1,5-Anhydro-D-fructose: Regioselective acylation with fatty acids

Article abstract of DOI:10.1016/S0008-6215(99)00163-9

Regioselective acylation of 1,5-anhydro-D-fructose was performed with dodecanoic acid to give 1,5-anhydro-6-O-dodecanoyl-D-fructose, chemically in 50% yield and enzymatically in quantitative yield. Quantitative conversions were also obtained using hexadecanoic and octadecanoic acids as acyl donors. Copyright (C) 1999 Elsevier Science Ltd.

Full text of DOI:10.1016/S0008-6215(99)00163-9

www.elsevier.nl/locate/carres
Carbohydrate Research 320 (1999) 250–256
Note
1,5-Anhydro-
D
-fructose: regioselective acylation with
fatty acids
Søren M. Andersen a,b, Inge Lundt a,*, Jan Marcussen b, Shukun Yu b
a Department of Organic Chemistry, Technical Uni6ersity of Denmark, Building 201, DK-2800 Lyngby, Denmark
b Danisco Biotechnology, Langebrogade 1, PO Box 17, DK-1001 Copenhagen K, Denmark
Received 5 March 1999; revised 15 June 1999; accepted 16 June 1999
Abstract
Regioselective acylation of 1,5-anhydro-D-fructose was performed with dodecanoic acid to give 1,5-anhydro-6-O-
dodecanoyl- -fructose, chemically in 50% yield and enzymatically in quantitative yield. Quantitative conversions
D
were also obtained using hexadecanoic and octadecanoic acids as acyl donors. © 1999 Elsevier Science Ltd. All rights
reserved.
Keywords: 1,5-Anhydro-D-fructose; Antioxidant; Regioselective acylation; Lipases
its hydrophilicity, 1,5-anhydro- -fructose (1)
D
1. Introduction
will partition into the aqueous phase, and thus
does not have the potential to protect the lipid
phase from oxidation. Since it is known that
fatty acid esters of carbohydrates have am-
phiphilic properties [5] and tend to be concen-
trated on the surface between an aqueous and
a lipid phase, these compounds might protect
a lipid phase better than 1 against oxidation.
This class of compounds, non-ionic surfac-
tants based upon sugars, has widespread uses
besides food applications, e.g., extraction of
proteins from membranes and solubilisation
of hydrophobic molecules. They have also
shown anti-HIV and anti-Aspergillus fumiga-
1,5-Anhydro-
into an enediol (Fig. 1) and might therefore be
a potential antioxidant (similar to -ascorbic
D
-fructose (1) can tautomerise
L
acid [1]). Antioxidants are important both in
food and biological systems. In food, they are
particularly advantageous to prevent oxida-
tion of unsaturated fatty acids, while in bio-
logical systems they protect cell components,
such as membranes, proteins and DNA, from
oxidative damage. Since 1,5-anhydro- -fruc-
D
tose is easily produced as a degradation
product from a-(1“4)-glucans such as starch
[2–4], it is an attractive alternative to already
existing antioxidants.
Food systems typically consist of a hetero-
geneous suspension of water and oil. Due to
* Corresponding author. Tel.: +45-4525-2133; fax: +45-
4593-3968.
E-mail address: okil@pop.dtu.dk (I. Lundt)
Fig. 1. Enolisation of 1,5-anhydro-D-fructose.
0008-6215/99/$ - see front matter © 1999 Elsevier Science Ltd. All rights reserved.
PII: S0008-6215(99)00163-9

S.M. Andersen et al. / Carbohydrate Research 320 (1999) 250–256
251
Scheme 1. Synthesis of 6-O-acyl-1,5-anhydro-D-fructose.
tus activity [6]. Furthermore, fatty acid esters
of carbohydrates are biodegradable, harmless
and inexpensively made from renewable
sources [7].
oxime (3b) (11%) were isolated. Alternatively
the carbonyl group was protected with O-ben-
zylhydroxylamine to give the corresponding
O-benzyloximes 2c and 3c. By the relative
downfield shift of the proton on the acyloxy-
lated carbon, the position of the acyl group
Monoacylation of 1,5-anhydro- -fructose
D
(1) could in principle be accomplished at OH-
3, OH-4 or OH-6, but to avoid tedious protec-
tion and deprotection of 1 as well as to
preserve the antioxidative activity [8,9], it is
desirable to selectively acylate the primary
hydroxyl group. It was recently found that
1
was assigned by H NMR spectroscopy. For
comparison, methyl a- -glucopyranoside can
D
be regioselectively acylated at the primary hy-
droxyl group with dodecanoyl chloride in
pyridine at low temperature in 27% yield [12].
The improved regioselectivity in the acylation
of 1 was probably due to steric hindrance in
1,5-anhydro- -fructose exists as a hydrated
D
monomer in aqueous solutions. In non-
aqueous systems, 1 exists as two dimeric
spiroketals in equilibrium with the monomeric
ketoform [10,11]. An incomplete regioselective
acylation of the primary hydroxyl group could
therefore theoretically afford seven products.
To make the monitoring of the reaction easier,
the reaction mixture was derivatised with hy-
droxylamine, which quantitatively converted
the products to the corresponding monomeric
oximes.
the dimeric forms of 1,5-anhydro- -fructose
D
(1). Employing activated esters such as N-lau-
roylthiazolidine-2-thiones [13] or Mitsunobu
conditions [14] did not improve the regioselec-
tivity, but caused some degradation of 1,5-an-
hydro- -fructose (1), probably caused by the
D
addition of NaH, DMAP or PPh3. The only
product isolated from the regioselective acyla-
tion of 1 with N-lauroylthiazolidine-2-thiones
followed by derivatisation was the 3-O-dode-
At low temperature, 1,5-anhydro-D-fructose
canoyl-1,5-anhydro- -fructose oxime (3b). 3-
D
(1) was regioselectively acylated with dode-
canoyl (lauroyl) chloride in pyridine. After
derivatisation with hydroxylamine and
chromatographic separation, 6-O-dodecanoyl-
O-Acylation therefore stabilises 3a against
degradation, as known for similar derivatives
of -ascorbic acid [8,9].
L
An attractive alternative to the chemical
1,5-anhydro-
D
-fructose oxime (2b) (50%)
procedures is the use of lipases and proteases
for the regioselective acylation of the primary
and 3-O-dodecanoyl-1,5-anhydro-
D-fructose

252
S.M. Andersen et al. / Carbohydrate Research 320 (1999) 250–256
Table 1
Enzymatic acylation of 1,5-anhydro-D
-fructose (1) with dodecanoic acid a
,
Dodecanoic acid
(mol/mol 1)
Solvent
3 A Molecular sieve
(w/w 1)
Temperature
(°C)
Reaction time
(h)
Conversion
1
1
1
1
1
1
1
3
3
3
3
tert-BuOH
tert-BuOH
tert-BuOH
acetone
tert-BuOH
tert-BuOH
tert-BuOH
tert-BuOH
tert-BuOH
tert-BuOH
acetone
–
40
40
40
20
45
45
45
45
45
45
20
24
24
72
24
24
24
24
24
24
48
72
21%
56%
62%
55%
56%
61%
66%
73%
1 (powd.)
1 (powd.)
1 (powd.)
5
10
20
20
20 (powd.)
20 (powd.)
20
78%
quantitative
quantitative
a A typical experiment: 1,5-anhydro-
D-fructose (200 mg), dodecanoic acid, Novozym 435 (200 mg), 3 A molecular sieves and
,
solvent (10 mL) were stirred for the time and at the temperature indicated. The solids were filtered off and washed with pyridine.
The filtrate was concentrated, pyridine (2 mL) and NH2OH, HCl (200 mg) were added and stirred for 1 h, followed by 13C NMR
analysis.
hydroxyl group of unprotected sugars with
fatty acids [15–17]. The enzymatic acylation
of sugars with fatty acids can be accomplished
in organic solvents [18], in a minimum of
organic solvent (adjuvant technique) [19] or
solvent-free [20,21]. Recently, supercritical
CO2 has also been used [22].
of drying agents (molecular sieves), tempera-
ture and reaction time (Table 1). Simply stir-
ring equimolar amounts of 1, lauric acid and
lipase in tert-BuOH at 40 °C for 24 h afforded
21% conversion. The reaction is reversible.
Several factors can shift the equilibrium to-
wards complete conversion: product crystalli-
sation, removal of water and the sugar–fatty
acid ratio. Since the acylation products from
Initially, commercially available lipases
from nine different organisms were screened
for the regioselective acylation of 1,5-anhydro-
1,5-anhydro- -fructose (1) did not crystallise
D
from the reaction mixture, the removal of
D-fructose (1) using dodecanoyl O-acetoxime
water and the 1–fatty acid ratio were investi-
as the acylating agent. In pyridine lipases iso-
lated from Candida antarctica, Pseudomonas
cepacia, Pseudomonas fluorescens and hog
pancreas were active. Since polar solvents
such as pyridine can inhibit the lipase activity,
the screening was also performed in a mixture
of 2:1 tert-BuOH–pyridine . It was found that
the same lipases were active together with the
lipase from Candida cylindracea. Since the li-
pase from C. antarctica showed higher activity
compared with the others in both screenings,
this lipase was chosen for further studies.
The enzymatic (Novozym 4351) acylation of
,
gated. Addition of excess drying agent (3 A
molecular sieves) increased the conversion to
66% after 24 h. It is noteworthy that a large
excess of molecular sieves was necessary in
order to optimise the enzymatic reaction,
much more than the capacity (20%, w/w) of
the molecular sieves. Increasing the amount of
fatty acid to 3 molar equiv, together with an
excess of molecular sieves, afforded 73% con-
version. By using powdered molecular sieves,
78% conversion was obtained after 24 h. Ex-
tending the reaction time to 48 h afforded a
quantitative conversion of 1 to the 6-O-dode-
1,5-anhydro-
D
-fructose (1) with dodecanoic
13
canoyl ester 2a, as shown by C NMR. When
acid to 1,5-anhydro-6-O-dodecanoyl-
D-fruc-
the enzymatic acylation was carried out in
refluxing tert-BuOH or simply in melted fatty
acid at 70 °C under low pressure in order to
remove water [20,21], several by-products
were formed, such as elimination and diacyla-
tion products.
tose (2a) was studied in different solvents by
varying the sugar–fatty acid ratio, the amount
1 Novozym 435 is the trademark for an immobilised C.
antarctica lipase from Novo Nordisk A/S.

S.M. Andersen et al. / Carbohydrate Research 320 (1999) 250–256
253
To further simplify the reaction, acetone
was used as solvent and the enzymatic acyla-
tion was performed at ambient temperature.
Under these reaction conditions, 1 was con-
verted in quantitative yields to the corre-
sponding 6-O-dodecanoyl-, 6-O-hexadecanoyl
and 6-O-octadecanoyl esters (2a, 4a, and 5a,
respectively) in three to seven days.
coated Kieselgel 60 F254 and spots were visu-
alised by spraying with a mixture of 1.5%
(w/w)
(NH4)6Mo7O24·4H2O,
1%
(w/w)
Ce(SO4)2·4H2O and 10% (v/v) H2SO4, fol-
lowed by heating. Flash chromatography was
performed on Silica Gel 60 (Grace AB Ami-
con, 35–70 mm). HPLC was performed on a
Waters 8NVC186 reverse-phase column, with
an evaporative light scattering detector
(ELSD). Evaporations were performed in
vacuo at temperatures below 45 °C. In the
screenings the lipases from Aspergillus niger,
C. antarctica, C. cylindracea, Mucor miehei, P.
cepacia, P. fluorescens, Rhizopus arrhizus, Rhi-
zopus ni6eus and hog pancreas were tested
(Lipase Basic Kit from Fluka).
Like anhydrous 1,5-anhydro- -fructose (1),
D
the 6-O-acylated products 2a, 4a, and 5a form
dimeric ketals in pyridine solution [10,11].
Thus the acylation products were present as
mixtures of monomeric ketone, dimer type 1
and dimer type 2 as seen by 13C NMR
(Scheme 1).
In conclusion, a promising enzymatic syn-
thesis of 6-O-acylated derivatives of 1,5-anhy-
1,5-Anhydro-6-O-dodecanoyl-
oxime (2b) and 1,5-anhydro-3-O-dodecanoyl-
-fructose oxime (3b).—1,5-Anhydro- -fruc-
D-fructose
dro-
D
-fructose has been achieved. These types
D
D
of biodegradable fatty acid esters of 1,5-anhy-
tose (1, 1.0 g, 6.3 mmol) was dissolved in
pyridine (30 mL) and the solution was cooled
to 0 °C. Dodecanoyl chloride (1.50 mL, 6.3
mmol) was added dropwise and the reaction
mixture was stirred for 1 h at 0 °C and 3 h at
ambient temperature. Then MeOH (1 mL)
was added and, after 1 h, NH2OH·HCl (660
mg, 9.5 mmol) was added and the mixture was
stirred for 24 h. The reaction mixture was
concentrated and the products were separated
by column chromatography (64 g silica, eluted
with 3:1 to 1:2 hexaneꢀEtOAc) to give 1,5-an-
dro- -fructose (1) will be tested in food sys-
D
tems both as antioxidants and emulsifiers.
2. Experimental
General methods.—1,5-Anhydro- -fructose
D
(1) was freeze-dried from an aqueous solution
and residual water co-evaporated with toluene
three times before use. Molecular sieves were
activated with a heatgun in a flow of Ar.
Pyridine and tert-BuOH were distilled from
hydro-6-O-dodecanoyl-
1.1 g, 50%), mp 89−96 °C, and 1,5-anhydro-
3-O-dodecanoyl- -fructose oxime as a syrup
D
-fructose oxime (2b,
,
CaH2 onto 3 A molecular sieves, acetone was
dried over MgSO4 and dodecanoyl chloride
was distilled prior to use. The enzymes used in
the screenings were dried in vacuo over ‘Blue
Gel’ for two days. The reactions were per-
formed in an Ar atmosphere, unless otherwise
D
(3b, 250 mg, 11%). Analytical samples of 2b
and 3b were prepared by further column
chromatography.
1
stated. H NMR and 13C NMR spectra were
Table 2
13C NMR of isolated compounds in pyridine-d5
recorded on Bruker instruments AC 250 (am-
bient temperature) or AM 500 (300 K). For
1H and 13C NMR spectra in pyridine the
solvent peak was used as reference. For 13C
NMR in D2O, acetone (l=30.89 ppm) was
used as internal reference. Reported melting
points are uncorrected. Optical rotations were
measured on a Perkin–Elmer 241 polarimeter.
Microanalyses were carried out by the Micro-
analytical Department, University of Copen-
hagen and Institut fu¨r Physikalische Chemie,
University of Vienna. HRMS was carried out
by the Chemistry Department, University of
Copenhagen. TLC was performed on pre-
a
Compound C-1
C-2
C-3
C-4
C-5
C-6
1b
1c
2b
2c
3b
3c
4b
5b
61.9
62.2
61.6
61.9
62.1
62.1
61.6
61.6
155.7
157.8
155.0
157.1
151.3
152.5
155.0
155.0
74.8
74.4
74.9
74.4
75.5
75.0
74.9
74.9
74.0
73.5
73.8
73.4
70.4
70.1
73.8
73.8
82.7
82.8
79.2
79.2
82.9
82.8
79.2
79.2
63.2
63.1
64.8
64.6
62.8
62.4
64.8
64.8
a l-Values (ppm) for 1b, 1c, 2b, 2c, 3b (125 MHz), 3c, 4b, 5b
(62.9 MHz); signals were assigned by CꢀH correlated NMR;
solvent peaks used as reference (l 149.9, 135.5 and 123.5 ppm).

254
S.M. Andersen et al. / Carbohydrate Research 320 (1999) 250–256
Table 3
1H NMR of isolated compounds in pyridine-d5
a
Compound
H-1a
H-1b
H-3
H-4
H-5
H-6a
H-6b
1b
4.27 (d)
J1a,1b 14.5
5.68 (d)
4.89 (d)
J3,4 7.5
4.38 (broad t)
J4,5 8.5
3.99 (ddd)
J5,6a 6.0
4.31 (dd)
J6a,6b 11.5
4.49 (dd)
J5,6b 2.5
3.95 (ddd)
J5,6a 5.5
J5,6b 2.5
4.06 (ddd)
J5,6a 6.5
J5,6b 2.0
4.02 (ddd)
J5,6a 6.5
J5,6b 2.5
3.95 (ddd)
J5,6a 5.5
J5,6b 2.5
3.92 (ddd)
J5,6a 5.0
J5,6b 2.5
4.07 (ddd)
J5,6a 6.5
J5,6b 2.0
4.05 (ddd)
J5,6a 6.5
1c
2b
2c
3b
3c
4b
5b
4.21 (d)
J1a,1b 15.0
5.42 (d)
5.68 (d)
5.40 (d)
5.63 (d)
5.38 (d)
5.70 (d)
5.68 (d)
4.83 (d)
J3,4 7.5
4.39 (dd)
J4,5 8.5
4.29 (dd)
J6a,6b 12.0
4.45 (dd)
4.97 (dd)
4.91 (dd)
4.41 (dd)
4.39 (dd)
4.98 (dd)
4.97 (dd)
4.24 (d)
J1a,1b 14.5
4.87 (d)
J3,4 8.0
4.22 (broad t)
J4,5 8.5
4.77 (dd)
J6a,6b 12.0
4.17 (d)
J1a,1b 14.5
4.80 (d)
J3,4 7.5
4.21 (broad t)
J4,5 8.5
4.75 (dd)
J6a,6b 11.5
4.24 (d)
J1a,1b 14.5
6.16 (d)
J3,4 7.5
4.53 (broad t)
J4,5 8.0
4.29 (dd)
J6a,6b 12.0
4.08 (d)
J1a,1b 15.0
6.10 (d)
J3,4 7.0
4.52 (broad t)
J4,5 8.5
4.28 (dd)
J6a,6b 12.0
4.23 (d)
J1a,1b 14.5
4.87 (d)
J3,4 8.0
4.23 (broad t)
J4,5 8.0
4.77 (dd)
J6a,6b 12.0
4.21 (d)
J1a,1b 14.5
4.86 (d)
J3,4 8.0
4.21 (broad t)
J4,5 8.5
4.76 (dd)
J6a,6b 12.0
J5,6b 2.0
a l-Values (ppm) and coupling constants (J, Hz) for 1b, 1c, 2b, 2c, 3b (500 MHz), 3c, 4b, 5b (250 MHz); signals were assigned
by 2D-COSY NMR; solvent peaks were used as reference (l 8.71, 7.55 and 7.19 ppm).
decanoyl-
D
-fructose O-benzyloxime (2c, 287
1,5-Anhydro-6-O-dodecanoyl-
D
-fructose
mg, 46%) and 1,5-anhydro-3-O-dodecanoyl-
D
-
oxime (2b): mp 90−101 °C; [h]D −13.5° (c
13
1
fructose O-benzyloxime (3c, 43 mg, 7%). Ana-
lytical samples of 2c and 3c were prepared by
further column chromatography.
1.1, MeOH); C and H NMR: Tables 2 and
3. Anal. Calcd for C18H33NO6: C, 60.14; H,
9.25; N, 3.90. Found: C, 60.35; H, 9.18; N,
4.10.
1,5-Anhydro-6-O-dodecanoyl- -fructose O-
D
benzyloxime (2c): mp 45−50 °C; [h]D −9.6°
1,5-Anhydro-3-O-dodecanoyl- -fructose
D
(c 1.0, MeOH); 13C and H NMR: Tables 2
1
oxime (3b): [h]D −45.7° (c 1.1, MeOH); 13C
and 3. HRMS Anal. Calcd for C25H39NO6:
1
and H NMR: Tables 2 and 3. HRMS Anal.
450.2856 [M+H+]. Found: 450.2854.
Calcd for C18H33NO6: 359.2309 [M]. Found:
359.2297.
1,5-Anhydro-3-O-dodecanoyl- -fructose O-
D
benzyloxime (3c): mp 48−51 °C; [h]D –51.7°
1,5-Anhydro-6-O-dodecanoyl-
benzyloxime (2c) and 1,5-anhydro-3-O-dode-
canoyl- -fructose O-benzyloxime (3c).—Using
the same procedure as above, 1,5-anhydro-
D
-fructose O-
(c 1.3, MeOH); 13C and H NMR: Tables 2
1
and 3. HRMS Anal. Calcd for C25H39NO6:
D
450.2856 [M+H+]. Found: 450.2845.
D-
Screening of lipases for the regioselecti6e
fructose (1, 227 mg, 1.4 mmol) was acylated
with lauroyl chloride (0.332 mL, 1.4 mmol) in
pyridine (10 mL) followed by addition of
BnONH2·HCl (236 mg, 1.5 mmol). After stir-
ring for 24 h, the mixture was concentrated.
The products were separated by column chro-
matography (20 g silica, eluted with 3:1 to 1:1
hexaneꢀEtOAc) to give 1,5-anhydro-6-O-do-
esterification of 1,5-anhydro-
D
-fructose (1).—
-fruc-
Screening 1: A solution of 1,5-anhydro-
D
tose (1, 308 mg, 1.9 mmol), dodecanoyl
O-acetoxime (480 mg, 1.9 mmol) and pyridine
(13 mL) was prepared. A volume of 1 mL of
this solution was added to approximately 10
mg of the corresponding lipase in an Eppen-
dorf tube, and the mixture was shaken (1000

S.M. Andersen et al. / Carbohydrate Research 320 (1999) 250–256
255
13
rpm) at 30 °C for four days and centrifuged.
A sample (0.5 mL) of the supernatant was
added to 0.5 M NH2OH in pyridine (0.5 mL)
and left for 24 h. The solution was analysed
by HPLC. The lipases isolated from C.
antarctica, P. cepacia, P. fluorescens and hog
pancreas were active. Screening 2: The enzy-
matic reactions were performed in 2:1 tert-
BuOHꢀpyridine using the same procedure as
screening 1. The four lipases above and the
lipase from C. cylindracea were active.
stirred for 1 h and was shown by C NMR to
consist of only 1,5-anhydro-6-O-hexade-
canoyl- -fructose oxime (4b). The remaining
D
syrup (6.4 g) was purified by column chro-
matography (100 g silica, eluted with 4:1:0
then 0:4:1 hexaneꢀEtOAcꢀEtOH) to afford
1,5 - anhydro - 6 - O - hexadecanoyl - - fructose
D
(4a) as an amorphous solid (1.6 g, 68%). This
13
preparation was shown by C NMR (relative
signal intensities) [11] to consist of monomeric
ketone (l 82.6, C-5) (20%), dimer type 1 (l
88.3, C-3%) (20%) and dimer type 2 (l 89.1,
C-3%) (60%). To a sample (58 mg) was added
pyridine (1 mL) and NH2OH·HCl (57 mg).
The solution was stirred for 1 h and was
Enzymatic regioselecti6e acylation of 1,5-an-
hydro-
D
-fructose (1). —1,5-Anhydro-6-O-do-
decanoyl-
D
-fructose (2a). 1,5-Anhydro- -fruc-
D
tose (1, 1.1 g, 6.7 mmol), dodecanoic acid (4.0
g, 20.2 mmol), Novozym 435 (1.1 g) and 3 A
13
,
shown by C NMR to consist of only 1,5-an-
molecular sieves (21.8 g) were suspended in
acetone (95 mL) and stirred for 72 h. Pyridine
(30 mL) was added to the reaction mixture
and the solids were filtered off and washed
with pyridine (2×30 mL). The filtrate was
concentrated in vacuo to a syrup (6.3 g). To a
sample (377 mg) was added pyridine (1.5 mL)
and NH2OH·HCl (116 mg). The solution was
hydro-6-O-hexadecanoyl-
D
-fructose
oxime
(4b). An analytical sample of 4b was prepared
by column chromatography (eluted with 2:1 to
1:1 hexaneꢀEtOAc): mp 97−104 °C; [h]D
−
13.4° (c 1.1, MeOH); 13C and H NMR: Ta-
bles 2 and 3. Anal. Calcd for C22H41NO6: C,
63.59; H, 9.94; N, 3.37. Found: C, 63.80; H,
9.69; N, 3.65.
1
13
stirred for 1 h and was shown by C NMR to
1,5-Anhydro-6-O-octadecanoyl- -fructose
D
consist of only 1,5-anhydro-6-O-dodecanoyl-
(5a). By the same procedure as used for the
preparation of 2a, a mixture of 1,5-anhydro-
D-fructose oxime (2b). The remaining syrup
(5.9 g) was purified by column chromatogra-
phy (100 g silica, eluted with 4:1:0 then 0:4:1
hexaneꢀEtOAcꢀEtOH) to afford 1,5-anhydro-
D
-fructose (1, 1.1 g, 6.6 mmol), octadecanoic
(stearic) acid (5.7 g, 20.0 mmol), Novozym
,
435 (1.1 g), acetone (95 mL) and 3 A molecu-
6-O-dodecanoyl-D-fructose (2a) as an amor-
lar sieves (21.7 g) was stirred for 168 h.
Workup gave a crude residue (7.5 g). To a
sample (230 mg) was added pyridine (1.0 mL)
and NH2OH·HCl (110 mg). The solution was
phous solid (1.5 g, 67%). This preparation was
shown by 13C NMR (relative signal intensities)
[11] to consist of monomeric ketone (l 82.6,
C-5) (20%), dimer type 1 (l 88.3, C-3%) (60%)
and dimer type 2 (l 89.1, C-3%) (20%). To a
sample of this mixture (72 mg) was added
NH2OH·HCl (75 mg) and pyridine (1 mL).
The solution was stirred for 1 h and was
13
stirred for 1 h and was shown by C NMR to
consist of only 1,5-anhydro-6-O-octade-
canoyl- -fructose oxime (5b). The remaining
D
residue (7.3 g) was purified by column chro-
matography (100 g silica, eluted with 8:1:0
then 0:4:1 hexaneꢀEtOAcꢀEtOH) to afford
13
shown by C NMR to consist of only 1,5-an-
hydro-6-O-dodecanoyl-D-fructose oxime (2b),
1,5-anhydro-6-O-octadecanoyl- -fructose (5a)
D
identical with the product described above.
1,5-Anhydro-6-O-hexadecanoyl- -fructose
(4a). Using the same procedure as for 2a, a
mixture of 1,5-anhydro-
as an amorphous solid (1.8 g, 65%). This
13
D
preparation was shown by C NMR (relative
signal intensities) [11] to consist of monomeric
ketone (l 82.6, C-5) (25%), dimer type 1 (l
88.3, C-3%) (50%) and dimer type 2 (l 89.1,
C-3%) (25%). To a sample (69 mg) was added
pyridine (0.7 mL) and NH2OH·HCl (73 mg).
The solution was stirred for 1 h and was
D
-fructose (1, 1.0 g,
6.3 mmol), hexadecanoic (palmitic) acid (4.9
g, 19.0 mmol), Novozym 435 (1.0 g), acetone
,
(95 mL) and 3 A molecular sieves (20.6 g) was
stirred for 72 h, followed by workup, to give a
crude residue (6.7 g). To a sample (340 mg)
was added pyridine (1.5 mL) and
NH2OH·HCl (116 mg). The solution was
13
shown by C NMR to consist of only 1,5-an-
hydro-6-O-octadecanoyl-
D
-fructose
oxime
(5b). An analytical sample of 5b was prepared

256
S.M. Andersen et al. / Carbohydrate Research 320 (1999) 250–256
124 (1997) 37–48.
[7] T. Madsen, G. Petersen, C. Seierø, J. Tørsløv, J. Am. Oil
Chem. Soc., 73 (1996) 929–933.
[8] K. Morisaki, S. Ozaki, Chem. Pharm. Bull., 44 (1996)
1647–1655.
[9] K. Morisaki, S. Ozaki, Carbohydr. Res., 286 (1996) 123–
138.
by column chromatography (eluted with 2:1 to
1:1 hexaneꢀEtOAc): mp 99−105 °C; [h]D
−
1
13.0° (c 1.1, MeOH); 13C and H NMR: Ta-
bles 2 and 3. Anal. Calcd for C24H45NO6: C,
64.98; H, 10.22; N, 3.16. Found: C, 64.70; H,
9.82; N, 3.15.
[10] S. Freimund, S. Ko¨pper, Carbohydr. Res., 308 (1998)
195–200.
[11] S.M. Andersen, I. Lundt, J. Marcussen, I. Søtofte, S. Yu,
J. Carbohydr. Chem., 17 (1998) 1027–1035.
[12] E. Reinefeld, H.-F. Korn, Die Sta¨rke, 20 (1968) 181–
189.
Acknowledgements
[13] D. Plusquellec, K. Baczko, Tetrahedron Lett., 28 (1987)
3809–3812.
[14] S. Bottle, I.D. Jenkins, J. Chem. Soc., Chem. Commun.,
(1984) 385.
[15] M. Therisod, A.M. Klibanov, J. Am. Chem. Soc., 108
(1986) 5638–5640.
The authors thank Novo Nordisk A/S for
kindly providing Novozym 435 and ATV for
their financial support.
[16] K. Faber, S. Riva, Synthesis, (1992) 895–910.
[17] N.B. Bashir, S.J. Phythian, A.J. Reason, S.M. Roberts,
J. Chem. Soc., Perkin Trans., 1 (1995) 2203–2222.
[18] J.A. Arcos, M. Bernabe´, C. Otero, Biotech. Bioeng., 57
(1998) 505–509.
[19] L. Cao, U.T. Bornscheuer, R.D. Schmid, Fett/Lipid, 98
(1996) 332–335.
References
[1] T. Kurata, N. Miyake, Y. Otsuka, Biosci. Biotech.
Biochem., 60 (1996) 1212–1214.
[2] S. Yu, M. Pederse´n, Planta, 191 (1993) 137–142.
[3] S. Yu, L. Kenne, M. Pederse´n, Biochim. Biophys. Acta,
1156 (1993) 313–320.
[20] F. Bjo¨rkling, S.E. Godtfredsen, O. Kirk, J. Chem. Soc.,
Chem. Commun., (1989) 934–935.
[4] S. Yu, T. Ahmad, L. Kenne, M. Pederse´n, Biochim.
Biophys. Acta, 1244 (1995) 1–9.
[5] A. Ducret, A. Giroux, M. Trani, R. Lortie, J. Am. Oil
Chem. Soc., 73 (1996) 109–113.
[21] K. Adelhorst, F. Bjo¨rkling, S.E. Godtfredsen, O. Kirk,
Synthesis, (1990) 112–115.
[22] C. Tsitsimpikou, H. Stamatis, V. Sereti, H. Daflos, F.N.
Kolisis, J. Chem. Technol. Biotechnol., 71 (1998) 309–
314.
[6] I. Rico-Lattes, A. Lattes, Colloids and Surfaces A, 123–
.