Total synthesis of (-)-stellettamide B and determination of its absolute stereochemistry.

Article abstract of DOI:10.1021/ol0003228

[figure: see text] The first total synthesis of (-)-stellettamide B has been achieved by a sequence based on amide coupling of the chiral 1-(aminomethyl)-indolizidine fragment, prepared by TiCl4-mediated asymmetric allylation of the tricyclic N-acyl-N,O-a

Full text of DOI:10.1021/ol0003228

ORGANIC
LETTERS
2001
Vol. 3, No. 2
193-196
Total Synthesis of (−)-Stellettamide B
and Determination of Its Absolute
Stereochemistry
Naoki Yamazaki, Wataru Dokoshi, and Chihiro Kibayashi*
School of Pharmacy, Tokyo UniVersity of Pharmacy and Life Science,
1432-1 Horinouchi, Hachioji, Tokyo 192-0392, Japan
kibayasi@ps.toyaku.ac.jp
Received October 23, 2000
ABSTRACT
The first total synthesis of (−)-stellettamide B has been achieved by a sequence based on amide coupling of the chiral 1-(aminomethyl)-
indolizidine fragment, prepared by TiCl4-mediated asymmetric allylation of the tricyclic N-acyl-N,O-acetal, with the chiral trienoic acid fragment
This synthesis led to revision of the published relative stereochemistry of the natural product and established its absolute stereochemistry
to be 1S,4S,8aR,6′′R.
Stellettamide A (1), which showed antifungal and cytotoxic
activity, was first isolated by Fusetani et al.1 in 1990 from a
marine sponge of the genus Stelletta collected in Shikine-
jima Island of Japan as the first representative of an
indolizidine class of marine alkaloids. Subsequently, Shin
et al.2 reported the isolation of the closely related alkaloid
stellettamide B from a Korean specimen of Stelletta sp.,
which was found to have antifungal and RNA-cleaving
activities. On the basis of spectroscopic data, the structure
of stellettamide B was formulated as 2 having the 6′′S
absolute configuration assigned by chemical correlation (see
below).3 Very recently, another antibacterial alkaloid, stel-
lettamide C (3), was isolated from the Japanese sponge
Stelletta sp.4
(1) Hirota, H.; Matsunaga, S.; Fusetani, N. Tetrahedron Lett. 1990, 31,
4163-4164.
(2) Shin, J.; Seo, Y.; Cho, K. W.; Rho, J.-R.; Sim, C. J. J. Nat. Prod.
1997, 60, 611-613.
(3) The absolute configuration of the C(6′′) chiral center in the structure
of stellettamide B reported in the literature (ref 2) is erroneously depicted
as R and should therefore be corrected to S (J. Shin, personal communica-
tion).
Stellettamides A-C comprise a common indolizidine
(4) Matsunaga, S.; Yamashita, T.; Tsukamoto, S.; Fusetani, N. J. Nat.
Prod. 1999, 62, 1202-1204.
skeleton, containing a quaternary nitrogen atom and a
10.1021/ol0003228 CCC: $20.00 © 2001 American Chemical Society
Published on Web 01/05/2001

variation of the unsaturated side chain, which are connected
via an amide linkage. The relative stereochemistry originally
proposed for the indolizidine part of stellettamide A and its
absolute stereochemistry have been established by the
synthesis5 of the antipodal stellettamide A as depicted in 1.
Moreover, the direct comparison of the degradation products
obtained from stellettamides A and C confirmed that the
indolizidine part of stellettamide C has the same absolute
configuration, shown in 3, as that of stellettamide A.4 On
the other hand, there is still no evidence for the absolute
configuration of the indolizidine unit of stellettamide B,
though the same relative stereochemistry as that for stellet-
tamides A and C has been assigned to it on the basis of NMR
study.2 The absolute stereochemistry of the asymmetric center
at C(6′′) in the norsesquiterpene side chain was, however,
assigned as S by chemical correlation of the degradation
product of stellettamide B to the known (S)-2-methylglutaric
acid.2
acid-mediated asymmetric allylation of cyclic N-acyl-N,O-
acetals, in which enantiomeric 2-(1-aminoethyl)phenol (9)
has been proven useful as a chiral auxiliary.6 Accordingly,
our initial study was application of this methodology to the
synthesis of 8. Thus, the glutarimide 10 incorporating (R)-9
as an auxiliary was converted to the tricyclic N,O-acetal 11
as a single diastereomer by partial reduction with Vitride
(toluene, -78 °C) followed by acid treatment. Upon treat-
ment of 11 with allyltrimethylsilane (3 equiv) and TiCl4
(3 equiv) at 50 °C, the reaction proceeded smoothly to give
allylated products in 98% yield and a ratio of 6.0:1 favoring
the desired (6R)-isomer 12. The stereochemical assignment
of the newly generated allyl-bearing stereogenic center in
12 by NMR analysis was actually difficult, and since 12 was
formed as an oil, we were unable to grow suitable crystals
for X-ray crystallography. However, racemic 12, prepared
following the same reaction sequence used in Scheme 2 via
In this paper, we disclose an enantioselective synthesis of
the structure 2 originally proposed for stellettamide B by a
strategy involving an asymmetric allylation of a cyclic
N-acyl-N,O-acetal as a key feature. We also report that
structure 2 does not correspond to natural stellettamide B,
while its epimer 4 at the C(6′′) position is identical to the
natural product, leading to revision of the original structure
2. Herein, we describe the first total synthesis of stellettamide
B, as its natural (-)-enantiomer, that establishes the relative
and absolute stereochemistry of this alkaloid as 4.
Scheme 2a
In our strategy, for the synthesis of the initial target 2, we
sought to utilize a simple and straightforward approach that
involves connecting the chiral aminomethylindolizidine frag-
ment 5 to the (S)-trienoic acid fragment 6 via amide coupling
as outlined retrosynthetically in Scheme 1. We envisioned
a (a) AcCl, toluene, reflux; (b) Vitride, toluene, -78 °C, then
HCl; (c) allyltrimethylsilane, TiCl4, toluene, 50 °C.
Scheme 1
the TiCl4-mediated allylation of racemic 11, was obtained
as a crystalline product7 suitable for X-ray crystallography
(Figure 1) which allowed assignment of the retentive
stereochemistry of 12. Preference for the formation of the
(6R)-isomer 12 in this case can be rationalized by a retentive
allylation process via an SN1-like mechanism consistent with
previous results from our laboratories.6 Accordingly, the
initially formed N-acyliminium ion 14 is expected to adopt
conformation A with the hydrogen atom in the inside position
and the bulky aromatic ring nearly perpendicular to the
iminium CdN plane to minimize the 1,3-allylic strain. In
this conformation, coordination of the titanium(IV) phenox-
ide with the carbonyl oxygen atom is likely. In such a
stereochemical arrangement, the silane nucleophile approach
can then occur from the face opposite to the aromatic ring
to generate the R configuration at the reaction center. In this
(5) Whitlock, G. A.; Carreira, E. M. J. Org. Chem. 1997, 62, 7916-
7917.
(6) (a) Yamazaki, N.; Ito, T.; Kibayashi, C. Tetrahedron Lett. 1999, 40,
739-742. (b) Yamazaki, N.; Ito, T.; Kibayashi, C. Org. Lett. 2000, 2, 465-
467.
(7) Prepared by recrystallization from benzene-hexane as colorless
needles having mp 104-106 °C.
that 5 would be formed via the piperidine derivative 7, which
might be derived from the (R)-piperidinyl acetate 8. On the
basis of this analysis, at the outset, studies were directed
toward the development of an effective enantioselective
synthesis of 8. Recently, our group has investigated Lewis
194
Org. Lett., Vol. 3, No. 2, 2001

Scheme 3a
Figure 1. X-ray crystallographic structure of racemic 12 presented
by one enantiomer.
case, the preferred axial attack by the silane nucleophile is
validated by Stevens’ stereoelectronic principle.8
a (a) MeI, K2CO3, acetone; (b) OsO4-NaIO4, dioxane-H2O; (c)
LiAlH4, Et2O, reflux; (d) H2/Pd-C, MeOH; (e) (Boc)2O, NaOH,
dioxane-H2O; (f) PDC, DMF; (g) CF3CO2H, CH2Cl2; (h) Me3Al-
NH4Cl, benzene, 50 °C; (i) LiAlH4, (i-Pr)2O, reflux.
to 21, transformation of the ester group to a cyano group to
form 22 occurred instead of amide formation. The resultant
nitrile 22 was converted to the aminomethylindolizidine 5
by LiAlH4 reduction by using diisopropyl ether as a solvent.12
As the synthesis of the aminomethylindolizidine fragment
5 was thus achieved, we then undertook the preparation of
the (S)-trienoic acid fragment 6. Starting from geraniol (23),
(2R,3S)-3,7-dimethyl-6-octene-1,2-diol (24) was prepared
following known procedures involving asymmetric epoxi-
dation13 with L-diethyl tartrate followed by reductive ring
opening (Scheme 4).14 Compound 24 was then subjected to
oxidative glycol cleavage to give the aldehyde 25. On the
other hand, ethyl (2E)-4-hydroxy-2-methyl-2-butenoate (26)15
was converted to the phosphonate ester 28 by bromination
(Ph3P, CBr4) followed by phosphorylation. Subsequent
Horner-Emmons olefination with the aldehyde 25 (BuLi,
-78 °C) resulted in the (2E,4E)-trienoate 29 as a geo-
metrically single product (71% yield), which was then
hydrolyzed to give the trienoic acid 6.
The (6R)-allyl-2-piperidinone 12 so obtained was subjected
to methylation of the phenolic hydroxyl group, followed by
oxidative cleavage of the olefin with OsO4 and NaIO4, to
give the aldehyde 15 (Scheme 3). After LiAlH4 reduction
of 15, the chiral auxiliary was cleaved by catalytic hydro-
genolysis to provide the amino alcohol 16, which was
converted to the (2R)-piperidinyl acetate 17 via a sequence
involving N-Boc protection, PDC oxidation in DMF,9 and
esterification. Diastereoselective allylation of 17 was per-
formed according to literature procedure10 by using allyl
bromide and LHMDS to furnish the desired 1′R product 18
(84% de, 76% yield). Oxidative cleavage of the olefin
converted 18 to the aldehyde 19. N-Boc deprotection of 19
followed by catalytic hydrogenation resulted in the indolizi-
dine 21 via intramolecular reductive amination as a single
isomer. When the procedure for aluminum-mediated amide
formation using trimethylaluminum-NH4Cl11 was applied
With the trienoic acid fragment 6 now in hand, the stage
was set to combine 6 with the above-described amine
(11) (a) Levin, J. I.; Turos, E.; Weinreb, S. M. Synth. Commun. 1982,
12, 989-993. (b) Wood, J. L.; Khatri, N. A.; Weinreb, S. M. Tetrahedron
Lett. 1979, 4907-4910.
(12) Considerable epimerization was observed when THF was used as
a solvent.
(13) Katsuki, T.; Sharpless, K. B. J. Am. Chem. Soc. 1980, 102, 5974-
5976.
(8) Stevens, R. V. Acc. Chem. Res. 1977, 10, 193-198.
(9) Corey, E. J.; Schmidt, G. Tetrahedron Lett. 1979, 399-402.
(10) Morley, C.; Knight, D. W.; Share, A. C. Tetrahedron: Asymmetry
1990, 1, 147-150.
(14) Taber, D. F.; Houze, J. B. J. Org. Chem. 1994, 59, 4004-4006.
(15) Nicolaou, K. C.; Liu, J.-J.; Yang, Z.; Ueno, H.; Sorensen, E. J.;
Claiborne, C. F.; Guy, R. K.; Hwang, C.-K.; Nakada, M.; Nantermet, P. G.
J. Am. Chem. Soc. 1995, 117, 634-644.
Org. Lett., Vol. 3, No. 2, 2001
195

natural product ([R]25D -24.2 (c 0.5, CHCl3)2). These results
revealed that the structure of stellettamide B is not as
originally formulated and suggested that the natural alkaloid
is epimeric at C(6′′) as depicted by structure 4 (or its antipode
ent-4). To clarify this issue, we undertook the synthesis of 4
containing the aminomethylindolizidine structure with the
same absolute stereochemistry as that found in natural
stellettamides A (1) and C (3). The synthesis followed the
above-described amide coupling protocol using the aminom-
ethylindolizidine and trienoic acid fragments. The synthesis
of the required (R)-trienoic acid ent-6 was carried out by
using the same sequence outlined in Scheme 4 except for
the use of the (R)-aldehyde ent-25 prepared via asymmetric
epoxidation using D-diethyl tartrate. The DCC-DMAP
condensation of 5 with ent-6 furnished the coupling product
31 (85% yield), which was converted into the chloromethy-
late 4 by sequential treatment with MeI and AgCl (Scheme
Scheme 4a
1
6). Both H and 13C spectra of the obtained compound 4
Scheme 6a
a (a) NaIO4, dioxane-H2O; (b) CBr4, Ph3P, MeCN; (c) (EtO)3P,
reflux; (d) BuLi, THF, -78 °C; (e) LiOH, THF-H2O, reflux.
fragment 5 via amide coupling for the synthesis of the
proposed structure of stellettamide B (2). Thus, the two
fragments 5 and 6 were condensed by using DCC-DMAP
to give the coupling product 30 in 83% yield (Scheme 5).
Subsequent quarternization with iodomethane led to the
iodomethylate which was, upon exposure to AgCl, converted
to the chloromethylate, providing 2 in 84% yield. The
synthetic sample of 2 exhibited a 13C NMR spectrum which
was quite superimposable on that of natural stellettamide B
kindly supplied by Professor Shin. Otherwise, the 1H NMR
spectrum of synthetic 2 was very similar to that of the natural
product, but careful inspection of these spectra revealed that
there are significant differences in the signal patterns
observed for the protons at C(3), C(5), and C(9) in the
1-(aminomethyl)indolizidine moiety. The most striking dif-
ference was found between the optical rotation of synthetic
a (a) BuLi, THF, -78 °C; (b) LiOH, THF-H2O, refl; (c) DCC,
DMAP, CH2Cl2; (d) MeI, MeOH, then AgCl.
were found to be completely identical to those of natural
stellettamide B. Additionally, the observed optical rotation
value, [R]27 -28.0 (c 0.71, CHCl3), agreed satisfactorily
D
2 ([R]24 +23.6 (c 0.58, CHCl3)) and that reported for the
with the reported value ([R]25 -24.2 (c 0.5, CHCl3)2).
D
D
In conclusion, the first total synthesis of (-)-stellettamide
B has been achieved by a sequence based on amide coupling
of the chiral 1-(aminomethyl)indolizidine fragment, prepared
by TiCl4-mediated asymmetric allylation of the tricyclic
N-acyl-N,O-acetal, with the chiral trienoic acid fragment This
synthesis led to revision of the published relative stereo-
chemistry of the natural product and established its absolute
stereochemistry to be 1S,4S,8aR,6′′R.
Scheme 5a
Acknowledgment. We are grateful to Professor J. Shin
(Korean Ocean Research and Development Institute) for
1
providing authentic H and 13C NMR spectra of natural
stellettamide B.
OL0003228
a (a) DCC, DMAP, CH2Cl2; (b) MeI, MeOH, then AgCl.
196
Org. Lett., Vol. 3, No. 2, 2001