Aziridines as a structural motif to conformational restriction of azasugars.

Article abstract of DOI:10.1039/b210038j

In order to investigate the hypothesis that the glycosidase inhibitor isofagomine was bound to alpha- or beta-glucosidase in a 1,4B conformation, a number of bicyclic aziridines that adopt the 1,4B or B1,4 conformations were synthesised and investigated. (1R)-2-endo,3-exo-2,3-Dihydroxy-4-endo-4-hydroxymethyl-6- azabicyclo[3.1.0]hexane (5) and its N-methyl and N-benzyl analogues and (1S)-2-exo-3-endo-2,3-dihydroxy-4- endo-4-hydroxymethyl-6-azabicyclo-[3.1.0]hexane (6) were synthesised. The aziridines 5 and 6 were found to be weak or not inhibitors of alpha-glucosidase, beta-glucosidase and alpha-fucosidase.

Full text of DOI:10.1039/b210038j

View Article Online / Journal Homepage / Table of Contents for this issue
Aziridines as a structural motif to conformational restriction of
azasugars
Oscar Lopez Lopez,a,b José G. Fernández-Bolaños,*b Vinni H. Lillelunda and Mikael Bols*a
a Department of Chemistry, University of Aarhus, Langelandsgade 140, DK-8000 Aarhus,
Denmark. E-mail: mb@chem.au.dk; Fax: ϩ4586196199; Tel: ϩ4589423963
b Department of Organic Chemistry, Faculty of Chemistry, University of Seville, PO Box 553,
41071 Seville, Spain. E-mail: bolanos@us.es; Fax: ϩ34 954624960; Tel: ϩ34 954557150
Received 11th October 2002, Accepted 20th November 2002
First published as an Advance Article on the web 13th January 2003
In order to investigate the hypothesis that the glycosidase inhibitor isofagomine was bound to α-or β-glucosidase in a
1,4B conformation, a number of bicyclic aziridines that adopt the 1,4B or B1,4 conformations were synthesised and
investigated. (1R)-2-endo,3-exo-2,3-Dihydroxy-4-endo-4-hydroxymethyl-6-azabicyclo[3.1.0]hexane (5) and its
N-methyl and N-benzyl analogues and (1S)-2-exo-3-endo-2,3-dihydroxy-4-endo-4-hydroxymethyl-6-azabicyclo-
[3.1.0]hexane (6) were synthesised. The aziridines 5 and 6 were found to be weak or not inhibitors of α-glucosidase,
β-glucosidase and α-fucosidase.
Introduction
While the aziridine group is known as a useful reaction inter-
mediate,1 it is also an interesting structural motif in bioactive
compounds. The aziridine’s proton accepting properties, its
rigidity and its potential reactivity can all contribute to specific
molecular interactions with proteins, and indeed several
important natural products such as Mitomycin C,2 Porfiro-
mycin3 and Carzinophilin A4 contain the aziridine functional-
ity. A number of saccharide derivatives containing the aziridine
group have been made, mostly as intermediates,5–8 but also as
glycosidase inhibitors.9–11 The aziridines 19 and 210 have been
reported to be irreversible inhibitors, while 3 was recently
shown to be a reversible competitive inhibitor.11
Fig. 1
Fig. 2
Our long-standing interest in glycosidase inhibitors of the
isofagomine-type12 (4, Fig. 2) has led us to consider con-
formationally restrained analogues13 as a means of providing
information about the binding of these inhibitors. The isofago-
mines are strong β-glycosidase inhibitors.12 However stereo-
electronic effects dictate that β-glycosides must adopt a
boat-like transition state during hydrolysis (A, Fig. 2).14 It may
therefore be considered paradoxical that 4 and analogues, while
themselves in a chair conformation, are particularly potent
against β-glycosidases. This led us to consider whether 4 might
be binding in a boat conformation and/or whether 4b, the boat
conformer of 4, would be a good inhibitor. Another reason to
suppose that inhibitor 4 might not be binding to certain glyco-
sidases in the favoured chair conformation is the peculiar slow-
onset binding observed, particularly to β-glucosidase.18,19 The
association of 4 to β-glucosidase is much slower than the rate of
diffusion control one would normally expect for such a process.
The slow rate would however be consistent with an energetic-
ally unfavourable and little-populated conformation binding to
the enzyme. In this paper we analyse the problem by synthesis
of the bicyclic aziridine 5, which adopts the desired boat 1,4B
conformation, and mimics 4b. We also report the synthesis of
an isomer 6 that adopts the B1,4 conformation.
Results and discussion
Our synthetic plan to obtain 5 and 6 relied on the chemistry
of Ferrier et al., who synthesised aziridine 12 from methyl
-glucopyranoside (7) as outlined in Scheme 1.5 Conversion of
7 to a protected 2,6-ditosylate 8 followed by nucleophilic substi-
tution with iodine and reductive elimination with Zn powder
gave the alkenal 10, which was shown to undergo 1,3-dipolar
cycloaddition upon treatment with N-methylhydroxylamine
to give 11.5 Reductive cleavage of the N–O bond resulted in
spontaneous formation of 12.5 Since deprotection of 12 would
give us the N-methyl analogue of the desired compound, our
immediate goal was to modify Ferrier’s synthesis to reach 5.
Initially, we uneventfully repeated the sequence to 12 and found
that the benzoyl groups can be successfully removed with
NaOMe in methanol to give new aziridine 13 in a 78% yield.
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 4 7 8 – 4 8 2
T h i s j o u r n a l i s © T h e R o y a l S o c i e t y o f C h e m i s t r y 2 0 0 3
478

View Article Online
Scheme 1
Scheme 3
Scheme 4
Scheme 2
and J3,4 clearly identifying the boat conformation as the pre-
dominant one. Similarly the NMR data for 13 and 16 reveal the
same conformational preference of these compounds.
The more tricky synthesis of 5 was carried out as outlined in
Scheme 2. The alkenal 10 was reacted with N-benzylhydroxyl-
amine, which gave the 1,2-oxazine 14 in 61% yield. It was found
that yields were improved when CaCO3 and toluene were used
in place of pyridine in this reaction. Reduction of the N–O
bond was carried out with Raney nickel at 1 atm of hydrogen
pressure and it was possible to avoid too much N-debenzyl-
ation. The aziridine 15 was obtained in 48% yield after chrom-
atography. Debenzoylation using Zemplen conditions gave the
N-benzylaziridine 16 in 81% yield. Finally hydrogenolysis in the
presence of palladium on carbon gave the target compound 5 in
57% yield (Scheme 2). The aziridine is very sensitive, and the
modest yield in this reaction is due to partial reductive opening
of the aziridine.
We anticipated the isomer 6 might be synthesised from
the methyl -mannoside 17 in a similar manner since the
intramolecular 1,3-dipolar cycloaddition has been shown to
give opposite stereoselectivity in the mannose case relative
to glucose.15 However selective 2,6-ditosylation of 17 is not
possible.16 We therefore used Ley’s method17 of selectively
protecting a diequatorial 1,2-diol. Tosylation of the crude 18
gave the new ditosylate 19 in 56% yield from 17. Reaction
of 19 with NaI in acetic anhydride gave the 6-iodo compound
20 in 74% yield. Hydrolysis of the diacetal with TFA to
diol 21 and benzoylation gave the dibenzoate 22 in 92% yield
from 20.
The aziridine 5 may adopt either a conformation having
The elimination–cycloaddition of 22 with N-benzylhydroxyl-
amine proceeded satisfactorily. After treatment with Zn
powder the formation of alkenal 23 was observed by NMR.
6
the piperidine ring in the desired 3,6B or in a C3 conformation
(Fig. 2). The NMR spectrum of 5 shows large couplings for J2,3
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 4 7 8 – 4 8 2
479

View Article Online
Table 1 Ki values in µM at pH 6.8, 25 ЊC (— = not investigated, NI = no inhibition)
13
16
5
26
6
α-Glucosidase (yeast)
β-Glucosidase (almonds)
α-Fucosidase (bovine kidney)
4000
240
—
4900
280
—
NI
NI
—
NI
NI
NI
NI
1220a
2780a
a At 32 ЊC.
The reaction of 23 with N-benzylhydroxylamine in the presence
of CaCO3 resulted in 24 being obtained in 41% yield from 22.
No stereoisomers were observed.
The oxazine 24 was hydrogenolysed in the presence of Raney
nickel giving the aziridine 25 in 46% yield. The internal nucleo-
philic substitution is in this case a relatively slow reaction and
the initially formed monocyclic amine could be observed.
Debenzoylation with NaOMe–MeOH gave the unprotected
aziridine 26 in 73% yield. Finally hydrogenolysis of 26 gave 6 in
an 83% yield.
(1R)-2-endo-3-exo-2,3-Dihydroxy-4-endo-4-hydroxymethyl-
N-methyl-6-azabicyclo[3.1.0]hexane (13). To a solution of 125
(50 mg, 0.14 mmol) in dry methanol (3 mL) was added methan-
olic NaOMe (pH 10). Reaction was kept at rt for 1 h. Then it
was neutralised with IR-120(Hϩ) resin and the resin was washed
with 5% aqueous ammonia and concentrated to dryness to give
1
13 (17 mg; 78%). H NMR (200 MHz, D2O): δ 3.98 (m, 1H,
J2,3 = 6.9 Hz, H-2), 3.78 (m, 2H, H-7a, H-7b), 3.12 (t, 1H,
J3,4 = 6.9 Hz, H-3), 2.22 (br s, 2H, H-1, H-5), 2.14 (s, 3H, Me),
2.0 (m, 1H, H-4).13C NMR (50 MHz, D2O): δ 77.8, 75.3 (C-2,
C-3), 60.6 (C-7), 46.3, 45.1, 43.5, 42.4 (Me, C-1, C-4, C-5).
HRFAB-MS: calcd. for (M ϩ H)ϩ C7H14NO3: 160.0974, found:
160.0978.
The aziridine 6 may adopt a conformation with the piper-
3
idine ring in either a B3,6 or in a C6 conformation (Fig. 2),
and indeed both conformations appear likely. However the
NMR spectrum of 6 shows J2,3 = 0 Hz, which is inconsistent
with the chair conformation. The boat conformation must
therefore be predominant. The NMR data for 26 reveal the
same conformational preference.
(1R,5R)-6-exo,7-endo-6,7-Bis(benzoyloxy)-N-benzyl-8-exo-8-
toluene-p-sulfonyloxy-3-oxa-2-azabicyclo[3.3.0]octane (14). To a
suspension of methyl 3,4-di-O-benzoyl-6-deoxy-6-iodo-2-O-p-
toluenesulfonyl-α--glucopyranoside (9)5 (1.0 g, 1.50 mmol)
in aqueous EtOH (20 mL, 96%) was added Zn dust (1.0 g,
15.30 mmol). The mixture was refluxed for 1 h and then filtered
through Celite and concentrated to dryness to give a yellow
oil that was dissolved in CH2Cl2 (20 mL), washed with water
(2 × 10 mL), dried over Na2SO4, filtered and concentrated to
dryness to give alkenal 10 as an oil. To a solution of com-
pound 10 in toluene (8 mL) were added N-benzylhydroxyl-
amine hydrochloride (359 mg, 2.25 mmol) and CaCO3 (225 mg,
2.25 mmol), and the reaction was heated at 50 ЊC for 2 h. The
mixture was then concentrated to dryness, and the residue was
dissolved in CH2Cl2 (15 mL), washed with water (2 × 10 mL),
dried over Na2SO4, filtered and concentrated to dryness. Com-
pound 14 was crystallised from methanol (558 mg; 61%). Mp:
The unprotected azidirines 13, 16, 5, 26 and 6 were investi-
gated for their ability to inhibit α- and β-glucosidase (Table 1).
Weak competitive inhibition of α-glucosidase and intermediate
inhibition of β-glucosidase were found for compounds 13 and
16, but, surprisingly, no inhibition at all for the unsubstituted
aziridine 5. This suggests that these compounds bind distinctly
differently from isofagomine (4) and similar compounds,
because N-substitution of 4 decreases inhibition significantly.
Inhibition was not found to be time-dependent for any of the
compounds. Compounds 26 and 6 were also investigated
for inhibition of an α-fucosidase due to their resemblance to
-fucose. Weak reversible, competitive inhibition was found for
both compounds. Again the N-substituted aziridine 26 was
more potent than unsubstituted 6 suggesting that binding
was different from that of the corresponding isofagomine.
Altogether the inhibition study disproves the hypothesis that
1-azasugars bind in a 1,4B conformation.
In summary we have synthesised and investigated bicyclic
aziridines as conformationally restricted analogues of isofago-
mine in a 1,4B conformation. The aziridines are, in contrast to
isofagomine, very poor or not glycosidase inhibitors and appear
to bind to enzymes in a different mode than isofagomine. The
results show that isofagomine does not bind the investigated
glycosidases in the 1,4B conformation.
148–150 ЊC. [α]2D2 Ϫ18 (c 1.0, CH2Cl2). H NMR (300 MHz,
1
CDCl3): δ 7.96–7.06 (m, 19 H, Ar–H), 5.91 (t, 1 H, J6,7 = 8.2 Hz,
J7,8 = 8.2 Hz, H-7), 5.17 (dd, 1H, J5,6 = 6.3 Hz, H-6), 5.10
(dd, 1H, J1,8 = 5.6 Hz, H-8), 4.25 (dd, 1H, J4a,5 = 4.2 Hz,
J4a,4b = 9.6 Hz, H-4a), 4.21 (dd, 1H, J4b,5 = 6.9 Hz, H-4b),
3.95, 3.83 (2d, 1H each, 2JH,H = 13.5 Hz, CH2Ph), 3.84 (dd, 1H,
J1,5 = 9.6 Hz, H-1), 3.24 (m, 1H, H-5), 2.18 (s, 3H, Me). 13C
NMR (75.5 MHz, CDCl3): δ165.8, 164.6 (CO), 144.6, 136.2,
133.2, 133.0, 129.6, 129.5, 129.4, 128.7, 128.6, 128.2, 128.1,
128.0, 127.6, 127.3 (24 C, Ar), 83.5 (C-8), 78.7 (C-6), 76.3 (C-7),
70.4 (C-1), 69.8 (C-4), 59.0 (CH2Ph), 49.8 (C-5), 21.2 (Me).
HRMS (ES) calcd. for (M ϩ Na)ϩ C34H31NNaO8S: 636.1668,
found: 636.1667.
Experimental section
General
(1R)-2-endo,3-exo-2,3-Bis(benzoyloxy)-N-benzyl-4-endo-4-
hydroxymethyl-6-azabicyclo[3.1.0]hexane (15). To a solution of
14 (211 mg, 0.34 mmol) in acetone (5 mL) was added Raney
nickel and the mixture was hydrogenated at atmospheric pres-
sure and room temperature for 4 days. Then it was filtered
through a Celite bed, concentrated to dryness and purified by
Solvents were distilled under anhydrous conditions. All
reagents were used as purchased without further purification.
Pyridine was dried over potassium hydroxide before use.
Evaporation was carried out on a rotary evaporator with the
temperature kept below 40 ЊC. Glassware used for water-free
reactions was dried for 2 hours min. at 130 ЊC before use.
Columns were packed with silica gel 60 (230–400 mesh) as the
stationary phase. TLC-plates (Merck, 60, F254) were visualised
by spraying with cerium sulfate (1%) and molybdic acid (1.5%)
in 10 % H2SO4 and heating until coloured spots appeared.
1H-NMR, 13C-NMR and COSY were carried out on a Varian
Gemini 200 instrument. In water the water-signal (δ 4.7)
was used as reference. Mass spectra were carried out on a
Micromass LC-TOF instrument. Optical rotations are given in
column chromatography (CH2Cl2
CH2Cl2–MeOH 80 : 1
gradient) to yield 15 (73 mg; 48%). [α]2D5 Ϫ96 (c 1.0, CH2Cl2).
1H NMR (200 MHz, CDCl3): δ 8.07–7.08 (m, 15H, Ar–H),
5.59 (dd, 1H, J1,2 = 2,8 Hz, J2,3 = 6.2 Hz, H-2), 5.45 (dd, 1H,
J3,4 = 6.8 Hz, H-3), 3.91 (2dd, 2H, J4,7a = 4.6, J4,7b = 6.2 Hz,
J7a,7b = 11.0 Hz, H-7a, H-7b), 3.60, 3.25 (2d, 1H each,
2JH,H = 13.4 Hz, CH2Ph), 2.66 (dd, 1H, J1,5 = 5.2 Hz, J4,5 = 3.4
Hz H-5), 2.54–2.39 (m, 2H, H-1, H-4). 13C NMR (50 MHz,
CDCl3): δ 166.4, 166.3 (2CO), 138.4, 133.2, 133.0, 129.8, 129.7,
129.5, 128.3, 128.2, 127.4, 127.0 (18C, Ar), 79.8, 76.6 (C-2,
10Ϫ1 deg cm2 gϪ1
.
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 4 7 8 – 4 8 2
480

View Article Online
C-3), 62.3 (CH2Ph), 61.1 (C-7), 46.2, 42.8, 42.4 (C-1, C-4, C-5).
HRMS (ES) calcd. for (M ϩ Na)ϩ C27H25NNaO5: 466.1630,
found: 466.1629.
dissolved in CH2Cl2 (40 mL) and washed with saturated aque-
ous Na2S2O3 (2 × 20 mL), water (20 mL), dried over MgSO4,
filtered and concentrated to dryness. Crystallisation from
methanol yielded 20 (2.78 g; 74%). Mp: 132–136 ЊC. [α]2D2
=
(1R)-N-Benzyl-2-endo,3-exo-2,3-dihydroxy-4-endo-4-hydroxy-
methyl-6-azabicyclo[3.1.0]hexane (16). To a solution of 15
(69 mg, 0.16 mmol) in dry methanol (5 mL) was added methan-
olic NaOMe (pH 10). The reaction mixture was kept at rt for
3 h. Then it was neutralised with IR-120(Hϩ) resin and the resin
was washed with 5% aqueous ammonia and concentrated to
ϩ125 (c 1.1, CH2Cl2). 1H NMR (300 MHz, CDCl3): δ 7.84–7.24
(m, 4H, Ar–H), 4.88 (d, 1H, J1,2 = 1.5 Hz, H-1), 4.51 (dd, 1H,
J2,3 = 3.1 Hz, H-2), 3.93 (dd, 1H, J3,4 = 9.7 Hz, H-3), 3.69 (t, 1H,
J4,5 = 9.7 Hz, H-4), 3.62 (ddd, 1H, J5,6a = 2.1 Hz, J5,6b = 8.2 Hz,
H-5), 3.49 (dd, 1H, J6a,6b = 10.6 Hz), 3.40, 3.18 (s, 3 H each,
2 OMe), 3.17 (dd, 1H, H-6b), 3.03 (s, 3H, OMe), 2.41 (s, 3H,
CH3C6H4), 1.18, 0.98 (s, 3 H each, 2 Me). 13C NMR (75.5 MHz,
CDCl3): δ 144.4, 133.6, 129.3, 128.4 (6 C, Ar). 100.0, 99.8 (C-2Ј,
C-3Ј), 99.5 (C-1), 76.8 (C-2), 70.4 (C-5), 67.0 (C-4), 65.43 (C-3),
55.3, 48.2, 47.9 (3 OMe), 21.6 (CH3C6H4), 17.6, 17.3 (2 Me),
4.18 (C-6). FAB-MS m/z: 595 ([M ϩ Na]ϩ). Anal. Calcd. for
C20H29IO9S: C, 41.96; H, 5.10; S, 5.60. Found: C, 42.11; H, 5.09;
S, 5.59%.
dryness to give 16 (30 mg; 81%). [α]2D5 ϩ39 (c 0.8, D2O). H
1
NMR (200 MHz, D2O): δ 7.36 (m, 5H, Ar–H), 4.05 (dd, 1H,
J1,2 = 2.7 Hz, H-2, J2,3 = 6.5 Hz, H-2), 3.70 (dd, 1H, J3,4 = 8.5 Hz,
H-3), 3.64 (dd, 1H, J4,7a = 4.8 Hz, J7a,7b = 11.8 Hz, H-7a),
3.47, 3.30 (2d, 1H each, 2JH,H = 14.0 Hz, CH2Ph), 3.22 (dd, 1H,
J4,7b = 2.7 Hz, H-7b), 2.50 (br s, 2H, H-1, H-5), 2.06 (m, 1H,
H-4).13C NMR (50 MHz, D2O): δ 138.0, 128.2, 127.6, 127.1
(5C, Ar), 77.8, 75.3 (C-2, C-3), 60.6 (CH2Ph), 60.2 (C-7), 46.3,
44.6, 42.0 (C-1, C-4, C-5). HRFAB-MS: calcd. for (M ϩ H)ϩ
C13H18NO3: 236.1287, found: 236.1288.
Methyl
3,4-di-O-benzoyl-6-deoxy-6-iodo-2-O-toluene-p-
sulfonyl-ꢀ-D-glucopyranoside (22). A solution of 20 (2.43 g, 4.25
mmol) in trifluoroacetic acid–H2O 9 : 1 (13 mL) was kept
at room temperature for 1.5 h. Concentration to dryness
and co-evapourating with ethanol gave methyl 6-deoxy-6-iodo-
2-O-toluene-p-sulfonyl-α--glucopyranoside (21), which was
directly used for the next reaction, without further purification.
Data for compound 21: HRCI-MS: calcd. for [M ϩ H]ϩ C14H20-
IO7S: 458.9974, found: 458.9974. To a solution of 21 in pyridine
(15 mL) was added benzoyl chloride (2.96 mL, 25.5 mmol) and
the reaction was kept at rt for 24 h. The solution was then
poured over water–ice, diluted with CH2Cl2 (25 mL) and the
organic layer was washed with 1 M aqueous HCl (2 × 20 mL),
saturated aqueous NaHCO3 (20 mL), water (20 mL), dried over
MgSO4, filtered and concentrated to dryness. Crystallisation
from methanol yielded 22 (2.62 g; 92%). Mp: 172–174 ЊC. [α]2D2
(1R)-2-endo,3-exo-2,3-Dihydroxy-4-endo-4-hydroxymethyl-6-
azabicyclo[3.1.0]hexane (5). To a solution of 16 (16 mg, 0.07
mmol) in water (5 mL) was added palladium over charcoal 10%
(50 mg) and the mixture was hydrogenated at atmospheric pres-
sure and rt for 20 minutes. Then it was filtered through Celite
1
and concentrated to dryness to yield 5 (6 mg; 57%). H NMR
(200 MHz, D2O): δ 4.07 (d, J2,3 = 6.6 Hz, H-2), 3.69 (m, 2H,
H-7a, H-7b), 3.19 (dd, 1H, J3,4 = 8.0 Hz, H-3), 2.69 (s, H-1,
H-5), 2.06 (m, 1H, H-4). HRCI-MS calcd. for [M ϩ H]ϩ
C6H12NO3: 146.0817, found: 146.0814.
(2ЈS,3ЈS )-Methyl 3,4-O-[2Ј,3Ј-dimethoxybutane-2Ј,3Ј-diyl]-
2,6-bis(O-toluene-p-sulfonyl)-ꢀ-D-mannopyranoside (19). To a
stirred solution of methyl α--mannopyranoside (17, 5.0 g,
25.75 mmol) and ( )-camphorsulfonic acid (692 mg, 2.98
mmol) in methanol (50 mL) were added trimethyl orthoformate
(12 mL, 110.22 mmol) and butane-2,3-dione (3.6 mL, 41.2
mmol) and the solution was refluxed for 16 h. On cooling to
room temperature, the mixture was quenched with Et3N
(0.5 mL) and then concentrated to dryness. To a stirred solution
of the residue in anhydrous pyridine (30 mL) at 0 ЊC was added
toluene-p-sulfonyl choride (14.7 g, 77.25 mmol). The mix-
ture was warmed to rt and stirring was continued for 24 h
before the reaction mixture was partitioned between ice-cold
water and dichloromethane. The organic layer was washed with
1 M aqueous HCl (3 × 30 mL), saturated aqueous NaHCO3
(20 mL), water (20 mL), dried over MgSO4, filtered and concen-
trated to dryness. Crystallisation from methanol yielded 19
1
Ϫ8 (c 1.2, CH2Cl2). H NMR (300 MHz, CDCl3): δ 7.90–6.86
(m, 14H, Ar–H), 5.53 (t, 1H, J4,5 = 10.2 Hz, H-4), 5.46 (dd, 1H,
J2,3 = 3.1 Hz, J3,4 = 10.1 Hz, H-3), 5.04 (d, 1H, J1,2 = 1.6 Hz,
H-1), 4.90 (dd, 1H, H-2), 4.00 (td, 1H, J5,6a = 2.6 Hz, J5,6b
=
9.2 Hz, H-5), 3.54 (s, 3H, OMe), 3.35 (dd, 1H, J6a,6b = 10.8 Hz,
H-6a), 3.27 (dd, 1H, H-6b), 2.09 (s, 3H, CH3C6H4). 13C NMR
(75.5 MHz, CDCl3): δ 164.5 (2 CO), 154.4, 133.5, 133.2, 129.7,
128.6, 128.4, 128.1, 127.8 (18C, Ar), 98.9 (C-1), 75.8 (C-2), 70.6
(C-5), 69.6 (C-4), 69.2 (C-3), 55.8 (OMe), 21.5 (CH3C6H4),
3.3 (C-6). FAB-MS m/z: 689 ([M ϩ Na]ϩ). Anal. Calcd. for
C28H27IO9S: C, 50.46; H, 4.08; S, 4.81. Found: C, 50.43, H, 3.96;
S, 4.90%.
(1S,5S )-6-endo,7-exo-6,7-Bis(benzoyloxy)-N-benzyl-8-exo-8-
toluene-p-sulfonyloxy-3-oxa-2-azabicyclo[3.3.0]octane (24). To a
suspension of 22 (355 mg, 0.53 mmol) in aqueous EtOH (8 mL,
96%) was added Zn dust (374 mg, 6.10 mmol). The mixture was
refluxed for 1 h and then filtered through Celite and concen-
trated to dryness to give an oil that was dissolved in CH2Cl2
(15 mL), washed with water (2 × 10 mL), dried over MgSO4,
filtered and concentrated to dryness to give (2S,3S,4R)-3,4-bis-
O-benzoyloxy-2-O-toluene-p-sulfonylhex-5-enal (23) as an oil,
that was used without further purification in the next step.
HRFAB-MS calcd. for [M ϩ H]ϩ C27H25O8S: 509.1263, found:
509.1270. To a solution of 23 in toluene (5 mL) were added
N-benzylhydroxylamine hydrochloride (131 mg, 0.80 mmol)
and CaCO3 (80 mg, 0.80 mmol) and the reaction mixture was
heated at 50 ЊC for 1.5 h. The mixture was then filtered and the
filtrate was concentrated to dryness. The residue was dissolved
in CH2Cl2 (10 mL), washed with water (2 × 10 mL), dried over
Na2SO4, filtered and concentrated to dryness. Column chrom-
atography (CH2Cl2) gave 24 (146 mg; 41%), which was crystal-
lised from methanol. Mp: 126–130 ЊC (decomp.). [α]2D2 Ϫ95
(8.85 g; 56%). Mp: 138–142 ЊC. [α]2D4 ϩ120 (c 1.0, CH2Cl2). H
1
NMR (300 MHz, CDCl3): δ 7.80–7.27 (m, 8H, Ar–H), 4.70
(d, 1H, J1,2 = 1.4 Hz, H-1), 4.44 (dd, 1H, J2,3 = 3.1 Hz, H-2), 4.25
(dd, 1H, J5,6a = 1.5 Hz, J6a,6b = 10.6 Hz, H-6a), 4.14 (dd, 1H,
J5,6b = 5.5 Hz, H-6b), 3.87 (dd, 1H, J3,4 = 10.0 Hz, H-3), 3.79
(m, 1H, H-5), 3.74 (t, 1H, J4,5 = 9.9 Hz, H-4), 3.23, 3.10, 2.99
(s, 3 H each, 3 OMe), 2.42, 2.40 (s, 3 H each, 2 CH3C6H4), 1.14,
0.94 (s, 3 H each, 2 Me). 13C NMR (75.5 MHz, CDCl3): δ 144.7,
144.4, 129.7, 129.2, 128.3, 127.9 (12C, Ar), 100.0, 99.7 (C-2Ј,
C-3Ј), 99.4 (C-1), 76.4 (C-2), 68.5 (C-5), 67.9 (C-6), 65.5 (C-3),
62.8 (C-4), 55.1 (CH3OC-1), 48.0, 47.8 (2 OMe), 21.6, 21.5
(2CH3C6H4), 17.5, 17.3 (2 Me). CI-MS m/z: 585 ([M Ϫ OMe]ϩ).
Anal. Calcd. for C27H36O12S2: C, 52.58, H, 5,88. Found: C,
52.59, H, 5.91%.
(2ЈS,3ЈS )-Methyl
6-deoxy-3,4-O-[2Ј,3Ј-dimethoxybutane-
2Ј,3Ј-diyl]-6-iodo-2-O-toluene-p-sulfonyl-ꢀ-D-mannopyranoside
(20). To a solution of 19 (4.06 g, 6.59 mmol) in acetic anhydride
(40 mL) was added sodium iodide (1.48 g, 9.87 mmol), and the
mixture was refluxed for 1.5 h. After filtration of the solids,
the filtrate was concentrated to dryness and the residue was
1
(c 0.8, DMSO). H NMR (300 MHz, CDCl3): δ 7.97–6.90 m,
19H, Ar–H), 5.72 (dd, 1H, J5,6 = 8.3 Hz, J6,7 = 9.9 Hz, H-6), 5.56
(dd, 1H, J7,8 = 4.5 Hz, H-7), 4.90 (d, 1H, J1,8 ≈ 0.0 Hz, H-8),
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 4 7 8 – 4 8 2
481

View Article Online
4.05, 3.95 (2d, 1H each, 2JH,H = 13.5 Hz, CH2Ph), 3.98 (dd, 1H,
J4a,5 = 4.6 Hz, J4a,4b = 9.5 Hz, H-4a), 3.93 (dd, 1H, J4b,5 = 6.4 Hz,
H-4b), 3.84 (d, 1H, J1,5 = 8.1 Hz, H-1), 3.76 (m, 1H, H-5), 2.20
(s, 3H, Me) . 13C NMR (75.5 MHz, CDCl3): δ 165.7, 165.1
(2 CO), 144.8, 136.4, 133.5, 133.3, 132.6, 129.9, 129.8, 129.7,
128.9, 128.5, 128.2, 127.7, 127.6 (24C, Ar), 79.6 (C-8), 74.3
(C-6), 73.8 (C-7), 72.2 (C-1), 65.4 (C-4), 60.5 (CH2Ph), 45.4
(C-5), 21.6 (Me). HRCI-MS calcd. for [M ϩ H]ϩ C34H32NO8S:
614.1849, found: 614.1834.
7.6 Hz, H-7b), 2.65 (br s, 1H, H-5), 2.60 (br s, 1H, H-1),
2.52 (m, 1H, H-4). 13C NMR (75.5 MHz, D2O): δ 77.1
(C-2), 75.7 (C-3), 59.0 (C-7), 44.7 (C-4), 37.5 (C-1), 36.1 (C-5).
HRCI-MS calcd. for [M ϩ H]ϩ C6H12NO3: 146.0817, found:
146.0814.
Enzyme kinetics
The enzyme assays were carried out as described previously.18
All assays were performed at pH 6.8 and 25 ЊC. Steady state
kinetics was performed and reaction rates were measured after
possible slow-onset inhibition was essentially complete. The
inhibition constants (Ki) were obtained from the formula
Ki = [I]/(KMЈ/KM-1), where KMЈ and KM are Michaelis–Menten
constants with and without inhibitor present respectively. KMЈ
and KM were obtained from a Hanes plot, which was also used
to ensure that inhibition was competitive. The following KM
values (without inhibitor) were obtained using 4-nitrophenyl
glycosides as substrates and the above conditions: α-glucosidase
(yeast): 0.25 mM, β-glucosidase (almonds): 4 mM, α-fucosidase
(bovine kidney): 0.24 mM.
(1S )-N-Benzyl-2-exo,3-endo-2,3-bis(benzoyloxy)-4-endo-4-
hydroxymethyl-6-azabicyclo[3.1.0]hexane (25). To a solution of
24 (70 mg, 0.11 mmol) in EtOAc–EtOH 3 : 1 (8 mL) was added
Raney nickel and the mixture was hydrogenated at atmospheric
pressure and rt for 18 h. Then it was filtered through Celite and
concentrated to dryness. The residue was dissolved in MeOH
(5 mL) and stirred at room temperature for 3 days. Finally, the
solution was concentrated to dryness and compound TLC
(CH2Cl2–MeOH 80 : 1) of the residue gave 25 (23 mg; 46%).
[α]2D2 Ϫ77 (c 1.0, CH2Cl2). 1H NMR (300 MHz, CDCl3): δ 8.10–
7.24 (m, 15 H, Ar–H), 5.57 (d, 1H, J2,3 ≈ 0.0 Hz, J3,4 = 7.6 Hz,
H-3), 5.42 (s, 1H, J1,2 ≈ 0.0 Hz, H-2), 3.73 (m, 2H, H-7a, H-7b),
3.73, 3.34 (2d, 2H each, 2JH,H = 13.5 Hz, CH2Ph), 2.82 (m, 1H,
J4,5 = 2.5 Hz, H-4), 2.51 (d, 1H, J1,5 = 4.2 Hz, H-1), 2.44 (dd, 1H,
H-5). 13C NMR (75.5 MHz, CDCl3): δ 166.4, 165.5 (2 CO),
138.9, 133.4, 129.9, 129.8, 129.7, 128.5, 127.8, 127.3, 123.0 (18
C, Ar), 79.5 (C-2), 78.3 (C-3), 61.4 (CH2Ph), 60.0 (C-7), 45.5
(C-4), 45.4 (C-1), 44.6 (C-5). HRFAB-MS calcd. for [M ϩ H]ϩ
C34H32NO8S 444.1805, found: 444.1811.
Acknowledgements
We thank the Lundbeck foundation and the EU Marie Curie
fellowship program for financial support.
References
1 J. B. Sweeney, Chem. Soc. Rev., 2002, 31, 247–258.
2 S. Wakaki, H. Marumo, K. Tomioka, G. Shimizu, E. Kato,
H. Kamada, S. Kudo and Y. Fujimoto, Antibiot. Chemother., 1958,
8, 228.
3 R. R. Herr, M. E. Bergy, T. E. Eble and H. K. Jahnke, Antimicrob.
Agents Ann., 1960, 8, 23.
4 T. Hata, F. Koga, Y. Sano, K. Kanamori, A. Matsumae,
R. Sugawara, T. Hoshi and T. Shima, J. Antibiot., 1954, 7A, 107.
5 R. J. Ferrier, P. Prasit and G. J. Gainsford, J. Chem. Soc.,
Perkin Trans. 1, 1983, 1629.
6 Y. Miyamoto and S. Ogawa, J. Chem. Soc., Perkin Trans. 1, 1989,
1013.
7 R. Blattner, R. H. Furneaux, T. Kemmitt, P. C. Tyler, R. J. Ferrier
and A.-K. Tiden, J. Chem. Soc., Perkin Trans. 1, 1994, 3411.
8 E. A. Acar, F. Glarner and U. Burger, Helv. Chim. Acta, 1998, 81,
1095.
9 M. Nakata, C. Chong, Y. Niwata, K. Toshima and K. Tatsuta,
J. Antibiot., 1993, 46, 1919.
(1S )-N-Benzyl-2-exo,3-endo-2,3-dihydroxy-4-endo-4-hydroxy-
methyl-6-azabicyclo[3.1.0]hexane (26). To a solution of 25
(96 mg, 0.22 mmol) in dry methanol (2 mL) was added methan-
olic NaOMe (pH 10). The reaction was kept at rt for 4 h. Then
1 M aqueous HCl was added (pH 8) and the solution was con-
centrated to dryness. Column chromatography of the residue
(CH2Cl2
CH2Cl2–MeOH 10 : 1 gradient) gave 26 (37 mg;
73%). [α]2D7 = Ϫ27 (c 0.4, CH3OH). 1H NMR (300 MHz,
CD3OD): δ 7.31 (m, 5H, Ar–H), 3.92 (s, 1H, J1,2 ≈ 0.0, Hz
J2,3 ≈ 0.0 Hz, H-2), 3.82 (dd, 1H, J4,7a = 7.1 Hz, J7a,7b = 11.0 Hz,
H-7a), 3.66 (ddd, 1H, J3,4 = 5.5 Hz, J1,3 = 1.8 Hz, J3,5 = 1.3 Hz,
H-3), 3.65 (dd, 1H, J4,7b = 8.1 Hz, H-7b), 3.42 (s, 2H, CH2Ph),
2.52 (dd, 1H, J1,5 = 4.3 Hz, J4,5 = 2.0 Hz, H-5), 2.42 (dd, 1H,
H-1), 2.37 (m, 1H, H-4). 13C NMR (75.5 MHz, CD3OD):
δ 129.5, 128.9, 128.3, 104.2 (6C, Ar), 77.6 (C-3), 77.5 (C-2), 61.8
(CH2Ph), 60.0 (C-7), 47.3 (C-1), 46.6 (C-4), 45.7 (C-5).
HRFAB-MS cald. for [M ϩ H]ϩ C13H18NO3: 236.1287, found:
236.1286.
10 M. K. Tong and B. Ganem, J. Am. Chem. Soc., 1988, 110, 312.
11 R. C. Schoenfeld, J.-P. Lumb and B. Ganem, Tetrahedron Lett.,
2001, 42, 6447.
12 M. Bols, Acc. Chem. Res., 1998, 31, 1.
13 K. Søndergaard, X. Liang and M. Bols, Chem. Eur. J., 2001, 7, 2324.
14 P. Deslongchamp, Stereoelectronic Effects in Organic Chemistry,
Oxford University Press, Oxford, 1983.
(1S )-2-exo,3-endo-2,3-Dihydroxy-4-endo-4-hydroxymethyl-6-
azabicyclo[3.1.0]hexane (6). To a solution of 26 (14 mg, 0.06
mmol) in ethanol (1.5 mL) was added palladium over charcoal
10% (30 mg) and the mixture was hydrogenated at atmospheric
pressure and room temperature for 10 minutes. Then it was
filtered through a Celite bed and concentrated to dryness to
15 R. A. Farr, N. P. Peet and M. S. Kang, Tetrahedron Lett., 1990, 31,
7109–12.
16 Y. Kondo, Carbohydr. Res., 1986, 154, 305–309.
17 S. V. Ley, D. R. Owen and K. E. Wesson, J. Chem. Soc.,
Perkin Trans. 1, 1997, 2805–2806.
18 M. Bols, R. G. Hazell and I. B. Thomsen, Chem. Eur. J., 1997, 3,
1
give 6 (7.2 mg; 83%). H NMR (300 MHz, D2O): δ 4.07 (s,
940–7.
1H, J2,3 ≈ 0.0 Hz, H-2), 3.91 (m, 1H, H-3), 3.86 (dd, 1H,
19 A. Lohse, T. Hardlei, A. Jensen, I. Plesner and M. Bols, Biochem. J.,
2000, 349, 211–215.
J4,7a = 7.3 Hz, J7a,7b = 11.0 Hz, H-7a), 3.79 (dd, 1H, J4,7b
=
O r g . B i o m o l . C h e m . , 2 0 0 3 , 1, 4 7 8 – 4 8 2
482