Enantioselective synthesis of epoxides by α-deprotonation - Electrophile trapping of achiral epoxides

Article abstract of DOI:10.1039/b309717j

Enantioselective α-deprotonation of achiral epoxides 1, 21, and 26 using organolithiums in the presence of (-)-sparteine 2 and subsequent electrophile trapping gives access to enantioenriched trisubstituted epoxides 9-17, 22, 23, 27 and 28 (in up to 86% ee).

Full text of DOI:10.1039/b309717j

View Article Online / Journal Homepage / Table of Contents for this issue
Enantioselective synthesis of epoxides by ꢀ-deprotonation—
electrophile trapping of achiral epoxides
David M. Hodgson,* Timothy J. Buxton, Iain D. Cameron, Emmanuel Gras† and
Eirene H. M. Kirton
Dyson Perrins Laboratory, Department of Chemistry, University of Oxford, South Parks Road,
Oxford, UK OX1 3QY. E-mail: david.hodgson@chem.ox.ac.uk; Fax: ϩ44(1865)275674
Received 13th August 2003, Accepted 8th October 2003
First published as an Advance Article on the web 24th October 2003
Enantioselective α-deprotonation of achiral epoxides 1, 21, and 26 using organolithiums in the presence of
(
2
Ϫ)-sparteine 2 and subsequent electrophile trapping gives access to enantioenriched trisubstituted epoxides 9–17, 22,
3, 27 and 28 (in up to 86% ee).
Introduction
generation of the oxiranyl anion) and deuteration (CD OD as
3
external electrophile). We have also recently communicated that
a similar strategy could be used to enantioselectively desym-
metrise epoxides of cycloalkenes and report here our studies in
The enantioselective desymmetrisation of an achiral compound
is an attractive and powerful concept; its application to meso-
epoxides has been developed in a number of ways, yet in these
approaches the epoxide functionality is removed in an enantio-
selective chemical transformation. As epoxides are widely
found both as versatile synthetic intermediates, and in a number
of interesting natural products, the development of effici-
ent (especially asymmetric) methods for the elaboration of
epoxides constitutes an important ongoing challenge.
a strategy could use the nucleophilic chemistry of epoxides (via
1
2
–8
16
detail.
Results and discussion
As previously observed during our studies on enantioselective
α-lithiation–rearrangement of meso-epoxides of medium-sized
cycloalkenes (e.g. 1 3, Scheme 2), rearrangement occurs effi-
9
10,11
Such
ciently upon warming of the reaction mixture from Ϫ90 ЊC in
4
12
oxiranyl anions), first studied by Eisch and Galle, which is far
less developed than the use of epoxides as electrophiles. A
current requirement with this latter strategy is that, in order to
avoid any subsequent rearrangement of the reactive α-metal-
17
Et O. Therefore, in order to trap the transient oxiranyl anion,
2
we carefully explored the time required for its formation at
Ϫ90 ЊC by quenching the reaction at increasing duration with
s
CD OD. The initial reaction conditions used [Bu Li (2.45
4
,5
3
lated epoxide, the epoxide must possess an activating sub-
stituent (electron-withdrawing, trialkylsilyl, or trialkylstannyl
group) attached to the epoxide ring. Electron-withdrawing and
trialkylsilyl substituents facilitate formation of oxiranyl anions
by promoting deprotonation (usually lithiation) and prolonging
the solution lifetime of these otherwise very labile intermedi-
ates. Trialkylstannyl- and sulfinyl-substituted epoxides react
with organolithiums (by transmetallation and desulfinylation,
equiv.), (Ϫ)-sparteine 2 (2.5 equiv.), Et O, Ϫ90 ЊC] were the
2
same as those for the rearrangement of (commercially avail-
17
able) cyclooctene oxide 1 (Scheme 2). As only two compounds
(
starting material and bicyclic alcohol 3 from transannular
rearrangement) were observed by TLC, the % D incorporation
1
in 1 was determined from the crude H NMR spectrum of the
reaction mixture after aqueous work-up. These results showed a
significant increase in D incorporation over the first 20 minutes
13,14
respectively, Scheme 1)
rapidly enough at low temperatures
(
Table 1). It is noteworthy that the amount of 3 increased at a
such that the resultant unstabilised oxiranyl anions can exhibit
synthetically useful nucleophilic (rather than carbene-type)
reactivity with a range of electrophiles.
slower rate, with it mainly arising between 5 and 10 minutes
s
after addition of the epoxide 1 to the Bu Li–2 mixture. Once
α-lithiation is complete, the lithiated epoxide is surprisingly
stable towards transannular C–H insertion with no decom-
position observed over a 2 h period; some reaction was
observed over 14 h, however (Table 1, entry 7). These results
indicate that the initial formation of 3 occurs during the de-
protonation, possibly due to local heating in the likely exo-
1
thermic lithiation. From the crude H NMR spectra, only
Scheme 1
cyclooctene oxide 1/1-D and bicyclic alcohol 3 were present
after aqueous work-up. Also of interest is the observation of
complete rearrangement to the bicyclic alcohol 3 after 3 h when
Such reactions demonstrate the value of oxiranyllithiums as
important intermediates in the elaboration of epoxides. How-
ever, they also indicate the potential limitation of requiring
an activated epoxide precursor to carry out the chemistry.
Recently, we showed that conservation of epoxide functionality
was possible after direct lithiation of terminal epoxides in the
presence of a diamine ligand, followed by electrophile trapping
the deuteration was attempted with D
O rather than CD OD,
2
3
15
of the non-stabilised oxiranyl anion intermediates. However,
the reaction was restricted to silylation (TMSCl present during
†
Present address: LSPCMIB, Université Paul Sabatier, 118, Route de
Narbonne, 31062 Toulouse, France. gras@chimie.ups-tlse.fr; Fax:
ϩ33(0)561556011.
Scheme 2
T h i s j o u r n a l i s © T h e R o y a l S o c i e t y o f C h e m i s t r y 2 0 0 3
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 4 2 9 3 – 4 3 0 1
4293

View Article Online
s
Table 1 Reaction profile as a function of time using epoxide 1 with Bu Li (2.45 equiv.) and (Ϫ)-sparteine 2 (2.5 equiv.) in Et O at Ϫ90 ЊC, followed
2
by the addition of CD OD after the time indicated
3
a
Entry
Reaction time
% D incorporation in 1
Isolated yield of 3 (%)
1
2
3
5 min
10 min
15 min
20 min
30 min
2 h 30 min
14 h
40
56
63
100
100
100
100
Traces
26
19
29
29
28
43
b
4
5
6
7
a
1
b
Crude H NMR spectra and GC analysis only show the presence of 1/1-D and 3. Cyclooctene oxide 1-D was isolated in 61% yield.
s
Table 2 Reaction profile as a function of time using epoxide 1 with Bu Li (1.25 equiv.) and (Ϫ)-sparteine 2 (1.3 equiv.) in Et O at Ϫ90 ЊC
2
Entry
Reaction time
% D incorporation in 1
Isolated yield of 3 (%)
1
2
3
4
5
30 min
1 h
1.5 h
2 h
78
100
100
100
100
0
3
5.5
10
17
a
14 h
a
Cyclooctene oxide 1-D was isolated in 85% yield.
s
Table 3 Reaction profile as a function of time using epoxide 1 with Bu Li (1.25 equiv.) and TMEDA (1.3 equiv.) in Et O at Ϫ90 ЊC
2
Entry
Reaction time
% D incorporation in 1
Isolated yield of 3 (%)
1
2
3
4
30 min
2 h
3 h
49
84
98
0
0
Traces
20
14 h
100
s
probably because of a miscibility issue combined with a lack of
temperature control due to localised heating while quenching
were carried out with reactions involving Bu Li (1.25 equiv.)
and (Ϫ)-sparteine 2 (1.3 equiv.). The results obtained as a
function of the reaction time are given in Table 2.
with D O.
2
With the nucleophilic (or strictly in this case, basic) nature of
the transient lithiated epoxide established, we tried to optimise
the efficiency of the reaction by accelerating the trapping reac-
tion and at the same time limiting the rate of the rearrange-
ment. The first objective could potentially be achieved by the
addition of a more polar co-solvent after generation the oxir-
anyl anion. This was attempted by producing the oxiranyl anion
under the above conditions and then after 20 min adding pre-
Not surprisingly, comparison of the results in Tables 1 and 2
show that using a lower quantity of base leads to a lower rate of
deprotonation of cyclooctene oxide 1. After 30 min, the deuter-
ation is not complete (Table 2, entry 1), this contrasts with the
case using 2.45 equiv. of Bu Li (Table 1, entries 4 and 5). How-
ever, the most interesting observation is the reduced amount of
s
bicyclic alcohol 3 obtained under the conditions reported in
s
Table 2. Indeed, when using 2.45 equiv. of Bu Li at least 20% of
cooled THF (1 ml when 10 ml Et O had been used) at Ϫ90 ЊC
3 was obtained, together with fully deuterated 1-D (Table 1);
with the smaller excess of alkyllithium only trace amounts of 3
are obtained after 1 h (Table 2, entry 2). The slower rate of
deprotonation (leading to less localised heating) and/or aggre-
gate modification (due to the lower quantity of Bu Li present)
may be the origin of the reduced quantity of bicyclic alcohol 3
2
and leaving for a further 5 min before quenching with CD OD.
3
This reaction clearly showed the high destabilising effect of
THF on the oxiranyl anion, since the bicyclic alcohol 3 was now
obtained in 64% yield. This effect could be due to displacement
of the (Ϫ)-sparteine 2 ligand around the lithium of the lithiated
s
15,18
epoxide by THF
and would support the hypothesis that the
formed.
(
Ϫ)-sparteine 2 is stabilising the lithiated oxiranyl anion. More-
The results shown in Table 2 are significant since not only do
they indicate a lower quantity of organolithium can be used,
but they also permit the use of a smaller excess of electrophile,
resulting in fewer by-products. As the products obtained by
extension of this methodology to the trapping of electrophiles
over, only 30% yield of cyclooctene oxide 1 was obtained, with
only 43% D incorporation. This poor yield and D incorpor-
ation rules out the use of a polar co-solvent. We then investi-
gated the influence of the presence of the diamine ligand on the
formation and potential stabilisation of the oxiranyl anion. It
was found that after stirring for 1 h without (Ϫ)-sparteine 2
under otherwise identical conditions, only 30% of the recovered
cyclooctene oxide 1 was deuterated and also the reaction
yielded 12% of bicyclic alcohol 3. This result demonstrates that
the presence of (Ϫ)-sparteine 2 not only facilitates the de-
protonation of cyclooctene oxide 1 but also helps to achieve a
better ratio 1-D : 3, this ratio being lowered from 5 : 1 with
ϩ
other than D would give enantioenriched products, for which
the determination of the ee would be important, we then sought
a way to access racemic samples for ee determinations. We first
examined the use of TMEDA as an achiral additive. The first
requirement was to establish the reaction time necessary to fully
s
lithiate cyclooctene oxide 1 using Bu Li (1.25 equiv.) with
TMEDA (1.3 equiv.) in Et O at Ϫ90 ЊC. The results obtained
2
after quenching with CD OD are summarised in Table 3. It is
3
(
Ϫ)-sparteine 2 to 2 : 1 without diamine. This also indicates that
noteworthy that after 3 h almost complete D incorporation was
there is likely to be a modification of the aggregate of the
achieved and only traces of bicyclic alcohol 3 was detected in
the crude H NMR spectrum.
We then sought to probe the effect of the structure of the
achiral diamine on the reaction profile. In order to mimic
more closely the structure of (Ϫ)-sparteine 2 we decided to use
1
lithiated epoxide in the presence of a diamine ligand (as has
19
been clearly established for alkyllithiums). Finally, we exam-
ined the quantity of alkyllithium and (Ϫ)-sparteine 2 required
to allow complete deprotonation. Deuteration experiments
4
294
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 4 2 9 3 – 4 3 0 1

View Article Online
3
,7-dibutyl-3,7-diazabicyclo[3.3.1]nonane 8 (also called dibutyl-
bispidine, DBB). This bicyclic diamine is classically obtained
via a Wolff–Kishner reduction of bispidinone 7, which itself
arises from a double Mannich reaction with 4-butylpiperidone
20–22
23
6
(Scheme 3).
We, and others, have found difficulties in this
synthetic route to 3,7-dialkylbispidines, mainly due to the harsh
conditions required. Our main concern was to reproduce the
Wolff–Kishner reduction using anhydrous hydrazine in a steel
bomb. We therefore considered alternative ways to reduce the
ketone. As Clemmensen reduction failed, we used the Wolff
procedure involving the isolation of the intermediate hydrazone
(
formed in this case by reaction of the bispidinone 7 with an
excess of hydrazine monohydrate at reflux) and its subsequent
reaction with potassium hydroxide; this procedure proved to be
reproducible in our hands allowing after distillation the iso-
lation of 49% of DBB 8 on up to 20 g scale. Also of interest is
that refluxing the distillation residue in conc. HCl allowed a
partial recovery of the starting bispidinone 7 permitting
recycling. The total conversion of bispidinone 7 to bispidine 8
then reaches 70%.
s
Scheme 4 Electrophile trapping of epoxide 1 using Bu Li (1.25 equiv.
unless indicated otherwise) with TMEDA, DBB 8, or (for the
enantioenriched products) (Ϫ)-sparteine 2 (1.3 equiv.).
with N,N-dimethylbenzamide (56%). Acylation using an amide
indicates that the destabilised oxiranyl anion possesses good
nucleophilicity toward relatively unreactive electrophiles. Reac-
tions with propanal, 3-pentanone and N,N-dimethylpropion-
amide yielded the corresponding alcohols 11 and 12 and ketone
1
3 (in 48%, 60% and 45% yields, respectively), showing that
Scheme 3
reaction with potentially enolisable aldehydes, ketones and
amides is achievable. It also proved possible to introduce an
alkyl group, using MeI to give α-methylcyclooctene oxide 18
(75%). The yield of the latter was determined by GC, as it
Having a significant quantity of bispidine 8 in hand, we
examined the deuteration of cyclooctene oxide 1 utilising this
diamine ligand 8 (1.3 equiv.). The reaction time required to
obtain complete D incorporation with epoxide 1 proved much
longer than previously observed using (Ϫ)-sparteine 2, indi-
cating a strong dependence of the rate of deprotonation on
the structure of the alkyllithium–diamine complex. After 16 h,
complete D incorporation was obtained with only 13% of
bicyclic alcohol 3 additionally present, indicating a good stabil-
ity in the presence of 8 of the transient oxiranyl anion with
respect to the rearrangement. This product profile is compar-
able with that obtained with (Ϫ)-sparteine 2 and indicates a
potential stabilisation of the lithiated epoxide. The stabilisation
arising from the presence of (Ϫ)-2 or 8 might be explained
either by ‘protection’ induced by a bulky complex, or by a more
important modification of the structure of the aggregate.
proved to have the same R by TLC as the starting material,
f
which made up the rest of the material. Attempts to trap the
anion with other alkyl halides such as butyl iodide and allyl
bromide proved to be unsuccessful by this procedure, and the
cyclooctene oxide 1 was recovered with no deuterium incorpor-
ation after quenching with deuterated methanol. Use of ethyl
chloroformate (15 equiv.) allowed isolation of the α,β-epoxy-
ester 14 (40%). α-Tributylstannyl epoxide 15 was obtained
(56%) from reaction with Bu SnCl. It is noteworthy that using
3
TMEDA in the direct trapping methodology gave important
variation in isolated yields: use of TMSCl as an external electro-
phile resulted in a poor yield of the α,β-epoxysilane 17.
However, when TMSCl was added with the epoxide a 38% yield
of 17 could be realised, which improved to 60% using DBB 8.
At this point, we examined trapping the transient oxiranyl
anion generated in the presence of (Ϫ)-sparteine 2 in order
to create new enantioenriched products containing carbon–
carbon or carbon–heteroatom bonds while retaining the
epoxide functionality (Scheme 4). Using the best experimental
conditions established for the deuteration of cyclooctene oxide
1, we examined trapping various electrophiles. We first explored
the possibility of creating a carbon-carbon bond in a single
step.
We then moved on to examine more useful electrophile trap-
ping reactions that would allow the creation of new carbon–
carbon and carbon–heteroatom bonds. Significantly, when
s
using TMEDA (1.3 equiv.) as the ligand with Bu Li (1.25
equiv.), introduction of a wide array of functionality at the
epoxide carbon of cyclooctene oxide 1 was found to be possible
using a range of electrophiles (1.5 equiv. unless indicated
otherwise) (Scheme 4).
For the electrophile trapping of epoxide 1 using TMEDA as
the diamine ligand, the reaction with benzaldehyde yielded the
corresponding benzylic alcohol 9 in good yield (80%, 1:1 mix-
Trapping the oxiranyl anion with benzaldehyde gave 80%
of the desired epoxyalcohol (ϩ)-9 as a 2 : 1 mixture of
diastereoisomers (by H NMR). After MnO oxidation to the
ketone (Ϫ)-10, a 76% ee was measured. Reaction with N,N-di-
24
1
ture of chromatographically inseparable diastereoisomers),
while the corresponding ketone 10 was obtained via reaction
2
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 4 2 9 3 – 4 3 0 1
4295

View Article Online
methylbenzamide led to the corresponding ketone (Ϫ)-10 in
Ϫ90 ЊC, followed by addition of MeI allowed the isolation of
i
s
6
8% yield with 77% ee. Using Pr Li instead of Bu Li gave a
72% of the α-methylated epoxide 18 (Scheme 5). It is also note-
worthy that this procedure allowed purification of the latter
(which was not possible by direct lithiation–trapping). Under
the same conditions, it was also possible to trap the lithiated
epoxide using butyl iodide, albeit in very low yield (11%).
significant increase of the ee of (Ϫ)-10 (86% compared to 77%).
i
This reflects the higher enantiotopic proton selection of Pr Li–
s
17
(
Ϫ)-2 vs. Bu Li–(Ϫ)-2 with epoxide 1. Addition to an enolis-
able aldehyde such as propionaldehyde also proved to be
successful. This led to the corresponding epoxyalcohol (Ϫ)-11
in 74% yield but with no diastereomeric excess. Dess–Martin
oxidation of epoxyalcohol (Ϫ)-11 to ketone (ϩ)-13 allowed the
ee to be determined (76%). Moving to more bulky carbonyl
compounds did not disfavour the reaction. Addition of
3
-pentanone gave a 75% yield of the corresponding tertiary
i
s
alcohol (Ϫ)-12 in 77% ee. Using Pr Li instead of Bu Li again
allowed a significant increase of the ee of (Ϫ)-12 (up to 86%).
Reaction with the corresponding amide, N,N-dimethylpropion-
amide led to the corresponding ketone (ϩ)-13 in 67% yield with
Scheme 5
We also examined the possible formation of a tin “ate” com-
plex from epoxystannane 15. This was performed via tin NMR
studies, by adding Bu Li (with and without diamine and polar
n
7
8% ee. We tried then to apply this methodology to the syn-
thesis of trialkyl-substituted epoxides. Trapping the lithiated
epoxide with methyl iodide yielded an inseparable mixture of
cyclooctene oxide 1 and α-methylcyclooctene oxide 18. Finally,
using a large excess (15 equiv.) of ethyl chloroformate we
managed to isolate a reasonable yield (58%) of the epoxyester
co-solvent) to a frozen (in liquid N ) solution of stannylated
2
119
epoxide 15. Following the reaction by Sn NMR did not allow
the observation of a potential signal (around Ϫ280 ppm) of an
“ate” complex, such as 20 (Fig. 1), likely ruling out the possible
existence of an equilibrium between such a complex and the
27
(
Ϫ)-14 in 81% ee. The use of a stoichiometric amount of this
lithiated epoxide.
electrophile yielded an inseparable 1 : 1 mixture of the epoxy-
ester (Ϫ)-14 and the diepoxyketone resulting from addition of
lithiated 1 to 14. When this reaction was carried out using 3 or
1
5 equivalents of ethyl chloroformate, none of the diepoxyke-
tone was observed. With 3 equiv. of electrophile, the epoxy ester
(
(
Ϫ)-14 was obtained in 51% yield. Alternatively, epoxyester
Ϫ)-14, could potentially be prepared using ethyl cyanoformate,
Fig. 1
but this proved unsuccessful and no identifiable products
were isolated. Attempts to trap the oxiranyl anion with benzo-
We next examined the scope of this new methodology with
alternative meso-epoxides. 5,6-Disubstituted epoxide 21 has
previously been successfully used in the enantioselective
desymmetrisation transannular C–H insertion reaction (cf.
phenone yielded
compounds.
a complex mixture of uncharacterised
In order to create valuable carbon–heteroatom bonds enan-
tioselectively, we first examined the use of trimethyltin chloride
28,29
formation of 3 in Scheme 2).
Epoxide 21 was fully deuter-
(
(
1.3 equiv.) as the electrophile, which gave epoxystannane
Ϫ)-16 in 65% yield and 78% ee (determined after BuLi-induced
ated under the same reaction conditions as cyclooctene oxide 1
s
using Bu Li (1.25 equiv.) and (Ϫ)-sparteine (1.3 equiv.), albeit
transmetallation 3-pentanone trapping). If tributyltin chloride
was used then the reaction afforded the corresponding epoxy-
stannane (Ϫ)-15 in similar yield and ee to (Ϫ)-16. We then
turned our attentions to trap the oxiranyl anion with TMSCl as
an internal electrophile (cf. the trapping using TMEDA as the
diamine ligand); this yielded 72% of the silylated epoxide
after a slightly longer reaction time (2.5 h). An increased stabil-
ity was observed for lithiated 21, since an overnight reaction at
Ϫ90 ЊC also produced complete D incorporation, with no by-
product observed. The increase in both lithiation time and
oxiranyl anion stability might be due to the slow equilibrium
previously established between the two enantiomeric con-
formers of 21. The steric bulk of the backside of the ring
might also account for the extent of stability of the lithiated
epoxide derived from 21. Having established the time required
for complete deprotonation of epoxide 21 with TMEDA (12 h),
we managed to trap the intermediate lithiated epoxide with
benzaldehyde and N,N-dimethylbenzamide to give the alcohol
28
(
Ϫ)-17 with 74% ee. This reaction is of significance since
α-silylated epoxides are valuable intermediates in organic syn-
25
thesis. Reactions with a more bulky silane, namely TBDPSCl,
were unsuccessful. The fact that similar levels of enantio-
enrichment were observed for all the electrophiles examined
reflects the fact that the enantiodiscriminating event is the
deprotonation before the electrophile is added. In the case of
TMSCl present in situ, the TMSCl does not appear to influence
the enantioselectivity of the deprotonation step. The absolute
configuration of the predominant enantiomers formed in the
enantioselective lithiation–electrophile trappings of epoxide 1
are as shown in Scheme 4, and are assigned from the known
sense of asymmetric induction in the deprotonation of epoxide
2
8
2 (1.5 : 1 mixture of diastereomers) and ketone 23 in 85% and
4% yields, respectively. In the presence of (Ϫ)-sparteine 2,
(
ϩ)-22 and (Ϫ)-23 were isolated in 79% and 84% yields with
7
7% and 74% ee, respectively (Scheme 6). The absolute
1
(at the R-epoxide stereocentre)–rearrangement to alcohol 3
17
using (Ϫ)-sparteine 2.
The synthesis of the epoxystannanes may allow incorpor-
ation of other electrophiles using subsequent cross-coupling
26
techniques. We have used the stannylated cyclooctene oxide 15
in transmetallation reactions. Pfaltz and co-workers originally
reported the tin–lithium exchange of an epoxystannane and
subsequent electrophile trapping of the intermediate oxiranyl-
13
lithium. Using the stannylated epoxide 15 allowed electrophile
trapping in a solvent other than Et O, and as the generation of
2
the oxiranyl anion is almost instantaneous, the reaction times
are considerably reduced. Indeed, running the reaction for 5
n
min with Bu Li (1.3 equiv.) and TMEDA (2 equiv.) in THF at
Scheme 6
4
296
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 4 2 9 3 – 4 3 0 1

View Article Online
configuration of (Ϫ)-23 is as shown in Scheme 6 and is
assigned from the known sense of asymmetric induction in
the deprotonation–rearrangement of the epoxide 21 using
selective α-deprotonation of epoxides using (Ϫ)-sparteine 2,
where proton removal at the R-epoxide stereocentre is consist-
4
ently seen.
28
(
Ϫ)-sparteine 2.
Attempts to generate and trap the oxiranyl anions from
cyclohexene and cyclopentene oxides were unsuccessful (the
presumed intermediate oxiranyllithium species underwent typi-
The last 8-membered cycloalkene epoxide examined was 1,2-
epoxycyclooct-5-ene 24 (Scheme 7). Alexakis and co-workers
s
4
have reacted epoxide 24 with Bu Li–(Ϫ)-sparteine in Et O at
cal carbenoid rearrangements and reductive alkylation). Trap-
2
3
2
Ϫ90 ЊC for 6 h, followed by warming to room temperature;
under these conditions they isolated the allylic alcohol 25 in
ping an acyclic alkene epoxide, such as cis-5-decene oxide
s
with D O resulted in 90% D incorporation using Bu Li (2.5
2
30
8
5% yield and 62% ee. We hoped that under slightly different
equiv.)–(Ϫ)-sparteine 2 (2.5 equiv.) in Et O at Ϫ90 ЊC for 4 h.
2
conditions the supposed intermediate lithiated epoxide could be
trapped with deuterated methanol. However, from the reactions
However, the isolated yield of recovered epoxide was low
(below 40%), and a significant amount of volatile (E)-5-sec-
33
carried out in Et O, it appears that it is not possible at this time
butyldec-5-ene (arising from reductive alkylation) was iso-
lated (34%). This trend was confirmed by an in situ trapping
experiment with TMSCl (3 equiv.) which only led to an 18%
2
to trap the oxiranyl anion of epoxide 24 with deuterium, as
s
rearrangement to the allylic alcohol 25 is too facile. Using Bu Li
34
(
1.3 equiv.) and (Ϫ)-sparteine 2 (1.3 equiv.) in Et O at Ϫ90 ЊC
yield of the corresponding epoxysilane.
2
and quenching after 10 min gave a ratio 1 : 13, 24 : 25; if
quenched after 20 min this increased to 1 : 46. Changing from
As cycloheptene oxide 27 was successful under the desym-
metrisation conditions, other 7-membered cycloalkene oxides
were examined that possessed additional functionality (Fig. 2).
Et O to hexane did slow down the rearrangement, and after 3 h
2
at Ϫ90 ЊC, 30% D incorporation was observed (but with a 1 : 1
ratio 24/24-D : 25). Changing the ligand from (Ϫ)-sparteine 2 to
DBB 8 also reduced the rate of formation of 25, but no D
incorporation was observed in 24 (Scheme 7).
Scheme 7
Fig. 2
This research has been extended to other cycloalkene
epoxides. It proved possible to trap the oxiranyl anion derived
from cycloheptene oxide 26 using DBB 8, which required 16 h
to obtain complete D incorporation (TMEDA did not give
satisfactory results with epoxide 26). Compared with cyclooctene
oxide 1, a significantly longer reaction time was required with
Deuteration studies using CD OD were undertaken with
3
35
epoxides 29 and 30 and it was found that in Et O after 18 h at
2
s
Ϫ90 ЊC the epoxides were inert to Bu Li (either 1.25 or 2.45
equiv.)–(Ϫ)-sparteine 2 (either 1.3 or 2.5 equiv.), since the
epoxides were recovered with no deuterium incorporation.
When the reaction was run at Ϫ78 ЊC, after 1 h 18% D
incorporation was observed with epoxide 29, however after 3 h
no D incorporation was observed and multiple products had
formed. Therefore epoxide 30 was not investigated at Ϫ78 ЊC.
Epoxide 31 also proved inert at Ϫ90 ЊC after 18 h, irrespective
2
6, when (Ϫ)-sparteine was used, to obtain complete D
incorporation (12 h vs. 1 h) and also side products [cyclo-
heptanone (18%) and alkene from reductive alkylation (8%)]
4
were obtained. The lower reactivity toward α-deprotonation,
when using (Ϫ)-sparteine, of cycloheptene oxide 26 compared
with cyclooctene oxide 1 may be due to the reduced (angle)
s
of whether 1.3 or 2.5 equiv. of Bu Li–(Ϫ)-sparteine 2 was used.
Unfortunately, increasing the reaction temperature to Ϫ78 ЊC
Ϫ1
strain in 26 (angle strain has been calculated at 3.4 kJ mol for
s
for 18 h (either 1.3 or 2.5 equiv. of Bu Li–(Ϫ)-sparteine 2) had a
Ϫ1
31
2
6 and 4.7 kJ mol for 1). With DBB as the ligand, trapping
detrimental effect, as epoxide 31 was mainly decomposed to a
number of unidentified products and as the small amount of
epoxide remaining did not have any D incorporation. Lastly,
epoxide 32 was examined. It had already been shown that the
neopentyl glycol protecting group was stable under similar reac-
with 3-pentanone and N,N-dimethylpropionamide allowed the
isolation of the corresponding substituted epoxides 27 and 28,
albeit in moderate yields (52% and 54% respectively, Scheme 8).
Using (Ϫ)-sparteine 2 instead of DBB 8 as the ligand gave
similar yields of (Ϫ)-27 and (Ϫ)-28 (48% and 57%, respectively)
and ees comparable to those obtained in the cyclooctene oxide 1
series (73% and 74% ee, respectively). The absolute configur-
ations of (Ϫ)-27 and (Ϫ)-28 are tentatively assigned as shown in
Scheme 8, in line with all our previous observations on enantio-
8b
tion conditions. However, as with epoxides 29–31 at Ϫ90 ЊC,
no reaction occurred, and at Ϫ78 ЊC epoxide 32 was not
stable. The investigation of epoxides 26 and 29–32 highlights
the point that modest changes in substrate structure can have a
significant affect on reactivity in the current chemistry. The
presence of the additional functionality in epoxides 29–32
s
appears to reduce the ability for Bu Li–(Ϫ)-sparteine to effect
their α-lithiation at a low enough temperature (Ϫ90 ЊC) for the
intermediate oxiranyl anions to survive to undergo subsequent
synthetically useful nucleophilic reactions.
28
Azacyclic epoxide 33 (Scheme 9, for which the precursor
alkene was prepared by a slightly improved method—see
s
Experimental section) required Bu Li (2.5 equiv.)–(Ϫ)-sparteine
Scheme 8
Scheme 9
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 4 2 9 3 – 4 3 0 1
4297

View Article Online
2
(2.5 equiv.) to observe any reaction at Ϫ90 ЊC. The require-
performed using a Daicel Chiralcel OD column (4.6 mm × 250
s
ment for 2.5 equiv. of Bu Li–(Ϫ)-sparteine 2 in order to observe
reactivity with azacyclic epoxide 33 may be due to complex-
ation of 1 equiv. of the base–ligand with the N-Boc function-
ality. However, lithiation at Ϫ90 ЊC for 12 h only led to the
mm) or Daicel Chiralpak AD column (4.6 mm × 250 mm) on a
Gilson System with 712 Controller Software and a 118 UV–Vis
detector set at 254 nm. Chiral GC was performed using a
ThermoQuest CE Instruments TRACE GC, running Chrom-
Card for TRACE software, fitted with a Chrompack-CP-
Chiralsil-DEX-CB column at the stated temperature/temper-
ature gradient. Retention times for major (tR mj) and minor
28
rearranged ester 34 (45%), previously reported (Scheme 9).
In summary, the present work illustrates a conceptually
different approach to previous asymmetric desymmetrisation
strategies of meso-epoxides: in the current work a chemical
transformation occurs selectively at one of the enantiotopic
epoxide termini, but the useful epoxide functional group is
retained in the product, resulting in a new method for chiral
epoxide synthesis. The results described herein demonstrate the
process is currently viable in cyclooctene-derived epoxides, and
for cycloheptene oxide. For the cyclooctene-derived epoxides,
the ligand used in combination with the organolithium to effect
deprotonation plays a key role in allowing the deprotonation to
occur at low temperature. The ligand may also sterically ‘shield’
the lithiated epoxide intermediate from (currently) undesired
decomposition pathways, such that the usual reaction pathway
for these particular intermediates (transannular C–H inser-
(t mn) enantiomers are given in minutes.
R
3,7-Dibutyl-3,7-diazabicyclo[3.3.1]nonane 8. To neat 3,7-
20
Dibutyl-3,7-diaza-9-oxobicyclo[3.3.1]nonane 7 (11.3 g, 44.7
mmol) was added hydrazine monohydrate (11.2 ml, 11.5 g,
223.5 mmol). The mixture was refluxed for 2 h, then cooled to
room temperature, diluted with Et O (200 ml), washed with
2
water (2 × 200 ml), brine (100 ml) and dried. To the resultant
yellow oil was added powdered KOH (6.76 g, 49 mmol). The
mixture was heated at 90 ЊC for 12 h and then distilled under
reduced pressure (0.08 mbar, 115 ЊC) to yield 3,7-dibutyl-3,7-
diazabicyclo[3.3.1]nonane 8 (5.2 g, 49%) as a colourless oil
whose spectral data were consistent with those reported
4
20
tion) does not occur. Another important factor in the success
earlier.
of the current chemistry is that the deprotonation can be
achieved with just over one equivalent of base. Compared with
the cyclooctene epoxides, cycloheptene oxide is somewhat less
susceptible to deprotonation, however in the presence of a
diamine this can be achieved without significant further reac-
tion of the intermediate lithiated epoxide. However, substituted
cycloheptene oxides, cyclohexene and cyclopentene oxide, and
a representative acyclic alkene oxide are currently not viable
substrates for this chemistry. With these substrates, the ligands
examined thus far are not capable of assisting α-lithiation to
give controlled generation of oxiranyl anions.
Once allowed to cool to 0 ЊC, the distillation residue was
carefully diluted with conc. HCl (50 ml). The mixture was
refluxed for 12 h, cooled to room temperature, diluted with
water and extracted with Et O. The aqueous layer was basified
2
to pH 11 with KOH and then extracted with Et O (3 × 100 ml).
2
The combined organic layers were washed with brine, dried
and evaporated under reduced pressure giving recovered
3,7-dibutyl-3,7-diaza-9-oxobicyclo[3.3.1]nonane 7 (3.3 g).
Typical procedure for lithiation–electrophile trapping of
cyclooctene oxide 1 in the presence of a diamine
While this paper reports the first steps toward a new
approach for the elaboration of simple achiral epoxides to
chiral epoxides, a significant challenge remains in expanding the
scope of the chemistry so that it becomes viable with a wider
range of substrates. Future progress in this area will likely
require a greater understanding of the factors that affect
epoxide deprotonation, along with those factors that contribute
to the stability of the oxiranyl anion intermediate, yet still
allow the latter to function effectively in electrophile trapping
reactions.
The diamine (2.6 mmol) was added dropwise to a solution of
RLi (1.4 M, 2.5 mmol) in Et O (8 ml) at Ϫ90 ЊC. This mixture
2
was stirred for 1 h at Ϫ90 ЊC. A solution of cyclooctene oxide 1
(
2.0 mmol) in Et O (2 ml), pre-cooled to Ϫ90 ЊC, was then
2
added rapidly by cannula to the solution of ligand–RLi and the
reaction mixture was then stirred at Ϫ90 ЊC for 3 h. Neat
electrophile (3.0 mmol) was then added dropwise, and the reac-
tion was then allowed to warm to room temperature over 5 h.
After this time H PO (0.5 M, 25 ml) was added, the organic
3
4
phase was washed with NaHCO (25 ml) and brine (25 ml). The
3
aqueous layers were extracted twice with Et O (25 ml) and the
2
Experimental
combined organic phases were dried and evaporated under
reduced pressure. Purification of the residue by column chrom-
atography (SiO , Et O in petrol) gave the substituted epoxide.
General directions
2
2
All reactions requiring anhydrous conditions were conducted in
flame- or oven-dried apparatus under an atmosphere of argon.
Syringes and needles for the transfer of reagents were dried at
(9-Oxabicyclo[6.1.0]non-1-yl)phenylmethanol 9. R (pentane–
f
Ϫ1
2
4
ether: 9 : 1) 0.05; [α]
2927s, 2858m, 1469m, 1453m, 1066m, 1043m, 1022m, 929m,
701m; δ (200 MHz) 7.41–7.26 (10 H, m), 4.78 (1 H, d, J 2.1),
D
ϩ20.1 (c = 1.0 in CHCl ); νmax/cm 3428s,
3
1
40 ЊC and allowed to cool in a desiccator over P O before use.
2 5
Ethers were distilled from sodium benzophenone ketyl under
argon; CH Cl and benzene from CaH under argon. External
H
4.70 (1 H, d, J 8.2), 3.34 (1 H, dd, J 10.1, 4.6), 2.98 (1 H, s),
2.70–2.63 (2 H, m), 2.33–2.23 (2 H, m), 2.08–2.02 (2 H, m),
2
2
2
reaction temperatures are reported unless stated otherwise.
Reactions were monitored by TLC using commercially avail-
able aluminium-backed plates, pre-coated with a 0.25 mm layer
of silica containing a fluorescent indicator (Merck). Organic
1.69–1.20 (20 H, m); δ
128.1, 128.0, 127.4, 127.3, 75.3, 72.1, 64.5, 64.1, 60.2, 57.8, 28.5,
27.8, 26.8, 26.7, 26.1, 26.0, 25.8, 25.2, 24.6; m/z [CI ϩ (NH )]
requires 250.1807), 233
(100%), 215 (45), 199 (32). The ee was determined to be 76%
after MnO oxidation [MnO (16 equiv.), CH Cl , 25 ЊC, 12 h]
to ketone 10; chiral GC as for 10 below.
(50 MHz) 141.3, 139.8, 128.7, 128.2,
C
3
ϩ
layers were dried over MgSO unless stated otherwise. Column
250.1804 (M ϩ NH . C15H24NO
4 2
4
chromatography was carried out on Kieselgel 60 (40–63 µm).
Light petroleum refers to the fraction with bp 40–60 ЊC. [α]
Values are given in 10 deg cm g . IR spectra were recorded
2
2
2
2
D
Ϫ1
2
Ϫ1
as thin films unless stated otherwise. Peak intensities are speci-
1
13
fied as strong (s), medium (m) or weak (w). H and C NMR
(9-Oxabicyclo[6.1.0]non-1-yl)phenylmethanone 10. R (pent-
f
24
spectra were recorded in CDCl unless stated otherwise with a
Bruker DPX 250, JEOL EX400 or Bruker AMX500 spectro-
ane–ether: 9 : 1) 0.7; [α] Ϫ91.5 (c = 1.0 in CHCl ) (from reaction
D
3
3
s Ϫ1
with Bu Li); νmax/cm 2929m, 1679s, 1468m, 1448m, 1209s,
meter. Chemical shifts are reported relative to CHCl [δ 7.26,
1177s, 710m; δ (200 MHz) 8.11–8.05 (2 H, m), 7.61–7.41 (3 H,
3
H
H
δC (central line of t) 77.0]. Coupling constants (J) are given in
m), 3.12 (1 H, dd, J 10.0, 4.0), 2.49–2.41 (1 H, m), 2.29–2.21
Hz. OH signals were assigned by the absence of cross-peaks in
H– C correlation spectra. Chiral stationary phase HPLC was
(1 H, m), 1.83–1.53 (10 H, m); δ (50 MHz) 197.8, 134.3, 133.3,
129.4, 128.4, 65.5, 59.8, 29.1, 26.7, 26.0, 25.8, 25.5, 25.0; m/z [CI
C
1
13
4
298
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 4 2 9 3 – 4 3 0 1

View Article Online
ϩ
4
ϩ (NH )] 248.1653 (M ϩ NH . C H NO requires 248.1651),
m), 0.15 (9 H, s); δ (50 MHz) 58.2, 47.6, 30.9, 30.4, 27.0, 26.9,
3
15 22
2
C
2
30 (25%), 214 (30), 105 (70), 77 (100). The ee was determined
24.9, 22.6, Ϫ10.9; δ (186.4 MHz) Ϫ2.6; m/z [CI ϩ (NH )]
Sn
3
Ϫ1
ϩ
to be 77% by chiral GC (130 ЊC, 0.7 ml min ): t mn 83.7, t mj
290.0689 (M . C H OSn requires 290.0693), 294 (20%), 292
R
R
11 22
8
5.6.
(15), 290 (100), 289 (35), 288 (75), 286 (45), 287 (30), 291 (10).
The ee was determined to be 78% after reaction with BuLi and
trapping with 3-pentanone to give alcohol 12 [Bu Li (1.3
n
1
-(9-Oxabicyclo[6.1.0]non-1-yl)propan-1-ol 11. R (pentane–
f
2
4
Ϫ1
ether: 9 : 1) 0.1; [α]
D
Ϫ25.3 (c = 1.0 in CHCl ); νmax/cm 3430m,
equiv.), TMEDA (2 equiv.), 3-pentanone (1.6 equiv.), THF,
3
2
920s, 2860m, 1465w, 1106m, 976m, 926m, 836m; δH (400
MHz) 3.66 (1 H, dd, J 8.8, 2.9), 3.51 (1 H, dd, J 9.5, 3.8), 3.14
1 H, dd, J 9.8, 4.8), 2.94 (1 H, dd, J 9.7, 4.6), 2.26–2.03 (6 H,
m), 1.77–1.29 (24 H, m), 1.00 (6 H, t, J 7.4); δ (50 MHz) 74.4,
Ϫ90 ЊC, 10 min]; chiral GC as for 12 above.
(
Trimethyl(9-oxabicyclo[6.1.0]non-1-yl)silane 17. R (pentane–
f
2
4
Ϫ1
ether: 9 : 1) 0.8; [α]
D
Ϫ33.1 (c = 1.0 in CHCl ); νmax/cm 2928s,
C
3
7
2
0.0, 64.0, 63.7, 60.5, 58.0, 28.3, 27.7, 27.0, 26.8, 26.4, 26.3,
1471m, 1448m, 1247m, 1019s; δH (200 MHz) 2.81–2.56 (1 H,
6.2, 26.1, 26.0, 25.8, 25.4, 25.3, 24.9, 10.7, 9.8; m/z [CI ϩ
m), 2.18–2.05 (2 H, m), 1.63–1.25 (10 H, m), 0.08 (9 H, s); δ (50
C
ϩ
(
NH )] 202.1805 (M ϩ NH . C H NO requires 202.1807),
MHz) 58.9, 55.7, 29.1, 26.8, 26.5, 26.4, 26.3, 25.2, Ϫ3.3; m/z [CI
3
4
11 24
2
ϩ
1
7
85 (20%), 167 (100), 151 (90). The ee was determined to be
6% after Dess–Martin oxidation [periodane (1.1 equiv.),
ϩ (NH )] 199.1517 (M ϩ H . C H OSi requires 199.1518),
3
11 22
183 (10%), 169 (7), 155 (10), 143 (17), 109 (40), 90 (100). The ee
Ϫ1
CH Cl , 25 ЊC, 20 min] to ketone 13; chiral GC as for 13 below.
was determined to be 74% by chiral GC (80 ЊC, 0.5 ml min )
2
2
t mn 97.4, t mj 101.4.
R
R
3
-(9-Oxabicyclo[6.1.0]non-1-yl)pentan-3-ol 12. R (pentane–
f
24
ether: 9 : 1) 0.2; [α]
with Bu Li); νmax/cm 3515s, 2968s, 2936m, 1461m, 1369w,
301m, 1151m, 1038s, 961m, 920m; δH (400 MHz) 2.97–2.93
D
Ϫ41.6 (c = 1.0 in CHCl ) (from reaction
[4,5-Bis(tert-butyldimethylsilyloxy)-9-oxabicyclo[6.1.0]-
3
s
Ϫ1
24
D
nonan-1-yl]phenylmethanol 22. R (pentane–ether: 8 : 2) 0.4; [α]
f
Ϫ1
1
ϩ3.0 (c = 1.0 in CHCl ); νmax/cm 3515m, 2960s, 2929m,
3
(
1 H, m), 2.55 (1 H, s), 2.13–2.03 (2 H, m), 1.65–1.31 (14 H, m),
2896m, 2847m, 1465m, 1252m, 1179m, 1030w, 890m, 710m;
0
.89 (3 H, t, J 7.4), 0.82 (3 H, t, J 7.4); δ (50 MHz) 73.4, 65.5,
δ (C D , 250 MHz, 90 ЊC) 7.38–7.30 (2 H, m), 7.25–7.11 (3 H,
C
H
7
8
5
6.2, 31.4, 28.8, 27.9, 27.6, 26.4, 26.3, 26.2, 24.7, 8.0, 7.3; m/z
m), 4.42–4.39 (1 H, br s), 4.05–3.90 (2 H, m), 2.66 (1 H, t, J 7.0),
1.99–1.40 (8 H, m), 0.97 (18 H, s), 0.14 (3 H, s), 0.08 (3 H, s),
0.07 (6 H, s); δ (C D , 62.5 MHz, 90 ЊC) 141.5, 128.3, 127.8,
ϩ
[CI ϩ (NH )] 230.2123 (M ϩ NH . C H NO requires
3
4
15 28
2
2
30.2120), 213 (57%), 195 (100), 179 (90). The ee was deter-
C
7
8
Ϫ1
mined to be 77% by chiral GC (130 ЊC, 0.5 ml min ): t mn
127.3, 78.4, 78.1, 76.2, 63.6, 60.3, 58.7, 31.5, 30.7, 26.0, 24.0,
R
2
8.8, t mj 30.7.
23.7, 18.3, 18.2, Ϫ4.5, Ϫ4.7, Ϫ4.8; m/z [CI ϩ (NH )] 493.3174
R
3
ϩ
(
M ϩ H . C H O Si requires 493.3169), 515 (100%), 493 (65),
27
49
4
2
1
-(9-Oxabicyclo[6.1.0]non-1-yl)propan-1-one 13. R (pent-
Ϫ1
454 (55), 413 (50), 391 (31), 361 (30), 342 (60), 229 (65). The ee
was determined to be 77% after MnO oxidation [MnO (16
f
2
4
ane–ether: 9 : 1) 0.6; [α] ϩ27.5 (c = 1.0 in CHCl ); νmax/cm
D
3
2
2
2
1
974s, 2919m, 2856m, 1708s, 1464m, 1448w, 1404m, 1117m,
equiv.), CH Cl , 25 ЊC, 12 h] to ketone 23; chiral HPLC as for 23
below.
2 2
027s, 924m; δ (200 MHz) 3.03 (1 H, dd, J 9.5, 3.8), 2.51–2.20
H
(
4 H, m), 1.61–1.24 (10 H, m), 0.94 (3 H, t, J 7.3); δ (50 MHz)
C
2
10.8, 64.1, 60.4, 30.2, 28.9, 27.9, 26.6, 26.4, 26.1, 24.6, 7.4;
[4,5-Bis(tert-butyldimethylsilyloxy)-9-oxabicyclo[6.1.0]-
ϩ
m/z [CI ϩ (NH )] 200.1653 (M ϩ NH . C H NO requires
2
to be 78% by chiral GC (130 ЊC, 0.7 ml min ): t mn 8.2, t mj
nonan-1-yl]phenylmethanone 23. R (pentane–ether: 96 : 4) 0.5;
3
4
11 22
2
f
2
4
Ϫ1
00.1651), 183 (7%), 167 (100), 137 (20). The ee was determined
[α]
D
Ϫ31.2 (c = 1.0 in CHCl ); νmax/cm 2955s, 2924m, 2889m,
3
Ϫ1
2853m, 1680s, 1468m, 1247m, 1174m, 1025w, 894m, 704m;
R
R
9
.4.
Ethyl 9-oxabicyclo[6.1.0]nonane-1-carboxylate 14. R (pent-
Ϫ1
δ (C D , 250 MHz, 90 ЊC) 8.12 (2 H, br d, J 7.5), 7.20 (1 H, br
H
7
8
d, J 6.8), 7.17 (2 H, br d, J 7.5), 4.35–4.19 (2 H, m), 2.92 (1 H,
dd, J 9.0, 4.3), 2.18–1.51 (8 H, m), 0.99 (9 H, s), 0.97 (9 H, s),
0.18 (6 H, s), 0.13 (6 H, s); δ (C D , 62.5 MHz, 90 ЊC) 197.5,
f
2
4
ane–ether: 9 : 1) 0.7; [α] Ϫ35.4 (c = 1.0 in CHCl ); νmax/cm
D
3
C
7
8
2
1
919s, 2859m, 1733s, 1467m, 1448w, 1306m, 1201m, 1156m,
051s, 935m, 849m; δH (200 MHz) 4.18 (2 H, q, J 7.2), 3.28
135.1, 133.0, 129.8, 128.4, 77.4, 76.4, 65.4, 59.3, 31.0, 30.8, 26.1,
25.2, 23.3, 18.3, 18.2, Ϫ4.5, Ϫ4.6, Ϫ4.8, Ϫ4.9; m/z [CI ϩ (NH )]
3
ϩ
(
1 H, dd, J 10.0, 4.1), 2.50–2.41 (1 H, m), 2.24–2.14 (1 H, m),
491.3020 (M ϩ H . C H O Si requires 491.3013), 475 (40%),
27
47
4
2
1
6
.79–1.29 (10 H, m), 1.26 (3 H, t, J 7.2); δ (50 MHz) 170.8,
(12%), 105 (100%), 90 (95%). The ee was determined to be 74%
3 Ϫ1
C
0.9, 60.0, 59.2, 27.5, 26.7, 26.6, 26.5, 26.0, 24.7, 14.0; m/z [CI ϩ
by chiral HPLC (OD column, 100% heptane, 0.8 cm min ):
ϩ
(
NH )] 216.1600 (M ϩ NH . C H NO requires 216.1600),
t mn 15.8, t mj 30.5.
3
4
11 22
3
R
R
1
99 (60%), 183 (70), 125 (20). The ee was determined to be 81%
Ϫ1
by chiral GC (110 ЊC, 0.7 ml min ): t mn 28.3, t mj 29.8.
3-(8-Oxabicyclo[5.1.0]oct-1-yl)pentan-3-ol 27. R (pentane–
R
R
f
2
4
Ϫ1
ether: 8 : 2) 0.4; [α]
D
Ϫ18.9 (c = 1.0 in CHCl ); νmax/cm 3475m,
3
Tributyl(9-oxabicyclo[6.1.0]non-1-yl)stannane 15. R (pent-
Ϫ1
2924s, 2860m, 1459w, 1165w, 957w, 916w; δH (250 MHz) 3.34
(1 H, dd, J 6.7, 1.9), 2.14–1.42 (14 H, m), 0.92 (3 H, t, J 7.5),
0.91 (3 H, t, J 7.5); δ (62.5 MHz, CDCl ) 74.6, 66.3, 56.6, 31.5,
f
2
4
ane–ether: 9 : 1) 0.8; [α] Ϫ38.2 (c = 1.0 in CHCl ); νmax/cm
D
3
2
1
(
958s, 2919s, 2848m, 1472m, 1456m, 1413m, 1373w, 1070s,
C
3
019m, 921m; δ (500 MHz) 2.81 (1 H, dd, J 10, 4.5), 2.18–2.15
31.3, 29.9, 28.3, 27.9, 24.4, 23.7, 8.3, 8.1; m/z [CI ϩ (NH )]
H
3
ϩ
2 H, m), 1.63–1.33 (23 H, m), 0.96–0.91 (14 H, m); δC (50
MHz) 62.5, 59.4, 32.3, 29.1, 27.4, 26.9, 26.7, 26.6, 26.2, 25.8,
3.6, 8.8; δ (186.4 MHz) Ϫ26.0; m/z [CI ϩ (NH )] 416.2098
199.1701 (M ϩ H . C H O requires 199.1698), 216 (40%),
12
23
2
199 (40), 198 (40), 181 (98), 165 (100). The ee was determined to
Ϫ1
1
(
(
be 73% by chiral GC (130 ЊC, 0.5 ml min ): t mn 20.5, t mj
Sn
3
R
R
ϩ
M . C H OSn requires 416.2101), 420 (20%), 418 (15), 417
23.1.
20
40
20), 416 (100), 415 (40), 414 (75), 413 (30), 412 (45). The ee was
determined to be 79% after reaction with BuLi and trapping
with 3-pentanone to give alcohol 12 [Bu Li (1.3 equiv.),
1-(8-Oxabicyclo[5.1.0]oct-1-yl)propan-1-one 28. R (pentane–
f
n
24
D
Ϫ1
ether: 96 : 4) 0.45; [α]
Ϫ10.5 (c = 1.0 in CHCl ); νmax/cm
3
TMEDA (2 equiv.), 3-pentanone (1.6 equiv.), THF, Ϫ90 ЊC,
2975s, 2927m, 2846m, 1705s, 1454m, 1366w, 1199m, 1090m;
1
0 min]; chiral GC as for 12 above.
δH (250 MHz) 3.19 (1 H, dd, J 6.0, 3.7), 2.55–2.25 (3 H, m),
2
.06–1.44 (9 H, m), 1.02 (3 H, t, J 7.4); δ (62.5 MHz) 212.6,
C
Trimethyl(9-oxabicyclo[6.1.0]non-1-yl)stannane 16. R (pent-
66.6, 60.6, 31.5, 29.7, 29.1, 28.8, 25.3, 24.4, 7.9; m/z [CI ϩ
f
24
Ϫ1
ϩ
ane–ether: 9 : 1) 0.8; [α]
D
Ϫ34.5 (c = 1.0 in CHCl ); νmax/cm
(NH )] 169.1232 (M ϩ H . C H O requires 169.1229), 186
3
3
10 17
2
2
925s, 2859s, 1474m, 1449m, 1184m, 1021s, 922m; δH (200
(100%), 169 (70). The ee was determined to be 74% by chiral
Ϫ1
MHz) 2.83–2.76 (1 H, m), 2.20–2.12 (2 H, m), 1.60–1.38 (10 H,
GC (110 ЊC, 0.5 ml min ): t mn 12.4, t mj 14.7.
R
R
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 4 2 9 3 – 4 3 0 1
4299

View Article Online
Typical procedure for the synthesis of ꢀ-alkylcyclooctene oxide
organic layers were washed with water (30 ml), saturated aq.
NaHCO (2 × 20 ml), brine (20 ml), dried and evaporated under
3
To a solution of tributyl(9-oxabicyclo[6.1.0]non-1-yl)stannane
reduced pressure to give a colourless oil that solidified on stand-
ing. The product was purified by repeated trituration with cold
light petroleum to give the epoxide 32 (399 mg, 63%) as a
1
5 (0.5 mmol) and TMEDA (2 equiv., 1.0 mmol) in THF (2.5
n
ml) at Ϫ90 ЊC was added Bu Li (1.3 equiv., 0.65 mmol). The
alkyl iodide (3 equiv., 1.5 mmol) was then added in one portion
after 0.5 min and the reaction mixture was stirred at this tem-
perature for 10 min. After quenching with H PO (0.5 M, 15
colourless solid (mp close to rt); R (petroleum ether–ether:
f
Ϫ1
7
9
2
4
6
: 3) 0.21; νmax/cm 3430s (hydroscopic), 2959s, 2857s, 1107m,
3
4
16m; δ (400 MHz) 3.44 (2 H, s, CH O), 3.40 (2 H, s, CH O),
H
2
2
ml), and extraction with Et O (3 × 10 ml), the combined
2
.98–3.04 (2 H, m, 2 × CH epoxide), 1.67–2.00 (8 H, m,
× CH ), 0.88 (6 H, s, 2 × CH ); δ (100 MHz) 99.4, 70.0,
organic phases were washed with brine (15 ml), dried and evap-
orated under reduced pressure. Purification of the residue by
column chromatography (SiO , 5% Et O in petrol) gave the
2
3
C
9.9, 55.1, 30.0, 29.6, 22.6, 21.5; m/z [CI ϩ (NH )] 213.1489
3
2
2
ϩ
(
M ϩ H . C H O requires 213.1491), 142 (40%), 128 (30).
12 21 3
alkylated epoxide.
(
Z )-N-(4-Methylbenzylsulfonyl)-2,3,4,7,8,9-hexahydroazo-
ꢀ
-Methylcyclooctene oxide 18. R (pentane–ether: 9 : 1) 0.50;
f
nine. A 3 l three-necked round-bottomed-flask was equipped
with a reflux condenser and one dropping funnel. The flask was
charged with toluene (600 ml), Bu NI (1.14 g, 3.1 mmol),
NaOH (18.6 g, 0.46 mol) and water (44 ml). The funnel was
charged with a solution of (Z)-1,8-bis(4-methylbenzylsulfonyl-
oxy)oct-4-ene (1 g, 2.21 mmol) in toluene (125 ml). The funnel
was allowed to discharge the ditosylate dropwise into the reac-
Ϫ1
νmax/cm 2930s, 1724w, 1263w, 1216w; δ (400 MHz) 2.66 (1 H,
dd, J 10.3, 4.4), 2.17–2.10 (1 H, m), 1.91–1.85 (1 H, m), 1.66–
.32 (10 H, m), 1.30 (3 H, s); δ (50 MHz) 63.2, 59.79, 31.28,
7.87, 26.67, 26.05, 25.31, 21.58; m/z [CI ϩ (NH )] 141.1276 (M
ϩ H . C H O requires 141.1279), 158 (100%), 141 (83), 140
65).
H
4
1
2
C
3
ϩ
9
17
(
tion flask over a period of 4 h and a solution of TsNH (644 mg,
2
ꢀ
-Butylcyclooctene oxide 19. R (pentane–ether: 95 : 5) 0.61;
f
3.76 mmol) in hot toluene (125 ml, 70 ЊC) was added to the
reaction flask by cannula at the same rate. A gentle reflux was
maintained over 12 h and a second addition of (Z)-1,8-bis-
Ϫ1
νmax/cm 2929s, 2859w, 1468w, 1260w, 1023w, 804w; δH (400
MHz) 2.69 (1 H, dd, J 9.9, 4.5), 2.03 (1 H, dt, J 13.8, 3.4), 2.19–
.11 (1 H, m), 1.94–1.85 (1 H, m), 1.71–1.17 (15 H, m), 0.91
3 H, t, J 7.0); δ (50 MHz) 65.8, 62.2, 33.7, 28.2, 27.9, 26.7,
6.3, 26.2, 25.8, 25.4, 23.0, 14.1; m/z [CI ϩ (NH )] 200.2014
3
ϩ
M ϩ NH . C H NO requires 200.2021), 183 (53%), 165
2
(
2
(
(
4-methylbenzylsulfonyloxy)oct-4-ene and TsNH was similarly
2
C
performed. After 10 h of gentle reflux, a third addition was
performed and a fourth after 12 h. After 24 h at reflux, the
solution was allowed to cool to room temperature. The organic
layer was separated and the aqueous phase was extracted with
4 12 26
(
100).
Et O (3 × 125 ml). The combined organic layers were washed
with brine (125 ml), dried and evaporated under reduced pres-
sure. Purification of the residue by column chromatography
2
tert-Butyl 8-oxa-4-azabicyclo[5.1.0]octane-4-carboxylate 31.
Peracetic acid (1 ml, 36% w/w in dilute AcOH; 4.8 mmol) was
added to a stirred mixture of tert-butyl 2,3,6,7-tetrahydro-
(from 0 to 30% Et O in petrol) gave a crystalline solid (1.8 g,
2
36
azepine-1-carboxylate (602 mg, 3 mmol), Na CO (1.3 g, 12
72%) for which spectral and physical data were identical with
2
3
28
mmol) and NaOAc (13 mg, 0.14 mmol) in CH Cl (20 ml) at 0
those reported earlier.
2
2
ЊC. The reaction was warmed to 25 ЊC and stirred for a further
5 h. Et O (30 ml) was added and the organic layers were
1
2
Acknowledgements
washed with water (30 ml), saturated aq. NaHCO (2 × 20 ml),
3
We thank the EPSRC for a Research Grant (GR/N26302), the
European Union for a Marie Curie Fellowship (HPMF-CT-
brine (20 ml), dried and evaporated under reduced pressure
to give epoxide 31 (520 mg, 81%) as a colourless oil; R (light
f
Ϫ1
2
000-00558), the EPSRC and Rhône-Poulenc Rorer for a
petroleum–ether: 7 : 3) 0.18; νmax/cm 2975s, 1694s, 1417s,
CASE award (to I. D. C.), and the EPSRC National Mass
Spectrometry Service Centre for mass spectra.
1
2
2
5
(
1
336s, 1172s, 944s; δ (400 MHz) 3.17 (2 H, br s, epoxide), 2.56–
H
.89 and 3.5–3.86 (4 H, 2 × m, 2 × CH N), 1.92–2.21 (4 H, m,
2
× CH ), 1.41 (9 H, s, 3 × CH ); δ (100 MHz) 56.0, 55.8,
2
3
C
3.4, 43.4, 43.1, 28.9, 28.4, 28.1; m/z [CI ϩ (NH )] 214.1438
3
References
ϩ
M ϩ H . C H NO requires 214.1443), 175 (40%), 158 (40),
11 20 3
14 (100).
1 M. C. Willis, J. Chem. Soc., Perkin Trans. 1, 1999, 1765–1784.
2
D. M. Hodgson, A. R. Gibbs and G. P. Lee, Tetrahedron, 1996, 52,
4361–14384.
A. Magnus, S. K. Bertilsson and P. G. Andersson, Chem. Soc. Rev.,
1
3
,3-Dimethyl-1,5-dioxaspiro[5.6]dodec-9-ene oxide 32. Neo-
3
pentyl glycol (832 mg, 8 mmol) was added to a stirred mixture
2
002, 31, 223–229.
37
of 4-cyclohepten-1-one (770 mg, 7 mmol) and PTSA (27 mg,
.14 mmol) in benzene (50 ml). The reaction was refluxed using
4 D. M. Hodgson and E. Gras, Synthesis, 2002, 1625–1642.
0
5 D. M. Hodgson, K. Tomooka and E. Gras, Top. Organomet. Chem.,
2
003, 5, 217–250.
a Dean–Stark trap for 24 h. Light Petroleum (30 ml) was added
and the organic layer was washed with saturated aq. NaHCO3
6
7
E. N. Jacobsen, Acc. Chem. Res., 2000, 33, 421–431.
S. Matsunaga, J. Das, J. Roels, E. M. Vogl, N. Yamamoto, T. Iida,
K. Yamaguchi and M. Shibasaki, J. Am. Chem. Soc., 2000, 122,
(
2 × 20 ml), brine (20 ml), dried and evaporated under reduced
pressure to give 3,3-dimethyl-1,5-dioxaspiro[5.6]dodec-9-ene
775 mg, 57%) as a colourless oil, which was used in the next
step without further purification; R (light petroleum–ether:
2
252–2260.
(
8 (a) D. M. Hodgson, C. R. Maxwell, T. J. Miles, E. Paruch, M. A. H.
Stent, I. R. Matthews, F. X. Wilson and J. Witherington, Angew.
Chem., Int. Ed., 2002, 41, 4313–4316; D. M. Hodgson, C. R.
Maxwell, T. J. Miles, E. Paruch, M. A. H. Stent, I. R. Matthews,
F. X. Wilson and J. Witherington, Angew. Chem., Int. Ed., 2002, 41,
f
Ϫ1
5
: 2) 0.75; νmax/cm 2953s, 2866s, 1110s, 912s, 733s; δH (400
MHz) 5.48–5.55 (2 H, m, 2 × CH᎐), 3.46 (4 H, s, 2 × CH O),
᎐
2
2
(
2
.02–2.09 (4 H, m, 2 × CH ), 1.85–1.93 (4 H, m, 2 × CH ), 0.92
2
2
611 (corrigendum); (b) D. M. Hodgson, M. A. H. Stent, B. Stefane
ˇ
4
6 H, s, 2 × CH ); δ (100 MHz) 132.3, 131.9, 106.9, 70.3, 34.1,
3
C
and F. X. Wilson, Org. Biomol. Chem., 2003, 1, 1139–1150.
9 I. Erden, in Comprehensive Heterocyclic Chemistry II, Vol. 1A,
eds. A. R. Katritzky, C. W. Rees and E. F. V. Scriven, Pergamon,
Oxford, 1996, pp. 97–171.
0 R. A. Johnson and K. B. Sharpless, in Catalytic Asymmetric
Synthesis, ed., I. Ojima, Wiley-VCH, New York, 2000, pp. 231–285.
1 E. N. Jacobsen and M. H. Wu, in Comprehensive Asymmetric
Catalysis, eds. E. N. Jacobsen, A. Pfaltz and H. Yamamoto,
Springer-Verlag, Berlin, 1999, pp. 649–677.
ϩ
8.4, 23.4, 23.0; m/z [CI ϩ (NH )] 197.1539 (M ϩ H . C H O
3 12 21 2
requires 197.1542), 128 (50%). Peracetic acid (1.4 ml, 36% w/w
in dilute AcOH; 6.4 mmol) was added to a stirred mixture of
,3-dimethyl-1,5-dioxaspiro[5.6]dodec-9-ene (775 mg, 4 mmol),
1
1
3
Na CO (1.7 g, 16 mmol) and NaOAc (18 mg, 0.19 mmol) in
2
3
CH Cl (20 ml) at 0 ЊC. The reaction was warmed to 25 ЊC and
2
2
stirred for a further 15 h. Et O (30 ml) was added and the
2
4
300
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 4 2 9 3 – 4 3 0 1

View Article Online
1
1
1
1
2 J. J. Eisch and J. E. Galle, J. Am. Chem. Soc., 1976, 98, 4646–
648.
3 P. Lohse, H. Loner, P. Acklin, F. Sternfeld and A. Pfaltz, Tetrahedron
Lett., 1991, 32, 615–618.
4 T. Satoh, S. Kobayashi, S. Nakanishi, K. Horigushi and S. Irisa,
Tetrahedron, 1999, 55, 2515–2528.
5 D. M. Hodgson and S. L. M. Norsikian, Org. Lett., 2001, 3, 461–
25 P. F. Hudrlik and A. M. Hudrlik, in Advances in Silicon Chemistry,
4
ed. G. L. Larson, JAI, Greenwich, 1993, pp. 1–89.
26 J. R. Falck, R. K. Bhatt, K. M. Reddy and J. Ye, Synlett, 1997, 481–
482.
27 H. J. Reich and N. H. Philips, J. Am. Chem. Soc., 1986, 108, 2102–
2103.
28 D. M. Hodgson, I. D. Cameron, M. Christlieb, R. Green, G. P. Lee
and L. A. Robinson, J. Chem. Soc., Perkin Trans. 1, 2001, 2161–
2174.
29 (a) D. M. Hodgson, J.-M. Galano and M. Christlieb, Chem.
Commun., 2002, 2436–2437; (b) D. M. Hodgson, J.-M. Galano and
M. Christlieb, Tetrahedron, 2003, in press.
4
63; also see D. M. Hodgson, N. J. Reynolds and S. J. Coote,
Tetrahedron Lett., 2002, 43, 7895–7897.
1
6 Preliminary communication: D. M. Hodgson and E. Gras, Angew.
Chem., Int. Ed., 2002, 41, 2376–2378.
1
7 D. M. Hodgson, G. P. Lee, R. E. Marriott, A. J. Thompson,
R. Wisedale and J. Witherington, J. Chem. Soc., Perkin Trans. 1,
30 A. Alexakis, E. Vrancken and P. Mangeney, J. Chem. Soc., Perkin
Trans. 1, 2000, 3354–3355.
1
998, 2151–2161.
1
1
8 D. B. Collum, Acc. Chem. Res., 1992, 25, 448–454.
9 J. Clayden, Organolithiums: Selectivity for Synthesis, Oxford, 2002,
pp. 3–8.
31 K. M. Morgan and S. Gronert, J. Org. Chem., 2000, 65, 1461–1466.
32 M. O. Brimeyer, A. Mehrota, S. Quici, A. Nigam and S. L. Regen,
J. Org. Chem., 1980, 45, 4254–4255.
33 (a) E. Doris, L. Dechoux and C. Mioskowski, Synlett, 1998, 337–
343; (b) E. Doris, L. Dechoux and C. Mioskowski, Tetrahedron
Lett., 1994, 35, 7943–7946.
34 K. Tamao and K. Maeda, Tetrahedron Lett., 1986, 27, 65–68.
35 D. M. Hodgson, L. A. Robinson and M. L. Jones, Tetrahedron Lett.,
1999, 40, 8637–8640.
36 D. I. MaGee and E. J. Beck, J. Org. Chem., 2000, 65, 8367–8371.
37 R. L. Danheiser, J. M. Morin, Jr. and E. J. Salaski, J. Am. Chem.
Soc., 1985, 107, 8066–8073.
2
0 K. M. B. Gross, Y. M. Jun and P. Beak, J. Org. Chem., 1997, 62,
7
679–7689.
2
1 J. Spieler, O. Huttenloch and H. Waldmann, Eur. J. Org. Chem.,
000, 391–399.
2
2
2
2 Y. Miyahara, K. Goto and T. Inazu, Synthesis, 2001, 364–366.
3 K.-J. Haack, R. Goddard and K.-R. Pörschke, J. Am. Chem. Soc.,
1
997, 119, 7992–7999.
4 G. A. Molander and K. Mautner, J. Org. Chem., 1989, 54, 4042–
050.
2
4
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 4 2 9 3 – 4 3 0 1
4301