Enantioselective total synthesis of (-)-neovibsanin G and (-)-14-epi-neovibsanin G

Article abstract of DOI:10.1039/c1cc15995j

The first total synthesis of vibsane-type diterpenoids neovibsanin G and 14-epi-neovibsanin G has been achieved. Key to this endeavour was a late stage EtAlCl₂ mediated skeletal rich cascade leading to the bicyclo[3.3.1]nonane core in one step.

Full text of DOI:10.1039/c1cc15995j

View Article Online / Journal Homepage / Table of Contents for this issue
Dynamic Article Links
ChemComm
Cite this: Chem. Commun., 2012, 48, 287–289
www.rsc.org/chemcomm
COMMUNICATION
Enantioselective total synthesis of (À)-neovibsanin G and
(À)-14-epi-neovibsanin Gw
Jeffrey Y. W. Mak and Craig M. Williams*
Received 27th September 2011, Accepted 1st November 2011
DOI: 10.1039/c1cc15995j
The first total synthesis of vibsane-type diterpenoids neovibsanin G
and 14-epi-neovibsanin G has been achieved. Key to this endeavour
was a late stage EtAlCl2 mediated skeletal rich cascade leading to
the bicyclo[3.3.1]nonane core in one step.
of biosynthetic intermediate 1. This was a role that 11 (Scheme 2),
an intermediate in the synthesis of racemic 2-O-methyl-
neovibsanin H9 (6), could well satisfy, if it could be accessed
in an enantioenriched form.
Towards the synthesis of key intermediate 11, ketone 10
was formed via an asymmetric Cu catalysed 1,4-addition of
Grignard reagent 8 to cyclohexenone 7 (70%), using the chiral
NHC ligand 912–14 (Scheme 2) [synthesised15 from commercially
available (S)-tert-leucinol] forming the all-carbon quaternary
centre in 91% ee. This reaction could be performed on a 12 g
scale without any reduction in yield or selectivity.
Vibsanins are rare diterpenes that exist exclusively in a selection
of the Viburnum species such as V. awabuki, V. odoratissimum
and V. suspensum.1 These natural products can be divided into
three subtypes based on their carbon skeleton; eleven membered
ring compounds, seven-membered ring compounds, and the
rearranged type, also known as the neovibsanins.
Having obtained 10 for the first time in an enantioenriched
form, key intermediate 11 was synthesised following the pro-
cedures prescribed in the racemic series,9 setting the stage for
The neovibsanins, certain members of which display potent
neurotrophic activity,2–7 can be further divided into three
subclasses; the (1) bicyclic, (2) ketal, and (3) caged neovibsanins.
According to Fukuyama’s elegantly proposed biosynthesis,3,8
the chemical fate of intermediate 1 dictates the formation of
each neovibsanin subtype. If 1 undergoes dihydrofuran cyclisa-
tion followed by hemiketalisation (Path A, Scheme 1), this leads
to the formation of the ketal neovibsanins (e.g. 2). If 1 under-
goes elimination instead (Path B), allylic cation 3 is formed.
Subsequent capture of this cation by water (Path D) gives rise to
the bicyclic neovibsanins (e.g. 6), while an intramolecular
capture of the cation by the endogenous olefin (Path C) leads
to the construction of the bicyclo[3.3.1]nonane ring system of
the caged neovibsanins (e.g. 4 and 5).
The natural products (Æ)-2-O-methylneovibsanin H9 (6)
and (Æ)-neovibsanin B (2)5–7,10,11 represent respective targets
from the bicyclic and ketal subclasses that have succumbed to
total synthesis following a biomimetic route. As the synthesis
of the caged neovibsanins had not previously been explored,
the synthesis of this subclass was needed in order to substantiate
the proposed biosynthetic pathway beyond doubt (Scheme 1).
It is in this light that the total synthesis of two targets belonging
to the caged subclass, (À)-neovibsanin G (4) and (À)-14-epi-
neovibsanin G (5), was undertaken, the results of which are
herein reported.
A central aim for the translation of this proposed biosynthesis
to a total synthesis would involve access to a synthetic equivalent
School of Chemistry and Molecular Biosciences, University of
Queensland, Brisbane, 4072, Australia. E-mail: c.williams3@uq.edu.au
w Electronic supplementary information (ESI) available: Experimental
procedures and copies of 1H and 13C NMR spectra. CCDC 841524.
See DOI: 10.1039/c1cc15995j
Scheme 1 Proposed biosynthesis of the three neovibsanin subclasses.
Chem. Commun., 2012, 48, 287–289 287
c
This journal is The Royal Society of Chemistry 2012

View Article Online
Scheme 2 Enantioselective synthesis of key intermediate 11 via chiral
NHC-chelated copper catalysed 1,4-addition to 7. (Inset: ORTEP
diagram of 11; ellipsoids are shown at 30% probability.)
the crucial bicyclo[3.3.1]nonane cyclisation. The absolute con-
figuration of 11 was established as 11R through X-ray crystallo-
graphy (Scheme 2, inset).
Scheme 3 Bicyclo[3.3.1]nonane cyclisation with H2SO4 and the Brønsted
acid promoted equilibrium between 13, 14 and 17.
In finding suitable conditions for the bicyclo[3.3.1]nonane
cyclisation, the use of H2SO4 was a logical starting point, as it
was known from the synthesis9 of 2-O-methylneovibsanin H (6)
that treatment of 11 with H2SO4 in MeOH at room temperature
gave the bicyclic core of 6 via cation 12. Although the polarity
of methanol was presumed to be important in promoting the
formation of cation 12, in accordance with classical carbocation
theory, in this case 12 must be generated in the absence of
a nucleophile in order to encourage intramolecular cation
capture. Thus, a polar non-nucleophilic solvent was required.
The solvent tetrahydrofuran appeared to be the best candi-
date, and in this vein, 11 was treated with H2SO4 in THF at
40 1C, followed by methylation with diazomethane (Scheme 3).
Although the cyclisation was successful, the selectivity was
poor (15 : 16 : 18 2 : 1 : 5.7), presumably due to the Brønsted acid
catalysed isomerisation of the desired terminal alkenes to the
thermodynamically more stable tetrasubstituted isomer.
All attempts to improve the selectivity of the Brønsted acid
cyclisation proved to be futile. Thus, the employment of a Lewis
acid to kinetically control the regiochemical outcome of the
final elimination step was a logical choice. To this end, 11 was
TBS deprotected with catalytic HCl, which also triggered an
intramolecular 1,4-addition to form the dihydrofuran ring of 19,
providing a substrate that allowed the critical cyclisation to be
probed using Lewis acids.
After extensive investigations, the only acid/solvent combination
that could coerce the highly oxygenated substrate to undergo
the desired cyclisation was EtAlCl2 in THF (Scheme 4).
Gratifyingly, the regioselective outcome (15 : 16 : 18 1 : 2 : 1.5)
was drastically improved over the reaction with H2SO4. Only
trace amounts of the products were observed when Et2O
was used as the solvent, while performing the reaction in
1,2-dichloroethane resulted in the formation of an unidentifi-
able mixture. Notably, when EtAlCl2 was employed, the ratio
of 16 : 15 increased relative to the H2SO4 reaction. While 16
was presumably the kinetically favoured product, thermo-
dynamically it was less stable than 15 due to the proximity
of its C-14 pendant group to the dihydrofuran ring. Thus,
when the cyclisation was carried out using a Brønsted acid, the
ratio of 16 : 15 was decreased due to the preferential isomeri-
sation of the former isomer to the tetrasubstituted olefin 18.
In this intriguing cascade, initial treatment of 19 with EtAlCl2
induced elimination of the carboxylate to form allylic carbo-
cation 20, which in the absence of a nucleophile, was captured
by the endogenous olefin to give 21. The change in hybridisation
at C-4 forced 21 to undergo C-5 epimerisation to 23, via 22, to
alleviate the steric strain between the two carbonyl functional-
ities. Finally, aqueous work up followed by methylation with
CH2N2 gave 15, 16 and 18 in a combined yield of 60% (73%
brsm). Separation of the three isomers was achieved with careful
silver nitrate chromatography.14
The Lewis acid TMSOTf12 and the Meerwein salts13 appeared
to be the ideal reagents for the desired transformation, but each
failed to induce the bicyclo[3.3.1]nonane cyclisation when tested
in a range of solvents. Treatment of 19 with BF3ÁEt2O in THF
returned starting material, while treatment with Et2AlCl in
the same solvent resulted in gradual substrate decomposition.
Individual conversion of 15 and 16 to their respective
aldehydes (24 and 25) was achieved through successive treat-
ment with LiAlH4 and DMP (Scheme 4). Trapping the sodium
enolates of 24 and 25 with 3,3-dimethylacryloyl chloride5
c
288 Chem. Commun., 2012, 48, 287–289
This journal is The Royal Society of Chemistry 2012

View Article Online
G [(À)-5] was achieved in an enantioselective manner. This
established the absolute configuration of neovibsanin G and
14-epi-neovibsanin G as 11S, which is amongst the strongest
evidence for Fukuyama’s proposal that the biosynthetic pro-
genitor of the neovibsanins is vibsanin B3 (which was also
shown to possess the 11S configuration16). The success of this
biogenetically inspired synthesis lends overwhelming support
for Fukuyama’s postulated biosynthesis3 of the neovibsanins.
The authors thank the Australian Research Council (DP
DP0666855) for financial support, Prof. P. V. Bernhardt and
Dr P. Y. Hayes (School of Chemistry and Molecular Bio-
sciences, University of Queensland [UQ]) for X-ray crystallo-
graphical data collection and chiral HPLC purification
(Enantioselective chromatography facility) respectively,
Dr G. Pierens (Centre for Advanced Imaging, UQ) for NMR
data collection, Prof. Y. Fukuyama (Faculty of Pharmaceutical
Sciences, Tokushima Bunri University) for the NMR spectra
of natural neovibsanin G and 14-epi-neovibsanin G, and
Dr M. Mauduit (Ecole Nationale Superieure de Chimie de
Rennes) and Prof. A. Alexakis (Departement de Chimie
Organique, Universite de Geneve) for their advice concerning
the asymmetric 1,4-addition.
Notes and references
1 Y. Fukuyama, M. Kubo, T. Esumi, K. Harada and H. Hioki,
Heterocycles, 2010, 81, 1571–1602.
2 Y. Fukuyama, H. Minami, K. Takeuchi, M. Kodama and
K. Kawazu, Tetrahedron Lett., 1996, 37, 6767–6770.
3 Y. Fukuyama, M. Kubo, H. Minami, H. Yuasa, A. Matsuo,
T. Fujii, M. Morisaki and K. Harada, Chem. Pharm. Bull., 2005,
53, 72–80.
4 M. Kubo, Y. Kishimoto, K. Harada, H. Hioki and Y. Fukuyama,
Bioorg. Med. Chem. Lett., 2010, 20, 2566–2571.
5 H. Imagawa, H. Saijo, T. Kurisaki, H. Yamamoto, M. Kubo,
Y. Fukuyama and M. Nishizawa, Org. Lett., 2009, 11, 1253–1255.
6 A. P. J. Chen, C. C. Muller, H. M. Cooper and C. M. Williams,
Org. Lett., 2009, 11, 3758–3761.
7 A. P. J. Chen, C. C. Mueller, H. M. Cooper and C. M. Williams,
Tetrahedron, 2010, 66, 6842–6850.
8 M. Kubo, Y. Minoshima, D. Arimoto, H. Minami, K. Harada,
H. Hioki and Y. Fukuyama, Heterocycles, 2009, 77, 539–546.
9 A. P. J. Chen and C. M. Williams, Org. Lett., 2008, 10, 3441–3443.
10 T. Esumi, T. Mori, M. Zhao, M. Toyota and Y. Fukuyama, Org.
Lett., 2010, 12, 888–891.
11 T. Esumi, M. Zhao, T. Kawakami, M. Fukumoto, M. Toyota and
Y. Fukuyama, Tetrahedron Lett., 2008, 49, 2692–2696.
12 M. F. Jung and M. A. Lyster, J. Am. Chem. Soc., 1977, 99,
968–969.
13 H. Meerwein, P. Borner, O. Fuchs, H. J. Sasse, H. Schrodt and
J. Spille, Chem. Ber., 1956, 89, 2060–2079.
14 C. M. Williams and L. N. Mander, Tetrahedron, 2001, 57,
425–447.
15 B. D. Schwartz, C. M. Williams, E. Anders and P. V. Bernhardt,
Tetrahedron, 2008, 64, 6482–6487.
Scheme 4 EtAlCl2 mediated bicyclo[3.3.1]nonane cyclisation and the
completion of the total synthesis.
completed the synthesis of (À)-5 and (À)-4 respectively. The
moderate yields obtained for the installation of the side-chain
was typical using this approach. Although we have developed
methodology15 to access this functional group in one step and
in higher yields, it was not applicable in this case.
In conclusion, six years after their isolation,3 the first total
16 Y. Fukuyama, H. Minami, S. Takaoka, M. Kodama, K. Kawazu
and H. Nemoto, Tetrahedron Lett., 1997, 38, 1435–1438.
synthesis of (À)-neovibsanin G [(À)-4] and (À)-14-epi-neovibsanin
c
This journal is The Royal Society of Chemistry 2012
Chem. Commun., 2012, 48, 287–289 289