4042 J. Agric. Food Chem., Vol. 51, No. 14, 2003
Villard et al.
Table 1. Synthesis of 5-(Hydroxymethyl)-2-furaldehyde (3) from
Table 2. 1H NMR Data of Racemic
D-Fructose
N-(1-Carboxyethyl)-6-(hydroxymethyl)pyridinium-3-ol (1) before and
after Refluxing under Alkaline Conditions in D O
2
reaction temp
trial time (h) (°C)
yield of
3a (%) ref
solvent(s)
purification
none
chromatography
none
1 at pD 4.7
1 after 24 h reflux at pD 9.5
1
2
3
16
2
100 DMSO
150 DMSO
32
24
0
nrb
24
nr
proton
δ (ppm)
multa
J (Hz)
δ (ppm)
multa
J (Hz)
CH
1.73
4.68
5.31
7.42
7.52
7.83
d
s
q
dd
d
7.2
1.55
s
3
40
50 H O/EtOAc
2
+
CH OH
2
H , H O
2
CH
H-4
H-5
H-2
7.2
9.0, 2.5
9.0
4
5
1
300
distillation
6
25
7.15
7.30
d
d
9.0
9.0
+
2.5
100 BuOAc, resin H , chromatography
52
nr
H O
2
d
2.5
a Yields of 3 were determined by HR/GC (FID detector). b Not reported.
a Multiplicity of signals: s, singlet; d, doublet; dd, double-doublet; q, quartet.
Amberlite 15 (10.0 g), and NEt3‚HCl (2.0 g) were stirred in a mixture
of n-butyl acetate (150 mL) and water (5 mL) and then heated at 100
°C under a nitrogen atmosphere for 2.5 h. After cooling to room
temperature, the mixture was decanted, and the organic layer was
concentrated under reduced pressure to yield 3 (3.3 g, 26 mmol, yield
) 52%, purity > 80%, GC) as an amorphous solid after filtration over
neutral alumina using a mixture of petroleum ether and ethyl acetate
(3:2, v/v) as the solvent.
water (35 mL) over a period of 30 min. After 1 h of stirring, the
resulting mixture was neutralized by adding a strongly basic ion-
exchange resin [Amberlite IRA-400 (OH-)] and filtered, and the filtrate
was freed from solvent in vacuo. The residue was dissolved in a formic
acid solution (0.1% in water) and applied to a 30 × 3 cm water-cooled
glass column filled with RP-18 (LiChroprep, 25-40 µm) in a formic
acid solution (0.1% in water). Using the same solvent as the mobile
phase and monitoring the effluent at 300 nm, (+)-(S)-1 was eluted
between 150 and 290 mL. Chromatography was repeated using
ammonium formate (10 mmol/L, pH 8.2) as the mobile phase, and the
fraction containing the target compound was freeze-dried twice to obtain
the (+)-(S)-1 inner salt (0.76 mmol) in 13% yield: 1H NMR (360 MHz,
D2O, pD 7.1, TSP) δ 1.86 (d, J ) 6.8 Hz, 3H), 4.84 (s, 2H), 5.42 (q,
J ) 6.8 Hz, 1H), 7.70 (dd, J ) 2.3, 8.6 Hz, 1H), 7.77 (d, J ) 8.6 Hz,
1H), 8.08 (d, J ) 2.3 Hz, 1H); LC-MS (ESI+), m/z (%) 198 (100, [M
+ 1]+), 220 (57, [M + Na]+); UV-vis (Millipore water, 12.5 mmol/
L, pH 7.1), λmax 248, 326 nm; [R]D20 +40.2° (c 0.01, Millipore water,
Racemic 1. A solution of L-alanine (1.0 equiv) and 3 (1.2 equiv) in
EtOH/water (50:50, v/v) was adjusted to pH 9.4 with an aqueous NaOH
solution (32%). After stirring at room temperature for 1.5 h, the mixture
was refluxed for 24 h and then cooled to room temperature. This
procedure was repeated twice after the addition of 0.4 and 0.3 equiv
of 3. The sample was then concentrated under reduced pressure and
purified by chromatography on silica gel 60 with a solvent mixture
composed of butanol, water, and acetic acid (2:2:1, v/v/v). Finally,
chromatography on RP18 reversed phase material with aqueous formic
acid (0.5% in water) and subsequent recrystallization from absolute
ethanol furnished racemic 1 in 65% yield with a purity of >98%
+
pH 7.0, NH4 salt).
(-)-(R)-1. Following the procedure described for the preparation of
(S)-4 and (+)-(S)-1, (-)-(R)-1 was synthesized using D-alanine. LC-
20
(NMR): mp (absolute ethanol), 91 °C; [R]D ) 0° (c 1.0, Millipore
water, pH 4.7, Na+ salt); 13C NMR (90 MHz, D2O, pD 4.7) δ 174.6
20
MS and NMR data were identical to those obtained for (+)-(S)-1: [R]D
(s), 160.9 (s), 142.8 (s), 133.8 (d), 133.1 (d), 128.6 (d), 65.5 (d), 59.7
+
-38.6° (c 0.01, Millipore water, pH 7.0, NH4 salt).
(t), 18.5 (q); IR (KBr) ν ) 3417, 3090, 3080, 3040, 2798, 1616 cm-1
;
UV-vis (ammonium formate, 10 mmol/L, pH 8.2), λmax 248, 326 nm;
UV-vis (formic acid, 0.1% in water, pH 2.5), λmax 300 nm; LC-MS
(ESI+), m/z (%) 198 (21, [M + 1]+), 220 (10, [M + Na]+), 395 (100,
[2M + 1]+), 417 (29, [2M + K]+), 592 (53, [3M + 1]+); crystal data,
light brown monoclinic prism, (C9H11NO4) (C9H10NO4)- Na+ (C2H6O),
Mr ) 462.5; µ ) 0.12 mm-1, dx ) 1.359 g‚cm-3, P 21/c, Z ) 4, a )
7.1705(3), b ) 22.7540(16), c ) 13.8627(7) Å, â ) 91.635(6)°, V )
2260.9(2) Å3.
RESULTS AND DISCUSSION
Syntheses. As shown in Figure 2, the synthesis of racemic
and enantiopure 1 is based on 5-(hydroxymethyl)-2-furanalde-
hyde 3 as the common intermediate. The synthesis of 3 was
studied in more detail, as it is a rather unstable and expensive
material. It has mainly been prepared from the readily available
D-fructose 2 by dehydration using Brønsted and Lewis acids as
catalysts (18-20). In general, the yields were low and strongly
dependent on the reaction conditions such as time and temper-
ature as well as the solvent used. The methods described for its
production so far suffered from the disadvantage that it was
obtained in aqueous or polar media from which isolation was a
challenging task.
To overcome these problems, we applied a phase transfer
procedure with n-butyl acetate/water using the strong acid resin
Amberlite 15 and ammonium salts as catalyst, thus resulting in
∼50% yield after chromatographic workup (Table 1, trial 5).
This procedure could easily be scaled up to 100 g without
decreasing the yield. Another advantage was the feasibility of
recovering the catalyst and solvents. Purification was achieved
by dry chromatography on alumina with petroleum ether/EtOAc
as eluent.
(S)-N-(1-Carboxyethyl)-2-(hydroxymethyl)-5-(methylamino)furan [(S)-
4]. A solution of 3 (30 mmol) and L-alanine (60 mmol) in water (30
mL) was adjusted to pH 8.5 with an NaOH solution (32% in water)
and was stirred in a hydrogenation vessel at room temperature for 30
min. After the addition of Raney nickel (0.75 g), the mixture was stirred
under a hydrogen atmosphere at 5 bar for 48 h. Another aliquot of 3
(10 mmol) was added, and hydrogenation was continued for an
additional 48 h. After filtration of the catalyst and a wash with methanol,
the filtrate was concentrated in vacuo. The residue was dissolved in
ammonium formate (10 mmol/L, pH 8.2) and applied on a water-cooled
30 × 4 cm glass column filled with a slurry of RP-18 (LiChroprep,
25-40 µm) in a mixture (99:1, v/v) of ammonium formate (10 mmol/
L, pH 8.2) and methanol. By using the same solvent mixture as the
mobile phase and monitoring the effluent at 220 nm, the fractions
eluting between 210 and 330 mL were collected to obtain (S)-4 (15.0
mmol) after freeze-drying in 38% yield: 1H NMR (250 MHz, D2O,
TSP) δ 1.43 (d, J ) 7.3 Hz, 3H), 3.57 (q, J ) 7.3 Hz, 1H), 4.18 (d, J
) 2.4 Hz, 2H), 4.56 (s, 2H), 6.39 (d, J ) 3.4 Hz, 1H), 6.51 (d, J ) 3.4
Hz, 1H); 13C NMR (41.7 MHz, D2O, TSP) δ 173.7 (s), 157.7 (s), 148.9
(s), 115.4 (d), 112.1 (d), 59.7 (d), 58.5 (t), 44.7 (t), 18.5 (q); LC-MS
(ESI+), m/z (%) 421 (100, [2 M + Na]+), 399 (95, [2 M + H]+), 222
(42, [M + Na]+), 200 (49, [M + H]+); UV-vis (ammonium formate,
10 mmol/L, pH 8.2), λmax 210 nm.
Alternatively, 3 was obtained at ∼30% yield without puri-
fication using DMSO as solvent (trial 1). However, a much
longer reaction time was required compared to trial 5. Increasing
the reaction temperature to 150 °C led to lower yields (trials 2
and 4). This is most likely due to polymerization reactions,
which we observed in most of the samples shown in Table 1,
known to be favored at high temperatures and in highly
(+)-(S)-1. A solution of bromine in methanol (6.0 mmol in 10 mL)
was added dropwise to a cooled (0 °C) solution of (S)-4 (7.5 mmol) in