Convergent synthesis of trisubstituted Z -allylic esters by wittig-schlosser reaction

Article abstract of DOI:10.1021/ol101843q

β-Lithiooxyphosphonium ylides, generated in situ from aldehydes and Wittig reagents, react readily with halomethyl esters to form trisubstituted Z-allylic esters. The methodology was applied to a total synthesis of the geranylgeraniol-derived diterpene (6S,7R,Z)-7-hydroxy-2-((E)-6-hydroxy-4- methylhex-4-enylidene)-6,10-dimethylundec-9-enyl acetate (12).

Full text of DOI:10.1021/ol101843q

ORGANIC
LETTERS
2010
Vol. 12, No. 18
4204-4207
Convergent Synthesis of Trisubstituted
Z-Allylic Esters by Wittig-Schlosser
Reaction
David M. Hodgson* and Tanzeel Arif
Department of Chemistry, Chemistry Research Laboratory, UniVersity of Oxford,
Mansfield Road, Oxford OX1 3TA, United Kingdom
daVid.hodgson@chem.ox.ac.uk
Received August 6, 2010
ABSTRACT
ꢀ-Lithiooxyphosphonium ylides, generated in situ from aldehydes and Wittig reagents, react readily with halomethyl esters to form trisubstituted
Z-allylic esters. The methodology was applied to a total synthesis of the geranylgeraniol-derived diterpene (6S,7R,Z)-7-hydroxy-2-((E)-6-hydroxy-
4-methylhex-4-enylidene)-6,10-dimethylundec-9-enyl acetate (12).
The Wittig reaction of aldehydes 1 with phosphonium ylides
2 continues to be a popular method for alkene synthesis.1
Scheme 1. Wittig-Schlosser Reaction
An important variant is the Schlosser modification,2 which
involves delaying the normal phosphine oxide elimination
from the initially formed oxaphosphetane intermediate 3, by
the presence of excess soluble lithium salts and the addition
of an organolithium (preferably PhLi)3 at low temperature
(Scheme 1). The lithium salts are believed to promote ring
opening of the oxaphosphetane 3 to give a betaine-salt
complex, which is susceptible to R-lithiation, resulting in a
new ylide. Although the latter is likely comprised of complex
species in solution,1b,4 it can be represented as ꢀ-lithiooxy-
phosphonium ylide 4 for reactivity purposes. Addition of
certain electrophiles to ylide 4 results in successful trapping
at the ylidic carbon, which is then followed by elimination
of phosphine oxide. Trapping with a proton source allows
access to trans-1,2-disubstituted alkenes 5 (E ) H) from
unstabilized phosphonium ylides 2 (R2 ) alkyl);2,3 the latter
would lead to cis-alkenes in a normal Wittig reaction.
Trapping with other electrophiles constitutes a potentially
powerful (triply convergent) approach to trisubstituted alk-
enes;5 however, synthetically useful yields and high stereo-
selectivities are generally only attained for reactions with
aldehydes6 or for halogenation.7 With excess dry paraform-
(1) (a) Maryanoff, B. E.; Reitz, A. B. Chem. ReV. 1989, 89, 863–927.
(b) Vedejs, E.; Peterson, M. J. Top. Stereochem. 1994, 21, 1–157. (c)
Schobert, R.; Ho¨lzel, C.; Barnickel, B. In Science of Synthesis; de Meijere,
A., Ed.; Thieme: Stuttgart, 2010; Vol. 47a, pp 9-84.
(2) Schlosser, M.; Christmann, K. F. Angew. Chem., Int. Ed. Engl. 1966,
5, 126.
(5) Schlosser, M.; Christmann, K.-F. Synthesis 1969, 38–39.
(6) Corey, E. J.; Yamamoto, H. J. Am. Chem. Soc. 1970, 92, 226–228,
corrigendum, 3523.
(3) Wang, Q.; Deredas, D.; Huynh, C.; Schlosser, M. Chem.sEur. J.
2003, 9, 570–574.
(4) Vedejs, E.; Meier, G. P. Angew. Chem., Int. Ed. Engl. 1983, 22,
(7) Hodgson, D. M.; Arif, T. J. Am. Chem. Soc. 2008, 130, 16500–
16501.
56–57
.
10.1021/ol101843q  2010 American Chemical Society
Published on Web 08/19/2010

aldehyde6 (or monomeric formaldehyde)8 as the electrophile,
the process leads to primary Z-allylic alcohols 5 (E )
CH2OH), and although this chemistry has found signifi-
cant use in natural product synthesis9 [mainly with
ethylidene(triphenyl)phosphorane9a,d-g ], the yields in those
cases are often only moderate. On the basis that a reactive
organohalide might efficiently trap the ylide 4, we examined
halomethyl esters as alternative electrophiles and report here
on their promising reactivity.
Table 1. Z-Allylic Acetates 6 from ꢀ-Lithiooxyphosphonium
Ylides
Under optimal Wittig-Schlosser conditions,3,7 we found
that the ꢀ-lithiooxy ylide derived from hydrocinnamaldehyde
(1a) and butylidene(triphenyl)phosphorane (2a) reacted with
bromomethyl acetate10 (1.1 equiv) to give the allylic acetate
6a in 80% yield and excellent stereoselectivity, Z > 99%
(Table 1, entry 1).11,12 While an allylic ester might be
required for a particular synthetic purpose (see later ex-
amples), if instead the corresponding allylic alcohol is
desired,13 then it can also be easily obtained; in an otherwise
identical experiment, the crude allylic acetate 6a was
hydrolyzed with NaOMe in MeOH to give the corresponding
allylic alcohol (77%) in a one-pot process. For comparison,
in an otherwise identical Wittig-Schlosser reaction, but using
dry paraformaldehyde as the electrophile, we obtained the
allylic alcohol in 52% yield (Z > 99%). The scope of the
allylic acetate synthesis was next examined with a range of
other aldehydes 1 and phosphoranes 2 (Table 1).
Allylic acetates 6 were generally obtained in good yields
and excellent Z-selectivities (Table 1); erosion in stereocon-
trol was only observed with benzaldehyde (entry 4) and also
with an aliphatic aldehyde and ethylidene(triphenyl)phos-
phorane (entry 8). R,ꢀ-Unsaturated aldehydes proved viable
substrates (entries 3, 6, and 9). It was of interest to examine
the potential scope of the process for generating allylic esters
other than acetates. Sterically more demanding and non-
enolizable halomethyl pivaloates and benzoates would, if
successful as electrophiles, provide more robustly protected
allylic alcohols. Also, although bromomethyl acetate is
commercially available, chloromethyl pivalate is considerably
less expensive (likely due to its use in pivaloyloxymethyl
ester prodrugs, of ꢀ-lactam antibiotics for example).14 In the
event, chloromethyl pivaloate and benzoate10 displayed
similar scope to bromomethyl acetate in terms of yield and
stereoselectivity for allylic pivalates 7 and benzoates 8
(Figure 1). Hydrolysis of dienyl pivaloate 7e using KOt-Bu/
H2O (2:1)15 gave the corresponding dienyl alcohol (92%),
previously used in a synthesis of 8-epi-dendrobine,9d and
(8) Schlosser, M.; Coffinet, D. Synthesis 1971, 380–381.
(9) For examples (target and yield for allylic alcohol indicated), see:
(a) Corey, E. J.; Yamamoto, H.; Herron, D. K.; Achiwa, K. J. Am. Chem.
Soc. 1970, 92, 6635-6636 [JH-I, 60%]. (b) Corey, E. J.; Yamamoto, H.
J. Am. Chem. Soc. 1970, 92, 6636-6637 [JH-I, 50%]. (c) Corey, E. J.;
Yamamoto, H. J. Am. Chem. Soc. 1970, 92, 6637-6638 [farnesol, 46%].
(d) Borch, R. F.; Evans, A. J.; Wade, J. J. J. Am. Chem. Soc. 1977, 99,
1612-1619 [8-epi-dendrobine, 40%]. (e) Takano, S.; Goto, E.; Ogasawara,
K. Tetrahedron Lett. 1982, 23, 5567-5570 [nuciferol, 42%]. (f) Liu, J.-
S.; Tao, Y. Tetrahedron 1992, 48, 6193-6798 [manwuweizic acid, 54%].
(g) Fu¨rstner, A.; Gastner, T.; Rust, J. Synlett 1999, 29-32 [iricinin-4, 40%].
(h) Imamura, Y.; Takikawa, H.; Sasaki, M.; Mori, K. Org. Biomol. Chem.
2004, 2, 2236-2244 [mispyric acid, 22%].
(10) Commercially available, but was prepared from the corresponding
acyl halide, paraformaldehyde and ZnCl2: Sosnovsky, G.; Rao, N. U. M.;
Li, S. W.; Swartz, H. M. J. Org. Chem. 1989, 54, 3667–3674.
(11) All yields reported are for chromatographically purified products.
All E/Z ratios reported were determined on crude reaction mixtures by GC/
MS. E/Z assignments were based on NOE studies.
(12) Chloromethyl acetate gave the allylic acetate 6a in 71% yield, >99%
Z.
Figure 1
. Z-Allylic pivalates 7, benzoates 8, and propionates 9
(13) For a review on allylic alcohols, see: Hodgson, D. M.; Humphreys,
P. G. In Science of Synthesis; Clayden, J., Ed.; Thieme: Stuttgart, 2007;
Vol. 32, pp 583-665.
prepared from ꢀ-lithiooxy ylides.11
Org. Lett., Vol. 12, No. 18, 2010
4205

provides improved access to this intermediate compared with
the original Wittig-Schlosser hydroxymethylation.
The successful generation of enolizable allylic esters higher
than acetates would provide substrates directly suitable for
diastereoselective (and potentially enantioselective) Ireland-
Claisen rearrangements.16 Figure 1 indicates that allylic
propionates 9 could be accessed with excellent stereocontrol,
albeit with, in general, slightly lower yields than found for
the previously examined esters. Ireland-Claisen rearrange-
ment of the E-silyl ketene acetal of allylic propionate 9a
gave γ,δ-unsaturated acid 11 in 80% yield (dr ) 95:5), where
the major diastereomer was assigned on the basis of the
reaction preferentially proceeding through a chairlike transi-
tion state 10 (Scheme 2).16
Scheme 4. Synthesis of Phosphonium Salt 14
Scheme 2. Ireland-Claisen Rearrangement of 9a
To demonstrate the utility of the above allylic ester
forming methodology, we focused on a synthesis of the
recently reported geranylgeraniol-derived diterpene (6S,7R,Z)-
7-hydroxy-2-((E)-6-hydroxy-4-methylhex-4-enylidene)-6,10-
dimethylundec-9-enyl acetate (12, Scheme 3), which was
alcohol 19 (96% ee by chiral GC) using Me2CuLi18 was
avoided by using MeMgBr and CuBr·Me2S.19 Three carbon
homologation of diol-derived iodide 23 with 1-chloro-3-
iodopropane using organocopper chemistry gave chloride 24
(74%), which was straightforwardly converted into phos-
phonium salt 14.
Scheme 3. Synthetic Analysis of Diterpene 12
Scheme 5. Completion of Synthesis of Diterpene 12
isolated from the seeds of Carpesium triste.17 Diterpene 12
is part of a family of Z-allylic acetates, some of which show
cytotoxic activity.
The phosphonium salt 14 for the central Wittig-Schlosser
step was prepared from the known diol 20,18 the latter being
readily available in four steps from propargyl alcohol (15)
and prenyl bromide (16, Scheme 4). The originally reported
modest (2:1) regioselectivity in the ring opening of epoxy
(14) Testa, B.; Mayer, J. M. Hydrolysis in Drug and Prodrug Metabo-
lism; Wiley-VCH: Weinheim, Germany, 2003.
In the key step, Z-allylic acetate 25 was formed in 63%
yield from phosphonium salt 14, aldehyde 13 (three steps
from geraniol),20 and bromomethyl acetate (Scheme 5).
(15) Gassman, P. G.; Schenk, W. N. J. Org. Chem. 1977, 42, 918–920.
(16) (a) Ireland, R. E.; Mueller, R. H.; Willard, A. K. J. Am. Chem.
Soc. 1976, 98, 2868–2877. (b) For a review, see: Mart´ın Castro, A. M.
Chem. ReV. 2004, 104, 2939–3002. (c) See also: Fleming, F. F.; Liu, W.;
Ghosh, S.; Steward, O. M. J. Org. Chem. 2008, 73, 2803–2810.
(17) Gao, X.; Lin, C.-J.; Jia, Z.-J. J. Nat. Prod. 2007, 70, 830–834.
(18) Kong, L.; Zhuang, Z.; Chen, Q.; Deng, H.; Tang, Z.; Jia, X.; Li,
Y.; Zhai, H. Tetrahedron: Asymmetry 2007, 18, 451–454.
(19) Mikami, K.; Koizumi, Y.; Osawa, A.; Terada, M.; Takayama, H.;
Nakagawa, K.; Okano, T. Synlett 1999, 1899–1902.
(20) See Supporting Information for details.
4206
Org. Lett., Vol. 12, No. 18, 2010

Subsequent PMB deprotection using DDQ gave diterpene
12 (73%), with spectral and specfic rotation data in accord
with that reported for the natural product.17 Our synthesis
confirms the original structural and stereochemical assign-
ment, where the latter was based on a modified Mosher’s
method.
thermore, we have demonstrated the methodology in an
asymmetric synthesis of the naturally occurring diterpene 12.
Acknowledgment. We thank the Higher Education Com-
mission of Pakistan for studentship support (to T.A.) and Prof.
Z.-J. Jia (Lanzhou) for NMR spectra of diterpene 12.
Supporting Information Available: Full experimental
procedures and characterization data. This material is avail-
able free of charge via the Internet at http://pubs.acs.org.
In conclusion, we have shown a new, experimentally
straightforward method for the stereocontrolled formation
of allylic esters by in situ trapping of ꢀ-lithiooxyphospho-
nium ylides with readily available halomethyl esters. Fur-
OL101843Q
Org. Lett., Vol. 12, No. 18, 2010
4207