Glycosylation catalyzed by a chiral bronsted acid

Article abstract of DOI:10.1021/ol1001895

"Chemical equation presented" The use of a chiral Bronsted acid catalyst for the activation of trichloroacetimidate glycosyl donors has been demonstrated for the first time. In toluene the chirality of the acid catalyst is seen to influence the stereochem

Full text of DOI:10.1021/ol1001895

ORGANIC
LETTERS
2010
Vol. 12, No. 7
1452-1455
Glycosylation Catalyzed by a Chiral
Brønsted Acid
Daniel J. Cox,† Martin D. Smith,† and Antony J. Fairbanks*,‡
Department of Chemistry, Chemistry Research Laboratory, UniVersity of Oxford,
Mansfield Road, Oxford, OX1 3TA, U.K., and Department of Chemistry, UniVersity of
Canterbury, PriVate Bag 4800, Christchurch 8140, New Zealand
antony.fairbanks@canterbury.ac.nz
Received January 25, 2010
ABSTRACT
The use of a chiral Brønsted acid catalyst for the activation of trichloroacetimidate glycosyl donors has been demonstrated for the first time.
In toluene the chirality of the acid catalyst is seen to influence the stereochemical outcome of the glycosylation processes, hinting that
perhaps diastereocontrol of glycosylation processes may become achievable through the judicious use of chiral organic catalysts.
The conceptually simple process of linking carbohydrate
units by glycosylation has proven to be one of the most
difficult synthetic processes to control from a stereochemical
perspective.1 Although significant advances have been made,
ranging from neighboring group participation of 2-O-acyl
protected glycosyl donors2 and recent refinements,3 to
intramolecular strategies,4 stereochemical control of glyco-
sylation is one of the last areas of diastereoselective synthesis
to be entirely resolved.
The techniques of asymmetric synthesis have frequently
been applied to control a variety of processes, hinting that
reagent control could perhaps be applied to influence the
diastereochemical outcome of glycosylation processes. How-
ever it is difficult to see how, for example, the use of
asymmetric transition metal complexes could be applied in
glycosylation chemistry, and correspondingly there are only
very scant reports in the literature of previous efforts that
have been made to control the stereochemical outcome of
such reactions using chiral catalysts. Recent developments
in the field of organocatalysis5 may represent a paradigm
shift since these principles could be applied to control the
stereochemistry of glycosylation processes.6 Among the
many emerging applications of organocatalysis, the use of
asymmetric Brønsted acid catalysis has been noteworthy. In
particular the levels of stereochemical control obtained using
† The University of Oxford.
‡ The University of Canterbury.
(1) Demchenko, A. V. In Handbook of Chemical Glycosylation:
AdVances in StereoselectiVity and Therapeutic ReleVance; Demchenko,
A. V., Ed.; Wiley-VCH: Weinheim, Germany, 2008; pp 1-27.
(2) Wulff, G.; Ro¨hle, G. Angew. Chem., Int. Ed. Engl. 1974, 13157-170.
Wulff, G.; Ro¨hle, G. Angew. Chem. 1974, 86, 173–187.
(3) (a) Kim, J. H.; Yang, H.; Boons, G. J. Angew. Chem., Int. Ed. 2005,
44, 947–949. (b) Kim, J. H.; Yang, H.; Park, J.; Boons, G. J. J. Am. Chem.
Soc. 2005, 127, 12090–12097. (c) Park, J.; Kawatkar, S.; Kim, J. H.; Boons,
G. J. Org. Lett. 2007, 9, 1959–1962. (d) Cox, D. J.; Fairbanks, A. J.
Tetrahedron: Asymmetry 2009, 20, 773–780.
(5) MacMillan, D. W. C. Nature 2008, 455, 304–308.
(6) For a recent approach to model glycosylation using a chiral thiourea
catalyst see: Reisman, S. E.; Doyle, A. G.; Jacobsen, E. N. J. Am. Chem.
Soc. 2008, 130, 7198–7199.
(4) (a) Jung, K.-H.; Mu¨ller, M.; Schmidt, R. R. Chem. ReV. 2000, 100,
4423–4442. (b) Fairbanks, A. J. Synlett 2003, 1945–1958.
10.1021/ol1001895  2010 American Chemical Society
Published on Web 03/03/2010

BINOL-derived phosphoric acid catalysts, such as 1, have
in many cases exceeded expectations in a quite spectacular
fashion.7
First the actual feasibility of activation of donor 2 by a
Brønsted acid was investigated using diacetone galactose 3
as the acceptor (Scheme 1, Table 1). Trichloroacetimidates
are most commonly activated using catalytic quantities of
Lewis acids, such as boron trifluoride etherate (BF3·OEt2)
or trimethylsilyl triflate (TMSOTf). Therefore to provide a
benchmark control process for investigation of any stereo-
chemical effects of using a chiral acid activator, activation
of donor 2 was first performed in the presence of 3 using
catalytic TMSOTf. The desired disaccharide 4 was rapidly
formed in excellent yield as an almost equimolar anomeric
mixture of compounds in which the R-anomer predominated
slightly (4r:4ꢀ, 1.2:1, Table 1, entry 1).
We therefore initiated an investigation to see if asymmetric
Brønsted acid catalysis could be used to control the stere-
ochemical outcome of glycosylation reactions. Considering
the wide variety of glycosyl donors that are available, acid-
catalyzed activation should be a straightforward process to
achieve. In particular, glycosyl trichloroacetimidates,8 one
of the most reliable and generally applicable classes of
glycosyl donors, should be amenable to activation by
catalytic quantities of a Brønsted acid. Toste recently reported
the activation of trichloroacetimidate leaving groups in a
noncarbohydrate context using a chiral Brønsted acid.9 We
therefore sought to investigate the feasibility of effecting
glycosylation of glycosyl trichloroacetimidate donors using
both enantiomers of the chiral BINOL-derived phosphoric
acid 110 as chiral catalytic activators.
Table 1. Yield and Stereoselectivity of Glycosylation of
Acceptor 3 by Donor 2a
entry
activator (15 mol %)
yield of 4 (%)
ratio of 4r:4ꢀ
1
2
3
4
TMSOTf
(R)-1
(S)-1
98b
88c
80c
19d
1.2:1
1:2
1:7
(PhO)2P(O)OH
1:1.9
a Anomeric ratio of donor 2r:2ꢀ, 7.9:1. b Reaction complete after 15
min at rt. c Reactions required 48 h at rt to reach completion. d Isolated
yield of product after 48 h at rt.
Glycosylation was then performed using a catalytic
quantity of (R)-1 (15 mol %). Although the reaction was
considerably slower than when TMSOTf was used as
activator, requiring 48 h to reach completion at rt, efficient
glycosylation did occur, and disaccharide 4 was isolated in
88% yield. Interestingly the anomeric ratio of the products
was altered, and the ꢀ-anomer now predominated (4r:4ꢀ,
1:2, Table 1, entry 2). To investigate whether the chirality
of the acid catalyst affected the stereochemical outcome of
the glycosylation reaction, activation of donor 2 in the
presence of 3 was then undertaken in the presence of a
catalytic quantity of the (S)-enantiomer of 1. The disaccharide
product 4 was isolated in similar yield to when (R)-1 had
been used, but significantly, the anomeric ratio of product
was altered considerablysthe ꢀ-anomer was now markedly
favored over the R-anomer (4r:4ꢀ, 1:7, Table 1, entry 3). A
second control experiment was then undertaken, namely
glycosylation of donor 2 with acceptor 3 catalyzed by an
achiral phosphoric acid, diphenylphosphoric acid being
selected as a suitable activator (Table 1, entry 4). Glycosy-
lation catalyzed by (PhO)2P(O)OH was considerably slower
than by either enantiomer of 1, and the product was formed
in only 19% yield after 48 h. Although the stereochemical
outcome of the reaction was similar to that obtained using
Figure 1. Chiral BINOL-derived Brønsted acids 1.
To investigate any dependence of the stereochemical
outcome of glycosylation upon the stereochemistry of the
catalytic activator, protection of the hydroxyl groups of the
glycosyl donor with nonparticipating protecting groups, such
as benzyl ethers, would be required so that neighboring group
participation did not predominate. Thus the tetrabenzyl-
protected galactose donor 2 was synthesized and used as a
donor in a variety of glycosylation processes.
Scheme 1. Glycosylation of a Trichloroacetimidate Donor
Catalyzed by Chiral Brønsted Acid
(7) (a) Terada, M. Chem. Commun. 2008, 4097–4112. (b) Akiyama, T.
Chem. ReV. 2007, 107, 5744–5758.
(8) (a) Schmidt, R. R.; Michel, J. Angew. Chem., Int. Ed. 1980, 19, 731.
(b) Schmidt, R. R. Angew. Chem., Int. Ed. 1986, 25, 212.
(9) Hamilton, G. L.; Kanai, T.; Toste, F. D. J. Am. Chem. Soc. 2008,
130, 14984–14986.
(10) Akiyama, T.; Morita, H.; Itoh, J.; Fuchibe, K. Org. Lett. 2005, 7,
2583–2585.
Org. Lett., Vol. 12, No. 7, 2010
1453

(R)-1 (4r:4ꢀ, 1:1.9), it was significantly different to that
obtained using (S)-1 as catalyst. These results demonstrate
that the stereochemical outcome of the glycosylation reaction
is dependent on the chirality of the acid catalyst used, and
indicate that the acid, or most likely its counterion, is
involved in the glycosylation step. On the basis of previous
explanations for the stereodirecting effects of chiral BINOL-
derived acids in other stereoselective processes, this is
presumably by way of ion-pair formation. Such tight ion pair
formation would be expected to be a solvent-dependent
process. Solvent effects are well-known in glycosylation
chemistry,11 and frequently solvent participation can be used
to increase the formation of one desired anomer over the
other. Therefore the effect of reaction solvent on the
stereochemical outcome of glycosylation of 2 with acceptor
3 was next investigated using the less selective (R)-
enantiomer of 1 (Table 2).
Table 3. Glycosylation of Donor 2a with a Variety of Glycosyl
Acceptors 5a-c
activator
entry acceptor (15 mol %)
product/%
yield
time
(h)
R:ꢀ ratio
1
2
3
4
5
6
7
8
9
5a
5a
5a
5b
5b
5b
5c
5c
5c
TMSOTf
(R)-1
(S)-1
TMSOTf
(R)-1
(S)-1
TMSOTf
(R)-1
(S)-1
6a/97
6a/87
6a/88
6b/99
6b/84
6b/88
6c/97
6c/71
6c/73
0.25
16
16
0.25
48
48
0.25
72
72
1:10
1:47
ꢀ only
1:1.2
1:5.7
1:70
1:3.9
1:6
Table 2. Effects of Solvent and Donor/Acceptor Concentrations
on Stereoselectivity of Glycosylation of 2a Using Acceptor 3
activator
yield of
ratio of
1:4.9
entry
solvent/concn
(15 mol %)
4 (%)
4r:4ꢀ
a Anomeric ratio of donor 2r:2ꢀ, 7.9:1.
1
2
3
4
5
toluene/73 mM
DCM
DCE
toluene/24 mM
toluene/12 mM
(R)-1
(R)-1
(R)-1
(R)-1
(R)-1
88
67
63
80
77
1:2
1:1.1
1:1
1:1.7
1:1.8
acceptors 5a-c, giving rise to products 6a-c, respectively.
In each case glycosylation was undertaken using TMSOTf
as a control achiral catalytic activator, and both the (R) and
(S) enantiomers of 1. Perhaps surprisingly the most stereo-
selective reactions were obtained using MeOH 5a as the
acceptor (Table 3 entries 1-3). With TMSOTf activation
these reactions were complete within 15 min, and the methyl
glycosides 6a were isolated in 97% yield as an anomeric
mixture in which the ꢀ-anomer predominated (5br:5bꢀ,
1:10). Use of the chiral acid 1 as catalytic activator again
led to a slower glycosylation reaction, though notably this
reaction was complete in 16 h, and was thus considerably
faster than when diacetone galactose 3 had been used as the
acceptor. Thus, use of (R)-1 led to the isolation of glycoside
5b in 87% yield as almost entirely the ꢀ-anomer (5br:5bꢀ,
1:47). Subsequently the use of (S)-1 as the activator led to
the formation of methyl glycoside 5b in 88% yield as pure
ꢀ-anomer. The variation in stereochemical outcome of these
reactions, although the ꢀ-anomer predominated in each case,
does again indicate an influence of the chirality of the catalyst
on the stereoselectivity of the glycosylation process.
a Anomeric ratio of donor 2r:2ꢀ, 7.9:1.
Changing the solvent from toluene to either dichlo-
romethane (DCM) or dichloroethane (DCE) led to an erosion
of stereoselectivity as compared to the equivalent reaction
in toluene. A plausible explanation for this observation is
that ion pair separation is more pronounced in the two more
polar chlorinated solvents, leading to a diminution of the
stereochemical influence of the chiral catalyst. Following on
from a recent report12 on the observation of interesting
concentration effects on the stereochemical outcome of
glycosylation processes led us to perform the glycosylation
of 2 with acceptor 3 as catalyzed by (R)-1 at three different
substrate concentrations in toluene. No effective changes
were seen in the stereochemical outcome of these reactions,
indicating no participation of the solvent in the glycosylation
process.
Attention then turned to the variation of the nature of the
glycosyl acceptor (Table 3). It is well-established that the
stereochemical outcome of glycosylation can be highly
dependent on the acceptor used in the glycosylation step; in
particular matching and mis-matching of donor and acceptor
may occur.13 Donor 2 was therefore glycosylated with
Attention then moved to other carbohydrate acceptors,
which themselves may have more of an inherent stereo-
chemical preference (i.e., match/mis-match with the donor).
The use of the gluco configured primary alcohol acceptor
5b with TMOSTf as the activator led to the formation of
disaccharide 6b in excellent yield as an almost equimolar
anomeric mixture (6br:6bꢀ, 1:1.2). Glycosylation catalyzed
by (R)-1, though slower (48 h to reach completion), produced
disaccharide 6b in good yield and with significantly increased
ꢀ-stereoselectivity (6br:6bꢀ, 1:7). The trend was continued
when (S)-1 was used as the catalytic activator, and disac-
charide 6b was produced as almost exclusively the ꢀ-anomer
(11) (a) Demchenko, A.; Stauch, T.; Boons, G.-J. Synlett 1997, 818.
(b) Braccini, I.; Derouet, C.; Esnault, J.; Herve´ du Penhoat, C.; Mallet,
J.-M.; Michon, V.; Sinay¨, P. Carbohydr. Res. 1993, 246, 23. (c) Sinay¨, P.
Pure Appl. Chem. 1991, 63, 519.
(12) Chao, C.-S.; Li, C.-W.; Chen, M.-C.; Chang, S.-S.; Mong, K.-K. T.
Chem.sEur. J. 2009, 15, 10972.
(13) Paulsen, H. Angew. Chem., Int. Ed. Engl. 1982, 21, 155.
1454
Org. Lett., Vol. 12, No. 7, 2010

essentially unchanged when the (R)-enantiomer of 1 was used
as activating catalyst (Table 4, entry 2). However when (S)-1
was used the reaction was again found to favor formation
of the ꢀ-anomer (ratio of 8r:8ꢀ, 1:3.4).
Table 4. Glycosylation of 4,6-Benzylidene Protected Donor 7a
with Acceptor 3
Overall the results contained herein demonstrate that the
stereochemical outcome of a glycosylation reaction can be
affected by the chirality of a catalytic acid that is used as an
activator. As compared to activation with an achiral Lewis
acid (TMSOTf) the reactions are generally seen to be more
ꢀ-selective. While the (R)-enantiomer of 1 induces quite low
levels of ꢀ-selectivity, the use of the (S)-enantiomer of 1
generally caused highly selective ꢀ-glycosylation, and in
certain cases the ꢀ-product was formed exclusively. A
corollary to these findings is that this ꢀ-selectivity is to some
extent still dependent on the identity of the glycosyl acceptor
employed. The ultimate objective of a research program such
as this would be to develop a chiral catalytic activating
system that was ꢀ-selective for one enantiomer of the
catalyst, and R-selective for the other, independently of the
identities of glycosyl donor and acceptor. Although such a
“Holy-Grail” of glycosylation chemistry remains very distant,
this first report on the feasibility of achieving effective
chemical glycosylation using a chiral catalytic activator,
together with the observation that the stereochemical outcome
of glycosylation is dependent on the chirality of the activating
catalyst, does indicate that this is certainly an avenue worthy
of further investigation.
entry
activator (15 mol %)
yield of 8 (%)
ratio of 8r:8ꢀ
1
2
3
TMSOTf
(R)-1
(S)-1
94b
76c
87c
1:1.1
1:1.2
1:3.4
a Anomeric ratio of donor 7r:7ꢀ, 10:1. b Reaction complete after 15
min at -40 °C. c Reactions required 16 h at rt to reach completion.
(6br:6bꢀ, 1:70, Table 3, entry 6). However when the manno
secondary alcohol acceptor 5c was used two things became
apparent. First the glycosylation catalyzed by TMSOTf was
itself ꢀ-selective (6cr:6cꢀ, 1:3.9). Second glycosylation
catalyzed by the chiral acids 1 was both considerably slower,
and, in this instance, no significant difference between the
stereochemical outcomes of the two reactions was observed.
In this one instance the (R)-enantiomer was the slightly more
ꢀ-selective of the two (6cr:6cꢀ, 1:6).
To conclude this initial study, the influence of the
protecting group pattern of the glycosyl donor on the
stereochemical outcome of the reaction was investigated
briefly. Thus the 4,6-benzylidene protected donor 7 was
accessed and subjected to glycosylation again using diacetone
galactose 3 as the glycosyl acceptor (Table 4). In this instance
the use of TMSOTf as activator produced disaccharide 8 as
an almost equimolar mixture of the R- and ꢀ-anomers (ratio
of 8r:8ꢀ, 1:1.1, Table 4 entry 1). This ratio remained
Acknowledgment. The authors gratefully acknowledge the
EPSRC (DTA award to D.J.C.) and the Royal Society (URF
to M.D.S.) for financial support.
Note Added in Proof. For a recent example of Pd
catalyzed activation of glycosyl trichloroacetimidate donors
which resulted in ꢀ-selective glycosylation in the absence
of neighboring group participation, see: Mensah, E. A.;
Azzarelli, J. M.; Nguyen, H. M. J. Org. Chem. 2009, 74,
1650. This paper was published ASAP March 3, 2010 and
reposted on March 11, 2010 with this Note.
Supporting Information Available: Synthetic procedures,
together with characterization and spectral data for all
compounds. This material is available free of charge via the
Internet at http://pubs.acs.org.
OL1001895
Org. Lett., Vol. 12, No. 7, 2010
1455