Highly enantioselective synthesis of glycidic amides using camphor-derived sulfonium salts. Mechanism and applications in synthesis

Article abstract of DOI:10.1021/ja0568345

The reactions of a range of amide-stabilized sulfur ylides derived from readily available camphor-derived sulfonium salts for the synthesis of glycidic amides have been studied. Primary, secondary, and tertiary amides were tested, and it was found that th

Full text of DOI:10.1021/ja0568345

Published on Web 01/25/2006
Highly Enantioselective Synthesis of Glycidic Amides Using
Camphor-Derived Sulfonium Salts. Mechanism and
Applications in Synthesis
Varinder K. Aggarwal,*,† Jonathan P. H. Charmant,‡ Daniel Fuentes,†
Jeremy N. Harvey,† George Hynd,† Diasuke Ohara,† Willy Picoul,†
Raphae¨l Robiette,† Catherine Smith,† Jean-Luc Vasse,† and Caroline L. Winn†
Contribution from the School of Chemistry and Structural Chemistry Laboratory, School of
Chemistry, UniVersity of Bristol, Cantock’s Close, Bristol BS8 1TS, U.K.
Received October 18, 2005; E-mail: v.aggarwal@bristol.ac.uk
Abstract: The reactions of a range of amide-stabilized sulfur ylides derived from readily available camphor-
derived sulfonium salts for the synthesis of glycidic amides have been studied. Primary, secondary, and
tertiary amides were tested, and it was found that the highest enantioselectivities were observed with tertiary
amides, which provided glycidic amides in good to excellent yields, exclusive trans selectivity, and excellent
enantioselectivities. The reaction was general for aromatic aldehydes, but aliphatic aldehydes gave more
variable enantioselectivities. The epoxy amides could be converted cleanly into epoxy ketones by treatment
with organolithium reagents. We were also able to effect selective ring opening of the epoxy amides with
a variety of nucleophiles, followed by hydrolysis of the amide to yield the corresponding carboxylic acid.
This methodology was applied to the total synthesis of the target compound SK&F 104353. A combination
of crossover experiments and theoretical calculations has revealed that the rate- and selectivity-determining
step is ring closure, not betaine formation as was the case for phenyl-stabilized ylides.
Scheme 1. Possible Routes toward the Synthesis of Epoxy
Amides and Esters
Introduction
Glycidic esters and amides are important and versatile
intermediates in organic synthesis, and a number of methods
exist for their preparation (Scheme 1).
One such route involves oxidation of the corresponding R,â-
unsaturated esters or amides.1 Most of the work in this area
involves asymmetric variations of the Weitz-Scheffer epoxi-
dation2 in which a variety of metals have been used.1
Some particularly elegant work has been carried out in this
area by Shibasaki et al.,3 using lanthanide-derived catalysts.
Epoxidation of R,â-unsaturated ketones employed a Las
BINOLsPh3AsdO complex generated in situ from La(OiPr)3,
BINOL, and Ph3AsdO in a 1:1:1 ratio, and when enantiopure
BINOL was used, epoxides were generated in moderate to
excellent yield (72-99%) with high enantiomeric excess (89-
99% ee).4 Although the methodology could not be directly
applied to the formation of glycidic esters, carboxylic acid
imidazolides4 and R,â-unsaturated-N-acyl-pyrroles,5 which act
as ester surrogates, underwent epoxidation in a highly selective
manner. In addition, R,â-unsaturated amides were also amenable
to epoxidation, yielding glycidic amides with excellent levels
of enantiocontrol.6
There are also examples in which chiral phase-transfer
catalysts derived from Cinchona alkaloids have been used to
induce asymmetry in the epoxidation of electron-deficient
alkenes,7-9 although unfortunately, this methodology has not
yet been applied to the synthesis of glycidic amides and esters.
In the early 1980s, Julia´ et al. reported the use of a triphasic
epoxidation system involving an aqueous solution of NaOH and
H2O2, a solution of chalcone in an organic solvent, and an
insoluble poly-(L)-alanine catalyst.10 A wide range of chalcone
derivatives were investigated, giving products with good to
excellent levels of enantioselectivity (50-99% ee). Recent
modifications by Roberts involved the use of a biphasic system
† School of Chemistry.
‡ Structural Chemistry Laboratory, School of Chemistry.
(1) Porter, M. J.; Skidmore, J. J. Chem. Soc., Chem. Commun. 2000, 1215-
1225.
(2) Weitz, E.; Scheffer, A. Ber. Dtsch. Chem. Ges. 1921, 52, 2327.
(3) Matsunaga, S.; Kinoshito, T.; Okada, S.; Harada, S.; Shibasaki, M. J. Am.
Chem. Soc. 2004, 126, 7559-7570.
(6) Nemoto, T.; Kakei, H.; Gnanadesikan, V.; Tosaki SY.; Ohshima, T.;
Shibasaki, M. J. Am. Chem. Soc. 2002, 124, 14544-14545.
(7) Lygo, B.; Wainwright, P. G. Tetrahedron 1999, 55, 6289-6300.
(8) Corey, E. J.; Zhang, F. Y. Org. Lett. 1999, 1, 1287-1290.
(9) Arai, S.; Tsuge, H.; Shioiri, T. Tetrahedron Lett. 1998, 39, 7563-7566.
(10) Julia, S.; Masana, J.; Vega, J. C. Angew. Chem., Int. Ed. Engl. 1980, 19,
929-931.
(4) Nemoto, T.; Ohshima, T.; Yamaguchi, K.; Shibasaki, M. J. Am. Chem.
Soc. 2001, 123, 2725-2732.
(5) Kinoshita, T.; Okada, S.; Park, S. R.; Matsunaga, S.; Shibasaki, M. Angew.
Chem., Int. Ed. 2003, 42, 4680-4684.
9
10.1021/ja0568345 CCC: $33.50 © 2006 American Chemical Society
J. AM. CHEM. SOC. 2006, 128, 2105-2114
2105

A R T I C L E S
Aggarwal et al.
Scheme 2. Dai et al.’s Asymmetric Synthesis of Epoxy Amides
that utilized the much cheaper sodium percarbonate as both
oxidant and base, while retaining a high level of enantiocontrol.11
This methodology is currently restricted to the epoxidation of
enones.
Using Chiral Sulfonium Salt 1a29
Chiral dioxiranes have been developed by a number of groups
as asymmetric oxidants,12 most notably by Shi and co-workers.13
Shi et al. employed ketones derived from (-)-quinic acid that
were oxidized by Oxone to form dioxiranes in situ, which could
carry out highly enantioselective epoxidation on a range of
enones (82-96% ee).14 R,â-Unsaturated esters have also been
employed as substrates using a fructose-derived ketone. Again,
high enantioselectivities (82-98%) were obtained for a number
of trans and trisubstituted substrates.15
and basic properties was required to complex the ester-stabilized
ylide on the solid surface and provide it with appropriate
reactivity.
Recent work in our own laboratory has shown that carboxy-
late-substituted sulfur ylides (thetin salts) can be used in
asymmetric epoxidation to yield glycidic acids,24 but unfortu-
nately, the enantioselectivities observed to date are modest (up
to 67% ee).
Another metal-free approach has been reported by Jørgensen
et al. in which R,â-unsaturated aldehydes underwent epoxidation
using an organocatalyst.16 Using a proline-derived chiral amine
with hydrogen peroxide as the oxidant, epoxidation occurred
in good yield and high diastereoselectivity with excellent levels
of enantiomeric excess (>94% ee) for a range of substrates.
An alternative to the oxidative method for the formation of
glycidic amides and esters is a Darzens reaction.17 Although
chiral phase-transfer catalysts have been explored in reactions
of R-haloketones (and R-halosulfones), the enantioselectivities
observed were only moderate (53-86% ee).18,19 Superior results
have been achieved using chiral reagents and chiral auxiliaries,
but in these cases, a two-step process (C-C bond formation
followed in a separate step by ring closure) is required. Corey
et al. have developed a highly enantioselective version of the
Darzens reaction in which an achiral aldehyde and t-butyl
bromoacetate in the presence of a chiral borane were converted
into an intermediate chiral R-bromo-â-hydroxy ester, which was
subsequently transformed into the corresponding epoxy ester
in up to 98% ee.20
Amide-stabilized sulfonium ylides, however, have attracted
more attention and have been shown to react with both
aldehydes and ketones to afford glycidic amides with a high
degree of trans selectivity.25 It has been demonstrated by Lo´pez-
Herrera et al. that good diastereoselectivity can also be achieved
when chiral aldehydes are employed (>92% de for a single trans
isomer).26
We previously reported the preparation of glycidic amides
in good yields and high diastereoselectivity (>95:5) using an
achiral sulfonium ylide that was prepared in situ via the reaction
between a sulfide and diazoacetamide in the presence of a copper
catalyst.27 More recently, Seki has shown that it is possible to
prepare glycidic amides with a moderate degree of enantiose-
lectivity (64% ee) using chiral sulfur ylides that were generated
in situ from diazoacetamide in the presence of catalytic amounts
of chiral binaphthylsulfide and copper(II) acetylacetone.28
Dai et al. also reported the preparation of enantiomerically
enriched glycidic amides using chiral sulfonium salt 1a with a
range of aromatic aldehydes (Scheme 2).29 The camphor-derived
sulfide30 that was employed had previously been used in the
formation of enantiomerically enriched 2,3-diarylepoxides, with
enantioselectivities reaching up to 74% when a stoichiometric
quantity of sulfide was employed.31
The glycidic amide products were obtained under either
solid-liquid phase-transfer conditions (conditions A: acetoni-
trile and KOH at room temperature) or liquid-liquid phase-
transfer conditions (conditions B: dichloromethane and aqueous
NaOH at 0 °C) for a range of aldehydes. Moderate to excellent
yields (49-94%) with moderate levels of asymmetric induction
(up to 72% ee) were achieved.29 It is interesting to note that,
although the sulfonium bromide salts were prepared and isolated
prior to the reaction, their diastereomeric purity was not
discussed. Indeed, we believe that the enantiomeric excesses
observed in this and related work by the same group could have
been improved by using a single diastereomer of sulfonium salt
(vide infra).
Sulfur ylide methodology has also been applied to the
synthesis of glycidic esters and amides. The reactivity of the
ester-stabilized ylides is too low to react with simple aldehydes
to form 2,3-substituted epoxy esters;21,22 they react only with
1,2-dicarbonyl compounds. However, Delmas has shown that
ester-stabilized ylides can react under specific conditions.23
Using either the solid base Ba(OH)2‚8H2O (C-O) or K2CO3‚
1.5H2O, ester-stabilized ylides underwent reaction with very
electrophilic aldehydes, whereas use of the strong base KOH
or a more hydrated barium hydroxide base (C-200) did not yield
any epoxide. It was proposed that a lattice of specific dimensions
(11) Allen, J. V.; Drauz, K.-H.; Flood, R. W.; Roberts, S. M.; Skidmore, J.
Tetrahedron Lett. 1999, 40, 5417-5420.
(12) Yang, D. Acc. Chem. Res. 2004, 37, 497-505.
(13) Shi, Y. Acc. Chem. Res. 2004, 37, 488-496.
(14) Wang, Z.-X.; Tu, Y.; Frohn, M.; Zhang, J. R.; Shi, Y. J. Am. Chem. Soc.
1997, 119, 11225-11235.
(15) Wu, X. Y.; She, X. G.; Shi, Y. A. J. Am. Chem. Soc. 2002, 124, 8792-
8793.
(16) Marigo, M.; Franze´n, J.; Poulsen, T. B.; Zhuang, W.; Jørgensen, K., A. J.
Am. Chem. Soc. 2005, 127, 6964-6965.
(24) Aggarwal, V. K.; Hebach, C. Org. Biomol. Chem. 2005, 3, 1419-1427.
(25) Ratts, K. W.; Yao, A. N. J. Org. Chem. 1966, 31, 1689-1693.
(26) Valpuesta Ferna´ndez, M.; Durante-Lanes, P.; Lo´pez-Herrera, F. J. Tetra-
hedron 1990, 46, 7911-7922.
(17) Rosen, T. In ComprehensiVe Organic Synthesis; Fleming, I., Ed.; Pergamon
Press: Oxford, U.K., 1991; Vol. 2, p 409.
(18) Arai, S.; Shirai, Y.; Ishida, T.; Shioiri, T. Tetrahedron 1999, 55, 6375-
(27) Aggarwal, V. K.; Blackburn, P.; Fieldhouse, R.; Jones, R. V. H. Tetrahedron
Lett. 1998, 39, 8517-8520.
6386.
(19) Arai, S.; Tokumaru, K.; Aoyama, T. Tetrahedron Lett. 2004, 45, 1845-
1848.
(28) Imashiro, R.; Yamanaka, T.; Seki, M. Tetrahedron: Asymmetry 1999, 10,
2845-2851.
(20) Corey, E. J.; Choi, S. Tetrahedron Lett. 1991, 32, 2857-2860.
(21) Trost, B. M.; Melvin, L. S. Sulfur Ylides; Academic Press: New York,
1975.
(29) Zhou, Y.-G.; Hou, X.-L.; Dai, L.-X.; Xia, L.-J.; Tang, M.-H. J. Chem.
Soc., Perkin Trans. 1 1999, 77-80.
(30) Goodridge, R. J.; Hambley, T. W.; Haynes, R. K.; Ridley, D. D. J. Org.
Chem. 1988, 53, 2881-2889.
(22) Payne, G. B. J. Org. Chem. 1968, 33, 3517.
(23) Borredon, E.; Clavellinas, F.; Delmas, M.; Gaset, A. J.; Sinisterra, V. J.
Org. Chem. 1990, 55, 501.
(31) Li, A.-H.; Dai, L.-X.; Hou, X.-L.; Huang, Y.-Z.; Li, F.-W. J. Org. Chem.
1996, 61, 489-493.
9
2106 J. AM. CHEM. SOC. VOL. 128, NO. 6, 2006

Highly Enantioselective Synthesis of Glycidic Amides
A R T I C L E S
Table 1. Asymmetric Synthesis of (2R,3S)-2,3-Epoxy Amides via Camphor-Derived Sulfonium Salts 1-3
a The enantiomeric excess was determined by chiral HPLC, and the absolute configuration of the product in entries 3 and 4 was determined by comparison
of HPLC elution orders and by comparison of [RD] values with the literature. All others are given by analogy. b These reactions were carried out using
p-Cl-PhCHO as the substrate. c Epoxidation was carried out at room temperature.
The same chiral sulfide was employed by Tang et al. for the
preparation of vinyl cyclopropanes which involved the addition
of an allylic sulfur ylide to R,â-unsaturated esters, amides, and
ketones.32 It was noted that the sulfonium salts employed on
this occasion were diastereomerically pure by NMR spectros-
copy, and enantioselectivities of >94% were achieved for a
range of substrates.
Dai et al.29 had reported that the hydroxyl group in 1a was
responsible for controlling the conformation of the ylide thereby
enhancing enantioselectivity. If this were correct, replacing the
OH group for OMe would be expected to lower the selectivity.
We elected to test this hypothesis, but instead of observing a
decrease in enantiomeric excess, we obtained much improved
enantioselectivity.33 In this article, we describe the features of
the ylide required to give high enantioselectivity, the application
of the products in synthesis, and mechanistic and computational
studies that have now established the rate- and therefore enantio-
determining step of the reaction.
values.29 The absolute stereochemistry of this product was found
to be (2R,3S), and the epoxide had a negative [RD] value. It
was assumed, by analogy, that all other epoxides prepared during
this study also had the absolute configuration (2R,3S).
Using the hydroxy-substituted sulfide 2a (R1 ) H, Table 1,
entries 1-3), a very poor yield was observed for the reaction
involving the primary amide (3a, entry 1), and the enantiose-
lectivity was also low. The secondary amide (4a) also gave very
low enantiomeric excess (entry 2). However, upon changing
the substrate to a tertiary amide, excellent yields and enanti-
oselectivities were observed (entry 3). In addition, it can be seen
that changing from the hydroxy sulfonium salt 1a to the
methoxy-substituted salt 1b led to a slightly higher yield and
an increase in enantioselectivity (entry 4).
Using methoxy-substituted sulfonium salt 1b, which gave the
best results in the original study, an investigation into substrate
scope was undertaken using a variety of aldehydes (Table 2).
The reaction was found to be general for a range of aromatic
and heteroaromatic aldehydes with equally high levels of
enantiomeric excess (see Supporting Information for full details).
Aliphatic aldehydes, however, were found to be more unpredict-
able. Whereas tertiary and monosubstituted aliphatic aldehydes
resulted in products with high enantioselectivity (93% and 63%
ee, respectively; entries 4 and 2), secondary aldehydes, rather
inexplicably, gave very low selectivity. In all cases, the trans
epoxide was formed exclusively.
Further functionalization of the glycidic amide products was
then carried out.33 It was possible to convert the epoxy amide
into the corresponding epoxy ketone with complete chemose-
lectivity using an organolithium reagent (Scheme 3). Subjecting
the epoxy ketone to m-CBPA gave the corresponding ester in
good yield, which could then be hydrolyzed to the corresponding
carboxylic acid (vide infra). Direct hydrolysis of the tertiary
amide to the acid in the presence of the sensitive epoxide moiety
was not possible.
In addition, it was possible to carry out a regioselective
nucleophilic ring-opening of the epoxides. It was initially
assumed that ring-opening would occur preferentially at the
benzylic position, but addition of thiophenol in a range of
Results and Discussion
Sulfide 2a was synthesized in high yield and only two steps
from D-camphor30 and was readily methylated to give sulfide
2b.31 With the desired hydroxy- and methoxy-substituted chiral
sulfides (2a and 2b) in hand, the corresponding sulfonium salts
were prepared in excellent yield using a variety of R-bromoa-
mides, including primary, secondary, and tertiary amides. In
some cases, a diastereomeric mixture of sulfonium salts, varying
by their stereochemistry at the sulfur atom, was obtained that
was recrystallized prior to use to furnish the diastereomerically
pure salt. Brief optimization of the reaction conditions showed
that high enantioselectivities could be obtained using KOH in
EtOH at -50 °C (for a discussion of base effects, see ref 33).
It should be noted that enantioselectivities were determined
by chiral HPLC, and the absolute configuration of the major
product from the addition of the sulfonium salt 1a to benzal-
dehyde was determined by [RD] and comparison with literature
(32) Ye, S.; Huang, Z.-Z.; Xia, C.-A.; Tang, Y.; Dai, L.-X. J. Am. Chem. Soc.
2002, 124, 2432-2433.
(33) Aggarwal, V. K.; Hynd, G.; Picoul, W.; Vasse, J.-L. J. Am. Chem. Soc.
2002, 124, 9964-9965.
9
J. AM. CHEM. SOC. VOL. 128, NO. 6, 2006 2107

A R T I C L E S
Aggarwal et al.
Table 2. Asymmetric Synthesis of (2R,3S)-2,3-Epoxy Amides
Using Camphor-Derived Sulfonium Salt 1b and a Variety of
Aldehydes
and -OMe-substituted sulfonium salts. In most cases, the
enantioselectivities observed for the OMe derivative were
superior.
Initially, salts with bromide as the counteranion were
employed, but investigations were also made into the use of
other counterions including -BPh4 and -ClO4. The respective
salts had different solubilities and, therefore, in some cases,
facilitated the recrystallization process. These salts could be
prepared via an ion-exchange method that involved stirring the
bromide sulfonium salt with, for example, NaBPh4 overnight
in a minimum volume of acetone. Complete ion exchange could
be observed by NMR studies. It was observed, however, that a
change in counteranion did not have much influence on the
observed enantioselectivity (see entries 2, 4, and 5). In general,
as previously observed, we saw slight increases in enantiose-
lectivity on going from the hydroxy sulfide to the methoxy-
derived sulfide (compare entries 5 and 6 and entries 12 and
13).
a The enantiomeric excess was determined by chiral HPLC, and the
absolute configuration of the product was determined by comparison of
HPLC elution orders and [RD] values with the literature (see Supporting
Information for further details and an expanded table of results). b Epoxi-
dation was carried out at -30 °C. c Epoxidation was carried out at -20
°C.
It is interesting to note that, in some cases, temperature does
not appear to have an effect on the enantioselectivity of the
reaction. Comparing entries 2 and 3 (which were carried out at
RT and -50 °C, respectively), it can be seen that decreasing
the temperature from RT to -50 °C gave only a 2% increase
in enantiomeric excess. However, there is a much larger effect
for salt 7a (entries 7 and 8). In this case, there was an increase
from 81% to 97% ee when the temperature was lowered. It is
not clear why some amides are more sensitive to temperature
effects than others.
solvents with different bases furnished a 1:1 mixture of
regioisomers. Reports by Berhens and Sharpless34 of regiose-
lective C3-ring opening of glycidic esters and secondary amides35
using Ti(OiPr)4 were also tested, but they also gave a 1:1 mixture
of regioisomers. Finally, it was discovered that Yb(OTf)3
catalyzed the ring-opening with complete regioselectivity at the
C3 position for both S and N nucleophiles (Scheme 3).
Scheme 3. Further Functionalization of Epoxy Amides
From Tables 1 and 3, we can conclude that, whereas dialkyl
and cyclic amides are good substrates for our epoxidation
process, primary and secondary amides and morpholine-
substituted amides are not useful substrates for this reaction.
Although formation of the desired sulfonium salts occurred
uneventfully and generally in high yield, in each case, a
diastereomeric mixture of sulfonium salts was obtained (ratio
1
was established by H NMR spectroscopy), ranging from 3:2
to 16:1 (except for 7a, which was obtained as a single
diastereomer after alkylation). In the design and synthesis of
chiral sulfides for use in asymmetric epoxidation reactions, the
diastereomeric purity of the intermediate sulfonium salt is
usually a major consideration.36,37 It is for this reason that the
sulfides employed often either are C2-symmetric,36,38,39 in which
alkylation of either lone pair of the sulfur leads to the same
diastereomeric product, or have a conformationally rigid
structure in which only one of the lone pairs of the sulfur is
available for alkylation.36,40 In the specific case of this work,
the importance of using a single diastereomer of sulfonium salt
was realized at an early stage of the investigation (Table 4).
Through analysis of the stereochemical outcomes of different
diastereomeric mixtures of sulfonium salts, it was shown that
the higher the diastereomeric purity of the salt, the higher the
enantioselectivity of the epoxide (entries 5 and 6). It is also
interesting to note that the pure minor diastereomer of salt 1b
Although epoxidation had been shown to work well with the
N,N-diethyl amide derivatives, the limitations that had been
observed, in terms of transformation into the corresponding
carboxylic acid, as well as our interest in the effect of the amide
functionality in the epoxidation reaction, prompted us to
investigate other amide functionalities (Table 3). We hoped that
this would help us identify the features required for selectivity,
but it was also thought that the synthesis of enantiopure epoxides
containing more versatile amide functionalities would improve
the applicability of this method in the synthesis of complex
target compounds. Because we had observed that primary and
secondary amides did not give useful levels of enantioselectivity,
our attention was focused on the use of a range of other tertiary
amide substrates that could be readily prepared, would give high
enantioselectivity in the epoxidation step, and might be more
easily hydrolyzed. In our efforts to understand the features
responsible for high enantioselectivity, we tested both the -OH-
(36) Aggarwal, V. K.; Winn, C. L. Acc. Chem. Res. 2004, 37, 611-620.
(37) Aggarwal, V. K.; Kalomiri, M.; Thomas, A. P. Tetrahedron: Asymmetry
1994, 5, 723-730.
(38) Winn, C. L.; Bellanie, B.; Goodman, J. M. Tetrahedron Lett. 2002, 43,
5427-5430.
(39) Julienne, K.; Metzner, P.; Henyron, V.; Greiner, A. J. Org. Chem. 1998,
63, 4532-4534.
(34) Berhens, C. H.; Sharpless, K. B. J. Org. Chem. 1985, 50, 5696-5704.
(35) Tertiary amides were also tested but gave a mixture of regioisomers.
(40) Aggarwal, V. K.; Charmant, J. P. H.; Dudin, L.; Porcelloni, M.; Richardson,
J. Proc. Natl. Acad. Sci. U.S.A. 2004, 101, 5467-5471.
9
2108 J. AM. CHEM. SOC. VOL. 128, NO. 6, 2006

Highly Enantioselective Synthesis of Glycidic Amides
A R T I C L E S
Table 3. Asymmetric Synthesis of (2R,3S)-2,3-Epoxy Amides Using a Range of Camphor-Derived Sulfonium Salts
a The enantiomeric excess was determined by chiral HPLC, and the absolute configuration of the product in entry 1 was determined by comparison of
HPLC elution order with the literature. All others are given by analogy. b These reactions were carried out using p-Cl-PhCHO as the substrate. c Epoxidation
was carried out at room temperature. d Epoxidation was carried out at 0 °C. e Pure minor diastereoisomer was used. n/d ) yield not determined.
yielded the same major enantiomer of epoxide but with lower
enantioselectivity (54% ee as opposed to 97% ee; compare
entries 2 and 4).33 Keeping this in mind and comparing our
results with those obtained by Dai et al. under very similar
reaction conditions (entry 3),41 we believe that our superior
enantioselectivity is due to the use of a single diastereomer of
sulfonium salt. Fortunately, in most cases, the major diastere-
omer of the sulfonium salt could be isolated either by recrys-
tallization (this was aided in some cases by counteranion
exchange, vide supra) or by column chromatography. The
stereochemistry at sulfur for the major isomer of sulfonium salt
6b was determined by X-ray analysis to be R (see Figure 1),
and thus, the stereochemistry at sulfur of the major diastereomer
of the rest of the sulfonium salts has been assigned by analogy.
Ring Opening of Epoxides and Applications. Whereas
diallyl amide was resistant to hydrolysis to the corresponding
glycidic acid, the epoxide could be opened by thiophenol. As
in the case of the diethyl amide, a mixture of regioisomers in a
1:1 ratio was obtained in the presence of Ti(OiPr)4.34 However,
use of Yb(OTf)3 led to ring opening of amide 12 with complete
regioselectivity at the C3 position with both sulfur and nitrogen
nucleophiles (Scheme 4). According to the literature, cleavage
of N-allyl substituents of amide groups can be effected using
either rhodium,42,43 ruthenium,44 or nickel45 catalysts, but in our
hands, all attempts to prepare either the primary or secondary
(41) Although our reaction was carried out at -50 °C rather than room
temperature, we believe that temperature does not have a great effect on
the enantioselectivity. Indeed, we have shown that, whereas epoxidation
at -50 °C gives a product with 92% ee, the corresponding reaction at room
temparture resulted in an epoxide with 87% ee (see Table 3 and ref 33).
(42) Lessen, T. A.; Demko, D. M.; Weinreb, S. M. Tetrahedron Lett. 1990, 31,
2105-2108.
(43) Moreau, B.; Lavielle, S.; Marquet, A. Tetrahedron Lett. 1977, 18, 2591-
2594.
(44) Alcaide, B.; Almendros, P.; Alonso, J. M.; Aly, M. F. Org. Lett. 2001, 3,
3781-3784.
(45) Taniguchi, T.; Ogasawara, K. Tetrahedron Lett. 1998, 39, 4679-4682.
9
J. AM. CHEM. SOC. VOL. 128, NO. 6, 2006 2109

A R T I C L E S
Aggarwal et al.
Table 4. Investigation into Importance of Diastereomeric Purity of Sulfonium Salt
a The enantiomeric excess was determined by chiral HPLC, and the absolute configuration of the product in entry 1 was determined by comparison of
HPLC elution order with the literature. All others are given by analogy. b Results from ref 29; reaction carried out in CH3CN at room temperature. c Pure
minor diastereoisomer was used. d These reactions were carried out using p-Cl-PhCHO as the substrate. e Epoxidation was carried out at 0 °C.
regioisomeric product (ring opened in the C3 position) according
to the method of Berhens and Sharpless (Scheme 5).34 The
selectivity for ring opening at C3 in this case was probably
facilitated by the much bulkier amide substituents. Unfortu-
nately, efforts to hydrolyze this amide, either before or after
ring opening, also proved to be unsuccessful.
Scheme 5. Ring Opening of Bis-PMB-Protected Epoxy Amide 14
Following the work of Shibasaki et al.,48 who successfully
employed an epoxy acyl pyrrole as a versatile intermediate and
glycidic ester surrogate, epoxide 16 was prepared in good yield
Figure 1. X-ray structure of sulfonium salt 6b.
and high enantioselectivity (entries 11 and 13, Table 3). The
product was then oxidized by DDQ to give the pyrrole
intermediate 17. It was not possible to prepare the pyrrole
epoxide 17 directly, as the corresponding pyrrolamido sulfonium
salt was unreactive toward the aldehyde. After ring opening
using PhSH in the presence of Yb(OTf)3, the pyrrole was readily
hydrolyzed to the corresponding acid (19) by treatment with
NaOH in good yield (Scheme 6).
amide proved to be fruitless both in the presence of the sensitive
epoxide functionality (12) and even after ring opening had
occurred (13).
Scheme 4. Ring Opening of Diallyl-Substituted Epoxy Amide 12
Scheme 6. Ring Opening and Hydrolysis of Epoxy Acyl Pyrrole
17
Efforts then focused on the bis-PMB-protected amide, as it
was anticipated that this could be cleaved by use of CAN46 or
DDQ.47 Again, good yields and excellent selectivities were
observed for the formation of the desired glycidic amide 14,
and this time, ring-opening of the epoxide with a sulfur
nucleophile in the presence of Ti(O-iPr)4 yielded a single
(46) Yamaura, M.; Suzuki, T.; Hashimoto, H.; Yoshimura, J.; Okamoto, T. Bull.
Chem. Soc. Jpn. 1985, 58, 1414.
(47) Grunder-Klotz, E.; Ehrhardt, J.-D. Tetrahedron Lett. 1991, 32, 751-752.
9
2110 J. AM. CHEM. SOC. VOL. 128, NO. 6, 2006

Highly Enantioselective Synthesis of Glycidic Amides
A R T I C L E S
Scheme 7. Two Approaches to the Total Synthesis of SK&F 104353a
a Reagents and conditions. (a) t-BuNH2, toluene, 110 °C, 93%. (b) (i) LDA, Ph(CH2)7Br, THF, -78 to 60 °C; (ii) HCl, 85%. (c) 22, KOH, MeOH, -30
°C, 24 h, 77%. (d) (i) PhLi, THF, -78 °C, 84%; (ii) m-CPBA, CH2Cl2, ∆, 68%. (e) HS-(CH2)2-CO2Me, Yb(OTf)3, CH2Cl2, -78 °C to room temperature,
64%. (f) NaOH, MeOH, H2O, room temperature, 16 h, 81%. (g) 22, KOH, EtOH, -50 °C, 48 h, 62%. (h) DDQ, dioxane, 85 °C, 8 h, 81%. (i) HS-
(CH2)2-CO2Me, Yb(OTf)3, CH2Cl2, -78 °C to room temperature, 58%. (j) NaOH, DME, H2O, room temperature, 16 h, 85%.
In summary, we have shown that it is possible to prepare
epoxy amides in good yield, complete diastereoselectivity, and
excellent enantioselectivity when secondary alkyl and cyclic
amides are employed as substrates. The importance of using a
diastereomerically pure sulfonium salt to obtain epoxides in high
enantioselectivity has been illustrated, and we have shown that,
although low temperature is an important consideration for some
amides, in other cases, equally high selectivities can also
obtained at room temperature. In addition, we have demonstrated
that the epoxy amide products can be further functionalized by
nucleophilic ring opening under Lewis acidic conditions. The
corresponding carboxylic acids can be obtained either by the
addition of organolithium reagents followed by Baeyer-Villiger
oxidation and subsequent hydrolysis or by use of an epoxy acyl
dihydropyrrole that can be oxidized and then subsequently
hydrolyzed to the carboxylic acid.
treatment of bronchial asthma.49 The shortest synthesis of SK&F
104353 to date was reported by Lantos et al.50 In their synthesis,
the desired epoxide was formed in 95% ee using the Julia´-
Collona epoxidation10 on an R,â-unsatutated ketone with a
polyleucine catalyst. The target compound was obtained in a
35% overall yield in eight steps.
We were able to prepare this molecule by two complementary
routes, both of which involved the sulfur ylide methodology
discussed above as the key step (Scheme 7). Route A involving
the diethylamide sulfonium salt 1b, was reported in a com-
munication33 and is also described in the Supporting Informa-
tion.
Route B involved the reaction of the diastereomerically pure
sulfonium salt 10b, instead of 1b, with the known aldehyde
2251 (Scheme 7, route B), and in this case, the glycidic amide
was obtained in 81% ee. In both routes A and B, the chiral
sulfide 2b was reisolated by column chromatography in >90%
Asymmetric Synthesis of SK&F 104353. The synthetic
applicability of our new methodology for the enantioselective
synthesis of glycidic amides was showcased in the synthesis of
SK&F 104353 (20), a leukotriene D4 inhibitor for the potential
(49) Mong, S.; Wu, H. L.; Miller, J.; Hall, R. F.; Gleason, J. G.; Crooke, S. T.
Mol. Pharmacol. 1988, 32, 223.
(50) Flisak, J. R.; Gombatz, K. J.; Holmes, M. M.; Jarmas, A. A.; Lantos, I.;
Mendelson, W. L.; Novack, V. J.; Renich, J. J.; Schneider, L. J. Org. Chem.
1993, 58, 6247.
(48) Nemoto, T.; Ohshima, T.; Shibasaki, M. J. Am. Chem. Soc. 2001, 123,
9474-9475.
(51) Firth, M. A.; Mitchell, M. B.; Smith, S. A. C. J. Org. Chem. 1995, 59,
2616.
9
J. AM. CHEM. SOC. VOL. 128, NO. 6, 2006 2111

A R T I C L E S
Aggarwal et al.
yield. Oxidation of 24 using DDQ afforded the pyrrole derivative
25 in high yield, and this epoxy acyl pyrrole was then opened
with 3-mercaptopropionate in the presence of Yb(OTf)3. The
ring-opened derivative 26 was formed as a 14:1 mixture of
regioisomers that were separable by silica gel chromatography.
The major isomer was obtained in moderate yield and was
subsequently hydrolyzed to give the target compound 20. Thus,
the target compound was formed via this complementary route
in a sequence of five linear steps with an overall yield of 23%.
Origins of Diastereoselectivity and Enantioselectivity. The
mechanism of the sulfur-ylide-mediated epoxidation reaction
involves two key steps along each of the diastereoisomeric
pathways leading to cis and trans epoxides (see Figure 2 for
semistabilized ylide epoxidations).52 The first is addition of the
ylide onto the aldehyde to form a betaine intermediate. This
initial addition step occurs via a quasi-[2 + 2] mode of addition
(i.e., where the sulfonium and oxy groups are gauche with
respect to each other), which makes favorable Coulombic
interactions between the dipoles of the polar reactants possible.52
The betaine formed initially is therefore in a cisoid conformation.
In order for the second key step to occur, this betaine needs to
undergo rotation around the newly formed carbon-carbon bond
to give the corresponding transoid conformer. This latter betaine
can then ring close to yield the desired epoxide.
independently generated by another route in the presence of a
more reactive aldehyde. If betaine formation were nonreversible,
direct ring closure would occur without incorporation of the
more reactive aldehyde and ring closure would be a stereospe-
cific process, whereas if betaine formation were reversible, the
more reactive aldehyde would be trapped and trans epoxide
would be preferentially obtained even in the case of the syn
betaine.
The substrates for the crossover experiments, 27 and 28, were
prepared by an aldol reaction, followed by separation of the
diastereoisomers and subsequent alkylation with Meerwein’s salt
(Scheme 8). Treatment of either the anti or syn â-hydroxy
sulfonium salt (27 or 28, respectively) with a base in the
presence of a more reactive aldehyde (p-NO2C6H4CHO) gave
only trans epoxide in which the more reactive aldehyde had
been incorporated (Scheme 8).
Scheme 8. Crossover Experiments Indicating that the Betaine
Formation Is Reversiblea
Two isomeric betaines can be formed during the addition
step: an anti and a syn diastereomer. After torsional rotation
and ring closure, these betaines lead, respectively, to a trans
and a cis epoxide. Through experimental53,54 and computa-
tional52 investigations, we recently showed that, in the reaction
of semistabilized ylides (R2SCHPh), the high trans selectivity
is a result of similar rates of formation for both syn and anti
betaines, where formation of the syn betaine is reversible but
the anti betaine does not reverse and leads directly to the trans
epoxide (Figure 2). This difference in reactivity between the
two diastereomeric pathways was found to be mainly due to
the high barrier for the torsional rotation step for the syn betaine.
a Reagents and conditions. (i) LDA. -78 °C, LiCl, PhCHO; (ii)
Me3OBF4, CH2Cl2; (iii) 3 equiv of p-NO2C6H4CHO, 1.5 equiv of NaOH,
CH2Cl2, H2O, room temperature.
This indicates that, in the case of amide-stabilized ylides, both
diastereomeric betaines reverse to reactants, which means that
the rate- and selectivity-determining step must lie after the
addition step in both cases. The exclusive trans selectivity
observed in these reactions must therefore result from a
significant barrier in either the bond rotation or ring-closure
step of the syn betaine, which results in reversal back to the
ylide and aldehyde. But which of the two steps (bond rotation
or ring closure) is rate-determining? Is it the same in both
diastereomeric pathways? What is the origin of the exclusive
trans epoxide formation? To answer these questions, we
investigated the energy profile of the reaction of ylide 29 with
benzaldehyde computationally (Figure 3). Calculations55 were
carried out at the B3LYP/6-311+G**//B3LYP/6-31G* level,
including a continuum description of ethanol as the solvent (see
the Supporting Information for computational details). Such
methods have already been shown to describe accurately energy
profiles of similar systems.52,56
Figure 2. Mechanism of semistabilized ylide epoxidation.
In the case of amide-stabilized ylides, one might expect
greater reversibility in betaine formation because the ylide is
more stable than in the case of phenyl-stabilized ylides. To test
whether betaine formation is reversible, a crossover experiment
was conducted in which the two diastereomeric betaines were
As previously observed for reaction of semistabilized ylides,52
the addition of the stabilized ylide 29 to benzaldehyde occurs
via a quasi-[2 + 2] approach of the reactants, leading to the
betaines in their cisoid conformation. Given the stabilization
(52) Aggarwal, V. K.; Harvey, J. N.; Richardson, J. J. Am. Chem. Soc. 2002,
124, 5747-5756.
(53) 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, 10926-10940.
(54) Aggarwal, V. K.; Richardson, J. Chem. Commun. 2003, 2644-2651.
(55) Calculations were carried out using the Jaguar 4.0 pseudospectral program
package (Jaguar 4.0; Schro¨dinger, Inc.: Portland, OR, 1991-2000).
Relative energies correspond to electronic energies at the indicated level
of theory. See Supporting Information for computational details.
(56) Aggarwal, V. K.; Harvey, J. N.; Robiette, R. Angew. Chem., Int. Ed. 2005,
44, 5468-5471.
9
2112 J. AM. CHEM. SOC. VOL. 128, NO. 6, 2006

Highly Enantioselective Synthesis of Glycidic Amides
A R T I C L E S
Figure 3. Computed potential energy profile for the epoxidation reaction between benzaldehyde and ylide 29. Energies were obtained at the B3LYP/6-
311+G**(ethanol)//B3LYP/6-31G*(ethanol) level and are given in kcal/mol relative to reactants.
of the reactants, this step is endothermic (6-8 kcal/mol; for
comparison, the energy of this step was between -2 and -3
kcal/mol for the semistabilized Me2SCHPh ylide52) and, con-
sequently, occurs with a significant energy barrier (10-12 kcal/
mol; the barrier is ∼4 kcal/mol for Me2SCHPh ylide52). The
syn diastereomer of the betaine is found to be more stable and
formed through a lower barrier than the anti one. However, for
both diastereomeric cisoid betaines, the energy barrier to
torsional rotation is close to the barrier to reverse to reactants.
Moreover, compared to a phenyl substituent, the amide group
imposed a higher barrier to ring closure,57 and the highest point
on the potential energy surface was found to be, in both cases,
the transition state of the ring-closure step. Therefore, the
addition and the rotation steps should be entirely reversible for
both diastereomeric pathways. In other words, the rate-
determining, and thus diastereoselectivity-determining, step is
ring closure (Figure 4).
addition to form the syn betaine is faster than formation of the
anti betaine. However, ring closure of the syn betaine to the cis
epoxide is highly disfavored because of developing steric
interactions between the phenyl and amide substituents, so that
syn betaine formation is unproductive. Reversion to reactants
is followed by addition to form the anti betaine, which undergoes
ring closure much more rapidly, leading to the observed
preferential formation of trans epoxide. It is harder to account
for the high enantioselectivity that is observed.
Scheme 9. Transition State Proposed by Dai et al.29 to Explain
the Enantioselectivity
Dai et al. proposed that sulfonium salt 1a reacts with
aldehydes via transition state 30 (Scheme 9), in which hydrogen
bonding controls the conformation of the ylide and nonbonding
interactions between the ylide and aldehyde are responsible for
the enantioselectivities observed.29 However, our calculations
and crossover experiments indicate that selectivity in the addition
step is unimportant in determining the enantioselectivity of the
overall process. If we assume similar energy profiles for the
reaction of the chiral ylide 1a and for 29, the enantioselectivity
must be determined by the relative energy of enantiogenic
transition states in the ring-closure step. The exact nature of
the interactions in these transition states that are responsible
for the high enantioselectivity observed is, however, very
difficult to identify because of the flexibility of the system.
Figure 4. Origin of diastereo- and enantioselectivity in stabilized ylide
reactions.
The exclusive formation of trans epoxide in reactions of
stabilized ylides can thus be explained as follows: Initial
Conclusions
(57) The high barrier to ring closure seems to be due to the substitution of the
electrophilic carbon by an electron-withdrawing group (an amide). This is
in contradiction to the high reactivity usually observed for R-halogenated
carbonyl compounds in SN2 reactions. Studies aiming at understanding this
dissimilarity are currently underway in our laboratories.
We have described a completely diastereoselective, highly
enantioselective, and practical process for the synthesis of
epoxy amides. To achieve high diastereoselectivity, single
9
J. AM. CHEM. SOC. VOL. 128, NO. 6, 2006 2113

A R T I C L E S
Aggarwal et al.
diastereomers of sulfonium salts bearing tertiary amides are
required. The epoxy amide products were further transformed
by ring opening with a variety of nucleophiles, followed by
hydrolysis of the amide to yield the corresponding carboxylic
acid. In addition, we have demonstrated how this methodology
can be applied to two complementary syntheses of SK&F
104353.
computational study allowed us to identify the rate- and
selectivity-determining step as being ring closure, not bond
rotation as was the case with phenyl-stabilized ylides.
Supporting Information Available: Full experimental details,
compound characterization data, X-ray structures, computational
details, and tables with optimized Cartesian coordinates and
corresponding energies for all species discussed in the text. This
material is available free of charge via the Internet at
http://pubs.acs.org.
Moreover, we have enhanced our mechanistic understanding
of this process. Crossover experiments indicated that the
formation of both the syn and anti betaines is reversible, and a
JA0568345
9
2114 J. AM. CHEM. SOC. VOL. 128, NO. 6, 2006