Extension of ring closing metathesis methodology to the synthesis of carbocyclic methyl and silyl enol ethers

Article abstract of DOI:10.1039/b208445g

Carbocyclic methyland silylenol ethers (TBDMS is optimum) can be synthesised regiospecifically by ring closing metathesis using the 'second generation' Grubbs ruthenium carbene catalyst incombination with molecular sieves.

Full text of DOI:10.1039/b208445g

Extension of ring closing metathesis methodology to the synthesis of
carbocyclic methyl and silyl enol ethers†
Varinder K. Aggarwal* and Adrian M. Daly
School of Chemistry, University of Bristol, Cantock’s Close, Bristol, UK BS8 1TS.
E-mail: v.aggarwal@bristol.ac.uk; Fax: +44-117-9298611; Tel: +44-117-9546315
Received (in Cambridge, UK) 30th August 2002, Accepted 26th September 2002
First published as an Advance Article on the web 9th October 2002
Table 1 RCM of methyl enol ethersa
Carbocyclic methyl and silyl enol ethers (TBDMS is
optimum) can be synthesised regiospecifically by ring
closing metathesis using the ‘second generation’ Grubbs
ruthenium carbene catalyst in combination with molecular
sieves.
In recent times ring closing metathesis (RCM) has become a
Entry
Substrate
Catalyst
Conc./M
Yield (%)
widely used and powerful tool in organic synthesis.1,2 We
considered the potential application of RCM to the synthesis of
regiospecific carbocyclic enol ethers/cyclic ketones through
cyclisation of a linear enol ether/alkene (Scheme 1). If this
1
2
3
4
3a
3a
3a
3b
1
2
2
2
0.005
0.05
0.005
0.005
0
14
53
50
a For experimental details see Supporting Information.
enol ethers 5a–g to use as substrates for our investigations.7 We
initially exposed 5a to 10 mol% 2 in refluxing benzene, but
Scheme 1
process could include the use of silyl enol ethers it would be of
great synthetic utility because of the ease of preparation of the
starting materials and the usefulness of the regiospecific silyl
enol ether products. The recent development of the ‘second
generation’ Grubbs ruthenium dihydroimidazole catalyst 2,3
which has increased reactivity with a broader range of
substrates than 1, further encouraged us to investigate this idea.
Although there are a few examples of RCM of enol ether/alkene
systems to give cyclic enol ether products,4 at the start of our
studies there were no examples of silyl enol ethers involved in
metathesis. However, during the course of our work, Shibasaki
and co-workers and Nakagawa and co-workers independently
reported that 2 was capable of catalysing the RCM of alkene/
silyl enol ether substrates.5 These results will be discussed later,
in conjunction with our own.
despite the substrate being completely consumed in 30 min the
major product observed was in fact the isomerised alkene 6a (as
a mixture of isomers) (Table 2, entry 1).8 We next examined the
TES derivative 5b due to the greater stability of this more
hindered silyl enol ether, but similar behaviour was observed.
However, this time a small amount ( < 5%) of the cyclised
product 7b was formed (Table 2, entry 2). We also observed the
Table 2 RCM of silyl enol ethersa
We initially synthesised the methyl enol ethers 36 and
investigated RCM. As expected catalyst 1 failed to effect
cyclisation of substrate 3a and complete decomposition to
unidentified products was observed (Table 1, entry 1). However
use of the second generation catalyst 2 gave low yields of
cyclopentene 4a. Simple variation of the reaction parameters
showed that optimum conditions were refluxing benzene with a
substrate concentration of 0.005 M. Under these conditions 4a
and 4b were synthesised in 53 and 50% yield, respectively, after
only 1 h. The remainder of the substrate appeared to decompose
under the reaction conditions, however we did not make
extensive further attempts to optimise the yield.
Mol%
2
Yield of Yield of
Entry
Substrate Conc./M
Methodb 6 (%)
7 (%)
1
2
3
4
5
6
7
8
9
5a
5b
5b
5b
5b
5c
5d
5e
5f
0.01
0.01
0.01
0.01
0.005
0.005
0.005
0.005
0.005
0.005
10
10
10
10
20
20
20
20
20
20
A
A
B
C
C
C
C
C
C
C
31
22
15
7
0
5
15
14
54
60
37
52
65
45c
5
15
< 5
< 5
—
—
10
5g
We then wished to extend the methodology to the more easily
accessible silyl enol ethers. Thus, we synthesised a range of silyl
a For experimental detail see ESI†. b A: catalyst and substrate heated in
benzene at reflux; B: substrate added over 30 min to a solution of the
catalyst in refluxing benzene; C: substrate added slowly over 30 min to
catalyst and powdered 4 Å mol. sieves heated at reflux in benzene. c 23% of
7f.
† Electronic supplementary information (ESI) available: full experimental
details. See http://www.rsc.org/suppdata/cc/b2/b208445g/
2490
CHEM. COMMUN., 2002, 2490–2491
This journal is © The Royal Society of Chemistry 2002

formation of some homodimer of the substrate and therefore
experimented with slow addition of the substrate to a solution of
the catalyst. Encouragingly, not only did this avoid the
formation of homodimer but it also gave a 1+1 mix of 6b and 7b
(Table 2, entry 3). At this stage we believed that the formation
of the isomerised product 6 was mediated by a ruthenium
hydride complex.8 We were unsure as to the source of this
complex but felt that the addition of molecular sieves to the
reaction might have a beneficial effect. This was indeed borne
out; addition of freshly activated 4 Å sieves to the reaction
combined with slow addition of the substrate gave a 2+1 ratio of
7b and 6b in a combined yield of 21% at a substrate
concentration of 0.01 M (Table 2, entry 4). The yield and
product ratio were further increased to a practical 54% and 11+1
respectively, by raising the catalyst loading to 20 mol% and
decreasing the concentration to 0.005 M (Table 2, entry 5).9
When these optimised conditions were applied to 5a, no
improvement in yield was observed. We then turned to other
silyl groups and were surprised to find that increasing the steric
bulk still further (TBDMS) led to an increase in yield to 60%
(Table 2, entry 6). Clearly the paradigm in RCM of decreasing
efficiency with increasing steric bulk around the olefins did not
apply in this case. Conventional wisdom prevailed in that
further increases in bulk of the silyl enol ether then led to lower
yields (Table 2, TIPS, entry 7). As TBDMS gave the highest
yield, this group was employed with other substrates (5e,f,g)
and was found to be generally useful. Replacing the phenyl
group by a cyclohexyl group gave similar results (Table 2, entry
8). Ring closure to form six- and seven-membered ring enol
ethers was also efficient although in the latter case a mixture of
the seven-membered ring 7g and six-membered ring 7f (formed
from isomerisation and subsequent ring closure)8 was ob-
tained.
which can react with the silyl enol ether in a slow cyclisation
step to give 12. Cyclisation is slow because there is an inherent
electronic preference for the enol ether to react with the metal
carbene with the opposite regioselectivity to that required for
ring closure.11 This electronic bias was revealed in the
intermolecular metathesis reaction of ruthenium alkylidenes
with enol ethers which lead exclusively to the ruthenium
Fischer carbenes; no cross metathesis product being formed.11
If the substrate bears groups which enhance cyclisation rates
(e.g, X = C(alkyl)2, C(ester)2), 12 is formed uneventfully but in
the absence of such groups (e.g. X = CHPh) 11 can undergo
reversion and give back 10. The metal carbene can now react
either with 10 in the same way as before or with the silyl enol
ether moiety and give the Fischer carbene 13 which will not
react further and may decompose. Indeed Grubbs has recently
indicated that the thermal stability of species analogous to 13 is
considerably lower than those such as 11.11 A decomposition
process will not only consume the substrate but also the catalyst.
TMS enol ethers, the least hindered of all, will react with the
metal carbene fastest and this could account for the poor results
obtained with 5a. TBDMS enol ethers may provide the
optimum steric hindrance to minimise reaction leading to 13 but
still allow cyclisation of 11 to give 12.12
In summary we have successfully synthesised carbocyclic
methyl enol ethers using RCM for the first time. We have also
developed RCM conditions under which alkene/silyl enol ethers
can be cyclised and have shown that for substrates with a high
propensity for cyclisation, TMS enol ethers can be employed
but for general substrates TBDMS enol ethers are optimum. The
formation of regiospecific silyl enol ethers which have high
synthetic utility renders this process very useful.
We thank Ian Holmes for initial investigations and Guy
Lloyd-Jones for valuable discussions. A. M. D. acknowledges
the support of the European Commision in the form of a Marie
Curie Fellowship (no. HPMF-CT-2000-00914).
As we were completing this study, Shibasaki and co-workers
published a report of their investigations into RCM of alkene/
silyl enol ether systems.5a Surprisingly they succeeded in
cyclising trimethylsilyl enol ether/alkenes in generally > 90%
yield using 7–10 mol% of 2 as catalyst. However, all of
Shibasaki’s substrates contained either gem diester substituents
on the carbon backbone or a cyclic acetal linker. According to
Jung these two groups in particular provide dramatically
enhanced rates of cyclisation, greater even than a gem dialkyl
group.10 In light of this we synthesised trimethylsilyl enol ether
8 bearing a simple gem dialkyl group and gratifyingly it gave an
89% yield of the corresponding cyclised derivative 9 using only
Notes and references
1 R. H. Grubbs and S. Chang, Tetrahedron, 1998, 54, 4413.
2 A. Fürstner, Angew. Chem., Int. Ed., 2000, 39, 3012.
3 M. Ulman and R. H. Grubbs, J. Org. Chem., 1999, 64, 7202.
4 J. R. Stille and R. H. Grubbs, J. Am. Chem. Soc., 1986, 108, 855; J. R.
Stille, B. D. Santarsiero and R. H. Grubbs, J. Org. Chem., 1990, 55, 843;
O. Fujimura, G. C. Fu and R. H. Grubbs, J. Org. Chem., 1994, 59, 4029;
J. S. Clark and J. G. Kettle, Tetrahedron, 1999, 55, 8231; J. D. Rainier,
S. P. Allwein and J. M. Cox, J. Org. Chem., 2001, 66, 1380; M. H. D.
Postema and D. Calimente, Tetrahedron Lett., 1999, 40, 4755; M. H. D.
Postema, D. Calimente, L. Liu and T. L. Behrmann, J. Org. Chem.,
2000, 65, 6061; J. S. Clark and O. Hamelin, Angew. Chem., Int. Ed.,
2000, 39, 372; O. Dirat, T. Vidal and Y. Langlois, Tetrahedron Lett.,
1999, 40, 4801; C. F. Sturino and J. C. Y. Wong, Tetrahedron Lett.,
1998, 39, 9623; S. P. Allwein, J. M. Cox, B. E. Howard, H. W. B.
Johnson and J. D. Rainier, Tetrahedron, 2002, 58, 1997; J. D. Rainier,
J. M. Cox and S. P. Allwein, Tetrahedron Lett., 2001, 42, 179; M.-P.
Heck, C. Baylon, S. P. Nolan and C. Mioskowski, Org. Lett., 2001, 3,
1989; A. K. Chatterjee, J. P. Morgan, M. Scholl and R. H. Grubbs, J.
Am. Chem. Soc., 2000, 122, 3783.
5 (a) A. Okada, T. Ohshima and M. Shibasaki, Tetrahedron Lett., 2001,
42, 8023; (b) M. Arisawa, C. Theeraladanon, A. Nishida and M.
Nakagawa, Tetrahedron Lett., 2001, 42, 8029.
6 We found the Petasis reagent (Cp2TiMe2) to be efficient for methylen-
ation of the precursor esters: N. A. Petasis and E. I. Bzowej, J. Am.
Chem. Soc., 1990, 112, 6392; other methods were unsatisfactory.
7 The substrates all contain a large group (Ph/Cy) to reduce volatility and
aid isolation.
10 mol% of 2. Slow addition or molecular sieves were not
required. This clearly shows the large gem-dialkyl effect
observable in RCM and illustrates the importance of our
discovery that molecular sieves, slow substrate addition and the
use of TBDMS enable the cyclisation of alkene/silyl enol ethers
which are not particularly prone to cyclisation using 2.
We propose a possible rationalisation of these results
(Scheme 2). Initial reaction of the metal carbene with 10 is most
likely to occur with the less sterically hindered alkene to give 11
8 Isomerisation of this type has been observed before: e.g. D. Joe and L.
E. Overman, Teyrahedron Lett., 1997, 38, 8635 and S. J. Miller, H. E.
Blackwell and R. H. Grubbs, J. Am. Chem. Soc., 1996, 118, 9606.
9 Solely raising the catalyst loading did not have the same effect.
10 M. E. Jung, Synlett, 1999, 843.
11 J. Louie and R. H. Grubbs, Organometallics, 2002, 21, 2153.
12 Ring closure with more hindered silyl enol ethers will also be slowed
down but the rate of the intermolecular reaction (10 ? 13) is expected
to be more greatly affected than the intramolecular reaction (11 ?
12).
Scheme 2
CHEM. COMMUN., 2002, 2490–2491
2491