Experimental demonstration of base-catalyzed interconversion of isomeric betaine intermediates in the corey-chaykovsky epoxidation

Article abstract of DOI:10.1021/ol070875j

The collapse of hydroxysutfonium salts has been examined as a model for the epoxidation of aldehydes. The anti diastereomer reacted with retention of stereochemistry and no crossover, while the syn diastereomer gave crossover products along with cis and trans epoxides. Deprotonation and reprotonation on the carbon of the α-hydroxy sulfonium ylide is presumed responsible for production of the trans epoxide. This reaction pathway has been proposed to explain losses of enantioselectivity but never directly observed.

Full text of DOI:10.1021/ol070875j

ORGANIC
LETTERS
2007
Vol. 9, No. 12
2397-2400
Experimental Demonstration of
Base-Catalyzed Interconversion of
Isomeric Betaine Intermediates in the
Corey−Chaykovsky Epoxidation
David R. Edwards,* Jenny Du, and Cathleen M. Crudden*
Department of Chemistry, Queen’s UniVersity, 90 Bader Lane,
Kingston, Ontario, Canada, K7M 3N6
cruddenc@chem.queensu.ca
Received April 15, 2007
ABSTRACT
The collapse of hydroxysulfonium salts has been examined as a model for the epoxidation of aldehydes. The anti diastereomer reacted with
retention of stereochemistry and no crossover, while the syn diastereomer gave crossover products along with cis and trans epoxides.
Deprotonation and reprotonation on the carbon of the
r-hydroxy sulfonium ylide is presumed responsible for production of the trans epoxide.
This reaction pathway has been proposed to explain losses of enantioselectivity but never directly observed.
The Corey-Chaykovsky reaction, in which an aldehyde and
an ylide are coupled to yield an epoxide has proven to be a
versatile and valuable method for the production of
epoxides.1 Since the initial discovery of this reaction,2
catalytic enantioselective variants have been reported which
have permitted the synthesis of a broad range of products
including epoxides, cyclopropanes, aziridines, glycidic amides,
and acids.1 From a variety of mechanistic probes, in addition
to theoretical calculations,3 a general consensus on the
mechanism of the reaction has been reached. In broad strokes,
the mechanism entails deprotonation of the sulfonium salt
to generate an ylide, attack on the aldehyde partner generating
a betaine (1, Scheme 1), and expulsion of the sulfide leaving
group to generate the epoxide.
The critical questions then surround identification of the
rate-determining step and the reversibility of the remaining
individual steps, both of which can have a significant impact
(1) (a) Aggarwal, V. K.; Winn, C. L. Acc. Chem. Res. 2004, 37, 611.
(b) Aggarwal, V. K.; Alonso, E.; Fang, G.; Ferrara, M.; Hynd, G.; Porcelloni,
M. Angew. Chem., Int. Ed. 2001, 40, 1433. (c) Aggarwal, V. K.; Charmant,
J. P. H.; Fuentes, D.; Harvey, J. N.; Hynd, G.; Ohara, D.; Picoul, W.;
Robiette, R. I.; Smith, C.; Vasse, J.-L.; Winn, C. L. J. Am. Chem. Soc.
2006, 128, 2105. (d) Aggarwal, V. K; Hebach, C. Org. Biomol. Chem. 2005,
3, 1419. (e) Aggarwal, V. K.; Richardson, J. Chem. Commun. 2003, 2644.
(2) (a) Johnson, A. W.; LaCount, R. B. Chem. Ind. (London) 1958, 1440.
(b) Johnson, A. W.; LaCount, R. B. J. Am. Chem. Soc. 1961, 417.
(3) (a) Aggarwal, V. K.; Harvey, J. N.; Richardson, J. J. Am. Chem.
Soc. 2002, 124, 5747. (b) Volatron, F.; Eisentein, O. J. Am. Chem. Soc.
1987, 109, 1. (c) Lindvall, M. K.; Koskinen, A. M. P. J. Org. Chem. 1999,
64, 4595. (d) Lindvall, M. K.; Koskinen, A. M. P. Tetrahedron 2001, 57,
4629.
10.1021/ol070875j CCC: $37.00
© 2007 American Chemical Society
Published on Web 05/16/2007

The syn- and anti-R hydroxy sulfonium salts (2) were
generated by the method of Aggarwal which involves
treatment of cis or trans stilbene oxide with MeSNa, followed
by methylation with methyl iodide.4a Deprotonation of the
anti diastereomer of 2 with NaOH in a mixed solvent system
of CH3CN/H2O (10:1) was carried out using p-NO2-benz-
aldehyde as the trapping agent (2 equiv), eq 1.
Scheme 1. General Mechanism for the CoreysChaykovsky
Epoxidation Reaction
on the control of relative stereochemistry and the introduction
of chirality. Answers to these questions often hinge upon
betaine crossover experiments.1c,d,4
The reactivity of p-NO2-benzaldehyde relative to benz-
aldehyde itself was determined by reacting both aldehydes
with ylide 4 under similar reaction conditions.5 p-NO2-
benzaldehyde reacted faster than benzaldehyde by a factor
of 62. Owing to the increased reactivity of p-NO2-benz-
aldehyde, and the fact that it is used in excess, it is reasonable
to assume that it will efficiently trap any free ylide that is
generated. Despite this fact, when anti-betaine 2 was treated
with base in the presence of excess p-NO2-benzaldehyde,
the only observable product is trans-stilbene oxide. This is
consistent with direct collapse of the anti-betaine to form
the expected trans-epoxide product. The absence of any
crossover products (5) indicates that the anti-betaine is
formed irreversibly in the typical epoxidation reaction, in
line with previous reports by Aggarwal.4a
In a typical crossover experiment, an independently
generated R-hydroxy sulfonium salt (2) is treated with base
to generate the corresponding betaine 1, Scheme 2. The
Scheme 2. Probing the Reactivity of Betaine 1 with an
Intermolecular Competition Experiment
Remarkably, reaction of the syn-betaine (2) under identical
conditions resulted in a vastly different outcome, eq 2. In
this case, significant amounts of the crossover products were
obtained; cis-NO2-stilbene oxide (cis-5) 10% and trans-NO2-
stilbene oxide (trans-5) 50%. The expected cis-stilbene oxide
(cis-3) was also obtained (31%) and trans-stilbene oxide
(trans-3) made up the remainder of the epoxide products at
9%, (Table 1, entry 1). The observation of crossover products
cis- and trans-5 is expected and indicates that the formation
of the syn-betaine is reversible. The appearance of trans-
stilbene oxide, albeit in minor amounts, is more difficult to
explain.
As noted previously, the krel for the reaction of 4 with para-
NO2-benzaldehyde and benzaldehyde is 62:1 under the
reaction conditions described. Despite this fact, the ratio of
trans-3 to trans-5 obtained in the experiment is closer to
5:1. This is an order of magnitude lower than expected if
both species are being produced via the free ylide, which
undergoes competitive trapping with both para-NO2-benz-
aldehyde and benzaldehyde present in solution (Scheme 1).
Repeating the experiment with 10 equiv of para-NO2-
benzaldehyde in an attempt to further bias formation of 5
gave the same result as with 2 equiv (entry 2). This indicates
deprotonation is conducted in the presence of a second
aldehyde, ArCHO, which is more reactive than PhCHO to
ascertain the preferred pathway for reaction of 1, that is,
collapse to 3 or reversion to 4 and aldehyde. Since any free
ylide (4) generated from 1 will react preferentially with
ArCHO, the observation of unsymmetrical arylepoxides (5)
in the experiment is interpreted as evidence for generation
of free ylide 4 in the reaction. In addition, since it is possible
to independently generate syn and anti sulfonium salts, this
experiment permits information to be obtained about the
reactivity of both syn and anti betaines under controlled
conditions.
Using this method, we have uncovered a previously
unknown and significant difference in reactivity of the two
diastereomers of 2. These results, which can have an impact
on the introduction of chirality in this system, are presented
herein.
(4) (a) Aggarwal, V. K.; Calamai, S.; Ford, J. G. J. Chem. Soc., Perkins
Trans. 1 1997, 593. (b) Aggarwal, V. K.; Charmant, J. P. H.; Ciampi, C.;
Hornby, J. M.; O’Brien, C. J.; Hynd, G.; Parsons, R. J. Chem. Soc., Perkins
Trans. 1 2001, 3159. (c) Yoshimine, M.; Hatch, M. J. J. Am. Chem. Soc.
1967, 89, 5831.
(5) krel determined by competition experiment between excess benz-
aldehyde and p-nitrobenzaldehyde reacting with a limiting amount of
sulfonium salt 4 (see Supporting Information for full details).
2398
Org. Lett., Vol. 9, No. 12, 2007

Table 1. Crossover Studies with syn-2a
Scheme 4. Isomerization of syn-2 via C-H Deprotonation
entry
ArCHO
NaI
cis-3 trans-3 cis-5 trans-5
1
2
3
2 equiv
10 equiv
2 equiv
31
31
33
9
8
9
10
6
11
50
55
47
either face producing anti-2. Ring closure of anti-2 would
then be responsible for the production of trans-stilbene oxide.
Evidence for the base catalyzed isomerization of betaines
was obtained from a deuterium labeling study. The crossover
experiment was carried out with syn-2 as in eq 3, but this
time using CH3CN/D2O/NaOD. In addition to the expected
crossover products cis- and trans-5, trans-stilbene oxide was
obtained with quantitative deuteration at one of the benzylic
carbons, eq 3. In contrast, only 5% deuteration was observed
in the cis product.
1.5 equiv
a Ar ) p-nitrophenyl
that all the free ylide generated in the experiment is being
efficiently trapped by para-NO2-benzaldehyde to form cis-
and trans-5.
To rule out the possibility that the trans-stilbene oxide
was generated by a counterion promoted inversion of the
betaine, the reaction was repeated in the presence of excess
iodide (entry 3). In this scenario, the iodide counterion would
first displace the sulfonium substituent with inversion at
carbon in an SN2-type process.6 The intermediate so gener-
ated would then undergo direct ring closure to provide
stilbene oxide with trans geometry, and regenerate the iodide
catalyst, Scheme 3.
These results indicate that syn-2 reacts via three path-
ways: (1) direct collapse to epoxide cis-3, (2) reversion to
benzaldehyde and free ylide and (3) deprotonation to 6,
which is converted into trans-3.8 On the other hand, the anti-
betaine generated from anti-2 reacted cleanly with no
crossover products and no deuterium incorporation, consis-
tent with the facile ring closure from the anti-betaine (eq
4).
Scheme 3. Possible Invertive Process Yielding trans-3 from
cis-1
If this pathway were responsible for production of trans-
stilbene oxide, the addition of exogenous iodide would be
expected to alter the product distribution in the crossover
experiment in favor of trans-stilbene oxide. As shown in
Table 1, entry 3 the addition of 1.5 equiv of NaI resulted in
essentially no change in the product distribution. Conse-
quently a double inversion process can be discounted.
Having firmly ruled out (a) free ylide trapping with
benzaldehyde and (b) counterion catalyzed isomerization of
syn-betaine as sources of trans-stilbene oxide, we next turned
to a third potential source for this product. Specifically, base
catalyzed isomerization of syn-betaine-2 to anti-2, which
could then undergo ring closure to afford trans-stilbene
oxide. In this process, the syn-betaine would undergo
deprotonation to furnish a new ylide species 6, Scheme 4.
Betaines such as 6 have previously been shown to be sp2
hybridized7 and therefore reprotonation of 6 could occur from
These findings support Aggarwal’s suggestion that the
diastereoselectivity in epoxidations of this type results from
a fast collapse of the anti-betaine to the corresponding trans-
epoxide, compared with the slow collapse of syn-betaines.3a,4a
It should be noted that the base catalyzed isomerization
of betaines has precedent in the elegant work of Vedejs and
co-workers with phosphonium betaine analogues.9 Further-
more, Aggarwal has proposed that the near identical product
distribution obtained in crossover experiments with anti- and
(8) Note that the low deuteration observed in cis-3 implies that once 6
is generated, deuteration occurs with high stereoselectivity to yield anti-
betaine 2 which collapses to yield trans-3.
(9) (a) Vedejs, E.; Fleck, T.; Hara, S. J. Org. Chem. 1987, 52, 4637. (b)
Vedejs, E.; Fleck, T. J. Am. Chem. Soc. 1989, 111, 5861.
(6) Ratts, K. W.; Yao, A. N. J. Org. Chem. 1966, 31, 1689.
(7) Aggarwal, V. K.; Schade, S.; Taylor, B. J. Chem. Soc., Perkins Trans.
1 1997, 2811.
Org. Lett., Vol. 9, No. 12, 2007
2399

syn-betaines of ester and amide stabilized N-tosyl-protected
imines may result from base-catalyzed equilibration of
betaines.4b However, since incomplete trapping of free ylide
in these experiments could also account for the product
distributions obtained, a definitive demonstration of inter-
conversion of syn and anti betaines is important.10 Consider-
ing that ester and amide stabilized betaines represent even
more activated substrates for base catalyzed isomerization,
it is entirely likely that a similar process is operative with
these substrates.
To probe the effect of the stabilizing group, substrate 7
which would generate an unstabilized betaine was subjected
to the combined crossover/labeling experiment, eq 5. No
crossover products were obtained indicating irreversible
formation of the betaine. However, significant deuteration
did occur, indicating that the betaine is sufficiently persistent
to allow proton transfer to occur, even when reversion to
free ylide does not occur.11
reaction can be subject to base catalyzed isomerization. All
classes of betaine; stabilized, semistabilized and unstabilized
classes, are capable of undergoing proton abstraction and, if
structurally feasible, isomerization. Understanding of this
process is critical since it is likely to increase diastereo-
selectiVity in favor of the preferred trans isomer, but decrease
enantioselectiVity. A similar effect was observed during
cyclopropanation studies.12 It should be noted that although
only 25% of the syn-betaine underwent isomerization in the
current system, variations in substrate structure and reaction
conditions could have a significant effect on the extent of
deprotonation observed.13 In particular, this possibility should
be considered for cases which exhibit low enantioselectivity.
Acknowledgment. The Natural Sciences and Engineering
Research Council of Canada (NSERC) is gratefully acknowl-
edged for support of this work in terms of operating and
equipment grants to C.M.C. D.R.E. thanks the government
of Ontario for graduate scholarships and both D.R.E. and
J.D. thank Queen’s University for Queen’s Graduate Schol-
arships.
Supporting Information Available: Experimental pro-
cedures, proton and deuterium NMR spectra for deuterium
labeling experiments, crossover experiments, and competition
experiments. This material is available free of charge via
the Internet at http://pubs.acs.org.
The present study convincingly demonstrates for the first
time that betaine intermediates in the Corey-Chaykovsky
OL070875J
(10) Relative rate data for the two competing substrates were not
available.
(12) Aggarwal, V. K.; Grange, E. Chem.sEur. J. 2006, 12, 568.
(13) For example, less than 5% isomerization is observed using DBU in
pure CH3CN.
(11) It is not clear if this results from a relatively slow ring closure or
faster H+ abstraction from the secondary carbon R to the sulfonium group.
2400
Org. Lett., Vol. 9, No. 12, 2007