Total synthesis of (±)-acetoxyodontoschismenol using zirconium chemistry

Article abstract of DOI:10.1039/b309848f

The dolabellane diterpene (±)-acetoxyodontoschismenol has been synthesised for the first time by a short route in which a three component coupling on zirconium is used to assemble all the carbons needed for the skeleton in one-pot.

Full text of DOI:10.1039/b309848f

Total synthesis of (±)-acetoxyodontoschismenol using zirconium
chemistry†
Ian R. Baldwin and Richard J. Whitby*
Department of Chemistry, University of Southampton, Southampton, UK SO17 1BJ.
E-mail: rjw1@soton.ac.uk
Received (in Cambridge, UK) 15th August 2003, Accepted 19th September 2003
First published as an Advance Article on the web 11th October 2003
The dolabellane diterpene ( )-acetoxyodontoschismenol has
been synthesised for the first time by a short route in which
a three component coupling on zirconium is used to
assemble all the carbons needed for the skeleton in one-
pot.
stereochemistry. Molecular modelling indicated that the desired
trans fused isomer of 2 was > 12 kJ mol21 more stable than the
cis and that addition of an isopropyl group should occur from
the desired side.9 The obvious cyclisation precursor to 2 was the
alcohol 4 but its zirconocene induced cyclisation failed,
presumably due to alkoxide elimination from an intermediate
2
We have described how zirconacyclopentanes, readily derived
by co-cyclisation of 1,n-dienes,1 could be elaborated by
insertion of metallated allyl chlorides to afford allylzircono-
cenes followed by addition of electrophiles (Scheme 1).2 We
now report the first total synthesis of the dolabellane diterpene
(±)-acetoxyodontoschismenol3 1 in which a sequence of
zirconium mediated transformations are used to assemble all the
carbons required for the basic skeleton in one pot.4
The dolabellanes are a widely distributed group of > 150
naturally occurring diterpenoids characterised by a trans-
bicyclo[9.3.0]tetradecane skeleton, many of which exhibit
antimicrobial, antitumor and antiviral activities. The family is
also important as the likely biogenetic precursors, via trans-
annular cyclisations, of several other diterpene classes including
the fusicoccanes and dolastanes. The chemistry of the dola-
bellanes has recently been reviewed.5 The total synthesis of
(±)-d-araneosene, dolabellatrienone, palominol, claenone, and
stolonidiol,6 and the closely related neodolabellenol and
4,5-deoxyneodolabelline7 have been reported. Acetoxyodon-
toschismenol 1 is the major dolabellane of five isolated from the
liverwort Odontoschisma denudatum and displayed moderate
growth-inhibitory activity on a series of plant pathogenic
fungi.3
h -alkene complex.10 Allylic silyl groups are compatible with
zirconocene induced co-cyclisations and may act as masked
hydroxyl groups. We chose the diphenylethoxysilyl substituted
diene 6 as our starting material since Tamao oxidation of an
alkoxysilane to a hydroxyl occurs under mild conditions.11 The
diene 6 was synthesised from the allylic chloride 512 by SN2A
displacement using the silylcuprate Ph2(Et2N)SiCu(CN)Li,11
subsequent ethanolysis furnishing an easily separated 22 : 1
mixture of the desired diene 6 and the regioisomer 7 resulting
from SN2 displacement (Scheme 2). The ratio of regioisomers 6
and 7 was sensitive to the quality and quantity of CuCN used, it
being crucial to avoid the presence of any [Ph2(Et2N)Si]2CuLi.
A suitable equivalent 8 for the aldehyde component required in
Eq 1 was synthesised from acetol (Scheme 2).
(1)
Retrosynthetic analysis, based on the transformation shown
in Scheme 1, suggested that acetoxyodontoschismenol could be
rapidly constructed from the three components shown in Eq 1.
Model studies indicated that the ring-junction methyl group in a
zirconacycle such as 3 would direct carbenoid insertion to the
adjacent C–Zr bond, as required, but also that the ring-junction
stereochemistry of 3 was likely to be predominantly cis.8 We
thus chose the ketone 2 as an advanced intermediate in our
synthesis to allow epimerisation of the adjacent ring junction
With all the components now available the key zirconocene
induced co-cyclisation, carbenoid insertion and electrophile
addition was attempted. Addition of the diene 6 to dibutylzirco-
nocene (generated in situ from zirconocene dichloride and 2
equiv. of BuLi) in THF at 278 °C followed by stirring at room
temperature for 12 h formed a zirconacyclopentane.1 After
cooling to 278 °C methallyl chloride was added, followed by
the dropwise addition of lithium 2,2,6,6-tetramethylpiperidide
(LiTMP). The in situ generated carbenoid inserted into the
zirconacycle to afford an allylzirconium species which was
further elaborated by the BCl3 promoted addition of aldehyde 8,
followed by iodinolytic cleavage of the final carbon–zirconium
bond. The resulting unstable iodide 9 was immediately reacted
with the sodium salt of benzenesulfinic acid to furnish sulfone
10 as a mixture of 4 diastereoisomers (4 : 4 : 1 : 1) in 51%
overall yield based on 6. The use of BCl3 rather than the usual
BF3·Et2O2 in the aldehyde addition step avoided by-products
probably resulting from cleavage of the C–SiPh2(OEt) bond. As
expected, complete relative control was observed between C-11
and C-12, but the remote centre at C-6 was formed as a 1 : 1
mixture (dolabellane numbering).13 We were surprised and
pleased to find that the major isomers of 10 had the desired
relative stereochemistry between the carbons C-1 and C-11
implying that the intermediate zirconacycle was formed as a 4
: 1 trans : cis mixture. Molecular modelling suggests that steric
Scheme 1 Tandem reactions on zirconocene.
† Electronic supplementary information (ESI) available: comparison of
reported 13C NMR data of natural 1 and 17 with that observed for synthetic
1 and 17 and their C-6 epimers. See http://www.rsc.org/suppdata/cc/b3/
b309848f/
2786
CHEM. COMMUN., 2003, 2786–2787
This journal is © The Royal Society of Chemistry 2003

View Article Online
in 16 by reaction with Na/NH3 at 270 °C furnished diol 17,
which is rac-(1R,6R,11R,12R)-6,12-dihydroxydolabella-
3E,7E-diene, one of the minor dolabellanes isolated from the
same source as acetoxyodontoschismenol.3
Selective acetylation of the secondary allylic alcohol over the
tertiary alcohol in 17 gave (±)-acetoxyodontoschismenol 1. The
NMR data of both 17 and 1 were in good agreement with the
literature.3† The C-6 epimer 14 was carried through the same
sequence as above to afford 6-epi-acetoxyodontoschismenol,
the NMR data of which was very different to that of the natural
product 1.†
In summary we have accomplished the first total synthesis of
the dollabellane diterpene acetoxyodontoschismenol in 12
steps, and 3.5% overall yield from the chloride 5. The synthesis
illustrates the use of tandem reactions on a zirconocene template
(cocyclisation, carbenoid insertion, aldehyde addition, iodinol-
ysis) to rapidly assemble complex compounds from simple
fragments. Macrocyclisation was achieved in excellent yield via
an a-lithiosulfone intramolecular displacement of an allylic
chloride.
We thank GSK, Pfizer, Oxford Molecular, and AstraZeneca
for funding the Society of the Chemical Industry Young
Chemists Award studentship to IRB. We also thank our
industrial collaborators in the work for help and advice: Dr.
Julian Blagg, Dr. John Harling, Dr. Mike Harris, Dr. Prem
Meghani, and Dr. Paul Smith.
Notes and references
1 C. J. Rousset, D. R. Swanson, F. Lamaty and E. Negishi, Tetrahedron
Lett., 1989, 30, 5105; W. A. Nugent and D. F. Taber, J. Am. Chem. Soc.,
1989, 111, 6435.
2 T. Luker and R. J. Whitby, Tetrahedron Lett., 1994, 35, 785; T. Luker
and R. J. Whitby, Tetrahedron Lett., 1994, 35, 9465; T. Luker and R. J.
Whitby, Tetrahedron Lett., 1995, 36, 4109.
3 A. Matsuo, K. I. Kamio, K. Uohama, K. I. Yoshida, J. D. Connolly and
G. A. Sim, Phytochemistry, 1988, 27, 1153; A. Matsuo, K. Uohama, S.
Hayashi and J. D. Connolly, Chem. Lett., 1984, 599.
4 T. Luker and R. J. Whitby, Tetrahedron Lett., 1996, 37, 7661.
5 A. D. Rodriguez, E. Gonzalez and C. Ramirez, Tetrahedron, 1998, 54,
11683.
6 L. Jenny and H. J. Borschberg, Helv. Chim. Acta, 1995, 78, 715; H.
Miyaoka, Y. Isaji, H. Mitome and Y. Yamada, Tetrahedron, 2003, 59,
61; H. Miyaoka, T. Baba, H. Mitome and Y. Yamada, Tetrahedron Lett.,
2001, 42, 9233; T. Hu and E. J. Corey, Org. Lett., 2002, 4, 2441; E. J.
Corey and R. S. Kania, J. Am. Chem. Soc., 1996, 118, 1229; E. J. Corey
and R. S. Kania, Tetrahedron Lett., 1998, 39, 741.
Scheme 2 Reagents and conditions: (i) Li dispersion, Ph2(Et2N)SiCl (3
equiv.), THF, 0 °C, then CuCN (3 equiv.), then 5 (1 equiv.), 278 °C–rt, 88%
t
of 6, 4% of 7; (ii) Me2 BuSiCl, imidazole, DMF, 94%; (iii) NaH,
(EtO)2P(O)CH2CO2C2H5, THF, 0–20 °C, 86% of 9, 12% of Z-isomer; (iv)
iBu2AlH, toluene, 0–20 °C, 80%; (v) BaMnO4, CH2Cl2, 20 °C, 81%; (vi) (1)
ZrCp2Cl2, 2 equiv. n-BuLi, THF, 278–20 °C. (2) CH2NC(CH3)CH2Cl. (3)
t
LiTMP, THF, 278 °C. (4) Me2 BuSiOCH2C(Me)NCHCHO, BCl3, 278–20
°C. (5) I2, 278–0 °C, 65%; (vii) C6H5SO2Na, HMPA/Et2O, 45 °C, 2 h,
78%; (viii) BOMCl, iPrEt2N, CH2Cl2, 0–20 °C, 83%; (ix) TBAF, THF, 20
°C, 1 h then H2O2, KHCO3, MeOH/THF, 12 h, 73%; (x) NCS, Me2S,
CH2Cl2, 220–0 °C, 94%; (xi) added to refluxing LiHMDS (0.05 M) in THF
over 5 h, 78%; (xii) Mg/MeOH, 50 °C, 3h, 72%; (xiii) Dess–Martin
i
periodinane, CH2Cl2, 0 °C, 93%; (xiv) PrMgCl, CeCl3, THF, 0 °C, 0.5h,
76%; (xv) Na, NH3, THF, EtOH, 270 °C, 86%; (xvi) Ac2O, pyridine, rt,
81%.
repulsion between the bulky silyl group and the methyl group at
C-1 in the intermediate zirconacycle accounts for the reversal in
ring junction stereochemistry compared with the model.8 The
secondary allylic alcohol was protected as its benzyloxymethyl
ether (BOM) then both cleavage of the tert-butyldimethylsilyl
ether and Tamao oxidation of the (EtO)Ph2Si–C bond were
accomplished with Bu4NF, KF, H2O2, KHCO3 to afford the diol
11 in 71% yield. The macrocyclisation precursor 12 was formed
by selective conversion of the primary allylic alcohol to a
chloride by reaction with N-chlorosuccinimide and Me2S in
CH2Cl2.14 Addition of 12 (0.02 M) to refluxing LiHMDS in
THF (0.05 M) over 5 h furnished the desired macrocycle as a
mixture of 8 diastereoisomers in an excellent 78% yield.
Cyclisation of the analogous allylic iodide occurred in much
lower yield. The sulfone moiety was removed with Mg/
MeOH15 to afford the now readily separable epimers at C-6, 13
and 14, each a 4 : 1 mixture of trans : cis ring junctions. Dess–
Martin oxidation of 13 afforded ketone 15 in good yield. All
attempts to alter the ring junction stereochemistry of 15 by
epimerisation failed, indeed we could not exchange the C-11
proton with MeOD/MeONa. Fortunately, reaction of ketone 15
with i-PrMgCl in the presence of cerium trichloride16 was
selective for the trans-isomer giving diastereoisomerically pure
16. In the cis-fused isomer of 15, both approaches to the ketone
are blocked, one by the C-1 methyl group, the other by being on
the endo-face, so it does not react. Cleavage of the BOM ether
7 D. R. Williams and P. J. Coleman, Tetrahedron Lett., 1995, 36, 35; D.
R. Williams and R. W. Heidebrecht, J. Am. Chem. Soc., 2003, 125,
1843.
8 G. J. Gordon, T. Luker, M. W. Tuckett and R. J. Whitby, Tetrahedron,
2000, 56, 2113.
9 Carried out using MM2 force field as implemented in the CaChe
Workstation 4.0, Oxford Molecular.
10 For conversion of allyl ethers to allylzirconium species see: H. Ito, T.
Nakamura, T. Taguchi and Y. Hanzawa, Tetrahedron, 1995, 51, 4507.
Interestingly the isomer of 4, 2-hydroxy-2-methyl-1,6-heptadiene,
cyclises well—only leaving groups allylic to the alkene which first
complexes to the zirconium cause problems.
11 K. Tamao, A. Kawachi, Y. Tanaka, H. Ohtani and Y. Ito, Tetrahedron,
1996, 52, 5765.
12 Chloride 5 was synthesised from 2-methyl-2-propen-1-ol via condensa-
tion/Claisen rearrangement with triethyl orthoacetate to give ethyl
4-methyl-4-pentenoate. Subsequent LiAlH4 reduction to 4-methyl-
4-penten-1-ol, PCC oxidation to the aldehyde and Wadsworth–Emmons
olefination with triethyl phosphonoacetate furnished ethyl 6-me-
thylhepta-2,6-dienoate which was reduced with DiBAl-H to the primary
alcohol and converted into chloride 5 via the mesylate.
13 B. L. Pagenkopf, E. C. Lund and T. Livinghouse, Tetrahedron, 1995,
51, 4421.
14 E. J. Corey, C. U. Kim and M. Takeda, Tetrahedron Lett., 1972, 13,
4339.
15 A. C. Brown and L. A. Carpino, J. Org. Chem, 1985, 50, 1749.
16 T. Imamoto, N. Takiyama, K. Nakamura, T. Hatajima and Y. Kamiya,
J. Am. Chem. Soc., 1989, 111, 4392.
CHEM. COMMUN., 2003, 2786–2787
2787