.
Angewandte
Communications
Table 1: Reaction of glycals with the model acceptor 3a.[a]
Scheme 2. Diastereomeric half-chair conformers of glucal oxacarbe-
nium ions and preferred nucleophilic attack to yield the corresponding
glycosides.
Entry
R1
R2
R3
Yield [%][b]
6a–j
Yield [%][b]
a/b[b] 7a–j a/b[b]
1
2
3
4
5
6
7
8
9
5a
5b TBS
5c TBS
5d TBS
5e TIPS
5 f
5g TIPS
5h
5i
Bn
Bn
TBS
Bn
Bn
ca. 74
ca. 60
ca. 60
6:1
5:1
ca. 11 7:1
philic addition to each conformer gives different diastereo-
meric products.[8,10]
TBS
TBS
TBS
34
20
4:1
4:1
4:1
n.d.
–
15:1
ca. 64[d] 15:1
2
Protecting groups can influence the reactivity of a glycosyl
donor[11] and the conformer equilibrium by influencing
thermodynamic factors.[12] For example, bulky silyl protecting
groups trans-vicinal to each other in a carbohydrate moiety
have been used to achieve higher reactivity, and in some cases
selectivity, by way of inducing conformational constraints
which favor molecules in axial-rich conformations.[12c,13]
Cyclic protecting groups can also influence the reactivity
and stereoselectivity of glycosylation reactions.[14] The effect
of 4,6-O-benzylidene acetals on the stereo-outcome of
glycosylations has been extensively studied by Crich and co-
workers[14a] in the synthesis of mannopyranosides, and 3,5-O-
di-tert-butylsilane and disiloxane acetals have been found to
favor the formation of b-arabinofuranosides.[15]
We hypothesized that protecting-group-induced confor-
mational constraints on the charged glucal-derived oxacarbe-
nium ion could be used to influence the stereoselectivity of
the glycosylation. To that end, we set out to explore the
reactivity and stereoselectivity of a range of differentially
protected glucals using the bench-stable protic acid
TsOH·H2O as an economical, easy-to-handle, and more
reactive glycosylation promoter, relative to the previously
used 1,[16] to efficiently activate the less reactive glucal
donors.[3j] Thus, glucals containing benzyl and bulky silyl
ethers (5a–e), as well as derivatives with cyclic protecting
groups such as 4,6-O-silyl and benzylidene acetal groups (5 f–
h) and 3,4-O-siloxane derivatives (5i,j) were synthesized and
reacted with 3a, as the model glycoside acceptor, in CH2Cl2
(Table 1).
Initial reactions with the perbenzylated glucal 5a and
acceptor 3a using 1 mol% of TsOH·H2O at room temper-
ature for 4 hours afforded an inseparable mixture of the
disaccharide 6a (ca. 74%, 6:1 a/b), the rearranged product 7a
(ca. 11%), and hydrolyzed starting material (Table 1,
entry 1). When the persilylated glucal 5b was subjected to
the same reaction conditions, the disaccharide 6b was formed
in yields of over 60% with similar a-selectivity as before,
alongside 34% of the rearranged 7b with a preference for the
a-product (entry 2). Reactions with the disilylated derivatives
5c–e, which bear a benzyl or allyl ether, or an acetate ester at
C4, afforded the disaccharides 6c–e in good yield (60–80%)
and with a strong preference for the a-glycosides. In these
instances, the rearranged products (7c-e) were also observed
albeit in small amounts (entries 3–5). These results show that
bulky silyl ether protecting groups can influence the stereo-
outcome of the glycosylation with glucals, as previously shown
for other glycoside donors.[12c,13] We propose that changing
Ac
Allyl
Si(tBu)2
CHPh
TIPS 80[c]
15:1
3:1
3:1
a
4[c]
TIPS
ca. 34
n.o.
n.o.
n.o.
n.o.
n.o.
ca. 39
–
–
–
–
O[Si(iPr)2]2
O[Si(iPr)2]2
O[Si(Me)2]2
Bn
76[c]
TIPS 85[c]
a
10 5j
TIPS ca. 61
a
[a] Used 1.2–1.5 equiv of 5a–j. [b] Measured by 1H NMR spectroscopy
unless stated otherwise. [c] Yield of isolated product. [d] Reaction run for
8 h. n.d.=not determined, n.o.=not observed, TBS=tert-butyldime-
thylsilyl, TIPS=triisopropylsilyl.
OR1 from a benzyl ether to a bulky silyl ether (entry 1 versus
2) hinders attack on the 3H4 conformer (Scheme 2), thus
a higher a-selectivity results.[8] Further increases in a-
selectivity are obtained by switching R2 to a small group
(entries 3–5). We speculate that this effect results from a shift
in the conformer equilibrium in the absence of vicinal bulky
groups,[13a] but we cannot rule out a change in the relative
reactivity of the conformers.
Encouraged by these results, we decided to explore the
effect of cyclic protecting groups. Reactions using the 4,6-O-
linked 5 f and 5g were slow and the disaccharides 6 f and 6g,
respectively, were produced in low yields (ca. 34–39%) along
with hydrolyzed starting materials (Table 1, entries 6 and 7).
However no rearranged products were observed . This result
was not completely unexpected as it has been shown[17] that
conformational restraints imposed by these type of trans-
fused bicyclic systems can disfavor the formation of oxacar-
benium ions, and thus such substrates tend to be less reactive
than the noncyclic derivatives.
Excitingly, reaction of the 3,4-O-tetraisopropyldisiloxane
derivatives 5h and 5i under the same reaction conditions as
before, afforded the disaccharides 6h (76%) and 6i (85%),
respectively, and exclusively as the a glycoside (Table 1,
entries 8 and 9). Complete a-selectivity was also observed
with the formation of 6j when the less bulky 3,4-O-
tetramethyldisiloxane 5j[18] was used as the glycoside donor,
thus suggesting that the high a-selectivity of the reaction
should be attributed to conformational constraints induced by
the cyclic nature of the protecting group rather than to the
steric bulk of the isopropyl substituents.
To determine the scope of the methodology, the glucal
derivative 5i was used as the glycosyl donor and reacted with
a range of differentially protected glycoside acceptors 3b–h
(Table 2). In all cases, the reactions proceeded in excellent
yields and with high a-selectivity. For instance, the thiogluco-
2
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2014, 53, 1 – 6
These are not the final page numbers!