The synthesis of a carbohydrate-like dihydrooxazine and tetrahydrooxazine as putative inhibitors of glycoside hydrolases: A direct synthesis of isofagomine

Article abstract of DOI:10.1139/v02-060

The treatment of benzyl 2,3-O-isopropylidene-β-L-xylopyranoside with N-hydroxyphthalimide under Mitsunobu conditions, followed by protecting-group interchange, gave benzyl 4-O-[(tert-butoxycarbonyl)amino]-2,3-O-isopropylidene-α-D-arabinoside. Mild acid hydrolysis and catalytic hydrogenolysis afforded 4-O-[(tert-butoxycarbonyl)amino]-D-arabinose that, upon heating in water, gave the dihydrooxazine [(4R,5S,6R)-5,6-dihydro-4,5-dihydroxy-6-hydroxymethyl-4H-1,2-oxazine] as a crystalline solid. A single-crystal structure determination of this solid showed it to exist in the ⁵H₆ conformation. Reduction of the dihydrooxazine gave the tetrahydrooxazine [(4R,5S,6R)-4,5-dihydroxy-6-hydroxymethyl-3,4,5,6-tetrahydro-2H-1,2-oxazine]. The dihydrooxazine was an effective inhibitor of two β-glucosidases (K_i = 27 and 35 μM). Benzyl 2,3-O-isopropylidene-β-L-xylopyranoside, via the derived imidazylate, was converted into a nitrile that, upon reduction and protecting-group manipulations, gave benzyl 4-C-aminomethyl-4-deoxy-α-D-arabinoside. Treatment of this amine with hydrogen and palladium-on-carbon gave isofagomine.

Full text of DOI:10.1139/v02-060

857
The synthesis of a carbohydrate-like
dihydrooxazine and tetrahydrooxazine as
putative inhibitors of glycoside hydrolases:
A direct synthesis of isofagomine
Wayne M. Best, James M. Macdonald, Brian W. Skelton, Robert V. Stick,
D. Matthew G. Tilbrook, and Allan H. White
Abstract: The treatment of benzyl 2,3-O-isopropylidene-ꢀ-L-xylopyranoside with N-hydroxyphthalimide under
Mitsunobu conditions, followed by protecting-group interchange, gave benzyl 4-O-[(tert-butoxycarbonyl)amino]-2,3-
O-isopropylidene-ꢁ-D-arabinoside. Mild acid hydrolysis and catalytic hydrogenolysis afforded 4-O-[(tert-butoxycar-
bonyl)amino]-D-arabinose that, upon heating in water, gave the dihydrooxazine [(4R,5S,6R)-5,6-dihydro-4,5-dihydroxy-
6-hydroxymethyl-4H-1,2-oxazine] as a crystalline solid. A single-crystal structure determination of this solid showed it
5
to exist in the H6 conformation. Reduction of the dihydrooxazine gave the tetrahydrooxazine [(4R,5S,6R)-4,5-dihy-
droxy-6-hydroxymethyl-3,4,5,6-tetrahydro-2H-1,2-oxazine]. The dihydrooxazine was an effective inhibitor of two ꢀ-
glucosidases (Ki = 27 and 35 M). Benzyl 2,3-O-isopropylidene-ꢀ-L-xylopyranoside, via the derived imidazylate, was
converted into a nitrile that, upon reduction and protecting-group manipulations, gave benzyl 4-C-aminomethyl-4-
deoxy-ꢁ-D-arabinoside. Treatment of this amine with hydrogen and palladium-on-carbon gave isofagomine.
Key words: dihydrooxazine, tetrahydrooxazine, isofagomine, iminosugars, glycosidase inhibitors.
Résumé : Le traitement du 2,3-O-isopropylidène-ꢀ-L-xylopyranoside de benzyle par du N-hydroxyphtalimide dans
des conditions de Mitsunobu, suivi d’un échange des groupes protecteurs, conduit à la formation du 4-O-[(tert-
butoxycarbonyl)amino]-2,3-O-isopropylidène-ꢁ-D-arabinoside de benzyle. Soumis à une hydrolyse acide légère et une
hydrogénolyse catalytique, celui-ci fournit alors le 4-O-[(tert-butoxycarbonyl)amino]-D-arabinose qui, par chauffage
dans l’eau, livre la dihydrooxazine [(4R,5S,6R)-5,6-dihydro-4,5-dihydroxy-6-hydroxyméthyl-4H-1,2-oxazine] sous la
forme de solide cristallin. Une détermination de structure sur un cristal unique montre qu’elle existe dans la conforma-
5
tion H6. La réduction de la dihydrooxazine fournit la tétrahydrooxazine [(4R,5S,6R)-4,5-dihydroxy-6-hydroxyméthyl-
3,4,5,6-tétrahydro-2H-1,2-oxazine]. La dihydrooxazine est un inhibiteur efficace de deux ꢀ-glucosidases (Ki = 27 et
35 M). Le 2,3-O-isopropylidène-ꢀ-L-xylopyranoside de benzyle, par le biais de l’imidazylate qui en dérive, a été
transformé en un nitrile qui, après réduction et des manipulations des groupes protecteurs, fournit le 4-C-aminométhyl-
4-désoxy-ꢁ-D-arabinoside de benzyle. Le traitement de cette amine avec de l’hydrogène et du palladium sur du charbon
conduit alors à l’isofagomine.
Mots clés : dihydrooxazine, tétrahydrooxazine, isofagomine, iminosucres, inhibiteurs de glycosidases.
[Traduit par la Rédaction] Best et al. 865
Introduction
glycoside hydrolases has received much attention in the last
half century (1–3).
The making and breaking of the glycosidic linkage are
events central to the survival of virtually all organisms. Both
processes are mediated by enzymes, generally described as
synthases or transferases for the former process and hydro-
lases for the latter. The study of the mechanism of action of
One way to study the action of an enzyme is to design
molecules that will inhibit the normal catalytic process. In
this regard, glycoside hydrolases are no exception and there
exists a class of inhibitor, generally described as “imino sug-
ars,” that encompasses such well-known molecules as swain-
sonine, castanospermine, nojirimycin, deoxynojirimycin, and
fagomine (4, 5). Bols (6) suggested, and subsequently syn-
thesised, an unnatural aza sugar, isofagomine (1), which was
shown to be a powerful inhibitor of some enzymes that hy-
drolyze ꢀ-D-glycosidic linkages with retention of configura-
tion.
Received 29 October 2001. Published on the NRC Research
Press Web site at http://canjchem.nrc.ca on 21 June 2002.
Dedicated to the memory of Professor Ray Lemieux.
W.M. Best, J.M. Macdonald, B.W. Skelton, R.V. Stick,1
D.M.G. Tilbrook, and A.H. White. Department of
Chemistry, 35 Stirling Highway, The University of Western
Australia, Crawley, WA 6009, Australia.
Some years ago, it occurred to us that tetrahydrooxazine 2
might better represent the natural substrate for the enzymes
of interest (or, rather, conjugate acid 3 would better repre-
sent the transition state commonly implicated in the
1Corresponding author (e-mail: rvs@chem.uwa.edu.au).
Can. J. Chem. 80: 857–865 (2002)
DOI: 10.1139/V02-060
© 2002 NRC Canada

858
Can. J. Chem. Vol. 80, 2002
Scheme 1. P, P > are protecting groups.
Fig. 2. Unit cell contents, projected down a, showing hydrogen
bonding.
O
OH
PO
O
L
-xylose
O
ONH2
OH
ONHP
'
O
O
O
PO
O
HO
HO
CH2OH
HO
HO
ONH2
(4)
(2)
CHO
Fig. 1. A single molecule (the asymmetric unit) of 4, showing
50% probability thermal ellipsoids for the non-hydrogen atoms,
hydrogen atoms having arbitrary radii of 0.1 Å.
phthalimide (6) was seen as a suitable source of the required
hydroxylamine residue. Thus, treatment of alcohol 5 with 6
under Mitsunobu conditions gave ꢁ-L-arabinoside 7. Mild
acid hydrolysis of 7 then gave diol 8 and a subsequent treat-
ment with phenylhydrazine gave hydroxylamine 9. Try as
we might, hydroxylamine 9 could never be converted into
any recognizable product under acidic conditions (there was
never any evidence for the formation of the desired product,
the enantiomer of the dihydrooxazine (4) [ent-(4)]).
The source of our problem seemed to reside with the
choice of a methyl glycoside for the protection of the
anomeric centre. The sequence from D-xylose was, therefore,
mimicked in both the allyl and 2-trimethylsilylethyl series,
resulting in ꢁ-L-arabinosides 10 and 11. But conditions
could not be found for the removal of either the allyl or 2-
trimethylsilylethyl glycoside with concomitant formation of
ent-(4).
hydrolytic process (1–3)). We thus started a synthetic pro-
gram that would furnish tetrahydrooxazine 2, via dihydro-
oxazine 4; subsequently, the synthetic route chosen also
supplied isofagomine (1).
In 1999, during the course of our investigations, Bols and
co-workers (7) published a synthesis of the racemic form of
2, not unexpectedly involving the cycloaddition of a nitroso
compound onto a diene. It was always our intention to pro-
duce the tetrahydrooxazine as one enantiomer and the suc-
cess of the process, starting with L-xylose and outlined in
Scheme 1, would ultimately depend upon the choice of pro-
tecting groups.
In some desperation, a benzyl glycoside was considered
and, subsequently, ꢁ-L-arabinoside 12 was synthesized. All
attempts at the hydrogenolysis–hydrogenation of 12, gener-
ally over palladium-on-carbon, failed to produce either ent-
(4) or ent-(2).
Much discussion, heartache, and experimentation were to
follow before we hit upon a successful twist for the synthe-
sis of 2 and 4. L-Xylose was converted into benzyl ꢀ-L-
xylopyranoside (we now employ trichloroacetimidate
methodology for the synthesis of the glycoside, rather than
the earlier Koenigs–Knorr approach) and in turn into alcohol
13. Normal Mitsunobu chemistry gave phthalimide 14, and
Results and discussion
In our initial investigations, D-xylose (in the early stages
of our work, it was much cheaper to use this enantiomer) was
converted into methyl ꢀ-D-xylopyranoside and then to methyl
2,3-O-isopropylidene-ꢀ-D-xylopyranoside (5). N-Hydroxy-
© 2002 NRC Canada

Best et al.
859
OH
OH
O
OH
O
OX
HO
HO
HO
HO
HO
HO
O
O
OX
NH
NH
OH
NH2
BnO
O
BnO
O
(1)
(2)
(3)
O
O
O
O
OH
O
HO
O
HO
HO
(14)
X =
N
(13)
(20)
X = H
OH
N
N
SO2Im
(4)
ent-(4)
(15)
(16)
NH2
O
O
O
NHCOOBut
HO
O
OMe
O
HO
N
ONHCOOBut
OH
O
OX
(5)
(6)
BnO
O
HO
O
O
OH
HO
HO
(18)
(19)
(17)
N
O
X = NHCOOBut
NH2
O
O
OMe
OX
X
O
O
OH
O
O
O
X
HO
OR
BnO
O
(7)
(R)
O
BnO
OH
HO
(24)
(25)
CH2NHCOOBut
X =
(21)
X = CN
CH2NH2
(8)
(9)
X =
R = Me
Me
N
(22)
(23)
CH2NH2
O
NH2
CH2NHCOOBut
NH2
NH2
NH2
CH2CH=CH2
(10)
CH2CH2SiMe3
CH2Ph
(11)
(12)
OH
OH
O
O
HO
HO
HO
NH
OH
subsequent manipulations gave hydroxylamine 15 and then
presumably carbamate 16. Mild acid hydrolysis of 16 gave
diol 17 and subsequent catalytic hydrogenolysis gave
hemiacetal 18. Treatment of 18 in boiling water presumably
liberated hydroxylamine 19, which subsequently formed the
desired and crystalline dihydrooxazine 4! A single-crystal
structure determination of 4 (Figs. 1 and 2, Table 1) showed
(26)
OH
NH
OH
HO
HO
HO
HO
NH
N
O
5
it to exist in the H6 conformation.
(27)
(28)
Dihydrooxazine 4 could be converted into desired tetrahy-
drooxazine 2 by treatment with sodium cyanoborohydride in
acetic acid. Our synthetic material (2) shared all of the phys-
ical attributes as described by Bols (7), especially in terms
seemed a logical choice.) Thus, alcohol 13 was converted
into imidazylate 20 (8) and a subsequent treatment with
trimethylsilyl cyanide and tetrabutylammonium fluoride
gave nitrile 21 (9), a far superior result to our efforts using
potassium cyanide in dimethylformamide.
Nitrile 21 was then reduced, presumably to amine 22, and
the product isolated as carbamate 23. Then, in virtually
quantitative yields, mild acid hydrolysis of 23 gave diol 24,
and subsequent treatment with anhydrous acid presumably
gave amine 25. Hydrogenolysis–hydrogenation of 25 over
palladium-on-carbon then gave isofagomine (1).
1
of NMR chemical-shift predictions (both H and 13C), and
possessed a negative value for the specific optical rotation.
Dihydrooxazine 4 acted as an efficient inhibitor of two ꢀ-
glucosidases isolated from Aspergillus niger and Clostridium
saccharolyticium — 4 had Ki values of 27 and 35 M, re-
spectively. Bols and co-workers (7) have previously reported
an inhibition constant (60 M) for (±)-(2) against almond ꢀ-
glucosidase.
With the euphoria that comes with the successful conclu-
sion to a long and protracted piece of work, we wondered if
our process could be adapted to a synthesis of isofagomine
(1). (All that was really required was to replace the oxygen
of the hydroxylamine by a carbon, and the cyano group
We have routinely used the synthesis described here to
produce appreciable amounts of isofagomine — the lowest
yielding step, the isopropylidenation of benzyl ꢀ-L-xylo-
pyranoside, is encountered early on in the sequence and the
© 2002 NRC Canada

860
Can. J. Chem. Vol. 80, 2002
Table 1. Molecular non-hydrogen geometries.
Distances (Å)
and N-hydroxyphthalimide (6) (2.50 g, 15.3 mmol) in THF
(20 mL) were cooled (10°C) and treated with diisopropyl
azodicarboxylate (3.60 mL, 18.3 mmol) for 3 days at rt. The
mixture was then poured into NaHCO3 solution and sub-
jected to normal workup (EtOAc). Flash chromatography
(EtOAc–petrol, containing 0.5% Et3N, 2:3) gave the
phthalimide (7) (3.10 g, 65%) as a colourless solid; mp 158–
O(1)—N(2)
O(1)—C(6)
N(2)—C(3)
C(3)—C(4)
C(4)—O(4)
1.405(2) C(4)—C(5)
1.450(2) C(5)—O(5)
1.269(2) C(5)—C(6)
1.506(2) C(6)—C(7)
1.417(2) C(7)—O(7)
1.516(2)
1.418(2)
1.521(2)
1.519(2)
1.425(2)
1
160°C; [ꢁ]D +7.8 (CHCl3). H NMR (500 MHz, CDCl3) :
1.35, 1.40 (2s, CMe2), 3.57 (s, OMe), 3.64 (dd, J2,3 = 9.8,
J3,4 = 2.4 Hz, H3), 3.66 (dd, J4,5 = 1.7, J5,5 = 13.6 Hz, H5),
4.30 (dd, J1,2 = 7.7 Hz, H2), 4.40 (dd, J4,5 = 1.5 Hz, H5),
4.49 (d, H1), 5.04 (m, H4), 7.72–7.73, 7.79–7.81 (2m, 4H,
Ar). 13C NMR (125.8 MHz, CDCl3) : 26.03, 26.52 (CMe2),
56.50 (OMe), 65.49 (C5), 72.88, 77.63, 79.64 (C2,3,4),
104.29 (Cl), 111.00 (CMe2), 123.45, 128.98, 134.42 (6C,
Ar), 163.04 (2C, CO).
Angles (°)
N(2)-O(1)-C(6)
O(1)-N(2)-C(3)
N(2)-C(3)-C(4)
C(3)-C(4)-O(4)
C(3)-C(4)-C(5)
O(4)-C(4)-C(5)
C(4)-C(5)-O(5)
119.3(1) C(4)-C(5)-C(6)
119.3(1) O(5)-C(5)-C(6)
127.0(1) O(1)-C(6)-C(5)
109.1(1) O(1)-C(6)-C(7)
109.4(1) C(5)-C(6)-C(7)
111.2(1) C(6)-C(7)-O(7)
111.3(1)
108.9(1)
107.8(1)
111.8(1)
103.9(1)
113.8(1)
111.0(1)
Ring torsion angles (°)
C(6)-O(1)-N(2)-C(3)
O(1)-N(2)-C(3)-C(4)
Methyl 4-O-(N-phthalimido)- -L-arabinoside (8)
–6.9(2) C(3)-C(4)-C(5)-C(6)
–2.1(2) C(4)-C(5)-C(6)-O(1) –55.9(2)
46.0(1)
Phthalimide 7 (1.01 g) in MeOH (40 mL) was treated with
hydrochloric acid (0.4 mL of 1 M) for 45 min. The mixture
was then concentrated and the residue dissolved in warm
CHCl3, filtered and concentrated to give diol 8 (900 mg) as a
colourless solid, a portion of which was recrystallized to
give colourless needles; mp 205–206°C (EtOH); [ꢁ]D –16.7
N(2)-C(3)-C(4)-C(5) –19.3(2) C(5)-C(6)-O(1)-N(2)
36.7(2)
Note: Hydrogen bonds: OH(4)···O(5) (x – 1/2, 1/2 – y, 2 – z) =
2.747(2), 2.03(2) Å; OH(5)···O(7) (2 – x, y – 1/2, 1.5 – z) = 2.737(2),
1.94(2) Å; OH(7)···N(2) (2 – x, y – 1/2, 1.5 – z) = 2.808(2), 2.05(3) Å.
1
(CHCl3). H NMR (500 MHz, DMSO-d6) : 3.39 (s, OMe),
whole process compares favourably with previously pub-
lished syntheses (10–16). It is our immediate intention to
produce glycosylated versions of isofagomine, such as 26,
and to see if our synthetic sequence can be adopted to pro-
duce 1-azafagomine (27) (17) and interesting nitrone 28.
3.58 (dd, J4,5 = 1.8, J5,5 = 12.7 Hz, H5), 3.64 (ddd, J1,2
=
6.0, J2,3 = 8.5, J2,OH = 4.5 Hz, H2), 3.68 (ddd, J3,4 = 3.1,
J3,OH = 5.4 Hz, H3), 4.08 (dd, J4,5 = 3.8 Hz, H5), 4.11 (d,
H1), 4.47ꢆ4.51 (m, H4), 4.89 (d, 3-OH), 5.26 (d, 2-OH),
7.86 (s, 4H, Ar). 13C NMR (125.8 MHz, DMSO-d6) : 55.76
(OMe), 61.52 (C5), 70.34, 71.36, 82.86 (C2,3,4), 104.00
(Cl), 123.19, 128.64, 134.71 (6C, Ar), 163.29 (2C, CO).
Anal. calcd. for C14H15NO7: C 54.4, H 4.9, N 4.5; found:
C 54.4, H 5.0, N 4.5.
Experimental
General experimental procedures have been given previ-
ously (18). HCl in MeOH was generated by the addition of
acetyl chloride to anhydrous MeOH.
Methyl 4-O-amino- -L-arabinoside (9)
Diol 8 (777 mg, 2.50 mmol) in EtOH (15 mL) was treated
with phenylhydrazine (0.35 mL, 3.6 mmol) and the mixture
stirred (45 min, rt), poured into water, and washed with
EtOAc. The aqueous phase was concentrated and the residue
triturated with EtOAc (which was discarded) to leave
hydroxylamine derivative 9 (357 mg, 79%) as a colourless
solid, a portion of which was recrystallized to give colour-
Structure determination of 4
A full sphere of CCD area-detector diffractometer data
was measured (Bruker AXS instrument; (ꢂ–scans, 2ꢃmax
=
58°; monochromatic Mo Kꢁ radiation, ꢄ = 0.71073 Å; T =
-153 K) yielding 6405 total reflections, merging to 939
unique (Rint = 0.023), 897 with F > 4ꢅ(F) being used in the
full matrix least squares refinement (C, N, and O thermal pa-
rameter forms anisotropic (x, y, z, Uiso)H refined). “Friedel
pair” data were not retained distinct, chirality being assigned
from the chemistry. Conventional residuals R and Rw
(weights: ꢅ2(F) + 0.0004 F 2)–1) on | F | at convergence were
0.028 and 0.039, respectively. Neutral-atom scattering fac-
tors were used within the context of the Xtal 3.6 program
system (19).
1
less needles; mp 144°C (EtOH); [ꢁ]D +32.3 (MeOH). H
NMR (300 MHz, DMSO-d6) : 3.25 (ddd, J1,2 = 6.7, J2,3
=
9.0, J2,OH = 4.8 Hz, H2), 3.28 (dd, J4,5 = 1.5, J5,5 = 12.3 Hz,
H5), 3.34 (s, OMe), 3.41–3.46 (m, H3), 3.57–3.60 (m, H4),
3.97 (d, H1), 3.99 (dd, J4,5 = 3.0 Hz, H5), 4.79 (br d, J3,OH
=
5.4 Hz, 3-OH), 5.00 (br d, 2-OH), 6.00 (br s, NH2).
13C NMR (75.5 MHz, DMSO-d6) : 55.7 (OMe), 61.5 (C5),
70.6, 71.7, 80.1 (C2,3,4), 104.4 (Cl). Anal. calcd. for C6H13NO5:
C 40.2, H 7.3, N 7.8; found: C 40.4, H 7.2, N 7.9.
Crystal data
C5H9NO4, M = 197.1, orthorhombic, space group P212121
(D42, No. 19), a = 7.091(1), b = 8.847(1), c = 8.928(1) Å,
Allyl 2,3-O-isopropylidene- -D-xylopyranoside
V = 631.2 Å3, Dcalcd. = 1.548 g cm–3, Z = 4,
= 1.3 cm–1
HCl in MeOH (4 M, 60 L) was added to allyl ꢀ-D-
xylopyranoside (21) (2.9 g, 15 mmol) suspended in DMF
(13 mL), the mixture cooled (0°C), and 2-methoxypropene
(4.0 mL, 42 mmol) added. The mixture was stirred (1 h, rt)
and normal work-up (CHCl3) gave a colourless oil. Flash
chromatography (EtOAc–petrol, containing 0.5% Et3N, 1:4)
gave allyl 3,4-O-isopropylidene-ꢀ-D-xyloside as a colourless
Mo
(no correction).
Methyl 2,3-O-isopropylidene-4-O-(N-phthalimido)- -L-
arabinoside (7)
Methyl 2,3-O-isopropylidene-ꢀ-D-xylopyranoside (20) (5)
(2.80 g, 13.7 mmol), triphenylphosphine (4.10 g, 15.6 mmol),
© 2002 NRC Canada

Best et al.
861
syrup (210 mg, 6%). A small sample was distilled (glass
Allyl 4-O-amino- -L-arabinoside (10)
T-piece, 0.05 mm Hg, heating block 110–130°C); [ꢁ]D –59
(CHCl3). H NMR (500 MHz, CDCl3) : 1.46, 1.47 (2s,
Allyl 4-O-(N-phthalimido)-ꢁ-L-arabinoside (1.6 g, 4.8 mmol)
in CHCl3 (60 mL) was treated with methylhydrazine
(0.4 mL, 7.5 mmol) and the mixture stirred (2 h, rt), ex-
tracted with water, and the extracts filtered and concentrated.
Recrystallization of the residue gave hydroxylamine deriva-
tive 10 (620 mg, 63%) as colourless plates; mp 120°C
1
CMe2), 3.47–3.54 (m, H3,5), 3.57–3.59 (m, H5), 3.72 (m,
H2), 4.11 (tdd, J1>,1> = 11.8, J1>,2> = 6.4, J1>,3> = 1.2 Hz, H1>),
4.17 (dd, J3,4 = 9.8, J4,5 = 4.3 Hz, H4), 4.29 (d, J1,2 = 6.9 Hz,
H1), 4.35ꢆ4.39 (m, H1>), 5.22–5.25 (m, H3>), 5.31 (dq,
J2>,3> = 17.1, J3>,3> = 1.3 Hz, H3>), 5.89–5.97 (m, H2>). 13C NMR
(125.8 MHz, CDCl3) : 26.51, 26.73 (CMe2), 64.63, 70.88
(C5,1>), 73.06, 73.61, 81.17 (C2,3,4), 103.64 (Cl), 112.44
(CMe2), 118.31 (C3>), 133.44 (C2>). Further elution gave
allyl 2,3-O-isopropylidene-ꢀ-D-xylopyranoside (2.7 g, 77%)
as a colourless syrup; [ꢁ]D ꢆ49 (CHCl3). 1H NMR
1
(EtOH); [ꢁ]D +24.5 (MeOH). H NMR (500 MHz, DMSO-
d6) : 3.24–3.28 (m, H2,5), 3.41–3.44 (m, H3), 3.56–3.57
(m, H4), 3.95ꢆ4.00 (m, H5,1>), 4.08 (d, J1,2 = 6.8 Hz, H1),
4.17 (tdd, J1>,1> = 13.4, J1>,2> = 5.0, J1>,3> = 1.6 Hz, H1>), 4.77 (d,
J3,OH = 5.8 Hz, 3-OH), 4.98 (d, J2,OH = 4.9 Hz, 2-OH), 5.12
(dq, J2>,3> = 10.5 Hz, H3>), 5.29 (dq, J2>,3> = 17.3 Hz, H3>),
5.83–5.91 (m, H2>), 5.99 (br s, NH2). 13C NMR
(125.8 MHz, DMSO-d6) : 61.60, 68.65 (C5,1>), 70.94,
71.94, 80.15 (C2,3,4), 102.61 (Cl), 116.43 (C3>), 134.97
(C2>). Anal. calcd. for C8H15NO5: C 46.8, H 7.4, N 6.8;
found: C 46.6, H 7.4, N 6.8.
(500 MHz, C6D6) : 1.32, 1.36 (2s, CMe2), 2.40 (d, J4,OH
=
2.6 Hz, OH), 3.07 (dd, J4,5 = 7.6, J5,5 = 11.9 Hz, H5), 3.46–
3.53 (m, H2,3), 3.69–3.73 (m, H4), 3.85 (dd, J4,5 = 5.3 Hz,
H5), 3.92 (tdd, J1>,1> = 13.1, J1>,2> = 5.7, J1>,3> = 1.5 Hz, H1>),
4.24 (tdd, J1>,2> = 5.0, J1>,3> = 1.6 Hz, H1>), 4.53 (d, J1,2
=
6.8 Hz, H1), 4.99 (dq, J2>,3> = 10.4, J3>,3> = 1.5 Hz, H3>), 5.25
(dq, J = 1.7, 17.2 Hz, H3>), 5.81 (dddd, H2>). 13C NMR
(125.8 MHz, C6D6) : 26.76, 26.87 (CMe2), 67.49, 69.38
(C5,1>), 69.60 (C4), 77.11, 81.63 (C2,3), 101.44 (C1),
111.19 (CMe2), 116.74 (C3>), 134.50 (C2>).
2>-Trimethylsilylethyl -D-xylopyranoside
HgO (11.5 g, 54.0 mmol), HgBr2 (1.16 g, 3.21 mmol),
CaSO4 (36 g), and 2-trimethylsilylethanol (8.0 mL, 56 mmol)
in CH2Cl2 (75 mL) were stirred (5 min) before tri-O-acetyl-
ꢁ-D-xylopyranosyl bromide (22) (18.3 g, 54.0 mmol) in
CH2Cl2 (60 mL) was added dropwise. The mixture was
stirred (2 days, rt), then filtered through silica with a pad of
Celite on top, washing the filter residue with CH2Cl2. The
combined filtrate and washings were washed with saturated
KI solution and brine, then dried, and concentrated to give a
yellow oil. Vacuum distillation (Kugelrohr, 30 mm Hg,
170°C) removed the excess 2-trimethylsilylethanol. The resi-
due was dissolved in MeOH (150 mL) and treated with so-
dium (200 mg). After 2 h, the mixture was neutralized with
resin (Dowex-50W, H+ form), then filtered, washed with
MeOH, and the combined filtrate and washings were con-
centrated to give a pale yellow solid. Recrystallization gave
2>-trimethylsilylethyl ꢀ-D-xylopyranoside (6.6 g, 49%) as
white flakes; mp 126–128°C (CH2Cl2–i-Pr2O); [ꢁ]D ꢆ46.9
Allyl 2,3-O-isopropylidene-4-O-(N-phthalimido)- -L-
arabinoside
Allyl 2,3-O-isopropylidene-ꢀ-D-xylopyranoside (2.3 g)
was treated with 6 as previously for 5 to give, after flash
chromatography (EtOAc–petrol, containing 0.5% Et3N, 1:4
then 1:2) and trituration with diisopropyl ether, the title
phthalimide (2.1 g, 56%) as a colourless solid; mp 148°C;
1
[ꢁ]D +0.5 (CHCl3). H NMR (500 MHz, CDCl3) : 1.35,
1.41 (2s, CMe2), 3.63 (dd, J3,4 = 2.4, J2,3 = 9.9 Hz, H3), 3.66
(dd, J4,5 = 1.9, J5,5 = 13.6 Hz, H5), 4.19 (tdd, J1>,1> = 12.6,
J1>,2> = 6.4, J1>,3> = 1.3 Hz, H1>), 4.34 (dd, J1,2 = 7.7 Hz, H2),
4.38ꢆ4.42 (m, H5,1>), 4.61 (d, H1), 5.04 (bq, J = 1.8 Hz,
H4), 5.21 (dq, J2>,3> = 10.3 Hz, H3>), 5.34 (dq, J2>,3> = 17.2 Hz,
H3>), 5.98 (dddd, J1>,2> = 5.4 Hz, H2>), 7.71–7.74, 7.79–7.83
(2m, 4H, Ar). 13C NMR (75.5 MHz, CDCl3) : 26.02, 26.53
(CMe2), 65.41, 69.53 (C5,1>), 72.88, 77.65, 79.59 (C2,3,4),
102.25 (Cl), 110.92 (CMe2), 117.94 (C3>), 123.41, 128.96,
134.36 (6C, Ar), 133.66 (C2>), 163.00 (2C, CO).
1
(MeOH). H NMR (300 MHz, CDCl3) : 0.02 (s, SiMe3),
0.91–1.11 (m, 2H, H2>), 3.27 (dd, J4,5 = 8.9, J5,5 = 11.8 Hz,
H5), 3.38 (dd, J1,2 = 6.9, J2,3 = 8.4 Hz, H2), 3.52 (t, J =
8.4 Hz, H3), 3.57 (ddd, J1>,1> = 11.4, J1>,2> = 6.2, 9.8 Hz, H1>),
3.68 (dt, J4,5 = 5.0 Hz, H4), 3.93 (ddd, J1>,2> = 6.2, 9.7 Hz
H1>), 3.95 (dd, H5), 4.29 (d, H1). 13C NMR (75.5 MHz,
CDCl3) : –1.44 (SiMe3), 18.32 (C2>), 65.26, 67.32 (C5,1>),
69.62, 72.94, 76.04 (C2,3,4), 102.50 (C1). Anal. calcd. for
C10H22O5Si: C 48.0, H 8.9; found: C 48.1, H 8.6.
Allyl 4-O-(N-phthalimido)- -L-arabinoside
Allyl 2,3-O-isopropylidene-4-O-(N-phthalimido)-ꢁ-L-ara-
binoside (2.0 g) was treated with hydrochloric acid as previ-
ously for 7 to give the title diol (1.9 g) as a colourless solid,
a small portion of which was recrystallized to give long,
2>-Trimethylsilylethyl 2,3-O-isopropylidene- -D-xylopy-
ranoside
fine, colourless needles; mp 145°C (EtOAc–i-Pr2O); [ꢁ]D –7.3
1
(MeOH). H NMR (500 MHz, CDCl3) : 3.69 (dd, J4,5
=
2>-Trimethylsilylethyl ꢀ-D-xylopyranoside (1.0 g) was
treated with 2-methoxypropene as previously for allyl ꢀ-D-
xylopyranoside to give a colourless oil (930 mg). Flash chro-
matography (EtOAc–petrol, containing 0.5% Et3N, 1:4 then
3:7) gave the title alcohol (580 mg, 50%) as an unstable,
colourless oil; [ꢁ]D ꢆ43.4 (CHCl3–1% pyridine). 1H
NMR (300 MHz, CDCl3) : –0.02 (s, SiMe3), 0.91–1.08 (m,
1.7, J5,5 = 13.0 Hz, H5), 3.78 (dd, J2,3 = 9.1, J3,4 = 3.4 Hz,
H3), 3.90 (dd, J1,2 = 6.7 Hz, H2), 4.15 (ddt, J1>,1> = 12.7,
J1>,2> = 6.5, J1>,3> = 1.2 Hz, H1>), 4.27ꢆ4.29 (m, H4), 4.35 (d,
H1), 4.40 (ddt, J1>,2> = 5.2, J1>,3> = 1.4 Hz, H1>), 4.60 (dd,
J
4,5 = 3.2 Hz, H5), 5.23 (dq, J2>,3> = 10.4, J3>,3> = 1.4 Hz, H3>),
5.34 (dq, J2>,3> = 17.2 Hz, H3>), 5.96 (dddd, H2>), 7.78–7.80,
7.84–7.87 (2m, 4H, Ar). 13C NMR (125.8 MHz, CDCl3) :
63.77 (C5), 69.93 (C1>), 71.98, 72.22, 86.77 (C2,3,4), 101.68
(Cl), 118.11 (C3>), 124.01, 128.49, 134.96 (6C, Ar), 133.67
(C2>), 164.02 (2C, CO). Anal. calcd. for C16H17NO7·H2O:
C 54.4, H 5.4, N 4.0; found: C 54.0, H 5.3, N 4.0.
2H, H2>), 1.42, 1.43 (2s, CMe2), 3.20 (dt, J4,5 = 3.3, J
11.1 Hz, H5), 3.24 (br s, OH), 3.29 (dd, J1,2 = 7.6, J
9.6 Hz, H2), 3.48 (t, J3,4 = 9.4 Hz, H3), 3.60 (ddd, J1>,1>
=
=
=
5,5
2,3
11.3, J1>,2> = 6.1, 9.5 Hz, H1>), 3.91 (ddd, J1>,2> = 6.1, 9.5 Hz,
© 2002 NRC Canada

862
Can. J. Chem. Vol. 80, 2002
H1>), 3.97ꢆ4.05 (m, H4,5), 4.60 (d, H1). 13C NMR
(75.5 MHz, CDCl3) : –1.58 (SiMe3), 18.12 (C2>), 26.47,
26.62 (CMe2), 66.73, 67.11 (C5,1>), 68.94, 76.42, 81.04
(C2,3,4), 101.16 (Cl), 111.42 (CMe2). Anal. calcd. for
C13H26O5Si: C 53.8, H 9.0; found: C 53.9, H 9.0.
J1>,2> = 6.1, 9.5 Hz, H1>), 4.19 (d, H1), 4.24 (dd, J4,5 = 2.4 Hz,
H5). 13C NMR (75.5 MHz, CDCl3) : –1.47 (SiMe3), 18.27
(C2>), 62.86, 67.19 (C5,1>), 72.28, 73.05 (C2,3,4), 102.59
(Cl). Anal. calcd. for C10H23NO5Si: C 45.3, H 8.7, N 5.3;
found: C 45.4, H 8.9, N 5.4.
2>-Trimethylsilylethyl 2,3-O-isopropylidene-4-O-(N-
phthalimido)- -L-arabinoside
Benzyl 2,3-O-isopropylidene- -D-xylopyranoside
Benzyl ꢀ-D-xylopyranoside (23) (7.60 g) was treated as
previously for allyl ꢀ-D-xylopyranoside to give, after flash
chromatography (EtOAc–petrol, containing 0.5% Et3N, 1:4),
benzyl 3,4-O-isopropylidene-ꢀ-D-xyloside (620 mg, 7%) as
2>-Trimethylsilylethyl 2,3-O-isopropylidene-ꢀ-D-xylopyra-
noside (383 mg, 1.30 mmol), triphenylphosphine (420 mg,
1.60 mmol), and 6 (260 mg, 1.60 mmol) in THF (12 mL)
were treated with diethyl azodicarboxylate (0.31 mL,
1.6 mmol) for 3 days (rt). Air was bubbled through the mix-
ture (2 h) before it was concentrated and the residue sub-
jected to flash chromatography (EtOAc–petrol, containing
0.5% Et3N, 1:9 then 1:4) to give the title phthalimide
(442 mg, 77%) as an unstable colourless oil. 1H NMR
(300 MHz, CDCl3) : 0.01 (s, SiMe3), 0.99–1.13 (m, 2H,
H2>), 1.34, 1.40 (2s, CMe2), 3.60–3.73 (m, H3,5,1>), 4.01
a
colourless solid, a small portion of which was
recrystallized; mp 129–130°C (i-Pr2O; lit. (24) 126°C); [ꢁ]D
–79 (CHCl3, lit. (24) –87). H NMR (500 MHz, CDCl3) :
1
1.46, 1.47 (2s, CMe2), 2.52 (br s, OH), 3.49 (dd, J2,3 = 9.6,
J3,4 = 9.0 Hz, H3), 3.54 (t, J4,5 W J5,5 = 10.1 Hz, H5), 3.62
(ddd, J4,5 = 4.3, 10.1 Hz, H4), 3.74 (ddd, J1,2 = 6.8 Hz, H2),
4.20 (dd, H5), 4.35 (d, H1), 4.59, 4.92 (ABq, J = 11.5 Hz,
PhCH2), 7.31–7.37 (m, Ph). 13C NMR (125.8 MHz, CDCl3)
: 26.54, 26.77 (CMe2), 64.72 (C5), 71.80 (PhCH2), 73.16,
73.66, 81.17 (C2,3,4), 103.67 (Cl), 112.45 (CMe2), 128.19,
128.24, 128.58, 136.76 (Ph). Further elution (EtOAc–petrol,
containing 0.5% Et3N, 1:2) gave benzyl 2,3-O-isopropy-
lidene-ꢀ-D-xylopyranoside (5.26 g, 59%) as a somewhat un-
stable, colourless syrup; [ꢁ]D –58.0 (CHCl3, lit. (24) –54).
1H NMR (500 MHz, CDCl3) : 1.45, 1.46 (2s, CMe2), 2.22
(d, J4,OH = 3.8 Hz, OH), 3.28 (dd, J4,5 = 7.7, J5,5 = 11.7 Hz,
H5), 3.43 (dd, J1,2 = 7.4, J2,3 = 9.6 Hz, H2), 3.53 (t, J =
9.2 Hz, H3), 4.04ꢆ4.08 (m, H4), 4.10 (dd, J4,5 = 5.3 Hz,
H5), 4.68, 4.89 (ABq, J = 11.7 Hz, PhCH2), 4.74 (d, H1),
7.28–7.39 (m, Ph). 13C NMR (125.8 MHz, CDCl3) : 26.58,
26.77 (CMe2), 67.26 (C5), 70.44 (PhCH2), 69.43, 76.56,
81.01 (C2,3,4), 100.77 (Cl), 111.71 (CMe2), 127.90, 128.12,
128.43, 136.91 (Ph).
(ddd, J1>,1> = 10.9, J1>,2> = 7.2, 9.4 Hz, H1>), 4.31 (dd, J1,2
=
7.4, J2,3 = 9.9 Hz, H2), 4.40 (dd, J4,5 = 1.6, J5,5 = 13.7 Hz,
H5), 4.58 (d, H1), 5.03 (dd, J3,4 = 3.8 Hz, H4), 7.70–8.81
(m, 4H, Ar). 13C NMR (75.5 MHz, CDCl3) : –1.48
(SiMe3), 18.17 (C2>), 26.06, 26.58 (CMe2), 65.43, 66.65
(C5,1>), 73.03, 77.74, 79.74 (C2,3,4), 102.93 (Cl), 110.87
(CMe2), 123.43, 128.99, 134.36 (6C, Ar), 163.05 (2C, CO).
2>-Trimethylsilylethyl 4-O-(N-phthalimido)- -L-arabinoside
2>-Trimethylsilylethyl 2,3-O-isopropylidene-4-O-(N-phtha-
limido)-ꢁ-L-arabinoside (253 mg) in THF (8 mL) was
treated with hydrochloric acid (2 mL of 1 M) for 2 days.
The mixture was neutralized with resin (Amberlite IRA 400,
OH– form), then filtered, and concentrated to give the title
diol (216 mg, 94%) as a colourless crystalline solid; mp
118–120°C; [ꢁ]D –10.7 (CHCl3). 1H NMR (300 MHz,
CDCl3) : 0.02 (s, SiMe3), 0.92–1.12 (m, 2H, H2>), 2.64
(br s, 2-OH), 3.60 (ddd, J1>,1> = 11.3, J1>,2> = 6.1, 9.6 Hz, H1>),
3.68 (dd, J4,5 = 1.6, J5,5 = 13.0 Hz, H5), 3.76 (dd, J2,3 = 9.5,
J3,4 = 3.2 Hz, H3), 3.84 (dd, J1,2 = 6.6 Hz, H2), 4.00 (ddd,
J1>,2> = 6.6, 9.5 Hz, H1>), 4.10 (d, J3,OH = 3.2 Hz, 3-OH), 4.26
(dt, J4,5 = 3.0 Hz, H4), 4.29 (d, H1), 4.61 (dd, H5), 7.76–
7.87 (m, 4H, Ar). 13C NMR (75.5 MHz, CDCl3) : –1.46
(SiMe3), 18.17 (C2>), 63.91, 67.13 (C5,1>), 72.05, 72.30
(C2,3), 86.96 (C4), 102.29 (Cl), 123.96, 128.47, 134.89 (6C,
Ar), 163.96 (2C, CO). Anal. calcd. for C18H25NO7Si: C 54.7,
H 6.4, N 3.5; found: C 54.7, H 6.4, N 3.5.
Benzyl 2,3-O-isopropylidene-4-O-(N-phthalimido)- -L-
arabinoside
Benzyl 2,3-O-isopropylidene-ꢀ-D-xylopyranoside (2.8 g)
was treated with 6 as previously for 5 to give after flash
chromatography (EtOAc–toluene, containing 0.5% Et3N,
1:5) an oil. Trituration of this oil with i-Pr2O then gave the
title phthalimide (3.4 g, 80%) as a colourless solid, a small
portion of which was recrystallized; mp 172–174°C
1
(MeOH); [ꢁ]D –19 (CDCl3). H NMR (500 MHz, CDCl3) :
1.35, 1.42 (2s, CMe2), 3.63 (dd, J2,3 = 9.9, J3,4 = 2.4 Hz,
H3), 3.67 (dd, J4,5 = 1.9, J5,5 = 13.6 Hz, H5), 4.39 (dd, J1,2
=
7.7 Hz, H2), 4.43 (dd, J4,5 = 1.6 Hz, H5), 4.66 (d, H1), 4.72,
4.97 (ABq, J = 11.8 Hz, PhCH2), 5.05 (m, H4), 7.27–7.43,
7.72–7.75, 7.80–7.83 (3m, 9H, Ar). 13C NMR (125.8 MHz,
CDCl3) : 26.08, 26.58 (CMe2), 65.48 (C5), 70.09 (PhCH2),
73.00, 77.68, 79.65 (C2,3,4), 102.20 (Cl), 110.97 (CMe2),
123.46, 127.76, 128.10, 128.35, 129.02, 134.40, 136.98
(12C, Ar), 163.05 (2C, CO). Anal. calcd. for C23H23NO7:
C 64.9, H 5.5, N 3.3; found: C 64.6, H 5.4, N 3.3.
2>-Trimethylsilylethyl 4-O-amino- -L-arabinoside (11)
2>-Trimethylsilylethyl 4-O-(N-phthalimido)-ꢁ-L-arabinoside
(26 mg) in CHCl3 (3 mL) was treated with methylhydrazine
(0.15 mL). After 5 min, TLC showed that no starting mate-
rial remained and the solution was applied to the top of a
small silica-gel column. Elution with CHCl3, then MeCN–
CHCl3 (2:3), gave a colourless solid. Recrystallization gave
hydroxylamine derivative 11 (16 mg, 92%) as a colourless
crystalline solid; mp 140–142°C (CHCl3–petrol); [ꢁ]D +14.2
Benzyl 4-O-(N-phthalimido)- -L-arabinoside
1
(MeOH). H NMR (300 MHz, CDCl3) : 0.17 (s, SiMe3),
Benzyl
2,3-O-isopropylidene-4-O-(N-phthalimido)-ꢁ-L-
0.94–1.12 (m, 2H, H2>), 1.62 (br s, 4H, OH,NH2), 3.43 (dd,
arabinoside (2.00 g) was treated with hydrochloric acid as
previously for 7 to give the title diol (1.82 g) as a colourless
solid, a small portion of which was recrystallized to give
colourless micro-needles; mp 183°C (EtOH); [ꢁ]D –20.0
J4,5 = 1.2, J5,5 = 13.2 Hz, H5), 3.56 (ddd, J1>,1> = 11.6, J1>,2>
=
6.6, 9.6 Hz, H1>), 3.63 (dd, J1,2 = 6.8, J2,3 = 9.3 Hz, H2),
3.70 (dd, J3,4 = 2.4 Hz, H3), 3.91–3.94 (m, H4), 3.97 (ddd,
© 2002 NRC Canada

Best et al.
863
1
(CHCl3). H NMR (500 MHz, DMSO-d6) : 3.61 (dd, J4,5
1.7, J5,5 = 12.6 Hz, H5), 3.70–3.75 (m, H2,3), 4.12 (dd,
4,5 = 3.8 Hz, H5), 4.32 (d, J1,2 = 5.8 Hz, H1), 4.50ꢆ4.52 (m,
=
(100 mL) were treated with diethyl azodicarboxylate
(4.90 mL, 15.0 mmol) for 4 days at rt. The mixture was con-
centrated and flash chromatography (EtOAc–petrol, contain-
ing 0.5% Et3N, 1:4 then 2:3) gave crude phthalimide 14
(4.62 g) as a pale yellow solid. This solid was dissolved in
CHCl3 and treated with N2H4·H2O (2 mL, 30 min). The mix-
ture was filtered and the residue washed with CHCl3. The
combined filtrate–washings were washed with H2O and
brine, then dried and concentrated to give a pale yellow oil
which crystallized. Recrystallization gave hydroxylamine 15
as colourless cubes (1.2 g, 62%); mp 85–87°C (Et2O–
J
H4), 4.57, 4.80 (ABq, J = 12.2 Hz, PhCH2), 4.93 (d, J =
5.1 Hz, OH), 5.32 (d, J = 4.4 Hz, OH), 7.28–7.43 (m, Ph),
7.86 (s, 4H, Ar). 13C NMR (125.8 MHz, DMSO-d6) : 61.60
(C5), 69.42 (PhCH2), 70.48, 71.35, 82.85 (C2,3,4), 102.25
(Cl), 123.21, 127.41, 127.61, 128.18, 128.66, 134.72, 138.01
(12C, Ar), 163.31 (2C, CO). Anal. calcd. for C20H19NO7:
C 62.3, H 5.0, N 3.6; found: C 62.5, H 5.2, N 3.8.
1
petrol); [ꢁ]D +1.9 (CHCl3). H NMR (500 MHz, CDCl3) :
1.44, 1.48 (2s, CMe2), 3.43 (dd, J4,5 = 1.5, J5,5 = 13.2 Hz,
Benzyl 4-O-amino- -L-arabinoside (12)
Benzyl 4-O-(N-phthalimido)-ꢁ-L-arabinoside (1.60 g) was
treated with methyl hydrazine, as previously for allyl 4-O-
(N-phthalimido)-ꢁ-L-arabinoside, to give hydroxylamine de-
rivative 12 (726 mg, 69%) as colourless plates; mp 155–
156°C, [ꢁ]436 ꢆ4.0 (MeOH). H NMR (500 MHz, DMSO-
d6) : 3.31 (dd, J4,5 = 1.4, J5,5 = 12.3 Hz, H5), 3.34 (ddd,
H5), 3.57 (dd, J2,3 = 9.8, J3,4 = 2.9 Hz, H3), 3.86 (dd, J1,2
=
7.7 Hz, H2), 4.18–4.20 (m, H4), 4.32 (dd, J4,5 = 1.8 Hz, H5),
4.61 (d, H1), 4.72, 4.97 (ABq, J = 11.8 Hz, PhCH2), 5.65
(br s, NH2), 7.27–7.41 (m, Ph). 13C NMR (125.8 MHz,
CDCl3) : 26.51, 26.58 (CMe2), 65.00 (C5), 70.17 (PhCH2),
73.47, 77.60, 78.36 (C2,3,4), 101.74 (Cl), 110.58 (CMe2),
127.77, 128.07, 128.34, 136.94 (Ph). Anal. calcd. for
C23H23NO7: C 61.0, H 7.2, N 4.7; found: C 61.2, H 7.1,
N 4.6.
1
J
1,2 = 6.7, J2,3 = 8.8, J2,OH = 4.8 Hz, H2), 3.46–3.48 (m, H3),
3.60–3.62 (m, H4), 4.03 (dd, J4,5 = 3.1 Hz, H5), 4.20 (d,
H1), 4.53, 4.75 (ABq, J = 12.2 Hz, PhCH2), 4.80 (br d, J =
6.2 Hz, 3-OH), 5.05 (d, 2-OH), 6.03 (br s, NH2), 7.27–7.39
(m, Ph). 13C NMR (125.8 MHz, DMSO-d6) : 61.59 (C5),
69.39 (PhCH2), 70.98, 71.86, 80.10 (C2,3,4), 102.64 (Cl),
127.33, 127.57, 128.12, 138.10 (Ph). Anal. calcd. for
C12H17NO5: C 56.5, H 6.7, N 5.5; found: C 56.2, H 6.7,
N 5.4.
Benzyl 4-O-[(tert-butoxycarbonyl)amino]- -D-arabinoside
(17)
Hydroxylamine 15 (1.12 g, 3.80 mmol) in CHCl3 (70 mL)
was treated with triethylamine (1.0 mL) and di-tert-butyl
dicarbonate (1.24 g, 5.69 mmol). The mixture was heated at
reflux overnight and then concentrated. The residue was dis-
solved in MeOH–CHCl3 (50 mL, 1:1), treated with
pyridinium p-toluenesulfonate (70 mg) and heated under re-
flux (1 h), then allowed to cool. The mixture was treated
with resin (Amberlite IRA 400, OH– form), filtered, and
then concentrated, co-evaporating with toluene, to afford
diol 17 (984 mg, 73%) as a colourless crystalline solid. A
small portion was recrystallized to give colourless micro-
crystals; mp 158–160°C (EtOAc–petrol); [ꢁ]D +5.8, [ꢁ]365
Benzyl -L-xylopyranoside
Tri-O-acetyl-L-xylopyranosyl trichloroacetimidate (25)
(51.1 g, 120 mmol), benzyl alcohol (38 mL, 0.37 mol), and
crushed molecular sieves (3 Å, 15 g) in CH2Cl2 (250 mL)
were stirred (30 min, 0°C). Et2O·BF3 (1 mL, 7.9 mmol) was
added and the reaction was stirred (4 h). Pyridine (120 mL)
and acetic anhydride (90 mL) were added and, after 2 h, the
reaction was quenched upon the addition of MeOH (70 mL).
After 1 h, the mixture was filtered (silica, washing with
EtOAc) and the filtrate concentrated. The residue was sub-
jected to normal work up (EtOAc), and rapid silica filtration
(EtOAc–petrol, 1:9 then 1:3) afforded a colourless solid
(44 g) that was taken up in MeOH and sodium (200 mg)
added. After 3 h, the mixture was neutralized with resin
(Dowex-50W, H+ form), then filtered, washed with MeOH,
and the combined filtrate and washings concentrated to give
a colourless solid. Recrystallization of the solid gave benzyl-
ꢀ-L-xylopyranoside (26.2 g, 90%) as colourless cubes; mp
112–114°C (EtOAc; lit. (23) 116–117°C).
1
+35.8 (MeOH). H NMR. (500 MHz, CDCl3) : 1.46 (s,
CMe3), 3.45 (d, J5,5 = 12.5 Hz, H5), 3.67 (dd, J3,4 = 3.7,
J2,3 = 9.2 Hz, H3), 3.71 (dd, J1,2 = 6.8 Hz, H2), 4.03ꢆ4.05
(m, H4), 4.16 (dd, J4,5 = 2.7 Hz, H5), 4.28 (d, H1), 4.59,
4.88 (ABq, J = 11.8 Hz, PhCH2), 7.26–7.36 (m, Ph). 13C
NMR (75.5 MHz, CDCl3) : 28.10 (CMe3), 63.16 (C5),
70.63 (PhCH2), 71.96, 72.14 (C2,3), 82.82 (CMe3), 83.25
(C4), 101.78 (Cl), 127.85, 128.05, 128.40, 137.02 (Ph),
157.96 (CO). Anal. calcd. for C17H25NO7: C 57.4, H 7.1;
found: C 57.5, H 7.1.
Benzyl 2,3-O-isopropylidene- -L-xylopyranoside (13)
Compound 13 was prepared as previously for the enantio-
mer, 60%; [ꢁ]D +65.9 (CHCl3).
4-O-[(tert-butoxycarbonyl)amino]-D-arabinose (18)
Diol 17 (850 mg), palladium-on-carbon (85 mg of 10%),
and hydrochloric acid (0.4 mL of 1 M) in H2O–THF (1:2,
30 mL) were treated with hydrogen (24 h at 1 atm). The
mixture was filtered through Celite, washing with H2O–THF
(1:1). The combined filtrate and washings were neutralized
with resin (Amberlite IRA 400, OH– form), filtered, and
then concentrated to give a colourless solid. Flash chroma-
tography (MeOH–CHCl3, 7:93 then 15:85) gave hemiacetal
18 (600 mg, 94%) as a colourless solid; mp 121–123°C;
[ꢁ]D +35.2 (MeOH). 1H NMR (500 MHz, MeOH-d4) : 1.50
(s, CMe3), 3.51 (dd, J1,2 = 4.0, J2,3 = 8.8 Hz, H2), 3.57–3.65
(m, H3,4), 3.68 (dd, J4,5 = 6.6, J5,5 = 12.8 Hz, H5), 3.88 (dd,
Benzyl 2,3-O-isopropylidene-4-O-(N-phthalimido)- -D-
arabinoside (14)
Compound 14 was prepared as previously for the enantio-
mer, 68%; mp 177.5–179°C (MeOH); [ꢁ]D +18.4 (CHCl3).
Benzyl 4-O-amino-2,3-O-isopropylidene- -D-arabinoside
(15)
Alcohol 13 (3.23 g, 11.5 mmol), triphenylphosphine
(3.93 g, 15.0 mmol), and 6 (2.45 g, 15.0 mmol) in THF
© 2002 NRC Canada

864
Can. J. Chem. Vol. 80, 2002
J4,5 = 1.5 Hz, H5), 5.52 (d, H1). 13C NMR (125.8 MHz,
MeOH-d4) : 28.37, (CMe3), 61.54 (C5), 67.86, 73.96,
80.92, 83.70 (C2,3,4,CMe3), 86.62 (C1), 155.85 (CO). FAB-
HRMS calcd. for C10H20NO7 (M + 1): 266.1240, found:
266.1251. Anal. calcd. for C10H19NO7: C 45.3, H 7.2; found:
C 45.1, H 7.1.
(dd, J1,2 = 6.9, J2,3 = 10.1 Hz, H2), 3.60 (dd, J4,5 = 4.7, J5,5
13.2 Hz, H5), 3.64 (dd, J3,4 = 8.2 Hz, H3), 4.17 (dd, J4,5
=
=
5.2 Hz, H5), 4.62, 4.81 (ABq, J = 11.6 Hz, PhCH2), 4.82
(ddd, H4), 4.83 (d, H1), 7.12–7.43 (m, 7H, Ph,Im), 7.97–
7.99 (m, Im). 13C NMR (75.5 MHz, CDCl3) : 26.29, 26.50
(CMe2), 63.40 (C5), 70.27 (PhCH2), 76.43, 76.66, 82.22
(C2,3,4), 99.77 (C1), 113.01 (CMe2), 117.80 (2C, Im),
128.05, 128.48, 131.17, 136.52 (Ph), 137.32 (Im). FAB-
HRMS calcd. for C18H23N2O7S (M + 1): 411.1226, found:
411.1221. Anal. calcd. for C18H22N2O7S: C 52.7, H 5.4;
found: C 53.0, H 5.5.
(4R,5S,6R)-5,6-Dihydro-4,5-dihydroxy-6-hydroxymethyl-
4H-1,2-oxazine (4)
Hemiacetal 18 (130 mg) was heated at reflux in water
(7 mL) for 10 min. Freeze drying gave dihydrooxazine 4
(72 mg) as a colourless solid, a portion of which was
recrystallized; mp 142–145°C (MeOH–EtOAc–Et2O); [ꢁ]D =
Benzyl 4-cyano-4-deoxy-2,3-O-isopropylidene-ꢁ-D-
arabinoside (21)
1
+18.3 (MeOH). H NMR (500 MHz, D2O) : 3.80–3.86 (m,
H5), 3.89ꢆ4.02 (m, 3H, H6,7), 4.29 (dd, J3,4 = 1.6, J4,5
=
Imidazole sulfonate 20 (4.58 g, 11.2 mmol) was treated
with trimethylsilyl cyanide (2.94 mL, 22.3 mmol) in MeCN
(100 mL) and the mixture heated under reflux.
Tetrabutylammonium fluoride (1.0 M in THF, 17 mmol) was
added dropwise over 30 min. After a further 1 h, the mixture
was cooled to room temperature, concentrated, and normal
work-up (EtOAc) gave a red oil. Flash chromatography
(EtOAc–petrol, containing 0.5% Et3N, 1:5 then 1:3) of the
oil gave nitrile 21 as a pale yellow oil (2.61 g, 81%), which
crystallized. A small portion was recrystallized to give off-
white needles; mp 89–90°C (Et2O–petrol); [ꢁ]D –21.2
(CHCl3). 1H NMR (300 MHz, CDCl3) : 1.47, 1.51 (s, CMe2),
8.3 Hz, H4), 7.29 (d, H3). 13C NMR (125.8 MHz, D2O) :
59.41 (C7), 65.83, 68.40 (C4,5), 80.51 (C6), 151.96 (C3).
EI-HRMS calcd. for C5H9NO4 (M+): 147.0532, found:
147.0534. Anal. calcd. for C5H9NO4: C 40.8, H 6.2; found:
C 41.0, H 6.2.
(4R,5S,6R)-4,5-Dihydroxy-6-hydroxymethyl-3,4,5,6-
tetrahydro-2H-1,2-oxazine (2)
Dihydrooxazine 4 (140 mg, 0.97 mmol) in AcOH (12 mL)
was treated with NaCNBH3 (64 mg, 95%, 0.97 mmol)
(40 min, rt). The mixture was concentrated and the residue
dissolved in hydrochloric acid (6 mL of 0.05 M) and then
applied to a column of cation-exchange resin (Dowex 50W-
X2, H+ form). The column was washed with water, then
eluted with hydrochloric acid (3 M). Concentration of the
eluate gave a colourless oil that was taken up in MeOH and
treated with resin (Amberlite IRA 400, OH– form). The mix-
ture was filtered and concentrated to give tetrahydrooxazine
2 (160 mg) as a colourless foam, a portion of which was
crystallized; mp 149–151°C (MeOH–EtOAc); [ꢁ]D –36.4
3.29 (ddd, J3,4 = 4.9, J4,5 = 1.8, 3.1 Hz, H4), 3.58 (dd, J2,3
9.4 Hz, H3), 3.64 (dd, J5,5 = 12.5 Hz, H5), 3.80 (dd, J1,2
=
=
7.6 Hz, H2), 4.30 (dd, H5), 4.68 (d, H1), 4.69, 4.93 (ABq,
J = 11.8 Hz, PhCH2), 7.27–7.42 (m, Ph). 13C NMR
(75.5 MHz, CDCl3) : 26.39, 26.65 (CMe2), 32.82 (C4),
63.66 (C5), 70.38 (PhCH2), 74.53, 76.00 (C2,3), 101.45
(C1), 111.97 (CMe2), 116.59 (CN), 120.10, 128.02, 128.48,
136.55 (Ph). FAB-HRMS calcd. for C16H20NO4 (M + 1):
290.1392, found: 290.1398. Anal. calcd. for C16H19NO4:
C 66.4, H 6.6; found: C 66.3, H 6.4.
1
(MeOH). H NMR (500 MHz, D2O) : 2.92 (dd, J3,3 = 13.2,
J3,4 = 10.7 Hz, H3), 3.28 (dd, J3,4 = 5.4 Hz, H3), 3.41 (t,
J4,5 W J5,6 = 9.3 Hz, H5), 3.58 (ddd, J6,7 = 2.1, 6.2 Hz, H6),
3.67–3.73 (m, H4,7), 3.90 (dd, J7,7 = 12.7 Hz, H7).
13C NMR (125.8 MHz, D2O) : 53.26 (C3), 60.54 (C7),
71.05, 71.07 (C4,5), 84.33 (C6) (7). EI-HRMS calcd. for
C5H11NO4 (M+): 149.0688, found: 149.0692. Anal. calcd.
for C5H11NO4: C 40.3, H 7.4, N 9.4; found: C 40.1, H 7.2,
N 9.2.
Benzyl 4-{[(tert-butoxycarbonyl)amino]methyl}-4-deoxy-
2,3-O-isopropylidene-ꢁ-D-arabinoside (23)
Nitrile 21 (2.0 g, 6.9 mmol) in ether (40 mL) was added
dropwise to a mixture of lithium aluminium hydride
(730 mg, 17 mmol) in ether (40 mL) and this mixture stirred
(1 h). The mixture was quenched with successive additions
of H2O (730 L), aqueous NaOH (3 M, 730 L), and H2O
(2.2 mL). The mixture was filtered, washing with ether, and
the combined filtrate and washings were concentrated to
give an oil (2.5 g). This oil was dissolved in CHCl3
(120 mL) and treated with triethylamine (2 mL) and di-tert-
butyl dicarbonate (1.7 g, 7.6 mmol). The mixture was heated
at reflux overnight, then concentrated to give a colourless
solid. Recrystallization gave carbamate 23 as fine needles
(1.52 g, 56%); mp 146–148°C (EtOAc–petrol); [ꢁ]D +29.9
Benzyl 4-O-(imidazyl-1-sulfonyl)-2,3-O-isopropylidene- -
L-xyloside (20)
Sodium hydride (60% in mineral oil, 1.0 g, 25 mmol) was
added to a chilled mixture of alcohol 13 (5.4 g, 20 mmol) in
DMF (30 mL, 0°C) and the mixture stirred (30 min, rt). The
mixture was cooled (ꢆ40°C) and N,N>-sulfuryldiimidazole
(5.0 g, 25 mmol) in DMF (40 mL) was added dropwise, then
the mixture allowed to stir (30 min). MeOH (3 mL) was
added and, after a further 30 min, the mixture was allowed
to warm to room temperature, then subjected to normal
workup (CH2Cl2) to give a pale yellow oil. Flash chromatog-
raphy (EtOAc–petrol, containing 0.5% Et3N, 1:3 then 1:2)
gave imidazole sulfonate 20 (5.75 g, 72%) as a colourless
solid. A small portion was recrystallized to give colourless
needles; mp 90–92°C (EtOAc–petrol); [ꢁ]D +86.6 (CHCl3).
1H NMR. (300 MHz, CDCl3) : 1.30, 1.37 (s, CMe2), 3.47
1
(CHCl3). H NMR (300 MHz, CDCl3) : 1.46 (s, 15H, Me),
2.33–2.43 (m, H4), 3.26–3.49 (m, CH2N), 3.58 (dd, J4,5
3.8, J5,5 = 12.6 Hz, H5), 3.61–3.69 (m, H2), 3.72 (dd, J2,3
=
=
9.8, J3,4 = 5.3 Hz, H3), 3.91 (dd, J4,5 = 2.6 Hz, H5), 4.61 (d,
J1,2 = 7.3 Hz, H1), 4.65, 4.85 (ABq, J = 11.7 Hz, PhCH2),
5.00–5.09 (m, NH), 7.27–7.41 (m, Ph). 13C NMR
(75.5 MHz, CDCl3) : 26.54, 26.63 (CMe2), 28.39 (CMe3),
38.08 (CH2N), 38.99 (C4), 64.93 (C5), 70.10 (PhCH2),
74.43, 77.54 (C2,3), 79.29 (CMe3), 101.84 (C1), 110.60
© 2002 NRC Canada

Best et al.
865
(CMe2), 127.79, 128.05, 128.37, 137.00 (Ph), 155.89 (CO).
FAB-HRMS calcd. for C21H32NO6 (M + 1): 394.2230,
found: 394.2247. Anal. calcd. for C21H31NO6: C 64.1, H 7.9;
found: C 64.0, H 7.7.
water and then eluted with aqueous NH3 (1.5 M). The eluate
was concentrated to give 1 as the free amine; [ꢁ]D +20.5
(EtOH, lit. (12) +19.6).
Acknowledgments
Benzyl 4-{[(tert-butoxycarbonyl)amino]methyl}-4-deoxy-
ꢁ-D-arabinoside (24)
We thank the Australian Research Council for financial
assistance and Dr. Jon Fairweather (working in the labora-
tory of Professor Marc Claeyssens, Gent, Belgium) for de-
termining the inhibition constants of the dihydrooxazine.
A mixture of carbamate 23 (1.50 g, 3.82 mmol) and
pyridinium p-toluenesulfonate (80 mg, 310 (mol) in MeOH–
CHCl3 (25 mL, 1:1) was heated under reflux (1 h) and al-
lowed to cool. The mixture was treated with resin
(Amberlite IRA 400, OH– form), filtered, and then concen-
trated, co-evaporating with toluene, to afford diol 24 as a
colourless solid (1.30 g). A small portion was recrystallized
to give colourless micro-needles; mp 159–162°C (EtOAc–
References
1. M.L. Sinnott. Chem. Rev. 90, 1171 (1990).
2. D.L. Zechel and S.G. Withers. Acc. Chem. Res. 33, 11 (2000).
3. G. Legler. Adv. Carbohydr. Chem. Biochem. 48, 319 (1990).
4. B. Ganem. Acc. Chem. Res. 29, 340 (1996).
5. A. E. Stütz (Editor). In Nojirimycin and beyond. Wiley-VCH,
Weinheim. 1999.
6. M. Bols. Acc. Chem. Res. 31, 1 (1998).
7. P. Bach and M. Bols. Tetrahedron Lett. 40, 3461 (1999).
8. S. Hanessian and J.-M. Vatèle. Tetrahedron Lett. 22, 3579
(1981).
9. E.D. Soli, A.S. Manoso, M.C. Patterson, P. DeShong, D.A. Fa-
vor, R. Hirschmann, and A.B. Smith, III. J. Org. Chem. 64,
3171 (1999).
10. G. Zhao, U.C. Deo, and B. Ganem. Org. Lett. 3, 201 (2001).
11. G. Pandey and M. Kapur. Tetrahedron Lett. 41, 8821 (2000).
12. T.M. Jespersen, M. Bols, M.R. Sierks, and T. Skrydstrup. Tet-
rahedron, 50, 13449 (1994).
1
petrol); [ꢁ]D +77.8 (CHCl3). H NMR (300 MHz, CDCl3) :
1.43 (s, CMe3), 1.98 (br s, OH), 2.16–2.27 (m, H4), 2.89
(br s, OH), 3.18–3.32 (m, CH2N), 3.53 (dd, J = 4.0, 11.9 Hz,
1H), 3.68 (dd, J = 3.9, 5.5 Hz, 1H), 3.76–3.87 (m, 2H), 4.53,
4.81 (ABq, J = 11.8 Hz, PhCH2), 4.59 (d, J1,2 = 3.7 Hz, H1),
5.00–5.08 (m, NH), 7.27–7.39 (m, Ph). 13C NMR
(75.5 MHz, CDCl3) : 28.39 (CMe3), 37.41 (C4), 38.95
(CH2N), 60.17 (C5), 69.46, 70.31 (C2,3), 70.04 (PhCH2),
79.49 (CMe3), 100.31 (C1), 128.08, 128.55, 136.71 (Ph),
156.34 (CO). FAB-HRMS calcd. for C18H28NO6 (M + 1):
354.1917, found: 354.1924. Anal. calcd. for C18H27NO6:
C 61.2, H 7.7; found: C 61.4, H 7.8.
Isofagomine (HCl salt) (1)
13. S.U. Hansen and M. Bols. J. Chem. Soc. Perkin Trans. 1, 911
(2000).
Diol 24 (1.3 g) was treated with trifluoroacetic acid (6 mL;
5 min, rt) and the mixture concentrated, co-evaporating with
H2O. The residue was dissolved in MeOH, treated with resin
(Amberlite IRA 400, OH– form), filtered, and then concen-
trated. MeOH (40 mL), AcOH (1.5 mL), and palladium-on-
carbon (150 mg of 10%) were added to the residue and this
mixture was treated with hydrogen (overnight, 1 atm). The
mixture was filtered through Celite, washing with MeOH,
followed by concentration, to give a pale-brown gum
(720 mg). This gum was dissolved in hydrochloric acid
(4 mL of 1 M) and applied to a cation-exchange column
(Dowex 50W-X2, H+ form). The column was washed with
water, then eluted with aqueous NH3 (1.5 M). The eluate was
concentrated and flash chromatography (i-PrOH–aqueous
NH3 (7 M), 6:1 then 4:1) gave a colourless oil. This oil was
treated with hydrochloric acid (10 mL of 1 M) and concen-
trated. Freeze drying of the residue gave isofagomine (1)
(HCl salt) as a colourless gum (660 mg, 96%); [ꢁ]D +16.5
(H2O). 1H NMR (500 MHz, D2O) and 13C NMR
(125.8 MHz, D2O) spectral data were in good agreement
with those reported previously (12, 14).
14. Y. Ichikawa, Y. Igarashi, M. Ichikawa, and Y. Suhara. J. Am.
Chem. Soc. 120, 3007, 5854 (1998).
15. Y.J. Kim, M. Ichikawa, and Y. Ichikawa. J. Org. Chem. 65,
2599 (2000).
16. J. Andersch and M. Bols. Chem. Eur. J. 7, 3744 (2001).
17. M. Bols, R.G. Hazell, and I.B. Thomsen. Chem. Eur. J. 3, 940
(1997).
18. J.C. McAuliffe and R.V. Stick. Aust. J. Chem. 50, 193 (1997).
19. S.R. Hall, D.J. du Boulay, and R. Olthof-Hazekamp (Editors).
The Xtal 3.6 system. University of Western Australia, Austra-
lia. 2000.
20. J.J. Naleway, C.R.H. Raetz, and L. Anderson. Carbohydr. Res.
179, 199 (1988).
21. R.F. Helm, J. Ralph, and L. Anderson. J. Org. Chem. 56, 7015
(1991).
22. L.R. Schroeder, K.M. Counts, and F.C. Haigh. Carbohydr. Res.
37, 368 (1974).
23. C.K. de Bruyne and G. van der Groen. Carbohydr. Res. 2, 173
(1966).
24. S. Rio, J.-M. Beau, and J.-C. Jacquinet. Carbohydr. Res. 219,
71 (1991).
25. M. Mori, Y. Ito, and T. Ogawa. Carbohydr. Res. 195, 199
(1990).
A sample of 1 was applied to a cation-exchange column
(Dowex 50W-X2, H+ form). The column was washed with
© 2002 NRC Canada