Synthesis of farnesol isomers via a modified wittig procedure

Article abstract of DOI:10.1021/ol0513239

(Chemical Equation Presented) The four olefin stereoisomers of farnesol have been synthesized from readily available nerylacetone or commercial geranylacetone. A new variation on the use of β-oxido ylides favored the (2Z)-stereoisomers, whereas the (2E)-isomers were obtained through a classical Horner-Wadsworth-Emmons condensation with triethyl phosphonoacetate and reduction of the resulting ester.

Full text of DOI:10.1021/ol0513239

ORGANIC
LETTERS
2005
Vol. 7, No. 22
4803-4806
Synthesis of Farnesol Isomers via a
Modified Wittig Procedure
Jose´ S. Yu, Troy S. Kleckley, and David F. Wiemer*
Department of Chemistry, UniVersity of Iowa, Iowa City, Iowa 52242
daVid-wiemer@uiowa.edu
Received June 7, 2005
ABSTRACT
The four olefin stereoisomers of farnesol have been synthesized from readily available nerylacetone or commercial geranylacetone. A new
variation on the use of -oxido ylides favored the (2Z)-stereoisomers, whereas the (2E)-isomers were obtained through a classical Horner
Wadsworth Emmons condensation with triethyl phosphonoacetate and reduction of the resulting ester.
â
−
−
Farnesol pyrophosphate (FPP) has been recognized as a key
intermediate in the biosynthesis of more complex sesquit-
erpenoids, higher terpenoids, and steroids for many years,1
and more recently many other roles have been discovered
for both this pyrophosphate and the corresponding alcohol
farnesol.2-6 For example, farnesol acts as a quorum-sensing
molecule to suppress filamentation in the fungus Candida
albicans,2 and farnesyl derivatives are involved in the a-factor
mating peptide of the dodecapeptide pheromone found in
Saccharomyces cereVisiae.3 Terpene alcohols such as farnesol
have antibacterial effects against Staphylococcus aureus,
including promotion of potassium ion leakage.4 Furthermore,
FPP plays a pivotal role in the posttranslational processing
of Ras proteins. Because mutated forms of Ras are associated
with ∼30% of all human cancers,5,6 many investigations on
the biological activity of farnesol and FPP analogues have
been reported. In general these studies were based upon the
most abundant farnesol isomer, (2E,6E)-farnesol, while the
noncommercial isomers were ignored or only tested as
mixtures.2-4 In many of these cases, it would be of interest
to examine the consequences of olefin isomerization on
biological activity, which has prompted new attention to the
synthesis of these compounds.7 In this communication we
report the synthesis of the four olefin stereoisomers of
farnesol through variations on Wittig chemistry.
Of the four possible isomers of farnesol (Figure 1, 1-4),
only (2E,6E)-farnesol (1) is commercially available in
relatively high isomeric purity. In contrast, the three isomers
containing at least one (Z)-olefin (2, 3, and 4) are available
only as minor components in the commercial material. At
one time it was necessary to perform repeated chromato-
(1) Liang, P. H.; Ko, T. P.; Wang, A. H. Eur. J. Biochem. 2002, 269,
3339-3354.
(2) Shchepin, R.; Hornby, J. M.; Burger, E.; Niessen, T.; Dussault, P.;
Nickerson, K. W. Chem. Biol. 2003, 10, 743-750.
(3) Xie, H.; Shao, Y.; Becker, J. M.; Naider, F.; Gibbs, R. A. J. Org.
Chem. 2000, 65, 8552-8563 and references therein.
(4) Inoue, Y,; Shiraishi, A.; Hada, T.; Hirose, K.; Hamashima, H.;
Shimada, J. FEMS Microbiol. Lett. 2004, 237, 325-331.
(5) Hohl, R. J.; Lewis, K. A.; Cermak, D. M.; Wiemer, D. F. Lipids
1998, 33, 39-46.
Figure 1. Farnesol isomers.
(6) Downward, J. Curr. Opin. Genet. DeV. 1998, 8, 49-54.
10.1021/ol0513239 CCC: $30.25
© 2005 American Chemical Society
Published on Web 09/27/2005

graphic separations in hopes to obtain an enriched sample
of a noncommercial farnesol isomer.8
initial carbonyl compound employed in the condensation was
an aldehyde. If the same type of procedure were used with
an unsymmetrical ketone, it might be possible to achieve
the desired substitution pattern from this reaction sequence,
the pattern required for a primary terpenoid allylic alcohol
(Scheme 2). However, whether a ketone (e.g., 11) would
Recently, Gibbs and co-workers have reported four-step
routes to alcohols 2, 3, and 4 that utilize Stille-type couplings
on intermediate vinyl triflates.7 Although these routes provide
the respective farnesol isomers and are amenable to synthesis
of analogues, they require both extensive chromatographic
separations and the use of tin-based reagents. Given the
known toxicity of tin compounds9 and our interest in the
biological activity of the farnesol isomers, development of
a route that avoided use of these reagents was attractive.
The central issue in synthesis of at least the (2Z,6E)- and
(2Z,6Z)-isomers (2 and 3) is construction of a cis-allylic
alcohol. The past work of Schlosser10 and Corey11 provides
some precedent for use of a modified Wittig reaction that
involves â-oxido ylides in construction of other types of (Z)-
allylic alcohols (Scheme 1). For example, Schlosser reported
Scheme 2. Pattern of Substitution with Unsymmetrical Ketone
favor the desired (Z)-olefin stereochemistry (12) was not at
first clear.
The initial investigations of this strategy (Scheme 3) were
conducted with commercially available geranylacetone (13).
Reaction of ketone 13 at low temperature with the Wittig
reagent derived from methyltriphenylphosphonium bromide
presumably yielded an oxaphosphetane intermediate, and
although some of the corresponding olefin was observed,
elimination of triphenylphosphine oxide is not facile at these
low reaction temperatures. While holding the reaction
temperature at -78 °C, addition of a second equivalent of
strong base should lead to formation of intermediate 14, and
subsequent trapping via reaction with formaldehyde should
give the alkoxide intermediate (15). Quenching of the
reaction mixture with a combination of water in hexanes,
followed by the elimination of triphenylphosphine oxide, then
results in formation of farnesol 2.7
Scheme 1. Representative Applications of â-Oxido Ylides
Scheme 3. Reaction Scheme for the Modified Wittig
Procedure
preparation of compound 7 through condensation of meth-
ylenetriphenylphosphorane (generated in situ from the phos-
phonium salt 5) with hexanal (6) at low temperature,
followed by addition of base to generate the â-oxido ylide,
addition of formaldehyde, and finally allowing the reaction
to reach a temperature where elimination of triphenylphos-
phine oxide proceeds.10 Homologation of the symmetrical
ketones cyclohexanone and acetone also was reported.10
Corey and co-workers employed an analogous procedure
with the phosphorane derived from phosphonium salt 8 and
aldehyde 9 to fabricate a key intermediate (10) for the
synthesis of juvenile hormone.11 In these past studies, the
Initial experiments with procedures parallel to those
reported afforded modest and variable yields of alcohol 2
(20-29%) along with trace amounts of the (2E,6E)-isomer
1 (∼10:1 ratio). Although a tertiary allylic alcohol might be
formed by a competing elimination involving the C-1
oxygen, none of this product was detected. Even with the
modest yield, the reaction is attractive considering that it is
a one-flask protocol with only minimal chromatography.10,11
(7) (a) Shao, Y.; Eummer, J. T.; Gibbs, R. A. Org. Lett. 1999, 1, 627-
630. (b) Zahn, T. J.; Whitney, J.; Weinbaum, C.; Gibbs, R. A. Bioorg. Med.
Chem. Lett. 2001, 11, 1605-1608.
(8) Cane, D. E.; Iyengar. R.; Shiao, M. S. J. Am. Chem. Soc. 1981, 103,
914-931.
(9) Krigman, M. R.; Silverman, A. P. Neurotoxicology 1984, 5, 129-
139.
(10) Schlosser, M.; Coffinet, D. Synthesis 1972, 575-576.
(11) Corey, E. J.; Yamamoto, H. J. J. Am. Chem. Soc. 1970, 92, 6636-
6637.
4804
Org. Lett., Vol. 7, No. 22, 2005

A number of other conditions were examined in an effort
to improve the yield. Use of s-BuLi for deprotonation of
the oxaphosphetane intermediate 14 resulted in a slight
increase in yield (32%). Further improvement was achieved
(42%) when monomeric formaldehyde in ether solution was
used with s-BuLi12,13 as an alternative to use of paraform-
aldehyde. Subsequent experiments with monomeric formal-
dehyde and n-BuLi provided a lower yield (34%), thereby
suggesting that both the stronger base and the more reactive
form of formaldehyde were helpful.
Horner-Wadsworth-Emmons (HWE) reaction15 of the two
previously described ketones (13 and 18, respectively,
Scheme 5). Reaction of nerylacetone (18) with triethyl
Scheme 5. Synthesis of (2E,6Z)-Farnesol
Still other trials at different reaction temperature also were
performed in hopes of further improvement of the yield.
During the course of these experiments, the order of addition
for the geranylacetone and formaldehyde also was reversed.
Although in theory this strategy also could lead to farnesol
2, repeated examination of such conditions provided com-
plicated mixtures of unknown products and no apparent
formation of the desired isoprenoid 2.
To apply this methodology to preparation of (2Z,6Z)-
farnesol (3) required starting with ketone 18, nerylacetone
(Scheme 4). Although this ketone is commercially available
phosphonoacetate gave rise to the expected conjugated ester
19 where the C-2 olefin was formed primarily as the (E)-
isomer (10:1, E:Z). Standard reduction of ester 19 with
lithium aluminum hydride under mild conditions generated
alcohol 4 in 55% yield over 2 steps. Similarly, preparation
of commercial (2E,6E)-farnesol (2) also can be achieved via
a parallel route beginning with geranylacetone (13), which
may be an attractive means to obtain this isomer in greater
isomeric purity than the commercial material.16
Scheme 4. Synthesis of Nerylacetone and (2Z,2E)-Farnesol
Identification of the four geometric isomers of farnesol
was realized through use of 13C NMR data in comparison
with literature values.17 The 13C NMR spectra revealed
γ-effect differences for the C-2 olefin substituents of the
major and the minor olefin isomers from all four reaction
sequences. By chance there were striking similarities between
the 13C spectra of (2Z,6E)-farnesol (2) and (2E,6Z)-farnesol
(4), where the two olefins effectively mimic each other and
thus produce nearly identical spectra. However, in each case
the stereochemistry of the C-6 olefin was established by
choice of starting material, and the stereochemistry of the
C-2 olefin could be verified by HMBC experiments.12
Finally, it should be recognized that the methods described
here can be applied to the synthesis of other acyclic
isoprenoids. For example, when commercial (5E,9E)-farne-
sylacetone (20) was employed as the starting material
(Scheme 6), it was possible to prepare the corresponding
(2Z,6E,10E)-geranylgeraniol (21) using this modified Wittig
protocol, and (2E,6E,10E)-geranylgeraniol (22) was available
through the HWE strategy, both through short reaction
sequences in reasonable yields.7b Even though this was not
at considerable cost, it also can be prepared easily from neryl
bromide (16).14 Thus alkylation of ethyl acetoacetate with
bromide 16 followed by successive hydrolysis and decar-
boxylation of the resulting â-keto ester 17 afforded neryl-
acetone (18) in ∼60% yield over three steps. The synthesis
of farnesol 3 then was achieved in a 40% yield via a one-
flask reaction sequence under conditions parallel to those
used for preparation of compound 2. Again, the major
product was the (Z)-isomer, and this was readily separated
from the major byproduct resulting from formation of the
terminal olefin from compound 18.
Synthesis of the two other farnesol isomers, (2E,6Z)-
farnesol (4) and (2E,6E)-farnesol (2), could be based on a
(15) Coates, R. M.; Ley, D. A.; Cavender, P. L. J. Org. Chem. 1978,
43, 4915-4922.
(12) Representative experimental procedures and additional spectral data
are available in Supporting Information.
(13) Schlosser, M.; Jenny, T.; Guggisberg, Y. Synlett 1990, 11, 704.
(14) Kato, T.; Suzuki, M.; Kobayashi, T.; Moore, B. P. J. Org. Chem.
1980, 45, 1126-1130.
(16) Commercial trans,trans-farnesol contained only ∼70% of the desired
isomer via GC analysis.
(17) Burrell, J. W. K.; Garwood, R. F.; Jackman, L. M.; Oskay, E.;
Weedon, B. C. L. J. Chem. Soc. C 1966, 2144-2154.
Org. Lett., Vol. 7, No. 22, 2005
4805

these reactions were completed via the HWE and â-oxido
ylide protocols in reasonable yields (61% and 38%, respec-
tively).
Scheme 6. Synthesis of (2Z,6E,10E)-Geranylgeraniol (21)
through a â-Oxido Ylide and Comparison with the (2E)-Isomer
22
In summary, we have developed a modified Wittig reaction
that exploits the use of â-oxido ylides to allow synthesis of
isoprenoid allylic alcohols with good selectivity for the (Z)-
olefin isomer. This methodology has been used along with
a complementary approach based on a classical HWE
condensation to prepare the four isomers of farnesol, and
this strategy has been shown to be applicable to other
isoprenoids as well. The techniques presented require fewer
steps than previous studies by Gibbs and co-workers7 and
minimize the extent of the chromatographic separations
required. Future studies in our laboratory will be directed to
exploration of the relative activity of isomeric isoprenoids
in various biological systems.
Acknowledgment. Financial support from the Roy J.
Carver Charitable Trust is gratefully acknowledged.
Supporting Information Available: Experimental pro-
cedures; 1H and 13C NMR spectra for compounds 1-4, 18,
21, 22, geraniol and nerol; HMBC spectra for compounds 2
and 4. This material is available free of charge via the Internet
at http://pubs.acs.org.
pursued further, with all four farnesol isomers available it
should be possible to generate all four farnesylacetone
isomers, which ultimately would allow access to all eight
geranylgeraniol isomers. Similarly, commercial prenylac-
etone could be used to prepare both geraniol and nerol, and
OL0513239
4806
Org. Lett., Vol. 7, No. 22, 2005