Studies on the origin of stereoselectivity in the synthesis of 1,2-trans glycofuranosyl azides

Article abstract of DOI:10.1016/S0008-6215(99)00293-1

The stereoselectivity of the 1,2-trans directed, Lewis acid-catalysed azidation of peracylated furanoses was found to depend on the reactivity of the azide donor (azide nucleophilicity) and the configuration at the anomeric centre relative to the neighbouring 2-O-acyl group. Reactions of 1,2-trans glycosyl esters with highly nucleophilic azide donors, generated from SnCl₄ and Me₃SiN₃, were stereospecific. The results are interpreted in terms of the rapid reaction of the azide species with bicyclic 1,2-acyloxonium (1,2-O-alkyliumdiyl-D-glycofuranose) ions, which were the primarily formed reactive intermediates. When using 1,2-cis glycosyl esters as starting materials the selectivity was reduced (90-94% de); the same is true with 1,2-trans counterparts if less nucleophilic Me₃SiN₃ in combination with Me₃SiOTf catalyst was used. This occurred due to the appearance of the more reactive but less selective oxocarbenium (glycofuranoxonium) ions either as primarily formed reactive intermediates in the former case or after equilibration with acyloxonium ions in the latter case. Protected 1,2-trans β-D-glycofuranosyl azides with ribo, xylo and 3-deoxy-erythro-pento configurations were best prepared from the corresponding glycosyl esters using 0.05 equivalents of SnCl₄, i.e., under anomerization-free conditions. Azidation of methyl glycofuranosides proceeds with inferior (80-90% de) and less predictable selectivity irrespective of the starting anomeric configuration. Copyright (C) 2000 Elsevier Science Ltd.

Full text of DOI:10.1016/S0008-6215(99)00293-1

www.elsevier.nl/locate/carres
Carbohydrate Research 324 (2000) 149–160
Studies on the origin of stereoselectivity in the synthesis
of 1,2-trans glycofuranosyl azides
1
&
Anton Stimac , Jozˇe Kobe *
National Institute of Chemistry, Hajdriho6a 19, 1115 Ljubljana, Slo6enia
Received 5 August 1999; accepted 24 October 1999
Abstract
The stereoselectivity of the 1,2-trans directed, Lewis acid-catalysed azidation of peracylated furanoses was found
to depend on the reactivity of the azide donor (azide nucleophilicity) and the configuration at the anomeric centre
relative to the neighbouring 2-O-acyl group. Reactions of 1,2-trans glycosyl esters with highly nucleophilic azide
donors, generated from SnCl4 and Me3SiN3, were stereospecific. The results are interpreted in terms of the rapid
reaction of the azide species with bicyclic 1,2-acyloxonium (1,2-O-alkyliumdiyl-D-glycofuranose) ions, which were the
primarily formed reactive intermediates. When using 1,2-cis glycosyl esters as starting materials the selectivity was
reduced (90–94% de); the same is true with 1,2-trans counterparts if less nucleophilic Me3SiN3 in combination with
Me3SiOTf catalyst was used. This occurred due to the appearance of the more reactive but less selective
oxocarbenium (glycofuranoxonium) ions either as primarily formed reactive intermediates in the former case or after
equilibration with acyloxonium ions in the latter case. Protected 1,2-trans b-D-glycofuranosyl azides with ribo, xylo
and 3-deoxy-erythro-pento configurations were best prepared from the corresponding glycosyl esters using 0.05
equivalents of SnCl4, i.e., under anomerization-free conditions. Azidation of methyl glycofuranosides proceeds with
inferior (80–90% de) and less predictable selectivity irrespective of the starting anomeric configuration. © 2000
Elsevier Science Ltd. All rights reserved.
Keywords: 1,2-trans Glycofuranosyl azides, synthesis; Azidation, stereoselective; Anomerization; Oxonium ions; Stannic chloride
catalysis; Trimethylsilyl triflate catalysis
tion at C-1 and C-2 of the sugar. As a part of
1. Introduction
our programme towards a linear synthesis of
1,2,3-triazolo[4,5-c]pyridine nucleoside ana-
logues [2] via the 1,2,3-triazole nucleosides,
Glycosyl azides [1] serve as valuable carbo-
hydrate building blocks, especially as precur-
sors to glycosylamines and heterocyclic
derivatives such as 1,2,3-triazoles. Accord-
ingly, several methods for their synthesis were
developed, which proved to be quite general
with respect to the desired relative configura-
several b- -pentofuranosyl azides were re-
D
quired, of which only those with the 1,2-trans
configuration will be considered in this article.
The Lewis acid-catalysed reaction of
trimethylsilyl azide (Me3SiN3) with glycosyl
esters possessing a participating 2-O-acyl
group proved to be the most efficient and
reliable method for the preparation of 1,2-
trans glycosyl azides [1]. In fact, the azidation
of pyranosyl esters [1,3,4] was consistently
reported to be stereospecific regardless of the
* Corresponding author. Tel.: +386-61-1760-200; fax: +
386-61-1259-244.
E-mail address: joze.kobe@ki.si (J. Kobe)
1
&
Present address: Krka, d. d., Novo mesto; Smarjesˇka c. 6,
8501 Novo mesto, Slovenia.
0008-6215/00/$ - see front matter © 2000 Elsevier Science Ltd. All rights reserved.
PII: S0008-6215(99)00293-1

&
A. Stimac, J. Kobe / Carbohydrate Research 324 (2000) 149–160
150
anomeric configuration of the starting ester,
while the selectivity of aldofuranosyl counter-
parts was reduced in several cases [5,6]. The
efficiency of this azidation approach in terms
of the chemical yield and selectivity makes it
undoubtedly the method of choice; however, a
rather wide selection of catalysts and their
amounts used (usually nearly equimolar) were
not so clearly evident. Recently, Matsubara
and Mukaiyama [4] reported for the first time
a truly catalyst-aided synthesis of pyranosyl
azides by taking advantage of SnCl4–AgClO4
systems in dichloromethane or ytterbium(III)
triflate in nitromethane (0.05–0.2 equivalents).
In addition, we [6] and others [5,7] have
pointed out that the azidation of furanosyl
esters could be carried out conveniently using
only catalytic amounts of selected Lewis acids
like SnCl4 or Me3SiOTf. Thus, of several cata-
lysts tested in substoichiometric amounts, the
azidation of 1-O-acetyl-2,3,5-tri-O-benzoyl-b-
D-ribofuranose (1b) with Me3SiN3 in
dichloromethane at ambient temperature was
completed only with SnCl4 (0.05 equivalents,
12 h) and Me3SiOTf (0.1 equivalents, 9 days)
[6]. These experiments revealed another rather
unexpected observation that, contrary to other
catalysts which gave the 1,2-trans azide 2b
exclusively, the Me3SiOTf reaction is much
less selective (b:a=17:3). In continuation of
the previous work [6], a more detailed mecha-
nism of azidation in non-nucleophilic solvents,
like dichloromethane, is proposed herein,
which provides a reasonable explanation for
the exclusive formation of the b-azide 2b in
the SnCl4-catalysed reaction as well as the
reduced selectivity in the case of the Me3SiOTf
reaction. Showing advantages over other cata-
lysts, SnCl4 was employed for the highly
stereoselective synthesis of 1,2-trans azides
(Table 1). It is evident from Table 1 that the
former is a considerably more efficient catalyst
for any comparable set of reaction conditions.
In view of the superior 1,2-trans selectivity
observed in SnCl4-catalysed reactions [1,3,4,6]
(formation of 1,2-cis azides from peracylated
aldoses has never been reported), it was rather
surprising to find that the initially formed
b-azide 2b underwent slow anomerization in
the presence of this Lewis acid, if the reaction
was continued after consumption of 1b. This
effect was particularly noticeable at higher
catalyst loadings (entries 2 and 3), but was
completely absent with our optimized reaction
conditions (0.05 equivalents, entry 1).
Anomerization of the initially formed b-azide
2b during the reaction may also give both
anomers in the Me3SiOTf-catalysed azidation.
However, this hypothesis was disproved after
several control anomerizations of 2b were car-
ried out by means of Me3SiOTf and SnCl4 in
with b-
3-deoxy-b-
-galactopyranosyl configurations.
D
-ribofuranosyl-, b-D-xylofuranosyl-,
D
-erythro-pentofuranosyl-, and b-
D
2. Results and discussion
Stereochemistry of the reaction of 1b with
Me3SiN3. Mechanistic considerations.—A set
of azidation experiments of 1b with Me3SiN3
in the presence of SnCl4 or Me3SiOTf was
carried out under various reaction conditions

&
A. Stimac, J. Kobe / Carbohydrate Research 324 (2000) 149–160
151
Table 1
Catalysed azidation of 1b with Me3SiN3 in CH2Cl2 at 23 °C
Entry Catalyst
(equiv.)
Time
Yield of 2b+2a 2b:2a
(%) a
1 b
2
SnCl4 (0.05)
SnCl4 (1)
SnCl4 (1)
Me3SiOTf (0.1) 9 days quant
Me3SiOTf (0.5) 6 days quant
Me3SiOTf (1)
Me3SiOTf (1)
Me3SiOTf (1)
12 h
10 min 98
7 days 90.5
quant
100:0
93:7
3
64:36
85:15
86:14
86:14
85:15
80:20
4 b
5
6
33 h
98
7
21 days 96
8 c
128 h d 95.5 e
a Crude product.
b Data from Ref. [6].
c Reaction carried out in
Me3SiOAc.
Scheme 1. Reagents and conditions: (a) Me3SiN3 (1.1 equiv),
Me3SiOTf (0.5 equiv), CH2Cl2, 3 h; then HCO3−, H2O; (b)
silica gel chromatography.
a
1:1 mixture of CH2Cl2–
d Reaction proceeded to ꢀ95% conversion.
e Contaminated with ꢀ5% of acetates 1b and 1a (3:2).
the presence of the unreacted starting material
1b and the product 2b, the Me3SiOTf reaction
appeared more complex (TLC data). When
the latter reaction in the presence of 0.5 equiv-
alents of Me3SiOTf was quenched at the initial
stage (after 3 h) and worked up, the crude
reaction mixture contained six components
the presence or absence of trimethylsilyl ac-
etate (Me3SiOAc), a side product of the glyco-
sylation of 1b with silylated nucleophiles [8].
The anomerization of 2a [6] as well as 2b (see
Table 2 and entry 3 of Table 1) indicates the
equilibrium composition of 64:3692 in fa-
vour of 2b at 23 °C in dichloromethane. This
value is quite different from that (b:a=
1
(from H NMR data) in fractions given in
Scheme 1.
In Scheme 1, the 2b:2a ratio of 84:16 was
estimated to be almost identical (within an
experimental error) to the ratio observed after
complete consumption of the starting material
(86:14; entry 5, Table 1). Virtually constant
ratios throughout the progress of azidation
suggests the existence of common intermedi-
ates for the unselective formation of both
azides with the anomerization of 1b taking
place as a side reaction. By analogy to the
anomerization of peracylated aldoses, e.g., b-
glucose pentaacetate [9], these data clearly
point to the intermediacy of the oxocarbenium
86:1491)
consistently
found
in the
Me3SiOTf-catalysed azidation of 1b, as well as
the anomerization of 2b, in the presence of
Me3SiOTf and Me3SiOAc (1 equivalent each).
Under conditions where the azidation of 1b is
complete within 1–1.5 days (Table 1, entry 6),
only ꢀ2% of the pure 2b would anomerize
within this time interval. Moreover, unlike the
SnCl4-catalysed reaction under optimized con-
ditions (entry 1, Table 1), which revealed only
Table 2
Anomerization of 2b in CH2Cl2 at 23 °C
ion B+ (2,3,5-tri-O-benzoyl-
-ribofuranoxo-
D
nium cation) formed from the more stable
Entry Catalyst
(equiv)
Additive
(equiv)
Time (days) 2b:2a
1,2-benzoxonium ion A+ (3,5-di-O-benzoyl-
1,2-O-phenylmethyliumdiyl-a- -ribofuranose
D
1
2
3
4
5
6
7
8
9
SnCl4 (0.5)
SnCl4 (0.5)
Me3SiOTf (1) none
Me3SiOTf (1) none
Me3SiOTf (1) none
Me3SiOTf (1) Me3SiOAc (1) 5.5
Me3SiOTf (1) Me3SiOAc (1) 11.3
Me3SiOTf (1) Me3SiOAc (1) 21
Me3SiOTf (1) Me3SiOAc (1) 40
none
Me3SiOAc (1)
8
8
1.5
5.5
11.3
64:36
62:38
67:33
61:39
62:38
92:8
cation) [8,10].
After chromatography on silica gel, the
third anomeric pair of products, 2,3,5-tri-O-
benzoyl-
as an inseparable mixture accompanied by a
-ribo-
D
-ribofuranose (3ab) [11] was isolated
90:10
86:14
87:13
small amount of 1,3,5-tri-O-benzoyl-a-
D
furanose (4) [12] in 11% combined yield. Com-
pound 4 is an artefact, which originated from

&
A. Stimac, J. Kobe / Carbohydrate Research 324 (2000) 149–160
152
furnish 3a, a product with retained stereo-
chemistry of the benzoxonium ion A+. Al-
though only 3a would be expected to appear
in the crude product mixture, the anomeric
hydroxy group is inevitably subjected to the
anomerization process during mildly alkaline
workup manipulations, resulting in the mix-
ture of anomers and, eventually, the rear-
ranged product 4 (Scheme 2). Since the
oxocarbenium ion B+ is apparently much
more reactive (less stable) and, consequently,
would be present in a negligible concentration,
it is of interest that the concentration of A+
amounted to ꢀ16% of the total product com-
position at the moment of trapping with
water.
Based on the fact that Me3SiN3 is a rela-
tively weak nucleophile and reacts with the
ion A+ very slowly or not at all, the mecha-
nism of the Me3SiOTf-catalysed azidation of
1b can now be rationalized, as shown in
Scheme 3.
The stereochemical course of azidation will
depend on the equilibrium composition of A+
and B+, their relative reactivities towards
Me3SiN3 and Me3SiOAc, and the concentra-
tion of both nucleophiles. The presence of
Me3SiOAc obviously triggers additional equi-
libria that are responsible for the observed
ratio of azides (b:a=43:7), which is substan-
tially shifted from the intrinsic equilibrium
value (b:a=16:9).
3a through 2“1 migration of the benzoyl
group during chromatographic separation.
The structures of these compounds were deter-
1
mined by comparison of H and 13C NMR
data with those reported [11,12b,c], as well as
by transformation of a mixture of 3b, 3a, and
4 into a mixture of known derivatives 1b, 1a
[11a,13], and 2-O-acetyl-1,3,5-tri-O-benzoyl-a-
D-ribofuranose [12b] by acetylation. The iden-
tification of compounds 3b and 3a provides
indirect evidence for the existence of a rela-
tively stable, long-lived 1,2-benzoxonium ion
1
A+ (for an indirect H NMR evidence of this
cation, see Ref. [14]). Cyclic acyloxonium (1,3-
dioxolan-2-ylium) ions [15] like A+ are am-
bident electrophiles [16], which can react with
nucleophiles to give products of either kinetic
or thermodynamic control. Unlike the azide
nucleophile, which reacts at the anomeric cen-
tre (thermodynamic product), water [15,16]
always reacts in a kinetic sense (at C-2 of the
1,3-dioxolan-2-ylium ring). In the case of the
A+ ion a cyclic hemiorthoester [12a,17] is
formed, which decomposes spontaneously to
Scheme 2.
Stereochemistry of the oxocarbenium triflate
ion pair B+OTf−.—The outcome of the
stereochemistry attributed to B+OTf− should
reside in the neighbouring group effect, which
favours the b-triflate ion pair b-B+OTf−.
This could be formed by the approach of the
triflate anion to the anomeric position of A+
from the b-face, while the CꢀO bond of the
1,3-dioxolan-2-ylium ring is being cleaved.
Since its lifetime is very short, the ion pair
b-B+OTf− reverts either to the bicyclic ion
A+OTf− or reacts with an external nucleo-
phile almost instantaneously with predomi-
nant a-selectivity, before efficient equilibration
between b- and a-triflate ion pairs could take
place (Scheme 4).
Scheme 3.
The above explanation may find support in
the allylation of our model substrate 1b with
allyltrimethylsilane [18], which was reported
Scheme 4.

&
A. Stimac, J. Kobe / Carbohydrate Research 324 (2000) 149–160
153
to be markedly a-selective under any condi-
tions. This fact may reflect the inability of the
cyclic ion A+ to react with weak carbon nu-
cleophiles; that is why the oxocarbenium ion
B+ could be responsible for the final product
distribution. A general trend of a-selectivity
deriving from oxocarbenium triflate ion pairs
indicates that the bulk of 2b should arise from
the attack of Me3SiN3 on the benzoxonium
ion A+.
Mechanistic alternati6es.—The formation
of 2a in the Me3SiOTf-catalysed azidation
could take place by at least three additional
mechanistic pathways; however, no firm evi-
dence exists in favour of any pathway.
Direct SN2 substitution at C-1. Direct dis-
placement of Me3SiOAc on the complex of the
substrate and Me3SiOTf by Me3SiN3, assum-
ing the absence of the anchimeric assistance
by the 2-O-benzoate group, hardly seems to
be a noticeable alternative to the more fa-
vourable anchimerically assisted intramolecu-
lar displacement. An efficient SN2 substitution
requires a powerful nucleophile, which is not
the case with Me3SiN3. Besides, as shown
below, stronger azide donors than Me3SiN3
drive the reaction towards the increased b-
selectivity.
2-Azido-1,3-dioxolane intermediates. To the
best of our knowledge, the isolation or detec-
tion of 2-azido-1,3-dioxolane intermediates
has never been reported in the synthesis of
glycosyl azides. The hypothetical kinetic
diastereoisomeric intermediate C could trans-
form thermally or in the presence of Lewis
acids, via the ion A+, into the thermodynamic
products 2b and 2a. However, its appearance
does not seem to be very likely because the
relatively high stability of some non-carbohy-
drate derived 2-azido-1,3-dioxolanes towards
protic acids [19] and temperature [20] suggests
comparable stability of C.
either an acyclic oxocarbenium ion or a
monocyclic benzoxonium ion. The azidation
of both with Me3SiN3 and subsequent Lewis
acid-promoted recyclization would occur with
a low selectivity and co-formation of acyclic
by-products [21]. It is of interest to point out
that no acyclic intermediates or by-products
were identified in any of the azidation reac-
tions of 1b. With the hope of trapping the
acyclic 4-O-silylated derivative, being more
resistant towards subsequent recyclization,
Me3SiOTf was replaced by tert-butyl-
dimethylsilyl triflate (0.5 equivalents) in a par-
ticular experiment. This reaction, however,
proceeded analogously and without major dif-
ferences from that given in Table 1 (entry 5).
For this reason, the endocyclic cleavage in the
catalysed azidation of glycosyl esters is not to
be expected.
Effect of the azide nucleophilicity.—In con-
trast to the poorly selective Me3SiOTf-
catalysed reaction mentioned above, the
SnCl4-catalysed azidation of 1b is stereospe-
cific and, therefore, must proceed exclusively
through the benzoxonium ion A+; Me3SiN3
acts as a nucleophile in both cases. The reac-
tion of Me3SiN3 with SnCl4 in dichloro-
methane at ambient temperature is known to
first result in a soluble dimer [Me3SiCl·
Cl2Sn(N3)2]2, which was then slowly trans-
formed into a final insoluble polymeric
product [Cl2Sn(N3)2]n [22]. The soluble dimer,
some higher oligomer, and/or their complex
with a leaving group comprise probable azide
donors of increased nucleophilicity in the
SnCl4-catalysed reaction of 1b with Me3SiN3.
The benzoxonium ion A+ is intercepted by
the highly reactive azide donor immediately
after its formation, before the isomerization to
the oxocarbenium ion B+ could take place.
Since the cation A+ has only a momentary
existence, it could never be detected in the
form of its hydrolysis products 3b and 3a by
TLC analysis of the azidation progress.
The observation that the highly nucleophilic
tin–azido species are responsible for the spe-
cificity observed has been further confirmed
by a set of experiments in which trimethyltin
azide (Me3SnN3) and tributyltin azide
(Bu3SnN3), with precisely defined structures,
were employed as azidation agents (Table 3).
Ring-opening mechanism. The competitive
complexation of Me3SiOTf with the furanose
ring oxygen results in the endocyclic cleavage
of the C(1)ꢀO(4) bond and the formation of

&
A. Stimac, J. Kobe / Carbohydrate Research 324 (2000) 149–160
154
Table 3
Azidation of 1b with trialkyltin azides
Entry
Catalyst (equiv)
Azide
Solvent
Time (days)
2b:2a
1
2
3
Me3SiOTf (1)
Me3SiOTf (1)
none
Me3SnN3
Bu3SnN3
Bu3SnN3
CH2Cl2
CH2Cl2
neat
5.5
8
8
96:4
97:3
100:0
The stereoselectivity of the Me3SiOTf-
catalysed azidation can be greatly improved
simply by substituting more nucleophilic tri-
alkyltin azides [23] for Me3SiN3. Bu3SnN3 as
solvent (70 °C, 8 days, ꢀ95% conversion) can
function as an azide donor and a weak Lewis
acid. Compound 2b was isolated in about 70%
yield along with a few unidentified side prod-
ucts. In this case, complete exclusion of either
comparable rates using much lower quantities
of SnCl4 than previously reported [1,3].
The azidation of methyl glycofuranosides
5b and 14b resulted in the predominant for-
mation of the b-azides 2b and 15b with 90 and
80% de, respectively, being much inferior to
those starting from the appropriate acetates
1b and 12b. The origin of minor quantities of
1,2-cis products 2a and 15a is not completely
clear and points to the involvement of alterna-
tive mechanisms, e.g., the furanose ring
opening.
Conclusions.—Furanose sugars turned out
to be more useful probes for studying stereo-
chemical effects of the azidation than their
pyranose counterparts. The Lewis acid-
catalysed reactions of pyranosyl esters with
Me3SiN3 were reported [1,3,4] to be stereospe-
cific without exception, while at least two
properties affect the stereoselectivity in the
furanosyl series, namely, the azide nucleo-
philicity and the nature of the substrate struc-
ture. When both effects are favourable, like in
the case of 1,2-trans glycosyl esters and a
Lewis acid capable of generating more nucle-
ophilic azidating species after reaction with
Me3SiN3 (e.g., SnCl4), the azidation was
stereospecific in all cases examined. This oc-
curs because the 1,2-acyloxonium ion is pri-
marily formed, which is rapidly captured by
the nucleophile before isomerization to less
selective reactive intermediates takes place.
Glycosyl esters are also preferred starting ma-
terials over glycosides in the presence of the
SnCl4 catalyst.
1
1a or 2a was determined by H NMR and
TLC analyses.
Stereoselecti6ity of the SnCl4-catalysed azi-
dation. Substrate structure effect.—Superior
selectivity and efficiency of the azidation of 1b
in comparison with other catalysts tested [6]
suggested SnCl4 (0.05 equivalents) for the 1,2-
trans directed azidation of some other glycosyl
substrates with Me3SiN3 (Scheme 5). In all
cases examined (Table 4), reactions proceeded
with 80–100% de selectivities in favour of
1,2-trans azides depending on the nature and
configuration of the anomeric substituent rela-
tive to the 2-O-acyl group. The available 1,2-
cis esters 1a and 8a displayed slightly reduced
selectivities in comparison with their 1,2-trans
counterparts 1b and 8b, because the corre-
sponding oxocarbenium ions are formed as
the first reactive intermediates after dissocia-
tion of the glycosidic bond. The highly nucleo-
philic azidation reagent intercepts the oxocar-
benium ions, to a small extent, before they
completely isomerize to the more stable acy-
loxonium ions. Trace amounts of a-azides 9a
and 13a were detected in case of the azidation
of 8b and 12b, presumably due to the contam-
ination of starting materials with 1,2-cis esters
8a (5%) and 12a (ꢀ10%), respectively. For
the same reason, a small amount of the a-
azide 11a could also be expected to form from
a,b- -xylofuranose tetraacetate (10ab, 2:1 a:b
D
mixture). The last example in Table 4 (entry
10) clearly demonstrates that the azidation of
furanosyl and pyranosyl esters proceeds at
Scheme 5.

&
A. Stimac, J. Kobe / Carbohydrate Research 324 (2000) 149–160
155
Table 4
SnCl4-catalysed azidation of various sugar derivatives with Me3SiN3 in CH2Cl2 at 23 °C a
Entry
1 e
Substrate
Time (h)
Product(s)
Yield (%) b
b:a c
[h]D (c) d (°)
−45 (3.11)
1b
1a
5b
6b
8a
12
30
30
12
4
2b
95
100
91
97
97
2.4
94
88
74
86
7
100:0
95:5
95:5
1:0 h
97:3
2
3
2b+2a f
2b+2a g
7b
4 e
5
−149 (1.0)
−73 (4.00)
9b
9a
9b
11b
13b
15b
15a
17
+170 (0.99)
−72 (4.00) m
−129 (2.20) n
−171 (1.11) m
−150 (4.00)
+85 (4.00)
6
7
8
9
8ab i
10ab j
12ab k
14b
2
24
12
12
\200:1
nd l
124:1
90:10
10
16
12
70
100:0
−14.6 (2.885) o
a Molar ratio of substrate:Me3SiN3:SnCl4=1:1.1:0.05, except for entries 3 and 9, where 1.5 mol equiv of Me3SiN3 was used.
b Yields after chromatography and/or crystallization, crude yields are given for entries 2 and 3.
1
c Anomeric ratio determined from the H NMR spectrum of the crude product mixture.
d Measured for solutions in CHCl3 at 25 °C, unless stated otherwise.
e Data from Ref. [6].
f Contaminated with 1a (8%), 1b hardly detectable.
g Contaminated with a mixture of 5b and 5a (11%) in a 7:4 ratio.
h The corresponding a anomer was not identified.
i b:a=19:1.
j a:b=2:1.
k b:a=9:1.
l Not determined, the corresponding a anomer was not identified.
m Measured at 26 °C.
n Measured in ethyl acetate at 24 °C.
o Measured at 20 °C; lit. [h]2D0 −16.2° (c 2.885) [33].
3. Experimental
[h]2D5 +167° (c 1.10, CHCl3), lit. [h]2D0 +170°
(c 1.1) [25]; H NMR [12b,26]) revealed that
1
General methods are those reported in Ref.
[6]. NMR spectra were recorded for solutions
in CDCl3. Optical rotations were measured
with a Perkin–Elmer 241 MC polarimeter. All
reactions were carried out in a dry inert
atmosphere.
both materials were configurationally homo-
geneous. Tetra-O-acetyl-a,b- -xylofuranose
D
(10ab) was obtained from the same supplier
and was bulb-to-bulb distilled before use at a
bath temperature of 175 °C and ꢀ0.01 mbar.
The distillate contained mainly 10a and 10b in
a 2:1 ratio. Methyl 3-deoxy-2,5-di-O-benzoyl-
Materials.—Dichloromethane (Fluka), kept
over molecular sieves, was used as solvent.
b- -erythro-pentofuranoside (14b: mp 82–
D
1-O-Acetyl-2,3,5-tri-O-benzoyl-b-
D
-ribofura-
-xylofura-
83 °C (hexane), lit. mp 80.5–81.5 °C [27b])
was conveniently prepared in five steps from
1,2-di-O-isopropylidene-5-O-methoxycarb-
nose (1b) and tetra-O-benzoyl-a-
D
nose (8a) were supplied by Pfanstiehl under
names that did not specify the anomeric
configuration. Analysis of commercial samples
(1b: ivory coloured powder with mp 127–
128.5 °C; [h]2D6 +41.8° (c 1.32, CHCl3);
colourless needles after crystallization from
isopropanol: mp 129–130 °C, lit. mp 131–
132 °C [24]; [h]2D6 +42.3° (c 1.32), lit. [h]2D5
onyl-3-O-tosyl-a-
following the reported procedures [27].
Methyl 2,3,5-tri-O-benzoyl-i- -ribofuran-
D-xylofuranose (Pfanstiehl),
D
oside (5b).—This compound was prepared
[28] from 7b [6] after final purification by
chromatography with CH2Cl2: [h]2D5 +58° (c
1
1
0.94, CHCl3), lit. [h]D +58° [29]; H NMR: l
+42.2° (c 1.32) [24]; H NMR [12b,13]. 8a:
3.42 (s, 3 H, OMe), 4.52 (dd, 1 H, H-5a),
mp 165–166.5 °C, lit. mp 164–165 °C [25];

&
A. Stimac, J. Kobe / Carbohydrate Research 324 (2000) 149–160
156
and two unknown contaminants in a 81:10:5:4
ratio. H NMR data for 12b: l 1.96 and 2.10
4.70–4.76 (m, 2 H, H-4, 5b), 5.16 (s, 1 H,
H-1), 5.68 (d, 1 H, H-2), 5.88 (dd, 1 H, H-3),
7.29–7.61, 7.87–7.91, and 8.0–8.10 (3 m, 15
H, ArꢀH), J2,3 4.9, J3,4 6.6, J4,5a 6.3, J5a,5b 12.8
Hz.
1
(2 s, each 3 H, 2 OCOCH3), 2.17–2.33 (m,
H-3a,3b), 4.34 (dd, 1 H, H-5a), 4.55 (dd, 1 H,
H-5b), 4.73 (dddd, 1 H, H-4), 5.24 (dd, 1 H,
H-2), 6.20 (s, 1 H, H-1), 7.42–7.49, 7.55–7.61,
and 8.04–8.10 (3 m, 5 H, ArꢀH), J2,3a 1.1, J2,3b
4.6, J3a,4 6.8, J3b,4 9.3, J4,5a 5.3, J4,5b 3.8, J5a,5b
Tetra-O-benzoyl-h,i-
—The crude anomeric mixture (b:a=7:10)
was prepared from 1,2-O-isopropylidene-a-
D-xylofuranose (8ab).
D
-
1
1
11.9 Hz; H NMR [30b]. H NMR data for
12a: l 6.44 (d, J1,2 4.2 Hz, H-1).
xylofuranose (Pfanstiehl) (30 g, 0.158 mol) by
the method of Holy [26] and subsequently
highly enriched in favour of the b anomer.
This was first flash chromatographed with
10:1 petroleum ether–EtOAc, followed by
CH2Cl2. A fraction, containing ꢀ50% of the
b anomer, was fractionally crystallized several
times from EtOAc, then several times from
acetone to remove the majority of 8a. The
concentrated mother liquor enriched with the
b isomer was freed from unknown by-prod-
ucts by chromatography with a 10:1“2:1 gra-
dient of petroleum ether–1% AcOH in
EtOAc. After concentration, the appropriate
fractions were crystallized twice, first from 5:1
MeOH–acetone, followed by MeOH to give
8b (1.20 g, 1.3%), contaminated with ꢀ5% of
the a anomer. The crystalline material con-
tained two kinds of crystals with a sharp
melting range, for the most part plates with
mp 111–113 °C along with some needles with
mp 151–153 °C, lit. mp 108–109 °C [25]; [h]D25
+18.2° (c 1.00, CHCl3); lit. [h]2D0 +11.2° (c
Azidation of 1b with Me3SiN3 (Table 1).—
To a solution of 1b (0.5–4 mmol scale) and
Me3SiN3 (1.1 equiv) in the appropriate anhy-
drous solvent, was syringed the required
amount of catalyst (neat Me3SiOTf or 1 M
solution of SnCl4 in CH2Cl2) adjusting the
final concentration to 3 mL of solvent per
mmol of 1b. After being stirred for the time
specified, the reaction mixture was quenched
with satd aq NaHCO3 and worked-up as de-
scribed previously [6] (see also the general
procedure below), except for entry 8, which
was first concentrated and rediluted with
CH2Cl2. The 2b:2a ratio was determined from
1
the H NMR spectrum of the crude mixture.
Anomerization of 2,3,5-tri-O-benzoyl-i- -
D
ribofuranosyl azide (2b) (Table 2).—To a solu-
tion of 2b (244 mg, 0.5 mmol) and, when
necessary, Me3SiOAc (75 mL, 0.5 mmol) in
CH2Cl2 was syringed either neat Me3SiOTf
(90 mL, 0.5 mmol) or 1 M solution of SnCl4 in
CH2Cl2 (0.25 mL, 0.25 mmol), adjusting the
final concentration to 3 mL of CH2Cl2 per
mmol of 2b. After being stirred for the time
specified, the reaction mixture (or 0.15 mL
aliquot) was quenched and worked-up as
usual. The 2b:2a ratio was determined from
1
1.0) [25]; H NMR: l 4.66–4.76 (m, 2 H,
H-5a,5b), 5.11 (pseudo q, 1 H, J 5.7 Hz, H-4),
5.82 (d, 1 H, H-2), 5.96 (dd, 1 H, H-3), 6.67 (s,
1 H, H-1), 7.33–7.66 and 7.95–8.12 (2 m, 20
H, ArꢀH), J2,3 1.0, J3,4 5.0 Hz; 1H NMR [12b];
13C NMR: l 62.84, 74.52, 80.03, 80.47, 99.49
(C-1 to C-5), 128.3–130.0 and 133.1–133.8
(ArꢀC), 164.73 (2 CO), 164.96 and 166.04 (2
CO). This material was used for the azidation
without further manipulations.
1
the H NMR spectrum of the crude mixture.
Product mixtures from the anomerizations in
the presence of Me3SiOAc contained in addi-
tion ꢀ2–5% of 1b and 1a.
Azidation of 1b with trialkyltin azides (Table
3)
With Me3SnN3 under Me3SiOTf catalysis
(entry 1). To a suspension of 1b (505 mg, 1
mmol) and Me3SnN3 (227 mg, 1.1 mmol) in
CH2Cl2 (3 mL) was syringed Me3SiOTf (180
mL, 1 mmol), whereby the azide dissolved
almost instantaneously. The solution was
stirred at room temperature (rt) for 5.5 days,
5-O-Benzoyl-3-deoxy-1,2-di-O-acetyl-h,i- -
D
erythro-pentofuranose (12ab).—This com-
pound was prepared [30] in five steps from
1,2-O-isopropylidene-a-
D
-xylofuranose
and
subsequently purified by chromatography
(50:1 CH2Cl2–Et2O): [h]2D5 −29.4° (c 1.25,
CHCl3), lit. [h]3D0 −27.2° [30b]. Although a
1
single spot by TLC analysis, the H NMR
spectrum revealed the presence of 12b, 12a,

&
A. Stimac, J. Kobe / Carbohydrate Research 324 (2000) 149–160
157
nose [1a: 179 mg, 9.5%; IR (film): w 1726
(CꢁO), 1268 (benzoate CꢀO), 1227 (acetate
CꢀO), 1134, 1110, 1024, and 708 cm−1 (ben-
then quenched and worked-up as usual. The
crude semisolid product mixture (589 mg) was
1
directly analysed by H NMR spectroscopy.
1
zoate); H NMR: l 2.15 (s, 3 H, COCH3),
Crystallization from MeOH afforded 2b (184
4.62 (dd, 1 H, H-5a), 4.73 (dd, 1 H, H-5b),
4.83 (ddd, 1 H, H-4), 5.64 (dd, 1 H, H-2), 5.84
(dd, 1 H, H-3), 6.72 (d, 1 H, H-1), 7.30–7.62,
7.86–7.91, and 8.04–8.12 (3 m, 15 H, 3 Ph),
J1,2 4.6, J2,3 6.4, J3,4 2.5, J4,5a 3.6, J4,5b 3.4,
mg, 38%) with mp 66.5–67.5 °C, lit. mp 66–
1
67.5 °C [6]; H and 13C NMR [6].
With Bu3SnN3 under Me3SiOTf catalysis
(entry 2). A similar reaction on the same scale
with Bu3SnN3 (365 mg, 1.1 mmol) took 8 days
at rt. The crude product mixture consisted of
two immiscible liquids and was further
purified by dissolving in MeCN (10 mL) and
washing with hexane (5×10 mL). Evapora-
tion of MeCN left a colourless viscous oil (314
13
J5a,5b 12.1 Hz; C NMR: l 21.02 (CH3), 63.89
(C-5), 70.67 (C-3), 71.06 (C-2), 82.19 (C-4),
94.27 (C-1), 128.4–129.7, 133.34, 133.48, and
133.53 (ArꢀC), 164.89, 165.56, and 166.02 (3
1
COPh), 169.52 (COCH3); H and 13C NMR
’
1
mg), which was analysed by H NMR spec-
[11a,13]; EIMS: m/z 461 ([M–COCH3]+ ,
troscopy. Crystallization from MeOH gave 2b
(231 mg, 47%) with mp 67.5–68 °C.
1.2%), 445 ([M–OCOCH3]+, 11), 105
([PhCO]+, 100)]. Last eluted was a mixture of
3b, 3a, and 4 in a 58:31:11 ratio (204 mg,
With Bu3SnN3 only (entry 3). A suspension
of 1b (101 mg, 0.2 mmol) and Bu3SnN3 (0.4
mL) was heated in an oil bath at 70 °C for 8
days. The reaction mixture was allowed to
cool, diluted with MeCN (2 mL) and washed
with hexane (5×2 mL). Evaporation of
MeCN left a brown viscous liquid (112 mg),
1
12%). Partial H NMR data for 3b: 3.87 (d,
OH), 5.64 (br d, J1,OH 3.5 Hz, H-1), 5.69 (dd,
J1,2 0.8 Hz, H-2), 5.91 (dd, J3,4 6.2, J2,3 5.0 Hz,
H-3); for 3a: 3.58 (d, OH), 5.49 (dd, J2,3 6.3
Hz, H-2), 5.79 (dd, J3,4 3.3 Hz, H-3), 5.82 (dd,
J1,OH 7.5, J1,2 4.3 Hz, H-1); for 4: 2.84 (d, J2,OH
10.5 Hz, OH), 5.60 (dd, J 6.5, 1.8 Hz, H-3),
1
which was analysed by H NMR spectroscopy
1
(ꢀ70% yield of 2b).
6.69 (d, J1,2 4.5 Hz, H-1); H and 13C NMR
Partial Me3SiOTf-catalysed azidation of 1b
with Me3SiN3.—To a solution of 1b (2.02 g, 4
mmol) and Me3SiN3 (0.58 mL, 4.4 mmol) in
CH2Cl2 (12 mL) was syringed Me3SiOTf (0.36
mL, 2 mmol). The mixture was stirred at rt for
3 h, then quenched with satd aq NaHCO3 (12
mL) and the aqueous phase was extracted
with CH2Cl2 (12 mL). The combined organic
solutions were dried (MgSO4) and concen-
trated to give a mixture of products in frac-
tions, given in Scheme 1, as a colourless oil
(2.02 g). A portion of this mixture (1.86 g) was
chromatographed on silica gel (125 g) with
CH2Cl2 (2 L) collecting 50–60 mL fractions,
followed by CH2Cl2–Et2O (19:1, 0.8 L and
10:1, 0.7 L) collecting 50 mL fractions. First
eluted was the pure compound 2b [717 mg,
40%; mp 67–69 °C (from MeOH)], followed
by mixtures of 2b and 2a (68:32; 274 mg,
15%), 2b and 2a (5:95; 93 mg, 5%), 1b and 1a
(96:4; 232 mg, 12.5%; crystallization from iso-
propanol gave pure 1b with mp 128.5–130 °C;
1H and 13C NMR spectra were in agreement
with those of the starting material), and pure
spectra are in agreement with those reported
[11,12b,c] for individual components. A fur-
ther confirmation of identity was achieved by
acetylation [31] of a portion (140 mg) of this
mixture by acetic anhydride (1 mL) in the
presence of 4-(N,N-dimethylamino)pyridine (2
mg) at rt for 1 day. Work-up as usual af-
forded a mixture of 1b, 1a, and 2-O-acetyl-
1,3,5-tri-O-benzoyl-a-D-ribofuranose
[12b]
(109 mg) in an 8:1:1 ratio as determined on
1
the basis of known H NMR spectra of indi-
vidual components and TLC analysis.
General procedure for the SnCl4-catalysed
azidation with Me3SiN3.—To a solution of
sugar derivative and Me3SiN3 (1.1 equiv for
glycosyl esters, 1.5 equiv for glycosides) in
CH2Cl2 (2 mL/mmol) was added 50 mM
SnCl4 in CH2Cl2 (1 mL/mmol). The mixture
was stirred at rt for the time specified in Table
4, the reaction was quenched with satd aq
NaHCO3
(1.5 mL/mmol) and the aq phase
extracted with an equal vol of CH2Cl2 (three
times for acetates 13b and 17). The combined
organic solutions were dried (MgSO4) and
concentrated to give a crude product or a
1-O-acetyl-2,3,5-tri-O-benzoyl-a- -ribofura-
D

&
A. Stimac, J. Kobe / Carbohydrate Research 324 (2000) 149–160
158
(NH3): m/z 505 ([M+NH4]+, 2.5%), 460
([M−N2]+, 4.5), 445 ([M−N3]+, 22), 105
([PhCO]+, 100). Anal. Calcd for C26H21N3O7
(487.47): C, 64.06; H, 4.34; N, 8.62. Found: C,
64.01; H, 4.50; N, 8.47.
mixture of anomers. Separation and purifica-
tion was achieved as described for individual
preparations.
Azidation of 1a.—The crude product mix-
ture, obtained from 1a (36 mg, 71 mmol) in
quantitative yield, contained compounds 2b,
From compound 8ab. Azidation of 8ab (567
mg, 1 mmol) afforded the crude product mix-
ture containing less than 0.5% of 9a, which
was purified by chromatography (CH2Cl2) to
give 9b (456 mg, 94%), identical to the com-
pound obtained from pure 8a above.
1
2a and 1a in a 88:4:8 ratio as shown by H
NMR analysis.
Azidation of 5b.—The crude product mix-
ture (444 mg), obtained from 5b (477 mg, 1
mmol), contained compounds 2b, 2a, 5b, and
methyl 2,3,5-tri-O-benzoyl-a-
D
-ribofuranoside
2,3,5-Tri-O-acetyl-i- -xylofuranosyl azide
D
1
(5a) [32] in a 85:4:7:4 ratio as shown by H
(11b).—From the distilled compound 10ab
(31.83 g, 0.1 mol), the reaction mixture was
quenched by satd aq NaHCO3 (120 mL) and
the phases were allowed to separate overnight.
The clear organic and the emulgated layers
were treated with Celite (40 g) and, after 30
min, MgSO4 (20 g). Stirring was continued for
1 h, the slurry was filtered through Celite and
washed with CH2Cl2 (3×200 mL). The com-
bined filtrate and washings were dried
(MgSO4) and concentrated to give the crude
product (29.9 g) as a yellow liquid. This was
purified by chromatography (50:1 CH2Cl2–
EtOAc) to yield pure 11b (26.56 g, 88%) as a
slightly yellow liquid: IR (film): w 2117 (N3),
1748 (CꢁO), 1373 (acetate CH3), and 1226
NMR analysis.
2,3,5-Tri-O-benzoyl-b- (9b) and h-
D-xylo-
furanosyl azide (9a)
From compound 8a. Azidation of 8a (5.66 g,
10 mmol) afforded the crude product mixture
(b:a=97:3), which was resolved by chro-
matography (CH2Cl2, H2O content50.01%).
Both products were isolated as colourless vis-
cous oils.
b-Azide 9b: 4.73 g (97%); Rf 0.59 (CH2Cl2);
IR (film): w 2115 (N3), 1727 (CꢁO), 1262 (ben-
zoate CꢀO), 1098, and 710 cm−1 (benzoate);
1H NMR: l 4.71 (d, 2 H, H-5a,5b), 4.98 (dt, 1
H, H-4), 5.46 (br s, 1 H, H-2), 5.55 (s, 1 H,
H-1), 5.88 (dd, 1 H, H-3), 7.37–7.63 and
8.00–8.10 (2 m, 15 H, ArꢀH), J2,3 1.2, J3,4 4.9,
1
cm−1 (acetate CꢀO); H NMR: l 2.10, 2.12,
J
4,5a =J4,5b 6.1 Hz; 13C NMR: l 62.42 (C-5),
and 2.14 (3 s, each 3 H, 3 OCOCH3), 4.26 (dd,
1 H, H-5a), 4.37 (dd, 1 H, H-5b), 4.59 (ddd, 1
H, H-4), 5.04 (br s, 1 H, H-2), 5.19 (s, 1 H,
H-1), 5.33 (dd, 1 H, H-3), J2,3 1.5, J3,4 4.6, J4,5a
7.1, J4,5b 4.9, J5a,5b 11.7 Hz; 13C NMR: l
20.48, 20.57, and 20.67 (3 OCOCH3), 61.76
(C-5), 74.05 (C-3), 79.49 (C-4), 80.32 (C-2),
74.82 (C-3), 80.13 (C-4), 80.91 (C-2), 93.73
(C-1), 128.3–129.9, 133.16, 133.76, and 133.84
(ArꢀC), 164.78, 165.00, 166.02 (3 CO); EIMS:
’
m/z 445 ([M−N3]+ , 10%), 105 ([PhCO]+,
100); CIMS (NH3): m/z 505 ([M+NH4]+,
30), 445 ([M−N3]+, 33), 105 (100); IR, NMR
and MS data are in agreement with the re-
ported values [5a]. Anal. Calcd for
C26H21N3O7 (487.47): C, 64.06; H, 4.34; N,
8.62. Found: C, 63.99; H, 4.45; N, 8.60.
93.31 (C-1), 169.07, 169.39, and 170.42 (3
’
OCOCH3); EIMS: m/z 259 ([M−N3]+ ,
50%), 139 ([M−N3−2 AcOH]+, 100), 128
(93). Anal. Calcd for C11H15N3O7 (301.26): C,
43.86; H, 5.02; N, 13.95. Found: C, 44.30; H,
4.95; N, 13.82.
a-Azide 9a: 118 mg (2.4%); Rf 0.52; IR
(film): w 2116 (N3), 1727 (CꢁO), 1268 (ben-
zoate CꢀO), 1110, and 710 cm−1 (benzoate);
1H NMR: l 4.53–4.64 (2 m, 2 H, H-5a,5b),
4.96 (ddd, 1 H, H-4), 5.61 (pseudo t, 1 H,
H-2), 5.96 (d, 1 H, H-1), 6.01 (dd, 1 H, H-3),
7.37–7.63 and 7.97–8.12 (2 m, 15 H, ArꢀH),
2-O-Acetyl-5-O-benzoyl-3-deoxy-i- -ery-
D
thro-pentofuranosyl azide (13b).—Azidation
of 12ab (322 mg, 1 mmol) afforded the crude
product mixture (b:a=124:1), which was
purified by chromatography (CH2Cl2) to give
pure 13b (222 mg, 74%) as a colourless oil; IR
(film): w 2113 (N3), 1748 (acetate CꢁO), 1723
(benzoate CꢁO), 1377 (acetate CH3), 1275
(benzoate CꢀO), 1231 (acetate CꢀO), 1125
13
J1,2 5.0, J2,3 5.3, J3,4 6.2 Hz; C NMR: l 62.12
(C-5), 74.86 (C-3), 75.62 (C-4), 77.03 (C-2),
89.67 (C-1), 128.4–130.0, 133.22, and 133.73
(ArꢀC), 165.42, 165.48, 165.96 (3 CO); CIMS

&
A. Stimac, J. Kobe / Carbohydrate Research 324 (2000) 149–160
159
9%; Rf 0.39 (CH2Cl2); mp 59–60.5 °C (from
MeOH)] and the a-azide 15a as a colourless
oil: 1.52 g (7%); Rf 0.30; IR (film): w 2114 (N3),
1724 (CꢁO), 1268 (benzoate CꢀO), 1116, and
(PhCOOꢀC), 1093, 1071, 1054, 1026, and 713
1
cm−1 (benzoate); H NMR: l 2.09 (s, 3 H,
OCOCH3), 2.12–2.20 (m, 1 H, H-3a), 2.24
(ddd, 1 H, H-3b), 4.38 (dd, 1 H, H-5a), 4.58
(dd, 1 H, H-5b), 4.72 (dddd, 1 H, H-4), 5.05
(dd, 1 H, H-2), 5.44 (s, 1 H, H-1), 7.43–7.49,
7.55–7.61, and 8.07–8.12 (3 m, 5 H, ArꢀH),
J2,3a 1.1, J2,3b 4.5, J3a,3b 14.0, J3a,4 6.6, J3b,4 9.5,
1
711 cm−1 (benzoate); H NMR: l 2.37–2.42
(m, 2 H, H-3a,3b), 4.38 (dd, 1 H, H-5a), 4.52
(dd, 1 H, H-5b), 4.77 (m, 1 H, H-4), 5.45 (dt,
1 H, H-2), 5.74 (d, 1 H, H-1), 7.42–7.49,
7.55–7.62 and 8.06–8.10 (3 m, 10 H, ArꢀH),
J1,2 4.5, J2,3a and J2,3b 7.6, J4,5a 4.9, J4,5b 3.4,
13
J4,5a 5.5, J4,5b 3.5, J5a,5b 12.0 Hz; C NMR: l
20.84 (CH3), 31.43 (C-3), 65.82 (C-5), 77.89
(C-2), 78.58 (C-4), 94.30 (C-1), 128.40 (C-
3%,5%), 129.61 (C-2%,6%), 129.67 (C-1%), 133.18
(C-4%), 166.28 and 169.97 (2 CO); FABMS:
m/z 306 ([MH+], 6%), 263 ([M−N3]+, 100),
105 ([PhCO]+, 92). Anal. Calcd for
C14H15N3O5 (305.29): C, 55.08; H, 4.95; N,
13.76. Found: C, 55.24; H, 4.79; N, 13.67.
a Anomer 13a could not be isolated due to
the paucity of this material. Nevertheless, its
presence in the crude mixture was identified
by the SnCl4-induced anomerization of pure
13b as described for 2b (entry 1 from Table 2).
Compound 13a displayed a doublet (J1,2 4.5
Hz) for the anomeric proton, centered at l
5.62 ppm.
13
J5a,5b 12.0 Hz; C NMR: l 30.78 (C-3), 65.80
(C-5), 73.42 (C-2), 75.39 (C-4), 90.94 (C-1),
128.4–129.8, 133.20, and 133.45 (PhꢀC),
165.72 and 166.18 (2 CO); CIMS (NH3): m/z
385 ([M+NH4]+, 2%), 340 ([MH−N2]+, 40),
325 ([M−N3]+, 100), 105 ([PhCO]+, 70).
Anal. Calcd for C19H17N3O5 (367.36): C,
62.12; H, 4.66; N, 11.44. Found: C, 61.98; H,
4.78; N, 11.47.
2,3,4,6-Tetra-O-acetyl-i- -galactopyran-
D
osyl azide (17).—Azidation of 16 (1.95 g, 5
mmol) afforded the crude product (1.67 g,
90%), which was crystallized from MeOH to
give 17 (1.31 g, 70%) as colourless cubes: mp
96.5 °C, lit. mp 96 °C [33]; IR (KBr): w 2165
and 2127 (N3), 1743 (CꢁO), 1379 (acetate
3-Deoxy-2,5-di-O-benzoyl-i- (15b) and -h-
-erythro-pentofuranosyl azide (15a).—Azida-
D
1
CH3), 1236 (acetate CꢀO), and 1216 cm−1; H
tion of 14b (21.0 g, 59 mmol) afforded the
crude product mixture (b:a=9:1), which was
crystallized from MeOH to give pure 15b
(16.75 g, 77%) as colourless crystals: mp 59–
61 °C; IR (film): w 2113 (N3), 1723 (CꢁO),
1268 (benzoate CꢀO), 1109, and 711 cm−1
NMR: l 1.99, 2.07, 2.10, and 2.18 (4 s, each 3
H, 4 OCOCH3), 4.03 (ddd, 1 H, H-5), 4.12–
4.23 (2 m, 2 H, H-6a,6b), 4.61 (d, 1 H, H-1),
5.04 (dd, 1 H, H-3), 5.17 (dd, 1 H, H-2), 5.43
(dd, 1 H, H-4), J1,2 8.6, J2,3 10.4, J3,4 3.3, J4,5
13
1
1.1, J5,6a 6, J5,6b 7 Hz; C NMR: l 20.45 and
(benzoate); H NMR: l 2.27–2.40 (m, 2 H,
20.55 (2 COCH3), 20.59 (2 COCH3), 61.16
(C-6), 66.77 (C-4), 67.96 (C-2), 70.64 (C-3),
72.76 (C-5), 88.19 (C-1), 169.30, 169.93,
H-3a,3b), 4.41 (dd, 1 H, H-5a), 4.62 (dd, 1 H,
H-5b), 4.82 (m, 1 H, H-4), 5.30 (m, 1 H, H-2),
5.60 (s, 1 H, H-1), 7.41–7.61 and 8.01–8.13 (2
m, 10 H, ArꢀH), J4,5a 5.6, J4,5b 3.7, J5a,5b 12.0
Hz; 13C NMR: l 31.54 (C-3), 65.76 (C-5),
78.30 (C-2), 78.62 (C-4), 94.31 (C-1), 128.3–
129.6, 133.09, and 133.45 (PhꢀC), 165.36 and
166.18 (2 CO); CIMS (NH3): m/z 385 ([M+
NH4]+, 3%), 340 ([MH−N2]+, 15), 325
([M−N3]+, 53), 105 ([PhCO]+, 100). Anal.
Calcd for C19H17N3O5 (367.36): C, 62.12; H,
4.66; N, 11.44. Found: C, 61.90; H, 4.49; N,
11.33.
170.06, and 170.32 (4 CO); EIMS: m/z 331
’
([M−N3]+ , 28%), 115 (100); FABMS: 374
(MH+, 1.2), 346 ([MH−N2]+, 3.7), 331
([M−N3]+, 100).
Acknowledgements
This investigation was supported by the
Ministry of Science and Technology of Slove-
nia (Grant Nos. 33-030 and CI-01513) and the
EU (BIOMED1, Agreement No. ERB-
CIPDCT 930194, PL 93-1112). We thank Dr
Bogdan Kralj at the Mass Spectrometry Cen
The residue, obtained after concentration of
the mother liquor, was resolved by chro-
matography (CH2Cl2, H2O content50.01%)
to give an additional amount of 15b [1.88 g,

&
A. Stimac, J. Kobe / Carbohydrate Research 324 (2000) 149–160
160
Perst, in G. Olah, P. von R. Schleyer (Eds.), Carbonium
Ions, Vol. 5, Wiley-Interscience, New York, 1976, pp.
1961–2047.
tre at the Jozˇef Stefan Institute (Ljubljana)
and Dr Janez Plavec at the Slovenian Na-
tional NMR Centre.
[16] S. Hu¨nig, Angew. Chem., Int. Ed. Engl., 3 (1964) 548–
560.
[17] P.R. Brodfuehrer, C. Sapino Jr., H.G. Howell, J. Org.
Chem., 50 (1985) 2597–2598.
[18] (a) T.L. Cupps, D.S. Wise, L.B. Townsend, J. Org.
Chem., 47 (1982) 5115–5120. (b) A.P. Kozikowski, K.L.
Sorgi, B.C. Wang, Z. Xu, Tetrahedron Lett., 24 (1983)
1563–1566.
[19] F. Becsi, E. Zbiral, Monatsh. Chem., 110 (1979) 955–978.
[20] W. Hartmann, H.-G. Heine, Tetrahedron Lett., (1979)
513–516.
[21] P.T. Jørgensen, E.B. Pedersen, C. Nielsen, Synthesis,
(1992) 1299–1306.
[22] N. Wiberg, K.H. Schmid, Chem. Ber., 100 (1967) 748–
754.
[23] S. Saito, S. Yamashita, T. Nishikawa, Y. Yokoyama, M.
Inaba, T. Moriwake, Tetrahedron Lett., 30 (1989) 4153–
4156.
[24] E.F. Recondo, H. Rinderknecht, Hel6. Chim. Acta, 42
(1959) 1171–1173.
[25] H.G. Fletcher Jr., H.W. Diehl, Carbohydr. Res., 4 (1967)
438–440.
References
[1] Z. Gyo¨rgydea´k, L. Szila´gyi, H. Paulsen, J. Carbohydr.
Chem., 12 (1993) 139–163.
&
[2] A. Stimac, I. Leban, J. Kobe, Synlett, (1999) 1069–1073.
[3] H. Paulsen, Z. Gyo¨rgydea´k, M. Friedmann, Chem. Ber.,
107 (1974) 1568–1578.
[4] K. Matsubara, T. Mukaiyama, Chem. Lett., (1994) 247–
250.
[5] (a) J. Hiebl, E. Zbiral, Liebigs Ann. Chem., (1988) 765–
774. (b) F. Hammerschmidt, J.-P. Polsterer, E. Zbiral,
Synthesis, (1995) 415–418.
&
[6] A. Stimac, J. Kobe, Carbohydr. Res., 232 (1992) 359–
365.
[7] (a) W. Scho¨rkhuber, E. Zbiral, Liebigs Ann. Chem.,
(1980) 1455–1469. (b) F.J. Schendel, J. Stubbe, Biochem-
istry, 25 (1986) 2256–2264.
[8] H. Vorbru¨ggen, G. Ho¨fle, Chem. Ber., 114 (1981) 1256–
1268.
[26] A. Holy, Collect. Czech. Chem. Commun., 42 (1977)
902–908.
[9] R.U. Lemieux, Ad6. Carbohydr. Chem., 9 (1954) 1–57.
[10] K.A. Watanabe, D.H. Hollenberg, J.J. Fox, J. Carbo-
hydr., Nucleosides, Nucleotides, 1 (1974) 1–37.
[11] (a) C. Chavis, F. Grodenic, J.-L. Imbach, Eur. J. Med.
Chem. Chim. Ther., 16 (1981) 219–227. (b) T. Itoh, H.
Takamura, K. Watanabe, Y. Araki, Y. Ishido, Carbo-
hydr. Res., 156 (1986) 241–246.
[27] (a) C.D. Anderson, L. Goodman, B.R. Baker, J. Am.
Chem. Soc., 80 (1958) 5247–5252. (b) W.W. Lee, G.L.
Tong, in W.W. Zorbach, R.S. Tipson (Eds.), Synthetic
Procedures in Nucleic Acid Chemistry, Vol. 1, Wiley–In-
terscience, New York, 1968, pp. 224–227.
[28] G. Biagi, I. Giorgi, O. Livi, V. Scartoni, Farmaco, 47
(1992) 525–536.
[12] (a) R.K. Ness, H.G. Fletcher Jr., J. Am. Chem. Soc., 78
(1956) 4710–4714. (b) J.D. Stevens, H.G. Fletcher Jr., J.
Org. Chem., 33 (1968) 1799–1805. (c) C. Chavis, F.
Dumont, J.-L. Imbach, J. Carbohydr., Nucleosides, Nu-
cleotides, 5 (1978) 133–147. (d) H.G. Howell, P.R. Brod-
fuehrer, S.P. Brundidge, D.A. Benigni, C. Sapino Jr., J.
Org. Chem., 53 (1988) 85–88.
[13] M. Popsavin, V. Popsavin, N. Vukojevic, D. Miljkovic,
Collect. Czech. Chem. Commun., 59 (1994) 1884–1888.
[14] R.L. Shone, Tetrahedron Lett., (1977) 993–996.
[15] (a) H. Paulsen, Ad6. Carbohydr. Chem. Biochem., 26
(1971) 127–195. (b) C.U. Pittman Jr., S.P. McManus,
J.W. Larsen, Chem. Re6., 72 (1972) 357–438. (c) H.
[29] S. Hanessian, J. Banoub, Carbohydr. Res., 59 (1977)
261–267.
[30] (a) P.J. Bolon, T.B. Sells, Z.M. Nuesca, D.F. Purdy, V.
Nair, Tetrahedron, 50 (1994) 7747–7764. (b) A. Kumar,
S.I. Khan, A. Manglani, Z.K. Khan, S.B. Katti, Nu-
cleosides, Nucleotides, 13 (1994) 1049–1058.
[31] M.J. Robins, M. MacCoss, S.R. Naik, G. Ramani, in
L.B. Townsend, R.S. Tipson (Eds.), Nucleic Acid Chem-
istry, Part 3, Wiley, New York, 1986, pp. 58–60.
[32] Y. Ishido, N. Sakairi, M. Sekiya, N. Nakazaki, Carbo-
hydr. Res., 97 (1981) 51–79.
[33] A. Bertho, J. Maier, Liebigs Ann. Chem., 498 (1932)
50–61.
.