Enantioselective total synthesis of the osteoclastogenesis inhibitor (+)-symbioimine

Article abstract of DOI:10.1002/anie.200605160

(Chemical Equation Presented) No bones about it: The first enantioselective total synthesis of the osteoclastogenesis inhibitor (+)-symbioimine (1) has been achieved. The synthesis features a convergent enol silane addition to a dimethyl acetal and a key,

Full text of DOI:10.1002/anie.200605160

Communications
DOI: 10.1002/anie.200605160
Biomimetic Synthesis
Enantioselective Total Synthesis of the Osteoclastogenesis Inhibitor
(+)-Symbioimine**
Justin Kim and Regan J. Thomson*
The isolation, structural determination, and biological activity
of the unusual tetracyclic alkaloid, symbioimine (1), was
reported by the research group of Uemura in 2004
(Scheme 1).[1] The alkaloid 1 was isolated from Symbiodi-
nium sp., a culture from a marine dinoflagellate that has a
Scheme 1. The structures of symbioimine (1) and neosymbioimine (2).
Scheme 2. Proposed biosynthetic pathways to symbioimine (1).
symbiotic relationship with the acoel flatworm Amphisco-
lops sp., and inhibits osteoclastogenesis in RAW264 cells
(EC50 = 44 mgmLꢀ1) while maintaining cell viability at con-
centrations up to 100 mgmLꢀ1. As such, symbioimine (1) or
analogues may offer a new opening to the development of
preventative treatments for osteoporosis. Further studies by
the research group of Uemura have revealed that 1 inhibits
cyclooxygenase-2 activity, suggesting that this zwitterion may
also find use as an anti-inflammatory therapeutic agent.
Neosymbioimine (2), a compound related to 1, was reported
in 2005.[1c]
Several hypotheses regarding the biogenetic origin of the
6,6,6-tricyclic iminium ring system present within 1 and 2 have
been suggested. In the initial study, Uemura and co-workers
suggested a plausible pathway based on an exo-selective
Diels–Alder reaction akin to the conversion of 3 into 1 as
shown in Scheme 2.[1a] It seemed unlikely to us that the
cyclization of 3 would proceed with exo selectivity in the
absence of an enzyme. Furthermore, we expected that face
selectivity of the dienophile that arises from the methyl group
would be poor. In light of these considerations we sought to
investigate an alternative route that would proceed with
greater diastereoselectivity in the key bond-forming event.
An endo-selective cyclization of the cyclic iminium species 4
should occur with high facial selectivity, and thus the
predicted major diastereomer 5 would arise from an addition
opposite the methyl substituent. Epimerization of the resul-
tant ring system from all-cis-fused to all-trans-fused arrange-
ments should be driven by favorable thermodynamics (5!1).
During the course of our own investigation, Uemura and co-
workers published a similar idea,[1c] and Snider and Che
reported a model study in which an N-acyl iminium ion
underwent a selective Diels–Alder reaction followed by
epimerization.[2] Two additional studies by the research
groups of Maier and Uemura have explored alternative
cycloaddition strategies,[3] and the former group recently
published a synthesis of racemic 1. Very recently Snider and
co-workers reported the second total synthesis of racemic 1
based on their strategy of the cycloaddition of an acyl iminium
ion.[4] Herein we report the results of our own efforts that
have culminated in the first enantioselective synthesis of
(+)-symbioimine (1).
[*] Prof. R. J. Thomson
Department of Chemistry
Northwestern University
2145 Sheridan Rd, Evanston, IL 60208 (USA)
Fax: (+1)847-467-2184
E-mail: r-thomson@northwestern.edu
J. Kim
Department of Chemistry & Chemical Biology
Harvard University
12 Oxford St, Cambridge, MA 02138 (USA)
[**] We thank Prof. D. Uemura and Dr. M. Kita for providing an authentic
sample of (+)-symbioimine. A portion of this work was conducted
at Harvard University and Prof. D. A. Evans is gratefully acknowl-
edged for support. This work was supported by Northwestern
University (NU), the NU Analytical Services Laboratory (NSF grants
DMR0114235 and CHE9871268), and the Mass Spectrometry
Laboratory at the University of Illinois, Chicago.
As a result of the ease with which dihydropyridine
derivatives are oxidized or can disproportionate,[5] we plan-
ned to access the desired cyclic iminium ion (4, R = Me) in a
one-pot cascade sequence than began with the tetrahydro-
pyridine 13. We hoped that exposure of 13 to either Brønsted
or Lewis acids would effect elimination, followed by cyclo-
addition and epimerization to afford the skeleton of the
natural product. From the beginning, we elected to protect
Supporting information for this article is available on the WWW
under http://www.angewandte.org or from the author.
3104
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2007, 46, 3104 –3106

Angewandte
Chemie
the phenolic groups as the corresponding methyl ethers, with
the reason that double deprotection and selective sulfation
conditions could be found once our key cycloaddition had
been achieved. The synthesis of imine 13 is outlined in
Scheme 3.
Our initial efforts sought to effect thermal elimination of
methanol from 13, but only decomposition products that
resulted from oxidation were observed. During this time,
Kishi and co-workers reported an intramolecular Diels–Alder
reaction mediated by a Brønsted acid of an exocyclic
unsaturated imine en route to the marine toxin gymnod-
imine.[12] However, no reaction was observed when we
exposed 13 to the conditions reported by Kishi and co-
workers (pH 6.5 buffer). Nevertheless, further exploration
revealed a set of conditions that did effect the desired cascade
sequence (Scheme 4). Heating 13 in a 0.5m solution of TFA at
Scheme 3. Synthesis of the precursor to the Diels–Alder reaction 13.
a) Dimethyl 3,5-dimethoxybenzylphosphonate, NaHMDS, 82%;
b) MeMgCl, NHMe(OMe)·HCl, 70%; c) iPrMgCl, NHMe(OMe)·HCl;
MeMgCl, 90% (2 steps) d) LDA, TMSCl, Et3N, 99%; e) LiBH4; TsCl, py,
61% (2 steps); f) 9 + 11, TMSOTf, 2,6-di-tert-butylpyridine, 58%;
g) NaN3; solid-supported Ph3P, 98% (2 steps). HMDS=hexamethyldi-
silazide, LDA=lithium diisopropylamide, py=pyridine, Tf=trifluoro-
methanesulfonyl, TMS=trimethylsilyl, Ts=toluenesulfonyl.
A Horner–Wadsworth–Emmons reaction between the
known aldehyde 6[6] and dimethyl 3,5-dimethoxybenzyl-
phosphonate[7] provided the desired diene 7 with high levels
of E/Z selectivity (> 11:1) and in good yield (82%). Prepa-
ration of methyl ketone 8 could be achieved in 70% yield by a
one-step protocol using NHMe(OMe)·HCl and excess
MeMgCl.[8] On a larger scale it was more convenient to use
a two-step procedure which involved the initial isolation of
the intermediate Weinreb amide, and its subsequent con-
version into the methyl ketone 8 (90% yield, two steps).
Silylation of 8 by using the Corey protocol[9] afforded enol
silane 9 (70% yield, two steps). We proposed to couple 9 with
dimethyl acetal 11, which was readily prepared from the
thione 10[10] (97% ee), by reduction and tosylation (61%
yield, two steps). In the event, exposure of a 2:1ratio of 9 and
11 to TMSOTf in the presence of 2,6-di-tert-butylpyridine
provided the desired adduct 12 in 58% yield (based on the
tosylate 11). Treatment of 12 with NaN3, followed by a
Staudinger reduction and aza-Wittig reaction delivered the
imine 13 in excellent yield (98%, two steps).[11]
Scheme 4. Diels–Alder cascade reaction and completion of the syn-
thesis. a) 0.5m TFA (4:1 water/THF), 508C; b) TFAA, Et3N, 25–37%
(2 steps); c) K2CO3/MeOH, 99%; d) 1.0m BBr3, 82%; e) SO3·py, py,
27%. TFA=trifluoroacetic acid, TFAA=trifluoroacetic anhydride.
508C for 24 h provided the desired cycloadduct 16, presum-
ably by the pathway outlined in Scheme 4. Analysis of the
1H NMR spectrum of the unpurified reaction mixture
revealed that the cyclization afforded a single diastereoiso-
mer. For ease of purification, the reaction mixture was treated
with trifluoroacetic anhydride, and in this manner acetamide
17 was isolated in 25–37% yield from 13. Tentative proof of
the stereochemistry was obtained at this juncture by observ-
ing key NOE signals within 17 (see Scheme 4). Trifluoro-
acetamide 17 could be converted back into 16 by using a mild
methanolysis (K2CO3/MeOH). Brief exposure of 16 to an
excess of 1.0m boron tribromide for one hour cleanly
provided the corresponding bisphenol derivative. Sulfation
using the modified conditions reported by Maier and co-
Angew. Chem. Int. Ed. 2007, 46, 3104 –3106ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
3105

Communications
4795; b) M. Kita, D. Uemura, Chem. Lett. 2005, 454 – 459; c) M.
Kita, N. Ohishi, K. Washida, M. Kondo, T. Koyama, K. Yamada,
D. Uemura, Bioorg. Med. Chem. 2005, 13, 5253 – 5258.
[2] B. B. Snider, Q. Che, Angew. Chem. 2006, 118, 946 – 949; Angew.
Chem. Int. Ed. 2006, 45, 932 – 935.
[3] a) G. N. Varseev, M. E. Maier, Angew. Chem. 2006, 118, 4885 –
4889; Angew. Chem. Int. Ed. 2006, 45, 4767 – 4771; b) E. Sakai,
K. Araki, H. Takamura, D. Uemura, Tetrahedron Lett. 2006, 47,
6343 – 6345.
[4] Y. Zou, Q. Che, B. B. Snider, Org. Lett. 2006, 8, 5605 – 5608.
[5] a) R. F. Francis, C. D. Crews, B. S. Scott, J. Org. Chem. 1978, 43,
3227 – 3230; b) R. J. Sundberg, D. S. Grierson, H.-P. Husson, J.
Org. Chem. 1984, 49, 2400 – 2404; c) N. Naiman, H. Rollema, E.
Johnson, N. Castagnoli, Jr., Chem. Res. Toxicol. 1990, 3, 133 –
138.
[6] M. A. Ciufolini, S. Zhu, J. Org. Chem. 1998, 63, 1668 – 1675.
[7] J. Drew, M. Letellier, P. Morand, A. G. Szabo, J. Org. Chem.
1987, 52, 4047 – 4052.
[8] J. M. Williams, R. B. Jobson, N. Yasuda, G. Marchesini, U.-H.
Dolling, E. J. J. Grabowski, Tetrahedron Lett. 1995, 36, 5461–
5464.
[9] E. J. Corey, A. W. Gross, Tetrahedron Lett. 1984, 25, 495 – 498.
[10] D. A. Evans, R. J. Thomson, J. Am. Chem. Soc. 2005, 127,
10506 – 10507.
[11] a) H. Staudinger, J. Meyer, Helv. Chim. Acta 1919, 2, 635 – 646;
b) M. Vaultier, N. Knouzi, R. Carrie, Tetrahedron Lett. 1983, 24,
763 – 764.
[12] J. W. Johannes, S. Wenglowsky, Y. Kishi, Org. Lett. 2005, 7, 3997 –
4000.
workers provided 1 in 27% yield ([a]D + 251( c = 0.15,
DMSO) [lit. [a]D + 245 (c = 0.1, DMSO)]). The spectroscopic
data (1H and 13C NMR, MS) of synthetic 1 corresponded with
those of the natural product.[13]
Our Diels–Alder sequence is distinctive relative to the
previously reported approaches to symbioimine (1) because it
proceeds via iminium ion intermediates[14,15] rather than acyl
iminium ions[2] or ketones.[3] As such, the reaction provides
direct support that such an intramolecular [4+2] cycloaddi-
tion reaction may be involved in the biosynthesis of 1. The
moderate yield obtained is probably a result of the generation
of pyridine derivatives that could not be completely avoided.
Notably we did not isolate any of the unusual structures that
were reported by Snider and co-workers[4] during the cyclo-
additions of the acyl iminium ions, although these cannot be
completely ruled out. More importantly, this example of a
rare cycloaddition of a dihydropyridinium ion is a step
towards the expansion of the use of these entities in complex
molecule synthesis. From a synthetic standpoint, the direct
conversion of 13 into 16 represents a significant increase in
structural complexity: two new rings and four stereocenters
are formed in a diastereoselective manner from the presence
of a single methyl substituent. The overall brevity of the
sequence detailed here suggests that it may be useful for the
synthesis of analogues of 1 and in the identification of
structure–activity relationships with respect to osteoclasto-
genesis inhibition. Such research efforts will become increas-
ingly more important because of the increase in elderly
populations around the world.
[13] See the Supporting Information for experimental procedures
and spectroscopic data for all compounds.
[14] For examples of 5,6-dihydropyridinium ions as dienophiles, see:
a) J. E. Baldwin, D. R. Spring, R. C. Whitehead, Tetrahedron
Lett. 1998, 39, 5417 – 5420; b) K. Jakubowicz, K. Ben Abdeljelil,
M. Herdemann, M.-T. Martin, A. Gateau-Olesker, A. Al-
Mourabit, C. Marazano, B. C. Das, J. Org. Chem. 1999, 64,
7381– 7387; c) J. E. Baldwin, T. D. W. Claridge, A. J. Culshaw,
F. A. Heupel, V. Lee, D. R. Spring, R. C. Whitehead, Chem. Eur.
J. 1999, 5, 3154 – 3161; d) M. Herdemann, A. Al-Mourabit, M.-T.
Martin, C. Marazano, J. Org. Chem. 2002, 67, 1890 – 1897.
[15] For the use of 3,4-dihydropyridinium ions as dienes in biomim-
etic Diels–Alder reactions, see: R. B. Ruggeri, M. M. Hansen,
C. H. Heathcock, J. Am. Chem. Soc. 1988, 110, 8734 – 8736.
Received: December 21, 2006
Revised: January 24, 2007
Published online: March 15, 2007
Keywords: biomimetic synthesis · cycloaddition ·
.
Diels–Alder reaction · natural products · total synthesis
[1] a) M. Kita, M. Kondo, T. Koyama, K. Yamada, T. Matsumoto, K.
Lee, J. Woo, D. Uemura, J. Am. Chem. Soc. 2004, 126, 4794 –
3106
www.angewandte.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2007, 46, 3104 –3106