Catalytic (asymmetric) methylene transfer to aldehydes

Article abstract of DOI:10.1021/ol902816w

[Chemical equation presented] An investigation Into the poor activity of sulfides as catalysts for sulfonium-ylide-mediated methylene transfer to aldehydes has indicated that ylide formation is the problematic catalytic cycle step. Alkylation with traditional electrophiles does not proceed with sufficient efficiency to allow the sulfide to be used catalytically. Methyl triflate rapidly alkylates cyclic thiolanes under mild conditions, allowing their use in efficient aldehyde epoxidation reactions (in conjunction with phosphazene bases) at loadings as low as 10 mol %.

Full text of DOI:10.1021/ol902816w

ORGANIC
LETTERS
2010
Vol. 12, No. 3
608-611
Catalytic (Asymmetric) Methylene
Transfer to Aldehydes
Alessandro Piccinini, Sarah A. Kavanagh, Paul B. Connon, and Stephen
J. Connon*
Centre for Synthesis and Chemical Biology, School of Chemistry, UniVersity of Dublin,
Trinity College, Dublin 2, Ireland
connons@tcd.ie
Received December 7, 2009
ABSTRACT
An investigation into the poor activity of sulfides as catalysts for sulfonium-ylide-mediated methylene transfer to aldehydes has indicated that
ylide formation is the problematic catalytic cycle step. Alkylation with traditional electrophiles does not proceed with sufficient efficiency to
allow the sulfide to be used catalytically. Methyl triflate rapidly alkylates cyclic thiolanes under mild conditions, allowing their use in efficient
aldehyde epoxidation reactions (in conjunction with phosphazene bases) at loadings as low as 10 mol %.
Over the past 15 years the Corey-Chaykovsky (CC) epoxi-
dation process1 involving the reaction between sulfonium
ylides and aldehydes has evolved into a highly synthetically
useful, catalytic asymmetric methodology.2 The use of
semistabilized ylides such as those derived from the reaction
of either benzyl halides or R-haloesters/amides with chiral
sulfides at catalytic loadings can result in excellent levels of
product enantiomeric excess.3 In stark contrast, however, the
asymmetric synthesis of terminal epoxides via methylene
transfer is characterized by moderate yields and a requirement
for (super)stoichiometric loadings of chiral sulfide cata-
lysts.4,5 For example, the benchmark literature procedures
(Scheme 1) for the asymmetric sulfonium ylide-mediated
formation of styrene oxide (2) from the archetypal substrate
benzaldehyde (1) developed independently by Aggarwal6a,b
and Goodman6c involve the use of 100-200 mol % of the
chiral sulfides 3-4 and produce 2 in ca. 50-60% yield and
<60% ee.7
While it has been (reasonably) proposed that the moderate
levels of stereocontrol observed in these reactions relative
to those involving more stabilized ylides are due to irrevers-
ible betaine formation in the case of the addition of the
methylide to aldehydes,2,8 we found the poor levels of
catalyst activity (high activity and low sulfide loadings are
normal in the corresponding benzylidene transfer reactions
(3) Selected representative examples: (a) Furukawa, N.; Sugihara, Y.;
Fujihara, H. J. Org. Chem. 1989, 54, 4222. (b) Aggarwal, V. K.; Abdel-
Rahman, H.; Jones, R. V. H.; Standen, M. C. H. J. Am. Chem. Soc. 1994,
116, 5973. (c) Aggarwal, V. K.; Ford, J. G.; Thompson, A.; Jones, R. V. H.;
Standen, M. C. H. J. Am. Chem. Soc. 1996, 118, 7004. (d) Zanardi, J.;
Leriverend, C.; Aubert, D.; Julienne, K.; Metzner, P. J. Org. Chem. 2001,
66, 5620. (e) Saito, T.; Akiba, D.; Sakairi, M.; Kanazawa, S. Tetrahedron
Lett. 2001, 42, 57. (f) Winn, C. L.; Bellenie, B. R.; Goodman, J. M.
Tetrahedron Lett. 2002, 43, 5427. (g) Aggarwal, V. K.; Alonso, E.; Bae,
I.; Hynd, G.; Lydon, K. M.; Palmer, M. J.; Patel, M.; Porcelloni, M.;
Richardson, J.; Stenson, R. A.; Studley, J. R.; Vasse, J.-L; Winn, C. L J. Am.
Chem. Soc. 2003, 125, 10926. (h) Davoust, M.; Brie`re, J.-F.; Jaffre`s, P.-
A.; Metzner, P. J. Org. Chem. 2005, 70, 4166. (i) Deng, X.-M.; Cai, P.;
Ye, S.; Sun, X.-L.; Liao, W.-W.; Li, K.; Tang, Y.; Wu, Y.-D.; Dai, L.-X
J. Am. Chem. Soc. 2006, 128, 9730. (j) Davoust, M.; Cantagrel, F.; Metzner,
P.; Brie`re, J.-F. Org. Biomol. Chem. 2008, 6, 1981. (k) Bi, J.; Aggarwal,
V. K. Chem. Commun. 2008, 120.
(1) (a) Johnson, A. W.; LaCount, R. B. J. Am. Chem. Soc. 1961, 83,
417. (b) Corey, E. J.; Chaykovsky, M. J. Am. Chem. Soc. 1962, 84, 867.
(c) Franzen, V.; Driesen, H.-E. Chem. Ber. 1963, 96, 1881. (d) Corey, E.
J; Chaykovsky, M. J. Am. Chem. Soc. 1965, 87, 1353.
(2) Recent reviews: (a) Li, A.-H.; Dai, L.-X.; Aggarwal, V. K. Chem.
ReV. 1997, 97, 2341. (b) Aggarwal, V. K.; Richardson, J. Chem. Commun.
2003, 2644. (c) Aggarwal, V. K.; Winn, C. L. Acc. Chem. Res. 2004, 37,
611. (d) McGarrigle, E. M.; Myers, E. L.; Illa, O.; Shaw, M. A.; Riches,
S. L.; Aggarwal, V. K. Chem. ReV. 2007, 107, 5841.
(4) (a) Trost, B. M.; Hammen, R. F. J. Am. Chem. Soc. 1973, 95, 962.
(b) Bellenie, B. R.; Goodman, J. M. Chem. Commun. 2004, 1076.
(5) Breau, L.; Durst, T. Tetrahedron: Asymmetry 1991, 2, 367.
10.1021/ol902816w  2010 American Chemical Society
Published on Web 01/07/2010

To test this hypothesis, we initially treated benzaldehyde
(1) with the unhindered sulfide 9 (20 mol %) in the presence
of stoichiometric amounts of solid KOH and MeI in CH2Cl2
(Scheme 2). Under these conditions we observed the forma-
tion of 2 in 18% yield. A repeat of this experiment using
100 mol % of the preformed sulfonium salt 10 however,
resulted in a 90% yield under otherwise identical conditions
(Scheme 2). Thus, it is clearly the rate of alkylation to form
the sulfonium salt 10 that is problematic in this process.11
Scheme 1. Benchmark Literature Procedures for the
Sulfonium-Ylide Mediated Methylene Transfer to Benzaldehyde
(1)
Scheme 2
.
Preliminary Experiments: Identification of Sulfonium
Salt Formation as the Problematic Step
With this in mind we carried out a screening study aimed
at the identification of an agent which would methylate the
sulfide at a rate conducive to smooth catalysis of the CC
reaction under these conditions. The results of these experi-
ments are outlined in Table 1. With the sulfide utilized at
the same concentration as that used in the attempted catalytic
CC reaction (see Scheme 2), we found that methyl iodide
furnished the corresponding iodide salt in 25% yield after
24 h (entry 1). Methyl tosylate and methyl mesylate failed
to alkyate 9 under these conditions (entries 2-3) and even
powerful methylating electrophiles such as Meerwein’s salt
and dimethylsulfate gave unsatisfactory alkylation rates in
for example) difficult to fathom. Given that CC reactions
are proposed to occur via betaine intermediates (8, Scheme
1) derived from attack of the ylide 7 on the aldehyde,9,10 it
seemed likely to us that betaine formation from a methylide
(7, R1 ) H, Scheme 1) should be more efficient (if anything)
than the corresponding reaction using less nucleophilic
(semi)-stabilized ylides (7, R1 ) aryl or COR). This leaves
either ring closure of the betaine or the formation of the
sulfonium ylide as the factor responsible for the low catalytic
activity associated with methylene transfer.
(8) Aggarwal has suggested that stereocontrol in the reaction of (semi)-
stablized ylides with aldehydes is influenced by the reversibility of betaine
formation and a rate-determining C-C bond rotation of the betaine to allow
the SN2 ring closure to occur (see ref 9). It is notable however that in silico
studies by Goodman (ref 10) on the enantioselective formation of styrene
oxide from a semistabilized ylide derived from 4 and benzaldehyde found
that the relative energies of the transition states of the intially formed betaines
were excellent predictors of experimentally observed product enantiose-
lectivity in the case studied. Thus, it cannot be ruled out that enantiose-
lectivity can derive from kinetic control alone (i.e., facial selectivity in the
attack of the ylide on the aldehyde) in these reactions, which would make
the modest observed enantioselectivites associated with methylene transfer
more difficult to explain.
Since it is also difficult to see how sterically unencumbered
methylide-derived betaines (8, R1 ) H) would fail to ring-
close where more substituted (albeit aryl substituted) ana-
logues succeed, we reasoned that inefficient formation of the
sulfonium ylide (i.e., 6 f 7, Scheme 1) was the most
plausible underlying cause of the observed sluggish catalysis
of methylene transfer. This thesis was also supported by
Goodman’s observation that treatment of 4 with MeI
followed by base did not lead to the formation of 2 from
1.6c
(9) (a) Aggarwal, V. K.; Calamai, S.; Ford, J. G. J. Chem. Soc. Perkin
Trans. 1 1997, 593. (b) Aggarwal, V. K.; Harvey, J. N.; Richardson, J.
J. Am. Chem. Soc. 2002, 124, 5747. (c) Aggarwal, V. K.; Alonso, E.; Bae,
I.; Hynd, G.; Lydon, K. M.; Palmer, M. J.; Patel, M.; Porcelloni, M.;
Richardson, J.; Stenson, R. A.; Studley, J. R.; Vasse, J.-L.; Winn, C. L.
J. Am. Chem. Soc. 2003, 125, 10926. (d) Aggarwal, V. K.; Richardson, J.
J. Chem. Commun. 2003, 2644. (e) Aggarwal, V. K.; Charmant, J. P. H.;
Fuentes, D.; Harvey, J. N.; Hynd, G.; Ohara, D.; Picoul, W.; Robiette, R.;
Smith, C.; Vasse, J.-L.; Winn, C. L J. Am. Chem. Soc. 2006, 128, 2105. (f)
Edwards, D. R.; Montoya-Peleaz, P.; Crudden, C. M. Org. Lett. 2007, 9,
5481.
(6) (a) Aggarwal, V. K.; Ali, A.; Coogan, M. P. J. Org. Chem. 1997,
62, 8628. (b) Aggarwal, V. K.; Coogan, M. P.; Stenson, R. A.; Jones,
R. V. H.; Fieldhouse, R.; Blacker, J. Eur. J. Org. Chem. 2002, 319. (c)
Bellenie, B. R.; Goodman, J. M. Chem. Commun. 2004, 1076
.
(7) Hiyama et al. (ref 7a) claimed that 2 could be synthesized in 97%
ee in the presence of a chiral phase transfer catalyst. However our group
(ref 7b) recently demonstrated that Hiyama was in error and has actually
observed the decomposition of the catalyst to a related epoxide (with a
relatively large specific rotation) during the formation of 2 as a racemate:
(a) Hiyama, T.; Mishima, T.; Sawada, H.; Nozaki, H. J. Am. Chem. Soc.
1975, 97, 1626. (b) Kavanagh, S. A.; Connon, S. J. Tetrahedron: Asymmetry
(10) Silva, M. A.; Bellenie, B. R.; Goodman, J. M. Org. Lett. 2004, 6,
2559.
(11) It is noteworthy that analysis of these reactions by 1H NMR
spectroscopy did not indicate the formation of methanol.
2008, 19, 1414
.
Org. Lett., Vol. 12, No. 3, 2010
609

dichloromethane solvent (entries 4-5). However, use of
commercially available MeOTf resulted in clean, rapid
formation of salt 10f inside 15 min at ambient temperature
(entries 6-7). MeOTf could be conveniently added to the
reaction Via syringe and could be also used in acetonitrile
solvent; however its high reactivity renders it incompatible
with THF (entries 8-9).
(entries 2-3). Commercially available P1 and P2 phosphazene
organic bases fared significantly better and led to the
formation of 2 in excellent yield, with the P2 base13 proving
superior (Scheme 3).
Scheme 3. Use of the Sulfide at Catalytic Loadings
Table 1. Identification of a Suitable Alkylating Agent
alkylating
agent
time
(h)
yield
(%)a
entry
product
solvent
1
2
3
4
5
6
7
8
9
a
MeI
10a
10b
10c
10d
10e
10f
10f
10f
10f
CH2Cl2
CH2Cl2
CH2Cl2
CH2Cl2
CH2Cl2
CH2Cl2
CH2Cl2
MeCN
THF
24
24
24
24
24
25
0
0
36
33
100
100
100
0
MeOTs
MeOMs
Me3O+ BF4
Me2SO4
MeOTf
MeOTf
MeOTf
MeOTf
1
-
We were now ready to bring both efficient alkylation and
deprotonation to bear on the methylene transfer process.
Unfortunately, treatment of 1 with stoichiometric amounts
of MeOTf and P2 base in the presence of 20 mol % of sulfide
9 resulted in poor product yield, despite considerable
conversion of the aldehyde. Interestingly, we also noted the
presence of phenylacetaldehyde (presumably a Meinwald
rearangement14 product) in the crude 1H NMR spectrum of
these reactions. This led us to speculate that (a) the alkylating
agent is not compatible with the P2 base when both are
present together in solution and (b) a protic species in
solution (possibly due to the presence of adventitious water)
is capable of catalyzing the deleterious Meinwald rearrange-
ment.
24
0.25
0.25
24
Determined by H NMR spectroscopy using an internal standard.
With catalytically promising alkylation conditions in hand,
we next required an appropriate base. Table 2 details the
results of a screening study to identify an active and easily
handled base which would transform triflate salt 10f into 2
in the presence of 1 in dichloromethane solvent.
To circumvent these two difficulties we developed a
modified protocol: use of proton sponge additive avoids the
Meinwald rearrangement, while addition of the alkylating
agent and base in 5 portions (with a ca. 20 min interval)
allows the consumption of the methyl triflate to occur prior
to introduction of each portion of base. While portion-wise
addition is obviously less desirable than a protocol involving
initial addition of all reagents, the excellent efficiency of both
the individual alkylation and deprotonation steps and the
liquid nature of both added reagents (i.e., they can be injected
via syringe) results in a convenient and remarkably repro-
ducible catalytic protocol which can be executed in 5 h and
furnished an average of 95% product yield over 3 experi-
ments.15
Table 2. Identification of a Suitable Base
entry
base
yield(%)a
1
2
3
NaOH (aq 50% v/v)
KHMDS (solution in PhMe)
tBuOK
56
29
59
90
95
4
P1
5
P2 (t-Bu)
a
We were next interested in investigating the issue of
substrate scope (Table 3). We were pleased to observe that
under these conditions 9 could promote methylene transfer
to 1 (at either 10 or 20 mol % loadings, entries 1-2), together
1
Determined by H NMR spectroscopy using an internal standard.
Fifty percent NaOHswhich had proven useful in an earlier
study12 on the influence of hydrogen-bond donating catalysts
on the CC reactionsgave moderate yields of 2 under our
standard conditions (entry 1), while neither KHMDS nor
KOtBu mediated the reaction to a synthetically useful degree
(13) Phosphazene bases have been previously used in stoichiometric CC
reactions involving semistabilsed ylides, for examples see: (a) Solladie´-
Cavallo, A.; Boue´rat, L.; Roje, M. Tetrahedron Lett. 2000, 41, 7309. (b)
Solladie´-Cavallo, A.; Roje, M.; Isarno, T.; Sunjic, V.; Vinkovic, V. Eur. J.
Org. Chem. 2000, 1077. (c) Aggarwal, V. K.; Bae, I.; Lee, H.-Y.;
Richardson, J.; Williams, D. T. Angew. Chem. Int. Ed 2003, 42, 3274.
(14) Meinwald, J.; Labana, S. S.; Chadha, M. S. J. Am. Chem. Soc.
1963, 85, 582.
(12) Kavanagh, S. A.; Piccinini, A.; Fleming, E. M.; Connon, S. J. Org.
Biomol. Chem. 2008, 6, 1339.
610
Org. Lett., Vol. 12, No. 3, 2010

ficiently; furnishing 2 in high yield. The isopropyl-substituted
catalyst 27 provided the epoxide product with very modest
levels of enantioselectivity,16 however use of the bulkier tert-
butyl analogue 28 allowed the isolation of (R)-2 in 43% ee.
While this is not on a par with the levels of selectivity
achievable using semistabilized ylides, this methodology
affords levels of product enantiomeric excess approaching
those obtainable using the benchmark literature procedures
for asymmetric methylene transfer, but requires 5-10 times
less catalyst and provides the product in significantly higher
yield (Schemes 1 and 4).
Table 3. Evaluation of Substrate Scope
Scheme 4. Catalytic Asymmetric Methylene Transfer
In summary, an investigation into the low catalyst activity
in CC reactions involving methylene transfer resulted in the
development of a methodology capable of converting
aromatic, R,ꢀ-unsaturated and aliphatic aldehydes to the
corresponding terminal epoxides in excellent yields using
10-20 mol % of a simple sulfide catalyst. An asymmetric
variant of this process was also developed which could
produce styrene oxide in up to 43% ee. This field (enanti-
oselective methylene transfer to aldehydes) has been rela-
tively dormant over the past 5 years. We would be optimistic
that this protocolsthe first not to require (super)stoichio-
metric amounts of chiral sulfide (prepared via multistep
syntheses)swill facilitate the development of new sulfide
catalyst systems, the methylides derived from which could
be better able to discriminate between aldehyde faces in the
betaine-forming reaction step.
a Isolated yield. b Determined by 1H NMR spectroscopy using an internal
standard due to product decomposition during chromatography.
Acknowledgment. This material was based on work
funded by Science Foundation Ireland. We dedicate this work
to the memory Paul B. Connon, an aspiring young scientist
and our co-worker in this endeavour who sadly passed away
before submission of this manuscript.
with a range of hindered- (entry 3), deactivated (entries 4-5)
and activated (entries 6-7) aromatic aldehydes to afford
epoxides 2 and 19-23 in excellent yields. R,ꢀ-Unsaturated
(entry 8) and R-branched aliphatic aldehydes (16 and 17
respectively, entries 8-9) are also good substrates, although
an unbranched aldehyde did not furnish epoxide 26, presum-
ably due to competeitive homoaldol reactions.
Finally, we then synthesized two novel chiral analogues
of 9 (i.e., 27 and 28) with a view toward catalysis of an
asymmetric CC reaction involving methylene transfer. Both
materials proved capable of promoting the reaction ef-
Note Added after ASAP Publication. Reference 12 was
updated on January 29, 2010.
Supporting Information Available: General experimental
procedures, 1H and 13C NMR spectra, characterization data,
HPLC assays. This material is available free of charge via
the Internet at http://pubs.acs.org.
OL902816W
(15) A key predictor of success in these experiments is having prior
knowledge of the interval required to ensure complete alkylation of 9 (20
mol%) by MeOTf (20 mol%) at the concentration to be used in the
epoxidation protocol.
(16) Enantioselectivity in these reactions is strongly dependent on the
steric bulk of the sulfide substituents; for instance, in an epoxidation
experiment using a stoichiometric amount of the methyltriflate salt of a
2,5-dimethyl analogue of 27, (S)-2 was isolated in 2% ee.
Org. Lett., Vol. 12, No. 3, 2010
611