Organic Letters
Letter
transformations.23 The basicity of fluoride is highly variable,
depending on its counterion, the nature of the solvent, and the
amount of water present.24 TBAF often has greater reactivity
than alkali metal fluorides due to the weakly coordinating
nature of its tetrabutylammonium counterion and its high
organic solubility. However, though further drying is possible,
TBAF is only commercially available in its hydrated form.25
Fluoride forms strong hydrogen bonds, so this water content
greatly attenuates the reagent’s nucleophilicity and basicity.
Attempted preparation of 6-fluoro-6-deoxy-GlcNAc by
treating tosylate 1 with undried commercial TBAF in DMF/
THF (Scheme 1) led to an excellent yield (87%) of the
Table 1. Screen of Reaction Conditions To Promote 3,6-
Anhydro Cyclization of Tosylate 1
a
base or additive
solvent
temp (°C) time (h)
yield (%)
e
b
b c
,
TBAF
DMF/THF
THF
THF
55
55
55
55
55
55
55
50
50
80
50
50
50
65
16
16
16
16
16
16
16
24
22
24
24
24
24
3
83 (87)
e
b
TBAF
69
e
f
d
TBAF , H2O
trace
e
g
TBAF , AcOH
THF
trace
trace
0
e
CsF
DMF/THF
DMF/THF
DMF/THF
DMF
Pyr
Pyr
Pyr/THF
Pyr/THF
Pyr/THF
MeOH
e
f
CsF , H2O
e
Et3N
Pyr
0
0
e
trace
0
Scheme 1. Unexpected 3,6-Anhydro Cyclization of C6-Tosyl
GlcNAc 1 Promoted by TBAF
e
TBABr
trace
trace
e
CsF
e
b
TBAF
65
0
h
NaOMe
a
All reactions were performed on 0.1 mmol scale unless otherwise
b
c
d
stated. Isolated yield. Reaction at 1 mmol scale. Yield <10% by
TLC. Using 7.5 equiv. Using 20 equiv. Using 3.75 equiv. At 0.3 M.
e
f
g
h
Furthermore, substituting CsF for TBAF gave poor conversion.
We noted that the CsF fully dissolved upon heating, so
reduced solubility was unlikely the source of the reduced
conversion. We hypothesized that the water content of
commercial TBAF may be beneficial for the transformation
but, to the contrary, adding water (20 equiv) to the CsF
reaction reduced the yield to zero. Other mild bases including
triethylamine and pyridine also failed to give appreciable
product formation. Similarly, employing TBABr or CsF as
additives in a pyridine/THF solvent system failed to enhance
the yield of product, though TBAF gave a dramatic
improvement. Taken together, these results suggest that both
components of the fluoride/tetrabutylammonium ion pair are
required for efficient 3,6 anhydro sugar formation.
unexpected 3,6-anhydrofuranoside 2. The structure of 2 was
verified by acetylation which resulted in a large downfield shift
in the H-5 absorbance in the 1H NMR spectrum of product 3,
consistent with acetylation of the C-5 hydroxyl and a furanose
ring configuration. An X-ray crystal structure of 3 was acquired
to unambiguously confirm its identity (Figure 2). Surprised by
the simplicity, high yield, and mild conditions for 3,6 anhydro
sugar formation in comparison with reported methods,9−22 the
scope of the reaction was explored.
Previous reports of 3,6-anhydro GlcNAc formation treated
methyl 2-acetamido-2-deoxy-6-O-tosyl-α-D-glucopyranoside
with sodium methoxide in refluxing methanol.3 For compar-
ison, we subjected tosylate 1 to these conditions. Substantial
decomposition was observed and the desired 3,6-anhydro
product could not be isolated, supporting the mild TBAF
promoted condition as a useful alternative to the reported
synthetic strategy.
We next explored the scope of this transformation to
determine if tosylate 1 was unusually susceptible to 3,6-
anhydro cyclization or if TBAF was a general reagent for
formation of these fused ring systems. We found that the
TBAF-based conditions gave a range of the 3,6-anhydrohexo-
ses in high yield, including glucose, N-acetyl-glucosamine, 2-
deoxyglucose, mannose, and galactose configurations (Table
2). Both α- and β-O-glycoside substrates consistently gave
pyranose products with retention of anomeric configuration.
Intriguingly, N,O-dimethyl hydroxylamine glycoside substrates
(1 and 5) rearranged to the 3,6-anhydrofuranose sugars. It is
known that the methyl glycosides of 3,6-anhydroglucose, 2-
deoxyglucose, and mannose undergo a rapid pyranose to
furanose rearrangement upon treatment with strong acids.9−11
It is likely that the lower energy barrier for ring opening N,O-
dimethyl hydroxylamine N-glycosides relative to O-glycosides
allows the facile rearrangement of these substrates. Consistent
with low barrier ring opening, we observed that N,O-dimethyl
Figure 2. X-ray crystal structure of α-anomer 3 (CCDC deposition
To understand the requirements for high yielding intra-
molecular ether formation, we examined the reaction
conditions (Table 1). Using DMF as a cosolvent was found
to be beneficial, giving a slight improvement over THF alone.
Attenuating the reactivity of TBAF by buffering it with acetic
acid or adding additional water (20 equiv) gave inferior yields.
B
Org. Lett. XXXX, XXX, XXX−XXX