Synthesis of 1,4-anhydro-D-xylitol heteroanalogues of the naturally occurring glycosidase inhibitor salacinol and their evaluation as glycosidase inhibitors

Article abstract of DOI:10.1139/v02-078

The syntheses of two 1,4-anhydro-D-xylitol heteroanalogues (8 and 9) of the naturally occurring sulfonium ion, salacinol (3), containing a sulfur or nitrogen atom in the ring are described. Salacinol (3) is one of the active principles in the aqueous extracts of Salacia reticulata that are traditionally used in Sri Lanka and India for the treatment of Type 2 diabetes. The synthetic strategy relies on the nucleophilic attack of sulfur or nitrogen analogues of 1,4-anhydro-D-xylitol at the least-hindered carbon of 2,4-O-benzylidene-L-erythritol-1,3-cyclic sulfate. The sulfonium ion 8 inhibited barley-α-amylase (AMY1) and porcine pancreatic-α-amylase (PPA), with K_i values of 109 ± 11 and 55 ± 5 μM, respectively. In contrast, the ammonium ion 9 showed no significant inhibition of either AMY1 or PPA. Compounds 8 and 9 also showed no significant inhibition of glucoamylase.

Full text of DOI:10.1139/v02-078

937
Synthesis of 1,4-anhydro-D-xylitol heteroanalogues
of the naturally occurring glycosidase inhibitor
salacinol and their evaluation as glycosidase
inhibitors
Ahmad Ghavami, Blair D. Johnston, Matthew D. Maddess, Sarah M. Chinapoo,
Morten T. Jensen, Birte Svensson, and B. Mario Pinto
Abstract: The syntheses of two 1,4-anhydro-D-xylitol heteroanalogues (8 and 9) of the naturally occurring sulfonium
ion, salacinol (3), containing a sulfur or nitrogen atom in the ring are described. Salacinol (3) is one of the active
principles in the aqueous extracts of Salacia reticulata that are traditionally used in Sri Lanka and India for the treat-
ment of Type 2 diabetes. The synthetic strategy relies on the nucleophilic attack of sulfur or nitrogen analogues of
1,4-anhydro-D-xylitol at the least-hindered carbon of 2,4-O-benzylidene-L-erythritol-1,3-cyclic sulfate. The sulfonium
ion 8 inhibited barley-ꢀ-amylase (AMY1) and porcine pancreatic-ꢀ-amylase (PPA), with Ki values of 109 ± 11 and
55 ± 5 M, respectively. In contrast, the ammonium ion 9 showed no significant inhibition of either AMY1 or PPA.
Compounds 8 and 9 also showed no significant inhibition of glucoamylase.
Key Words: glycosidase inhibitors, salacinol analogues, anhydro-D-xylitol heteroanalogues, enzyme inhibition.
Résumé : On décrit les synthèses de deux hétéroanalogues du 1,4-anhydro-D-xylitol (8 et 9) qui comportent un atome
de soufre ou d’azote dans le cycle de l’ion sulfonium, salacinol (3), que l’on retrouve dans la nature. Le salacinol (3)
est un des principes actifs que l’on retrouve dans la phase d’extraction aqueuse du Salacia reticulata utilisée au Sri-
Lanka et aux Indes dans le traitement du diabète de type 2. La stratégie de la synthèse repose sur une attaque
nucléophile d’analogues soufrés ou azotés du 1,4-anhydro-D-xylitol au niveau du carbone le moins encombré du sulfate
1,3-cyclique du 2,4-O-benzylidène-L-érythritol. L’ion sulfonium 8 inhibe l’ꢀ-amylase de l’orge (AMY1) et de l’ꢀ-amy-
lase du pancréas de porc (PPA) avec des valeurs de Ki de 109 ± 11 et 55 ± 5 M, respectivement. Par ailleurs, l’ion
ammonium 9 ne provoque pas d’inhibition significative de l’AMY1 ou de la PPA. Les composés 8 et 9 n’inhibent pas
de façon significative la glucoamylase.
Mots clés : inhibiteurs de glycosidase, analogues du salacinol, hétéroanalogues de l’anhydro-D-xylitol, inhibition
enzymatique.
[Traduit par la Rédaction] Ghavami et al. 942
Introduction
body directed against the Shigella flexneri Y O-antigen (2).
In the latter study, the crystal structures of the Ab-peptide
mimetic and Ab-pentasaccharide complexes were compared.
The results indicated that although both ligands engage
some common groups on the Ab receptor in H-bonding and
hydrophobic interactions, each ligand also displays unique
interactions with groups on the protein, lending support to
our previous hypothesis that “functional” and not “struc-
tural” mimicry might be the mode of mimicry with peptide
mimetics (1).
We have also studied the mimicry of carbohydrates by
glycomimetics as potential glycosidase inhibitors. Thus, for
example, we have described the synthesis and conforma-
tional analysis of a sulfonium-ion analogue (1) of the
glycosidase inhibitor castanospermine (2) (3). Our reasoning
was inspired by the pioneering work of the late B. Belleau
who synthesized sulfonium-ion analogues of the morphinans,
levorphanol, and isolevorphanol, and showed that they were
agonists or antagonists of morphine for the opiate receptor
(4a–d). Recently, a new class of glycosidase inhibitor with
an intriguing inner-salt sulfonium-sulfate structure was
A program of research to investigate the nature and origin
of carbohydrate mimicry is in progress in our laboratory.
Thus, we have recently reported the study of the peptide
mimicry of carbohydrates recognized by antibodies directed
against the Group A Streptococcus cell-wall polysaccharide
(1) and have recently communicated our results with an anti-
Received 21 January 2002. Published on the NRC Research
Press Web site at http://canjchem.nrc.ca on 19 July 2002.
This work is dedicated, with respect and affection, to the
memory of R.U. Lemieux.
A. Ghavami, B.D. Johnston, M.D. Maddess, S.M. Chinapoo,
B.M. Pinto.1 Department of Chemistry, Simon Fraser University,
Burnaby, BC V5A 1S6, Canada.
M.T. Jensen and B. Svensson. Department of Chemistry,
Carlsberg Laboratory, Gamle Carlsberg Vej 10, DK-2500,
Valby, Copenhagen, Denmark.
1Corresponding author (e-mail: bpinto@sfu.ca).
Can. J. Chem. 80: 937–942 (2002)
DOI: 10.1139/V02-078
© 2002 NRC Canada

938
Can. J. Chem. Vol. 80, 2002
Scheme 1.
isolated from the roots and stems of the plant Salacia
reticulata. Extracts of this plant have been traditionally used
in the Ayurdevic method of Indian medicine as a treatment
OH
OH
OH
4'
3'
X
OSO3
X
2'
OH
OH
1'
HO
RO
L
OSO3
OH
HO
OH
(A)
RO
R = Bn, H
OR
(B)
HO
HO
S
OSO3
(C)
ClO4
S
N
HO
HO
HO
X = S , NH
HO
OH
OH
2'
1'
HO
OH
O
O
Ph
O
O
HO
HO
O
O
O
O
S
S
1
3
2
4'
3'
10
O
(D)
O
for diabetes. One of the active ingredients of these extracts
is the sulfonium salt salacinol (3) (5).
Scheme 2.
We (6) and others (7) have recently reported the synthesis
of salacinol (3) and its stereoisomers (4, 5), and provided
conclusive proof of structure of the natural product. We have
also reported the syntheses of the hitherto unknown nitrogen
congeners (6, 7) as potential glycosidase inhibitors (8). En-
zyme inhibition assays indicated that salacinol (3) is a weak
(Ki = 1.7 mM) inhibitor of glucoamylase, whereas com-
pounds 6 and 7 inhibit glucoamylase with Ki values in the
range about 10-fold higher. The nitrogen analogues 6 and 7
showed no significant inhibitory effect of either barley
ꢀ-amylase (AMY1) or porcine pancreatic ꢀ-amylase (PPA)
at concentrations of 5 mM. In contrast, salacinol (3) inhib-
OH
4 steps ref(10)
OMs
OH
OMs
MsCl
BnO
L-arabinose
BnO
Pyridine
BnO
OBn
BnO
OBn
12
11
NaN3 / DMF
Na2S / DMF
H
N
OMs
S
N3
HO
BnO
BnO
H2, Pd/C
HO
OH
BnO
OBn
BnO
OBn
15
14
13
enzymes. Therefore, to probe these structure–function stud-
ies further, we now report the syntheses of the thio and
iminoxylitol analogues (8, 9) of salacinol (3) and their eval-
uation as glycosidase inhibitors of AMY1, PPA, or
glucoamylase.
OH
OH
OH
OH
S
OSO3
S
OSO3
HO
HO
HO
OH
OH
HO
OH
OH
OH
OH
4'
4'
5
4
3'
1'
3'
1'
2'
2'
OH
OH
5
5
NH OSO3
1
S
OSO3
1
2
HO
HO
4
4
2
OH
9
OH
OH
3
3
NH OSO3
HO
NH OSO3
HO
OH
HO
HO
8
HO
OH
HO
OH
7
6
Results and discussion
ited AMY1 and PPA in the micromolar range, with Ki values
of 15 ± 1 M and 10 ± 2 M, respectively (8).
Yuasa and et al. (9) have also investigated the glucosidase
inhibitory activities of compounds (3) and (5) and showed
Retrosynthetic analysis indicated that salacinol (3) or its
analogues (A) could be obtained by alkylation of anhydro-
alditol derivatives at the ring heteroatom (Scheme 1). As in
our previous work (6, 8), the benzylidene acetal 10 of D was
chosen as the alkylating agent. We envisaged that selective
attack of the heteroatom at the least-hindered primary
centre would afford the desired sulfonium or ammonium
ions.
The cyclic sulfate (10) was synthesized in five steps start-
ing from L-glucose (6). The thio- and iminoxylitols (13, 15)
were synthesized from L-arabinose. Thus, the diol (11) was
synthesized from L-arabinose in four steps according to the
procedure used by van der Klein et al. (11) to synthesize its
enantiomer (Scheme 2). Treatment of the diol (11) with
methanesulfonyl chloride in pyridine then afforded the di-
mesylated compound (12) (88% yield). Compound (12) was
used as a key intermediate to synthesize both the thio- and
iminoxylitols. Treatment of (12) with sodium sulfide in
DMF produced compound (13) in 95% yield, whereas
treatment with sodium azide in DMF followed by
that although salacinol (3) is a better inhibitor (IC50
1.1 M) of rice ꢀ-mannosidase than its diastereomer 5 (IC50
=
=
0.38 mM), the inhibitory activities are comparable for al-
mond ꢀ-glucosidase (IC50 = 2.1 mM for 3; 3.6 mM for 5). In
the case of almond ꢁ-glucosidase, (5) is a better inhibitor
than (3) ((3) showed no activity; IC50 = 3.4 mM for (5)).
In a recent study, Muraoka et al. (10) showed that
salacinol (3) inhibits intestinal ꢀ-glucosidases: maltase,
sucrase, and isomaltase with IC50 values of 9.6 M, 2.5 M,
and 1.8 M, respectively, whereas compound (6) inhibits
these enzymes with IC50 values of 306 M, 44 M, and
136 M, respectively.
The results described above suggest that the stereo-
chemistry at the different stereogenic centres and the nature
of the ring heteroatom in the candidate inhibitors play a sig-
nificant role in discriminating between different glycosidase
© 2002 NRC Canada

Ghavami et al.
939
Scheme 3.
Table 1. Ki (mM) values of compounds 1 and 3–9 against barley
ꢀ-amylase (AMY1), porcine pancreatic ꢀ-amylase (PPA), and
glucoamylase.
Ph
O
O
OH
Compound
AMY1
PPA
Glucoamylase
OH
S
OSO3
PdOH/C
AcOH
S
OSO3
1
3
4
5
6
7
8
9
>5
0.015
>5
>5
>5
>5
0.109
>5
>5
0.01
>5
>5
>5
>5
0.052
>5
1.32
1.71
2.17
1.06
>2.5
>8
10
13
+
BnO
HO
BnO
OBn
HO
OH
OH
16
8
Ph
O
O
>5
>30
OH
NH OSO3
Pd/C
NH OSO3
10
15
+
HO
HO
AcOH
tively, as compared to salacinol (3), with Ki values of 15 ± 1
and 10 ± 2 M, respectively. In contrast, the ammonium ion
9 showed no significant inhibition of either AMY1 or PPA
(Table 1). It would appear then that analogues 8 and 9 and
salacinol (3) show discrimination or selectivity for certain
glycosidase enzymes, and further testing against a wider
panel of enzymes that includes human small intestinal
maltase–glucoamylase (16) and human pancreatic ꢀ-amylase
(17) is planned to map the enzyme selectivity profiles of
these compounds.
HO
OH
HO
OH
17
9
hydrogenolysis afforded the iminoxylitol (15) in 55% yield
for the two steps (12).
To synthesize the target sulfonium ion 8, compound 16
was first synthesized by alkylation of 1,4-anhydro-2,3,5-tri-
O-benzyl-4-thio-D-xylitol (13) with the cyclic sulfate 10
(1.2 equiv) in acetone containing K2CO3 at 60–70°C in 72%
yield. Compound 16 was obtained as the sole coupled prod-
uct (Scheme 3). The stereochemistry at the stereogenic sul-
fonium centre in 16 was established by means of a NOESY
experiment. Thus, a correlation between H-1> and H-4, con-
firmed the trans relationship between the erythritol side
chain and C-4 substituent on the anhydroxylitol moiety,
which is similar to the stereochemistry at the stereogenic
sulfur atom in salacinol (3). Deprotection of 16 by hydro-
genolysis over a palladium hydroxide catalyst on carbon
was problematic because of poisoning of the catalyst but af-
forded compound 8 in 50% yield.
The corresponding nitrogen congener 17 was synthesized
in an analogous manner although, in this case, the increased
nucleophilicity of the nitrogen atom did not necessitate
benzylation of the hydroxyl groups. Thus, alkylation of 1,4-
dideoxy-1,4-imino-D-xylitol (15) with the cyclic sulfate 10
(1.2 equiv) in methanol containing K2CO3 at 60–70°C af-
forded compound 17 in 63% yield. The stereochemistry at
the stereogenic nitrogen centre in 17 was established by
means of a NOESY experiment, as above. In this case, a
correlation between H-1> and H-3, confirmed the trans rela-
tionship between the erythritol side chain and the C-3
substituent on the iminoxylitol moiety. Deprotection of 17
by hydrogenolysis over a Pd/C catalyst afforded compound 9
in 83% yield.
Experimental section
General
1
Optical rotations were measured at 23°C. H and 13C
NMR spectra were recorded at 400.13 and 100.6 MHz. All
assignments were confirmed with the aid of two-dimensional
1
1
1H, H (COSYDFTP), or H, 13C (INVBTP) experiments us-
ing standard Bruker pulse programs. Column chromatogra-
phy was performed with Merck Silica gel 60 (230–400 mesh).
High resolution mass spectra were measured with liquid sec-
ondary ionization fast atom bombardment (LSI-MS (FAB)),
run on a Kratos Concept H double focussing mass spectrom-
eter at 10 000 RP, using meta-NO2-benzyl alcohol as the ma-
trix.
Enzyme inhibition assays
The glucoamylase G2 form from Aspergillus niger was
purified from a commercial enzyme (Novo Nordisk,
Bagsvaerd, Denmark) as described (13, 14). The initial rates
of glucoamylase G2-catalyzed hydrolysis of maltose was
tested with 1 mM maltose as substrate in 0.1 M sodium ace-
tate pH 4.5 at 45°C using an enzyme concentration of 7.0 ×
10–8 M and five inhibitor concentrations in the range 1 m –
5 mM. The effect of the inhibition on rates of substrate hy-
drolysis were compared for the different compounds. The
glucose released was analyzed in aliquots removed at appro-
priate time intervals using a glucose oxidase assay adapted
to microtiter plate reading and using a total reaction volume
for the enzyme reaction mixtures of 150 or 300 L (18). The
Ki values were calculated assuming competitive inhibition
from 1/v = (1/Vmax) + [(Km)/(Vmax[S]Ki)] × [I], where v is the
rate measured in the presence or absence of inhibitor, [I] and
[S] the concentrations of inhibitor and substrate, Km 1.6 mM,
and kcat 11.3 s–1, using ENZFITTER (19).
Enzyme inhibition assays
Compounds 8 and 9 were tested for their inhibition of
three glycosidase enzymes, namely glucoamylase G2, (13,
14) porcine pancreatic ꢀ-amylase, and barley ꢀ-amylase (15).
The effects were compared to those of salacinol (3). Gluco-
amylase G2 was weakly inhibited by salacinol (3) (Ki =
1.7 mM) whereas compounds 8 and 9 showed no significant
inhibition of glucoamylase. The sulfonium ion 8 inhibited
barley ꢀ-amylase (AMY1) and porcine pancreatic ꢀ-amylase
(PPA), with Ki values of 109 ± 11 and 55 ± 5 M, respec-
© 2002 NRC Canada

940
Can. J. Chem. Vol. 80, 2002
1
Porcine pancreatic ꢀ-amylase (PPA) and bovine serum al-
(c 1.5, CH2Cl2). H NMR (CD2Cl2) ꢂ: 7.42–7.20 (15H, m,
Ar), 5.01 (1H, ddd, J4,5b = 6.8, J4,5a = J3,4 = 3.2 Hz, H-4),
4.72 and 4.59 (2H, 2d, JA,B = 11.1 Hz, CH2Ph), 4.65 and
4.61 (2H, 2d, JA,B = 11.0 Hz, CH2Ph), 4.53 (2H, s, CH2Ph),
4.34–4.29 (2H, m, H-1a, H-1b), 3.96–3.91 (2H, m, H-2, H-3),
3.87 (1H, dd, J5a,5b = 11.3 Hz, H-5a), 3.81 (1H, dd, H-5b),
3.00 (3H, s, OSO2CH3), 2.93 (3H, s, OSO2CH3). 13C NMR
(CD2Cl2) ꢂ: 137.95, 137.87, 137.72 (3Cipso), 128.84–128.29
(15CAr), 81.52 (C-4), 77.98 (C-2), 77.32 (C-3), 74.75,
73.96, 73.77 (3CH2Ph), 6916 (C-5), 68.64 (C-1), 39.04
(OSO2CH3), 37.67(OSO2CH3). Anal. calcd. for C28H34O9S2:
C 58.12, H 5.92; found: C 58.21, H 6.02.
bumin (BSA) were purchased from Sigma. Amylose EX-1
(DP17; average degree of polymerization 17) was purchased
from Hayashibara Chemical Laboratories (Okayama, Japan).
Recombinant barley ꢀ–amylase isozyme 1 (AMY1) was pro-
duced and purified as described (15). An aliquot of the por-
cine pancreatic ꢀ-amylase (PPA) crystalline suspension (in
ammonium sulfate) was dialyzed extensively against the as-
say buffer without BSA. The enzyme concentration was de-
termined by aid of amino acid analysis as determined using
an LKB model Alpha Plus amino acid analyzer. The inhibi-
tion of AMY1 (3 × 10–9 M) and PPA (9 × 10–9 M) activity
towards DP17 amylose was measured at 37°C in 20 mM so-
dium acetate (pH 5.5, 5 mM CaCl2, 0.005% BSA (for
AMY1)) and 20 mM sodium phosphate (pH 6.9, 10 mM
NaCl, 0.1 mM CaCl2, 0.005% BSA (for PPA)). Six different
final inhibitor concentrations were used in the range 1 M –
5 mM. The inhibitor was preincubated with enzyme for
5 min at 37°C before addition of substrate. Initial rates were
determined by measuring reducing sugar by the copper-
bicinchoninate method as described (15, 20). The Ki values
were calculated assuming competitive inhibition, as de-
scribed above for the case of glucoamylase, and a Km of
0.57 mg mL–1 and kcat of 165 s–1 for AMY1 and 1 mg mL–1
and 1200 s–1 for PPA, as determined in the substrate concen-
tration range 0.03–10 mg mL–1 using ENZFITTER (19). For
the Ki determinations, [S] = 0.7 mg mL–1 amylose DP 17 for
the AMY1 binding and [S] = 2.5 mg mL–1 amylose DP 17
for the PPA binding.
1,4-Anhydro-2,3,5-tri-O-benzyl-4-thio-D-xylitol (13)
Compound 12 (1.6 g, 2.8 mmol) was dissolved in DMF
(10 mL) and Na2S·H2O (1.1 g, 1.5 equiv) was added. The
mixture was stirred at 100°C until TLC (hexanes–EtOAc,
4:1) showed complete disappearance of the starting material.
The solvent was removed under high vacuum, and the resi-
due was dissolved in EtOAc (100 mL) and washed with H2O
(30 mL). The organic phase was dried (Na2SO4) and concen-
trated on a rotary evaporator. The product was purified by
flash chromatography (hexanes–EtOAc, 4:1) to give a
colourless syrup (1.1 g, 95%). [ꢀ]2D2 +67° (c 1.3, CH2Cl2).
1H NMR (CD2Cl2) ꢂ: 7.37–7.25 (15H, m, Ar), 4.57–4.47
(6H, m, 3CH2Ph), 4.21 (1H, ddd, H-2), 4.13 (1H, dd, J2,3
3.8, J3,4 = 3.5 Hz, H-3), 3.85 (1H, dd, J4,5a = 7.5 Hz, H-5a),
3.80 (1H, ddd, H-4), 3.60 (1H, dd, J5a,5b = 8.2, J4,5b
=
=
5.5 Hz, H-5b), 3.10 (1H, dd, J1a,1b = 11.4, J1a,2 = 4.4 Hz,
H-1a), 2.85 (1H, dd, J1b,2 = 2.1 Hz, H-1b). 13C NMR
(CD2Cl2) ꢂ: 138.87, 138.56 (3Cipso), 128.70–127.86 (15CAr),
83.52 (C-3), 83.47 (C-2), 73.52, 72.95, 71.60 (3CH2Ph),
69.85 (C-5), 48.55 (C-4), 33.46 (C-1). Anal. calcd. for
C26H28O3S: C 74.25, H 6.71; found: C 74.05, H 6.63.
2,3,5-Tri-O-benzyl-L-arabinitol (11)
The diol 11 was synthesized from L-arabinose according
to the procedure used by van der Klein et al. (11) to synthe-
size its enantiomer. [ꢀ]2D2 –4.6° (c 1.0, CH2Cl2) (lit. (10)
1
value +6.8° (c 1, CHCl3)). H NMR (CD2Cl2) ꢂ: 7.38–7.25
(15H, m, Ar), 4.65 and 4.61 (2H, 2d, JA,B = 11.4 Hz,
CH2Ph), 4.60 and 4.57 (2H, 2d, JA,B = 11.2 Hz, CH2Ph),
4.55 and 4.51 (2H, 2d, JA,B = 11.9 Hz, CH2Ph), 4.00 (1H,
dddd, H-4), 3.81–3.74 (3H, m, H-2, H-1a, H-1b), 3.70 (1H,
dd, J3,4 = 7.0, J2,3 = 3.6 Hz, H-3), 3.67 (1H, dd, J5a,5b = 12.2,
J4,5a = 3.9 Hz, H-5a), 3.63 (1H, dd, J4,5b = 5.2 Hz, H-5b),
2,3,5-Tri-O-benzyl-1,4-dideoxy-1,4-[[(2>S,3>S)-2>,4>-O-
benzylidene-3>-(sulfooxy)butyl]-episulfoniumylidene]-D-
xylitol inner salt (16)
A mixture of the thioxylitol 13 (100 mg, 0.24 mmol) and
2,4-O-benzylidene-L-erythritol-1,3-cyclic sulfate (10) (80 mg,
1.2 equiv) was dissolved in dry acetone (0.5 mL) and anhyd
K2CO3 (15 mg) was added. The mixture was stirred in a
sealed tube in an oil bath (60–70°C) overnight. The solvent
was removed under reduced pressure and column chroma-
tography (CHCl3–MeOH, 10:1 + 0.1% Et3N) of the crude
product gave an amorphous solid (120 mg, 72%). [ꢀ]D22
+28° (c 0.3, CH2Cl2). 1H NMR (CD2Cl2) ꢂ: 7.54–7.07
(20H, m, Ar), 5.52 (1H, s, CHPh), 4.60 and 4.49 (2H, 2d,
JA,B = 11.7 Hz, CH2Ph), 4.57–4.47 (2H, m, H-3>, H-4> eq),
4.47 and 4.44 (2H, 2d, JA,B = 11.3 Hz, CH2Ph), 4.41 (1H,
dd, H-1>a), 4.41–4.36 (2H, m, H-2, H-3), 4.27 (1H, ddd,
2.90 (1H, d, JOH,4 = 5.4 Hz, 4-OH), 2.20 (1H, dd, JOH,1a
=
6.1, JOH,1b = 4.6 Hz, 1-OH). 13C NMR (CD2Cl2) ꢂ: 138.68,
138.58 (3Cipso), 128.76–128.07 (15CAr), 80.08 (C-2), 79.16
(C-3), 74.14, 73.72, 73.15 (3CH2Ph), 71.63 (C-5), 70.84 (C-
4), 61.82 (C-1).
2,3,5-Tri-O-benzyl-1,4-di-O-methanesulfonyl-L-arabinitol
(12)
To a solution of the diol 11 (4.0 g, 9.5 mmol) in pyridine
(20 mL) at 0°C was added a solution of methanesulfonyl
chloride (1.8 mL, 2.5 equiv) in pyridine (3.0 mL). Stirring
was continued at 0°C, and under an N2 atmosphere until
TLC (hexanes–EtOAc, 3:2) showed complete disappearance
of the starting material. The solvent was removed under high
vacuum, and the residue was dissolved in CH2Cl2 (100 mL)
and washed with 1 M aq HCl (2 × 30 mL), H2O (30 mL),
and sat. aq NaHCO3 (30 mL), and dried (Na2SO4). The solu-
tion was concentrated on a rotary evaporator and the product
was purified by flash chromatography (hexanes–EtOAc, 3:2)
to give 12 as a colourless oil (4.8 g, 88%). [ꢀ]2D2 +4.1°
J2>,3> = 9.1, J1>a,2> = J1>b,2> = 3.3 Hz, H-2>), 4.09 (1H, ddd, J4,5a
9.4, J4,5b = 6.0, J3,4 = 3.2 Hz, H-4), 4.02 and 3.96 (2H, 2d,
A,B = 11.4 Hz, CH2Ph), 3.92 (1H, dd, J1>b,1>a = 13.6 Hz, H-1>b),
3.87 (1H, dd, J1a,1b = 14.4, J1a,2 = 3.1 Hz, H-1a), 3.82 (1H,
dd, J5a,5b = 9.4 Hz, H-5a), 3.76 (1H, dd, J4>ax,4>eq = J3>,4>ax
=
J
=
12.2 Hz, H-4> ax), 3.71 (1H, brd, H-1b), 3.67 (1H, dd, H-5b).
13C NMR (CD2Cl2) ꢂ: 137.42, 137.17, 136.46 (4Cipso),
129.82–126.65 (20CAr), 101.76 (CHPh), 82.64 (C-3), 81.74
(C-2), 76.63 (C-2>), 73.92, 73.84, 72.64 (3CH2Ph), 69.54
(C-4>), 66.56 (C-4), 66.40 (C-3>), 64.00 (C-5), 51.51 (C-1>),
© 2002 NRC Canada

Ghavami et al.
941
47.42 (C-1). Anal. calcd. for C37H40O9S2: C 64.14, H 5.82;
found: C 64.45, H 5.85.
phy (CHCl3–MeOH–H2O, 7:3:1) to give an amorphous solid
(49 mg, 83%). [ꢀ]2D2 –5.3° (c 0.8, H2O). HR-FAB-MS calcd.
for C9H18O9SN (M + H): 318.0859; found: 318.0859.
1H NMR (D2O) ꢂ: 4.59–4.40 (4H, m, H-3, H-2, H-2>, H-3>),
4.27–4.17 (2H, m, H-5a, H-5b), 4.18–4.06 (2H, m, H-1a,
H-4), 4.11 (1H, dd, J4>a,4>b = 12.6, J3>,4>a = 2.0 Hz, H-4>a),
4.02 (1H, dd, J3>,4>b = 3.3 Hz, H-4>b), 4.11–3.95 (1H, m,
H-1>a), 3.68–3.57 (1H, m, H-1b), 3.57–3.45 (1H, m, H-1>b).
13C NMR (D2O) ꢂ: 82.52 (C-3>), 77.46 (C-3), 76.56 (C-2>),
73.20 (C-4), 68.42 (C-2), 63.09 (C-1), 62.38 (C-1>), 61.93
(C-4>), 59.38 (C-5).
1,4-Dideoxy-1,4-[[(2>S,3>S)-2>,4>-dihydroxy-3>-(sul-
fooxy)butyl]-episulfoniumylidene]-D-xylitol inner salt (8)
The protected compound 16 (150 mg, 0.22 mmol) was
dissolved in AcOH–H2O (4:1, 3 mL) and stirred with palla-
dium hydroxide catalyst on carbon (100 mg) under H2 (52
psi). After 72 h, the reaction mixture was filtered through a
pad of Celite, which was subsequently washed with MeOH.
The combined filtrates were concentrated and the residue
was purified by column chromatography (CHCl3–MeOH–
H2O, 7:3:1) to give an amorphous solid (36 mg, 50%). [ꢀ]D22
+20° (c 1.2, MeOH). HR-MS calcd. for C9H18O9S2 (M + H):
Acknowledgments
1
335.0471; found: 335.0471. H NMR (CD3OD) ꢂ: 4.64–4.59
We are grateful to Sidsel Ehlers for technical assistance
with the glucoamylase assays. We also thank the Natural
Sciences and Engineering Research Council of Canada
(NSERC) for financial support and the Michael Smith Foun-
dation for Health Research for a trainee fellowship (to AG).
(1H, m, H-2), 4.55 (1H, dd, J2,3 = J3,4 = 2.9 Hz, H-3), 4.37
(1H, ddd, H-4), 4.35 (1H, ddd, J1> a,2> = 3.5 Hz, H-2>), 4.25
(1H, ddd, J2>,3> = 7.2, J3>,4>a = J3>,4>b = 3.4 Hz, H-3>), 4.16
(1H, dd, J5a,5b = 11.7, J4,5a = 6.0 Hz, H-5a), 4.07 (1H, dd,
J4,5b = 9.0 Hz, H-5b), 3.95 (1H, dd, J1>a,1>b = 13.4 Hz, H-1> a)
3.93 (1H, dd, H-4> a), 3.91 (1H, dd, H-1a) 3.88 (1H, dd,
J1>b,2> = 8.0 Hz, H-1> b), 3.81 (1H, dd, J4 >b,4>a = 12.1 Hz, H-4
> b), 3.58 (1H, brd, J1b,1a = 13.7 Hz, H-1b). 13C NMR
(CD3OD) ꢂ: 81.05 (C-3>), 79.55 (C-2), 78.56 (C-3), 71.69
(C-4), 67.71 (C-2>), 61.66 (C-4>), 58.92 (C-5), 52.60 (C-1>),
49.26 (C-1).
References
1. S.L. Harris, L. Craig, J.S. Mehroke, M. Rashed, M.B. Zwick,
K. Kenar, E.J. Toone, N. Greenspan, F.-I. Auzanneau,
J.-R. Marino-Albernas, B.M. Pinto, and J.K. Scott. Proc. Natl.
Acad. Sci. U.S.A. 94, 2434 (1997).
2. B.M. Pinto, J.K. Scott, S.L. Harris, M.A. Johnson, D.R. Bundle,
M.C. Chervenak, M.N. Vyas, N.K. Vyas, and F.A. Quiocho.
XIX International Carbohydrate Symposium, San Diego,
U.S.A., (August 1998), Abstr. CP 132, CO 019; B.M. Pinto.
Meeting on anti-idiotypes and mimotopes in vaccine develop-
ment, Vibo Valentia, Italy, (May 2000); B.M. Pinto.
Glycobiology Symposium, Pacifichem 2000 Conference,
Honolulu, Hawaii, U.S.A., (December 2000) Abstr. 607.
3. L. Svansson, B.D. Johnston, J-H. Gu, B. Patrick, and B.M. Pinto.
J. Am. Chem. Soc. 122, 10 769 (2000).
4. (a) B. Belleau, U. Gulini, B.J. Gour-Salin, and F.R. Ahmed.
Can. J. Chem. 63, 1268 (1985); (b) B. Belleau, U. Gulini,
R. Camicioli, B.J. Gour-Salin, and G. Sauvé. Can. J. Chem.
64, 110 (1986); (c) S. Lemaire, B. Belleau, and F. Jolicoeur.
Adv. Biosci. (Oxford), 75, 105 (1989); (d) F.B. Jolicoeur,
D. Ménard, B. Belleau, and S. Lemaire. Int. Narcotics Res.
Conf. 89, 49 (1990).
5. M. Yoshikawa, T. Murakami, H. Shimada, H. Matsuda. J. Yamahara,
G. Tanabe, and O. Muraoka. Tetrahedron Lett. 38, 8367
(1997).
6. A. Ghavami, B.D. Johnston, and B.M. Pinto. J. Org. Chem. 66,
2312 (2001).
7. H. Yuasa, J. Takada, and H. Hashimoto. Tetrahedron Lett. 41,
6615 (2000).
8. A. Ghavami, B.D. Johnston, M. Jensen, B. Svensson, and
B.M. Pinto. J. Am. Chem. Soc. 123, 6268 (2001).
9. H. Yuasa, J. Takada, and H. Hashimoto. Bioorg. Med. Chem.
Lett. 11, 1137 (2001).
1>-((1,4-Dideoxy-1,4-imino-D-xylitol)-4-N-ammonium)-
2>,4>-O-benzylidene-1>-deoxy-L-erythritol-3>-sulfate (17)
A mixture of 1,4-dideoxy-1,4-imino-D-arabinitol (15)
(100 mg, 0.74 mmol) and 2,4-O-benzylidene-L-erythritol-
1,3-cyclic sulfate (10) (240 mg, 1.2 equiv) was dissolved in
dry MeOH (0.5 mL) and anhyd K2CO3 (15 mg) was added.
The mixture was stirred in a sealed tube in an oil bath
(60–70°C) overnight. The solvent was removed under re-
duced pressure, and column chromatography (CH2Cl2–MeOH,
5:1) of the crude product gave a white solid (191 mg, 63%)
that was recrystallized from methanol; mp 202–204°C, [ꢀ]D22
1
+30° (c 0.5, H2O). H NMR (D2O) ꢂ: 7.73–7.58 (5H, m,
Ar), 5.93 (1H, s, CHPh), 4.68 (1H, dd, J4>eq,4>ax = 11.0,
J
3>,4>eq = 5.5 Hz, H-4>eq), 4.54 (1H, brt, J2>,3> = J1>b,2 > = 9.9 Hz,
H-2>), 4.50–4.41 (3H, m, H-3, H-3>, H-2), 4.20 (1H, dd,
J5a,5b = 12.6, J4,5a = 5.5 Hz, H-5a), 4.18–4.12 (2H, m, H-1>a,
H-5b), 4.10 (1H, dd, J3>,4>ax = 11.0 Hz, H-4>ax), 4.11–4.01
(2H, m, H-1a, H-4), 3.66 (1H, br dd, J1>b,1>a = 11.6 Hz,
H-1>b), 3.56–3.50 (1H, brd, H-1b). 13C NMR (D2O) ꢂ:
138.50 (Cipso), 132.45 (Cpara), 131.26 (2C) and 128.62 (2C)
(Cortho and Cmeta), 103.38 (CHPh), 78.08 (C-2>), 77.38 (C-3),
76.70 (C-2), 73.31 (C-4), 70.77 (C-4>), 70.41 (C-3>), 63.18
(C-1), 60.01 (C-1>), 59.55 (C-5). Anal. calcd. for
C16H22O9SN: C 47.51, H 5.49, N, 3.47; found: C 47.29,
H 5.80, N 3.22.
1>-((1,4-Dideoxy-1,4-imino-D-xylitol)-4-N-ammonium)-1>-
deoxy-L-erythritol-3>-sulfate (9)
10. O. Muraoka, S. Ying, K. Yoshikai, Y. Matsuura, E. Yamada,
T. Minematsu, G. Tanabe, H. Matsuda, and M. Yoshikawa.
Chem. Pharm. Bull. 49, 1503 (2001).
11. P.A.M. van der Klein, A.E.J. de Nooy, G.A. van der Marel,
and J.H. van Boom. Synthesis, 347 (1991).
12. P.A.M. van der Klein, W. Filemon, H.J.G. Broxterman,
G.A. van der Marel, and J.H. van Boom. Synth. Commun. 22,
1763 (1992).
The protected compound 17 (75 mg, 0.18 mmol) was dis-
solved in AcOH–H2O (4:1, 5 mL) and stirred with Pd/C
(30 mg) under H2. After 16 h, the reaction mixture was
filtered through a pad of Celite, which was subsequently
washed with MeOH. The combined filtrates were concen-
trated and the residue was purified by column chromatogra-
© 2002 NRC Canada

942
Can. J. Chem. Vol. 80, 2002
13. B. Svensson, T. Pedersen, I. Svendsen, T. Sakai, and M. Ottesen.
Carlsberg Res. Commun. 47, 55 (1982).
17. C. Braun, G.D. Brayer, and S.G. Withers. J. Biol. Chem. 270,
26 778 (1995).
14. B. Stoffer, T.P. Frandsen, P.K. Busk, P. Schneider, I. Svendsen,
and B. Svensson. Biochem. J. 292, 197 (1993).
15. N. Juge, J.S. Andersen, D. Tull, P. Roepstorff, and B. Svensson.
Protein Expression Purif. 8, 204 (1996).
16. B.L. Nichols, J. Eldering, S. Avery, D. Hahn, A. Quaroni, and
E. Sterchi. J. Biol. Chem. 273, 3076 (1998).
18. T.P. Frandsen, C. Dupont, J. Lehmbeck, B. Stoffer, M.R. Sierks,
R.B. Honzatko, and B.Svensson. Biochemistry, 33, 13 808 (1994).
19. R.J. Leatherbarrow. Enzfitter, a nonlinear regression data anal-
ysis program for IBM PC. Elsevier Science Publishers BV,
Amsterdam. 1987.
20. J.D. Fox and J.F. Robyt. Anal. Biochem. 195, 93 (1991).
© 2002 NRC Canada