Mild synthesis of disaccharidic 2,3-enopyranosyl cyanides and 2-C-2-deoxy pyranosyl cyanides with Hg(CN)2/HgBr2/TMSCN

Article abstract of DOI:10.1021/jo0111661

Lewis acid-catalyzed dimerization of mono- and disaccharidic per-O-acetylated glycals gave di- and tetrasaccharidic O-acetylated C-glycosides, respectively. 2,3-Enopyranosyl cyanides were obtained from per-O-acetylated glycals by a new, mild anomeric S_N′-acetoxy displacement with Hg(CN)₂/HgBr₂/TMSCN. Per-O-acetylated 2-C-2-deoxy-pyranoses were converted into pyranosyl cyanides by the same reagent. An unprecedented acetic acid elimination from dimers with D-galacto- and L-fuco-configurations accompanied the S_N-displacement under those conditions. A new set of ¹H NMR coupling constants for 2,3-enopyranosyl systems was used for configurational assignment of complicated tetrasaccharide mimics.

Full text of DOI:10.1021/jo0111661

Mild Syn th esis of Disa cch a r id ic 2,3-En op yr a n osyl Cya n id es a n d
2-C-2-Deoxy P yr a n osyl Cya n id es w ith Hg(CN)2/HgBr 2/TMSCN†
Andreas H. Franz, YiQiu Wei, Vyacheslav V. Samoshin, and Paul H. Gross*
Department of Chemistry, University of the Pacific, 3601 Pacific Avenue, Stockton, California 95211
Received December 20, 2001
Lewis acid-catalyzed dimerization of mono- and disaccharidic per-O-acetylated glycals gave di- and
tetrasaccharidic O-acetylated C-glycosides, respectively. 2,3-Enopyranosyl cyanides were obtained
from per-O-acetylated glycals by a new, mild anomeric SN′-acetoxy displacement with Hg(CN)2/
HgBr2/TMSCN. Per-O-acetylated 2-C-2-deoxy-pyranoses were converted into pyranosyl cyanides
by the same reagent. An unprecedented acetic acid elimination from dimers with D-galacto- and
1
L-fuco-configurations accompanied the SN-displacement under those conditions. A new set of H
NMR coupling constants for 2,3-enopyranosyl systems was used for configurational assignment of
complicated tetrasaccharide mimics.
In tr od u ction
All literature procedures for the preparation of glycals
via glycosyl bromides have several problems. For ex-
ample, 1 had been synthesized from per-O-acetylated
D-glucose in a one-pot, two-step sequence by the method
of Roth/Pigman9 in 60-70% yield. Shafizadeh10 had
synthesized 3 from the O-acetylated glycosyl bromide in
32% yield. To purify 3, a vacuum distillation was de-
scribed, which might contribute to the overall low yield.
Compound 5 had been obtained from the O-acetylated
starting material in 40% yield by the method of El
Khadem and co-workers.11 Most problematic in the above
methods had been the solvolysis of the intermediate
glycosyl bromides during elimination.
In published procedures,9-20 the formation of glycosyl
bromides involved the formation of HBr in situ of a
glacial acetic acid solution of the per-O-acetylated start-
ing material. Red phosphorus had been treated with
elemental bromine to give PBr3, which was hydrolyzed
with H2O. Already here, a substantial amount of the
glycosyl bromide may have been lost due to solvolysis.
The elimination step usually had been performed in
aqueous medium (AcOH/NaOAc buffer) adding unwanted
hydrolysis of the intermediate bromide. The zinc-pro-
moted elimination of acetylated glycosyl bromides, as
Synthetic oligosaccharides derived from glycals have
been tested as antigens in anticancer vaccination.1,2 The
role of oligosaccharides in biological systems has become
evident within the past decade.3-8 The potential infor-
mational content of oligosaccharides due to monosaccha-
ride sequence and permutations of glycosidic linkages
most likely exceeds that of proteins. However, a distinct
disadvantage of O-glycosidically linked oligosaccharides
is their metabolic instability and limited lifetime in
biological systems. Therefore, much effort has been spent
in the past two decades on the development of feasible
pathways towards C-glycosidic sugars, many of which
exhibit biological activities similar to their O-glycosidic
counterparts.
Resu lts a n d Discu ssion
Syn th esis of Glyca ls. Glycals are versatile chiral
building blocks in the synthesis of natural products and
biologically relevant oligosaccharides2 because of the ease
of their chemical transformation.
† From the Ph.D. Dissertation of A. H. Franz, University of the
Pacific, Stockton, CA. Preliminary partial versions (p ) poster, o )
oral) were presented at the XIX International Carbohydrate Sympo-
sium, San Diego, 1998 (2p, Abstracts B386 and B387); 39th Experi-
mental Nuclear Magnetic Resonance Conference, Pacific Grove, CA,
March 1998 (1p, Abstract 070); 34th American Chemical Society
Western Regional Meeting, San Francisco, CA, October 1998 (1o);
219th National Meeting of the American Chemical Society, San
Francisco, CA, March 2000 (1o, Abstract CARB34); and 36th American
Chemical Society Western Regional Meeting, October 2000 (1o, 2p).
(1) Danishefsky, S. J .; Allen, J . R. Angew. Chem., Int. Ed. 2000, 39,
836.
(9) Roth, W.; Pigman, W. In Methods Carbohydrate Chemistry; 1963;
p 405.
(10) Shafizadeh, F. In Methods Carbohydrate Chemistry; 1963; p
409.
(11) El Khadem, H. S.; Swartz, D. L.; Nelson, J . K.; Berry, L. A.
Carbohydr. Res. 1977, 58, 230.
(12) Hall, L. D.; J ohnson, L. F. Tetrahedron 1964, 20, 883.
(13) Chalmers, A. A.; Hall, R. H. J . Chem. Soc., Perkin Trans. 2
1974, 728.
(2) Danishefsky, S. J .; Bilodeau, M. T. Angew. Chem., Int. Ed. Engl.
1996, 35, 1380.
(3) Simanek, E. E.; McGarvey, G. J .; J ablonowski, J . A.; Wong,
C.-H. Chem. Rev. 1998, 98, 833.
(4) Dwek, R. A. Chem. Rev. 1996, 96, 683.
(5) Mammen, M.; Chio, S.-K.; Whitesides, G. M. Angew. Chem., Int.
Ed. 1998, 37, 2755.
(14) Munesada, K.; Ogihara, K.; Suga, T. J . Sci. Hiroshima Uni-
versity, Ser. A: Phys. Chem. 1989, 53, 1.
(15) Iselin, B.; Reichstein, T. Helv. Chim. Acta 1944, 27, 1200.
(16) Iselin, B.; Reichstein, T. Helv. Chim. Acta 1944, 27, 1146.
(17) Iselin, B.; Reichstein, T. Chem. Abstr. 1945, 39, 4845.
(18) Haworth, W. N.; Hirst, E. L.; Reynolds, R. J . W. J . Chem. Soc.
C 1934, 302.
(6) Lis, H.; Sharon, N. Chem. Rev. 1998, 98, 637.
(7) Weis, W. I.; Drickamer, K. Annu. Rev. Biochem. 1996, 65, 441.
(8) Rini, J . M. Annu. Rev. Biophys. Biomol. Struct. 1995, 24, 551.
(19) Watters, A. J .; Hudson, C. S. J . Am. Chem. Soc. 1930, 64, 1852.
(20) Haskins, W. T.; Hann, R. M.; Hudson, C. S. J . Am. Chem. Soc.
1942, 64, 1852.
10.1021/jo0111661 CCC: $22.00 © 2002 American Chemical Society
Published on Web 10/08/2002
7662
J . Org. Chem. 2002, 67, 7662-7669

Mild Synthesis of Pyranosyl Cyanides
TABLE 1. Glyca ls Syn th esized u n d er Ap r otic
Con d ition s
SCHEME 1. Typ ica l Syn th esis of a Glyca l
(Tr i-O-a cetyl-D-glu ca l)
described by Somsa´k and Ne´meth,21 avoided the possibil-
ity of their hydrolysis. However, Somsa´k and Ne´meth still
found it necessary to isolate the notoriously unstable
glycosyl bromides as starting materials.
In the present work we describe a new two-step
synthesis of glycals in an aprotic solvent followed by a
simple workup procedure, with good to excellent yields.
Side products are minor. The intermediate glycosyl
bromide does not need to be isolated. Commercially
available D-glucose, D-galactose, and L-fucose were per-
O-acetylated according to standard literature procedures.
Some of the per-O-acetylated saccharides are commer-
cially available.
In our procedure (see Experimental Section), TiBr4 in
CH2Cl2 was chosen as a brominating reagent because of
titanium’s tendency to form a stable titanyl TiO2+ cation,
which is a well-known aspect of titanium chemistry.22
The reaction was run with 0.6 equiv of TiBr4 (20%
excess). We reasoned that the bromination would be
based on a catalytic cycle (Scheme 1). Small-scale reac-
tions were run solely with TiBr4. Residual traces of water
on the surface of small containers caused hydrolysis of
TiBr4 and furnished HBr. Upon scale-up, yields of the
intermediate bromides decreased. However, addition of
a catalytic amount of glacial acetic acid overcame this
problem. In the overall balance, Ac2O is produced as a
byproduct instead of potentially solvolytic AcOH.
The inertness of CH2Cl2 toward TiBr4 in our procedure
and the high solubility of starting materials, intermedi-
ates, and products in combination with the ease of
removal of byproducts from CH2Cl2 made it possible to
directly use glycosyl bromides in CH2Cl2. Running the
elimination at the boiling point of CH2Cl2 (40 °C) favored
the reaction entropically. Lower temperatures resulted
in no reaction. We found none of the reported inconsis-
tencies when we used freshly activated zinc dust.35
Per-O-acetylated D-galactose, L-fucose, D-maltose,
D-lactose, and D-gentiobiose were treated analogously to
penta-O-acetyl-D-glucose (see Table 1). Compounds 1, 3,
and 5 crystallized at -20 °C from a minimal amount of
warm Et2O (∼2.5 mL/g), with pentane added (∼8 mL/g)
to opalescence. Compounds 9 and 11 crystallized from a
minimal amount of hot ethanol (∼5 mL/g), with water
added (∼12 mL/g) to opalescence at 0 °C. Compound 12
was further purified by flash column chromatography on
silica gel by gradient elution with hexane with increasing
amounts of EtOAc. All glycals were obtained in >90%
purity prior to purification, as estimated by NMR.
Syn th esis of Dim er s. One way to render carbohydrate-
derived structures metabolically more stable is the
(30) Barfield, M.; Spear, R. J .; Sternhell, S. Chem. Rev. 1976, 76,
593.
(31) Casiraghi, G.; Cornia, M.; Colombo, L.; Rassu, G.; Fava, G. G.;
Belicchi, M. F.; Zetta, L. Tetrahedron Lett. 1988, 29, 5549.
(32) Collins, P. M., Ed. Carbohydrates; Chapman & Hall, Ltd.:
London, 1987.
(33) Advantageous for the handling of TiBr4 is a glovebox. In the
method described herein, an “atmosbag” (Aldrich) was used to maintain
an inert gas atmosphere and solid TiBr4 was employed.
(34) A system 4:6 hexane/EtOAc was used (Rf starting material )
0.54, Rf bromide ) 0.71). Before spotting the glycosyl bromides on the
TLC plate, it was necessary to prespot with THF to prevent their
hydrolysis due to local heat of adsorption. Prolonged exposure to air
humidity (>3 h) resulted in complete hydrolysis of the bromide on the
plate. TLC detection was done by spraying with 10% H2SO4/MeOH
and charring at 150 °C.
(35) Inconsistencies due to different grades of zinc reported by E.
Erdik (Tetrahedron 1987, 43, 2203) were not noted if a large excess of
zinc powder (i.e., 8 g, Aldrich, 98+%, <10 µm) was activated properly
prior to use as follows. The zinc was suspended in water (10 mL). With
stirring was added dropwise CuSO4 (10% solution, 5 mL). After the
mixture was stirred for 10 min, the Zn/Cu couple was filtered off,
washed with water (2 × 10 mL), ethanol (2 × 10 mL), and diethyl
ether (3 × 10 mL), subsequently dried in vacuo over P2O5 for 6 h, and
used promptly, as it may loose its activity by aging in air over several
days.
(21) Somsak, L.; Nemeth, I. J . Carbohydr. Chem. 1993, 12, 679.
(22) Gmelin Handbuch der Anorganischen Chemie-Titan; 8th ed.;
Verlag Carl Winter: Heidelberg; Vol. 41.
(23) Ferrier, R. J .; Prasad, N. J . Chem. Soc. C 1969, 581.
(24) Franz, A. H.; Gross, P. H. Carbohydr. Lett. 1997, 2, 371.
(25) March, J . Advanced Organic Chemistry: Reactions, Mecha-
nisms, and Structure, 4th ed.; Wiley-Interscience: 1992; p 327.
(26) Wei, Y. M. S. Thesis, University of the Pacific, Stockton, CA,
1991.
(27) De las Heras, F. G.; San Felix, A.; Fernandez-Resa, P. Tetra-
hedron 1983, 39, 1617.
(28) Franz, A. H.; Zhdankin, V. V.; Samoshin, V. V.; Minch, M. J .;
Young, V. G.; Gross, P. H. Mendeleev Commun. 1999, 45.
(29) Lemieux, R. U.; Fraga, E.; Watanabe, K. A. Can. J . Chem. 1968,
46, 61.
J . Org. Chem, Vol. 67, No. 22, 2002 7663

Franz et al.
SCHEME 2. Str u ctu r es of P r od u cts fr om Glyca l Dim er iza tion a n d Cya n a tion a
a
Ring A (nucleophile) and ring B (electrophile) are used throughout the text. The molecular structure of 20 was unambiguously
established by X-ray crystallography, details of which will be published.
SCHEME 3. Dim er iza tion of P er -O-a cetyl-cellobia l a n d P er -O-a cetyl-la cta la
a
(i) BF3/Ac2O/CH2Cl2; (ii) Hg(CN)2/HgBr2/TMSCN, 2:3 THF/H3CCN.
dimerization of glycals, which has been described first
for per-O-acetylated glucal by Ferrier and Prasad.23 We
synthesized similar products (Scheme 2), which are
directly C-linked. We also applied an improved dimer-
ization method24 with BF3/Ac2O catalyst to disaccharidic
glycals in order to obtain tetrasaccharidic glycomimetics
(Scheme 3). Such compounds may potentially show
altered and/or enhanced bioactivity due to a well-
documented5 cluster effect. The nonphysiological nature
of the new C-C-linkages of our compounds may transfer
metabolic stability also to O-glycosidic bonds still present
in their structures. Dimerizations that lead from disac-
charidic glycals to tetramers gave much lower yields.
Their syntheses required careful optimization of the
reaction to minimize cleavage of the disaccharidic bonds
by the BF3-catalyst.
Cya n a tion P r od u cts. Another way to render carbo-
hydrate-derived structures metabolically more stable is
introducing a C-glycosidic cyanide functionality into
glycals by the SN′-reaction (Scheme 4). The complex
stereospecificity, if any, of such reactions has been
discussed by March.25
In an experimental procedure investigated previously
in this laboratory with tri-O-acetyl-D-glucal,26 a mixture
of Hg(CN)2, HgBr2, and TMSCN was used as a mild
source of cyanide in a mixture of dry THF and MeCN
(2:3). We could use this reagent not only to convert
O-glycosidic glycals into 2,3-enopyranosyl cyanides
(Scheme 4) but also to obtain anomeric cyanides from
glycal dimers and tetramers (Schemes 2 and 3), which
are 2-C-2-deoxy-sugars. The reagent was compatible with
1,4- and 1,6-O-glycosidic linkages, which are normally
7664 J . Org. Chem., Vol. 67, No. 22, 2002

Mild Synthesis of Pyranosyl Cyanides
SCHEME 4. Cya n a tion of P er -O-a cetyl-glu ca l a n d
Disa cch a r id ic P er -O-a cetyl-glyca ls
F IGURE 1. Coupling constants for 5HO-28 and OH5-conforma-
tions29,31 of 2,3-dideoxy-2-enopyranose systems (ring “B”).
NMR P r oof of Str u ctu r e. The assignment of config-
uration and conformation of the obtained dimers was
accomplished by NMR spectroscopy. Small coupling
constants (0-2 Hz) between the allylic protons and the
vinylic protons in ring B of products 2, 4, and 6 were
indicative of an ∼80° angle. Larger coupling constants
(3-5 Hz) indicated a ∼40° angle (Figure 1).28
SCHEME 5. P ossible Mech a n ism for th e
F or m a tion of 19 a n d 20
Compound 18 was obtained from 2 as a 1:2 R/â
mixture. A sample of the â-anomer could be obtained by
1
flash chromatography. In the H NMR spectrum and in
1H-1H COSY correlation of the â-anomer, H2A showed
3
coupling constants of 3J 2,1 ) 5.1 Hz, J 2,3 ) 11.4 Hz, and
3J 2,1B ) 5.1 Hz. In turn, the equatorial H1A showed a
coupling constant of 3J 1,2 ) 4.8 Hz at a chemical shift of
δ ) 4.88 ppm. The values of the coupling constants for
the remainder of ring A are consistent with the gluco-
configuration as assigned previously.24 The signal of H1B
is well exposed as a double doublet at δ ) 4.39 ppm with
3
values of 3J 1,2 < 1 Hz, 4J 1,3 ) 2.1 Hz, and J 1,2A ) 4.8 Hz.
5
Ring B was assigned a H0 conformation on the basis of
a set of standard coupling constants for 2,3-enopyranosyl
systems published by us.28
Compounds 19 and 20 showed a close resemblance to
each other in their NMR spectra (Figure 2). However,
significant differences compared to 18 were discovered.
1
In the H NMR spectrum of 19 only four signals for the
prone to cleavage in the presence of BF3.27 In contrast to
the sequence 2 f 18, the introduction of the cyanide
group at the anomeric center of ring A in the galactal
and fucal dimers (4 and 6, respectively) introduced also
a double bond into ring B for either anomer (Scheme 2).
acetyl groups were observed. On the basis of 1H-1H
COSY correlation and an unusually large chemical shift,
the double triplet at δ ) 6.06 ppm was assigned to a
vinylic H3A. The absence of an H2A signal showed that
one molecule of acetic acid had been eliminated between
C2A and C3A. The signal at 5.19 ppm was assigned to
H4A with vinylic coupling of 5.7 Hz to H3A, consis-
tent with a pseudoequatorial orientation. The signal at
δ ) 5.39 ppm represents H1A with allylic coupling to
H3A of 1.2 Hz. The configuration at C1A was inferred
by NOEDIF experiments. Selective irradiation of H1A
gave a NOE at H1B and H5B. No other transannular
NOEs were observed, consistent with both H1A and H1B
being pseudoequatorial. The configuration at C1A was
concluded to be R. In ring B, H1B showed a coupling
constant of 3J 1,2 ) 3.3 Hz and H4B a coupling constant of
3J 4,3 ) 5.1 Hz, which put both protons into a pseudoequa-
torial position.28 It was concluded that ring B favored the
5H0 conformation.
The example 6 f 20 (Scheme 5) shows a possible
mechanism for the formation of 19 or 20 from 4 or 6,
respectively. Initial activation of the anomeric acetate
from 6 results in a partially flattened oxonium ion (6a ).
The axial acetoxy group at C4A helps to stabilize the
positive charge at C1A-O. The increased lifetime of 6a
allows CN- to react as a base, removing H2A to give an
intermediate glycal (6b). This removal may be aided by
the carbonyl oxygen of the 4-acetoxy group acting as an
intramolecular base, transporting the proton into the
medium, like a “molecular crane”. Subsequently, 6b and
cyanide react in an SN2′ reaction.
For 2 f 18, the acetoxy group at C4A is in an equatorial
position and, therefore, cannot stabilize the oxonium ion.
The nucleophilic cyanide traps the oxonium ion prior to
proton abstraction.
1
In the very similar H NMR spectrum of 20, the region
at δ ) 3.8-4.5 ppm was greatly simplified due to the low
J . Org. Chem, Vol. 67, No. 22, 2002 7665

Franz et al.
F IGURE 3. Compound 14: (a) 1H NMR spectrum (detail);
(b) configuration and conformation; (c and d) unusual five-bond
coupling between H1B and H4B.
TABLE 2. Cou p lin g Con sta n ts (h er tz), Con figu r a tion s
(C1), a n d P r efer r ed Con for m a tion s for 2,3-En op yr a n osyl
Rin gs (CDCl3)
configuration conformation
compd
3J 1,2
3J 3,4
at C1B
of ring Bc
2
<1
3.3
3.3
3.0
3.0
2.4
5.1
4.5
5.1
1.8
1.8
2.1
1.8
2.1
1.5
1.5
4.8
4.8
R
R
R
R
R
â
5HO (D-config)
OH5 (D-config)
5HO (L-config)
OH5 (D-config)
OH5 (D-config)
OH5 (D-config)
OH5 (D-config)
OH5 (D-config)
OH5 (D-config)
OH5 (D-config)
5HO (D-config)
OH5 (D-config)
5HO (L-config)
5HO (L-config)
5HO (D-config)
4a
6a
8
10
14
15a b 3.3
R
â
15bb 1.8
16
17
18
19
20
2.5
2.1
< 1
3.3
R
R
R
R
-
R
R
F IGURE 2. Compounds 19 (a) and 20 (b): 1H NMR spectrum
(detail).
- (ring A)
3.5 (ring B) 5.1 (ring B)
1.5 3.3
5.4 (ring A)
δ-values of H6 (CH3 group). Only two acetyl signals were
observed. Also here, the signal for H2A was absent, and
20 had a double bond between C2A and C3A. Assignments
of both rings in 20 were completely analogous. The axial
disposition of the sterically demanding ring A in 19 and
20 is possible because the acetyl group at C-4 is on the
opposite side of the ring compared to that in the gluco-
configuration of 18.
21
a
b
Measured in acetone-d6. Fractions of both R- and â-anomer
were isolated in pure form. c Conformation symbol convention as
shown in Figure 1.
R-anomers. The Casiraghi rule also states that the
anomer with the greater J 4,5 coupling constant (8.7-10
3
Compound 14 (Scheme 3) was obtained almost exclu-
sively as the â-anomer (>20:1). The values of the three
bond couplings in ring A of 14 were all in agreement with
a â-D-gluco-configuration and a 4C1-conformation. The
trans-diaxial couplings for the 1-5 had values of 8.1-
9.9 Hz. The â-configuration at C1 in ring B was supported
Hz) in the unsaturated ring has the â-configuration as
opposed to the R-configuration (3-9 Hz).31 However, in
our experiments, the values of 3J 1,2 in ring B were a more
reliable indicator. The coupling constants are generally
on the order of 3.3 and 1.8 Hz for the R- and â-anomers,
respectively (Figure 4).28
3
3
by coupling constants of J 1B,2B ) 2.4 Hz and J 3B,4B
)
The obvious change of the splitting pattern of H2B
2.1 Hz (Table 2).28 Protons 1B and 4B showed a five-bond
(R-anomer) with one large coupling constant J 2,3 ) 10.2
3
5
3
coupling with a value of J 1,4 ) 1.8-2.4 Hz. Proton 1B
Hz, one medium constant J 2,1 ) 3.3 Hz, and one small
caused a quartet at δ ) 5.06 ppm (Figure 3). Proton 4B
caused a double quartet at δ ) 4.13 ppm. Allylic inter-
proton coupling has been thoroughly reviewed.30
constant J 2,4 ) 1.8 Hz is evidence that H1B is in a
4
pseudoequatorial position. The coupling pattern of H3B
did not change. Therefore, the configuration was assigned
as R. The ring is in a OH5 conformation. The unusual
coupling pattern observed for H1B and H4B in the
â-anomer of 14 was observed here also (Figure 4). H1B
generates a quartet at δ ) 5.08 ppm with a five-bond
coupling of 2.4 Hz toward H4B, which could not be
Both, the R- and â-anomers of 15 could be isolated by
flash chromatography. The values of optical rotation for
15 are consistent with the Casiraghi rule, which predicts
for the â-anomer of monosaccharidic 2,3-enopyranose
cyanides higher absolute optical rotations than for the
7666 J . Org. Chem., Vol. 67, No. 22, 2002

Mild Synthesis of Pyranosyl Cyanides
F IGURE 5. Compound 17: (a) 1H NMR spectrum (detail);
(b) configuration and conformation; (c and d) unusual five-bond
coupling between H1B and H4B (R-anomer)
F IGURE 4. Compound 15: (a) 1H NMR spectrum (detail, R);
(b) 1H NMR spectrum (detail, â); (c) configuration and confor-
mation (R); (d) configuration and conformation (â); (e) unusual
five-bond coupling between H1B and H4B (â-anomer).
investigated because its signal at ∼4.1 ppm heavily
overlaps with other signals.
Compound 16 was obtained as a 1:1 mixture of R- and
â-anomers. The coupling constants for H2B and H3B are
virtually identical to those of 15. The unusual coupling
pattern observed for H1B and H4B in the â-anomer of 14
was observed here also. H1B causes a quartet at δ ) 5.05
ppm including a five-bond coupling of 2.1 Hz toward H4B,
which could not be investigated because its signal at
∼4.26 ppm heavily overlaps with other signals.
Compound 17 was assigned the R-configuration analo-
gously. The unusual coupling pattern observed for H1B
and H4B in the â-anomer of 14 was observed here also
(Figure 5). H1B generates a quartet at δ ) 5.06 ppm with
a five-bond coupling of 2.1 Hz toward H4B. Proton 4B gave
a double quartet at δ ) 5.32 ppm, which is shifted
downfield because the OH group at carbon 4B is acety-
lated.
The NMR values found for vinylic and allylic coupling
constants in the 2,3-enopyranosyl systems of 2, 6, 14-
17, 19, and 20 subsequently allowed us to assign the
configuration of the C-C linkage and conformation of
rings A and B in 8 and 10. The NMR spectra of the
tetramers that resulted from dimerization of per-O-acetyl
cellobial and per-O-acetyl lactal were characterized by
heavily overlapped signals except for H1A, H2A, H2B, and
H3B.
F IGURE 6. Compound 8 (a and b): 1H NMR spectrum, details,
and conformation. Compare splitting patterns of H3B and H2B
to Figure 4a.
ppm) and a double triplet (δ ) 5.91 ppm), respectively.
3
The coupling constants of H2B were measured with J 2,1
3
4
) 3.0 Hz, J 2,3 ) 10.8 Hz, and J 2,4 ) 1.8 Hz. The values
for H3B were J 3,1 ) 1.8 Hz, J 2,3 ) 10.8 Hz, and J 3,4
)
4
3
3
The signal of H2A in 8 showed coupling constants of
1.8 Hz. The coupling constants of H2B (3J 2,1 and J 2,4) and
4
3
3
3J 2,1 ) 3.6 Hz, 3J 2,3 ) 11.1 Hz, and J 2,1B ) 9.6 Hz (Figure
H3B (4J 3,1 and J 3,4) were compared to vinylic and allylic
6). This was consistent with an R-configuration at C1A
coupling constants found in 2, 6, 14-17, 19, and 20. The
configuration at C1B in 8 was assigned as R with a OH5-
conformation of ring B.
4
and a C1-conformation of ring A. The vinylic protons H2B
and H3B displayed a double double doublet (δ ) 5.63
J . Org. Chem, Vol. 67, No. 22, 2002 7667

Franz et al.
Similarly, 10 showed coupling constants for H2B (δ )
previously dried over 4 Å molecular sieves overnight). The
slurry was heated to reflux in an ultrasonic bath; 1-methyl
imidazole (1.4 mL, 1.2 equiv) and molecular sieves (4 Å, 1 g)
were added, and the slurry was kept at reflux in the ultrasonic
bath for another 0.5 h. The solution of the glycosyl bromide in
CH2Cl2 was added via dropping funnel into the refluxing
mixture over the course of 1 h. After the mixture had been
sonicated at reflux for 12 h, the reaction was normally
complete.36 The mixture, cooled to rt, was filtered through a
pad of Celite. Evaporation of the solvent in vacuo and crystal-
lization of the residue from the minimal amount of hot MeOH
(∼20 mL) gave 3,6-di-O-acetyl-4-[2,3,4,6-tetra-O-acetyl-â-D-
glucopyranosyl]-2-deoxy-D-arabino-hex-1-enopyranose (hexa-
O-acetyl-D-cellobial, 7, 7.4 g, 90%, mp 135-137 °C (lit.32 137
3
3
4
5.66 ppm) with J 2,1 ) 3.0 Hz, J 2,3 ) 10.8 Hz, and J 2,4
) 1.5 Hz. The signal of H3B (δ ) 5.96 ppm) was split
4
into a double triplet with coupling constants of J 3,1
)
3
3
1.8 Hz, J 2,3 ) 10.8 Hz, and J 3,4 ) 1.8 Hz. On the basis
of the same arguments outlined above, the configuration
at C1B in 10 was R and the conformation of ring B was
OH5.
Con clu sion s
New and mild methods for the dimerization of acety-
lated glycals and for the introduction of a cyanide
functionality into acetylated glycals and into acetylated
2-C-2-deoxy pyranoses have been developed. 2,3-Enopy-
ranosyl cyanides and 2-C-2-deoxy pyranosyl cyanides
were isolated in moderate to excellent yields. The reagent
Hg(CN)2/HgBr2/TMSCN is compatible with 1,4- and 1,6-
O-glycosidic bonds contrary to the method published by
DeLas Heras.27 The unprecedented additional elimina-
tion of acetic acid in 19 and 20 as opposed to a plain
displacement of the anomeric acetate in 18 may be
attributed to a delicate balance of nucleophilicity and
basicity of the cyanide anion on one hand and the steric
position of the C4B-acetoxy group as a potential internal
base on the other hand. NMR analysis of the vinylic and
allylic coupling constants in the 2,3-enopyranosyl rings
of the products gave a new diagnostic tool for the
assignment of configuration and conformation of such
ring systems.
°C), TLC system 4:6 hex/EtOAc, Rf 0.72, [R]18 -19° (c 1.2;
D
lit.32 -20°).37
Other glycals (1, 3, 5, 9, 11, 12) synthesized under aprotic
conditions along with yields and physical properties are listed
in Table 1.
Im p r oved P r oced u r e for Glyca l Dim er iza tion (8 a n d
10). A fresh solution of CH2Cl2 (5 mL), Ac2O (1.0 equiv), and
BF3‚Et2O (0.1 equiv) was stirred at 0 °C for 15 min with a
solution of the glycal (1.0 g) in CH2Cl2 (5 mL) and took on a
deep purple color when allowed to reach rt over the next hour.
The color changed to light yellow when the solution was stirred
vigorously with saturated aqueous NaHCO3 (2 equiv) for 0.5
h. The separated organic phase was dried (Na2SO4). After
removal of the solvent in vacuo, flash chromatography as
specified gave the products.
1,3,6-Tr i-O-a cetyl-2-C-[6-O-a cetyl-2,3-d id eoxy-4-(2,3,4,6-
t e t r a -O-a ce t yl-â-D-glu cop yr a n osyl)-r-D-er yth r o-h e x-2-
en op yr a n osyl]-4-(2,3,4,6-tetr a -O-a cetyl-â-D-glu cop yr a n o-
syl)-2-d eoxy-r-D-glu cop yr a n ose (8). Obtained from 7 (0.5
g, 0.89 mmol) by dimerization, the product was isolated by
flash chromatography (20 g SiO2, EtOAc/hexane: [2:3] 250 mL,
[2:1] 2 × 250 mL) and crystallized from the solvent after 1
day. Colorless crystals: yield 20%, mp > 250 °C, no decom-
Exp er im en ta l Section
Flash chromatography was performed on silica gel from J .
T. Baker (40 µm, F ) 0.5 g/cm3). Solvents of HPLC grade were
used without further purification unless stated otherwise. TLC
was done on SiO2 plates from Analtech with the same solvents,
unless stated otherwise. 1H NMR and 13C NMR data were
acquired at 300 MHz. Melting points are uncorrected. Optical
rotations were measured against the sodium 589 nm line in
CHCl3, unless stated otherwise. Elemental analysis was done
by Desert Analytics, Tucson, AZ.
Mass spectral data were obtained with an infusion pump
in positive and negative atmospheric pressure chemical ioniza-
tion mode. Solutions of c ) 100 nM in MeCN were prepared.
A 250 µL syringe was used for continuous injection at a rate
of 0.3 mL/h. N2 was used as a desolvating gas at a pressure of
700 psi. Protonated, sodiated, and potassiated molecules are
abbreviated as (M + H)+, (M + Na)+ and (M + K)+, respec-
tiveley. Radical anions are abbreviated as (M•-).
position, Rf (7:3 EtOAc/hexane) 0.56, [R]25 -51° (c 1).
D
1,3,6-Tr i-O-a cetyl-2-C-[6-O-a cetyl-2,3-did eoxy-4-(2,3,4,6-
t et r a -O-a cet yl-â-D-ga la ct op yr a n osyl)-r-D-er yth r o-h ex-2-
en op yr a n osyl]-4-(2,3,4,6-tetr a -O-a cetyl-â-D-ga la ctop yr a -
n osyl)-2-d eoxy-r-D-glu cop yr a n ose (10). Obtained from 9
(1.0 g, 1.78 mmol) by dimerization, the product was isolated
by flash chromatography (40 g SiO2, EtOAc/hexane: [2:3] 250
mL, [2:1] 2 × 250 mL). Rf (4:3 EtOAc/hexane) ) 0.42 and 0.29.
Crystallization from 1:1 THF/(i-Pr)2O gave colorless crystals
(mp 103-105 °C, yield 40 mg, 4%), which were a mixture of
four compounds (1:1:1:1). The NMR spectrum revealed eight
vinylic carbons, four C2B, and four C1B. Careful crystallization
of the remainder from THF/(i-Pr)2O (added to turbidity) gave
(36) All glycals char noticeably faster than the starting material or
bromides. The acetylated glycals of the monosaccharides run faster
on TLC than the bromides. The acetylated glycals of the disaccharides
run only slightly faster than the bromides, possibly leading to
misinterpretations of TLC data. Incomplete conversion from disaccha-
ridic bromides to glycals could be detected as follows. A TLC plate was
spotted with the reaction mixture. After moistening the plate with
water vapor, it was allowed to sit for 3 h on the benchtop. In case of
incomplete reaction, any remaining bromide was hydrolyzed and
showed up after development and charring with a lower Rf-value than
the glycal. The acetylated glycals of the disaccharides, when isolated
as syrups, were not stable for a long time. If possible, they should be
used directly in subsequent chemical transformations. However, they
can be stored in a freezer for several days without serious decomposi-
tion. In crystalline form they can be stored for prolonged times at rt.
(37) Spectral properties are as follows: 1H NMR (300 MHz, CDCl3,
primed protons refer to the unsaturated ring) δ 2.00 (s, 3H, CH3ac),
2.02 (s, 3H, CH3ac), 2.05 (s, 3H, CH3ac), 2.06 (s, 3H, CH3ac), 2.10 (s, 3H,
CH3ac), 2.13 (s, 3H, CH3ac), 3.69 (ddd, 1H, H5), 3.97-4.22 (m, 4H), 4.32
(dd, 1H, H4′), 4.45 (dd, 1H, H6a′), 4.69 (d, 1H, H1), 4.83 (dd, 1H, H2′),
4.98 (dd, 1H, H2), 5.09 (t, 1H, H3), 5.19 (t, 1H, H4), 5.43 (ddd, 1H,
H3′), 6.41 (dd, 1H, H1′); 13C NMR (300 MHz, CDCl3) δ 20.99, 21.10,
21.27, 21.42, 62.14, 68.41, 68.93, 71.70, 72.35, 73.08, 74.67, 74.99, 99.35,
100.83, 145.63, 169.35, 169.45, 170.42.
Chemicals were purchased commercially and used without
further purification unless specified.
Gen er a l P r oced u r e for P r ep a r a tion of Glyca ls (1, 3,
5, 7, 9, 11, 12). A 500 mL round-bottom flask was charged
under nitrogen with CH2Cl2 (100 mL, previously dried over 4
Å molecular sieves overnight) and TiBr4 (3.23 g, 0.6 equiv).33
Glacial acetic acid (10 drops) was added with stirring followed,
after 10 min, by the dropwise addition of D-cellobiose-octa-
acetate (10 g, 14.7 mmol) in CH2Cl2 (100 mL) over 0.5 h. The
solution was stirred at rt for 24 h. Emergence of the glycosyl
bromide was monitored by TLC.34 The reaction was quenched
with portions of NaHCO3 (5 × 1 g). Each portion was followed
by H2O (5 mL) over the course of 0.5 h. Vigorous stirring (CO2
evolution) was continued for 2 h. The separated organic layer
was filtered through a pad of Celite. The filter cake was
washed with CH2Cl2 (2 × 20 mL), and the combined CH2Cl2
solutions were dried (Na2SO4) with stirring for 2 h.
In a three-necked 500 mL round-bottom flask, a Zn/Cu
couple (8 g, 8 equiv)35 was suspended in CH2Cl2 (20 mL,
7668 J . Org. Chem., Vol. 67, No. 22, 2002

Mild Synthesis of Pyranosyl Cyanides
colorless crystals after several days: yield 100 mg, 10% with
mmol, 62%). The R-anomer was crystallized from i-PrOH:
yield (R-anomer) 200 mg (0.72 mmol, 31%), colorless crystals,
mp 165-167 °C, [R]25D -45° (c 1), Rf 0.57 (9:1 CH2Cl2/EtOAc).
Gen er a l Syn th etic P r oced u r e for 2-C-2-Deoxy-p yr a -
n osyl Cya n id es (18-21). The suspension of the Ferrier
dimerization product (0.07 mmol), Hg(CN)2 (0.1 g, 0.35 mmol),
HgBr2 (0.002 g, 0.006 mmol), and TMSCN (0.014 mL, 0.1
mmol) in absolute THF (0.3 mL) and absolute MeCN (0.6 mL)
was stirred at rt for 24 h. After completion of the reaction
(TLC), the solvent was removed in vacuo. The suspension of
the residue in CH2Cl2 (20 mL) was filtered through Celite.
Isolation and purification are described under the individual
preparations.
respect to starting material, mp 186-188 °C, [R]25 + 43° (c
D
0.5).
4,6-Di-O-a cetyl-2,3-d id eoxy-r/â-D-er yth r o-h ex-2-en op y-
r a n osyl cya n id e (13). To a mixture of Hg(CN)2 (25.1 g, 97.6
mmol) and HgBr2 (2.14 g, 5.94 mmol) in MeCN (60 mL) and
THF (30 mL) at 0 °C was added TMS-CN (8 mL, 60.0 mmol)
followed by 3,4,6-tri-O-acetyl-D-glucal 1 (4.0 g, 14.7 mmol). The
colorless suspension was stirred at room temperature under
N2 for 20 h. The catalyst was filtered off, and the filtrate was
concentrated in vacuo to a brown residue that was extracted
with CH2Cl2 (100 mL). The extract was washed with saturated
aqueous KHCO3 (100 mL), H2O (100 mL), and saturated
aqueous NaCl (100 mL), dried (MgSO4), and concentrated in
vacuo to an oil. The residual oil was crystallized from THF
and (i-Pr)2O to give 13: yield 1.51 g (6.3 mmol, 43%), mp 58-
3,4,6-Tr i-O-a cetyl-2-C-(4,6-d i-O-a cetyl-2,3-d id eoxy-r-D-
er yth r o-h ex-2-en op yr a n osyl)-2-d eoxy-r/â-D-glu cop yr a n o-
syl Cya n id e (18). Compound 224 (80 mg, 0.147 mmol) was
completely consumed after 36 h at rt [TLC (4:3 EtOAc/hexane)]
and gave one apparent product, which was crystallized from
EtOH (2 days at -20 °C): yield 70 mg (93%), mp 114-116 °C,
86 °C, [R]20 -14.5° (c 1).
D
The mother liquor from the crystallization was evaporated.
The residue was subjected to column chromatography (100 g
SiO2, 1:2 EtOAc/hexane) to give the â-anomer: yield 1.4 g (5.7
mmol, 39%). Recrystallization from (i-Pr)2O at -78 °C afforded
[R]20 +225° (c 0.5). This product was an anomeric mixture
D
(NMR). A sample of the â-anomer was isolated by flash
chromatography (5 g SiO2, EtOAc/hexane: [1:6] 20 mL, [3:4]
20 mL, [1:1] 20 mL) as an oil, which aided in the NMR
spectroscopic assignment of both anomers in the mixture.
3,4,6-Tetr a -O-a cetyl-2-C-(4,6-d i-O-a cetyl-2,3-d id eoxy-r-
D-th r eo-h ex-2-en op yr a n osyl)-2-d eoxy-r-D-ga la ctop yr a n o-
syl Cya n id e (19). Compound 424 (3.0 g, 5.51 mmol) was
completely consumed after 36 h at rt [TLC (4:3 EtOAc/hexane)
and gave an anomeric mixture: yield 2.28 g (4.46 mmol, 81%,
R/â ratio 10:1, Rf 0.74 and 0.54, 4:3 EtOAc/hexane) without
apparent side products. The R-anomer was crystallized from
i-PrOH to give white crystals: yield 2.0 g (79%), mp 94-96
colorless crystals: mp 60-60.5 °C, [R]20 +192° (c 1).
D
Gen er a l Syn th etic P r oced u r e for 2,3-En op yr a n osyl
Cya n id es (14-17). The suspension of the glycal (7.6 mmol),
Hg(CN)2 (9.7 g, 38 mmol), HgBr2 (0.2 g, 0.6 mmol), and
TMSCN (1.4 mL, 10 mmol) in absolute THF (16 mL) and
absolute acetonitrile (30 mL) was warmed (60 °C bath) for 4
h. The solvents were removed in vacuo. The suspension of the
residue in CH2Cl2 (20 mL) was filtered through Celite to give
the products’ solution, which was subjected to flash chroma-
tography on silica gel with hexane/EtOAc gradients specified
for the individual preparations.
6-O-Acetyl-4-(2,3,4,6-tetr a-O-acetyl-â-D-glu copyr an osyl)-
1,2,3-tr id eoxy-â-D-2-en op yr a n osyl Cya n id e (14). From 7
(4.25 g, 7.6 mmol). Flash-column chromatography (120 g SiO2,
EtOAc/hexane: [1:6] 300 mL, [3:4] 300 mL, [1:1] 300 mL) of
the product solution gave the â-anomer, which crystallized
upon removal of the solvent: yield 2.1 g (4.0 mmol, 52%), mp
°C, [R]25 -125° (c 1).
D
4-O-Acetyl-2-C-(4-O-a cetyl-2,3-d id eoxy-r-L-th r eo-h ex-2-
en op yr a n osyl)-2,3-d id eoxy-2-en o-r-L-fu cop yr a n osyl Cya -
n id e (20). Crude 624 (2.3 g, 5.37 mmol, R/â 4:1) was completely
consumed after 36 h at rt. Two products [TLC (4:3 EtOAc/
hexane)] were observed: yield 0.61 g (1.82 mmol, 34%, R/â 10:
1). The R-anomer was crystallized from MeOH: yield 0.59 g
142-144 °C, [R]25 +10° (c 1), Rf 0.55 (9:1 CH2Cl2/EtOAc).
D
6-O-Acetyl-4-(2,3,4,6-tetr a -O-a cetyl-â-D-ga la ctop yr a n o-
syl)-1,2,3-tr ideoxy-r/â-D-2-en opyr an osyl Cyan ide (15). From
9 (2.3 g, 4.1 mmol). Flash-column chromatography (90 g SiO2,
EtOAc/hexane: [1:6] 300 mL, [3:4] 300 mL, [1:1] 300 mL) of
the product solution gave an R/â-mixture: yield 1.42 g (2.7
mmol, 65%). The predominant â-cyanide (3:1, 0.3 g) could be
crystallized from i-PrOH (3 mL): mp 147-149 °C, [R]25D +134°
(c 1), Rf 0.73 (9:1 CH2Cl2/EtOAc). A later fraction gave the
R-anomer as an oil (0.11 g): [R]25D +16° (c 1), Rf 0.69 (9:1 CH2-
Cl2/EtOAc).
(30%), mp 137-139 °C, [R]25 +413° (c 1).
D
3,6-Di-O-a cet yl-2-C-[6-O-a cet yl-2,3-d id eoxy-4-(2,3,4,6-
t e t r a -O -a c e t yl-â-D -glu c op y r a n o syl)-r-D -er yt h r o-h e x -
2-en op yr a n osyl]-4-(2,3,4,6-t et r a -O-a cet yl-â-D-glu cop y-
r a n osyl)-1,2-d id eoxy-r-D-glu cop yr a n osyl Cya n id e (21).
Compound 8 (80 mg, 0.07 mmol) was completely consumed
after 24 h [TLC (7:3 EtOAc/hexane)]. Removal of the solvent
in vacuo gave a syrup, which was digested with dry Et2O (5
mL) for 2 h at reflux to give the R-anomer as colorless, cotton-
like crystals: yield 40 mg (52%), mp 173-175 °C, [R]25 -16°
D
6-O-Acetyl-4-(2,3,4,6-tetr a-O-acetyl-r-D-glu copyr an osyl)-
1,2,3-tr id eoxy-r/â-D-2-en op yr a n osyl Cya n id e (16). From
11 (2.3 g, 4.1 mmol). Flash-column chromatography (90 g SiO2,
EtOAc/hexane: [1:6] 300 mL, [3:4] 300 mL, [1:1] 300 mL) of
the product solution gave an R/â-mixture (1:1): yield 1.12 g
(2.1 mmol, 52%). The R-anomer could be crystallized from
i-PrOH: yield 120 mg (0.23 mmol, 5.5%), colorless crystals,
mp 130-132 °C, [R]25D +105° (c 1), Rf 0.68 (9:1 CH2Cl2/EtOAc).
4-O-Acetyl-6-(2,3,4,6-tetr a-O-acetyl-â-D-glu copyr an osyl)-
1,2,3-tr id eoxy-r/â-D-2-en op yr a n osyl Cya n id e (17). From
12 (1.2 g, 2.1 mmol). Flash-column chromatography (90 g SiO2,
EtOAc/hexane: [1:6] 300 mL, [3:4] 300 mL, [1:1] 300 mL) of
the product solution gave a 4:1 R/â mixture: yield 0.68 g (1.3
(c 1).
Ack n ow led gm en t. We thank Dr. Allen D. Hunter
for helpful discussion of molecular structures. Financial
support by the Chemistry Department of the University
of the Pacific is gratefully acknowledged.
Su p p or tin g In for m a tion Ava ila ble: Complete NMR
assignment of compounds 8, 10, and 13-21, elemental analy-
ses, and mass spectrometric data. This material is available
free of charge via the Internet at http://pubs.acs.org.
J O0111661
J . Org. Chem, Vol. 67, No. 22, 2002 7669