Stereocontrolled allylation of 2-amino-2-deoxy sugar derivatives by a free-radical procedure

Article abstract of DOI:10.1016/S0008-6215(98)00141-4

Preparative routes for anomerically specific 1-C-allylation of 2-amino-2-deoxy sugars have been evaluated in a comparative study of various N-substituents and aglycons as precursors for glycosyl radicals that effectively capture an allyl group from allylt

Full text of DOI:10.1016/S0008-6215(98)00141-4

Carbohydrate Research 309 (1998) 319±330
Stereocontrolled allylation of 2-amino-2-deoxy
sugar derivatives by a free-radical procedure 1
Jingrong Cui 2, Derek Horton 3,*
Department of Chemistry, The Ohio State University, Columbus, Ohio 43210, USA
Received 13 March 1998; accepted 26 May 1998
Abstract
Preparative routes for anomerically speci®c 1-C-allylation of 2-amino-2-deoxy sugars have been
evaluated in a comparative study of various N-substituents and aglycons as precursors for glycosyl
radicals that e€ectively capture an allyl group from allyltributyltin. The crystalline triacetate 4 of
3-(2-acetamido-2-deoxy-ꢀ-d-glucopyranosyl)-1-propene (6) was obtained in 70% yield when 2-
acetamido-3,4,6-tri-O-acetyl-2-deoxy-ꢀ-d-glucopyranosyl chloride (1) was treated with allyl-
tributyltin under free-radical conditions, whereas the corresponding bromide 3 led only to an
oxazolidine derivative; the ꢁ-1-ethylxanthate analogue of 1 gave 4, but in only 25% yield. The 2-
tri¯uoroacetamido 1-bromide analogue of 1 was also an e€ective radical source, giving the 2-tri-
¯uoroacetamido analogue 8 of 4 in 60% yield. The free amino analogue 7 of 4 was conveniently
obtained via the 2-p-methoxybenzylideneamino 1-bromide analogue of 1. Use of 3,4,6-tri-O-
acetyl-2-deoxy-2-phthalimido-ꢁ-d-glucopyranosyl bromide as radical precursor allowed stereo-
speci®c access to ꢁ-1-C-allyl derivatives of the amino sugar. The crystalline galactosamine analo-
gue 12 of 4 was obtained by using the galacto analogue of chloride 1, but the corresponding manno
chloride gave only an oxazoline product. The 1-C-allylated amino sugar derivatives are con-
formationally more mobile than derivatives not having a 1-C-linked substituent. # 1998 Elsevier
Science Ltd. All rights reserved
Keywords: C-Glycosyl derivatives; 2-Amino sugars; Free-radical reactions; Glycosyl radicals
1. Introduction
anomeric oxygen atom of a glycoside is replaced by
a carbon atom, have received much attention dur-
ing the past decade. They occur as subunits in a
variety of natural products [2] and act as enzyme
inhibitors [3]. Because they are stable toward both
acidic and enzymic hydrolysis, there is high interest
in the use of C-linked disaccharides for enzymic
and metabolic studies [4]. C-Linked glycosyl com-
pounds also serve as useful chiral synthons for
organic synthesis [5]. A wide range of methods for
carbon±carbon bond formation at the anomeric
carbon of neutral sugars has been developed [6±8].
1-C-Linked glycosyl derivatives (often termed,
incorrectly, `C-glycosides'), formed when the
* Corresponding author. Fax: 202-885-1095;
e-mail: carbchm@american.edu
1
For preliminary reports on this work, see ref [1].
Current address: Corvas International, 3030 Science
2
Park Road, San Diego, CA 92121, USA.
3
Current address: Department of Chemistry, The
American University, 4400 Massachusetts Ave., NW,
Washington, DC 20016, USA.
0008-6215/98/$Ðsee front matter # 1998 Elsevier Science Ltd. All rights reserved
PII S0008-6215(98)00141-4

320
J. Cui, D. Horton/Carbohydrate Research 309 (1998) 319±330
The most important of these include the Lewis
acid-catalyzed nucleophilic addition [6] and radical-
promoted addition [7] to the anomeric center of an
appropriately activated carbohydrate derivative.
The importance of amino sugars as constituents
of aminoglycoside antibiotics, antigenic determi-
nants, glycoproteins, and glycolipids is well recog-
nized. Analogues linked through carbon at C-1
may constitute useful glycomimetics, stable to gly-
coprocessing enzymes. However, synthetic access
to these structures is limited by the diculty of
preparing such carbon-linked analogues of 2-
amino-2-deoxy sugars. The Lewis acid-catalyzed
C-glycosylation of suitably activated sugar deriva-
tives fails for 2-amino-2-deoxy sugars [6b], pre-
sumably because the nitrogen atom binds the acid
catalyst. Few examples of successful synthesis of 2-
amino-2-deoxy C-glycosyl derivatives have been
reported [9]. The ®rst such instance, ethyl 2-(2-
acetamido-4,6-O-benzylidene-2-deoxy-ꢀ-d-gluco-
pyranosyl)acetate, was prepared through a Wittig
reaction of 2-acetamido-4,6-O-benzylidene-2-
deoxy-ꢀ-d-glucopyranose with (ethoxycarbonyl-
methylene)triphenylphosphorane, followed by a
Michael addition reaction [9a,b]. Nicotra and co-
workers reported that an ꢀ-C-glycosyl derivative of
2-amino-2-deoxy-d-glucose can be produced ste-
reoselectively through the reaction of (2,3,5-tri-
The methods published thus far for the synthesis
of 2-amino-2-deoxy-C-glycosyl compounds mostly
su€er the disadvantages of several steps, expensive
starting materials, and low stereoselectivity. We
®rst reported in 1992 the stereocontrolled ꢀ- and
ꢁ-allylation of 2-amino-2-deoxy sugars at the
anomeric center through a free-radical allylation
reaction [1a]. The free-radical allylation reaction of
suitably activated 2-amino-2-deoxy sugars has the
advantages of few steps, high stereoselectivities,
and the use of inexpensive and abundant natural
amino sugars. Bertozzi and co-workers published
related experimental results for the synthesis of ꢁ-
C-glycosyl derivatives of N-acetylglucosamine in
1996 [9j].
In this paper we provide a full report of our
research [1] on the introduction, highly stereo-
selectively, of a C±C bond at the anomeric center
of 2-amino-2-deoxy sugar derivatives by a free-
radical procedure. The course of the free-radical
reaction, and the in¯uence on product yield and
stereochemical outcome of di€erent protecting
groups on the amino function and substituents at
the anomeric center, is evaluated in detail.
2. Results and discussion
O-benzyl-d-arabinofuranosyl)benzylamine
with
The stereocontrolled synthesis of 3-(2-ace-
tamido-3,4,6-tri-O-acetyl-2-deoxy-ꢀ-d-glucopyra-
nosyl)-1-propene (4) was achieved by treating 2-
acetamido-3,4,6-tri-O-acetyl-2-deoxy-ꢀ-d-gluco-
pyranosyl chloride [10] (1) with allyltributyltin
(3 mol equiv) in the presence of the radical initiator
AIBN (0.15 mol equiv) in dry toluene for 8 h at
vinylmagnesium bromide and subsequent mercur-
iocyclization of the open-chain product thereby
obtained [9c]. A di€erent approach has been
described that involves the hetero [4+2] cycload-
dition of azo compounds with a 1-alkylglucal and
subsequent conversions [9d]. Bertozzi and
Bednarski reported the synthesis of 2-azido-2-
deoxy-ꢀ-C-glycosyl compounds by the direct Lewis
acid-catalyzed addition of alkylsilanes to 2-azido-
2-deoxyglycosyl nitrates [9e] (compare ref. [9g]).
A C-2 oxidation, oximation, reduction sequence
performed on a C-glycosyl derivative of a neutral
sugar provides access to C-glycosyl derivatives of 2-
amino-2-deoxy-d-glucose and -mannose, as recently
reported by Nicotra and co-workers [9k]. Coupling
of 2-acetamido-3,4,6-tri-O-acetyl-2-deoxy-ꢀ-d-glu-
copyranosyl chloride with potassium diethylmalo-
nate was shown by Kim and Hollingsworth [9h] to
give a ꢁ-C-alkylated product in 21% yield. A novel
and potentially generally useful approach involves
reaction of a lithiated dianion derived from 2-acet-
amido-3,4,6-tri-O-benzyl-2-deoxy-d-glucopyranose
with various electrophiles [9i].
ꢀ
85 C under argon. Puri®cation on a column of
silica gel provided 4 as a white solid in 67% yield.
Recrystallization from 3:1 hexanes±ethyl acetate
ꢀ
a€orded white needles having mp 109±110 C, [ꢀ]d
+47ꢀ in chloroform. The ꢁ anomer of 4 was
detected by NMR methods, formed in 2% yield,
although it could not be obtained pure through
conventional recrystallization and chromatog-
raphy of the mother liquor. Polymerization was
the major competing reaction and a polymeric
material was obtained in 24% yield. The mass
spectrum of the polymer showed peaks for the
mass of a dimer and a trimer of the C-glycosyl
derivative 4. The oxazoline 5 was a side-product,
isolated in 5% yield.
The preparation was readily conducted on a 10-
gram scale. The best yield was obtained when the

J. Cui, D. Horton/Carbohydrate Research 309 (1998) 319±330
321
4
exclusively adopt the C1 conformation, but exists
in conformational equilibrium [11] between the
1
4C1 and C4 conformers, with the former being
preponderant.
The in¯uence of di€erent substituents at the
anomeric center of 2-amino-2-deoxy sugars on the
course and stereoselectivity of the free-radical
reaction was investigated. 2-Acetamido-3,4,6-tri-O-
acetyl-2-deoxy-ꢁ-d-glucopyranosyl ethylxanthate
[12] (2) in toluene was treated with 2 molar
equivalents of allyltributyltin and 0.15 molar
equivalent of AIBN at 80 ꢀC overnight under
argon. The procedure a€orded the same glycosyl-
alkene 4, but in only 25% yield. Other products
could not be puri®ed and identi®ed.
The 1-ethylxanthate substituent was thus not as
e€ective as the 1-chloride group for the production
of glycosyl radicals, and it took overnight reaction
for the starting material 2 to be consumed. Pre-
cursors 1 and 2 both have the same 2-substituent,
but the ꢀ-substituted chloride and the ꢁ-substituted
xanthate each give rise to the same ꢀ-C-glycosyl
product, indicating that the stereoselectivity at the
anomeric center during the free-radical reaction is
evidently not in¯uenced by the con®guration of the
activating group on the anomeric center in the
precursor, at least when the same protecting group
was present on the amino function.
2-Acetamido-3,4,6-tri-O-acetyl-2-deoxy-ꢀ-d-gluco-
pyranosyl bromide [13] (3) was also evaluated as a
possible free-radical precursor. However, under
reaction conditions similar to those successfully
used with chloride 1 to produce the ꢀ-C-glycosyl
product 4, compound 3 gave the oxazoline deriva-
tive 5 as the major product (88% yield), and the C-
allylated compound 4 could not be found in the
reaction mixture. The excellent leaving-group
character of bromide can be expected to generate a
transient oxonium ion that is rapidly stabilized
through participation by the acetamido group,
with a 1,2-cyclization reaction leading to the oxa-
zoline product 5. In contrast, chloride is a less
e€ective leaving group, and production of an
intermediate glycosyl radical rather than an oxo-
nium ion is the favored pathway, leading to the
radical-capture product 4 as the major product,
with the oxazoline 5 being obtained in only 5%
yield.
starting material, 2-acetamido-3,4,6-tri-O-acetyl-2-
deoxy-ꢀ-d-glucopyranosyl chloride, was freshly
prepared. The yield of product 4 was increased
from 40 to 67% by increasing the amount of allyl-
tributyltin from 2 to 3 molar equivalents. When the
amount of initiator (AIBN) was decreased from
0.15 to 0.075 mol equiv, the yield was decreased
from 67 to 26%. When the volume of toluene sol-
vent was doubled, the yield decreased from 67 to
49%.
Assignment of the structure of 4 was based on its
satisfactory elemental analysis, NMR spectra, spe-
ci®c optical rotation, and mass spectrum. The ꢀ-
anomeric con®guration of 4 was assigned on the
0
0
basis of the small J1 ,2 value (4.5 Hz) and a lower
down®eld chemical shift of H-10 (at ꢂ 4.21) as
0
0
compared with the large J1 ,2 value of 10.1 Hz and
higher up®eld chemical shift of H-10 (at ꢂ 3.16) for
1
the ꢁ anomer of 4. The H NMR spectrum of 4
showed an apparent quintet (ddd, J 4.5, 4.8,
10.0 Hz) for H-10 at ꢂ 4.21 and a doubled doublet
of doublets (J 4.5, 8.2, 8.7 Hz) for H-20 at ꢂ 4.52.
Deacetylation of 4 with a catalytic amount of
sodium methoxide in methanol gave 3-(2-ace-
tamido-2-deoxy-ꢀ-d-glucopyranosyl)-1-propene (6)
in quantitative yield; recrystallized from ethanol it
0
The relatively small values of J2 ,3 , J3 ,4 , and
0
0
0
0
J4 ,5 (8.2, 6.8, and 6.8 Hz) for compound 4 in
0
C6D6 suggest that this glycosylalkene does not

322
J. Cui, D. Horton/Carbohydrate Research 309 (1998) 319±330
was ꢀobtained as white crystals having mp 204±
206 C and [ꢀ]d +109ꢀ in methanol.
Dissolution of the crude product in hot 4:1 hex-
anes±ethyl acetate and collecting the initially pre-
cipitated syrup a€orded pure compound 8 as a
colorless syrup, which had [ꢀ]d +22ꢀ in chloro-
form; the ꢁ anomer of 8 remained in solution along
with additional 8.
The in¯uence of the amino protecting group
on the free-radical reaction was studied with
three additional N-substituents. Treatment of
3,4,6-tri-O-acetyl-2-p-methoxybenzylideneamino-
2-deoxy-ꢀ-d-glucopyranosyl bromide [14] with
allyltributyltin (3 mol equiv) and AIBN (0.30 mol
1
The H NMR spectrum of 8 in C6D6 showed
1
quite unusual H±1H coupling constants for the
ꢀ
0 0
hydrogen atoms in the tetrahydropyran ring (J1 ,2
equiv) for 24 h at 85 C under argon a€orded 3-
0
0
0
0
0
0
(3,4,6-tri-O-acetyl-2-deoxy-2-p-methoxybenzylidene-
amino-ꢀ-d-glucopyranosyl)-1-propene. The N-sub-
stituent was then cleaved by hydrolysis in acidi®ed
acetone to give 3-(3,4,6-tri-O-acetyl-2-amino-2-
3.8 Hz, J2 ,3 6.0 Hz, J3 ,4 5.8 Hz, and J4 ,5 4.6 Hz).
The anomeric con®guration of 8 could not be
0
0
attributed unambiguously by the J1 ,2 coupling
constant and the [ꢀ]d value. To clarify this point,
the syrupy product 8 was fully deprotected in
saturated methanolic HCl solution and then acety-
lated with acetic anhydride and pyridine. Recrys-
tallization of the product gave white needles that
were con®rmed to be 3-(2-acetamido-3,4,6-tri-O-
acetyl-2-deoxy-ꢀ-d-glucopyranosyl)-1-propene (4)
by melting point, speci®c rotation, and also 1H and
13C NMR spectra. The ꢀ-anomeric con®guration of
8 was thus clearly demonstrated by its chemical
transformation into the known propene derivative
deoxy-ꢀ-d-glucopyranosyl)-1-propene
hydro-
chloride (7) in 45% yield as a crystalline white
solid, mp 220±222 ꢀC (dec), [ꢀ]d +59ꢀ in methanol;
the corresponding free base was a syrup. The ꢀ-
anomeric con®gurational assignment of 7 was
0
0
based on the small J1 ,2 coupling constant of
4.8 Hz. The 1H NMR spectrum of 7 showed the N-
H signal at the lowest ®eld (ꢂ 8.66) as a very broad
singlet. The H-10 signal appeared at ꢂ 4.44 as a
0
0
0
doubled doublet of doublets (J1 ,2 4.8 Hz, J1 ,3a
0
10.5 Hz, J1 ,3b 4.9 Hz) and H-20 at ꢂ 3.79 as a dou-
4. The coupling constants of J1 ,2 3.8 Hz, J2 ,3
0
0
0
0
0
0
0
0
0
0
0
0
bled doulet (J1 ,2 4.8 and J2 ,3 8.8 Hz). The cou-
pling constants of compound 7 showed that the
compound in chloroform solution exists essentially
6.0 Hz, J3 ,4 5.8 Hz, and J4 ,5 4.6 Hz suggest [1c] that
compound 8 exists as a conformational mixture [11]
in rapid equilibrium between the 4C1 and 1C4 chair
conformers, with signi®cant population of the lat-
ter, having all substituents axial except that at C-1.
Protection at the 2-position as a phthalimido
group is a well established procedure with 2-
amino-2-deoxy sugars for the introduction of a 1,2-
trans-glycosidic linkage. The in¯uence of this pro-
tecting group on the anomeric stereoselectivity in
the free-radical allylation reaction was here inves-
tigated. The reaction of 3,4,6-tri-O-acetyl-2-deoxy-
2-phthalimido-ꢁ-d-glucopyranosyl bromide [16]
with allyltributyltin (2 mol equiv) and AIBN
4
0
0
0
in the C1 chair conformation (J1 ,2 4.8 Hz, J2 ,3
0
8.8 Hz, J3 ,4 7.9 Hz, and J4 ,5 8.0 Hz).
0
0
0
0
The strongly electron-withdrawing character of
the tri¯uoromethyl group can be expected to
greatly diminish the neighboring-group participa-
tory e€ect of the amino function in a tri-
¯uoroacetyl protecting group as compared with the
acetamido group. This consideration made 3,4,6-
tri-O-acetyl-2-deoxy-2-tri¯uoroacetamido-ꢀ-d-
glucopyranosyl bromide attractive as a possible
free-radical precursor, as compared with 2-acet-
amido-3,4,6-tri-O-acetyl-2-deoxy-ꢀ-d-glucopyrano-
syl bromide (3), which produced exclusively the
oxazoline 5 through 1,2-cyclization during the
attempted free-radical allylation reaction. The
reaction of 3,4,6-tri-O-acetyl-2-deoxy-2-tri¯uoro-
acetamido-ꢀ-d-glucopyranosyl bromide [15] with
allyltributyltin (2 mol equiv) and AIBN (0.15 mol
ꢀ
(0.3 mol equiv) in dry toluene for 24 h at 85 C
under argon gave 3-(3,4,6-tri-O-acetyl-2-deoxy-2-
phthalimido-ꢁ-d-glucopyranosyl)-1-propene (9) in
40% yield. The observed ꢁ-stereospeci®city was
expected as a consequence of the bulky and
strongly participating 2-substituent. This large
blocking group at C-2 decreased the rate of the
free-radical reaction, and two portions of AIBN
(0.15 mol equiv) were added to the solution. Pur-
i®cation on a column of silica gel provided 9 as a
white solid, which could be crystallized from 3:1
hexanes±ethyl acetate toꢀa€ord very beautiful nee-
dle crystals, mp 79±81 C, [ꢀ]d +70ꢀ in chloro-
form. 1H NMR spectra of the initial fractions from
ꢀ
equiv) in dry toluene under argon for 6 h at 85 C
gave the desired product, 3-(3,4,6-tri-O-acetyl-2-
deoxy-2-tri¯uoroacetamido-ꢀ-d-glucopyranosyl)-
1-propene (8) in 60% yield, after chromatographic
puri®cation on a column of silica gel. The ꢁ
anomer of 8 was produced in about 5% yield in the
crude product, but it could not be obtained pure.

J. Cui, D. Horton/Carbohydrate Research 309 (1998) 319±330
323
the column showed exclusively the ꢁ anomer. The
1H NMR spectrum of the needle crystals showed
that the ꢁ-allyl C-glycosyl derivative 9 co-crystal-
lized with the solvents ethyl acetate and hexanes in
the ratio of 1:0.25:0.2. The solvents could not be
removed under vacuum at room temperature, and
the crystals were very stable in air. Even on drying
under diminished pressure at room temperature for
5 days, the ratio of 9 to solvents did not change.
free-radical allylation reaction of 2-amino-2-deoxy-
d-glucopyranosyl derivatives with allyltributyltin.
2-Amino-2-deoxy-d-glucopyranosyl
derivatives
having acetyl, tri¯uoroacetyl, and p-methoxy-
benzylidene protecting groups on the amino func-
tion favored formation of ꢀ-allylation products
during the free-radical reaction. The ꢀ:ꢁ stereo-
selectivity for the N-acetyl protecting group was
34:1, and 11:1 for the N-tri¯uoroacetyl derivative;
the N-p-methoxybenzylidene derivative gave
exclusively the ꢀ product. The 2-phthalimido-2-
deoxy-d-glucopyranosyl derivative gave exclusively
the ꢁ-allylation product. The ®nal steric outcome
was not in¯uenced by the con®guration of the
anomeric substituent. For 2-acetamido-2-deoxy-d-
glucopyranosyl derivatives, both the ꢀ-substituted
chloride 1 and the ꢁ-substituted xanthate deriva-
tive 2 gave preponderantly the ꢀ-allylation prod-
ucts. The nature of the anomeric substituent
in¯uenced the rate of the free-radical reaction. It
took only 6 h for 3,4,6-tri-O-acetyl-2-deoxy-2-tri-
¯uoroacetamido-ꢀ-d-glucopyranosyl bromide to
react completely, whereas the chloride derivative 1
required 8 h, and the xanthate derivative 2 required
overnight reaction. The N-2 protecting group also
in¯uenced the rate of reaction. For 3,4,5-tri-O-acetyl-
2-deoxy-2-phthalimido-ꢁ-d-glucopyranosyl bromide,
it was necessary to add AIBN twice, and it took 24 h
for all of the starting material to be consumed.
ꢀ
Eventually, the solvents were removed at 50 C
under vacuum to leave a glassy solid, which was
1
demonstrated to be pure 9 by its H NMR spec-
trum and elemental analyses.
A 1,2-elimination process was the major side-
reaction in the preparation of 9, and led to the
glycal 10 in 16% yield. Because the elimination
reaction generated hydrogen bromide in the sys-
tem, the acetylated 2-deoxy-2-phthalimido-d-glu-
copyranosyl radical could also abstract a hydrogen
radical from HBr, leading to the anhydro alditol
11, formed in 8% yield. The glycal 10 and the
alditol 11 had very similar Rf values. Recrystalli-
zation gave the pure white crystalline 10, having
mp 126±127 ꢀC and [ꢀ]d 11ꢀ in chloroform.
Ogawa et al. [17] reported 10 as a syrup and [ꢀ]d
15ꢀ. A pure preparation of 11 could not be
obtained.
The ꢁ-anomeric con®guration of 9 was assigned
0
1
on the basis of H NMR data for H-2 , which
0
0
resonated as a triplet at ꢂ 4.45 and showed J1 ,2
The stereoselectivity in reactions of glycosyl
radicals can be expected to be controlled by both
electronic and steric e€ects. According to the ESR
studies of Giese [20], d-glycopyranosyl radicals
may exist partially in the B2,5 boat conformation
and partially in the 4C1 chair conformation of their
precursors. Glycosyl radicals are fundamentally p-
radicals and the unpaired electron occupies an
orbital that has mainly p-character. The stereo-
electronic e€ect makes attack from the ꢀ-face to
0
0
0
0
0
10.1 Hz. The large values of J2 ,3 , J3 ,4 and J4 ,5
0
(9.9, 9.3, and 10 Hz) were in accordance with the
4C1 d-glycopyranosyl chair conformation. The ꢁ-
C-glycosylpropene 9 showed an abnormally high
dextrorotation ([ꢀ]d +70ꢀ in chloroform). Such
rotatory anomalies are known [18] in glycosides
having certain aryl groups at C-2.
Although the acetylated 2-acetamido-2-deoxy-ꢀ-
d-glucopyranosyl chloride (1) reacted with allyl-
tributyltin in the presence of AIBN to produce the
allylated product 4 in 67% yield, 3,4,5-tri-O-acetyl-
2-deoxy-2-phthalimido-ꢁ-d-glucopyranosyl chlor-
ide [19], in contrast to the bromide analogue, did
not give any allylation product under the same
free-radical allylation conditions. It appears that
e€ective formation of an intermediate glycosyl
radical requires a suitable combination of both the
anomeric substituent and also the substituent
group at C-2.
4
the glycosyl radical in the C1 chair conformation
more favorable, because interaction is maintained
in the transition state between the unpaired elec-
tron in the radical and the lone-pair electrons on
the ring oxygen atom, and the ꢀ-product is
obtained exclusively. The stereoselectivity is rela-
tively low for glycosyl radicals in the B2,5 boat
conformation, even though ꢀ-attack is favored on
stereoelectronic grounds. ꢁ-Attack is more favored
for a glycosyl radical in the 1,4B boat conformation.
There is only a small energy di€erence between the
B2,5 boat and the 1,4B boat, and this would account
for the observed stereoselectivity [20].
It is evident that both the anomeric substituent
and the C-2 substituent play a major role in
determining the course and stereoselectivity of the

324
J. Cui, D. Horton/Carbohydrate Research 309 (1998) 319±330
Direct ESR investigation of 2-amino-2-deoxy-d-
The reaction of 2-acetamido-3,4,6-tri-O-acetyl-2-
deoxy-ꢀ-d-mannopyranosyl chloride (prepared
from 2-acetamido-2-deoxy-d-mannose by the gen-
eral procedure [10]) with allyltributyltin (3 mol
equiv) and AIBN (0.15 mol equiv) in dry toluene
did not give the desired allylation product, and
instead the known [21] oxazoline 14 was obtained
glucopyranosyl radicals was not pursued. Based on
Giese's results, it was concluded that the acetylated
2-deoxy-2-p-methoxybenzylideneamino-d-gluco-
4
pyranosyl radical should adopt the C1 chair con-
formation, because of the steric strain caused by
the large protecting group, and lead exclusively to
the ꢀ-allylation product. The acetylated 2-ace-
tamido-2-deoxy-d-glucopyranosyl and 2-tri¯uoro-
ꢀ
in 82% yield, mp 127±128 C (from ether), [ꢀ]d
31ꢀ in chloroform. Evidently the axial 2-ace-
tamido group exerts a stronger neighboring-group
participation e€ect than the equatorial group in the
gluco and galacto series. The 1,2-ring-closure reac-
tion is consequently much faster than the free-radi-
cal reaction, and so only the neighboring-group
participating product, oxazoline 14, is formed.
acetamido-2-deoxy-d-glucopyranosyl
radicals
should exist in the B2,5 boat conformation and lead
to an ꢀ,ꢁ mixture of products, with the ꢀ-product
being preponderant. It is considered that the
acetylated 2-phthalimido-2-deoxy-d-glucopyrano-
4
syl radical adopts the C1 chair conformation, and
the observed complete ꢁ-selectivity is attributable
mainly to the steric e€ect. ꢀ-Attack is totally
impeded by the rigid 2-phthalimido protecting
group and the ꢁ-allylation product is formed
exclusively.
In conclusion, a carbon±carbon bond has been
successfully introduced with high stereoselectivity
at the anomeric center of activated 2-amino-2-
deoxy-d-glycopyranosyl derivatives via a free-radi-
cal process. The stereoselectivity is controlled by
the substituents at C-2. The acetyl, tri¯uoroacetyl
and p-methoxybenzylidene protecting groups on
the 2-amino function a€ord ꢀ-allylation products,
whereas the phthalimido function provides a ꢁ-
allylation product. The stereoselectivity is not
in¯uenced by the con®guration of the substituent
at the anomeric center. The rate of the free-radical
reaction is controlled by the nature of the 2-sub-
stituent and the substituent at the anomeric center.
The ꢀ-C-allylated glucopyranosyl compounds in
solution have [1c] an appreciable proportion of
The foregoing results show that the allyl group can
be successfully introduced at the anomeric center of
2-amino-2-deoxy-d-glucopyranose
derivatives
through free-radical reaction using allyl-
a
tributyltin, with reasonable yields and short steps.
The generality of this methodology was further
evaluated with 2-amino-2-deoxy-d-galactopyra-
nose and 2-amino-2-deoxy-d-mannopyranose.
Treatment of 2-acetamido-2-deoxy-d-galactose
with acetyl chloride by the procedure [10] that
converts the gluco analogue exclusively into the
acetylated glycosyl chloride leads to 2-acetamido-
3,4,6-tri-O-acetyl-ꢀ-d-galactopyranosyl chloride in
admixture with a smaller proportion of an oxazo-
line cyclization product 13. This mixture was used
directly in reaction with allyltributyltin and AIBN
1
4
the C4 conformers in equilibrium with the C1
conformers.
3. Experimental
ꢀ
in dry toluene for 7 h at 85 C under argon to
a€ord 3-(2-acetamido-3,4,6-tri-O-acetyl-2-deoxy-ꢀ-
d-galactopyranosyl)-1-propene (12) in 56% yield
(based on the glycosyl chloride precursor). The
oxazoline 13, which had an Rf value very close to
that of the product 12, was readily removed during
the isolation procedure through hydrolysis in acid-
i®ed acetone to a water-soluble product. Com-
pound 12 was crystallized from 2:1 hexanes±ethyl
General methods.ÐEvaporations were con-
ducted under diminished pressure. Reaction sol-
vents were puri®ed and dried by distillation as
recommended. TLC was performed on precoated
plates of Silica Gel 60F-254 (E. Merck); compo-
nents were detected by UV light and by spraying
the plates with 10% H2SO4 and subsequent heat-
ing. Flash-column chromatography was performed
on 230±240 mesh silica gel (E. Merck). Melting
points were determined in open glass capillaries in
a Thomas±Hoover apparatus, and are uncorrected.
Speci®c rotations were determined with ꢀa Perkin±
Elmer Model 241MC polarimeter at 20 C, unless
ꢀ
acetate as needles, mp 129±130 C; [ꢀ]d +81ꢀ in
chloroform. The ꢀ-anomeric con®guration of 12
0
was con®rmed by the J1 ,2 coupling constant of
0
4.8 Hz. The coupling constants between protons in
0
0
0
0
0
0
0
0
the ring (J1 ,2 4.8 Hz, J2 ,3 9.4 Hz, and J3 ,4 =J4 ,5
4
1
3.2 Hz) indicate that 12 adopts the C1 conforma-
tion in solution.
otherwise noted. H NMR and 13C NMR spectra
1
were recorded with Bruker AM-250 (250 MHz H,

J. Cui, D. Horton/Carbohydrate Research 309 (1998) 319±330
325
1
62.5 MHz 13C), AM-300 (300 MHz H, 75.5 MHz
procedure was followed. A suspension of 2-acet-
amido-3,4,6-tri-O-acetyl-2-deoxy-ꢀ-d-glucopyranosyl
chloride [10] (1, 10.97 g, 30 mmol), allyltributyltin
(29 mL, 91 mmol) and AIBN (738 mg, 4.5 mmol) in
dry toluene (60 mL) was degassedꢀfor 10 min, and
the mixture was then stirred at 85 C under Ar for
8 h. TLC showed the complete absence of starting
material (Rf 0.64, EtOAc) and the presence of three
new spots having Rf 0.49, 0.46, and 0.08 (EtOAc).
The crude products were chromatographed on a
column of silica gel, eluting with 4:1 EtOAc±hex-
anes to give the product 4 as a white solid (7.46 g,
67%). This solid was crystallized from 1:3 EtOAc±
hexanes, and further recrystallization a€orded the
13C), and AM-500 (500 MHz H, 125 MHz 13C)
1
spectrometers. Chemical shifts (ppm) are relative
to Me4Si as the internal standard. Splitting pat-
terns are designated: s, singlet; d, doublet; dd,
double doublet; t, triplet; m, multiplet. Spectra
with the AM-500 instrument at The Ohio State
University Instrument Center were recorded by Dr.
C. E. Cottrell. All signal assignments were veri®ed
by H±1H and 13C±1H correlation spectra. Fast-
1
atom-bombardment (FAB) mass spectra were
recorded at The Ohio State University Instrument
Center with Kratos VG 70-250S mass spectro-
meters by D. Chang. Elemental analyses were per-
formed by Atlantic Microlabs, Atlanta, Georgia.
General procedure for free-radical allylation of
activated 2-amino-2-deoxy sugars.ÐA suspension
or solution of the activated 2-amino-2-deoxy sugar
derivative (1 mol equiv), allyltributyltin (2±3 mol
equiv), and AIBN (0.15 mol equiv) in dry toluene
was degassed for 10 min and the mixture was then
ꢀ
pure product 4 as needles having mp 109±110 C;
[ꢀ]d +47ꢀ (c 1.2, CHCl3); Rf 0.49 (EtOAc); NMR
data see Tables 1±3; MS: m/z 372 (M+H)+, 330
(M allyl)+, 312 (M+H HOAc)+, 270 (M-
HOAc allyl)+. Anal. Calcd for C17H25NO8
(371.383): C, 54.98; H, 6.79; N, 3.77. Found: C,
54.82; H, 6.85; N, 3.73.
ꢀ
stirred at 85 C under Ar. A clear solution was
Evaporation of the mother liquor from the
recrystallization gave a syrup, which was demon-
strated by its 1H NMR spectrum to be a mixture of
4 and its ꢁ anomer. A total of 0.22 g of the ꢁ
anomer of 4 was present (NMR integration) in the
mother liquor (2% yield), which had the same Rf
value as 4. It was not found possible to obtain the
ꢁ anomer of 4 pure through chromatography and
obtained during the ®rst hour of heating. After the
starting material had been consumed completely,
the toluene solvent was evaporated o€ and the
residue was dissolved in MeCN. The solution was
washed three times with pentane and evaporated.
The residue was puri®ed on a column of silica gel
to provide the pure allylation product.
3-(2-Acetamido-3,4,6-tri-O-acetyl-2-deoxy-a-d-
glucopyranosyl)-1-propene (4).Ð(a) From 2-acet-
amido-3,4,6-tri-O-acetyl-2-deoxy-a-d-glucopyranosyl
chloride (1). The general free-radical allylation
1
recrystallization. H NMR data for the ꢁ anomer
of 4 (300 MHz, C6D6): ꢂ 5.93 (tdd, 1 H, J1a,2
10.3 Hz, J1b,2 17.1 Hz, J2,3a=J2,3b 6.90Hz, H-2), 5.24
0
(dd, 1 H, J3 ,4 9.4 Hz, J4 ,5 9.8 Hz, H-4 ), 5.19 (d, 1 H,
0
0
0
Table 1
1H NMR chemical shifts (ꢂ) and multiplicities of 3-(2-acetamido-3,4,6-tri-O-acetyl-2-deoxy-ꢀ-d-glucopyranosyl)-1-propenea (4), 3-
(3,4,6-tri-O-acetyl-2-amino-2-deoxy-ꢀ-d-glucopyranosyl)-1-propene hydrochlorideb (7), 3-(3,4,6-tri-O-acetyl-2-deoxy-2-tri¯uoro-
acetamido-ꢀ-d-glucopyranosyl)-1-propenec (8), 3-(3,4,6-tri-O-acetyl-2-deoxy-2-phthalimido-ꢁ-d-glucopyranosyl)-1-propenea (9)
and 3-(2-acetamido-3,4,6-tri-O-acetyl-2-deoxy-ꢀ-d-galactopyranosyl)-1-propenea (12)
Compound
H-1a H-1b H-2 H-3a H-3b H-10 H-20 H-30 H-40 H-50 H-60a
H-60b N-H
Ac
4
4.98
dq
5.00 5.68 2.27
dq tdd ddd
2.07
ddd
4.21 4.52 5.19 5.08 3.88
ddd ddd dd ddd
4.35
dd
4.11
dd
5.80
d
1.70,1.62
1.60,1.58
t
7
8
5.14
dq
5.20 5.76 2.61
dq tdd
2.61
m
4.44 3.79 5.40 4.95 3.89
ddd dd dd dd ddd
4.26
dd
4.06
dd
8.66
bs
2.16,2.08
2.06
m
4.96
dq
4.99 5.61 2.25
dq tdd ddd
2.03
ddd
4.08 4.30 5.12 4.93 3.93
ddd ddd dd dd ddd
4.43
dd
4.02
dd
7.07
d
1.69,1.59
1.57
Phth
7.46
6.86
9
4.81
dq
4.84 5.73 2.17
dq tdd
2.17
m
4.53 4.44 6.08 5.31 3.41
ddd dd dd ddd
4.26
dd
4.01
dd
1.70,1.63
1.46
m
t
12
4.98
dq
5.00 5.67 2.19
dq tdd
2.00
m
4.37 4.76 5.16 5.40 3.94
ddd dd
4.36
dd
4.21
dd
5.44
d
1.67,1.66
1.63,1.50
m
m
t
m
aAt 300 MHz in C6D6.
bAt 300 MHz in CDCl3.
cAt 500 MHz in C6D6.

326
J. Cui, D. Horton/Carbohydrate Research 309 (1998) 319±330
Table 2
The ®rst-order 1H±1H coupling constants (Hz) of 3-(2-acetamido-3,4,6-tri-O-acetyl-2-deoxy-ꢀ-d-glucopyranosyl)-1-propenea (4), 3-
(3,4,6-tri-O-acetyl-2-amino-2-deoxy-ꢀ-d-glucopyranosyl)-1-propene hydrochlorideb (7), 3-(3,4,6-tri-O-acetyl-2-deoxy-2-tri¯uoro-
acetamido-ꢀ-d-glucopyranosyl)-1-propenec (8), 3-(3,4,6-tri-O-acetyl-2-deoxy-2-phthalimido-ꢁ-d-glucopyranosyl)-1-propenea (9)
and 3-(2-acetamido-3,4,6-tri-O-acetyl-2-deoxy-ꢀ-d-galactopyranosyl)-1-propenea (12)
0
0
Compound
J1a,2
J1b,2
J1a,3a
J1b,3a
J1a,3b
J1b,3b
J2,3a
J2,3b
J3a,3b
J3a,1
J3b,1
4
7
8
9
12
10.2
10.3
10.2
10.2
10.2
17.0
17.4
17.1
17.1
17.0
ꢁ1
ꢁ1
ꢁ1
ꢁ1
ꢁ1
ꢁ1
ꢁ1
ꢁ1
ꢁ1
ꢁ1
ꢁ1
ꢁ1
ꢁ1
ꢁ1
ꢁ1
ꢁ1
ꢁ1
ꢁ1
ꢁ1
ꢁ1
6.8
6.8
6.4
6.9
6.9
6.8
11.5
6.4
14.5
10.0
10.5
9.0
4.8
4.9
5.3
5.2
6.9
6.9
6.8
5.2
0
J1 ,2
0
0
J2 ,3
0
0
0
0
0
0
0
0
0
0
0
0
JNH,2
8.7
Compound
4
7
8
9
12
J3 ,4
J4 ,5
J5 ,6 a
6.5
6.0
J5 ,6 b
3.8
3.0
J6 a,6 b
11.9
12.2
11.9
12.3
11.5
4.5
4.8
3.8
10.2
4.8
8.2
8.8
6.0
10.2
9.4
6.8
7.9
5.8
9.0
3.2
6.8
8.0
4.6
10.3
3.2
7.3
4.5
7.1
4.5
2.1
5.4
9.2
8.4
aAt 300 MHz in C6D6.
bAt 300 MHz in CDCl3.
cAt 500 MHz in C6D6.
Table 3
13C NMR chemical shifts (ꢂ) of 3-(2-acetamido-3,4,6-tri-O-acetyl-2-deoxy-ꢀ-d-glucopyranosyl)-1-propenea (4), 3-(3,4,6-tri-O-acetyl-
2-amino-2-deoxy-ꢀ-d-glucopyranosyl)-1-propene hydrochlorideb (7), 3-(3,4,6-tri-O-acetyl-2-deoxy-2-tri¯uoroacetamido-ꢀ-d-gluco-
pyranosyl)-1-propenec (8), 3-(3,4,6-tri-O-acetyl-2-deoxy-2-phthalimido-ꢁ-d-glucopyranosyl)-1-propenec (9) and 3-(2-acetamido-
3,4,6-tri-O-acetyl-2-deoxy-ꢀ-d-galactopyranosyl)-1-propened (12)
Compound
C-1
C-2
C-3
C-10
C-20
C-30
C-40
C-50
C-60
4
7
8
9
12
117.56
118.67
118.00
117.67
117.38
133.26
133.93
133.20
133.12
133.42
32.05
32.91
33.71
36.94
31.38
70.86
70.42
69.22
74.73
71.09
50.47
51.78
50.02
55.21
48.74
70.01
69.65
68.87
72.38
68.16
67.85
68.96
67.36
69.47
66.80
70.58
72.25
72.45
76.22
68.77
61.58
62.23
60.88
62.28
61.25
Compound
COCH3
23.10
20.78
20.20
20.31
COCH3
20.74
20.64
19.98
20.15
COCH3
20.64
20.64
19.98
19.93
COCH3
20.64
COCF3
C=O
C=O
C=O
C=O
Phth
Phth
4
7
8
9
170.86 170.55 169.59 168.92
172.22 171.18 171.01
170.10 169.84 168.53 156.91
170.00 169.16 168.00 167.58 134.13 134.00
134.07 133.90
123.63 123.27
116.52
12
23.02
20.71
20.58
20.53
170.61 170.44 169.99 169.90
aAt 63 MHz in CDCl3.
bAt 75 MHz in CD3OD.
cAt 75 MHz in C6D6.
dAt 75 MHz in CDCl3.
0
0
0
JNH,2 10.5 Hz, N-H), 5.12 (dd, 1 H, J2 ,3 9.2 Hz, H-
1,3-oxazole (5, Rf 0.46, EtOAc) was obtained as a
syrup (0.49 g, 5%), [ꢀ]d +13ꢀ (c 1.3, CHCl3); lit.
[22]. syrup, [ꢀ]d +12ꢀ (c 1.0, CHCl3). The fraction
having Rf 0.08 (EtOAc) was evaporated to a white
solid, which was demonstrated by its mass spec-
trum to be an oligomeric mixture (2.69 g, 24%),
MS: m/z 1114 (3M+H)+, 743 (2M+H)+ (M:
molecular mass of compound 4).
(b) From 2-acetamido-3,4,6-tri-O-acetyl-2-deoxy-
b-d-glucopyranosyl ethylxanthate (2). The general
free-radical allylation procedure was followed. A
suspension of 2-acetamido-3,4,6-tri-O-acetyl-2-
30), 5.09 (d, 1 H, H-1a), 5.06 (d, 1 H, H-1b), 4.34 (dd,
1 H, J5 ,6 a 4.9 Hz, J6 a,6 b 12.2 Hz, H-60a), 4.16 (m, 2
0
0
0
0
H, H-20 and H-60b), 3.51 (ddd, 1 H, J5 ,6 b 2.3 Hz, H-
0
0
50), 3.16 (ddd, 1 H, J1 ,2 10.1 Hz, J1 ,3a 3.8 Hz, J1 ,3b
7.4 Hz, H-10), 2.32 (m, 2 H, H-3a and H-3b), 1.74,
1.73, 1.71, and 1.63 (4s, 4Â3H, 4-COCH3).
Preparation of an inseparable 10:1 mixture of 4
and its ꢁ anomer as a waxy solid has been reported
[9j].
0
0
0
0
The side-product, 4,5-dihydro-2-methyl-(3,4,6-
tri-O-acetyl-1,2-dideoxy-ꢀ-d-glucopyranoso)-[2,1-d]-

J. Cui, D. Horton/Carbohydrate Research 309 (1998) 319±330
327
deoxy-ꢁ-d-glucopyranosyl ethylxanthate [12] (2,
903 mg, 2 mmol), allyltributyltin (1.24 mL, 4 mmol),
and AIBN (49.3 mg, 0.3 mmol) in dry toluene
(4 mL) was degassed for 10 min. The mixture was
3-(3,4,6-Tri-O-acetyl-2-amino-2-deoxy-a-d-gluco-
pyranosyl)-1-propene hydrochloride (7).ÐThe gen-
eral free-radical allylation procedure was followed.
A
solution of 3,4,6-tri-O-acetyl-2-p-methoxy-
ꢀ
stirred for 24 h at 80 C under Ar. The crude pro-
benzylideneamino-2-deoxy-ꢀ-d -glucopyranosyl
bromide [14] (928 mg, 1.91 mmol), allyltributyltin
(1.80 mL, 5.8 mmol), and AIBN (66.1 mg,
0.4 mmol) in dry toluene (6 mL) was degassed for
ducts were chromatographed on a column of silica
gel (4:1 EtOAc±hexanes) to give 4 as a white solid
(184 mg, 25%).
ꢀ
(c) Attempted preparation from 2-acetamido-
3,4,6-tri-O-acetyl-2-deoxy-a-d-glucopyranosyl
bromide (3). The general free-radical allylation
procedure was followed. To a solution of 2-acet-
amido-3,4,6-tri-O-acetyl-2-deoxy-ꢀ-d-glucopyrano-
syl bromide [13] (3, 1.04 g, 2.52 mmol) and AIBN
(62.1 mg, 0.38 mmol) in dry toluene (4 mL) and
1,2-dimethoxyethane (DME, 2 mL) was added
allyltributyltin (1.56 mL, 4.88 mmol). The solu-
tion was degassed for 10 min and then heated for
10 min. The mixture was heated for 9 h at 90 C
under Ar and then cooled to room temperature. To
the solution was added additional AIBN (62 mg,
0.37 mmol). The solution ꢀwas degassed for another
10 min and heated at 80 C overnight again under
Ar. The crude products were dissolved in Et3N
(0.5 mL) and EtOAc (0.5 mL), and puri®ed on a
column of silica gel with 1:1 EtOAc±hexanes to
give a light-yellow syrupy product: 3-(3,4,6-tri-O-
acetyl-2-deoxy-2-p-methoxybenzylideneamino-ꢀ-d-
glucopyranosyl)-1-propene (387 mg, 45%). This
syrup was dissolved in acetone (4 mL) and to the
solution were added 4 drops of 6 M HCl. A white
precipitate formed immediately that was identi®ed
as compound 7: mp 220±222 ꢀC (dec); [ꢀ]d +59ꢀ (c
0.53, MeOH); NMR data see Tables 1±3; MS: m/z
330 (M Cl)+, 288 (M HCl-allyl)+, 270
(M HCl OAc)+. Anal. Calcd for C15H24ClNO7
(365.807): N, 3.83. Found: N, 4.24.
ꢀ
18 h at 80 C under Ar. TLC (EtOAc) showed a
single spot (Rf 0.44) for the oxazoline 5. The
crude product was puri®ed on a short column of
silica gel, eluting with EtOAc to a€ord com-
pound 5 as a syrup (735.6 mg, 88%).
3-(2-Acetamido-2-deoxy-a-d-glucopyranosyl)-1-
propene (6).ÐA solution of compound 4 (1.11 g,
3 mmol) in dry MeOH (15 mL) was treated with
25 wt% of NaOMe in MeOH (0.03 mL) for 10 min
at room temperature, and then made neutral with
Amberlite IR-120 (H+) resin. The resin was ®ltered
o€, washed with MeOH, and the ®ltrate was eva-
porated to a€ord a white solid (0.73 g, 100%),
which was recrystallized from EtOH to a€ord
Compound 7 was suspended in CH2Cl2 and then
washed with satd. NaHCO3 and brine, and dried
over Na2SO4. After ®ltration and evaporation,
compound 7a was obtained as a syrup; NMR data
(1 H, 300 MHz, CDCl3): ꢂ 5.81 (tdd, 1 H, J1a,2 10.1,
J1b,2 16.9, J2,3a=J2,3b 7.0 Hz, H-2), 5.16 (dq, 1 H,
J1b,3 1, J1a,1b 1 Hz, H-1b), 5.13 (dq, 1 H, J1a,3 1 Hz,
ꢀ
white crystalline 6: mp 204±206 C; [ꢀ]d +109ꢀ (c
1.0, MeOH); Rf 0.38 (2:1 EtOAc±EtOH); 1H NMR
(300 MHz, CD3SOCD3): ꢂ 7.70 (d, 1 H, J2 ,NH 7.6 Hz,
H-1a), 5.05 (dd, 1 H, J2 ,3 9.4, J3 ,4 8.8 Hz, H-30),
0
0 0 0 0
NHAc), 5.75 (tdd, 1 H, J1a,2 10.2 Hz, J1b,2 17.1 Hz,
J2,3a=J2,3b 6.9 Hz, H-2), 5.05 (bd, 1 H, H-1b), 4.97
4.89 (t, 1 H, J4 ,5 8.8 Hz, H-40), 4.24 (dd, 1 H,
0 0
J5 ,6 a, 5.7, J6 a,6 b 12.0 Hz, H-60a), 4.08 (m, 1 H, H-
0
0
0
0
(dd, J1b,3ꢁ1.0 Hz, H-1a), 4.89 (d, 1 H, J4 ,OH
10), 4.06 (dd, 1 H, J5 ,6 b 2.7 Hz, H-60b), 3.86 (ddd, 1
0
0 0
4.1 Hz, OH-40), 4.75 (bs, 1 H, OH-30), 4.35 (t, 1 H,
H, H-50), 3.14 (dd, 1 H, J1 ,2 5.3 Hz, H-20), 2.47 (m,
2 H, H-3a,3b), 2.09, 2.07, and 2.04 (3s, 3Â3 H, 3
OAc), 1.30 (bs, 2 H, NH2).
0
0
J6 a,OH=J6 b,OH 5.9 Hz, OH-60), 3.86 (td, 1 H,
0
0
J1 ,2 =J1 ,3b 4.8 Hz, J1 ,3a 10.0 Hz, H-10), 3.69 (m, 1 H,
0
0
0
0
J2 ,3 10.9 Hz, H-20), 3.56 (ddd, 1 H, J5 ,6 a 2.5 Hz,
3-(3,4,6-Tri-O-acetyl-2-deoxy-2-tri¯uoroacetam-
ido-a-d-glucopyranosyl)-1-propene (8).ÐThe gen-
eral free-radical allylation procedure was followed.
3,4,6-Tri-O-acetyl-2-deoxy-2-tri¯uoroacetamido-ꢀ-
d-glucopyranosyl bromide [15] (931.4 mg, 2 mmol),
allyltributyltin (1.24 mL, 4 mmol), and AIBN
(57.8 mg, 0.35 mmol) in dry toluene (6 mL) was
degassed for 15 min, and the mixture was then
0
0
0
0
J6 a,6 b 11.5 Hz, H-60a), 3.44 (m, 2 H, H-30, H-60b),
3.27 (m, 1 H, H-50), 3.13 (m, 1 H, H-40), 2.33 (m, 1 H,
H-3a), 2.11 (m, 1 H, H-3b), 1.81 (s, 3 H, NHCO
CH3); 13C NMR (75 MHz, CD3SOCD3): ꢂ 169.16
(NHCOCH3),135.56 (C-2), 116.23 (C-1), 73.81 (C-50),
72.36 (C-10), 71.00 (C-30), 70.37 (C-40), 61.03 (C-60),
53.11 (C-20), 30.19 (C-3), 22.64 (NHCOCH3); MS:
m/z 268 (M+Na)+, 246 (M+H)+, 204 (M allyl)+.
Anal. Calcd for C11H19NO5 (245.273): C, 53.87; H,
7.81; N, 5.71. Found: C, 53.89; H, 7.76; N, 5.64.
0
0
ꢀ
heated for 6 h at 80±85 C under Ar. The crude
products were puri®ed on a column of silica gel
that was eluted with 1:2 EtOAc±hexanes to a€ord

328
J. Cui, D. Horton/Carbohydrate Research 309 (1998) 319±330
a mixture of the ꢀ and ꢁ anomers of 8 as a colorless
syrup (559.4 mg, 65%, ꢀ:ꢁ=12:1). The mixture was
dissolved in hot 1:4 EtOAc±hexanes, and then a
syrup that separated was collected, which was pure
8: [ꢀ]d +22ꢀ (c 1.1, CHCl3); Rf 0.52 (1:1 EtOAc±
hexanes); NMR data see Tables 1±3; MS: m/z
426 (M+H)+, 384 (M allyl)+, 366 (M OAc)+,
324 (M-HOAc allyl)+, 306 (M+H 2HOAc)+,
264 (M 2HOAc allyl)+. Anal. Calcd for
C17H22F3NO8 (425.354): C, 48.00; H, 5.21; N, 3.29.
Found: C, 48.10; H, 5.18; N, 3.30.
Conversion of compound 8 into its 2-acetamido
analogue 4.ÐA solution of 8 in satd methanolic
HCl was stirred at room temperature until TLC
showed complete disappearance of 8, whereupon
the soln was evaporated. The residue was dissolved
in pyridine and treated with an excess of Ac2O.
The solution was kept overnight and then poured
into ice-water. Conventional processing gave com-
pound 4 as needles identical with authentic 4 by
mp, [ꢀ]d, and 1H and 13C NMR spectra.
NMR data see Tables 1±3; MS: m/z 460 (M+H)+,
418 (M-allyl)+, 400 (M OAc)+, 358 (M HOAc
allyl)+, 340 (M+H 2HOAc)+, 298 (M 2HOAc
allyl)+, 280 (M+H 3HOAc)+. Anal. Calcd for
C23H25NO9 (459.448): C, 60.13; H, 5.48; N, 3.05.
Found: C, 60.02; H, 5.56; N, 3.10.
A nonsolvated form of 9 has been reported [9j]
to have mp 105±106 ꢀC.
The second fraction (Rf 0.29, 1:1 EtOAc±hex-
1
anes) was evaporated to a syrup, whose H NMR
spectrum indicated it to be a mixture of the glycal
elimination product 10 (16%) and a reduction
product, 3,4,6-tri-O-acetyl-1,5-anhydro-2-deoxy-2-
phthalimido-d-glucitol (11) (8%). After recrystalli-
zation from 1:4 EtOAc±hexanes, 3,4,6-tri-O-acetyl-
1,5-anhydro-2-deoxy-2-phthalimido-d-arabino-hex-
1-enitol (10) was obtained as colorless needles: mp
ꢀ
126±127 C; [ꢀ]d 11ꢀ (c 0.57, CHCl3); lit [17].
syrup, [ꢀ]d 15ꢀ (c 0.20, CHCl3); H NMR data
1
for compound 10 (300 MHz, C6D6): ꢂ 7.44 and 6.83
(m, 4 H, Phth), 6.44 (s, 1 H, H-1), 5.95 (d, 1 H, J3,4
4.0 Hz, H-3), 5.36 (dd, 1 H, J4,5 4.7 Hz, H-4), 4.48
(dd, 1 H, J5,6a 6.4 Hz, J6a,6b 11.8 Hz, H-6a), 4.27
(m, 1 H, H-5), 4.20 (dd, 1 H, J5,6b 3.8 Hz, H-6b),
3-(3,4,6-Tri-O-acetyl-2-deoxy-2-phthalimido-b-
d-glucopyranosyl)-1-propene (9).Ð(a) From 3,4,6-
tri-O-acetyl-2-deoxy-2-phthalimido-b-d-glucopyrano-
syl bromide. The general free-radical allylation
procedure was followed. To a solution of 3,4,6-tri-
O-acetyl-2-deoxy-2-phthalimido-ꢁ-d-glucopyranosyl
bromide [16] (1.49 g, 3 mmol) and AIBN (73.9 mg,
0.45 mmol) in dry toluene (6 mL) was added allyl-
tributyltin (1.86 mL, 6 mmol). The solution was
1
1.65, 1.62, 1.47 (3s, 3Â3 H, 3OAc); H NMR data
for 11 (300 MHz, C6D6): ꢂ 7.43 and 6.87 (m, 4 H,
Phth), 6.06 (dd, 1 H, J2,3 10.4 Hz, J3,4 9.1 Hz, H-3),
5.28 (dd, 1 H, J4,5 10.1 Hz, H-4), 4.63 (ddd, 1 H,
J1a,2 11.4 Hz, J1e,2 5.5 Hz, H-2), 4.26 (t, 1 H, J1a,1e
11.2 Hz, H-1a), 4.25 (dd, 1 H, J5,6a 4.6 Hz, J6a,6b
12.3 Hz, H-6a), 4.05 (dd, 1 H, J5,6b 2.1 Hz, H-6b),
3.56 (dd, 1 H, H-1e), 3.35 (ddd, 1 H, H-5), 1.70,
1.61, and 1.44 (3s, 3OAc).
ꢀ
degassed for 15 min, then heated for 20 h at 85 C
under Ar, and then cooled to room temperature. A
second portion of AIBN (74.8 mg, 0.45 mmol) was
added to the solution, which was again degasꢀsed
for 15 min and heated for 4 h under Ar at 90 C.
TLC showed the complete disappearance of the
strongly charring spot at Rf 0.47 (1:1 EtOAc±hex-
anes) for the starting material. The product 9 had
the same Rf value as the starting material, but
showed a very light color after spraying with 10%
sulfuric acid and subsequent heating, and another
new spot at Rf 0.29 was present. The crude pro-
ducts were puri®ed on a column of silica gel, elut-
ing with 1:1 EtOAc±hexanes, to give compound 9
as a white solid (543.1 mg, 40%), which crystallized
from 1:3 EtOAc±hexanes to provide white crystals;
(b) Attempted C-allylation of 3,4,6-tri-O-acetyl-
2-deoxy-2-phthalimido-b-d-glucopyranosyl chloride.
The same reaction conditions as used for the free-
radical reaction of 3,4,6-tri-O-acetyl-2-deoxy-2-
phthalimido-ꢁ-d-glucopyranosyl bromide with
allyltributytin initiated with AIBN in toluene sol-
vent were used with 3,4,6-tri-O-acetyl-2-deoxy-2-
phthalimido-ꢁ-d-glucopyranosyl chloride [19].
However, no compound 9 was detected in the pro-
duct mixture.
3-(2-Acetamido-3,4,6-tri-O-acetyl-2-deoxy-a-d-
galactopyranosyl)-1-propene (12).ÐThe general
free-radical allylation procedure was followed. To
a soln of a mixture of 2-acetamido-3,4,6-tri-O-
acetyl-2-deoxy-ꢀ-d-galactopyranosyl chloride and
ꢀ
ꢀ
1
mp 79±81 C; [ꢀ]d +70 (c 1.2, CHCl3); The H
NMR spectrum showed that the white crystalline
product was a solvate with EtOAc and hexanes in
the ratio of 1:0.25:0.20, which was very stableꢀin
air. The solvating solvents were removed at 50 C
under vacuum to a€ord a pure glassy product 9.
1
oxazoline 13 (1.2456 g, 1:0.85 by H NMR spec-
trum, 1.93 mmol of the chloride derivative, pre-
pared from 2-acetamido-2-deoxy-d-galactose by
the general procedure [10]) in dry toluene (7 mL)

J. Cui, D. Horton/Carbohydrate Research 309 (1998) 319±330
329
were added allyltributyltin (3.27 mL, 10.7 mmol)
and AIBN (87.4 mg, 0.53 mmol). The soln was
degassed for 10 min and then heated for 7 h at 85 ꢀC
under Ar. The crude products were chromato-
graphed on a column of silica gel with 4:1 EtOAc±
hexanes to a€ord a mixture of product 12 and
oxazoline 13. To a soln of the mixture in acetone
(20 mL) was added a 1% aqueous soln of HCl
(1 mL) to convert the oxazoline 13 into a more
polar compound, 1,3,4,6-tetra-O-acetyl-2-amino-2-
deoxy-ꢀ-d-galactopyranose hydrochloride, to
facilitate separation. The solvent was evaporated,
and the residue was dissolved in CH2Cl2. The soln
was washed successively with satd NaHCO3, brine,
dried (Na2SO4) and evaporated to a syrup, which
was chromatographed on a column of silica gel
with 4:1 EtOAc±hexanes to give 12 as a white solid
(293.5 mg, 56%). The white solid was recrystallized
from 1:2 EtOAc±hexanes to provide white crystals
of 12: mp 129±130 ꢀC; [ꢀ]d +81ꢀ (c 1.0, CHCl3); Rf
0.39 (EtOAc); NMR data see Tables 1±3; MS: m/z
372 (M+H)+, 330 (M allyl)+, 312 (M OAc)+,
270 (M HOAc allyl)+, 252 (M+H 2HOAc)+.
Anal. Calcd for C17H25NO8 (371.383): C, 54.98; H,
6.79; N, 3.77. Found: C, 54.90; H, 6.76; N, 3.75.
4,5-Dihydro-2-methyl-(3,4,6-tri-O-acetyl-1,2-
dideoxy-b-d-mannopyranoso)-[2,1-d]-1,3-oxazole
(14).ÐThe general free-radical allylation procedure
was followed. To a solution of 2-acetamido-3,4,6-
tri-O-acetyl-2-deoxy-ꢀ-d-mannopyranosyl chloride
(prepared from 2-acetamido-2-deoxy-d-mannose
by the general procedure [10]) (620.7 mg, 1.7 mmol)
and AIBN (41.8 mg, 0.25 mmol) in dry toluene
(4 mL) was added allyltributyltin (1.63 mL,
5.1 mmol). The solution was degassed for 15 min
University, 1994; Diss. Abstr. Int., 55B (1994) 899.
(c) J. Cui and D. Horton, Abstracts of Papers, 19th
International Carbohydrate Symposium, San
Diego, CA, August 9±14, 1998.
[2] (a) R.E. Moore and G. Bartolini, J. Am. Chem.
Soc., 103 (1981), 2491±2494. (b) D.T. Connor,
R.C. Greenough, and M. Strandtmann, J. Org.
Chem., 42 (1977) 3664±3669. (c) S. Hanessian, T.J.
Liak, and D.M. Dixit, Carbohydr. Res., 88 (1981)
C14±C19. (d) R.E. Ireland, S. Thaisrivongs, N.
Vanier, and C.S. Wilcox, J. Org. Chem., 45 (1980)
48±61. (e) M. Sezaki, S. Kondo, K. Maeda, H.
Umezawa, and M. Ohno, Tetrahedron, 26 (1970)
5171±5190.
[3] (a) M. Chmielewski, J.N. BeMiller, and D.P. Cer-
retti, Carbohydr. Res., 97 (1981) C1±C4. (b) M.L.
Shulman, S.D. Shiyan, and A.Y. Khorlin, Carbo-
hydr. Res., 33 (1974) 229±235. (c) J.N. BeMiller,
M.P. Yadav, V.N. Kalabokis, and R.W. Myers,
Carbohydr. Res., 200 (1990) 111±126. (d) R.W.
Myers and Y.C. Lee, Carbohydr. Res., 152 (1986)
143±158.
[4] (a) M.L. Shulman, D. Svetlana, S.D. Shiyan, and
A.Y. Khorlin, Carbohydr. Res., 33 (1974) 229±235.
(b) D. Rouzaud and P. Sinay, J. Chem. Soc., Chem.
Commun., (1983) 1353±1354. (c) B. Giese, M.
Hoch, C. Lamberth, and R.R. Schmidt, Tetra-
hedron Lett., 29 (1988) 1375±1378. (d) U.C. Dyer
and Y. Kishi, J. Org. Chem., 53 (1988) 3384±3386.
(e) M. Carcano, F. Nicotra, L. Panza, and G.
Russo, J. Chem. Soc., Chem. Commun., (1989)
642±643. (f) A. de Raadt and A.E. Stutz, Carbo-
hydr. Res., 220 (1991) 101±115.
[5] (a) T.D. Inch, Tetrahedron, 40 (1984) 3161±3213.
(b) D. Horton, T. Machinami, and Y. Takagi,
Carbohydr. Res., 121 (1983) 135±161. (c) F. Reus-
ser, Infect. Immun., 2 (1970) 77±81. (d) R.E. Ire-
land, P.G.M. Wuts, and B. Ernst, J. Am. Chem.
Soc., 103 (1981) 3205±3207. (e) K. Tatsuta, H.
Takahashi, Y. Amemiya, and M. Kinoshita, J. Am.
Chem. Soc., 105 (1983) 4096±4097.
ꢀ
and then heated for 6 h at 85 C under Ar. The
crude products were applied to a short column of
silica gel. Washing the column with EtOAc and
evaporation of the eluate gave a crystalline residue
of 14 (458.0 mg, 82%). Recrystallization from ether
a€orded needles of 14: mp 127±128 ꢀC; [ꢀ]d 31ꢀ (c
[6] (a) D. Horton and T. Miyake, Carbohydr. Res.,
184 (1988) 221±229. (b) M.G. Garcia Martin, and
D. Horton, Carbohydr. Res., 191 (1989) 223±228.
(c) M.D. Lewis, J.K. Cha, and Y. Kishi, J. Am.
Chem. Soc., 104 (1982) 4976±4978. (d) A.P. Kozi-
kowski, K.L. Sorgi, B.C. Wang, and Z.-B. Xu,
Tetrahedron Lett., 24 (1983) 1563±1566. (e) A.
Giannis and K. Sandho€, Tetrahedron Lett., 26
(1985) 1479±1482. (f) J.S. Panek and M.A. Sparks,
J. Org. Chem., 54 (1989) 2034±2038. (g) A. Hos-
omi, Y. Sakata, and H. Sakurai, Tetrahedron Lett.,
25 (1984) 2383±2386.
ꢀ
1.0, CHCl3); lit [22]. mp 132±133 C; [ꢀ]d 30ꢀ (c
1.0, CHCl3).
References
[1] For preliminary communications, see: (a) J. Cui
and D. Horton, Abstracts of Papers, 203rd
National Meeting, American Chemical Society,
San Francisco, CA, April, 1992, CARB-34. (b) Ph.
D. Dissertation, Jingrong Cui, The Ohio State
[7] (a) B. Giese and J. Dupuis, Angew Chem., Int. Ed.
Engl., 22 (1983) 622±623. (b) R.M. Addlington,
J.E. Baldwin, A. Basak, and R.P. Kozyrod, J.

330
J. Cui, D. Horton/Carbohydrate Research 309 (1998) 319±330
Chem. Soc., Chem. Commun., (1983) 944±945. (c)
[11] P.L. Durette and D. Horton, J. Org. Chem., 36
(1971) 2658±2669; Adv. Carbohydr. Chem. Bio-
chem., 26 (1971) 49±125.
[12] D. Horton and M.L. Wolfrom, J. Org. Chem., 27
(1962) 1794±1798.
[13] T. Osawa, Carbohydr. Res., 7 (1968) 217±220.
[14] S. Umezawa, H. Sano, and T. Tsuchiya, Bull.
Chem. Soc. Jpn., 48 (1975) 556±559.
[15] M.L. Wolfrom and H.B. Bhat, J. Org. Chem., 32
(1967) 1821±1823; M.L. Wolfrom and P.J. Con-
igliaro, Carbohydr. Res., 11 (1969) 63±76.
F. Baumberger and A. Vasella, Helv. Chim. Acta,
66 (1983) 2210±2222. (d) G.E. Keck and E.J.
Enholm, Tetrahedron, 41 (1985) 4074±4079.
[8] (a) M.H.D. Postema, Tetrahedron, 48 (1992) 8545±
8599. (b) J. Herscovici and K. Antonakis, Stud. Nat.
Prod. Chem., 10 (1992) 337±403. (c) M.H.D. Post-
ema, C-Glycoside Synthesis, CRC, Boca Raton,
1995. (d) D.E. Levy and C. Tang, The Chemistry of
C-Glycosides, Pergamon, Tarrytown, 1995.
[9] (a) F. Nicotra, G. Russo, F. Ronchetti, and T.
Toma, Carbohydr. Res., 124 (1983) C5±C7. (b) A.
Giannis and K. Sandho€, Carbohydr. Res., 171
(1987) 201±210. (a) M. Carcano, F. Nicotra, L.
Panza, and G. Russo, J. Chem. Soc., Chem. Com-
mun., (1989) 297±298. (d) R. Grondin, Y. Leblanc,
and K. Hoogsteen, Tetrahedron Lett., 32 (1991)
5021±5024. (e) C.R. Bertozzi and M.D. Bednarski,
Tetrahedron Lett., 33 (1992) 3109±3111. (f) S. Abel,
T. Linker and B. Giese, Synlett, (1991) 171±172. (g)
C. Leteux and A Veyrieres, J. Chem. Soc., Perkin
Trans. 1, (1994) 2647±2655. (h) K. Kim and R.I.
Hollingsworth, Tetrahedron. Lett., 35 (1994) 1031±
1032. (i) M. Ho€man and H. Kessler, Tetrahedron
Lett., 35 (1994) 6067±6070. (j) B.A. Roe, C.G.
Boojamra, J.L. Griggs, and C.R. Bertozzi, J. Org.
Chem., 61 (1996) 6442±6445. (k) L. Cipolla, L.
Lay, and F. Nicotra, J. Org. Chem., 62 (1997)
6678±6681.
[16] R.U. Lemieux, T. Takeda, and B.Y. Chung, ACS
Symp. Ser., 39 (1976) 91±115.
[17] T. Ogawa, S. Nakabayashi, and K. Sasajima, Car-
bohydr. Res., 96 (1981) 29±39.
[18] D. Horton, J. Org. Chem., 29 (1964) 1776±1782; S.
Guberman and D. Horton, ibid., 32 (1967) 294±296.
[19] S. Akiya and T. Osawa, Chem. Pharm. Bull., 8
(1960) 583±587.
[20] (a) B. Giese, Angew Chem., Int. Ed. Engl., 28 (1989)
969±980. (b) J. Dupuis, B. Giese, D. Ruegge, H.
Fischer, H.-G Korth, and R. Sustmann, Angew.
Chem., Int. Ed. Engl., 23 (1984) 896±898. (c) H.-G.
Korth, R. Sustmann, J. Dupuis, and B. Giese, J.
Chem. Soc., Perkin Trans. 2, (1986) 1453±1459. (d)
B. Giese, Pure Appl. Chem., 60 (1988) 1655±1658.
[21] T. Yamazaki, C.D. Warren, A. Herscovics, and
R.W. Jeanloz, Can. J. Chem., 59 (1981) 2247±2252.
[22] B.R. Baker, J.P. Joseph, R.E. Schaub, and J.H.
Williams, J. Org. Chem., 19 (1954) 1786±1792.
[10] D. Horton, Org. Syn., 46 (1966) 1±5.