman, W.; Satake, M.; Yasumoto, T. Toxicon 2002, 40, 909.
(2) Satake, M.; Ofuji, K.; Naoki, H.; James, K. J.; Furey, A.; McMahon,
T.; Silke, J.; Yasumoto, T. J. Am. Chem. Soc. 1998, 120, 9967.
(3) (a) Ofuji, K.; Satake, M.; McMahon, T.; Silke, J.; James, K. J.; Naoki,
H.; Oshima, Y.; Yasumoto, T. Nat. Toxins 1999, 7, 99. (b) Ofuji, K.; Satake,
M.; McMahon, T.; James, K. J.; Naoki, H.; Oshima, Y.; Yasumoto, T.
Biosci. Biotechnol. Biochem. 2001, 65, 740.
10.1021/ol0613766 CCC: $33.50
© 2006 American Chemical Society
Published on Web 08/01/2006

efforts toward this goal and the synthesis of the C21-C40
EFGHI-ring fragment 2, the lower half of 1.8
Scheme 2. Preparation of the E-Ring Allylic Stannane 3
Our synthetic plan toward azaspiracid-1 (1) is shown in
Scheme 1. Retrosynthetic disconnection at the C20-C21 bond
Scheme 1. Synthetic Strategy for Azaspiracid-1 (1)
generates two C1-C20 and C21-C40 (2) fragments. These
fragments would be assembled by the coupling between a
benzotriazole amide and a sulfonylpyran, also employed in
the synthesis of altohyrtin C by Evans et al.9 The C21-C40
EFGHI-ring fragment 2 was further divided retrosynthetically
into two fragments, corresponding to the C21-C27 E-ring (3)
and the C28-C40 FGHI-ring (4) domains, which were to be
coupled by the reaction between an aldehyde and an allylic
stannane under mild conditions.
The synthesis of the E-ring allylic stannane 3 started with
the known optically active acetoxyalcohol 5, which can be
readily prepared from methylmalonic acid ester over six steps
via the key enzymatic desymmetrization of meso-2,4-
dimethyl-1,5-pentanediol (Scheme 2).10 At first, the hydroxy
group was oxidized with IBX11 and the aldehyde was treated
with thiophenol and BF3‚OEt2 to give the dithioacetal 6 in
83% yield. Removal of the acetyl group followed by
oxidation provided aldehyde 7, which in turn was reacted
with the vinylic anion, generated from 2-bromopropen-1-
ol12 and t-BuLi. The reaction proceeded quite smoothly at
-78 °C to give diol 8 in 76% yield as a chromatographically
inseparable, diastereomeric mixture (R/S ) 1.2:1). We then
enriched the desired (25S)-isomer. Thus, the four-step
transformation, including (1) the temporary protection of the
primary hydroxy group, (2) oxidation of the remaining
hydroxy group by IBX, (3) stereoselective reduction of the
ketone by using DIBAL-H, and (4) the removal of the
TBDPS protecting group by TBAF, allowed enrichment of
the diastereomeric ratio of 8 to R/S ) 1:5.0.13 To construct
the E-ring tetrahydropyran, 8 was next treated with BF3‚OEt2
in CH2Cl2. The (25S)-isomer preferentially reacted to give
9 in 71% yield as a diastereomeric mixture at the acetal
carbon center (21R/21S ) 1:2.4), and the dithioacetal with
unnatural (R) stereochemistry at the C25 position was
recovered and recycled for the same transformation again.
Careful chromatographic separation of the alcohol 9 gave
the pure (21S)-isomer, which was finally converted to the
allylic stannane (21S)-3, the E-ring fragment, by bromination
followed by stannylation (Bu3SnLi, CuBr).14
The synthesis of the FGHI-ring fragment, corresponding
to the C28-C40 position, was next explored. The stereose-
lective construction of the unique azaspiro HI-ring is the
major concern in this synthesis. Our first-generation synthesis
employed the Boc protecting group for the C40-amino
functionality.7 However, a partial decomposition of the Boc
group was observed under the conditions for the HI-ring
formation using Yb(OTf)3. Here, we decided to explore an
(6) Nicolaou, K. C.; Frederick, M. O.; Petrovic, G.; Cole, K. P.; Loizidou,
E. Z. Angew. Chem., Int. Ed. 2006, 45, 2609.
(7) Sasaki, M.; Iwamuro, Y.; Nemoto, J.; Oikawa, M. Tetrahedron Lett.
2003, 44, 6199.
(8) For synthetic studies on the E-, F-, G-, H-, and/or I-ring domain of
1, see: (a) Carter, R. G.; Weldon, D. J. Org. Lett. 2000, 2, 3913. (b)
Aiguade, J.; Hao, J. L.; Forsyth, C. J. Org. Lett. 2001, 3, 979. (c) Aiguade,
J.; Hao, J. L.; Forsyth, C. J. Tetrahedron Lett. 2001, 42, 817. (d) Forsyth,
C. J.; Hao, J. L.; Aiguade, J. Angew. Chem., Int. Ed. 2001, 40, 3663. (e)
Zhou, X. T.; Carter, R. G. Chem. Commun. 2004, 2138. (f) Zhou, X. T.;
Carter, R. G. Angew. Chem., Int. Ed. 2006, 45, 1787. (g) Nguyen, S.; Xu,
J.; Forsyth, C. J. Tetrahedron 2006, 62, 5338.
(9) Evans, D. A.; Trotter, B. W.; Coleman, P. J.; Cote, B.; Dias, L. C.;
Rajapakse, H. A.; Tyler, A. N. Tetrahedron 1999, 55, 8671.
(10) (a) Lin, G. Q.; Xu, W. C. Tetrahedron 1996, 52, 5907. (b) Fujita,
K.; Mori, K. Eur. J. Org. Chem. 2001, 493.
(11) Frigerio, M.; Santagostino, M. Tetrahedron Lett. 1994, 35, 8019.
(12) Li, K. Q.; Du, W. S.; Que, N. L. S.; Liu, H. W. J. Am. Chem. Soc.
1996, 118, 8763.
(13) The stereochemistry of these products was determined by the
modified Mosher ester analysis of undesired (R)-8, which was obtained by
recrystallization.
(14) (a) Piers, E.; Chong, J. M.; Gustafson, K.; Andersen, R. J. Can. J.
Chem. 1984, 62, 1. (b) Williams, D. R.; Plummer, S. V.; Patnaik, S. Angew.
Chem., Int. Ed. 2003, 42, 3934.
3944
Org. Lett., Vol. 8, No. 18, 2006

alternative, stable protecting group, which also supports the
stereoselective cyclization of the azaspiro ring.
The 2-nitrobenzensulfonyl (nosyl, Ns) group15 was at first
employed because it is electron withdrawing and hence is
expected to promote the spiroaminal formation smoothly
(Scheme 3). However, the reaction on 1016 by using Yb-
Scheme 4. Preparation of the C28-C40 FGHI-Ring Domain 4
with Correct Spiroaminal Stereochemistrya
Scheme 3. Spiroaminal Formation Leading to an Unnatural
Isomera
a Dashed arrows indicate the NOESY correlations observed.
(OTf)3 was found to be very slow at room temperature, and
the stereochemistry of the product 11, obtained in 25% as
the sole product, was inconsistent with that of the natural
azaspiracid-1 (1). That is, from the NOEs observed at H35-
(pro-S)/C37-Me and H35(pro-S)/H38ax, the stereochemistry
at C36 was determined to be R, which is not desired. When
BF3‚OEt2 was used for the Lewis acid, the cyclization
proceeded quite smoothly in 89% yield, but the stereoselec-
tivity was not improved.17
We next examined the alkylcarbamate protecting group,
which is more stable than the Boc group previously used
for the C40-amino functionality.7 The alcohol 127 was at
first protected, and a 2-nosylamino group was introduced to
the C40 position by desilylation with TBAF and the Mit-
sunobu reaction,18 to afford 13 in 78% overall yield (Scheme
4). After the amide was allyloxycarbonylated (Alloc), the
Ns group was removed by the Fukuyama protocol using
2-mercaptoethanol to give 14. Then, the H-ring moiety was
constructed by successive deprotection of the MOM19 and
the dithiane groups,20 followed by the furanose formation
in MeOH to give 15 in 28% yield (47% based on recovered
materials) for three steps. At this time, the Alloc protecting
group was hydrogenated with a less acidic catalyst21 to
provide the propyl carbamate 16 in 91% yield, ready for the
next crucial cyclization. The reduction was necessary for the
selective removal of the benzyl groups at the C28, C32, and
C34 positions later.22 With the cyclization precursor 16 in
hand, Yb(OTf)3-catalyzed spiroaminal formation8d was at-
a Dashed arrow indicates the NOESY correlations observed.
tempted in CH2Cl2 at 0 °C. As expected, the cyclization
proceeded smoothly in 1 h, to give spiroaminal 17 in 78%
yield. The cyclization was highly stereoselective, and the
stereochemistry at the spirocenter was unambiguously de-
termined to be identical to that of the natural 1 by the NOE
at H35(pro-R)/C37-Me. The spiroaminal 17 was further
converted to alcohol 18 by protecting group manipulations,
which was then oxidized by tetrapropylammonium perruth-
enate (TPAP)23 to give aldehyde 4 for the coupling with the
E-ring fragment 3 in 71% yield over four steps.
With the key fragments 3 and 4 in hand, their coupling
reactions leading to the desired EFGHI-ring domain were
next explored. In our first-generation synthesis, where simple
(15) (a) Fukuyama, T.; Jow, C. K.; Cheung, M. Tetrahedron Lett. 1995,
36, 6373. (b) Kan, T.; Fukuyama, T. Chem. Commun. 2004, 353.
(16) Prepared from 12 in 54% yield over five steps.
(17) The stereoselectivity observed in the spiroaminal formation was
unexpected but was reasonably explained by calculation at the MMFF94S
force field. The details will be reported in the full account of this work.
(18) Henry, J. R.; Marcin, L. R.; McIntosh, M. C.; Scola, P. M.; Harris,
G. D.; Weinreb, S. M. Tetrahedron Lett. 1989, 30, 5709.
(19) Hanessian, S.; Delorme, D.; Dufresne, Y. Tetrahedron Lett. 1984,
25, 2515.
(22) An alkyl carbamate protecting group can be generally removed with
various nucleophiles; see: (a) Tius, M. A.; Kerr, M. A. J. Am. Chem. Soc.
1992, 114, 5959. (b) Wenkert, E.; Hudlicky, T.; Showalter, H. D. H. J.
Am. Chem. Soc. 1978, 100, 4893. (c) Corey, E. J.; Weigel, L. O.; Floyd,
D.; Bock, M. G. J. Am. Chem. Soc. 1978, 100, 2916.
(20) Stork, G.; Zhao, K. Tetrahedron Lett. 1989, 30, 287.
(21) Sajiki, H.; Ikawa, T.; Hirota, K. Tetrahedron Lett. 2003, 44, 7407.
(23) Griffith, W. P.; Ley, S. V.; Whitcombe, G. P.; White, A. D. J. Chem.
Soc. Chem. Commun. 1987, 1625.
Org. Lett., Vol. 8, No. 18, 2006
3945

methallyl stannane was used as a preliminary study, MgBr2‚
OEt2 was employed for this coupling.7 However, it was found
from another preliminary model study using more complex
allylic stannane 1924 that neither MgBr2‚OEt2 nor LiClO4
worked for this coupling reaction, only recovering the
substrates (Table 1, runs 1 and 2). More reactive BF3‚OEt2
fragments 3 (5.6 equiv) and 4 (1.0 equiv) also proceeded
smoothly at room temperature to give 21, which has the C21-
C40 entire chain of azaspiracid (1) with complete function-
alities, in 88% yield (Scheme 5). After mild oxidation with
Scheme 5. Synthesis of the EFGHI-Ring Fragment of
Azaspiracid-1 (1)
Table 1. Preliminary Model Studies for the Coupling of 4 with
Allylic Stannane 19
IBX, the FG-ring was stereoselectively constructed upon
exposure to HF‚pyridine, which induced desilylation and
intramolecular acetalization, to give the desired EFGHI-ring
fragment 2 of azaspiracid-1 (1) in 26% yield. The stereo-
chemistry was unambiguously confirmed by NOESY experi-
ments (see the Supporting Information).
In summary, we have successfully synthesized the suitably
protected, lower half C21-C40 fragment 2 of azaspiracid-1
(1) in 0.025% yield for the longest linear pathway from
D-glutamic acid (37 steps). The synthesis features (1)
stereoselective construction of the HI-ring spiroaminal and
(2) high-yield coupling of the C21-C27 and C28-C40 frag-
ments by allylic stannane chemistry. Work is in progress
toward the total synthesis of 1 in our laboratory.
yield for 20 (%)/
recovery of 4 (%)
run
conditions
1
2
3
4
5
MgBr2‚OEt2 (4.0 equiv), CH2Cl2, -20 °C
LiClO4 (3.0 equiv), Et2O, rt
BF3‚OEt2 (2.0 equiv), MS4A, CH2Cl2, -78 °C
InCl3 (3.0 equiv), acetone, rt
0/56
0/95
0/<30
0/88
100/-
InCl3 (3.0 equiv), THF, rt
induced decomposition even at -78 °C (run 3). From the
1H NMR spectra of the product mixture, the decomposition
was supposed to occur at the HI-ring azaspiro moiety. To
avoid the decomposition, we investigated another Lewis acid,
which would hopefully work in a Lewis basic solvent. It
was then found that, when InCl3 was used in acetone, the
allylic stannane 19 reacted with acetone cleanly (run 4).25
We therefore explored other solvents, and THF was finally
found to be satisfactory to give the desired product 20 in a
quantitative yield (run 5). With this protocol using InCl3 (3.0
equiv) in THF, the coupling reaction between the key
Acknowledgment. We thank Mr. Kitago of our labora-
tory for assistance with the experiment. Financial support
was provided by the Novartis Foundation (Japan) for the
Promotion of Science and by the Suntory Institute for
Bioorganic Research. We also acknowledge a Grant-in-Aid
for Scientific Research from the Ministry of Education,
Science, Sports, Culture and Technology, Japan.
Supporting Information Available: Experimental details
and spectroscopic data for all new compounds. This material
is available free of charge via the Internet at http://pubs.acs.org.
(24) The allylic stannane 19 was prepared from 2-bromo-3-trimethylsi-
lylpropene in 37% yield over nine steps. See the Supporting Information.
The details will be given in the full account of this work.
(25) Marshall, J. A.; Hinkle, K. W. J. Org. Chem. 1995, 60, 1920.
OL0613766
3946
Org. Lett., Vol. 8, No. 18, 2006