Total synthesis of (-)-dysiherbaine

Article abstract of DOI:10.1039/c2cc32736h

The enantioselective synthesis of (-)-dysiherbaine (1) has been established with efficiency via a unique synthetic strategy involving the desymmetrization of 2-substituted glycerol to install a quaternary chiral carbon, which induces further stereochemistry in the bicyclic perhydrofuropyran through mercuriocyclization and epoxidation. The Royal Society of Chemistry 2012.

Full text of DOI:10.1039/c2cc32736h

View Article Online / Journal Homepage / Table of Contents for this issue
Dynamic Article Links
ChemComm
Cite this: Chem. Commun., 2012, 48, 6295–6297
www.rsc.org/chemcomm
COMMUNICATION
Total synthesis of (ꢀ)-dysiherbainew
Nelma Carurucan Celindro, Tae Woo Kim and Sung Ho Kang*
Received 18th April 2012, Accepted 3rd May 2012
DOI: 10.1039/c2cc32736h
The enantioselective synthesis of (ꢀ)-dysiherbaine (1) has been
established with efficiency via a unique synthetic strategy involving
the desymmetrization of 2-substituted glycerol to install a quaternary
chiral carbon, which induces further stereochemistry in the bicyclic
perhydrofuropyran through mercuriocyclization and epoxidation.
Scheme 1 Retrosynthetic analysis of (ꢀ)-dysiherbaine 1. PMP =
4-methoxyphenyl, Boc = tert-butoxycarbonyl.
(ꢀ)-Dysiherbaine was isolated in 1997 by Sakai and his
collaborators from extracts of the Micronesian marine sponge
Dysidea herbacea.1 Later, they corrected the original source as
Synechocystis cyanobacteria harbored in the sponge Lendenfeldia
chondrodes.2 It has a characteristic molecular structure based on
a cis-fused hexahydrofuro[3,2-b]pyran bicyclic core and a
glutamate subunit with four contiguous stereogenic centers
and a quaternary asymmetric carbon. Ligand binding assay
studies revealed that dysiherbaine is a strong agonist of
KA/AMPA type glutamate receptors pertaining to ionotropic
glutamate receptors, which generate ligand-gated channels to
convey fast excitatory synaptic transmission in the mammalian
central nerve system.3 Its binding to the receptors activates ion
channels, disturbs ionic equilibria, and incurs glutamate
excitotoxicity to cause neuronal overactivation and eventual
cell death.4 Dysiherbaine displays the most potent convulsant
activity among the known excitatory amino acids such as
domoic acid and kainic acid.5 Since it induces epilepsy-like
seizures in mice, it is considered in brain research as a potential
surrogate of the current seizurogenic drugs.6 Its beneficial
pharmacological profiles and unique molecular architecture
led us to be involved in synthetic studies on dysiherbaine.
Furthermore, although many reports have been published
concerning its synthesis, most of them are not practical enough
to supply the scarce natural product.7,8 Herein, we describe an
asymmetric total synthesis of dysiherbaine with efficiency.
Our present retrosynthetic analysis of dysiherbaine (1)
breaks the C–O bonds of the tetrahydropyranyl (THP) and
tetrahydrofuranyl (THF) rings, and involves a diene alcohol
intermediate 2 (Scheme 1). Conceivably, the intermediate
evolves through mercurioetherification, with stereoselective
uncertainty, and later acid-catalyzed cyclization via an epoxy
alcohol to restore the THP and THF rings. The initial chiral
invocation guides the fate of the remaining contiguous centers.
At the design stage, we assumed that (i) the tertiary allylic oxygen of
2 might exert an electrostatic stabilizing effect on the mercuronium
cation and/or (ii) the allylic silyloxy group could sterically shield the
a-face to favour the desired asymmetry in the 6-membered
oxacycle. The construction of 2 was planned through Stille
coupling with the vinyl iodide 8 and vinyltin 10.9 Recently, we
developed an asymmetric desymmetrization of 2-substituted
glycerols to enantioselectively form the corresponding tertiary
alcohols.10 The application of the desymmetrization protocol to
the triol 3 was expected to install the required quaternary chiral
carbon, which would be carried to the intermediate 2 through 8.
The synthesis of 1 commenced with preparation of 3. After
transmetallation of the commercial iodide 4, the organozinc
iodide generated was sonicated with the vinyl iodide 5 (readily
derived from propargyl alcohol)11 in the presence of Pd12 to
afford the coupled allylic alcohol in 87% yield (Scheme 2). The
coupling reaction should be carried out at lower than 35 1C to
suppress the decomposition of the organozinc iodide. The
allylic alcohol was dihydroxylated to the triol 3, and then
benzoylated under established desymmetrization conditions
using an imine monooxazoline–CuCl2 catalyst 610a to furnish
the predicted monobenzoate 7 with higher than 97% de in
96% yield. After the diastereomeric separation, 7 was sequentially
converted into benzylidene acetals with 4-methoxybenzaldehyde
dimethyl acetal, into alcohols by reduction (LiAlH4), into aldehydes
under Dess–Martin oxidation conditions,13 and into vinyl iodides 8
with Stork’s phosphonium salt14 in 72% overall yield. All
benzylidene acetal products were 1.5 : 1 diastereomeric mixtures.
It is worthwhile to mention that while only the (Z)-iodoalkenes 8
could be detected with the crude aldehydes from the Dess–Martin
oxidation in the olefination reaction, a 2–3 : 1 mixture of the
(Z)- and (E)-geometric isomers were formed consistently with
the chromatographically purified aldehydes. We supposed that
the observed stereoselectivity difference would be attributed to
the incompletely removed pyridine. Indeed, the Wittig olefination
of the purified aldehydes generated the (Z)-stereoisomers exclusively
in the presence of a few equivalents of pyridine. The production of
vinyltin 10 as the Stille coupling partner15 to 8 was initiated by
the hydrolysis of the acetonide group of the known vinyl iodide 9.
MIRC, Dept. of Chemistry, KAIST, Daejeon 305-701, Korea.
E-mail: shkang@kaist.ac.kr; Fax: +82-42-350-2810;
Tel: +82-42-350-2825
w Electronic supplementary information (ESI) available. See DOI:
10.1039/c2cc32736h
c
This journal is The Royal Society of Chemistry 2012
Chem. Commun., 2012, 48, 6295–6297 6295

View Article Online
Scheme 2 Preparation of the dihydropyran 12. (a) Zn–Cu, TMSCl,
TMEDA, THF, sonication (150 W), RT, then 5, Pd(PPh3)4, sonication
(150 W), o35 1C, 87%. (b) OsO4, NMO, H2O, acetone, 0 1C, 99%.
(c) 6, BzCl, Et3N, THF, RT, 95%. (d) PPTS, PMPCH(OMe)2, PhH,
RT, 97%. (e) LiAlH4, Et2O, 0 1C, 97%. (f) Dess–Martin periodinane,
pyridine, CH2Cl2, 0 1C. (g) ICH2+PPh3Iꢀ, NaHMDS, HMPA, THF,
ꢀ78 1C, 81% (over 2 steps). (h) 6N HCl, MeOH, RT, 95%. (i) TBSCl,
DMAP, Et3N, CH2Cl2, RT, 92%. (j) (Me3Sn)2, Pd(PPh3)2Cl2, THF,
RT, 90% (based on 10% recovered SM). (k) PPTS, MeOH, 0 1C, 88%
(based on 15% recovered SM). (l) Pd(PPh3)4, CuI, CsF, DMF, RT,
85%. (m) Hg(CF3CO2)2, K2CO3, PhMe, ꢀ30 1C, then Et3B, LiBH4,
THF, ꢀ78 1C, 70%. (n) DDQ, 4A MS, PhH, RT. (o) K2CO3, MeOH,
RT, 95% (over 2 steps). TMS = trimethylsilyl, TMEDA = N,N,N0,N0-
Scheme 3 Preparation of 14 and NOE measurement. (a) CF3COCH3,
OXONEs, Na2EDTA, NaHCO3, MeCN, 0 1C. (b) CDCl3, RT (87% for
12 to 14, 75% for 11 to 13). (c) K2CO3, MeOH, RT, 98%. (d) InCl3, wet
MeCN, RT, 95%. (e) Ac2O, DMAP, Et3N, CH2Cl2, RT, 100%.
OXONEs = potassium peroxymonosulfate, EDTA = ethylenediamine-
tetraacetate, s = strong, m = medium, w = weak.
We contemplated introducing an oxygen instead of the iodo
functional group. Accordingly, 12 was subjected to epoxidation
using mCPBA, dimethyldioxirane and other peracids; it was
found to be unreactive under mild conditions and decompose
under harsh conditions. Eventually, the employment of trifluoro-
methylmethyldioxirane18 allowed for the formation of the desired
unstable epoxide, a portion of which (about 20%) was cyclized to
the tetrahydrofuran derivative 14 under epoxidation conditions
(Scheme 3). Treatment of this crude mixture with CDCl3 gave the
cis-fused bicyclic alcohol 14 as a single stereoisomer in 87%
overall yield without any detection of the corresponding
perhydropyranopyran. Its structure was manifested through
conversion of the benzoate 11 to 14 via the bicyclic benzoate
13 via sequential epoxidation, cyclization and debenzoylation. In
order to elucidate the stereochemistry of 14 at the ring junction,
installed by the mercuriocyclization of 2 and epoxidation of 12,
several related derivatives were prepared for the purposes of
discrete 1H NMR spectral assignments that support its proposed
three-dimensional structure. The bicyclic diacetate 15 gave
tetramethylethylenediamine, NMO
Bz = benzoyl, PPTS = pyridinium p-toluenesulfonate, HMDS =
bis(trimethylsilyl)amide, HMPA hexamethylphosphoramide,
= 4-methylmorpholine N-oxide,
=
TBS = tert-butyldimethylsilyl, DMAP = 4-dimethylaminopyridine,
DMF = N,N-dimethylformamide, DDQ = 2,3-dichloro-5,6-dicyano-
1,4-benzoquinone, MS = molecular sieves.
Subsequently, the two hydroxyl groups of the demasked diol
were silylated, its iodo group was substituted by Pd-mediated
stannylation,15 and finally the primary silyloxy group was
desilylated chemoselectively to offer 10 in 69% overall yield from
9. Clean Stille coupling between 8 and 10 provided the diene
alcohol 2 in 85% yield. Coupling using the corresponding
tri-n-butyltin instead of the trimethyltin 10 was less effective
(o20% yield) probably due to steric effects. Next, diastereoselective
formation of the 6-membered heterocycle from 2 was needed,
but involved a stereochemistry challenge. The survey of various
electrophile-promoted cyclization conditions led us to resort to the
one-pot process of mercuriocyclization followed by reductive
demercuration.16 Importantly, when this process was applied to
2, the desired cis-3,6-disubstituted dihydropyran was produced in
70% yield, along with 1–2% of the trans-diastereomer. The acetals
were cleaved with DDQ17 to give a 20–25 : 1 mixture of primary
benzoate 11 and tertiary benzoate, which was debenzoylated to the
diol 12 in 95% overall yield. After separating the diastereomeric
benzylidene acetals prepared from 7, each diastereomer was
also converted to 12, respectively, as aforementioned; they
were confirmed to have similar reactivity and stereoselectivity.
The bicyclic precursor 12 was exposed to NIS to give only
the cis-fused hexahydrofuropyran in 90% yield. At this juncture,
the intermolecular substitution reaction of the iodo group by the
azide anion was attempted under various conditions. Also, after the
removal of the TBS group, we intended to deliver the requisite
nitrogen nucleophile intramolecularly through the secondary
hydroxyl group. However, all trials proved unrewarding.
1
adequate H NMR peak separations: 11 was converted to 15
via 13, in which the acetonide group was removed and
acetylated. The NOE values of 15 were obtained to support
the desired connectivity (Scheme 3). NOE signals between two
H2s/H3, one H2/H6, H3/H4 and H4/H5 suggest a tetrahydro-
pyran chair conformation; whereas H3 is positioned equatorially,
H6 is situated axially. Since much stronger NOE crosspeaks were
observed between H6/Ha7 than were for H6/He7, H6 and Ha7
are assigned to be in the same face of the tetrahydrofuran ring.
Also, crosspeaks for Ha7/H10, H5/H10 and H5/Ha7 support
that H5, H6, Ha7 and the C10 side chain occupy the b-face of the
5-membered heterocycle. Spectroscopically, it is concluded that
14 is a cis-fused bicyclic compound with a cis relationship
between C3 and C6 (Scheme 3). The stereochemical arrangement
of 14 was also confirmed by synthesizing the known dimethyl
ester 20 as well as (ꢀ)-dysiherbaine 1 (vide infra).
The remaining synthetic issue is the critical installation of
the 4-amino functionality. In this context, 14 was oxidized to
the unstable keto carboxylic acid 1619 before being converted
c
6296 Chem. Commun., 2012, 48, 6295–6297
This journal is The Royal Society of Chemistry 2012

View Article Online
to the oxime 1720 in 88% overall yield. Due to its relative
instability, it was immediately hydrogenated in the presence of
Boc2O; the resulting unstable carbamate carboxylic acid was
methylated to give 18 in 81% overall yield. The appended
acetonide, silyl and t-butoxycarbonyl groups are plausible
factors for the instability of all the carboxylic acid derivatives,
due to their acidic lability. The acidic hydrolysis of 18 followed by
the chemoselective oxidation of the primary hydroxyl group of the
generated diol gave the stable carboxylic acid 19 in 95% overall
yield, which was confirmed downstream upon its conversion to the
corresponding known methyl ester 20.7d It is noted that under
identical deprotection conditions, only the acetonide group was
removed from 13, whereas both the acetonide and silyl groups were
removed from 18. Finally, the global acidic deprotection of 19
produced the dysiherbaine hydrochloride 1ꢁ2HCl quantitatively,
which was briefly treated with aqueous NaOH and then purified on
an ion exchange resin to afford (ꢀ)-dysiherbaine 1 in 95% yield.
Other significant operations in the synthetic sequence include
Stille coupling of 8 and 10, the stereoselective mercuriocyclization
of 2, and the successful epoxidation of the unreactive alkene 12.
This research was supported by the NRF grant funded by
the MEST (2012-0000912 and 2012-0000098). We thank Prof.
David G. Churchill at KAIST for his assistance during the
preparation of this manuscript.
Notes and references
1 R. Sakai, H. Kamiya, M. Murata and K. Shimamoto, J. Am.
Chem. Soc., 1997, 119, 4112.
2 R. Sakai, K. Yoshida, A. Kimura, K. Koike, M. Jimbo, K. Koike,
A. Kobiyama and H. Kamiya, ChemBioChem, 2008, 9, 543.
3 (a) R. Dingledine, K. Borges, D. Bowie and S. F. Trynelis, Pharmacol.
Rev., 1999, 51, 7; (b) D. R. Madden, Nat. Rev. Neurosci., 2002, 3, 91;
(c) G. T. Swanson, T. Green, R. Sakai, A. Contractor, W. Che,
H. Kamiya and S. F. Heinemann, Neuron, 2002, 34, 589; (d) M. L.
Mayer, Curr. Opin. Neurobiol., 2005, 15, 282; (e) J. M. Sanders, K. Ito,
L. Settimo, O. T. Pentikainen, M. Shoji, M. Sasaki, M. S. Johnson,
R. Sakai and G. T. Swanson, J. Pharmacol. Exp. Ther., 2005, 314, 1068;
(f) M. L. Mayer, Nature, 2006, 440, 456.
4 J. M. Sanders, O. T. Pentikainen, L. Settimo, U. Pentikainen,
M. Shoji, M. Sasaki, R. Sakai, M. S. Johnson and G. T. Swanson,
Mol. Pharmacol., 2006, 69, 1849.
5 R. Sakai, G. T. Swanson, M. Sasaki, K. Shimamoto and
H. Kamiya, Cent. Nerv. Syst. Agents Med. Chem., 2006, 6, 83.
6 R. Sakai, G. T. Swanson, K. Shimamoto, A. Contractor,
A. Ghetti, Y. Tamura-Horikawa, C. Oiwa and H. Kamiya,
J. Pharmacol. Exp. Ther., 2001, 296, 650.
7 Total syntheses: (a) B. B. Snider and N. A. Hawryluk, Org. Lett.,
2000, 2, 635; (b) M. Sasaki, T. Koike, R. Sakai and K. Tachibana,
Tetrahedron Lett., 2000, 41, 3923; (c) H. Masaki, J. Maeyama,
K. Kamada, T. Esumi, Y. Iwabuchi and S. Hatakeyama, J. Am.
Chem. Soc., 2000, 122, 5216; (d) D. Phillips and A. R. Chamberlin,
J. Org. Chem., 2002, 67, 3194; (e) M. Sasaki, N. Akiyama,
K. Tsubone, M. Shoji, M. Oikawa and R. Sakai, Tetrahedron Lett.,
2007, 48, 5697; (f) K. Takahashi, T. Matsumura, J. Ishihara and
S. Hatakeyama, Chem. Commun., 2007, 4158; (g) M. Sasaki,
K. Tsubone, K. Aoki, N. Akiyama, M. Shoji, M. Oikawa,
R. Sakai and K. Shimamoto, J. Org. Chem., 2008, 73, 264.
8 Formal and partial syntheses: (a) T. Naito, J. S. Nair, A. Nishiki,
K. Yamashita and T. Kiguchi, Heterocycles, 2000, 53, 2611;
(b) S. H. Kang and Y. M. Lee, Synlett, 2003, 993; (c) O. Miyata,
R. Iba, J. Hashimoto and T. Naito, Org. Biomol. Chem., 2003, 1, 772;
(d) J. Huang, K. Xu and T. Loh, Synthesis, 2003, 755; (e) J. L. Cohen,
A. Limon, R. Miledi and A. R. Chamberlin, Bioorg. Med. Chem.
Lett., 2006, 16, 2189; (f) W. R. F. Goundry, V. Lee and J. E. Baldwin,
Synlett, 2006, 2407; (g) J. L. Cohen and A. R. Chamberlin, Tetra-
hedron Lett., 2007, 48, 2533; (h) R. Fernandez de la Pradilla, N. Lwoff
and A. Viso, Tetrahedron Lett., 2007, 48, 8141; (i) C. L. Carroll and
A. R. Chamberlin, Tetrahedron Lett., 2011, 52, 3995.
1
The H NMR spectral data of 1ꢁ2HCl are highly dependent on
sample concentration, and are similar to that reported by Snider
and Hawryluk (0.005 M)7a and by Hatakeyama et al. (0.03 M).7c
While its optical rotation is known to be [a]D = ꢀ51 (c 0.06, H2O)
by Snider and Hawryluk7a and [a]D = +7.01 (c 0.53, H2O) by
23
Hatakeyama et al.,7f our value is [a]D23 = +7.01 (c 0.54, H2O).
In the case of (ꢀ)-dysiherbaine 1, the 1H NMR spectra are
almost identical with those previously published,7c,d,f its optical
rotation found by us, [a]D24 = ꢀ7.31 (c 0.38, H2O), is in good
agreement with the reported value, [a]D23 = ꢀ7.51 (c 0.52, H2O)
by Hatakeyama et al., (Scheme 4).7f
We report an efficient asymmetric total synthesis of (ꢀ)-dysi-
herbaine from the readily available chiral building block 4. The
synthesis has demonstrated the utility of the enantioselective
desymmetrization of the 2-substituted glycerol 3 to install a
quaternary chiral center, which is thought to guide further stereo-
selectivity in the formation of the bicyclic perhydro-furanopyran.
9 S. P. H. Mee, V. Lee and J. E. Baldwin, Angew. Chem., 2004, 116, 1152.
10 (a) B. Jung and S. H. Kang, Proc. Natl. Acad. Sci. U. S. A., 2007,
104, 1471; (b) B. Jung, M. S. Hong and S. H. Kang, Angew. Chem.,
Int. Ed., 2007, 46, 2616.
11 N. Kamiya, Y. Chikami and Y. Ishii, Synlett, 1990, 675.
12 G.-Q. Shi, X.-H. Huang and F. Hong, J. Org. Chem., 1996, 61, 3200.
13 M. J. Batchelor, R. J. Gillespie, J. M. C. Golec and C. J. R.
Hedgecock, Tetrahedron Lett., 1993, 34, 167.
14 K. Zhao and G. Stork, Tetrahedron Lett., 1989, 30, 2173.
15 J. D. White, P. R. Blakemore, N. J. Green, E. B. Hauser, M. A.
Holoboski, L. E. Koewn, C. S. N. Kolz and B. W. Phillips, J. Org.
Chem., 2002, 67, 7750.
16 S. H. Kang, J. H. Lee and S. B. Lee, Tetrahedron Lett., 1998, 39, 59.
17 Z. Zhang and G. Magnusson, J. Org. Chem., 1996, 61, 2394.
18 D. Yang, M.-K. Wong and Y.-C. Yip, J. Org. Chem., 1995, 60, 3887.
19 (a) M. Shibuya, M. Tomizawa, I. Suzuki and Y. Iwabuchi, J. Am.
Chem. Soc., 2006, 128, 8412; (b) M. Inai, T. Goto, T. Furuta,
T. Wakimoto and T. Kan, Tetrahedron: Asymmetry, 2008, 19, 2771.
20 O. Miyata, R. Iba, J. Hashimoto and T. Naito, Org. Biomol.
Chem., 2003, 1, 772.
Scheme 4 Preparation of dysiherbaine 1. (a) AZADO, PhI(OAc)2,
NaClO2, pH 6.2 phosphate buffer, MeCN, RT. (b) MeONH2ꢁHCl,
pyridine, 0 1C to RT, 88% (over 2 steps). (c) H2 (1 atm), Boc2O,
NaHCO3, Raney Ni, MeOH, RT. (d) NaH, MeI, DMF, 0 1C, 81%
(over 2 steps). (e) InCl3, wet MeCN, RT (96%). (f) AZADO,
PhI(OAc)2, NaClO2, pH 5.8 phosphate buffer, MeCN, RT, 99%.
(g) TMSCHN2, MeOH, PhH, RT, 100%. (h) 6 M HCl, 80 1C, 99%.
(i) 10 M aq NaOH, RT, then Amberlite, weakly acidic cation
exchanger, H form, 95%. AZADO = 2-azaadamantane N-oxyl.
c
This journal is The Royal Society of Chemistry 2012
Chem. Commun., 2012, 48, 6295–6297 6297