196
J. Chang, Y. Zhang / Process Biochemistry 47 (2012) 195–200
3. Results
3.1. Degradation of amygdalin
observed by HPLC. To identify the amygdalin degradation products,
chromatographic characteristics of standards of mandelonitrile,
prunasin and benzaldehyde were determined under the same HPLC
conditions. The HPLC retention times of peaks 1, 3 and 4 were 3.3,
16.37 and 22.75 min, respectively (Fig. 2), which were the same
as the retention times of the mandelonitrile, prunasin and ben-
zaldehyde standards. Thus, peaks 1, 3 and 4 were identified as
mandelonitrile, prunasin and benzaldehyde.
The time course of the catalytic degradation of amygdalin (Fig. 3)
demonstrated that amygdalin rapidly degraded into prunasin, an
unknown peak 2, mandelonitrile and benzaldehyde. The concentra-
tion of prunasin dramatically increased to 0.7 mg/ml from 0 to 1.5 h
and then slowly decreased to 0.3 mg/ml from 1.5 to 4 h. The concen-
trations of the unknown peak 2, mandelonitrile and benzaldehyde
steadily increased to 0.346, 0.251 and 0.520 mg/ml, respectively,
within a time range of 0 to 4 h. After 4 h, no amygdalin was observed,
indicating that amygdalin was fully degraded.
Fig. 1. Structure of amygdalin.
for 12 h), lyophilized and stored at 4 ◦C for future use as a mixture of extracellular
enzymes.
2.4. Catalysis of amygdalin and HPLC assay
3.2. Separation and purification
Amygdalin (0.5 ml of a 10 mg/ml stock solution) was added into 2 ml of a mixture
of extracellular enzymes (82 U/ml, where U refers to a unit of -glucosidase activity)
and incubated at 37 ◦C for 4 h. Aliquots of the assay were analyzed by HPLC every
0.5 h. Chromatographic separation (with an injection volume of 10 l) was per-
formed on an Agilent HPLC system equipped with a C18 ODS column (250 × 4.6 mm
i.d.) by elution with 35% (v/v) methanol in water at a flow rate of 1 ml/min. The
detection wavelength was 210 nm, and the column temperature was maintained at
25 ◦C.
amygdalin such that the reaction was carried out in a 1000 ml reac-
tion bulb containing 600 ml of a mixture of extracellular enzymes
(82 U/ml). After 4 h, the reaction mixture was separated and puri-
fied as mentioned above. The results of purification are shown in
Table 1. After the pooled chloroform layer was evaporated to dry-
ness under vacuum, 2.3 g of residue was obtained and the purity
of peak 2 was estimated to be 50% by HPLC. The elution of the dis-
solved residue on the macroporous resin column with chloroform
and methanol as gradient mixture yielded 1.6 g of residue contain-
ing 85% of peak 2. The eluate was dried, purified on silica gel column,
and 1.1 g of peak 2 (98% in purity) was obtained.
2.5. Separation and purification of products
The reaction solution was extracted three times with an equal volume of
petroleum (60–90 ◦C) and subsequently extracted three times with an equal vol-
ume of chloroform. The chloroform layers were pooled, dried over sodium sulfate
and evaporated to dryness under vacuum. The residue was then separated on a
macroporous resin column using chloroform and methanol as the gradient elution
followed by a silica gel column using methanol and water as the gradient elution
(30–60%).
3.3. Structure elucidation of compound 2
Peak 2 was isolated as white needles (CHCl3–MeOH); mp:
137–140 ◦C; [˛]D20 −37.0 (CH3COCH3); vmax: 3470 cm−1; ESI-MS
m/z: 341.2986 [M+H]+; 13C NMR (CDCOCD3, 125 MHz): ıC 106.7
(C-1ꢀ), 69.6 (C-2ꢀ), 118.4 (C-3ꢀ), 125.3 (C-4ꢀ), 108.9 (C-5ꢀ), 154.4 (C-
6ꢀ); 1H NMR (CDCOCD3, 500 MHz): ıH 6.57 (2H, m, H-5, 7), 7.05 (2H,
m, H-4, 8), ıH 5.91 (1H, s, H-2), ıH 4.30 (1H, d, J = 7.3 Hz, H-1ꢀ); ele-
ment analysis: Anal. C 49.32%, H 5.36%, O 41.21%, N 4.14%, calcd for
2.6. Structure determination
High resolution mass spectrum (MS) was performed on a Bruker Apex II FI-ICR
mass spectrometer with ESI (Agilent Corp., Palo Alto, USA) as the ion source. Ions
were monitored in the positive ion mode. The purified products were dissolved in
CDCOCD3 for NMR analysis. The 1H- and 13C NMR were performed on a Bruker AV400
(Bruker BioSpin Group, Faellanden, Switzerland). Chemical shifts were reported as
ı values relative to the TMS internal standard. The melting point was determined
using a Yanaco MP-3 micro melting point apparatus (Yanaco Corporation, Kyoto,
Japan) and uncorrected. IR spectra were acquired on a FT-IR spectrophotometer
(PerkinElmer, Altham, USA) using KBr disks.
C14H18O7N, C 49.41%, H 5.33%, O 41.13%, N 4.11%.
On the base of the elemental analysis (C 49.32%, H 5.36%, O
41.21% and N 4.14%), the molecular formula of 2 was established to
be C14H18O7N, which is consistent with the high resolution ESI-MS
([M+H]+ at m/z 341.2986) results. In the IR spectra, one hydroxyl
group was observed. The 1H NMR data exhibited one glycogenic
proton at ıH 4.30 and four olefinic protons at ıH 6.57 (2H, m, H-5,
7) and 7.05 (2H, m, H-4, 8). The 13C NMR data on ıC 106.7, ıC 69.6, ıC
118.4, ıC 125.3, ıC 108.9 and ıC 154.4 suggest that there is one glu-
copyranosyl unit in compound 2. Based on the 1H NMR and 13C NMR
data of compound 2 and prunasin [4], we concluded that the 6-H of
prunasin is substituted by one hydroxyl group in compound 2. Thus,
the unknown peak 2 was identified as phenyl-(3,4,5-trihydroxy-6-
methyl-tetrahydro-pyran-2-yloxy)-acetonitrile, a novel hydroxyl
derivative of prunasin. The structures of phenyl-(3,4,5-trihydroxy-
6-methyl-tetrahydro-pyran-2-yloxy)-acetonitrile (compound 2),
mandelonitrile, prunasin and benzaldehyde are shown in Fig. 4.
2.7. Assay of anti-tumor activity
S-18 cells were cultured in 1640 medium (Gibco, Grand Island, NY, USA), sup-
plemented with 10% fetal bovine serum (FBS, Gibco) and maintained at 37 ◦C in
a humidified chamber with 5% CO2 for 7 d. The tumor cells (107/ml) were then
transferred into mice (45 males and 45 females). The tested compound was orally
administered at 5, 10 and 20 mg/kg doses once every day for 10 d. Water and
cyclophosphamide were used as negative and positive controls, respectively. On day
11, all mice were killed, the weights of tumors were measured, and the inhibition
rates were calculated.
Data obtained from animal experiments were expressed as mean standard
error (SEM). Statistical differences between the treatments and the controls were
tested by a one-way analysis of variance (ANOVA) and Student–Newman–Keuls post
hoc test. A value of P < 0.05 was considered to be significant.