Cyclizations versus rearrangements in the reactions of some epoxyolefins with Lewis acids

Article abstract of DOI:10.1039/b010101j

Treatment of various substituted epoxyolefins A with B F₃·OEt₂ and other reagents that could be expected to induce carbocyclization to give cyclohexanes was investigated. It turned out that the general reaction of these systems was t

Full text of DOI:10.1039/b010101j

View Article Online / Journal Homepage / Table of Contents for this issue
Cyclizations versus rearrangements in the reactions of some
epoxyolefins with Lewis acids†
Lars Pettersson and Torbjörn Frejd*
1
Division of Organic Chemistry, Centre for Chemistry and Chemical Engineering, Lund
University, P.O.B. 124, S-221 00 Lund, Sweden. E-mail: Torbjorn.Frejd@orgk1.lu.se;
Fax: ϩ46-46-2224119
Received (in Cambridge, UK) 18th December 2000, Accepted 29th January 2001
First published as an Advance Article on the web 2nd March 2001
Treatment of various substituted epoxyolefins A with BF3ؒOEt2 and other reagents that could be expected to induce
carbocyclization to give cyclohexanes was investigated. It turned out that the general reaction of these systems was
the epoxide to ketone rearrangement, while the carbocyclization was only a rare event. Only substrates carrying the
allylsilane grouping underwent carbocyclization and in addition the protecting groups and the stereochemistry of
the system had a decisive influence on whether ring closure or rearrangement were to take place.
Introduction
Electrophilic olefin cyclizations as formalised in Fig. 1(a) have
been studied for over 40 years and have been employed in many
biomimetic syntheses of carbocyclic compounds.1–9 The initiat-
ing electrophilic center may be formed by the action of Lewis
acids on a variety of structural units (starter groups) such as
double bonds, epoxides, alcohols, acetals, etc. and the cycliz-
ation can be terminated by nucleophilic trapping or loss of a
cationic species. In the latter case an olefin is formed. As a
terminating group the 3-silylpropenyl group [allylic silane, Fig. 1
(b) and (c)] is very attractive due to its mild nucleophilicity and
the controlled way in which the silyl moiety is eliminated.10–20
During our previous work with the synthesis of taxol A-ring
building units we noticed that the diastereomeric epoxyallyl-
silanes 1 and 3 gave completely different products on treatment
with BF3ؒOEt2.21,22 While 1 gave the cyclohexane derivative 2
(Scheme 1), compound 3 with the other configuration at the
epoxide unit produced the rearranged product 4 and fluoro-
hydrin 5 (Scheme 1). Related examples of diastereoselective
ring closure of substituted epoxyallylsilane have been
reported.14,23–25
Even though electrophilic carbocyclizations,4–6,26–29 and
Scheme 1
rearrangements of epoxide to carbonyl compounds7,30–33 and
allylic alcohols31,34,35 are well known synthetic transformations,
the examples presented here seem to be the first cases where
tetra-substituted epoxides have been used as starter units for the
electrophile formation, and where electrophilic olefin cycliz-
ation could be expected to compete with epoxide rearrange-
ments. A comparative investigation seemed appropriate and in
this paper we report the reactions of several different substi-
tuted epoxyolefins A (Fig. 2) on treatment with BF3ؒOEt2 and
other reagents that were intended to induce carbocyclization.
The epoxyolefins used (A) can be broadly divided into two
categories; those carrying and those not carrying the allylic
TMS-group i.e. R = TMS and H, respectively.
Results
Fig. 1
Our first attempts at ring closure of the epoxyolefins were made
using the compounds lacking the allylic TMS group. This was
†
13C NMR and elemental analysis data for compounds 6–40 and full
reasonable since there were several such examples in the liter-
ature.29,30 Thus, the bis-TBDMS protected epoxyolefin 6 was
subjected to BF3ؒOEt2 at rt, which, however, resulted in the
experimental details for compounds 42–65 are available as supplemen-
tary data. For direct electronic access see http://www.rsc.org/suppdata/
p1/b0/b010101j/
DOI: 10.1039/b010101j
J. Chem. Soc., Perkin Trans. 1, 2001, 789–800
This journal is © The Royal Society of Chemistry 2001
789

View Article Online
rearrangement was observed but the oxirane ring was opened
by the nucleophilic CH2O moiety, presumably formed during
desilylation of the SEM-group, to give 9. The six-membered
structure of the cyclic acetal 9 was indicated by the coupling
constant of the geminal acetal protons, Jgem = 6.3 Hz. In the
alternative five-membered acetal the corresponding coupling
constant was expected to be ca. 0 Hz.36 The epoxide to allylic
alcohol rearrangement31,34,35 occurred to give the SEM-
deprotected allylic alcohol 10 when 8 was treated with TiCl4 at
0 ЊC. Compound 10 was most likely an intermediate in the
reaction when 8 was adsorbed and heated at 170 ЊC on silica
gel, although 10 now cyclized to give 11 as a 1 : 5 mixture of
diastereomers via an intramolecular Michael addition. A simi-
lar reaction has been described by Bartlett.37 However, the yield
of 11 was difficult to reproduce and an analytically pure sample
could not be obtained.
Anionic nucleophilic cyclization of epoxy ester enolates was
examined as an alternative, since it has been shown that such
reactions may be useful for ring-constructions.31,38 However,
attempted carbocyclization via the enolate ion of 8, generated
by LDA treatment, resulted solely in deconjugation to give the
β,γ-unsaturated ester 12.
As was earlier shown by Fleming et al.10 and extended by
Armstrong and Weiler14,15 ring-closure of epoxyolefins was
greatly enhanced if the olefin part was in the form of an allylic
silane. Indeed, subjected to BF3ؒOEt2 the bis-OTBDMS
protected allylic silane 13 (Scheme 4) cyclized to give 14 as the
Fig. 2
Scheme 2
rearranged product 7 in good yield (Scheme 2). No ring-closed
product could be identified.
In order to see if the migratory aptitude of the oxymethylene
grouping could be reduced in favour of ring closure we instead
tried the trimethylsilylethoxymethyl (SEM) protected 8 as
substrate (Scheme 3). When reacting 8 with BF3ؒOEt2 no
Scheme 4
major reaction, although some rearrangement also took place
to give 15.
As we noticed earlier, the migration was not observed when
the primary TBS protection was absent as for 1. This substrate
gave a quite high yield of the desired cyclohexane derivative 2
(Scheme 1) without any traces of the rearranged product.21,22
We assume that in this case the hydroxy did not react with
BF3ؒOEt2 to give a borate but perhaps only gave a coordin-
ation complex. Thus, the BF3-coordinated HOCH2-group
(BF3ؒHOCH2–) seemed to be less prone to migrate than
TBDMSOCH2–. Migration of the TBDMSOCH2-group has
been observed also in other compounds.39,40
Acidic conditions combined with fluoride ion attack on the
allylic silane was a logical experiment in order to achieve ring
closure. However, in contrast to the facile cyclization of 1 under
BF3ؒOEt2-conditions (Scheme 1) the use of Bu4NFؒ3H2O in the
presence of HOAc in THF at 0 ЊC resulted in the exclusive
formation of the deconjugated ester 16 (Scheme 5). The fluor-
ide ion attack on the TMS-group probably facilitated the
protiodesilylation, but it should be noted that the TBDMS-
group was not removed nor was the epoxide ring cleaved.
Aprotic fluoride ion conditions, which may result in ring
closure via the enolate ion of 16,41 gave only a low yield of a
mixture of the conjugated and the deconjugated trans-silylated
esters 17a,b. The primary TMS-ether was probably formed
Scheme 3
790
J. Chem. Soc., Perkin Trans. 1, 2001, 789–800

View Article Online
Scheme 5
by intra- or intermolecular transfer of the TMS-group from the
starting material. Thus, fluoride ion-induced cyclizations were
not successful with 1.
Since 18,22 carrying an allylic alcohol moiety as an alternative
starter group, was the starting material for 1 it was natural to
test 18 for cyclization (Scheme 6). Being a primary allylic
Scheme 7
to give keto-alcohol 29 in 77% isolated yield. Whether the
rearrangement took place solely on silica gel, or if it was
induced by the Ti-catalyst during epoxidation as well, was not
possible to decide, since the reaction was monitored by silica gel
TLC.
Interestingly, compound 28 having a free alcohol group at
C-4 could not be isolated or even detected on Sharpless
epoxidation of the corresponding allylic alcohol 24. The reac-
tion was slow and resulted in the formation of the tetrahydro-
furan derivative 31 as the only isolable product. This compound
may have been formed via 28, which under the Lewis acidic
reaction conditions (Ti(OiPr)4) gave 31, by attack of O-4 at the
epoxide gem-dimethyl carbon. Similar acid catalyzed ring
closures of hydroxy-epoxides have been developed previously
for the syntheses of a variety of 5- and 6-membered hetero-
cycles present as substructures in natural products.44,45
Since the reaction was very slow it was quenched before
completion, hence the low yield of 31. Titanate formation,
32a/32b (Scheme 8) between the diol precursor 24 and the
Scheme 6
alcohol it would have a tendency to rearrange to the tertiary
allylic alcohol 18a possibly via cation 19 or an equivalent
species. This seemed to be the case since on treatment of 18
with BF3ؒOEt2 a small amount of 18a was detected by NMR
spectroscopy and to judge from TLC analysis the cyclization
was slow and occurred at least partly via 18a to give a 3 : 1
mixture of 20 and 21. Obviously the regiochemistry of the
attack on the electrophilic allylic cation moiety was not domin-
ated by the steric hindrance of the gem-dimethyl group. No
diastereomeric compounds were detected and the configuration
at the α-C was assumed to be the same as in 2.
A rather unexpected influence of the protecting groups at
O-4 was observed for 26 and 27 (Scheme 7). Instead of ring
closure in parallel to 1, only the hydroxymethyl migration took
place to give the keto-alcohols 29 and 30, respectively, in high
yields. In fact, epoxides 26 and 27, prepared via Sharpless
epoxidation42,43 of the corresponding allylic alcohols 22 and 23
using (Ϫ)-DET, could not be obtained pure by chromatography
on silica gel due to this rearrangement. Moreover, according to
NMR analysis of the crude product, a small amount (6%) of an
allylic alcohol with a tentative structure similar to 10 (Scheme
3) was also formed.
An even more rearrangement-prone situation was found
when 22 was epoxidized using (ϩ)-DET, which would lead to
25, the epoxide portion with the opposite configuration as
compared with 26. To judge from TLC analysis, the reaction
mixture contained approximately equal amounts of epoxide
and ketone. However, the epoxide 25 could not be isolated since
it was totally rearranged upon chromatography on silica gel
Scheme 8
Ti-catalyst may have caused the slow reaction already at the
epoxidation stage.46 Alternatively, the formation of the tetra-
hydrofuran derivative may have occurred from 28 in contact
with the silica gel used both in monitoring the reaction (TLC)
and in the isolation of 31. Further details were not investigated.
The compounds having cis-configuration at the α,β-unsatur-
ated ester moiety rearranged to some extent but also underwent
other reactions. However, no carbocyclization was observed.
Thus, compound 33 (Scheme 9) gave, besides the rearrange-
J. Chem. Soc., Perkin Trans. 1, 2001, 789–800
791

View Article Online
structural features as in A (Fig. 2) undergo three major
reactions under BF3-acidic conditions; epoxide rearrangement,
carbocyclization and tetrahydrofuran ring closure.
Epoxide rearrangement by migration of the hydroxymethyl
and TBDMSOCH2-groups seems to be the generally preferred
reaction. The high migratory aptitude of the TBDMSOCH2-
group was demonstrated by Maruoka et al. for rearrangement
of TBDMS-protected epoxyalcohols to TBDMS-protected
β-hydroxyaldehydes using methylaluminium bis(4-bromo-2,6-
di-tert-butylphenoxide).40 However, these authors pointed out
that this rearrangement could not easily be accomplished with
conventional Lewis acids such as BF3ؒOEt2. Our results show
that it obviously can, provided that the substrate has the
appropriate structure. One should note that the substrates used
by Maruoka et al. did not have the oxy-function at C-4, as was
the case in all of our substrates.
Migration of the hydroxymethyl group was not unexpected
but there are cases reported in the literature where hydroxyalkyl
groups do not migrate, but instead, an alkyl group or an aryl
group positioned at the same carbon as the hydroxy group
migrates, in analogy with the pinacol rearrangement.52–54 In the
reaction with BF3ؒOEt2 it seems likely that the epoxide oxygen
of substrate A acts as the Lewis basic centre. Polarisation of the
oxirane O–C bond would then result in the build-up of positive
charge at C-6 and C-7. A five-membered arrangement involving
C-7 as in 1b and 3b (Fig. 2) would therefore be favoured, lead-
ing to oxolanes as has been demonstrated in many similar
cases31,44,55 and also occurred with A when O-4 carried a proton
as leaving group (Scheme 7, 31). The corresponding MeO and
BnO-derivatives 25–27 did not form the corresponding oxolane
derivatives; instead the migration dominated to give 29 and 30,
respectively.
On the other hand, the bulky TBDMSO-group may prevent
the oxygen atom from coming close to the C-7 cationic centre,
which makes it possible for carbocyclization to compete, pos-
sibly via transition state 1a. In addition to this steric effect, the
complexing ability of the oxygen lone-pair electrons in silyl
ethers is reduced by the p-accepting character of the empty
d-orbitals of silicon.56,57 Hence, the observed diminished ability
to stabilise a positive charge by co-ordination for silyloxy-
groups as compared to alkoxy-groups.
Scheme 9
ment product 35, the [4.2.1]bicyclic orthoester 37, the carbon
skeleton of which was not rearranged. The alternative [4.2.2]-
bicyclo compound (epoxide-opening at C-7) was ruled out on
spectroscopic grounds. The coupling constant of the geminal
protons in the smaller ring of 37, Jgem = 7.1 Hz, indicated that
this ring was five-membered.46,47 The 13C NMR-signal of the
quaternary orthoester carbon appeared as a singlet at 122.4
ppm, which was consistent with reported values for similar
structures.47–50 An attempted acetylation using standard
acetylation conditions of 37 failed, which again indicated that
the free hydroxy was tertiary and not primary.
Also in the case of 34, with a methoxy instead of a
TBDMSO-grouping at C-13, rearrangement took place to
give 36, which was formed in a small amount directly during
epoxidation of the precursor allylic alcohol as well. Two com-
pounds of unknown structures were also formed (ca. 25%).
Thus, despite the presence of the allylic TMS-moiety no
carbocyclization was observed for the cis-olefins.
It is also worth mentioning that all ketones isolated were
formed by migration of the hydroxymethyl grouping, and not
by migration of the carbon chain residue (C-6 migration). If we
assume that 1b and 3b are contributing to the stabilisation of
the developing positive charge at C-7, the C-6 migration would
have to be syn-periplanar with respect to the co-ordinating C-4
oxygen. However, the hydroxymethyl moiety can migrate in an
anti-periplanar fashion, which seems more favourable. Interest-
ingly, it was recently shown by computational methods, that the
boron alcoholate moiety resulting in the opening of the epoxide
ring was positioned in the same plane as the carbon frame work
in the BF3-induced hydride migration of methylpropene oxide
to the corresponding aldehyde.58 These results, transferred to
our case, would give a conformation as shown in B (Fig. 2) and
seems reasonable to explain the preferred migration of the
hydroxymethyl group as depicted.
Based on the Stork–Eschenmoser hypothesis for the electro-
philic ring-closure of olefins59,60 and previous discussions of
the transition state for carbocyclization of epoxy-allylsilanes
by Armstrong and Weiler14 it seems necessary to achieve the
correct alignment of the electrophilic centre, the double bond
and the TMS-group as depicted in 1a and 3a (Fig. 2). The C–Si
bond should be perpendicular to the plane of the double bond
and also anti-periplanar to the epoxide C–C bond. When com-
paring 1 with 3 in their chair-like conformations 1a and 3a there
was no obvious reason for their different behaviour, except that
the steric bulk of the TBDMS-group may prevent the allylic
TMS-group from aligning properly for carbocyclization to
occur. This has been discussed by Prestwich et al.23 for a similar
The rearrangement of the “des-TMS“ derivative 38 gave 39,
which lactonized to give the 9-membered lactone 40 after
extended reaction time. Formation of 9-membered lactones is
generally not a feasible process, but in this particular case the
combination of the syn-geometry of the reacting terminals of
39 and the gem-dimethyl effect51 may have contributed to this
lactonization.
Discussion
The results presented show that compounds possessing the
792
J. Chem. Soc., Perkin Trans. 1, 2001, 789–800

View Article Online
substrate and our experimental results are quite in line with
their arguments. However, preliminary molecular mechanics
calculations of the steric energies of a number of geometries
of 1 and 3 do not point strongly in this direction. For example,
1a and 3a had essentially the same steric energies in the
lowest energy conformations hitherto found. Also very similar
energies were found for the ring closed forms 1aЈ and 3aЈ, which
may be taken as gross approximations of possible transition
state models. Moreover, there was no indication that the align-
ment of the C–Si bond would be disturbed in 3a as compared to
the situation for 1a, an almost perfect 90 degree angle between
the C–Si bond and the plane of the double bond was found for
both arrangements. This was also found for the cyclohexene
models 1aЈ and 3aЈ. The computations did indicate that the
chair-like arrangements were of lower energies than the corre-
sponding boat conformations but only by 1–2 kcal molϪ1.
Obviously, more sophisticated and detailed computational
work will be necessary in order to understand the choice of
the competitive pathways of carbocyclization and epoxide
rearrangement for compounds such as A and related deriv-
atives. But at present it seems uncertain that the buttressing
effect on the C–Si alignment of the allylic silane moiety by large
neighbouring groups is a major factor.
The first step in the orthoester formation (37, Scheme 9)
resembles the cyclization of an epoxy allylsilane. One may
speculate that a seven-membered transition state and a
nucleophilic attack by oxygen as in 37a was a consequence of
steric congestion between the carboxylate group and the
hydroxymethyl group as shown in 33 (Scheme 9), making a
6-membered carbocycle formation impossible.
A mechanistic possibility of fluorohydrins being general
intermediates can not be ruled out. Formation of fluorohydrins
in epoxyolefin cyclizations, when BF3ؒOEt2 was used as reagent,
was reported to be favoured in Et2O as opposed to benzene, and
in Et2O further reaction of the fluorohydrins to give ketones
occurred.61 In our experiments CH2Cl2 was routinely used as
solvent so no direct comparisons can be made. However, minor
amounts of fluorohydrins were probably formed in most of our
experiments since they were tentatively identified by TLC and
NMR spectroscopy of crude reaction products. In one case a
fluorohydrin 5 was isolated.
Syntheses of the starting materials
Scheme 10
The common starting material for the allylsilanes was the earl-
ier described doubly protected isopropyl ester 4122 (Scheme 10).
After selective removal of the TBDMS-protection to give 42
the methyl- and benzyl-protecting groups were introduced
by simple alkylation. Thus, methylation (MeI and Ag2O in
DMF)62,63 and benzylation (BnBr, t-BuOK and Bu4NI in
THF)62 of 42 gave 43 and 44, respectively. These compounds
were transformed into Z,E-mixtures of enolphosphates (47a,b
and 48a,b) via enolate trapping with dimethylphosphoric
chloride of the corresponding β-ketoesters 45 and 46, in their
turn obtained by treatment of 43 and 44 with the lithium
enolate of ethyl acetate. The Z- and E-isomers 47 and 48 were
separated by column chromatography. Interestingly the Z : E
ratio was dependent on the protecting group at O-4. While the
methyl ether 45 gave a 45 : 55 Z : E ratio the more bulky benzyl
ether 46 gave 7 : 3 and the even more bulky TBDMS-ether22
gave a 9 : 1 ratio.
In the Ni-catalyzed coupling of 47a and 48a with TMSCH2-
MgCl both compounds gave, in addition to the expected allylic
silanes 50a and 51a, respectively, the corresponding minor
coupling products 50b (8%) and 51b (14%), resulting from
coupling at the allylic THP ether site (Felkin-type coupling64).
Since the desired coupling reaction was favoured by the use of
an excess of the Grignard reagent a compromise was to use 2
equivalents of this reagent, which, still gave some Felkin-
coupling. After separation of the Felkin products by column
chromatography and removal of the THP protection using
PPTS in EtOH, the resulting allylic alcohols 22 and 23 were
epoxidized by the Sharpless asymmetric epoxidation condi-
tions42 using -(Ϫ)-DET as ligand to give the epoxides 26 and
27 in 71 and 77% yield, respectively. Minor amounts of the
rearrangement products, ketones 29 and 30 (Scheme 7) as well
as the allylic alcohols corresponding to 10, were also formed.
These contaminants could not be removed from the main prod-
uct by chromatography on silica gel since the silica gel itself
induced the rearrangements. Thus, both 26 and 27 were used in
their slightly contaminated forms in the BF3ؒOEt2 treatments.
Ni-catalyzed cross-coupling of Me3SiCH2MgCl with the
E-enolphosphates 4922 and 47b (48b was not further trans-
formed) gave lower yields of allylsilanes than the analogous
Z-enolphosphates. Furthermore, Felkin coupling of the
THP-protected allyl alcohol moieties in the E-enolphosphates
was not observed. Isomerization of the double bond of the
α,β-unsaturated ester in the reaction of 49 to give 53 (E : Z 6 : 1
according to NMR analysis) was noticed and for the further
transformation the Z-isomer was removed by chromatography.
Removal of the THP protection of 52 was performed by
using PPTS in EtOH to give 54. The TBDMS-protected deriv-
ative 53 was, however, reacted in propan-2-ol to suppress
solvolysis of the silylether. This gave 55 in 40% yield. Epoxid-
J. Chem. Soc., Perkin Trans. 1, 2001, 789–800
793

View Article Online
ation of 55 provided the epoxy allylsilane 33 in high yield. The
epimeric epoxide was not prepared and the diastereomeric
excess of the epoxidation was not determined with precision
of the presumed silylketone 60: a direct protiodesilylation or,
alternatively, by C–O rearrangement of the α-TMS group to
give the corresponding silylenol ether, which after hydrolytic
work-up or during chromatography resulted in methylketone
61.71
1
but was estimated to be 90–95% by H NMR analysis. The
epoxidation of 54 to give 34 was accompanied by rearrange-
ment to ketone 36 (8%), which could not be separated from 34
by column chromatography. The rearrangement was caused by
the silica gel, or by the reaction conditions in the epoxidation
e.g. Ti(OiPr)4.
The other reaction, the intramolecular acetalization leading
to the bicyclic acetals 63 and 64, took place on silica gel and on
dry neutral alumina. The isolation and purification of 61 there-
fore had to be conducted on wet alumina (10% water). During
NMR analysis it was found that 61 could be quantitatively con-
verted into acetal 63 by standing in dry CDCl3 (most likely
containing small amounts of DCl). Thus, the acetalization did
not proceed via a diol obtained by hydrolysis of the epoxide but
rather directly via a keto-epoxide to acetal conversion.72 The
SEM-protected 59 behaved analogously on treatment with the
TMSCH2Li reagent and gave 62 together with 65. Interestingly,
the acetals 63–65 were diastereomerically pure, but we do not
know the configurations of the bridgehead carbons, i.e. we do
not know which of the two epoxide carbons was attacked by the
carbonyl oxygen.
The synthesis of the des-TMS derivatives 6, 8 and 38 is
outlined in Scheme 11. Epoxidation using the Sharpless AE
The trans-α,β-unsaturated esters 6 and 8 were prepared in
good yields from 61 and 62, respectively, by the reaction with
the sodium salt of (MeO)2P(O)CH2CO2Me in 1,2-dimethoxy-
ethane (DME).73 Reaction of 61 with the lithium-salt of
Me3SiCH2CO2Et74,75 provided the cis-ester 38. Small
amounts of the trans-isomer (detected by TLC) were
probably formed in this reaction, but the purification of 38
was quite easy due to the large difference in Rf values on silica
gel.
The stereochemistry of the double bond at the α,β-unsatur-
1
ated ester was determined by H NMR spectroscopy. A differ-
ence of 0.21 ppm was noticed for 3-CH3 in 38 and 6/8 (1.91 ppm
and 2.12 ppm, respectively) while a much larger effect (a differ-
ence of 1.55 ppm) was noticed for H-4. In 38 (5.88 ppm), this
proton is probably forced to be in close proximity to the
deshielding sector of the ester carbonyl76,77 by the bulky
TBDMSO-group and by the carbon-chain. This method of
determining the stereochemistry of α,β-unsaturated esters was
used for all similar derivatives in this work.
Conclusions
To the best of our knowledge this is the first time a series
of tetra-substituted epoxides have been investigated under
conditions where carbocyclizations of epoxyolefins to give
6-membered rings could be expected. Our experiments showed
that a particularly facile reaction for these kinds of compounds
was the 1,2-rearrangement of the –CH2OH and –CH2-
OTBDMS moieties to give 1,3-hydroxyketones. Other compet-
ing reactions were the epoxide to allylic alcohol rearrangement
and the formation of THF-derivatives. The SEM-protecting
group makes other reaction paths possible and is obviously
Scheme 11
conditions (-(Ϫ)-DET) of the known allylic alcohol 5622 pro-
vided epoxy alcohol 57 (92%, 92% de). Short reaction times
should be used since stirring the reaction mixture at rt overnight
resulted in epoxide-opening and lactonization. After protecting
the epoxy alcohol as the TBDMS-ether65,66 and as the trimethyl-
silylethoxymethyl(SEM)-ether67 to give 58 and 59, respectively,
we originally had in mind to synthesise allylic silanes similar to
1 and therefore sought methods to attach a TMSCH2-group at
C-1 of these compounds. Thus, isopropyl esters 58 and 59 were
treated with an excess of Me3SiCH2Li, formed in situ from
Me3SiCH2SnBu3 and n-BuLi68,69 (lithium–halogen exchange of
Me3SiCH2Cl with Li was not effective), in the presence of
TMEDA. The use of the Li-reagent, in favour of the more
easily prepared Grignard-reagent, was indicated by the
reported examples of bis-addition of the Grignard-reagent to
esters.11 We reasoned that the Li-reagent was probably capable
of deprotonating the expected silylketone 60 at the α-position
adjacent to silicon, thus protecting the carbonyl group from
further addition.70 However, silylketone 60 (R = TBDMS)
was not detected (its presence was only indicated by the
TMS-moiety in 64). Instead, the desilylated ketone 61 together
with the bicyclic acetals 63 and 64 were formed. These
compounds could have been formed via two different reactions
a
rather reactive protecting group under Lewis acidic
conditions.78,79
It could be concluded that the epoxyolefins lacking the allylic
TMS-group were unable to form carbocyclic products. Thus,
the allylsilane moiety was crucial for the epoxyolefin carbo-
cyclization in these cases. However, the structural tolerance is
very narrow and the presence of the allylic silane moiety is not
enough to achieve carbocyclization. A further feature bene-
ficial for the carbocyclization was the low nucleophilicity of the
4-OTBDMS oxygen combined with the large bulk of the
TBDMS-group, which prevented the cation stabilizing effect
of the 4-O coordination, thus allowing a folding of the carbon
chain leading to carbocyclization. Substrates with a more
electron rich 4-O in combination with a smaller oxygen sub-
stituent probably folded in a fashion favouring a five-membered
arrangement, which prevented carbocyclization. In these cases
the 1,2-rearrangement of the –CH2OH and –CH2OTBDMS
moieties or oxolane formation occurred.
794
J. Chem. Soc., Perkin Trans. 1, 2001, 789–800

View Article Online
mg, 0.929 mmol) as starting material. Column chromatography
Experimental
(H–E, 10 : 1) gave 8 (oil, 410 mg, 90%): Rf (1 : 1) 0.68; [α]D20 Ϫ33
(c 0.92, CDCl3); δH 0.00 (s, 3H, –Si(CH3)2–), 0.02 (s, 9H, TMS),
0.08 (s, 3H, –Si(CH3)2–), 0.90 (s, 9H, t-Bu), 0.95 (m, 2H, –CH2–
TMS), 1.35 (s, 6H, >C(CH3)2), 1.84 (dd, 1H, JAB = 14.2 Hz,
J = 9.5 Hz, H-5), 1.94 (dd, 1H, JAB = 14.2 Hz, J = 3.1 Hz,
Column chromatography separations were performed using
Merck SiO2 60 (0.040–0.063 mm) silica gel. TLC analyses were
done on Merck SiO2 60 F254 precoated aluminium sheets and
the spots were visualized by charring with 10% aqueous H2SO4
or by Merck molybdophosphoric acid spray reagent. The
chromatographic eluents were heptane–ethyl acetate (H–E)
mixtures throughout and the ratios are given in parentheses in
this order. Melting points were determined with a Reichert
microscope and are uncorrected. Optical rotations were meas-
ured with a Perkin-Elmer 141 polarimeter. Mass spectra were
recorded on a Finnigan 4021 spectrometer (electron impact
mode at 70 eV). NMR spectra were recorded at 23 ЊC with a
Varian XL-300 spectrometer using CDCl3 as solvent and
CHCl3 as internal standard if not otherwise stated (δ (1H)
7.26 ppm). The following abbreviations are used: br (broad),
s (singlet), d (doublet), t (triplet), q (quartet), sept (septet),
m (multiplet or complex signal). Heptane and EtOAc were
distilled before use. Dry CH2Cl2 and DMF were purified by
distillation and stored over molecular sieves (4 Å). Dry toluene,
Et2O, THF and DME were prepared by drying the commercial
solvents (p.a. grade) over molecular sieves (4 Å). The molecular
sieves were activated at 170 ЊC under vacuum. All other
reagents were used as delivered. Organic extracts were dried
using either Na2SO4 or MgSO4 throughout. 13C NMR data for
all compounds are available as Supplementary material
together with full details of the syntheses of compounds 42–
65.†
H-5), 2.12 (d, 3H, J = 1.6 Hz, –C(CH )᎐CH–), 3.64 (m, 2H,
᎐
3
–CH2CH2–TMS), 3.68 (d, 1H, JAB = 10.4 Hz, –CH2-OSEM),
3.70 (s, 3H, –OCH3), 3.78 (d, 1H, JAB = 10.4 Hz, –CH2-OSEM),
4.32 (dd, 1H, J = 9.5, 3.1 Hz, H-4), 4.70 (s, 2H, –OCH2O–), 5.88
(m, 1H, vinyl) (Found: C, 58.94; H, 9.96. C24H48O6Si2 requires
C, 58.97; H, 9.90%).
(5R,2ЈS,3ЈE)-5-Hydroxy-4,4-dimethyl-5-(4Ј-methoxycarbonyl-
3Ј-methyl-2Ј-tert-butyldimethylsilyloxybut-3Ј-enyl)-1,3-dioxane
(9)
The experiment was performed as described for 14 using 8 (50
mg, 0.10 mmol) as starting material. Column chromatography
(H–E, 10 : 1) gave 9 (oil, 19 mg, 48%): Rf (1 : 1) 0.53; [α]D20 Ϫ41
(c 0.55, CDCl3); δH Ϫ0.03, 0.11 (2s, 6H, –Si(CH3)2–), 0.91 (s,
9H, t-Bu), 1.20, 1.33 (2s, 6H, >C(CH3)2), 1.40 (ddd, 1H,
JAB = 14.4 Hz, J = 8.8 Hz, 1.9 Hz, H-5), 1.53 (dd, 1H, JAB = 14.4
Hz, J = 2.0 Hz, H-5), 2.11 (d, 3H, J = 1.1 Hz, –C(CH )᎐CH–),
᎐
3
3.23 (d, 1H, J = 1.9 Hz, –OH), 3.70 (s, 3H, –OCH3), 3.83 (d, 1H,
JAB = 12.3 Hz, –CH2O–), 4.05 (dd, 1H, JAB = 12.3, 1.0 Hz,
–CH2O–), 4.52 (ddd, 1H, J = 8.8, 2.0, 0.8 Hz, H-4), 4.80 (dd,
1H, JAB = 6.3 Hz, J = 1.0 Hz, acetal), 4.95 (d, 1H, JAB = 6.3 Hz,
acetal), 5.91 (dq, 1H, J = 0.8 Hz, 1.1 Hz, vinyl) (Found: C,
58.68; H, 9.30. C19H36O6Si requires C, 58.73; H, 9.34%).
Methyl (4S,6R,2E)-4-(tert-butyldimethylsilyloxy)-6-[(tert-butyl-
dimethylsilyloxy)methyl]-6,7-epoxy-3,7-dimethyloct-2-enoate
(6)
Methyl (4S,6S,2E)-4-(tert-butyldimethylsilyloxy)-6-hydroxy-6-
hydroxymethyl-3,7-dimethylocta-2,7-dienoate (10)
Trimethyl phosphonoacetate (0.159 mL, 1.1 mmol) was added
to a slurry of NaH (38 mg, ca. 1.0 mmol, ca. 63% in mineral oil)
in dry DME (3.0 mL) under nitrogen. The slurry was stirred for
30 min at rt and then a solution of 61 (159 mg, 0.382 mmol) in
dry DME (1.0 mL) was added. The temperature was raised to
90 ЊC and stirring was continued for 3 h. After cooling, ether
(100 mL) was added and the solution was washed with water,
dried and the solvent was evaporated at reduced pressure.
Column chromatography (H–E, 50 : 1) gave 6 (oil, 164 mg,
91%): Rf (10 : 1) 0.33; [α]D20 Ϫ25 (c 1.23, CDCl3); δH Ϫ0.01, 0.06,
0.07, 0.08 (4s, 12H, –Si(CH3)2–), 0.90, 0.91 (2s, 18H, t-Bu), 1.31,
1.35 (2s, 6H, >C(CH3)2), 1.77 (dd, 1H, JAB = 14.2 Hz, J = 10.0
Hz, H-5), 2.02 (dd, 1H, JAB = 14.2 Hz, J = 3.0 Hz, H-5), 2.12 (d,
TiCl4 (0.102 mL, 0.10 mmol, 1.0 M in CH2Cl2) was added to a
solution of 8 (50 mg, 0.10 mmol) in dry CH2Cl2 (0.5 mL) under
nitrogen at 0 ЊC. The reaction mixture was stirred for 15 min
and NaHCO3 (sat., 0.5 mL) was then added followed by tartaric
acid (10% aq., 0.5 mL) and ether (10 mL). The organic phase
was washed with water, dried and the solvent was evaporated at
reduced pressure to give crude 10 which was slightly decom-
posed when standing at rt overnight. Column chromatography
(H–E, 3 : 1) gave 10 (oil, 14 mg, 38%): Rf (1 : 1) 0.43; [α]D20 Ϫ5
(c 0.35, CDCl3); δH Ϫ0.01, 0.08 (2s, 6H, –Si(CH3)2–), 0.92 (s,
9H, t-Bu), 1.69 (dd, 1H, JAB = 14.9 Hz, J = 1.7 Hz, H-5), 1.75
(dd, 3H, J = 1.7, 0.7 Hz, H-8), 2.07 (dd, 1H, JAB = 14.9 Hz,
J = 10.7 Hz, H-5), 2.13 (d, 3H, J = 1.1 Hz, –C(CH )᎐CH–), 3.43
᎐
3
3H, J = 1.6 Hz, –C(CH )᎐CH–), 3.70 (s, 3H, –OCH ), 3.78, 3.80
᎐
3
(2d, 2H, JAB = 11.1 Hz, –CH2O–), 4.33 (dd, 1H, J =310.0 Hz, 3.0
Hz, H-4), 5.87 (m, 1H, vinyl) (Found: C, 61.13; H, 10.31.
C24H48O5Si2 requires C, 60.97; H, 10.23%).
(s, 2H, –CH2OH), 3.71 (s, 3H, –OCH3), 4.32 (ddd, 1H, J = 1.7,
10.7, 0.7 Hz, H-4), 4.38 (s, 1H, –OH), 5.11 (dq, 1H, J = 2.0, 1.7
Hz, ᎐CH ), 5.25 (dq, 1H, J = 2.0, 0.7 Hz, ᎐CH ), 5.82 (dq, 1H,
᎐
᎐
2
2
J = 0.7, 1.1 Hz, H-2) (Found: C, 60.39; H, 9.54. C18H34O5Si
requires C, 60.30; H, 9.56%).
Methyl (4S,2E)-4,8-bis(tert-butyldimethylsilyloxy)-3,7,7-tri-
methyl-6-oxooct-2-enoate (7)
Methyl 2-[2-methyl-3-tert-butyldimethylsilyloxy-5-(propen-2-
yl)-5-(trimethylsilylethoxymethoxymethyl)tetrahydrofuran-2-
yl]acetate (11)
The experiment was performed as described for 14 using 6 (116
mg, 0.245 mmol) as starting material. Column chromatography
(H–E, 50 : 1) gave 7 (oil, 90 mg, 78%): Rf (10 : 1) 0.42; [α]D20 Ϫ32
(c 1.16, CDCl3); δH 0.01 (s, 3H, –Si(CH3)2–), 0.02 (s, 6H,
–Si(CH3)2–), 0.03 (s, 3H, –Si(CH3)2–), 0.85, 0.85 (2s, 18H, t-Bu),
Silica gel (Merck SiO2 60, 660 mg) was added to a solution of 8
(22 mg) in ether (ca. 2 mL). The solvent was removed at reduced
pressure and the residue was then heated under nitrogen at
170 ЊC for 45 min. After cooling, EtOAc (5 mL) was added and
the slurry was stirred for 2 min. Filtration and evaporation of
the solvent at reduced pressure was followed by column
chromatography (H–E, 20 : 1) to give 11 (oil, 9 mg, 41%) as a
diastereomeric mixture (1 : 5). The reaction gave variable yields
on repeated experiments and analytically pure samples could
not be obtained.
1.06 (s, 6H, –C(CH ) –), 2.08 (d, 3H, J = 1.3 Hz, –C(CH )᎐
᎐
3
3
2
CH–), 2.43 (dd, 1H, JAB = 17.3 Hz, J = 3.1 Hz, H-5), 2.90 (dd,
1H, JAB = 17.3 Hz, J = 8.3 Hz, H-5), 3.50, 3.59 (2d, 2H,
JAB = 9.9 Hz, –CH2O–), 3.70 (s, 3H, –OCH3), 4.68 (ddd, 1H,
J = 3.1 Hz, 8.3 Hz, 0.9 Hz, H-4), 5.96 (m, 1H, vinyl) (Found: C,
61.29; H, 10.29. C24H48O5Si2 requires C, 60.97; H, 10.23%).
Methyl (4S,6R,2E)-4-(tert-butyldimethylsilyloxy)-6,7-epoxy-
3,7-dimethyl-6-[(2-trimethylsilylethoxy)methoxymethyl]oct-2-
enoate (8)
For 11a (major isomer): δH 0.01 (s, 9H, TMS), 0.06, 0.06
(2s, 6H, –Si(CH3)2–), 0.87 (s, 9H, t-Bu), 0.92 (m, 2H, –CH2–
TMS), 1.22 (s, 3H, >C(CH3)O–), 1.75 (d, 3H, J = 0.8 Hz,
The experiment was performed as described for 6 using 62 (402
–C(CH )᎐CH ), 1.95, 2.77 (2dd, 2H, JAB = 12.9 Hz, J = 6.6 Hz,
᎐
3
2
J. Chem. Soc., Perkin Trans. 1, 2001, 789–800
795

View Article Online
–CH2–), 2.52, 2.58 (2d, 2H, JAB = 13.9 Hz, –CH2-CO2–), 3.41,
3.43 (2d, 2H, JAB = 10.5 Hz, –CH2-OSEM), 3.62 (m, 2H,
–CH2CH2–TMS), 3.65 (s, 3H, –OCH3), 4.48 (t, 1H, J = 6.6 Hz,
–CH(OTBDMS)–), 4.66 (s, 2H, –OCH2O–), 4.84, 5.09 (2m, 2H,
min and NaHCO3 (sat., 5 mL) was then added followed by
ether (25 mL). The organic phase was washed with brine and
dried (MgSO4). Evaporation of the solvent at reduced pressure
and column chromatography (H–E, 120 : 1 then 60 : 1) gave 14
(oil, 73 mg, 62%) and 15 (oil, 23 mg, 17%).
᎐CH ).
᎐
2
For 11b (minor isomer): δH 0.0 (s, 9H, TMS), 0.04, 0.05 (2s,
6H, –Si(CH3)2–), 0.86 (s, 9H, t-Bu), 0.90 (m, 2H, –CH2–TMS),
1.43 (s, 3H, >C(CH )O–), 1.76 (s, 3H, –C(CH )᎐CH ), 1.99 (dd,
For 14: Rf (40 : 1) 0.11; [α]D20 ϩ48 (c 1.36, CDCl3); δH 0.06, 0.06,
0.09, 0.10 (4s, 12H, –Si(CH3)2–), 0.89, 0.92 (2s, 18H, t-Bu), 1.00,
1.14 (2s, 6H, >C(CH3)2), 1.26 (t, 3H, J = 7.0 Hz, Et), 1.70 (dd,
1H, JAB = 13.2 Hz, J = 10.8 Hz, H-4), 1.94 (dd, 1H, JAB = 13.2
Hz, J = 5.4 Hz, H-4), 3.08 (s, 1H, H-1), 3.52, 3.55 (2d, 2H,
JAB = 10.0 Hz, –CH2-OTBDMS), 4.13, 4.15 (2dq, 2H,
JAB = 10.9 Hz, J = 7.0 Hz, Et), 4.83 (dddd, 1H, J = 10.8, 5.4, 2.0,
1.7 Hz, H-5), 4.90 (dd, 1H, J = 2.0 Hz, 1.7 Hz, vinyl), 5.22 (dd,
1H, J = 2.0 Hz, 2.0 Hz, vinyl) (Found: C, 61.63; H, 10.39.
C25H50O5Si2 requires C, 61.68; H, 10.35%).
᎐
3
3
2
2H, JAB = 12.7 Hz, J = 6.8 Hz, –CH2–), 2.30 (dd, 2H, JAB
=
12.7 Hz, J = 6.1 Hz, –CH2–), 2.52, 2.64 (2d, 2H, JAB = 15.6
Hz, –CH2-CO2–), 3.43 (s, 2H, –CH2-OSEM), 3.61 (m, 2H,
–CH2CH2–TMS), 3.64 (s, 3H, –OCH3), 4.20 (dd, 1H, J = 6.1,
6.8 Hz, –CH(OTBDMS)–), 4.68 (s, 2H, –OCH2O–), 4.85, 5.10
(2m, 2H, ᎐CH ).
᎐
2
For 15: Rf (40 : 1) 0.23; [α]D20 ϩ47 (c 0.30, CDCl3); δH 0.02
(s, 6H, –Si(CH3)2–), 0.04, 0.06 (2s, 6H, –Si(CH3)2–), 0.08 (s, 9H,
TMS), 0.86, 0.87 (2s, 18H, t-Bu), 1.08 (s, 6H, –C(CH3)2–),
1.27 (t, 3H, J = 7.1 Hz, Et), 1.41 (d, 1H, JAB = 11.3 Hz,
–C(H)H-TMS), 2.57 (dd, 1H, JAB = 17.8 Hz, J = 2.2 Hz, H-5),
2.83 (dd, 1H, JAB = 17.8 Hz, J = 8.4 Hz, H-5), 2.97 (dd, 1H,
JAB = 11.3 Hz, J = 1.0 Hz, –C(H)H-TMS), 3.51, 3.59 (2d, 2H,
JAB = 10.1 Hz, H-8), 4.13 (q, 2H, J = 7.1 Hz, Et), 4.66 (ddd, 1H,
J = 8.4, 2.2, 1.0 Hz, H-4), 5.95 (dd, 1H, J = 1.0, 1.0 Hz, vinyl)
(Found: C, 60.23; H, 10.50. C28H58O5Si3 requires C, 60.16; H,
10.46%).
Methyl (4S,6R)-4-(tert-butyldimethylsilyloxy)-6,7-epoxy-7-
methyl-3-methylene-6-[(2-trimethylsilylethoxy)methoxymethyl]-
octanoate (12)
n-BuLi (0.100 mL, 0.15 mmol, 1.5 M in hexane) was added to
a solution of diisopropylamine (0.028 mL, 0.20 mmol) in dry
THF (0.2 mL) under nitrogen at 0 ЊC. After stirring for 10 min,
a solution of 8 (50 mg, 0.10 mmol) in dry THF (0.1 mL) was
added. After another 10 min stirring at 0 ЊC, ether (10 mL) and
HCl (1.0 mL, 1.0 M) was added and the organic phase was
washed with water, dried and the solvent was evaporated at
reduced pressure. Column chromatography (H–E, 10 : 1) gave
12 (oil, 32 mg, 64%): Rf (1 : 1) 0.65; [α]D20 Ϫ18 (c 0.52, CDCl3);
δH 0.02 (s, 12H, TMS, –Si(CH3)2–), 0.08 (s, 3H, –Si(CH3)2–),
0.89 (s, 9H, t-Bu), 0.95 (m, 2H, –CH2–TMS), 1.35, 1.35 (2s, 6H,
>C(CH3)2), 1.85 (dd, 1H, JAB = 14.4 Hz, J = 9.2 Hz, H-5), 1.94
(dd, 1H, JAB = 14.4 Hz, J = 3.7 Hz, H-5), 3.08 (s, 2H, H-2),
3.56–3.71 (m, 2H, –CH2CH2–TMS), 3.67 (d, 1H, JAB = 10.9 Hz,
–CH2-OSEM), 3.68 (s, 3H, –OCH3), 3.77 (d, 1H, JAB = 10.9 Hz,
–CH2-OSEM), 4.42 (dd, 1H, J = 9.2, 3.7 Hz, H-4), 4.70 (s, 2H,
–OCH2O–), 5.01 (m, 1H, vinyl), 5.18 (s, 1H, vinyl) (Found: C,
59.00; H, 9.96. C24H48O6Si2 requires C, 58.97; H, 9.90%).
Ethyl (4S,6R)-4-(tert-butyldimethylsilyloxy)-6,7-epoxy-6-
hydroxymethyl-7-methyl-3-methyleneoctanoate (16)
A solution of Bu4NFؒ(H2O)3 (95 mg, 0.30 mmol) in dry THF
(0.3 mL) was added to a solution of 1 (131 mg, 0.295 mmol)
and HOAc (0.050 mL, 0.87 mmol) in dry THF (1.5 mL) under
nitrogen at 0 ЊC. After stirring for 2 h, the reaction mixture was
diluted with ether (20 mL) and washed with NaHCO3 (sat.) and
brine. The organic phase was dried and the solvent was evapor-
ated at reduced pressure to give crude 16 (109 mg, 99%) which
was pure on TLC (H–E, 3 : 1). Column chromatography (H–E,
5 : 1) gave 16 (oil, 87 mg, 79%): Rf (3 : 1) 0.13; [α]D20 Ϫ9 (c 1.41,
CDCl3); δH 0.05, 0.12 (2s, 6H, –Si(CH3)2–), 0.91 (s, 9H, t-Bu),
1.26 (t, 3H, J = 7.1 Hz, Et), 1.33 (s, 6H, >C(CH3)2), 1.94 (dd,
1H, JAB = 14.6 Hz, J = 4.0 Hz, H-5), 1.99 (dd, 1H, JAB = 14.6
Hz, J = 8.9 Hz, H-5), 3.07 (s, 2H, H-2), 3.67 (d, 1H, JAB = 12.2
Hz, broad, –C(H)HOH), 3.80 (dd, 1H, JAB = 12.2 Hz, J = 5.6
Hz, broad, –C(H)HOH), 4.15 (q, 2H, J = 7.1 Hz, Et), 4.39 (dd,
1H, J = 4.0 Hz, 8.9 Hz, H-4), 5.05 (d, 1H, J = 1.0 Hz, vinyl),
5.21 (s, 1H, vinyl) (Found: C, 61.19; H, 9.75. C19H36O5Si
requires C, 61.25; H, 9.74%).
Ethyl (4S,6R,2Z)-4-(tert-butyldimethylsilyloxy)-6-[tert-butyl-
dimethylsilyloxymethyl]-6,7-epoxy-7-methyl-3-[(trimethylsilyl)-
methyl]oct-2-enoate (13)
Compound 122 (133 mg, 0.300 mmol) was added to a solution
of TBDMSCl (54 mg, 0.360 mmol) and imidazole (51 mg, 0.75
mmol) in DMF (1 mL). The reaction mixture was stirred at rt
for 1 h and then CH2Cl2 (5 mL) was added. The solution was
washed with water, 1.0 M HCl, NaHCO3 (sat.), again with
water and dried. The solvent was evaporated at reduced
pressure followed by column chromatography of the residue
(H–E, 10 : 1) to give 13 (oil, 163 mg, 98%): Rf (3 : 1) 0.64; [α]D20
ϩ53 (c 0.96, CDCl3); δH 0.03 (s, 3H, –Si(CH3)2–), 0.08 (s, 6H,
–Si(CH3)2–), 0.09 (s, 12H, –Si(CH3)2–, TMS), 0.91, 0.93 (2s,
18H, t-Bu), 1.27 (t, 3H, J = 7.1 Hz, Et), 1.32, 1.35 (2s, 6H,
>C(CH3)2), 1.67 (dd, 1H, JAB = 14.2 Hz, J = 10.0 Hz, H-5),
1.75 (d, 1H, JAB = 11.7 Hz, –C(H)H-TMS), 2.17 (dd, 1H,
JAB = 14.2 Hz, J = 2.2 Hz, H-5), 2.89 (dd, 1H, JAB = 11.7 Hz,
J = 1.0 Hz, –C(H)H-TMS), 3.82 (s, 2H, –CH2-OTBDMS),
4.11, 4.14 (2dq, 2H, JAB = 10.6 Hz, J = 7.1 Hz, Et), 4.28 (ddd,
1H, J = 10.0, 2.2, 1.0 Hz, H-4), 5.85 (dd, 1H, J = 1.0 Hz, 1.0
Hz, vinyl) (Found: C, 60.20; H, 10.44. C28H58O5Si3 requires C,
60.16; H, 10.46%).
Ethyl (2E,4S,6R)-4-(tert-butyldimethylsilyloxy)-6,7-epoxy-3,7-
dimethyl-6-[(trimethylsilyloxy)methyl]oct-2-enoate (17a) and
ethyl (4S,6R)-4-(tert-butyldimethylsilyloxy)-6,7-epoxy-7-
methyl-3-methylene-6-[(trimethylsilyloxy)methyl]octanoate
(17b)
Anhydrous Bu4NF (1.0 mL, 0.21 mmol, 0.21 M in dry THF)
was added to a solution of 1 (93 mg, 0.21 mmol) in dry THF
(4.0 mL) under nitrogen at 0 ЊC. TLC analysis (H–E, 3 : 1) of
the reaction mixture directly after the addition showed mostly
nonpolar products. After stirring for 15 min at the same tem-
perature, ether (20 mL) was added and the solution was washed
with water and dried (MgSO4). The solvent was evaporated at
reduced pressure to give 97 mg of crude product. Column
chromatography (H–E, 50 : 1 then 20 : 1) gave 17a (oil, 8 mg,
9%) containing ca. 7% of 17b, and 17b (oil, 6 mg, 7%) contain-
ing ca. 15% of 17a. The low yields might be due to the instabil-
ity of TMS-ethers on silica gel. The elemental analysis was
carried out on a mixture of 17a and 17b.
Ethyl (1R,3S,5S)-5-(tert-butyldimethylsilyloxy)-3-[(tert-butyl-
dimethylsilyloxy)methyl]-3-hydroxy-2,2-dimethyl-6-methylene-
cyclohexanecarboxylate (14) and ethyl (4S,2Z)-4,8-bis(tert-
butyldimethylsilyloxy)-7,7-dimethyl-6-oxo-3-[(trimethylsilyl)-
methyl]oct-2-enoate (15)
BF3ؒOEt2 (0.243 mL, 0.243 mmol, 1.0 M in CH2Cl2) was added
to a solution of 13 (136 mg, 0.243 mmol) in dry CH2Cl2 (12 mL)
under nitrogen at 0 ЊC. The reaction mixture was stirred for 10
For 17a: Rf (10 : 1) 0.31; δH 0.00, 0.06 (2s, 6H, –Si(CH3)2–),
0.14 (s, 9H, TMS), 0.90 (s, 9H, t-Bu), 1.28 (t, 3H, J = 7.1 Hz,
Et), 1.32, 1.35 (2s, 6H, >C(CH3)2), 1.76 (dd, 1H, JAB = 14.2 Hz,
796
J. Chem. Soc., Perkin Trans. 1, 2001, 789–800

View Article Online
J = 9.8 Hz, H-5), 2.02 (dd, 1H, JAB = 14.2 Hz, J = 3.1 Hz, H-5),
JAB = 11.5 Hz, –C(H)H-TMS), 1.74 (s, 3H, ᎐C(CH )CH ),
᎐
3 3
2.12 (d, 3H, J = 1.3 Hz, –C(CH )᎐CH–), 3.75, 3.78 (2d, 2H,
2.24 (dd, 1H, JAB = 14.5 Hz, J = 9.9 Hz, H-5), 2.65 (d, 1H,
JAB = 14.5 Hz, broad, H-5), 3.02 (s, 1H, broad, –OH), 3.06 (dd,
1H, JAB = 11.5 Hz, J = 1.0 Hz, –C(H)H-TMS), 3.82 (ddd, 1H,
J = 9.9, 2.3, 0.7 Hz, H-4), 3.99 (d, 1H, JAB = 12.1 Hz, –C(H)-
HOH), 4.16 (q, 2H, J = 7.1 Hz, Et), 4.24 (d, 1H, JAB = 11.7 Hz,
benzyl), 4.28 (d, 1H, JAB = 12.1 Hz, –C(H)HOH), 4.64 (d, 1H,
JAB = 11.7 Hz, benzyl), 5.95 (s, 1H, vinyl), 7.25–7.38 (m, 5H,
–Ph) (Found: C, 68.30; H, 8.84. C23H36O4Si requires C, 68.27;
H, 8.97%).
᎐
3
JAB = 11.2 Hz, –CH2-OTMS), 4.16 (q, 2H, J = 7.1 Hz, Et), 4.33
(dd, 1H, J = 9.8, 3.1 Hz, H-4), 5.86 (m, 1H, vinyl).
For 17b: Rf (10 : 1) 0.26; δH 0.03, 0.07 (2s, 6H, –Si(CH3)2–),
0.13 (s, 9H, TMS), 0.90 (s, 9H, t-Bu), 1.26 (t, 3H, J = 7.1 Hz,
Et), 1.32, 1.35 (2s, 6H, >C(CH3)2), 1.79 (dd, 1H, JAB = 14.2 Hz,
J = 9.6 Hz, H-5), 1.99 (dd, 1H, JAB = 14.2 Hz, J = 3.6 Hz, H-5),
3.06 (s, 2H, H-2), 3.72, 3.78 (2d, 2H, JAB = 11.0 Hz, –CH2-
OTMS), 4.14 (q, 2H, J = 7.1 Hz, Et), 4.44 (dd, 1H, J = 9.6, 3.6
Hz, H-4), 5.01 (m, 1H, vinyl), 5.17 (s, 1H, broad, vinyl) (Found:
C, 59.50; H, 10.10. C24H44O5Si2 requires (mixture of 17a and
17b) C, 59.41; H, 9.97%).
Ethyl (2Z,4S,6R)-6,7-epoxy-6-hydroxymethyl-4-methoxy-7-
methyl-3-[(trimethylsilyl)methyl]oct-2-enoate (26)
-(Ϫ)-Diethyl tartrate (20.6 mg, 0.100 mmol) and Ti(OiPr)4
(0.025 mL, 0.084 mmol) were added to a mixture of 22 (100 mg,
0.305 mmol) and powdered molecular sieves (50 mg, 4 Å, acti-
vated at 170 ЊC in vacuo) in dry CH2Cl2 (2 mL) under nitro-
gen at Ϫ15 ЊC. The mixture was stirred for 10–15 min at Ϫ15 ЊC
and the temperature was then lowered to Ϫ40 ЊC. At this tem-
perature tert-butyl hydroperoxide (anhydrous, 0.16 mL, 0.48
mmol, 3.0 M in toluene) was slowly added. After 30 min stirring
at this temperature a solution of FeSO4 (150 mg) and tartaric
acid (70 mg) in water (1 mL) was added followed by ether
(5 mL). Stirring was continued at rt for 20 min and the organic
phase was then washed with brine, dried and the solvent was
evaporated at reduced pressure. Column chromatography
(H–E, 5 : 1) gave 26 (oil, 103 mg, 98% recovery) contaminated
by the epoxide rearranged allylic alcohol (4%) and by ketone
29 (23%) according to 1H NMR analysis.
Ethyl (1S,3S)-3-(tert-butyldimethylsilyloxy)-6,6-dimethyl-2,5-
dimethylenecyclohexanecarboxylate (20) and ethyl (1S,3S)-3-
(tert-butyldimethylsilyloxy)-2-methylene-5-isopropylidenecyclo-
hexanecarboxylate (21)
The experiment was performed as described for 14 using 1822
(107 mg, 0.250 mmol) as starting material. The reaction mix-
ture was stirred at rt for 3 h. Column chromatography (H–E,
150 : 1) gave a mixture of 20 and 21 (3 : 1, oil, 62 mg, 73%) and
18a (<2 mg). The proposal of structure 18a was based on its
NMR data: δH 1.30, 1.38 (2s, 6H, –C(CH3)2OH), 4.81, 4.89 (2m,
2H, ᎐CH ); δ 27.46, 28.79, 82.50 (–C(CH3)2OH), 103.96,
᎐
2
C
155.18 (>C᎐CH ).
᎐
2
For 20: Rf (3 : 1) 0.64; [α]D20 ϩ78 (c 0.86, CDCl3); δH 0.08 (s, 6H,
–Si(CH3)2–), 0.92 (s, 9H, t-Bu), 1.09, 1.14 (2s, 6H, >C(CH3)2),
1.23 (t, 3H, J = 7.0 Hz, Et), 2.33 (dd, 1H, JAB = 12.9 Hz,
J = 10.8 Hz, H-4), 2.53 (dddd, 1H, JAB = 12.9 Hz, J = 5.9, 1.7,
1.7 Hz, H-4), 3.14 (s, 1H, H-1), 4.06, 4.10 (2dq, 2H, JAB = 10.7
Hz, J = 7.0 Hz, Et), 4.67 (dddd, 1H, J = 10.8, 5.9, 2.0, 2.0 Hz,
H-3), 4.76, 4.84, 4.96, 5.26 (4dd, 4H, J = 2 Hz, 2 Hz, vinyl).
For 21: δH 0.10 (s, 6H, –Si(CH3)2–), 0.93 (s, 9H, t-Bu), 1.24 (t,
For 26 (contaminated): Rf (1 : 1) 0.40; [α]D20 ϩ77 (c 1.25,
CDCl3); δH 0.09 (s, 9H, TMS), 1.28 (t, 3H, J = 7.1 Hz, Et), 1.32,
1.36 (2s, 6H, >C(CH3)2), 1.54 (d, 1H, JAB = 11.5 Hz, –C(H)H-
TMS), 1.91 (dd, 1H, JAB = 15.1 Hz, J = 2.9 Hz, H-5), 1.98 (dd,
1H, JAB = 15.1 Hz, J = 9.0 Hz, H-5), 2.97 (dd, 1H, JAB = 11.5
Hz, J = 1.0 Hz, –C(H)H-TMS), 3.34 (s, 3H, –OCH3), 3.67–3.76
(m, 3H, –CH2OH, H-4), 4.14 (q, 2H, J = 7.1 Hz, Et), 5.85
(s, 1H, broad, vinyl) (Found: C, 59.23; H, 9.41. C17H32O5Si
requires C, 59.27; H, 9.36%).
3H, J = 7.0 Hz, Et), 1.70 (s, 6H, broad, ᎐C(CH ) ), 1.85–1.95,
᎐
3
2
2.06–2.15 (2m, 2H, H-4), 2.76 (dd, 1H, JAB = 13.5 Hz, J = 5.1
Hz, H-6), 2.99 (dd, 1H, JAB = 13.5 Hz, J = 4.1 Hz, H-6), 3.45
(dd, 1H, J = 5.1 Hz, 4.1 Hz, H-1), 4.02–4.20 (m, 2H, Et), 4.29
(dddd, 1H, J = 11, 4.9, 2, 2 Hz, H-3), 4.86, 5.19 (2dd, 2H, J = 2,
2 Hz, vinyl) (Found: C, 67.45; H, 10.16. C19H34O3Si (mixture of
20 and 21) requires C, 67.41; H, 10.12%).
Ethyl (2Z,4S,6R)-4-benzyloxy-6,7-epoxy-6-hydroxymethyl-7-
methyl-3-[(trimethylsilyl)methyl]oct-2-enoate (27)
The experiment was performed as described for 26 using 23 (96
mg, 0.237 mmol) as starting material. Column chromatography
(H–E, 6 : 1) gave 27 (93 mg, 93%) as an oil which crystallized
on standing. This product was contaminated by the epoxide
rearranged allylic alcohol (6%) and by ketone 30 (10%).
For 27 (contaminated): Rf (3 : 1) 0.23; mp 53–70 ЊC; [α]D20 ϩ69
(c 0.45, CDCl3); δH 0.12 (s, 9H, TMS), 1.30 (t, 3H, J = 7.1 Hz,
Et), 1.33, 1.34 (2s, 6H, >C(CH3)2), 1.60 (d, 1H, JAB = 11.5 Hz,
–C(H)H-TMS), 1.96 (dd, 1H, JAB = 14.9 Hz, J = 9.5 Hz, H-5),
2.04 (dd, 1H, JAB = 14.9 Hz, J = 2.7 Hz, H-5), 2.71 (t, 1H,
J = 6.2 Hz, broad, –OH), 3.03 (dd, 1H, JAB = 11.5 Hz, J = 1.0
Hz, –C(H)H-TMS), 3.66 (d, 2H, J = 6.2 Hz, –CH2OH), 3.99
(dd, 1H, J = 9.5, 2.7 Hz, H-4), 4.16 (q, 2H, J = 7.1 Hz, Et), 4.31,
4.59 (2d, 2H, JAB = 10.7 Hz, benzyl), 5.99 (s, 1H, vinyl), 7.25–
7.42 (m, 5H, –Ph) (Found: C, 65.60; H, 8.72. C23H36O5Si
requires C, 65.68; H, 8.63%).
Ethyl (2Z,4S)-6-hydroxymethyl-4-methoxy-7-methyl-3-
[(trimethylsilyl)methyl]octa-2,6-dienoate (22)
Pyridinium tosylate (19 mg, 0.075 mmol) was added to 50a (297
mg, 0.720 mmol) dissolved in EtOH (absolute, 6.0 mL). The
reaction mixture was stirred at 50 ЊC for 24 h and after cooling,
ether (ca. 100 mL) was added. The solution was washed with
NaHCO3 (sat.) and brine, dried and the solvent was evaporated
at reduced pressure. Column chromatography (H–E, 10 : 1) gave
22 (oil, 219 mg, 93%): Rf (3 : 1) 0.31; [α]D20 ϩ108 (c 1.10, CDCl3);
δH 0.10 (s, 9H, TMS), 1.29 (t, 3H, J = 7.0 Hz, Et), 1.65 (d, 1H,
JAB = 11.5 Hz, –C(H)H-TMS), 1.72, 1.77 (2s, 6H, ᎐C(CH ) ),
᎐
3
2
2.13 (dd, 1H, JAB = 14.7 Hz, J = 9.6 Hz, H-5), 2.65 (dd, 1H,
JAB = 14.7 Hz, broad, H-5), 3.05 (dd, 1H, JAB = 11.5 Hz, J = 1.0
Hz, –C(H)H-TMS), 3.28 (s, 3H, –OCH3), 3.57 (dd, 1H, J = 9.6,
2.5 Hz, H-4), 3.89 (d, 1H, JAB = 12.1 Hz, –C(H)HOH), 4.14
(q, 2H, J = 7.0 Hz, Et), 4.29 (d, 1H, JAB = 12.1 Hz, –C(H)-
HOH), 5.81 (s, 1H, broad, vinyl) (Found: C, 62.14; H, 9.85.
C17H32O4Si requires C, 62.20; H, 9.82%).
Ethyl (2Z,4S)-8-hydroxy-4-methoxy-7,7-dimethyl-6-oxo-
3-[(trimethylsilyl)methyl]oct-2-enoate (29)
The experiment was performed as described for 14 using 26 (60
mg, 0.175 mmol) as starting material. Column chromatography
(H–E, 6 : 1) gave 29 (oil, 55 mg, 92%): Rf (1 : 1) 0.45; [α]D20 ϩ50
(c 0.72, CDCl3); δH 0.11 (s, 9H, TMS), 1.11, 1.14 (2s, 6H,
–C(CH3)2–), 1.29 (t, 3H, J = 7.0 Hz, Et), 1.53 (d, 1H, JAB = 11.4
Hz, –C(H)H-TMS), 2.35 (dd, 1H, JAB = 16.4 Hz, J = 2.2
Hz, H-5), 2.89 (dd, 1H, JAB = 16.4 Hz, J = 9.7 Hz, H-5), 2.98
(dd, 1H, JAB = 11.4 Hz, J = 0.9 Hz, –C(H)H-TMS), 3.25 (s,
Ethyl (2Z,4S)-4-benzyloxy-6-hydroxymethyl-7-methyl-3-[(tri-
methylsilyl)methyl]octa-2,6-dienoate (23)
The experiment was performed as described for 22 using 51a
(150 mg, 0.307 mmol) as starting material. Column chrom-
atography (H–E, 10 : 1) gave 23 (oil, 104 mg, 84%): Rf (3 : 1)
0.33; [α]D20 ϩ18 (c 1.20, CDCl3); δH 0.09 (s, 9H, TMS), 1.30 (t,
3H, J = 7.1 Hz, Et), 1.62 (s, 3H, ᎐C(CH )CH ), 1.69 (d, 1H,
᎐
3
3
J. Chem. Soc., Perkin Trans. 1, 2001, 789–800
797

View Article Online
3H, –OCH3), 3.48, 3.71 (2d, 2H, JAB = 11.5 Hz, H-8), 4.12
(m, 1H, H-4), 4.15 (q, 2H, J = 7.0 Hz, Et), 5.87 (s, 1H, vinyl)
(Found: C, 59.22; H, 9.42. C17H32O5Si requires C, 59.27; H,
9.36%).
1.0 Hz, H-4) (Found: C, 59.28; H, 9.93. C22H44O5Si2 requires C,
59.41; H, 9.97%).
Ethyl (2E,4S,6R)-6,7-epoxy-6-hydroxymethyl-4-methoxy-7-
methyl-3-[(trimethylsilyl)methyl]oct-2-enoate (34)
Ethyl (2Z,4S)-8-hydroxy-4-methoxy-7,7-dimethyl-6-oxo-3-
[(trimethylsilyl)methyl]oct-2-enoate (29) via (25)
The experiment was performed as described for 26 using 54 (78
mg, 0.24 mmol) as starting material. The reaction mixture was
stirred for 1 h 20 min. Column chromatography (H–E, 10 : 1)
gave 34 (oil, 71 mg, 87%) contaminated by ketone 36 (8%).
For 34 (contaminated with 36): Rf (3 : 1) 0.18; [α]D20 Ϫ40 (c 1.29,
CDCl3); δH 0.09 (s, 9H, TMS), 1.27 (t, 3H, J = 7.1 Hz, Et), 1.35,
1.38 (2s, 6H, >C(CH3)2), 1.75 (dd, 1H, JAB = 13.5 Hz, J = 1.3
Hz, –CH2–TMS), 1.86 (d, 1H, JAB = 13.5 Hz, –CH2–TMS),
1.86–1.99 (m, 2H, H-5), 3.27 (s, 3H, –OCH3), 3.68 (m, 1H,
–C(H)OH, addition of D2O gives δ at 3.66 ppm, JAB = 12.4 Hz),
3.69 (s, 1H, –OH, disappears on addition of D2O), 3.98 (m,
1H, –C(H)HOH, addition of D2O gives δ at 3.97 ppm,
JAB = 12.4 Hz), 4.12, 4.15 (2dq, 2H, JAB = 11.0 Hz, J = 7.1 Hz,
Et), 5.22 (ddd, 1H, J = 9.3, 4.2, 1 Hz, H-4), 5.63 (m, 1H,
vinyl) (Found: C, 59.25; H, 9.40. C17H32O5Si requires C,
59.27; H, 9.36%).
The experiment was performed as described for 26 using 22 (85
mg, 0.26 mmol) as starting material and -(ϩ)-diethyl tartrate
instead of -(Ϫ)-diethyl tartrate. The reaction mixture was
stirred for 2 h 40 min. The crude product contained approxi-
mately equal amounts of epoxide 25 and ketone 29 as judged
from TLC (3 : 1). However, after column chromatography on
silica gel (H–E, 10 : 1) the only product isolated was ketone 29
(69 mg, 77%). Recovered starting material (10 mg, 12%) was
also isolated.
Ethyl (2Z,4S)-4-benzyloxy-8-hydroxy-7,7-dimethyl-6-oxo-3-
[(trimethylsilyl)methyl]oct-2-enoate (30)
The experiment was performed as described for 14 using 27 (40
mg, 0.095 mmol) as starting material. Column chromatography
(H–E, 7 : 1) gave 30 (31 mg, 78%) as an oil which crystallized on
standing: Rf (3 : 1) 0.24; mp 43–49 ЊC; [α]D20 ϩ33 (c 0.47, CDCl3);
δH 0.12 (s, 9H, TMS), 1.10, 1.12 (2s, 6H, –C(CH3)2–), 1.30 (t,
3H, J = 7.1 Hz, Et), 1.56 (d, 1H, JAB = 11.5 Hz, –C(H)H-TMS),
2.39 (dd, 1H, J = 7.8, 6.3 Hz, –OH), 2.40 (dd, 1H, JAB = 16.4
Hz, J = 2.2 Hz, H-5), 2.97 (dd, 1H, JAB = 16.4 Hz, J = 10.0 Hz,
H-5), 3.03 (dd, 1H, JAB = 11.5 Hz, J = 0.9 Hz, –C(H)H-TMS),
3.45 (dd, 1H, JAB = 11.5 Hz, J = 7.8 Hz, H-8), 3.64 (dd, 1H,
JAB = 11.5 Hz, J = 6.3 Hz, H-8), 4.16 (q, 2H, J = 7.1 Hz, Et),
4.28 (d, 1H, JAB = 10.7 Hz, benzyl), 4.41 (dd, 1H, J = 2.2 Hz,
10.0 Hz, H-4), 4.52 (d, 1H, JAB = 10.7 Hz, benzyl), 6.00 (s, 1H,
vinyl), 7.25–7.38 (m, 5H, –Ph) (Found: C, 65.48; H, 8.69.
C23H36O5Si requires C, 65.68; H, 8.63%).
Ethyl (2E,4R)-4-(tert-butyldimethylsilyloxy)-8-hydroxy-7,7-
dimethyl-6-oxo-3-[(trimethylsilyl)methyl]oct-2-enoate (35) and
(1S,4S,6S)-1-ethoxy-3-methylene-4-tert-butyldimethylsilyloxy-
6-(2-hydroxypropan-2-yl)-8,9-dioxa[4.2.1]bicyclononane (37)
The experiment was performed as described for 14 using 33
(100 mg, 0.225 mmol) as starting material. Column chromato-
graphy (H–E, 10 : 1 then 7 : 1) gave 35 (oil, 31 mg, 31%) and 37
(28 mg, of ca. 80% purity as determined by 1H NMR analysis)
which was further purified by acetylation of the by-products
(Ac2O and 4-pyrrolidinopyridine in pyridine) followed by
column chromatography (H–E, 10 : 1) to give 37 (oil, 14 mg,
17%).
For 35: Rf (3 : 1) 0.24; [α]D20 ϩ33 (c 0.35, CDCl3); δH 0.01, 0.05
(2s, 6H, –Si(CH3)2–), 0.12 (s, 9H, TMS), 0.85 (s, 9H, t-Bu), 1.11,
1.13 (2s, 6H, –C(CH3)2–), 1.26 (t, 3H, J = 7.1 Hz, Et), 1.61 (dd,
1H, JAB = 13.5 Hz, J = 1.3 Hz, –CH2–TMS), 2.24 (d, 1H,
JAB = 13.5 Hz, J = 1.2 Hz, –CH2–TMS), 2.27 (dd, 1H,
JAB = 16.4 Hz, J = 2.2 Hz, H-5), 2.95 (dd, 1H, JAB = 16.4 Hz,
J = 9.3 Hz, H-5), 3.52, 3.55 (2d, 2H, JAB = 11.4 Hz, H-8),
4.11, 4.14 (2dq, 2H, JAB = 10.9 Hz, J = 7.1 Hz, Et), 5.50 (s,
1H, broad, vinyl), 6.02 (d, 1H, J = 9.3 Hz, broad, H-4)
(Found: C, 59.50; H, 10.06. C22H44O5Si2 requires C, 59.41; H,
9.97%).
Ethyl (2Z,2ЈS,4ЈR)-3-(5Ј,5Ј-dimethyl-4Ј-hydroxy-4Ј-hydroxy-
methyltetrahydrofuran-2Ј-yl)-4-trimethylsilylbut-2-enoate (31)
The experiment was performed as described for 26 using 2422
(100 mg, 0.318 mmol) as starting material. The reaction mix-
ture was stirred at Ϫ40 ЊC for 30 min and then at Ϫ20 ЊC for
4 h. Column chromatography (H–E, 3 : 1 then 1 : 1) gave crystal-
line 31 (40 mg, 38%). Recovered starting material (58 mg, 58%)
was also isolated.
For 31: Rf (2 : 1) 0.11; mp 59–60 ЊC; [α]D20 ϩ110 (c 0.20,
CDCl3); δH 0.06 (s, 9H, TMS), 1.21 (s, 3H, –C(CH3)CH3–), 1.27
(t, 3H, J = 7.1 Hz, Et), 1.37 (s, 3H, –C(CH3)CH3–), 1.57 (d, 1H,
JAB = 11.6 Hz, –C(H)H-TMS), 1.94 (dd, 1H, JAB = 13.7 Hz,
J = 5.1 Hz, –CH2–), 2.52 (dd, 1H, JAB = 13.7 Hz, J = 9.8 Hz,
–CH2–), 3.06 (dd, 1H, JAB = 11.6 Hz, J = 0.8 Hz, –C(H)H-
TMS), 3.63 (s, 2H, –CH2OH), 4.11, 4.16 (2dq, 2H, JAB = 10.9
Hz, J = 7.1 Hz, Et), 4.35 (ddd, 1H, J = 5.1, 9.8, 1.2 Hz, >CHO–),
6.03 (s, 1H, vinyl) (Found: C, 58.10; H, 9.18. C16H30O5Si
requires C, 58.15; H, 9.15%).
For 37: Rf (3 : 1) 0.35; [α]D20 Ϫ21 (c 0.74, CDCl3); δH 0.02, 0.04
(2s, 6H, –Si(CH3)2–), 0.89 (s, 9H, t-Bu), 1.14 (s, 3H, >C(CH3)-
CH3), 1.21 (t, 3H, J = 7.1 Hz, Et), 1.33 (s, 3H, >C(CH3)CH3),
1.56 (dd, 1H, JAB = 13.4 Hz, J = 9.4 Hz, –CH2–), 2.49 (dd, 1H,
JAB = 13.4 Hz, J = 5.9 Hz, –CH2–), 2.59 (d, 1H, JAB = 13.9 Hz,
broad, –CH2C(O–)3), 2.76 (d, 1H, JAB = 13.9 Hz, –CH2C(O–)3),
3.69, 3.78 (2dq, 2H, JAB = 9.3 Hz, J = 7.1 Hz, Et), 3.80, 4.25 (2d,
2H, JAB = 7.1 Hz, –CH2O–), 4.34 (dd, 1H, J = 9.4 Hz, 5.9 Hz,
broad, –CH(OTBDMS)–), 5.03, 5.34 (2m, 2H, vinyl) (Found:
C, 61.07; H, 9.82. C19H36O5Si requires C, 61.25; H, 9.74%).
Ethyl (2E,4S,6R)-4-(tert-butyldimethylsilyloxy)-6,7-epoxy-6-
hydroxymethyl-7-methyl-3-[(trimethylsilyl)methyl]oct-2-enoate
(33)
Ethyl (2E,4S)-8-hydroxy-4-methoxy-7,7-dimethyl-6-oxo-
3-[(trimethylsilyl)methyl]oct-2-enoate (36)
The experiment was performed as for 26 using 55 (242 mg,
0.564 mmol) as starting material. Column chromatography
(H–E, 6 : 1) of the crude product gave 33 (oil, 230 mg, 92%): Rf
(3 : 1) 0.34; [α]D20 ϩ18 (c 1.43, CDCl3); δH Ϫ0.01, 0.08 (2s, 6H,
–Si(CH3)2–), 0.10 (s, 9H, TMS), 0.91 (s, 9H, t-Bu), 1.27 (t, 3H,
J = 7.1 Hz, Et), 1.33, 1.40 (2s, 6H, >C(CH3)2), 1.61 (dd, 1H,
JAB = 14.0 Hz, J = 1.7 Hz, –C(H)H-TMS), 1.67 (s, 1H, broad,
–OH), 1.78 (dd, 1H, JAB = 14.2 Hz, J = 10.9 Hz, H-5), 1.99 (dd,
1H, JAB = 14.2 Hz, J = 2.3 Hz, H-5), 2.17 (d, 1H, JAB = 14.0 Hz,
–C(H)H-TMS), 3.56–3.73 (m, 2H, –CH2OH), 4.09–4.20 (m,
2H, Et), 5.52 (s, 1H, broad, vinyl), 5.79 (dd, 1H, J = 10.9, 2.3,
The experiment was performed as described for 14 using 34 (45
mg, 0.13 mmol) as starting material. Column chromatography
(H–E, 10 : 1 then 5 : 1) gave 36 (oil, 16 mg, 36%) and two other
compounds with unknown structures.
For 36: Rf (3 : 1) 0.15; [α]D20 Ϫ5 (c 0.52, CDCl3); δH 0.11 (s, 9H,
TMS), 1.14, 1.14 (2s, 6H, –C(CH3)2–), 1.27 (t, 3H, J = 7.1 Hz,
Et), 1.75 (dd, 1H, JAB = 13.1 Hz, J = 1.0 Hz, –CH2–TMS), 1.90
(d, 1H, JAB = 13.1 Hz, –CH2–TMS), 2.39 (dd, 1H, JAB = 16.4
Hz, J = 2.4 Hz, H-5), 2.95 (dd, 1H, JAB = 16.4 Hz, J = 9.5 Hz,
H-5), 3.25 (s, 3H, –OCH3), 3.50, 3.64 (2d, 2H, JAB = 11.5 Hz,
798
J. Chem. Soc., Perkin Trans. 1, 2001, 789–800

View Article Online
3 W. S. Johnson, Bioorg. Chem., 1976, 5, 51.
H-8), 4.13 (q, 2H, J = 7.1 Hz, Et), 5.42 (ddd, 1H, J = 2.4, 9.5,
1 Hz, H-4), 5.61 (m, 1H, vinyl) (Found: C, 59.31; H, 9.38.
C17H32O5Si requires C, 59.27; H, 9.36%).
4 P. A. Bartlett, in Asymmetric Synthesis, ed. J. D. Morrison,
Academic Press, Orlando, 1984, vol. 3, p. 341.
5 D. Goldsmith, Fortschr. Chem. Org. Naturst., 1971, 29, 363.
6 J. K. Sutherland, Chem. Soc. Rev., 1980, 265.
7 N. Murase, K. Maruoka, T. Ooi and H. Yamamoto, Bull. Chem.
Soc. Jpn., 1997, 70, 707.
Ethyl (2Z,4S,6R)-4-(tert-butyldimethylsilyloxy)-6-[(tert-butyl-
dimethylsilyloxy)methyl]-6,7-epoxy-3,7-dimethyloct-2-enoate
(38)
8 E. E. van Tamelen, Acc. Chem. Res., 1968, 1, 111.
9 J. K. Sutherland, in Comprehensive Organic Synthesis: Polyene
Cyclizations, ed. B. Trost, I. Flemming and G. Pattenden, Pergamon
Press, Oxford, 1991, vol. 3, p. 341.
10 I. Fleming, A. Pearce and R. L. Snowden, J. Chem. Soc., Chem.
Commun., 1976, 182.
11 I. Fleming and A. Pearce, J. Chem. Soc., Perkin Trans. 1, 1981, 251.
12 H. F. Chow and I. Fleming, J. Chem. Soc., Perkin Trans. 1, 1984,
1815.
13 G. Majetich, in Allylsilanes in Organic Synthesis, ed. T. Hudlicky,
JAI Press Inc., Greenwich, Connecticut, 1989, vol. 1, p. 175.
14 R. J. Armstrong and L. Weiler, Can. J. Chem., 1983, 61, 2530.
15 R. J. Armstrong, F. L. Harris and L. Weiler, Can. J. Chem., 1982, 60,
673.
16 W. S. Johnson, J. D. Elliott and G. J. Hanson, J. Am. Chem. Soc.,
1984, 106, 1138.
17 R. J. Armstrong and L. Weiler, Can. J. Chem., 1986, 64, 584.
18 W. S. Johnson, S. D. Lindell and J. Steele, J. Am. Chem. Soc., 1987,
109, 5852.
19 D. Guay, W. S. Johnson and U. Schubert, J. Org. Chem., 1989, 54,
4731.
20 S. Hatakeyama, H. Numata, K. Osanai and S. Takano, J. Chem.
Soc., Chem. Commun., 1989, 1893.
n-BuLi (2.0 mL, 3.0 mmol, 1.50 M in hexane) was added to a
solution of diisopropylamine (0.434 mL, 3.1 mmol) in dry THF
(1.5 mL) under nitrogen at 0 ЊC. After stirring for 10 min,
Me3SiCH2CO2Et (481 mg, 3.0 mmol) was added and stirring
was continued for another 10 min. A solution of 61 (300 mg,
0.720 mmol) in dry THF (0.5 mL) was then added and the
reaction mixture was stirred for 30 min at 0 ЊC. Heptane (100
mL) was added and the solution was washed with water, 1.0 M
HCl, NaHCO3 (sat.) again with water and dried (MgSO4).
Evaporation of the solvent at reduced pressure was followed by
column chromatography (H–E, 50 : 1) to give 38 (oil, 231 mg,
66%): Rf (10 : 1) 0.45; [α]D20 Ϫ2 (c 1.17, CDCl3); δH Ϫ0.02, 0.05 (2s,
6H, –Si(CH3)2–), 0.08 (s, 6H, –Si(CH3)2–), 0.88, 0.89 (2s, 18H,
t-Bu), 1.26 (t, 3H, J = 7.1 Hz, Et), 1.32, 1.38 (2s, 6H,
>C(CH ) ), 1.91 (d, 3H, J = 1.1 Hz, –C(CH )᎐CH–), 1.95, 1.95
᎐
3
3
2
(2d, 2H, J = 5.8, 7.5 Hz, H-5), 3.58, 3.91 (2d, 2H, JAB = 10.6 Hz,
–CH2O–), 4.12 (q, 2H, J = 7.1 Hz, Et), 5.61 (m, 1H, vinyl), 5.88
(dd, 1H, J = 5.8, 7.5 Hz, broad, H-4) (Found: C, 61.59; H,
10.29. C25H50O5Si2 requires C, 61.68; H, 10.35%).
21 L. Pettersson, T. Frejd and G. Magnusson, Tetrahedron Lett., 1987,
28, 2753.
22 L. Pettersson, T. Frejd and G. Magnusson, Acta Chem. Scand.,
1993, 47, 196.
23 X. Xiao, S.-K. Park and G. D. Prestwich, J. Org. Chem., 1988, 53,
4869.
24 D. Serramedan, B. Delmond, G. Deleris, J. Dunogues, M. Pereyre
and C. Filliatre, J. Organomet. Chem., 1990, 398, 79.
25 M. E. Jung, Y. M. Cho and Y. H. Jung, Tetrahedron Lett., 1996, 37,
3.
Ethyl (2Z,4S)-4,8-bis(tert-butyldimethylsilyloxy)-3,7,7-tri-
methyl-6-oxooct-2-enoate (39)
The experiment was performed as described for 14 using 38
(50.0 mg, 0.102 mmol) as starting material. Column chrom-
atography (H–E, 50 : 1) gave 39 (oil, 34 mg, 68%): Rf (10 : 1)
0.45; [α]D20 ϩ3 (c 0.75, CDCl3); δH 0.02 (s, 9H, –Si(CH3)2–), 0.06
(s, 3H, –Si(CH3)2–), 0.83, 0.86 (2s, 18H, t-Bu), 1.08 (s, 6H,
–C(CH3)2–), 1.27 (t, 3H, J = 7.1 Hz, Et), 1.90 (d, 3H, J = 1.4 Hz,
26 S. K. Taylor, S. A. May and E. S. Stansby, J. Org. Chem., 1996, 61,
2075.
27 T. Doi, J. Robertson, G. Stork and A. Yamashita, Tetrahedron Lett.,
1994, 35, 1481.
–C(CH )᎐ CH–), 2.34 (dd, 1H, JAB = 16.8 Hz, J = 3.0 Hz, H-5),
᎐
3
3.05 (dd, 1H, JAB = 16.8 Hz, J = 8.9 Hz, H-5), 3.51, 3.58 (2d,
2H, JAB = 9.8 Hz, –CH2O–), 4.13, 4.16 (2dq, 2H, JAB = 11.0 Hz,
J = 7.1 Hz, Et), 5.60 (m, 1H, vinyl), 6.04 (ddd, 1H, J = 3.0, 8.9,
0.8 Hz, H-4) (Found: C, 61.84; H, 10.40. C25H50O5Si2 requires
C, 61.68; H, 10.35%).
28 N. K. Yee and R. M. Coates, J. Org. Chem., 1992, 57, 4598.
29 S. K. Taylor, Org. Prep. Proced. Int., 1992, 24, 245.
30 B. Rickborn, in Comprehensive Organic Synthesis; Carbon-Carbon
σ-Bond Formation, ed. B. M. Trost, I. Fleming and G. Pattenden,
Pergamon Press, Oxford, 1991, vol. 3, ch. 3.3.
31 A. S. Rao, S. K. Paknikar and J. G. Kirtane, Tetrahedron, 1983,
39, 2323.
32 R. E. Parker and N. S. Isaacs, Chem. Rev., 1959, 59, 737.
33 B. C. Ranu and U. Jana, J. Org. Chem., 1998, 63, 8212.
34 H. Itokawa, H. Nakanishi and S. Mihashi, Chem. Pharm. Bull. Jpn.,
1983, 31, 1991.
35 G. H. Magnusson, Org. Prep. Proced. Int., 1990, 22, 547.
36 Atta-ur-Rahman, Nuclear Magnetic Resonance, Springer-Verlag,
New York, 1986.
37 P. A. Bartlett, in Asymmetric Synthesis, ed. J. D. Morrison,
Academic Press, Orlando, 1984, vol. 3 (part B), p. 411.
38 G. Stork, L. D. Cama and D. R. Coulson, J. Am. Chem. Soc., 1974,
96, 5268.
(2Z,4S)-4-(tert-Butyldimethylsilyloxy)-3,7,7-trimethyl-6-
oxooct-2-en-8-olide (40)
The experiment was performed as described for 14 using 38
(100 mg, 0.205 mmol) as starting material, except that the reac-
tion mixture was stirred for 4 h. Column chromatography
(H–E, 10 : 1 then 5 : 1) gave 40 (oil, 36 mg, 54%): Rf (1 : 1) 0.47;
[α]D20 ϩ10 (c 0.64, CDCl3); δH 0.03 (s, 6H, –Si(CH3)2–), 0.86 (s,
9H, t-Bu), 1.11, 1.12 (2s, 6H, –C(CH3)2–), 2.01 (dd, 3H, J = 1.5,
0.8 Hz, –C(CH )᎐CH–), 2.84 (dd, 1H, JAB = 17.9 Hz, J = 5.6
᎐
3
Hz, H-5), 3.00 (dd, 1H, JAB = 17.9 Hz, J = 6.7 Hz, H-5), 3.56,
3.58 (2d, 2H, JAB = 9.9 Hz, –CH2O–), 5.38 (dddd, 1H, J = 5.6,
6.7, 1.6, 0.8 Hz, H-4), 5.80 (dq, 1H, J = 1.6, 1.5 Hz, vinyl)
(Found: C, 62.63; H, 9.27. C17H30O4Si requires C, 62.54; H,
9.26%).
39 S. P. Tanis, Y.-H. Chuang and D. B. Head, J. Org. Chem., 1988, 53,
4929.
40 K. Maruoka, T. Ooi and H. Yamamoto, J. Am. Chem. Soc., 1989,
111, 6431.
41 G. Majetich, A. Casares, D. Chapman and M. Behnke, J. Org.
Chem., 1986, 51, 1745.
42 T. Katsuki and K. B. Sharpless, J. Am. Chem. Soc., 1980, 102, 5974.
43 R. M. Hanson and K. B. Sharpless, J. Org. Chem., 1986, 51, 1922.
44 K. C. Nicolaou, C. V. C. Prasad, P. K. Somers and C. K. Hwang,
J. Am. Chem. Soc., 1989, 111, 5330.
45 P. G. Wuts, R. D’Costa and W. Butler, J. Org. Chem., 1984, 49, 2582.
46 M. G. Finn and K. B. Sharpless, in Asymmetric Synthesis, ed. J. D.
Morrison, Academic Press, Orlando, 1985, vol. 5, p. 247.
47 S. D. Jolad, J. J. Hoffmann, K. H. Schram, J. R. Cole, M. S.
Tempesta and R. B. Bates, J. Org. Chem., 1981, 46, 641.
48 A. E. de Jesus, R. M. Horak, P. S. Steyn and R. Vleggaar, J. Chem.
Soc., Perkin Trans. 1, 1987, 2253.
Acknowledgements
To the memory of Professor Göran Magnusson, Lund Uni-
versity. We thank the Swedish Natural Science Research
Council, the Crafoord Foundation, the Knut and Alice
Wallenberg Foundation, and the Royal Physiographic Society
in Lund for financial support.
References
1 W. S. Johnson, Acc. Chem. Res., 1968, 1, 1.
2 W. S. Johnson, Angew. Chem., 1976, 88, 33.
49 E. Taskinen, J. Org. Chem., 1978, 43, 2773.
50 R. Radeglia, H. Schuetz, L. Haase and E. Klemm, Z. Chem., 1985,
25, 66.
J. Chem. Soc., Perkin Trans. 1, 2001, 789–800
799

View Article Online
51 C. Galli, G. Giovannelli, G. Illuminati and L. Mandolini, J. Org.
Chem., 1979, 44, 1258.
52 S. W. Baldwin, P. Chen, N. Nikolic and D. C. Weinseimer, Org. Lett.,
2000, 2, 1193.
53 M. Shimazaki, H. Hara, K. Suzuki and G. Tsuchihashi, Tetrahedron
Lett., 1987, 28, 5891.
54 K. Maruoka, M. Hasegawa and H. Yamamoto, J. Am. Chem. Soc.,
1986, 108, 3827.
55 A. M. Doherty and S. V. Ley, Tetrahedron Lett., 1986, 27, 107.
56 G. E. Keck and S. Castellino, Tetrahedron Lett., 1987, 28, 281.
57 S. D. Kahn, G. E. Keck and W. J. Hehre, Tetrahedron Lett., 1987, 28,
279.
65 E. J. Corey and A. Venkateswarlu, J. Am. Chem. Soc., 1972, 94,
6190.
66 G. van Look, Silylating Agents, Fluka Chemie, 1988.
67 B. H. Lipshutz and J. J. Pegram, Tetrahedron Lett., 1980, 21, 3343.
68 D. E. Seitz and A. Zapata, Synthesis, 1981, 557.
69 D. E. Seitz and A. Zapata, Tetrahedron Lett., 1980, 21, 3451.
70 M. Demuth, Helv. Chim. Acta, 1978, 61, 3136.
71 G. L. Larson, D. Hernandez, I. Montes de Lopez-Cepero and
L. E. Torres, J. Org. Chem., 1985, 50, 5260.
72 A. Meskens, Synthesis, 1981, 501.
73 W. S. Wadsworth and D. W. Emmons, J. Am. Chem. Soc., 1961, 83,
1733.
58 J. M. Coxon, A. J. Thorpe and W. B. Smith, J. Org. Chem., 1999, 64,
74 K. Shimoji, H. Taguchi, K. Oshima, H. Yamamoto and H. Nozaki,
J. Am. Chem. Soc., 1974, 96.
9575.
59 A. Eschenmoser, L. Ruzicka, O. Jeger and D. Arigoni, Helv. Chim.
Acta, 1955, 38, 1890.
60 G. Stork and A. W. Burgstahler, J. Am. Chem. Soc., 1955, 77,
5068.
75 H. Taguchi, K. Shimoji, H. Yamamoto and H. Nozaki, Bull. Chem.
Soc. Jpn., 1974, 47, 2529.
76 G. L. Larson, J. A. Preito and A. Hernandes, Tetrahedron Lett.,
1981, 22, 1575.
61 D. J. Goldsmith, J. Am. Chem. Soc., 1962, 84, 3913.
62 T. Purdie and J. C. Irvine, J. Chem. Soc., 1903, 83, 1021.
63 R. Kuhn, H. Trischmann and I. Löw, Angew. Chem., 1955, 67, 32.
64 A. Yamamoto, Organotransition Metal Chemistry. Fundamental
Concepts and Applications, John Wiley & Sons, 1986, 391.
77 L. M. Jackman and R. H. Wiley, J. Chem. Soc., 1960, 2886.
78 K. Jansson, T. Frejd, J. Kihlberg and G. Magnusson, Tetrahedron
Lett., 1986, 27, 753.
79 K. Jansson, S. Ahlfors, T. Frejd, J. Kihlberg and G. Magnusson,
J. Org. Chem., 1988, 53, 5629.
800
J. Chem. Soc., Perkin Trans. 1, 2001, 789–800