Towards α- or β-D-C-glycosyl compounds by tin-catalyzed addition of glycosyl radicals to acrylonitrile and vinylphosphonate, and flexible reduction of tetra-O-acetyl-α-D-glucopyranosyl bromide with cyanoborohydride

Article abstract of DOI:10.1016/S0008-6215(02)00052-6

Photo-induced radical addition of acetylated α-D-glucopyranosyl bromide (1) to acrylonitrile or diethyl vinylphosphonate, in the presence of catalytic amounts of tri-n-butyltin chloride and sodium (or tetra-n-butylammonium) cyanoborohydride in excess, all

Full text of DOI:10.1016/S0008-6215(02)00052-6

Carbohydrate Research 337 (2002) 1623–1632
www.elsevier.com/locate/carres
Towards a- or b-
D
-C-glycosyl compounds by tin-catalyzed
addition of glycosyl radicals to acrylonitrile and vinylphosphonate,
and flexible reduction of tetra-O-acetyl-a- -glucopyranosyl
D
bromide with cyanoborohydrideꢀ
Jean-Pierre Praly,* Azin Salek Ardakani, Isabelle Bruye`re, Chrystelle Marie-Luce,
Bing Bing Qin
Uni6ersite´ Claude-Bernard Lyon I, UMR CNRS-Uni6ersite´ 5622, ESCPE-Lyon, Baˆtiment 308, 43, Boule6ard du 11 No6embre 1918,
F-69622 Villeurbanne, France
Received 5 November 2001; received in revised form 14 January 2002; accepted 15 February 2002
Abstract
Photo-induced radical addition of acetylated a-
presence of catalytic amounts of tri-n-butyltin chloride and sodium (or tetra-n-butylammonium) cyanoborohydride in excess,
-glucopyranosyl)-ethylphosphonate (79, 70% yields,
D-glucopyranosyl bromide (1) to acrylonitrile or diethyl vinylphosphonate, in the
allowed efficient preparations of a-configurated nonononitrile and 2-(a-
D
respectively). These conditions led to 2-(a-
D
-manno-, and galactopyranosyl)-ethylphosphonates in 68 and 76% yields. Similarly,
radical addition of acetylated 1-bromo-b-
D
-glucopyranosyl chloride (2) to acrylonitrile or diethyl vinylphosphonate afforded
mainly intermediate chlorides which, upon radical reduction with excess tri-n-butyltin hydride, afforded the corresponding b
anomers (40 and 38%, respectively) by sequential CꢀC and CꢀH bond formation. Stereocontrol relies on the a-stereoselective
quenching of
afforded either 1,3,4,6-tetra-O-acetyl-2-deoxy-a-
was added, 2,3,4,6-tetra-O-acetyl-1,5-anhydro- -glucitol (79%). These are simple, tin-free, and easily controlled conditions, which
D
-glycopyranos-1-yl radicals. We found also that UV light irradiation of 1 with excess NaBH3CN in tert-butanol
D
-arabino-hexopyranose (65% after crystallization) or, when 10% mol thiophenol
D
compare well with known preparations of these reduced sugars. © 2002 Elsevier Science Ltd. All rights reserved.
Keywords: Glycopyranos-1-yl radicals; C-Glycosyl compounds; Acetylated sugar nitriles; Acetylated (a- and b-D-glycopyran-
osyl)ethylphosphonates; Acetylated 1,5-anhydro-
D
-glucitol; Acetylated 2-deoxy-a-D-arabino-hexopyranose
otide diphosphate derivatives (in Leloir pathways),
while dolichyl diphosphate derivatives take part in the
biosynthesis of N-linked glycoproteins. Each of these
glycosyl phosphates transfers a glycosyl moiety, upon
cleavage of their glycosidic bond. In connection with
the interest in carbohydrate mimics,1 analogs of natural
glycosyl phosphates have been investigated, and among
them glycosyl phosphonates2 which, depending on the
modifications introduced, are classified as isosteric or
non-isosteric to natural molecules.2,3
Several synthetic routes to glycosylphosphonates
have been proposed. For example, access to glycosyl-
methylphosphonates in two steps from benzyl-protected
glyconolactones, by conversion with the lithium anion
of dimethyl methylphosphonate to lactol-phosphonates
subsequently reduced with triethylsilane, first consid-
ered as inefficient,2 appears now promising in the light
1. Introduction
Glycosylphosphates are essential for the bioconver-
sions of carbohydrates, being involved in the primary
metabolism, for example in glycolysis and gluconeogen-
esis, and in the biosynthesis of oligosaccharides,
polysaccharides, and varied glycoconjugates. Among
the glyco-processing enzymes mobilized by such
metabolic pathways, glycosyltransferases are responsi-
ble for glycosylations. As substrates, they use either
glycosyl phosphates (in non-Leloir pathways), or nucle-
ꢀ This work was presented at EUROCARB XI, Lisbon,
Portugal, 2–7, September, 2001 (abstract OC51).
* Corresponding author. Fax: +33-472-448349.
E-mail address: jean-pierre.praly@univ-lyon1.fr (J.-P.
Praly).
0008-6215/02/$ - see front matter © 2002 Elsevier Science Ltd. All rights reserved.
PII: S0008-6215(02)00052-6

1624
J.-P. Praly et al. / Carbohydrate Research 337 (2002) 1623–1632
of improved reduction conditions.4 Use of aldoses in
Wittig-type reactions as another approach to glyco-
sylphosphonates showed limitations in terms of yields,
and stereoselectivity.2 Nucleophilic substitution of a
glycosyl bromide with the lithium anion of dimethyl
methylphosphonate was claimed to result in 1,2-elimi-
nation to afford a glycal, instead of the desired glyco-
sylmethylphosphonate.2 Therefore, indirect methods
have been explored early,5 based on multi-step synthe-
ses of C-glycosyl compounds, either as halogenated
precursors for substitution by trialkylphosphite accord-
ing to the Arbuzov reaction,2 or with an aldehyde
group prone to nucleophilic attack by dialkylphos-
phite.6 In the particular case of 2-glycosyl-ethylphos-
phonates, the base-catalyzed Michael addition of
1-deoxy-1-nitrosugars7 or the addition of glycopyran-
osyl radicals8,9 to protected vinylphosphonates are
known, having a precedent with a radical-based access
to nucleoside derivatives.10 Being short routes, they
appear superior to indirect approaches.11,12 Coupling of
modified glycosylphosphates and glycosylphosphonates
to activated nucleotides led to stable analogs mimicking
substrates of glycosyltransferases. Some were found to
be good inhibitors of specific glycosyltransferases,13,14
and their applications as regards to carbohydrate-medi-
ated biological recognition have been discussed.13,14
In this context, sugar anomeric dihalides15,16 ap-
peared suitable precursors for preparing unknown
range 17–47% from the acetobromosugars), under
three related conditions (boiling Et2O) using either trib-
utyltin hydride (n-Bu3SnH, either irradiation with halo-
gen lamps or heating in the presence of AIBN) or
tris(trimethylsilyl)silane (irradiation). Due to the
stereoselective attack of glycohexopyranos-1-yl radi-
cals,17 the observed a/b ratio ranged from 85:15 to
\98:2.8 Optimization of this procedure in terms of
yield appeared desirable, before its extension to dihalide
2. Improvements were expected from photo-stimulated
conditions due to milder reaction temperatures, and
from the use of n-Bu3SnH in low concentration to
minimize competing hydrogen atom abstraction by gly-
cosyl radicals, yielding reduced sugars.17 This can be
achieved easily by using n-Bu3SnH (or preferably n-
Bu3SnCl) in catalytic amounts in the presence of such
reducing agents17 as sodium cyanoborohydride (in tert-
butanol) or tetrabutylammonium cyanoborohydride (in
benzene). In order to favor radical addition (CꢀC bond
formation) over radical reduction (CꢀH bond forma-
tion), use of alkene in excess (5–12-fold excess) is also
recommended.
Acrylonitrile was chosen for optimizing the radical-
mediated synthesis of C-glycosyl compounds since the
reaction products obtained from 120–22 and 223 are
known, in part from our study of the radical addition
of 2 to acrylonitrile under thermal conditions.23 As seen
from Scheme 1 and Table 1, the products obtained
under UV light irradiation arose mainly from radical-
mediated CꢀC bond formation. When n-Bu3SnH was
avoided, being replaced by n-Bu3SnCl (entry 4), the
acetylated
2-(b-D-glycopyranosyl)-ethylphosphonates
by radical addition to diethyl vinylphosphonate. This
idea was based on the expectation that radical addition
would lead to intermediate chlorides prone to a-
stereoselective radical reduction with tributyltin hy-
dride.17–19 Besides the b anomers expected from this
approach, a anomers can be obtained in a single radical
step from glycopyranosyl halides, as reported.8,9 Each
of them would be accessible precursors for preparing,
upon reaction with activated nucleotides, mimics of
substrates of glycosyltransferases, potentially interesting
for enzymatic studies, because their opposite configura-
tions might result in modulated bioactivities. This paper
describes stereocontrolled preparations of diethyl 2-
yield of 5a reached 79%, a value found higher than
21,22
those published,
while the reduced sugars were
isolated in a ꢀ10% yield only. Therefore, a molar ratio
sugar–n-Bu3SnCl–reducing agent–alkene equal to
1:0.3:ꢀ2:5 (see Tables 1 and 2) was fixed as a good
compromise throughout this work. 2,3,4,6-Tetra-O-ace-
tyl-1,5-anhydro-
acetyl-2-deoxy-a-
D
-glucitol (3)24,25 and 1,3,4,6-tetra-O-
-arabino-hexopyranose
D
(4),26–28
having similar TLC mobilities, were distinguished by
1H NMR. In the absence of n-Bu3SnH, the rearranged
2-deoxy sugar 4 was clearly observed (entry 5).
(2,3,4,6-tetra-O-acetyl-
D
-glycopyranosyl)-ethylphos-
Treatment of dihalide 2 under the same conditions
(entry 6) afforded compound 8 in 57% isolated yield.
The product distribution showed that radical addition
to acrylonitrile (5a: 5%; 8: 57%) predominated over
mono-reduction (6: 6%; 7: 4%) and double-reduction
(3/4: 4%). One simple explanation to the formation of 8
involves hydrolysis of a labile halogenated precur-
sor.19,23 So, in further experiments (entries 7–11) the
reaction mixture was reacted with excess n-Bu3SnH,
added either directly to the reaction medium (entry 11),
or after concentration under reduced pressure and addi-
tion of benzene as solvent (entries 7–10). These experi-
ments led to 5b in varying yields (up to 40%), with no
detectable trace (NMR) of the a anomer in accordance
phonates under radical conditions (either a, or b
anomer) and preliminary observations, collected en
route, concerning unprecedented photo-stimulated re-
ductions of 2,3,4,6-tetra-O-acetyl-a-D-glucopyranosyl
bromide with sodium cyanoborohydride.
2. Results and discussion
Addition of peracetylated glycopyranos-1-yl radicals
to diethyl vinylphosphonate has been reported8 to af-
ford predominantly the corresponding a-configurated
2-glycopyranosyl-ethylphosphonates (yields in the

J.-P. Praly et al. / Carbohydrate Research 337 (2002) 1623–1632
1625
with the high a-stereoselectivity of hydrogen atom ab-
straction by C-1-substituted hexopyranosyl radicals.17,18
The intermediate chloride underwent other competing
reactions to afford 8 and 9 whose structure was estab-
lished by NMR (nOe enhancements pointed to the
endo orientation of the H3C-11 group and vicinal cou-
radical addition to diethyl vinylphosphonate predomi-
nated, 12a being isolated in 70% yield (entry 3). With
other acetylated a-
D
-hexopyranosyl bromides (10:
D
-
-
manno, 11: -galacto), the corresponding 2-(a-
D
D
glycopyranosyl)-ethylphosphonates 13 and 14 were
obtained in 68 and 76% yield, respectively, significantly
higher than those reported.8 Isolation of a 1,2-O-ethyli-
plings J5,6 2.6, J6,7 1.7 Hz indicated a distorted
D-
glucopyranosyl ring, as showed in Scheme 1).29,30 Use
of NaBH3CN from a recently purchased batch favored
formation of 9, and did not prevent that of the product
8 (entries 10, 11), which might arise also from radical
quenching with molecular oxygen.31 Formation of 9 via
radical intermediates seems improbable, because of the
absence of any synthetic precedent based on radical
approaches.18,27,28,32 The conversion of 1,2-trans acyl-
glycosyl halides into 1,2-O-alkylidene33–35 by NaBH4-
mediated reductive attack of 1,3-dioxolenium ions is
well known. Due to the a-stereoselective quenching of
dene derivative was achieved in the
only, (R/S)-3,4,6-tri-O-acetyl-1,2-O-ethylidene-b-
mannopyranose34 being obtained in 12% yield, from the
1,2-trans related 2,3,4,6-tetra-O-acetyl-a- -mannopy-
D-manno series
D
-
D
ranosyl bromide (10). This data is in keeping with the
proposed reaction path to 9. When dihalide 2 was
reacted with diethyl vinylphosphonate in the presence
of n-Bu3SnCl and either NaBH3CN or n-Bu4N
BH3CN, radical addition led to 15 and 16 in 28–36%
total yield (entries 5, 6). The b anomer 12b could be
obtained from 2 by a sequential one-pot procedure
(entries 7–9) in up to 38% yield (entry 8). All these
results showed that addition of hexopyranos-1-yl radi-
cals to such electron-poor alkenes as acrylonitrile or
diethyl vinylphosphonate can be controlled to afford in
uniform yields either a-configurated C-glycosyl
derivatives17 (5a: 79%, 12a: 70%, 13: 68%, 14: 76%
from acetobromosugars) or the b anomers18,19 in two
steps from the anomeric dihalide 2 (5b: 40%, 12b: 38%).
Syntheses of stable substrate analogs of glycosyltrans-
ferases from the glycosylphosphonates described herein
are underway and will be reported in due course.
D
-glucopyranos-1-yl radicals, the chloride initially
formed in the radical addition is expected to have a
b-oriented chlorine atom. Hence, chloride displacement
by the vicinal acetoxy group could afford a 1,3-dioxole-
nium ion easily. Its reduction to 9 should involve
NaBH3CN rather than n-Bu3SnH, being present in low
concentration.
Photo-initiated addition of radicals derived from
halides 1 and 2 to diethyl vinylphosphonate was investi-
gated next (Scheme 2, Table 2). Use of n-Bu3SnCl with
varying amounts of NaBH3CN in tert-butanol (entries
1–3) showed that, with 1.5 equiv of the reducing agent,
Scheme 1. (a) Acrylonitrile, 5 equiv; n-Bu3SnCl, 0.3 equiv; NaBH3CN, 2 equiv; tert-butanol; AIBN; UV light irradiation (Vycor
filter); reaction temperature, ꢀ30–35 °C. (b) (1) Acrylonitrile, 5 equiv; n-Bu3SnCl, 0.3 equiv; NaBH3CN, 2 equiv; tert-butanol;
AIBN; UV light irradiation (Vycor filter); 30 min; (2) tert-butanol replaced by benzene; n-Bu3SnH, 4 equiv; AIBN; UV light
irradiation (Vycor filter); 15 min.

1626
J.-P. Praly et al. / Carbohydrate Research 337 (2002) 1623–1632
Table 1
Addition of radicals derived from 1 and 2 to acrylonitrile a with either n-Bu3SnH (A) or catalytic n-Bu3SnCl (B) and NaBH3CN
(C) b
Entry Halide (mmol) Reductant (equiv)
Solvent (mL)
Time (h)
Products (%) c
3–4 d
5a 5b
6
7
8
9
1
2
3
1, 10
1, 5
1, 0.5
A: 1+0.4
A: 1.1+0.4
A+C
Et2O, 20 e
4+39
4+6
2
19 (\95/5)
15 (\95/5)
12
37
49
56
3
2
1.5
Et2O, 10 f
t-BuOH, 12 g
0.3+2
B+C
0.3+2
B+C
0.3+2
B+C
0.3+2
B+C
0.3+2
B+C
0.3+2
B+C
0.3+2
B+C l
0.3+2
B+C l
0.3+1.5
4
5
1, 0.5
t-BuOH, 12 g
t-BuOH, 3 g
t-BuOH, 12 h
2.5
9 (3/2)
79
60
5
1, 0.12
2, 0.5
2.5
13 (ꢀ3/2)
6
0.75
4
9.5
11
6
4
57
24
30
7
2, 0.5
t-BuOH, 11.5 g 0.5 i
benzene, 7
0.35 j
t-BuOH, 11.5 g 0.5 i
benzene, 12
0.25 j
40
8
2, 0.5k
2, 0.5 k
2, 0.25
2, 0.5 k
24
9
t-BuOH, 11.5 g 0.5 i
40 24
28 27
21 32 m
benzene, 12
t-BuOH, 5 g
benzene, 8
t-BuOH, 8
0.25 j
0.75 i
0.5 j
10
11
12
6
0.5 i
10 m
0.5 j
a Unless indicated otherwise, all reactions were carried out at ꢀ35 °C under argon, using a medium pressure mercury lamp
(Hanovia, 450 W), equipped with a Vycor filter (u\254 nm) and inserted in a two-wall jacket made of quartz with a tap water
flow for cooling. The reaction mixture was introduced into a quartz tube (ꢀ13 mL volume, 13 mm external diameter) equipped
with a stirring bar. Acrylonitrile was used in excess (5 equiv, unless otherwise noted) while AIBN was used in catalytic amounts,
as indicated. Except for run 6, the transformation of 2 involved two steps: first, radical addition to acrylonitrile in t-BuOH and
then radical reduction of an intermediate chloride with n-Bu3SnH (4 equiv) plus AIBN (0.6 equiv) in benzene. Therefore, unless
otherwise stated, after conversion of 2, t-BuOH was evaporated under reduced pressure and replaced with benzene.
b First, NaBH3CN taken from a batch kept for an undetermined period in an excicator was used, but control experiments were
carried out with NaBH3CN from a recently purchased batch, as indicated.
c Isolated yields.
1
d The 3:4 ratio, when estimated by H NMR, is given in parenthesis.
e The reaction medium was introduced in a round-bottomed pyrex flask and refluxed with a 250 W sun lamp placed below the
flask. After boiling the ethereal solution for 4 h, the solids were filtered off. More acrylonitrile (2.4 equiv), n-Bu3SnH (0.4 equiv),
and AIBN (0.03 equiv) were added to the mixture before heating for 39 h.
f After irradiation with filtered UV light for 4 h, the solids formed were filtered off. More acrylonitrile (2.4 equiv), n-Bu3SnH
(0.4 equiv), and AIBN (0.03 equiv) were added to the mixture before exposure to UV light for 6 h.
g With 0.6 equiv AIBN.
h With 0.3 equiv AIBN.
i One-pot reaction: irradiation time for the addition step.
j One-pot reaction: irradiation time for the reduction step.
k A larger quartz tube was used (ꢀ94 mL volume; 27 mm external diameter) for exposure to unfiltered UV light.
l NaBH3CN from a recently purchased batch.
m Estimated yields based on the 1H NMR analysis of a mixture (81 mg, 5b–9 in 3:1 ratio) collected after column
chromatography.
n-Bu3SnH–n-Bu3SnCl added, compound 5a was iso-
lated in low yield (9%). Interestingly, 1 was converted
mainly into 3–4 (53% yield), with the rearranged
product 4 being predominant (3–4 ratio: 17:83). There-
fore, under such tin-free conditions, formation of 4 and
The aforementioned addition reactions were believed
to be mediated mainly by tin radicals, and that view
was confirmed by a blank experiment. When the addi-
tion of 1 to acrylonitrile was attempted under irradia-
tion (4 h) with NaBH3CN and AIBN, but without any

J.-P. Praly et al. / Carbohydrate Research 337 (2002) 1623–1632
1627
required,27,28 difficulties due to toxicity, and elimination
of organotin compounds18) or costly tris(trimethylsi-
lyl)silane,36 and compared also with the conversion of 1
5a occurred, suggesting the operation of competing
radical processes, not based on tin. Without acryloni-
trile, enhanced yield of 3–4 was expected. Indeed,
exposure of a mixture of 1 and NaBH3CN in tert-bu-
tanol to filtered UV light for ꢀ0.5 h resulted in a clean
reaction, visualized by a new single spot on TLC plates.
Simple workup led to crude products, shown by 1H
NMR spectroscopy to contain essentially 4, in admix-
ture with 3 (4–3 ratio: ꢀ95:5). These experiments led
reproducibly to pure 4 as nice colorless prisms in
60–65% yield (Scheme 3), under simple and inexpensive
conditions. In spite of somewhat lower yields, the pro-
posed method is advantageous, as compared to radical
methods using either tin hydride (slow hydride addition
into 3,4,6-tri-O-acetyl-1,5-anhydro-2-deoxy-
hex-1-enitol (triacetyl- -glucal), followed by addition of
either acetic acid (crude 4, 98.4%, a/b ratio: 71:29)37 or
D-arabino-
D
water, to give 3,4,6-tri-O-acetyl-2-deoxy-
D-arabino-
hexopyranose (67% yield, a/b mixture).38
The suspicion that conversion of 1 into 4 proceeded
via radical intermediates27,28 was reinforced by a further
experiment based on the concept of polarity reversal
catalysis.39 While the electrophilic character of
D-
glycopyranos-2-yl radicals makes them prone to hydro-
gen atom abstraction from cyanoborohydride anion or
Table 2
Addition of radicals derived from 1 and 2 to diethyl vinylphosphonate a with catalytic n-Bu3SnCl (B) and either NaBH3CN b (C)
or n-Bu4NBH3CN (D)
Entry Halide (mmol)
Reductant (equiv)
Solvent (mL)
Time (h)
Products (%) c
3–4 d
7
12a 12b 15 16
1
2
3
4
5
6
7
8
9
1, 0.49
1, 0.49
1, 1.5
B+C
0.3+2
B+C
0.3+1.5
B+C e
0.3+1.5
B+D
0.3+1.5
B+C
0.3+1.5
B+D
0.3+2
B+D
0.3+2
B+D
0.3+2
B+D
t-BuOH, 8
t-BuOH, 8
t-BuOH, 24
benzene, 8
t-BuOH, 8
benzene, 6
benzene, 9
benzene, 8 i
benzene, 8 i
0.6
0.8
1
26 (47/53)
13 (47/53)
8 (39/61)
8
3
3
3
55
63
70
1, 0.49
2, 0.45
2, 0.45
2, 0.68 f
2, 0.68 f
2, 0.68 f
9
25
38
0.6
1
10 18
18 18
1 g
21 (\95/5)
29 (\95/5)
24 (\95/5)
34
9
27
38
20
8
7
2
0.25 h
1.5 g
0.5 j
2.5 g
0.35 j
5
0.3+1.5
a The reaction mixture, containing diethyl vinylphosphonate (5 equiv), n-Bu3SnCl (0.3 equiv), AIBN (0.6 equiv), sodium or
tetra-n-butyl ammonium cyanoborohydride in the selected solvent, as indicated, was introduced into a quartz tube (ꢀ94 mL
volume, 27 mm external diameter) equipped with a stirring bar. Argon was bubbled into the liquid until the sugar halide get
dissolved (ꢀ0.5/1.5 h). All reactions were carried out at ꢀ35 °C under argon, using a medium pressure mercury lamp (Hanovia,
450 W), equipped with a Vycor filter (u\254 nm) and inserted into a two-wall jacket made of quartz and cooled with a tap water
flow. Some one-pot type experiments carried out with 2 involved first a radical addition to diethyl vinylphosphonate, then radical
reduction of the intermediate chloride 15 in benzene with excess n-Bu3SnH (3 equiv, or more), as indicated.
b As indicated, NaBH3CN was taken either from a recently purchased batch or from a batch kept for an undetermined period
in an excicator.
c Isolated yields.
1
d The 3:4 ratio, given in parenthesis, was estimated by H NMR.
e NaBH3CN was taken from a recently purchased batch.
f Radical-based addition and reduction carried out in one pot.
g Irradiation time for the addition step.
h Irradiation time for the reduction step with 10 equiv n-Bu3SnH.
i Argon was bubbled through the liquid for ꢀ2 h.
j Irradiation time for the reduction step with 3 equiv n-Bu3SnH.

1628
J.-P. Praly et al. / Carbohydrate Research 337 (2002) 1623–1632
Scheme 2. (a) Diethyl vinylphosphonate, 5 equiv; n-Bu3SnCl, 0.3 equiv; NaBH3CN, 1.5 equiv; tert-butanol; AIBN; UV light
irradiation (Vycor filter); temperature, ꢀ30–35 °C. (b) (1) Diethyl vinylphosphonate, 5 equiv; n-Bu3SnCl, 0.3 equiv; n-
Bu4NBH3CN, 2 equiv; benzene; AIBN; UV light irradiation (Vycor filter); (2) n-Bu3SnH, 3 equiv; AIBN; UV light irradiation
(Vycor filter).
related species, anomeric radicals are nucleophilic and
abstract hydrogen atom from thiols and selenols read-
ily.18 This accounts to the fact that the presence of
catalytic amounts of thiols/selenols prevents formation
of 2-deoxy sugars by radical rearrangement under
poorly hydrogen-donating conditions, thus favoring
formation of anhydro-itols.40 This expectation was
founded, since under the aforementioned conditions,
but with 10 mol% thiophenol added, almost pure 1,5-
resulting in a simple and efficient synthesis of acetylated
1,5-anhydro- -glucitol from tetra-O-acetyl-a- -gluco-
pyranosyl bromide, whereas the rearranged 1,3,4,6-
tetra-O-acetyl-2-deoxy-a- -arabino-hexopyranose was
formed with NaBH3CN alone (Scheme 3). Worth of
note are the high flexibility of these reductions, and the
unprecedented use of NaBH3CN. Its photo-reactivity
appears to be poorly developed.41 Therefore, ongoing
investigations to establish the scope, limitations, and
mechanisms of such reductions are in progress and will
be reported soon.
D
D
D
anhydro-D-glucitol was isolated by chromatography in
79% yield (3–4 ratio: ꢀ98:2), a result which compares
well with other routes to 3.18,22 Hence, adding 10 mol%
thiophenol modified dramatically the reduction issue,
3. Experimental
General methods.—Benzene and tert-butanol were
distilled over CaH2 and under argon atmosphere. Di-
ethyl vinylphosphonate (97% grade), NaBH3CN (95%
grade) and Bu4NBH3CN were purchased from Aldrich,
and used as received. Air was removed from the reac-
tion media by bubbling argon into organic solutions for
0.5–1.5 h. For photo-induced transformations, except
that mentioned in Table 1, entry 1, deoxygenated solu-
tions introduced into a quartz tube (small one: ꢀ13
mL volume, 13 mm external diameter or large one:
ꢀ94 mL volume, 27 mm external diameter) placed
near (ꢀ1 cm) a medium pressure mercury lamp
(Hanovia, 450 W) equipped with a Vycor filter (u\254
nm) were irradiated under an argon atmosphere and
Scheme 3.

J.-P. Praly et al. / Carbohydrate Research 337 (2002) 1623–1632
1629
magnetic stirring (the lamp was inserted into a two-wall
device made of quartz and allowing circulation of tap
water for cooling purpose). NMR spectra were
recorded with Bruker spectrometers AC 200, DRX 300,
and DRX 500 for solutions in CDCl3. Chemical shifts
(l) are in ppm (reference: Me4Si); coupling constants
170.0, 169.9 (CꢁO), 119.8 (C-1), 96.5 (C-4), 72.5 (C-5),
71.5 (C-6), 69.1 (C-8), 68.5 (C-7), 62.1 (C-9), 34.2 (C-3),
21.15, 21.03, 21.02, 21.00 (acetyl), 11.2 (C-2). Anal.
Calcd for C17H23NO10 (387): C, 50.87; H, 5.78; N, 3.49;
O, 39.86. Found: C, 50.95; H, 5.71; N, 3.49; O, 39.59.
(10R)-6,7,9-Tri-O-acetyl-2,3-dideoxy-4,5-ethylidene-
3
(Hz) other than J are specified with the appropriate
h-D-gluco-4-nonulopyranononitrile (9).—A solution
superscript. 31P NMR spectra were recorded with
Bruker spectrometers at 81 MHz (AC 200) or 121.5
MHz (DRX 300) from CDCl3 solutions with H3PO4 as
the external reference. Mass spectra were measured
with a Finnigan Mat 95 XL spectrometer. Other details
have been published recently.19,31
prepared as before for the radical addition of 2 (0.5
mmol) to acrylonitrile was irradiated for 30 min, where-
upon the solvent was evaporated under reduced pres-
sure. The residue was dissolved in benzene (12 mL),
and after adding n-Bu3SnH (580 mg, 2 mmol) and
AIBN (25 mg, 0.15 mmol) to the solution introduced
into a quartz tube, irradiation was applied for 15 min.
Benzene was evaporated under diminished pressure and
the residue was diluted in CH2Cl2, washed with water
and dried (MgSO4). After filtration and concentration,
the product mixture was freed from organotin com-
pounds by washing an MeCN solution with pentane
(3×15 mL). Chromatography on silica gel with 2:3
EtOAc–petroleum ether as the mobile phase afforded
3–4, 5b, 8, and 9 (see Table 1). Compound 9 being only
slightly more mobile than 5b could be obtained in pure
state after repeated chromatography.
5,6,7,9-Tetra-O-acetyl-2,3-dideoxy-h-
ulopyranononitrile (8).—A solution containing 2,3,4,6-
tetra-O-acetyl-1-bromo-b- -glucopyranosyl chloride
D-gluco-4-non-
D
(2)15 (222 mg, 0.5 mmol), acrylonitrile (132.5 mg, 2.5
mmol), n-Bu3SnCl (48.8 mg, 0.15 mmol), NaBH3CN
(62.8 mg, 1 mmol), and AIBN (25 mg, 0.15 mmol) in
tert-butanol (12 mL) was introduced into a quartz tube
(13 mm external diameter). While stirring the solution
under an argon atmosphere, irradiation with filtered
UV light was applied for 45 min, whereupon TLC
monitoring showed the transformation of the starting
material into five products. After concentration under
reduced pressure, the residue was taken up in CH2Cl2.
The organic phase was washed with water, then dried
(Na2SO4), and concentrated to a residue which was
dissolved in MeCN. Organotin compounds were re-
moved by washing the MeCN phase with pentane
(3×15 mL). Upon concentration and chromatography
of the mixture on a column of silica gel eluted with 2:3
EtOAc–petroleum ether, concentration of homoge-
neous fractions led to the following products: 6 (Rf
0.65, 11 mg, 0.03 mmol, 6% yield), 3–4 (Rf 0.3, 6.2 mg,
0.02 mmol, 4% yield), 5a (Rf 0.18, 10.1 mg, 0.026
mmol, 5% yield), 7 (Rf 0.14, 7 mg, 0.02 mmol, 4%
yield), and 8 (Rf 0.1, 115.2 mg, 0.28 mmol, 57.4% yield).
8: colorless crystals, mp 102–105 °C (Et2O–
petroleum ether); Rf 0.1 in 2:3 EtOAc–petroleum ether;
[h]2D0 +33° (c 0.6, CH2Cl2); IR (KBr) w 3400 (OH), w
9: Colorless oil; [h]2D2 +31.6° (c 1.2, CH2Cl2); 1H
NMR (200.13 MHz, CDCl3): l 5.17 (dd, 1 H, J5,6 2.6,
J6,7 1.7 Hz, H-6), 5.03 (q, 1 H, J10,11 4.9 Hz, H-10), 4.83
(dt, 1 H, J7,8 9.5 Hz, H-7), 4.17 (d, 2 H, J 4 Hz, H-9a,
H-9b), 4.07 (dt, 1 H, J8,9a=J8,9b 4.1 Hz, H-8), 3.79 (dd,
4
1 H, J5,7 0.9 Hz, H-5), 2.61 (dd, 2 H, J 6.9, J 8.8 Hz,
H-2a, H-2b), ꢀ2.10 (m, 2 H, H-3a, H-3b), 2.13, 2.10,
2.09 (acetyl), 1.49 (d, 3 H, endo-H3C-11); nOe differ-
ence spectra recorded for a CDCl3 solution of 9
showed, on selective saturation at 1.5 ppm (H3C-11),
enhancements (ppm, %) of the following resonances:
HC-10 (5.03, 5.7%) and H-8 (4.07, 3%). Selective satu-
ration at 5.03 ppm (HC-10) resulted in enhancements
(ppm, %) of the H-5 (3.79, 7.1%) and H3C-11 (1.5,
7.3%) resonances, in support of the endo and exo
orientations of, respectively, the H3C-11 group and the
HC-10 hydrogen atom; 13C NMR (50.32 MHz, CDCl3):
l 170.7, 169.8, 169.0 (CꢁO), 119.6 (CN), 102.7 (C-4),
100.9 (C-10), 77.5 (C-5), 70.3 (C-6), 68.4 (C-7), 67.0
(C-8), 63.1 (C-9), 33.9 (C-3), 20.9, 20.8, 20.8 (acetyl),
19.2 (C-11), 11.4 (C-2). MS (CI, isobutane): m/z 386
(17) [M+H]+, 342 (53) [M+H−CH3CHO]+, 282
(6.5) [M+H−CH3CHO−CH3COOH]+, 222 (100)
[M+H−CH3CHO−2CH3COOH]+, 162 (34.6) [M+
H−CH3CHO−3CH3COOH]+. HRMS Calcd for
C17H24NO9 [M+H] 386.145106. Found: 386.14503.
1
2240 (CN), and w 1750 cm−1 (CO); H NMR (200.13
MHz, CDCl3): l 5.45 (t, 1 H, J6,7 9.6 Hz, H-6), 5.09 (t,
1 H, J7,8 9.7 Hz, H-7), 4.96 (d, 1 H, J5,6 9.7 Hz, H-5),
4.25 (dd, 1 H, J9a,9b 11.6, J8,9a 5.8 Hz, H-9a), 4.12 (dd,
1 H, H-9b), ꢀ4.2 (m, 1 H, H-8), ꢀ3.5 (br s, 1 H,
OH), 2.56 (m, 2 H, H-2a, H-2b), 2.11, 2.09, 2.03, 2.00
(4s, 12 H, acetyl), ꢀ2.0 (m, 2 H, H-3a, H-3b); nOe
difference spectra recorded for a CDCl3 solution of 8
showed, on selective saturation at 2.5 ppm (H2C-2),
enhancements (ppm, %) of the following resonances:
H2C-3 (ꢀ2.0, 6%) and H-5 (4.96, 4%). Selective satura-
tion at ꢀ2.0 (H2C-3) resulted in enhancements (ppm,
%) of the H2C-2 (2.56, 16%) and H-5 (4.96, 9%) reso-
nances, in support of the methylene carbons attribu-
tion; 13C NMR (75.46 MHz, CDCl3): l 171.2, 170.6,
Diethyl
2-(2,3,4,6-tetra-O-acetyl-h-D-glucopyran-
osyl)-ethylphosphonate (12a).—A solution of 1 (604.1
mg, 1.47 mmol), n-Bu3SnCl (125 mL, 0.44 mmol, 0.3
equiv), NaBH3CN (95% grade, 148.5 mg, 2.21 mmol,
1.5 equiv), diethyl vinylphosphonate (1.17 mL, 7.35
mmol, 5 equiv), and AIBN (146.9 mg, 0.88 mmol, 0.6

1630
J.-P. Praly et al. / Carbohydrate Research 337 (2002) 1623–1632
1
equiv) in freshly distilled tert-butanol (24 mL) was
introduced in a quartz tube (ꢀ94 mL volume, 27 mm
external diameter) equipped with a magnetic bar. While
stirring under an argon atmosphere, the reaction mix-
ture was irradiated with filtered UV light for 50 min,
whereupon TLC monitoring (1:1 EtOAc–petroleum
ether) showed complete transformation of 1 (Rf 0.70)
into reduced products 3–4 (Rf 0.65), 2,3,4,6-tetra-O-
13: syrup; [h]2D5 +8° (c 1, CHCl3); H NMR (500
MHz, CDCl3): l 5.24 (dd, 1 H, J4,5 3.3, J5,6 8.3 Hz,
H-5), 5.17 (t, 1 H, J6,7 8.3 Hz, H-6), 5.15 (dd, 1 H, J3,4
3.4 Hz, H-4), 4.39 (dd, 1 H, J7,8a 6.6, J8a,8b 12.3 Hz,
H-8a), 4.19–4.05 (m, 4 H, 2 OCH2CH3), 4.10 (dd, 1 H,
J7,8b 2.5 Hz, H-8b), 3.95 (dt, 1 H, J2,3 10.9 Hz, H-3),
3.87 (ddd, 1 H, H-7), 2.12, 2.10, 2.08, 2.05 (4s, 12 H,
acetyl), 2.10–1.60 (m, 4 H, H-1a, H-1b, H-2a, H-2b),
4
acetyl-
(2,3,4,6-tetra-O-acetyl-a-
D
-glucopyranose (7) (Rf 0.48), and diethyl 2-
-glucopyranosyl)-ethylphos-
1.34 (dt, 6 H, JH,P 2.3, JH,H 7.1 Hz, 2 OCH2CH3); 13C
D
NMR (75.47 MHz, CDCl3): l 170.4, 170.0, 169.6, 169.5
(4s, CꢁO), 73.5 (d, JC-3,P 16.8 Hz, C-3), 70.7 (C-7), 69.9
(C-4), 68.4 (C-5), 66.9 (C-6), 62.0 (C-8), 61.6 (d, 2 C,
phonate (Rf 0.03). After concentration under dimin-
ished pressure, the oily residue was dissolved in CH2Cl2
(ꢀ30 mL), and after washing the organic phase with
water (3×30 mL), it was dried (MgSO4), filtered, and
concentrated under diminished pressure. The residue
was taken up in CH3CN (ꢀ30 mL), and the organotin
compounds were removed by washing with hexane
(3×30 mL). Concentration and chromatography on a
column of silica gel eluted with 1:1.5 EtOAc–petroleum
ether afforded 3–4 (40.8 mg, 0.123 mmol, 8% yield,
3–4 ratio: 39:61). Further elution with 20:1 EtOAc–
ethanol as the mobile phase gave 7 (17 mg, 0.05 mmol,
3%), and 12a (510.0 mg, 1.03 mmol, 70%) as an oil.
12a: Rf 0.36 in 20:1 EtOAc–ethanol; [h]2D5 +47° (c 1,
2JC,P 6.2 Hz, OCH2CH3), 22.1 (d, JC,P 4.3 Hz, C-2),
2
1
21.3 (d, JC-1,P 143.6 Hz, C-1), 20.7, 20.6, 20.6, 20.5
(acetyl), 16.3 (d, 2 C, JCH-3,P 5.6 Hz, OCH2CH3); 31P
NMR (121.5 MHz, CDCl3): l 32.05; Anal. Calcd for
C20H33O12P (496.44): C, 48.39; H, 6.70; P, 6.24. Found:
C, 47.05; H, 6.83; P, 6.14.
Diethyl 2-(2,3,4,6-tetra-O-acetyl-h-
D-galactopyran-
osyl)-ethylphosphonate (14).—As indicated for 12a, this
product was prepared from 2,3,4,6-tetra-O-acetyl-a-D-
galactopyranosyl bromide (11) (579.3 mg, 1.41 mmol,
reaction time: 40 min) in 76% yield. A ꢀ7:3 insepara-
ble mixture of 2,3,4,6-tetra-O-acetyl-1,5-anhydro-D-
CH2Cl2); H NMR (200 MHz, CDCl3):8 l 5.31 (t, 1 H,
galactitol25 and 1,3,4,6-tetra-O-acetyl-a-
D-lyxo-hexo-
1
J4,5 9.5 Hz, H-5), 5.10 (dd, 1 H, J3,4 5.7 Hz, H-4), 4.98
(t, 1 H, J6,5 9.5 Hz, H-6), 4.24 (dd, 1 H, J7,8a 5.5, J8a,8b
11.9 Hz, H-8a), 4.16–4.05 (m, 6 H, H-3, H-8b, 2
OCH2CH3), 3.81 (m, 1 H, H-7), 2.09, 2.08, 2.06, 2.03
(4s, 12 H, acetyl), 1.92–1.60 (m, 4 H, H-1a, H-1b,
pyranose28 (Rf 0.78 in 1:1 EtOAc–petroleum ether, 59
mg, 0.18 mmol, 13%) was recovered first from the
column. 14: syrup; Rf 0.48 in 20:1 EtOAc–ethanol; [h]D25
1
+61° (c 1, CHCl3); H NMR (300 MHz, CDCl3): l
5.41 (dd, 1 H, J5,6 3.3, J6,7 2.5 Hz, H-6), 5.29 (dd, 1 H,
J3,4 5.2, J4,5 9.2 Hz, H-4), 5.19 (dd, 1 H, H-5), 4.23 (dd,
1 H, J7,8a 7.7, J8a,8b 11.4 Hz, H-8a), 4.23–4.14 (m, 6 H,
H-3, H-8b, 2 OCH2CH3), 4.00 (ddd, 1 H, J7,8b 4,8 Hz,
H-7), 2.13, 2.09, 2.06, 2.03 (4s, acetyl), 1.34 (dt, 6 H,
4JH,P 1.8, JCH-2,CH-3 7.0 Hz, 2 OCH2CH3); 13C NMR
(125 MHz, CDCl3): l 170.3, 170.1, 169.8, 169.7 (4
CꢁO), 71.7 (d, JC,P 16.5 Hz, C-3), 68.0 (C-7), 67.9 (C-4),
67.6 (C-5), 67.3 (C-6), 61.6 (d, 2JC,P 6.5 Hz, OCH2CH3),
4
H-2a, H-2b), 1.34 (dt, 6 H, J 7, JH,P 1.3 Hz, 2
OCH2CH3); 13C NMR (50 MHz, CDCl3): l 170.4,
169.9, 169.5, 169.4 (4s, CꢁO), 72.4 (d, JC-3,P 16.8 Hz,
C-3), 70.1 (C-5), 70.0 (C-4), 68.7 (C-7), 68.6 (C-6), 62.2
2
(C-8), 61.6 (d, 2 C, JC,P 6.5 Hz, OCH2CH3), 21.1 (d,
1JC-1,P 144.3 Hz, C-1), 20.6, 20.6, 20.5, 20.5 (acetyl),
2
18.9 (d, JC-2,P 3.9 Hz, C-2), 16.4 (d, 2C, JCH-3,P 5.9 Hz,
OCH2CH3); 31P NMR (81 MHz, CDCl3): l 31.54; MS
(FAB+, glycerol+LiCl): m/z 503.2 [M+Li]+; Anal.
Calcd for C20H33O12P (496.44): C, 48.39; H, 6.70; P,
6.24. Found: C, 48.43; H, 6.72; P, 6.31.
2
61.6 (d, JC,P 6.6 Hz, OCH2CH3), 61.3 (C-8), 21.2 (d,
1JC,P 143.9 Hz, C-1), 20.6, 20.5 (2 C), 20.45 (acetyl),
2
19.6 (d, JC,P 3.5 Hz, C-2), 16.3 (d, 2 C, JC,P 6.0 Hz,
Diethyl
osyl)-ethylphosphonate (13).—As described for 12a, this
product was prepared from 2,3,4,6-tetra-O-acetyl-a-
2-(2,3,4,6-tetra-O-acetyl-h-
D
-mannopyran-
OCH2CH3); 31P NMR (81 MHz, CDCl3): l 31.7; Anal.
Calcd for C20H33O12P (496.44): C, 48.39; H, 6.70; P,
6.24. Found: C, 48.48; H, 7.54; P, 6.28.
D
-
mannopyranosyl bromide (10) (550.8 mg, 1.34 mmol,
reaction time: 55 min) in 68% yield. The following
products were recovered successively from the column
(with 2:3 EtOAc–petroleum ether as the mobile phase):
Diethyl
2-(2,3,4,6-tetra-O-acetyl-i-D-glucopyran-
osyl)-ethylphosphonate (12b).—A solution of 215 (300
mg, 0.676 mmol), n-Bu3SnCl (60 mL, 0.2 mmol, 0.3
equiv), Bu4NBH3CN (382 mg, 1.15 mmol, 2 equiv),
diethyl vinylphosphonate (540 mL, 3.38 mmol, 5 equiv),
and AIBN (67 mg, 0.41 mmol, 0.6 equiv) in dry oxy-
gen-free benzene (8 mL) was introduced in a quartz
tube (ꢀ94 mL volume; 27 mm external diameter)
equipped with a magnetic bar. While stirring under an
argon atmosphere, the reaction mixture was irradiated
with filtered UV light for 1.5 h, whereupon complete
transformation of 2 was indicated by TLC monitoring.
3,4,6-tri-O-acetyl-1,2-O-ethylidene-b-
nose34 as a ꢀ9:1 R/S mixture (Rf 0.39, 49.5 mg, 0.15
mmol, 12%), 2,3,4,6-tetra-O-acetyl-1,5-anhydro-
mannitol42,43 (Rf 0.33, 50.0 mg, 0.151 mmol, 12%),
2,3,4,6-tetra-O-acetyl- -mannopyranose (Rf 0.20, 29
D-mannopyra-
D
-
D
mg, 0.083 mmol, 6%), and upon elution with 20:1
EtOAc–ethanol, 13 (Rf 0.57, 454 mg, 0.915 mmol,
68%).

J.-P. Praly et al. / Carbohydrate Research 337 (2002) 1623–1632
1631
JCH-3,P 6.1 Hz, OCH2CH3), 16.7 (d, JCH-3,P 6.1 Hz,
OCH2CH3); 31P NMR (81 MHz, CDCl3): l 34.67;
Anal. Calcd for C20H33O13P: C, 46.88; H, 6.49; P, 6.04.
Found: C, 46.94; H, 6.54; P, 6.03.
After adding n-Bu3SnH (565 mL, 2 mmol, 3 equiv)
irradiation was applied for 30 min. TLC monitoring
showed the formation of 6 (Rf 0.7 in 1:1 EtOAc–
petroleum ether), 3 (Rf 0.65 in 1:1 EtOAc–petroleum
ether), 7 (Rf 0.48 in 1:1 EtOAc–petroleum ether), 16 (Rf
0.48 in 20:1 EtOAc–EtOH), and 12b as the major
product (Rf 0.46 in 20:1 EtOAc–EtOH). The reaction
mixture was filtered through a bed of Celite, concen-
trated under reduced pressure, and the residue was
dissolved in MeCN (30 mL). The resulting organic
phase was washed with hexanes (3×30 mL) in order to
remove the tin compounds. Then, the MeCN phase was
concentrated under reduced pressure and the residue
was applied to a column of silica gel developed follow-
ing a gradient technique (1:1.5–1:1 EtOAc–petroleum
ether then 40:1–20:1 EtOAc–EtOH) to afford the fol-
lowing compounds: 6 (3.5 mg, 1%), 3 (65.7 mg, 29%, no
2,3,4,6-Tetra-O-acetyl-1,5-anhydro- -glucitol (3).—
D
To a mixture of 1 (200 mg, 0.486 mmol) and
NaBH3CN (60.3 mg, 1 mmol, 2 equiv) in tert-butanol
(8 mL) in a quartz tube (ꢀ94 mL volume, 27 mm
external diameter) and under argon, thiophenol (5 mL,
0.05 mmol, 0.1 equiv) was added and the medium was
irradiated with filtered UV light for 1.75 h. TLC Mon-
itoring showed the conversion of the starting bromide
(Rf 0.70) into 3–4 (Rf 0.65) and 7 (Rf 0.34 in 1:1
EtOAc–petroleum ether). After the mixture was con-
centrated under reduced pressure, the oily residue dis-
solved in CH2Cl2 (30 mL) was washed with water
(3×30 mL). After workup, the crude product (166.4
1
1
rearranged product detected by H NMR), 7 (21.6 mg,
mg) was shown by H NMR to contain 3, 4, and 7 in
9%), 16 (24.7 mg, 7%), and 12b (126.6 mg, 38%).
Compound 12b: white solid, mp: 70–74 °C (CH2Cl2–
a ꢀ97:0.5:2.5 ratio (estimated weight of 3–4: ꢀ139
mg). Concentration of homogeneous fractions obtained
from a column of chromatography eluted with 1.5:3
EtOAc–petroleum ether afforded 3 in admixture with 4
1
cyclohexane); [h]2D5 −11° (c 1, acetone); H NMR (500
MHz, CDCl3): l 5.08 (t, 1 H, J4,5 9.7 Hz, H-5), 4.96 (t,
1 H, J5,6 9.8 Hz, H-6), 4.79 (t, 1 H, J3,4 9.2 Hz, H-4),
4.14 (dd, 1 H, J7,8a 4.9, J8a,8b 12.3 Hz, H-8a), 4.15–3.9
(m, 5 H, H-8b, OCH2CH3), 3.55 (ddd, 1 H, J6,7 9.8,
J7,8b 2.4 Hz, H-7), 3.41 (ddd, 1 H, J2a,3 2.2, J2b,3 9.2 Hz,
H-3), 2.1–1.5 (m, 4 H, H-1a, H-1b, H-2a, H-2b), 2.02,
1.98, 1.96, 1.93 (4s, 12 H, acetyl), 1.23 (t, 6 H, JCH-2,CH-3
7.0 Hz, OCH2CH3); 13C NMR (125 MHz, CDCl3): l
171.0, 170.7, 170.0, 169.9 (CꢁO), 77.1 (d, JC-3,P 15.4 Hz,
C-3), 75.8 (C-7), 74.4 (C-5), 71.6 (C-4), 68.7 (C-6), 62.4
1
(128.3 mg, 0.386 mmol, 79%, 3–4 ratio: 98:2 by H
NMR), and 7 (11.1 mg, 0.032 mmol, 7%). Compound 3
crystallized as colorless crystals (126.6 mg, 0.381 mmol,
78%), mp 68–70 °C (1:3 CH2Cl2–petroleum ether),
1
lit.24 68–69 °C; H NMR data in agreement with Ref.
25 but not with Ref. 44; 13C NMR (50 MHz, CDCl3):45
l 170.4, 170.1, 169.5, 169.3 (CꢁO), 76.2 (C-5), 73.5
(C-3), 68.7 C-2), 68.2 (C-4), 66.6 (C-1), 62.0 (C-6),
20.55, 20.5 (2 C), 20.4 (acetyl).
2
2
(C-8), 61.8 (d, JC,P 6.7 Hz, OCH2CH3), 61.8 (d, JC,P
1,3,4,6-Tetra-O-acetyl-2-deoxy-h-D-arabino-hexopy-
2
6.7 Hz, OCH2CH3), 24.7 (d, JC-2,P 3.9 Hz, C-2), 21.0
ranose (4).—A mixture of 1 (200 mg, 0.486 mmol) and
NaBH3CN (60.3 mg, 1 mmol, 2 equiv) in tert-butanol
(8 mL) was irradiated with filtered UV light for 35 min
and treated as before. The crude mixture (133.2 mg,
1
(d, JC,P 143 Hz, C-1), 21.0, 20.9, 20.84, 20.82 (4s,
acetyl), 16.6 (d, 2 C, JC,P 6.2 Hz, OCH2CH3); 31P
3
NMR (81 MHz, CDCl3): l 31.70; MS (FAB+, glyc-
erol+LiCl): m/z 503 [M+Li]+; Anal. Calcd for
C20H33O12P: C, 48.39; H, 6.70; P, 6.24. Found: C,
48.54; H, 6.87; P, 5.60.
1
82%) which was shown by H NMR to contain only 3
and 4 in a ratio B5:95 crystallized from acetone–cy-
clohexane to afford 4 (104.6 mg, 65%) as colorless
prisms, mp 108–110 °C lit.28 109–110 °C; 13C NMR (50
MHz, CDCl3):45 l 170.7, 170.3, 169.7, 168.9 (CꢁO),
90.9 (C-1), 70.2 (C-5), 68.7 (C-4), 68.5 (C-3), 62.0 (C-6),
33.9 (C-2), 21.0, 20.9, 20.7, 20.7 (acetyl).
Compound 16: white solid, mp: 122 °C (CH2Cl2–cy-
1
clohexane); [h]D25 +13° (c 1, CH2Cl2); H NMR (500
MHz, CDCl3): l 6.66 (s, 1 H, OH), 5.48 (t, 1 H, J4,5 9.6
Hz, H-5), 5.07 (t, 1 H, J5,6 9.6 Hz, H-6), 4.88 (d, 1 H,
J3,4 9.8 Hz, H-4), 4.29–4.24 (m, 2 H, H-7, H-8a),
4.18–4.03 (m, 5 H, H-8b, 2 OCH2CH3), ꢀ1.95–1.75
(m, 2 H, H-1a, H-1b), ꢀ1.75–1.55 (m, 2 H, H-2a,
H-2b), 2.09, 2.08, 2.02, 1.97 (4s, 12 H, acetyl), 1.33 (dt,
Acknowledgements
4
6 H, JCH-2,CH-3 7.1, JCH-3,P 5.1 Hz, 2 OCH2CH3); 13C
A stipend (to B.B. Qin) in the frame of a collabora-
tion between ‘la Re´gion Rhoˆne-Alpes’ and Shangha¨ı
City (P.R. China) is gratefully acknowledged (TEM-
PRA program), as well as financial supports, from ‘le
Ministe`re de l’Enseignement Supe´rieur, de la Recherche
et de la Technologie’ (to I.B.) and from the Islamic
Republic of Iran (to A.S.A.).
NMR (125 MHz, CDCl3): l 171.1, 170.6, 170.4, 170.1
3
(CꢁO), 96.4 (d, JC-3,P 6.5 Hz, C-3), 74.1 (C-4), 72.0
2
(C-5), 69.3 (C-6), 68.3 (C-7), 62.8 (d, JC,P 6.7 Hz,
2
OCH2CH3), 62.6 (d, JC,P 6.7 Hz, OCH2CH3), 62.5
2
(C-8), 30.7 (d, JC-2,P 4 Hz, C-2), 21.2, 21.1, 21.1, 21.0
1
(4s, acetyl), 18.6 (d, JC-1,P 141.1 Hz, C-1), 16.8 (d,

1632
J.-P. Praly et al. / Carbohydrate Research 337 (2002) 1623–1632
22. Readman, S. K.; Marsden, S. P.; Hodgson, A. Synlett
References
2000, 1628–1630 and references therein.
23. Praly, J.-P.; El Kharraf, Z.; Descotes, G. Carbohydr. Res.
1992, 232, 117–123.
24. Praly, J.-P. Tetrahedron Lett. 1983, 24, 3075–3078.
25. Bennek, J. A.; Gray, G. R. J. Org. Chem. 1987, 52,
892–897.
1. Chapleur, Y., Ed. Carbohydrate Mimics Concepts and
Methods; Wiley-VCH: Weinheim, 1998.
2. Nicotra, F. in Ref. 1, pp. 67–85 and references therein.
3. Engel, R. Chem. Re6. 1977, 77, 349–367.
4. Dondoni, A.; Marra, A.; Pasti, C. Tetrahedron: Asymme-
try 2000, 11, 305–317.
5. Chmielewski, M.; BeMiller, J. N.; Cerretti, D. P. Carbo-
hydr. Res. 1981, 97, C1–C4.
6. Meuwly, R.; Vasella, A. Hel6. Chim. Acta 1986, 69,
26. Monneret, C.; Choay, P. Carbohydr. Res. 1981, 96, 299–
305.
27. Giese, B.; Gilges, S.; Gro¨ninger, K. S.; Lamberth, C.;
Witzel, T. Liebigs Ann. Chem. 1988, 615–617.
28. Giese, B.; Gro¨ninger, K. S. Org. Synth. 1990, 69, 66–71.
29. Lemieux, R. U.; Hindsgaul, O. Carbohydr. Res. 1980, 82,
195–206.
30. Jimenez-Barbero, J.; Bernabe, M.; Martin-Lomas, M.
Tetrahedron 1988, 44, 1441–1447.
31. Chen, G.-R.; Fei, Z. B.; Huang, X.-T.; Xie, Y.-Y.; Xu,
J.-L.; Gola, J.; Steng, M.; Praly, J.-P. Eur. J. Org. Chem.
2001, 2939–2946 and references therein.
32. Beckwith, A. L. J.; Crich, D.; Duggan, P. J.; Yao, Q.
Chem. Re6. 1997, 97, 3273–3312.
33. Dick, W. E., Jr.; Weisleder, D.; Hodge, J. E. Carbohydr.
Res. 1972, 23, 229–242.
34. Betaneli, V. I.; Ovchinnikov, M. V.; Backinowsky, L. V.;
Kochetkov, N. K. Carbohydr. Res. 1982, 107, 285–291.
35. Wang, W.; Kong, F. Carbohydr. Res. 1999, 315, 117–
127.
751–760.
7. Mirza, S.; Vasella, A. Hel6. Chim. Acta 1984, 67, 1562–
1567.
8. Junker, H.-D.; Fessner, W.-D. Tetrahedron Lett. 1998,
39, 269–272.
9. Junker, H.-D.; Phung, N.; Fessner, W.-D. Tetrahedron
Lett. 1999, 40, 7063–7066.
10. Barton, D. H. R.; Ge´ro, S. D.; Quiclet-Sire, B.; Samadi,
M. Tetrahedron 1992, 48, 1627–1636.
11. Gaurat, O.; Xie, J.; Vale´ry, J.-M. Tetrahedron Lett. 2000,
41, 1187–1189.
12. Guanti, G.; Banfi, L.; Zannetti, M. T. Tetrahedron Lett.
2000, 41, 3181–3185.
13. Wang, R.; Steensma, D. H.; Takaoka, Y.; Yun, J. W.;
Kajimoto, T.; Wong, C.-H. Bioorg. Med. Chem. 1997, 5,
661–672 and references therein.
36. Giese, B; Kopping, B.; Chatgilialoglu, C. Tetrahedron
Lett. 1989, 30, 681–684.
14. Sears, P.; Wong, C.-H. Angew. Chem. Int. Ed. 1999, 38,
2300–2324 and references therein.
37. Bonner, W. A. J. Org. Chem. 1961, 26, 908–911.
38. Sabesan, S.; Neira, S. J. Org. Chem. 1991, 56, 5468–5472.
39. Roberts, B. P. Chem. Soc. Re6. 1999, 28, 25–35 and
references therein.
40. Crich, D.; Yao, Q. J. Org. Chem. 1995, 60, 84–88.
41. Pelter, A.; Pardasani, R. T.; Pardasani, P. Tetrahedron
2000, 56, 7339–7369.
42. Kocienski, P.; Pant, C. Carbohydr. Res. 1982, 110, 330–
332.
43. Haque, M. B.; Roberts, B. P.; Tocher, D. A. J. Chem.
Soc., Perkin Trans. 1 1998, 2881–2889.
15. Praly, J.-P.; Brard, L.; Descotes, G.; Toupet, L. Tetra-
hedron 1989, 45, 4141–4152.
16. Praly, J.-P.; Brendle´, J.-C.; Klett, J.; Pe´query, F. C.R.
Acad. Sci. Paris, Chem. 2001, 4, 611–617.
17. Giese, B. Radicals in Organic Synthesis: Formation of
Carbon–Carbon Bonds; Pergamon Press: Oxford, 1986.
18. Praly, J.-P. Ad6. Carbohydr. Chem. Biochem. 2000, 56,
65–151 and references therein.
19. Praly, J.-P.; Chen, G.-R.; Gola, J.; Hetzer, G. Eur. J.
Org. Chem. 2000, 2831–2838.
20. Giese, B.; Dupuis, J.; Leising, M.; Nix, M.; Lindner, H.
J. Carbohydr. Res. 1987, 171, 329–341.
44. Beckwith, A. L. J.; Duggan, P. J. J. Chem. Soc., Perkin
Trans. 1 1993, 1673–1679.
45. Bock, K.; Pedersen, C. Ad6. Carbohydr. Chem. Biochem.
1983, 41, 27–66 and references therein.
21. Gotanda, K.; Matsugi, M.; Suemura, M.; Ohira, C.;
Sano, A.; Oka, M.; Kita, Y. Tetrahedron 1999, 55,
10315–10324 and references therein.