Total synthesis of (-)-dysiherbaine, a novel neuroexcitotoxic amino acid

Article abstract of DOI:10.1016/S0040-4039(00)00518-9

A total synthesis of (-)-dysiherbaine, a neuroexcitotoxic amino acid isolated from the Micronesian marine sponge Dysidea herbacea, starting from a carbohydrate precursor, is described. The present synthesis features a palladium(0)-catalyzed cross-coupling of the 6/5-bicyclic core with an amino acid residue. (C) 2000 Elsevier Science Ltd.

Full text of DOI:10.1016/S0040-4039(00)00518-9

TETRAHEDRON
LETTERS
Pergamon
Tetrahedron Letters 41 (2000) 3923–3926
Total synthesis of (−)-dysiherbaine, a novel neuroexcitotoxic
amino acid
Makoto Sasaki,a,∗ Tatsuki Koike,a Ryuichi Sakai b and Kazuo Tachibana a
aDepartment of Chemistry, Graduate School of Science, The University of Tokyo, and CREST, Japan Science and Technology
Corporation (JST), Hongo, Bunkyo-ku, Tokyo 113-0033, Japan
bSchool of Fisheries Science, Kitasato University, Sanriku-cho, Iwate 022-0100, Japan
Received 1 March 2000; revised 21 March 2000; accepted 24 March 2000
Abstract
A total synthesis of (−)-dysiherbaine, a neuroexcitotoxic amino acid isolated from the Micronesian marine
sponge Dysidea herbacea, starting from a carbohydrate precursor, is described. The present synthesis features a
palladium(0)-catalyzed cross-coupling of the 6/5-bicyclic core with an amino acid residue. © 2000 Elsevier Science
Ltd. All rights reserved.
Keywords: amino acids and derivatives; marine metabolites; natural products; toxins.
Dysiherbaine (1) was recently isolated as a neuroexcitotoxin from the Micronesian marine sponge
Dysidea herbacea.1 Radioligand binding assay of 1 towards ionotropic glutamate receptors and elec-
trophysiological experiments indicated that 1 is a potent agonist of non-NMDA (N-methyl-D-aspartate)
subtype receptors in the central nervous system (CNS). On the basis of extensive spectroscopic studies
including long-range carbon–proton coupling constants (2,3JC,H) analysis,2 the structure of 1 was deter-
mined as an unprecedented diamino dicarboxylic acid, which is characterized by a structurally novel
cis-fused hexahydrofuro[3,2-b]pyran ring system containing a glutamate substructure.1 Due to its unique
skeletal structure and potent neuroexcitatory activity, dysiherbaine may become a useful lead compound
for the design of selective and powerful agonists or antagonists of glutamate receptors; however, its
supply from natural sources is limited. Therefore, total synthesis of 1 and its designed analogues is
required for further physiological studies. In connection with the synthetic studies on dysiherbaine, we
have already reported the synthesis of a dysiherbaine model lacking the hydroxyl and methylamino
groups.3 Very recently, total syntheses of 1 have been achieved by two independent groups.4 In this
letter, we describe a total synthesis of dysiherbaine (1) by a convergent strategy that can be adopted to
facilitate the synthesis of a variety of analogues.
Our retrosynthetic analysis for dysiherbaine (1) is shown in Scheme 1. We planned to construct the 6/5-
cis-fused bicyclic ring system of 1 via a 5-exo ring closure of hydroxy exo-olefin 2. The key intermediate
∗
Corresponding author. Fax: +81-3-5841-8380; e-mail: msasaki@chem.s.u-tokyo.ac.jp (M. Sasaki)
0040-4039/00/$ - see front matter © 2000 Elsevier Science Ltd. All rights reserved.
PII: S0040-4039(00)00518-9
tetl 16797

3924
2 was envisioned to be synthesized by palladium(0)-catalyzed cross-coupling reaction of the known
organozinc reagent 35 with vinyl iodide 4, which could be accessed from dianhydrohexose 5. The use
of a carbohydrate as a chiral starting material would allow for the synthesis of various analogues with
respect to the bicyclic core structure.
Scheme 1. Structure and retrosynthesis of dysiherbaine (1)
Synthesis of vinyl iodide 4 commenced with the known 1,6:2,3-dianhydrohexose 56 (Scheme 2).
Benzylation of 5 under the usual conditions was accompanied with migration of the epoxide ring6 to give
benzyl ether 6 in 86% yield. Epoxide ring-opening in 6 with sodium azide proceeded regioselectively to
give azide alcohol 7, which was then treated with acetic anhydride in the presence of BF3·OEt2 to yield
triacetate 8 as a single stereoisomer in 84% yield for the two steps. Reduction of the anomeric acetoxyl
group was effected with Et3SiH in the presence of TMSOTf and BF3·OEt2 to give diacetate 9 in 78%
yield. The acetyl groups were exchanged to t-butyldimethylsilyl (TBS) groups and reduction of the azide
group with PPh3, followed by protection of the resulting amino group as the t-butyl carbamate, gave N-
Boc derivative 10 in 84% overall yield for the four steps. N-Methylation of 10 and selective cleavage of
the primary silyl group gave alcohol 11 in 96% yield, which was converted to the corresponding triflate
and then treated with lithium trimethylsilylacetylide in THF–HMPA to give, after desilylation, alkyne 12
in 76% overall yield. The hydroxyl group of 12 was inverted by using an oxidation–reduction sequence to
give alcohol 13 as a single stereoisomer in 95% yield. After silylation of 13, iodoboration of the resulting
alkyne with B-iodo-9-BBN followed by treatment with acetic acid was accompanied with deprotection
of the Boc group to produce vinyl iodide 4 in 60% yield.7
Organozinc reagent 35 was prepared from the protected iodoalanine (5 equiv.) following the procedure
of Jackson et al.8 and in situ treated with 4 in the presence of PdCl2(PPh3)2 (17 mol%) at 35°C to furnish
the desired cross-coupled product 14 (Scheme 3). Removal of the TBS group followed by protection
of the amino group gave alcohol 2 in 58% overall yield from 4. Epoxidation of the exo-olefin of 2
with m-chloroperbenzoic acid (mCPBA) in CH2Cl2:pH 7.0 phosphate buffer (1:1) provided epoxide 15
as an approximately 1:1 mixture of diastereomers in 92% yield. Upon treatment of this mixture with
camphorsulfonic acid (CSA), 5-exo-ring closure proceeded smoothly to yield an inseparable mixture of
6/5-bicyclic alcohol 16 and its δ-lactone, which was then hydrolyzed with 1 N aqueous NaOH to give
169 as a mixture of diastereomers at C4.
Conversion of 16 into dysiherbaine (1) and its C4 epimer 19 required oxidation of the primary hydroxyl
group to the carboxylic acid and removal of the protective groups. We found unexpected transformation
of 16 into fully protected dysiherbaine 17 and its C4 epimer 18. Thus, oxidation of crude 16 with
TPAP/NMO,10 concentration of the reaction mixture under reduced pressure, and treatment of the residue
with trimethylsilyldiazomethane provided 17 and 18 (24 and 16% isolated yields, respectively, from 15),
which were separated by chromatography on silica gel.11 Finally, hydrogenolysis of the benzyl group

3925
Scheme 2. Synthesis of vinyl iodide 4. Reagents and conditions: (a) NaH, BnBr, DMF, 0°C→rt, 86%; (b) NaN3, NH4Cl,
MeOCH2CH2OH–H2O, reflux; (c) BF3·OEt2, Ac2O, rt, 84% (two steps); (d) Et3SiH, TMSOTf, BF3·OEt2, CH2Cl2–CH3CN,
rt, 78%; (e) NaOMe, MeOH, rt; (f) TBSOTf, 2,6-lutidine, CH2Cl2; (g) Ph3P, THF, rt, then H2O, 45°C; (h) Boc2O, Et3N,
CH2Cl2, rt, 84% (four steps); (i) NaH, Mel, DMF, 0°C, 96%; (j) CSA, CH2Cl2–MeOH, rt, 99%; (k) Tf2O, 2,6-lutidine, CH2Cl2,
−78°C; (l) trimethylsilylacetylene, BuLi, THF–HMPA, −78°C; (m) Bu4NF, THF, rt, 76% (three steps); (n) (COCl)2, DMSO,
i-Pr2NEt, CH2Cl2, −78°C→rt; (o) NaBH4, THF–MeOH, −78→0°C, 95% (two steps); (p) TBSOTf, 2,6-lutidine, CH2Cl2, 0°C;
(q) B-1-9-BBN, pentane, −20°C, then AcOH, 60% (two steps)
Scheme 3. Total synthesis of dysiherbaine (1) and its C4-epimer 19. Reagents and conditions: (a) PdCl2(PPh3)2, THF–DMA,
35°C; (b) Bu4NF, THF, rt; (c) Boc2O, Et3N, CH2Cl2, rt, 58% (three steps); (d) mCPBA, CH2Cl2–pH 7.0 phosphate buffer,
rt, 92%; (e) CSA, CH2Cl2, rt; (f) 1N NaOH, THF, rt; (g) TPAP, NMO, 4 Å molecular sieves, CH3CN, rt; (h) TMSCHN2 (20
equiv.), MeOH, rt, 17: 24% from 15, 18: 16% from 15; (i) H2, Pd/C, MeOH, rt; (j) 6N HCl, 120°C, 1: 90% (two steps), 19: 89%
(two steps)
followed by acid hydrolysis with 6N HCl furnished synthetic dysiherbaine (1) and its C4 epimer 19
as their hydrochloride salts in 90 and 89% yields, respectively. Each compound 1 and 19 was further
purified by TOSO HW40 column (1.5×160 cm; eluent H2O; UV 210 nm; flow rate 0.25 mL/min).12 The
synthetic dysiherbaine was confirmed to be identical to natural dysiherbaine by 1H NMR, ESIMS, and
HPLC analysis.
The convulsant activity of synthetic dysiherbaine (1) in mice after an intracerebral injection (ED50=9.8

3926
pmol/mouse) was virtually identical with that of the natural sample (ED50=13 pmol/mouse). The
diastereomeric compound 19 was also found to be active in vivo although the potency (ED50=385
pmol/mouse) was about 40 times less than that of 1. These results are somewhat surprising since the
C4 epimer of dysiherbaine model compound was virtually inactive, whereas the model compound with
native stereochemistry at C4 displayed CNS activity.3 Detailed pharmacological properties of 19 along
with the model compounds will be published elsewhere.
In summary, a total synthesis of (−)-dysiherbaine (1) was accomplished from readily available
dianhydrohexose 5. The synthesis described herein should allow for the preparation of a variety of
analogues of this neuroexcitotoxin for further neurobiological studies. Further synthetic studies along
this line are currently underway.
Acknowledgements
This work was financially supported in part by a Grant-in-Aid for Scientific Research from the Ministry
of Education, Science, Sports and Culture, of Japan.
References
1. Sakai, R.; Kamiya, H.; Murata, M.; Shimamoto, K. J. Am. Chem. Soc. 1997, 119, 4112–4116.
2. Matsumori, N.; Kaneno, D.; Murata, M.; Nakamura, H.; Tachibana, K. J. Org. Chem. 1999, 64, 866–876, and references
cited therein.
3. Sasaki, M.; Maruyama, T.; Sakai, R.; Tachibana, K. Tetrahedron Lett. 1999, 40, 3195–3198.
4. A first total synthesis of dysiherbaine (1) has been achieved by the Iwabuchi and Hatakeyama group of Nagasaki University.
They completed the total synthesis by a cross-coupling based strategy very similar to ours and established the absolute
configuration of dysiherbaine as shown in structure 1: Masaki, H.; Maeyama, J.; Kamada, K.; Iwabuchi, Y.; Hatakeyama,
S. In 41th Symposium on the Chemistry of Natural Products; Symposium papers, 1999; pp. 13–18. Very recently, the second
total synthesis has been reported: Snider, B. B.; Hawryluk, N. A. Org. Lett. 2000, 2, 635–638.
5. Palladium(0)-catalyzed cross-coupling reaction of 3 with aryl halides has been successfully used for the preparation of
various amino acids: Jackson, R. F. W.; Wishart, N.; Wood, A.; James, K.; Wythes, M. J. J. Org. Chem. 1992, 57,
3397–3404.
6. Trnka, T.; Cerny, M. Collect.Czech. Chem. Commun. 1971, 36, 2216–2225.
7. Hara, S.; Dojo, H.; Takinami, S.; Suzuki, A. Tetrahedron Lett. 1983, 24, 731–734.
8. Dunn, M. J.; Jackson, R. F. W.; Pietruszka, J.; Wishart, N.; Ellis, D.; Wythes, M. J. Synlett 1993, 499–500.
9. Hydroxy carboxylic acid 16 was easily cyclized to the corresponding δ-lactone under weakly acidic conditions. Therefore,
crude 16 was passed through a short column of neutral SiO2 (10% MeOH–CHCl3 containing 1% Et3N) and immediately
used in the next oxidation reaction.
10. Ley, S. V.; Norman, J.; Griffith, W. P.; Marsden, S. P. Synthesis 1994, 639–666.
11. Following the oxidation of 16 with TPAP/NMO, the usual workup followed by purification by chromatography on silica
gel and esterification with trimethylsilyldiazomethane provided a mixture of spirolactam methyl esters i and ii.
1
12. H NMR data (400 MHz, D2O) for compound 19: δ 4.04 (1H, brdd, J=5.6, 1.7 Hz, 6-H), 3.96 (1H, brs, 9-H), 3.96 (1H,
brd, J=4.3 Hz, 7-H), 3.83 (1H, d, J=13.2 Hz, 10-H), 3.80 (1H, dd, J=11.7, 1.5 Hz, 2-H), 3.49 (1H, m, 8-H), 3.46 (1H, d,
J=13.2 Hz, 10-H), 2.64 (3H, s, NMe), 2.50 (1H, dd, J=15.4, 5.6 Hz, 5-H), 2.36 (1H, dd, J=15.8, 1.5 Hz, 3-H), 2.10 (1H, d,
J=15.4 Hz, 5-H), 2.01 (1H, dd, J=15.8, 11.7 Hz, 3-H).